Analysis of Personal and Company-Based Carpooling Incentives in Belgium

UNIVERSITEIT GENT

FACULTEIT ECONOMIE EN BEDRIJFSKUNDE

ACADEMIEJAAR 2014-2015

Analysis of personal and company-based carpooling incentives in Belgium

Masterproef voorgedragen tot het bekomen van de graad van

Master of Science in de Bedrijfseconomie

Xander Steenbrugge en Maxime Dedecker

onder leiding van

Prof. Philippe van Cauwenberge

Permission for consultation

Ondergetekenden verklaren dat de inhoud van deze masterproef mag geraadpleegd en of gereproduceerd worden, mits bronvermelding.

Naam studenten: ......

2 Nederlandstalige samenvatting

België staat ocieel op de eerste plaats in de rangschikking van landen met het meeste les. Om het leprobleem in België op te lossen zijn er verschillende mo- gelijkheden gaande van een verbeterd openbaar vervoer tot het invoeren van reken- ingrijden. Een ander alternatief waar op wordt ingezet is carpoolen. In deze thesis werpen we ons licht op welke factoren de mensen aanzetten tot carpoolen en wat het in de toekomst te bieden heeft. Carpool focust op het verhogen van de bezettingsgraad van auto's en levert zowel nanciële, sociale en milieuvoordelen op vergeleken met het pendelen naar het werk alleen in de auto. Om carpoolen te stimuleren zijn er verschillende incentives die bedrijven kunnen invoeren zoals betalende parkings, preferente parkeerplaatsen of toegang tot een carpoolplatform aanbieden. Om de factoren die het succes van carpoolen beïnvloeden hebben we een uitgebreide analyse gedaan op de resultaten van de enquête voor het woon-werkverkeer van 2011.

Na analyse van deze enquête konden we besluiten dat slechts 2% van de mensen carpoolt naar het werk tegenover 51% die alleen met de auto naar het werk komt. Carpoolen is het populairst in het oosten en het zuiden van het land maar beslaat zelden meer dan 5% van het aantal pendelaars. Binnen grote stadskernen wordt carpoolen minder gebruikt dan elders. Het drukke verkeer maakt dat een klein ver- schil in afstand een groot tijdsverschil oplevert waardoor mensen oppikken lastiger wordt. Ook zijn er in grote steden dikwijls aantrekkelijkere alternatieven zoals een goed bereikbaar openbaar vervoer. Carpoolen lijkt dus eerder aangewezen in meer landelijke gebieden. Er blijkt ook een omgekeerd verband te zijn tussen het percentage mensen dat pendelt naar het werk al carpoolend en het aantal inwoners van een gemeente of stad.

De huidige incentives van grote bedrijven om werknemers al carpoolend naar het werk te laten komen lijkt voorlopig geen vruchten af te werpen, waardoor de overheid en bedrijven een grotere inspanning zouden moeten leveren om carpoolen verder te promoten. Uit een interview met de mobiliteitsverantwoordelijken van Electrabel, een bedrijf met veel aandacht voor mobiliteit, werd duidelijk dat grote bedrijven een voorbeeldfunctie te vervullen hebben. Hierbij zijn aanwezigheid van openbaar vervoer rond een bedrijfsvesting en mobiliteitsplannen een grote troef. In een overgang naar een mobieler België is de uitbouw van een goed openbaar vervoer onontbeerlijk. Carpoolen kan vooral een waardig alternatief zijn in ondernemingen gelegen buiten dichtbevolkte gebieden .

3 Contents

1 Introduction 5

I Literature study 6

2 What is carpooling? 6 2.1 Denitions ...... 6 2.2 A brief history of Carpooling ...... 7 2.3 Advantages of carpooling ...... 10 2.3.1 Financial advantages ...... 10 2.3.2 Environmental advantages ...... 13 2.3.3 Other advantages ...... 16 2.4 Carpooling incentives ...... 17 2.5 Dynamic carpooling ...... 18

II Practical analysis 20

3 Data set analysis 20 3.1 First analysis ...... 20 3.2 Custom geocoding ...... 26

4 Sustainable mobility 31 4.1 Case study ...... 31 4.2 Carpooling ...... 32 4.3 Better public transportation ...... 33 4.4 Mobility packages ...... 35 4.5 Eectiveness of the mobility management ...... 35 4.6 Future ...... 38

