MSc Program Engineering Management

Accelerated crypto currencies: A systematic analysis of Silvio Gesell’s theories and its implementation on local crypto currencies

A Master's Thesis submitted for the degree of “Master of Science”

supervised by Dr. Larry Stapleton PhD, MA, BSc (Hons), CPIM

Fabian Janisch, BSc

01226430

Vienna, 15.03.2021

Affidavit

I, FABIAN JANISCH, BSC, hereby declare

1. that I am the sole author of the present Master’s Thesis, "ACCELERATED CRYPTO CURRENCIES: A SYSTEMATIC ANALYSIS OF SILVIO GESELL’S THEORIES AND ITS IMPLEMENTATION ON LOCAL CRYPTO CURRENCIES", 127 pages, bound, and that I have not used any source or tool other than those referenced or any other illicit aid or tool, and 2. that I have not prior to this date submitted the topic of this Master’s Thesis or parts of it in any form for assessment as an examination paper, either in Austria or abroad.

Vienna, 15.03.2021 ______Signature Abstract

Thismaster’s thesisevaluated the acceptance of crypto currencies, under the impact an existing localFreigeld crypto currency.The intersectionsofthe underlying technology of crypto currencies with aspecific part of Silvio Gesell’s Freigeldtheory namedeffective demand,builtthe basis of alocal Freigeld cryptocurrency scenario. Thisscenario wasthen used to evaluate the level of acceptancebycomparing thescenario to differentcrypto currencies. Participantsofthis study were selected by snowball sampling.The data wasgatheredbyawebbrowser based self-administered survey. The results showedthat the level of acceptancefor thelocal Freigeld cryptocurrency exceedsthe level of every presented cryptocurrency.Additionally,this studyanalyzed factors, to explainthe cause of thedifferent level of acceptance of thelocal Freigeld crypto currency.The evaluation showedthat thesupport of the local economywas theonly significant impact which caused the increasedacceptance of thelocal Freigeld crypto currency.

Kurzfassung

DieseMasterarbeit evaluierte die Akzeptanz vonKryptow¨ahrungen, unter dem Einfluss der Verf¨u gbarkeit einer lokalen Freigeld-Kryptow¨a hrung. Die Ub¨ erschneidungen,bez¨u glichder zu- grundeliegenden Technologie vonKryptow¨ahrungen,und eines Teils der Freigeld-Theorie, genannt effectivedemand,bildeten dabeidie Basis des entwickelten Szenarios einer lokalen Freigeld- Kryptow¨ahrung. DieseswurdeimZug einerUmfrage verwendet, um die Auswirkungen auf die Akzeptanz vonKryptow¨ahrungenzuuntersuchen.Die Daten wurden mithilfe einer eigenst¨andig auszuf¨u llenden,onlineUmfrage erhoben. DieVerteilung erfolgt durchdas Schneeballsystem. Das Ergebnisder Studiezeigte, dass dieAkzeptanz der lokalen Freigeld-Kryptow¨ahrung h¨oher ist alsdie Akzeptanz aller evaluiertenKryptow¨ahrungen.Dar¨u berhinausuntersuchtediese Ar- beit Faktoren, die eine And¨ erung der Akzeptanz erkl¨arenk¨o nnten.Dabei zeigtelediglichdie Unterst¨u tzungder lokalen Wirtschafteinensignifikanten Einfluss aufdie erh¨o hteAkzeptanz der lokalenFreigeld-Kryptow¨ahrung.

I Acknowledgements

First of all, Iwould liketothank my supervising professor, Larry Stapleton. He gave me the freedom to develop atopic accordingtomyown ideas and theconfidence to write this master’s thesis.Furthermore, besides hisadvice, whichwas consistentlyonpoint,healwaysfound time to answermyquestionsand e-mails.Ireally enjoyedworkingwith him.

Ialso would liketothank everyone who helpedmetocomplete this study.Iam reallygrateful forsomanygoodfriends and Iamsorry for bothering youwith the questionnaire as well as with the revision of my thesis. Youreallyearned your spot in this acknowledgments.

Manythanks to:Albert M.,DanielJ., Julia K.,Norbert H.,David R., DanielV., Andreas B., Christian K., Andreas M.,StefanI., Judith L., Vera S., Elisha I., Naemi L., Johannes T.,Manon M.,KatrinL., PatrickB., Michael G., Nikolas A., Lukas H., Katja S.,NikolasR., Tiki B., Nadja R., Monika S., Laurin T., KatharinaS., BenjaminL., Andreas P.,Victoria S, Markus S., Natalie S.,Nicole I., PatrickS., andofcourse to everyone who Iunfortunately overlooked whilst creating this list.

My deepest gratitude goes to my sister Yara whotruly supports me just by being herself and my parents Monika and Hans who supported me notonly during this master’s degree,but during my wholelife.

THANK YOU!

II TableofContents

1Introduction 1 1.1Motivation ...... 1 1.2Why SilvioGesell is applicabletoday ...... 2 1.3Problem statement ...... 3 1.4Aims of the study ...... 3 1.5Structure ...... 3

2Background 5 2.1Freigeld ...... 5 2.1.1 Structure ...... 5 2.1.2 Silvio Gesell’s problem statement ...... 6 2.1.3 The origin of money ...... 7 2.1.4 The trade with money (andothercommodities) ...... 7 2.1.5 The three premises of money ...... 8 2.1.6 The backing of money ...... 9 2.1.7 The value of money ...... 9 2.1.8 Supply, demand andneeds ...... 11 2.1.9 The quantity equation ...... 13 2.1.10 The idea of Freigeld ...... 16 2.1.11 The effective demand ...... 17 2.1.12 Postulated impactsofFreigeld ...... 18 2.2Blockchain and cryptocurrencies ...... 20 2.2.1 Structure ...... 20 2.2.2 Blockchain ...... 21 2.2.3 Ledger ...... 21 2.2.4 Hash code ...... 21 2.2.5 Distributed ledger ...... 22 2.2.6 Consensus mechanism ...... 22 2.2.7 Synchronizeddistributed ledger ...... 23 2.2.8 Open synchronizeddistributedledger ...... 23 2.2.9 Cryptocurrencies ...... 23 2.2.10 A“trustless”system ...... 24 2.3Sustainabilityand complementarycurrencies ...... 24 2.3.1 Quantitative Economic Development ...... 24 2.3.2 Complementary currencies ...... 25

III 2.4 Accelerated crypto currencies ...... 26 2.4.1Structure ...... 26 2.4.2Evaluationofthe three premises of money ...... 26 2.4.3Applicabilityofthe effectivedemand ...... 27 2.4.4The circle coin -alocalFreigeldcrypto currency ...... 28

3Methodology 31 3.1 Structure ...... 31 3.2 Sampledesign ...... 31 3.2.1Sample design workflow ...... 31 3.2.2Survey objective ...... 32 3.2.3Effectivityand efficiency ...... 33 3.2.4Total surveydesign ...... 34 3.2.5Representativesamples ...... 37 3.2.6Errorand bias ...... 38 3.2.7Validity ...... 40 3.3 Sampling techniques ...... 41 3.3.1Probabilitysampling techniques ...... 41 3.3.2Non probabilitysamplingtechniques ...... 43 3.3.3Single stage andmulti stage sampling ...... 45 3.3.4Selfadministeredonlinesurveys ...... 47 3.4 Definitions ...... 49

4Implementation 53 4.1 Originalsample ...... 53 4.1.1Total surveydesign: Original sample ...... 53 4.1.2Sampling design:Originalsample ...... 54 4.1.3Summary: Original sample ...... 61 4.1.4Limitations: Original sample ...... 62 4.2 Statement: Change of thesurveydesign ...... 62 4.3 Executed sample ...... 63 4.3.1Total surveydesign: Executed sample ...... 63 4.3.2Sample design: Executed sample ...... 63 4.3.3Summary: Executedsample ...... 66 4.4 Questionnaire design ...... 66 4.4.1Platform ...... 66 4.4.2General framework ...... 66 4.4.3Core concept ...... 68 4.4.4Classification ...... 71

5Results 73 5.1 Timeline,samplesize and response rate ...... 73 5.2 Demographics ...... 77 5.3 Findings concerning theuse of electronic payments ...... 80 5.4 The acceptance of cryptocurrencies ...... 83 5.4.1The readinesstouse cryptocurrencies ...... 83 5.4.2The trust in crypto currencies ...... 86 5.4.3The acceptance of cryptocurrencies ...... 86

IV 5.5Details concerningthe acceptance of crypto currencies ...... 88 5.5.1 The impact of theobjective named “long try-out” ...... 88 5.5.2 The impact of governmental support on cryptocurrencies ...... 88 5.5.3 The impact of differentparameters introducedbythe localFreigeld crypto currency ...... 89

6Discussion 93 6.1The readiness to usecrypto currencies ...... 93 6.2The trust in crypto currencies ...... 93 6.3The acceptanceofcrypto currencies ...... 94 6.4Impactsonthe levelofacceptance of cryptocurrencies ...... 95

7Conclusion 96 7.1Introduction ...... 96 7.2ResearchObjectives ...... 96 7.3Limitations ...... 97 7.3.1 The effective demand ...... 97 7.3.2 The executed sample ...... 98 7.4Broad implications and future work ...... 98 7.5Summary and closing ...... 99

AAppendix 102 A.1 Questionnaire ...... 102

Literature 116

Internet References 117

V Abbreviations

SME small and medium sizedenterprises

DLT distributed ledger technology

POW proofofwork

GDP grossdomesticproduct

RPI retailprices index

CPI consumer price index

QED quantitativeeconomicdevelopment

ZMR zentrales melderegister

GDPR general data protection regulation

VI Symbols

This table contains all symbols used in mathematical expressions.

Symbol Definition s Second t Time M Money in circulation V Velocityofcirculationofmoney M Bank deposits(depositcurrency) V Velocityofcirculationofdepositcurrency V · M Active quantityofmoney T Volumeoftrade P General price level R Resilience E Efficiency C Development capacity S Level of sustainability

VII 1Introduction

1.1 Motivation

Caprio et al.(1996) stated 96 bankingcrisesbetween 1980 and 1996. [CK96]Margrit Kennedy (2011) referred to the world bankwhichdocumented 124 bankingcrises, 326 currencycrisesand 64 governmentdebt crisesbetween 1970 and 2007. [Ken11,p.11] Hence, it can be saidthatthe recentpast shows aconsiderable amountofeconomiccrises. Althoughboomand bust cycles are widely acknowledged andacceptedwithin the economicscholarly society, the questioning of our capabilities to ease their impact increased considerably over therecentyears. After the outbreak of theeconomiccrisis in 2008, The Economist wrote:“With aflaweddiagnosis of thecauses of the crisis, it is hardlysurprisingthat many policymakers have failedtounderstand its progression.” [eco,p.13] Economic cruncheshaveanimpact on avast majorityofmarket participants, this alsoincludes small and medium sized enterprises(SME). In 2017, 99.6 %ofall Austrian enterprises were identi- fied as SME (less than 250 employees),employing 67 %ofall employees in Austriaand generating 61 %ofthe Austriangrossvalue added. [Aus19,p.16f]This shows the considerable importance of SMEsfor theAustrian society. Furthermore, their importance wasparticularlyoutlinedbythe international restrictionsduring the emergingcorona crisis in March 2020. Nonetheless, global- ization and competitiveinternational players continuetosuccessfully challenge this backbone of the Austriansociety,leaving thequestion of howthis process will effect the society’s resilience against the frequentcrisesand howthis will generally impact thesocietyovertime.One possible outcome mightbethat theincreasinginfluenceofbig internationalcompanies willstreamlineour society towardsmore efficiency.Thereby, the SME’s whichmight have been an useful balance weightfor the seemingly frequent economic crises are diminished. [UGLG09] Under thepremise of this effect, approaches have to be found to tackle the described issue of an efficiency driven global economy, which exhibits frequentboomand bust cycles.Looking back again on theeconomichistory of thepast50years, it can be said that theeconomicglobalization increased and added pressure on SMEs, but the implementedeconomicconcepts didnot change the recurrence of economiccrunches. Therefore the personal conclusion wasdrawn thatconcepts aside theconventional position might be useful forthe furtherdevelopmentofour monetary system. In this regard,theories which are determinedasmarginal region of economicscience should be reevaluated and therefore atheory published around 1900 by the entrepreneurSilvio Gesell waschosenasresearchtopic of this thesis.

1 Introduction

1.2 WhySilvioGesellisapplicable today

Silvio Gesellwas born in 1862 in St.Vith,Belgium and movedtoArgentina in 1887.The ongoing economicand socialcrisis in Argentina made him questionthe economic system.This resulted in his first book “Die ReformationimM¨u nzwesen als Br¨u cke zumsozialen Staat”.Itwas published in 1891 and already illustrated one of hiskey ideas to secure acontinuous moneycirculation by a concept named “stamped money”.Between 1891 and 1916 he livedinSwitzerland,Bavaria and Argentina.The dominant constantduring this periodofhis life were recurring (financial) crises. [Ilg15]Silvio Gesell broadened and updated histheories over time and published his most famous book “The natural economicorder” (originally in German “Dienat¨u rliche Wirtschaftsordnung”) in 1916. [Ges49]

Influenced by the economicsituation,Gesell tried to provide atheory to prevent (economic) crises and secure price stability. He arguesthat stabilitypromotes afeeling of contentmentinhuman beingsand afeeling of contentmentisthe basis forpeace between nations, people and the world. Instabilitywill lead to crises,whichinturnlead to sorrowand war.

In his monetary theory called “Freigeld” he ties (economic)crises and wars back to an unbalance between money and productsand thereby outlines that this unbalance is caused by the design of the conventional currency.Therefore he proposes an adaption of the latterand consequently of the monetary system, to ensure (price) stability.

Silvio Gesell’s monetary theorywas and is successfullyapplied in localareas. The most famous experiment wasexecuted in an Austriantownand is therefore called thewonder of ‘W¨orgl’. Despite itssuccess it wasshutdownbythe Austrian national bank in 1933. [Sch08,p.72] Since then Gesell’s ideas and alternative monetary concepts in general, didnot gain alot of attention.

Roughly 100years after thepublication of hismain bookand proposals,the basesofour monetary system, the economic boomand bust cycles including alltheir collateralsocialinstabilities, haven’t changed.The monetary system seemedindisputable.

This changed withthe recenteconomiccrisesof2001 and 2008 and thesimultaneous disruption of the financial world by the development of the distributed ledger technology (DLT)(blockchain and cryptocurrencies) in 2009. The situationofaneconomiccrunchcombinedwith the new concept of DLT’s created an environment-of-change.

Bigcentralized and/or global corporations (including banks) are challenged by this environment. Financial technology companies (called FinTechs) are emerging with newslim and fast financial solutions. DLTsconstantlyincreasetheirpresence in society.The idea of supportingthe local economyseems to experience an upswing withinthe thoughtsofthe young European generation. Common buzzwords are decentralization, regionalityand stability.

By bringing all theabove mentionedcircumstances together, it is evidentwhy the contemplated discussion of Silvio Gesell’s (Freigeld)theory matches thecurrent periodoftime. Furthermore, thenew developmentsenable to study his theories from anew perspective. Complementarylocal cryptocurrencies,basedonthe theories of Silvio Gesell, mayhave the potential to counter the issuesarising fromanefficiency drivenglobalizedeconomyand are therefore elaborated in more detail.

2 Introduction

1.3 Problem statement

SilvioGesell’s theories were classifiedapproximately100 years ago as insufficientand are therefor seen as marginal region of economicscience. Only afew well groundedresearch papers existand discussiononanacademiclevel is rare. Sincethe majorityofthe academicfielddoesn’tacknowledgeGesell’stheories apossiblere- consideration aside from thescholarly field of economics is considered to be likely.This refers especially to the domainofcrypto currenciesand theFreigeld wasthereforealready pickedupby acrypto currencynamed Freicoin [2]Inthis regard,itisassumed that furtherdevelopmentof the economic systemwithout scholarly support, is caused by the failure of the academic society to provide solutions demandedofthe prevailing environment-of-change. The altering of thecomplex economicsystem, without thesupport of the scholarsocietyincreases the riskofmisinterpretations or wrong assumptions.This mighthinderaprogressivedevelopment of our monetary system.

1.4 Aims of the study

The overallobjectiveofthis thesisistosupport the progressive furtherdevelopmentofthe monetary system by an accurateand justacademicdiscussion of the formerly rejected Freigeld theoryofSilvio Gesell. In detail, the applicabilityofeffective demand (a specificpart of the Freigeld theory)oncrypto currencies is evaluated. Basedonthese findings, abasicscenario of a local Freigeld cryptocurrency is developed.The scenario is then usedinasocial survey to assess the acceptance of thelocal Freigeldcrypto currency by comparingittothe acceptanceofother crypto currencies. Hence, this thesis discusses two mainparts:

1. The applicability of theconceptofeffective demand on cryptocurrencies.

2. Apreliminary investigation on howthe level of acceptance of crypto currencies is influenced by theavailabilityofalocal Freigeldcrypto currency.

Amore detailed description of the research objectives and testedhypotheses is providedinsec- tion 3.2.2 (Surveyobjective).

1.5 Structure

The ‘Background’ofthis thesis starts by describing the conceptofthe effective demand.There- fore, the line of argument from Silvio Gesell, whichultimately resulted in the developmentofthis concept is outlinedindetail. The aim wastoallowanaccurate assessmentofthe original work. Next, thetheoreticalbackground of blockchains and crypto currencies is provided. Thethird part of thebackgrounddetermines the notation of sustainability,whichenables an exact demar- cation and implementation of a complementary currency.The backgroundendsbydiscussing intersections of the previously outlined conceptsand the development of acomplementarycrypto currency,whichincludes the concept of effective demand.

3 Introduction

The‘Methodology’ofthe conducted socialsurvey providesdefinitions, proceedings and theoretical background for two sample designs. The reason for thedevelopmentoftwo designswas the situation in Austria, caused by the SARS-CoV-2virus,whichleadtoanecessary adaptionof the originally developed approach. Nonetheless, due to the noveltyofthe originaldesign it was decided to present both to the full extent. Adetaileddescriptionofbothsamplesetups (namedoriginalsample and executed sample)and astatement concerning the design change is providedinchapter 4,named‘Implementation’. The ‘Results’ofthe obtainedsurveydata are presented in chapter 5,followedbythe ‘Discussion’ in chapter 6.The thesis endswith the final ‘Conclusion’inchapter 7.The onlinesurvey is presented in appendix A.

4 2Background

2.1 Freigeld

2.1.1Structure

“Freigeld” is amonetary theorydeveloped by the entrepreneurSilvio Gesell around1900and comprehensivelypresented in hismost famous book”The naturaleconomicorder” published in 1916. [Ges16] The bookisseparated into twomain parts: “Freigeld” and “Freiland”.The nomenclature hasits origin in theGerman language, and itsliteraltranslation means “Free money” and “Free land”. Thetheories of Free money address theproperties of money and theaccompanying monetary system and thetheories of Free land addressland ownership and privateproperty. Asuitable translation of Freigeld is the English term demurragemoney or demurrage currency.Used to a lesser extentare the terms acceleratedmoney, acceleratedcurrency or stamped money.[Bla98] As outlined in thefollowing sections, eachofthe terms statedshows aconsiderable overlap with thecore conceptofFreigeld.Nonetheless, duetothe commonliteraltranslation of the term Free money,itisoften wrongly equated with the so called unconditional basic income.Tobeclear, neitherFreigeld, nor Freilandhaveanything to do with anyforms of unconditionalbasic income. After studying both theories it canbesaidthat Freigeld and Freiland arethe opposite of an unconditional basic income. The Freigeld theory canbedividedintotwo concepts: effective demand and Freigeld interest theory or capital theory.[Ges16]The latteristhe more prominent part of theFreigeldtheory and as emphasized by Ilgmann(2011), theconceptofeffectivedemand is often overlooked or neglected in scholarlyelaborations.[Ilg15,p.1] The term effectivedemand wasnot directlyassigned by Gesell. Nonetheless, it seems to be very accurateand is therefore usedfor this report. Consideringthe elaborations in Gesell’s bookand the aspired impact of the effectivedemand, it is evidentthatitbuildsthe basis of the Freigeld- interest-theory.Due to its relevance forthe stated researchquestion the concept of effective demand is specifically discussed by thisstudy.Inthis regard,the considerable criticism about the viewpointsortheories of Gesell have to be acknowledged. Therefore,particular emphasis wasplaced on presenting the effectivedemand in amethodical and structuredway,startingwith asubstantial descriptionofthe viewpoints of Gesellonmoney. This allows adetailed review of theunderlyingassumptions.Itwas also decided to support the elaborations witchaconsiderable amountofdirectcitationstoaccuratelyillustrate Gesell’s line of argument.

5 Background

It has be mentionedthat the outlined effects of Freigeldare notexhaustive. Acompletediscus- sion of all consequences would have exceeded thescopeofthis study. Nonetheless, the impacts presented concerning the effective demand are carefullydiscussed andenable aprofound under- standing of the concept of Freigeld.Additionally,ithas to be acknowledged thatthe statements and conclusions of Gesell’shave been writtenmore than 100 years ago and should be seeninthe contextofthe respectivetime. In general,this thesis does not evaluate anytheories of Gesell and no conclusions of its actual applicability are drawn.

2.1.2 Silvio Gesell’s problem statement

Duringthe introductionofhis Freigeld theory,Gesellargues thatmoneywas regularlycriticized and condemned throughout human history and still no preciseand lasting answer has ever been provided on whymoney should be the condemnation of humanity.This alsoincludes historically famous peoplelikePythagoras or Goethe.The failure to provide thisansweris, to hisbelief, caused by external circumstances, which directly influencedthe scientific society.Heelaborates thate.g. religion or natural sciences featurefar moreprestige than thescienceofmoney.Money wassomething an aspiring scientific mindshouldn’t worry about.Heoutlined that only afew fact driven scientistscould be botheredtoinvestigatethe (as he calledit) “stepchild of science”. He alsostatesthat this circumstance wasworsenedbythe widespread belief in the“value doctrine of money”. As described by Gesell, the doctrine backthenwas predominate. It postulates that money needs an “intrinsic physical value” to work. Commonly,this “intrinsic physical value” referstosilver or gold. [Ges49,p.90] Gesellargues thatnosupporter of this theory hasever provided adefinition which allowsanexact determination of this intrinsic or inner value.He therefore summed thevaluedoctrineupasfollows: “Gold has an attribute -the so called ‘value’, whichisgrown together withthe material in the same wayasits physical weight. This attribute is boundtogold, similarasits chemicalattributes, inseparable from gold (‘inner value’), unchangeable and indestructible (‘stable value’).The exis- tenceofthe ‘value’ is determinedbyweighting [Assumption: Against other products]. Currently no otherpossible way has beendeveloped.Itisnot effectedbythe magnetic needle or thehighest known degree of heat, but is thereand of fundamental importancefor scienceand life.” [Ges49, p. 96] Gesell’sfalsification of this doctrine strongly emphasizes that thereisnothing more real than economicvalidation of amonetarytheory by the individualand by the nation state. Everything thatdefersfromthis truevalidation can onlybeaproduct of imagination and sincethe inner value can’tbedefinedexactly through economicvalidation, thestatedvalue doctrineshould be seenasimagination.[Ges49,p.96] The current fiat money seems to prove him right, although thediscussion about the “value doctrine of money”isstill notcompletelysettled. According to Gesell,the described situation shifted the research about currencies away froma profound academicdiscussion(referring especially to the “innervalue doctrine”) to the imprecise handling of the currency questionand monetaryaffairs.Healso argues that delicatetopicsare seldomtouchedatuniversitylevel,whichaggravated the situation even more. At last,heoutlined the fact that, to his belief,acertain part of knowledge has to be acquired by practice in commerce, to get acomprehensiveunderstanding of themonetary system. He concludes that for this reason,

6 Background over 4000 years and 100 human generations, no substantivedefinition of money has ever been providedand therefore it is still handled by old habitsand without scientific reasoning.[Ges49, p. 88] Gesell endedhis introductionwith ashort comment on theworksofKarl Marx and Proudhon. He criticizes Marxfor writing “threethickbooks about interest(capital) and dedicatingintotal only5minutes of it on the fundamental theory of money”. He also states that most nameable economists, especially those with socialist origins, have dedicated there work against interest, failingtosee thatinterest is amere symptom whilethe origin of injustice lies in the design of the currency itself. He positivelyacknowledgesthe work of Proudhonfor not ignoringthis factand providinganidea after whichGesell wasable develophis Freigeld theory.[Ges49,p.89]

2.1.3The originofmoney

Geselloutlinedthatwithout the divisionoflabor, the predominantwealthwouldn’t be possible and money wasdeveloped to overcome thebottleneckofthe ineffectivebarter trade. He argues thatexcept for the products of farmers,almost all manufactured items are basic commodities, onlyusablefor themeans of trade and exchange. [Ges49,p.92]

Definition 2.1.1 (commodity). An itemcan be seen as commodity if it featuresthe following two attributes: 1. Ademand exists forthe item. 2. The offered item has to be useful forthe buyer (notthe seller). [Ges49,p.107]

Definition 2.1.2 (money). Moneyisamedium of exchange for basic commoditiesand it “is aprecondition of thedivisionoflabor, as soon itsextentdisqualifiesthe barter trade.” [Ges49, p. 91]

Gesell emphasized that there is aclear demarcation between basiccommodities (objects of utility) and themonetarycommodity called money.Heexplains this statementbysaying thatmoney (e.g. coins)can’tbeconsumed and doesn’t disappearinterms of adecay-it can only be transferred fromone individual to another. [Ges49,p.103] In this regardregard, theauthor of this study acknowledges the similarities to the lawofconservation of energy,whichstates: Energy can neither be created nor destroyed; energy can onlybetransferredorchanged from one formto another. Gesellstates: Thedivision of labor(and the therebyinduced growing specialization of society) creates acompulsive demand formoney. [Ges49,p.92]

2.1.4 The trade with money (and other commodities)

Hence, according to Gesell, the only dutyofmoney is to be amediumofexchange. The resulting trade of money or trade market is described as follows: “Everyone who brings something to the trade market bears thesame spirittodemandthe highest possible price, which is permitted by the market conditions”. [Ges49,p.102] “He has the right to make demands with his slipofpaper as far as the trade allows it.” [Ges49,p.100] “The ambition to get for aminimum of serviceamaximum of serviceinreturnisthe force which guidesand restrains thetrade of goods.” [Ges49,p.102]

7 Background

Nonetheless, Gesell argues that thereisnomarketwithout the peoplewho trade. He criticizes trading people who use the marketconditions and respectively escalated and depressed prices as an excuse. He states “What arethe market conditions,including the economic cycles, without the demand andsupply of thetrading people?” [Ges49,p.102] By this statement,heemphasises that the issuesofthe marketare influenced by the participatingindividuals.The individuals should therefore be included in the assessment of theissue and if an actual issue is present, it hastobe solved.

