Αποτελέσματα εισοδήματος στην Οριακή Επιθυμία για Πληρωμή μέσα στο πλαίσιο του Κοινοτικού Πλαισίου για τους Υδάτινους Πόρους
Εμμανουήλ Τυλλιανάκης
Υποψήφιος Διδάκτορας
Πανεπιστήμιο Πατρών
Σχολή Διοίκησης Επιχειρήσεων
Σχολή Οικονομικών Επιστημών
Επιβλέπων καθηγητής: Δημήτρης Σκούρας
Καθηγητής
Πανεπιστήμιο Πατρών
Σχολή Διοίκησης Επιχειρήσεων
Σχολή Οικονομικών Επιστημών
© Emmanouil Tyllianakis 2017 Επιβλέπουσα Επιτροπή:
κ. Δημήτριος Σκούρας, Καθηγητής, Σχολή Διοίκησης Επιχειρήσεων, Τμήμα Οικονομικών Επιστημών, κ. Δημήτριος Τζελέπης, Επίκουρος Καθηγητής, Σχολή Διοίκησης Επιχειρήσεων, Τμήμα Οικονομικών Επιστημών, κ. Κωνσταντίνος Κουνετάς, Επίκουρος Καθηγητής, Σχολή Διοίκησης Επιχειρήσεων, Τμήμα Οικονομικών Επιστημών,
Η παρούσα διατριβή κατατέθηκε ως μέρος των απαιτήσεων για τον τίτλο του Διδάκτορα της Οικονομικής Επιστήμης στο Τμήμα Οικονομικών Επιστημών του Πανεπιστημίου Πατρών και εξετάσθηκε και ενεκρίθη από την παρακάτω επιτροπή την 4η Ιουλίου 2017 :
κ. Ηλία Κουρλίουρο, Καθηγητής, Σχολή Διοίκησης Επιχειρήσεων, Τμήμα Οικονομικών Επιστημών, κ. Ιωάννη Βενέτη, Αναπληρωτής Καθηγητής, Σχολή Διοίκησης Επιχειρήσεων, Τμήμα Οικονομικών Επιστημών, κ. Φοίβη Κουντούρη, Καθηγήτρια, Οικονομικό Πανεπιστήμιο Αθηνών, κ. Tiziana Luisetti, Λέκτορας, Πανεπιστήμιο East Anglia, Ηνωμένο Βασίλειο
Ευχαριστίες
Στον Θεό, δοτήρα παντός αγαθού,
Στους γονείς μου, για την αγάπη, την ποικιλότροπη στήριξη, την παρότρυνση, τον ιδρώτα και τα δάκρυα για να με φέρουν μέχρι εδώ,
Στην ευρύτερη οικογένειά μου που προπορεύθηκε στον αγώνα της ζωής επιτρέποντάς μου απλά να ακολουθήσω το όμορφό παράδειγμά τους,
Στον καθηγητή και επιβλέποντα της διατριβής κ. Δημήτρη Σκούρα που μου έμαθε σχεδόν όλα όσα ξέρω για τα οικονομικά και την χρήση τους στο διάστημα μιας δεκαετίας που μου έκανε την τιμή να με έχει κοντά του σε διαφορετικά πόστα,
Στους φίλους μου, εντός και εκτός ακαδημαϊκού χώρου, για την υποστήριξή τους, την φιλοξενία, την αγάπη και την κατανόηση που έδειξαν για όλα αυτά τα χρόνια έρευνας και μελέτης και
Τους καθηγητές του Τμήματος Οικονομικών Επιστημών του Πανεπιστημίου Πατρών που μου ενεφύσησαν την αγάπη για την Οικονομική επιστήμη και με εξόπλισαν ώστε να ανταπεξέλθω στις απαιτήσεις του επαγγέλματος.
Ευρεία Περίληψη
Η παρούσα διδακτορική διατριβή επιχειρεί να αναλύσει και να σχολιάσει την επίδραση ευρωπαϊκών περιβαλλοντολογικών και αγροτικών πολιτικών και προγραμμάτων μεγάλης εμβέλειας επάνω στο επίπεδο ευημερίας των πολιτών της Ευρωπαϊκής Ένωσης. Το πρόγραμμα που εξετάζεται με ενδελέχεια στην διατριβή αυτή είναι η «Κοινοτική Οδηγία- Πλαίσιο για τα Νερά» (2000/60/EC) που ως στόχο της είχε τον ορισμό ενός κοινού πλαισίου για την Κοινότητα στον τομέα των υδάτων ώστε να σχηματισθεί ένα ολοκληρωμένο και αναλυτικό πλαίσιο δράσεων, πολιτικών και κατηγοριοποίησης όλων των υδάτινων πόρων της Ευρωπαϊκής Ένωσης. Η ανάλυση επιχειρεί να «ισορροπήσει» μεταξύ της οικονομικής επιστήμης και των κοινωνικών επιστημών χρησιμοποιώντας την οικονομική θεωρία και τα εργαλεία της χωρίς όμως να αγνοούνται οι περιορισμοί της. Ο στόχος της διατριβής είναι να δοθεί μια ανθρωποκεντρική οπτική στον σχεδιασμό πολιτικών, κάτι που συχνά μπορεί να απουσιάζει όταν οικονομικά εργαλεία και νεοκλασικές πρακτικές χρησιμοποιούνται στο έπακρο. Επιπλέον, η διατριβή αυτή εξετάζει τις επιπτώσεις της οδηγίας 2000/60/EC κατά την διάρκεια της περιόδου εφαρμογής της αλλά και μετά την περίοδο που τα αποτελέσματα της εφαρμογής έπρεπε να γίνουν ορατά. Τέλος, σχολιάζονται τα ευρήματα από τις πλευρές των φυσικών, κοινωνικών και οικονομικών επιστημών.
Οι παρακάτω ερωτήσεις θα γίνει προσπάθεια να απαντηθούν στο σύνολο της διατριβής¨
1. Ποιές είναι οι πραγματικές επιπτώσεις μεγάλων «κεφαλαίων» πολιτικής όπως η Κοινοτική Οδηγία-Πλαίσιο για τα Νερά (εφεξής, ΚοπΥ); 2. Ποιά είναι, εάν υπάρχει, η επίπτωση των επιστημονικών μοντέλων στον σχεδιασμό αγρο-περιβαλλοντικών πολιτικών σε συγκεκριμένες λεκάνες απορροής; 3. Είναι αποτελεσματικά από πλευράς κόστους, τέτοια μεγάλα «κεφάλαια» πολιτικής; 4. Ποιός ωφελείται περισσότερο από αυτά, τα πλούσια ή τα λιγότερο πλούσια νοικοκυριά; 5. Εάν ληφθούν υπ’οψιν οι ατομικές απόψεις για τον κίνδυνο, πόσο καλά αντιπροσωπεύουν τα πραγματικά προβλήματα και κινδύνους στα υδάτινα περιβάλλοντα και πόσο πρέπει να επηρεάζουν αυτές οι απόψεις τον σχεδιασμό πολικής; 6. Πώς οι οικολογικοί και περιβαλλοντικοί δείκτες που ορίζονται από την Ευρωπαϊκή Ένωση για τα νερά βοηθούν τους πολίτες της να κατανοήσουν τους κινδύνους στα υδάτινα περιβάλλοντα, με δεδομένη την επιθυμία της Ευρωπαϊκής Ένωσης να συμπεριλάβει όλο και περισσότερο δημόσιες διαβουλεύσεις κατά τον σχεδιασμό μεγάλων «κεφαλαίων» πολιτικής στα θέματα του περιβάλλοντος;
Οι υδάτινοι πόροι ενδιαφέροντος της ΚοπΥ επικεντρώνονται σε εσωτερικούς επιφανειακούς, υπόγειους και παράκτιους πόρους καθώς και σε υδάτινους πόρους κοινής ιδιοκτησίας μεταξύ χωρών-μελών. Ο στόχος της ΚοπΥ ήταν να εισαγάγει μια νέα μορφή διαχείρισης υδάτινων πόρων που θα βασίζεται επάνω σε κάθε μια λεκάνη απορροής ποταμού ξεχωριστά. Η θέσπιση συγκεκριμένων προθεσμιών για την επίτευξη περιβαλλοντικών στόχων για υδάτινα οικοσυστήματα στόχευε στην βαθμηδόν επίτευξη των διάφορων «φιλόδοξων» στόχων.
Η ΚοπΥ επίσης εισήγαγε και υιοθέτησε μια ολοκληρωμένη προσέγγιση για τις στατικές και- μη πηγές (συνήθως αναφερόμενες ως «σημειακές και διάχυτες» πηγές) μόλυνσης των υδάτινων πόρων, προσπάθεια που βασίστηκε επάνω στα ευρήματα προηγούμενων κοινοτικών οδηγιών.
Η ολοκληρωμένη διαχείριση σε επίπεδο λεκανών απορροής επαφίεται στις κατά τόπους επιτροπές που συστάθηκαν ανά Υδατικό Διαμέρισμα που υπάγονται στην αντίστοιχη Εθνική Επιτροπή Υδάτων. Κάθε επιτροπή Υδατικού Διαμερίσματος όφειλε να εκπονήσει σχέδια διαχείρισης για κάθε λεκάνη απορροής ποταμών που υπάγονται στο συγκεκριμένο υδατικό διαμέρισμα που θα έπρεπε στην συνέχεια να εγκριθεί από την εθνική επιτροπή.
Ο στόχος της ΚοπΥ ήταν η επίτευξη του «καλού οικολογικού επιπέδου», ο «χαμένος κρίκος» μεταξύ περιβαλλοντικών πολιτικών, ελέγχου εκπομπών μόλυνσης και επίτευξης περιβαλλοντικών στόχων (χημικών και οικολογικών). Πιο συγκεκριμένα, η επίτευξη της «Καλής Οικολογικής Κατάστασης» (ΚΟΚ) μέχρι το 2015 ήταν ο γενικός στόχος, επικεντρώνοντας την προσοχή σε βιολογικούς δείκτες που θα κατέτασσαν τα νερά σε πέντε τάξεις (Εξαιρετική, πολύ καλή, καλή, μέτρια και κακή). Η ΚοπΥ συνέδεε υπάρχουσες κοινοτικές οδηγίες, επέτασσε την συνεργασία μεταξύ κρατών-μελών λόγω μεταβατικών υδάτινων πόρων, επεκτεινόταν σε θέματα στους τομείς της ενέργειας, αγροτικής παραγωγής, αλιείας, τουρισμού και μεταφορών, απαιτούσε την νομοθετική θέσπιση κανόνων καλής περιβαλλοντικής χρήσης από όλες τις χώρες-μέλη της ΕΕ και τέλος την ανάλυση και ταξινόμηση κάθε υδάτινου όγκου στην ΕΕ και σχεδιασμό της διαχείρισής του. Ο απώτερος στόχος της ΚοπΥ είναι η εξάλειψη των πιο σημαντικών μολυσματικών στοιχείων στο υδάτινο και το θαλάσσιο περιβάλλον, δίνοντας την ευχέρεια κινήσεων και εφαρμογών των πολιτικών σε κάθε χώρα-μέλος.
Κάθε χώρα-μέλος όφειλε να δώσει απαντήσεις στα παρακάτω ερωτήματα για την οικονομική αποτίμηση της εφαρμογής της ΚοπΥ σε κάθε υδάτινο όγκο:
- Ποιές είναι οι χρήσεις του νερού σε κάθε λεκάνη απορροής ποταμού; - Πληρώνει ο καθένας ανάλογα με την χρήση που κάνει ή ο καθένας που μολύνει ανάλογα με την ζημιά που προκαλεί, στο υδάτινο περιβάλλον; - Εμφανίζονται όλα τα είδη κόστους στην ανάλυση (περιβαλλοντικά και οικονομικά); - Είναι αυτά τα κόστη ρεαλιστικά με δεδομένη την παρούσα κατάσταση του εκάστοτε πόρου;
Όλα τα παραπάνω έπρεπε να χρησιμοποιηθούν για την ανάλυση Κόστους Οφέλους προσαρμοσμένη για περιβαλλοντικά αγαθά. Η ανάλυση αυτή περιλαμβάνει κόστη (άμεσα και έμεσσα), οφέλη (διοικητικά, κοινωνικά, περιβαλλοντικά, έμμεσα και προσόδους σπανιότητας). Τα κόστη και τα οφέλη έπειτα πρέπει να αθροιστούν χωρικά και χρονικά.
Η οικονομική ανάλυση δεν ήτανε το τελικό στάδιο ανάλυσης, αντιθέτως η ΚοπΥ εισήγαγε την ανάλυση της δυσαναλογίας (disproportionality analysis). Η ποιοτική ανάλυση αυτή καθορίζει ποιές τοπικές κοινωνικές ομάδες πιθανολογείται να σηκώνουν δυσανάλογα οικονομικά βάρη σε σχέση με άλλες για να επιτευχθεί η ΚΟΚ στον εκάστοτε υδάτινο όγκο. Εάν αυτό συνέβαινε, η διοικούσα αρχή έχει την δυνατότητα να μην εφαρμόσει την ΚοπΥ.
Τα παραπάνω αναλύονται στην παρακάτω περίπτωση προς μελέτη στον ποταμό Λούρο στην Ελλάδα. Ο ποταμός Λούρος βρίσκεται κοντά σε δύο μετρίου μεγέθους αστικά κέντρα (Πρέβεζα και Άρτα) με συνολικό πληθυσμό 67,429 (ΕΛ.ΣΤΑΤ 2011). Η λεκάνη απορροής του Λούρου βρίσκεται σε περιοχές που λαμβάνουν χώρα οικονομικές δραστηριότητες από ετήσιες αγροτικές καλλιέργειες, εκτροφή ζωοειδών και γενικές οικονομικές δραστηριότητες αστικών κέντρων.
Κατά συνέπεια η μόλυνση και οι αιτίες της εντοπίζονται στην εναπόθεση χημικών στοιχείων στα νερά από λιπάσματα και ζωικά εκρίμματα, εντομοκτόνα (που οδηγούν στον ευτροφισμό των υδάτων όπου διαβιούν πληθυσμοί σολομού) και μη-επαρκή διαχείριση ανθρώπινων λυμάτων. Οι πηγές μόλυνσης όμως δεν φαίνονται από τα διαθέσιμα δεδομένα να επιβάλλουν άμεση εφαρμογή αγρο-περιβαλλόντικών μέτρων μιας και το επίπεδο των νερών είναι καλό. Ως εκ τούτου η οικονομική ανάλυση κρίθηκε απαραίτητη. Για την οικονομική ανάλυση, 4 είδη ετήσιων καλλιεργειών λήφθηκαν υπ’ όψιν (μηδική, βαμβάκι, αραβόσιτος και εσπεριδοειδή) και 4 μορφές διαχείρισης (μειώσεις σε νιτρικά και φωσφορικά λιπάσματα, αγρανάπαυση ποτισμένης γης, εναλλαγή καλλιεργειών και ζώνες αγρανάπαυσης). Το κόστος υπολογίστηκε για κάθε μορφή καλλιέργειας και κάθε μορφής διαχείρισης. Στην συνέχεια υπολογίστηκαν οι οριακές καμπύλες μείωσης/αποφυγής κόστους για κάθε μορφή/σενάριο διαχείρισης. Οι μορφές διαχείρισης που επέτρεπαν την αγρανάπαυση ήταν οι λιγότερο κοστοβόρες αλλά ήταν σε πολύ υψηλά επίπεδα. Με άλλα λόγια, για να επιτευχθούν τα επίπεδα συγκέντρωσης νιτρικών και φωσφωρικών στα νερά του Λούρου θα έπρεπε να καταναλωθούν δυσανάλογα μεγάλα ποσά σε σχέση με τα οφέλη και αυτό κατά πάσα πιθανότητα οφείλεται κυρίως στην ήδη καλή ποιότητα των νερών της περιοχής.
Εάν η κλιματική αλλαγή και τα σενάρια του IPCC (2007) συμπεριληφθούν στα μοντέλα INCA της ανάλυσης μας, οι μειώσεις στην βροχόπτωση και οι υψηλότερες θερμοκρασίες φαίνονται προβλέπονται να μειώσουν την μεταφορά ιζημάτων και να αυξήσουν την απορρόφηση θρεπτικών από τα φυτά, βραχυπρόθεσμα. Συμπεριλαμβάνοντας την κλιματική αλλαγή στα μοντέλα πρόβλεψης δεν αλλάζουν πολλά στο πιο μοντέλο διαχείρισης είναι το λιγότερο κοστοβόρο.
Τα οφέλη από την εφαρμογή της ΚοπΥ στον Λούρο μετρήθηκαν στην συνέχεια. Σαν οφέλη θεωρήθηκε η Επιθυμία για Πληρωμή που εκδήλωσαν συμμετέχοντες σε παρόμοιες πρωτογενείς έρευνες για την εφαρμογή της ΚοπΥ σε άλλες χώρες της ΕΕ και το αποτέλεσμα ήταν 40.93 Ευρώ ανά νοικοκυριό για να επιτευχθεί Καλή Οικολογική Κατάσταση στο Λούρο. Εάν συνυπολογιστεί το σύνολο των νοικοκυριών της περιοχής και ένας χρονικός ορίζοντας 6 ετών (μέχρι την επίτευξης της ΚοπΥ, δηλαδή μέχρι το έτος 2016) το σύνολο είναι 11,150,948 ευρώ (με χρήση ενός χαμηλού και σταθερού προεξοφλητικού επιτοκίου 1,5%). Τα οφέλη είναι σημαντικά μικρότερα από το κόστος και το φυσικό ερώτημα είναι πως μπορούν να αποφευχθούν λάθη στον σχεδιασμό και την εφαρμογή μεγάλων «κεφαλαίων» πολιτικής, με δεδομένα τις σοβαρές επιπτώσεις εάν εφαρμοστούν ενώ δεν χρειάζονται ή να μην εφαρμοστούν ενώ είναι αναγκαία.
Τα λάθη αυτά ονομάστηκαν λάθη Τύπου Ένα (υιοθέτηση προγράμματος ενώ δεν είναι απαραίτητο) και Τύπου Δύο (μη υιοθέτηση προγράμματος ενώ είναι απαραίτητο) και η διαδικασία εντοπισμού τους απαρτίζεται από 8 βήματα:
1. Ορισμός και περιγραφή του προβλήματος που απαιτεί το εκάστοτε αγρο- περιβαλλοντικό πρόγραμμα να εφαρμοστεί 2. Χρήση υπαρχόντων δεδομένων (κοινωνικο-οικονομικών και μη) για την κατασκευή μοντέλων μεταφοράς θρεπτικών στην εκάστοτε λεκάνη απορροής ποταμού 3. Επισήμανση πιθανών λαθών τύπου Ένα και Δύο 4. Πρόταση μέτρων και πολιτικών αποφυγής 5. Ένταξη των μέτρων αποφυγής στα μοντέλα μεταφοράς θρεπτικών και ορισμός των μετρήσεων βάσης 6. Ανάλυση των επιπτώσεων εφαρμογής των μέτρων αποφυγής και πώς η λανθασμένη εφαρμογή τους μπορεί να οδηγήσει σε λάθη πολιτικής τύπου Ένα και Δύο 7. Επαν-υπολογισμός των μοντέλων συμπεριλαμβάνοντας τις περιπτώσεις της κλιματικής αλλαγής και αλλαγής στην χρήση γης 8. Προσδιορισμός σεναρίων όπου η εφαρμογή αγρο-περιβαλλοντικών μέτρων στο παρόν μπορεί να μοιάζει δικαιολογημένη ή όχι αλλά στο μέλλον λόγω κλιματικής αλλαγής να καταστεί περιττή ή αναγκαία
Η εφαρμογή της ΚοπΥ στις χώρες της ΕΕ πυροδότησε μεγάλο αριθμό μελετών στις οικονομικές επιστήμες για την αποτίμηση του κόστους και του οφέλους από την επίτευξης της ΚΟΚ. Η διατριβή αυτή προσπάθησε να αποτιμήσει την συνεισφορά όλων αυτών των μελετών στους τομείς της αγρο-περιβαλλοντικής πολιτικής και στο κατά πόσον τέτοια προγράμματα συμβάλλουν στην προσπάθεια επίτευξης μιας πιο δίκαιης κοινωνίας. Η ποσοτική μέθοδος που επιλέχθηκε ήταν η μέτα-ανάλυση (meta-analysis), όπου συναφείς μελέτες που υπολογίζουν τις ίδιες μονάδες συγκεντρώνονται και βάση με την στατιστική εγκυρότητά τους επηρεάζουν λιγότερο ή περισσότερο το συνολικό αθροιστικό αποτέλεσμα της μέτα-ανάλυσης. Επίσης, η Ελαστικότητα Εισοδήματος της Επιθυμίας για Πληρωμή μετρήθηκε μιας και είναι δείκτης του κατα πόσον το αγαθό που αποτιμάται (η βελτίωση της οικολογικής ποιότητας των νερών/ η επίτευξη του ΚΟΚ) θα ωφελήσει περισσότερο τα φτωχά ή τα πλούσια νοικοκυριά.
Από τις μελέτες που αναλύθηκαν και συγκεντρώθηκαν, 21 μελέτες και 32 ξεχωριστά πληθυσμιακά δείγματα συμπεριλήφθησαν στην μέτα-ανάλυση από διαφορετικές περιοχές και χώρες της ΕΕ και το effect size της μέτα-ανάλυσης υπολογίσθηκε σε 41,8 ευρώ ανά νοικοκυριό ανά έτος. Στην συνέχεια, υπολογίσθηκε μια μέτα-παλινδρόμηση για να εξετασθεί η επίπτωση των διαφόρων μεταβλητών στην ετερογένεια μεταξύ των διαφορετικών Επιθυμιών για Πληρωμή, χρησιμοποιώντας κάθε δυνατή ανάλυση και διαχείριση των δεδομένων (δομή ως panel και ως cross-section) και με τρείς διαφορετικές μορφές. Σαν ανεξάρτητη μεταβλητή χρησιμοποιήθηκαν και τρείς επίσημοι δείκτες της Eurostat για τα επίπεδα εισοδήματος ανά περιοχή και χώρα της ΕΕ (κατά κεφαλή ΑΕΠ, Διαθέσιμο εισόδημα Νοικοκυριών και Εισόδημα Νοικοκυριών). Όλες οι μετρήσεις μεταβλήθηκαν σε νομισματικές τιμές 2005 και κάθε ένας από τους 3 δείκτες υπολογίστηκε και σε μονάδες αγοραστικής δύναμης. Τα αποτελέσματα έδειξαν ότι η Ελαστικότητα Εισοδήματος της Επιθυμίας για Πληρωμή είναι χαμηλότερη της μονάδας όταν σαν επεξηγηματική μεταβλητή χρησιμοποιούνται τα εισοδήματα που δήλωσαν οι συμμετέχοντες στης έρευνες που συμπεριλήφθησαν στην μέτα-ανάλυση, κάτι που δεν συνέβη εάν χρησιμοποιήσουμε σαν επεξηγηματικές μεταβλητές τους επίσημους δείκτες της Eurostat. Αυτό σημαίνει ότι στην πρώτη περίπτωση, η επίτευξη του ΚΟΚ στις λεκάνες απορροής της ΕΕ θα ωφελήσει περισσότερο τα φτωχά νοικοκυριά απ’ ότι τα πλούσια ενώ στην δεύτερη το αντίθετο, κάτι που σημαίνει ότι τα πλούσια νοικοκυριά είναι διατεθειμένα να πληρώσουν περισσότερο, ένδειξη προοδευτικής αναδιανεμητικής πολιτικής. Τέλος, υπολογίστηκε ότι για κάθε 1000 ευρώ αύξηση στο μέσο εισόδημα ενός νοικοκυριού, αυξάνεται η Επιθυμία για Πληρωμή για την επίτευξη του ΚΟΚ κατά 10,3%.
Η μεγάλη ετερογένεια μεταξύ των μελετών αλλά και το σχετικά χαμηλό επίπεδο της Επιθυμίας για Πληρωμή για την επίτευξη του ΚΟΚ ήγειρε ερωτηματικά για τον τρόπο που οι πολίτες της ΕΕ αντιλαμβάνονται τους κινδύνους που αντιμετωπίζει το υδάτινο περιβάλλον, κάτι που αποτέλεσε το τελευταίο ζήτημα που αναλύθηκε σε αυτή την διατριβή. Η κατανόηση των κινδύνων αυτών όπως και η διάθεση των πολιτών να υιοθετήσουν και να συμβιβαστούν με κάποια επίπεδα κινδύνου αποτελεί σημαντικό παράγοντα πολιτικής στην σύγχρονη ΕΕ. Ο λόγος βρίσκεται στην ολοένα και αυξανόμενη επιθυμία της ΕΕ να συμπεριλάβει δημόσιες διαβουλεύσεις και γενικότερα τη γνώμη του κοινού κατά τον σχεδιασμό πολιτικών σχετικών με το περιβάλλον.
Για να αναλυθούν οι θεωρήσεις των πολιτών της ΕΕ σχετικά με το τι αποτελεί κίνδυνο για το υδάτινο περιβάλλον και ποιοί είναι οι παράγοντες που επηρεάζουν την εκλογή διαφορετικών κινδύνων για το υδάτινο περιβάλλον ακολουθήθηκε μια ενδελεχής διαδικασία που παρουσιάζεται αμέσως μετά. Χρησιμοποιήθηκαν δύο ειδών δεδομένα, επίσημα δεδομένα της ΕΕ σχετικά με τις απόψεις των πολιτών της ΕΕ για προβλήματα στα νερά και επίσημα δεδομένα σχετικά με την οικολογική, χημική και βιολογική κατάσταση των νερών (βάσεις δεδομένων που σχηματίστηκαν για να βοηθήσουν στην εφαρμογή της ΚοπΥ). Τα πρώτα δεδομένα ήρθαν από την σχετική μελέτη αρ. 334 του Ευρωβαρόμετρου του 2012 και τα δεύτερα από την βάση δεδομένων της WISE (The Water Information System for Europe). Τα δεδομένα που χρησιμοποιήθηκαν από το Ευρωβαρόμετρο σχετίζονταν με την ερώτηση για το τι θεωρούν οι πολίτες των 28 χωρών-μελών της ΕΕ ως απειλή στα νερά και ήταν τα εξής: Αύξηση Άλγης, Χημική Μόλυνση, Λειψυδρία, Πλημμύρες, Αλλαγές στα Υδάτινα Οικοσυστήματα, Φράγματα, Κανάλια και Άλλες Φυσικές Αλλαγές, Κλιματική Αλλαγή, Δεν Ενδιαφέρομαι και Άλλο. Επιπλέον η μελέτη του Ευρωβαρόμετρου συνέλεξε κοινωνικο-οικονομικές μεταβλητές σχετικά με τις περιβαλλοντικές τάσεις των συμμετεχόντων, το επίπεδο εκπαίδευσης, την περιοχή κατοικίας, το είδος εργασίας, φύλο, ηλικία κ.ο.κ. . Τα δεδομένα της βάσης WISE χρησιμοποιήθηκαν σε συνδυασμό με αυτά του Ευρωβαρόμετρου για να σχηματίζουν έναν «καμβά» όπου συνδυάζονταν οι απαντήσεις από το Ευρωβαρόμετρο σταθμισμένες ανά γεωγραφική περιφέρεια της ΕΕ (ανά NUTS 2 περιοχές) και συνδυασμένες με τους επίσημους δείκτες για τα ποσοστά των υδάτινων πόρων ανά υδατικό διαμέρισμα που πάσχουν από α) αύξηση άλγης, β) χημική μόλυνση και γ) αλλαγές στα υδάτινα οικοσυστήματα. Η ανάλυση έπειτα δημιούργησε ένα πρωτοποριακό χάρτη όπου η χημική, βιολογική και οικολογική κατάσταση των υδάτινων πόρων παρουσιαζόταν σε επίπεδο NUTS 2 και όχι πλέον σε επίπεδο υδατικού διαμερίσματος. Αυτό επιτρέπει να επικεντρωθεί η ανάλυση στις περιοχές όπου υπάρχει μεγαλύτερη συγκέντρωση απαντήσεων (και κατοίκων). Η ανάλυση των δεδομένων έγινε μια random-intercept παλινδρόμηση σε ένα Generalized Linear Mixed Model (GLMM) με fixed και random effects και αναλύθηκαν τα αποτελέσματα σε δύο επίπεδα. Το πρώτο και πιο γενικό είναι το επίπεδο χώρας και το δεύτερο και πιο συγκεκριμένο είναι το επίπεδο περιφέρειας ώστε να επισημανθούν τυχόν διαφορές μεταξύ περιφερειών και χωρών αλλά και το ποσοστό ετερογένειας στις απαντήσεις μεταξύ περιφερειών και χώρας. Το τελικό μοντέλο συμπεριλάμβανε 16 μεταβλητές και για να επιτευχθεί η καλύτερη ακρίβεια στην μέθοδο Μεγίστης Πιθανοφάνειας χρησιμοποιήθηκε μια adaptive quadrature approximation. Τα αποτελέσματα έδειξαν ότι η κατάσταση των νερών στην γεωγραφική περιοχή όπου κατοικούν οι συμμετέχοντες στην έρευνα δεν επηρεάζει την επιλογή αυτών των προβλημάτων ως θεωρούμενους κινδύνους για τα νερά, τόσο όσο άλλοι προσδιοριστικοί παράγοντες. Αυτοί οι παράγοντες ήταν εάν οι συμμετέχοντες είχαν λάβει υψηλή εκπαίδευση και η αυξημένη ευαισθησία σε περιβαλλοντικά προβλήματα, που ωθούσαν τους συμμετέχοντες να διαλέξουν και τις τρείς απειλές των υδάτινων οικοσυστημάτων ως θεωρούμενους κινδύνους. Πιο αναλυτικά, συμμετέχοντες με φιλο-περιβαλλοντικές απόψεις είχαν 40% περισσότερες πιθανότητες να επιλέξουν ως θεωρούμενο κίνδυνο την χημική μόλυνση στην χώρα τους. Επίσης, εάν η οικολογική κατάσταση των υδάτινων πόρων στην περιοχή κατοικίας των συμμετεχόντων ήταν σε κακό ή μέτριο επίπεδο τότε, η πιθανότητα να διαλέξουν την χημική μόλυνση και τις αλλαγές στα υδάτινα οικοσυστήματα ως θεωρούμενες απειλές αυξανόταν (και ήταν στατιστικά σημαντική) αλλά δεν επηρέαζε όσο άλλοι παράγοντες. Τα ευρήματα αυτά βρίσκονται σε συνάφεια με την ευρύτερη βιβλιογραφία στον κλάδο των risk perceptions και τα αποτελέσματα ήταν πολύ ισχυρά ανεξαρτήτως εάν αφαιρούνταν συγκεκριμένες χώρες ή περιοχές μέσα στην ΕΕ.
Εν κατακλείδι, η διατριβή αυτή στόχευσε στο να συμμετάσχει στον συνεχιζόμενο επιστημονικό διάλογο και έρευνα σχετικά με τις ακόλουθες κατηγορίες: α) σχεδιασμό αγρο- περιβαλλοντικής πολιτικής, β) οικονομική θεωρία και γ) σχεδιασμό αγρο-περιβαλλοντικών μοντέλων.
Σχετικά με τον σχεδιασμό αγρο-περιβαλλοντικών μέτρων η μελέτη στον ποταμό Λούρο ανέδειξε τις πρακτικές αγρανάπαυσης ως τις λιγότερο κοστοβόρες αλλά επίσης ότι η επίτευξη της ΚοπΥ θα είχε δυσανάλογα μεγάλα βάρη επάνω στους παραγωγούς αγροτικών προϊόντων. Επίσης, ο λανθασμένος σχεδιασμός και ανάλυση της παρούσας και μελλοντικής κατάστασης μπορεί να οδηγήσει σε σφάλματα πολιτικής (τύπου Ένα και Δύο) που μπορεί να έχουν σοβαρό οικονομικό αντίκτυπο (μιας και αποδείχθηκε ακριβή η γενικότερη εφαρμογή τους) ή/και σοβαρό οικονομικό και περιβαλλοντικό αντίκτυπο (εάν δεν ληφθούν εγκαίρως και με τον σωστό τρόπο, ειδικά λαμβάνοντας υπόψη την κλιματική αλλαγή). Η περίπτωση του Λούρου είναι μια περίπτωση μιας λεκάνης απορροής όπου η οικολογική της κατάσταση είναι αρκετά καλή και ως εκ τούτου η εφαρμογή της ΚοπΥ θα είχε δυσανάλογα μικρά οφέλη δεδομένου του κόστους εφαρμογής της.
Αναφορικά με την συνεισφορά της διατριβής στην οικονομική επιστήμη, χρησιμοποιήθηκε ποιοτικά και ποσοτικά η σχετική βιβλιογραφία και αποδείχθηκε ότι μπορεί να υπάρξει αποτελεσματική και οικονομικά ορθή ανάλυση των οικονομικών επιπτώσεων πολιτικών περιβαλλοντικών ζητημάτων μέσω της μεθόδου της μέτα-ανάλυσης. Επιπροσθέτως, η χρήση επίσημων δεδομένων σε σύγκριση με την χρήση δεδομένων για το εισόδημα σε μελέτες Πιθανολογικής Αποτίμησης απεδείχθη ότι μπορεί να προσφέρει σημαντικά και ενδιαφέροντα αποτελέσματα για την αποτελεσματικότητα και την αναδιανομή εισοδήματος μέσω αγρο-περιβαλλοντικών πολιτικών. Τέλος, η εφαρμογή της ΚοπΥ μπορεί να ωφελήσει περισσότερο τα φτωχά νοικοκυριά απ’ ότι τα πλούσια, κάτι που είναι επιθυμητό όταν αναλογιστεί κάποιος την αξία του νερού στην ανθρώπινη ζωή, παραγωγή και κοινωνία. Η συμμετοχή του κοινού στην δημόσια λήψη αποφάσεων πρέπει να γίνεται με προσοχή σύμφωνα με τα αποτελέσματα της διατριβής μιας και οι πολίτες της ΕΕ φαίνονται να σχηματίζουν απόψεις για τα περιβαλλοντικά θέματα χωρίς να λαμβάνουν τόσο υπόψην τους τα πραγματικά προβλήματα των περιοχών που διαμένουν όσο τα προβλήματα στους υδάτινους πόρους στο σύνολο της χώρας. Τέλος, η σωστή εφαρμογή και χρήση μοντέλων μεταφοράς θρεπτικών σε συνδυασμό με μοντέλα κλιματικής αλλαγής εφαρμόστηκε στην περίπτωση του Λούρου και απέδειξε την χρησιμότητα τέτοιων μεταβλητών και των ανάλογων περιβαλλοντικών μετρήσεων για την αποφυγή λαθών πολιτικής αλλά και για την ορθή αποτίμηση της οικολογικής κατάστασης κάθε υδάτινου πόρου.
Ακολουθεί το σύνολο της διατριβής στην αγγλική γλώσσα Income Effects on Willingness-to-Pay measures in the Water Directive
Emmanouil Tyllianakis
PhD Candidate University of Patras School of Business Administration School of Economics
Supervisor: Dimitris Skuras Professor University of Patras School of Business Administration School of Economics
Κύριε, προς τίνα θέλομεν υπάγει; λόγους ζωής αιωνίου έχεις·
- John, Chapter 6, verse 68
Η δε γη ήτο άμορφος και έρημος· και σκότος επί του προσώπου της αβύσσου. Και πνεύμα Θεού εφέρετο επί της επιφανείας των υδάτων. […] Και εκάλεσεν ο Θεός την ξηράν, γήν· και το σύναγμα των υδάτων εκάλεσε, Θαλάσσας· και είδεν ο Θεός ότι ήτο καλόν.
- Genesis, chapter 1: verses 2 and 10
“The rule of no realm is mine, neither of Gondor nor any other, great or small. But all worthy things that are in peril as the world now stands, those are my care. And for my part, I shall not wholly fail of my task, though Gondor should perish, if anything passes through this night that can still grow fair or bear fruit and flower again in days to come. For I also am a steward. Did you not know?”
- J.R.R. Tolkien, The Return of the King
Contents Chapter one: Introduction ...... 17 Chapter Two: Water Framework Directive: an overview, a time frame in Europe and Greece and the role of economics ...... 22 1. An overview of the Directive ...... 23 1.1. Related Directives to the Water Framework Directive and their links to it ...... 25 1.2. A Historical Overview ...... 28 1.3. Requirements and targets of the Directive in a European framework ...... 29 2. The Greek story of implementing the Directive ...... 30 2.1. Articles of importance ...... 32 2.2. Timeline of Greece’s Implementation of the WFD...... 33 3. The Economic Perspective in Water Management ...... 34 4. Cost-Benefit Analysis in the WFD ...... 35 4.1. Defining costs ...... 36 4.1.1. Direct Costs ...... 37 4.1.2. Indirect Costs ...... 37 4.2. Defining benefits ...... 38 4.2.1. Valuing Environmental Benefits ...... 39 4.2.2. Valuing Scarcity rents ...... 41 5. Aggregating and comparing costs and benefits ...... 42 5.1 Aggregating over time ...... 42 5.2. Aggregating over water bodies ...... 43 5.3 Determinants of costs and benefits ...... 43 5.4 Determinants of the level of spatial analysis ...... 43 5.5 Determinants of equity ...... 44 6. Disproportionality: definition, use and applications ...... 45 7. Conclusions ...... 46 8. Reference List ...... 47 Chapter Three: Case Study area: Louros watershed, a case of the implementation of the Water Framework Directive ...... 49 1. Introduction to the key features of Louros’ area ...... 50 2. River and floodplain characteristics ...... 52 2.1. Climate, morphology and significance of Louros ...... 53 2.2. Louros’ current water status ...... 54 3. History of human activity in the Louros’ area and current status ...... 55 3.1. History of human activity ...... 55 3.2. Current population statistics and demographics ...... 55 4. Identification of key pressures and sources ...... 56 4.1. Sectors of emphasis for Louros ...... 57 4.1.1. Agriculture ...... 57 4.1.2. Farming ...... 60 4.2. Water Quality ...... 62 4.2.1. Water quality and geomorphology ...... 62 5. Water Framework Directive in Louros ...... 62 5.1. Chemicals in the Louros catchment ...... 63 5.2. Diffuse pollution sources ...... 64 5.3. Point pollution and its causes ...... 66 6. Conclusions ...... 66 References ...... 67 Chapter Four: Measuring costs in the context of the Water Framework Directive: The case of Louros, Greece ...... 68 1. Introduction ...... 69 2. Alternative agri-environmental schemes and cost elements ...... 69 2.1. Income forgone ...... 70 2.2. Additional costs ...... 70 2.3. Cost forgone ...... 71 2.4. Transaction Costs ...... 71 3. Identification of cost bearers in Louros catchment ...... 78 3.1. Estimation of Marginal Abatement Cost curves for the four available scenarios ... 79 4. Scenario analysis: Incorporating climate change ...... 84 4.1. Climate change in Greece and the Louros catchment ...... 84 4.2. Climate change scenarios and drivers ...... 85 4.3. Land use and climate change ...... 87 4.4. Nutrient application and costs under land use change scenarios ...... 90 5. Scenario setting ...... 93 6. Discussion and conclusions ...... 97 References ...... 100 Chapter Five: Measuring Benefits: The case of the Louros catchment, benefit transfer and an application of meta-analysis ...... 101 1. Introduction ...... 102 2. Identification of Beneficiaries in the Louros catchment ...... 102 3. Benefits assessment ...... 104 3.1. Sensitivity analysis of aggregate benefit estimates ...... 105 4. Meta-analysis: an application ...... 108 4.1. Why conduct a meta-analysis on GES? ...... 109 5. The Water Framework Directive and the Income Elasticity of Willingness to Pay for GES 110 6. Data and Methods ...... 114 6.1. Data ...... 114 6.2. Methods ...... 117 7. Results ...... 118 8. Conclusions ...... 121 Bibliography ...... 123 Appendix A ...... 132 Appendix B...... 143 Chapter Six: A how-to-do Cost-Benefit analysis in the context of the Water Framework Directive manual: combining results from primary and non-primary sources with a sensitivity analysis ...... 146 1. Introduction ...... 147 2. Sources of false positives and negatives in the design of agri-environmental policy 150 2.1. The occurrence of a false positive ...... 158 2.2. The cost of a false positive ...... 159 2.3. A how-to-do guide on decision-making processes ...... 161 3. Sensitivity analysis ...... 162 3.1. On benefits and time horizons ...... 162 4. Results ...... 164 5. Discussion ...... 167 6. Conclusions ...... 168 Appendix ...... 171 References ...... 172 ANNEX A. Determining the selection of 3,5% discount rate in environmental projects.. 176 Chapter Seven: Risk perceptions and environmental policy in the European Union: assessing perceptions on pressures on water bodies using ecological and chemical indicators through an integrated model ...... 178 1. Introduction ...... 179 2. Previous research: perceptions on pressures on water bodies ...... 181 2.3. Previous research: Chemical pollution ...... 185 2.4. Previous research: Changes in water ecosystems ...... 187 4. Methods ...... 197 5. Results ...... 198 6. Conclusion and discussion ...... 204 Reference list ...... 207 Appendix ...... 212 Chapter Eight: Conclusion and contributions: Concluding remarks on Economic Theory and findings of the thesis ...... 213 1. Contribution to policy-making: Abatement cost, Cost-benefit analysis and analysis of disproportionality ...... 214 2. Contribution to economic theory: Risk perceptions and environmental goods ...... 216 3. Contribution of this thesis on research design and focus: modeling ...... 218 References ...... 219
Chapter one: Introduction
‘Making sense of the world’; this was perhaps, and still is, the primary interest of human beings throughout history. Making sense of the natural world around them but also making sense of the living beings inhabiting it. Although humans are considered to be at the top of the “food chain” this “achievement” still doesn’t unlock all the mysteries of the natural world. And if the natural world is still covered with a veil of mystery, the human mind and it’ endeavors are even more unknown and mysterious.
Due to the mysterious nature of the unknown, nature, and in many cases water, were prescribed divine attributes. In ancient Egyptian religion, it was through conflict that the world was created and through the god of water that humanity was placed on earth (Hughes 2012, pp. 23). In Japanese tradition, the waters spirits were responsible for creating and sustaining life but they also demanded life in forms of sacrifices in order to be satisfied and not claim back the life they gave in the first place. The New Testament records Jesus’ words of Him being the ‘living water’ that drives spiritual thirst away, drawing on the image of water springs in the Palestine, an unusual but still extremely important for sustaining life. Rivers, wells and springs have in every religion and area in the world a deep connection with life, birth and death, as well as connection with the underworld. The mysticism surrounding water was initially only perceived and explained through the supernatural and perhaps the continuous search for knowledge and information is what most people believe will be the “water that will make them thirst no more”, paraphrasing the words of Jesus to the Samaritan woman by the well in the New Testament.
The mysterious and the new in the natural world attracted and “kindled the fire” of scientific discoveries and advances in science for centuries, from Archimedes, to Newton and Pascal all the way to Marie Curie and Albert Einstein. The mysterious and the unknown in the world fueled adventurers such as David Livingston and Roald Amundsen to discover new lands and passages with the utmost purpose to satiate their thirst for knowledge but also to make scientific advances that would benefit humanity. Discoveries and knowledge are not always celebrated or profitable though. Lesser known figures craved for new discoveries experienced that, figures such as David Thomson, the man who escaped poverty and child labor through studying mathematics and science and ended up mapping two million square miles of Northern America and Canada, only to die penniless or Alexander von Humboldt, who in the 17th century was a true child of the Renaissance, a rendition of Homo Universalis , a man self-taught of geometry, mathematics, geology, anatomy and mapping who set for South America famously stating “I shall endeavour to find how nature’s forces act upon one another.” Others tried to combine their thirst for discoveries and knowledge with learning more about the human race, like Colonel Percy Harrison Fawcett who set out to explore the Amazon in 1925, only to disappear in the jungle to live, what is believed, a life amongst the natives whom he found fascinating. If only the mind is to wonder and not the body, then, we need to look no further to the works of Jules Verne who populated his works with heroes who were longing for adventure while possessing both physical treats and scientific knowledge to be envied. Discoveries do not take place in the natural world only but also in the world of science and intellectualism. Economics long have been trying to walk the fine line between natural sciences and political sciences, more notably in the frame of thoughts of neo-classical economists. Neo-classical economics tried to move from “mere theory” to mathematics and statistics, economic models became econometric models and people and societies were substituted by “representative agents” and “endogenous choices”. The curiosity remains though; human behavior in an ever-changing world remains a mystery, especially when one considers the implications of human interactions and our dependence, with nature.
Economics take pride into doing exactly that; incorporating human behavior in mathematical and statistical modeling. At least this has been the aspiration of neo-classical economics (Gowdy 2004). Economics maintain that they provide a more concrete basis for conclusions that psychology or sociology when it comes to solving societies’ problems that are caused by scarcity, the fact that not all goods and services are in infinite quantities for all people to enjoy and consume. Economics also claim that they provide this basis which can be (more or less) objective when it comes to policy-making decisions; outcomes of economic analysis are not based on ethics or cultural norms, they are rather based on the “rationality model”, that people/economic agents make decisions that are endogenous (described and determined by factors that exist inside the economic system under examination) and rational (given the alternatives people make choices that would make them, and the people around them, happier by consuming a certain amount of finite goods). Whether neo-classical economics are the best means of representing a society of economic agents and also the best medium of designing policy-making is highly challenged by non-neoclassical economists, sociologists, psychologists and (for what concerns this thesis) climate-change scientists, but, economics (or better, neo-classical economics) do “partake” in both disciplines (social sciences and natural sciences) and therefore should be considered as one guideline to policy making.
Welfare economics are the branch of economics that are focusing on the distribution of goods in a society with the restrictions of social preferences and income. Baumol in his 1952 book “Welfare Economics and the Theory of the State” is perhaps the first to note that social preferences are impacted deeply and in various ways by social factors and that externalities in production and consumption are more common than we expect. Baumol was first to “draw the line” for the subject of the role of the state, by presenting the two opposite sides of the same coin: a state which is limited in its role to restrict and co- ordinate economic activities and individuals who through individual profit maximization attain the most efficient economic outcome. When it comes to examining issues of the natural world, the focus shifts on goods stemming directly from nature, sometimes called “public goods” due to their nature of non-exclusivity and non-rivalry in their consumption. The distinction between the Neoclassical approach and the New Welfare Economics’ (NWE) approach to welfare economics that are occupied with public goods is the effect of preferences in utility, substituting the cardinal utility (measurable by intervals’ scale) to the NWE’ ordinal utility (based on comparing bundles of goods and choosing the better one, but with the level of improvement or loss being unknown. The works of Bernoulli, Edgeworth, Sidgwick, Marshall, Jevons, Walras, Marshall (these last three can also be described as ‘Marginalists’) and Pigou (with the latest spearheading later on with his research the transition to ordinal utility) were responsible for the existence of cardinal utility. Debreau in 1954 proved the existence of the ordinal utility, based on the work of Pareto, and claimed that the difference between various utilities cannot be measured and that matters not since the movement to a new utility curve is what matters most.
The implications of the use of ordinal or cardinal utility affect the ‘expected utility theory’ which defines choices under risk, which is always the framework concerning public goods. From the expected utility theory stem the theories of risk aversion and the measuring of risk. Amongst the criticisms of the expected utility theory, the highest place belongs to the prospect theory of Kahneman and Tversky (1979) who via experiments proved that choices are not always consistent and depend to elements that the expected utility theory considers exogenous in its approach.
Welfare economics have two fundamental theorems: the first states that all agents in the system are price takers with selfish motives to increase the satisfaction from consuming goods while reaching a competitive equilibrium where the allocation of goods is Pareto optimal. The second theorem states that there are potential deviations from the initial equilibrium which can also be Pareto optimal if winners compensate losers. The second fundamental theorem is the basis of non-market valuation, where various techniques try to create hypothetical markets for public/ environmental goods and estimate the expected utility, the “market prices” determined by WTP according to the preferences of economic agents. The critics of non-market valuation, with the leaders being the heterodox economists, amongst others, critique the creation of hypothetical markets altogether. They claim that numerous distortions that have to do with the emotional state of the respondents (Hanley et al. 2016), the unfamiliarity or the abstract nature of the study by asking WTP questions on issues that people have no experience in (Gowdy 2004), the absence of market substitutes for environmental goods in order for real trade-offs to take place or the inability to include preferences and altruistic behaviors in the optimization exercise (Ackerman et al. 2009). The absence of markets and of true substitute goods is making the actual valuation process extremely difficult and some large assumptions must be made in order for the assumptions of the expected utility theory to hold. This is especially true when the environmental good in question is water quality, which is the case in this thesis. It is perhaps easier for economists to turn a blind eye to these (and the many, many other) complications of neo-classical economic theory and continue in their previous course. In this thesis, there is a considerable effort to eliminate these distortions by following careful inspection and design of the research and clearly defining the environmental good under inspection. Still, many other distortions still exist and potential biases are acknowledged to linger.
This thesis aims to walk this fine line between economics and social sciences; utilizing economic theory and tools to their maximum without overlooking their restrictions, but in the meantime have an “anthropocentric” view of policy-making and decision making. In this thesis, no matter the writer’ passion for economics, the focus is fixed on improving human welfare. The aforementioned distortions on welfare economics are attempted to be solved by using rigorous econometric and natural science modeling and alleviate, as much as possible, potential biasness. Therefore, the Water Framework Directive and its implications are thoroughly examined in a dual manner, from both an anthropocentric and an economic point of view. This thesis also contains work that is both “ex” and “post-ante”. This is due to the inclusion of policy work carried out for informing policy-making and the Water Framework Directive in particular but also includes work undertaken to evaluate the findings of the primary work but of the impact of the Water Framework Directive in total, from scientific, political, welfare and economic perspectives.
The following questions are attempted to be answered in this thesis:
1. What is the pragmatic effect of large policy pieces of work such as the Water Framework Directive on income and welfare? 2. What is, if any, the contribution of modeling in designing agri-environmental measures adapted to specific watersheds ? 3. Are large policy mandates cost-effective when specific case studies are concerned? 4. Who is benefited more by such mandates, poorer or richer households? 5. When individuals’ perceptions of risk are considered, how well they represent true problems concerning the water environment and how they should affect policy design and policy making? 6. How do the ecological and environmental indicators set by EU mandates for the water environment help EU citizens define perceived risks, given the growing willingness of the Commission to include public consultation in the design and implementation of such directives
References
Ackerman, F., DeCanio, S. J., Howarth, R. B., & Sheeran, K. (2009). Limitations of integrated assessment models of climate change. Climatic change, 95(3-4), 297-315.
Baumol 1952. Welfare Economics and the Theory of the State
Debreu, G., 1954. Representation of a preference ordering by a numerical function. Decision processes, 3, pp.159-165
Hughes R.A. (2012) Religion, Law, and the Present Water Crisis American University Studies
Gowdy, J. M. (2004). The revolution in welfare economics and its implications for environmental valuation and policy. Land economics, 80(2), 239-257.
Hanley, N., Boyce, C., Czajkowski, M., Tucker, S., Noussair, C., & Townsend, M. (2016). Sad or happy? The effects of emotions on stated preferences for environmental goods. Environmental and Resource Economics, 1-26.
Kahneman, D. and Tversky, A., 1979. Prospect theory: An analysis of decision under risk. Econometrica: Journal of the econometric society, pp.263-291.