5 Carpooling, practical implementation 40

6 Cooperation strategies 42

7 Anonymous carpooling platform 43

8 Conclusion 44

4 1 Introduction

The country of Belgium suers from the most severe trac congestions in the world. The main sources of this problem are the high population density and the fact that most commuters travel to work alone in their car. The government tries to solve this problem by attempting to reduce the number of vehicles on the road and by creating new public transportation infrastructure that can eciently move more people on a daily basis. One of the solutions focuses on the promotion of commuter carpooling to increase the average car occupancy rate during peak hours. This thesis examines the main incentives that lead to successful carpooling and analyzes the role of the government and large companies in achieving these goals.

The rst part of this thesis claries some useful concepts about carpooling and gives a general overview of the nancial and legal systems currently in place to promote carpooling in Belgium. In the second part, a large data set of commuter trac information is processed using graphical representations made in Tableau that allow the visualization of the commuter trac across Belgium. After analyzing the data set, all acquired insights are tested with a practical example of an implemented mobility strategy in a large company located in the heart of the capital: Electrabel. Finally, some new ideas and possible solutions are presented based on this work that can lead the country towards a more sustainable and congestion free future.

5 Part I Literature study

2 What is carpooling?

2.1 Denitions

Before starting this literature study concerning carpooling, one should clearly know what the word carpooling really means. Many dierent denitions of carpooling have been adopted. To clarify their meanings, each of them has been presented below to form a quick overview of the used terminology:

Carpooling is the sharing of a ride in the private car of a driver with one or more passengers, usually for commuting [1].

Two or more employees drive to work together in a private or company car [2].

The transportation of two or more individuals in a motor vehicle with a capacity not exceeding 15 passengers, when such transportation is incidental to the principal purpose of the driver, which is to reach a destination, and when such transportation does not seek to transport persons for prot [3].

Due to the abundance of denitions it is not always easy to separate one form of carpooling from another and when dierent institutes adopt dierent denitions it can be hard to compare research results. To counteract possible misunderstandings, the following types of carpooling are distinguished.

Company-based carpooling Company-based carpooling involves all kinds of carpooling that are encouraged by the company. This kind of carpooling is usually for commuting (home to work travel), but can also be the use of a carpooling app to ex- ecute rides for work between cities or even countries. This thesis particularly treats commuting and the incentives for this type of carpooling.

Personal-based carpooling Personal-based carpooling is the use of carpooling tools at own initiative from an individual without any extra incentive but to reduce costs.

6 Internal (household-based) carpooling The most obvious form of carpooling is internal carpooling. This stands for the form of carpooling when individuals from the same household travel together for example to school or to go on a trip. It should be watched carefully before reading any data or article on carpooling whether internal carpooling is also considered in the numbers. When this is the case gures can be drastically dierent from researches where internal carpooling is not taken into account. Internal carpooling is characterized by the fact that no detour and consequently no time is lost for picking up the passenger and that the people that ride along are well-known, which is on its turn an important parameter to facilitate carpooling.

External (non-household based) carpooling External carpooling is the op- posite of internal carpooling. Here people that share the ride with each other are not living in the same place, excluding the required detour to pick up passengers. These people can however be family or colleagues and thus possibly have a form of connection.

Carpooling is not the same as car sharing Often the word `carpooling' is mixed up with `car sharing' and `ridesharing'. While ridesharing is in theory just the sharing of empty seats in a vehicle, it can also be used intermittently for carpooling. Car sharing however is the joint ownership of a car or car park by its users. In simple words, carpooling exploits the cars already owned by people, while car sharing exploits cars bought precisely for and by the car sharing group [4] ". A famous example of car sharing in Belgium is Cambio. To rent a car a subscription fee has to be paid monthly to get access to the car park, while an extra fee has to be paid for the actual distance that you have covered with the car. This implies that it is actually possible (for the sake of correct terminology) to carpool in a car-shared car. Hitchhiking cannot be considered as carpooling since the people that carpool have agreed on beforehand to the terms where they meet and how to share the costs.