2.1.5 The three premises of money

The first premiseofmoney: Money has to be indifferentfor the user, with aminimum amountofphysical at- tributes. “Money doesn’t rust,doesn’t rot, doesn’t grow, doesn’t dissolve, doesn’t scrape, burn, or cut. (...). We search fornoactive, but only to alldirections inactive attributes. The society demands the minimum of allembodiedcharacteristics for the substantial part of money.” [Ges49,p.111] “Money has of allthe useful objects the fewestattributes and the lowestapplicability in industry and agriculture. We show no higher indifferencetoany other object. Thereforeitwas so easy to declare gold as money.” [Ges49,p.112] “For money it is sufficient that itiscountable, the rest is dead weight.” [Ges49,p.110]Nonetheless, “the mediation of trade between basic commodities is asufficient power” [Ges49,p.110] Giventhe firstpremise, Gesell acknowledgesthatthe visible propertiesofmoney are only tangible due to the necessary material of transportation.Heargues thatautilityindependentofthe material mightbethe reasonwhy it is so difficult to grasp itsconcept. “If someone would remove thephysical attributes,noone would miss them”. [Ges49,p.110] The secondpremise of money: Theeconomicfreedomtocreatemoney has to be prohibitedand the quantityof availablemoney has to be regulated. “It shouldnever be free of charge. Money fulfills its purpose duetoits nature, whichisaconstant search for it and the law that it canonly be obtainedbytrade” [Ges49,p.103] Geselloutlinesbyvarious examples that the freecreation of money will lead to acollapseofthe currencyand thereforeconcludes, “the economic freedom to create money has to be prohibited”. [Ges49,p.103] He acknowledges thatmoneyispart of societyand societyisasetofrules on whichpeople have agreed to live.Given the powerofthe nation states at thebeginning of the20th century,he proposes that onlynation states are able to prohibit the freecreationofmoneyand alsoregulate itsemission. [Ges49,p.104] The thirdpremise of money: Money depicts acircle whichittransverses forever; it returns to its origin. [Ges49, p. 104] Referring to the definition of money (definition 2.1.2), Gesellargues thatdue to the division of labor, theproduced basiccommodities are bound to be traded. The same goes for money, both “are only usefulasbartering object.” [Ges49,p.103] The difference is that money can’tbe consumedand will therefore circulate withinsociety.

8 Background

To explainthis claim,Gesell providedthe example of teaand salt during colonialtime.Atthat time, both where usedasmeansofexchange, but duetotheir attributes, both were consumed over time.According to Gesell’s definition of money,both examplescan’t be seenassuch. [Ges49, p. 104]

2.1.6 The backing of money

“A commodity is backedaslong as thereissomeone whoiswillingtotrade for it, with another basic commodity or money. In other words, acommodity is backedaslong as the demand doesn’t fade. Acommoditycan’t back itself.” [Ges49,p.117] “Besides thedivision of labor thereisno other backing for money.The divisionoflabor creates an uninterruptedflow of basic commodities, which generate an uninterrupteddemand for means of exchange, formoney”. [Ges49,p.118] “The money is notbackedbyits material,itisbackedbythe need forit(...) andits function as means of exchange.” [Ges49,p.118] “The mandatoryneed of money,byasimultaneousmandatory need of itscontrol by the nation state, gives the state unrestrictedpower over money and compared to this absolute power the metallicsecurity of the coin is achaff in the wind. Hence, the money can’tbesavedagainst the misuseofpower by the fabric, just as littleasthe constitution of anation state written on parchmentcan safe it from an arbitrary rule (Note: the powerlies in the people). Only thewillof those in power (absoluterulerorthe parliament as representation of the people) canprotect the money from carelessness, imposters and thieves -preconditionedthat thoseinpower purposeful know how to useit, which has unfortunatelynever and nowherebeenthe case.” [Ges49,p.116] Despite hiscriticismofthe nationstate,Gesell sees no other possibilityfor money than the administration by law, whichiscontrolled by the state. He declares, “the money collapses with the collapse of the state and the state collapses with the collapseofthe society”.[Ges49,p.112] The provided insight of Gesell’s thoughts on the backing of money,can be summarized by the following three conditions:

- “Thematerial of the money can’tinfluencethe demand for it; it canthereforenever serve as abacking for money.” [Ges49,p.118]

- “The backingofmoney(...) is exclusively guaranteed by the divisionoflabor.” [Ges49, p. 118]

- “The protection of money canonlybeaccomplished, if asoundopinion about monetary policiesbecomescommon property of the peopleand of the thoseinpower.” [Ges49,p.118]

2.1.7The valueofmoney

Thevalue of money is determined by the exchange relationshipbetween commodities.Ashift of therelation results in ahigher(or lower) quantity of basic commodities obtained by the same quantityofmoney or ahigher (or lower) quantityofmoneyobtainedbythe same quantityofbasic commodities. “Fromthis point of view this would be an indifferent matter,ifthe obtainedmoney would againbespent instantly,but this is notthe case.” [Ges49,p.119] Gesellelaborates this issuebyanexampleofcreditorsand debtors. He outlines that by withholding money over time

9 Background and presence of an altering exchange rate, eitherthe creditor or thedebtorwill have anegative return. [Ges49,p.119] Hence, in asocietywith an overall altering exchangerelationship, thequestion who to support by thedetermined money regulation needstobeanswered. [Ges49,p.119] Gesellstates: “We don’twant to betray anyone and the single case whichisonly benefiting forone individual must not be incorporated in the administration of money.Money mustbeadministered by national, not private accounts. Money should bare the samevalueover timeand space. The price which was paid today(...)should alsobeclaimablethe next day, or in one or in ten years. Thereby the debtor pays back what he received and thecreditor gets back what he borrowed: No penny moreorless.” [Ges49,p.120] In this regard, Gesellargues thatsuchasystemcan only work if aproof for aconstantvalue of moneyisprovided, andstates: “Thepriceofmoney canonly be expressedbythe commodities.” He explains thattocalculatethe currentvalueofmoney (the price of goods), “all bills and price lists have to be read backwards.” [Ges49,p.120] Nonetheless, the aspired stateisaconstantprice over time and therefore, thecurrent priceisnot crucial.[Ges49,p.121] Thus, Gesellproposes a methodwhichisknown todayasretailprices index (RPI), consumer priceindex(CPI)or“basket of commodities”.[Atk98,p.368] With this method, aset of basic commoditiesismonitored over time. To be precise, thepricefor which the basic commodities are soldand quantityisrecorded. Thisisdone forspecific periods, with the aimtocomparethe total average pricefor all goodssold. An example of the described methodispresented in table 2.1.Tofacilitate acomparison of these recordings,astandardization of the quantitiessoldisnecessary.Therefore, the totalquantityofgoodssold in eachperiodis adjusted to be equalfor allperiods.Inthis regard, areferenceperiodhas to be determined beforehand and commonly,the first (earliest)period is chosen.After this standardization, it is possible to compare theresulting average prices and to evaluate if the value of money has increased (deflation)ordecreased(inflation)overtime.Ithas to be acknowledged thatdue to the lawof supplyand demand, the prices and quantities are likely to change for individual commodities, but the average price of allgoodssold may still be constant. [Ges49,p.121] Definition 2.1.3 (inflation). “(...)ageneral and continuing increase in prices.”[Atk98,p.368]

“Thereare anumberofpoints to note about thisdefinition. First, inflation is about ageneral increase in prices. An increase in thepriceofone or two goodsisnot considered inflation, because most goodsmake up onlyatiny part of total spending. An increase in thepriceofafew goods does affect relative pricesand have implications for the allocation of resources, butitdoesnot affect the pricelevel as awhole, unlessthe commoditiesare extremelyimportant.Inthe more distant past, an increase in thepriceofafew basic foodscould have affectedprices generally; more recently,the huge increases in the priceofoil in 1973 had knock-on effects that put up prices as awhole. In general, however, price risesfor afew goodsdonot constitute inflation. Thesecond point to note is that inflation is aprocess that continues over aperiodoftime; aone-off risein pricesisnot usually called inflation.” [Atk98,p.368] Definition 2.1.4 (deflation). Ageneral and continuing decrease in prices. Thisequals anegative inflation.

Regarding the definitions of inflation and deflation,the following equivalences have to be acknowl- edged:

10 Background

-(Inflation) Apriceincrease of basic commoditiesequals adecrease of thevalue of money.

-(Deflation) Apricedecreaseofthe basic commodities equals an increase of thevalueof money.

Table to determinethe average price of indexed goods 1860 1880 1900 a. b. c. a. b. c. a. b. c. Price Quantity Total Price Quantity Total Price Quantity Total Wool 1.00 100 100.00 0.80 90 72.00 0.70 40 28.00 Sugar 1.00 20 20.00 0.90 90 81.00 0.80 110 88.00 Linen 1.00 70 70.00 1.10 40 44.00 1.20 10 12.00 Cotton 1.00 20 20.00 0.90 40 36.00 0.80 60 48.00 Wood 1.00 150 150.00 1.20 100 120.00 1.30 80 104.00 Iron 1.00 50 50.00 0.80 100 80.00 0.70 130 91.00 Grain 1.00 400 400.00 0.80 300 240.00 0.75 260 195.00 Meant 1.00 150 150.00 1.20 200 240.00 1.40 260 364.00 Indigo 1.00 30 30.00 0.80 5 4.00 0.75 1 0.75 Oil 1.00 10 10.00 1.10 35 38.50 1.20 49 58.80 Total 1000 1000.00 1000 955.50 1000 989.55

Table 2.1: This tableshows an example of thecalculation of theaverageprice of goods. Acomparison between differentyearsisenabled by standardizing the quantity. This allows to assessifthe value of money (theprice of goods) has altered. Themethodisnowadays called RPI, CPI or basket of commodities andisused to assess the degree of inflation or deflation.The table was derived from [Ges49,p.122]

Getting backtothe example of table 2.1,the totalpricehas to be raised (deflation)for the year 1880 and 1900 to ensureaconstantmoneyvalue. To be precise, araise by 4.66 %for 1880 and of 1.06 %for 1900. “This consistent price raisefor allgoodscan only be done by acommon origin (...)and an equal affectonall pricescan solely have money. We only need to put as much money into circulation untilthe pricesrise (...)”[Ges49,p.121] Regardingthe collection of data,Gesellacknowledges the considerable effort,especiallyinrespect to theformertime.Nonetheless, he emphasizes on theimportance of this measure and proposes asystemofdata collection withaveryhighindependence of thepredominant political system, combinedwith thedevelopmentofindexprices. [Ges49,p.123-126] He concludesthatthe integrityofthe assembly,the correctnessofthe price evaluation and the proportionally correct weighting of theindividualcommodities will ultimately provide an accurate result.1 [Ges49,p.123]

2.1.8Supply,demand andneeds

To ensure aclearnotation,itisnecessarytodefine theterms “supply” and “demand” in more detail.The reason is the ambiguityofthe term “supply”, which can either be seen as “stock” or “offer” and theambiguity of theterm“demand”, which can either be seen as “need” or “request”.

1As outlined in thefollowingsections, thisstatement is only valid if the concept of effective demand is present.

11 Background

Forthe translationofthe subsequent quotes,the implied meaningofthe terms supply and demand had to be interpretedand translatedcorrectly.Tooutline this personal impact theterms were marked by an ‘*’ (asterisk).

As shownbythe example in table 2.1 and elaborated in more detailinsection 2.1.9 (The quantity equation), the pricedepends on the offer and requestofboth -basiccommodities and money. In this regard,Gesellemphasizes that both shouldnot be equated with theneeds of people. To clarifythis circumstance,hepresents the following classification: request formoney,needof money, requestfor goods, need of goods.

“The request*for money has itsorigin withinthe basic commodities andisinduced by the division of labor. Thedivisionoflabor is the motherofcommodities,the inexhaustible source of the request* for money, (...)” [Ges49,p.108]

He arguesthat natural decayprevents long-term stocking of basic commodities. Therefore,stock and offer of basic commoditiescan be seenasequal. This equalization allowsapreciseand unambiguous use of theterminology “supply of basic commodities”. Giventhat the production of basiccommoditiesrequires thetrade for money,itcan be saidthat thesupplyofbasic commodities naturally embodiesthe request formoney.[Ges49,p.108] (seefig. 2.1)

Gesellargues, that “the entrepreneur, who needs* money from thebank doesn’t exchange anything, he gives nothing butthe promise to returnthe money.Heborrows, buthedoesn’t exchange. He gives money for money.”“This is no request*, which agreeswith thedetermination of amedium of exchangeand the money should be particularly amedium of exchange.” [Ges49,p.126] “No trade, no exchange; thereisnomention of prices. It is called interest. Also thenation state doesn’t request* amedium of exchange with its government loans.Itexchanges currentfor future money. Thereisnorequest*for amedium of exchange.” [Ges49,p.126] “The need*ofmoney originates from aperson, the request* for money from athing, from abasiccommodity”.[Ges49, p. 127]

As outlined,Gesell therefore strictly distinguishes between therequestfor credit(baring interest), whichwas named “need formoney”, and thetradeofcommodities,whichwas named “request for money”.

Thisargumentation facilitates the separation of the request for money from theneed of money and ultimately from theneeds of people.

Asimilarargumentation is providedfor the separation between therequestfor basic commodities and the need of basic commodities.Gesellgives the followingexample: Anewlydiscovered gold mine would generate ahigher availability of money,which wouldresultsinstantly in ahigher requestfor basic commodities.Withoutthe discovery of the goldmine, the availabilityofmoney would be lower, whichwouldresult in alower requestfor basic commodities.2 [Ges49,p.128]

“Request*for basic commodities existsdue to thesupply of money; those who have no money, offer no request* for commoditiesand those who have money, request* the basic commodities. Need forbasiccommodities is expressedofthose in need (beggars,poorpeople or people in trouble without money)”. “The request* for basic commodities, commonlyreferred to as ‘demand’,is representedexclusively by the offer for money”. [Ges49,p.128]

The demarcation stated is illustrated in fig. 2.1 and can be summarized as follows:

2The limitations to this statementare discussed in section 2.1.9 (Thequantity equation).

12 Background

1. Requestfor money =Supply of basiccommodities

2. Requestfor commodities = Offer of money

Thisoutlineddemarcation emphasizes on thecircumstance thatthe supply of basic commodities equals therequestfor money,but the supply of money doesn’t equal therequestfor commodities. Only the offer of money represents therequestfor commodities and the originofthis difference lies within theneeds of people. Therefore,Gesellargues: “Ifthe offer* of money would be uninterruptedand equaltothe fixed quantity* of money,the price (the exchangerelation between commoditiesand money) wouldbe independent of thehuman factor” (Note: Independent of peoples needs). “Money wouldthen be the embodied, precise shapeofthe demand, similartothe embodied, weighableand calculable supply of the basic commodities.” [Ges49,p.129] Thisled Geselltothe keyquestion:Isitpossibleand correct to declare theoffer of money equal to thequantityofmoney?Itwas decided to answer this questionbyThe quantityequation, whichisdiscussed in the followingsection and it mayalready be noted thatthe solution builds the basis of the Freigeld currency.

Figure 2.1: Thisfigureillustrates the process of supply anddemand(offer andrequest). As briefly intro- duced in section 2.1.8 andoutlined in moredetailinsection 2.1.9,the supply of commodities equals the demand formoney,the offer of money equals the demand forcommodities,but the totalquantityofthe money available doesn’t ultimately equal themoney offered.Hence, to correctly depict the marketconditions the term‘supply of money’was changed to ‘offer of money’.This circumstance is indicated by the dashed arrow.

2.1.9 The quantityequation

Definition 2.1.5 (quantity equation). P · T = M · V + M · V [Fis20]

The quantity equation, defined in definition 2.1.5 and illustrated in fig. 2.2,was publishedin1911 by Irvin Fisher in his book“The Purchasing PowerofMoney”. [Fis20]Itdetermines that the average price level (P)dependsonfive other variables [P.11,p.1]:

-M=moneyincirculation. [M] =1e;

trades 1 -V=velocity of circulation of money. [V] =1 s = s ; -M=bank deposits subjecttocheck(deposit currency). [M]=1e;

13 Background

trades 1 -V=velocityofcirculationofdepositcurrency.[V]=1 s = s ;

piece 1 -T=volume of trade.[T] =1 s = s ;

e -(P=general price level. [P]= 1 piece =1e;)

In reference to the “basketofcommodities” (section 2.1.7), every variableexcept the velocity of circulation of money has already been introduced.Itdescribes the “number of timesthat a money in circulation is ‘turned over’ in ayear”.[P.11,p.1]Inotherwords, this meanshow often the moneyisexchanged forbasic commodities in agiven periodoftime and thus changing its owner.

In definition 2.1.5,the active quantity of money (M · V)isdescribed as themultiplication of the quantityofmoney (M)timesits circulation velocity(V) plus thequantityofdepositcurrency (M)timesits velocityofcirculation (V). Hence,the active quantity of money equalsthe average level of price (P)times the volumeoftrade (T). To specify,Trefers to the number of products (pieces) sold during the chosentime period.This is illustrated in fig. 2.2.

Figure 2.2: This figure illustrates the dependency of thegeneral price level(P) on five(to some extent) independent variables as definedinthe quantityequationbyIrvin Fisher.[P.11]. Theidea for this illustrationwas taken from [Ben96].

As already indicated, Geselldescribes thebasiccommodities as acontinuous stream, originating fromthe divisionoflabor and running dry in thehomes of theconsumers. Themoney however transverses themarket withoutwearingoff. He concludesthat “money is no stream, it is a circulating object”[Ges49,p.135], and by this comparison, it can be saidthatthe “offerof money obeys to other laws then theoffer of commodities”. [Ges49,p.135]

“Duetothe constitution of the commonmoney it is possible that the demand (the offer of money) is delayedfor one day, one week or even ayear, without anyimmediatelosses–whereatthe supply (the supply of basic commodities) can’tbewithholdwithoutconsiderable expanses of allsorts.” [Ges49,p.138] “The basic commoditiescommand without the allowanceofany contradiction. The willofthe owner is so powerless that we canleave it unconsidered.Incontrast, the offer of money follows the willofthe owner,itisacompliant servant of his master.” [Ges49,p.141]

14 Background

Geselloutlines thatthe commonrequirement fortrade is aprofit, atribute or an interest.3 Without those, thereisnotrade. Money howeverallows theownertopostpone the trade as long as he wantswithoutconsiderable losses.[Ges49,p.142] Hence, it is evident thatthe velocityof circulation depends on the freewill of all individualswho participate on the trading market. This circumstance wascomprehensively discussed by Gesell and his argumentation canbesummarized as follows [Ges49,p.129-160]: It can be saidthatmoney will onlybeoffered if themarketprovidesenoughsecurity against losses and offers acertain amount of profit for themoney offered.[Ges49,p.142] If the market featuresthese good conditions,more money is spentand more productsare bought.This in turn stimulatesproductionand due to the increased spending, thevelocityofcirculationofmoney increases, too. Thisagain stimulates the productionofgoods, but due to the faster response of thevelocityofcirculation, the active quantityofmoney(M · V )exceeds the increase of production.This shifts the weighttothe activequantityofmoney(seefig. 2.3 -bottom). The delayed production and thereactiveactivequantityofmoney leads to aprice raise forthe basic commodities (moremoney then products). In this regard,Gesell argues thatthe increasedvelocity of circulation of moneyinduces ahigher availability of money,whichtheoretically leads to lower interest rates forcredits (lowerrequest for money).However, thedemand for money,pushed by the production,actually increases theinterestrates, generatingmore revenue forcreditors. The debtors approve the higher interestrates on thebasisofthe prospering outlook of theeconomy. Thisincreases deposit currency and itsvelocityofcirculation, which additionally amplifiesthe situation. Productionstartstoincrease, but duetothe higher active quantity of money,prices increase or at best do not fall. If the described situationofincreasing priceslastsfor alonger periodoftime,itiscalled an inflation.This is illustrated by the quantityequationinfig. 2.3 (bottom). Nevertheless, therewill be apointintime when the production catches up, or for some reason (e.g. controversial politicaldecisions),the velocityofcirculation remains constant or decreases. Thus, therevenues (ofthe products) shrink, whichcauses themarket participants to lessentheir investments and therefore, money will not be spenttothe same extent as before. Theowners of money will use theirrighttoobserveand wait and therebypostpone some of their investments. Thisprocesswillmost probably lead to adecrease of thevelocityofcirculationof money.Ifthis circumstance extends acertain point, the weight will be shifted to thesupply of commodities, whichleads to apricedecrease. (see fig. 2.3 -top) Thisshiftcauses thevalue of money to increase by itself,challenging thenecessityofinvestments and ultimatelyleading to an additional decrease of theactivequantityofmoney. Thislowers interest rates forcredits. It hasto be acknowledged that there isatheoreticalchancethat the lowerinterests, will lead to an increase of depositcurrency (andits velocityofcirculation)and thereby counterthe possibility of astarting recession. Nonetheless, Gesellargues thatwith the economic outlook presentinthis situation,the majorityofinvestorswill nottakethis chance and therebythe totalquantityofdepositcurrency and itsvelocityofcirculation will also decrease.This continuous cycle aggravates over time. Additionally,the fall of theprices putspressure on thejob market.The induced stagnation of production restrains the creation of newjobs and thedecrease of priceschallengesthe payment of employees and the sustainability of companies. Over time, these circumstances increases the unemploymentand leadtobankrupted companies,whichdecreasesthe velocityofcirculation of money even more. As outlined in definition 2.1.4,the situationofdecreasingpricesoveralonger periodoftime is called deflation.

3In this regard,Gesell specifies that thisdoesn’t include the general tradingprofit and describesitasadditional revenue. It shouldbeacknowledgedthat thisalso the starting pointofhis second monetarytheory, the capital theory.Inreferencetothe researchobjectivestated, this theory,whichincludes the topic of credit and interest, will notbecovered in this elaboration.

15 Background

In referencetothe evaluationoutlineditisreasonable that Gesell came to thefollowingconclusion: “Economiccrises, thus stagnationoftrade and unemployment, with alltheir accessory phe- nomenons, is only thinkable withretreating prices”. [Ges49,p.149] “How do we prevent these stagnations of trade? (...)The explanation of its origin already gives the condition to prevent the stagnations of trade: The pricesmust notfallunder any conditions”. [Ges49,p.150] By examiningdefinition 2.1.5,itcan be saidthat if the price(P) (the value of money) is aspired to be kept constant,the other variablesmustalsobekeptinaconstantrelationship. To ensure this,the exactvalue has to be known for thequantityofmoney (M)and deposit currency (M’) in circulation, the volume of trade (T)(the supplyofbasic commodities), the velocityofcirculation of money (V) and thevelocity of circulation of depositcurrency (V’). As outlined in section 2.1.7 (the basket of commodities), compilingthe totalvolume of trade (T) is extensive, but feasible. Sinceone of the main attributes of money is to be countable, Mand M’ arecomparativelyeasy to assess. Nonetheless, andasoutlined by Gesell, the velocityofmoneyisnot determinableto adegree which allowsamonetary policy of stable long-term prices.Hence, coming backtothe exampleofthe basket of commodities(table 2.1), themerechange of the quantityofmoney to ensureaconstantmoneyvalue has no mathematical basis to be successful. Therefore, to ensurepricestabilityand aconstantvalue of money,Gesell provided asolution by proposingademurrage currencynamed Freigeld.

Figure 2.3: This figure interprets inflation and deflation on thebasisofthe quantityequation.[P.11]A descriptionofthe variables indicatedisgiven in fig. 2.2.The idea for this illustration was taken from [Ben96].

2.1.10The idea of Freigeld

Geselldefined theunbound circulation velocityasthe origin forvarying average prices. Hence,to ensureaconstantaverage pricefor commodities(and aconstantmoney value), the incapability of

16 Background measuring or controlling the circulation velocityofmoney(and depositcurrency) hadtobesolved. He traced this issue backtothe difference between basic commodities and moneybyarguing that basic commoditiesdecay and therebyshowanabsolute demandtobetraded. Money however doesn’t decay and thus featuresonly an unbound demand for trade. This leads to an unbound velocityofcirculation,whichcreates the stated instabilityonthe trade market. “Thus, get ridofthe privileges of money!The money as commodity shouldn’tbebetter for anyone as the stock of the markets,shops and railroadwarehouses,also forsavers, speculators andcapitalists4.Hence, if it isn’t allowedtopossess anyprivilegescompared to its counterpart thebasic commodities, themoney should alsorust,get moldy and decay, it shoulderode, sicken, run away and if it dies theowner should paythe wage of the knacker. Onlythen we willbeable to say,money and basic commoditieshavethe same rank and areequivalent things -asProudhon would have liked it to be.” [Ges49,p.13] Thus,Freigeld wasdeveloped as acurrency whichdemands feesfor inactivity. The economic term for this tributefor extralay days is demurrage and the correspondingcurrencyisnamed demurrage currency (seedefinition 2.1.6).

Definition2.1.6 (demurrage currency). “A currency in whichthe value of unitsofcurrency is designedtofallovertime at afixed rate.” [Lie19,p.9]

To be more precise, Gesellproposed aweekly demurrage of 0.1%at theexpenseofthe owner. [Ges49,p.184] According to the records of one of themost famous Freigeldexperiments known as “the wonder of W¨orgl”, aslightlydifferentdemurrage of 1%(due forpaymentonthe first dayofeachmonth)was established. [Sch08,p.45] Giventhe elaborations of Gesell, the regulator of themonetary system (presumably the state) will receivethe demurrage paid by the owners. To prevent misinterpretations,ithas to be acknowledged that this is no inflation.Asstated in definition 2.1.3,inflation is“ageneral and continuing increase in prices”.[Atk98,p.368] The money paid as demurrage is just somewhereelse,but its value stays the same.Unfortunately,this fact is often overlooked.Italsohas to be acknowledged that thedemurrage inducesthe balance between basic commoditiesand money.Hence, an exaggeration of the demurrage willshift the balance towards the basic commodities, whichmight lead to new issues and therefore,Freigeld should not be used as means of taxation(or similar).[Sch08,p.45] Gesellacknowledges thatthe demurrage will probably be defined more accuratelyovertime,similarasfor thecompilation of datafor thebasket of commodities. Nonetheless, he emphasizes that this shouldnot hinder the general start. [Ges49,p.188] Fordetails about thedistributionorimplementation of Freigeld, it is referred to thewritings of Gesell[Ges49]orFritz Schwarz, who documented the Freigeld experimentof“W¨orgl”inhis book“Das Experimentvon W¨orgl”[Sch08].

2.1.11 Theeffectivedemand

The effective demand describes theexact adjustment of thesupply of money to the demand for money (orrespectively, to the supply of commodities)(see fig. 2.4). Hence, Gesell claimsthatademurrage currency is apreconditionfor the effective demand, since otherwise, theunboundcirculation velocityofmoney distortsthe adjustmentstated and con- tradictsalong-term constantvalueofmoney.Asemphasized by Gesell, the demurrage allows

4 Gesell defined capitalism as follows: “Capitalism =aneconomic state, wherethe demand forcredit andreal capital exceedsthe supply, andtherefore causes interest.” [Ges49,p.150]

17 Background to separatethe needs of the peoplefromthe request for money.Money will then be tradedon the market equally to its counterpart -the basic commodities.Hence,the supply of money will generateademand for basic commodities in the same way as thesupplyofbasiccommodities generates ademand for money.This allows to effectivelyregulatethe supply of money according to its demand. Thisattribute of Freigeld is presumably thereason whyIlgmann(2011) referred to this part of Gesell’stheory as thetheory of effectivedemand.[Ilg15]Inreference to sec- tion 2.1.8,itcan be saidthatthe demand formoney willthen be equalthe supply of money. In circumstances where the economyresemblesthe outlinedsituation, the continuous production (divisionoflabor) and thecontinuous offer of money induces acontinuous economiccycle.This is illustrated in fig. 2.4.

Figure 2.4: This figuredepictsthe continuous circulation of money and uninterruptedtrade,whichis the desired goalofFreigeld. Compared to fig. 2.1,Freigeldallows to reject the stated double meaning of supply and demand(see section 2.1.8), yieldinganunambiguous useofthe common terminology.