Chapter Two: Water Framework Directive: an overview, a time frame in Europe and Greece and the role of economics
1. An overview of the Directive
In the 20th of October 2000 the European Parliament singed the directive 2000/60/EC to “establish a framework for Community action in the field of water policy” in an effort to have an integrated and informed framework for actions, policy, classification and management of the water bodies in the European Union. The water bodies that the directive named Water Framework Directive (WFD) was concerned were inland surface waters, transitional waters, coastal waters as well as groundwater. The directive aimed at establishing an innovative approach for water management based on river basins, the natural geographical and hydrological units, and set specific deadlines for Member States to achieve ambitious environmental objectives for aquatic ecosystems. The directive also adopted a “combined approach for point and diffuse sources” and referred back to several related and pre-existing directives. The various directives of the Union drew upon an integrated river basin management plan that facilitates the implementation of the WFD by providing information of the progress, status and management practices in each river basin of the Member States. Each river basin would then be assigned to river basin districts (RBD). RBDs were defined as the basic management units according to article 3 of the Directive. Coastal waters would be assigned to the nearest or most appropriate river basin district and if a water resource is shared by two or more countries, international river basins would be created. Responsibility to develop appropriate basin management plans befalls to each Member Stat in order to achieve the main goals of the Directive, concerning proper monitoring of the water status and developing and array of programs of measures to deal with pollution control and produce progressive improvements on the quality of surface waters, groundwater and protected areas within each RBD.
The protection of European waters has been a priority for the European Union (EU) since the mid-1970’s when the first directives were produced focusing on the area of environmental and human health protection. These directives referred specifically to drinking waters, bathing waters, shellfish waters, groundwater and water for human consumption (not referring to drinking water exclusively). Later on came directives concerned with discharges of hazardous substances in the water sources that were the cornerstone and the main instrument for controlling emissions from industrial sources.
The Directive of interest aims at tackling the various threats of the water environment within the EU. Reasons behind the fact that a new management plan was needed for European waters were the rise of new pressures such as climate change effects and accelerated economic growth and practices. Particularly, climate change has been rather impactful in the European waters in the form of eutrophication. Eutrophication is a chemical process in the waters due to increased sedimentation of nutrients and is also combined with the phenomenon of ocean acidification. From the one hand there is increased sedimentation of nutrients such as nitrogen and phosphorus compounds that do not dissolve adequately and therefore end up stimulating algae growth. Then algae growth in order to take place requires increased amount of oxygen which then is found in lower quantities for other living organisms like fish which end up dying of hypoxia, the absence of oxygen in the waters. Eutrophication and the resulting algae growth, or algae bloom as it is known, also decreases water quality and increases water temperatures. European waters suffer from eutrophication, particularly the Baltic, North and Mediterranean seas. It was estimated that in 2008 around 40% of all European rivers and lakes would show signs of eutrophication. (Waternote 9, 2008).
The nutrients responsible for the eutrophication comprise 50 to 80% of all water pollution and can come from a variety of sources such as agriculture, farming and soil erosion. In the agricultural sector, fertilizers from nitrogen are applied to the fields and in the farming sector manure from the rearing of livestock is rich in phosphorus. Finally, soil erosion allows phosphoric and nitrogen nutrients to reach unfiltered the groundwater. These are the primary pollutants to the water environment to be followed by wastewater discharges from sewage treatment plants (Waternote 9, 2008).
Despite the good intentions of the pre-existing directives, there was still a gap in the integration and the actual protection of the water bodies. The quality standard approach that previous directives were targeting appeared also to falter in the face of new pressures in the water environment, the product of accelerated economic and industrial growth and of climate change-related pressures such as eutrophication. Advances in science that shed light to effects of pollutant in human health and particularly in the water environment made a new approach to water resources management within the EU more of a priority. Consequently, the EU decided that an approach focusing on the source of pollutants would be more suitable to tackle water quality issues and moved away from the quality standard approach.
This shift in policy led to an array of Directives such as the Urban Wastewater Treatment (UWWT) Directive, the Nitrates Directive and the Integrated Pollution Prevention and Control Directive. The UWWT Directive requires Member States to invest, heavily in times, to infrastructural changes and policies to collect and treat sewage discharges in urban areas. Also, the Nitrates Directive was put in action in order to monitor the amount of nitrogen- based fertilizers applied from farmers to their fields. Finally, the Directive of Integrated Pollution Prevention and Control (IPPC) that came a few years later has as its aim to control large emissions from large industrial facilities (Waternote 9, 2008).
The new, source of pollutants approach needed an integrated set of policies and legislation to operate. This need led to the introduction of the WFD in 2000, which brought forth a new and unified approach to water legislation and management. The WFD operated as an umbrella for existing Directives that their implementation was a minimum requirement for the WFD to put in place in all Member States. They directives that operate under the WFD’s “umbrella” are:
Bathing Water (76/160) (now replaced by 2006/7) Drinking Water (80/778, as amended by 98/83) Urban Wastewater Treatment (91/271) Nitrates (91/676) Integrated Pollution Prevention & Control (96/61, codified as Directive 2008/1/EC). Sewage Sludge (86/278)
The EU with the implementation of the WFD ushered an era of combined approaches in the management of water resources. The combined approach refers to the side of the Member States which need to establish programmes of measures for their respective river basins (or parts of them if they belong to international river basin districts) alongside with the parameters set by the WFD. The combined approach also refers to the point and diffuses sources links and monitoring that the Member States must put in action, alongside with the mandates from the other EU directives.
The diffuse sources are of particular value in the WFD since the emission controls set in the IPPC, UWWT and other directives are partnered with monitoring the use of chemicals in agriculture by requiring the best possible environmental practices and product controls.
The WFD also coordinates the attainment of environmental objectives by providing a new set of quality standards for all water sources. This is called “good status objective”, the “missing” link in other words between emission controls and the achievement of environmental quality objectives, both ecological and chemical. Depending on their current status, country members may be required to obtain more severe controls and legislations to meet WFD mandates and standards. In order for the “good status” to be achieved both in ecological and chemical terms, the need for high-quality monitoring and listing of priority substances is of great importance. The risks of priority substances for human health and their environmental impact are a key feature of the directive (Waternote 9, 2008).
Finally, the WFD has led and influenced other Directives that came after it, such as the Marine Strategy Framework Directive (MSFD) which aims to bring the coastal and marine waters of the EU to Good Environmental Status (GES) by 2020. The approach of the MSFD is not that of integrated management or of quality standard approach but an ecosystem-based approach. This approach takes into account the sustainable management of marine and coastal resources and ecosystems (management that does not further deteriorate the status of those resources) while managing human activities (fishing, transport and recreation are the main themes of interest). Along the framework of the Common Fisheries Policy (CFP) and the mandates of the WFD, the MSFD targets Maximum Sustainable Yield (MSY) status for all EU fisheries. Based on Annex’s X of the WFD priority substances, the MSFD defines GES and targets eutrophication as well. Finally, it focuses on the impact of food webs and ecosystems to human health and supports the creation of Marine Protected Areas (MPAs) and Special Areas of Conservation (SACs) in an effort to attain GES (European Environment Agency 2010)
1.1. Related Directives to the Water Framework Directive and their links to it
The Bathing Directive
It was introduced in 2006 and as it is evident by its name has as its main target to ensure clean and safe waters for the European residents to enjoy by bathing and playing in the,. The Directive came in replacement of a previous directive, dating back in 1976 and introduced a more sophisticated system of monitoring and classification of the water quality of potential bathing sites with its utmost goal to protect human health (Water note 9 European Commission DG Environment 2008).
Following the footsteps of the WFD, the Bathing Directive aims to minimize the risks to bathers by requiring from Member States to draw up management plans for each bathing site separately. Furthermore, the Bathing Directive attempts to include the public as much as possible in the decision-making process by encouraging citizens to participate in the drawing up of the management plans alongside with providing them with extensive information about the risks and the benefits of clean bathing sites.
The Bathing Directive is closely linked with the UWWT Directive due to the fact that the primary danger the bathers are facing is faecal contamination coming from sewage discharges which are inadequately treated from treatment plants or their absence thereof. Sewage discharges and animal waste are the major threats to bathers according to the Bathing Directive. The Directive sets up four distinct categories of bathing sites, ranging from “excellent” to “poor” referring to their quality and the decisive factors for them are the levels of two microbiological factors in the waters: E. Coli and Intestinal Enterococci (Water note 9 European Commission DG Environment 2008)
The Drinking Water Directive
The Drinking Water Directive (98/8/EC) has as its primary aim to ensure high drinking water quality for all EU residents and protect them from any health hazards. The standards set by the directive aim for achieving clean and healthy water for all citizens while abiding with the WHO guidelines.
Again, it’s in the jurisdiction of Member States to access and monitor the quality of their drinking water supply as well as the supply of water used in food production. The directive is comprised of 48 microbiological and chemical parameters that need to be in place for a water supply to be assigned the “useable” tag. Thus, testing of the water supply of member states in the supply network starts straight from the tap at private and public premises. It is in the discretion of Member States to adopt more strict and high standards than those set by the Drinking Water Directive but its required standards are the minimum acceptable ones. (Water note 9 European Commission DG Environment 2008)
The Nitrates Directive
The Nitrates Directive (91/676/EEC) targeted the prevention of nitrated coming from agricultural sources to enter the ground and surface waters.
Member States are required to monitor and detect areas where waters that are already affected by nitrate pollution or likely to be affected and define all these areas as “vulnerable zones”. Furthermore, Member States were endowed with the responsibility to develop action programmes within these vulnerable zones to alleviate the pressures on the water environment and finally access and monitor the success of the action programmes put in place if they achieve the goals of the directive. It is evident that the Nitrates Directive follows in the footsteps of the WFD and the other directives as well.
Member States were demanded to form codes od good agricultural practices that farmers were required to abide to on a voluntary basis. In the cases of vulnerable zones, the Member States were obliged to invest in livestock manure storage facilities to limit the accumulation of nitrogen fertilizers to the soil and consecutively to the water environment (Waternote 9 European Commission DG Environment 2008).
The Urban Wastewater Treatment Directive
The Urban Wastewater Treatment Directive (UWWT) is the final “sister” directive to the WFD and by far the most costly piece of EU legislation in terms of implementation. It refers to 22,000 urban areas across Europe and consists of requirements for pre-treatment plants for industrial and human waste water entering collective systems as well as for the disposal of sewage sludge.
The UWWT Directive is functioning regarding of the sizes of “urban agglomeration” areas which refer to populated areas or of economic activities and on how sensitive are the waters where these areas discharge to. The areas are then classified to “sensitive areas” depending on the levels or of the risks of eutrophication, areas where drinking water is extracted or areas where further treatment is needed to fulfill other directive such as the Bathing Directive. Wastewater discharges are a serious threat to both the environment and to human health. All urban areas with residents exceeding the 2,000 mark are required to conduct at least secondary biological treatment of their wastewater. Areas with 10,000 inhabitants or more, more stringent treatment is required. (Water note 9 European Commission DG Environment 2008)
The cost of implementing the UWWT Directive was extremely high for the new member states of 2004 and 2007 with an estimated cost of 35 billion Euros for the new counties that entered the Union in these years. Poland and Romania alone were estimated that they needed to invest approximately 10 billion Euros each. For that reason the EU established its Structural and Cohesion funds that were estimated to offer 22 billion Euros for investments in wastewater treatment plants, sewage networks and construction of drinking water facilities. The EU required for all the Member States to have water prices that reflected the level of delivery and quality of drinking water to their residents. (Water note 9 European Commission DG Environment 2008)
Measures taken from the EU to protect water quality Source: (Waternote 9, 2008)
1.2. A Historical Overview
It is essential to underline the role and the position of the directive with the numerous agreements between member states and neighbouring countries of the Union. International agreements signed during the 70’s, 80’s and 90’s such as the Convention on the Protection of the Marine Environment of the Baltic Sea Area, signed in Helsinki on 9 April 1992, the Convention for the Protection of the Marine Environment of the North-East Atlantic, the Convention for the Protection of the Mediterranean Sea Against Pollution and finally the Protocol for the Protection of the Mediterranean Sea Against Pollution from Land-Based Sources. The purpose of the directive was to better enable the country members to fulfil the obligations from these agreements and not to overshadow them.
Historically, the first Community address of the issue of water quality and the need for an ecological framework to be established was articulated in 1988 in Frankfurt at the Water Policy Ministerial Seminar. At the conclusion of the seminar, the Commission was urged to assess and propose plans for improving the ecological quality of the surface waters of the E.U. Further recognition to the issue was given by the Ministerial Seminar on groundwater held at The Hague in 1991. There, the need for establishing a framework specifically targeting the quality and quantity of freshwater resources was articulated.
It was only later on in February 1992 and in February 1995 where the Council requested for an action programme for groundwaters to be issued. In these dates the revision of the Council Directive 80/68/EEC that dated all the way back to 1979 on the protection of groundwaters was also revised.
In November 1995 the European Environment Agency (EEA) in a report named “Environment in the European Union” urged for actions for protecting the Community waters both in terms of quantitative and qualitative terms. Shortly after, on December 1995, the Council concluded that a new framework Directive should be established and put in action, therefore drawing the basic pillars for sustainable water policy within the EU and also suggested that the Commission should address it with a proposal. The final activity from the 1995 year was on May 29th when the Commission adopted a communication to the European Parliament and the Council in which the importance of the water bodies have on various functions in the society and thus, the need for their conservation .
On 1996, on different times of the year, the Committee of the Regions, the Economic and Social Committee and the European Parliament asked the Commission for a directive with a clear framework on water policy. It was finally on the 9th of September 1996 when the Commission presented a proposal for a Decision of the European Parliament and of the Council for a programme that referred to groundwater management with sufficient emphasis on regulation on the abstraction and the monitoring of freshwater, while ensuring for the quantity and quality of it. The official launch date came later on, in 2000.
1.3. Requirements and targets of the Directive in a European framework
The directive had as a main target to achieve in a fifteen year period the attainment of “good status” in all water bodies. The driving force behind the directive was a unilateral consensus for tackling further environmental degradation, for an integrated and holistic river basin management plans but also for enough space for Member States to come up with ways and initiatives to define and manage their respective river basins (Hanley and Black 2006). This main target was strengthened by the acknowledgement that aquatic ecosystems, estuaries gulfs in closed seas and inland waters are inter-related and as such, they are also in need for protection to ensure economic and social benefits. The economic benefits will be presented in detail in the end of this chapter.
Holding together this endeavour was the apparent need for close cooperation between the country members of the Union, due to transboundary water bodies. In addition to that, with the launch of the programme, it became apparent that the management and protection of water within the Union required actions in other areas as well. These areas were those of energy, agriculture, fisheries, tourism, regional policies and transport. Legislative actions also had to be undertaken by the country-members to ensure that the framework proposed was integrated and functioning in a Union-wide level. The provisions for the attainment of the good status were laid in the Annex V of the Directive where the foundations of discerning in which state a river basin was, were laid. The ecological potential of every water body had to be assessed and then later classified according to whether it was in a heavily or not modified state. The ecological assessment was not an “one-time-effort” but the directive was the initial start for constant monitoring in terms of quantity and quality due to the serious connections between water bodies and their status with terrestrial ecosystems. The ultimate aim of the Directive was to eventually contribute to the elimination of priority hazardous substances in the water bodies of member states and also to achieve concentrations in the marine environment of normal values to naturally occurring substances that can cause environmental hazards and damages. These substances were classified in Annex III of the Directive and minimum acceptable levels of them were announced. Following the line of the directive, Member States should devise their own measures to eliminate pollution of surface waters from priority substances and gradually reduce pollution from other substances that were in Annex III. Time and patience were of the essence in implementing the directive since early action was advised when dealing with renewable natural resources as well as constant monitoring to ensure that the good status of waters, when achieved, would be sustained and an “upwards trend” would start in terms of the concentration of pollutants. Monitoring did not only extended to mere monitoring of water quality but also in preventing pollution and facilitating the control of it.
The key components of the Directive were as follows. Firstly, water was defined not as a commercial product that can be treated as such, but rather as a heritage to and from people and thus the need to be protected and treated accordingly. Consequently, the increasing pressures on water bodies were acknowledged due to the increasing demand for the production of goods. Then, water bodies used for abstraction used in drinking waters had to be identified in order to ensure compliance with the Council Directive 80/778/EEC of 15th July of 1980 relating to water quality for potable water. Communicating the principles, targets, the process, progress and benefits of the directive to the general public was also required in order to ensure their active involvement to it. Finally, the Commission was also required to provide Member States with annual updates and improvements for any initiative that it intended to propose for the waters in order to assist the Member States to be better prepared and organized to incorporate them to their legislative efforts and measures.
2. The Greek story of implementing the Directive
Greece implemented the WFD starting from December 22, 2000. Greece anticipated that the implementation of the Directive would have only positive effects on the environment, human health and activities and in the management of its water resources. The legislative action that came forth from the Greek parliament was comprised of 17 articles that some included the specific mandates of the WFD and others referred to its implementation in Greece.
The ensuing action was the creation of a National Water Committee that was responsible for devising the policies required for the protection and management of the water environment, monitoring and observing the implementation of the river management plans of each prefecture while informing the relevant minister of the Ministry of Energy and of Environment. The National Water committee included the ministers of Finances, Energy and the Environment, Internal Affairs, Growth, Healthcare and Agriculture and occasionally the minister of External Affairs when international waters are in discussion, which in their turn assigned specific scientific committees responsible of assisting the works of the National Water Committee.
A National Counsel on Water was also formed with the minister of Energy and the Environment as its president where member where also representatives from each municipal and prefectural area, as well as representatives from the Federal Water Service and of the Federal Electric Company. The above two counsels delegated the management duties to the newly founded Central Water Service which is a part of the ministry of Energy and the Environment which was now responsible for:
- Devising the national plans for the protection and management of the water resources, monitor and coordinate their implementation. These national programmes were divided in long term and short term, regarding of their implementation period of six and two years respectively depending on the level of management and requirements in each prefecture. These programmes before implemented had to be authorized by the National Water Committee.
- The Central Water Service is also responsible for the annual report on the status of the implementation of the Directive in the country.
- Coordinate the local governmental services responsible for the implementation of the directive and also participates in any decision-making process from their part.
- Is responsible for devising and distributing the mandates of calculating the implementation costs of each river management plan and is also responsible for the monitoring of their fair usage and implementation.
- Proposing needed legislative actions needed for furthering the cause of the directive. - Monitoring in a national level the quality and quantity of the water sources in the country and takes actions to develop and improve the national network of monitoring of water quality and quantity.
- Using hydrological and meteorological data to access the level of water quality and quantity. Civil services that are involved in monitoring are responsible in communicating their findings with the counsel free of charge as well as any private agent that is operating in the market of water distribution and/or usage.
- Monitoring the functioning of the water services in each prefecture and provides them with counselling and advice.
- To have devised a complete and analytic report on the characteristics of each river basin in each of the 14 Greek prefectures until the implementation of the mandates of the directive in each prefecture. After that the prefectures were responsible at continuing monitoring and accessing the quality and quantity of the waters through their respective agencies and services.
- Creating the National registry of protected areas until the 22nd of December 2004.
- Attaining the improvement, repair and monitoring of all surface and groundwater in good ecological and chemical status, according to the mandates of the WFD by the year 2015. Only in cases of heavily modified waters or of natural disasters occurring that halter the process of implementing the directive, an extension to the programme will be issued for a specific prefecture and a specific water body.
- Founding the Special Secretariat of Water responsible for the development and implementation of all programs related to the protection and management of the water resources of Greece and the coordination of all competent authorities dealing with the aquatic environment.
The implementation of the Water Framework and the Marine Strategy Directives as well of the related daughter Directives fall within the scope of the activities of the Secretariat. The Secretariat, in collaboration with the Regional Water Authorities, formulates and, upon approval by the National Council for Water, implements the River Basin Management Plans and the national monitoring program. The Secretariat is composed of four Directorates and is headed by a Special Secretary, appointed by the Ministry of Environment, Energy and Climate Change and the Government. The Secretariat is responsible for:
- the coordination of all agencies and state institutions, related to water issues and the regional Water Directorates - the implementation of the Water Framework Directive - the implementation of the Marine Strategy Directive - the implementation of the national monitoring program - the implementation of the Floods Directive (Directive 2007/60/EC) - the implementation of the Urban Wastewater Directive and reuse programs - the implementation of the Nitrates Directive - the implementation of the Bathing Waters Directive - Transboundary and international water issues
Practices of water resources were also issued by the government, according with the mandates of the directive, in a 4-chapter, 14-article law.
2.1. Articles of importance
The law, apart from the targets of the directive, described the implementation process and mechanisms adopted by the Greek government and the rules of water use were included in this law passed in December 2004 in the Greek Parliament.
In article 10 of the law the uses of water were identified as water supply uses, irrigation, industrial uses, energy uses and recreational uses. Water supply use is prioritized in terms of quantity and quality from all the other uses. Uses of all water types had to ensure the sustainable fullfillement of needs of individuals and of associations as well as ensuring long term protection of water resources in terms of stocks and of their quality as well. Water demand has to be regulated according to the level of water deposits and of the need of sustaining ecosystems in terms of replenishing deposits and feeding the water supply. As much as possible, the water needs in an area should be fulfilled by local river basins while their sustainable use is ensured. Article 11 defines the system of tapping in the water supply and the system of water usage permits. Permits for any activity that requires water uses must be issued after sufficient proof of water availability is produced by the local authorities and after the purposes of each activity are assessed. Depending on the status of water resources and river basins in a prefecture, extra perquisites may be asked for individuals or associations as well as extra controls and checks for the issuance of a water usage permit. All permits regarding of their uses include a time horizon for their use in their issuance.
Article 13 defines the penalties ensuing improper use of water resources or of degradation of their existing water quality. Depending on the severity of the transgression in the issuance of water usage permits, fines can be in place ranging from 200 to 600,000 Euros. Fines are issued from the Secretariat Office of each Prefecture. In cases of extreme pollution that damages severely the water resources and places human health in great danger, penalties may reach up to 1,500,000 Euros. Businesses that pollute or damage water quality may be ordered to cease their activities until proper measures to stop the pollution are taken and until verification of actual ceasing of the polluting practices are provided in the local authorities. Again, order for such action is given only form the Secretariat Office of each prefecture. Depending on the severity of the pollution, alongside with the ordering for the ceasing of activities, fines ranging from 500 to 50,000 Euros per day may be issued for each day that the business disobeyed the order of ceasing its activities.
2.2. Timeline of Greece’s Implementation of the WFD
Greece proceeded with issuing river management plans for its river basin in the country, which was the main requirement from the EU. The process of designing the river management plans started in the 15 of October 2011 with the river basins of Thessaly, Epirus and Sterea Ellada. In November 18 of the same year, the river management plans of Easter Macedonia and Thrakis were also devised. Following that, in the 21 of the same month, the whole of Pelloponisos’ prefectures issued their own plans and finally, on January 2012 Attiki and Anatoliki Sterea Ellada issued their own river management plans. The remaining prefectures of West and Central Macedonia came later in 2012 with their own plans and the final prefecture was Kriti, in July 2013 when all 14 prefectures of Greece had published and devised their own river management plan. Alongside with the plan, studies on the environmental impact of each river management plan were constructed by each respective prefectural authority. In the 5th and 6th of June 2014 in Crete, Greece hosted the annual meeting of the Managers of Water and Sea water of Europe as well as those of the countries in the process of joining the EU
Concluding, although Greece adopted an adequate number of measures as it was required from the EU it is apparent that the level of communication between the various committees, councils and secretariats makes the decision-mechanism slightly slow. Greece was apparently late, not in starting but, in finishing the process of implementing the WFD as still in 2014 river management plans were still being issued with the deadline of 2015 set by the EU of achieving good ecological status in all water bodies. From October 2000 when the law concerning the EU passed in the Greek Parliament to the last river management plan being issued, there was a lot of period of inactivity from the all sides, both governmental and prefectural.
3. The Economic Perspective in Water Management
The rise in threats to the water environment from both economic activity and of climate change forced the EU to recognize the need for economic analysis and assessment of water quality and quantity in the formulation of the WFD and for that purpose articles 5 and 91 of the directive alongside with annex III of the Directive are dedicated to it. The WFD aims to introduce to Member States a new approach in water management. It is the first piece of EU legislation that truly introduces and integrates economics and economic analyses in its measures (Waternote 5, 2008). Alongside with the ecological analysis of each river basin from each prefectural area an economic analysis of the water use must be issued. WFD outlined that three different levels of management and assessment had to be undertaken. They key feature the WFD emphasized was the cost-effective analysis of each river basin and its resulting management plans. Firstly each river basin depending on the economic activities that are taking place in it. Seasonality in the water supply as well as known pressures to it had also to be integrated in the analysis as well as recovery costs for the water services. Secondly, water bodies that their present state does not abide with the mandates of the Directive had to be identified and lastly, tailored and cost-effective management programmes had to be undertaken to alleviate ecological pollution and account for potential pollution in the future. Finally, the liberty for each country to manage and assess their river basins was again an integral part in the economic analysis of the WFD.
WFD introduced two key economic principles. The first was the “user pays” principle, meaning that users of water resources, either household, farmers or industries, pay for the full services they receive. The second was that Member States are to use economic analysis to formulate the management practices for their water resources, with the cost- effectiveness being the key feature there.
While free to operate in their implementation of the directive, certain principles must be followed by Member States. The purpose of the economic analysis in the directive is for appropriate policies to be developed with strong incentives for civil and corporate users that promote sustainable use of water resources. In addition, there should be no disproportionate burdens on the recovery costs distributed across the industry sector, households and the agricultural sector. The efficacy of this will be analyzed in a subsequent chapter of this thesis. Finally, Member States should evaluate the cost of the application of the different management programmes and again choose the most cost-effective. (Koundouri and Skianis 2013)
In more detail, Member States should focus on five different areas of economic analysis:
1 Source: http://ec.europa.eu/environment/water/water- framework/economics/pdf/pricing_policies.pdf - Get in-depth understanding of the important trade-offs in each river basin and identify those with a positive impact on local and the national economy, - Identify and accordingly assess the most cost-efficient means of achieving the environmental objective of the WFD with regard of the available budgets in each prefecture - Identify the economic recipients of each management plan, most importantly the main beneficiaries and the main victims of each measure, - In case it is needed and for achieving the over-arching target of sustainability, relax environmental objectives to absorb severe economic and social impacts on the population or parts of it, - Develop appropriate economic instruments to account and deal with pollution such as (environmental taxes, pollution charges and fines etc.) to ensure WFD mandates are met more effectively Source : Koundouri and Skianis 2013
Thus, Member States are obliged to incorporate the environmental and resource recovery cost for water services as well as considering the socio-economic and environmental effects of the recovery of the water resources. Waternote 5 (2008) from the European Commission expands on the innovations of the WFD by focusing on the principle and the methodological approach to be followed by the Member States.
The primary target of Member States is to recover the costs of water services. In the above section, the “user pays” principle is referred as a key point but the Directive also stressed out the importance of having realistic costs in the water services. As it is evident from the set of laws passed in the Greek Parliament, operational costs and maintenance costs for the water supply and treatment must be indicative of the water quality they offer. Environmental costs as well as resource costs must also be reflected in the water prices, all in an effort to balance the scales with polluters and users together. No one pays more than the actual natural resources he uses or the damage he creates to them. This principle is further explained by the emphasis on the disproportionality measures taken by the Directive that will be expanded below. Environmental costs include damages caused to ecosystems such as pollution in the waters that damages wildlife and fish altogether and water extraction that makes water scarcer for local flora and fauna. These costs do not appear in financial sheets but it can be incorporated with economic tools such as the cost-benefit analysis and the cost-effectiveness analysis.
To achieve all the above, Member States are obliged to consider all possible water uses and the resulting activities. The three main practices proposed by the European Commission for economic valuation in the WFD, cost-benefit analysis, disproportionality and water pricing are presented in the following section.
4. Cost-Benefit Analysis in the WFD
The general steps that are needed for the Cost-Benefit Analysis (CBA) to be performed are presented in Figure 1. The problem that the CBA is facing is the improvement of the water status by 2015 in a good ecological status. The “Business-as-usual” scenario will be the baseline scenario that would imply the absence of the mandates of the Directive. The likely developments that take place between the time of the analysis and the end year of 2015 were considered afterwards, as were exogenous factors that might influence the income such as changes in pollution emissions due to accelerated economic growth. How the sister directives (Nitrates Directive, UUWT Directive etc) to the WFD interact with the problem are also considered and finally the result is devised by the WFD requirements and the required program of measures. The principle of cost-effectiveness is also present in the analysis, since the most cost-effective selection of measures must be selected. Cost-effectiveness refers to choosing from the plethora of measures the one which has the least negative effect on the economy and the environment. (Water Note 5, 2008)
CBA must aggregate all the costs and benefits over the time frame of the measures to be implemented:
By 2015 all measures had to be put in place, which was the year of achieving good ecological status. The time horizon of the benefits had to be from 40 to 100 years Costs and benefits are aggregated over time, with the use of the appropriate discount rate
After costs and benefits are aggregated, discounted and estimated, the net benefit of the measures to be implemented had to be assessed. In order for robust estimates to be produced, a sensitivity analysis should be conducted, concerning the time horizon, the discount rate and changes in benefits and/or costs from the measures that had to be implemented.
Figure 1. Step-by-step approach for a CBA, source Hanley and Spash 1993.
4.1. Defining costs
There are only two categories of costs in the CBA, the direct costs and the other economic costs which have a wider effect such as changes in prices, external pressures and so on. Direct costs refer to the actual financial costs of implementing the measures of the Directive. They need to be corrected for transfer in the case of VAT (which is transferred to the consumers as a cost) and also the cost of subsidies that are excluded from the CBA. Apart from these direct costs, CBA rests on the inclusion of other forms of costs, usually neglected in mere financial analyses. These costs can be compliance costs, welfare losses or gains to consumers from changes in water quality and availability, spill-over effects in the wider economy as sectors re-adjust to new realities in the water supply sector, transaction costs in the structural changes the Directive causes, environmental costs from the measures being implemented and governments costs in designing and monitoring and enforcing the measures, among many other types of costs.
Each type of cost has to be adjusted though to each analysis separately and every assumption of the level of a cost must be justified properly. Market prices and interest rates may vary from country to country and costs may vary as well. Especially New Member States faced much higher implementation and government costs than other Member States. In addition, when Cost Effectiveness Analysis (CEA) is considered, lost opportunity costs can be substantial for a country when the implementation costs are too high or when some measures cause more harm (disproportionately) to some society groups than benefits.
4.1.1. Direct Costs
Direct costs or otherwise financial costs, refer to the expenditure needed for the measures of the Directive to be implemented. Two over-arching criteria are present in this type of costs: environmental improvements and non-profitability. Environmental improvements refer to the targets of the measures, which need to ensure and promote environmental improvements that can be tangible for the public. Non-profitability refers to the criteria that the measures taken must lead to additional costs that ensure that the measures selected are not meant to provide profits to the local governments but they rather pursue the improvements of the first criterion. Operational costs, maintenance costs and investments are all added in the Present Value of all future costs with the appropriate discount rate. Moreover, in the CBA, the benefits and the costs need to have the same time horizon and discount rate applied to them to ensure uniformity.
4.1.2. Indirect Costs
Financial costs may be the more “tangible” from all the available costs for a policy maker to observe but the existence and the incorporation of indirect costs may render a set of measures plausible or not. A good example for the indirect costs is a possible increase in water outages until measures undertaken to ensure ecological improvements according to the Drinking Water Directive are put in place and the water network in an area is improved.
Indirect costs are the most difficult and critical to assess in the CBA since they may vary and many times be hidden. The best way to elaborate further in this is through examples. The implementation of measures at a river may cause fish migration since better water quality will be present in the waters. Higher fish migration may result to higher recreational activities from anglers and recreational fishermen, an effect that can be considered a benefit, apart from the other benefits that higher water quality brings. On the other hand, higher fish migration may cause a potential local hydropower plant to reduce its production since the presence of fish presents dangers for the operations of the plant. An obstructed functioning of the plant will result to lower production of the plant that leads to lower income. Lower income and reduced production will result to higher electricity prices due to reduced production. The excessive demand for electricity that is created now may lead to more polluting alternatives that may attempt to satisfy the demand for electricity (i.e. burning of fuel). All these may result to higher costs, by implementing measures to achieve the targets of the WFD.
All costs, either direct or indirect, need to be well defined as section 4.1 states but also need to be evaluated after they have been implemented. Potential over-or-under estimation of costs needs to be assessed and if located, corrected with the appropriate techniques and measures.
4.2. Defining benefits
Benefits that are considered in a CBA in the context of the WFD can be various and of a great range. According to the European Commission’s document on CBA in a WFD context, 2 the types of benefits considered are:
Environmental, Scarcity rents, Administrative, Indirect Social They most important benefits for this scope are the environmental and scarcity rents benefits that will be elaborated in the following paragraphs. References to administrative benefits, indirect and social benefits can be found in various sections of this volume, particularly chapters three and four.
Environmental benefits refer to increases in the welfare from mere gains to avoided costs for people, companies and local governments due to the improvement of all the goods and services associated with water. Goods can be drinking water, bathing water and fish population to use our previous example and services refer to recreation and pollution control to name a few. Admittedly, most of these costs fall to the non-market category, such as increased carbon capturing from wetlands and related ecosystem services as well as better aesthetic values due to improvements in the water quality that affect coastal and water ecosystems, making them more appealing to visiting and viewing. Following the
2 Source: http://ec.europa.eu/environment/water/water- framework/economics/pdf/framework_directive_economic_benefits.pdf Millennial Ecosystem Assessment Approach, the European Commission classifies the environmental services provided by water sources as provisioning, regulating, supporting and cultural, the benefits are analyzed accordingly. There is much debate in the literature concerning the definition of Ecosystem Services (ES) and multiple definitions given but works such as the UKNEAFO (2014), Fisher et al (2009), Luisseti et al (2014) and Turner et al. (2014) claim that ecosystem services are, either intermediate or final, the functions and the processes from the natural stock that can provide welfare to the population. The total of the ES is not the total economic value of a specific environment since there are functions and processes that cannot be assessed and valued economically. In other words, the value of the environment is always higher than the monetary value of all the ecosystem services combined, contrary to what Constanza et al. (1997) claim. Regardless, the ecosystem approach presented below is the one adopted by the European Commission regarding the valuation of benefits from implementing the WFD.
- Provisioning services are all the functions that lead to food provision such as better and higher number in the fish populations and better water quality. - Regulating services refer to all functions that lead to improvements in better quality and monitoring of the water environment, such as improved water retention and drainage, providing a healthier ecosystem, saving of costs in the water management and higher capturing of CO² emissions. - Supporting services refer to all functions that lead to provisioning services such as water cycling and carbon cycling that healthier ecosystems provide - Cultural services are all the functions and processes of ecosystems that present a cultural, spiritual and religious value to people. Cultural values in the water environment may possess high values (Jobstvogt et al. 2014)
4.2.1. Valuing Environmental Benefits
The various water-related goods and the ecosystem services that provide them need to be valued. In order to do so, prices must be assigned to them, reflecting the use and value consumers, users and non-users ascribe to the benefits they enjoy from them. These benefits are presented to Figure 2 and according to the European Commission are classified as:
Direct Use benefits Indirect Use benefits and, Non-Use benefits
Direct use benefits can be enjoyed from specific goods and services from water bodies. Better water quality that does not cause health issues to consumers, better quality and higher fish numbers for fishers that lead to higher market prices and anglers and less damages to water companies due to improved water quality that lead to higher income. The means to measure the increased welfare to direct users is through measuring both their willingness to pay (WTP) for enjoying these improved goods and services and their consumer surplus (CS) that is the higher price they would actually be willing to pay to enjoy these goods and services. WTP and CS can be estimated through direct market prices (differences in the value of fish produced in healthy and less healthy waters, increase or decrease in permits for recreational fishing, entrance fees in swimming facilities etc). Indirect market prices such as higher travel costs for anglers to visit water sources, higher housing values for houses located close to improved wetlands and water bodies etc. Finally, contingent valuation surveys (CVM) were respondents state their WTP for enjoying improved goods and services related to water (such WTP to improve the performance of a water treatment plant and improvements in water quality for recreational fishers etc).
Indirect use benefits mainly affect the regulating services and functions of a water body. Better water cycling and improved carbon storage from wetlands are the results of improved water quality due to the mandates of the WFD. In addition, better water drainage and water recycling are regulating services that result from the WFD mandates. All of the above have values that can be captured by decreases in costs in the market. Decreases in costs for flooding control and for water recycling plants are the most evident ones. Increased carbon storage and improved water cycling can be partially observed in decreases in health expenses for respiratory issues. The Millennium Ecosystem Assessment Approach possesses some fallacies due to the fact that many benefits can be either attributed to direct or indirect use benefits and therefore befall in the danger of being double counted. Luisseti et al. (2014) stress out the danger of doing so in a Contingent Valuation study, thus all resulting debate in the literature. Later monitoring and assessing of any method can help avoid such issues.
Non-use benefits engulf all the benefits that do not associate with any type of use. Non-use values in economic literature define option, existence and bequest values Option values refer to potential future use of a good or a service, in other words people place value on preserving or improving the status of a water body due to the fact they believe they might have an actual use value in the future (moving to a place close to it for example). Existence values refer to the value people bestow upon goods and services due to the fact that they merely exist and regardless to the fact if they use them or if they are planning on using hem in the future. Existence values exist for example for improving the ecological status of the Danube river which is a historic landmark for many countries through history. Bequest values or inheritance value as they are known refer to the value people place on goods and services they wish to preserve and pass on to future generations. This is also a notion found in the principles of sustainability meaning that goods and services must be kept in the same, if not better, state and passed to future generations. The following figure adapted from Hanley et al. 1993 shows the different types of values and the appropriate valuation methodologies.
Figure 2. Types of economic benefits
4.2.2. Valuing Scarcity rents
Scarcity rents refer to a fundamental principle in economics. Scarcity is the driving force behind all economic transactions, markets and values. The fact that a good or a service is not in abundance in the world creates demand for it. The scarcer it is, the higher possible value it has. Scarcity rents therefore in economics refer to the cost (otherwise known as opportunity cost) that using a good or a service has, instead of using the next-best alternative. In the context of WFD scarcity rents refer to the cost of not applying the practices of the WFD and thus losing the services from the “good status” of the water resource. Since water resources are used, even though not efficiently even without the WFD, they include costs as does the implementation of measures to reach the targets of the Directive.
The target therefore is resource efficiency in an environment of scarce resources. With the fact that water resources are in many cases mismanaged and overexploited, efficiency in their management is key. Irrigational, transportation and recreational uses of the water resources are the most frequent cases of mismanagement of water resources. The European Commission maintains that WFD and the status it ushers for water sources is the best and optimal use of all water bodies in Europe which ensures that all water sources are allocated efficiently and the economic and environmental values that they produce are the highest possible. This was a rather ambitious declaration that we will challenge in this thesis.
Scarcity rents are assessed accordingly to the differences in their state without the WFD and with it. Water uses referring to agricultural uses, industrial uses and recreational uses if there are any must be valued for pre and ante the implementation of the WFD. Total outputs in a sector (i.e. agriculture) need to be accounted and estimated for a future under the measures of the Directive. The interaction between sectors (i.e. lowest irrigation rates for agriculture may result to more available water volumes for the industry etc) also need to be estimated. This process might be strenuous since it requires modelling of the current and future water uses from all involving sectors and the overall effect on human welfare need to be put in monetary values as accurately as possible. Again, the danger of “double counting” needs to be addressed. Spill-over effects in particular that extend the boundaries of sectors must also be considered. The improvements in drinking water quality for example not only improve the quality of the food industry, bathing and drinking but also to other lower value- added practices such as cleaning, toilet use etc.
5. Aggregating and comparing costs and benefits
The final step of conducting a CBA, after costs and benefits are being identified and quantified, is the aggregation. Aggregation must be applied over a specific time horizon and over the number of water bodies involved. Determining factors of costs and benefits have also to be considered, as well as the spatial level of the analysis and issues of equity as well.
5.1 Aggregating over time
In order for the Present Value (PV) of a management plan to be estimated, a time frame must be established. The time frame refers to how long the estimated costs and benefits are supposed to last and on the discount rates that will be used in the estimation process.
Costs of a specific policy are perhaps the easiest to calculate and is suggested that every CBA starts with valuating them first. Costs of a management start before implementing it in the form of potential investments in order to amass the required resources to implement the measures and also include operation and maintenance costs, until the replacement investments are put in action. Benefits on the other hand can only start to be estimated after the program is being implemented, thus a period of only costs and delayed benefits must be inserted in the analysis. Consequently the time horizon must be long enough for all of the above to be accounted for. Especially for environmental benefits, the time horizon must be considerably long since ecological restoration, even with human intervention, might be a long way away from achieving good status. This is particularly the case in heavily modified water bodies that are quite often the case in the EU.
The time horizon in itself must consider the existence or surfacing of future costs and benefits due to the implementation of measures of water management, therefore discount rates are used. Discount rates reveal the way the future is valued, generally as being “better” or “worse” than the present. Specifically for the WFD, the European Commission suggests the use of a social discount rate to be applied to reflect the “social preference rate”, which is lower than financial discount rates (HM Treasury, 2003). Due to the nature of environmental benefits primarily, a declining discount rate can be applied in the latter years of the time horizon as it better reflects the sustainability goals of the Directive and of environmental management in general (Turner et al 2007). Generally, time horizons in the CBA are from the range of 40 to 100 years time and discount rates vary from 3% to 6%. A sensitivity analysis may be applied for lower discount rates if one considers the current state of climate change and of financial instability in the area of Europe.
5.2. Aggregating over water bodies
Since the WFD required the creation of RBMDs, aggregating over a number of water bodies may be required. Usually primary valuation studies include one or few water bodies in their assessment but the requirements of the WFD require the whole of a river basin to be included in an analysis. Distance decay valuations may also be put in place since the more far a person lives from a water body the more likely is to have a lower use value for it and therefore a lower WTP for environmental and ecological improvements to it. The aggregation of benefits from a water body cannot simply include a whole municipal area but rather people from a specific distance from the river basin in question. Again, this cannot be a matter of simply distance and value since indirect use benefits may rise from large improvements in a water body. For example, a big environmental improvement in a river may attract even more anglers from more distant areas and thus having a higher WTP for it.
5.3 Determinants of costs and benefits
CBA as a methodology uses various indicators to estimate the value of a management program. The Internal Rate of Return (IRR), or the Net Present Value (NPV) can be used, each with its own strengths and limitations. Benefit Cost Ratio can also be used in the earlier stages of designing a project, especially when one wants to compare potential projects with each other, not so much in estimating the benefits and costs of a specific program, due to the fact that that can lead to accounting errors. Another determinant of the costs and benefits is merely on which side some measures are considered. The avoidance costs from a project, or cost savings, can be accounted as a benefit (i.e. more efficient water uses that leads to lower total costs of WFD measures and lower emission abatement costs due to the existence of more pure water bodies etc).
5.4 Determinants of the level of spatial analysis
Water bodies are flowing usually “free” and do not take into account administrative and municipal units. Although in many cases river basins have been the focal point of designing and dividing administrative units, in many times this is not the case. Potential collaboration between administrative units may imply higher management and compliance costs and the need for balance is great in these cases. The topography and geography of every region surrounding a river basin must also be considered alongside with each region’s specific characteristics. Potential gaps between communication, data collection and availability as well as coordination between authorities might increase, in cases dramatically, the costs of management a river basin. The present geology and topography also affects the recipients of the measures (i.e. a river that flows through two different municipal units, with the first being mountainous and not very populated and the second being close to the river’s delta and being much more populated) and thus the level of management needed.
5.5 Determinants of equity
Distributional effects of a policy must always be considered. The European Commission’s declaration that the WFD is the most efficient and the most equity-dedicated management program has to be supported. Therefore, matters of equity must be evaluated. Equity refers to how costs benefits are distributed amongst recipients. In addition to that, the distribution of benefits and costs must be adapted for different income groups, keeping in mind that the lowest income groups are the most vulnerable in terms of costs. On the other hand, benefits might be of greater importance for lower income groups than those for higher ones. Disproportionality analysis will be the topic of the next section where the issues of equity will be addressed thoroughly. Equity also refers to different sectors of the economy such as industry and agriculture. For example, banning or reducing the use of certain nitrate-based nutrients might greatly affect the costs farmers are burdened with but not industrial facilities at all. All of the above must be carefully considered so that WFD does not result to another “top-down management policy”.
The attainment of “good status” for all water bodies in 2015 was the overarching objective of the WFD. Given the nature of the Union however, with new Member States being welcomed, great differences in economic and welfare terms between Member States and also given the fact that Europe is a continent were vast differences between populations exist, some measures needed to implement the WFD might be disproportionately higher for some countries or municipal areas. These “disproportionately expensive” measures as the Commission described them3 , if they are assessed as being disproportionately expensive or not technically feasible at the time, can be substituted with less strict measures as stated in the article 4 of the WFD. This is particularly the case for “Heavily Modified Bodies” and the management practices needed for them. In order for disproportionality as a status to define a water body a Cost Benefit Analysis must be performed in order for the costs and the benefits to be assessed (otherwise walled “Program of Measures” PoM). Not only the costs might be very high compared to benefits but also benefits might be extremely small, regardless of the benefits. If benefits exceed marginally the costs, “disproportionality” might be the term to describe the effects of a potential implementation of the WFD’s mandates.
3 Source : http://ec.europa.eu/environment/water/participation/pdf/waternotes/water_note5_economics. pdf
6. Disproportionality: definition, use and applications
Martin-Ortega et al. (2014) lay out the steps to be followed in order for a Disproportionality Analysis (DA) to be conducted. DA is generally considered as an expensive practice due to the involvement of several political decision-makers and since it has to be performed before any measures or decisions are made. The case of Martin-Ortega et al. (2014) for the Thames catchment in the UK demonstrates the methodology and the results of DA . Similar case studies are conducted in Italy (Galioto et al. 2013), Greece (Psaltopoulos et al. 2017) and Scotland (Hanley and Black 2006).. In addition, the Commission did not disclose reports or guidelines on how to perform a disproportionality analysis, thus this work acted as a reference point in this regard and will be treated as such in this chapter. The methodological approach used in Martin-Ortega et al. (2014) and in Psaltopoulos et al. (2017)are of particular importance since they was part of the REFRESH programme of “Adaptive Strategies to Mitigate the Impacts of Climate Change on European Freshwater Ecosystems” part of an FP7 programme which intended on assessing the economic implications from the implementation of the WFD.
Generally, the DA follows the “blueprint” for performing a CBA in the context of WFD. Since no particular guidelines were given by the Commission, we present below the process followed in Martin-Ortega et al. (2014).
The first steps included pointing out potential benefits from performing a DA, addressing a panel of experts on the specific characteristics of the specific catchment and identifying an initial list of benefits and beneficiaries and invite them in a workshop. The workshop had as its primary goal to inform local stakeholders on potential benefits and costs to society an improvement in water quality would bring. The interesting outcome of that workshop was that farmers were identified as the major cost-bearers in case of efforts to improve the water quality. Local councils, the European Environment Agency (EEA) and a local water company were the other cost bearers.
The spatial and temporal scales of the analysis of disproportionality were identified next, again following the guidelines for CBA. The selected level of analysis was chosen to be the sub-catchment level in order to include the best possible number of beneficiaries and the more accurate number of estimations. The spatial aspect was decided to include the whole of Thames and the temporal scale incorporated the planning cycles of the WFD. 2015 was the year of achieving GES and later stages involved the years 2021 and 2027 with the later being the year where the Thames sub-catchments reach the policy needs of the whole country and those that the WFD demands.
The estimation of costs was the ensuing action taken was market and non-market costs were estimated. The cost estimates used were imported from another study conducted in the UK in 2007 using the benefit transfer method.
Benefits, market and non-market alike, were also calculated. Benefits included the improvements in market goods and services (recreational uses included angling, boating and non-contact water sports) and were estimated from the WTP estimates to pay for environmental improvements in the area. Households’ WTP estimates were estimated using a payment card method (PCM) in a stated preference survey where several mean WTP were reported. Choosing the appropriate value for the DA depended on:
- Improvement level - WTP values at different scales - Targeted population - Elicitation format
Improvement level described eliciting households’ WTP for achieving improvements in water quality according to the WFD proposed policies. The various levels of water quality that respondents were asked to value were classified as low, medium and high that were to be achieved by 2015, 2021 and 2027 respectively. The survey used open-ended format in order to elicit the WTP values, meaning that respondents were asked to state their maximum WTP for securing various levels of water quality.
WTP at different scales refers to the study design that was used for the value transfer method performed by DEFRA and had two levels of interest. The first was regional, asking respondents to state their WTP for local improvements in the water quality and the second was national, for improvements in the water in the whole of UK and Wales. This distinction was made due to the fact that related literature indicates that local changes in water quality usher country-wide improvements.
Targeted population was the designation of people to their respective river basins in order to allow for average WTP estimations, per household per year.
Elicitation format refers to the different contingent valuation techniques used. A payment card question and a dichotomous choice question (dichotomous choice question uses a payment card format but respondents are allowed to select different “paths” according to their initial answers concerning their WTP). In addition, a Choice Experiment (CE) technique was used. Payment card questions usually lead to more downward biased responses as dichotomous choice questions as well.
7. Conclusions
The WFD ushered a new era in water resource management when it was introduced in 2000. The aspiration to combine ecosystem-based approached, integrated river basin management and devise a comprehensive set of quality measures to be fulfilled in regional level was as innovative as it was ambitious. The insertion of economic instruments and analysis in evaluation costs and benefits spurred a surge of case studies employing a variety of economic methodologies in trying to assess the benefits, the costs and the overall reception of the changes the WFD introduced. Given the pressures from a variety of sources such as climate change, diffuse source pollution and anthropogenic activities the WFD appears as a necessity as well as a riddle on how Member States would choose to enforce it.
Reference List
Costanza, R., d’Arge, R., De Groot, R., Farber, S., Grasso, M., Hannon, B., Limburg, K., Naeem, S., O’neill, R.V., Paruelo, J. and Raskin, R.G., 2016. The Value of the World’s Ecosystem Services and Natural Capital (1997). The Globalization and Environment Reader, p.117.
European Commission (DG Environment) December 2008. Water Notes on the Implementation of the Water Framework Directive.Water Note 5 : ‘Economics in Water Policy: The value of Europe's waters’
European Commission (DG Environment) December 2008. Water Notes on the Implementation of the Water Framework Directive.Water Note 9 ‘Integrating water policy: Linking all EU water legislation within a single framework’
European Environment Agency: The European Environment State and Outlook 2010. Marine and Coastal Environment
Fisher, B., Turner, R.K., Morling, P., 2009. Defining and classifying ecosystem services for decision making. Ecol. Econ. 648, 643–653
Galioto, F., Marconi, V., Raggi, M. and Viaggi, D., 2013. An assessment of disproportionate costs in WFD: The experience of Emilia-Romagna. Water, 5(4), pp.1967-1995.
Hanley, N. and Black, A.R., 2006. Cost‐benefit analysis and the water framework directive in Scotland. Integrated Environmental Assessment and Management, 2(2), pp.156-165.
Hanley, N. and Spash, C.L., 1993. Cost-benefit analysis and the environment (Vol. 9). Cheltenham: Edward Elgar.
Jobstvogt, N., Watson, V., & Kenter, J. O. (2014). Looking below the surface: The cultural ecosystem service values of UK marine protected areas (MPAs). Ecosystem Services, 10, 97- 110.
Koundouri P. and Skianis V. 2013 “ Socio-Economics and Water Management: Revisiting the Contribution of Economics in the Implementation of the Water Framework Directive in Greece and Cyprus “ Working Paper ICRE8
Luisetti T., Turner R.K., Jickells T., Andrews J., Elliott M., Schaafsma M., Beaumont N., Malcolm S., Burdon D., Adams C., Watts W. 2014: Coastal Zone Ecosystem Services: From science to values and decision making; a case study, Science of The Total Environment, Volume 493, 15 September 2014, Pages 682-693, ISSN 0048-9697, http://dx.doi.org/10.1016/j.scitotenv.2014.05.099
Martin-Ortega, J., Skuras, D., Perni, A., Holen, S. and Psaltopoulos, D., 2014. The Disproportionality Principle in the WFD: How to Actually Apply it (pp. 214-249). Science Publishers: Enfield, New Hampshire, USA.