2.2 A brief history of Carpooling

Sources of early carpooling are rather scarce due to the small line between private automobile traveling and carpooling and the diculty in measuring ridesharing in the early days. This is especially the case for the development of carpooling in Eu- rope. Consequently, the majority of the information provided below was extracted from an article from the Massachusetts Institute of Technology (MIT) that summa- rizes all important carpooling periods in America [3].

7 Carpooling has existed almost as long as the car itself. Generally Carl Benz is looked upon as the inventor of the car. His rst automobile was built in 1885 in Mannheim, which is also considered the birth year of the car as we know it [5]. From that moment on, lots of people built cars and rened the techniques to improve the travel distance and speed. Until 1905 cars were only for the very rich. From that period on until 1914, cars became gradually cheaper and more aordable for the middle class. At the start of the First World War, car production plants shifted their attention more towards military cars and private cars were again too expensive for the biggest part of society. In 1919, immediately after the rst world war, cars were still very rare, however before the great depression in 1930, cars were already common sight in the streets. Carpooling has it earliest roots in this period.

In 1914, the U.S. experienced the so called Jitney Craze. At the beginning of World War I there was a recession in the U.S., making cars no longer aordable for many people. Entrepreneurs and rich people took advantage by picking up passengers along the street in exchange for a so called jitney. The jitney was the name of the street car fare of 5 cents. By the end of 1915 the jitney attitude had spread across the whole nation. It is hard to appropriately indicate if the jitney service is rather an unregulated taxi service or the rst emerging form of carpooling. Either way, the fact remains that while some people intentionally bought a car and tried to pay it o with the money they got for transporting people (which is more like a taxi service), others were hoping to divide the costs for rides that were going to happen anyway. From a survey in 1915 it was seen that 62% of the jitney drivers in Houston did it as a form of commuting to share the personal costs of ownership and gasoline [6]. Due to the great success of these jitney services, the street cars (until then the main source of public transport) experienced serious losses of income. The government, earning taxes on the gain of street cars, was also suering from this diminished cash ow. That is why the end of the jitneys came as fast as its success had started [7]. The government was obliging jitney drivers to have a drivers license and to join a liability bond. This was not an unnecessary measure since many jitney drivers were known to drive very irresponsible and fast, causing accidents. By the end of 1918 the Jitney hype was over.

From this earliest phase of carpooling it appears that nancial pressure was the main reason to commute together rather than ecological awareness or social matters. The major problems were that the liability and the security of the trip could never be

8 guaranteed. Interestingly, those exact same problems still draw back people from carpooling these days.

During World War II traditional carpooling was born. Contrary to the jitney craze, government supported carpooling in order to save resources for war purposes. At rst, limited eort was done by the government to convince people to carpool and save rides. Later, when it was seen that the impact was not sucient, a more serious program was set up in cooperation with the oil industry to educate the people on the importance of saving fuel and sharing rides. Results of this program were however not clear since there were no good means to measure carpooling participation at that time.

During the 1970's a third carpooling period of interest took place. In 1973 the Arab Oil Embargo was raised against Canada, the US, the UK, the Netherlands and Japan due to their involvement in the Yum Kippur War. The US President Nixon imposed new rules including a maximum speed limit of 55 MPH on highways in order to save fuel. During this period nancial compensation was accredited to initiatives that supported ridesharing. In that period, ecological concern was another reason to encourage carpooling. As a consequence of the Clean Air Act Amendments of 1970, more stringent rules were imposed to improve the air quality. Ridesharing was seen as one of the methods to reduce air pollution. High Occupancy Vehicle (HOV) lanes were installed on highways and/or bus lanes in cities where, depending on the period, vehicles with 4+ or 3+ passengers could ride on. Despite the fact that these lanes are very useful, there weren't many built. Only a few cities implemented them. During the period of the Oil Embargo, the rst studies about carpooling arised. It was quickly seen that employers would have to fulll an important role to successfully encourage carpooling. At that time there was no good method to determine the success rate of the carpooling incentives.

Carpooling kept on being subsidized by the American State, leading to about 23.5 % of people carpooling in 1980. Carpool research was now started to be done system- atically and the eectiveness of measures was carefully examined. However, after 1980, oil prices fell back, income was again rising and the government support for carpooling disappeared.