To summarize, theeffective demand features two preconditions or results:

-Money is excluded as amean for saving and itssupply is adaptedtothe requirements of the market. [Ges49,p.150]

-The design of money ensures that it will be offered for commoditiesunderevery circum- stance, even if “theinterest on capital (...) and the interestonrealcapital decreases and disappears”.[Ges49,p.150]

2.1.12 PostulatedimpactsofFreigeld

In hisbook, Gesell discussed abroad variety of theeffects of Freigeld. In reference to the discussed topic of effectivedemand,itwas decided to provide alisting of all impacts, while onlyimpacts closely related to the effective demand are described by aconcise statement. Foracomprehensive descriptionofall factors, it is referredtothe original bookofGesell.

1. Money circulation Impact:

-Acontinuous money circulationand trade of commodities. -Aconstantmoney value. -The economic crises which arecaused by theunbalance between money and commodi- ties are eliminated.

18 Background

Statement: The aim of thedemurrage is to induceaconstanthighcirculationvelocityofmoney, ideally the maximum. Thetherebypostulated nearly unstoppable circulation of money facilitates the regulation of thevalueofmoney.The value is then kept constant by changing the quantity of moneyascontemplatedbythe exampleinsection 2.1.7 (basket of commodities). Thereby, the weightingbetween average price (P) times the volume of trade (T) and the activelycirculating money showsaclear balance as illustrated in fig. 2.2. 2. Interestand speculations (Capital theory) [Ges49,p.179 ff.] Impact: -The eliminationofinterest. -The eliminationofstock exchange speculations. Statement: As outlinedinsection 2.1.1,the theorynamed Freigeld-interest-theory or capital theory is not elaborated in this report.Nonetheless, due to the connection of both theories and the contrast to our presentmonetary system, it wasdecided to giveashort statement about the claimed elimination of interest. Gesellpostulates that duetothe effectiveregulation of themoney supplyand the thereby resulting fulfillmentofthe demand,the interestrates will diminish untilthey are completely eliminated. In this regard, thechange for thebankingsector can be considered tremendous. Geselloutlines thatthe availabilityofmoneywill eliminate the necessityofdebit money and therebyalso diminishand eliminate speculationsatthe stock exchange. “He who has money,has no morerightastobuy commoditiesinstantaneously. Aright for interest contradicts theideaofmoney,sincethis right would be similar to aprivate taxation of the exchange of goodswiththe support of government facilities.” [Ges49,p.153] 3. Trade facilities Impact: -Areductionoftrading facilities in terms of quantity and their shareatthe gross domestic product (GDP). Statement: Although Gesellextensively outlines thenecessity of trade, he argues that thequantityof trading facilities and theirshare in the grossdomestic product canbeused as an indicator forthe qualityofthe trade market. He acknowledgesthatthe progress of specializations also increasesthe need fortrading facilities.Nonetheless, he arguesthatthe trading efficiencyoffacilities also increases with the further developmentofsociety. Hence, theshareoftrade market facilities,can be used as ameasurementfor the qualityofmoney.Thereby,Gesellargues thatahighshare of trading facilities in the GDP is an indicator for alow qualityofthe money.Inreference to theissuesofconventionalmoneyoutlinedpreviously,Gesell states that theuse of Freigeld would reduce shareand quantityoftradingfacilities.Intotal,asimplification and reduction for thecost of trade is predicted. [Ges49,p.157 ff.] 4. Savings Impact:

19 Background

-The separation of money as ameans of trade and as ameans of saving. -The facilitationofsaving. Statement: Albeitthe statedseparation, Gesellargues thatFreigeld will facilitate saving by outlining thefollowingnew circumstances: -The capitalbears no burdens of interest. -Pricereduction of manufactured items due to reducedtradingcosts. -Noeconomic stagnations. To facilitateameansofsavingthe approachsuggested by Margrit Kennedy in her book “Occupymoney” (2012) is noteworthy.Long-termsaving can be achieved by lending money to someone with adetermining date of repayment. This allowsindividuals who ownmoney without an urgentneed to and enables long-term savingplans.Therebythe intermediate owner bears theliabilitytopay the demurrage (e.g. on the first dayofthe week or month). The process caneither be facilitated by banks or privately. Eitherway,the demurrage has to be paid by the current(intermediate) ownerofthe money.[Ken11,p.47] Nonetheless, theelimination of money as ameans of saving will shift the focus to products and especiallytorealestates. AccordingtoGesell,this leads to various issues, which presumably ledhim to hissecondmaintheory named Free land. “Money shouldbeacommodityand amaterial for personalneeds. They want ahybrid, something impossible.” [Ges49,p.117] “Therefor Idemand (...) acompleteand factual separation of the means of exchange from the means of saving.” [Ges49,p.151] “According to the presentedevaluations, Freigeldroughly and recklesslycuts the commonunion of money as means of exchange and saving. The money becomes thepuremeans of exchange, freed from the will of its owner,independent of the material and chemicallypuredemand.” [Ges49, p. 157] 5. Impacts beyond the areaofthe effectivedemand [Ges49,p.184 f.]

- “The suspension of ‘useless’ protective tariffs and the transition to free trade” - “Thesuspension of economicoriginsfor war” - “The mediation of aglobalmonetaryagreement fortrade, given that this is beneficial forall nations” - “The gradual increase of salaries” - “The lasting elimination of unemployment”

2.2 Blockchain and crypto currencies

2.2.1 Structure

In general,crypto currencies have been developed as an incentivetosecure theunderlying blockchain. In this regard, it can be said that crypto currenciesbuild the top layer(1=top), followedbyaclientprotocol as amiddlelayer (2 =middle) and the blockchain as thebase layer (3 =bottom).Hence,the followingchapter will describethe given parts frombottom to top ([Swa15,p.6]):

20 Background

1. Cryptocurrency: e.g.Bitcoin (BTC)

2. Protocol and client:Software programs conducting transactions

3. Blockchain: Underlyingdecentralizedledger

2.2.2 Blockchain

Ablockchain stores databyconnectingblocksofinformation. Thereby, eachblock is linked to the previous block and new information is stored by connectinganew blocktothe lastblock of thechain.

The original idea of blockchain wasdeveloped in 1991tocreate digitaltimestampsfor documents. Nonetheless, theconcept wasnot particularlyusefuluntil the adaptation of an “unknown person or entity”named Satoshi Nakamoto. The published new framework allowedthe creationofa different blockchain, which includedacrypto currency named Bitcoin (released on January9, 2009). [Swa15,p.2]

Over the last years, new developmentsinthe field of crypto currenciesand blockchains ap- peared.However, due to the sustainedimportance of theoriginalblockchain also known as Bitcoin blockchain,this chapter willintroduce the topic by elaboratingthe attributes of the lat- ter.[Swa15,p.6]Manyofthese attributes are valid for other blockchains,yet inferences have to be done carefully.

TheBitcoin blockchain is known as open synchronizeddistributed ledger technology(DLT).Each of these adjectives describes acrucial attribute of the Bitcoin blockchain. Hence, thefollowing elaboration will usethisfact to describeits functionality.

2.2.3Ledger

As indicated above, theblockchain connects blocksofdatatostore information. Theoriginal ideaofthe Bitcoinblockchain wastostore transactions of thecrypto currency namedBitcoin. Hence, thedescription of theblockchain features theterm ledger,presumablyinreferenceto an accountorledgeratausual bank whichalso depicts transactions.The dataiscollected in blocks andeachblock features(among other attributes -see fig. 2.5)areference to the previous block and aset of transactions.Asuccessfullyadded block is immutably storedinthe Bitcoin blockchain andthe included data can’t be manipulated or erased.This outstanding featureis secured by amathematical function named Hash-function,followedbyavalidation of the newly added blockthrough aconsensusmechanism.

2.2.4Hash code

The hash function is partofthe scientific field namedcryptography. Acommonly used algo- rithmisthe SHA-256-Hash-Algorithm.Ingeneral, independently of thesize of thedata file, the function transfersthe dataintoa64-character-code.[Swa15,p.1]Under theterms of acorrect implementation, theHash-function has the following features[FM20,p.5f.]:

21 Background

-Anexplicit representation of theinputinformation. Thisalso includes ahigh collision resistance (a lowpossibility to create equalresultsfor differentinputs) and asufficient variance (a smallchangeofthe inputresultsinconsiderable variation of theoutput).

-The infeasibilitytoderive the input data set fromthe present hash code.

-Without additional constraints, the functionshould be quicktocompute.

Figure 2.5: This figure depicts the attributes of data blocksand the connectionbetween them,resulting in achain known as blockchain.The figure wasretrieved from [FM20,p.12].

Eachblock includes ahash code knownasroothash (see fig. 2.5). The generation of this root hash includesthe complete data set of thepresent block.Hence, amodification (e.g. anyaltering of transactions) wouldresultinadifferentroothash,whichallows an easy verification of the integrityofthe whole block.Additionally,the data set of the present block also includesthe hash code of theprevious block,whichintroducesaconnection (chain)between both blocks. Hence, the change of anydata in ablock willresultinadifferenthash code and thereference from the subsequentblock will show aconflict. Thisisillustrated in fig. 2.5.[FM20,p.5f.]

2.2.5 Distributed ledger

The Bitcoin blockchain itselfisstored in apeer-to-peer network. Thismeans that everypar- ticipantofthe networkstores acompletecopyofthe Bitcoin blockchain.Thereby,nocentral organization is needed to organizethe storage, which allowsacompletedecentralized implemen- tation. Thisapproachistherefore referredtoasdistributed ledger.[KD20,p.5f.] Combining the aspectsofablockchain stated before, asuccessful attackonablock (e.g.the change of one bitorevenatransaction) wouldleadtoanecessaryadaptationofall hash values in all subsequentblocksofthe Bitcoin blockchain. Additionally,thishas to be validated by the distributed network of participants. To be precise, each newly added block(or in regard of an attack: changedblock)has to be approved by adistributed consensus mechanism.[KD20,p.3f.]

2.2.6 Consensus mechanism

The origin of theconsensusmechanism lies in acryptographic puzzle basedonthe hash-function (a hashpuzzle). Forthe process of validation, every participant of thedistributednetwork tries to

22 Background solve the puzzle by an automated computation.The block is added to the blockchain if theproof is successful. Thismeansthatthe transactions are evaluated as correct,the puzzle is solved and the majorityofthe network agreesonthe solution. Thecompulsive validation by the network is known as consensus mechanism.

The design of thepuzzleonly allowstofind the solutionbytrialand error. Theunderlying algorithm ensures acontinuous level of computation effort by automatically adapting thecom- plexityofthe puzzle. Thereby, the creation timeofone block is set to “nearly”10minutes. The totalcomputation effortfor theparticipants who chose to work on the solutionofthe puzzle is considerable, which is whythe consensus mechanism forthe Bitcoin blockchain is named proof of work (POW). It can therefore be concluded that cryptography induces alevel of computation time and effort to secure that data stored in the blockchain. [KD20,p.4][FM20,p.6f.]

2.2.7Synchronized distributed ledger

To enable ajustand simultaneous competitionfor thesolution of apuzzle, the timings within the network have to be synchronized, whichiswhy the Bitcoin blockchainisalso called a syn- chronised ledger. As soon as aset of transactions is determined to be validated, thedataset is distributed to the whole network as apotential new block.The participants who choose to work on thevalidation of thepuzzle are called miners.Hence, theminers validatethe correctness of thetransactions and trytosolvethe puzzle.The first minerwho successfullysolves the puzzle broadcasts thesolution to the wholenetwork. As stated, thehash puzzlerequires considerable efforttobesolved, but due to itsdesign, thevalidationofacorrect solution is effortless. Thus, it is fast and not limited by aminimum timefor calculation. After asuccessful validation by the consensus mechanism, theblock is added to theblockchain and thefirst successful minerwho solved thepuzzlegetsrewarded with Cryptocurrencies (see section 2.2.9)[KD20,p.6]

2.2.8Opensynchronized distributed ledger

The Bitcoin blockchain “isopenand public to everyone present in the network”.[KD20,p.5]Its data is immutable,but notencrypted.This means that everyone in the network of theBitcoin blockchain can trackevery transactionand the amountofBitcoins stored in different wallets. The Bitcoin blockchain is thereforecalled an open ledger.

2.2.9Crypto currencies

As stated in section 2.2.7,crypto currencies have been developed as an incentivefor thevalidation of blocks fromthe Bitcoin blockchain. Forevery successfully solved puzzle and validation process, acertain amountofcrypto currencies is transferred to the account of thesuccessful miner. This caneither be an individual,acompanyoraconsortium. The crypto currency for theBitcoin blockchain is named Bitcoin,whichisdisplayedand stored in an electronic wallet.For asmall transaction fee,Bitcoinscan be used as ameansofexchangefor goodsand services.These transactions are conducted by aclientprotocol(the middle layer). [Swa15,p.2]

23 Background

2.2.10 A“trustless” system

Good cooperation of individuals or organizations has always been one of theprinciples of progress. But human interactionhas alwaysbeen accompaniedbythe issue of trust. Is the opponent trustworthy? Whathappens in case of adeception,anerrororsome other kind of problem? As asolution to this issueand driven by the growing interconnection of thehuman world, society developedframeworks suchaslaws, regulationsand governance processes. Alongwith these developments, institutions emerged whichnarrowed down frameworks, streamlined processes, solveddisputesand, above all, served as liableauthority. Especiallyinstitutions whoare used as liableauthoritiesare knowntoday as trustedthirdparties.Inthe case of currencies, this has longbeen to an extensive degree thefield of private and national banks. In thisregard,itcan be saidthat our societydemanded trustworthyinteractionsand thebanksutilized this demand by establishingatrustworthymiddleman. Nonetheless, in reference to the recenteconomic crises, theiracting as an intermediaryhas been considerably questioned. This particular circumstance is intensified by the developmentofcrypto currenciesand their incorporated mechanisms. [BAI+18, p. 69 f.]

Cryptocurrencies(Bitcoins) are regulated by open,distributedproof mechanismsand the data is stored completely transparent. Hence, theBitcoin Blockchain claims to be freeofhuman influence and equal for every participant. The decentralized approachdistributes decision making to the participantsofthe network,and therebyloweringthe threat of the influence of acentralized power. Thetrustworthy middleman (e.g. the bank) formerly deemed necessary is replaced by a transparentmechanism.Hence, trust is grantedbytransparentcodeand not by intermediaries. [BAI+18,p.69f.] Theblockchain and crypto currencytechnology is thereforeoften called a trustless system, whichmeans “at its most basic level, intermediary-free transactions.” [Swa15, p. 3]

2.3 Sustainability andcomplementary currencies

2.3.1 QuantitativeEconomic Development

Quantitative economicdevelopment(QED)isaconcept developedbyGoerner S., Lietaer B. and UlanowiczR.(2009) and allowsa“quantitativemeasure of sustainabilityfor anycomplex, matter/energy flowsystem”. [UGLG09,p.76]The conceptisillustrated infig. 2.6.

During theirresearchonecological systems Ulanowicz et al. foundarelationbetween what wasnamed resilience (R)and efficiency (E). [UGLG09,p.77]Systems whofeatured along-term vitality were placed in aregion, which showed acertain balancebetween Rand E. Thisregion wasnamed “The Window of Vitiality/Viability”. [UGLG09,p.77]

Definition 2.3.1 (efficiency). “The network’s capacitytoperform in asufficiently organizedand efficient manner as to maintainits integrityovertime (May, 1972).”[UGLG09,p.77]

Definition 2.3.2 (resilience). “Itsreserveofflexible fall-backpositions and diversityofactions that can be usedtomeetthe exigenciesofnovel disturbancesand the noveltyneeded for on-going developmentand evolution(Holling, 1973, 1986; Walker et al., 2006).” [UGLG09,p.77]

24 Background

Furthermore,Ulanowicz et al.(2009) defined sustainability as “an ecosystem’sability to maintain its own vitality over long periods”. Afactor named Development Capacity (C) wasdetermined to assess the level of sustainability (S).The relation of all mentioned factors is shownineq. (2.1). [UGLG09,p.78]

C = S = E + R (2.1)

Theresearch by Ulanowicz et al.(2009) canbeportrayedasfollows:Asystem is sustainable over along periodoftime if the relationofEand R is situated within the theWindow of Vitality/Viability (see fig. 2.6). Atilt towards toomuchefficiency willresult in heightened fragility, atilt towards toomuchresilience (or diversity) will move thesystem towardsstagnation. Onceasystemistilted towards toomuchefficiencyanincreased“ability to pullmoreand more energy into its sway,whilereducing extraneous diversity”can be observed. This introducesaself- amplifying loop, which willresult in acollapse of thesystem. In generalitshould be acknowledged thatthe function of the DevelopmentCapacity(C) is not symmetricand shows aslighttendency towards the factorofresilience.(see fig. 2.6)[UGLG09,p.77] Ulanowicz et al.(2009) transferredthe findings derivedfromecosystems to our presenteco- nomic/monetary system, which enabled anew interpretation of thelatterasflow system. The conclusion ultimately drawn states thatthe notable frequency of economiccrashes is causedby our presentmonetary systemwhichseems to exhibitaconsiderabledislocation towards efficiency.

Figure 2.6: Thisfigure depicts sustainability as afunction of efficiency and resilience andwas retrieved from [UGLG09,p.77].

2.3.2 Complementary currencies

As co-founderofthe theory named QED the economistBernard Lietaer strengthened hispoint of view about thenecessityofcomplementarycurrencies for our economicsystem. In his book “Towards asustainable world”,heemphasizes on thenecessity of acleardemarcation of com- plementary currencies and arguesthatcomplementarycurrencies should be designedtosupport “the conventional monetarysystem where it fails to assign money to meet the need”.Thereby,he especially referstothe social needs of thesociety. [Lie19,p.30] Additionally,inregardtoour efficiency driven monetary system, BernardLietaer outlines thata complementarycurrency can only be identified as such if it featuresatleast one of the following attributes [Lie19,p.30]:

25 Background

- “Dampen business cyclesofboomand bust, supporting sustainability of businesses,countries and even banks.”

- “Support longer termthinking rather than short-term thinking.”

- “Are interest free,thus do not drive exponential growth.”

- “Bear no interest, or even better, charge fees for storingmoney, thus discouraging accumu- lation of money for creating wealth.”

- “Discourageaccumulation of acurrency, for instancewith a‘demurrage fee’ (...)”

- “Avoid the ego-centriceffects of ‘money priming’.”

In reference to the discussed crypto currencies, Bernard Lietaer stresses thepoint thatalbeit their distributed decentralizednature, cryptocurrenciesare usually designed to increase efficiency. Their mere usage as asecondcurrencywould therefore notchange anyofthe issuesoutlined by the theoryofQED.

2.4 Accelerated cryptocurrencies

2.4.1 Structure

Thissection discusses possibleintersections of theeffectivedemand and crypto currencies. Based on this evaluation ahypothetical Freigeldcrypto currency scenario is developed. The developed currency is then used forthe subsequentsocialsurvey.

2.4.2 Evaluation of thethree premises of money

As outlined in section 2.1.5,Gesellstated three premisesofmoney:

1. Money has to be indifferent for theuser,with aminimum amountofphysicalattributes.

2. The economic freedom to create money has to be prohibited and the quantityofavailable money has to be regulated.

3. Money depicts acircle whichittransverses forever;itreturnstoits origin.

Comparingthe stated premises to the attributes of crypto currencies, it wasconcluded that the abstract nature of adigitalcurrency fulfillsthe first premise(minimal physical attributes)toan exceptional high degree. The underlyingmechanismsofcrypto currencies allow atransparent and clear regulation of the supplyofmoney, implemented by adecentralizedsystem ensuring little to no influenceofthe respectiveholders of power. Therefore, the second premise (quantity regulation)was evaluatedasfulfilled.The thirdpremise (circulation)isalready partly fulfilled by the abstract nature of crypto currencies.Nonetheless, in reference to theoutlinedhistory of gold as material for money,the lowapplicabilityisanecessary, but no sufficient condition for a continuous circulation of money.Hence,similarasoutlinedinthe elaboration of Freigeld,the underlying double natureofcrypto currencies as ameans of exchange and as ameans of saving

26 Background mightcause thesame issues as with the presentfiat currency.Itwas therefore concludedthat the compliancefor thethird premise depends on the underlyingdesign of thecrypto currency. Thethird premisewas therefore evaluatedaspartly fulfilled, while it can be saidthatthe origin of this partial fulfillment lies in the design and not in the nature of crypto currencies.

2.4.3Applicabilityofthe effective demand

Following up on thestated partialfulfillment of thethird premise, the application of the effective demand might solve this issue and is therefore an evidentquestion. Basedonthe previouselaborationsand similar to thepresentlydeployedfiat currency,crypto currenciesrequirethe followingtwo features to implementthe effective demand:

1. Ademurrage to ensure ahigh to maximum velocityofcirculation of money.

2. Acontrolledsupply of money,which is solelydefinedbythe demand (notthe need) for money.

The implementation of thefirstrequirement(maximumvelocity of circulation of money)issimply achievablebyademurrage fee.Thisresultsinvarious effects, whichhavetobehandled carefully to ensurethe fulfillment of thesecondrequirement(controlled money supply).Ingeneral, the digital format of thecrypto currencyallows arelativelyeasy assessmentofthe CPI (basket of commodities). Even if paymentsbycard and not onlybysmartphone app are madepossiblefor better accessibility, the Freigeld crypto currencystoresthe data of theexchanged goodsdirectly in one data base, the blockchain. The dataistherefore readily availabletoprove apossible inflation or deflation.Depending on theoutcome of the CPI,money hastobebrought into or withdrawnfromthe system. Theprocess of anecessary withdrawal is ensured by the demurrage. Accordingtohistorically useddesigns, the demurrage rangesfrom0.5 %to1%per month,which wasassessed as sufficientmechanism to reduce thequantityofcirculating money.[Sch08] This leads to the question of howtoincrease the quantity of therespective crypto currency in the system.Hence, three pillars were determined as emission possibilities,whereatitisvital to balance the total emission of allthree possibilities withthe withdrawalbythe demurrage and the demand formoney. First, as outlinedinsection 2.2.9 (Crypto currencies), crypto currencies areused as areward for miners to secure the (Bitcoin-) blockchain.Hence, to enable adecentralizedapproach, this concept is crucialand will perdesign emitcrypto currencies into the system. Second, participants who joinorparticipants who leavethe network willincrease or lowerthe moneysupply.Tocounter this riskofhigh variabilityasubscription modeliscontemplated.Thereby, participantsagree to exchange afixed amount of fiat currenciestothe crypto currency for afixed periodoftime. Thesubscriptions should be predetermined, includingamaximumofpossiblesubscriptions to prevent an overflowofthe system. Furthermore, theamountofpossiblesubscriptionsshould be adapted to the conditions inside and outside of thecurrency system. Third,ifthe totalmining reward doesn’tequalthe amountofthe withdrawn demurrage, theremaining coinshavetobe remitted into the system (if no inflation is present).One way to execute this distribution could be an equal paymentfor everyparticipant. Regardingthis option, it has to be mentionedthat Gesell emphasized on the necessityofbringing moneyintocirculation by buying commodities. He arguesthat forcredits, it is notpossibletowithdrawthe money from the system anymore

27 Background in case of arecession, thus limiting thepossibilities to regulatethe money supply.Nocredit is provided forthe stated option, whichrenders it feasible. Nonetheless, thiswould eliminatethe fixed withdrawalofmoney by the demurrage stated before,hinderingthe mediation of inflation. Therefore, and in reference to theidea of Gesell,anotherway of (re-)distributing money is to purchase commodities as donation for projects which are chosenbythe network. This option wasalso implemented duringthe successful Freigeld experiment“the wonder of W¨orgl”.Besides the positivesocial impact, this would allow ahigher marginofquantitycontrol, whichmightbe beneficial forasuccessful longtermregulation of themoney supply. It is therefore concluded that this would be asufficientway to facilitatethe emissionofcrypto currenciesintothe system. Aprecisecalculation of therelations exceeded theextentofthisthesis. Regarding the exchange of fiat money for Freigeldcrypto currencies, it is necessary to determine an exchange rate. This is acrucialquestion, especiallyifthe Freigeld currencyisdesigned as complementarycurrency(see section 2.3.2). In this regard, it has to be ensured that the variabilityofthe conventional monetary systemdoesnot directly effect thevalue of the Freigeld cryptocurrency throughthe exchangerate.Neglectingotherdistorting impacts, it can be stated that, forasystem which features more than one currency, the value of money can be held constant by adapting theexchange rateaccordingtothe market or the exchangerate can be held constant which willlead to an adaptation of thevalue of money.Thus, it wasconcluded that,inthis particularexample, the aspired constant moneyvaluedemandsaflexible exchange rate.One solutionmightbethe comparison of the CPI of different currencies, whereatanexistent inflation or deflation of the currency and (if available) aprofound longterm and short term outlook should be included in the calculation. Nonetheless, this is only asuggestion andshouldbeanswered by academicexperts. In reference to the elaboration provideditwas concludedthatunderthe premiseofacorrect implementation, the effective demandisapplicablefor crypto currencies. It has to be acknowledged thatacryptocurrency named Freecoin already implemented parts of theFreigeld idea. [2]Itwas not possible to assessifthe statedcurrency complies accurately with the theoryofthe effectivedemand,nor wasitpossible to assess if it is currently in use. A considerableamountofofficial information is providedonline, but abalance as emphasizedby the effectivedemand is not mentioned.The outlinedcircumstanceswere considered significant enough to elaboratethe applicationofthe effectivedemand on crypto currencies in more detail.

2.4.4The circle coin -alocalFreigeldcrypto currency

As concluded in the previous chapter, it is theoretically possibletodevelop acrypto currency whichresembles theideasofGesell’s effective demand. Hence,onthis basis ahypothetical local Freigeldcrypto currency scenario wasdeveloped.Due to its assessment by asocial survey,the information provided waskept very concisetominimize the riskofaresponse bias.Therefore, the concept of the localFreigeld cryptocurrency wassplit into three parts. The first partintroduced the conceptofcrypto currencies (see fig. 2.7), the second part introducedthe conceptofFreigeld (see fig. 2.8)and the thirdpart combinedbothconcepts, enablingthe descriptionofthe local Freigeld crypto currency named circle coin (see fig. 2.9). The design of circle coin implemented the theoryofeffectivedemand as outlinedinsection 2.4.3 and agreedwiththe stated definitions of acomplementarycurrency.Two attributes were added to the conceptstated before:

28 Background

1. The support of thelocal economy.

2. An incentive forprivate business who use the circle coin.

Thedecision to support the localeconomyoriginates from thegoal to implementacomplementary currency. In detail, the circle coin encourages regionalitybylimiting its areal validity. Hence, it was contemplatedthatthe valid area(50 -100 km) for thecoin holderisdeterminedbythe respective registered residence. Additionally, the coin can only be spentatregistered partners. In general, onlyafew facilities were banned frombeing registeredbusiness partner.This includesfacilities who providebasic needs suchaswater or electricity.Thesepartnersare presumablyinnoneedof support by acomplementarycurrencies and aregistration of such partners would favorindividuals livingnear thehead office. Thesame goes for taxation bills. The circle coinfollowsthe ideaof Gesell: money forproducts. Thedecision to provide incentives originates from theidea to increase thegrowth rate during the early stage of the regional crypto currency.Therefore, thescenario outlines aparticular option for registered private businesses, which allowsabeneficialexchangeoffiat currenciesfor circle coin. To be precise,individuals who signeduptothe circle coin communityagreed to exchange 10 %oftheir salary into circle coin. Hence,topromote the useofcirclecoins and to lowerthe barriers for privatebusiness, thelatter were allowed to exchange 9%ofthe salaryinconventional money for10%ofcircle coins.Amoredetailed conceptofdifferentsubscriptions exceeded the timeframe of the survey.The description of circle coin usedfor thesurvey is given in fig. 2.9.

Figure2.7: Thisfigureshows the introductionofcrypto currenciespresented as first partofthe conducted survey.