Psaltopoulos, D., Wade, A.J., Skuras, D., Kernan, M., Tyllianakis, E. and Erlandsson, M., 2017. False positive and false negative errors in the design and implementation of agri- environmental policies: A case study on water quality and agricultural nutrients. Science of The Total Environment, 575, pp.1087-1099.
Turner, K., Schaafsma, M., Elliott, M., Burdon, D., Atkins, J., Jickells, T., Tett, P., Mee, L., van Leeuwen, S., Barnard, S., Luisetti, T., Paltriguera, L., Palmieri, G., & Andrews, J. (2014) UK National Ecosystem Assessment Follow-on. Work Package Report 4: Coastal and marine ecosystem services: principles and practice. UNEP-WCMC, LWEC, UK.
UKNEAFO Albon, S., Turner, K., Watson, B., Anger, A., Baker, J., Bateman, I., Bentley, S., Blyth, N., Bowles-Newark, N., Brown, C. and Brown, I., 2014. UK National Ecosystem Assessment follow-on: Synthesis of key findings. UNEP World Conservation Monitoring Centre.
Chapter Three: Case Study area: Louros watershed, a case of the implementation of the Water Framework Directive
The current chapter is heavily based on the cumulative work of Dr. Dimitris Skuras, Dr Alexandra Kontolaimou and Dr. Demetrios Psaltopoulos as part of a deliverable for the FP7 programme of REFRESH “Adaptive Strategies to Mitigate the Impacts of Climate Change on European Freshwater Ecosystems.” titled as “The profile of Louros (GR) catchment” in 2010. This contributed to the final deliverable of a FP7 REFRESH report titled “Deliverable 6.12: “Cost effectiveness analysis report for the Louros catchment including analysis of disproportionality” Grant Agreement 244121 by Demetris Psaltopoulos, Dimitris Skuras, and Emmanouil Tyllianakis in 2014.
1. Introduction to the key features of Louros’ area
The Louros water catchment is situated in the central-southern part of the Epirus (NUTS 2 region) which is defined as water district GR05 for the purposes of the Water Framework Directive and the Louros catchment is defined as river district GR13. Map 1 shows the location of the four prefectures in the region of Epirus. Louros extends to 926 Km² in total and to three of the four prefectures of the Epirus region (Preveza, Ioannina and Arta). The Louros river is adjacent to Acherontas river district on the north and the Arachthos river district on east and north. The total length of the river is 72 Km. The density of the hydrographic network in the catchment is 0.69Km/Km2. At the Louros catchment there are two lakes, one artificial due to a hydroelectric power dam at the river’s major drinking water springs, and a natural one contributing to a relatively well-preserved, excellent ecosystem. The Amvrakikos gulf is the natural acceptor of all waters from the three catchments of Louros (north), Arachthos (east) and Acheloos (south). The total discharge to Amvrakikos is about 3 billion 푚3 per year from all adjacent watersheds.
Map 1. location of the four prefectures in the region of Epirus
Map 2. Louros’ watershed
The river’s first springs are in the Tomaros mountain near to the place of ancient Dodoni, the oldest reported ancient Greek oracle. Major rivers contributing to Louros are Xiropotamos of Thesprotiko and Skala (average allotment of 2.8 m3/sec) and Vosa (average allotment of 3.7- 4.0 m3/sec). Vosa collects the water springing from the Hanopoulo springs. These springs have chlorine-sodium (NaCl) and hydrosulfur (H2S) composition and a steady allotment of 3.7 m3/sec and thus influence the quality and quantity of Louros river downstream. The quality of Louros water downstream is laso influenced by Vathi spring which contributes sulfur oxide (SO4). At the commune of St. Georgios (north of the town of Fillipiada) the Louros river accepts water from the springs of St. Georgios. These springs are situated 120m above the sea level and are used mainly for supplying through gravity municipal water that is used by the towns of Arta, Preveza, and the island of Lefkas that altogether account for almost 100,000 inhabitants. At St. Georgios location there is also the first hydroelectric power plant built by the Public Power Corporation S.A (DEH AE) in 1963 with a dam and an associated artificial lake.
Important water bodies in the area include the Zirou lake that share the same acquifers as the Louros river and wetlands at the Amvrakikos Gulf that are formed by the estuaries and deltas of Louros and Arachthos rivers. The Amvrakikos Gulf is a designated Natura 2000 site (GR 2110001). By the estuaries of Louros river are formulated the lagoons of Tsoukalio and Rodia and part of its drained waters contribute to the Loggarou lagoon.
2. River and floodplain characteristics
Louros river has its main springs at the area of Tomaros and Xerovounio mountains and formulates a medium to low density hydrographic network without well defined direction of the secondary streams. The Louros watershed has an area of 926 Km2 (in another study and following another geographic boundary the area is indicated as 685 Km2. The watershed can be viewed in Map 2. The secondary streams of Xeropotamos Lourou, Xeropotamos Thesprotikou and Triantafyllias contribute to Louros’ main stream. At the point where Louros changes its direction from North to South to East to West, the secondary stream of Vosa contributes to Louros river. The drainage system of Kommenou contributes to Louros at the area Voida‐Mavri. The overall density of the network is 0,69 Km/Km2. The length of the river is about 72‐73 Km. The major springs of Louros include the springs of Terovo, Saint Georgios, Kambi, Skala, Priala, Louros springs and Hanopoulo.
Due to the geology of the watershed, Louros presents a very steady allotment with relatively low fluctuations even during the summer months. At the lowlands and the site Petra bridge, the average allotment of the river is measured at 60,866 m3/h, with a maximum allotments during the wet season measured at 161,676 m3/h and a minimum during the driest season at 144 m3/h. At the upland area and the Zita site the respective allotment figures are 13,105 m3/h average, 34,812 m3/h maximum and 2,362 m3/h minimum.
At the Louros catchment there are two lakes, one artificial behind the Louros dam and a natural one, the Ziros lake. The Louros dam with a length of 100m and a height of about 23m creates an artificial lake of about 320 Km2. Due to deposits, the dam has no longer depth and has lost its ability to withhold water. The dam channels the water overflowing the crest to the hydroelectric power plant. The first plant was established in 1954 (Louros I and II) with a power of 5 MW and was followed by Louros III in 1964 which added 5MW to the existing power. The water is released back to the river from the power plant. As such, the hydroelectric power does not consume water. However, the operation of the dam is important for the water environment downstream. The dam withholds the supply of the lowlands with deposits and thus alters the formation of embankments and affects the operation of the lagoons. The natural lake of Ziros (map 8) is found wets of the springs of Saint Georgions and is the natural tensiometer of the Louros catchment. Geologists argue that the lake was a karstic lake‐cave that collapsed 10,000 years ago and formed the Ziros lake. Its surface covers 0,25 Km2 with an average depth of 49 meters and a measured 10 maximum depth of 70m. Ziros lake is considered an excellent ecosystem, relatively well preserved.
Map 2. The river and hydrological network of the Louros catchment.
The Amvrakikos gulf is the natural acceptor of all waters from the three catchments of Louros (north), Arachthos (east) and Acheloos on the south.It is estimated that the total discharge to Amvrakikos is about 3 billion m3 per year from all neighboring watersheds.
2.1. Climate, morphology and significance of Louros
The general climate of the Louros area is temperate Mediterranean with mild rains (with the exception of the mountains) and extensive sunshine. The geology in the catchment exhibits a very indicative karstic relief. The area is dominated by permeable limestone that formulates a dense sub-surface karstic hydro network. Due to the geology of the watershed, Louros presents a very steady water allotment with relatively low fluctuations even during the summer months. The soils of the catchment are classified into two categories; those in the upland are typical soils of limestone and flysch. On sites with no or medium slopes the soils are rich and suitable for forestry and agricultural cultivation. However, on sites with high slopes, erosion due to grazing and infrastructure development has had serious impacts on soil and in several sites the mother limestone rock is revealed. In the plains of Preveza – Arta, tertiary depositions have formulated a very deep and rich soil.
Average monthly temperatures range from 9.3 C˚ to 25.3 Co for Preveza and 7.7 C˚ to 25.3 Co for Arta. The average maximum temperature for Preveza is recorded in August (29.4 Co). The average minimum for Preveza is recorded in January (5.9 C˚). Absolute minimum and maximum temperatures have been recorded at -4 C˚ and 38 C˚ respectively. Annual precipitation ranges from 1,033 mm per year in Arta to 868 mm per year in Preveza. Almost 70% of the precipitation occurs between October and March. The wet season for Arta extends from January to May and from September to December while the dry season (the season when plants need irrigation) is restricted only to June, July and August. For Preveza the dry and wet seasons are respectively almost the same.
In the Louros catchment area we may identify 4 bioclimatic types. In the uplands and near the Louros’ springs we identify the wet-cold type with abundance of precipitation and average minimum temperatures of the coldest month range from 0 to 3 C˚. The sub-wet cold type is identified at the south and east uplands of the catchment. The sub-wet temperate type is identified in the medium between uplands and lowland heights (up to 200 m) with temperatures of the coldest month ranging from 3 to 7 C. Finally the lowlands surrounding Amvrakikos gulf are characterized by a wet-temperate climate with significant precipitation and average temperatures of the coldest month ranging from 3 to 7 C. As concerns the bioclimatic level according to UNESCO – FAO the uplands (over 500m) are characterized as mild medium Mediterranean with biologically dry days of about 40 to 75 while the levels (less than 500 m) are characterized as intense medium Mediterranean with biologically dry days extended to 75 to 100.
Louros is a river of great significance at local, national and international level due to its multiple uses and environmental value. The river contributes with its delta and estuaries to one of the most important Natura 2000 sites, namely “Amvrakikos Gulf, Louros and Arachthos Delta”, while the wider wetland area forms ecosystems with particularly high ecological value. Farming (agricultural and livestock production) is the main land use around the Louros catchment, whilst there is also some limited manufacturing activity. In addition, common uses of the river’s waters include fisheries and fish farming units, hydroelectric power/energy production as well as water abstraction for municipal purposes.
2.2. Louros’ current water status
Although there is a lack of systematic environmental monitoring of the quality and quantity of the river’s surface and sub-surface water, fragmented scientific work (e.g. Albanis and Hela, 1998; Kotti et al., 2005, Konstantinou et al., 2006; Tsangaris et al., 2009) shows that the state of the Louros river has been affected in both qualitative and quantitative terms. Chemical analyses undertaken in several monitoring points indicate high conductivity and high concentrations of pollutants mostly at the river’s estuaries and the Petra Bridge located at a point after the Vosa drainage channel has returned drained irrigation water to Louros.
3. History of human activity in the Louros’ area and current status
3.1. History of human activity
The prefecture of Arta was annexed by the Greek state in 1881 while the prefecture of Preveza was annexed in 1912. From that time on the development of Arta and Preveza starts to put pressure on the area’s resources and especially the timber used for construction. Trade also develops and the dairy products of the mountainous areas of Preveza and Arta are transported to other Greek towns and especially in Patras and Athens. During the period between the two world wars the port of Patras gradually overpowered the port of Preveza in the total of trade performed which envoked negative impacts on the urban centres of Preveza and Arta. After World War II, significant pressures arose on land from human activity that had as a result the extensive marsh drainage and the conversion of marshes to fertile agricultural land. At the same time grubbing and land clearing of forests and partly forested areas in the mountainous resulted to new agricultural land and pastures for the then growing livestock sector. In 1947 the establishment of the first dam in Saint Georgios area restricted Louros’ overflows and started influencing the rate and quantity of sediment transportation that allowed for the construction of the first land reclamation (irrigation and drainage) works. The restriction of Louros’ floods was finalized with activities on the river’ side embankments. However, the reduction of sediment transportation affected the Tsoukalio and Loggarou lagoons. During the 70s the lowland areas experienced urban and tourism development while agriculture and especially the orange tree plantations increased significantly. Food processing plants and increased livestock capital complemented agricultural development in the 70s. In 1985, the establishment of Pournari dam and the extensive land reclamation projects in Arachthos river affected the south-east Louros catchment as most of this area is now irrigated by Arachthos and not by Louros. After the 90s, population ceased from moving to large cities and the livestock capital stabilized as a result of stable employment. The agricultural sector went through changes due to the implementation of the EU’s Common Agricultural Policy (CAP) reforms while manufacturing industry shrunk.
3.2. Current population statistics and demographics
The major urban settlements in the Louros catchment is Fillipiada with a population of 4,321 in 2001, followed by the commune of Louros (2,152 inhabitants) and the commune of Thesprotiko (1,678) and the commune of Anezi (1,619). The Louros catchment is between two major cities of Epirus (Preveza with 18,289 inhabitants and Arta with 24,725 inhabitants) and at close distance to Ioannina (75,550 inhabitants), proximity that allows for trade and daily travelling for some of the communes of the Louros catchment. Inside the crest line as we have defined it, there are 34,165 inhabitants but with the communes that are administratively part of the catchment the total population almost doubles at 67,429. The age structure of the population shows that people older than 65 have a higher share in the mountainous communes indicating ageing population. Mountainous communes in the prefecture of Ioannina really present an ageing population and an extremely low rejuvenation index. All other areas have a higher share of older residents than the Greek average but this is typical of rural areas in Greece and the situation is slightly better in the lowland areas of Louros in Arta and Preveza.
Unemployment is high and activity rates lower in the mountainous and more remote areas of the communes in the prefecture of Ioannina than in the other areas of the Louros catchment. As for sectoral economic activity, the primary sector is well developed in all areas and significantly higher than the respective Greek average, and the tertiary sector less developed.
Manufacturing activity is really restricted to food processing with the most significant units being a meat processing and a tomato juice unit, one slaughter house and small olive oil extraction and refinery units and a low capacity cheese making unit. The tourism sector is not well developed within the geographic boundaries of the Louros water catchment and thus not significant. The only area that has been developed for tourism is the area of the commune of Nikopolis with a capacity of almost 500 beds.
4. Identification of key pressures and sources
Relevant studies indicate agricultural and livestock activity, and population settlements as the principal sources of pollution however, the exercised pressures are not of the same nature and intensity all over the catchment. Following GeoData procedures, the Louros catchment is divided into three very distinct sub-catchments . Sub-catchment A is the uplands including the high altitude space which is the major water supplier to the river extending up to the Louros hydroelectric dam. Sub-catchment B, i.e. the south-eastern part enclosed between the Louros river and the Arachthos water crest includes the Arta plain, while sub-catchment C has its boundaries with the Acherontas water crest and the Louros river itself comprising the Preveza plain.
The uplands, i.e. sub-catchment A are largely pollution-free because of the very sparse resident population and very limited economic activity. The riparian zone and the physical and morphological characteristics of the river in sub-catchment A are also of a good status. On the other hand, water quality in the Arta plain sub-catchment is moderate due to nutrient loads caused by mainly agricultural and livestock activity, and inappropriate waste management. Waste from hog and poultry farming constitutes the major source of point pollution in this sub-catchment. Abstraction rates by agricultural activity are very high, while abstraction rates for municipal use are growing.
Water quality in the Preveza plain sub- catchment is also moderate due to intense agricultural activity and extensive irrigation and drainage works. In the estuaries of Louros in this sub-catchment, it was thought that pollution from nutrients and high conductivity were present. For this reason, the Arta-Preveza plain was characterized a Nitrate Vulnerable Zone by the Greek state in 1999 (article 2 of Common Ministerial Decision 19652/1906/1999 - Β΄1575 as this was later completed by article 2 paragraph β−5 of Common Ministerial decision 20419/2522/2001 - Β΄ 1212.
However, and despite the notable lack of supporting monitoring data, the Ministry of Rural Development and Food announced an agri-environmental programme to confront nitrate pollution in the Arta-Preveza plain in 2006 with Common Ministerial Decision 50981/2308 B’ 1895. Abstraction rates for municipal use grow at a very significant rate while many small villages and medium sized town do not treat their municipal wastes. Overall, with respect to water quality, nitrification problems are identified in the sub-catchments of Arta and Preveza plains, and phosphorus pollution is found to be particularly relevant to the sub- catchment of the Arta plain.
4.1. Sectors of emphasis for Louros
4.1.1. Agriculture
The sole source of information about agriculture is the agricultural censuses due to them being the only statistical source that reports results to commune level and thus, we are able to devise the current status of the sector in the Louros catchment by adding up the outputs of the communes. Other sources of information such as the Community’s Agricultural Structures Survey that also are most recent do not provide such detailed spatial information and its use can be misleading.
Table 1 shows the distribution of farms in the different farm types for 1991 and 2001. In the Louros catchment, specialized agricultural farms and mixed farms share an almost equal proportion while in Greek agriculture the proportion of mixed farms is less than a quarter of total farms. This shows the significance that the livestock sector has in the area despite the fact that the number of specialized livestock farms is not significant. Between 1991 and 2001 the number of farms dropped significantly especially in the mountain areas (all farms are located in the Ioannina prefecture) and increased in the lowland areas of the prefecture of Arta. In the mountainous areas of Ioannina the proportion of owned to total land is almost 90% but in the most economically viable areas of the lowland Arta and Preveza the respective proportion drops to around 75% which, however, remains a strong sign of family owned farming.
Table 2 shows the average farm size in the area in comparison to the Greek average farm size. The average farm size of farms in the Louros catchment is almost half of the average farm size in Greece indicating the small and non-viable farms in the area. However, one should take into account that agricultural specialized farms include tree plantations and livestock farms include intensive hog farming enterprises and poultry farms.
The distribution of land in the prefectures of Louros is dominated from annual cultivations that occupy the majority of the agricultural land and then by grazing lands, followed closely by tree plantations. Vineyards are the last section of land use and they occupy small and stable percentage of the land through the years. In the mountain areas of the communes of the prefecture of Ioannina grazing land has increased its share reflecting the gradual abandonment of cultivations. In the communes located in the prefecture of Preveza we witness a gradual shift to tree plantations while in the lowlands of Arta tree plantations prevail. Table 3 shows the percentage distributions of land uses.
Unfortunately there are not recent estimates of water demand for agriculture. However, the most recent study titled “Pilot Study for the Management of Water Resources of the Water District of Epirus by the Ministry of Industry, Energy and Technology, in 1993, provides some estimates of water supply and demand for the Louros catchment. The average (smoothed over a 12 month period) supply of Louros is estimated to 5.46 m3/sec for medium precipitation conditions (the same figure decreases to 4.2 m3/sec for dry years and 3.74 m3/sec for very dry years. Demand for agriculture, livestock and fish farming is estimated at 3.13 m3/sec which is higher than the river’s supply in the driest year. This demand is estimated on assumptions concerning water needs of the various cultivations irrigated by the land reclamation projects of the Louros catchment. Fish farming is assumed with no effects on water demand.
Table 1. Farm distribution in the Louros’ catchement
Table 2. Average size of farms in the Louros catchment.
Table 3. Distribution of land uses in the Louros’ prefectures
4.1.2. Farming
Farm fragmentation is very important because it imposes certain restrictions to mechanization and increases production costs and secondly it makes monitoring of agricultural programmes difficult and time and resource consuming. At the same time farm fragmentation maintains the ability of a farm household to farm under various micro- environments and differentiate its production base. Farm fragmentation is reduced after land reclamation projects are performed due to compulsory land consolidation projects. Table 4 shows the status of farm fragmentation in the Louros’ catchment and in relation to Greece for 1991 and 2001. It is noted here that the fragmentation of the communes located in the mountainous areas of Ioannina are close to the Greek average with every farm cultivating an average of almost 6 plots while for communes located in the lowland areas of Arta this figure drops to almost half the Greek average. This is due to the consolidation projects followed land reclamation in the area. However, the average size per plot still remains at about the national average for the farms in the lowlands and significantly behind the national average for farms in the mountains.
Table 4. Number of farms, size of plots and plots per farm for the Louros’ catchment
Table 5 shows the areas irrigated and their percentages in the Louros catchment in 1991 and 2001. Agriculture in Louros catchment presents higher levels of irrigation than the Greek agriculture. However, areas in the mountains present very low degree of irrigation indicating first that most of the area is non-irrigated grazing land and secondly the shortage of water in the mountains is due to lack of springs because of the karstic geology of the area that drains most waters to lower areas. In the lowlands of Arta the land irrigated is almost 80% of U.A.A. It is also important to observe that the land capable of irrigation is significantly higher than the area actually irrigated showing that some of the land cultivation is not economically viable or, at least is marginal.
Unfortunately we have no detailed information on the structure of cultivations in the Louros catchment because these are only published at prefectural level. However, we will present some information that concerns with the prefecture of Preveza and that covers a significant part of the Louros catchment from the annual agricultural statistics of 2004 published in 2009. In 2004, arable land in Preveza covered 5,921 ha of which 5,186 was corn and 542 ha of oats. As concerns with tree plantations, these covered 10,892 ha of which 779 ha are covered by citrus trees and 9,982 by olive trees. Preveza produced in 2004, 3,638 tones of oranges and 1,381 tones of lemons. In the livestock sector, Preveza produces a very significant amount of milk (43,898 tones) mainly from sheep and goat (almost 60% of it). Meat production also is considerable with almost 13,000 tones but cheese making and dairy product production is low indicating that milk processing takes place outside the prefecture, in the nearby prefectures of Arta and Ioannina.
Table 5. Irrigation of agricultural land in Louros’ catchment
4.2. Water Quality
4.2.1. Water quality and geomorphology
Water quality is one of the key points of emphasis for the entire WFD. The first point of sample collections was at the river’s estuary near the town of Preveza, the second at Petra Bridge after the Vosa drainage channel has returned drained irrigation water to Louros, the third at Saint Georgios where there are the major drinking water springs and the hydroelectric power dam, and the fourth at the river’s upland springs.
The results of sampling show that the water at the river’s estuary has extremely high conductivity which indicates the presence of salt water and intrusion of sea water in the fresh water. Conductivity is still high in the Petra point indicating again salt water coming from the drainage channel of Vossa and decreases as one moves to the uplands reaching normal values 27 at Zita. Cations and especially those coming from agricultural activity are extremely high at the river’s estuary and at the Petra bridge due to the water drained from the plains of Vossa. Anions and especially sulphur show again the effects of agricultural activity in the area. Finally the high concentrations of chlorine at the estuaries should be attributed to the drinking water before treatment.
5. Water Framework Directive in Louros
Unfortunately, the application of the WFD in Greece has shown a considerable lag. The Louros catchment has been recognised as river district GR13 within the Epirus water region (GR05). Due to the considerable lag in the application of the WFD, no managing authority is in place and yet, there is not any management plan. The monitoring of surface and subsurface waters is rather fragmented and piecemeal. The Prefectural Administration of Preveza is implementing (since early 2009) a Master Plan on the Water Catchments of Louros, Acherontas and Amvrakikos Gulf. The Prefectural Administration of Arta started a project concerning with the Creation of a Monitoring System and Assessment of Environmental Situation of Louros, Arachthos and Amvrakikos Gulf involving scientists from the University of Ioannina. Table 6 summarizes the actions that have been taken by 2012.
Table 6. Actions under the WFD for the Louros catchment
5.1. Chemicals in the Louros catchment
Greece adopted the WFD 2000/60 by Greek Law 3199/2003 and Presidential Decree 51/2007 which set up the “measures and processes for the integrated protection and management of waters”. The Common Ministerial Decision 39626/2208/2009 defined “measures for the protection of sub-surface waters from contamination and degradation in accordance to Directive 2006/118/EC”. In 2010, Common Ministerial Decision 51354/2641/Ε103 adopted Environmental Quality Standards for the concentration of certain pollutants and priority substances in accordance to Directive 2008/105/EC. In 2011,Ministerial Decision 1811/3322/Β’/30.12.2011 defined standards for the concentration of certain pollutants. Still, the Greek law has not adopted standards for coastal and transition waters.
Concentrations for total phosphorous are not defined by environmental standards but, taking into account experience in other European countries with comparable concentration levels for the other substances, we may propose a level of 0.1 mg/l for Soluble Reactive Phosphorous (SRP).
In addition to the aforementioned environmental standards, article 3 of the Common Ministerial Decision 39626/2208/Ε130 allows for setting stricter maximum allowed concentrations for a sub-catchment or a group of sub-surface waters with the decision of the Regional Secretary. It is envisaged that, following the finalization of water management plans, many Regions will set up stricter standards for sensitive waters that are related either to human health (potable water) or to environmental reasons.
5.2. Diffuse pollution sources
Unfortunately there is not a complete monitoring system in place and thus, there are no complete studies identifying the key diffused pollution problems and pressures in the Louros catchment. Various fragmented studies, however, reveal that the major problems concern with:
1) Nutrient enrichment from agricultural fertilizers (Kotti et al., 2005), manure leaching, sewage disposal and inappropriate waste management (Kotti et al., 2005) especially in the remote and small communes. 2) Pollution from inappropriate application of pesticides (Albanis et al., 1995; Albanis and Hela, 1998; Konstantinou et al., 2006; Tsangaris et al., 2009) and inadequate waste management Kotti et al. (2005) have found that while BOD, COD and nitrates were within some acceptable limits, the phosphate content was much higher than the upper limiting criteria for eutrophication for salmonid waters allowed by the Fresh water Fisheries Directive (75/659/EEC). The same researchers argue that the inorganic nutrient load should be attributed to sites that drain agricultural areas, especially during winter and spring while the organic matter should be attributed to urban activities especially during autumn. Tsangaris et al. (2010) argue that biochemical markers in mussels in the Amvrakikos gulf reveal stress conditions and provide early warning signals of pesticide contamination. The researchers found exposure to organophosphate and carbamate pesticides and possible contamination from organoclorine pesticides.
Konstantinou et al. (2006) found that the compounds most frequently detected were atrazine, simazine, alachlor, etolachlor and trifluralin of the herbicides, diazinon, parathion methyl of the insecticides and lindane, endosulfan and aldrin of the organochlorine pesticides. They also found that the detected concentrations of most pesticides follow a seasonal variation, with maximum values occurring during the late spring and summer period followed by a decrease during winter. Albanis et al. (1995) that the herbicides atrazine, simazine, alachlor, metolachlor, trifluralin and diuron and the organochlorine insecticides (u-BHC, lindane and 4,4’-DDE) were detected in river estuaries of Louros and Arachthos rivers and wetlands of the Amvrakikos gulf. The highest concentrations of herbicides, atrazine, simazine,a lachlor, metolachlor, diuron and trifluralin in water samplesw ere detected during the period from March to August. The same herbicides and organochlorine insecticides as well as the P-BHC and 4,4’-DDT were found in significant amounts in sediments of river estuaries and wetlands of the Amvrakikos Gulf. The percentage of total amount of detected pesticides released through the rivers into Amvrakikos Gulf are estimated as 3.1% for atrazine, 1.7% for simazine, 1.9% for diuron, 0.3% for metolachlor, 0.9% for alachlor, 0.3% for lindane and 0.6% for trifluralin. Seasonal variations and riverine input of pesticide residues to the coastal zone were determined by Albanis and Hela (1998) in Louros River for the period of 1995 and 1996. The sampling sites for the 28 determination of fluxes were located at the main river flow and its estuary at the boundary between freshwater and the brackish zone. The inputs of the five major herbicides, atrazine, simazine, alachlor, metolachlor and desethyl-atrazine (DEA) to the Louros River are mainly from tributaries and the agricultural area draining to the river estuary. The highest concentrations of these pesticides occurred in May and June seasons, just after their application. The seasonal variation studies showed a continuous presence of triazines, alachlor and metolachlor, which were detected in the dissolved phase throughout the year at 0.024.27 pgL. whereas the other pesticides exhibited a sporadic occurrence related to agricultural and imgation practices.
Atrazine and its degradation product DEA are the most abundant herbicides discharged into Amvrakikos Gulf, followed by metolachlor, simazine and alachlor. Their annual mean flux was estimated as 122.7 g/day for atrazine, 127.5g/day for DEA, 49.1 g/day for metolachlor, 43.9 g/day for simazine and 11.2 g/day for alachlor.
Trace metals (A1, Pb, Zn, Cr, Cu, Ni, Fe, Mn) were studied in waters (dissolved and particulate phases) and sediments of the Louros Estuary in the Amvrakikos Gulf by Scoulos et al., 1995. The study system of Louros, despite small for such a study and with a relatively narrow mixing zone, is typical for Mediterranean estuaries. Particular emphasis was given by the researchers to understanding the conditions prevailing in the summer. The authors argue that: “…During this season saline water intrudes the estuary along the river bed, despite the existing shallow sill, and forms a thin saltwedge water mass, which occupies the near bottom layer with its thin end pointed upstream. Particulate metal concentrations within this saline bottom layer are considerably higher than in the riverine and marine sections of the estuary. Since the metal content of particles collected upstream is higher than that of the marine ones, there is a clear evidence that the salt-wedge acts as a 'sink' for most metals during the summer. Coexistance in the same zone of high dissolved metal concentrations indicate that loosely associated metals are desorbed from riverine particles, whereas newly formed suspended matter is deposited together with particles, transported by the river. The accumulation of metals in the near bottom layer affects directly their distribution in the sediments. The maximum concentrations of the metal fraction which is loosely held in sediments are found primarily at the same site. The distribution of the 'nonlabile' metal fraction of the sediments (particularly for Cu and Pb) is broadly constant throughout the estuary, confirming the absence of any significant natural or industrial point sources, at the lower part of the river. However, the analysis of sediment cores reveals an enrichment of this metal fraction at the top, near surface sections of the mouth area, indicating relatively recent pollution.”
5.3. Point pollution and its causes
Point pollution in the Louros catchment has three distinct sources: the livestock sector, manufacturing industry and municipal activity. Wastes from hog farming constituted the major pollution problem in the catchment up until 1991 when most of the hog production units relocated to sites further away from Louros river and gradually adopted liquid waste treatment installations. Treated wastes are directed to the soil as a fertilizer. Poultry production does not constitute a major waste pollution problem but there is an odor issue. In the Louros catchment there is the commune of Kostakion at the administrative boundaries of the catchment (outside the catchment’s crest) where almost 60% of poultry production is located. Pollution from manufacturing activity is not extremely important. One meat processing industry has a 29 license to direct treated wastes to Louros. Slaughter units and olive oil extraction units do not constitute a major issue as concerns pollution. Dairy firms cause some pollution especially the older ones but the size of operation is not significant. All major urban centres in the catchment have integrated municipal wastewater treatment plants and do not pollute. However, municipal solid wastes remain an important issue and cause pollution of the aquifer.
6. Conclusions
The Louros catchment is a river that shows a great range of disparities across it’s spectrum. It is not a large water body accommodating irrigating-intensive practices like Pineios or Acheloos and it is not situated close to large population settlements. It faces pressures though from farming and especially units that operate in a micro-level and in the agricultural sector’s units that follow non-intensive practices. Louros’ lands that are dominated by season-long crops and grazing. In general, the water quality of the Louros catchment was in good environmental status and under any definition of environmental standards. Thus, there was not a need to apply a catchment-wide agri-environmental programme, or at least, this was not justified on the basis of non-point source pollution from agricultural activity. The following two chapters will focus on the economic approach of measuring costs and benefits (and the beneficiaries and bearers) for the implementation of the WFD in the Louros catchment.
References
Major National Studies including the Louros catchment Under the Water Framework Directive Louros has not as yet a management plan but only some, rather dated, studies. These include the study by the Institute of Geology and Mineral Exploration on the aquifers of Epirus (Nikolaou, 2001), the Specific Environmental Study of Amvrakikos (the then Ministry of Environment, Planning and Public Works, 1997) and the Pilot Study for the Management of Water Resources of the Water District of Epirus (Ministry of Industry, Energy and Technology, 1993). To these studies one should add the study of the National Technical University of Athens (NTUA) that was carried out to support the actions establishing water regions within the Water Framework Directive in 2008.
Ministry of Environment, Planning and Public Works. (1997). Specific Environmental Study of Amvrakikos. Athens.
Ministry of Industry, Energy and Technology. (1993). Pilot Study for the Management of Water Resources of the Water District of Epirus. Athens.
Nikolaou, E. (2001). Study of the Aquifers of Epirus, Volume II. Institute of Geology and Mineral Extraction, Preveza.
Koutsoyiannis, D., Andreadakis, A. Mavrodimou, A. et al. 2008. National Programme for the Management and Protection of Water Resources, Pages 748, Section of Water Resources and Environment – National Technical University, February 2008, Athens.
Other scientific references:
Albanis, T., Danis, Th., and Hela, D. (1995). Transportation of pesticides in estuaries of Louros and Arachthos rives (Amvrakikos Gulf, N.W. Greece), The Science of the Total Environment, 171, 85-93.
Albanis, T. and Hela, D. (1998). Pesticide concentrations in Louros river and their fluxes into the marine environment, International Journal of Environmental Analytical Chemistry, 70(1-4), 105-120.
Konstantinou, I., Hela, D. and Albanis, T. (2006). The status of pesticide pollution in surface waters (rivers and lakes) of Greece. Part I. Review on occurence and levels, Environmental Pollution, 141, 555-570.
Kotti, M., Vlessidis, A., Thanasoulias, N. and Evmiridis, N. (2005). Assessment of River Watre Quality in Northwestern Greece, Water Resources Management, 19, 77-94.
Tsangaris, C., Cotou, E., Papathanassiou, E. and Nicolaidou, A. (2009). Assessment of contaminant impacts in a semi-enclosed estuary (Amvrakikos Gulf, NW Greece): Bioenergetics and biochemical biomarkers in mussels, Environmental Monitoring and Assessment, 161, 259-269.
Chapter Four: Measuring costs in the context of the Water Framework Directive: The case of Louros, Greece
The current chapter is based upon “Deliverable 6.12: Cost effectiveness analysis report for the Louros catchment including analysis of disproportionality” which was part of the “Adaptive strategies to Mitigate the Impacts of Climate Change on European Freshwater Ecosystems Collaborative Project” (large-scale integrating project) (FP7),Grant Agreement 244121, duration: February 1st, 2010 – January 31st, 2014 Work package: 6 with contributors: Dr. Demetris Psaltopoulos, Dr. Dimitris Skuras and Emmanouil Tyllianakis
1. Introduction
In the previous chapters the targets and the requirements of the Water Framework Directive were presented thoroughly and the Greek case of Louros was analysed. The directive lays out specific guidelines on the means of managing water resources (creation of River Management Plans), the targets of the measures taken (reach Good Ecological Status by 2015, employ the most cost-effective measures and identify the list of cost-bearers and beneficiaries) apart from assessing the status of the water body in terms of concentration of chemicals, pollution sources and driving forces of pressures and pollution. Although the target for reaching GES by 2015 was not met by the member states and it will be reconsidered in 2017 (Gonzalez Davila O., et al. 2016), the targets and measures still stand as before.
The Louros catchment as it was put forth in the previous chapter is a catchment that did not experience large pollutant activities and thus the water status is good but it serves as a great example of firstly how “hands-on” economic instruments affect the environment and livelihoods. This chapter aims to present the various costs from various sets of measures from implementing measures to tackle future problems in the Louros catchment that are the product of climate change, in order for the catchment to continue to abide by the mandates of the WFD.
The various agricultural practices and products engulf various costs and under different cultivation practices produce different costs for the local economy. Depending on the agri- environmental schemes we devise Marginal Abatement Cost Curves for different measures and cultivations. In addition, the primary cost bearers in the catchment area are identified, following stakeholder meetings.
2. Alternative agri-environmental schemes and cost elements
Agri-environmental measures under consideration were defined based on the agri- environmental programme designed for the plains of Arta and Preveza (Common Ministerial Decision 50981/2308) which have been characterised as Nitrate Vulnerable Zones. Since farmers are identified as the key stakeholder at the critical sub-catchments, their contribution to the final choice of measures was also significant (Skuras et al., 2012). In this context, the relevant measures generally focus on changes in agricultural practices for main cultivations in the sub-catchments, i.e. maize, cotton, medic and citrus tree plantations, involving reductions in the use of N/P fertilizers, setting aside irrigated land, crop rotation and uncultivated strips or belts (buffer zones). Combinations of the aforementioned measures constitute the mitigation schemes considered to potentially achieve compliance with WFD requirements under the base-line conditions and alternative climate change scenarios.
In the case of maize, cotton and clover, four alternative agri-environmental schemes are considered:
1. Set aside 25% of irrigated land, reduce N fertilizers (including manure) by 25% to the rest of the 75% of irrigated land. 2. Crop rotation with non-irrigated nitrogen-fixing legumes on 20% of irrigated area; no fertilization to the 20% of land that was rotated with nitrogen-fixing legumes during the previous year for cotton and half of the fertilization for maize; leave 5% of land (strips or belts) uncultivated; reduce N fertilizer (including manure) by 25% to the rest of the 55% of irrigated land. 3. Set aside 30% of irrigated land, reduce N fertilizers (including manure) by 30% to the rest of the 70% of irrigated land. 4. Crop rotation with non-irrigated nitrogen-fixing legumes on 25% of irrigated area, no fertilization to the 25% of land that was rotated with nitrogen-fixing legumes during the previous year for cotton and half of the fertilization for maize, leave 5% of land (strips or belts) uncultivated, reduce N fertilizer (including manure) by 30% to the rest of the 45% of irrigated land.
In the case of citrus two mitigation measures are considered:
1. Reduce N fertilizers (including manure) by 25% to the whole of the plantation. 2. Reduce N fertilizers (including manure) by 30% to the whole of the plantation.
For each one of the cultivations and for each of the mitigation measures under consideration the average cost per hectare was estimated. The chosen cost model is based on Standard Gross Margins (SGMs) for each of the cultivations under consideration. The standard Gross Margin (SGM) of a cultivation is defined as the value of output from one hectare less the cost of variable inputs required to produce that output. We assume that agri-environmental policies induce only temporary changes to farm practices and thus, the constant cost of fixed assets such as capital, land, and buildings is not affected and should not enter the cost calculations.
SGMs from the RICA/FADN database are readily available for all Greek regions up to 2004 and provide a ground for estimating foregone income due to either no cultivation or reduced cultivation. The initial yields and the yields from reduced cultivation come from scientific evidence and the literature review presented in part 2.1.1 of this report and cross validated with stakeholders. Stakeholders also cross-referenced our RICA/FADN estimates or completed these estimates where needed.
In more detail, the cost elements examined in this work include:
2.1. Income forgone Income is forgone either by no cultivation or by lower yields due to lower fertilization. In order to calculate the income forgone per hectare for maize, cotton and medic for no- cultivation and for all cultivations for lower yields, gross margins per hectare are derived for the region of West of Greece for 2004 according to RICA/FADN estimations. As concerns reduced yields, a consensus between scientific agronomic data and stakeholders’ information assumed a reduction of N and P fertilization by 25% or 30% to reduce yields by 21% and 25% respectively for maize, cotton and medic and 20% and 25% for citrus tree plantations.
2.2. Additional costs Additional costs include the costs incurred from implementing the proposed land management activity that is required for alternative measures. This includes the cultivation of nitrogen fixing legumes that incurs cost but no benefit because the plants are incorporated into the soil and are not used as a fodder. The cost of nitrogen fixing legumes per hectare is also derived from the average 2003-2005 RICA/FADN estimations.
2.3. Cost forgone This is a benefit from reduced use of fertilization or from no fertilization on set aside land or on land that was cultivated with nitrogen fixing legumes during the previous year. The cost of fertilization per hectare is derived from the average 2003-2005 RICA/FADN estimations. We assume that the cost for plant protection remains the same and is not affected by reduced fertilization. The cost of energy for fertilization application practically remains the same as fertilizer is applied (i.e., the tractor works) but at lower quantities. Finally, the cost of irrigation remains the same as it depends on irrigation properties paid to the Local Organization for Land Reclamation.
2.4. Transaction Costs Transaction costs from the farmer’s perspective relate to drawing the Environmental Management Plan that is estimated at market prices and the transaction cost related to the work days lost and the payment for required legal documents. The annual equivalent of a seven-year plan for an average farm of 6.5 hectares is estimated at 30 euros including soil analyses and consultancy. The annual equivalent of the shadow price of work days forgone is estimated at 2 euros bringing the total transaction cost for an average farm to 32 euros per year (224 euros for a seven-year plan and an average 6.5 hectares farm).
Tables 15 to 18 present the detailed cost estimates for each mitigation measure and cultivation and the estimated SGMs per hectare, cost of fertilization per hectare and the reduction in yields incurred by reduced fertilization.
The cost of applying the alternative mitigation measures depends on the cultivation’s SGM. Citrus fruits and cotton have the highest SGM among the four cultivations but for citrus fruits no set aside practices can be applied. Thus, cotton shows the heist costs. However, this may be misleading when the cost is considered at the catchment’s level because cotton is cultivated on 337 ha out of the 9,745 ha occupied by the four cultivations together. Between two mitigation measures that opt at the same level of fertilizer reduction, the one that allows the set aside land to be cultivated with nitrogen fixing legumes is always cheaper because during the next cultivation period, the area under nitrogen fixing legumes provides full yields with half of its standard fertilization requirements. For example, the cost of applying mitigation measure 1 (25% set aside – 25% reduction in fertilization on the rest 75% of the land) for cotton is almost as high as 790 euros per hectare. Mitigation measure 2 which requires that the 25% set aside is broken down into 5% set aside and 20% under nitrogen fixing legumes has lower cost because the 25% reduction in fertilization is not applied to the rest of 75% of the land but to the rest 55% of the land. This is due to the fact that nitrogen fixing legumes cultivated in the past cultivation period need only half of the fertilization to provide full yields.
Finally, we may observe that the cost of mitigation measure 4 which implies 30% reductions, is of almost the same cost as mitigation measure 1 which implies reductions of 25% but does not allow the utilization of the set aside land by nitrogen fixing legumes. Table 1. Average cost estimates per hectare and cultivation for Mitigation Measure 1
Table 2. Average cost estimates per hectare and cultivation for Mitigation Measure 2
Table 3. Average cost estimates per hectare and cultivation for Mitigation Measure 3
Table 4. Average cost estimates per hectare and cultivation for Mitigation Measure 4 The cost per hectare and per cultivation is used to estimate the total cost of the measures at sub- catchment and catchment wide levels. Figure 1 shows the total cost for the whole sub-catchment, which, in the case of mitigation measure 3, the most expensive measure, is over 6 million Euros per year.
Figure 1. The total cost of the proposed mitigation measures.
Figure 2. The average cost per hectare for each mitigation measure. The average cost per hectare assumes that, each cultivation, participates in one hectare by the same share it participates in the watershed. In other words, the average hectare of land is made up by 4% of cotton, 34% of maize, 41% of medic and 21% of citrus fruit. As is shown in Figure 2, mitigation measure 2 is the less costly at 495 euros, followed by mitigation measure 1 at 542 euros per year.
The average cost per Kg of nitrogen or phosphorous not applied to the watershed is shown in figures 8 and 9. This is estimated again by assuming that cultivations join each measure proportionally to their shares in the catchment. We see that mitigation measures 4 and 2 are the less costly as concerns reduction of nitrogen and phosphorous. The fact that mitigation measure 4 achieves an average cost of almost 7 euro per Kg of nitrogen not applied and an average of 9 Euros per Kg of phosphorous not applied shows that there are scale effects as concerns the level of reduced application. In other words, the higher the percentage of land in which no fertilizer is applied and the higher the reduction of fertilizers in the remaining cultivated land, the lower the per cost per kg of fertilizer that is not applied. This is more evident between mitigation measures 1 and 3 which are exactly the same mitigation measures with the only difference being the proportion of fertilizer not applied to the field.
Figure 3. The average cost per Kg of N reduced for each mitigation measure.
Figure 4. The average cost per Kg of P reduced for each mitigation measure.
3. Identification of cost bearers in Louros catchment
According to workshop participants, the list of costs and cost bearers of mitigation measures in Louros presented by the research team (Table 5) was quite comprehensive and perhaps too detailed. They argued that the farm sector and especially cotton, maize, clover and citrus farmers, which have already been facing significant competition, would be significantly affected by the chosen mitigation measures. The application of the chosen mitigation measures would lead to a significant decline in yields, production and farm incomes in these particular crop activities. Taking into account the decoupled nature of CAP Pillar 1 support and the recent cost-price squeeze that farmers are facing, it could lead to the abandonment of farming activity; in turn, such a development would have significantly negative economy-wide knock on effects. Farm production costs could also increase if integrated pest management is applied, but according to workshop participants, local farmers do not have the skills to engage in such a production technique. Also, the lack of a comprehensive and accurate soil map for the area, which would have facilitated the targeting of fertilizer reduction makes costs even higher.
Other types of cost included in the Patras list did not seem to be very high in the stakeholders’ agenda. Stakeholders agreed that there would be a cost to taxpayers if some of the chosen mitigation measures were supported through agri-environmental measures; however, they pointed out that this potential support is rather irrelevant, as it would be very far from being able to compensate yield losses. In other words, they pointed out that the shock on farm incomes would be much higher than agri-environmental measures support, and hence, interest by farmers to apply for such measures would be marginal especially for farmers in the cotton and maize cultivations. It was pointed out that “….most farmers will abandon…”. Also, it is worth mentioning that according to stakeholder opinion, the structural characteristics of local farmers (small holders, rather aged) and their very limited feasibility to borrow in order to invest in the modernization of their farms constitute additional factors which make abandonment very likely. Also, participants were far from convinced about the efficiency and competence of monitoring system, so they did not pay any notable attention to the possible increase of administration costs, which was included in the list.
Benefits and the respective beneficiaries will be analysed in the next chapters.
Table 5. List of cost bearers and beneficiaries in the Louros catchment
3.1. Estimation of Marginal Abatement Cost curves for the four available scenarios In this work the INCA-P model allows us to simulate the P concentrations in water under the baseline and the alternative mitigation measures. Thus, we have the chance to re(?)-consider carefully the meaning of abatement. In conventional climate change studies abatement is related to the reduction (or moderation) of tonnes of CO2. However, climate change policy targets set a one-to-one relationship between the act of abatement (reduction in CO2 emissions) and target levels for the environment. In nitrate water pollution however, the act of abatement usually refers to kg of nitrate fertilization that are abated (per farm or per hectare of land) but the environmental policy (e.g. the nitrates directive) sets standards that refer to nitrate concentration in water (in mg/l or μg/l for nitrates and phosphorous correspondingly). Thus, there is not an obvious one-to-one (or even linear) relationship between the act of abatement and nutrient concentration in water. This is due to the fact that either nitrates or phosphorous added to the environment from fertilization or other agricultural activities (e.g. animal waste, plant residues, etc.) are just added to the existing stock of nutrients and to the nutrients added by physical processes. What ends up to water courses is the result of a very complex process that depends on climate (especially precipitation and temperature), physical conditions (slopes, elevation), soil conditions (especially soil chemistry), hydrography (especially the density of watercourses, and manmade interventions such as the irrigation network and drainage ditches), etc. Figure 5 depicts this complex process as it is captured by the INCA-P model used in this work.
We can consider mitigation measures 1 and 3 and 2 and 4 as two pairs of different technology measures (simple set aside or set aside cultivated with nitrogen fixing legumes). Reductions in phosphorous applied to the field do not directly and linearly correspond to abated phosphorous. Thus, we construct two kinds of abatement cost curves. The first depicts marginal abatement cost as euros per hectare for applying a specific mitigation measure related to kg of nitrogen not applied (abated) for the whole Louros catchment. The vertical (y axis) represents the total cost of the mitigation measure divided by the surface of land resulting to a measure of cost as euro per hectare. The horizontal axis (x axis) represents the abatement measured in kg of fertilizer not applied divided by the surface of land resulting to an abatement measure of Kg of abated nitrogen per hectare of land. This is depicted in Figure 6. The second estimates the unit abatement cost as the cost of reducing the concentration of phosphorous (TP) by 1 μg/l measured in euros for the whole Louros catchment. The environmental benefit is measured in the quantity of TP that is progressively reduced resulting from the adoption of mitigation measures. Figure 7 shows the MACCs for mitigation measures 1 and 3 (blue line) which assume a simple set aside and reduction of fertilization to the rest of the cultivated land at progressively higher levels of 25% (mitigation measure 1)and 30% (mitigation measure 2). The vertical axis (y axis) measures the cost of reducing TP concentration in water (abating TP) by 1 μg/l. It is derived by dividing the total cost of the measure by the reduction in TP concentration measured in μg/l achieved by applying the measure. For example, mitigation measure 1 for the whole Louros catchment is estimated to cost 5.3 million euros and INCA-P modelling simulated that if the measure is applied we have a reduction in TP concentration of about 11.7 μg/l (i.e., from the 48.2 μg/l at baseline down to 36.5 μg/l if the measure is applied). Thus the x- axis measurement for mitigation measure 1 is 451,384 euros per μg/l of reduced TP concentration. The horizontal axis measures the reduced concentration from the baseline concentration of 48.2 μg/l of TP. Figure 7 shows a crude linear approximation of the MACC. The area under the MACC represents the total abatement cost in both figures 7 and 8.
Figure 5. The INCA-P Model nutrient flows and process controls (after Crossman et al, 2013)
Figure 6. Marginal Abatement Cost Curves based on abated phosphorous for simple set aside mitigation measures (1 and 3) and nitrogen fixing legumes mitigation measures (2 and 4).
Figure 7. Marginal Abatement Cost Curves based on TP concentration for simple set aside mitigation measures (1 and 3) and nitrogen fixing legumes mitigation measures (2 and 4)
From figures 7 and 8, and despite the fact that we are not able to know the exact functional form of the MAC, we may risk drawing two conclusions: . First, if we consider abatement in the sense of phosphorous fertilization that is not applied to the field (Figure 6), the MACC for the measures allowing set aside land to be cultivated with nitrogen fixing legumes are under the MACC for the measures with set aside land only (i.e., more efficient) and to the right. The latter means that mitigation measures allowing for nitrogen fixing legumes are not only less costly but achieve more benefits (higher abatement). However, if we look at figure 20 which measures abatement in terms of TP concentrations, this is not the case. Mitigation measures allowing for nitrogen fixing legumes to be cultivated on set aside land achieve exactly the same abatement levels as the mitigation measures that involve only set aside land. This is due to the fact that the cultivation with nitrogen fixing legumes and their incorporation into the soil adds phosphorous through plant residues. Still, the measures allowing for nitrogen fixing legumes are more efficient.
. Second, probably due to the fact that we are at low pollution levels, marginal cost is increasing indicating that economies of scale in controlling phosphorous pollution have been exhausted and the proposed phosphorous pollution control measures operate as diseconomies
4. Scenario analysis: Incorporating climate change 4.1. Climate change in Greece and the Louros catchment In 2009, the Bank of Greece set up the “Climate Change Impacts Study Committee” with the mandate to draft a report presenting the foreseen climatic and environmental changes and estimating the cost of these changes on the Greek economy as well as the cost of the adjustment measures. The report was presented on 1st June 2011 and is the most complete study on climate change impacts in Greece and in line with the IPCC story lines scenarios.
In brief, the overall story lines of the four major scenarios refer to:
A1 storyline and scenario family: a future world of very rapid economic growth, global population that peaks in mid-century and declines thereafter, and rapid introduction of new and more efficient technologies.
A2 storyline and scenario family: a very heterogeneous world with continuously increasing global population and regionally oriented economic growth that is more fragmented and slower than in other storylines.
B1 storyline and scenario family: a convergent world with the same global population as in the A1 storyline but with rapid changes in economic structures toward a service and information economy, with reductions in material intensity, and the introduction of clean and resource-efficient technologies. B2 storyline and scenario family: a world in which the emphasis is on local solutions to economic, social, and environmental sustainability, with continuously increasing population (lower than A2) and intermediate economic development.