Today ridesharing activities seem to boom again. Everywhere new carpool platforms are set up and other green alternatives are promoted as well. Especially in America,

9 Figure 1: Car occupancy rate in Europe just as during the jitney craze, all this happens largely without support of the government. In Europe more carpool platforms have been raised by the support of public nancing.

2.3 Advantages of carpooling

The massive ownership of private and company cars and the continuous drop in household size result in declining car occupancy rates [8]. Figure 1 indicates that the average car occupancy in Europe between 2004-2008 is between 1 and 2 passengers. Low average car occupancy generates environmental, economic and social problems whereas carpooling is known to oer nancial, environmental and social advantages. In this section the most important advantages of carpooling are pointed out.

2.3.1 Financial advantages

The nancial advantage consists mainly of 2 factors: the shared costs of the ride and the scal compensation.

The costs of a carpool ride are usually shared. This oers nancial advantage to

10 the driver since the ride would be undertaken anyway. In Belgium, the average cost of a family car is calculated at ¿ 0,3468/km [9]. This price includes the fuel cost, insurance, taxes and depreciation of the car. Because a family car can comfortably seat 4 passenger a carpool remuneration of about ¿ 0,08/km is considered normal. If the driver has to make a detour to pick up the passenger, a higher compensation of about ¿ 0,21/km can be asked for those kilometers [10].

At all times it should be watched over that the amount of money only covers the costs. When prot is made it is no longer carpooling, but a kind of taxi service which is subject to strict regulations. In this regard the arrival of the popular Uber- service in Belgium is particularly interesting. With the Uber-service everyone can simply sign in for a ride with a privately owned car. Uber can denitely not be seen as a carpooling app since the driver is making prot from the ride and not just covering his costs. In 2014 the Brussels court decided that Uber could not operate in the capital for this reason. Recently an Uber driver was convicted for the rst time in Belgium. The drivers car was conscated [11]. Dozens of similar cases are still in the waiting line and the Brussels Minister of Mobility is currently looking at taxi company regulations and how to integrate these new emerging taxi applications within the existing transportation services.

Cost sharing is not the only nancial incentive. In Belgium an employee can get a remuneration for commuting between home and work. This remuneration is paid by the employer and is part of the income and consequently subject to taxes. Part of the remuneration is exempted from taxes. For a commuter traveling alone in his private car a maximal exemption of ¿380 is foreseen on the remuneration paid by his employer. Of course the exemption is never higher than the amount of the remuneration. For carpoolers the maximal exemption can be higher than ¿380 if certain conditions are fullled. These conditions are the choice of tax systems, whether a carpooling program is present in the company and the ownership of the car. Sources that were used to obtain the scal information were the scal guidelines for the carpooler [12] and the site of the FOD nancien and related gure [13]. All amounts used in this paragraph are valid in Belgium in the scal year 2015.

In Belgium there are 2 dierent tax systems. One is called the xed job expenses, while the other one is called the real job expenses. In the rst system a xed amount - based on the income of the employee - is tax-deductible and covers the costs that have been made for work. Sometimes, the amount is not sucient to cover all costs.

11 In this case it is better to opt for the real job expenses system where you have to prove all costs [14].

When a carpooling program is organized within the company and this company registers you as an ocial carpooler, it is possible to obtain a higher tax exemption than when the carpooling is organized at own initiative. In a recognized carpooling program the employer has to draft a carpooling code and has to control the carpooling employees because they take a declaration on oath while signing in for the program.

Ownership of the car is important as well because when the car is in private ownership, the company has to pay a remuneration for commuting. However, when the company provides a car for its employees, this remuneration does not have to be paid and a substitute income is foreseen under `advantages of all kind'.

Private car Fixed job expenses system Carpooling program in the company The maximum scal exemption on the remuneration for commuting is equal to the price of a rst class train ticket for a week covering the distance of a single ride between home and work. This price is multiplied by the days of carpooling and then divided by 5. It is highly unlikely that the employer's remuneration for commuting reaches this amount. The whole remuneration is thus exempted from taxes. If carpooling is only used some days and the other days the employee commutes by car, it is still possible to prot from the `normal' exemption of ¿380 for this part of the remuneration. If the carpooler chooses to come by public transport, the remuneration for this part will be completely exempted from taxes.