29 Background

Figure 2.8: This figure showsthe introduction of Freigeld presented as second partofthe conducted survey.

Figure 2.9: This figure shows the Freigeld crypto currencyscenario presentedasthirdpart of thecon- ducted survey.

30 3Methodology

3.1 Structure

Section 3.2 starts with adescription of thesampledesign workflow, followed by the reviewed researchobjectives, the researchquestions and thetested hypotheses (‘Survey objective’). Next, the‘Effectivityand efficiency’ofasample andthe conducted core strategyofthe sample design, named ’Total survey design’are determined. The latteraims to ensure ahighquality sample by emphasizingonfive crucial categories. Duetoits extent, the category named’The sampling design’isdescribed in the subsequentsection 3.3. Section 3.2 ends withaconcise elaboration of thefollowing important notations:

1. The nature of ’Errorand bias’ 2. ’Representative samples’ 3. ’Validity’ofasample

3.2 Sample design

3.2.1Sample design workflow

The workflow forthe developed sample design is based on theflow chart providedinFigure 3.1. Preceding to the stated workflow, the original research ideawas repeatedly adjusted and specified according to findings from differentscholarly literature. The intention wastoestablishaclearly defined and up-to-date researchquestion. Experience showedthatitwas worth emphasizing on aclear definition, although adjustments were still necessary duringthe sample designprocess. Startingfromthe base of aclearresearchquestion, atarget population,asampleframe (see section 3.2.4)and apossible sampling technique (see section 3.3)were derived. Foreachofthese given stepsare-evaluation of thepreceding stepswas donetotest for thelasting validityof designdecisions previously made. (this is indicated in fig. 3.1 by the dashedlines). Necessary adjustments were made from top to bottomuntil asamplingdesign wasfoundthat matched the chosen sampleframe,the target population and the researchquestion. Afterwards, the sample size (see section 3.2.4)was determined and the data collected accordingtothe selected sampling technique.The implemented designprocessends with the assessmentofthe response rate (see section 3.2.4).

31 Methodology

Figure 3.1: This figureshows the differentstages of the executed sample design process. Thedashed arrowsindicaterevision processes.Information taken from[Tah16,p.19].

3.2.2 Survey objective

The general researchobjectivewas to assess thelevel of trustofindividuals concerningcrypto currencies and the readiness to use them. Thiswas done by asample survey. Thetargetpopula- tion were people living in Austria.The obtained sample aims for conclusions about thegeneral acceptanceofcrypto currencies and whether alocal “Freigeld” crypto currency influences the outcome. The followingresearch objectives and questionsare investigated.

ResearchObjective

The research objectiveofthe survey wasapreliminary investigation into ...

RO1: ...whetherthe acceptance of acrypto currencywouldchange foragovernmental support.

RO2: ...whetherindividuals would accept acrypto currency for everydaypayments.

RO3: ...whetheralocal Freigeld cryptocurrency changesthe acceptance of crypto currencies.

ResearchQuestion

Three researchquestions wherederivedfromthe stated researchobjectives.

32 Methodology

RQ1: To what extentwould agovernmentalsupport change thelevel of acceptanceofcrypto currencies?

RQ2: To what extentwouldindividuals accept acrypto currency solutionfor everyday payments?

RQ3: To what extentdoesthe availability of alocal Freigeldcrypto currency change theaccep- tance of crypto currencies?

Hypotheses

As aconsequence of thestated researchquestions,three hypothesesare aimed to be tested during the evaluation of theobtained data.

H1: Governmental support increasesthe acceptanceofacrypto currency.

H2: Individualsdonot accept using crypto currenciesfor everydaypayments.

H3: The local Freigeld crypto currency exceeds the acceptance of othercrypto currencies.

3.2.3Effectivityand efficiency

The goal of a sample survey is to gather dataabout thecharacteristics of apopulation.To achievethis goal, thequalityofthe collected dataiscrucial. Thisissue can be described by the terms effectivity and efficiency.Bothare often used in business administration. Nonetheless, the definitionofeffectivityand efficiency allowaneasy and overall classification of differentsampling designs. Duetothe commonequivalentuse of both terms, an exactdefinition is provided in eq. (3.1)and eq. (3.2).

Result Effectivity= (3.1) Goal

Result Efficiency = (3.2) Effort

The general goal in social researchistogather information about apopulation. Ifthe research resultsufficiently fulfills thisgoal, eq.(3.1)will show aresultnear ’one’ and the design can be classified as effective.Leaving other biases aside, gathering information from eachunit of the whole populationwill provide aperfect result. This means we truly know 100 %ofthe opinionor characteristics of apopulation.The goal is completed and in referencetoeq. (3.1), an effectivity of 1or100 %isobtained.Nonetheless, gathering information from each individual will result in averyhigh effort. Accordingtoeq. (3.2), thisleads to lowerefficiency. Accordingtothe definitionofeffectivity(eq. (3.1)) and efficiency (eq. (3.2), it canbesaidthata surveyshould providethe highest possibleeffectivity and additionally,efficiencyshould be kept as high as possible. Thisisaimed forbythe useofsamples. Thus, sampling is an efficientway to deriveinformation from apopulation.Incase of a “good” sample,effectivityiskepthighand “good”results are obtained. Thereby,itispossible to obtain

33 Methodology informationand drawconclusions1 about thetarget population.Bychoosing asuitable sampling technique,the effort for sampling is kept low, whichkeepsthe efficiency high. The keyisthe term“good”sample. A “bad” sample willprovide results whichare less repre- sentative for thepopulation. In otherwords,incase of a “bad” sample,effectivitydecreases and projections from thesample deviate from the actual target population. (seealso Representative samples section 3.2.5, Validity section 3.2.7, Error and bias section 3.2.6) To satisfy the keypropertyofahigh effectivity, aspecial emphasishas to be placed on the quality of thesample design.Nonetheless, findingareasonable balance betweensample effectivity and sample efficiencyisone of the majorchallenges aresearcher has to face.

3.2.4 Total survey design

The introduced totalsurveydesign is basedonthe model of FowlerF.J.describedinthe book “Survey research methods“(2013). [Fow13,p.6ff.] Theauthoremphasizes that it is quite common to overlook critical issuesduring a sample design.Healsostressesthe necessityof high-qualityprocedures in “all salientareas” to minimizeerrors and biases of the obtained data. [Fow13,p.5]Therefore, he proposed an assemblyof“critical issues”, wherebyeachofthese issues is labeled “crucial”for ahigh-quality(effective) sample. This assembly is referred to as the total survey design.[Fow13,p.6ff.] Thetotal surveydesign includes:

1. The decision to implementaprobability sample or not. (Probability sampling or non-probabilitysampling)

2. The sample frame

3. The sampling design

4. The sample size

5. The rate of response

Probabilitysamplingornon-probabilitysampling

Sampling can be dividedintotwo broad types:

1. probabilitysampling (also called: random sampling)

2. non probabilitysampling (also called: non-random sampling)

Before defining asampling technique, adecision regarding which type is more suitable forthe target populationhas to be made. The distinction between probability sampling and non-probability samplingisdone by aknown or unknownselection probabilityfromafinite population. The ultimately obtained sample is called

1The extent of reasonable conclusionsishighly dependentonthe sampling process.

34 Methodology probabilitysample for aknown, or non-probabilitysample for aunknown selectionprobability. The terms probability and random,aswellasnon-probability and non random,are equivalent.

The denomination finite population can either refertothe target populationorthe sample frame (alsosampling frame).[Fow13,p.42] As stated in definition 3.4.8 (probabilitysampling), it is not necessarythateachitem hasthe same probability. However, if differentprobabilitiesare present, theyneedtobeadjusted by weighting.[Fow13,p.21]

Thus,asamplecan onlybedescribedasrandom if theprobabilityofselection is known. Asample is called non-random,ifthe probability is not known. Thismightseemcontradictory in thefirst instance, whichcould be the reason for atendencyofthe impreciseuse of theterm random.

Only for probability samples (see definition 3.4.8), it is possibleto“estimate accurately therela- tionship between the sample statistics and the population from which it was drawn” [Fow13,p.16]. With probability samples, it is also possible to minimize the personal bias through therandom selection process and to drawconclusions about the target population.

Fornon-probability samples (seedefinition 3.4.10), it is onlypossibletodescribethe obtained data. Thereby, it has to be emphasizedthat non-probability samples feature presumablylittle to no statisticalsupport for conclusions about the population.[Fow13,p.2f.]

Compared to probabilitysampling, non-probability samplingtends to be lesstime-consuming. Especially in social surveys, the effort of selectingindividuals by arandomized samplingtech- niqueisconsiderablyhigh. Hence,atendency forahigher efficiency of non-probabilitysamples is observable.Ingeneral, consulting individuals can be seenasacommon bottleneckfor social surveys. In probabilitysampling, only afew options are availabletoestablish contactwith indi- viduals.Therefore, in specific scenarios, conducting aprobabilitysample mightnot be feasible.

However, dependingonthe scenario,other biases thanthe personal biasmight be moredecisive, e.g. theaccessibility of individuals. Thismay lead to better results withanon-probability sample design. Nevertheless,the general recommendation for scientific research is to implement aprobability samplewhenever possible.Ifsubstantialdrawbacksoccur, non-probabilitysamples mayalso be evaluated. However, in the end,the specific scenario definesthe best suited approach.

Figure 3.2 showsexamples of common samplingtechniques. Amore detaileddiscussionofthe specific techniques stated is giveninthe sections ’Probability sampling techniques’(section 3.3.1) and ’Non probabilitysampling techniques’(section 3.3.2).

The sample frame

The sample frame (also sampling frame) representsthe setofpeople who have thechance to be sampledfromadefinedpopulation. It provides aclearpicture of theincludedand excluded units. Asurvey is only representativefor thepart of thepopulation which is includedinthe sample frame.Inversely,the higher thedeviation of thesample framefrom the target population, the higherthe introduced sample frame bias.

Anarrowing of thesample frame tends to increase sampling efficiency,but alsoincreasesthe extent of excluded units.Additionally,the boundaries of thesample frame are strongly influenced by availablemonetary fundingand timeframe.Considering all thegiven aspects, it can be said that thesample frame is very decisive for thefinal sampling design. [Fow13,p.16]

35 Methodology

Figure 3.2: This figure shows examplesofdifferentsampling techniques, separated into twomain types: probabilitysamplingand non-probabilitysampling. Data takenfrom[Tah16,p.20].

Thesamplingdesign

The sampledesign (orsampling technique) describes howthe sample is drawn from thechosen population.

Forsocial surveys, this canalso be described as howpeople are grouped, contacted and ultimately sampled. Differences in e.g. surveyduration, accessibility of participants or financial support may lead to differentmethodsofchoice.

More details about differenttechniques are giveninProbabilitysampling techniques (section 3.3.1) and Non probabilitysampling techniques (section 3.3.2). Additionalinformation concerning the selection procedure of asamplingtechniqueisprovided in the subsequent sections Representative samples, Error and bias and Validity (section 3.2.5,section 3.2.6 and section 3.2.7).

Thesample size

The sample size can alsobedescribed as thetotal number of units chosen to be evaluated. It is not advisedtodefine thetotal number of samples beforehand,especially without theconsideration of the wholesurvey design. Numerousaspects (including the goal of the survey) influence the sample size and vice versa.Ingeneral, the sample size hasaconsiderableeffect on the sampling error and the standarderror of mean.[Bry12,p.196] Nonetheless, focusingonapure minimizationof both errors might increase other errors to ahigherextent. [Fow13,p.38]

Therate of response

The calculation of the response rate(or rate of response) wasnot strictly defined in thereviewed literature ([Bry12, ES10, Kel15, Fow13]).Equation(3.3)shows thedefinition, whichwas ulti- matelyused to calculate theresponserate.Thus, onlyusable data setsand people who have

36 Methodology been accessibleare includedintothe calculation. Incomplete or evidently wrong data setsare rejected. Equation (3.3)isbased on [Bry12,p.199].

number of usable datasets rateofresponse= (3.3) selected units − unsuitableorunreachable units

Therate of response provides information about theproportionofunits who didn’t provide usabledata sets. Thisincludes participantswho refused to provideinformation, or failedto correctly executethe questionnaire. If asignificantpart of theselected sampledoesn’tprovide usabledata, asignificant amountofinformationisexcluded from thesample.This systematic error is addressed by the nonresponsebias.(see ’Error and bias’and definition 3.4.19)For eachrejectedresponse, thenonresponsebiasincreasesfor theselected sample.This effects the representativeness of thesample and the validity of theconclusions -drawn fromthe sample about thetarget population-islikely to decrease.Hence, theresponserate provides akey figure to evaluate thenonresponse bias of asample. (see also ’Validity’and ’Representative samples’) [Kel15,p.170]

3.2.5 Representativesamples

A representative sample is desired for every study.Nonetheless, thereasonings whether asample is representative, are very versatile. Awidespreadmain criterion for arepresentativesample is thataselected unitmusthaveaknown probabilityofselection. Hence, under this premise, a representative sample canonly be achievedbyprobability samples. Although aprobabilitysample is commonly referred to as necessary, it is notsufficient.Dependingonother incorporated errors and biases,asample maystill not be representative.[Fow13,p.38f.]

In general, if the amount of errors and biases includedinthe sample increases, the representa- tivenessofasampledecreases and the generalization from thesample to thepopulation leads to wrong assumptions. Thiscircumstance canalso be described by the previously introduced sampling effectivity. If theeffectivityofasampling processdeclines, theobtained conclusions dissociate more andmorefromthe truecharacteristics of the targetpopulation. Unfortunately, it is notpossible to defineaspecificlevel of effectivitywhich guaranteesarepresentative sample. It has to be saidthat, depending on other errors and biases,anon-probabilitysample might be even morerepresentative than aprobability sample, but since someofthe errorsand biases are just not measurable, this is not possible to verify.(seealso Errorand bias,section 3.2.6) Duetothis fact, scientists sometimes choose to comparetheir studies with other similar studies, whichclaim to be representative. This mayhelptounderlinethe representativeness of the own sample,but incorporates otherdrawbacks. Comparingsamples to derivethe representativeness is complex,sincethe wholesample design has to be included in the evaluation. (The sample size is not enough.) Other factorsmight be decisivesuchasthe sampled population, the appearance of thecorporationand manymore. [Fow13,p.38f.]

To counterthis problemand to ensure maximum efficiency,(standardized) strategies, approaches and processhavebeendeveloped. Oneofthem is the already introduced total surveydesign. Nonetheless, significantweightisput on theresponsiblescientist to ensure the representativeness of asampleand to clearly outline the limitations, allowingthe correctinterpretation of the obtained data.

37 Methodology

3.2.6 Errorand bias

As outlined, a sample survey generates statistics about atargetpopulation. Thesamplestatistics are then used todrawconclusionsabout the characteristics of the target population. During the sample design and the evaluationofthe data,different typesoferrors mayoccur. Hence, survey methodology is used to minimizeand measure these errors (as far as possible).[Fow13,p.8ff.] In general, two fundamental premises are of importance for asamplesurvey.[Fow13,p.8ff.]

1. The samplerepresents the characteristicsofatargetpopulation.(see Representative sam- ples,section 3.2.5) 2. The informationgathered from thesampled units accuratelydescribes thecharacteristics of theunits.

Thegiven fundamental premiseslead to twofundamental sources of error.[Fow13,p.8ff.]

1. The distributionofunits in the sample is notrepresenting the target population. 2. The information gathered fromthe units is no accurate measure.

Forsocial surveys, thesefundamental errorscan also be summarizedbytwo mainissues.[Fow13, p. 8ff.]

1. Howwellthe givenanswers measure thecharacteristics examined. Fowler (2014) describes this with the umbrella term: errorsassociatedwith answers (p. 11). 2. Howclosely the sample mirrors the population. Fowler (2014) describes this with the umbrella term: errors associatedwith who answers (p. 9).

The fundamental premisesand their relation are summarizedbyFigure 3.3 . Thefirst main issue - errors associatedwith answers -covers the difference of what was intendedtobemeasuredand what wasactuallymeasured by the receiveddata. [Fow13,p.8ff.] Manypossibilities exist to commit such errors. Compared to thenonresponse bias, units do try to provide information inthe firstplace, butstill fail to providethe requested data.Toname afew examples for social surveys, this could be amisunderstanding of thequestion,not enough information to answerthe questionoradeceivinganswertolookgood. In general,this type of error is present in twodifferentcategories: objective facts and subjectivestates. In theory,itis possible to verify objective facts by measurement(e.g. thegiven height of unit).Inpractice this is often notfeasible (e.g. theevaluation of howmanycigarettessomeone smokes).Incontrast, thereisnoway to verify the subjectivestate of aunit (e.g. Howmanyhoursaperson felt tired during the last days). This systematic difference between theobtained data fromanindividual and the“truth” is defined as response bias.[Fow13,p.8ff.] Thesecondmain issue - the error associatedwith whoanswers -isrelated to the sampled populationand howclosely it mirrors the target population. [Fow13,p.8ff.] This issuefeatures two distinctivecategories: Errorsbased on variation and errors based on systematicdeviations. [Fow13,p.8ff.]

38 Methodology

Figure 3.3: Thisfigureshows the dependence of thetwo fundamental premisesofsocial surveys in aflow chart.The two fundamental (possible) errors (or issues) are stated on the right. Theflow chart itselfshows the desired outcome of asample. Theinformationwas takenfrom[Fow13, p. 8ff.].

1. Errorsbased on variation. Independently of the samplingtechnique, thereisapossibility thatthe sample characteris- ticsrandomlydiffer from thepopulation. This is knownassamplingerror. Forexample, apopulation consists of 50 malesand 50 females,fromwhichasampleof10 people is randomlydrawn.The deviationofthe sample proportionfromthe population proportion(50 %males and 50 %females) is the sampling error.

2. Errorsbased on systematicdeviations. Thiscategoryoferrors is referredtoassampling bias. A biased sample leads to results, who on average, consistently differfrom thetargetpopu- lation. Thereare three possiblewaysabias can be introduced: [Fow13,p.10]

(a) The sample frame bias If certain unitsofthe target population are excluded by the presentsample frame, the result will showacontinuous deviation of thesample results from the actualcharac- teristics of thepopulation. (b) The selection bias An imperfect processofselection ultimatelyleads to aselection bias. Hence,every form of structuring or stratification is likely to introduce some amountofselection bias. Thiscan be due to unknown connections in the processofgroupingunits or due to the use of non-probability samples.Depending on the design, especially non-probability samplestendtofeature significantlyhigher selectionbias compared to probability samples. [Fow13,p.42] (c)The nonresponse bias The nonresponsebiasdescribes unitswhichare partofthe defined sampleframe,but due to the propertiesofsome units it is notpossible to collect thedata. Forsocial surveys, this might be the case,because unitsare unwillingtoprovide information.

39 Methodology

Another examplemight be thatpeoplejustdon’twanttointeractpersonally with thesurveyprovider. If the non-responding units are differentfromthe rest, abias is introducedintothe sample. The higher the rateofunitswithout aresponse,the higher is the possibility of an included nonresponse bias. [ES10,p.305]

Errors(arising from variations) and biases (arising from systematicdeviations) are not necessarily related to eachother. It is absolutely possibletoobtainsurveyresultswith lowsample variations, but aconsistentlyhighbiasand vice versa. [Fow13,p.11]

Giventhe stated definitions, it can be concludedthat unreachable units are not part of the nonresponse bias. Unreachable units are systematically excluded from thesampled population and should,accordingtodefinition 3.4.17,betreated as sample frame bias.

3.2.7 Validity

The dataobtainedfromasingle unit (xi)can be described as avalue consistingoftwo factors: The true value of thedata(ti)and thedeviation (di)introducedbyaunit. [Fow13,p.11]

xi = ti + di (3.4)

The accumulated deviation can result in arandom error, originating fromindividual data sets whichsometimesindicate onedirection and sometimes another. In this regard,the terms less/low certaintyorconfidence are used.Yet it mightalsobethe case thatthe data sets continuously deviate in onedirectionfromthe true value. This is consequentlyreferredtoasbias. [Fow13, p. 11]

Bothdeviations, the random error and thecontinuous bias, are collectivelyrepresented by the validity of asample.Itisanindicator of howaccurate the intended variablesare measured. [ES10,p.444]

Hence,all calculations concerning validity aim to assess howwellthe answers representthe value whichisdesired to be measured. The validity is then usedtodescribethe relation betweenthe received and the true value.

Thelargerthe deviation di,the higher the chancethataunit will be in error. Viceversa, ahigh validitymeansalowdeviation di and thus alow possibility thataunit deviatesfromthe true value.

It should be noted thatthe validity of subjectivestates can’tbeassesseddirectly.Commonly, this circumstance is addressed by incorporatingsimilar measures and comparisons. Nonetheless, sinceitisnot possible to assess thetrue value of asubjective fact, it is also notpossibletoassess thepresumably accompanyingsystematic bias.

40 Methodology

3.3 Samplingtechniques

3.3.1 Probabilitysampling techniques

Simple random sampling

The most basic sampling form is simple random sampling.It“approximates drawingasample outofthe hat: Members of apopulation areselectedone at atime,independent of one another and withoutreplacement; onceaunit is selected, it has no further chancetobeselected.” [Fow13, p. 18]

To executesimplerandomsampling, it is essential to have alist of the examined population. If a list can’t be provided,simple random sampling(as single technique) can’t be executed.[Fow13, p. 18 ff.]

The process of selection has two featuresworthmentioning. First,the random selection process from alistalmost entirely eliminates selection bias.Second, sincethe sampling is done from an existing list, the process does not dependonthe immediate accessibilityofthe selected unit. [Bry12,p.191 f.]

Especially in social surveys, the accessibility of units is an importantfactor. In acompanyfor example,alistofall employees is readily available, but due to the holidayseason,most of the employees are not accessible.

If no information of theconstitution of apopulation is provided, it is often suggestedtouse simple randomsampling. [Fli15,p.101 f.]

Systematic sampling

In general, systematic sampling can be usedintwo ways. [Fow13,p.18ff.]

First, if alist exists and the totalnumberofthe population is known,the defined sample size canbeused to calculatethe step size of thesystematic sampling process. This means thatifthe population is represented by alistof1000 entries andthe sample size is 100, every tenthperson is pickedfromthe list. (see eq. (3.5)) In doing so, two things have to be considered.

1. The list has to be randomized. Anon-randomized list implies aprecedingsorting or group- ingprocess. Properly done,asorting process mightbenefitthe sample, but only if the conductedprocessisknown and includedinthe calculation. [Fow13,p.18ff.]

2. The randomized startingpoint mightcause an overflow. Forexample,for alistwith 1,000 entries,the entry number 992 wasrandomlydeterminedasstartingpoint. With astep sizeoften,the next selected number wouldbe1,002. Thus, theoverflowis2and therefore the ultimatelysampledunit is the second entry in the list. From there on, the sampling continues with the definedstep sizeuntil thesample size is reached.Anoverflowcan be avoided by setting the random starting pointbetween thestep size and zero. [Fow13, p. 18 ff.]

41 Methodology

population step size= (3.5) sample size

Second,the units are directly selected fromthe sample frame.This facilitates the big advantage of generatingalist while sampling. Similar to asystematic samplefrom alist, it is importantthat this approachdoesn’t followa previoussorting or groupingprocess thatisnot includedinthe sampling design. If asorting or groupingisdone, theprobability of selection must be known and includedinthe succeeding calculations.Hence,acorrect implementation retains thevalidity of a’probabilitysample’. Generally,ifmore than one methodisused,itisreferred to as multi stagesampling, which is discussed in Section 3.3.3.One example for generatingalistwhile samplingisthe sampling of individuals in ashopping center.The shoppingcenter has to be randomlyselected,whichincludes aknown probability. Next,e.g. every thirdcustomer (step size) wholeavesthe shoppingcenter is selected for thesample.[Fow13,p.18ff.] (Note: If more than one exit is available, it has to be included in the probability calculation. [BHL96])

Stratified random sampling

The aim of stratified random sampling is to maximize the heterogeneity and representatives of the samplebygroupingand/or sortingofthe population priortothe sampling.The process itself is called stratification and therebyformedsub-populations arecalled stratum (pl. strata or stratums). Hence,toexecute astratification some characteristics of apopulation must be known. [Fow13,p.18ff.] In simple randomsamplingorsystematic sampling,the selection processisindependentofthe previousselections.This caneither leadtosample characteristics whichare equaltothe popula- tion or to characteristicswhich differ from thepopulation. Sampling10peoplefrom agroup of 50 malesand 50 females maycontain 10 malesand 0females, or 0males and 10 females.This potential samplingerror is lessened or eliminated by stratification. [Fow13,p.18ff.] Thedifferentgroups obtained are sub-populations and ideallyrepresent the different proportions of the total population. If apopulation is stratified for example by male and female, theresult will be two groups of males and females. From eachsubgroup, (e.g.)asimple random sample is drawn. Thecombined sample of all sub-population willalwaysfeature the same proportions of males and females as thetotal population.Hence, information concerning minorities mightbe especiallyusefulfor stratification. In general,asmall degree of stratification is simple to implementand if done thoroughly,itmight be beneficialfor thesample.Nonetheless, due to the stratification,itisimportanttohandle the selection probabilityproperly.The latter has to be eitherequal forall sub-populations or adjusted by weighting.The latter maybeusedtominimizethe samplingeffort and thus increase efficiency. The differentpossibilities are subsequently discussed by two examples[Fow13,p.18ff.]: Example A-Equal probabilities of selection:The population of students on acampusconsistsof 80 %malesand 20 %females. Astratification has been doneaccording to the given percentages, 1 resultinginamale and femalesub-population. Forthe stratifiedrandom sample, 10 of eachsub- populationisrandomly selected. Hence, theselection probabilities are equalfor eachsubgroup and thecombinedsample obtained willfeature the characteristics of thepopulation. [Fow13, p. 18 ff.]

42 Methodology

Example B-Different probabilities of selection:The population of studentsonacampus consists of 80 %males and 20 %females. The goal is to select at least40students from eachgroup. 1 Apopulation of 1,000 studentsissampled viastratifiedrandom sampling at arate of 10 .This means that 100 students are selectedfromthe population -80malesand 20 females. To obtainthe additional 20 femalestudents, it is necessary to drawanother100 studentsfromthe population. Thereby, the goal of 40 femalestudentsisreached, but it also adds80males to thesurvey which wouldn’t have been necessary forthe surveygoal. Thisprocess canbeimprovedbydifferent selection probabilities.Inthis case,one option is to doublethe sampling rate of femalestudents 1 -the groupofmale students is sampled at arate of 10 ,the groupoffemalestudents at arate 2 of 10 .Thereby,80maleand 40 female students are selected. In total,instead of 200 students, only120 have to be sampled.The efficiency increases considerably. To combine samples of differentselection rates, it is importanttocompensatefor thedifferentrates by weighting.In this particularcase,sincefemalestudentswere sampled at adoubledselection rate, the weight of thefemalestudents has to be half the weightofthe malestudents. [Fow13,p.18ff.] Stratification with differentselection probabilities can be used forparticular variablegroups. This oversampling mayhelptoimprove the overallresult.For moredetails it is referred to Groves (2004) [Gro04]. Furthermore,stratificationcan be done with more thanone criterion, forexample forstudents: Gender and Faculty. Nonetheless,itisonly feasible,ifitispossibletoidentify and allocatethe characteristicsofagroup. [Bry12,p.192 f.]