The Bank of Greece study for Greece considered four scenarios according to the 3rd UN report on climate change (Nakicenovic et al., 2010):
A2: the IPCC story line with rapid concentration of CO2 reaching 850 ppm in 2100
A1B: the IPCC story line with intense concentration of CO2 reaching 720 ppm in 2100
B2: the IPCC story line with moderate but steady increase in the concentration of CO2 reaching 620 ppm in 2100
B1: the IPCC story line with mild increase in the concentration of CO2 reaching 550 ppm in 2100
4.2. Climate change scenarios and drivers For the A2, B2 and A1B scenarios simulations estimated the likely impacts on six climatic parameters. For the B1 scenario and due to data constraints only the likely impacts on temperature were considered. More specifically, the impacts for the decade 2091-2100 are forecasted as follows:
Temperature:
For A2, it is expected an average 4.5 oC rise that will reach 5.4oC during summer and may be up to 6oC and 7oC in mainland Greece. For A1B the corresponding average rise is estimated to be 3.5oC with 4.5-5oC rise during summer and in mainland Greece. For B2 the rise will be around 3.2oC while for B2 the corresponding rise is expected even milder at around 2.4oC. Especially for the climatic region of the Louros watershed (region Western Greece – WG), and for the A2 scenario, the temperature is expected to rise to 18.24±1.87 in 2091-2100 from an average historic record of 12.94±1.52 in 1961- 1990. For the B2 scenario the corresponding changes will be 16.54±1.53.
Sea level rise:
Sea level rise is closely related to temperature rise. According to the conservative B1 scenario, air temperature will rise from 1.1 to 2.9 and according to the most unfavourable scenario (A1F1) the temperature will raise by 2.4-6.4. According to B1 scenario, sea level is expected to rise by 0.18-0.38 m and by 0.26-0.59 under the A1F1 scenario. Others however, (Rahmstorf, 2007) argue that according to the empirical relation between temperature and sea level rise with sea level rising by 10-30 cm for 1 degree Celsius rise in temperature, the impacts may be stronger. According to the SRES scenarios (IPCC (2007) for temperature rising between 1.4 and 5.8 degrees, sea level may rise between 0.5 and 1.4 meters by 2100. More adverse forecasts are provided by Pfefferet al. (2008) with expected sea level rise between 0.8 and 2 meters.
For the Louros watershed and especially for the Amvrakikos Bay coastline, sea level rise may have adverse impacts if it is related to reduced rainfall (see below) and decreased sediment transfer. For example, Poulos et al (2009a and 2009b) have estimated that an average yearly sea level rise of 4.3 mm can be reduced to 3,5 mm if we subtract a tectonic rise of about 0.8 mm per year due to sediment deposition especially in deltas. Taking into account the soft nature of alluvium depositions and the fact that sediment transfer and deposition has been significantly reduced due to the water dam, sea level rise is expected to cause adverse impacts on the Nature 2000 site of the Amvrakikos Bay area and especially in the Louros delta site.
Rainfall:
The reduction in forecasted precipitation (rainfall) is very significant in the cases of A2 and A1B scenarios and milder for B2 (no estimates provided for B1). For A2 this reduction will reach a yearly average of 17% with a peak during summer of 47%. However, this is not a very significant number in absolute terms because Greece does not receive important amounts of precipitation during summer.
Especially for the climatic region of the Louros watershed (region Western Greece – WG), and for the A2 scenario, the rainfall is expected to decrease by an average of more than 20% in 2071-2100 from the historic record in 1961-1990. For the B2 scenario the only significant change is expected for the autumn rainfall by almost 8%.
Relative humidity at 2 meters above earth surface:
Relative humidity is expected to decrease for the three scenarios (A2, A1B and B2) with changes in B2 been the mildest. For A2 the average yearly change will be reduced by 4.5% by 2091-2100. For A1B and B2 changes will be smaller.
Especially for the climatic region of the Louros watershed (region Western Greece – WG), the decrease in the average yearly relative humidity will be almost 10% (20% during summer, 3.5% during winter and 10% during spring). For the A1B scenario changes will be smaller while for the B2 scenario changes are expected to be non-significant.
Overcast:
For the A2 scenario is expected an average reduction of about 16% while for scenarios A1B and B2 this is expected to be 8%. There are not significant spatial variations in overcast and so the national forecasted figures hold for the Louros watershed.
Incoming short wave radiation (sun radiation): Incoming radiation will present a small increase that is, up to certain degree, related to the reduction in overcast. For the A2 scenario it is expected that incoming radiation will increase by 4.5 W per square meter in 2091-2100.
Especially for the climatic region of the Louros watershed (region Western Greece – WG), the increase in radiation may be as high as 8-10 W per square meter under the A2 scenario, 2-4.5 under the A1B scenario and 2.3 under the B2 scenario.
Wind velocity:
In general, no significant changes are expected. However, some seasonal variation may be significant. Especially for Western Greece (where Louros is located) it is expected a7% decrease in winter under the A2 scenario and 5% under the A1B scenario.
In our study the modelling work used meteorological data from three different climate models to define the meterological time series for the scenario period: . KNMI-RACMO2-ECHAM5 (abb. KNMI), . SMHIRCA-BCM (abb. SMHI) and . HadRM3-HadCM3Q (abb. Hadley)
4.3. Land use and climate change
The “Climate Change Impacts Study Committee” of the Bank of Greece has devised a table of likely impacts of the various scenarios on selected agricultural production for the Western Greece as shown below in Figure 9.
Figure 9. Likely Climate Change Impacts on Western Greece Agriculture
The more adverse effects are expected under the A2 scenario primarily for wheat and secondly for maize. For wheat, adverse effects also are expected under the A1B scenario. Under the B2 scenario production is expected to rise for all cultivations.
Besides the results derived from the “Climate Change Impacts Study Committee”, two more research projects have dealt with climate change impacts on wider areas, including Greece. The PESETA project, and for the most adverse climatic change scenario based on a 3.9 0C rise in temperature, estimates the likely impacts of climate change on european agriculture including Greece. The whole of Greece is under the most adverse zone with decreases in agricultural production of over 15%. However, under a scenario with a slightly lower rise in temperature (3.10C), in the western parts of Greece, including the Louros watershed, changes are expected to be slightly positive or no significant.
The second study conducted by Giannakopoulos et al (2009), has assessed the impacts of climate change on cultivation in Mediterranean basin using the HadCM3 model and for the Α2 and Β2 scenarios implying a global rise in temperature by 2⁰C for the period 2031‐2060. According to this study and for the North Mediterranean region, the joint implications of temperature rise and increased CO2 concentration the expected changes range from ‐9,33% for tuber crops (Α2) to +12,49% for cereals (Β2).
The major driving force behind land use in agriculture is the price received by the farmer and the way production costs are formulated and especially the price of oil affecting the price of fertilizers and the cost of irrigation through energy consumption. On top of these, and especially in Greek agriculture, production subsidies (now decoupled payments) also play a major role.
In general, areas planted follow market price fluctuations with a one or two year average lags that are used to form price expectations. Especially for cereals, international markets (and specifically shortages) have been very turbulent in recent years. Cereal shortages have appeared due to extreme weather phenomena in major cereal producing countries such as Russia and Australia that reduced yields and total production coupled by increased demand for cereals as foodstuff (especially by growing economies such as China). Figure 8 shows, for the typical case of cotton, that areas planted follow in general the market prices received by farmers with a time lag.
Policy has two types of impacts on agricultural production. Firstly, direct impacts coming from agricultural and rural development policies. Secondly, indirect impacts coming from non-agricultural policies but affecting agricultural production. Agricultural policies affect the cost of production through area subsidies. However, the general trend is that the EU will gradually reduce subsidies of any form moving towards market oriented agriculture. At the same time environmental subsidies and soil conservation restrictions will increase in Greek agriculture (Barbayiannis et al., 2011) and this will reduce production and increase production costs.
From the non-agricultural policies the most influential on agriculture is energy policy. Green energy policies have affected Greek agriculture in two ways. First, they have increased the price of many cultivations used for the production of biofuels such as cotton, maize and sunflower. Secondly the increase of areas under biomass production has increased and especially on soils that are of marginal importance for agriculture (previously cultivated with fodder plants or dry wheat). This policy has kept high prices for cotton and maize in the past.
Secondly, energy policy affects land available for cultivation through the wind and solar energy utilization programmes. Currently, the area withdrawn under these programmes is negligible, however, under a B1 scenario with intense introduction of clean and resource efficient technologies, the area withdrawn from production may be considerable, especially in Greece where solar power utilization is efficient.
Figure 8. Area cultivated by cotton and weighted average price, 2002-2009 in Greece.
The aforementioned discussion shows that in the case of a global, economy driven scenario (A1) prices are likely to increase due to increased food demand and areas under cultivation of cereals will also increase. At the same time, the introduction of efficient technologies may increase yields that will absorb some of the area increase. Thus, under scenario A1 a moderate 5% increase in cultivated land is the highest we may expect. Forest land will decrease and grass land will increase due to higher fodder plant prices
Under A2 scenario the increase in land cultivated may be as high as 10% due to the absence efficient technologies that increase yields. Forest land will decrease and grass land will increase due to higher fodder plant prices.
Under the B1 scenario, prices are likely to increase due to increased food demand and areas under cultivation of cereals will also increase. However, the introduction of efficient and clean energy technologies such as solar power and wind energy installations will withdraw land from agriculture or will absorb land abandonment at marginal soils. Thus, the increase in cultivated land will not be over 5%. Forest land is likely to decrease and grass land will increase due to higher fodder plant prices.
Under the B2 scenario, with stable food demand and stable prices areas cultivated will be likely reduced due to more ecologically sustainable management of resources, soil and water conservation policies and land abandonment. Forest land will increase and grass land will remain stable but may switch to more extensive forms of management.
The aforementioned analysis was translated into a land use change model induced by IPCC storylines, i.e., combined effects of climate change, technology, trade and consumption-production patterns. These expected land use changes are shown in Table 6. Table 6: Expected land use changes under IPCC scenarios for Louros
More specifically, we assumed a change in cotton, maize and medic of 15%, 10%, 10% and 8% for A2, A1, B2 and B1 correspondingly and all other cultivations to remain unchanged so as to have the changes shown in Table 6 for arable land in general. For citrus fruit we assumed a uniform increase of 5% for all storylines, with olive groves and vineyards having changes that will result to overall changes for permanent crops as in Table 6. Forests areas contract in order to accommodate increases in arable land, permanent crops and grassland.
4.4. Nutrient application and costs under land use change scenarios In this work it is assumed that the long-run application of nutrients will be affected only by climate change induced land use changes and not by climate change per se, i.e., the ability of plants to use more nitrogen, etc. In accordance we assume the same response to mitigation measures. Furthermore, it is assumed that the basic elements of cost and benefit estimations will remain the same. Thus, the cost per hectare and the Kg of fertilizer per hectare not applied for each mitigation measure remain the same. However, due to land changes, the total amount of applied fertilizers increases and the total cost of the mitigation measures also increases.
Tables 7 and 8 show the applied nitrogen and phosphorous fertilizers correspondingly under the no land use change hypothesis and the four land use changes induced by storylines A2, A1, B2 and B1. It is important to note that, under the A2 storyline, the total amount of nitrogen fertilization reaches the 2,000 tonnes (Table 6) while phosphorous fertilization is about 1,324 tonnes (Table 6).
The total cost and the reductions in applied fertilizers are shown in tables 9 and 10. Table 6: Application of Nitrogen fertilizers (kg) to the whole Louros catchment.
Table 8: Application of Phosphorous fertilizers (Kg) to the whole Louros catchment
Table 9: The cost of mitigation measures in Euros for the Louros catchment under four land use change storylines.
Table 10: Reduced fertilizers in Kg for the Louros catchment under the four land use change storylines.
5. Scenario setting
The models were run for the scenario period of 2031-2060, using all possible combinations of the three climate models (i.e., the KNMI, SMHI and Hadley), and the four land use scenarios, plus a fifth land use scenario representing a baseline (i.e. no land use change). The no land use change scenario, however, is modelled under climate change. The total number of scenarios ran for each model was thus 15 (3 by 5).
Instead of running all possible combinations of climate models, land use scenarios and mitigation measures, only the “best- and worst-case” combinations of climate models and land use scenarios were considered. For nitrogen, the “best-case” was the combination of the KNMI climate model and the B1 land cover scenario and the “worst-case” was the combination of the Hadley climate model and the A2 land cover scenario. For phosphorus, the “best-case” was the combination of the Hadley climate model and the B1 land cover scenario and the “worst-case” was the combination of the SMHI climate model and the A2 land cover scenario. All four mitigation measures were then simulated for the baseline period (1981-2010), the best- case and the worst-case scenarios. In figures 10 and 11 we replicate the Marginal Abatement Costs captured by the cost of reducing the concentration of phosphorous (TP) by 1 μg/l measured in euros for the whole Louros catchment. The environmental benefit is measured in the quantity of TP that is progressively reduced resulting from the adoption of mitigation measures. Figure 10 shows the MACCs for mitigation measures 1 and 3 in the baseline (blue line), best case (red line) and worst case (green line) which assume a simple set aside and reduction of fertilization to the rest of the cultivated land at progressively higher levels of 25% (mitigation measure 1)and 30% (mitigation measure 3). The vertical axis (y axis) measures the cost of reducing TP concentration in water (abating TP) by 1 μg/l. It is derived by dividing the total cost of the measure by the reduction in TP concentration measured in μg/l achieved by applying the measure. For example, mitigation measure 1 for the whole Louros catchment is estimated to cost 5,281,198 euros and INCA-P modelling simulated that if the measure is applied we have a reduction in TP concentration of about 11.7 μg/l (i.e., from the 48.2 μg/l at baseline down to 36.5 μg/l if the measure is applied). Thus the x-axis measurement for mitigation measure 1 is 451,384 euros per μg/l of reduced TP concentration (5,281,198 euros divided by 11.7). Under the best case (Hadley – B1), the cost of measure 1 is 5.7 million euros (see Table 9) and the reduction of TP is 12.1 μg/l (i.e., from 48.8 to 36.7 μg/l). Thus the x-axis measurement for mitigation measure 1 under the best case is 468,572 euros per μg/l of reduced TP concentration. The horizontal axis measures the reduced concentration from the baseline concentration of 48.2 μg/l of TP. Figure 11 shows the same linear approximation of the MACC for mitigation measures 2 and 4. The area under the MACC represents the total abatement cost in both figures 10 and 11.
From figures 10 and 1, and despite the fact that we are not able to know the exact functional form of the MAC, we can draw two conclusions:
. First, the worst case for the mitigation measures 1 and 3 (i.e., simple set aside) clearly shifts the MACC to the left implying increased cost for the same abatement levels. The best case almost coincides with the baseline. For measures 2 and 4, both the best and worst case MACCs are shifted to the right, i.e. are even more efficient than the baseline, i.e., achieve higher abatement at the same cost. Thus, the strategy of cultivating nitrogen fixing legumes on set aside land is a cost proof strategy in any climate change scenario. . Second, probably due to the fact that we are at low pollution levels, marginal cost is increasing in all climate change scenarios indicating that economies of scale in controlling phosphorous pollution have been exhausted and the proposed phosphorous pollution control measures operate as diseconomies.
Figure 10. Marginal Abatement Cost Curves based on TP concentration for simple set aside mitigation measures (1 and 3).
Figure 11. Marginal Abatement Cost Curves based on TP concentration for nitrogen fixing set aside mitigation measures (2 and 4). 6. Discussion and conclusions
This chapter aimed at presenting the various costs and mitigation alternatives in a Greek catchment under various climate change scenarios. In particular, the costs of applying nitrogen and phosphorous reducing mitigation measures at the Louros catchment in Greece were considered as they attract the highest interest from the European Commission. Louros river creates a very important catchment from both an environmental and economic perspective. Louros estuaries form very important lagoons not only for Greece but for the whole of Europe due to its location on a main corridor for migratory species. Louros river, its estuaries and the Amvrakikos Gulf are protected under the Habitats-Natura 2000 Directive and the Ramsar Convention. Louros waters are used as drinking water for major municipal centres in the area, are used by aquaculture activities at the upland and industrial activities and fish farming at the lowlands while the Amvrakikos gulf has the potential to develop important tourism activities. Furthermore, amenities including eco or agro-tourism, bird watching and game can be further developed.
The three lagoons formed at the Louros estuaries are, for a long time, assumed to suffer from nitrification and agriculture was held responsible for the issue. Sporadic and fragmented monitoring activities undertaken by the Ministry of Rural development and Food or by independent researchers failed to reveal any significant, persistent and consistent nitrification or phosphates pollution issues. However, in certain monitoring points and in certain times, some pollution spikes were observed. Following public pressure, and not hard scientific evidence, the Ministry of Rural Development and Food declared the Arta-Preveza plain that includes the Louros and Arachthos watersheds a Nitrate Vulnerable Zone as early as 1999. In 2006, the Ministry of Rural Development and Food announced a nitrification control programme for the Arta-Preveza plain that would be financed by the 2007-2013 Common Support Framework for Greece and specifically by its “Baltatzis Programme” for Rural Development. Due to the 2008/09 recession and following serious budgetary constraints, the idea of applying the programme was practically abandoned. This incident offered us a unique opportunity of examining the likely effects of an expressed and fully articulated agri-environmental policy that was never applied in practice. In general, the water quality in Louros is at good status under all environmental thresholds and all alternative simulation
The mitigation measures proposed for this study imitate the measures originally offered by the agri-environmental programme for Louros with two differences. First, we included, besides the cultivations of cotton and maize, the cultivations of medic and citrus fruit that could formulate potentially serious polluting activities. Second, we assumed different levels of abatement under two different production processes (“technologies”). The first production process allows for reductions in fertilizer application by means of set-aside land, reduction in fertilization to the cultivated land and reduction in irrigation. The second production process allows for equal levels of reductions in fertilizers but demands 5% set aside margins, allows the set aside land to be rain fed cultivated by nitrogen fixing legumes, and reduces irrigation. Following the formulation of measures, and due to the lack of access to farm specific data, the cost of measures was estimated for a “typical – average” farm using aggregate FADN-RICA accounting and production data informed by a widespread consultation process with all stakeholders. The cost of reducing nitrogen and phosphorous was estimated as cost per Kg of the reduced nutrient and as cost per hectare of land under each mitigation measure and cultivation. These were aggregated to the whole Louros catchment taking into account the distribution of cultivations in the catchment.
In parallel, likely land use changes induced by climate change were drawn. Climate change for the Louros catchment and its likely impacts on plant productivity and land use was drawn from a complete and coherent assessment study of climate change effects carried out by the Central bank of Greece under the supervision of the Academy of Athens. Following the IPCC story lines, four climate change scenarios were devised. For each climate change scenario the future costs for applying the mitigation measures were re-estimated. Thus, we ended up with the cost of reducing nitrogen and phosphorous under climate change estimated as cost per Kg of the reduced nutrient and as cost per hectare of land under each mitigation measure and cultivation. These were again aggregated to the whole Louros catchment taking into account the distribution of cultivations in the catchment under climate change induced land use changes.
Modelling of nutrients and sediment transport was based on INCA-N and INCA-models. Modelling was based on geological, climatic, hydrograph and agricultural activity data for the Louros catchment and monitoring data for nutrients. Modelling provided a baseline (calibrated) estimate of nutrient concentrations without any mitigation measures or land use and climate change. This showed very clearly, and in accordance to monitoring data, that the water quality of the Louros catchment was in good environmental status and under any definition of environmental standards. Thus, there was not a need to apply a catchment wide agri-environmental programme, or at least, this was not justified on the basis of non- point source pollution from agricultural activity. Modelling also produced simulated concentrations for nutrients under the mitigation measures and without any land use and climate change. These simulations showed that the application of mitigation measures marginally improve the water quality. The baseline scenario (i.e., no mitigation measures) also was simulated for climate change induced land use changes. Due to the complexity and the burden of estimations for all alternative combinations of climate change and mitigation measures models, only the baseline, best and worst scenarios are presented. The best scenario follows the B1 IPCC storyline while the worst scenario follows the A2 storyline. Climate change, increases nutrient concentrations but not as much as it would have been expected from foreseen land use changes. This is due to the fact that climate change, and especially expected higher temperatures, lower precipitation and decrease runoff, reduce sediment and nutrient transport and increase the use of nutrients by plants. Thus, the quality of water at Louros remains, under any environmental threshold levels, at good status. When mitigation measures are applied to the climate change baseline scenario, the reduction in nutrient concentration is marginal for nitrogen and more significant (but still low) for phosphorous.
A major issue is defining the abatement cost. The nitrates directive defines environmental targets as the concentration of nitrogen in water with a paramount target of 50 mg/l. Environmental chemical standards for the WFD application in Greece also define good status according to the concentration of nutrients and other chemical inorganic and organic substances or certain physical properties (e.g., pH and conductivity). In most environmental studies, abatement efforts have a one-to-one or a linear relationship to environmental standards. In other words, efforts to reduce CO2 emissions (abatement process) can be directly translated to efforts for achieving environmental standards because environmental targets are defined on the same scale, i.e., reduction of emissions measured in tonnes. Reduction of noise by using appropriate controls again is measured in reduction of dB that is directly translated to environmental standards that are defined as levels of dB. However, when we deal with non-point source pollution from transported nutrients, the relation between abatement and environmental standards is not straight forward. To this end, some researchers have used “leaching functions” to link fertilizer application to amount of nutrients leaching from the filed into the watercourse. This approach, however, misses the amount of nutrient lost with water transportation (e.g., deposited in the waterbed) and assumes a homogenous leaching behaviour. In this work, we made use of the INCA-N and INCA-P simulations to relate, empirically, fertilizer application (current and under mitigation measures) to average yearly nutrient concentrations in water. Thus, we defined abatement cost in its conventional form as euro per Kg of substance (N or P) not applied and as euro per mg of N/l and μg of P/l reduced in the water concentration. The constructed abatement cost curves relate abatement (cost per unit reduced in concentration) to environmental status (concentration). Marginal abatement cost curves are increasing, mitigation measures which allow for the cultivation of the set aside land with nitrogen fixing legumes are less costly for the same abatement levels or abate more at the same cost levels. However, when abatement is measured by the concentration of the nutrient, both “technologies” abate the same but the cultivation with nitrogen fixing legumes is less costly.
For the least cost solution for phosphorous (mitigation measure 2), abatement cost is at just over 42 euros per hectare abating 54 Kg of phosphorous (not of fertilizer) per hectare. Furthermore, watershed wide, for the reduction of Total Phosphorous concentration in water by one μg/l to the (very low) level of 12 μg/l, it costs just above 412 thousand euros. Thus, the total cost of applying the most cost effective mitigation measure 2 is extremely high at almost 4.8 million euros. When climate and land use changes are simulated, again the most cost effective solution remains the mitigation measure that allows for nitrogen fixing legumes to be cultivated on set aside land. The Marginal Abatement Cost Curves for either the worst or the best climate change scenarios are to the right of the Marginal Abatement Cost Curve for the baseline scenario when mitigation measures that allow for nitrogen fixing legumes are considered. This implies that at the same cost, both under the worst and the best scenario, abatement will be higher. Thus, the “technology” of cultivating nitrogen fixing legumes on set aside land ensures that, under any climate change scenario, abatement will be higher at any chosen cost level.
References
Barbayiannis, N., Panayotopoulos, K.,Psaltopoulos, D. and Skuras, D.2011. The influence of policy on soil conservation: A case study from Greece. Land Degradation & Development 22(1), 47-57.
Giannakopoulos, C., Le Sager, P., Bindi, M., Moriondo, M., Kostopoulou, E., & Goodess, C. M. (2009). Climatic changes and associated impacts in the Mediterranean resulting from a 2 C global warming. Global and Planetary Change, 68(3), 209-224.
Dávila, O. G., Koundouri, P., Pantelidis, T., & Papandreou, A. (2016). Do agents' characteristics affect their valuation of ‘common pool’resources? A full-preference ranking analysis for the value of sustainable river basin management. Science of The Total Environment.
Nakićenović, N., Alcamo, J., Davis, G., De Vries, B., Fenhann, J., Gaffin, S., ... & Kram, T. (2000). IPCC special report on emissions scenarios (SRES). Working Group III, Intergovernmental Panel on Climate Change(IPCC). Cambridge University Press, Cambridge (http://www. grida. no/climate/ipcc/emission/index. htm).
Poulos, Serafim E., George Ghionis, and Hampik Maroukian. "Sea-level rise trends in the Attico–Cycladic region (Aegean Sea) during the last 5000 years." Geomorphology 107.1 (2009): 10-17.
Poulos, Serafim E., George Ghionis, and Hampik Maroukian. "The consequences of a future eustatic sea-level rise on the deltaic coasts of Inner Thermaikos Gulf (Aegean Sea) and Kyparissiakos Gulf (Ionian Sea), Greece." Geomorphology 107.1 (2009): 18-24.
Rahmstorf, Stefan. "A semi-empirical approach to projecting future sea-level rise." Science 315.5810 (2007): 368-370.
Skuras, D., Kontolaimou, A., Psaltopoulos, D. (2012), Workshop Proceedings on Collaborative Scoping of Solutions, Louros Catchment, Greece, Deliverable 6.6. Refresh Research Project, Department of Economics, University of Patras.
Chapter Five: Measuring Benefits: The case of the Louros catchment, benefit transfer and an application of meta-analysis
This chapter is based on the “Cost effectiveness analysis report for the Louros catchment including analysis of disproportionality” for the FP7 REFRESH programme by Demetris Psaltopoulos, Dimitris Skuras and Emmanouil Tyllianakis (2014) and on the paper “The income elasticity of Willingness-to-Pay (WTP) revisited: A meta- analysis of studies for restoring Good Ecological Status (GES) of water bodies under the Water Framework Directive (WFD)” published in the Journal Of Environmental Management by Emmanouil Tyllianakis and Dimitris Skuras (November 2016).
1. Introduction
After assessing costs in the context of the WFD, benefits are assessed. For a more detailed view of the nature of benefits and the methodologies used to assess and monetize them go to chapter two of this thesis.
In general, benefits from improved water quality (reaching or maintaining a GES) would be summarized as De Nocker et al (2007) point:
a. avoided costs for treatment of drinking water; b. reduction of disposal costs for contaminated dredging material; c. more and better opportunities for informal recreation (walking, cycling) and water sports; and, d. improved health and living environments.
The magnitude of these benefits are location and case specific, depending on the type and magnitude of the improvements under WFD compared to the reference situation, the number of people affected, their income and their preferences. The current information indicates that no particular type of benefit seems to dominate others, and that total benefits may be substantial.
Ghermandi et al (2010) find that wetlands, both human-made and natural are the ones valued highest among all other ecosystems as their services are more tangible to the population and also because they impact human welfare directly through primary and auxiliary ecosystem services. These can be flood and disease control, storm buffering and water quality. According to the Millennium Ecosystem Assessment (2005), ecosystem services are classified in four categories, provisional, supporting, cultural and regulating. Depending on their classification, ecosystem services are described as either intermediate or final and the water supply is a final ecosystem service provided by the provisional services of water ecosystems such as wetlands, river basins or coastal environments. The value and implications of the quality of water is the focus of this chapter, as are the policy implications of measures taken to improve water quality and the effect they have on human welfare.
In this chapter the focus will be on identifying the beneficiaries of the Louros catchment and the nature of their benefits, the assessment of benefits to these beneficiaries and a sensitivity analysis is also performed to examine how pressure factors such as land use, climate change affect the benefits from achieving GES. Following that, we broaden our scope and view the overall impact of the WFD in the valuation literature on water quality benefits. We conduct a systematic review (meta-analysis) on the findings of similar studies on water quality on inland waters and try to estimate the effects of income on WTP for water quality improvements.
The results point to income being a good explanatory variable for WTP of water quality, that improved water quality is a policy benefiting more the poorer households than the rich ones and that using income estimates from official statistic sources may yield more consistent estimates, a fact that upgrades the role of benefit transfer and meta-analysis.
2. Identification of Beneficiaries in the Louros catchment
Though opinion on costs and cost bearers was quite unanimous, the same cannot be argued as far as benefits and beneficiaries are concerned. Some participants pointed out that few farmers would benefit from agri-environmental support, if of course some of the chosen measures were to be supported by the Greek RDP. On the other hand, there was a consensus on the benefits associated with the local fish population and hence, local fish farmers and fishermen. As nitrates increase plant vegetation and bind oxygen in underground water, a decrease in the application of fertilizer would lead to an increase in the fish population. Also, it was argued that environmental degradation in Louros has led to a decrease of islets and dykes and thus, to a decline in the surface used for fish farming. Thus, an improvement facilitated by amongst others, these specific measures, would improve prospects for fish farms. Finally, stakeholders argued that an increase in fish farming and fishing activity would lead to an increase in fish tourism in the area and ultimately, to significantly positive income and employment repercussions for the local economy.
Regarding benefits associated with recreation of the local population, participants argued that the river generates the prospects for such benefits (e.g. canoe, swimming, etc.), but these prospects are not really utilized. As far as tourism from urban areas is concerned, most stakeholders suggested that there is a perception (by urban visitors) that environmental conditions in the Louros area are very satisfactory; thus, the mitigation measures are not really expected to affect these specific tourism flows and hence, would not lead to benefits for local businesses associated with tourism. In this context, stakeholders emphasized the lack of a coherent tourism development action plan for the area. Indicatively, they indicated that institutional shortcomings in terms of tourism policy design has led into a very low diffusion of tourism benefits in the Louros villages, which lack the necessary infrastructure (e.g. accommodation and other facilities for tourism). As a result, the tourism flow in the area is mostly associated with day visitors and, as a result, very few extra jobs have been generated in the Louros area due to this activity.
A contrasting opinion on the above issue was expressed by the representatives of the Natura 2000 Amvrakikos Management Body. Representatives of this institution argued that the environmental degradation of the river is a source of disrepute for the wider area and has considerably negative effects for foreign tourism. They argued “that these environmental problems are well-known to foreign visitors and thus, any efforts to market ecotourism do not lead to the desired effects”. It is worth noting here that according to these stakeholders, this negative reputation is not solely attributed to the environmental repercussions of farm activity, but also (perhaps mostly) to illegal activities such as hunting and fishing. Also, both lack of funding (i.e. institutional constraints) and local attitudes were argued to be the reasons behind the inefficient monitoring of such activities.
Also, another positive effect with the chosen mitigation measures seems to be associated with an improvement in the market prospects of farm products produced in the area through environmentally friendly methods. Stakeholders argued that such an improvement could be important especially for citrus fruit. However, a vast restructuring of the marketing strategies and distribution channels was specified as a sine qua non condition for any positive market prospects for these products. It was also argued that the above-mentioned benefits associated with agricultural activity and fish farming could have important positive effects for the local economy and also sustain the rural fabric in the area, as fishermen and fish farmers can at the same time be pluriactive farmers and thus, be able to sustain their living in the Louros area.
Finally, stakeholders agreed that an improvement in water quality would increase water supply for the local population; regarding this issue it worth mentioning that there is a shortage of drinking water in the area, especially in the summer tourism season.
A list of the beneficiaries and cost bearers is presented in Table 1 below.
Table 1. Cost bearers and beneficiaries in the Louros catchment
3. Benefits assessment
Due to budgetary constraints, it was not possible to run a proper benefit assessment, thus, the use of the Benefit Transfer method was selected. During the scoping stage of the benefit transfer method, a number of studies were used in order to isolate the most relevant ones. The examined researches used stated preferences methods and were selected due to their similarities with the Louros catchments as concerns social, economic and administrative reasons. Specifically the most relevant works included the work of Koundouri (2011) for Cyprus, Birol et. al (2008; 2010) that refer again to Cyprus water bodies and others in the vicinity of the Mediterranean region like Martin-Ortega et. al (2009) for Spain, and Birol et. al. (2006) for Greece. Other recent works such as Kataria et al, (2009) for Denmark, were excluded due to the non-similarity of the Greek socio-economic context with the one in more developed European economies such as the Danish one.
The most relevant studies report Average WTP estimates per household in the area of 50 to 85 euros. For Spain, Martin-Ortega et al. (2009) found an Average WTP 81.2 € and 61.3€ for moving to an excellent and good environmental status correspondingly. The research was carried out for watershed and river. For Cyprus, Birol et al. (2008) found, for the Akrotiri lake an average WTP of 18,25 CYP (approximately 31,39 Euros) for moving to an excellent status. For Cyprus, Koundouri (2011) started with an environmental target of good status and estimated average WTP to move to this target from bad and medium environmental conditions. For Bad to Good she estimated an average WTP of 61.4 euros per household while for medium to good the corresponding WTP estimate is 40.93. Taking into account that Koundouri’s work is the most recent and most relevant to our case study and also the fact that the Louros water body is already at good environmental status, we will use the 40.93 euros per household as a first estimate.
Table 2. Number of households in the regional units of Arta and Preveza
Environmental benefits are assumed to accrue equally to all households, thus, the total benefits for one year are estimated to be 1,957,273 euros (by multiplying 47,820 households with an average WTP per household at 40.93 euros representing a move from medium to good environmental status). This is assumed to be a constant stream of benefits each year from 2010 to 2015 (six years). The present value of the benefits can be calculated by using an appropriate discount rate. Taking into account the long-run discount rates in Greece and the social time preference for environmental projects, a 1.5% discount rate was used as a starting point. This produced an aggregate benefit of 11,150,948 euros for the six years under consideration.
3.1. Sensitivity analysis of aggregate benefit estimates
For the Greek case study, the following three major sources of uncertainty deserve further investigation and a sensitivity analysis:
. the chosen benefit measure as this is related to the Louros catchment current status . the discounting rate . the effects of recession and austerity programmes on benefits
Furthermore, the Louros case, combines water quality and habitats. The quality of the water has a direct impact on the Natura 2000 site and thus, WTP for clean water also implies higher nature conservation and biodiversity. It is well known from the relevant literature that use values dominate total economic value for most environmental goods such as clean water, reduced noise or cleaner air, while, non use values dominate total economic value of actions concerning with biodiversity and habitats. Thus, we should bear in mind that, benefit transfer from similar water quality related studies may underestimate the benefits at Louros by overlooking the contribution of water quality to habitats and biodiversity. Thus, benefits may be adjusted upwards to reflect this situation that is particularly true for Louros.
The benefit measure
The chosen benefit measure refers to the average WTP per household found in Cyprus for moving from the medium to the good environmental status. One justified line of sensitivity analysis would be to search the benefits for higher average WTP estimates, i.e., a 61.4 euros as estimated for Cyprus and for moving from the bad to the medium environmental status or the 81.2 euros estimated for Spain and for moving to the excellent environmental status. Figure 1 shows the estimated aggregate benefits for the period 2010-2015 discounted with a rate of 1.5% for the three alternative measures. Due to the very low discount rate that practically leaves the present value of estimates unchanged, a doubling in the benefit estimate, i.e. from 40.93 to 81.20 euros almost doubles the net present value of estimates.
Figure 1. Aggregate benefits under alternative benefits
As it was stated above, the use of progressively higher benefits is justified by the fact that the chosen benefits estimates concern solely with water quality improvements while, in the Louros case, these directly contribute to improved habitats conditions and biodiversity.
Discount rates
Despite the fact that ECB’s long-term rates are approaching zero, and taking into account the economic situation in Greece and the social time preference to environmental projects we adopted a basic discount rate of 1.5%. Figure 2 shows the aggregate discounted benefit estimates of the basic scenario (average WTP per household at 40.93) with varying discount rates up to 3.5%. It is evident that the choice of the discount rates especially at such low levels, does not really affect the aggregate discounted benefits estimates. Even when doubling the discount rate from 1.5% to 3.0%, the decrease in the aggregate discounted benefits is almost 5%.
Recession and austerity programmes
One of the major effects of recession and of the consequent application of strict austerity programmes is the very abrupt loss of income both in terms of GDP as well as in terms of income per person. In Greece, between 2008 and 2012, almost 20% of the country’s GDP was lost, bringing the level of GDP back to 2002 levels (Figure 3). Furthermore, unemployment is rising to almost 30% by today. A recent research work attempted to examine the income effects on global WTP for biodiversity using 145 WTP estimates from 46 contingent valuation studies (Jacobsen and Hanley, 2009). The authors purposefully include studies from all over the world in order to capture the effects of differences in GDP (or income) on WTP. They find an income elasticity of 0.38 when using either GDP per capita or household’s personal income. This is in general agreement that WTP income elasticities are in between 0 and 1 (Hokby and Soderqvist, 2003).
Figure 2. The evolution of Greek GDP (2000-2012) in 2005 prices
Even if it is too risky, we will attempt to examine the effects of recession and the consequent loss of income on WTP estimates. Figure 33 shows the effects of recession by assuming a 0.38 and a 0.50 income elasticity on the benefit transfer of 40.93 and a 20% loss in GDP, directly translated into 20% loss in income. These elasticities translate the estimated WTP benefit transfer into 37.82 and 36.84 euros per household per year correspondingly.
Figure 3. Aggregate discounted benefits under recession
4. Meta-analysis: an application
The WFD gave birth to a great number of studies concerning water quality. The announcement for the implementation of the WFD in 2003 coincided with the release of the Green Book from the Department of Environment, Food and Rural Affairs (DEFRA) in the UK, another effort to lay down practical measures and guidelines on how to perform environmental valuation and most importantly, how policies and results can be monetized in a coherent and widely accepted manner.
Economists started applying these economic tools and numerous studies emerged with the WFD either as the target of the proposed policies or at least as a reference point for standards in water quality through measuring the public’s WTP for improving water quality. Carrying out these valuation studies was part of the intention of the Commission to include the public into shaping the policies of the WFD by stating their preferences and finally, their WTP for the proposed strategies. This spurt of research was furthered by the availability of EU and government-issued funding on issues concerning water quality. Closely related phenomena and research opportunities were also brought fourth such as valuation of coastal and marine ecosystem services, values of recreational swimming, fishing and angling and pressures from wastewater in water bodies.
This great number of studies, produced in a rather small time window (15 years) gave ground on performing both benefit transfer exercises and systematic reviews, otherwise known as meta-analyses. Benefit transfer was further favoured due to the geographical proximity of countries (within Europe) and the fact that many of the countries shared common economic and social characteristics. Meta-analyses on the other hand were long known to provide a good starting point for primary studies by giving an overall estimate that could be used as a reference point for the forthcoming results of a study. Although many surveys were carried out, surprisingly enough, many chose not to include household or personal income as an explanatory variable, for reasons later on explained. Income is widely considered as a determinant of WTP and also as one of the main drivers in environmental valuation studies.
4.1. Why conduct a meta-analysis on GES? This work therefore examines the relationship between Willingness to Pay (WTP) for achieving a Good Ecological Status for surface water and income. By definition, WTP is income money that an individual is willing to sacrifice for the production of the environmental good. As such, income must be an obvious empirical factor influencing WTP, beyond solid microeconomic theoretical grounds that are discussed latter in this paper. However, this is not revealed by recent review studies and meta-analyses. Indicatively, in Schläpfer’s (2006) meta-analysis of WTP estimates for environment-related public goods, out of the 83 estimates recorded in 64 studies only 47 (or 56.6%) include an explanatory income variable and only 30 (or 36.1%) record a significant income effect. Similarly, in Jacobsen and Hanley’s (2009) meta-analysis of WTP for biodiversity conservation, out of the 145 estimates recorded in 46 studies only 95 (or 65.5%) include in their analyses an income variable with only 56 of them (or 38.6% of total) recording a significant income effect. Disparate meta-analyses results may be attributed to various reasons. Frequently, the good/service under evaluation is very diverse and/or broadly defined thus creating extreme variation of the average WTP recorded by the studies included in the meta-analysis. Researchers do not share a common understanding of the environmental good or service to be evaluated and thus, the information conveyed to the participants is highly heterogeneous. Participants, on the other hand, if they accept to reveal their true income, may under or over report it for a wide range of reasons. This may be one of the reasons why the overwhelming majority of valuation studies despite they find their sample to be representative of the socio-demographic characteristics of the population, they fail to find it representative of the population’s income or other economic characteristics.
Meta-analysis as a process of quantitative research synthesis is a powerful tool. It is used for summarizing and combining the findings of past research that often reports diverse results but also for comparing and evaluating the associations between the effect size under consideration and the moderator variables that capture heterogeneity through different study characteristics. Taking into account the wealth of meta-analyses concerning WTP estimates up today, the value added by yet another meta-analysis can be significant only if a new perspective is provided. This paper performs a meta-analysis of WTP estimates for attaining Good Ecological Status of surface water in Europe since the adoption of the Water Framework Directive (WFD). As such we aim to reduce heterogeneity coming from diverse definitions of the environmental good under examination. The discussions preceding the adoption of the WFD in 2000 and its application in all Member States, associated countries or countries in the process of accession to the EU, has created, at least among the scientific community, a common understanding of the meaning and pre-conditions for attaining Good Ecological Status (GES) for surface and underground water bodies. So this common scientific understanding is more homogenously transmitted and translated to the general public in valuation studies. Surveys show that the general public in Europe also is informed of the processes contemplated by the WFD in the way to attain GES. In addition, by restricting the survey to European regions or countries we can utilize very detailed and precise EU data sources for wealth and income and incorporate them into our analyses. Thus, by assuming that the respective survey samples are representative of the population of the geographic area in which they are carried out, detailed estimates of the various measures of income can replace missing or probably miss-reported income estimates recorded by surveys.
5. The Water Framework Directive and the Income Elasticity of Willingness to Pay for GES
The EU has more than 100,000 surface water bodies. Of them, 80% are rivers, 15% are lakes and 5% are coastal and transitional waters. The European Union adopted the WFD in 2000 with a view to restore all European waters into Good Ecological Status (GES) by 2015. The definition of ecological status is multidimensional as it looks simultaneously at biological factors (the abundance of aquatic species, phytoplankton, macroalgae and angiosperms in transitional and coastal waters, macrophytes and phytobenthos, benthic invertebrate fauna and fish fauna), chemical and physicochemical elements (thermal conditions, oxygenation, salinity, nutrient status, acidification status and the concentration of specific pollutants including priority substances). The ecological status of water bodies is classified into five classes, i.e., High (class I), Good (II), Moderate (III), Poor (IV) and Bad (V). By the end of 2015, 47 % of EU surface waters have not reached GES (class II). Furthermore, the chemical status of 40 % of surface waters is unknown, showing that monitoring is inadequate in many Member States. As concerns groundwaters, about 25% of them have poor chemical status due to human activities. Pollution of inland surface and ground waters is a long standing environmental concern among European citizens. The latest Eurobarometer conducted in 2012 on water related issues recorded that chemical pollution is cited as the biggest threat to water resources by the majority of Europeans (84%) followed by climate change (55%) and changes in water ecosystems (49%) but respondents were not aware of EU initiatives to water management and especially of the River Basin Management Plans foreseen by the WFD (Eurobarometer, 2012).
The Environmental Kuznets Curve depicts a statistical relationship between national income and pollutants (Grossman and Krueger, 1995). Pollution increases with income up to a level and then decreases. This disputed empirical relationship (Arrow et al., 1995) may be supported by two processes. First, in low income levels, individuals are unwilling to trade consumption for investment in environmental protection and thus environmental quality is low. As the income of individuals grows above the “income turning point”, individuals begin to demand increasing investments in an improved environment and, as a result, decreased pollution and environmental degradation is observed. Second, in the process of economic growth, economies undergo structural economic changes towards less polluting industries and acquire higher levels of technological development in production and abatement which restrict pollution. Concerning contingent valuation studies as the ones presented in this work, the goal is to formulate a survey that revolves around a well-defined commodity so that respondents may be perfectly aware of what they are asked to value, in a way that they might actually feel that they might attain and understand the commodity (Hanemman 1994 ).This is secured here by the well-defined structures of the WFD where the good in question (improved water quality) is clearly stated and “within reach” for the respondent. By eliminating bathing waters, coastal waters etc, the narrow scope made the good far more understandable and conceivable for the respondents. According to Hanemman (1994), the vaguer and less specific the good, the more likely the respondents will consider their monetary contribution as symbolic. To avoid doing so, there is a need for clarity in the details of the payment mechanisms and how they are tied to the realization of the goal or target in question (Hanemman (1994). By having most of the payment vehicles identified as increases in water bills or increases in community taxes, payment appears to be clear for the realization of the improvements in water quality.
Contingent valuation may tread in dangerous waters when valuing public goods, especially in a national scale (Hanemman 1994). The existence of local studies though, as well as considering the fact that this study is restricted to the European Union, which in a way possess a sense of locality, as respondents consider themselves not only members of their country or state but as Europeans as well.
Following Barbier et al. (2015), who provide the most integrated theoretical framework of the relationship between Willingness to Pay (WTP) for pollution control and income, and Hökby and Söderqvist (2003), the income elasticity of demand for an environmental service 푧 with virtual price 푝, and consumer’s income 푦, is:
푦 휕퐷푧 휕(ln 퐷푧) 휀 = = 푦 푧 휕푦 휕(ln 푦)
As it is well known, contingent valuation studies do not allow the estimation of demand functions and, consequently, the estimation of income elasticities (Hökby and Söderqvist, 2003). Contingent valuation studies end up by estimating a WTP function, WTP most often representing a compensation variation function or, the marginal willingness to pay for pollution control. In this setting, variation in WTP is explained by, among others, individual’s income. Thus, the income elasticity of WTP for pollution control 휀푤 becomes:
푦 휕푊푇푃 휕(ln 푊푇푃) 휀 = = 푤 푊푇푃 휕푦 휕(ln 푦)
Estimates of the above elasticity are not estimates of income elasticity of demand and cannot be used as such. However, they are extremely important in the case of the WFD. Kriström and Riera (1996) construct a function of the share of the income directed to WTP as 푠 = (푊푇푃(푦)⁄푦). If this function is decreasing, then, the share of income that is assigned to WTP for controlling pollution decreases as income increases. Then, pollution control is said to be distributed regressively, i.e., would be relatively more beneficial for low-income groups than for high-income groups. Thus, pollution control will be regressively distributed if 휕푠⁄휕푦 < 0 which, by applying the chain rule, results to: 푦 휕푊푇푃 휀 = < 1 푤 푊푇푃 휕푦
Correspondingly, if 푠 = (푊푇푃(푦)⁄푦) is increasing, pollution control is distributed progressively and the income elasticity of WTP for pollution control is greater than one
(휀푤 > 1). The WFD aims to bring all European water bodies at least to a Good Environmental Status. This means that the decision to take all water bodies from a “Bad”, “Poor” or “Moderate” ecological status to “Good” environmental status, irrespective of the social profitability of the project, will have significant implications for poorer households which are more constrained by income than richer households if 휀푤 < 1.
Baumgarten et al (2012) build on the work of Jacobsen and Hanley (2006) and use their meta-analysis estimates to access the distribution of income and the respective effects on average WTP for ecosystem services. The results point to similar direction as Jacobsen and Hanley (2006) and Hokby and Sonderqvist (2003) on the nature of the good, with average WTP for the ecosystem service to increase with mean household income, if the ecosystem service and the consumption good are substitutes. Furthermore, they note that income inequality when it increases, WTP for the good declines. Finally they show that average WTP for the ecosystem service changes more elastically with mean household income than with income inequality and that the elasticity of WTP for ecosystem services in relation with mean household income is three times higher than the elasticity for income inequality.
Many individual studies and meta-analyses report contradicting income elasticity of WTP results or they do not report on income at all. The reason for this may be searched, among others, to data collection mechanisms and responses for individual’s income which is one of the most sensitive pieces of filed collected data. In addition, most surveys report on the representativeness of the collected sample by comparing collected socio-demographic and economic data with data reported by official statistical sources for the same geographic area covered by the survey. The overwhelming majority of surveys end up with a sample that is representative of the wider area or the country at least as concerns socio-demographics. Collected income data, however, may not be recorded or may not be representative of the area for, mainly, two reasons: nonresponding or misreporting. Nonresponse to income questions may be very high. For example, nonresponse to the household net income question in the European Social Survey ranges from close to or above 40% for Portugal, Austria, Spain and Slovakia, and in the range of 20-40% for the Czech Republic, Ireland, Switzerland, Germany and the U.K. (Micklewright and Schnepf, 2010). In the US official Survey for Consumers (SCA) response rates to an open ended income question ranged from 11.7% to 26.2% and for a bracketed question from 5.2% to 15.4% over the 1986-2005 period (Yan et al., 2010). Moore et al. (1999) report income nonresponse rates ranging from a minimum of 20% up to almost 50% for some types of income included in the US Current Population Survey, with questions about income from assets showing the highest rates of nonresponse. Income is still considered as sensitive a topic as the ones related to receipt of welfare, alcohol use, drug use, criminal history, embarrassing medical conditions, etc. and thus many respondents object reporting them on ethical and cognitive grounds. Nonresponse to the income question(s) in contingent valuation field surveys is an everlasting issue, but, unfortunately, it attracts very little attention and is rarely revealed or discussed by researchers. For example, Mourato et al. (2005) report that only 19% of the respondents were not willing to reveal their income while Bateman et al. (2006) report that 50% of the respondents either did not know or refused to reveal their income. in a most recent study undertaken by Doherty et al (2014) 399 respondents out of a sample of 853 did not report their income. A high nonresponse rate to the income variable may result either to the income variable being dropped out of the analyses or to some kind of bias if nonresponses are not randomly distributed among participants.
Misreporting may be as widespread as nonresponse. In the U.S., income underreporting by the self-employed in household surveys is estimated is estimated to be about 30%, and is remarkably robust across data sources and alternative model specifications (Hurst et al., 2014). Misreporting income information may be due to different reasons. Moore et al. (2000) highlight the many possible contributors to inaccurate income reporting, “including lack of knowledge, misunderstanding and other definitional issues, recall problems and confusion, as well as the much-suspected issue of sensitivity”. Respondents tend to unintentionally forget or intentionally suppress income from sources other than salary or wage and especially income coming from intangible or tangible assets, social benefits or remittances. Furthermore, respondents are not always fully aware of the household’s income, let alone the fact that the persons living under the same roof are not always a unified household in terms of expenditures. Finally, respondents are not always clear between the distinction of net and gross income, especially the monthly one, and tend to report as net income the money they receive without taking account of (adding or subtracting) the yearly tax clearance statement. Misreporting income may introduce bias and unreliability that threatens the validity of WTP elasticity estimates. Although some argue that bias may be low as under- or over-reporting center near zero, the effects of low reliability defined as the proportion of measurement error variance, may be significant. Marquis et al. (1986) report that “reliabilities below 70% imply that the correlation of the sensitive topic data with another (perfectly measured) variable will be attenuated by more than 20%, and regression coefficients on the sensitive topic measure by more than 30%”.
Most surveys of willingness to pay for restoring inland waters to GES are localized, are very specific as concerns the water body under consideration, e.g., a lake or the stretch of a river, and are usually conducted to the residents of the area surrounding the resource. However, some nation-wide surveys were conducted, especially in the framework of the WFD. In addition, the distance decay of WTP for use and even for non-use values shows that those who benefit can be easily bordered by larger administrative areas (Hanley et al., 2003; Bateman et al., 2006; Jørgensen et al., 2013). For non-use values, the explanation, although it might appear straightforward, according to Schaafsma et al. (2013) lies to the hypothesis that if the scenarios present future option value for the non-users, then distance decay makes its appearance. It is also true that, in countries which are relatively smaller in terms of spatial coverage, as for example the Netherlands, Denmark, Estonia or even Ireland, surveys of a national coverage are more common. Taking these into account we may assume that, for a survey that is representative of its area, the recorded average WTP can be matched with the official records of the area’s average income measures. Thus, in a meta-regression analysis of the surveys’ WTP estimates with income, officially recorded income values may be used to account for both income misreporting and nonresponse to income questions.