Private car xed job expenses system no carpooling program in the company The carpooler only can get a tax exemption of ¿ 380 on the remuneration for commuting from his employer. The extra advantage compared to commuters traveling alone by car is that they get a compensation from there passengers.

Private car real job expenses system Employees choosing to prove their job expenses can exempt ¿ 0.15/km from their taxes. This is valid for both the driver and the passengers. For passengers the limit is set at 100 kilometers for a single ride. In the real job expenses system the remuneration of the employer for commuting or

12 the compensation for carpooling is completely taxable.

Company car xed job expenses system no carpooling program For employees with a company car no tax exemption is possible. A company car is considered as a part of income under `advantages of all kind'. Taxes have to be paid onto this amount. Since January 2012 carpool passengers cannot longer prot from an exemption while occupying a company car.

Company car xed job expenses system carpooling program Only in one case it is possible to get a kind of tax advantage while carpooling with a company car. Some companies set available company cars only for carpooling. This means that at least 80% of the travel distance is covered carpooling. In this case there is no need to pay `advantages of all kind'.

Company car real job expenses system The same rules are applicable as with a private car. Remark that there is no remuneration for commuting since the car is company-owned. An overview of all the scal measures involving commuting by carpooling can be found in Attachments Figure 21 [13]. This gure is in Dutch and gives a schematic indication of what the Belgian rules are.

2.3.2 Environmental advantages

CO2 Since most cars in Belgium use non-renewable resources such as gasoline or diesel, they are responsible for a big part of the total consumption of these products. Gasoline and diesel are not only very costly; the combustion of these fossil fuels also causes the emission of other harmful substances that damage overall human health.

One of these combustion gases is CO2.

In Belgium an average car exhausted about 121.33 grams of CO2 per kilometer in 2014 [15] (see gure 2). This is slightly less than the European average of 123.4 grams CO2/km. With this number, Europe has succeeded in obtaining its target of 2015 which was set on 130 grams CO2/km. The next target is set on 95 grams

CO2/km in 2020. [16]. Despite the fact that the amount of CO2 exhausted per car is declining, the total amount of CO2 exhausted in Europe is hardly going down as

13 Figure 2: Average CO2 emissions from personal vehicles in Belgium. the number of cars riding on European roads is still rising. In 2014, a total of 12.5 million cars was registered in Europe, which is the rst overall increase since the crisis of 2007 [15]. In Belgium however, in 2014 the amount of cars was stable for the rst time in years. Only when the number of cars is stable or declining, the impact of the reduced CO2 norms has a clear impact on the environment and more specically on global warming. A side mark that has to be placed with the declining trend in

CO2 exhaustion levels is that many car constructors have specically tuned their cars so that the amount of CO2 is theoretically less, but when tests are performed in practice, it seems that this theoretical amount of exhaustion is quite unreasonable to obtain [17].

Other harmful combustion gases Other substances that are introduced into the atmosphere through the combustion of gasoline and diesel (and that are perhaps less known by the common public) are NOx, particulate matter (PM) and CO. NOx can transform into acid rain through complex atmospheric reactions. It is responsible for the formation of ozone [18] and is known to cause multiple forms of respiratory problems. The primary source for NOx is road trac as can be seen in gure 3.

PM is even more directly dangerous to human health. PM is normally followed by a number indicating the particulate's diameter in µm. The smaller the diameter of

14 Figure 3: Nitrogen oxide emissions by sector.

Figure 4: Fine particulate matter emissions in Belgium. the particulate matter, the deeper it can penetrate trough the respiratory system, causing serious breathing problems and making it more vulnerable for infections [18]. Fine particulate matter is also known to penetrate in the blood, causing blood and heart diseases. Currently intensive research is being undertaken to examine the eects of PM because they are thought to be underestimated heavily. In Belgium, road transport is responsible for about 27% of the PM2,5 and 24% of the PM10 emissions (see gure 4) [18].