3.3.2Non probabilitysamplingtechniques

Convenience sampling

Aconvenience sample is selected based on the mereavailabilityofparticipants. [Bry12,p.201 ff.] Typical samplingscenarios are events (e.g. concerts, sports),locations (e.g.shopping malls) or streetselection. Another frequentapproachisthe selection based on social relations (e.g. family, colleagues, friends, ...).[VTS16,p.328] The high availabilityofparticipants has time and cost advantages and mayalso feature avery high response rate. Nonetheless, given the selection process, it is not possible to knowthe population of thesample. Therefore, it is not possible to generalize the results. [Bry12,p.201 ff.] Convenience samples are commonly used in specific situations,suchaspilot testing, in situations where an outstanding sampling opportunityoccursorwhere anysample is better thannosample. [Bry12,p.201 ff.] Albeittheserestrictions, convenience samplingplays abigger role than often expected.The received data might not be representative, but it canbeused as afirst trail forfurtherresearch or to showconnections between different areas.[Bry12,p.201 ff.]

Snowball sampling

The name snowballsampling originates from thegraphicalinterpretation of thedistribution and is showninfig. 3.4 (picturetaken from [1]). Eachlayer of distributors leadstoanincreasing number of sampled subjects. The resulting picture is known fromacommon avalanche type, which initially starts with asmallsnowball.

43 Methodology

Figure 3.4: Theselection processofunits in snowball sampling. Each circledepicts aselected and sampled unit.Picture taken from [1].

In theory, the starting pointfor snowballsampling is randomly chosen. In practice, this is hard to achieve. Therefore the startingpoint is usually determined by convenience sampling and oftenthe researcher himselfstarts to distribute the sample. Nevertheless, alsofor random startingpoints, snowball sampling is no probabilitysamplesince it is not possible to determine the selection probabilityofaunit. This compromises representativeness,especiallyfor statistical evaluations of quantitativesurveys.Therefore,snowball sampling is more presentinthe areaofqualitative research. [BB12,p.60] Either way,the sample highly dependsonthe initial startingpoint and is biased towardsthe individualpersonal network and cooperativesubjects. Thisselection bias can be lessened by strategiestolimit the number of “possible furtherdistributions persubject” (e.g. alimited number of coupons perperson).However, an easy and fastdistributionisone of the main benefits of snowball sampling andadecisiontolimitthis benefitshould be made carefully.[BB12,p.60] Atkinson and Flint(2001,p.2)suggest that:“[...]the problem of aselection bias may be partiallyaddressed, firstly through thegeneration of alarge sample and secondly by the replication of resultstostrengthen any generalizations.” [AF01,p.2] Besides the selection bias,otherbiases associated with thedata collection such as, sampling bias (see definition 3.4.15), samplingbias and response bias arise.[BB12,p.60] [Bry12,p.202 f.] Comparing asnowballsample with randomized samples of thesame population can helpto identify and tackle thementioned errorand to improve the representativeness (Witte et al., 2000). Nonetheless, the representativeness of aprobabilitysamplewill neverbereachedsinceitisnot possible to calculatethe preciseinfluenceofthe conductedmeasures on therepresentativeness. [BB12][Bry12,p.201 ff.] All things considered, in acertain situation, snowballsampling still has its value. Especially if thepopulation is hard to determine, snowballsamplingmay be the onlyfeasible technique (e.g.asinstudies about the usage of drugs).Commonly, snowballsampling is usedinsettings where respondents are hard to reach,hard to involveorfew in numbers. Moreover, it is used in population where trustplays amajor roleorinstudies wherethe aim is to showrelationsbetween respondents.[BB12,p.60] [Bry12,p.202 f.]

44 Methodology

Quotasampling

Quota sampling usesasimilarapproachasstratifiedsampling. Its aimistoproduce asample that ideallyshows the same proportionsasthe targetpopulation. Similar as to stratifiedsampling, the categories are defined preceding the sampling process. This also includesthe number of sampled unitsfor eachcategory,e.g.50males, between 30 and 40 years with amaster’s degree and 50 females between 40 and 50 years and adoctoral degree.Inthis regard,the categories are also referred to as quotas.Unlikeinstratifiedsampling, theselection of theunits is notcarriedout randomly.The interviewer is liabletosearch forunits whomatch the defined categories. This introduces convenience sampling into thequotasampling approach, which is the main difference to stratifiedsampling. Forthis reason, quotasamples are classified as non-randomized samples. [Bry12,p.203 f.] Similar to snowball sampling, theselection bias strongly correlates with the interviewer. It is again biasedtowardsthe personal network and cooperativesubjects,althoughinthis case, even more weight maybeplaced on the personal network. [Bry12,p.203 f.] Duetoits design,quota sampling includes commondrawbacks of non-probability samples.The user of quotasampling has to argueagainst the lowrepresentativeness of thesample caused by personal biasand the under-representation of units who were not easytoreach. Thejudgmental influence of theintervieweralso introducesthe errorofunits assignedtowrong categories. It is alsonot feasible to calculatethe standard error of mean for aquota sample since the population is not known (a non-randomsample).[Bry12,p.203 f.] Quota sampling featuresthe typical benefits of non-probability samples such as high efficiency and lowcost.The main advantage of quota samplingisthe lowdurationofthe sampling process, especially comparedtoprobability samples.[Bry12,p.203 f.]

3.3.3 Single stage andmulti stage sampling

In single stagesampling,only one technique is used to createasample. [Fow13,p.18] In multi stage sampling more techniques are used, which can either be the same ineachstage(e.g. more layers of stratification) or differentfromstage to stage (e.g. acombination of stratification and systematic sampling). As long as more than one sampling layerisimplemented, it is called multi stage sampling. [Bry12,p.193 f.][Fli15,p.102 f.][Fow13,p.21] Multistagesampling is oftenusedifthe populationisnot available on alist and whenitis particularlydifficult to accessthe population.[Fow13,p.21] Oneexample might be the population of all Austrianuniversitystudents. Alist of students might notbeavailable and if so, driving throughthe whole countrytosamplethe studentsisnot very efficient. Hence,inmulti stage sampling, alist of universities canbeused to sample the universitiesinthe first stage. In thesecond stage, thestudents are sampledfrom theuniversities selected.Compared to theinitialsampling from thelist, the efficiency is increased, whereby, whenimplementedcorrectly theeffectivityiskept high. In this regard, threeterms are often used: cluster, clustersampling and areasampling. In accountwithdefinition 3.4.25 (stratum)and definition 3.4.29 (cluster), it canbesaidthat groupings,clusters and strataofapopulation are synonyms.The same canbesaidfor cluster sampling and stratifiedrandom sampling withmore than one layer. Also, no definition omitsthe

45 Methodology

Probabilityof Probabilityof selection at selection at Overall stage 1 stage 2 samplingrate (universities) (dormitories) 20 universities 1 1 ( 5 ), and 2 of 1 1 1 a) thehousing 5 x 2 = 10 unitsare selected 10 universities 1 ( 10 ),and all 1 1 1 b) belonging 10 x 1 = 10 housing units are selected

Table 3.1: Thistableshows an example of area probabilitysampling (regional stratification). The effort of sampling case a)(20 universities) is considerable higherthen the effort ofcase b) (10 universities). Thisresults in ahigher sampling efficiency for b). By adapting theselection probability of differentstages, the overallsamplingrate can be held constant, whichenables acomparison of differentsampling designs. Informationtaken from [Fow13,p.21ff.] usage of thegiven termsinnon-probability samplesetting. Still, the term strata has astrong relation to stratified random samplingand therefore to probabilitysampling. According to definition 3.4.30 (cluster sampling)and definition 3.4.28 (multi stage sampling), cluster sampling is asynonym formulti stage sampling. [ES10,p.292] To simplifythe terminology, cluster sampling and cluster are omitted and theterm stratum (pl. strataorstratums) is only usedfor probabilitysamples. Anotherfrequent samplingissueisthe lowaccessibility of units due to aregionally widespread population. This is addressed by amulti stage sampling approachcalled area probability sampling. Similar to astratificationprocess, the target area is divided into subareas withclearboundaries. These subareas (orsub-populations) are then randomly selected untilthe last sub-population consisting of the target units is reached.Theseunits are then sampled. Again,ademarcation to regional stratification or area stratification is notstated in thereviewed literature. ([Fow13, Fli15, Bry12, ES10]) It is therefore assumed, that areaprobabilitysamplingisasynonym for regional stratification. Independentofthe name, the overallselection probabilityisanimportant factor forcomparingdifferentarea probability samplingdesigns. Thisisoutlined in the example below. (Example C) Example C-Area probability sampling:(Thisexample is illustrated in table 3.1) The goal is to sample thepopulation of economicstudents which liveincorresponding dormitories. Alist of universities and respectivedormitories is available. In total, 100 universitiesare sampled. Table 3.1 showstwo sampledesigns with differentprobabilitiesofselection, butanoverall equal samplingrate. Area probability sampling (or regional stratification) is therebyused to lower the overallsampling effort by lowering the necessitytotravel. Approachb)shows the lowest traveleffort.Depending on the diversityofthe universities, approacha)islikely to have sample estimates of highervalidity, sincethe attributes of studentsare likely to vary foreachuniversity. Moreover, not onlydifferentattributes, but alsodifferentsizes of universitiesand housing units have not been considered. This mayhaveaneffect on thevalidityofthe samplealthough, as

46 Methodology stated in [Fow13,p.21ff.], samplestendtohaveahigher validity,ifthe laststage(here stage 2) featuresanequalnumberofunits. Theparticular proceedings are case-specific, but especiallyif area probabilitysampling is includedinthe sample design, the issues given should be included in the evaluation. Similar to clusters and clustersampling,itwas decided to omit theterm area probability sampling and use regional/areal stratification from nowonfor reasons of simplicity.

3.3.4Self administered onlinesurveys

The conventional formsofself-administeredquestionnaires, or also referredtoasself-completion questionnaires, are postal,mail or onlinequestionnaires. Especially over the last 20 years, the methodofself-administered online surveys increased itstotal shareinself-administered question- naires.[EM05]

Advantages

-Globalreachand largesample size Due to widespreadinternetaccess especially in industrializedcountries,online question- naireshavethe advantageofasuperior reachbeyond theborders of nation states. This alsoenablesthe assembling of alarge samplesize with highefficiency.

-Flexibility Differentformats andadaptedversions of questionnaires, individuallydesigned for respon- dents are possible. Common examplesare languagechoice, order control of questions (or answers), the skipping of questionsand theuse of high question diversitysuchasmultiple choice, scales, open-ended questionsetc.

-Speed and timeliness Online surveys allowatime-efficientadministration,whichultimatelylowers thetime needed to collect and processthe data.

-Technological innovations Simplifiedoptions and moreattractivepresentations are possible,evenfor respondents with less expertise (e.g. instruction videos instead of texts)

-Convenience “respondentscan answer at aconvenient timefor themselves;” [BB12,p.62]

-Ease of data entry andanalysis Thecollected dataisdirectly stored in adatabase, which is easily accessible for subsequent evaluations.

-Low administration cost Personal interviewers are not required.Itismoreover possibletooutsourcethe whole process to specialized onlinequestionnairecompanies.

47 Methodology

-Ease of follow-up An online follow-up of non-respondents is easier duetothe lowexpense of an additional contact. -Controlled sampling Thisrefers to databases,ormailing lists, created by opt-insurveys. One prominentexample is the feedback collection of customers,whichalso allows to enhance customer relationship. By gatheringknowledge of respondents, it is alsopossible to conduct estimations about the characteristicsofnon-respondents(marketingresearch).

Disadvantages

-Perceptionasjunk mail Thisalso includesrequests coming from trustedsources. Additionally,there is apossibility of rejectionsbypre-filters,suchasmail servers. [BB12,p.62f.] -Technological variations Problems and differentuser experiencesdue to differentinternet connections and technical set-ups, suchaspersonal computers, smartphones,monitors, etc. -Skewedattributesofthe internet population The selectionbias between theofflineand online target population and their relatedat- tributeslikegender, age, education level, etc.Nonetheless, due to the advancing digital- ization, thegap between both sub-populationsseems to decrease. [BB12,p.63] [EM05, p. 209] -Questions about sampleselection and implementation Depending on thesampledesign, some selection methods mightbebased on volunteer samples,where participants e.g. go on awebsite and participateproactively. Others might be based on randomized selection with aself-administered survey during the last sampling stage. Duetothis broad varietyofsample selection and implementation in self-administered online questionnaires, manyquestions concerning representativenessand validityhavetobe answeredduringdesign. -Ambiguousansweringinstructions Instructions must be extremelyclear to all participants. -Impersonal Brown et al. (2001)state thatthe potentialfor participants to pause and reflect is higher in telephonesurveysthan inonline surveys. [BBM01] -Privacy issues It has to be taken careofsecuredata transmission, datahandling, datastorage and data usage. -Low response rate The stated disadvantages often lead to lowresponserates if no emphasis is laid on their moderation.

48 Methodology

Theoutlined main advantages and disadvantages of onlinequestionnaires were derived fromthe researchofEvans and Mathur (2005). [EM05]

3.4 Definitions

Section: Sampledesign workflow

Definition3.4.1 (sample). “A selectedsubsetofapopulation chosen by some process usually withthe objectiveofinvestigating particular properties of theparentpopulation.” [ES10,p.376]

Definition 3.4.2 (population). “(...)the universeofunits,fromwhichthe sample is to be selected.”[Bry12,p.187]

The term unit is themost general notation for objects of interest. [Bry12,p.187] In social surveys the terms‘people’,‘individuals’or‘subjects’are also commonly used. [Bry12, Fli15, Fow13]

Definition 3.4.3 (sampling). “The process of selecting some partofapopulation to observe so as to estimatesomethingofinterest about thewhole population.” [ES10,p.377]

Definition 3.4.4 (samplingdesign). “The procedure by which asample of units is selected from thepopulation.” [ES10,p.377]

Definition 3.4.5 (sampledesign). “The process which leads to the developmentand selection of aspecific samplingdesign and theultimatelyobtained sample.” [ES10,p.377]

Accordingtodefinition 3.4.4,itiscorrect to equally usethe terms samplingdesign and sampling technique(s),asbothrefer to theactual selection strategy.Toclarify, the sample is drawnby sampling,whichisdone according to the samplingdesign that is part of the sample design.All of themhaveinfluenceonthe characteristics obtained by the sample.

Section: Survey objective

Definition 3.4.6 (sample survey). “A study that collectsplanned information fromasample of individuals about their history,habits, knowledge,attitudes or behaviour in order to estimate particularpopulation characteristics.” [ES10,p.377]

Section: Effectivityand efficiency

Definition 3.4.7 (efficiency). “A term applied in thecontext of comparing differentmethods of estimatingthe same parameter; the estimate with the lowest variance beingregarded as the most efficient. Also used whencomparing competingexperimental designs,with one designbeing more efficientthananotherifitcan achievewith fewer resources.” [ES10,p.149]

49 Methodology

Section: Total survey design

Definition 3.4.8 (probability sampling). “A sampleobtainedbyamethodinwhichevery indi- vidualinafinite population hasaknown (but notnecessarilyequal) chanceofbeing included in the sample.” [ES10,p.340]

Definition 3.4.9 (sampleframe(also sampling frame)). “The portionofthe population from which thesample is selected.” [ES10,p.377]

Definition 3.4.10 (non probabilitysampling). Asampling methodwhere the selectionprobabil- ityofanitem is notknown or the unequal probabilities can’tbeadjusted by weighting. [Bry12, p. 187]

Definition 3.4.11 (samplesize). “The number of individuals to be includedinaninvestigation. Usuallychosensothatthe studyhas aparticular powerofdetectinganeffectofaparticular size.” [ES10,p.376]

Definition 3.4.12 (responserate (or rate of response)). Refers to the proportionbetween the number of usabledata setsand thenumberofeligibleunits of aselected sample.[Fow13,p.42] [Bry12,p.199]

Section: Representativesamples

Definition 3.4.13 (representative sample). “A sample reflectingthe populationsoaccurately that it is amicrocosm of the population.” [Bry12,p.187]

Section: Error andbias

Definition 3.4.14 (responsebias). “The systematic component of thedifference between infor- mation provided by survey respondentand the ‘truth’.” [ES10,p.369]

Definition 3.4.15 (sampling error). “The difference between thesampleresultand the popula- tion characteristic being estimated.” [ES10,p.377] Its origin lies in therandom variationofthe sampling process and the sole fact thatthe sample is only asmallpart of thepopulation. It is therefore categorizedasrandom (sampling) error. [Fow13,p.10]

Definition 3.4.16 (sampling bias). Asystematicdeviation of therelation between thesample and thepopulation. [Fow13,p.10] It is thereforecategorizedasnon-random (sampling) error.

Definition 3.4.17 (sample frame bias). Abias originating fromthe difference between the population and the sample frame.Itisproportionaltothe excluded part of thetargetpopulation. [Fow13,p.10]

Definition 3.4.18 (selection bias). The bias that maybeintroduced into all typesofscientific investigations, whenever aselection is done by the involved unit or it is subject to constraints thatgounobserved by the researcher.[Fow13,p.42] [ES10,p.386]

Definition 3.4.19 (nonresponsebias). Afailure introducedbyunits whoare unable or refuse to providerelevantinformation. [Fow13,p.42] [ES10,p.305]

50 Methodology

Section: Validity

Definition3.4.20 (validity). “Theextent to whichameasuring instrumentismeasuring what wasintended.”[ES10,p.444]

Section: Probability sampling techniques

Definition 3.4.21 (simple random sampling). “A formofsampling design in which ndistinct units are selected form the Nunits in a finitepopulation in such away thatevery possible combination of nunits is equally likely to be the sample selected. (...)” [ES10,p.394] Definition 3.4.22 (systematic sampling). Systematic samplingprovidesanequal precision(va- lidity)assimple randomsampling. Thereby, afixed stepsize and arandom startingpointis definedand the population(or thesampleframe) is sampled by the defined step size. [Fow13, p. 18 ff.] Definition 3.4.23 (stratified random sampling). Astratified random samplingapproachuses “thepossibility of structuringthe samplingprocess to reduce the normal sampling variation, therebyproducing asamplethatismore likelytolooklikethe totalpopulation than asimple randomsample.” [Fow13,p.18] Definition 3.4.24 (stratification). The processofgroupingand/or sortingofapopulation, according to knowncharacteristics of population, prior to asampling process.[Fow13,p.18ff.] Definition3.4.25 (stratum). Groups or sub-populationsformedbystratification are also called stratum(pl. strata or stratums). [Bry12,p.192 f.][Fli15,p.102 f.]

Section: Non probabilitysampling techniques

Definition 3.4.26 (snowballsampling). “[. ..]atechniquefor finding research subjects.One subject gives the researcher the name of another subject, whointurn providesthe name of a third,and so on. Thisstrategycan be viewed as aresponsetoovercoming the problems associated with sampling concealed hard to reach populations [. ..]”[AF01,p.1]

Section: Single stageand multistagesampling

Definition 3.4.27 (single stage sampling). In singlestagesampling, only one samplingtechnique is used to create asample. [Fow13,p.18] Definition 3.4.28 (multi stage sampling). Multi stage sampling combines multiple sampling techniques to improvethe effectivityand efficiency of asample. It isn’t defined as one spe- cific sampling technique -the mix of sampling techniquesisadjusted according to the sampling scenario. [Bry12,p.193 f.][Fli15,p.102 f.] [Fow13,p.21] Definition 3.4.29 (cluster). In cluster sampling, theunits of apopulation are arranged in groups. These groups are referred to as clusters.[ES10,p.87] Definition 3.4.30 (clustersampling). “Withcluster sampling,the primary sampling unit(...) is not the units of population to be sampled butgroupings of those units.Itisthe latter groupings or aggregations of population units that areknown as clusters.” [Bry12,p.193]

51 Methodology

Section: Self administered online surveys

Definition 3.4.31 (self-administered online surveys). “Respondents answerquestionsbycom- pletingthe questionnaire themselves.” [Bry12,p.232 ff.] The term onlinerefers to the compelling necessityofatemporary or continuousinternet access, which also implies theavailability and cor- rect use of electronicsrequired.

52 4Implementation

4.1 Original sample

Theoriginal sample design usesamulti stage sampling strategy to conductarandomized sample of theAustrian population. Differentstages of stratification are implementedand an emphasis waslaidoneffectivityand efficiency. The approachsamplesthe Austrian population via food retailers.The main idea wastoutilize apresumedstrong overlap of thepopulation of food-retailer-visitors with the total Austrian population. Nonetheless, it has to be saidthatall the peoplewho arenot abletoorrefuse to go to food retailers are therebyomittedfromthe sampling process. This results in in a sample frame bias which is accepted by the executionofthis approach. It can be saidthatthis approach is quite common. The noveltyofthe developed sample design lies in the preceding multi stage stratification process. Hence,the followingchapterpresents adetailed description of thedifferent stratification layers. It startswiththe first stratification of thetotal sampled populationand ends with the smallestobtained strata from which asystematic randomized sample (see ‘Systematic sampling’) is drawn.

4.1.1Total survey design: Original sample

1. Probabilitysampling or non-probability sampling In reference to the outlined benefitsofprobabilitysampling(see section 3.2 and section 3.3), it wasdecided to implement aprobabilitysample.

2. The sample frame The sampleframe wasset to peoplewith their country of residenceinAustria, whovisit food retailers in Austria. The target populationwas set to peoplewith their countryofresidence in Austria.

3. The sampling design Aself-administered,systematicprobabilitysamplingapproach, usingdifferentstages of stratification. Adetailed description is provided in section 4.1.2.

4. The sample size The aspired samplesize wasset to 150 usabledata sets.

53 Implementation

5. The rateofresponse Therate of response is assessedbythe number of approached individuals and the usable data sets ultimately received.

4.1.2 Sampling design: Original sample

Layer1.1:Stratification of Austrian food retailers

In Austria, themarket of food retailers is dominated by two corporations, the “Spar Os¨ terreichis- cheWarenhandels-AG” and the“REWEInternational AG”. Thislowers thenumberofdifferent retailersand simplifiesthe sampling process. [3]

Table 4.1 showsthe differentmarket sharesofthe top10foodretailers (data wastaken from [3]). The top retailersnamed “Billa”, “Merkur”,“Penny” and “ADEG”are part of “REWE International AG“. [4]“Spar/Eurospar” and “Interspar/Maximarkt” arepartofthe “Spar Os¨ ter- reichischeWarenhandels-AG“. [5]

Marketshares of Austrian food retailers # Name [%] 1 Spar/Eurospar 24.3 2 Hofer 19.3 3 Billa 17.4 4 Merkur 8.8 5 Interspar/Maximarkt 7.9 6 Lidl 7.0 7 Penny 4.1 8 M-Preis/T&G 3.7 9 Adeg 2.1 10 Unimarkt 1.3

Table 4.1: In this table, thetop 10 Austrian food retailersare listed and ordered in termsoftheir market share(in %). The data wastaken from [3].

By mergingthe retailers according to their superiorcooperation,itisvisible (seetable 4.2)that four main companiesreachacombined market share of 90.90 %. It wastherefore decidedto sample onlyfromthe four dominant companies.Based on thegiven market coverage, it was assumed that theeffectivityloss is acceptable compared to the efficiency gainofthe sampling process. Hence, initial stratification wasdone accordingtothe given fourdominantcorporations depicted in table 4.2.

To ensure an integrated approach, threeruleswhereestablished for stratification:

Definition 4.1.1 (firststratification rule). The stratification aimstoresult in aminimum number of stores.

Definition 4.1.2 (second stratification rule). If the first stratification rule leadstoadistinctive situation,the result with higher relation to the original shares of thepopulation is favored.

54 Implementation

Number Market Number Corporation/Company of stores Share share of stores (rounded) REWEInternational AG 32.40 % 4.63 5 36 % Spar Ost¨ erreichische Warenhandels-AG 32.20 % 4.6 5 36 % Hofer 19.30 % 2.76 3 21 % Lidl 7.00 % 1 1 7% Total 90.90 % 12.99 14 100 %

Table 4.2: Thistableshows the top 10 Austrian food retailers. The differentretailersfromtable 4.1 were merged according to their superiorcorporation.The column“Number of stores” shows the intermediate stratificationstep, the column “Numberofstores(rounded))” showsthe result of the initialstratificationofthe Austrian food retailers.

Definition4.1.3 (third stratification rule). In caseswhere both rules are present, more weight is put on rulenumberone.

To optimize thesampling process, thegoal wastofind the minimum number of storestobe sampledand simultaneously ensuring arepresentativesamplefor thecalculated four dominant corporations. In reference to this goal,the lowest number of stores to be sampled from atop cor- poration is one.Therefore, this number wasassigned to the company withthe lowest percentage ratefromthe topfour dominantcompanies. In this case,the number one wasassigned to the company called “Lidl”. From thenon, the proportionswere extrapolated.The outcome is shown in table 4.2.The obtainednumbers of stores to sample were rounded to avaluewithout decimal points.The result is in line with both outlinedrules forstratification. First, the number of stores is minimized by startingfromthe lowest shares.Second, no otherproportionwas found withless deviation as theshares givenintable 4.2. Concluding, thefirst layerofstratification results in 14 stores to be sampled, covering 90.90 %of theAustrianpopulation that visitsfood retailers.

Layer1.2: Stratification of corporations

As outlinedabove,the twomain corporations are “Spar Oste¨ rreichischeWarenhandels-AG“ and “REWEInternational AG“. Bothinclude differentstore types. Theirstratification is shownin this sub-chapter.

Layer1.2.1: Stratification of “REWE International AG”

The description of thefollowing stratification is supported by table 4.3 and table 4.4. Four stores (brands) of the “REWEInternational AG”are part of the top10foodretailers (see table 4.2 and table 4.3). As obtained during the first levelofstratification, fivestoresofthe “REWEInternational AG”havetobesampled. Table 4.3 shows from left to righteachstore type of theREWEGroup, their associatedmarket shares and the number of stores to sample. Stratification stepone: If the store with the lowest share is sampled once, therateslead to 15 stores to be sampledintotal. Thisexceeds the maximumoffive stores from thefirststratification layer(see table 4.3). Thus, inreference to the firststratificationrule,the element wasset to zero.

55 Implementation

Stratification steptwo:Ifthe store with the second lowest shareissampledonce,the total amountofstores to be sampledisseven.This againexceeds themaximumoffive stores again. By following thismethod(basedonfirst stratification rule), two possible outcomes areobtained. Result A,sampling threestoresof“Billa” and two storesof“Merkur” or ResultB,sampling four stores of “Billa” and one store of “Merkur”.

Followingthe second stratification rule,Result Awas determined as final stratification result. Theconfrontation of thedifferentshares of Result Aand Result Bisdepicted in table 4.4.

Stores Stores Stores Stores Stores Stores Rewe Market to to to to to to Group share sample sample sample sample sample sample Billa 17.40 % 8 4 2 4 3 4 Merkur 8.80 % 4 2 1 2 2 1 Penny 4.10 % 2 1 0 0 0 0 ADEG 2.10 % 1 0 0 0 0 0 Total 32.40 % 15 7 3 6 5 5

Table 4.3: The tableshows the first stepofthe stratification process for sub-companies of the REWE International AG.The differentmarketshares are derived from table 4.1.The goal wasto find adistributionfor atotal number of fivestores (see table 4.2). Thecalculation starts with the company withthe lowest market share,setting its “stores to sample count” to one. The relativenumberofstores are then derived fromthe market share. If the total number of stores to sampleexceeds the requested number (of fivestores), the “storestosamplecount”ofthe companywith the lowest market share is set to zero forcontinuing calculation. In the end, this processresultedintwo possible strata. The finalcomparison is depicted in table 4.4.