6. Data and Methods 6.1. Data
The search for relevant primary studies was based on a set of initial criteria: 1) it was carried out after 1999 when widespread discussion for the then coming WFD started; 2) it concerned with surface water and provided a full written documentation; 3) it was carried out in one of the EU member states or countries formally associated with the EU; 4) it provided at least one WTP estimate concerning with the ecological condition and restoration of the water body under consideration. The search was carried out in all major databases including the Web of Science (ISI), Scopus and Google Scholar with a range of keywords and their combinations. The initial search returned approximately 130 academic papers in journals, collected works and conferences. From these papers those that reported estimates of a Benefit Transfer method and not of a primary survey, were excluded. This stage reduced the number of eligible works to 71. Furthermore, studies that did not report WTP estimates although they performed valuation methods and studies that identified the Total Economic Value of a watershed for management purposes without reporting average WTP for water quality improvements were excluded. Furthermore, for some works it was impossible to retrieve crucial pieces of statistical information such as sample sizes, data collection criteria, standard errors, p-values or standard deviations from the mean WTP and were also excluded from the present study. This reduced the number of eligible works to 49. These works contained results from 63 different and geographically distinct surveys. For example, Del Saz-Salazar et al. (2009) report on two distinct surveys in the regions of Alcoy and Gandia in Spain while Ahtiainen et al. 2012 report results from 8 different national surveys carried out in the Baltic Sea countries. From these we excluded surveys that focused solely on WTP for drinking water and on WTP for drinking water infrastructures, e.g., Jones et al. 2008, studies that focused solely on bathing water and associated health risks (Brouwer and Bronda, 2005) or solely on marine water (Eggert and Olsson, 2009). Studies that valued a very specific issue and not the status of the water body in the spirit of the WFD, were also excluded. These, for example, included a study on the value of cleaning sediments in water courses (Brouwer, 2004a as reported by Brouwer, 2006) or a study on preferences to land use management in relation to the water body’s status (Garcia-Liorente et al., 2012). When WTP estimates were not statistically significant from zero due to a combination of sampling and methodological issues, the case study was excluded. This is, for example, the case in Hanley et al. (2006) where the sample of the case study area of River Wear in the UK produces statistically significant WTP estimates under two alternative econometric methods while the sample in the River Clyde area fails to do so due to uncontrolled preference heterogeneity. Finally, studies that analysed the same data set or sub-sets of the original dataset but from a different research perspective were also excluded. After that, the final dataset included 21 distinct studies which accounted for 32 distinct surveys that took place on well-defined national or regional spatial units. Table A at the Appendix B of this work records all included surveys, their respective WTP estimates and most important characteristics. In this survey we use one measure of wealth and two of income: Gross Domestic Product (GDP), Income of Households and Disposable Income of Households. We use the Gross Domestic Product per capita for the wider geographic area (NUTS2 or nation) in the year the study was conducted as a measure of the region’s wealth. Eurostat measures GDP by the expenditure approach which is defined as the total value of the final consumption expenditures of households, NPISHs (Non-profit institutions serving households) and general government plus gross capital formation plus the balance of exports and imports. This is divided by the total resident population to produce an index of wealth as euro of GDP per inhabitant abbreviated in this work as GDP/CA (Eurostat variable: nama_10r_2gdp). The same index is produced by Eurostat using Purchasing Power Parity (PPP). PPP may be understood as a currency conversion rate that converts economic indicators such as GDP expressed in a national currency to an artificial common currency that equalises the purchasing power of different national currencies. As such PPP is both a price deflator and a currency converter and thus it eliminates the differences in price levels between countries. In this work that compares monetary values across EU member states which may have the same or different currency but definitely different purchasing power, this index is extremely useful and is abbreviated, in this work as GDP/CA-PPP.
As stated before, income is measured by two indicators. First the income of households derived by the balance of primary income of national accounts including compensation of employees (wages, salaries and employer’s social contribution), mixed income from activities including agriculture and rental income less cost, and property income including interest and dividends (Eurostat variable: nama_r_ehh2inc). This is divided by Eurostat with the resident population to produce income of households per capita abbreviated in this work as HINC/CA. The same index is produced by Eurostat using Purchasing Power Parity (PPP) and is abbreviated in this work as HINC/CA-PPP. The second income indicator refers to disposable income which is produced by income of households deduced by taxes, net social contributions and non-taxed net transfer income (Eurostat variable: nama_r_ehh2inc). This is divided by Eurostat with the resident population to produce disposable income of households per capita abbreviated in this work as HDINC/CA. The same index is produced by Eurostat using Purchasing Power Parity (PPP) and is abbreviated in this work as HINC/CA- PPP. Both indices of income are very important for the scope of this work as they reflect the ability of households in an area to generate earnings (HINC/CA) and the income they have to allocate between consumption goods and environmental goods, harmonized across countries using PPP.
Some of the surveys included in this meta-analysis record multiple estimates of WTP. Multiple estimates come from the fact that researchers either report estimates from different options or scenarios presented to respondents or estimates of the same option/scenario using different econometric approaches, most frequently the Multinomial Logit (MNL) and the Random Parameters Logit (RPL). From the 32 case studies, 21 report multiple estimates and 11 report single estimates. Our estimation strategy involves three alternative measures of WTP estimates. The first measure is a single WTP estimate, for those case studies that record multiple estimates, which is judged to represent best the WFD’s aim to restore the water body to GES. Following the terminology in Vista and Rosenberger (2013) we name this set of data the “best-set”. This did not always involve a judgment decision because many of the surveys asked their participants this very exact question. For example, the two case studies reported by Del Saz-Salazar et al. (2009) in the areas of Alcoy and Gandia in Spain provide only one WTP estimate which explicitly addresses the issue of restoring the respective water bodies to GES. In another case study Martin-Ortega and Berber (2010) divide the watershed in zones and all respondents state their WTP for upgrading the ecological status of each zone to GES. However, in this case, and based on the environmental status of each particular zone, restoring could refer to an upgrading from moderate to good or very good, from good to very good or from poor to moderate or good. In this particular case we chosen as best estimate the poor to good option from all available and statistically significant estimates.
The second measure is again a single estimate based on the average of the multiple estimates recorded by each study. We name this set of data the “average set”. In this case, the variance of the mean 푊푇푃 for a case study that reports 푚 estimates with each estimate having variance 푉푖 , is computed as:
푚 푚 푚 1 1 2 1 2 푣푎푟 ( ∑ 푊푇푃 ) = ( ) 푣푎푟 (∑ 푊푇푃 ) = ( ) (∑ 푉 + ∑(푟 √푉 √푉 )) 푚 푖 푚 푖 푚 푖 푖푗 푖 푗 푖=1 푖=1 푖=1 푖≠푗 where 푟푖푗 is the correlation between any pair of estimates 푖 and 푗 recorded by one case study. As this correlation coefficient is unknown or not recorded by researchers, we performed sensitivity analysis producing results assuming coefficients ranging from 0.5 to 0.9 without any important change in the calculated variance. The third measure of WTP is constructed by including all the relevant recorded estimates of one case study, irrespective of method or scenario, as individual observations. We name this set “all relevant set”. From each survey we isolated all estimates resultant from scenarios in the spirit of the WFD. For example, in the study by Stithou et al. (2012) recording estimates from scenarios referring to restoring river ecology, restoring condition of river banks and providing for more recreational activities the estimates on recreational activities were left out while the estimates of WTP on restoring river life and river banks were included as multiple estimates from the same study. Conceptually, there is no point to include in an analysis of the WTP for restoring GES, estimates that are far away from the aims and objectives of the WFD. Multiple observations are stratified under each study to produce either clustered observations or a panel.
All monetary values included in this work are converted to constant 2005 euros using first the average exchange rate of the year they are reported (Eurostat variable: ert_bil_eur_a) and then the Harmonized Index of Consumer Prices (HICP) reported by Eurostat (variable: prc_hicp_aind). Certain WTP were provided in PPP (e.g., Brouwer et al., 2016) or in market and PPP values (e.g., Czajkowski and Ščasný, 2010) and were all converted to market 2005 deflated values using the HICP. The income variables are provided by Eurostat directly in 2005 euro values. For each case study area, a range of “control” variables pertaining to the survey and the area were recorded. Survey specific control variables include the type of estimation method, if non-users and/or protest voters are included, the specified payment vehicle, dummy variables capturing if income was recorded, whether the sample’s representativeness was tested (statistically or not), if the WFD was explicitly stated in deriving alternative scenarios and storylines presented to respondents and if these scenarios more closely resembled a “moderate to good” or a “poor to good” environmental status upgrade. Area specific control variables include the scale of the survey measured in Km2, if it included transitional waters, if the watershed spanned across two countries, if in addition to the GES also valued issues related to water availability, water related recreation and angling, if the area is located within the Eurozone or outside, within the European South or outside and if the survey was carried out before or after the 2008 financial crisis. Table 1 shows all variables and their descriptive statistics. All variables are estimated on a size of 32 for the best set data or on 70 values for the all set data, except of the Income variable which is estimated on 13 and 33 values respectively and the case study’s area which is estimated on 11 and 25 values respectively. The descriptive statistics for the average relevant set is the same as the best set data except for the WTP variable with a mean of 47.13.
6.2. Methods Meta analysis was first used outside the medical sciences in the field of water demand (Esprey et al 1997, Gaudin (2006), environmental conservation (Jacobsen and Hanley 2009, Ghermandi et al. (2010)) and environmental valuation. The so-called “mean effect” of meta- analysis is the variable of interest where changes are depicted, either between “treated” and “non-treated” or between “change” or “no-change” scenarios. In this case, WTP is selected as the effect size.
WTP estimates are true effect sizes in the conventional sense of meta-analysis because they estimate the effect that a proposed environmental change will have on consumer welfare. As such, WTP estimates measure a difference (before-after) resultant from a treatment (environmental action undertaken to GES). Thus, the single WTP estimates, either the best- set estimate or the average set estimate will be treated using conventional meta-analysis and meta-regression techniques. In search for the overall effect size across all included case studies, we may pursue a fixed or a random effects meta-analysis. If we assume that all surveys share the same true WTP estimate noted by 휃, it follows that the observed WTP varies from case study area to case study area only because of the random error inherent in each survey. So, the observed WTP of a survey 푖, 푊푇푃푖 is given by the population mean 휃 plus the sampling error 휀푖 of that survey as:
푊푇푃푖 = 휃 + 휖푖 (1)
2 with 휀푖 ~푁(0, 휎푖 ) which is the fixed effects formulation. In many survey data meta-analysis, and surely in WTP surveys, the assumption that the true WTP is the same in all studies is implausible. WTP surveys differ in many respects even when the environmental service under consideration is narrowly and strictly defined or when the geographic coverage of the surveys is more homogenous. One way to acknowledge and address the variation across studies is to employ a random effects meta-analysis in which the true WTP of the surveys
(휃푖) have a normal distribution around a grand mean 휃, a true distance from the grand mean (푢푖) and the survey’s sampling error to be (휀푖) as:
푊푇푃푖 = 휃 + 푢푖 + 휖푖 (2) 2 2 2 with 휃푖 ~푁(휃, 휏 ) and 푢푖 ~푁(0, 휏 ) and 휀푖 ~푁(0, 휎푖 ). The variation in the true WTP estimates (휏2) is the heterogeneity among WTP surveys. There are certain mechanisms that allow us to extract the true between surveys variation in WTP from the observed variation. The measure 퐼2 reflects the proportion of the observed variance that is due to real differences in WTP survey estimates and is estimated as a ratio of excess dispersion to total dispersion.
Meta-analysis regression, or meta-regression, is an analytic extension to standard meta- analysis that examines the extent to which heterogeneity between WTP estimates of multiple studies can be related to one or more characteristics of the studies (Thompson and Higgins, 2002). In this work, meta-regression is used to investigate the effects of alternative notions of income on WTP. As such, random-effects meta-regression can be considered as an extension to random-effects meta-analysis presented in equation (2) that includes study- level covariates. This assumes that the true WTP follow a normal distribution around a linear 2 2 predictor, 푊푇푃푖|휃푖 ~푁(휃푖, 휎푖 ) where 휃푖 ~푁(푥푖훽, 휏 ) as:
푊푇푃푖 = 푥푖훽 + 푢푖 + 휖푖 (3)
2 2 with 푢푖 ~푁(0, 휏 ) and 휀푖 ~푁(0, 휎푖 ). Among the 푥 vector of covariates alternative variables capturing survey or region specific income will be used in different functional forms.
In the case where multiple estimates are recorded per study, we take two alternative estimation routes. First, we treat each estimate as a separate observation ignoring the panel structure of the data and estimate it by OLS with cluster corrected standard errors. Previous work has found that the panel structure (fixed or random) of multiple estimates per case study can be rejected in favor of an equal-effect (no panel effects) specification (Rosenberger and Loomis, 2000). The second route is to estimate a panel random effects model with or without weights. In the simple random effects panel model we use the Generalized Least Squares (GLS) estimator to address heteroscedasticity issues resultant from the unbalanced structure of the data. When the panel is weighted, the weight assigned to each study is the inverse of the within study variance with the weights for the observations in each study adding up to the study’s weight in the random effects meta- regression of equation 3. The model without weights was estimated using the Stata xtreg command and the model with weights was estimated using the Stata gllamm command.
We choose to show only the results from the log-log models both in the cross-section and the panel structure of the data for reasons of cohesiveness and length of the study. We note though that the results from the lin-lin models and the log-lin models for all 6 income estimates, were in their vast majority statistically significant (at least at the 5% level) and their respective signs were the same, regardless the form of the model. The results are provided upon request to the authors.
7. Results
The method used in this work is an extension of the non-iterative method of moments approach originally proposed by DerSimonian et al (1986). Alternative and more computationally intensive parametric random-effects models for estimating between-study heterogeneity in meta-analysis include, but are not necessarily limited to, maximum likelihood, restricted maximum likelihood and profile likelihood (Thompson and Sharp, 1999; Hardy and Thompson, 1996; Morris, 1983). The nonparametric random-effects model, known as the permutations model, has also been proposed for estimating the between- study variance in an aggregate data meta-analysis (Follmann and Prooschan, 1999). All methods can be implemented in STATA using the metan routine (Harris et al., 2008) or the metaan routine proposed by Kontopantelis and Reeves (2010). The overall mean of the single best set estimate for the 32 studies involving 20,065 respondents over a period of 14 years is estimated by the DerSimonian-Laird random effects model to be 41.75 euro per year [CI 95% at 26.71, 56.80]. The corresponding fixed effects estimate, taking into account the highly different weights assigned to the studies, is surprisingly close to the fixed effects estimate at 31.36 euro per year [31.25, 31.47]. The corresponding random and fixed estimate for the average set reported WTP are 46.44 [26.42, 66.45] and 34.57 [34.46, 34.68] euro per year. These estimates are notably robust to changes as concerns the sampled surveys. If we exclude 25% of the surveys with the lowest WTP estimates or the 25% of surveys with the highest WTP estimates, the corresponding estimated overall averages remain within the confidence intervals estimated for the full sample of surveys. The same holds true if the multi-country work of the eight Baltic sea surveys is excluded or if surveys carried out before 2006 (mid-period year) are excluded.
Figure 1 shows the forest plot of all included surveys and the estimated mean WTP of the random-effects model of equation (2). In this figure, each study is represented by a line. The centre of the line is the mean WTP reported by the study and converted to 2005 euro values, the bars represent the 95% confidence intervals of the study and the size of the shaded rectangle represents the weight assigned to each study according to the study’s variance and the between studies variance. The studies are presented in the chronological order of the time they carried out the survey and not the time of publication. A simple observation of the forest plot shows the very high heterogeneity among WTP estimates. The I2 value is almost 100% showing that this observed heterogeneity is due to real differences in WTP survey estimates. Figure 2 shows a plot of the cumulative meta-analysis of the best set data. It is notable that the overall random effects average converges to its final value of around 42 euros per year as early as the start of 2008 and based on highly diverse and geographically dispersed studies. This occurs before the incorporation of the large study carried out in the eight Baltic Sea countries (Russia is excluded from the analysis).
In order to examine why WTP estimates differ so much we run a series of sub-group analyses for various groups of surveys and tested whether the observed differences of the mean WTP of the various subgroups are statistically significant. First, we examined the difference in the best set mean WTP between the 21 studies that refer explicitly to the WFD and the 11 studies that do not refer explicitly to the WFD but evaluate the ecological status of the water body under consideration. The former have an average WTP of 46.45 euro [22.80, 70.11] and the later of 26.72 euro [22.64, 30.79]. The difference is statistically significant at 1%. Second, we examined the difference between the studies that report or discuss sample representativeness and we did not find any significant difference. Then we examined the difference in the mean WTP between all sub-groups formed by the control dummy variables. None of the differences was found to be statistically significant. Thus, all the study design variables, the case study characteristics and the location of the case study area in a European context (North-South, Eurozone or not) were insignificant determinants of WTP for the restoration of water bodies to GES. In a meta-regression framework, Table 2 shows that all study design variables, except from the one reflecting if the study explicitly refers to the WFD, were insignificant determinants of WTP. The left column of Table 2 shows at the upper part the results when the relationship of WTP to the income recorded by the studies is linear. The lower part shows the results of the same functional form (lin-lin) with the GDP/CA imputed variable. The right part of the table depicts the same results but with a log-log functional form. Of course, one should note that, with a sample size of 32 studies, the number of dummy control variables that could be simultaneously included in a meta- regression was limited by multicollinearity. However, the non-significance of most of the explanatory variables also has been in other relevant meta-analyses (Jacobsen and Hanley, 2009).
In the second stage we run a number of meta-regressions for the best set WTP, the average WTP and the all relevant set WTP estimates, against each one of the seven measures of income and for three functional forms specifications, i.e., lin-lin, log-lin and log-log. Table 3 shows the results of the log-log form which is of interest in this study since it gives the income elasticity of WTP. Tables 3 to 5 present the results for the best choice and the average WTP and each table for each of the three functional forms. Tables 6 to 8 present the results for the all-relevant estimates of WTP and each table for each of the three functional forms. The coefficient estimates of the plain OLS are exactly the same as the meta- regression. The coefficient estimates of Tables 3 to 5 are produced using the Knapp and Hartung variance estimator implemented by the metareg Stata command (Harbor and Higgins, 2008). The estimated standard errors remain almost the same when estimated by a permutation test proposed by Higgins and Thompson (2004). The assumption of normal random effects is adequate for all specifications as this is revealed by the relevant normal probability plots of the standardized shrunken residuals. The log-log models provide the best fit with the adjusted R2 being in some models as high as almost 50%.
The coefficients of the semilog (log-lin) models provide an estimate of the rate of change of the estimated WTP when the respective income variable changes by 1,000 euro. The highest growth rate is observed for the income variable measuring Household disposable income per capita adjusted for purchasing power (HDINC/CA-PPP) at 10.3% per 1,000 euro. All the coefficients of the purchasing power adjusted measures are higher than the respective non- adjusted measures. These estimates are very robust across different specifications of the dependent variable (best-choice or average WTP), different econometric estimation (OLS or random effects meta-regression) and different measures of wealth (GDP) or income.
The coefficients of the log-log models are measures of the elasticity of WTP with respect to income. All estimated coefficients are statistically significant even the one for the income variable as this is reported in the analyzed studies (for 13 out of the 32 studies). Tables 5 and 8 show that the variability of the estimated elasticities with respect to the income measure used in the model is considerable but the estimates among alternative econometric approaches are notably close. Elasticities range from around 0.75 when the independent variable is the income recorded by the studies to around 1 for wealth and income measures and up to around 1.5 when the same measures are PPP adjusted. All the above results, for cross sections or panels and for all functional forms remain notably robust in the magnitude, sign and level of significance of the coefficients, even when the sample is trimmed by as much as 15% from below or above or when certain influential case studies are withdrawn. For the results from all the estimations and handling of the data (both panel and cross- section estimations, in all functional forms and for all four models used, see in the Appendix A, tables 5 to 8).
The overall result is that income is a significant and substantive determinant of WTP. Depending on the income measure used an independent variable, achieving GES through the WFD can be said to be distributed regressively (income recorded by the studies), or progressively (wealth and income by official sources, especially when PPP adjusted). When the elasticity is greater than 1 then the proportion of WTP to the income measure is increasing as income rise and thus, achieving GES has proportionately higher benefits to richer households than to poorer households. Thus, if the transition to GES was not obligatory through the WFD, it is possible that this would occur in regions populated by higher income households. If this is true, the WFD, by restoring GES, supports an economic allocation that is more efficient (addresses the externality) and more fair as it supports lower income households. Such a discussion, however, is out of the scope of this work.
8. Conclusions
This work examined the empirical relationship between income and WTP for restoring GES of European water bodies within the WFD. Published research since the early 2000 allows for data to be treated with meta-analyses methodologies. We offer one alternative explanation for the seemingly disparate results related to the income elasticity of WTP. These may be due either to the heterogeneity of the valuated good or service or to no reported or miss-reported income. To do this we traded large sample properties for more narrowly but precisely defined valuation studies. The decision to select for meta-analysis the case studies related to the application of the WFD and valuing restoration to GES served our aim in two respects. First, it included studies valuing a well defined service, i.e., restoration to GES, which is commonly understood by the scientific community in Europe. As such, it may be more easily conveyed to the respondents and the risk to be misperceived by them is lower. Second, all study areas were located, by definition, in Europe and addressed well defined water bodies in terms of geography. This allowed access to Eurostat wealth and income data for the relevant geographic units (NUTS2 regions or countries). Eurostat data guarantee validity and, to a large extent, homogeneity. Thus, the use of official data alleviates no reporting or miss-reporting of income under the assumption that the samples are representative of the corresponding NUTS 2 regions or countries in which the field study took place. We believe that it is likely to have a correct measurement of the sample’s true income irrespective of whether the sample is representative of the area or not.
In this work we find elasticities that are both numerically substantial and statistically significant. In certain functional forms, income variables account for as much as 50% of WTP’s variability with estimates ranging from below one to well above one if the wealth and income indicators are PPP adjusted. These results are, of course, only indicative and should not be used as a policy guideline. However, they highlight the need for future research in three directions. First, to apply the same practice on larger and more geographically disperse samples in order to overcome issues of data sparseness. This, of course, should not be done on the expense of a well defined and narrowly framed environmental good or service. European Union wide applications of environmental regulations and directives offer such an opportunity. Second, if in larger and more geographically dispersed samples within the EU, the log of wealth or income remains a good predictor of WTP, its use in Benefit Transfer should be examined, at least as a first approximation. In this respect, as more studies are carried out or become publicly available, meta-analyses may be used for assessing the welfare implications of the application of the WFD together with efficiency (cost-benefit) calculations. A third line of future research could be related to the use of cumulative meta- analysis not on time but on income to reveal the relationship between WTP and income as the latter cumulates to higher levels.
Bibliography
Ahtiainen H., Hasselström L., Artell J., Angeli D., Czajkowski M., Meyerhoff J., Alemu M., Dahlbo K., Fleming-Lehtinen V., Hasler B., Hyytiäinen K., Karlõseva A., Khaleeva Y., Maar M., Martinsen L., Nõmmann T., Oskolokaite I., Rastrigina O., Pakalniete K., Semeniena D., Smart J. and Söderqvist T. 2012. Benefits of meeting the Baltic Sea nutrient reduction targets - Combining ecological modelling and contingent valuation in the nine littoral states. Discussion paper MTT 2012. http://www.mtt.fi/dp/DP2012_1.pdf Arrow, K., Bolin, B., Constanza, R., Dasgupta, P., Folke, C., Holling, C. S., Jansson, B., Levin, S., Mäler, K. G., Perrings, C. and Pimentel, D. 1995. Economic Growth, Carrying Capacity, and the Environment. Science, 268, 520-521. DOI: 10.1126/science.268.5210.520 Assessment, M.E., 2005. Millennium ecosystem assessment. Ecosystems and Human Well- Being: Biodiversity Synthesis, Published by World Resources Institute, Washington, DC. Barbier, E.D., Czajkowski, N. and Hanley, N. 2015. Is the income elasticity of the willingness to pay for pollution control constant? University of Warsaw, Faculty of Economic Sciences, Working paper 7/2015, Warsaw. DOI: 10.1007/s10640-016-0040-4 http://www.wne.uw.edu.pl/index.php/download_file/1542/ Bateman, I.J., Day, B., Georgiou, S. and Lake, I. 2006. The aggregation of environmental benefit values: Welfare measures, distance decay and total WTP. Ecological Economics, 60(2), 450-460 . DOI:10.1016/j.ecolecon.2006.04.003 Baumgärtner, S., Drupp, M., Meya, J., Munz, J., & Quaas, M. F. Income distribution and willingness to pay for ecosystem services. Birol E. and Cox V. 2007. Using choice experiments to design wetland management programmes: The case of Severn Estuary Wetland, UK., Journal of Environmental Planning and Management, 50(3), 363-380. DOI:10.1080/09640560701261661 Birol, E., N. Hanley, P. Koundouri, and Y. Kountouris (2009), Optimal management of wetlands: Quantifying trade-offs between flood risks, recreation, and biodiversity conservation, Water Resources Research, 45, W11426, 1-11. DOI: 10.1029/2008WR006955 Bliem M. and Getzner M. 2008, Valuation of ecological restoration benefits in the Danube River basin using stated preference methods – Report on the Austrian case study results. Institute for Advanced Studies Carinthia / Department of Economics, Klagenfurt University, Klagenfurt, Austria. http://www.ivm.vu.nl/en/Images/D35a_Case_study_report_Danube_Austria_tcm23 4-188719.pdf Brouwer, R. 2004a. What is the Economic Value of Cleaner Water as a Result of the Implementation of the EU Water Framework Directive? [in Dutch], RIZA report 2004.013, Lelystad, The Netherlands: Institute for Integrated Water Management and Wastewater Treatment (RIZA). Brouwer, R. 2004b. Valuing Biodiversity in and around Water Bodies as a Result of Contaminated Sediment Clean-up with the Contingent Valuation Method [in Dutch], RIZA report 2004–022, Lelystad, The Netherlands: Institute for Integrated Water Management and Wastewater Treatment (RIZA). Brouwer, R. 2006. Valuing water quality changes in the Netherlands using stated preference techniques, in D. Pearce (ed.), Environmental valuation in developed countries: case studies, pp. 132-147. Cheltenham, UK: Edward Elgar Publishing. Brouwer, R. and Bronda, R. 2005. The costs and benefits of a revised European Bathing Water Directive in the Netherlands, in R. Brouwer and D.W. Pearce (eds), Cost– Benefit Analysis and Water Resources Management, pp. 251-269. Cheltenham, UK: Edward Elgar Publishing. Brouwer R., Bliem M., Getzner M., Kerekes S., Milton S., Palarie T., Szerényi Z., Vadineanu A. and Wagtendonk A. 2016. Valuation and transferability of the non-market benefits of river restoration in the Danube river basin using a choice experiment. Ecological Engineering, 87, 20-29. DOI:10.1016/j.ecoleng.2015.11.018 Czajkowski M. and Ščasný M. 2010. Study on benefit transfer in an international setting. How to improve welfare estimates in the case of the countries' income heterogeneity? Ecological Economics, 69, 2409–2416. DOI :10.1016/j.ecolecon.2010.07.008 DerSimonian, R., and Laird, N. 1986. Meta-analysis in clinical trials. Controlled Clinical Trials, 7, 177–188. DOI:10.1016/0197-2456(86)90046-2 Doherty E., Murphy G., Hynes S. and Buckley C. 2014. Valuing ecosystem services across water bodies: Results from a discrete choice experiment. Ecosystems Services, 7, 89–97. DOI:10.1016/j.ecoser.2013.09.003 Eggert H. and Olsson B. 2009. Valuing multi-attribute marine water quality. Marine Policy, 33, 201– 206. doi:10.1016/j.marpol.2008.05.011 Espey, M., Espey, J., & Shaw, W. D. (1997). Price elasticity of residential demand for water: a meta‐analysis. Water resources research, 33(6), 1369-1374. Eurobarometer, 2012. Attitudes of Europeans towards water – related issues. Flash Eurobarometer 344 - TNS Political & Social. http://ec.europa.eu/public_opinion/flash/fl_344_en.pdf Follmann, D. A., and Proschan, M.A. 1999. Valid inference in random effects meta-analysis. Biometrics, 55, 732–737. García-Llorente M., Martín-López B., Nunes P.A.L.D., Castro A.J. and Montes C. 2012. A choice experiment study for land-use scenarios in semi-arid watershed environments. Journal of the Arid Environments, 87, 219- 230. doi:10.1016/j.jaridenv.2012.07.015 Gaudin, S. (2006). Effect of price information on residential water demand. Applied economics, 38(4), 383-393. Ghermandi, A., Van Den Bergh, J. C., Brander, L. M., De Groot, H. L., & Nunes, P. A. (2010). Values of natural and human‐made wetlands: A meta‐analysis. Water Resources Research, 46(12). Grossman, G.M. and Krueger, A.B. 1995. Economic growth and the environment. Quarterly Journal of Economics, 110, 353-377. Hanemann, W. M. (1994). Valuing the environment through contingent valuation. The Journal of Economic Perspectives, 8(4), 19-43. Hanley, N., Schläpfer, F. and Spurgeon, J. 2003. Aggregating the benefits of environmental improvements: distance-decay functions for use and non-use values. Journal of Environmental Management, 68, 297–304. doi:10.1016/S0301-4797(03)00084-7 Hanley N., Wright R. E. and Alvarez-Farizo B. 2006. Estimating the economic value of improvements in river ecology using choice experiments: an application to the water framework directive. Journal of Environmental Management, 78, 183–193. DOI: 10.1007/1-4020-5405-X_7 Hanley N., S. Colombo, D. Tinch, A. Black and A. Aftab 2006. Estimating the benefits of water quality improvements under the Water Framework Directive: are benefits transferable? European Review of Agricultural Economics, 33(3), 391–413. doi: 10.1093/eurrag/jbl019 Harbord, R. and Higgins, J. 2008. Meta-regression in Stata. The Stata Journal, 8, 493–519. Hardy, R. J. and Thompson, S.G. 1996. A likelihood approach to meta-analysis with random effects. Statistics in Medicine, 15, 619–629. DOI:10.1002/(SICI)10970258(19960330)15:6<619::AID-SIM188>3.0.CO;2-A Hökby, S. and Söderqvist T. 2003. Elasticities of Demand and Willingness to Pay for Environmental Services in Sweden. Environmental and Resource Economics, 26(3), 361-383. DOI:10.1023/B:EARE.0000003581.97411.75 Higgins, J. P. T., and Thompson, S. G. 2004. Controlling the risk of spurious findings from meta-regression. Statistics in Medicine, 23, 1663–1682. DOI: 10.1002/sim.1752 Hurst, E., Li, G. and Pugsley, B. 2014. Are Household Surveys Like Tax Forms: Evidence from Income Underreporting of the Self Employed. Review of Economics and Statistics, 96(1), 19-33. Jacobsen J. B. and Hanley N. 2009 Are There Income Effects on Global Willingness to Pay for Biodiversity Conservation? Environmental and Resource Economics, 43, 137–160. DOI: 10.1007/s10640-008-9226-8 Jones N. , Sophoulis C. and Malesios C. 2008. Economic valuation of coastal water quality and protest responses: A case study in Mitilini, Greece. The Journal of Socio- Economics Vol. 37 pp. 2478–2491 doi:10.1016/j.socec.2007.06.002 Jørgensen, S.L., Olsen, S.B., Ladenburg, J., Martinsen, L. and Svenningsen, S.R. 2013. Spatially induced disparities in users' and non-users' WTP for water quality improvements—Testing the effect of multiple substitutes and distance decay. Ecological Economics, 92, 58–66. doi:10.1016/j.ecolecon.2012.07.015 Kontopantelis, E. and Reeves, D. 2010. metaan: Random-effects meta-analysis. The Stata Journal, 10, 395–407. Kriström, B., and Riera, P. 1996. Is the income elasticity of environmental improvements less than one? Environmental and Resource Economics, 7,(1), 45-55. DOI:10.1007/BF00420426 Marquis, K.H., Marquis, M.S. and Polich, M.J. 1986. Response Bias and Reliability in Sensitive Topic Surveys. Journal of the American Statistical Association, Vol. 81(394), 381-389. DOI: 10.2307/2289227 Martin-Ortega J. and Berbel J. 2010. Using multi-criteria analysis to explore non-market monetary values of water quality changes in the context of the Water Framework Directive. Science of the Total Environment, 408, 3990–3997. doi:10.1016/j.scitotenv.2010.03.048 Micklewright, J. and Schnepf, S.V. 2010. How Reliable are Income Data Collected with a Single Question? Journal of the Royal Statistical Society: Series A, 173(2), 409–429. DOI: 10.1111/j.1467-985X.2009.00632.x Moore, J., Stinson, L. and Welniak, E. 1999. Income Reporting in Surveys: Cognitive Issues and Measurement Error. in Sirkin, M., Herrmann, D., Schechter, S., Schwarz, N., Tanur, J. and Tourangeau, R. (eds.), Cognition and Survey Research, pp. 155–173. New York: John Wiley and Sons. Moore, J.C., Stinson, L. and Welniak, E. 2000. Income measurement error in surveys: a review. Journal of Official Statistics, 16(4), 331–361. Morris, C. N. 1983. Parametric empirical Bayes inference: Theory and applications. Journal of the American Statistical Association, 78, 47–55. DOI: 10.2307/2287098 Mourato, S., Atkinson, G., Ozdemiroglu, E., Newcombe, J. and de Garis, Y. 2005. Does a Cleaner Thames Pass an Economic Appraisal?. Water International, 30, 174-183, DOI: 10.1080/02508060508691858. De Nocker L., Broekx S., Liekens I., Görlach B., Jantzen J., Campling P. 2007 Costs and Benefits associated with the implementation of the Water Framework Directive, with a special focus on agriculture: Final Report
Poirier J. and Fleuret A. 2010 .Using the choice experiment method for valuing improvements in water quality: a simultaneous application to four recreation sites of a river basin. Discussion paper.https://feem- projectnet.serversicuro.it/exiopol/userfiles/POIRER_paper%20exiopol.pdf
Ramajo-Hernandez J.and del Saz-Salazar S. 2012. Estimating the non-market benefits of water quality improvement for a case study in Spain: A contingent valuation approach.Environmental Science Policy, 22, 44-59. doi:10.1016/j.envsci.2012.05.006
Rosenberger, R.S. and Loomis, J.B. 2000. Panel stratification in meta-analysis of economic studies: An investigation of its effects in the recreation valuation literature. Journal of Agricultural and Applied Economics, 32, 459–470. DOI: http://dx.doi.org/10.1017/S1074070800020563 Schläpfer F. , 2006, Survey protocol and income effects in the contingent valuation of public goods: A meta-analysis. Ecological Economics vol. 57(3), 415-429. doi:10.1016/j.ecolecon.2005.04.019 del Saz-Salazar S., F. Hernández-Sancho and R. Sala-Garrido 2009. The social benefits of restoring water quality in the context of the Water Framework Directive: A comparison of willingness to pay and willingness to accept. Science of the Total Environment, 407, 4574–4583. del Saz Salazar S., García-Rubio M.A., González-Gómez F., Picazo -Tadeo A.J. 2012. A willingness to pay scenario for valuing water supply infrastructure improvements: A case study for Southern Spain. Working paper. http://www.revecap.com/encuentros/trabajos/s/pdf/114.pdf Schaafsma M., Brouwer R. and Rose J. 2013. Directional heterogeneity in WTP models for environmental valuation. Ecological Economics, 79 , 21–31. doi:10.1016/j.ecolecon.2012.04.013 Stithou M., Hynes, S., Hanley, N. and Campbell, D. 2012. Estimating the Value of Achieving “Good Ecological Status” in the Boyne River Catchment in Ireland Using Choice Experiments. The Economic and Social Review, 43, 397–422. Thompson, S. G. and Higgins, J.P.T. 2002. How should meta-regression analyses be undertaken and interpreted?. Statistics in Medicine, 21, 1559–1573. DOI: 10.1002/sim.1187 Thompson, S. G., and Sharp, S. J. 1999. Explaining heterogeneity in meta-analysis: A comparison of methods. Statistics in Medicine, 18, 2693–2708. DOI: 10.1002/(SICI)1097-0258(19991030)18:20<2693::AID-SIM235>3.0.CO;2-V Vista, A. and Rosenberger, R. 2013. Addressing dependency in the sportfishing valuation literature: Implications for meta-regression analysis and benefit transfer. Ecological Economics, 96, 181-189. doi:10.1016/j.ecolecon.2013.10.016 Yan, T., Curtin, R. and Jans, M. 2010. Trends in income nonresponse over two decades. Journal of Official Statistics, 26(1), 145-164.
Table 1. Variables and descriptive statistics Variable Best set All Set
WTP The WTP value recorded by the study in euro at 2005 prices 43.179 (34.039) 48.834 (38.861)
Sample size The sample size recorded for each study area 627.031 (492.545) 572.271 (469.601)
Year of Study The year the study’s field work started (all studies were complete within the same year) 2007.188 (3.596) 2006.929 (3.560)
Income The average income recorded by the study 23,005.87 (15,092.62) 27,489.63 (1,5174.62)
GDP/CA GDP per capita of the NUTS2 region or the country of the study 25,151.93 (12,351.03) 26,470.34 (1,2215.34)
GDP/CA-PPP The variable GDP/CA above in PPP 24,819.36 (7,756.88) 25,190.33 (7,973.934)
HINC/CA The income of households recorded by the balance of primary income of national accounts per capita 16,388.91 (7,436.01) 17,263.47 (7,071.502)
HINC/CA-PPP The variable HINC/CA above in PPP 15,661 (4,933.05) 16,058.89 (4,725.924)
HDINC/CA Disposable income of households per capita 14,582.11 (5,709.75) 15,348.19 (5,251.108)
HDINC/CA-PPP The variable HDINC/CA above in PPP 13,743.59 (4,342.86) 14,110.44 (4,077.784)
WFD Qualitative variable; 1if the WFD was explicitly referenced by the study’s design; 0 otherwise 0.656 (0.482) 0.728 (0.448) REPR-TEST Qualitative variable; 1if the study’s sample has been tested for representativeness; 0 otherwise 0.469 (0.507) 0.557 (0.500)
NON-USERS Qualitative variable; 1if non-users were included in the study’s sample; 0 otherwise 0.750 (0.440) 0.771 (0.423)
B-TO-G Qualitative variable; 1if the survey’s major aim was to evaluate from Bad to GES; 0 otherwise 0.937 (0.246) 0.957 (0.204)
G-TO-VG Qualitative variable; 1if the survey’s major aim was to evaluate from GES to Very Good; 0 otherwise 0.750 (0.440) 0.814 (0.392)
EUROZONE Qualitative variable; 1if the survey was carried out in a Eurozone country; 0 otherwise 0.437 (0.504) 0.357 (0.483)
SOUTH Qualitative variable; 1if the survey was carried out in Italy, Spain, Portugal or Greece; 0 otherwise 0.406 (0.499) 0.386 (0.490)
F-CRISIS Qualitative variable; 1if the survey was carried after the start of the 2008 financial crisis; 0 otherwise 0.562 (0.504) 0.557 (0.500)
WTP-TYPE Qualitative variable; 1if the recorded WTP measure is average and not marginal; 0 otherwise 0.969 (0.177) 0.943 (0.234)
VEHICLE Qualitative variable; 1if the survey’s vehicle was a type of increase in utilities bill; 0 otherwise 0.500 (0.501) 0.401 (0.461)
AREA The area of the case study recorded by the study in Km2 5,103.71 (15,733.4) 12,805.31 (22,784.37)
RECREATION Qualitative variable; 1if the survey evaluated recreational benefits; 0 otherwise 0.281 (0.457) 0.228 (0.423)
Note: Numbers in parentheses are standard deviations.
Table 2. Estimation results for the best set model including all study design variables.
Dependent=WTP Dependent=ln(WTP)
Constant 17.786 (44.657) Constant 1.155 (1.153)
Income 0.522 (0.760) Ln(Income) 0.646 (0.293)**
WFD 25.167 (41.428) WFD 0.381 (0.646)
NON-USERS -32.570 (39.080) NON-USERS -0.464 (0.609)
G-TO-VG 33.459 (33.037) G-TO-VG 0.814 (0.550)
VEHICLE 38.310 34.77468 VEHICLE 0.704 (0.545)
EUROZONE -32.6244 31.40867 EUROZONE -0.653 (0.508)
F(6,6)= 1.35 F(6,6)= 3.33
Prob>F= 0.3628 Prob>F= 0.0843*
Dependent=WTP Dependent=ln(WTP)
Constant 3.762 (23.188) Constant -0.080 (0.771)
GDP/CA 1.345 (0.449)*** Ln(GDP/CA) 1.057 (0.194)***
WFD 17.2902 (11.983) WFD 0.500 (0.239)**
NON-USERS -12.047 (12.695) NON-USERS -0.270 (0.244)
G-TO-VG -0.579 (13.249) G-TO-VG 0.065 (0.268)
VEHICLE 14.762 (14.501) VEHICLE 0.338 (0.278)
EUROZONE -1.381 (11.458) EUROZONE 0.102 (0.230)
F(6,25)= 1.84 F(6,25)= 5.70
Prob>F= 0.1322 Prob>F= 0.002
Note: Numbers in parentheses are standard errors. One, two and three asterisks stand for significance at the 10%, 5% and 1% respectively.
Appendix A Table 3. Meta-regression estimation results of log WTP as dependent and alternative log income variables as independent: Log to log functional form.
Best Set Average Relevant Set
Log-Log Models Coefficient s.e. t Coefficient s.e. t
Income 0.785** 0.330 2.38 0.837** 0.344 2.43
WFD -0.248 0.572 -0.43 -0.179 0.597 -0.30
Constant 1.542 1.143 1.35 1.378 1.194 1.15
F(2,10) 3.03 - - 3.08 - -
Adjusted R² 26.28% - - 26.34% - -
GDP/CA 1.031*** 0.184 5.59 1.036*** 0.176 5.87
WFD 0.580** 0.215 2.69 0.409* 0.205 1.99
Constant -0.099 0.624 -0.16 0.098 0.597 0.17
F(2,29) 16.73% - - 17.49 - -
Adjusted R² 51.68% - - 52.75% - -
GDP/CA-PPP 1.646*** 0.381 4.32 1.651*** 0.367 4.49
WFD 0.736*** 0.257 2.87 0.567** 0.247 2.29
Constant -2.224* 1.285 -1.73 -2.028 1.239 -1.64
F(2,29) 10.17 - - 10.27 - -
Adjusted R² 37.93% - - 38.18% - -
HINC/CA 1.057*** 0.170 6.21 1.061*** 0.162 6.55
WFD 0.529* 0.201 2.63 0.355* 0.191 1.86
Constant 0.301 0.502 0.60 0.503 0.478 1.05
F(2,29) 57.53 - - 21.75 - -
Adjusted R² 20.52% - - 58.78% - -
HINC/CA-PPP 1.656 0.276 5.99 1.662*** 0.264 6.30 Note: one, two and three WFD 0.558 0.206 2.70 0.386* 0.197 1.96 asterisks stand for Constant -1.365 0.789 -1.73 -1.166 0.753 -1.55 significance at the 10%, 5% F(2,29) 19.10 - - 20.07 - - and 1% respectively. Adjusted R² 55.12% - - 56.27% - -
HDINC/CA 1.074*** 0.235 4.56 1.089*** 0.225 4.84
WFD 0.488** 0.234 2.08 0.316 0.224 1.41
Constant 0.369 0.660 0.56 0.543 0.631 0.86
F(2,29) 11.31 - - 11.92 - -
Adjusted R² 41.38% - - 42.55% - -
HDINC/CA-PPP 1.397*** 0.274 5.09 1.352*** 0.272 4.96
WFD 0.594** 0.227 2.62 0.410* 0.225 1.82
Constant -0.491 0.758 -0.65 -0.155 0.753 -0.21
F(2,29) 13.94 - - 21.48 - -
Adjusted R² 46.94% - - 43.65% - - Table 4. Panel regression estimation results of log WTP as dependent and alternative log income variables as independent: Log to log functional form.
OLS clustered Random effects Weighted random effects
Log-Log models Coeff. s.e. t Coeff. s.e. t Coeff. s.e. t
Income 0.597* 0.281 2.12 0.785** 0.347 2.26 0.792*** 0.189 4.17
WFD -0.195 0.330 -0.59 -0.051 0.451 -0.11 -0.224 0.421 -0.53
Constant 2.121 0.866 2.45 1.401 1.118 1.25 1.350 0.803 1.68
F or Wald χ² 2.25 - - 5.10 - - -29.971 - -
Adjusted R² or R² overall 0.153 - - 0.151 - - 0.750 - -
GDP/CA 0.961*** 0.184 5.22 0.984*** 0.173 5.69 1.042*** 0.138 7.55
WFD 0.358* 0.181 1.97 0.377** 0.192 1.96 0.368* 0.192 1.91
Constant 0.296 0.598 0.50 0.242 0.579 0.42 0.061 0.479 0.13
F or Wald χ² 13.67 - - 33.76 - - -52.557 - -
Adjusted R² or R² overall 0.422 - - 0.421 - - 0.643 - -
GDP/CA-PPP 1.484*** 0.437 3.39 0.152*** 0.347 4.39 1.305*** 0.310 4.21
WFD 0.549** 0.246 2.23 0.490** 0.219 0.219 0.491** 0.237 2.07
Constant -1.522 1.458 -1.04 -1.603 1.157 -1.38 -0.877 1.054 -0.83
F or Wald χ² 5.78 - - 20.29 - - -56.773 - -
Adjusted R² or R² overall 0.361 - - 0.359 - - 0.612 - -
HINC/CA 0.998*** 0.151 6.61 1.023*** 0.162 6.30 1.08*** 0.133 8.12
WFD 0.295 0.180 1.64 0.341* 0.182 1.87 0.348** 0.169 2.06
Constant 0.638 0.408 1.56 0.578 0.474 1.22 0.482 0.374 1.29
F or Wald χ² 21.82 - - 41.26 - - -49.997 - -
Adjusted R² or R² overall 0.435 - - 0.435 - - 0.651 - -
HINC/CA-PPP 1.600*** 0.308 5.19 1.613*** 0.263 6.13 1.682*** 0.277 6.07
WFD 0.376* 0.187 2.00 0.378** 0.185 2.04 0.356* 0.187 1.91
Constant -1.050 0.826 -1.27 -1.036 0.746 -1.43 -1.233 0.769 -1.60
F or Wald χ² 13.60 - - 39.10 - - -51.545 - -
Adjusted R² or R² overall 0.449 - - 0.449 - - 0.582 - -
HDINC/CA 1.072*** 0.235 4.56 1.077*** 0.224 4.77 1.185*** 0.211 5.59
WFD 0.257 0.191 1.35 0.304 0.205 1.48 0.351* 0.209 1.68
Constant 0.552 0.614 0.90 0.562 0.622 0.90 0.330 0.526 0.63
F or Wald χ² 10.74 - - 23.88 - - -54.531 - -
Adjusted R² or R² overall 0.336 - - 0.335 - - 0.669 - -
HDINC/CA-PPP 1.267*** 0.281 4.51 1.304*** 0.266 4.90 1.266*** 0.241 5.25 WFD 0.388* 0.214 1.81 0.398* 0.205 1.94 0.434** 0.171 2.54
Constant 0.225 0.735 0.03 -0.055 0.728 -0.08 0.413 0.637 0.06
F or Wald χ² 10.24 - - 25.15 - - -54.503 - -
Adjusted R² or R² overall 0.349 - - 0.349 - - 0.612 - -
Note: one, two and three asterisks stand for significance at the 10%, 5% and 1% respectively.