Rewe Market Relative Result Relative Result Relative Group share share A Share A B Share B Billa 17.40 % 66.41 % 3 60 % 4 80 % Merkur 8.80 % 33.59 % 2 40 % 1 20 % Penny 4.10 % 0 0 ADEG 2.10 % 0 0 Total 32.40 % 5 5

Table 4.4: Thistableshows the second stratificationprocess of the“REWE International AG”corpora- tion. The top resultsfrom table 4.3 arecomparedtothe relativemarket share of theremaining companies (brands). As depicted, the relativeshare of Result A exhibits the highestoverlap with the relativeshare of theremaining companies. Accordingtothe second stratification rule,Result Awas determined as final stratification result.

Layer1.2.2:Stratification of “Spar Ost¨ erreichische Warenhandels-AG”

The stratification of stores from “Spar Oster¨ reichische Warenhandels-AG” follows the samepro- cedureasthe stratification of “REWE International AG”. Duringthe weighting process,two possible results whereobtained. By following the second stratification rule,Result Awas deter- minedasthe finalstratificationresult. The processisshown by table 4.5,followedbytable 4.6.

56 Implementation

Stores Stores Spar Market to to Group share sample sample Spar/Eurospar 24,30 % 4 3 Interspar/Maximarkt 7,90 % 1 2 Total 32,20% 5 5

Table 4.5: Firststratificationprocess forsub-companies of the “Spar Oste¨ rreichischeWarenhandels- AG“,resulting in twopossible outcomes.Anillustration of thefinalcomparison is provided in table 4.6 .

Spar Market Relative Result Relative Result Relative Groupe share share A shareA B shareB Spar/Eurospar 24,30 % 75,47 % 4 80 % 3 60 % Interspar/Maximarkt 7,90 % 24,53 % 1 20 % 2 40 % Total 32,20 % 5 5

Table 4.6: Secondstratificationprocess of the “Spar Oster¨ reichische Warenhandels-AG“. Accordingto the second stratificationrule,ResultAwasdetermined as final stratification result. (A more detailed descriptionofthe processisgiven in table 4.4.)

Result Layer1:Stratification of Austrian food retailers

After thestratification of both layers,14storesare arranged to be sampledintotal.During the whole process, thedeterminedrulesofstratification wherefollowed. The distribution of the stores finally calculated is shownintable 4.7.

Stratification of Austrian food retailers Stores Name to sample Spar/Eurospar 4 Billa 3 Hofer 3 Merkur 2 Interspar/Maximarkt 1 Lidl 1 Total 14

Table4.7: This table shows the finalstratificationresult of Austrian food retailers.

Layer2:Stratification of theurbanand rural population

The demographics of Austria exhibitonly six citieswith over 100,000 inhabitants.“Vienna”, thecapital of Austria, has by farthe biggest population of 1,911,191 inhabitants[6],followed by “Graz” with291,072 inhabitants [6]. Due to this populationdistribution, astratification

57 Implementation between theurbanand ruralpopulation of Austria can be considered reasonable. Additionally, it is considerednot uncommon to find diverse opinions between rural and urban inhabitantsin Austria, although no evidence canbeprovided to verifythis claim. Forthe given research topic, it wasdecided to stratify the Austrian populationof8,901,064inhab- itants [6]intotwo groupstominimize apossiblesamplingerror. Thefirstgroup are people living in cities, while the secondgroup are people living in rural areas. Eachmunicipalitywith 10,000 inhabitantsormore wasconsidered acity. The obtained result shows apopulation distribution where 52 % 1 of theAustrianpopulation live in ruralaresand 48 % 1 live in urban areas, with a considerable shareofVienna (21 % 1,see table 4.8). Datataken from [6].

Headcount Share 1 Total Austrian population 8,901,064 100 % Rural population 4,632,350 52 % Urban population 2,357,523 26 % Vienna 1,911,191 21 %

Table 4.8: Urban and rural population distributionofAustria. Asoutlined above,municipalitiesover 10,000 inhabitantswere consideredascities and assigned to the urban population. Vienna exhibits an outstanding share within the Austrian urbanpopulation and wastherefore specified as separategroup.

It wastherefore decidedtostratify the 14 food retailers previously defined into two groups. According to thestratification rules stated before, 7food retailershavetobesampled in urban areas and 7food retailersinrural areas. As visibleintable 4.8,the capital“Vienna” has a dominantweightinthe urbanarea subgroup.Itwas thereforedecidedtoadditionally stratify the urban area into two strata. Accordingtothe givenshares andinreference to the second stratification rule,4stores will be sampled in urbanareas(excl. “Vienna”) and 3storesin “Vienna”. The detailedstratification of thefood retailers and theAustrianpopulation wasdone in two steps. The first layerofstratification settles the distributionbetween rural areas (7 food retailers)and urbanareasincludingVienna(7food retailers). Thesecondlayer stratifies theseven food retailers for the urban area between “Vienna” and theremainingurban areas. Startingwith the first layerofstratification, the even numbers of stores to be sampled were split up between theurban and rural areas.Stores with oddsamplingnumbers were then distributed in reference to the secondstratification rule (weighting of shares). As visibleintable 4.9 (column number 2), stratification conflicts for thecompanies “Hofer”, “Billa”, “Lidl”and “Interspar/- Maximarkt” had to be solved. “Billa” and “Interspar/Maximarkt” arepartofthe nationwide competingcorporations“REWE Internationl AG”and “Spar Oste¨ rreichische Warenhandels-AG”. “Hofer” (part of theInternational corporation named“ALDI”) and “Lidl”are both setupascom- peting discounters on an international level. Concerning the first competitorsand in reference to the second stratification rule (weighting of shares),more weight wasput on the market leader “REWEInternational AG”. Thus, thecompany“Billa” will be sampled in theareawith the higher population share, therural area. Its counterpart,the company“Interspar/Maximarkt”, will consequently be sampled in urban areas.The same strategy wasusedfor “Hofer”, which

1 The numbers include roundingerrors.

58 Implementation has thehigher market share and will thereforebesampled in rural areas.Thus, “Lidl”will be sampledinurban areas. Theresultare shown in table 4.9,columns3and 4. The second layerofstratification concerning theurbanareasstartedagain by distributingthe even numbers of stores to be sampled. Next, theinternal competitors named “Billa” and “Merkur”of “REWEInternational AG”wereallocated. According to definition 4.1.2,the market-share-leader “Billa” will be sampled in the area with the higher population share, the urbanarea excluding “Vienna”. “Merkur” willtherefore be sampled in “Vienna”. Similarly,“Hofer” will be sampled in theurbanarea excluding Vienna and“Lidl”in“Vienna”. Since four companies in total have to be sampledinurban areasexcluding Vienna, thecompany“Interspar/Maximarkt” will be sampled in theremaining urban area, as this area features ahigher market share. (second stratification rule -weighting by shares). The finalresult is depicted in table 4.9 -columns3,5and 6.

Strataof Rural Urban area Urban Austria Vienna food retailers area (incl. Vienna) area Spar/Eurospar 4 2 2 1 1 Billa 3 2 1 1 Hofer 3 2 1 1 Merkur 2 1 1 1 Interspar/Maximarkt 1 0 1 1 Lidl 1 0 1 1 Total 14 7 7 4 3

Table 4.9: This figureshows the stratification processoffoodretailersinAustria between the rural population, the urbanpopulationexcl. Vienna and Vienna.First, the 14 stores previously determined were stratified between therural areaand the totalurban area (including Vienna). Next, the totalurbanarea wasstratified betweenVienna the remaining urban areas.The wholeprocess followedthe stratificationrules of definition 4.1.1 and definition 4.1.2.

Layer3:Areaprobabilitysampling

As outlined in section 3.3.3,area probabilitysampling is commonly used to increase sampling efficiency.For the given research objective,differentregional stratifications are possible. The approachfor this studyuses thefederal states of Austriatoestablish aregional stratifica- tion. In general,the nation state “Austria” exhibits ninefederalstatesnamed “Burgenland”, “Steiermark”, “Wien” (Vienna),“Nieder¨o sterreich”, “Ober¨osterreich”, “Salzburg”, “Tirol”, “Vo- rarlberg”, “K¨arnten”. Duetoits weight,the capitalVienna wasalready determined in ‘Layer2: Stratification of theurbanand ruralpopulation’asstrata with 3stores. Hence, adecision had to be made on howtosamplethe eleven remaining food retailers fromthe eight remainingfederal states.Inthis regard, four options maybeconsidered:

1. Sampling eleven storesfromone federalstate, randomly chosen from theeight remaining federal states.

2. Sampling eleven stores from two (or more) federalstates, randomly chosenfromthe eight remaining federal states. Similar to theprevious stratification layers,astratification has to be done according to the stratificationrules. Thus, ahigherweighthas to be placed on the federal state exhibitingahigher share of theAustrianpopulation.

59 Implementation

3. Samplingeleven stores from one federal state, conveniently chosen from theeightremaining federal states. Thismay be the federalstate of e.g. themain residenceorhometownofthe researching person.

4. Samplingeleven stores from more than one federal state, conveniently chosen from theeight remaining federal states.

The positive impact of drawing aconvenience sample frommore than one federal stateoneffectiv- ityisassumed to be insignificantlylow.This strategy would onlyresult in a confirmed efficiency loss. It is therefore not advised to drawaconveniencesample from more than one federal state. The decision betweenthe remaining threeoptions reliesonthe time available to the executing person and thesignificance of arandomized sample. In reference to section 3.2.4 and section 3.3.1, an emphasis should alwaysbeplaced on establishing aprobabilitysample.Still, the time is often alimitingfactor. Including thedecisive factoroftime available, it wasdecided to choose option one.The reasonisthe preservationofaprobabilitysample,combinedwith an agreeabledegree of sampling efficiency.

Selection of food retailers and systematic randomized sampling

The sampling methodultimatelyconducted is called systematic random sampling. Thereby, it is possibletodrawarandomized samplewithout the existence of acompletelistofthe target population. In detail, it wasdecided that every thirdperson who leavesthe selected food retailer is chosen for thesample.Thus, alist of thesampled target population is createdsimultaneously to the sampling process. The selectedperson is then asked to be part of the self-administered survey. As emphasized by Sudman (1980, p. 426),the peopleare chosen from lefttoright in case of a tie or necessarysorting. Thiscan happen e.g. if agroupofpeopleoracoupleleavesthe store simultaneously. Forfoodretailers withmore thanone entrance,aselection bias mayoccur. To be precise,ifthe opinion of visitors from one entrancediffers fromthat of another entrance, the sampling time per entrance has to be adjusted in proportiontothe visitor flowofeachentrance. This mayresult in considerableeffort. Differentfactors should be includedinthe considerationofthe necessityof the outlinedpartitioning. If one combined parking areawith only one roadaccess existsfor two entrances for example, an even split will most probably be sufficient. Additionally,experience showedthat the majorityofstores has onlyone entrance. Nonetheless, special cases mightcause additional effort to keep the selection bias as lowaspossible. In this regard,the researchof Bruweret. al. (1996) mightbeparticularly useful. [BHL96] In general,the sampletime perstorewas set to three hours, excluding arrival and set-up time. Thereby,itshould be possibletosampletwo stores perday.Hence, 14 stores can be sampled evenly distributed over one week. Three time slots were determined -morning, noon and after- noon. The slotswere randomly assignedunder the premise of two samplingslots perweekday. Next, under the assumption of asuccessfully obtained list of stores, thestoresare assigned to the weekdays in atwo step process. Instep one, thestores, whichare categorized in federalstates and brands, are randomly selectedfromthe list according to the stratificationcalculated before (stores to sample). In steptwo,the storesselected are randomly assigned to timeslots (and weekdays).

60 Implementation

Setup of theself-administered survey

-AnA2flip chart (or similar) and astanding table are placed at avisible, but notdisturbing position near the entranceofthe food retailer. The postershould provideconcise infor- mation about theplannedselection of leaving visitors.Thereby,potentially chosen people have the chance to prepare to be selected. Universitylogos (or similar) shouldbeshown on thechart to underline the professional purpose and to increase the level of trust of the visitors.

-The survey wasdesigned to be completed online or offline (onpaper).

-The selected person is asked to either do the offline survey now, on the standing table, or onlineataconvenient pointintime. If the participant chooses to do it online, aflyerwith the login details is provided.

-Similar to theA2flip chart,the flyer shouldfeatureimportantlogos, thename and thelink (alsoasQR-code) of thesurvey.

-Dependingonthe situation concerning theSARS-CoV-2virus,glovesand amask should be worn and new pencils and disinfectants should be provided.

-Toassess theresponserate,itisvital to countthe number of all selectedpeople. Combined with thetotal dataset (online and offline), the response rate can then be readily calculated.

Note:The peoplewith the main residence in Austriaare filtered by aspecific question during theself-administeredsurvey.

Apreliminary investigation during theafternoon opening hours showedthat within 20 minutes, 27 visitors left theconveniently chosen food retailer.The frequency of visitors wasconsidered moderate (two open cash registers,with an average of 2-3 visitors waitingfor thecheckout).If 27 visitorsleave astore per20minutesand every third is selected forthe survey,27visitors are selected perhour. Three hours of samplingtime perstore leads to 81 selected units perstore. Thereby, 162 unitsare selectedper dayand 1,134 perweek. Includinganestimated response rate of 25 %, asampleof283 (rounded)visitors is obtained over theperiodofone week.Ithas to be noted thatthis calculation is very approximativeand the number ultimatelyreceivedisvery likely to deviate from this estimation. Nonetheless, it wasdecided to include the given calculation to shareinitialconsiderations.

4.1.3Summary: Original sample

The original surveydesign features aprobabilitysampleofthe Austrian population. Thesample is limitedbythe sample frame, whichisdefined by allpeoplewho are able and choose to go to Austrian food retailers.Different layers of stratification are usedtomaximize the efficiency of thesampling process. In theend, arandomized sample is drawn by systematicsampling. The dataiscollectedonlineand offline by aself-administered survey to decrease personal contact.

61 Implementation

4.1.4 Limitations: Original sample

The high level of samplingefficiency features ahigh riskofincluding an unforeseen selection biases. Due to the rejectionofthe original design,itwas not possibletoassess theresponse rate and consequentlyitwas not possibletoassess thenon-response bias. Nonetheless, the research of (Bush &Hair,(1985, p. 165) showsthatthe non-response rate varies between 10 %and 50 %. Thisindicates the possibility of aconsiderablenon-responsebias. In this particularexample, the sampleframe includes90.90 %ofall customersofAustrian food retailers. Thetargetpopulation were people withtheir main residence in Austria. In reference to the givendescription, it wasassumed that thesample frame includesalarge extentofthe target population andtherefore legitimates the general approach. It wasalso assumedthatthe opinion of people who do notgotomain food retailers onlylightlydeviates from theopinion of people who do visit main food retailers.The sample frame biaswas thereforeconsidered insignificant. Nonetheless,this is onlyanassumption and up to now, it wasnot measuredhow closelythe sample frame(visitors of food retailers with main residenceinAustria)matches the targetpopulation (people with main residence in Austria).The same goes forthe deviationof opinions between visitors of main food retailers and theremaining food retailers.Hence,no evidencecan be providedtosupport both claims and further research is needed to verifythese decisions. Thesystematic randomsampling methodfinally conductedfeaturessome additional drawbacks which have to be mentioned. First,frequentshoppers have agreater probabilityofbeingselected. Second,there is the possibility of arespondentunnoticeablyavoiding selection. Third,due to differentlocations and selection times, theparticipants feature differentprobabilitiesofselection. Thisintroduces asignificantrisk of a(time-based) selectionbias.[BHL96]

4.2Statement: Change of the surveydesign

The possibilities to contact theAustrianpopulation for randomized sample design are limited. Optionslikerandom-digit dialingorrandom mail surveyspresume the availabilityofand accessto an exhaustivedata base. The best (and presumableonly) data base in this regard is the Austrian centralpopulation register (German: zentrales melderegister (ZMR)).Its access is very restricted and onlyallowedfor certaininstitutionsand/or individuals, excluding theTUWien.Itwas also not possibletoevaluate the integrityofpublic phone books, whichweretherefore classified insufficient. This leftpresumable onlyone possibility -the public approachofindividuals, e.g. as outline in the originalsurvey design.Irrespectiveofthe contemplated arrangements, this option wasnot feasible duetothe SARS-CoV-2 containmentmeasures. Although an uprise in COVID-19 cases wasexpected, the lowimpact of theSARS-CoV-2 virusduring thepreceding (summer) monthsled to theconclusionofaratherunlikely second nationwide lockdown (severe contact restrictions).Unfortunately,during apress conference on 14th.November2020, the secondnationwide lockdown wasannounced by chancellorSebastian Kurz. His statement “Meet nobody, every social contact is toomuch.” punctuated the formersituationand finalized the decision to adaptthe survey design. The change of thesurvey design increasesdifferenterrors and biasesconsiderably. Nevertheless, it wasestimated that theexpectedtremendous non-response rate and lowaccessibilityofindividuals in the oldstudy design excels theissuesand drawbacks of thenew design. Summarizing, the expectedseveresampleframe biasand lowresponse rate, caused by the lowaccessibilitywas themain reason to adaptthe samplingdesign.

62 Implementation

4.3 Executed sample

4.3.1 Total survey design: Executed sample

1. Probabilitysampling or non-probability sampling In reference to the outlinedissues of theoriginal sampling design (seesection 4.2), anon- probability sampling approachwas chosen as samplingapproach.

2. The sample frame The sample framewas settopeople who are able to accessthe internet via aweb browser and have their country of residence in Austria. The approximated target population wasset to peoplewho have their countryofresidence in Austria.

3. The sampling design Aself-administered, webbrowser basedonline survey distributedbysnowballsampling. A detailed description is provided in section 4.3.2.

4. The sample size The aspired samplesize was300 usable datasets.

5. The rateofresponse The rateofresponsewas assessedbythe number of individuals approached and theusable datasetsultimately received. Adetaileddescription is providedinsection 4.3.2.

4.3.2Sample design: Executed sample

The adapted sample design uses snowballsampling to create asampleofthe target population. Theapproachfollowsthe principals described in section 3.3.2. As alreadyemphasized, arandomly choseninitialseed for snowball sampling is hard to implement. This is due to the considerable challenge of selectingarandom seed unit. Furthermore,ifarandom selection wassuccessful, theoutcome reliesstrongly on thededication of theselected seed unit.It wastherefore decidedthatthe initial seedwill be conveniently selectedand it wasdetermined to be the researcher himself. Thedistribution of the surveywas done exclusively digital. The survey wasdeveloped as webbrowser based, online self-administered survey.Thereby,itwas possibleto createaweb link to accessthe survey, which complementsthe survey distribution of the chosen sampling technique. The distributionofthe surveywas done in two ways. First, friends, family members, colleagues, etc.were contactedwith the request to participate in the survey.Individuals were approached solely by apersonal message. Themajority was contacted by the FacebookMessenger App,followed by WhatsApp. Thelink wasn’t distributed to anyFacebookgroups, WhatsApp groups (or similar) by the initial seed. Other peoplewho distributedthe survey were allowed to post it in groups.Inthis regard,the headcountofthe group wascountedasindividualsultimately contacted.Avery highpercentage of all known peoplewas contacted, with onlythree very lowprimary filters: 1. Individuals with internetwere not asked

63 Implementation to fill in the survey.2.Individuals with known little to no possibilities to use asmartphone or personal computer.3.Individuals with aknown country of residencedifferentfrom Austria. The lowpre-filterwas established to prevent an additional personalbias. Second,specific people were conveniently chosentoserve as an additional distributor, called “helper”.These people were asked to participate in the surveyand to distribute it in their own personal environment. Additionally,itwas emphasized to themthatitwouldbeverybeneficial for thesurvey if helpers chose “two or threepeople” from their ownpersonal network whocan again functionashelper.Inthis regard, associated peoplesuchasgoodfriends or family members were given as an example. If theconveniently chosen helpers agreedtodoso, further instructions were provided. This also included twopreparedtexts.One short text for thedistributionofthe survey in their ownpersonal network and a long texttoexplain importantdistribution rules. The more detailed text wasdesigned to provide aconcise introduction to the helpers of the helper. All of theselected individuals were reminded to participate in the survey one-time and all of the helpers were reminded at least once to remind theirown network to participateinthe surveyas well (including their helpers). All of thehelpers were initially contacted by apersonal phonecall to avoid misunderstandingsand stress the importance of theirassistance.Afterthe initialphone call, the texts were forwardedbythe channeloftheir choice. The permission to alter the texts according to their requirements wasgiven. All helpers were advised to countthe number of approachedindividuals. At the end of the survey period, all helpers were askedtocollect thetotal number of peoplecontacted from their helpers, add it their ownnumberofcontacted peopleand then forwardthe total number of their branch to their contact person. The aimofthis procedurewas to calculate an overallresponse rateofthe survey.Nonetheless, this strategy has its limitations. Similar to theconductedsnowball sampling strategy, the accuracy of the ultimatelyreceived number relies on theaccuracy of theindividuals to ahigh degree. It has to be admitted that thefollowing biasesare thereby introduced in the calculated response rate:

-The biasoriginating frommisconceptions duringthe introduction and the biasoriginat- ing fromimprecise operations concerning the counting of approached individuals and the distributiontogroups.

-The personal bias to increase thenumberofapproached individualstosatisfy theperson the number is reported to.

-The biasintroduced by peoplewho have been counted twice. This is due to thedesign of the conducted snowballsampling strategy,whichshows aconsiderablepossibility of overlapping personal networks.

-The bias introducedbypeoplewho distribute the survey without being listed as helpers. Thereby,the approachedindividuals are not countedand reported arespective helper.

Thefollowing measures were conducted to limit the biases stated above:

-All helpers were approached by means of apersonal and extensivephone call describing and answering questionsabout the distribution rules. An emphasis wasplaced on the topic of accuratecounting by stressing theimportance of exact measurements.

64 Implementation

-Similar to item one,itwas outlined duringthe introductionand right before the final collectionofthe total number thatamanipulated number compromises theevaluation.It wasalso outlined thatthe general support of thehelperisofvery high value,regardless of the number of peopleapproachedfinally reported.

-Ifsomeonereported that he/she alreadyhas participated in thesurvey, thehelpers were instructedtosubtractthe respectivepeoplefromtheir totalnumberofapproached individ- uals.

-The biasofsurveyredistribution without notice wasconsideredreasonablysmall. It was assumed that peoplewho makethe effortofredistributingthe surveywill either askfor permission or at least mention theirdoingtoreceive acknowledgementfor their effort. Nonetheless, no scientificallyprovenevidence can be providedfor this claim. This results in an uncontrollablebias whichdirectly influences the responserate.

It has to be admittedthat therewouldhavebeen apossibility to limit aconsiderable amountof the above-mentionedbiasesbyimplementing restrictedaccess to the survey. Thereby, individual passwordswouldhavebeen provided for each helper, sub-helper and so on.This would have allowed to monitor peculiar distribution patterns and in case of aconfirmederror, the data of the respectivebranch couldhavebeen filtered. Nonetheless, it wasdecided to omitthis option to keep theeffort for thehelpers as lowaspossible. Thereby,abetter distributionand an increased range of the surveywas anticipated. Additionally,thisapproachalso increasedthe number of potentialhelpers. To summarize, it wasnot possible to determine theextent of thetotal bias and thereforethe calculated response rateshould be considered amereguidevalue. Albeit the outlined drawbacks, it wasdecided thatamereguide value is more beneficialfor theevaluation of thesurvey thanno response rate at all. Prepared text -short: Hey, Afriend of my minedoesasurveyfor his master’sthesis. The topic is “digital money”. The duration is about 12 minutes. Would you be so kindtofill in thesurvey? https://www.soscisurvey.de/digitalesgeld/ Prepared text -long: Hey, Afriend of my mine does asurveyfor his master’sthesis. The topic is “digital money” and he would need alittle bit of support for the distribution. Ideal wouldbe, if youcan fillinthis surveyand also sendthis link to some friends,relatives or just familiar people. Important forthe analysis is, that you countthe amount of people you’ve sent the link -I’llcheck in on you in aview weeksand askfor the number, that’s it :) WhatsApp-groups aresimilarare ok,but please no groups of 30 ormoreparticipants where no-oneresponse. Of course,the best would be aprivate message. Thesurvey duration is about 12 minutes, thedeadlineis27.01.2021. Would you be so kindto help? https://www.soscisurvey.de/digitalesgeld/

65 Implementation

4.3.3 Summary:Executed sample

The sampleultimately executedisbasedonaconvenientsnowball samplingapproachwhichis commonly usedfor populations hard to access. As outlined in section 4.2,the circumstances concerning the SARS-CoV-2 virus led to this kind of environment. Complementarytothe chosen sampling technique, aweb browserbased self-administered questionnairewas developed toobtain theinformationaspired fromthe sampledindividuals. As described in section 3.3.2,snowball sampling is based on cooperatingindividuals who help to redistribute the questionnaire. Those helpers were conveniently chosen and theirredistribution carefullyinstructed. Additionally, these helpers were assigned to countand report thenumberofapproached individuals. Thisallowsan estimation of theresponserate. The whole process wasdesigned to featurenonecessity of any personal contact.

4.4 Questionnairedesign

4.4.1 Platform

To implementthe self-administered, webbrowser based online survey, theplatform of theGerman company named “SoSciSurveyGmbH” wasused.Itfeaturesavery comprehensive amountof basic and advanced options to implementaprofessional questionnaire design. Additionally, the provider allowsanunlimitedand free of charge access for non-commercial researchbyuniversity students. Unfortunately, it has to be mentionedthat the page navigation forthe questionnaire design wasavailable only in German. However, anon functioninglanguage button indicates developmentefforts.

4.4.2 General framework

The conductedself-administered online surveyfeatured13pages, wherebythree main parts named “Conventional money andcrypto currencies”, “The ‘Freigeld’ idea”and “The circle coin -alocal ‘Freigeld’ crypto currency (scenario)”werepresented to the participants. The implementedpage layout reflects the usedframework. Therefore, aconcise description of eachpage is given below. Thecomplete questionnaire is providedinappendix A.1. The wholequestionnaire only allowed to continuetothe next page if all of thequestions presented have been answered. Thisdecision wasmade to encourage completedata sets. The thereby potentially incorporatedbias,introduced by frustration which mayhaveled to wrongly executed questions, wasconsidered as lowand therefore accepted. Excluding introductory questions and those regardingdemographics,the same Likert scale was usedinthe whole questionnaire.The rangeofthe scale wasset from “fully disagree” (-3) to “fully agree(+3). It wasdecidedtouse this particularscaletoenable the value four (4) forundecided or neutral participants. Thereby, it waspossible to implement this optionwithout the necessity of an additional (separated) checkbox.Anexampleisgiven in fig. 4.1.

-Page 1: Language Amandatorychoice between theGerman or English versionofthe survey.

66 Implementation

Figure4.1: An example of theimplemented Likert scale.The chosenapproachfeatures sevenvalues, wherebythe value four (middle) is intended to serveasneutral option.

-Page 2: Author Contact detailsand information about theliableinstitutionofthe master’s thesis.

-Page 3: general data protection regulation(GDPR) Details and declarations concerning the GDPR. It wasonly allowedtocontinuethe questionnaire,ifthe participant activelyagreedtothe presented GDPR rules.

-Page 4: Introduction Theintroductionincluded the aim,durationand structure of thesurvey as well as abrief descriptionofits three mainparts. It wasonly possibletocontinuetothe next page if the participant agreedtofill in thesurvey by his “bestknowledgeand believe” and without “incomplete, nor consciously wrongly executed questions”.

-Page 5: Participant demographics In total,the survey askedfor fourspecifications:

-Sex -Age -Countryofresidence -Highesteducationcompleted

-Page 6: Part1: Conventional money and crypto currencies First,ashort introductiontooutline thedifferencesbetween conventionalmoney and crypto currencies wasgiven (see fig. 2.7), followedbythe assessment of the participant’slevel of acceptance(levels of “readiness-to-use”and “trust”) for crypto currencies. This included the assessment of theacceptance of cryptocurrencieswith and without governmental support, as defined in the Coreconcept of thesurvey(see section 4.4.3).

-Page 7: An optional page for participants who had already purchased or obtained cryptocurrencies. Forthis particularsubgroup,additionalquestions concerningtrust and usage of crypto currencies were provided.

-Page 8: Continued theassessment of theacceptance for agovernmentallysupported crypto currency.

-Page 9: Part2:The Freigeld idea Thispage wasused to provideaconcise introductionabout thegeneral Freigeld idea. (see fig. 2.8)Hence,the questions presented subsequently enquirethe postulated features of the

67 Implementation

Freigeldidea.These includedthe “monthly tribute” (the demurrage),the “constant money value”and aquestion concerningthe satisfaction with our current monetaryand economic system.

-Page 10: Ashort information about theremaining duration of thequestionnairewas provided.

-Page 11: Part 3: The circle coin -alocal Freigeldcrypto currency Ascenario describingthe featuresofthe localFreigeld crypto currency named “circle coin” wasoutlined.(see fig. 2.9)The acceptanceofthe circle coin wasthen assessedaccording to the Core concept (see section 4.4.3).

-Page 12: The questionnaireended by presenting specific questions with the aim to assess the influence of differentparameterswhichmay have hadasignificantimpact on the overallresults. Therefore, general questions from Part 2wereexpandedtothe localFreigeld scenario. Additionally,the influenceof“regionality” wasincludedinthe evaluation.

-Page 13: The last pagepresented ashort acknowledgmentand contactdetails. The German version included areference for additional information concerning the Freigeld topic.

4.4.3 Core concept

According to the stated Research Objective and thederived Research Question the core issue wasdetermined as theassessment of differentlevelsofacceptance and thefollowingconcept was therefore developed. To enableanincremental assessmentofthe subjectivevalue of “acceptance”, it wasseparated into two fractions.The “readiness-to-use” the presentedcurrencyand the belonging levelof“trust”. Theinitialideawas: If someone decides to useaproduct, some amountofacceptance has to be present. Theacceptanceofthe product riseswith the level of usage, but no matterhow highthe level of usage is, the productwill neverbefully accepted as long as theuser doesn’t “(fully) trust” the product. Therefore, it wasconcludedthat the acceptanceofdifferentcurrency scenarios has to be measured as thesum of the two parameters: Thelevel of “usage” (or the “readiness-to-use”) and thelevel of “trust”.

The“readiness-to-use”crypto currencies

Thereadinesstouse crypto currencies wasassessed by presumable four incremental objectives.

1. Icould imagine tryingout crypto currencies forone day. Info: I’llget atest account andwithit, Ican tryout transactions and the general function- ality.

2. Icould imagine tryingout crypto currencies as an additional means of payment for one year. Info: I’ll be introduced to the topic for one year and I’m able to testtransactions with small amounts of money.

68 Implementation

3. Icould imagine settlingasmall part of my digital everydaypayments with crypto currencies. Info: ForexampleinsteadofDebit/Credit card payments,Apple Pay, ...

4. Icouldimagine receiving asmallpart of my salary in acrypto currency.

Hence, thefour presumable incremental steps were named:“shorttry-out”,“longtry-out”,“ev- eryday payments” and “salary”. To summarize, thefour presumableincremental stepsenquirethe following objectives of the participant:

1. The participantisready for a short try-out.

2. The participantisready for a long try-out.

3. The participantisready to use thecurrencyfor everyday payments.

4. The participantisready to receive acrypto currency as part of his/her salary.

The fourobjectiveswere enquired forevery currency. Thus, thethree presented scenarios (crypto currencies, governmental crypto currenciesand alocal Freigeld cryptocurrency)and the four presumable incremental objectives resultedina4x3 question matrix. Theresult is shown in table 4.10.Eachelementrepresentsone objective and one currencyscenario. The distinct question number belongstothe questionofthe implementedquestionnaire.Bytransferringthe outcome of eachquestionintothe questionmatrix, atwo dimensional variable is obtained. Thisvariable is then used to assessthe “readiness-to-use” in agradual manner over allscenarios and objectives.

The level of “trust” for crypto currencies

The level of “trust”was assessed by the Likert scale previously described. Similartothe assess- mentofthe “readiness-to-use”,the same question waspresented forcrypto currencies, govern- mentally supported crypto currencies and aregional “Freigeld” crypto currency.Anexample is shown in fig. 4.1.

Impacts of differentfactors

Supporting questionswereprovidedtoassess thereason for theindividuallevel of “trust”and the “readiness-to-use”and whyitmay differamong thethreegiven scenarios. Especially the Freigeld scenariointroduced aconsiderable amountofnew information. Therefore, the majority of thesupporting questions related to theFreigeld scenario, with theaim to assess theinfluence of differentparameters.

69 Implementation

Table 4.10: Thisfigure showsthe developedquestion matrix whichwas used to assess the level of “us- age” for differentcurrency scenarios. Therefore, thevariable named “readiness-to-use”was split into four presumableincremental steps shown on theleft, labeled as “Objective”. The three currency scenarioswereseparated into three columns. The result is amatrixinwhich each elementexhibits adistinct questionnumberand the dedicated questionofthe imple- mented questionnaire. Additionally,eachelementrepresents oneobjectiveand onecurrency scenario. Agradual, two dimensional variable of theoverall“readiness-to-use”for different crypto currencieswas obtainedbytransferring the outcome of each questionintothe question matrix.

In detail, the following influencing variableswereassessed.

1. The impact of theobjectivenamed ‘long try-out”

2. The impact of governmental support of cryptocurrencies

3. The impact of thefollowing differentparametersintroduced by the Local”Freigeld” crypto currency:

-The impact of supporting the localeconomy -The impact of aconstantmoney value and thesatisfaction withthe currentmonetary and economicsystem -The impact of the generalideaofalocal”Freigeld” cryptocurrency -The impact of the ”Freigeld”tribute

70 Implementation

4.4.4 Classification

By assemblingthe data of theconductedquestionnaire, the separation of “acceptance” into “readiness-to-use” and “trust”previously done wasused to classifythe differentscenarios and theirbelonging objectives. The initial classification idea can be described as follows:

-Maximum/Fullacceptance: The participant “fully agrees” (+3 on theLikert scale)toreceive asmallpart of hissalary in the presentedcurrency andalso exhibits the highest leveloftrust(+3 on the Likertscale) for thesaidcurrency. -Minimum/Noacceptance: The participant “fullydisagrees” (-3 on theLikertscale) to do ashort try out of the presentedcurrency andalsoexhibits the lowest level of trust(-3 on theLikert scale)for the saidcurrency.

Following this concept, the rangeofthe acceptance can be determined foreachcrypto currency. In general,three values were assessedfor everysurvey question: Themean value,the median and the mode.2 Eachquestionwas classified from insignificant(neutral) to very significant,by the evaluation of themean value, the medianand the mode in referencetothe obtaineddata. (Themodewas only used to indicate strong insignificance). Figure 4.2 showsanexampleofdata obtained by asurvey question. As indicated before, each questionrepresents amatrixelement. The obtained dataisthen transferred to table 4.10 which usedtoevaluate and illustrate thefindings. (seefor example fig. 5.15)Color coding wasused to highlightthe classification of matrixelements. (see table 4.11)

Figure4.2: Thisfigure showsanexample of dataobtained by asurvey question. The numbers1-7 refer to theoptions provided by theLikert scale (see fig. 4.1). The ‘Frequency’ depicts the quantity of votesfor eachoption. Mean value: 4.37(orange line);Median value: 4; Mode: 7;

Neutral, insignificant or not significant wasdefined as mean value bellow4.5 and above 3.5, and amedian of 4. Thisresembles aposition of both valuesaround the center of theLikert scale (seefig. 4.2). Additionally,ifthe mode features avalue of 4, thedatawas categorized as strongly neutral or strongly insignificant. Consequently, if at least one value exceedsthe neutraldefinition,the data wascategorizedas significant. The exactnomination is givenintable 4.11.Orange colorsindicate anegative significance (a rejection),blue colorsindicate apositive significance (an approval).

2Theinterquartile range (Q3 -Q1) wasnot includeddue to thelimited range of the implemented Likert scale.

71 Implementation

Table 4.11: The nomination of different categories for thedatasetsobtained. The mode wasincluded only for the verificationofastronglyneutral position and theinterquartilerangewas omitted, duetothe limitedrange of theimplemented Likert scale.

The presented differentcategories were used throughout thewhole evaluation,assuring aconsis- tent classification of allfindings. In case of thenecessity of an accumulation or merge of different categories, theassigned “Value” serves as basis for basic arithmetic operations.

72 5Results

5.1 Timeline, sample size and response rate

Thesampling periodcomprised32days. It started at 31st December2020, 00:01 and ended at 31st January2021, 23:59. Atimelineofthoseindividuals successfully contacted is shownin fig. 5.1.Intotal 940 datasetswereobtainedbythe sampling process. On average, participants needed about 9minutes 45 seconds to completethe survey.774 (82 %) of all theindividuals participating finishedthe questionnaire. The majority(77 %) of participantsused themobile version. Aslightly higher completion ratewas measured for participants that used thedesktop version. The dropout ratewas approximately evenly distributed ratebetween terminations during and after the introduction. The maximum wasassessed for terminations during the introduction forthe mobile version, but itsextentwas classified as insignificant. Thisisillustrated in table 5.1 and fig. 5.1. The questionnaire providedtwo language options: German and English.96%chose the German and 4%the English language. An evaluation of thedropout rateshowednosignificantdifference between both languages. Thisisillustrated in table 5.2. The obtained responserate was40.8 %. Its calculation wasdone according to eq.(3.3). Hence, three variableshad to be evaluated beforehand: The number of usabledata sets,the number of selected unitsand the number of unreachable units. An illustration of this calculation and the included and excludedvariables is provided in table 5.3. The number of usable datasetsfeatures thefollowingproperties:

-The participant completed 100 %ofthe given questionnaire.

-The participant’s country of residenceisAustria.

-Noextraordinarypatterns were found during the evaluation of the data set.

As definedbythe sample frame,only participants withtheircountryofresidence in Austria were selected forevaluation. Additionally,all datasetswere tested by color coding. Eachoptionofthe implemented Likert scale (-3 to +3) wasassigned to aspecific number and color. Datasetswith alow color variabilityorrepetitivecolor patterns were selected forrevision and subsequently rejected if necessary. The online platform also provided an additionalvariable which assessed theindividualsurvey duration of participants. On this basis, the platform provided ameansto

73 Results outline particularly fast participants,indicatinganinferiordata set.Intotal,10participants were categorised as inferior. Given the fact,that 14 of the selected helpers alreadycompleted the survey during the pretest phase, therewas ahigh probabilityofadataset falselyidentifiedas inferior. It wastherefore decided to include all 10 inferior data sets albeit they showedapartially conspicuouscolor coding. The calculatednumberofdata setsfinally used forthe subsequent evaluation was739. The number of units selected or contactedwas determinedbythe sumofindividuals contacted by helpers plus those individuals contacted the researcher.35selected individuals agreed to distributethe survey and serveashelper.Four out of these 35 helpers successfully recruited additional helpers.1,345individuals were contacted by helpers and 468 by the researcher. In total1,813individuals were contacted during thesampling process. Units were defined as unreachable if the forwarded privatemessage wasneither markedasreceived, nor as read. Thiscategorization wasdone only by the researcher.Hence, people were only classified as unreachable by the researcher and notbythe helpers.This decisionwas made to lowerthe effortfor the helpers,while accepting the integration of abiasand alowerresponse rate. As outlinedinsection 4.3.2,eachselected unitwas reminded at least once to participate in the survey.

Figure 5.1: This graph shows the time line of successfullycontacted individuals. The samplingprocess startedat31st December 2020,00:01 and endedat31st January 2021, 23:59.

74 Results

Both Both Mobile Mobile Desktop Desktop versions versions version version version version [#] [%] [#] [%] [#] [%] Total 940 100 % 724 77 % 216 23 % data sets Finished 774 82 % 584 81 % 190 88 % datasets Cancelled 166 18 % 140 19 % 26 12 % datasets Total 940 100 % 724 100 % 216 100 % Cancelled during introduction 93 56 % 80 57 % 13 50 % (page 1-3) Cancelled after introduction 73 44 % 60 43 % 13 50 % (page4-11) Total 166 100 % 140 100 % 216 100 %

Table 5.1: Thistableillustrates the obtainedtotalsample and elaboratesthe specific numbersand related fractions of canceled and completed datasets. Thetableisseparated into threeparts by twin lines.The first part shows thetotalamount of datasets. Thesecondshows the numbersand proportionsbetween finished andcancelled datasets. Thethird section shows aseparation of terminationsduring and after the introductionofthe questionnaire. Forall parts, the numbers and sharesare outlined forthe mobile and thedesktopversion of thequestionnaire.

Total Total Finished Finished data sets data sets data sets data sets [#] [%] [#] [%] German 907 96 % 750 97 % English 33 4% 24 3%

Table 5.2: Thistableshows frequencies andproportionsregarding thechosen language options of the survey.

75 Results

Usabledata sets Total usable datasets 774 Countryofresidence: OTHERS 35 Country of residence:AUSTRIA 739 Individualsselected/contacted By theResearcher 468 By helpers 1,345 Total 1,813 Response rate Response rate 40.8 %

Table 5.3: Thistable shows the number of usable data sets andthe number of individuals ultimately contacted. It also shows the response rate,whichwas calculatedaccording to eq.(3.3)and the definedsample frame (target population). To be precise,the response rate wascalculated by dividing the number of usable datasets (Countryofresidence: AUSTRIA)bythe total number of contacted individuals.

76 Results

5.2 Demographics

Thissectionshows the demographics of thesample obtained, whichincludes the following char- acteristics:

-Gender

-Age

-Highesteducationcompleted

-Countryofresidence

The sampleobtainedfeatures an almost equal distribution between males (50.00 %) and females (49.95 %). Thisisalso true for theseparation of malesand females by the highesteducation completed (fig. 5.7)and by the age (fig. 5.6). The majorityspecified theirage between 16 and 30 years (55.21 %), followedbythe category of 31 to 45 (20.57 %) and 46 to 60 years (20.43 %). For 32.12 %ofthe participants,the Leaving Certificateisthe highest education completed.Intotal 50.84 %obtainedauniversitydegree, withalmost equal shares between the Bachelor’s (22.77 %) and theMaster’s degree (25.98%). 14.66 %stated thattheir highest education completed is an apprenticeship. An illustration of theoutlineddemographics is provided by different figures below.

Figure 5.2: The sample characteristics of age and gender in percent

77 Results

Figure 5.3: Bar chartdepictingthe specific frequenciesofthe ageofparticipants

Figure 5.4: Thesample characteristic of the highesteducation completed in percent

Figure 5.5: Bar chart depicting the categories of thehighest education completed and the specific fre- quency of the sample

78 Results

Figure 5.6: Population pyramid by age and gender

Figure 5.7: Population pyramid by highest education completed and gender

79 Results

5.3 Findings concerning the use of electronic payments

The evaluation of the use of electronic payments showedthat61.98 %ofall participants feature apositive tendency towards cashless payments. 16.78 %remained neutral,while only 21.25 % showedanegativetendency to use cashless payments (see fig. 5.8). 71.85 %ofall participants us electronicpaymentsseveral times perweekand only 5.82 %barelyornever use electronic payments. (see fig. 5.9) 92.21 %had heard about cryptocurrencies before this questionnaire,but onlyaminority of 24.23 %stated to understand or somewhat understandthem.18.13 %chose to remain neu- tral withoutatendencytowards apositive or negativeunderstanding of crypto currencies(see fig. 5.10). 12.31 %ofall participantshave obtained or purchased crypto currencies. Thisspecificgroup wasadditionally asked about theirusage of cryptocurrencies.Data showedthatthe majority (73.91 %) uses theircrypto currencies as assetsand plans to sellthem at afavorablepoint in time,whereat only 20.65 %haveused theircrypto currencies as means of payment.Inthe questionnaire,itwas possible to choose either one, or both of the stated options (see fig. 5.12).

Figure 5.8: Frequencyand percentagesofthe general“use of cashlesspayments in everydaylife” .

Figure 5.9: Frequencyand percentagesofthe useof“electronicpayments methods”.

80 Results

Figure5.10: Proportionsofparticipants whohad heard aboutFreigeld before the questionnaire.

Figure 5.11: Proportions of participants whohad heard about crypto currenciesbeforethe questionnaire.

Figure 5.12: Cryptocurrenciesand the stated level of understanding depictedinahistogram andapie chart.

81 Results

Figure 5.13: The proportion of participantswho already had obtained or purchased crypto currencies is shown in the top pie chart. Theseparticipants were additionally askedabout their usageof crypto currencies as anassets,ameans of payments or both. The middle pie chart shows the proportions forthe use as asset,the bottom pie chart shows theproportionfor the use as ameans of payment.

82 Results

5.4 The acceptance of crypto currencies

5.4.1 The readiness to use crypto currencies

The variable “readiness-to-use” wasevaluated by the question matrixpreviously presented (see Core concept and Classification). The resultsare shown in table 5.4.Asillustrated, the level of thereadiness to use thelocal Freigeld crypto currency exceeds thatofcrypto currenciesand governmentally supported cryptocurrencies for every objective.Comparing cryptocurrenciesand governmentally supportedcrypto currencies,all four objectives featurethe same level of signifi- cance.Byexaminingthe total significancecomparison matrix it is also visible that themajority of elements show alow to very lowsignificantvariation aroundthe neutral level.Receiving crypto currencies and governmentallysupported crypto currenciesassalaryare the onlytwo objectives who shows asignificantrejection.

Table 5.4: Thismatrixshows the differentlevels of “readiness-to-use” forthe discussed crypto currencies. The matrix wasderived by combining the findings of the mean comparisonmatrix (table 5.5), the mediancomparison matrix(table 5.6)and the mode comparisonmatrix(table 5.7)ac- cording to table 4.11.

Table 5.5: Thismatrix shows themean comparison matrix. It wasderivedbytransferring the survey data into the questionmatrix(see table 4.10). The matrix depicts the differentmean values of the related questions. The results were color codedaccordingtothe classificationgiven in table 4.11.

83 Results

Table 5.6: Thismatrixshows the median comparisonmatrix. It wasderived by transferring the survey data into the questionmatrix(seetable 4.10). The matrix depicts the differentmedian values of the related questions. The results were color coded according to the classificationgiven in table 4.11.

Table 5.7: Thismatrixshows the mode comparisonmatrix.Itwas derivedbytransferring the survey data into the questionmatrix(seetable 4.10). The matrix depicts the differentmodes of therelatedquestions. The results were color coded accordingtothe classification givenin table 4.11.

84 Results

Figure5.14: Thismatrixdepictsthe surveyresultsconcerningthe “readiness-to-use”crypto currencies forthreedifferentscenarios. Eachdiagram representsamatrixelement. Thus, thenumber aboveeachdiagramreferstoaspecific questioninthe survey.The related questions are givenintable 4.10.The numbers 1-7refertothe optionsprovidedbythe Likert scale (see fig. 4.1). The ‘Frequency’ depicts the quantity of votesfor eachoption. The arrowonthe left (labeled “readiness-to-use”)referstothe incrementalobjectives, starting with “short try-out” andending “as salary” (see section 4.4.3). Thefindingsderivedfromthis matrix resulted in the mean, the medianand the mode comparisonmatrix(see table 5.5,table 5.6, table 5.7.The ultimate findingsare depictedintable 5.4.

85 Results

5.4.2 The trust in crypto currencies

Table 5.8: Thistableshows theresultsfor thedifferentlevels of trust forall discussed crypto currencies. Three out offour questions originally included the phrase “don’t trust”(or similar). The scale wasconverted to increase thereadabilityand phrase waschanged to “trust”.Thisis indicatedbyan‘*’ (asterisk).The results were color codedaccordingtothe classification given in table 4.11.

Table 5.8 illustrates the findings concerning thedifferentlevelsof“trust”.The data wasclassified according to table 4.11.Inadditiontothe threecurrenciespreviouslystated, the level of trust was also evaluated for participants whohad already purchased or obtained crypto currencies. This scenariofeaturesthe phrase“My crypto currencies”. To be precise,the scenario “Cryptocur- rencies”includes thedata of allparticipants,while the scenario “My cryptocurrencies” includes only participants who already purchased or obtained crypto currencies. Participantswho had already purchased or obtained crypto currenciesfeature the highest levelof trustintheir cryptocurrencies.Nonetheless, even their level of trustshows only alow positive significance. The localFreigeld crypto currency is ranked second, followedbythe governmentally supportedcryptocurrency,yet for both currencies, the variable trust shows no significantvalue, neither inpositive nor in negative direction. Thedata obtained showsthe lowest valueoftrust for crypto currencies and indicates alow significancetowardsnotrust.

5.4.3 The acceptance of crypto currencies

As outlined in section 4.4.3,the totalacceptanceofcrypto currencies is calculated as thesum of the variables “readiness-to-use” and “trust”.The former calculations led to twomatrixvariables with adifferentnumberofcolumns. Adirect summation of both variables is therefore not possible. Hence,tocalculatethe acceptance fromthe previously outlined findings, the meanvalues were converted by the following steps:

-The Likert scale(featuring ascale from one to seven) wasshifted by one to set theorigin to zero. -The scale wasnormed and convertedtopercent. -For thesummation,the total levelof“readiness-to-use” (the four combined incremental objectives)was assigned thesame weight as thelevel of “trust”. Hence, to obey this defined weighting,the value of each of thefourpresumable incremental objectives wasdividedby four. -The weighted variablesof“readiness-to-use” and “trust”were then separatelyaccumulated for eachcurrency.The result is shown in fig. 5.15,depicting acomparableillustration of the acceptanceofall crypto currencies presented.

86 Results

The converted scale cannow be interpreted as follows: 0%equals “Fully disagree”, 16.67 % equals “disagree”, 33.33 %equals “Somewhat disagree”, 50 %equals “Neutral”, 66.67 %equals “Somewhat agree”, 83.33 %equals “Agree”and 100 %equals “Fully agree”.

Figure 5.15: Overallacceptance of crypto currencies

Thesame calculation wasdone for themedian values. Compared to the mean values, the result showedanaverage deviation of 3.7 %ofthe totalacceptance of eachcurrency discussed. This difference wasconsideredasinsignificantand therefore the medianvaluewas neglected forfurther evaluations. Thisisillustrated in fig. 5.15 and alsointable 5.9. Participants showedahigher readiness-to-use governmentally supported crypto currencies (22.8 %) comparedtoregular crypto currencies (20.4 %).However, the participants showedahigher leveloftrust in regular crypto currencies (21.8 %) than in governmentally supported crypto currencies(19.3 %). The total overalllevel of acceptance for agovernmentally supported crypto currency(42.1 %)and aregular crypto currency (41.4 %) showsonly an insignificant difference. The acceptance for thelocal Freigeld crypto currency(55.1 %) exceedsevery other discussed currency.The readiness-to-use the currency (26.7 %) is higher then for every othercurrency presented, thelevel of trust is rankedsecond (28.3%). Only participants who had already purchased or obtainedcrypto currenciesfeature ahigherlevel of trust (28.3%)intheir obtained currencies.

87 Results

Acceptanceof Crypto Governmentallysupported My crypto Local Freigeld cryptocurrencies currencies crypto currencies currencies crypto currency short 7.2% 7.0 % 7.2 % 8.2 % try-out long 5.5% 5.5 % 5.5 % 6.5 % try-out for everyday 5.5% 4.8 % 5.5 % 6.5 % payments as salary 4.0% 3.1 % 4.0 % 5.5 % “trust” 19.3 % 21.8 % 30.9 % 28.3 % Total 41.4 % 42.1 % 53.0 % 55.1 %

Table 5.9: Thistable showsthe differentpercentages of acceptance for different objectivesand crypto currencies.

5.5 Details concerning the acceptance of crypto currencies

5.5.1 The impact of the objective named “long try-out”

The implemented concept featured theobjectivenamed long try-out as partthe of the readiness- to-use variable. It wasnot possibletoassess beforehand if theparticipants generally refusethis option.Two possible reasonsfor suchrefusal are alow selectivityinregardtothe objectivecalled short try-out or the general refusal of yearlong try-outs. Question15evaluated theattitude of participants concerning the stated objective. According the mean value of 4.37and the median of 4itwas concludedthatnosignificantgeneral rejection of this objective wasfound. The differentresultsofthe survey(shown in table 5.5)also verify this claim.

Figure 5.16: Thisfigure showsthe findingsofthe evaluation of theobjective named “long try-out”. According to amean value of 4.37 (orange line), amedian of 4and amodeof7.

5.5.2 The impact of governmental support on crypto currencies

The survey includedfour additional questionstoverify findings concerning governmentalsupport. The obtaineddataisillustrated in table 5.10 and shows no significant impact of government support on the level of trustincrypto currencies (Question 24) and on thepersonal preference

88 Results for crypto currencies (Question25). Forthe localFreigeld crypto currency a very low positive impact of governmental supportonthe leveloftrust wasfound(Question38) and a very low positive significance forthe requestofgovernmentsupport (Question37). The results indicate thatgovernmental support hasaverylow to no impact on the personal preference of theparticipantsand on thelevel of trust forall cryptocurrencies. Thepresent results confirm the previous findings(see fig. 5.15 and related elaborations).

Table 5.10: This table shows the supporting questions to validate the impactofgovernmentalsupport on thelocal Freigeldcrypto currency andonplaincrypto currencies. Theresult wascolor coded accordingtothe classification given in table 4.11.

5.5.3The impact of differentparametersintroduced by thelocal Freigeld crypto currency

The parametersand conclusions, orderedfromhighimpact to no significantimpact,are as follows:

1. (Table 5.11) The impact of involving the local economy The datashowedthat for a very significant amountofparticipants, it is importantto support the localeconomy(Question45). A significant partofparticipants thinksthat the circle coin would help support the local economy(Question40). Additionally,asignificant partofparticipants indicatedanin- creasedacceptance of the circle coin,caused by the promised support of thelocal economy (Question 40).

2. (Table 5.12) The impact of aconstantmoney valueand the satisfactionwith the currentmonetaryand economic system Theconstantvalue of money is importantfor a very significant partofall participants (Question 27).Wheninquired ifthe circle coinwill succeedtooffer aconstantmoney

89 Results

value,participants showed (a strongly)neutral position. (Question 44) Thus, theconstant money value promised by the localFreigeldcrypto currency showed no significantimpact on the level of acceptance.The currentmonetaryand economic system wasclassified as neutral (neither significantlygood,nor significantly bad) (Question 28). Question 41 indicates a very lowsignificance forrejecting additional currencies.

3. (Table 5.13) The impact of the general idea of alocal ”Freigeld”crypto currency Question32showedthat participants hadatendency to “somewhat agree”(Likertscale) to theidea of alocal Freigeld crypto currency. Nonetheless, according to table 4.11,the impact of this parameterhad averylow significance for the overallacceptanceofthe local Freigeldcrypto currency.

4. (Table 5.14) The impact of the“Freigeld tribute” The following variables, which mayhaveinfluenced the position of theparticipants con- cerning theFreigeld tribute(the demurrage),weredetermined beforehand:

-(Question29) The tributeinregardtothe promise of more economicstability. → insignificant -(Question30) Ageneral refusal of thetribute itself. → (strongly) insignificant -(Question31) No preference regardingthe tribute. → very lowsignificance (negative) -(Question42) Moreacceptanceofthe tribute duetomore information provided by the scenario. → (strongly) insignificant -(Question43) The local Freigeld crypto currencyisbetterwithout the tribute. → (strongly) insignificant

90 Results

Table 5.11: This table illustratesthe impactofinvolving the local economy into thedesign of acrypto currency.Asvisible,the parameter “supportofthe localeconomy” had an significantimpact on theacceptance of the localFreigeld crypto currency. Theresultwas color coded according to theclassification givenintable 4.11.