PANEL OLS not clustered OLS clustered Random Effects not weighted Random Effects weighted
Model type Lin-Lin Log-Lin Log-Log Lin-Lin Log-Lin Log-Log Lin-Lin Log-Lin Log-Log Lin-Lin Log-Lin Log-Log
Y from .318 .014 .597** .318 .014 .597* .829 .024 .785** .553** .022*** .792*** databases (.551) (.009) (.257) (.602) (.010) (.281) (.926) (.016) (.347) (.222) (.005) (.189)
WFD dummy -5.418 -.155 -.195 -5.41 -.155 -.195 .180 -.030 -.051 1.31 -.259 -.224 1 (34.553) (.598) (572) (29.673) (.423) (.330) (22.105) (.462) (.451) (13.823) (.382) (.421)
GDP 1.557*** .038*** .961*** 1.557*** .038*** .961*** .1577*** .040*** .984*** 1.431*** .006*** 1.042*** (.339) (006) (.140) (.497) (.009) (.184) (.452) (.008) (.173) (.222) (.006) (.138)
WFD dummy 18.994*** .382** .358** 18.994* .382** .358* 17.287* .369* .377** 6.815 .402** .368* 2 (9.248) (.174) (.165) (10.417) (.187) (.181) (10.500) (.208) (.192) (5.751) (.176) (.192)
GDP (PPP) 2.258*** .054*** 1.484*** 2.258** .054** 1.484*** 2.146*** .056*** .1523*** 2.123*** .045*** 1.305*** (.548) (.010) (.246) (1.078) (.021) (.437) (.779) (.015) (.347) (.262) (.009) (.310)
WFD dummy 23.901** .497** .549*** 23.901* .497* .549** 19.304* .430* .490** 20.244*** .351 .491** 3 (9.744) (.189) (.181) (13.240) (.253) (.246) (11.211) (.230) (.219) (4.832) (.220) (.237)
Balance of Y 2.884*** .071*** .998*** 2.884*** .071*** .998*** 2.920*** .074*** 1.023*** 3.071*** .071*** 1.08*** (.568) (.010) (.141) (.728) (.013) (.151) (.715) (.013) (.162) (.600) (.010) (.133)
WFD dummy 18.047** .359*** .295* 18.047* .359** .295 17.489* .372* .341* 14.622* .453*** .348** 4 (8.976) (.166) (.162) (10.309) (.181) (.180) (10.068) (.191) (.182) (8.562) (.176) (.169)
Balance of 4.330*** .106*** 1.600*** 4.330*** .106*** 1.600*** 4.218*** .109*** 1.613*** 5.271*** .113*** 1.682*** Y(PPP) (.858) (.016) (.220) (1.507) (.027) (.308) (1.109) (.020) (.263) (.970) (.028) (.277)
WFD 20.404** .415** .376** 20.404* .415** .376* 17.933* .393** .378** 19.270** .313* .356* dummy5 (9.059) (.169) (.161) (11.277) (.198) (.187) (10.294) (.197) (.185) (8.015) (.249) (.187)
Disposable Y 4.367*** .109*** .1266*** 4.367** .109*** 1.267*** 4.220*** .113*** 1.304*** 9.748*** .127*** 1.266*** (19.608) (.019) (.216) (1.667) (.031) (.281) (1.336) (.025) (.266) (.777) (.015) (.241)
WFD 19.608** .399** .388** 19.608 .399* .388* 17.507 .388* .398* 30.655*** .303** .434** dummy6 (9.478) (.180) (.176) (11.832) (.221) (.214) (10.756) (.212) (.205) (5.823) (.143) (.171)
Disposable Y 3.390*** .086*** 1.072*** 3.490*** .086*** 1.072*** 3.460*** .088*** .1.07*** 3.664*** .092*** 1.185*** (PPP) (.790) (.015) (.188) (1.017) (.019) (.235) (.983) (.018) (.224) (.728) (.014) (.211)
WFD 16.167* .312* .257 16.167 .312 .257 15.978 .331 .304 16.990* .420** .351* dummy7 (9.262) (.176) (.176) (10.369) (.189) (.191) (10.432) (.206) (.205) (9.265) (.198) (.209)
Table 6: Model Diagnostics for the panel structure
OLS not clustered OLS clustered Random Effects not weighted Random Effects weighted
Model Lin-Lin (R², Log-Lin(R², Log-Log(R², Lin-Lin(R², Log-Lin R², Log-Log R², Lin-Lin (Wald Log-Lin(Wald Log-Log(Wald Lin-Lin (log- Log-Lin(log- Log-Log(log- type Root MSE, Root MSE, Root MSE, Root MSE, Root MSE, Root MSE, χ²,R² overall, χ²,R² overall χ²,R² overall likelihood, likelihood, likelihood, F(2,67) F(2,67) F(2,67) F(2,31) F(2,31) F(2,31) rho) rho) rho) ICC) ICC) ICC)
Y from -5%, 47.349, 1%, .82075, 9%, .784, 1%, 47.349, 7%, .82075, 15%, .784, 0.80, 1%, 2.83, 7%, 5.10, 15%, -157.239 -30.72727 -29.97139 databases 0.18 1.18 2.73 0.14) 0.89 2.25 .842 .771 .719
GDP 23%, 34%, 40%, 61199, 22%, 34.025, 36%, .64295, 42%, 61199, 13.41, 25%, 23.22, 36%, 33.76, 42%, -323.76133 -54.698351 -52.556786 34.025, .64295, 24.44 4.90 8.48 13.67 .726 .633 .566 11.50 19.00
GDP (PPP) 19%, 27%, 34%, .64353, 22%, 34.847, 29%, .67631, 36%, .64353, 8.62, 21%, 14.04, 29%, 20.29, 36%, -326.6671 -57.878862 -56.773277 34.847, 9.41 .67631, 18.90 2.24 3.36 5.78 .753 .688 .646 13.95
Balance of 27%, 40%, 42%, .60451, 29%, 33.163, 42%, .61383, 43%, .60451, 18.11, 29%, 34.63, 41%, 41.26, 43%, -324.46294 -52.53702 -49.99665 Y 33.163, .61383, 25.89 7.86 15.57 21.82 .699 56% .520 13.87 24.09
Balance of 27%, 38%, 43%, .59745, 29%, 33.216, 40%, .62103, 45%, .59745, 15.81, 29%, 30.46, 40%, 39.10, 45%, -325.35781 -53.691116 -51.54499 Y (PPP) 33.216, .62103, 27.30 4.20 7.99 13.60 .713 .586 .535 13.72 22.77
Disposable 20%, 30%, .6622, 33%, .6496, 22%, 34.722, 32%, .6622, 34%, .6496, 11.14, 22%, 21.15, 32%, 25.15, 35%, -330.0671 -55.542815 -54.503494 Y 34.722, 9.72 15.99 17.93 3.47 6,41 10.25, .739 .643 .617
Disposable 22%, 31%, .6547, 31%, .65569, 24%, 34.335, 33%, .6547, 33%, .65569, 13.63, 24%, 23.40, 33%, 23.88, 33%, -325.40666 -54.963419 -54.531343 Y (PPP) 34.335, 17.13 16.98 5.88 10.45 10.74 .726 .632 .629 10.70
N 70 70 70 70 70 70 70 70 70 70 70
Table 7: For the cross-section structure of the data, results from the three functional forms
Model type Lin-Lin Log-Lin Log-Log Lin-Lin Log-Lin Log-Log
Y from .747 .0232 .785** .882 .026 .837** Table 8: For the cross-section structure of the data, results from the Best Estimate set and the (.776) (.0163) (.330) (.906) (.016) (.344) databases Average Estimate set, in the three functional forms
WFD dummy 1 -15.211 -.283 -.248 -9.451 -.213 -.179 (.31.794) (.656) (.572) (26.696) (.678) (.597) Cross-Section WTP Best-Estimate WTP Average-Estimate GDP 1.256*** .042*** 1.031*** 1.553*** .043*** 1.036*** (.420) (.009) (.184) (.444) (.008) (.176) Model type Lin-Lin Log- Log- Lin-Lin(R², Log-Lin R², Log-Log R², (adj.R², Lin(adj. R², Log(adj. R², F(2,31) F(2,31) F(2,31) WFD dummy 2 19.446* .580** .580** 17.212 .415* .409* F(2,29) F(2,29) F(2,29) (10.771) (.241) (.215) (11.287) (.229) (.205) Y from -7.30%, 2.68%, 26.28%, -8.58%, 4.65%, 26.34%, GDP (PPP) 1.773** .061*** 1.646*** 2.168** .061*** 1.651*** databases 0.61 1.17 3.03 0.53 1.29 3.08 (.735) (.0170) (.381) (.792) (.016) (.367)
22.32%, 39.38%, 51.68%, 28.68% 41.81%, 52.75%, 22.256* .681** .736*** 20.623 .512* .567** GDP WFD dummy 3 5.06 10.97 16.73 6.32 11.77 17.49 (11.825) (.273) (.257) (12.763) (.264) (.247)
13.80%, 29.32%, 37.93%, 16.93%, 28.77%, 38.18%, 2.302*** .076*** 1.057*** 2.818*** .078*** 1.061*** GDP (PPP) Balance of Y 3.45 7.27 10.17 3.93 7.09 10.27 (.661) (.014) (.170) (.691) (.013) (.162)
29.62%, 49.44%, 57.53%, 36.96% 53.11%, 58.78%, 19.117* .563** .529** 16.547 .398* .355* Balance of Y WFD dummy 4 6.70 15.48 20.52 8.52 17.79 21.75 (10.216) (.219) (.201) (10.673) (.204) (.191)
25.90%, 46.79%, 55.12%, 31.51%, 49.35%, 56.27%, .343*** .113*** 1.656*** 4.058*** .115*** 1.662*** Balance of Y Balance of 5.93 14.10 19.10 7.20 15.16 20.07 (1.025) (.022) (.276) (1.085) (.020) (.264) (PPP) Y(PPP)
20.74%, 40.24%, 46.94%, 22.06%, 38,91%, 43.65%, 19.617* .582** .558** 17.142 .416* .386* Disposable Y WFD dummy5 4.55 10.98 13.94 4.86 10.48 12.48 (10.491) (.226) (.206) (11.101) (.213) (.197)
22.68%, 39.67%, 41.38%, 28.39%, 42.12%, 42.55%, 3.415*** .119*** 1.397*** 4.020*** .118*** 1.352*** Disposable Y Disposable Y 4.93 10.66 11.31 6.17 11.72 11.92 (1.211) (.026) (.274) (1.316) (.026) (.272) (PPP)
WFD dummy6 19.421* .586** 594** 16.583 .408* .410* N 32 32 32 32 32 32 (10.952) (.240) (.227) (11.877) (.234) (.225)
Disposable Y 2.655*** .089*** 1.074*** 3.293*** .092*** 1.089*** (.901) (.020) (.235) (.953) (.019) (.225) (PPP) WFD dummy7 17.701 .520** .488** 14.954 .352 .316 (10.677) (.239) (234) (11.288) (.226) (.224)
Study ID
Bateman et al. (2006a) Hanley et al. (2006) -Wear River Brouwer (2004) Mourato et al. (2005) Brouwer (2004a) Bateman et al. (2006b) Birol and Cox (2007) Czajkowski and Ščasný (2010) - Czech Republic Czajkowski and Ščasný (2010) - Poland Del Saz-Salazar et al. (2009) - Alcoy Del Salazar et al. (2009) - Gandia Martin-Ortega and Berbel (2010) Hanley et al. (2006) - Brothock Hanley et al. (2006) - Motray Brouwer et al. (2016) - Romania Brouwer et al. (2016) - Hungary Birol et al. (2009) Bliem and Getzner (2008) Jones et al. (2008) Hasler et al. (2009) Stithou et al. (2012) Poirier and Fleureut (2010) Ahtiainen et al. (2012) - Latvia Ahtiainen et al. (2012) - Poland Ahtiainen et al. (2012) - Lithuania Ahtiainen et al. (2012) - Germany Ahtiainen et al. (2012) - Estonia Ahtiainen et al. (2012) - Finland Ahtiainen et al. (2012) - Denmark Ahtiainen et al. (2012) - Sweden Doherty et al. (2014) Ramajo-Hernandez et al. (2012) Overall (I-squared = 100.0%, p = 0.000)
0 20 41.8 60 80 100 120 140
Figure 1. Forest plot of random effects estimates and of overall mean WTP
Appendix B Table A. Studies included in the meta-analysis
Case Study – Year of Publication Year of WTP Estimate Geographic WFD Payment Vehicle Test for study Reference Relevanceb Representativeness of Best Set Average All Re- Area Samplec a Set levant (NUTS)
Bateman et al. (2006a) 1999 27.719 27.719 1 UKG3 0 Council tax 0
Hanley et al. (2006) -Wear River 2001 20.440 20.301 6 UKC1 1 Water bill 1
Brouwer (2004) 2002 75.000 75.000 1 NL12 1 Council tax 0
Mourato et al. (2005) 2002 113.780 112.410 3 UKI 0 Water bill 1
Brouwer (2004a) 2003 105.000 105.000 1 NL 1 Council tax 0
Bateman et al. (2006b) 2003 117.640 117.640 1 UKH1 1 Council tax 1
Birol and Cox (2007) 2004 20.879 28.225 4 UKK 1 One-off payment 1
Czajkowski and Ščasný (2010) – Poland 2005 7.409 8.590 2 PL41 1 Increase in sewage charge 1
Czajkowski and Ščasný (2010) - Czech 2005 18.020 22.040 2 CZ05 1 Increase in sewage charge 1 Republic
Del Saz-Salazar et al. (2009) - Alcoy 2006 27.400 27.400 1 ES53 1 Water bill 0
Del Salazar et al. (2009) - Gandia 2006 33.599 33.599 1 ES52 1 Water bill 0
Martin-Ortega and Berbel (2010) 2006 59.160 49.046 6 ES61 1 Water bill 1 Hanley et al. (2006) - Brothock 2006 85.117 105.897 3 UKM3 1 Water bill 1
Hanley et al. (2006) - Motray 2006 123.250 149.826 3 UKM6 1 Water bill 1
Brouwer et al. (2016) - Romania 2007 6.530 6.530 1 RO 1 Water bill 1
Brouwer et al. (2016) - Hungary 2007 16.200 16.200 1 HU 1 Water bill 1
Birol et al. (2009) 2007 21.562 21.562 1 PL22 1 Water bill 1
Bliem and Getzner (2008) 2007 27.959 27.966 3 AT13 1 Water bill 1
Jones et al. (2008) 2007 16.840 16.840 1 EL41 0 Water bill 0
Hasler et al. (2009) 2008 39.180 51.145 4 DK03 1 Water bill 0
Stithou et al. (2012) 2008 27.60 32.056 3 IE01 1 Council tax 0
Poirier and Fleureut (2010) 2010 52.099 52.099 1 FR25 1 Council tax 1
Ahtiainen et al. (2012) - Latvia 2010 4.909 7.790 2 LV 0 Voluntary contribution 0
Ahtiainen et al. (2012) - Poland 2011 11.150 15.200 2 PL 0 Voluntary contribution 0
Ahtiainen et al. (2012) - Lithuania 2011 12.420 13.035 2 LT 0 Voluntary contribution 0
Ahtiainen et al. (2012) - Germany 2011 18.240 25.815 2 DE 0 Voluntary contribution 0
Ahtiainen et al. (2012) - Estonia 2011 19.950 27.830 2 ES 0 Voluntary contribution 0
Ahtiainen et al. (2012) - Finland 2011 32.400 43.299 2 FI 0 Voluntary contribution 0
Ahtiainen et al. (2012) - Denmark 2011 35.150 49.415 2 DK 0 Voluntary contribution 0 Ahtiainen et al. (2012) - Sweden 2011 63.060 76.889 2 SE 0 Voluntary contribution 0
Doherty et al. (2014) 2012 25.000 20.000 3 IE0 1 Council tax 1
Ramajo-Hernandez et al. (2012) 2012 66.120 61.020 2 ES61 1 Water bill 0
Notes: a) The number of estimates included in the panel specification is reported. b) 1 if the WFD was explicitly referenced by the study’s design; 0 otherwise c) 1 if the study’s sample has been tested for representativeness; 0 otherwise
Chapter Six: A how-to-do Cost-Benefit analysis in the context of the Water Framework Directive manual: combining results from primary and non-primary sources with a sensitivity analysis
This chapter is partly based on the paper “False Positive and False Negative Errors in the Design and Implementation of Agri-environmental Policies: A Case Study on Water Quality and Agricultural Nutrients” by Demetrios Psaltopoulos, Andrew Wade, Dimitris Skuras, Martin Kernan, Emmanouil Tyllianakis and Martin Erlandsson published in the Science of the Total Environment, November (2016).
1. Introduction
This purpose of this chapter is to go a step forwards from the previous chapters for what a cost-benefit analysis should look like, within the WFD line of mandates. We present a comprehensive tool for decision-making in terms of policy-making and the potential lapses of policy design, in this case of agri-environmental measures.
In the European Union, agri-environment measures (AEM) constitute one of the main types of policy response for meeting society's demand for environmental outcomes provided by agriculture. AEM in the Common Agricultural Policy (CAP) were introduced as an option for EU Member States in the late 1980s and made a compulsory component of Rural Development Programmes (RDPs) since the 1992 reform of the CAP. AEM in the EU are made up of two components; the first follows the Polluter-Pays-Principle and ensures that farmers comply with mandatory national and European environmental standards which form part of the cross-compliance regime at their own costs (European Commission, 2003). The second follows the Provider-gets-Principle and is designed to remunerate farmers for voluntary environmental commitments going beyond legal requirements of the Polluter- Pays-Principle. As such, the application of AEM is compulsory at the Member State level, but optional at the farmer level. Thus, the design of AEM is foreseen to meet public demand for environmental goods under the budgetary constraint of payments to farmers that aim to cover the costs incurred and income forgone as resulting from voluntary environmental commitments. The involvement of farmers is usually medium to long-term with a minimum participation of five years. Over the years, agri-environment policy has emerged as one of the most important elements of the CAP in terms of its budgetary size and the proportion of participating farmers and farmland in the EU. Indicatively, in the 2007-2013 programming period the main AEM (measure 214 of the RDPs) had a total spending of €36.6 billion (27.9 per cent of total public expenditure on Pillar 2) and addressed nearly 43 million hectares or almost a quarter of the EU’s utilized agricultural area.
Following the assessment of the AEM that were proposed to be undertaken in the Louros’ catchment, we conduct a cost benefit analysis using the findings of chapter 5 to compare them with the primary findings of the Louros catchment in chapter 4 and see the range of changes due to changes in discount rates and economic variables.
Cost-Benefit Analysis (CBA) is a tool to perform an economic evaluation of alternative management and policy options by comparing the predicted beneficial against the expected adverse effects of that action, both assessed against the same reference situation. We can distinguish the 6 essential steps of a CBA (Eigenraam, 2000) below:
1. A good definition of the scope of the analysis. 2. A clear definition of the project to be evaluated. definition of the baseline scenario definition of the objectives (qualitative/quantitative) definition-selection of the package of measures 3. Tools and data to assess the costs of the measures. 4. Methodologies for the assessment of costs and benefit 5. Tools and data to assess the benefits. 6. Comparison of costs and measures in a CBA.
The general steps necessary to perform a CBA of the Water Framework Directive are presented in Figure 1. All CBAs start with a problem definition and the determination of the baseline scenario. Complications arise from the requirements stipulated in the WFD. The baseline scenario is the business as usual case or the position where we would like to be in the absence of the Directive. This means that the likely developments that take place between now and 2015 are included. Likely developments include exogenous developments (for example, change in industrial emissions due to economic growth), the impact of “old” water directives such as the Nitrates and Urban Wastewater Directives, and the impact of national policies which are in the pipeline. Based on the gap between the baseline scenario and the WFD requirements, a Program of Measures is designed. Article 11 and Annex III of the WFD require that the most cost-effective selection of measures is implemented, which means that economic efficiency is an important basis to select measures. A next step in the selection process consists of the assessment of the impacts of this program. Cost-benefit analysis should aim to identify, assess and monetise all impacts: · Costs should refer to the total economic costs which is a measure of the welfare losses due to the implementation of policies or projects. It includes the direct, financial costs that relate to expenditures for the additional investments, operational and administration costs of additional measures. In addition, there may be direct effects which are reflected in expenditures, such as foregone opportunities and furthermore there may be indirect effects in different sectors of the economy (see Section 2.1 and 2.2).
Benefits are the welfare gains realised by implementing these measures. They can take the form of market effects (e.g. avoided treatment costs, returns in commercial fisheries) or of non-market effects (e.g. improved amenity, informal recreation) through a better provision of “goods and services” by water bodies reaching a better status (for example recreation, bathing) or through improvements in use of scarce resources. Cost-benefit analysis requires the aggregation of costs and benefits over time: · The assessment of costs and benefits needs to be defined over a particular time period. Achievement of good status is designated as the year 2015. However, many actions need to be taken before that to achieve this target. Therefore the base year to start the comparison needs come early in the process.
· the time horizon needs to be long enough: typical 40 to 100 years4.
· both costs and benefits are aggregated over time, using discount rates.
After the estimation of both costs and benefits, they are compared systematically to assess the net benefit of implementing the planned measures. The robustness of these conclusions are tested by conducting an uncertainty and sensitivity analysis.
4 Some environmental projects have a shorter time horizon (i.e. 15 years) in their appraisal.
Mineral fertilizers and livestock are the main sources of nutrients which, very often, are out of balance with land availability and crop needs. This imbalance creates a surplus of nutrients, some of which may be lost to water (nitrates and phosphates) and air (ammonia and nitrogen oxides). The agri-environment policy has an embedded “Nitrates” component in its mandatory part, i.e. the Nitrates Directive (EEC, 1991), and implements action programmes for controlling nutrients balance that are voluntary for farmers within the so called Nitrates Vulnerable Zones (NVZs) and through the national and regional RDPs. The Nitrates Directive is an important building block of the wider European environmental and nature conservation policy as it is directly connected to the Water Framework (WFD) and the Habitats and Birds Directives.
The WFD establishes a comprehensive, cross-border approach to water protection with the aim of achieving good status for European bodies of water by 2015. It introduced two important economic perspectives related to water protection from pollution or other environmentally undesired actions. First, it requires the cost-effectiveness of environmental programmes applied at river basin level. Second, it introduces the notion of disproportionality, i.e. the situation in which the cost of actions aiming at protecting or restoring water bodies to good environmental status are, by large, exceeding the expected benefits. Disproportionate costs have important policy implications because their presence allows the relevant authorities to extend deadlines for achieving good environmental status beyond 2015 and up to 2027 or, alternatively, to adopt less stringent environmental objectives. In more extreme situations, disproportionate costs allow authorities to designate Heavily Modified Water Bodies or consider new modifications that cause status deterioration but are of overriding public interest, and ‘sustainable’. The Habitats Directive protects almost 21,500 Sites of Community Importance (SCIs), covering 13% of the EU’s total surface area, many are wetlands such as lagoons, Mediterranean temporary ponds, active bogs, wet meadows and calcareous fens including wetland habitats and species such as amphibians, water plants and invertebrates, that are also listed as Ramsar Wetlands of International Importance. One of wetlands’ major threats is eutrophication from agricultural and livestock nutrients and pollution from agricultural, urban and industrial runoffs.
Taking into account the very significant amount of public funds devoted to agri-environment policy in the EU, the long term commitments undertaken by farmers and society’s demand for environmental outcomes from agriculture it is important to assure policy decision makers that such funds are directed to the areas in need and in a cost effective way. In decision-making, a false positive, known in statistics as Type I error, refers to the situation where the presence of a condition is assumed when in reality there is not. A false negative, known in statistics as Type II error, refers to the situation where no presence of a condition is assumed when in reality there is one. In the process of designing agri-environmental policies controlling for nutrients, policy decision makers may fall into false positives and false negatives. In other words, action programmes may address areas where the nutrients issue is not currently very important and is not likely to become such in the future (false positive) and fail to address areas where the nutrients issue is currently very important or may likely become so in the future (false negative).
The aim of this section is to propose an integrated decision making framework for designing and establishing AEMs targeting nutrient reduction. This decision making framework reduces the risk of committing false positives and wasting financial resources or the risk of committing false negatives and not protecting the environment.
2. Sources of false positives and negatives in the design of agri- environmental policy
Mandatory and voluntary AEM have as aims, among others, to reduce nutrient concentrations in downstream rivers, lakes and wetlands. Most frequently, such measures directly target nutrient deposition (inputs) to land by setting maximum application rates. For example, the Nitrates Directive states that the amount of livestock manure applied on agricultural land each year, including that applied by animals themselves, should not exceed a maximum of 170 kg of nitrogen per hectare. Other measures attempt to manage nutrients on the field, by promoting favourable farm practices such as crop rotation systems, while others aim at restricting leaching of nutrients from the field, through (e.g.) the maintenance of buffer strips. The design and implementation of agri-environment action programmes for nutrient control is based on information about nutrient deposition from agricultural and livestock activity measured in kg per hectare and the concentration of nutrients in surface and groundwater measured in mg/L. This practice of setting policy targets presumes a direct relationship between nutrient deposition on the field and downstream nutrient concentration. As such, it fails to take account of the static abiotic environment (geology and soil) and the dynamics of human activity and climate change, as well as of the changing and fluctuating water supply. Thus, we should not presume a linear and static relationship between deposition and concentration on which to build solid and robust AEM.
Fig. 1 below attempts to sketch how false positives and false negatives may be generated in agri-environmental policy-making. The upper part of the diagram provides a coarse picture of the nutrient deposition- leaching-transportation-concentration process and how this process is influenced by abiotic, biotic, human activity and climate change factors. Under abiotic factors we refer to those physical processes pertinent to the geology, topography, soil physical and chemical properties. Under biotic factors, we refer to the whole range of sources that contribute to nutrient deposition such as land uses other than agriculture and animal activity other than livestock and/or grazing. Under human activity factors, we refer to agriculture and other activities contributing nutrients and including municipal and industrial sources coming from septic tanks or other devices of establishments that are not connected to municipal wastewater networks, animal wastes, food processing, etc. In addition, activities other than agriculture, may have an impact on the hydrology and especially on the quantity and frequency of water provided to water courses. Beyond irrigation and its corresponding drainage networks, examples include water extraction for municipal and industrial uses and sometimes small or large scale energy production from hydro electrical power plants. Finally, climate and especially temperature and runoff are important factors determining nutrient cycling and transport (Howarth et al., 2012).
The relationship between agricultural inputs and instream nutrient concentrations is not a simple one, for example because of a large groundwater store that may act as nutrient reservoir or because water flows may change, or plant uptake may increase, or land use may change (Jackson et al., 2008; Howden et al., 2010). Thus, the underlying relationship between nutrient deposition by agriculture and its impact on nutrient concentration in downstream water bodies may be important (yes-positive) or not (no-negative). However, without an integrated approach modelling the relationship between nutrient input and instream concentrations there will be uncertainty as to whether the policy can address the input-output relationship accurately, and therefore avoid the risk of false positives and false negatives. Table 1 provides an indicative list of false positive or negative decisions along with connotative reasons causing these deceptive decisions. A similar table may be generated if dynamic changes caused by land use and climate change are taken into account. In this context, dynamics may generate an agricultural pollution issue in areas that currently have not such an issue and vice versa.
In this work we focus on two broad areas within the policy design process which may contribute to the generation of false positives and false negatives:
• Appropriate baseline monitoring and modelling of the nutrient deposition- concentration function and the resultant baseline abatement function measured in terms of nutrient concentration in the water downstream and, • Forecasting and incorporation of changes resultant from human activity and climate change and the resultant dynamic abatement function again measured in terms of nutrient concentration in the water downstream
Taking into account the long-term horizon for implementing an agri-environment programme, policy design, and especially baseline modelling, should consider dynamic changes that may considerably alter the initial conditions that lead to the adoption or the rejection of an agri-environment programme in a specific area. For example, within a seven year agri-environment planning horizon, several changes may
occur in land use, in agricultural production or/and even climatic conditions. Land use changes may be instigated by agricultural policy changes such as the CAP, which may lead to the abandonment of agricultural production or to the drastic change in the adoption of cultivations with different nutrient applications (Barbayiannis et al., 2011). One vivid example is the decoupling of Pillar 1 subsidies, which in some EU areas, has induced the abandonment of several cultivations or the shift to other crops, including nitrogen fixing legumes and the consequent reduction in nitrogen deposition. At the same time other, economy-wide developments, may affect (increase) agricultural input prices resulting to a rationalization and the consequent reduction of nutrient deposition.
In conventional policy design, targets are set on deposition, assuming that a proportional reduction will be achieved in the corresponding concentration of nutrients. The Nitrates Directive and several other EU, national and regional policies set such targets. This approach promotes “one-size fits all” policy and fails to take account of the aforementioned specificities of the environment and of human activity in the target area( s), that call for a case-specific and “tailor made” approach to agri-environmental target setting. In this work it is suggested that the baseline situation should be modelled according to an integrated framework accounting for dynamic changes. In this respect, we minimize the risk of false positives and false negatives. To this end we advocate a procedure that uses a dynamic, mass-balance water quality model to help explain the input (deposition) – output relationship and integrates science and socio-economic models to protect policy design from committing false positives or false negatives (Skuras et al., 2014). Fig. 2 depicts this approach in a sequence of policy design steps supported by science and social science methods and models. Once the non-compliance issue is recognized and defined (step 1) with the support of existing data and socio-economic public participation models, an integrated model of nutrient and sediment transportation within the catchment is proposed to be constructed (step 2). This step is supported by scientific models of nutrients and/or sediment transport that calibrate a baseline situation based on flow and hydrochemistry conditions of the catchment depicted by meteorological, soil-geological, flow, land use and water quality data. In step three, decision maker swill have the capacity to avoid false negatives and false positives. False positives are usually generated by failing to take into account the whole range of sources contributing nutrients to the watershed and overestimate the contribution and impact of agriculture. In this context, adopting a policy to control nitrogen deposition from agriculture will not have an effect. Potentially, false positives may be generated by situations in which high nutrient deposition fails to showup inwater nutrient concentrations for various reason including geology, e.g., extensive carstic phenomena that redirect nutrient rich water to neighbouring watersheds or to underground water reservoirs, soil conditions that favour high denitrification, deposition at river banks, etc.
In steps 4 and 5, mitigation measures are proposed and their effect is examined according to the calibrated baseline model. This will allow the examination of the simulated effectiveness of the mitigation measures and hence, the prevention of false positives, by adopting measures that will not be effective or the prevention of false negatives, by rejecting measures that will be effective (step 6). In step 7 the baseline condition and the mitigation measures are re-estimated and simulated against changing conditions including climate, land use and production. This will allow the prevention of false positives in the sense that a deposition-concentration situation that seems positive today may be most likely ameliorated in the near future due to changing conditions, without the need of mitigation measures and thus, adopting a programme would be less appropriate (step 8). The same step will allow the prevention of false negatives in the sense that a seemingly unrelated deposition-concentration situation today may be most likely aggravated in the near future due to changing conditions and adopting mitigation measures under an agri-environment programme would be appropriate.
2.1. The occurrence of a false positive
Taking into account only the supply of nutrients, and especially those from agricultural activity, it is estimated that the watershed accepts an amount of 2594 t of active N substances and 1578 t of active TP per annum from which agriculture is responsible for almost 1780 t of N and 1163 t of P for the major cultivations within the watershed. These amounts of active fertilizer substance alone are enough to trigger public concerns over agricultural activity in relation to the high nature value of the lagoon and its importance for European biodiversity, despite the fact that monitoring data were sparse and showed at most moderate nutrient concentrations and few signs of eutrophication. The simulated average and monthly concentrations for nitrates, ammonium and Total Phosphorous (TP) are shown in Fig. 4. Simulated nitrate concentrations near the estuary range between 0.8 and 1.0 mg N/L with an average at 0.9 mg N/L, while ammonium concentrations range from 0.04–0.13 mg N/L with an average of 0.08 mg N/L. TP concentration ranges from 0.02–0.11 mg TP/L with an average of 0.05 mg TP/L and SRP concentration from 0.01–0.11. mg SRP/L with an average of 0.04 mg SRP/L. Skoulikidis et al. (2006) have proposed a Nutrient Classification System (NCS) for small/medium sized rivers in Greece based on annual average concentrations from 36 sites throughout Greece. According to this system, the river is classified as of moderate quality in relation to nitrates (0.6–1.3 mg N/L) and ammonium (0.06–0.20 mg N/L) and of high quality in relation to TP (0.17– 0.22mg TP/L). Under other classifications, e.g., the nutrient quality classes in French and Italian rivers (Skoulikidis et al., 2006), the Louros river would be placed between a “Good” and “Moderate” class. At the same river and sub-catchments, macrophyte data (taxon name and abundance class) were collected and the IBR (Indice Biologique Macrophytique en Rivière - Macrophyte Biological Index for Rivers) was calculated by Manolaki et al. (2011) according to the methodology proposed by Haury et al. (2006). Of the 17 sites they studied, eight are characterized as having “High” ecological status, three as “Good”, four as “Moderate” and two as “Poor”. The best predictors for the decrease in IBMR values were salinity and water temperature, while SRP was also found to be correlated with IBMR but able to explain only 47% of the variability in IBMR values. The classification of the river's estuaries based on the aforementioned simulated results was re-confirmed in 2013 by the Management Plan drawn for Epirus' water resources. Thus, assuming that there is a direct positive relationship between agricultural activity and nutrient concentration would be a false positive, i.e., assuming a direct relation that does not exist. This further supported by the fact that nitrate concentrations tend to be highest in the upper reaches, which are not affected by agriculture, while the poor ecological status for macrophytes can obviously not be attributed to nutrient concentrations. There are several alternative explanations of why nutrient deposition rates do not really contribute to high nutrient concentrations downstream. As local stakeholders argue, due to cost minimization strategies and the rising price of fertilizers and energy, farmers take very good care of the time of fertilizer application, of the appropriate amount of fertilizer and of irrigation. This may contribute to a more balanced nutrient deposition and nutrient uptake by plants leaving less residual nutrients on the soil. In the framework of cost minimization there is also reduced and more precisely applied irrigation for reducing the cost of energy. Thus, higher uptake by plants also may be supported by longer water residence time in the soil brought about by more modern irrigation schemes (drop irrigation) that are gradually replacing sprinklers. This practice also reduces leaching and nutrient transportation.
Finally, there are well-documented physical and biological processes that may contribute to lower nitrogen levels despite higher deposition rates. Denitrification and nitrogen immobilization in excess of mineralization, at least temporarily when temperature is high and the concentration of soil C is high (Saggar et al., 2013). High spring and summer temperatures enhance aerobic respiration and denitrification while aerobic respiration further enhances denitrification by consuming oxygen, resulting in strong sensitivity of denitrification to temperature though substrate type and soil moisture may limit microbial processing (Butterbach-Bahl et al., 2013; Luo et al., 2013). Finally, sediment and thus nutrient transportation has been reduced in the area due to the extensive drainage and river bank stabilization works that have been undertaken throughout the watershed in the last 30 years.
2.2. The cost of a false positive
For each one of the four major cultivations in the watershed the cost of the mitigation measures was estimated. In order to proceed in our calculations we carried out two focus groups with stakeholders and elite interviews with agronomists in the area. Farmers' income from the different cultivations was estimated from the Standard Gross Margins derived by the FADN database for the region of Epirus where the Louros watershed is situated. From the FADN database we also calculated initial estimates of the cost of fertilization, and the cost of cultivating lentils, as well as the SGM of the lentil for fodder. Elite interviews with agronomists were utilised to estimate the loss in production due to reduced fertilization and irrigation. Consequently, stakeholders were presented with the initial estimates during a focus group with the aim to discuss and adapt initial estimates of the exact effects of reduced fertilization and irrigation on production and on farmer's income. In the context of this focus group, the transaction cost for submitting an environmental plan and subscribe to an agri-environmental programme were also estimated. The cost of the mitigation measures for each one of the four major cultivations was estimated as income forgone from reduced production plus transaction costs minus cost avoided from reduced fertilization and irrigation and fodder production. For citrus fruit plantations only income forgone was estimated, as there is no way to have land under set-aside.
The average cost estimates for abating nitrates and TP for the different cultivations in the area and the watershed as a whole, should the Mitigation 1 scheme be adopted by all farmers located within the hydrological boundaries of the watershed, is presented in Table 3. The upper part of Table 3 provides average cost estimates for fertilizer reduction per hectare (ha) and kilogram (kg) of active substance for the four major cultivations and the watershed as a whole. The cost per hectare varies significantly from437.2 €/ha for the less profitable cultivation of medic to 657.2 €/ha for the most profitable cultivation of cotton. The cost of abating one Kg of pure nitrogen ranges from 4.5 €/kg for corn to 12.5 €/kg for medic. For phosphorous, the cost of abatement per Kg is much higher than for nitrogen ranging from a high of almost 54 €kg for citrus fruit cultivation to a low of 5.4 €/kg for corn.
These estimates can be compared with past estimates of abatement costs for seven EU Member States carried out in the framework of a study estimating the ex post costs of implementing the Nitrates Directive in Europe (Kuik, 2006). In this study, cost estimates at 2004 prices range from a high 236 €/ha in the Netherlands to a low of 6 €/ha in the UK, which, however, refer to livestock and grasslands respectively. As concerns the cost per Kg this range from a low 0.4 €/kg for Croatia, then not a member State of the EU, to a high of 3.5 €/kg for the Netherlands. Taking into account that these estimates were derived with the Nitrates Directive in focus, they refer to grasslands and livestock which are not as intense activities as, for example, cotton. They also target a nitrogen concentration of 50 mg/l set up by the Nitrates Directive for sub-surface waters. In our study, the nutrient loads are already low and thus the marginal abatement cost is at its steeply rising part. An indirect way to measure abatement cost is through prohibitive standards, penalties and/or taxes. In the Netherlands, between 1998 and 2005, penalty-free thresholds were gradually reduced – for example, for nitrogen from300 kg/ha to 140 kg/ha for grassland farms (Goffe, 2013). Penalties, in the Netherlands were fixed at €0.68/kg for nitrogen and €2.60–€10.40/kg for phosphorous in 1998, and were increased to €2.53–€5.07/kg and €20.60/kg respectively (Goffe, 2013) while levy taxes in 2003 were set to 2.3 €/Kg for nitrogen and 9.1 €/Kg for phosphate (Söderholm and Christiernsson, 2008) which compare with the results of our study.
The focus of this study, however, is to reveal the high abatement cost when this is measured in terms of reduced nutrient concentration downstream. The cost estimate for nitrates is unreal at the unprecedented levels of just over 300 thousand euro per reduced microgram per litre €/[(μg/l)]. For phosphorous this is at 412,398 €/[(μg/l)]. So, a false positive decision to comply with the WFD and attain a “Good” status as concerns nutrient loads would be obviously unacceptable by any taxpayer in Europe. Which are the reasons for this case? First, the nutrient status is already at “Moderate” to “Good”, i.e., the nutrient concentration is already low compared to the 50 mg/l threshold of the Nitrates Directive.
Thus, the marginal cost to attain an even lower level of concentration is extremely high. Second, at this level of concentration, the simulations showed that even the withdrawal of 30% of the cultivated land will not reduce nitrate concentration by N0.02 mg/l and TP concentration by 0.01 mg/l. These are negligible achievements at a highly disproportional cost.
To summarize the discussion so far, it can be argued that the abatement cost of agri- environmental programmes aiming to manage nutrient loads should be measured as a change in nutrient concentration and not at levels of abated fertilizer. In other words, the targets of such agri-environmental programmes and policies should be set at nutrient concentration levels and not at quantities of abated substance either in mineral fertilizer or in manure and slurry. This can be attained if, during the design of agri-environmental programmes, the status quo (baseline), the impacts of the mitigation measures and the impacts from likely future changes are simulated. Then, false positives and false negatives can be avoided, the cost-effectiveness of mitigation measures can be assessed and an appropriate monitoring system can be set up.
2.3. A how-to-do guide on decision-making processes
The EU's agri-environmental policy is a response to the growing public concern over the environmental impacts of agriculture. As such, agri-environmental policy attempts to meet requirements from the WFD, the Nitrates Directive and the Habitats Directive, the cornerstone of environmental conservation in Europe. Agri-environmental policy has grown to a tremendous budget (€36.6 billion spent in the 2007–2013 programming period across the EU) and power by affecting almost a quarter of the EU's utilised agricultural area. This work concerned only with programmes managing nutrient loads in freshwaters and not with other forms of agri-environmental programmes. Results showed that, under public pressure and seemingly high rates of mineral fertilization, decision makers may falsely adopt an agri- environmental programme that may be both, ineffective in reducing nutrient loads and cost inefficient. Furthermore, they fail to take account of future changes that may inactivate the proposed mitigation measures, aggravate or reverse the baseline situation.
The present work suggests that the design of such agri-environment programmes should evolve to a thoroughly designed, interdisciplinary exercise integrating science and social- science models in a step-wise procedure. This process will ensure decision makers with the highest possible information from scientific sources and models and from local knowledge. This information can be used by appropriate simulation models to calibrate the baseline scenario. Once an appropriately calibrated model is derived, further scenarios simulating policy, land use and climate changes can be simulated. Based on these results the effectiveness and cost efficiency of the proposed actions and of envisaged changes can be assessed. As a result, decision makers will be able to grasp an ex-ante evaluation of the current situation and of the proposed mitigation actions, if needed. This will allow decision makers to monitor the current situation and respond by adopting new measures or adapting existing ones to the changing physical, social and policy environment.
Under this proposal, the cost of the design phase of an agri-environmental programme will increase. But, in view of the high cost of mitigation measures, such an increase in the design stage of the agri-environmental policy should be considered as an insurance against the commitment of very expensive false positive and false negative decisions.
Finally, in this work it is proposed that the targets of agri-environmental policy and consequently, the measurement of abatement cost should be done in terms of nutrient concentrations and loads in water and not in terms of physical quantities of abated substance in the field. This will provide the cost efficiency exercise with a wider perspective as concerns the sources of nutrients and abiotic, biotic and anthropogenic activities that contribute the nutrient loads. In turn, this will force agricultural policy decision makers to coordinate their actions with other environmental policy makers for achieving maximum results and avoiding internal contradictions.
3. Sensitivity analysis
The costs of a false positive or a false negative are presented above. Given the fact that economics are not always used as a first choice of measures when AEM are designed, it is essential to examine the extent of these costs. In this section we are going to use the results from Chapters 4 and 5 of this volume to examine in an actual case study the implications of false negative or positives occurring. In particular we are going to use the WTP results of the meta-analysis and compare them initially with the results from the case study in the Louros catchment and comment of potential policy implications and policy design mechanisms.
3.1. On benefits and time horizons
Key policy areas where climate change is considered include: “investment appraisal for long- term planning and infrastructure projects, regulatory and planning frameworks, contingency planning and long-term policy frameworks.” (DEFRA 2011)
Cost-Benefit Analysis (usually found in the literature as CBA but also as BCA; to underline the coming of benefits before costs) is the primary medium of policy evaluation. Ever since 2009 the UK and the US governments have issued legislative actions for CBA to be used it as a mandatory exercise for large-scale projects (Dietz and Hepburn 2010). Since 1997, the UK government established the UK Climate Impacts Programme (UKCIP) for assisting public and private organisations in assessing their vulnerability to climate change. UKCIP, together with the Department of the Environment, Food and rural Affairs in the UK (DEFRA) aims through its “Green Book” to provide a uniform set of information on climate change predictions and assessments that stem from the appropriate literature. The Green Book includes guidance on identifying and assessing climate change risks and provides a methodology for costing the impacts of climate change. This includes the establishment of a declining discount rate practice on environmental projects. This is further explained in Appendix A and on how the usage of a 3,5% as the starting point of a discount rate is chosen. DEFRA using the Ramsay formula suggests using a 3,5% discount rate for the first 30 years of a project that afterwards declines to 3% for the remainder to the time horizon. The Ramsay formula is described as the following:
r = ρ + μ.g (1)
When g increases then wealth for future generations and therefore consumption, grows. ρ, if it is 0 it implies that society (or the social planner designing the CBA) assumes that the utility of the present and future generations is considered to contribute equally to the social welfare of the present (Arrow et al 2014). The more ρ increases the more present consumption is discounted in favour of future consumption. There is a large ongoing discussion in the literature on the issue of discount rates and whether declining discount rates must be used and on the levels of the rates as well. It appears to be a consensus that if the time horizon for a project is to be considered as long (exceeding 100 years) then, using a declining discount rate is advisable. The levels of interest rates evoke much more debate from economists and policy makers.
Interest rates reflect the trade-off that is made between consumption in year t, over consumption of the good or benefit provided from the realization of a project in the present. This discount rate is called “consumption discount rate” and mainly it splits into two schools of thought (Arrow et al. 2014). The first allows for the incorporation of uncertainty in the rate of consumption growth when designing the CBA. Dasgupta et al. (1972) . This approach uses the Ramsay formula (see Annex A) and depending on shocks or unexpected events in the economy it can incorporate a declining discount rate. The second wave of thought follows Weitzman (1998, 2001, 2007) which suggests using a decreasing term structure of discount rates. In other words, future uncertainty is discounted in the present in the form of decreased discount rates, while using a constant exponential discounting.
The level of discount rates are also one of, if not the, most discussed issues in the literature of environmental economics and policy. Commenting in length on this is beyond the scope of this study and of this thesis. For referencing reasons only we present some of the existing views.
Using the Ramsey formula or its extensive form can help determining the level of an interest rate in order for a project to increase social welfare. Naturally, if the discounted net benefits are positive then the project is assumed to be beneficial for the public due to the increase of social welfare. Uncertainty about the future is perplexing the establishment of a fixed interest rate. Particularly when climate change is concerned, agreements such as the Kyoto Protocol and the latest Paris Agreement that make the case of a future shaped by climate change and therefore efforts should concentrate on minimizing the effects but admit some costs are going to be present. Even achieving limiting temperature increase to a 2˚ by the end of the century will entail large costs. Therefore, in view of the negative effects of climate change researchers such as Stern (2007) make the case for small discount rates (Stern suggested 1.4% as the appropriate level). Small discount rates echo the prediction of a decreased future consumption (future generation will consume less due to negative effects of climate change). Gollier (Chapter 8, 2013) argues that catastrophic events will make decrease the wealth of future generations and suggests negative discount rates for lengthy CBAs. Negative discount rates although sound theoretically present challenges mathematically (Arrow et al. 2014).
Countries such as the UK (Arrow et al 2013) and France (Lebegue 2005) use a declining social discount rate after 30 years in the life of an environmental project. Time horizon can go until 500 years where the interest rate goes from 4% to 2% at the end of the life of a project (Gollier et al 2008) or to 250 years starting from 4% declining to 2% in the end of the life of the project (Lebegue 2005). In general, discount rates under the Ramsay formula in its simple form or in the extended one are subject to fluctuations due to several factors. These are catastrophic risks (big economic recessions such as the Great Depression or mega- viruses such as the Great Plague), uncertainties regarding the consumption of future generations ( instead of a constant consumption rate of discount having more uncertainties for the future making more cautious the consumption in the present) and subjective uncertainty (when assuming that the extended Ramsay formula and the stochastic variable of the consumption-growth can be explained by econometric models and historical data) that challenges the assumptions of the extended Ramsay formula (Arrow et al 2014).
Climate-change models such as DICE (Nordhaus 2007, 2013) predict a decrease in r, starting from 6,5% in 2007 to 4,5% in 2095 due to the slowing-down of consumption over time.
Starting all the way from Dasgupta et al. (1972), the focus of CBA has been on projects that will not change relative prices (referring to prices of goods in question, such as the price of water that is affected from the building of the Aswan dam in Egypt as Dasgupta presents). These changes in other words would be marginal. This basis of this assertion is somehow debatable since many of the projects examined in the literature and in policy nowadays are subject to small or big relative price changes (Dietz and Hepburn 2010).
In this case we maintain that the environmental project to be undertaken in the Louros’ catchment would be truly a marginal project since water prices would not be changed as this was one of the requirements of the WFD. Especially in the case of Louros where water quality was relatively high and no relative changes in water prices are expected, additional economic burdens are considered to be minimal. Changes in the prices of agricultural products that are the results of mitigation policies such as set-aside practices and reductions in N and P-rich fertilizers are considered to be exogenous in this research work.
In terms of discount rates we take a few steps forward in considering time horizons and levels of discount rates from chapter 4. One of the aims of conducting a CBA in chapter 4 was to determine whether reaching GES by 2016 was a policy beneficial to the population of Preveza and Arta in Greece. The aim in this section is to examine the impacts of reaching GES as a policy reform and incorporate uncertainty concerning future consumption due to climate change. The various IPCC scenarios presented in Chapter 4 and 5, although having variations in the impact of climate change they assume increases in temperatures and in extreme weather phenomena, what might offset the estimated increases in productivity of certain agricultural sectors under certain scenarios (particularly B1 is projected to have the most positive impact on various agricultural projects in Western Greece, see chapter 4).
Additionally, we assume that the cumulative effects of the recession in the Greek economy by the deepening of the economic crisis will affect future consumption as the country’ debt repayments are projected to increase for the future generations. Increased burden for repaying current debt will affect future consumption and therefore we aim to examine the effects of decreasing interest rates and that of small interest rates as well.
4. Results
This section presents the results from conducting a CBA considering the various mitigation costs and the estimated benefits from achieving GES in the Louros’ catchment. The results are presented in Tables 2 and 3.
Firstly, we estimate the NPV of the net benefits from the four different mitigation measures using the initial estimated benefits (the case study of Louros’ catchment used a WTP estimate from Koundouri et al (2011) of 40.93 Euros that was multiplied with the total number of households in the area to estimate the benefits per year to achieve GES) but for a 30 year time-horizon. Following the guidelines of the Green Book (DEFRA 2011) we use a stable discount rate for a 30-year period. The results can be found in Table 2.
Achieving GES was the target for 2016 for the whole of Europe. The interest now lies on how enjoying GES for an extended period of time would affect the NPV of such projects. Therefore, we conduct a CBA keeping the same estimates for the four different mitigation measures but estimating different aggregate benefits for the Louros’ catchment. Agri- environmental measures do take a large period of time for the benefits to be realized and asserted in the environment. Thus it is safe to assume and consequently examine the benefits from moving to an even better water status after reaching GES. With the establishment of RDBs and the appropriate River Management plans, every water body cab have its water quality improved. Thus, examining a “Good to Very Good” scenario seems appropriate.
From the meta-analysis study in chapter 5 we use the WTP estimates for achieving GES, or otherwise the WTP for achieving water quality from “bad to good”. Other WTP estimates were estimated in some of the studies, estimating WTP to achieve “Very Good Status”. We conduct a meta-analysis on these studies using the same methodology as chapter 5, with WTP to achieve “Very Good Status” as our effect size. In order to conduct such an exercise, we use the studies available from chapter 5. As stated in chapter 5, some of the studies reported WTP estimates for achieving higher water quality status than GES. The results from the meta-analysis from 32 distinct studies across the EU and between 15 years on achieving GES yielded WTP estimates of 46.44 Euros [26.42, 66.45] for the random-effects and 34.57 Euros [34.46, 34.68] for the fixed-effects. Using the latest available data from HELSTAT on the number of households in the regional units of Arta and Preveza we have 25,967 and 21,853 for Arta and Preveza respectively (Census of population, 2011). Using a discount rate of 1,5% (usually used for environmental projects and long-run projects) and for a 6 year period (2010-2015) the results yielded an aggregate benefit of 11,150,948 Euros for these six years. 18 studies included WTP that were according with the mandates of the WFD for additive measures, above the requirements of achieving GES. Following the same process as chapter 5 we assume that the factors affecting WTP in every study are identical but with varying and unknown levels of impact. Therefore, using a Random Effects (RE) estimate than a Fixed Effects-one (FE) is more appropriate. Conducting a Random Effects meta-analysis, we estimated a 31.591 Euros per household per year, notably lower than the estimate for achieving GES as shown in chapter 4. We assume that 2016 (otherwise the 6th year from the start of the time horizon) is the year that GES is achieved and benefits are enjoyed. For improvements in the existing water quality status, changes to “Very Good Status” must take place, therefore we estimate different benefits from year 7, those achieved by a different WTP. This WTP is the 31,591 per person, per year. The mid-point of the time horizon is first 30 years and then 100 years. In the cases where the time horizon exceeded 30 years we also examined the results using a declining discount rate, from 3,5% up until year 30 and thereafter 3% until year 100. The results can be seen in table 4 in Appendix A.
For the 30-year time horizon we examine three different scenarios for the four mitigation measures. A baseline, where we use a stable 3,5% rate following the suggestions from the Green Book for the whole period (called here “baseline 3,5% “), a scenario where from year 7 and onwards we estimate different benefits using the R.E. WTP estimate (called here “Meta-analysis estimates 3,5%” ), and finally a scenario where we use a stable 1,5% interest rate like the one used in chapter 5 for the whole time horizon while considering different benefits from year 7 onwards (called here “Meta-analysis estimates, 1,5%” ) and which aligns with the social discount rates used in most environmental projects (see Appendix A, Table 4).
For the 100-year horizon we use the previous scenarios but we add a declining discount rate after year 30. In detail, we estimate 5 different scenarios, one with the stable 3,5% discount rate for the whole period (called “baseline 3,5% “), another with the benefits changing from year 7 onwards (called here “Meta-analysis estimates 3,5%” ), a scenario with the interest rate declining from 3,5% in year 30 to 3% until the end of the life of the project (called here “Baseline 3,5% declining” ), a scenario that assumes different benefits from year 7 when GES is achieved (called here “ Meta-analysis estimates, declining 3,5%” ) and finally, a scenario with a steady social discount rate of 1,5% with the benefits changing from year 7 onwards (called here “Meta-analysis estimates, 1,5%”). The results can be found in table 3.
CBA 30 years
Baseline 3,5% Meta-analysis Meta-analysis estimates, 1,5% estimates 3,5%
Mitigation 1 -84,037,696.21 € -77,777,149.69 € -59,564,062.21 €
Mitigation 2 -72,505,335.33 € -66,244,788.81 € -50,732,236.11 €
Mitigation 3 -07,702,224.02 € -101,441,677.49 € -77,687,063.78 €
Mitigation 4 -90,516,152.65 € -84,255,606.12 € -64,525,457.48 €
Table 2. NPV results from different scenarios for the four mitigation measures for a 30-year horizon
CBA 100 years
Baseline Meta- Baseline Meta- Meta-analysis 3,5% analysis 3,5% analysis estimates, meta estimates declining estimates, 1,5% 3,5% declining3,5%
Mitigation -606,963,046 -559,115,771 -552,337,607 -508,796,490. -299,519,094 € 1 € € € €
Mitigation -523,670,462 -475,823,187 -476,541,186 -433,000,068 -254,899,141 € 2 € € € €
Mitigation -777,880,319 -730,033,044 -707,872,673 -664,331,556 -391,079,715€ 3 € € € €
Mitigation -653,753,758 -605,906,483 -594,917,250 -551,376,133 -324,584,944€ 4 € € € €
Table 3. NPV results from different scenarios for the four mitigation measures for a 100-year horizon
5. Discussion
The results are not different from those of chapter 4. The use of the benefits for achieve very good status yield slightly lower results since the R.E. WTP estimate for achieving GES was higher. This maybe is due to the fact that most of the studies had lower estimates for achieving very good status but what is interesting is the fact that all of those 18 studies were more concerned in estimating WTP estimates that reflected the WFD requirements on policy changes and therefore their estimates might be considered altogether more reliable. Overall, lower WTP for improved water quality may be attributed to the fact that most respondents believe that the water quality in their respective regions or countries was at a good enough status therefore paying for an even better water status would not be reasonable.