Table 5.12: This table depicts the findings concerning the importance of aconstantmoney value and thesatisfactionwith the currentmonetary and economic system.Italso indicatesthat both factors had averylow to no significant impact on the level of acceptance for the local Freigeld crypto currency.The result wascolor coded accordingtothe classificationgiven in table 4.11.

91 Results

Table 5.13: Thistable shows that the general idea of alocal Freigeldcrypto currencyhad averylow positive impactonits acceptance.The result wascolor coded according to theclassification given in table 4.11.

Table 5.14: Thistable showsdifferentvariableswhichmay have influenced the position of participants concerning thetribute.Overall, it is observable thatthe introduction of the tributehad no significantimpact. The result wascolor coded accordingtothe classificationgiven in table 4.11.

92 6Discussion

6.1 Thereadiness to usecrypto currencies

Based on thefindings of table 5.4 thefollowingconclusions were drawn aboutthe “readiness-to- use” crypto currencies:

1. The “readiness-to-use”isequalfor cryptocurrencies with and without governmental sup- port.

2. The data showed avery lowpositive significancethatparticipants woulduse agovernmen- tallysupported crypto currency or a regular crypto currency fora“shorttry-out”, but showednosignificant indicationtouse it beyond this objective.

3. The readiness to do a“short try-out” waspositive withalow to very lowsignificance for all crypto currencies presented.

4. The participants showed avery lowsignificance to usethe localFreigeld crypto currency up to the level of “everydaypayments”, but showednosignificantindicationtoreceive it “as asalary”. Conclusion:The second hypothesis is false. Second hypothesis: Individuals do not acceptusing crypto currenciesfor everyday payments. Statement: As indicated by obtained survey data, theparticipants showedavery lowbut positive significance to use thelocal Freigeld cryptocurrency for everydaypayments. Hence, theparticipants do not generallyreject an increased inclusion of crypto currencies in society. Given the still emerging crypto currencies, the extenttouse one of them up to thelevel of everydaypayments is considerable. Furthermore, theparticipants featured an age,which wasbelow theaverage andaneducation, which wasabove the average of theAustrian society. The datacan thereforebeinterpreted as preliminaryoutlook on the thoughtsof the next generation.

6.2 The trustincrypto currencies

The leveloftrust wasassessed for three kinds of cryptocurrencies: regularcrypto currencies, governmentallysupported crypto currencies and alocal Freigeld crypto currency.According to

93 Discussion table 5.8,the level of trust foragovernmentally supported crypto currency and for the local Freigeld crypto currency showed no significance,neither towards trust, nortowardsnotrust.

Theparticipants showedalowsignificance towards no trust for regularcrypto currencies,but the opposite wasfoundfor individualswho already obtainedcrypto currencies.Suchindividuals showedalowsignificance towards trustfor regular crypto currencies.

Conclusion: The influenceofgovernmental support on thelevel of trust in crypto currencies is insignificant.

Statement: The leveloftrust differentiatessignificantlybetween theparticipants, but not be- tween thecrypto currencies. Given the difference of trust in regular crypto currencies between participants who alreadyobtainedand didn’t obtaincrypto currenciesbefore,itcan be said that the leveloftrust is either influenced by detailed technicalattributesorbythe personal environ- ment of theindividual. Ipersonally hold the impression thatbothplayasignificantrole and that aperfect technical solution mightnot be sufficienttoachieveahighlevel of trust.Nonetheless, it has to be taken into account that only 12.33 %(fig. 5.13)ofall participantshaveeverobtained crypto currencies. Due to this limited amountofdata further researchisnecessary to verify this claim.

6.3 The acceptance of crypto currencies

Section 5.4.3 featured adetailedelaboration of theoverall levelofacceptance for governmentally supported crypto currencies(42.1 %) and regularcrypto currencies (41.4 %).According to this elaboration, whichwas mainly based on fig. 5.15 and table 5.9,aninsignificantdifference between both currencieswas determined.

Conclusion:The firsthypothesis is false.

First hypothesis: Governmental support increases the acceptance of acrypto currency.

The acceptance of thelocal Freigeldcrypto currencyincludes the highestlevel of “readiness-to- use” (26.7 %) and thesecondhighestlevel of trust (28.3 %). Compared to everyother crypto currency, this resultedinthe highestlevel of acceptance (55.1%)for thelocal Freigeld crypto currency.

Conclusion: The third hypothesis is true.

Third Hypothesis: The local Freigeld crypto currencyexceeds the acceptance of othercrypto currencies.

Statement: Compared to the other currencies, thelocal Freigeld crypto currency showedthe highest levelofacceptance. To allow abetter evaluation, the acceptance wasalso assessedfor the subgroup of participantswho already obtainedcrypto currencies. This currencyoptionwas named “mycrypto currency” and this particular subgroup featured aconsiderablehigher level of acceptancefor cryptocurrencies (with or without governmental support). Nonetheless, thelocal Freigeld crypto currencystill slightly exceeded the acceptanceof“my cryptocurrency”.This underlines the considerablehighlevel of acceptance of thelocal option,sincethe participants who already obtainedcrypto currenciescan be categorized as aparticularly positive subgroup.

94 Discussion

6.4 Impacts on thelevel of acceptance of crypto currencies

1. The impact of theobjectivenamed ”longtry-out” The concern about the introduced response bias by the objectivenamed “longtry-out” was not confirmed by the obtaineddata(see fig. 5.16). The overallresultofthe assessmentofthis particularobjectivecan be interpreted as follows: Ayear-longtry-outischallengingbut, in the end,itdependsonthe object to be tested. Conclusion: The objective“long try-out”did notresultinaresponse bias. 2. The impact of agovernmentalsupport on crypto currencies The surveyincludedadditionalquestions to verify the findings of governmental support in more detail. Theresults were outlinedinsection 5.5.2. The verificationofgovernmental support showedneitherapositive nor anegativeimpact on thelevel of trust or preference of regular crypto currencies. In case of thelocal Freigeld crypto currencythe impact changed to avery lowpositive attitude towardsgovernmental support. Conclusion: Theverificationconfirmed the very lowtonosignificantimpact of govern- mental support on the level of trustand the personalpreference of theparticipants. 3. The followingconclusionsweredrawn aboutthe impact of differentattributes introduced by thelocal Freigeld crypto currency: -(Table 5.11)The impact of supporting the localeconomy Conclusion: The parameter “support of the localeconomy” had asignificant impact on theincreasedacceptance of thelocal Freigeld crypto currency. -(Table 5.12)The impact of aconstantmoneyvalue and thesatisfaction with the currentmonetary and economicsystem Conclusion: The promisedconstantmoney value and thesatisfaction with thecurrent monetary/economic system showedvery lowtonoimpact on the levelofacceptance of thelocal Freigeld crypto currency. -(Table 5.13)The impact of thegeneral ideaofalocalFreigeld crypto currency Conclusion: Themereidea of thelocal Freigeld cryptocurrency hadavery low positive significance for itsoverall acceptance. -(Table 5.14)The impact of theFreigeld tribute (demurrage fee) Conclusion: The introductionofthe Freigeld tribute showed no significantimpact on itsresultingoverall acceptance.

Statement: Theassessmentofthe attributes introducedbythe localFreigeld crypto currency showed that thepromised support of thelocal economywas theonly significantparameter. Henceitwas therefore concluded that this attribute ledtothe high level of acceptance.Inother words, compared to allthe otherpresented crypto currencies, the local Freigeld crypto currency featured thehighestlevel of acceptance, because it supported the localeconomy. The two other attributes of thelocal currency,namelyaconstantmoney value and atribute (demurrage fee), hadaverylow to no impact on theacceptance. The same applies for apossible dissatisfaction withthe current monetary/economic system or ageneral rejection of thelocal Freigeld crypto currency idea. Evidently,ifabroad penetration of thesocietybycrypto currencies is aspired, the supportofthe local economyidentifieditself as aconsiderablelever.

95 7Conclusion

7.1 Introduction

The effectivedemand is partofthe Freigeld theorydeveloped around 1900 by the entrepreneur Silvio Gesell.Ifexecutedcorrectly,itpostulates aconstantvalue of money.This attributewas contemplated to suit thecurrentperiodoftime in twoways. Firstly,the stressed patienceofthe societyinregard to the recent economiccrises(boomand bustcycles) and secondly,the potential disruption of the bankingsector and themonetarysystembythe developmentofDLTs. Under this premise, thebalancing effectofacomplementarystablecurrency combined with the features of cryptocurrenciesmeetsthe spirit of thetime and wastherefore chosen as researchtopic. Based on Gesell’s framework,the applicability of the effective demand on cryptocurrencieswas evaluated and ultimately abasiclocal Freigeld cryptocurrency concept (named circle coin)was developed. Besides the assessmentofthis newcurrency, the survey alsoinquired the levelof acceptance of regular cryptocurrenciesand governmentallysupported crypto currencies. This enabled to compare thefindingsofdifferentcurrency set-upsand allowed to broadenevaluation.

7.2 ResearchObjectives

The thesishad two mainobjectives. Firstly,toassess theapplicabilityofthe theory of effective demand on crypto currencies.Secondly, to conductapreliminary investigation of thecrypto currency acceptance by theAustrian society-ingeneral and in reference to thetheory of effective demand. Duetothe considerable criticismofSilvio Gesell’s theoriesitwas decided to presenthis concept of effective demand to ahighextentbyhis ownoriginaldescription and framework. The aim was to allow aclearclassification of the original work, without anyinfluence of different conclusions drawnoverthe last 100years. Thisalsotookawaythe complexityofthe presenteconomic system which evolved over the20th century and narrowedthe framework down to the former time. Thepreliminary investigation of theacceptance of differentcrypto currencieswas originally planned to be executed by aprobability sample. Thiswould have allowedtoderivethe find- ingstoamajor part of theAustrian society. Unfortunately, as outlined in section 4.2,the survey design hadtobechanged due to the SARS-CoV-2 containment measures. The newdesign fea- turedaself-administered, webbrowserbased, onlinesurvey whichwas distributed by snowball

96 Conclusion

sampling. Hence,sincethe new design prohibits anygeneralization, the survey objectivehad to be changed from gathering information of theAustriansocietytoonly describing the data of Austrian people who participated at thesurvey. The followingresearch objectives, research questions and hypotheses were aimed to be assessedbythe social survey:

Theresearch objectivewas apreliminary investigation into ...

RO1: ...whether the acceptance of acrypto currency would change for agovernmental support.

RO2: ...whether individuals would accept acrypto currencyfor everyday payments.

RO3: ...whether alocal Freigeldcrypto currencychanges the acceptance of crypto currencies.

Three researchquestions were derived from the statedresearchobjectives.

RQ1: To what extentwould agovernmental support change thelevel of acceptanceofcrypto currencies?

RQ2: To what extentwouldindividuals accept acrypto currency solutionfor everyday payments?

RQ3: To what extentdoesthe availability of alocal Freigeldcrypto currency change theaccep- tanceofcrypto currencies?

As consequenceofthe stated researchquestionsthreehypotheseswere determined to be tested.

H1: Governmental support increases theacceptanceofacrypto currency.

H2: Individualsdonot accept using crypto currenciesfor everydaypayments.

H3: The local Freigeld cryptocurrency exceeds the acceptance of othercrypto currencies.

7.3 Limitations

7.3.1The effectivedemand

The evaluation of theapplicability of effective demand on cryptocurrencies is based on the former framework outlined by Silvio Gesell. In this regard it has to be acknowledged thatonly the clearregulation of an economic subsystem, as possiblefor crypto currencies, justifies the implementationofthe effective demand (basedonthe framework providedbySilvio Gesell) on crypto currencies.Conclusions drawnabout the applicabilityofthe effectivedemand outside of this limitedframework and unrestricteddeductions to the currenteconomicsystemare not supported by this thesis.

97 Conclusion

7.3.2 The executed sample

The sample features thefollowing attributes:

-Sample design:Aself-administered,web browser based, onlinesurveydistributedbysnow- ball sampling

-Sample frame: Individuals withinternetaccess and withtheir country of residence in Aus- tria

-Sample size:739 (100%complete datasetsand revised)

-Responserate: 40.8 %

-Sampling period: 32 days (31st December2020, 00:01 -31st January2021, 23:59)

-Average survey duration: 9min 45 sec

The obtained dataexhibits avery highrisk of severe errors andbiases,caused by the natureof the executed nonprobabilitysample. Therefore, ageneralization fromthe sample statistics to a broader population is notsupported by the designand theresultsshould be seen as descriptive statistic. Concluding, giventhe nonprobabilitysample design and thesevere restrictionofthe Austrian population by the chosensampleframe, the presented results were categorized as preliminary investigation. Limitations concerning theoriginalsample design are providedinsection 4.1.4.

7.4Broad implicationsand future work

Thefurther development of themonetary systemisadelicate topic, since ahighpercentage of theentire human societyrelies on itsoperation. Hence, adaptations should be madewith great caution and should ideally be approvedbythe scientific society. Nonetheless, theemerging DLT’s continue to challenge thecurrentmonetarysystem, whichrenders it crucial for thescholarly societytoproactivelyparticipate in this process. Theassessment of thetheoreticalapplicability of theeffective demand on crypto currencies showed thatitisworth revising previously rejectedconcepts.The fastdevelopmentofdifferenteconomic theories over thelast 100 years induced by the globalization of theeconomic system, shows aconsiderable potentialfor other reevaluations. ComingbacktoSilvio Gesell, the successful reviewofthe effective demand supports afurther assessmentofhis theories, theircomparison to the intermediatetheoriesofthe 20th century as well as to the economictheoriesofour present time.Ipersonally holdthe impression that ageneral reevaluation of former thoughts would broaden the horizon of scholarlyeconomics. Thus, if Ulanowicz R. et al. (2009) are correct with theirinterpretation,the therebyincreaseddiversityineconomicsciencewould be benefitingnot onlyfor thescholarlysociety, but also for themonetary system and consequentlyfor thehuman society. [UGLG09] The obtained data of the developedlocal Freigeldcrypto currency showedpotential foranactual application within thesociety.Nonetheless, aconsiderableamount of research and development

98 Conclusion is still required. In thisregard, the development of an accuratemathematical theory to balance theemission and withdrawalofcoinswithinthe acrypto currency systemisessential. This alsoincludes the determination of subscription modelsand above all, thetheoreticalbackground for reasonableexchangerates. Consideringthe boomand bust cycles of our economy, thevast complexityofthe generaleconomysystemand the volatile market of cryptocurrencies,thiscan be aconsiderablechallenge. Albeitthe assessed insignificance of governmentsupportonthe acceptance of crypto currencies (seechapter 6), it is assumedthatalocal currencywould benefitfromthe support of thelocal authorities. Nevertheless, no scholarly evidence can be provided to supportthis claim. Hence, it is suggested to add this attributetosuccessiveinvestigations.Additionally,itwas not assessed if anyincentives would increase the acceptance of alocal cryptocurrency.Albeit thechallenge of balancingsuchmeasures,itissuggested to alsoinclude this attribute for successivesurveyto verify the potentialimpact. Overall, further investigations arenecessarytosupport thestated claims. In reference to the present sample, it should subsequentlybeaimedfor arepresentativesample to allowagener- alization to theAustrian society. Investigation in differentsocietiesorcountries mightalso be useful. If an actual application of alocal Freigeld crypto currency is aspired, the suggested broader investigations would assist thedeterminationofanoptimal starting point.

7.5 Summary andclosing

Thisthesis didn’t find anyconstraints within the framework of Silvio Gesell whichrestrict the application of the effective demand on cryptocurrencies. This resulted in the developmentofa basic complementary local Freigeld crypto currency,named circle coin. Forthe second main objectiveofthis thesis, the acceptance of regular crypto currencies, govern- mentallysupported crypto currencies and circle coin wasassessed by asocial survey. Theevaluationofthe obtained survey data ledtothe following results:

H1: Governmental supportincreases theacceptance of acrypto currency. H1 = FALSE;

RO1: Governmental support hadaninsignificantimpact on the level of trust in cryptocur- rencies andnoimpact on the participants to change their“readiness-to-use”it. RQ1: In reference to the Core concept,whichwas usedtoderivethe acceptance fromthe level of trust and thelevel of “readiness-to-use”, it wasconcluded that theoverall acceptanceofacryptocurrency is notinfluenced by governmental support.

H2: Individualsdonot accept using crypto currenciesfor everydaypayments. H2 = FALSE;

RO2: The obtained datashowedavery lowpositive significance to use thelocal Freigeld cryptocurrencyuptothe level of everydaypayments.For crypto currencieswith or without governmental support,the participantswere onlyreadytodoashort try-out. Four main attributes were determinedbeforehand as potentialorigin of an alteredlevel of acceptance of thelocal Freigeldcryptocurrency.Accordingtothe evaluation,the

99 Conclusion

support of thelocal economywas theonly attributewhichhad asignificantimpact on the increased acceptance of the circle coin.The other attributes namely, aconstant money value,atribute (demurrage fee) and ageneral rejection of thelocal Freigeld crypto currency ideafeatured averylow to no significance. RQ2: Hence,the participantswereready to usethe local Freigeld crypto currency for every- daypayments underthe premise of supporting the local economy.

H3: The local Freigeld crypto currency exceedsthe acceptance of other cryptocurrencies. H3 = TRUE;

RO3: As indicated above, theacceptance of thelocal Freigeld crypto currency wasconsid- erablyhigher thanthe acceptance of regular cryptocurrenciesand governmentally supported crypto currencies.Additionally,the acceptanceofcryptocurrencies was separatelyassessed for participants who alreadyobtainedorpurchasedcrypto curren- cies. This currency option wasnamed“my crypto currency”.The level of acceptance of this particular subgroupwas considerablyhigherthanthe general levelofaccep- tance of cryptocurrencies(withorwithout governmental support). Nonetheless, the derived acceptance of circle coin excelledthe acceptance of my crypto currency with avery lowtonopositive significance. In this regard,ithas to be acknowledged that thelevel of acceptance of “mycrypto currency” presumablyrepresents aparticular positivesubgroup, but still didn’t manage to excel thetotal level of acceptance of circle coin. RQ3: It wastherefore concluded thatthe mere availabilityofthe localFreigeld crypto cur- rency had asignificantpositive impact on thecrypto currencyacceptance of the par- ticipants.

The research showed that thesupportofthe localeconomyisasignificantlever for local currencies. In reference to the work of Ulanowicz R. et al. (2009),this is avital information if theaspired goal is to stabilize ourmonetary system by increasingits diversity. [UGLG09]The outlined applicability of theeffective demand on crypto currencies seems to enable acomplementary currency whichfeatures:

-Aconstantmoney value

-Atransparentregulation

-Adecentralisedapproach, whichrejects theinfluence of single entities

-The support of the local economy

-The support of charity projects

-Acost efficientadministration

-And collectively,complementarycurrency as balance weighttoour presentmonetary system

Nonetheless, as indicated in section 7.4 moreresearchisnecessary to facilitate an actualimple- mentation of circle coin.Especially due to its digitalnature an emphasishas alsotobelaid on lowering the barriers of accessibility.However,the participantsofthe survey featured an

100 Conclusion age, which wasbelow the average and an education, which wasabove average of the Austrian society. Hence,the data cantherefore be interpreted as preliminary outlook on the thoughtsof the next generation. Ipersonallysee thisasapositive indicatorfor thepotential of circle coin and of cryptocurrenciesingeneral. Ihopethis thesis made valuable contributiontothe further developmentofour monetary system.

101 AAppendix

A.1 Questionnaire

Figure A.1: Page 1: Language selection

102 Appendix

Figure A.2: Page 2: Contact details about the author and the supervisor. Information about the liable institution.

Figure A.3: Page 3: Detailsand declarations concerning the GDPR.

103 Appendix

Figure A.4: Page 4: Introduction

104 Appendix

Figure A.5: Page 5: Participant demographics

105 Appendix

Figure A.6: Page 6: Part1 -Conventionalmoney and crypto currencies

106 Appendix

Figure A.7: Page 7: Part1 -Anoptionalpage, for participants whoalready purchasedorobtained crypto currencies

107 Appendix

Figure A.8: Page 8: Part1 -Continued theassessmentofthe acceptance for agovernmentally supported crypto currency

108 Appendix

Figure A.9: Page 8(2): Part1-Continued theassessment of theacceptance for agovernmentally sup- ported crypto currency

109 Appendix

FigureA.10: Page 9: Part 2-The”Freigeld” idea

110 Appendix

Figure A.11: Page 9(2):Part 2-The ”Freigeld” idea

Figure A.12: Page 10: Ashortinformationwas givenabout the remainingdurationofthe questionnaire

111 Appendix

Figure A.13: Page 11: Part 3-The circle coin -alocal ”Freigeld” crypto currency

112 Appendix

Figure A.14: Page 11 (2): Part 3-The circle coin-alocal”Freigeld” crypto currency

113 Appendix

Figure A.15: Page 12: Specific questions withthe aimtoassess theinfluence of differentparameters, whichmay have had asignificantimpact at the overall results

114 Appendix

Figure A.16: Page 13: Lastpage -Acknowledgmentsand contact details

115 Literature

[AF01] Atkinson,R.;Flint,J.: Accessing hidden and hard-to-reach populations: Snowball research strategies.2001. –1–4 S [Atk98] Atkinson,B.: AppliedEconomics.CRC Press, 1998. –368–385 S [Aus19] Austrian FederalMinistryfor Digital and Economic Affairs: KMUim Fokus 2019 -Bericht ¨u berdie Situation und Entwicklung kleiner und mittlerer Un- ternehmen der ¨o sterreichischen Wirtschaft.2019. –137 S [BAI+18] Bambara,J.;Allen,R.;Iyer,K.;Madsen,R.;Lederer,S.;Wuehler,M.: Blockchain:Apracticalguidetodeveloping business, law, and technology solutions. McGrawHill Professional, 2018. –302 S [BB12] Baltar,F.;Brunet,I.: Socialresearch 2.0: virtual snowball samplingmethodusing Facebook.Emerald Group Publishing Limited, 2012. –19S [BBM01] Bound,J.;Brown,C.;Mathiowetz,N.: Measurementerror in survey data / Handbookofeconometrics.Elsevier, 2001. –3705–3843 S [Ben96] Benjes,H.: Werhat Angst vor Silvio Gesell.1996. –360 S [BHL96] Bruwer,J.;Haydam,N.;Lin,B.: Reducingbiasinshoppingmall-intercept surveys:The time-basedsystematicsampling method.1996. –9–16 S [Bla98] Blanc,J.: Free Money forSocial Progress: Theoryand PracticeofGesell’sAccel- eratedMoney.American Journal of Economics and Sociology,Inc., Wiley, 1998. – 469–483 S [Bry12] Bryman,A.: Socialresearch methods.Oxford universitypress,2012. –766 S [CK96] Caprio,G.;Klingebiel,D.: Bankinsolvencies cross-country experience.World Bank Publications, 1996. –52S [eco]Savingthe system.In: TheEconomist .–published:2008 October 11 [EM05] Evans,J.R.;Mathur,A.: The value of online surveys.EmeraldGroup Publishing Limited, 2005. –195–219 S [ES10] Everitt,B.;Skrondal,A.: TheCambridge dictionaryofstatistics.Cambridge UniversityPressCambridge,2010. –468 S [Fis20] Fisher,H.G.: ThePurchasing Power of Money: Its Determination and Relationto Credit,Interest and Crises.The Macmillan Company, 1920 [Fli15] Flick,U.: Introducingresearch methodology: Abeginner’s guidetodoing aresearch project.Sage, 2015. –295 S [FM20] Fill,H.-G. ; Meier,A.: Blockchain: Grundlagen, Anwendungsszenarien und Nutzungspotenziale.Springer Verlag, 2020. –379 S [Fow13] Fowler,F.J.: Surveyresearchmethods.Sage publications, 2013. –172 S [Ges16] Gesell,Silvio: Die nat¨u rliche Wirtschaftsordnung durch Freiland und Freigeld.R.

116 LITERATURE LITERATURE

Gesell, 1916 [Ges49] Gesell,S.: Die nat¨u rliche Wirtschaftsordnung.ed. 9. Lauf beiNuernberg :Rudolf Zitzmann Verlag, Karl Walker (editor: .pdf by Florian Seiffert 2003), 1949. –285 S [Gro04] Groves,R.M.: Surveyerrors andsurveycosts.John Wiley&Sons, 2004. –590 S [Ilg15] Ilgmann,C.: Silvio Gesell:“Astrange, undulyneglected”monetarytheorist.Taylor &Francis, 2015.–532–564 S [KD20] Kim,S.;Deka,G.C.: Advanced applicationsofblockchaintechnology.Springer, 2020. –278 S [Kel15] Keller,G.: Statisticsfor Management and Economics.Cengage Learning, 2015. – 909 S [Ken11] Kennedy,Margrit: Occupy Money: damit wirzuk¨u nftig ALLE die Gewinner sind. J. Kamphausen Verlag, 2011. –107 S [Lie19] Lietaer,Beranrd: Towards asustainable world.Delta Institue -DietarLegat E.U., 2019. –86S [P.11] P.,Warren M.: Fisher’s ”The Purchasing Power of Money”.American Statistical Association,Taylor&Francis, Ltd., 1911. –818–829 S [Sch08] Schwarz,F.: Das Experiment vonW¨o rgl: Ein Wegaus derWirtschaftskrise.Syn- ergia,2008. –89S [Swa15] Swan,Melanie: Blockchain:Blueprintfor anew economy.O’Reilly Media,Inc., 2015. –130 S [Tah16] Taherdoost,H.: Sampling methods in research methodology;how to choose asam- pling techniquefor research.2016. –18–27 S [UGLG09] Ulanowicz,R.E.;Goerner,S.J.;Lietaer,B.;Gomez,R.: Quantifying sustainability:resilience, efficiency and thereturnofinformation theory.Elsevier, 2009. –27–36 S [VTS16] Vehovar,V.;Toepoel,V.;Steinmetz,S.: TheSage HandbookofSurvey Method- ology.Sage, 2016. –329–346 S

117 Internet References

[1]Z.Bobbitt.SnowballSampling: Definition +Examples. https://www.statology.org/ snowball-sampling/.(last access: 22 January 2021). [2] M. Friedenbach and J. Timon. Freicoin. http://freico.in/.(last access: 03 March2021). [3] REGAL Verlagsgesellschaftm.b.H. DieStruktur des LEH f¨ur dasJahr 2019 (Regal 09-2020). https://regal.at/media/pages/marktdaten-studien/handel/ regiodata-branchenkennzahlen-2020/a879f19c61-1601321118/leh-2019.pdf.(last access: 22 January2021). [4] REWE International AG.Sales formatsinAustria. https://www.rewe-group.at/en/ company/sales-formats-in-austria.(last access: 24 January 2021). [5] SPAR Oste¨ rreichischeWarenhandels-AG. Konzernabschluss 2019 ifrs. https: //www.spar.at/content/dam/spardam/at/spar-website/unternehmen/daten-fakten/ spar-oesterreichische-warenhandels-ag-konzernabschluss-2019-ifrs.pdf.(last access: 24 January2021). [6] STATISTIK AUSTRIA.Statistik des Bev¨olkerungsstandes. http://www. statistik.at/wcm/idc/idcplg?IdcService=GET_PDF_FILE&RevisionSelectionMethod= LatestReleased&dDocName=123069.(last access:05February2021).

118