The amounts of 40, 44 or 31 euros per person per year can be considered considerably low for such an important service such as water quality. If they are compared to donations made to charities (since some of the payment vehicles used in the meta-analysis described in this chapter are donations to institutions that are created to protect the environment and therefore can be viewed as charitable giving) in the UK only, 44% of the population claimed it donates monthly to charity organizations with an average monthly donation around 14£ (16.98 Euros) according to the yearly report of UK Giving (2014) with the average being in the last ten years of UK Giving from £10 to £15. When compared to an average water bill in the UK, a 1-2 bedroom apartment payed monthly in the UK an amount of 35 £ (41.28 Euros) which is almost identical with the RE estimate from the meta-analysis. Using Eurostat data, in 2012 (latest year with data available) the average expenditure per inhabitant per year (including expenses for maintenance) for all 28 EU member states was approximately 300 Euros5. This number as an average is highly variant since in large countries such as Germany, France, Italy, Spain and the UK drive the average higher than that of other countries. Even with this rough estimate though, the WTP estimate from the meta-analysis is the 1/10 of the average expenditure. Understandably then, when compared with the expenses for the water bill (increases in the water bill were the most used payment vehicle in the CVM studies to obtain WTP estimates and people can use the level of their water bill as reference6) the WTP amount is viewed as reasonable. In the case of the Louros’ catchment it is not enough to render NPV positive and therefore we can state that enforcing the implementation of the WFD in the area would decrease social welfare. Due to these results Skuras et al. (2013) conducted a disproportionality exercise to point out the groups of the population that are required to withstand the bulk of the cost of the policy.
Net costs are considerably higher than net benefits in the present analysis which can be attributed to the estimation of benefits. The environmental service that is valued is increased water quality, whose value is captured by WTP estimates. There are though multiple goods and services that are affected by achieving GES in a water body such as higher crop productivity, higher recreational uses (that could affect rafting, recreational angling and hiking experiences), lower water treatment costs (due to existing higher water quality) to name a few. The lack of incorporating such benefits attributes greatly to benefits being under-valued.
6. Conclusions
The aim of this chapter is to present a unified analysis for policy-makers and social planners. First, we presented the dangers of faulted policy design. The occurrence of a false positive is less serious than that of a false negative but both are of high importance. Given the level of the negative NPV regardless the time horizon, discount rate and benefits used is evident that “top-down “ policies and decision-making can lead to high losses of welfare.
5 Source: http://appsso.eurostat.ec.europa.eu/nui/submitViewTableAction.do last accessed 4/17/2017 6 It is possible to assume that respondents used “anchoring” techniques to reveal their WTP. Since there is no actual “market” for water quality, people in order to answer turn to what they consider to be the closest “reference point”. In this case that might be the level of their water bill. Using that, people could state WTP amounts that had their water bill as a starting point. The phenomenon of anchoring is present in CVM studies and it is hard to distinguish, especially in the case of the literature used in the meta-analysis exercise that is mentioned in this thesis since the payment vehicle used was the water bill. Anchoring may not pose a problem though in unearthing the true WTP of a respondent if his preferences are actually determined by market mechanisms (people place value to values and objects according to the value that society places on them; an expensive mobile device is more important due to the status it ensues and it is more desirable than spending more for eating healthy. Rationality of the individual, one of the fundamental principles of economic theory may be debated here.). Additionally, respondents might have lexicographic preferences, stating environmental friendly opinions, which they do not necessarily hold true in their own lives, but they comply with their sets of beliefs. For a further discussion on the subject, see chapter 7, section 9 of this thesis. Additionally, failing to incorporate the true extent of benefits from environmental goods and ecosystem services provided leads to significant underestimations of economic benefits.
Results also showed that, under public pressure and seemingly high rates of mineral fertilization, decision makers may falsely adopt an agri-environmental programme that may be both, ineffective in reducing nutrient loads and cost inefficient. Furthermore, they fail to take account of future changes that may inactivate the proposed mitigation measures, aggravate or reverse the baseline situation.
The present work suggests that the design of such agri-environment programmes should evolve to a thoroughly designed, interdisciplinary exercise integrating science and social- science models in a step-wise procedure. This process will ensure decision makers with the highest possible information from scientific sources and models and from local knowledge. This information can be used by appropriate simulation models to calibrate the baseline situation. Once an appropriately calibrated model is derived, further scenarios simulating policy, land use and climate changes can be simulated. Based on these results the effectiveness and cost efficiency of the proposed actions and of envisaged changes can be assessed. As a result, decision makers will be able to grasp an ex-ante evaluation of the current situation and of the proposed mitigation actions, if needed. This will allow decision makers to monitor the current situation and respond by adopting new measures or adapting existing ones to the changing physical, social and policy environment.
Under this proposal, the cost of the design phase of an agri-environmental programme will increase. But, in view of the high cost of mitigation measures, such an increase in the design stage of the agri-environmental policy should be considered as an insurance against the commitment of very expensive false positive and false negative decisions.
Finally, in this work it is proposed that the targets of agri-environmental policy and consequently, the measurement of abatement cost should be done in terms of nutrient concentrations and loads in water and not in terms of physical quantities of abated substance in the field. This will provide the cost efficiency exercise with a wider perspective as concerns the sources of nutrients and abiotic, biotic and anthropogenic activities that contribute the nutrient loads. In turn, this will force agricultural policy decision makers to coordinate their actions with other environmental policy makers for achieving maximum results and avoiding internal contradictions.
Appendix Meta Analysis
Study name Statistics for each study Mean and 95% CI Standard Lower Upper Mean error Variance limit limit Z-Value p-Value Bateman et al 21.200 1.895 3.592 17.486 24.915 11.186 0.000 Mourato et al 89.423 5.787 33.493 78.080 100.766 15.452 0.000 Czakofski and Scazni 1 9.770 0.893 0.797 8.020 11.520 10.942 0.000 Czakofski and Scazni 2 26.020 1.929 3.719 22.240 29.800 13.492 0.000 Martin ortega and berbel 10.030 5.340 28.516 -0.436 20.496 1.878 0.060 hanley bothrok 41.260 7.678 58.945 26.212 56.307 5.374 0.000 hanley mortray 35.084 4.674 21.848 25.922 44.245 7.506 0.000 Brouwer Romania 22.640 1.811 3.280 19.090 26.190 12.501 0.000 Brouwer Hungary 34.330 1.617 2.613 31.162 37.498 21.236 0.000 Hasler et al. (2009) 59.250 0.636 0.404 58.004 60.496 93.163 0.000 Ahtiainen et al. (2012) - Latvia 5.890 0.115 0.013 5.665 6.115 51.307 0.000 Ahtiainen et al. (2012) - Poland 13.390 0.176 0.031 13.045 13.735 76.036 0.000 Ahtiainen et al. (2012) - Lithuania 16.510 0.943 0.890 14.661 18.359 17.502 0.000 Ahtiainen et al. (2012) - Germany 25.150 0.802 0.643 23.578 26.722 31.363 0.000 Ahtiainen et al. (2012) - Estonia 25.760 2.641 6.973 20.584 30.936 9.755 0.000 Ahtiainen et al. (2012) - Finland 42.490 1.754 3.077 39.052 45.928 24.222 0.000 Ahtiainen et al. (2012) - Denmark 36.270 2.526 6.379 31.320 41.220 14.361 0.000 Ahtiainen et al. (2012) - Sweden 77.140 8.225 67.657 61.019 93.261 9.378 0.000 31.591 3.072 9.438 25.569 37.612 10.283 0.000 -150.00 -75.00 0.00 75.00 150.00
Table 4 : Meta-analysis on WTP to achieve Very Good status in the EU between years 2000-2016 Favours A Favours B
Meta Analysis References
Arrow, K.J., Cropper, M.L., Gollier, C., Groom, B., Heal, G.M., Newell, R.G., Nordhaus, W.D., Pindyck, R.S., Pizer, W.A., Portney, P.R. and Sterner, T., 2014. Should governments use a declining discount rate in project analysis?. Review of Environmental Economics and Policy, p.reu008.
Barbayiannis, N., Panayotopoulos, K., Psaltopoulos, D., Skuras, D., 2011. The influence of policy on soil conservation: a case study from Greece. Land Degrad. Dev. 22 (1), 47–57. http://dx.doi.org/10.1002/ldr.1053
Batáry, P.V., Dicks, L., Kleijn, D., Sutherland, W., 2015. The role of agri-environment schemes in conservation and environmental management. Conserv. Biol. 29 (4), 1006–1016. http://dx.doi.org/10.1111/cobi.12536
Buckley, C.,Wall, D.P., Moran, B., O'Neill, S., Murphy, P.N.C., 2016. Farm gate level nitrogen balance and use efficiency changes post implementation of the EU nitrates directive. Nutr. Cycl. Agroecosyst. 104 (1), 1–13. http://dx.doi.org/10.1007/s10705-015-9753-y
Butterbach-Bahl, K., Baggs, E.M., Dannenmann, M., Kiese, R., Zechmeister-Boltenstern, S., 2013. Nitrous oxide emissions from soils: how well do we understand the processes and their controls? Philos. Trans. R. Soc. B 368, 1–13. http://dx.doi.org/10.1098/rstb.2013.0122
Christensen, J.H., Rummukainen, M., Lenderink, G., 2009. Formulation of very-high- resolution climate model ensembles for Europe. In: Van der Linden, P., Mitchell, J.F.B. (Eds.), ENSEMBLES: Climate Change and Its Impacts: Summary of Research and Results From ENSEMBLES Project. Met Office Hadley Centre, FitzRoy Road, Exeter EX1 3PB, UK, pp. 47–58 (Chapter 5, 160 pp).
DasguptaP., Marglin, S., and Sen, A. (1972) Guidelines for project evaluation, New York, UNIDO
DEFRA 2011. The Green Book: Appraisal and Evaluation In Central Government Treasury Guidance London
Dietz S, Hepburn C. 2010. On non-marginal cost-benefit analysis, working paper, LSE
Doorenboos, J., Kassam, A.H., 1979. “Yield response to Water”, Irrigation and Drainage Paper No. 33. FAO, Rome Italy (193 pp).
European Economic Community, 1991. Council Directive of 12 December 1991 concerning the protection of waters against pollution caused by nitrates from agricultural sources. Off. J. Eur. Communities 375, 1–8.
Fowler, D., Steadman, C.E., Stevenson, D., Coyle, M., Rees, R.M., Skiba, U.M., Sutton, M.A., Cape, J.N., Dore, A.J., Vieno, M., Simpson, D., Zaehle, S., Stocker, B.D., Rinaldi, M., Facchini, M.C., Flechard, C.R., Nemitz, E., Twigg, M., Erisman, J.W., Butterbach-Bahl, K., Galloway, J.N., 2015. Effects of global change during the 21st century on the nitrogen cycle. Atmos. Chem. Phys. 15, 13849–13893. http://dx.doi.org/10.5194/acp-15-13849-2015
Futter, M.N., Erlandsson, M.A., Butterfield, D., Whitehead, P.G., Oni, S.K., Wade, A.J., 2013. PERSiST: the precipitation, evapotranspiration and runoff simulator for solute transport. Hydrol. Earth Syst. Sci. Discuss. 10, 8635–8681. http://dx.doi.org/10.5194/hessd-10-8635- 2013
Goffe, P., 2013. The Nitrates Directive, Incompatible with Livestock Farming? Policy Paper No. 93, 30th May. Jacques Delors Institute, Notre Europe (Downloadable from: www.institutdelors.eu/media/nitratesdirective-legoffe-ne-jdi-may13.pdf ).
Gollier, C., Koundouri, P., & Pantelidis, T. (2008). Declining discount rates: Economic justifications and implications for long-run policy. Economic Policy, 23(56), 758-795.
Gollier, C. (2013). Pricing the planet's future: the economics of discounting in an uncertain world. Princeton University Press.
Grinsven van, H.J.M., Tiktaka, A., Rougoorb, C.W., 2016. Evaluation of the Dutch implementation of the nitrates directive, the water framework directive and the national emission ceilings directive. NJAS Wageningen J. Life Sci. 78, 69–84. http://dx.doi.org/10.1016/j.njas.2016.03.010
Haury, J., Peltre, M.C., Tremolieres, M., Barbe, J., Thiebaut, G., Bernez, I., Daniel, H., Chatenel, P., Haan-Archipof, G., Muller, S., Dutartre, A., Laplace-Treyture, C., Cazaubon, A., Lambert-Servien, E., 2006. A new method to assess water trophy and organic pollution - the macrophyte biological index for rivers (IBMR): its application to different types of river and pollution. Hydrobiologia 570, 153–158. http://dx.doi.org/10.1007/s10750-006-0175-3
Howarth, R., Swaney, D., Billen, G., Garnier, J., Hong, B.G., Humborg, C., Johnes, P., Morth, C.M., Marino, R., 2012. Nitrogen fluxes from the landscape are controlled by net anthropogenic nitrogen inputs and by climate. Front. Ecol. Environ. 10, 37–43. http://dx.doi.org/10.1890/100178
Howden, N.J.K., Burt, T.P.,Worrall, F.,Whelan, M.J., Bieroza, M., 2010. Nitrate concentrations and fluxes in the River Thames over 140 years (1868–2008): are increases irreversible? Hydrol. Process. 24, 2657–2662. http://dx.doi.org/10.1002/hyp.7835
Jackson, B.M., Browne, C.A., Butler, A.P., Peach, D., Wade, A.J., Wheater, H.S., 2008. Nitrate transport in chalk catchments: monitoring, modelling and policy implications. Environ. Sci. Pol. 11, 125–135. http://dx.doi.org/10.1016/j.envsci.2007.10.006
Kleijn, D., Sutherland, W.J., 2003. How effective are European agri-environment schemes in conserving and promoting biodiversity? J. Appl. Ecol. 40, 947–969. http://dx.doi.org/10.1111/j.1365-2664.2003.00868.x
Kotti, M.E., Vlessidis, A.G., Thanasoulias, N.C., Evmiridis, N.P., 2005. Assessment of river water quality in northwestern Greece. Water Resour. Manag. 19 (1), 77–94. http:// dx.doi.org/10.1007/s11269-005-0294-z.
Kuik, O., 2006. Ex ante and Ex Post Costs of Implementing the Nitrates Directive: Case Study in the Framework of the Project ‘Ex Post Estimates of Costs to Business of EU Environmental Policies’. Commissioned by European Commission, DG Environment, Unit G.1 Sustainable Development and Economic Analysis. Framework Contract No ENV.G.1/FRA/2004/0081 (Downloadable from: ec./environment/enveco/ex_post/pdf/nitrates.pdf)
Lebègue. D. (2005). Révision du taux d'actualisation des investissements publics. rapport de général du Plan, C
Luo, G.J., Kiese, R., Wolf, B., Butterbach-Bahl, K., 2013. Effects of soil temperature and moisture on methane uptake and nitrous oxide emissions across three different ecosystem types. Biogeosciences 10, 3205–3219. http://dx.doi.org/10.5194/bg-10-3205-2013
Macdonald, G.K., Bennett, E.M., Potter, P.A., Ramankutty, N., 2011. Agronomic phosphorus imbalances across the world's croplands. Proceedings of the National Academy of Sciences of the United States of America. 108, pp. 3086–3091. http://dx.doi.org/10.1073/pnas.1010808108
Manolaki, P., Tsakiri, E., Papastergiadou, E., 2011. Inventory of aquatic and riparian flora of Acheron and Louros rivers, and Zirou Lake inWestern Greece. Fresenius Environ. Bull.20 (4), 861–874.
Matzdorf, B., Lorenz, J., 2010. How cost-effective are result-oriented agri-environmental measures?—an empirical analysis in Germany. Land Use Policy 27, 535–544. http://dx.doi.org/10.1016/j.landusepol.2009.07.011
Nakićenović, N., Alcamo, J., Davis, G., de Vries, B., Fenhann, J., Gaffin, S., Gregory, K., Grübler, A., Jung, T.Y., Kram, T., La Rovere, E.L., Michaelis, L., Mori, S., Morita, T., Pepper, W., Pitcher, H., Price, L., Raihi, K., Roehrl, A., Rogner, H.-H., Sankovski, A., Schlesinger, M., Shukla, P., Smith, S., Swart, R., van Rooijen, S., Victor, N., Dadi, Z., 2010. IPCC Special Report on Emissions Scenarios. Cambridge University Press, Cambridge, UK.
Nordhaus, W. D. (2007). A review of the Stern review on the economics of climate change. Journal of economic literature, 45(3), 686-702.
Nordhaus, W. D. (2013). The climate casino: Risk, uncertainty, and economics for a warming world. Yale University Press.
Ovezikoglou, V., Ladakis,M., Dassenakis, M., Skoullos, M., 2003. Nitrogen, phosphorus and organic carbon in main rivers of the western Greece. Global NEST J. 5 (3), 147–156.
D., Rankinen, K., Lepisto, A., 2002a. A nitrogen model for European catchments: INCA, new model structure and equations. Hydrol. Earth Syst. Sci. 6 (559–582), 2002. http://dx.doi.org/10.5194/hess-6-559-2002
Randall, N.P., Donnison, L.M., Lewis, P.J., James, K.L., 2015. How effective are on-farm mitigation measures for delivering an improved water environment? A systematic map. Environ. Evid. 4, 18. http://dx.doi.org/10.1186/s13750-015-0044-5
Saggar, S., Jha, N., Deslippe, J., Bolan, N.S., Luo, J., Giltrap, D.L., Kim, D.G., Zaman, M., Tillman, R.W., 2013. Denitrification and N2O:N2 production in temperate grasslands: processes,measurements, modelling andmitigating negative impacts. Sci. Total Environ. 465 (1), 173–195. http://dx.doi.org/10.1016/j.scitotenv
Skoulikidis, N., Amaxidis, Y., Bertahas, I., Laschou, S., Gritzalis, K., 2006. Analysis of factors driving stream water composition and synthesis of management tools—a case study on small/medium Greek catchments. Sci. Total Environ. 362, 205–241. http://dx.doi.org/10.1016/j.scitotenv.2005.05.018
Skuras, D., Wade, A., Psaltopoulos, D., Whitehead, P., Kontolaimou, A., Erlandsson, M., 2014. An interdisciplinary modelling approach assessing the cost-effectiveness of agri- environmental measures on reducing nutrient concentration to WFD thresholds under climate change: the case of the Louros catchment. Oper. Res. 14 (2), 205–224. http://dx.doi.org/10.1007/s12351-014-0158-5
Söderholm, P., Christiernsson, A., 2008. Policy effectiveness and acceptance in the taxation of environmentally damaging chemical compounds. Environ. Sci. Pol. 11, 240–252. http://dx.doi.org/10.1016/j.envsci.2007.10.003
Stern, N. H. (2007). The economics of climate change: the Stern review. Cambridge University press.
UK Giving 2014 : An overview of charitable giving in the UK during 2014
Verhoeven, J.T.A., Arheimer, B., Yin, C.Q., Hefting, M.M., 2006. Regional and global concerns over wetlands and water quality. Trends Ecol. Evol. 21, 96–103. http://dx.doi.org/10.1016/j.tree.2005.11.015
Wade, A.J., Durand, P., Beaujouan, V.,Wessel,W.W., Raat, K.J.,Whitehead, P.G., Butterfield,
Wade, A.J., Whitehead, P.G., Butterfield, D., 2002b. The integrated catchments model of phosphorus dynamics (INCA-P), a new approach for multiple source assessment in heterogeneous river systems: model structure and equations. Hydrol. Earth Syst.Sci. 6, 583– 606.
Wang, H.J., Wang, H.Z., 2009. Mitigation of lake eutrophication: loosen nitrogen control and focus on phosphorus abatement. Prog. Nat. Sci. 19 (10), 1445–1451. http://dx.doi.org/10.1016/j.pnsc.2009.03.009
Weitzman, M. L. (1998). Why the far-distant future should be discounted at its lowest possible rate. Journal of environmental economics and management, 36(3), 201-208.
Weitzman, M. L. (2001). Gamma discounting. American Economic Review, 260-271.
Weitzman, M. L. (2007). Subjective expectations and asset-return puzzles. The American Economic Review, 97(4), 1102-1130.
Whitehead, P.G.,Wilson, E.J., Butterfield, D., 2008. A semi-distributed Integrated Nitrogen model for multiple source assessment in Catchments (INCA): Part I—model structure and process equations. Sci. Total Environ. 210–211, 547–558. http://dx.doi.org/10.1016/S0048- 9697(98)00037-0
Zerefos, C., Capros, P., Natsis, A., Papandreou, A., Sabethai, I., Yfantopoulos, I. (Eds.), 2011. The Environmental, Economic and Social Impacts of Climate Change in Greece. Bank of Greece (in Greek). ANNEX A. Determining the selection of 3,5% discount rate in environmental projects
According to the Green Book of Defra (2011 and 2013), discount rate is derived from the social time preference rate. This is the value society places to the present, when compared with the future consumption of the good or service in question. The Social Time Preference Rate (STPR) is based on comparisons of utility across time and generations. The Green Book advices that STPR is used as the standard real discount rate. STPR comprises of two parts, the rate which individuals discount future consumption of the good or service in question and an additional element that points to the lower marginal utility of future consumption patterns in case they grow (This effect is represented by the product of the annual growth in per capita consumption (g) and the elasticity of marginal utility of consumption (μ) with respect to utility).
Therefore, STPR is defined as: r = ρ + μ.g
Ρ depends on i) catastrophe risk (L) and pure-time preference (δ). The first element (L) is the likelihood a catastrophic event occurs, so devastating that all policies and programmes refereeing to the good or service in question are terminated or drastically altered. These events would be natural disasters, wars or major technological advances that make the good or service obsolete. Naturally, such an estimate is hard to calculate. The second element (δ) reflects the preferences of individuals concerning consumption in the present, rather than the future, while maintaining a stable level of consumption per capita over time. Evidence from surveys suggest that ρ hoovers around 1,5 per cent a year for the near future. The “near future” term refers to projects and analyses that refer to time horizons no more than 30 years. After 30 years the Green Book suggests different STPR rates, lower than those of the near future.
Concerning the elasticity of the marginal utility of consumption μ, bibliographic evidence cited in the Green Book calculates it around 1. This implies that a marginal increment in consumption to a generation that has twice the consumption of the current generation will reduce the utility by half.
Concerning the estimates of g, Maddison (2001) demonstrates how growth per capita in the UK during the period of 195-1998 was around 2,1 per cent. This is also the position of the British Treasury’ paper Trend Growth: Recent Developments and Prospects. The Green Book suggests a 2 per cent rate would be the estimate of choice for g.
To calculate STPR we would now have: g = 2 per cent, ρ = 1.5 per cent, μ = 1.0, then from equation (1) the STPR to be used as the real discount rate is 0.015 +1.0*0.02 = 3.5 per cent
Chapter Seven: Risk perceptions and environmental policy in the European Union: assessing perceptions on pressures on water bodies using ecological and chemical indicators through an integrated model
1. Introduction
Risk perception on climate change appears as a rather straightforward subject. The existence of risks is accompanied by the risk perception of the people affected by them and also by their actions to be shielded from them or preventing their occurrence in the first place. According to the Stern Review (2006), climate change can be viewed as an externality distinguishable from other externalities as it entails uncertainties on its effects and risks while having pervasive economic impacts. William Nordhaus in “The Climate Casino” clearly states “global warming is a major threat to humans and the natural world” (p. 3)
According to European Environment Agency’s database7, EU Member States (MS) are mostly surrounded by water and have a total of 127,540 water bodies (rivers, lakes, transitional waters and coastal waters) with a total length of 1,753,867 km. Denmark, France, Sweden and the United Kingdom possess the highest number of water bodies from all 28 countries. European fresh and marine water environments are facing multiple threats by a variety of stressors stemming from both climate change events and human activities. In an effort to respond to these challenges and create a unified ecological and environmental status for marine and freshwater ecosystems, the European Union (EU) first launched the Water Framework Directive (WFD, 2000/60/EC) in 2000 and in 2008 the Marine Strategy Framework Directive (MSFD, 2008/56/EC). The WFD had as its primary goal for all EU MS to achieve by 2015 Good Ecological Status in their surface waters and groundwater when the MSFD aims for MSto achieve Good Environmental Status in their marine waters by 2020. Both directives address the issues of pollutants and of nutrients concentration in these environments and advises MS on the process of managing and monitoring these resources by assessing a range of biological communities instead of focusing only on chemical quality in these waters (Hering et al. 2010).
Climate change risks according to the AR5 IPCC (2014), are pervasive in physical, biological and human systems. According to the Summary for Policy-makers (SPM) article 2.3 (IPCC 2014), climate change risks in Europe depend on the projected temperature rise and affect water-related systems such as lakes and rivers on the one hand and events such as floods, droughts and so on the other. The consequent need for climate change adaptation can usher negative trade-offs such as chemical pollution and alterations to water ecosystems (IPCC 2014, SPM 1.4.). Human activities and systems such as livelihoods and economic activities face risks levels that may vary from “medium” “to very high” (IPCC Assessment Report 2014 2.3). According to the report’s SPM: “In urban areas climate change is projected to increase risks for people, assets, economies and ecosystems, including risks from heat stress, storms and extreme precipitation, inland and coastal flooding, landslides, air pollution, drought, water scarcity, sea level rise and storm surges […]. These risks are amplified for those lacking essential infrastructure and services or living in exposed areas.” (IPCC Summary for Policymakers 2014. pp 15).
7 Source: http://www.eea.europa.eu/themes/water/european-waters last accessed: 28/04/2017 Risk perception concerning the issue of climate change is not a new topic in scientific research or in the public’s discussion agenda, but it is now getting established also within decision and policy makers'. The 2014 World Economic Forum carried out a survey of global risks among the participants in the annual World Economic Forum in Davos, Switzerland that had as its primary topic the number and hierarchisation of potential threats for the economy worldwide. Some 900 participants, were policy and decision-makers, which added extra importance and validity to the results. In respect to the impact of the various threats, water crises, failure of climate-change adaptation and biodiversity loss and ecosystems collapse were ranked number one, five and ten respectively. The threats to Europe in particular according to the Impacts of Europe’s changing climate (2008) mainly refer to:
i. Sea surface temperature in European seas is increasing more rapidly than in the global oceans. The rate of increase in sea surface temperature in all European seas during the past 25 years has been about 10 times faster than the average rate of increase during more than the past century. ii. The rate of increase is higher in the northern European seas and lower in the Mediterranean Sea. iii. The rate of increase observed in the past 25 years is the largest ever measured in any previous 25-year period.
Temperature increase is evident worldwide, with the five warmest years ever recorded by NOAA all having occurred after 2010 (NOAA 20178). For European countries in particular, land temperatures demonstrated high volatility. Although this is somewhat expected as land possesses a lower threshold for high temperatures than oceans, still Europe experienced its third warmest year (NOAA 2017) in 2016 with 2014 being the record year and 2015 being the third warmest. The year of interest for this paper (2012) was the 9th warmest recorded since 1880 by NOAA.
The combination of high pressures from climate change on the water environments in the EU on the one hand and on the high detail and ambitious EU Directives ascribing certain environmental and ecological levels to water ecosystems which entail large economic costs and impacts on the other hand requires further investigation. The aim of this paper is to assess how ecological estimates, produced for the purposes of top-down policies such as the WFD or the MSFD, affect public views and perceptions on the issues these directives are concerned with. Risk perceptions are then compared with the actual water threats of the respondents’ area of living, as these are reported in the official chemical and biological indexes of the DG Environment. The analysis also attempts to distinguish between country- wide and region-wide (in NUTS 29 regions) effects by employing a two-level statistical analysis examining country-and-region-wide effects.
8 Source: https://www.ncdc.noaa.gov/sotc/global Last accessed: 4/5/2017 9 Nomenclature of territorial units for statistics
Pressures to water ecosystems are inter-related as water ecosystems are connected and interacting with each other. It would be scientifically naïve to try and separate the phenomena and their analysis from one another as if their triggering factors and their impacts were not inter-related with other phenomena, some of them possibly being the results of adaptation efforts (i.e. chemical pollution), the combination of human activities and climate change (i.e. changes to water ecosystems) or simply phenomena exacerbated by climate change (i.e. algae growth). Thus, describing or analysing one distinct type of events (i.e. chemical pollution) will also require taking into account droughts, water scarcity, water usage and abstraction and so on. We choose to treat climate change as one of the main pressures to the water environment, while acknowledging it as not the only one. Our focus will be on the impacts of climate change in all water bodies and in specific sectors, assessed through specific indexes, in an extended but yet well-defined geographical area. We narrow our focus on three perceived threats to the water environment, as these were stated by approximately 25,500 EU citizens from 168 NUTS 2 areas in the 2012 Eurobarometer 344, specifically algae growth, chemical pollution and changes to water ecosystems.
2. Previous research: perceptions on pressures on water bodies
Literature on the subject of climate change and risk can be found in a great spectrum of scientific fields, such as meteorology, environmental ethics and behaviour, agricultural sciences, hydrology, energy and environmental economics to name a few. However, there are distinctions regarding the literature around risk which corresponds to the ability of understanding risk by the respondents (the ability to synthesize knowledge and reach cohesive applications), the engagement around risk issues by the individuals (how the assimilated information drives respondents to action) and the perceptions around risk of the individuals (views and the interpretation of the influx of information coming from various sources) (Wolf and Moser 2011). In this paper we focus on research addressing the latter as understanding of risks is harder to be captured by researchers and behavioural responses such as engagement in risk issues is beyond the scope of this study. Risk perceptions are analyzed by the use using of a Eurostat dataset where the survey questions concerned the understanding and the amount of information the respondents had with regard to water- related issues. We investigate an obvious influencing parameter: the positive relationship between proximity to risk and past experiences. We start by observing that the literature is divided on the existence of this relationship and, in some cases, it has been shown that exposure to risk does not impact risk perceptions consistently (Mayer, 2016), or that proximity to elements and ecosystems affected by climate change impacts risk perception both negatively (Brody et al., 2008) and positively (Howel et al. 2002).
The relevant literature’s results of risk perception depend on an array of factors. These can be grouped into two broad categories; those dependent on the time when the research was conducted and those dependent on survey design, facing issues such as lack of individual- level controls for behavior-underlining factors such as political ideology as reported for example in McCright et al. (2013). There is a notable change through time in the findings of studies reporting risk perceptions on climate change, with those conducted in recent years yielding higher rates of risk perception. An example of studies conducted in the nineties is Bord et al. (1998) who found in a country-wide survey in the U.S. that global warming was perceived as a threat by most respondents. But, the study reports that global warming was mostly perceived in a distorted way due to lack of information and consequently, and was not seen as a primary threat or danger to the population. Dunlap (1998), in a six-country survey (Canada, USA, Mexico, Russia, Brazil and Portugal), observed similar findings. Some surveys carried out during the early 2000’s reveal a moderate risk perception as is the study of Leiserowitz (2006), where conflicting results are shown due to lack of clear public perception of the object of investigation in Leiserowitz (2003). The shifting of perceptions is clearly demonstrated in Nisbet and Myers (2007) who present combined surveys results from the United States in the subject of public opinions on climate change, starting with studies from the 80’s that revealed that only 39% of respondents had heard about greenhouse emissions or the greenhouse effects due to the lack of available information through the news. Nisbet and Myers claim that as extreme climate events continued to increase, people started expressing increased levels of awareness. On average, awareness of climate change floated around 80% in the 1990’s reaching a steady 90% in the year 2006 (Nisbet and Myers 2007). Similar results were produced by studies concerning awareness of the Kyoto protocol and the truthfulness of the event of global warming, despite the sometimes-offsetting remarks from think tanks in the 1990’s (Dunlap and Jacques 2013).
The focus of studies and their subsequent design influenced studies such as Whitmarsh (2008) where uniformal results, regardless of previous experience of climate-change-induced phenomena such as floods are reported by. Although there is scientific consensus that climate change is heavily due to anthropogenic activities (IPCC 2014), general public does not seem to support such an argumentation. When climate change is seen as a natural process, then respondents do not perceive it as a threat to them or their livelihoods, as noted by Asplung (2014). However, disseminating information is not as simple as opposing natural processes to anthropogenic activities as socio-economic factors also influence perceptions. For instance, Bord et al (1998) claim that globally, people are concerned with climate change and that they will support actions and initiatives to prevent it from becoming catastrophic but not in an extent that they will change their lifestyles, at least voluntarily. It is worth noting that most of the literature on the subject of climate change risk perception between the 80’s and the beginning of the 21st century is originated from the United States. Therefore, some uniformity in the results is expected, as well as some distorted findings since only in the recent years American authorities have shown interest in such issues. American population in general appears to be not as interested as Europeans in the issues, dangers and challenges climate change presents. Surveys on risk perception in European countries started after the 2000’s, especially in the United Kingdom (Wolf and Mosler 2007). But it is only after extreme weather phenomena such as record heat waves and catastrophic wildfires and water scarcity issues, that other European countries started to be concerned with the issues related to climate change (Whitemarsh 2008, Lorenzoni and Hulme 2009, Morton et al. 2011 etc.), both at a science and governance level. The sample sizes of respondents in such studies is also of interest. Literature over the subject of climate change perceptions is divided between large-scale studies that are conducted usually in large developed countries with population representative samples, and small-scale studies that use focus groups and experimental study designs (Wolf and Moser 2011). Regarding the perceptions of different groups on climate change, familiarity with climate change-related issues is believed to increase risk perception, as well as education and science literacy. In our analysis, we endeavour to specifically test those assertions in a European context, where large developed and financially strong countries exist along with emerging economies and countries facing economic struggles, exist.
Finally, we point that according to Sagoff (1994) there is a difference between choices and preferences. Choices are actual actions with legal and moral consequences, whereas preferences are theoretical constructs. We follow Sagoff’s argument and include the notion of risk perceptions with regard to threats to the water environment and actual dangers presented in the form of bad or poor ecological status of freshwater ecosystems. The danger of individuals stating preferences that would possible be different if the respondents were to make choices on these issues is present in this analysis.
2.1. Previous research: anthropological pressures on water bodies
The EU has demonstrated in its latest directives the willingness to include public consultation to as many steps as possible in the process of their design, with the WFD being the most prominent of these directives. Therefore, the ability of EU citizens to identify threats and issues in the water bodies of the Union is becoming a factor both in terms of the structure and aims of such directives and of the level of participation and engagement demonstrated by EU citizens.
To document the perceptions the EU deployed in 2012 the Flash Eurobarometer 344 focusing on the subject of perceived threats to the water environment, a continuation of a similar survey (Flash Eurobarometer 2009, no 241). The comparative results can be found in Appendix 1.
We focus on the pressures originating from algae growth, chemical pollution and changes to water ecosystem as these were the indicators with the most comprehensive data available in the WISE10 database which allows us to combine ecological indicators for water resources and population distribution in the regional and country-wide level.
Algae growth as a threat to the water environment was mentioned by 41% of the respondents in the 2012 wave of the Flash Eurobarometer (European Commission, 2012), an increase of 11% from the first Flash Eurobarometer in 2009. It was the 5th most selected
10 WISE is a partnership between the European Commission (DG Environment, Joint Research Centre and Eurostat) and The European Environment Agency, known as “the Group of Four” (Go4), Source: http://water.europa.eu/info option but it is of interest since it is visually comprehensive by everyone and therefore people are more familiar with it and are more likely to select it after having witnessed such an event.
The dangers of chemical pollution were of particular interest to the respondents of the Flash Eurobarometer studies of 2009 and 2012, with chemical pollution being characterized as the number one threat to the water environment both in 2009 and 2012, with the 2012 percentage being 84%.
In the 2012 Flash Eurobarometer, 49% of respondents indicated that changes in water ecosystems are a threat to the water environment, a change of 16 points since 2009, making it the third highest perceived risk for European water environments. Respondents from countries in central and western Europe appeared to be more concerned with that issue, while countries in the north appeared less concerned.
2.2. Previous research: Algae Growth
The rapid growth and accumulations of phytoplankton/ micro-algae, which result to colourisation in the waters, are called algae blooms (Paerl et al. 2001). This rapid growth of algae is the product of eutrophication, which is direct result of climate change. When excessive nutrients and sediments (which may be attributed to flooding), mainly nitrogen and phosphorus compounds (used, for example, in agriculture and aquaculture practices, manure from farming, and untreated sewage runoffs) are found in a waterbed, this leads to algae bloom, resulting to the depletion of oxygen from the water body (Spalding, et al. 2016). The main source of excessive nutrient deposition is agriculture (nitrogen fertilizers, manure from livestock and soil erosion) and waste water coming from sewage treatment plants. Increased depositions of nitrogen can also be attributed to increased energy production which leads to increased atmospheric inputs from NOx emissions (Glibert et al 2005). Correlations also exist between the appearance of algae blooms and increases in mean water temperature (Glibert et al 2005) both in marine and freshwater ecosystems (Miraglia et al. 2009). Eutrophication is one of the major threats of water quality in Europe, affecting approximately 40% of rivers and lakes across Europe (DG Environment 2008).
Algae blooms in EU marine waters now occur one month earlier than 30-40 years ago (Jol et al. 2009) and the growth period of algae is expected to last longer in the North Sea, affecting the harvesting of shellfish (Peperzak 2005). Changes in sea surface temperature also accelerate algae growth, along with nutrient and light abundance, and increases in algae appearances have been documented in the North Sea and eastern North Atlantic while decreases have been detected in the north-west of the European Shelf (Jol et al. 2009). Algae blooms are not necessarily harmful but excessive accumulation of micro-algae can initiate the cycle of eutrophication. It starts by the bloom of algae which does not allow sunlight from entering the water and reach the river’ seabed where seagrass or seaweed usually grow, diminishing the biodiversity and the potential habitats and source of food for smaller organisms and fishes. In addition, when algae decompose, oxygen is absorbed out of the water and this leads to resulting in anoxia or hypoxia with negative consequences for sea life (Miraglia et al. 2009). Fish and other animals consequently either die or they flee the area (Spalding et al. 2016). Lower water flows or reduced water flow velocities also can prevent water movement and therefore increase the probability of toxic algae blooms and reduce dissolved oxygen levels (Whitehead et al. 2009). This would lead to higher stress to water treatment plants as they would receive increased flows of dissolved organic carbon (Whitehead et al. 2009). This is alarming as by the early 2000s, more algae blooms with more toxic effects are appearing than the past decades (Glibert et al 2005).
Algae growth can also have an impact on tourism when water bodies such as rivers and lakes are used for recreational activities (IPCC 2013), for which water clarity is an important feature for the users of the resource. Harmful algae growth can have various colours (red, green yellow, depending on the type of algae present and their pigmentation (Glibert et al 2005) which can decrease recreational activities such as nature watching.
Eutrophication is documented to impact the industry sector as well. Martin-Ortega et al. (2016) note that in the Thames river in London, excessive algae growth, which was the result of water temperature increases in water reservoirs, caused cooling machines from various industries around the river to break down due to insertion of algae-rich waters in their systems. Similarly, murky waters due to excessive algae bloom and dead algae, may accentuate delays in transportation via water routes because the caution taken in navigation due to the reduce visibility of water depth. Algae blooms also have negative direct impact on human and animal health due to the deterioration of drinking water quality (Paerl et al. 2001).
2.3. Previous research: Chemical pollution
Chemical pollution was one of the main targets of the WFD. The EU’s purpose in that case is to safeguard a non-toxic environment (European Union, 2013). In particular, in order for a good chemical status of waters to be achieved, the WFD is providing a list of priority substances that pose dangers for human health and the environment and determines the necessary actions that need to be taken (Waternote 8, 2008). In order for this to be achieved, each Member State must ensure that any additional pollutants that have national relevance are controlled.
The EU is experiencing a significant amount of pollutants reaching the water systems and the environment, out of 100,000 man-made chemicals. Despite numerous efforts by the EU, still, some mixture effects still appear, requiring immediate attention and need to revise the process of monitoring and classification of chemicals as threats to human and environmental health (Brack et al. 2017). Malaj et al. (2014) examined data from 4,000 European monitoring sites and discovered high concentrations of organic chemicals in 14% of the sites in 14% and 42% of the sites were likely to be long-term affected by organic chemicals, with impact to at least on group of organisms. The consequences of concentrations of those chemicals were likely to have lethal and chronic-ling-term effects on sensitive fish, invertebrate, or algae species. In Loos et al. (2009), only 10% of samples in rivers across the EU were described as “very clean”.
Chemical pollution was the number one perceived threat for both waves of the Eurobarometer studies of 2009 and 2012. Chemical pollution in the water, either ocean waters, coastal waters or inland waters has been receiving a lot of attention from news outlets and media. Historically, long contaminated rivers were very common in the European mainland, rivers such as the Danube, Thames and the Volga that cross heavily populated areas and were the catalyst of economic growth for centuries up until the late 19th century. What was the main source of pollution in the early centuries of human civilization in Europe were human waste that ended up in the nearby rivers untreated, being the causes of diseases and pandemics. The strange paradox that functions that supported life such as economic activities of transportation of goods were carried in the same area were threats to human life existed in abundance. Technological advances and pervasive scientific knowledge permitted the swift in the 20th century towards striving for water bodies free from chemical pollution. Large investments were and are made in treating human waste and in collecting hazardous chemicals before ending up in the water column. Despite all the efforts in that domain, chemical pollution is a common phenomenon in many countries within Europe, either due to the lack of infrastructure, lack of appropriate monitoring or by accident. As a consequence, it is not particularly unexpected for the Eurobarometer’s respondents to state chemical pollution as the number one perceived threat to the water environment.
The level of alertness for the issue of chemical pollution can be supported, apart from information shared by news outlets, by scientific evidence as chemical pollution is evident in surface, groundwater and coastal waters across Europe. Vethaak et al. (2005) in various case studies across the Dutch aquatic environment found estrogenic contamination in all water bodies examined. Some specific areas are showing elevated levels of contamination with effects on wild fish as in the case of Kirby et al. (2004) in the UK. Halling-Sørensen et al. (1997) where among the first to pinpoint the cycle of the medical substances (pharmaceuticals) used in the primary (growth promoters and therapeutics for livestock production, feed additives in fish farms), industrial (pharmaceutical components ending in human waste which are later on disposed to sewage sludge) and processing sector follow, to end up in the water column.
Another main stressor to the water environment is sewage pollution, with untreated or poorly treated sewage water entering the water column. The severity of the problem goes hand in hand with each area’ dependence on water, with arid or semi-arid areas being more exposed to such ecological threats (Rivetti et al. 2017). Pollutants from urban and rural point sources represent biotic risks associated with eutrophication and the dispersion of pathogens (Anza et al., 2014) that, in the cases of floodplains or river basins, increase the dangers of pollution. Arid, or semi-arid areas such as Spain are also prone to flooding which, apart from being a danger for local ecosystems and human welfare, can carry pollutants and dispose them in large urban and rural areas.
Pesticides in Malaj et al. (2014) were the main contributors to chemical risk in 81%, 87%, and 96% of exceedances in fish, invertebrates and algae populations, respectively. The types of impact found by Malaj et al were different in invertebrates and algae than in fish, with the first ones being severely affected when fish suffered from problems in reproduction, development and fitness. Following that, land use is closely related to chemical risk Carpenter et al. (2011), although the interconnectedness of freshwater ecosystems (or the lack of it) is also a determinant, of the level of chemical risk an area is subject to (Malaj et al. 2014).
The dangers from chemical pollution differ between “closed” and “open” wetlands, with the latter ones being in danger from the accumulation of persistent chemicals with longstanding effects on flora and fauna (Rivetti et al. 2017). Close wetlands are considered those where no influx of water and nutrients occurs but instead sustain activities that produce chemical pollutants such as human and agricultural activities.
Halling-Sørensen et al. (1997) by using data from the Danish Drug Administration report that a total of 110 tonnes of antibiotics were used in 1995 as growth promoters in livestock, feed additives in fish farms and as medicine for poultry production and almost 30% of them used were of active substance. Chemical pollution can be categorized in river water pollution (Wauchope 1978), sediment pollution, ocean pollution, soil pollution (the application of manure to agricultural soils, drug resistant chemicals end up in the water column or to the intestinal flora of animals according to Halling-Sørensen et al. (1997) and ocean pollution and pollution to fauna (Vos et al. 2000). Chemical pollution to fauna results to abnormalities, subtle or permanent alterations, all the way to disturbed sex differentiation (Vos et al. 2000).
The EU is viewing the issue of water quality and quantity rather seriously and only between 2007 and 2013 it invested a total of 22 billion Euros for investments concerning construction of drinking water facilities, wastewater treatment plants and sewerage networks through its Structural and Cohesion Funds (DG Environment 2008).
2.4. Previous research: Changes in water ecosystems
Water ecosystems have been the subject of severe physical changes in areas of the world where intense agricultural practices and surface and groundwater extraction take place simultaneously (Rivetti et al. 2017). Freshwater ecosystems, along with land converted to agricultural land, are the two most rapidly changing ecosystems in the world (Carpenter et al. 2011). Most river floodplains in Europe have been functionally eradicated (Brinson and Malvárez, 2002), with the striking example of Spain who has lost 79% of its wetland area since the 19th century (Casado et al., 1992). Changes are attributed either to changing climate and to anthropogenic activities. In 41 large rivers in Europe, dams are present, strongly affecting 74% of their total area (Nilsson et al. 2005).
Additionally, changes to water ecosystems due to climate change refer to changes to ice sheet covers in lakes and rivers, particularly in the northern hemisphere, have been recorded (Jol et al. 2009). Ice sheet are formed when air temperatures are below 0˚C and are of importance for biodiversity, oxygen conditions, nutrient cycling and underwater light levels (Jol et al. 2009). During the last century, the mean number of days of ice sheet coverage has reduced by 12 days, with ice cover being reduced and ice break-up occurring earlier (Jol et al. 2009). Franssen and Scherrer (2008) found a reduction of ice sheet covers in 11 alpine Swiss lakes in the past century, as do Blenckner et al. (2007) in 196 lakes in Sweden.
Bunn and Arthington (2002) demonstrate the importance of the shape and size of river channels on the health of habitats, namely obstructions and erratic water flow patterns, which affect biotic elements such as fish (reduction in fish populations), and reductions in species richness and of benthic macroinvertebrates. Additionally, Nilsson and Berggren (2000) and Bunn and Arthington (2002) stress out the importance for aquatic organisms and species to be able to navigate through river networks. Water abstraction and dam construction impact negatively migratory species by not allowing them to escape predators or dry seasons as well as helping introduce invasive species.
Nilsson and Berggren (2000) address the trapping of waterborne sediments by water reservoirs which reduces sediment cycling which leads to the alteration of the very nature of rivers. Soil that is waterlogged becomes anoxic and that results to damages to the primary root system, with varying damaging effects depending on the type of plants and their response mechanisms. Rivers, when managed with the purpose of facilitating water abstraction and utilisation, may face reduced flows and result to permanently or intermittently dry channels. Water reservoirs created by dams in rivers may also lead to permanent loss of habitats by inundation of soils and vegetation, an effect that is more impacting dry or mountainous areas, reaching all the way to extinguish entire populations, according to Nilsson and Berggren (2000). This can lead to releasing greenhouse gases by the decomposing organic material in the inundated areas (Louis et al. 2000).
A probable connection exists between anthropogenic changes and climate change dynamics in alteration to water-level fluctuations. According to Carpenter et al (2011), human impacts follow certain patterns, namely overharvest which creates decline in fish population which then is followed by increases in pollution, elimination of habitats and invasive species. Extreme changes in water levels of shallow lakes are responsible for tampering with ecosystem dynamics and functions and such cases are reported in Greece where a combination of anthropogenic and climate change impacts is present and the Netherlands where anthropogenic impacts are prevalent (Coops et al. 2003). These alterations in lake ecosystems take different forms, depending on geographical areas. More temperate areas are projected to experience higher precipitation which would lead to higher water levels in spring when Mediterranean water environments are expected to receive lower precipitation and extended drought periods. In Mediterranean lakes that could lead to high evaporation in the summer which would enhance increases in vegetation around those lakes (Beklioglu et al. 2001). In temperate lakes, high water-level fluctuations would increase the volume of run-off due to higher phosphorus concentration in the water column (Beklioglu et al. 2001). Additionally, such high water-level fluctuations and extended dry periods might promote Mediterranean shallow lakes to shift into more brackish or saline state in their effort to adapt to increased temperatures and drought periods (Coops et al. 2003). Alterations are also occurring in the form of new shorelines being created due to water level fluctuations, leading to the creation of new riparian zones (Johansson and Nilsson 2002). In northern Sweden, these shorelines were found along water storage reservoirs and demonstrated lack of biodiversity (Nilsson et al. 1997).
3. Data and Methods
Eurobarometer Data
Based on the literature review on people’s perception on pressures to water bodies and the ecological as well as anthropological pressures to the water bodies, for this[…]
Data for this work were extrapolated from the Flash Eurobarometer entitled “Attitudes of Europeans towards water-related issues” a European survey conducted in 2012, otherwise known as Flash Eurobarometer 344 (European Commission 2012). About 25,000 European citizens located in almost all EU territories responded to a questionnaire addressing issues related to water. We used the data collected for Question 5, which asked the respondents about their perceived major threats to the water environment:
“I am going to read out a list of threats. Can you please tell me which you believe are the main threats to the water environment in (name of the country)? “. The possible pre-coded answers included Algae growth, Chemical pollution, Water shortage, Floods, Change to water ecosystems, Dams, canals and other physical changes, Climate change, Other (for the respondent to state one that is not included in the pre-coded answers), You do not care about this issue, Don’t Know or Not Answer.
The survey coded the response to each one of the above-mentioned threats as a dummy variable (1 if the threat is perceived as such and mentioned by the respondent and 0 if it was not mentioned). The survey recorded the respondents’ place of residence (see following section) and specific characteristics, which included demographics (age, gender, household size), human capital and employment characteristics (education, employment status), psychographic characteristics (e.g. political orientation, religiosity and environmentalism), and behavioral activities related to water. Behavioural activities related to water include everyday water conservation practices such as not leaving taps running, taking shower instead of bath, installing water saving appliances, etc., harvest rainwater and avoiding the use of pesticides and fertilizers in the garden. But they also include a wider spectrum of environmental activities such as the preference for organic farm products, recycling specific waste like oil waste, unused pharmaceuticals, unused household chemicals, paints, solvents and batteries. The survey did not directly capture psychographic characteristics but it recorded several “environmentally and water conservation friendly” activities undertaken by the respondents. Therefore, participation to the reported behavioural activities towards water had helped us to build a basic psychographic profile of the respondents defined as “environmentalism”.
Behavioural characteristics instead include the respondent’s ability to assess water related issues and their own actual behaviour towards water use. For example, the respondent’s personal judgment on the level of information related to problems facing groundwater, lakes, rivers and coastal waters in their country, and their perception of the state of the water environment in the place they live, but also their consumer behaviour, such as drinking either tap or bottled water. A limitation of the survey is that it only provides behavioural characteristics related to household uses. For example, it does not report the respondents’ use of water bodies for recreation, sports, or environmental activities (e.g., birdwatching, nature photography, etc.) or as an element of their everyday landscape. Table 1 summarises the names, definitions and descriptive statistics of the survey variables capturing perceived threats to the water environment, socio-demographic, psychographic and behavioural characteristics.
Table 1. Names, definitions and descriptive statistics of variables capturing perceived threat to the water environment and respondent characteristics (source: European Commission, 2012).
Perceptions’ statistical analysis
Variable Description Mean St.Dev Perceived Threats ALGAE Dummy variable, 1 if respondent perceives algae 0.39 0.49 growth to be a threat to the water environment, 0 otherwise POLLUTION Dummy variable, 1 if respondent perceives chemical 0.82 0.38 pollution to be a threat to the water environment, 0 otherwise SHORTAGE Dummy variable, 1 if respondent perceives water 0.37 0.48 shortage to be a threat to the water environment, 0 otherwise ECOSYSTEM Dummy variable, 1 if respondent perceives change 0.41 0.49 to water ecosystems be a threat to the water environment, 0 otherwise MORPHOLOGY Dummy variable, 1 if respondent perceives dams, 0.27 0.44 canals and other physical changes to be a threat to the water environment, 0 otherwise Respondents Socio-Demographic Characteristics AGE Continuous variable of respondent’s age in years 50.03 17.09 GENDER Dummy variable, 1 if respondent is male, 0 if female 0.43 0.50 HOUSEHOLD_SIZE Number of people aged 15 years or more that live 2.78 6.43 in the household, the respondent included EDUCATION0 Dummy variable, 1 if respondent has not completed 0.07 0.15 ever or as yet, full time education, 0 otherwise EDUCATION1 Dummy variable, 1 if respondent completed 0.48 0.50 education before the age of 18, 0 otherwise EDUCATION2 Dummy variable, 1 if respondent completed 0.44 0.50 education after the age of 18, 0 otherwise SELF_EMPLOYED Dummy variable, 1 if respondent declares to be self- 0.10 0.30 employed, 0 otherwise EMPLOYEE Dummy variable, 1 if respondent declares to be an 0.33 0.47 employee, 0 otherwise RETIRED Dummy variable, 1 if respondent declares to be out 0.29 0.45 of the labour market due to retirement, 0 otherwise OTHER_EMPL Dummy variable, 1 if respondent declares any other 0.27 0.41 occupation including housekeeping, 0 otherwise RURAL Dummy variable, 1 if respondent lives in a rural area 0.33 0.47 or village, 0 otherwise MIDDLE_TOWN Dummy variable, 1 if respondent lives in small or 0.30 0.46 medium sized town, 0 otherwise LARGE_TOWN Dummy variable, 1 if respondent lives in a large 0.31 0.46 town or city, 0 otherwise Respondents “Environmentalism” WATER_USE Dummy variable, 1 if respondent actively limited 0.80 0.41 the amounts of water used, 0 otherwise WATER_CHEM Dummy variable, 1 if respondent used eco-friendly 0.54 0.51 household chemicals, 0 otherwise WATER_HARVEST Dummy variable, 1 if respondent harvests rain 0.39 0.49 water, 0 otherwise PESTICIDES Dummy variable, 1 if respondent avoids the use of 0.60 0.49 pesticides and fertilizers in her garden, 0 otherwise ORGANIC Dummy variable, 1 if respondent choses organic 0.46 0.50 farming products, 0 otherwise RECYCLE Dummy variable, 1 if respondent recycles 0.69 0.46 household dangerous substances and used goods, 0 otherwise OTHER Dummy variable, 1 if respondent stated another 0.01 0.08 activity not included in the above, 0 otherwise ENVIRONMENT Quasi-continuous or ordinal variable that scores the 3.48 1.57 respondents strength of “environmentalism” on a theoretical 0 (no activity) to 7 (all activities) scale Respondents “Behavioural” Characteristics INFORM Dummy variable, 1 if respondent feels "very well or 0.39 0.49 well informed about water problems", 0 otherwise TAP_WATER Dummy variable, 1 if respondent usually drinks tap 0.58 0.49 water, 0 if the respondent drinks bottled water or both
Geospatial variables data
The Flash Eurobarometer 344 (European Commission 2012) survey recorded the place of residence in terms of the European hierarchy of spatial administrative units called NUTS (from the French - Nomenclature des Unités Territoriales Statistiques); specifically at the NUTS2 level. NUTS2 corresponds to the major administrative units of the EU and the basic regions for the application of policies. This does not include Germany and the UK, which has been reported at the NUTS1 level (the larger agglomerate of NUTS2 regions), and some smaller countries that has been reported at the country level, i.e, Luxemburg, Cyprus, Malta, Estonia and Latvia. All the countries that don’t report their findings in NUTS2 level have been excluded from our analysis (see Results section). A great variability in responses both within citizens living in the same place, but also and between citizens’ responses living in different MS was reported in the survey results. For example, the average proportion of citizens that perceive “Changes in water ecosystems” as a main threat to the water environment of their country ranges from 7% to 76% and, as is shown in map 1, this also varies within the same country. In our analysis, we assume that that such differences are due to differences in the respondents’ characteristics as well as differences of the state of the environment within which the respondents work and live. This can be used as a “contextual” or experiential variable same for all respondents living in the same region. Information of various aspects of the water environment is not available at the administrative level, which prevent us to do a finer analysis of the survey results for our purposes of clearly depicting the ecological, morphological and chemical status of Eu’s water bodies.
Map 1: Average proportion of respondents living in a region and perceiving “changes in water ecosystems” as a main threat to water environment.
Another set of spatially related data used in our analysis is the EU’s Water Information System for Europe (WISE), which reports various water related indicators for 2012 as the EU member states report them. Based on the Water Framework Directive (WFD) requirements, the ecological status of water bodies is classified into five categories, i.e., high, good, moderate, poor and bad reflecting the abundance of aquatic flora and fauna, the availability of nutrients, aspects of salinity, temperature, pollution by chemical pollutants and alterations to morphological features such as water quantity, flow and depth and the structure of the river beds. WISE records the percentage of water bodies (area or length) that are classified into these categories at the spatial level of the River Basin District (RBD). The RBD levels are spatial units for the management of river basins as delineated by Member States under Article 3 of the Water Framework Directive (WFD). WISE also records the proportion of water bodies at the RBD level affected by different impacts (e.g. acidification, altered habitats, contaminated sediments, contaminated water by priority substances, nutrient and organic enrichment) and pressures (e.g. diffuse and point source pollution, water abstraction, and morphological alterations).
Combining perceptions and spatial statistical analysis
A major complication we had to solve is that RBDs do not spatially coincide with the NUTS levels. In some instances, and depending on the NUTS level, RBDs are spatially enclosed within a NUTS, or a RBD include a NUTS; frequently RBDs intersect with NUTS.
In map 2 we provide a visual example of the process we followed to create NUTS specific water environment indicators utilizing the WISE database at the RDB level. Map 2 shows the boundaries of the NUTS2 region of Yugozapaden (BG41) in Bulgaria. This region comprises of the city and province of Sofia, Bulgaria’s capital and largest city, and the regions of Blagoevgrad, Pernik and Kyustendil, with the corresponding cities are shown on the map. The region of Yugozapaden intersects with three RBDs, the boundaries of which are in light grey, and forms three intersected areas namely BG41-BG1000, BG41-BG3000 and BG41- BG4000. We assume that a water indicator, e.g. the proportion of rivers’ length that is in good ecological status or the proportion of water bodies impacted by nutrient enrichment, is evenly spread within an RBD and corresponds to the RBD’s area. Thus, if in the Danube River Basin (in the North of Sofia) (RBD BG1000) 36% of the water bodies are impacted by nutrient enrichment, then we assume that 36% of the BG41-BG1000 intersection water bodies and area are impacted by nutrient enrichment. The same applies to the intersections BG41- BG3000 and BG41-BG4000 with the proportions: 28% and 32%. By summing up the intersections, we then calculate the proportion of water bodies and the corresponding proportion of the NUTS2 BG41 area that is impacted by nutrient enrichment.
Map 2. An example of a European NUTS2 administrative region (BG41) intersected with three different River Basin Districts (BG1000, BG3000 and BG4000) in Bulgaria.
We also calculate population weighted water indicators. Our analysis has so far assumed that the population is evenly spread around the BG41 NUTS area. However, this is not realistic as the BG41-BG1000 intersection includes the capital of the country and its surroundings. So, to overcome this limitation of our model, we utilize a population canvas layer that is provided by Eurostat and records the population at 1x1 Km all over Europe. We find that, still with reference to map 2, in the BG41-BG1000 intersection there are around 1.5 million inhabitants while in the BG41-BG3000 and BG41-BG4000 intersections there are about 60,000 and 600,000 inhabitants respectively. On these figures, we calculate population weights of 0.69, 0.03 and 0.28 to be applied to the respective water indicator of the intersections, which result to a population weighted water indicator for the whole NUTS region.
The variability around European NUTS regions is very wide. Indicatively, Map 3 shows the percentage of the length of water bodies whose ecological status has been classified as poor or bad upscaled to the NUTS level with the procedure described above. It is worth notice that the largest part of Poland and of Italy are missing because almost all Polish and most of the Italian RBDs had classified less than half of their water bodies to ecological status classes. Table 2 shows the proportions for various environmental indicators, as reported by WISE, averaged through our model over 205 NUTS regions.
Map 3: Average proportion of the length of water bodies whose ecological status is classified as poor or bad at the NUTS level.
Table 2. Names, definitions and descriptive statistics of variables capturing water environment indicators at the NUTS level.
Variable Description Mean St.Dev Ecological status H_pc % of the length of water bodies classified 3.96 8.17 “High” G_pc % of the length of water bodies classified 22.26 16.57 “Good” M_pc % of the length of water bodies classified 34.29 16.81 “Moderate” P_pc % of the length of water bodies classified 17.42 17.21 “Poor” B_pc % of the length of water bodies classified “Bad” 6.19 9.90 Unk_pc % of the length of water bodies classified 15.88 25.04 “Unknown” H_pc_pw % of the length of water bodies classified 3.84 8.46 “High” weighted by the population G_pc_pw % of the length of water bodies classified 22.11 16.58 “Good” M_pc_pw % of the length of water bodies classified 34.33 16.91 “Moderate” weighted by the population P_pc_pw % of the length of water bodies classified 17.59 17.18 “Poor” weighted by the population B_pc_pw % of the length of water bodies classified “Bad” 6.30 10.05 weighted by the population Unk_pc_pw % of the length of water bodies classified 15.82 24.91 “Unknown” weighted by the population Important Impacts Acid_pc % of waterbodies impacted by acidification 3.16 8.72 Altered_pc % of waterbodies impacted by altered habitats 27.58 28.72 as a result of hydromorphological alterations Cont_Sed_pc % of waterbodies impacted by contaminated 5.43 13.36 sediments Cont_Sub_pc % of waterbodies impacted by contamination 15.62 24.85 by priority substances or other specific pollutants Elev_pc % of waterbodies impacted by elevated 1.07 6.63 temperatures Nutr_pc % of waterbodies impacted by nutrient 30.79 26.16 enrichment at risk of becoming eutrophic Org_pc % of waterbodies impacted by organic 18.43 20.05 enrichment Sal_pc % of waterbodies impacted by saline intrusion 1.48 7.86 Important Pressures Point_pc % of waterbodies under pressure from point 30.44 21.57 source pollution Diff_pc % of waterbodies under pressure from diffused 46.08 31.42 source pollution Abstr_pc % of waterbodies under pressure from water 11.28 14.56 abstraction Flow_pcn % of waterbodies under pressure from water 36.06 27.54 flow regulations and morphological alterations Manag_pc % of waterbodies under pressure from river 18.63 25.97 management Trans_pc % of waterbodies under pressure from 0.70 2.01 transitional and coastal water management Morph_pc % of waterbodies under pressure from other 5.99 16.61 morphological alterations Other_Press_pc % of waterbodies under pressure from all other 13.42 19.99 Pressures
4. Methods
The random-intercept logistic regression is presented in Equation (1) with a respondent- specific intercept 휁푗which is assumed to be independent and identically distributed across respondents and covariates풙푖푗. This is a generalized linear mixed model with both fixed effects (훽1 to 훽4 ) and a random intercept휁푗. Given 휋푖푗, responses 푦푖푗 for respondent j at different affecting factors i are independently Bernoulli distributed. As heterogeneity is suspected to occur between respondents by unobserved factors, the addition of the random intercept allows capturing this effect.
푙표푔푖푡(휋푖푗) = 훽1 + 훽2푥2푗 + 훽3푥3푖푗 + 훽4푥2푗푥3푖푗 + 휁푗 (1)
푦푖푗|휋푖푗 ~ Binomial (1, 휋푖푗)
휁푖 ~ N (0,ψ)
Since we want to observe the differences between to observation levels, both in the individual and in a NUTS 2 level, we construct a level-2 model according to Raudenbush and Bryk (2002). The level-1 model has to contain a sampling model, a link function and a structural model, as seen below respectively.
푦푖푗 ~ Bernoulli (휑푖푗)
logit(휑푖푗 ) = 휂푖푗
휂푖푗 = 훽0푗 + 훽1푥2푗 + 훽2푥3푖푗 + 훽3푥2푗푥3푖푗+. . . +훽13푥14푖푗 + 훽14푥13푗푥14푖푗 (2)
The level-2 model for the intercept 훽0푗 is as follows:
훽0푗 = 훾00 + 푢0푗 (3) where 훾00 is a fixed intercept and 푢0푗 is the random intercept. The remaining coefficients have no residuals for the random-intercept model,
훽푛푗 = 훾푛0, n=1,2, ... ,14
The combination of the level-2 model to the level-1 structural model we obtain the full form of the model:
휂푖푗 = 훾00 + 푢0푗 + 훾01푥2푗 + 훾02푥3푖푗 + 훾03푥2푗푥3푖푗. . . +훾013푥14푖푗 + 훾014푥13푗푥14푖푗 (4)
≡ 훽1 + 휁0푗 + 훽2푥2푗 + 훽3푥3푖푗 + 훽4푥2푗푥3푖푗 +. . . +훽13푥14푖푗 + 훽14푥13푗 (5)
Using the gllamm command in Stata14, we estimate the level-2 random-intercept logistic regression model of equation (4), which uses an adaptive quadrature approximation to increase the accuracy of the maximum likelihood estimations (Rabe-Hesketh et al. 2005). Almost the exact results are produced when using the meglm command.
The model used to estimate the factors affecting risk perceptions on the three selected pressures to the water environment can be seen in equation (6). The dependant variable in each model was either algae growth, chemical pollution or changes to water ecosystems:
q5_= 훽1+훽2*AGE + 훽3*GENDER + 훽4*EDUCATION1 + 훽5*EDUCATION2 + 훽6 ∗RETIRED + 훽7*EMPLOYEE + 훽8*SELF_EMPLOYED + 훽9*RURAL + 훽10*LARGE_TOWN + 훽11*ENVIRONMENT + 훽12*INFORM + 훽13*TAP_WATER + 훽14 ∗ 퐷푖푓푓_푝푐 +
훽15*Good_Chem_pc+ 훽16* P_B_pc (6)
5. Results
The “Null Model”
First, we fit a “null model”, i.e., a model without covariates. The null model is very illuminating as it allows us to estimate the odds ratio and the probability of a perceived threat to the water environment as well as the between regions variation of the responses. Table 3 provides the estimations of the constant term and its standard error and the random effects, i.e., the between regions variance and its standard error. Chemical pollution is the top perceived threat to the water environment with an astonishing probability of 84.8%. This is followed by the threat of changes to water ecosystems with a probability of 46.3%. The threats of algae and water shortage follow with probabilities of around 42%. Physical changes is the less perceived threat to the water environment (28.2%). The Variance Partition Coefficient (VPC) shows the residual variation in the propensity to perceive a threat that is attributable to unobserved regional characteristics. Water shortage and algae show the largest between regions variability. Almost 21% of the response variance for water shortage and 14% for algae is attributed to unobserved regional features. The threat of physical changes and of chemical pollution show the lowest VPC, but even in these cases the multilevel model is superior to a simple logistic regression model as indicated by the likelihood ratio statistic of table 1.
Table 3. Estimates of the “null model” for the five perceived threats to the water environment
Algae Chemical Pollution Changes to ecosystems Estimate Std. Err. Estimate Std. Err. Estimate Std. Err.
Constant (훽0) -0.335 0.061 1.716 0.049 -0.148 0.050 Variance of 0.543 0.061 0.246 0.040 0.353 0.045 (훽0)
Odds-ratio 0.716 5.561 0.862 Probability 0.417 0.848 0.463
VPC 0.142 0.070 0.097
Log-Likelihood -14,582.449 -9,577.175 -14,752.598
LR vs. logistic 1,872.86 472.48 1,657.48
Observations 22,954 20,881 22,954
Regions 166 151 166
The constant term portrays the grand average, log odds and probability for a typical region
(uj=0). For a specific region j the corresponding log odds and probability are calculated from
the term 훽0 + 푢푗. If, for example, we consider the case of algae as a perceived threat to the −0.335 water environment, the log odds for a “typical” region 푢푗 = 0, are 푒 = 0.716 and the corresponding average probability is 푒−0.335⁄(1 + 푒−0.335) = 0.417. However, for a region
with an estimated 푢푗 below 0, at for example, -1.5 the corresponding region specific estimated log odds become 푒−0.335+(−1.5) = 0.16 and the corresponding region specific estimated probability becomes 푒−0.335+(−1.5)⁄(1 + 푒−0.335+(−1.5)) = 0.138 or almost three times lower than the average. To portray the between regions variability in responses we can graph the regions according to their rank in the estimated random effect together with their 95% confidence intervals. A region whose confidence interval does not overlap the line at zero differs significantly from the average at 5%.
Figure 1 shows the considerable variability between regions as concerns the perceived threat of algae. The regions with the largest negative values of the estimated random effect are the regions with the lowest probability of perceiving algae as a threat. These include regions in Bulgaria, Spain Austria, Greece and Slovenia. The regions with the largest positive values of the estimated random effect are the regions with the highest probability of perceiving algae as a threat to the water environment. These include many French regions as well as regions from the Czech Republic, Belgium and Finland. Figure 2 shows the same graph but for the “changes to ecosystems” response variable. In this figure, the regions with the lowest probability of perceiving “ecosystem changes” as a main threat to the water environment, i.e., regions with the largest negative values of the estimated random effect, include Lithuania and regions from Finland, Denmark, Bulgaria, the UK, Ireland and the Netherlands. Conversely, regions with the highest probability, i.e. regions with large positive random effects, include regions from France, Germany, Portugal and the Czech Republic.
Figure 1. Between region variability of random effect estimates for the “algae” response.
Figure 2. Between region variability of random effect estimates for the “changes to ecosystems” response.
Random intercept and random slope models
Table 4 shows the estimates from all variables affecting perceived threats as estimated by the model specified in equation 6. The coefficients of the level-1 variables are used to calculate the log-odds of perceiving the threat for an individual with all her characteristics at the average and living in an average region (uj=0). Thus, the coefficient of 0.005 for the centered variable of age at the “algae” threat model is interpreted as an increase in the probability of perceiving algae as a main threat to the water environment by age. On the contrary, for the threats “chemical pollution” and “changes to ecosystems”, age exerts a negative impact, i.e., the younger the respondent from the centered value the higher the probability to perceive this threat as a main threat. There is a unanimous and strong gender effect for all threats. Female respondents have higher probabilities to perceive a threat as being “main threat” in comparison to male respondents. Education, also exerts a significant positive effect on the probabilities to perceive threats. The employment status does not exercise statistically significant effects except from the case of employees who show higher probabilities to perceive “chemical pollution”. Living in a rural place reduces the probability to perceive all threats as a main threat to the water environment.
The psychographic variable capturing how environmentally active is the respondent exerts an extremely significant positive effect in the probability of perceiving all threats. If the respondent feels that she is very well or well informed about water related issues then the probability of perceiving chemical pollution or water shortage as a main threat decreases. It is notable that the level of information was not a significant predictor of the “algae” or “ecosystem change” threats. Finally, the respondents who usually drink tap water as water- related consumption activity have lower probabilities of perceiving a threat as a main threat.
Level 2 variables only take one value in each region and thus it is necessary to compare respondents with different regional residuals. As such living in areas where a higher percentage of water bodies is characterized by diffused source pollution increases the individual probability to perceive “algae” as a main threat to the water environment. Living in areas where a higher percentage of water bodies length is characterized by good chemical condition decreases the individual probability to perceive “chemical pollution” as a main threat to the water environment. Finally, living in areas where a higher percentage of water bodies length is characterized by poor or bad ecological status increases the individual probability to perceive “Changes to ecosystems” as a threat to the water environment.
For “algae” threat model presented in table 2 we also found statistically significant random slope effects for the “environment” variable. The negative intercept-slope covariance estimate for “algae” implies that regions with above average perceived percentages of “algae” as a main threat (intercept residuals positive) tend also to have below average effects of environmental activity (slope residuals negative). Figures 3 shows how more pronounced this relationship is.
Algae Chemical Pollution Changes to ecosystems Estimate Std. Err. Estimate Std. Err. Estimate Std. Err. Level-1 variables AGE (centered) 0.005 0.001 -0.013 0.002 -0.012 0.001 GENDER -0.029 0.030 -0.294 0.039 -0.111 0.031 EDUCATION1 0.229 0.067 0.145 0.084 0.074 0.067 EDUCATION2 0.250 0.067 0.173 0.085 0.282 0.068 RETIRED -0.094 0.054 0.059 0.069 -0.018 0.055 EMPLOYEE -0.014 0.041 0.109 0.055 0.064 0.041 SELF_EMPLOYED -0.055 0.056 -0.103 0.073 0.006 0.057 RURAL -0.042 0.036 -0.100 0.047 -0.142 0.037 LARGE_TOWN 0.049 0.039 0.131 0.050 0.031 0.039 ENVIRONMENT 0.206 0.013 0.208 0.013 0.258 0.011 (centered) INFORM 0.052 0.032 -0.197 0.041 -0.006 0.032 TAP_WATER -0.038 0.033 -0.095 0.042 0.018 0.033 Level-2 variables Diff_pc 0.006 0.002 ------Good_Chem_pc ------0.005 0.002 ------P_B_pc ------0.004 0.002 Constant (훽0) -0.873 0.131 2.080 0.145 -0.418 0.101 Random effects Parameters Intercept 0.005 0.003 0.182 0.033 0.283 0.038 variance Slope variance 0.428 0.057 ------Interc.-slope -0.013 0.010 ------covar. Log-Likelihood -13,808.778 -8,971.01 -13,423.328 LR interc vs. 8.464 slope Observations 22,186 20,141 22,186 Regions 166 151 166 Table 4. Model estimates for all perceived threats
Figure 3. Intercept-slope residuals for the model estimating “algae” as a main threat and the variable environment with a variable slope.
6. Conclusion and discussion
Our analysis investigated what are the factors that drive European citizens to consider algae growth, chemical pollution and changes to water ecosystems as the main threats to the water environment in their countries and if that corresponds to the actual ecological status of their water ecosystems. The study uses official data from the 2012 European Commission’s 344 Flash Eurobarometer on water. Combining official European Environment Agency’s data and from WISE we project the status of water bodies we in comprehensive values on their ecological status and chemical status. on water quality status and weighted population distribution we estimated the percentage of population living in areas described as having water in either poor or bad ecological status. We find that the status of the water bodies in the respondents’ area of residence does not affect the respondents’ selection of algae growth, chemical pollution, and changes to water ecosystems, as threats to the water environment. Conversely, higher education, increased sensitivity and perceived informedness towards environmental issues increase the chances of respondents selecting the above three pressures as perceived threats. Our results are extremely robust, even when removing the higher or the lower 25% of the observations.
Our analysis confirms the disparities previously observed between the perceptions of policy- makers and of EU citizens. Characteristic to that is the fact that 73% of respondents in the 2012 Flash Eurobarometer stated that the EU should propose additional measures for water problems to be addressed, while only 7% claimed to be aware of the policies the EU is planning to put place to address water policy issues. This result highlight substantiates the need for a proper communication of policy reforms to European general public by the EU, which should be underpinned by a more credible evaluation of climate-change-related risks.
The results of the multilevel analysis are very robust even when uncertainty is high as in the cases including respondents living in areas with 25% or more of their water environments classified as of unknown status, or when including residents coded as living in NUTS 1 areas, for all three stressors to the water environment (although not presented here because of space).
In our analysis, respondents who showed environmental-friendly attitudes, for example by buying eco-friendly house chemicals, had above 40% chances of selecting either algae growth or chemical pollution as a perceived threat to the water ecosystem in their country, which could support a claim of the existence of lexicographic preferences in the study sample. The use of eco-friendly house chemicals as well as the existing perception that water bodies have deteriorated affected positively all three pressures to the water environment being selected by respondents as threats, while being significant at the 1% level. At a lower degree, the percentage of water bodies being in bad or poor status in the area where the respondents live also increased the likelihood of selecting chemical pollution and changes to the water ecosystems as pressures to water but less than other deciding factors.
The findings of this study are overall consistent with the relevant literature. Results are consistent with Weber (2010) in that we also found that familiarity with water related issues does not increase significantly risk perceptions of climate change. However, we found that a higher level of education increases the probability of a respondent to select at least one of the three stressors to the water environment in their risk perceptions, as found by Kahan et al. (2012), but contrary to Etkin and Ho (2007) findings.
Furthermore, Sagoff (1994) claims that people when faced with decisions that concern their whole community act in an altruistic way choosing what is better for all instead of what is best for them only. According to Sagoff, such an argument contradicts the proposition that individuals aim to maximize their own individual welfare since, in certain contexts, they can act in an altruistic manner instead of only in a deterministic one. If we assume that a Eurobarometer study is seen by respondents as a lever to raise awareness on threats to the water environment to the EU and the member states, then Sagoff’s argument may explain why when citizens formulate their perceptions of the risks to the national water resources because of, for example, chemical pollution, the actual status of regional water bodies weighs less in that decision than other factors (despite being statistically significant).
This can be also explained by the so-called lexicographic preferences (van der Bergh 2002, pp 815); when individuals are faced with environmental decisions or statements, they have been observed to state conflict-avoiding opinions and choices. If respondents are favourable towards environmental protection, their ethical attitudes have been observed to promote lexicographic preferences in questions about environmental issues and quality (van der Bergh 2002, pp 815).
Assessing risks related to the pressures placed on the environment due to climate change is a procedure that requires rigor. Zenghelis and Stern (2016) in their report propose to policy- makers to “articulate and unbundle material risks” related to climate change and “tackle the marginalisation of nonphysical risk” among other suggestions. Material risks, such as the ones described as pressures in this paper, can be rather complicated and therefore difficult for individual businesses and economies as a whole to manage. Therefore, Zenghelis and Stern call for a more coherent taxonomy of terms related to climate risks and impacts. Although full hierarchisation is not possible, some assessment of likely risks is both required and possible. At the same time, there is a need of pricing plans and costs that embody non- physical risks. Zenghelis and Stern also claim that the magnitude of the disparity between policymakers and businesses on how risks are understood is so large that either asymmetric information or a lack of credibility in climate mitigation and risk-evaluation policies are responsible. Hence, a more coherent approach on climate change risks and perceptions is required.
The overall result of our analysis is that risk perceptions for EU citizens are not affected by the status of the water bodies in their region too much, but rather by the overall pressures in the water environment of their country, and by their pre-existing beliefs and cognitive processing methods. With an ever-growing disparity between the perceptions of the public and those of policy and decision makers on what constitutes a threat to the environment, this study aims to shed light on how to reduce this gap in different perceptions. In a European framework where high environmental quality standards are set for coastal, transitional, groundwater and freshwater ecosystems, citizen participation and engagement is key. Although respondents might respond altruistically or have lexicographic preferences, it still appears that nation-wide issues of the water environment affect risk perceptions of EU citizens instead of region-wide ones. Communicating effectively information about climate change and possible future threats is paramount given the need of the European Commission to include more public consultation and encourage public participation to its EU-wide policies. Our analysis aims to inform and help policy makers in their policy design and regional focus of top-down policy measures and directives.
Reference list
Beklioglu, M., Altinayar, G., & Tan, C. O. (2001, October). Role of water level fluctuations, nutrients and fish in determining the macrophytedominated clear-water states in five Turkish shallow lakes. In International Workshop (pp. 28-30).
Blenckner, T., Adrian, R., Livingstone, D.M., Jennings, E., Weyhenmeyer, G.A., George, D., Jankowski, T., Järvinen, M., Aonghusa, C.N., Nõges, T. and Straile, D., 2007. Large‐scale climatic signatures in lakes across Europe: A meta‐analysis. Global Change Biology, 13(7), pp.1314-1326.
Brack, W., Dulio, V., Ågerstrand, M., Allan, I., Altenburger, R., Brinkmann, M., Bunke, D., Burgess, R.M., Cousins, I., Escher, B.I. and Hernández, F.J., 2017. Towards the review of the European Union Water Framework Directive: Recommendations for more efficient assessment and management of chemical contamination in European surface water resources. Science of the Total Environment, 576, pp.720-737.
Brinson, M.M., Malvárez, A.I., 2002. Temperate freshwater wetlands: types, status, and threats. Environ. Conserv. 29, 115–133.
Brody, S.D., Zahran, S., Vedlitz, A. and Grover, H., 2008. Examining the relationship between physical vulnerability and public perceptions of global climate change in the United States. Environment and behavior, 40(1), pp.72-95.
Bunn, S. E., & Arthington, A. H. (2002). Basic principles and ecological consequences of altered flow regimes for aquatic biodiversity. Environmental management, 30(4), 492-507.
Carpenter, S. R., Stanley, E. H., & Vander Zanden, M. J. (2011). State of the world's freshwater ecosystems: physical, chemical, and biological changes. Annual review of Environment and Resources, 36, 75-99.
Casado, S., Florín, M., Mollá, S., Montes, C., 1992. Current status of Spanish wetlands. In: Finlayson, C.M., Hollis, T., Davis, T. (Eds.), Managing Mediterranean Wetlands and their Birds for the Year 2000 and Beyond IWRB Special Publication No. 20. International Waterfowl and Wetlands Research Bureau, Gloucester, UK, pp. 56–58
DG Environment (European Commission): Water Note 9, Water Notes on the implementation of the Water Framework Directive, December 2008,
Coops, H., Beklioglu, M., & Crisman, T. L. (2003). The role of water-level fluctuations in shallow lake ecosystems–workshop conclusions. Hydrobiologia, 506(1-3), 23-27.
Dunlap R.. “Lay perceptions of global risk: public views of global warming in cross-national context”. International Sociological Association 1998, 13:473–498
Dunlap, R. E., & Jacques, P. J. (2013). Climate change denial books and conservative think tanks exploring the connection. American Behavioral Scientist, 57(6), 699-731.
Etkin, D., & Ho, E. (2007). Climate change: perceptions and discourses of risk. Journal of risk research, 10(5), 623-641.
European Commission: Flash Eurobarometer 344: Attitudes of Europeans towards water- related issues. March 2012.TNS Political & Social [Producer];GESIS Data Archive: ZA5779, dataset version 1.0.0. doi: 10.4232/1.11585.
European Union, 2013a. Decision No 1386/2013/EU of the European Parliament and the Council of 20 November 2013 on a General Union Environment Action Programme to 2020 “Living Well, Within the Limits of Our Planet”. OJ L345. 28.12.2013. p. 171.
Franssen, H.J. and Scherrer, S.C., 2008. Freezing of lakes on the Swiss Plateau in the period 1901–2006. International Journal of Climatology, 28(4), pp.421-433.
Glibert, P. M., Anderson, D. M., Gentien, P., Granéli, E., & Sellner, K. G. (2005). The global, complex phenomena of harmful algal blooms.
Hering, D., Borja, A., Carstensen, J., Carvalho, L., Elliott, M., Feld, C. K., ... & Solheim, A. L. (2010). The European Water Framework Directive at the age of 10: a critical review of the achievements with recommendations for the future. Science of the total Environment, 408(19), 4007-4019.
Halling-Sørensen, B., Nielsen, S.N., Lanzky, P.F., Ingerslev, F., Lützhøft, H.H. and Jørgensen, S.E., 1998. Occurrence, fate and effects of pharmaceutical substances in the environment-A review. Chemosphere, 36(2), pp.357-393.
Howel, D., Moffatt, S., Prince, H., Bush, J. and Dunn, C.E., 2002. Urban air quality in North‐East England: Exploring the influences on local views and perceptions. Risk Analysis, 22(1), pp.121-130.
Intergovernmental Panel on Climate Change (IPCC) (2013) Climate Change 2013. The Physical Science Basis (Working Group I Contribution to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change; Summary for Policymakers)
Johansson, M. E., & Nilsson, C. (2002). Responses of riparian plants to flooding in free‐flowing and regulated boreal rivers: an experimental study. Journal of Applied Ecology, 39(6), 971-986.
Joint, E. E. A. J. R. C. W. H. "Impacts of Europe's changing climate—2008 indicator-based assessment." (2008).
Jol, A., Raes, F. and Menne, B., 2009. Impacts of Europe's changing climate–2008 indicator based assessment. In IOP Conference Series: Earth and Environmental Science (Vol. 6, No. 29, p. 292042). IOP Publishing.
Leiserowitz, A. A. (2003). Global warming in the American mind: The roles of affect, imagery, and worldviews in risk perception, policy preferences and behavior (Doctoral dissertation, University of Oregon).
Leiserowitz, A. (2006). Climate change risk perception and policy preferences: The role of affect, imagery, and values. Climatic change, 77(1), 45-72.
Kahan DM, Peters E, Wittlin M, Slovic P, Ouellette LL, Braman D, et al. (2012) The polarizing impact of science literacy and numeracy on perceived climate change risks. Nature Climate Change 2(10): 732–735
Kirby, M.F., Allen, Y.T., Dyer, R.A., Feist, S.W., Katsiadaki, I., Matthiessen, P., Scott, A.P., Smith, A., Stentiford, G.D., Thain, J.E. and Thomas, K.V., 2004. Surveys of plasma vitellogenin and intersex in male flounder (Platichthys flesus) as measures of endocrine disruption by estrogenic contamination in United Kingdom estuaries: temporal trends, 1996 to 2001. Environmental toxicology and chemistry, 23(3), pp.748-758
Loos, R., Gawlik, B. M., Locoro, G., Rimaviciute, E., Contini, S., & Bidoglio, G. (2009). EU-wide survey of polar organic persistent pollutants in European river waters. Environmental Pollution, 157(2), 561-568.
Lorenzoni, I., & Hulme, M. (2009). Believing is seeing: laypeople's views of future socio- economic and climate change in England and in Italy. Public Understanding of Science, 18(4), 383-400.
Malaj, E., Peter, C., Grote, M., Kühne, R., Mondy, C. P., Usseglio-Polatera, P., ... & Schäfer, R. B. (2014). Organic chemicals jeopardize the health of freshwater ecosystems on the continental scale. Proceedings of the National Academy of Sciences, 111(26), 9549-9554.
Martin-Ortega, J., Balana, B., Perni, A. and Slee, B., 2012. Cost-effectiveness analysis report for the Thame sub-catchment including analysis of disproportionality.
Mayer, A., 2016. Risk and benefits in a fracking boom: Evidence from Colorado. The Extractive Industries and Society, 3(3), pp.744-753.
McCright A. The effects of gender on climate change knowledge and concern in the American public. PopulEnviron 2010, 31:66–87.
McCright, A.M., Dunlap, R.E. and Xiao, C., 2013. Perceived scientific agreement and support for government action on climate change in the USA. Climatic Change, 119(2), pp.511-518.
McNeeley S, Huntington O. “Postcards from the (not so) frozen North: talking about climate change in Alaska.” In: Moser SC, Dilling L, eds. Creating a Climate for Change: Communicating Climate Change and Facilitating Social Change. Cambridge: Cambridge University Press; 2007, 139–152.
Miraglia, M., Marvin, H. J. P., Kleter, G. A., Battilani, P., Brera, C., Coni, E., ... & Filippi, L. (2009). Climate change and food safety: an emerging issue with special focus on Europe. Food and chemical toxicology, 47(5), 1009-1021.
Morton TA, Rabinovich A, Marshall D, Bretschneider P. “The future that may (or may not) come: how framing changes responses to uncertainty in climate change communications”. Global Environ Change 2011, 21:103–109
Nilsson C, Jansson R, Zinko U. 1997. Long-term responses of river-margin vegetation to water-level regulation. Science 276: 798–800.
Nilsson, C., & Berggren, K. (2000). Alterations of Riparian Ecosystems Caused by River Regulation: Dam operations have caused global-scale ecological changes in riparian ecosystems. How to protect river environments and human needs of rivers remains one of the most important questions of our time. BioScience, 50(9), 783-792.
Nilsson, C., Reidy, C. A., Dynesius, M., & Revenga, C. (2005). Fragmentation and flow regulation of the world's large river systems. Science, 308(5720), 405-408.
Nisbet MC, Myers T. : “Twenty years of public opinion about global warming.” Public Opinion Quart2007, 71:444–470
NOAA National Centers for Environmental Information, State of the Climate: Global Analysis for Annual 2016, published online January 2017, retrieved on January 30, 2017 from http://www.ncdc.noaa.gov/sotc/global/201613
Nordhaus, William D. The climate casino: Risk, uncertainty, and economics for a warming world. Yale University Press, 2013.
Paerl, H.W., Fulton, R.S., Moisander, P.H. and Dyble, J., 2001. Harmful freshwater algal blooms, with an emphasis on cyanobacteria. The Scientific World Journal, 1, pp.76-113.
Peperzak, L. (2005). Future increase in harmful algal blooms in the North Sea due to climate change. Water Science and Technology, 51(5), 31-36.
Rabe-Hesketh, S., Skrondal, A. and Pickles, A., 2005. Maximum likelihood estimation of limited and discrete dependent variable models with nested random effects. Journal of Econometrics, 128(2), pp.301-323.
Raudenbush, S.W. and Bryk, A.S., 2002. Hierarchical linear models: Applications and data analysis methods (Vol. 1). Sage.
Rivetti, C., López-Perea, J.J., Laguna, C., Piña, B., Mateo, R., Eljarrat, E., Barceló, D. and Barata, C., 2017. Integrated environmental risk assessment of chemical pollution in a Mediterranean floodplain by combining chemical and biological methods. Science of The Total Environment.
Sagoff, M. (1994). Should preferences count?. Land Economics, 127-144.
Spalding, MD; Brumbaugh RD; and Landis, E (2016). Atlas of Ocean Wealth. The Nature Conservancy. Arlington, VA
St. Louis, V. L., Kelly, C. A., Duchemin, É., Rudd, J. W., & Rosenberg, D. M. (2000). Reservoir Surfaces as Sources of Greenhouse Gases to the Atmosphere: A Global Estimate: Reservoirs are sources of greenhouse gases to the atmosphere, and their surface areas have increased to the point where they should be included in global inventories of anthropogenic emissions of greenhouse gases. BioScience, 50(9), 766-775.
Stern N. (2006). "Stern Review on The Economics of Climate Change (pre-publication edition). Executive Summary"
Wauchope, R.D., 1978. The pesticide content of surface water draining from agricultural fields—a review. Journal of environmental quality, 7(4), pp.459-472.
Weber, E. U. (2010). What shapes perceptions of climate change?. Wiley Interdisciplinary Reviews: Climate Change, 1(3), 332-342.
Whitehead, P.G., Wilby, R.L., Battarbee, R.W., Kernan, M. and Wade, A.J., 2009. A review of the potential impacts of climate change on surface water quality. Hydrological Sciences Journal, 54(1), pp.101-123.
Wolf, J., & Moser, S. C. (2011). Individual understandings, perceptions, and engagement with climate change: insights from in‐depth studies across the world. Wiley Interdisciplinary Reviews: Climate Change, 2(4), 547-569.
Van den Bergh, J.C. ed., 2002. Handbook of environmental and resource economics. Edward Elgar Publishing.
Vethaak, A.D., Lahr, J., Schrap, S.M., Belfroid, A.C., Rijs, G.B., Gerritsen, A., de Boer, J., Bulder, A.S., Grinwis, G.C., Kuiper, R.V. and Legler, J., 2005. An integrated assessment of estrogenic contamination and biological effects in the aquatic environment of The Netherlands. Chemosphere, 59(4), pp.511-524.
Vos, J.G., Dybing, E., Greim, H.A., Ladefoged, O., Lambré, C., Tarazona, J.V., Brandt, I. and Vethaak, A.D., 2000. Health effects of endocrine-disrupting chemicals on wildlife, with special reference to the European situation. Critical reviews in toxicology, 30(1), pp.71-133.
Zenghelis, D. and Stern, N., 2016. The importance of looking forward to manage risks: submission to the Task Force on Climate-Related Financial Disclosures.
Appendix
Source: Flash Eurobarometer on Water, Analytical Report, 2012
Chapter Eight: Conclusion and contributions: Concluding remarks on Economic Theory and findings of the thesis
1. Contribution to policy-making: Abatement cost, Cost- benefit analysis and analysis of disproportionality
Economic policy and the subsequent decision-making progress were two of the main points of emphasis of this thesis. Given the novelty of the layout of the WFD which included and demanded a heavy “portion” of policy-making to be informed by economic theory, welfare changes and economic theory were considered in this volume. As the review of the WFD is coming in 2019, this thesis aims to contribute to the evaluation of the Directive and to inform future Directives and measures between the member states. The results from the abatement costs for the four different mitigation measures presented in chapter 4 showed that abatement costs, regardless of mitigation measures and/or climate change scenarios, proved that costs were too high. Costs were high in terms of implementing the various mitigation measures and they also burdened disproportionately more farmers than any other type of stakeholders. Interestingly enough, the WFD followed an approach that can be described “anthropocentric” in its core, by placing human welfare ultimately as the decisive factor in decision-making processes related with the WFD. Although the attainment of GES across the EU and in all water bodies was the ultimate goal of the Directive, the primary focus was human welfare. WFD mandates make clear that if the costs are too high for some groups of stakeholders, if achieving GES proves to disruptive for a part of an RDB or it requires serious investment from the side of local governments that would lead to substantial welfare loss for local stakeholders, then, implementing the WFD was not required.
Furthermore, Marginal Abatement Cost Curves (MACCs) in every mitigation measure were high and reductions in fertilizers (thus, costs) had to be substantial in order for their respective MACCs to deviate from any other agri-environmental measure proposed in earlier EU legislation, such as the Common Agricultural Policy (CAP). This raises the question on the real impact of the WFD, both in terms of novelty in the policy side but also in terms of efficiency given the extent of research, funds and time devoted to the designing, assessing, valuing, implementing and monitoring the whole directive and its progress across Europe. These results though are also indicative of the pre-existing water status of the Louros’ watershed which was in a good status and achieving GES required significant start-up costs that “derailed” MACCs. Of extreme interest is the focus of the EU to fund water-related issues in member states. It is evident that the E.U. invested heavily on the implementation of the WFD. There are no available figures of the total spending from the E.U.’ or from member states’ part. Still, evident of how invested the Community was to the Directive is the total allocation of funds that member states could “tap into” to fund projects aiming at various issues related to water resources. There were various funding opportunities, each aiming to provide financial aid to member states as they dealt with the implementation of the WFD and with issues to their water resources. The Cohesion Policy Fund for the period 2007-2013 had a total budget of 344 billion Euros with the goal to fund “capital-intensive investments in water infrastructure and help EU Member States comply with water legislation.”11. The primary focus of the Fund was to divert funds to countries whose Gross National Product (GNP) was below 90% of the E.U. average. The Fund’s focus was mainly on water management (namely: prevention of disaster and risks and improving efficiency). Similar to the Cohesion Policy Fund but with a considerably smaller budget is the “Life + Funds” initiative. This focuses on supporting the implementation of the WFD as well as raising awareness on water-related issues. The total budget for the period between 2007 and 2013 was 1,7 billion, with small country-wide projects being awarded sums up to 1 million Euros. It needs to be noted that these sums of money were not ultimately spent by member states, there were available, if specific terms were met. Their significance is that they demonstrate the level of financial commitment from the E.U.’s part to the issue of water.
The findings of the disproportionality analysis revealed which groups were the primary cost bearers and the hydrological modeling along with the various climate change scenarios provided useful insights on the types of stressors on the environment. Although the proposed mitigation measure (mitigation measure 2) was the least “invasive” to the environment, still it yielded a negative benefit in the CBA. This can be attributed to the nature of the mitigation practices and to the existing water quality of the case study area. The mitigation practices were rather “vertical” in their nature, being expensive for the farmers and all of their costs being mandatory to occur every year. The water quality of the Louros catchment was not in a bad status in terms of ecological status but still, for it to achieve GES required considerable economic costs. Hypothetically, if other types of mitigation practices were more appropriate given the economic practices taking place in the Louros catchment, MACCs could be less steep and yielding considerably different values than the “no-change” scenario. Thus, the “technology” of cultivating nitrogen fixing legumes on set aside land ensures that, under any climate change scenario, abatement will be higher at any chosen cost level. The cost-benefit analysis in all scenarios did not yield positive Net Present Value (NPV). Although this is not desirable in environmental policy and economics, still, these results can be of great use. One of the most evident deficiencies in the CBA in chapter 6 of this thesis was the misrepresentation of benefits. Only the residents of the three municipal units were considered to be benefited from achieving GES when the river offers a multitude of services to non-residents as well. Apart from the energy generation from the hydroelectric dam at St. Georgios which provides electricity to an area exceeding the boundaries of the Louros catchment, the river has recreational value since activities such as rafting, canoeing, angling and hiking take place there. One can only assume that with the Ionia Odos highway being almost complete, even more visitors will be able to access Western Greece and Louros will attract even more visitors. Furthermore, even direct benefits from achieving GES could be expanded to residents of areas more distant than Arta and Preveza, such as residents from Ioannina, the largest residential area of northwestern Greece.
11 Source: http://ec.europa.eu/environment/water/quantity/instruments.htm 2. Contribution to economic theory: Risk perceptions and environmental goods
Environmental goods are notoriously difficult to assess as, usually, there are no markets for them to be bought and sold. Neoclassical economic theory allows for the creation of “quasi” markets by assuming that there are close substitute goods that can be traded in the hypothetical markets that are created. This allows for the exchange of goods, achieving equilibrium and the estimation of prices and surpluses.
The focus of this thesis is on the environmental good of water quality, (not just water, either drinking, bating, saline or freshwater) with a particular focus on freshwater (although the WFD includes coastal waters in its assessment as it requires the GES to be achieved at least up to 6 nautical miles from the coast of member states). The reason behind that was to establish a well-defined and easily conceivable good whose introduction to a hypothetical market would be relatively easier than other more broadly defined environmental goods. Additionally, as the Eurobarometer studies of 2009 and 2012 show, there is an interest from the public on the issues that affect water quality, amongst others. We also were able to utilize the substantial literature already existing in the field of environmental goods, ecosystem services and environmental economics and valuation, especially the fact that several primary studies on the valuation of freshwater quality already existed in a European context.
Environmental goods are greatly debated in terms of their nature, if they are normal, inferior or luxury goods. Depending on the value of the price elasticity, goods are classified as above. The focus of this thesis though revolves around policy measures and how these impact human welfare and therefore examining distributional effects of policy measures such as the WFD was more appropriate. This is achieved through the calculation of the elasticity of WTP which informs on the specific issue of the distribution of benefits across poor and rich households, determining which policy measures are “pro-poor” (elasticity of WTP lower than 1), “pro-rich” (elasticity of WTP higher than 1), or equally distributed (elasticity of WTP equal to 1).
The results from chapter 5 of this thesis contribute on this discussion by yielding robust results in terms of magnitude, sign and level of significance of the coefficients, even when the sample is trimmed by as much as 15% from below or above or when certain influential case studies are withdrawn. In most of the econometric methods used, elasticities of WTP were above 1 and up to around 1,5. These results though are dependent on the income measured used to estimate WTP. When official data were used, obtained by Eurostat for the respective NUTS 2 regions of the primary studies, income elasticity of WTP were above 1 meaning that the proportion of WTP to the income measure is increasing as income rises and thus, achieving GES has proportionately higher benefits to richer households than to poorer households. If the explanatory variable of WTP is the income reported by the primary studies, then, income elasticity is below the unit and benefits from achieving GES benefit more the poorer households than the rich ones. These findings reveal the serious policy implications of selecting reliable income estimates and contribute to the debate of whether selecting income as an explanatory variable of WTP for environmental goods and services is required. Furthermore, the policy implications about the distributional effects on income by achieving GES strengthen the views of Hokby and Sorderquvist (2003) that when such distributional concerns are the subject of a cost-benefit analysis, the introduction of weights or at least a sensitivity analysis is required for policy appraisal. This finding also aligns with the broader discussion started by Aaron and McGuire (1970) on the distributional nature of public goods and the wider benefits and positive externalities that they offer. Aaron and McGuire also make the case that the very nature of public goods affects more the “working poor”, an issue to be considered in public budget design and spending.
This thesis went a step “backwards in its last chapter, attempting to combine environmental services and the results of environmental and ecological monitoring in an effort to explain human behaviour. We did so by using EU-wide data concerning perceptions of EU citizens and biophysical, chemical, geomorphological and ecological data (the product of new and more demanding monitoring practices as new EU directives such as the Water Framework Directive, the Nitrates Directive and the Marine Strategy Framework Directive). The purpose behind it was to have an overall view on how these ambitious mandates which are deployed to augment human welfare, while having a great financing, implementation and monitoring costs, fare in the perceptions of individuals. Given the heightened willingness of the EU Commission to include public consultation in the designing stages of such directives, citizens’ input becomes more valuable. The simple question behind chapter eight was “what do people view as risks to the water environment of their country, given the actual status of their region’s water bodies”?
The responses gave us robust estimates of risk perceptions. Risk perceptions on water issues appear to be determined by various factors, all of them not depending significantly though on the status of the respondents’ water bodies. Instead, pre-existing environmental awareness, high levels of education and sensitivity towards the environment appear to be more determining factors of formulating risk perceptions. Overall, risk perceptions for EU citizens are not affected by the status of the water bodies in their region so much, but rather by the overall pressures in the water environment of their country, and by their pre-existing beliefs and cognitive processing methods.
Such findings are mostly consistent with the wide literature of pressures to the environment, most notably perceived pressures from climate change, and shed light in the assessment of top-down policies such as the Water Framework Directive. The key recipients of the benefits of such policies appear to be little informed of the actual pressures of their regions, despite the fact that they perceive themselves as being informed and aware. Additionally, although management of water bodies is mostly assigned to regional bodies (such as the relevant water management plans designed and delivered by each EU prefecture having water bodies), the design of those directives is conducted in an EU-level and many times appears distant of what EU citizens demand. The inclusion of public consultation as an important factor in such policies appears at first to be solving such problems but the findings of chapter 7 demonstrate that the public doesn’t take much into account the actual problems in their regions’ water bodies, or that there is still considerable lack of information for those issues in EU prefectural-level.
3. Contribution of this thesis on research design and focus: modeling
This thesis heavily relies on the results from two hydrological and land-use models. It used official data on water quality indicators, collaborating with scientific university departments and applying them to the Louros catchment. Although the availability of biophysical and chemical data was not always up-to-date (few data collection points and data sometimes coming from older studies), still, they were enough to inform the INCA (for nitrates and phosphorus) hydrological model to simulate the P concentrations in water under the baseline and the alternative mitigation measures. The INCA( N and P) model prides itself for incorporating as many aspects as possible that affect the water environment such as climate conditions (especially precipitation and temperature), physical conditions (slopes, elevation), soil conditions (especially soil chemistry), hydrography (density of watercourses and manmade interventions such as the irrigation network and drainage ditches). The main drivers and focus of the model was nutrient and sediment transport. The complexity of the model also allowed for climate change scenarios to be incorporated later on on the analysis to project the effects of policy measures according with the four main IPCC (2007) scenarios.
Land-use was modeled after combining the four different mitigation measures that focused both on agricultural practices but also in the types of cultivations, combining “practice” and “technology”. These mitigation measures and their future projections were modeled after the IPCC scenarios and the country-specific measures published from the Central Bank of Greece.
The results were particularly interesting, especially in terms of water quality. Implementing the WFD mandates would yield only marginal improvements in water quality across the catchment as it was already in good status. Additionally, regardless of the climate change projections, all mitigation measures reduced, again, only marginally the nitrate and phosphorus concentrations.
References
Aaron, H. and McGuire, M., 1970. Public goods and income distribution. Econometrica: Journal of the Econometric Society, pp.907-920.
Hökby, S. and Söderqvist T. 2003. Elasticities of Demand and Willingness to Pay for Environmental Services in Sweden. Environmental and Resource Economics, 26(3), 361-383. DOI:10.1023/B:EARE.0000003581.97411.75