15 2.3.3 Other advantages

Reduce trac congestion

Belgium is famous for its trac jams and has the 2 cities with the most congested trac in the world. These cities are Antwerp and Brussels [19]. This according to the INRIX index which is the general measure for trac congestion. The principle of the INRIX index is simple. A road with no congestion gets a score of zero. Every extra point is a percent increase in the average travel time above free-ow conditions during peak hours. The peak hours are set between 6am till 10 am and 3pm till 7 pm from Monday till Friday [20].

According to this INRIX index Belgium is rst with an index of 20.2, indicating that averagely one fth of our travel time should be added to free ow conditions. Belgium leaves behind countries like the UK (INRIX 16.5), the Netherlands (INRIX 15.8), Italy (INRIX 15.6) and Canada (INRIX 14.2) [21]. In Europe there is no country where people have longer home-work travel distances than in Belgium [22]. One of the important reasons according to OECD is the scal regime around company cars. These are heavily supported by the Belgian government by tax benets.

Commuters confronted regularly with trac jams are known to suer from higher stress levels [23]. From the same research it was seen that automobile commuters traveling regularly on more congested routes than their colleagues showed increased decits in task performance. This was conrmed by later research [24].

Trac jams do also cost a lot of money. Although economic costs of trac congestion are hard to estimate, VOKA calculated an amount of ¿ 2.000.000 euros every working day, adding up to a signicant ¿ 500.000.000 euros a year [25]. The OECD wrote in its report that it is estimated that the congestion cost for Belgium is about 2% of the GDP (OECD) indicating a better use of infrastructure could lead to growth [26].

Meeting new people A social advantage of carpooling to work is that one can meet new people in an entertaining way. This does not only provide a nice chat when stuck in trac jams, but can also ameliorate the way in which colleagues are appreciated resulting in better performances on the work oor.

The problems mentioned in 2.3.2 and 2.3.3 can only be solved by a transition to a less car-intensive road transport. Accomplishing this is only possible by a more accessible

16 and intensied public transport or by more car sharing initiatives. Overnancing of car usage in Belgium seems to be a major problem [26]. The utopia where everyone lives near his work seems nearly impossible to implement in Belgium.

2.4 Carpooling incentives

Aside from the advantages already highlighted above for encouraging carpooling, governments and companies can launch other incentives from which the employee can benet by carpooling. This section highlights some incentives and their eectiveness that have been adopted by companies or by Belgian government.

The main decision level for mobility in Belgium are the three regions. This means that in Brussels, Wallonia and Flanders dierent approaches are possible. Gov- ernments in Belgium have a signicant inuence on the mobility management in companies by encouraging them (either nancially or legally) to think about the mobility of their employees. In Brussels companies with over 200 employees are obliged to give a mobility plan for every employee. In Flanders, a specic Commut- ing Fund subsidizes projects that enable more sustainable commuting. The Walloon region promotes mobility plans for business parks and in such way encourages smaller companies to communicate and to coordinate their mobility plans [2].

Parking space Most cities with a lot of job activity have problems with parking space. Companies have to build expensive parking lots or underground garages to stall the cars of its employees. Carpooling helps to reduce parking space, resulting in a cost saving for the employer. Employees also benet from this fact since less cars mean less pressure on the available parking. Another parking related incentive is the introduction of preferred parking spaces. Similar to privileged parking spaces for disabled people, companies can locate some of the available parking space close to the company and make it accessible only for people that commute by carpooling. However, the most eective carpooling incentive still remains the exemption of parking charges for carpoolers while people driving alone have to pay [27] [2].

Guaranteed ride home Carpooling people have to travel back home together as well. Companies working with a lot of exitime jobs need to do more eort in promoting carpooling by assuring that employees can get home. Sometimes the company accounts for the costs for a taxi in case of emergency. The impact of these guaranteed ride home measures on the carpooling behaviour of employees is still being debated, and many studies report dierent results on the subject [2].

17 Matching service Some companies rely on already existing carpooling organizations to nd matches for their personnel. By actively working together with such organizations it is easier for people that want to carpool to nd a match. This incentive is considered to be minor.

2.5 Dynamic carpooling

The new technological evolutions might give carpooling a new boost. Most problems associated with carpooling relate to the gathering of a stable pool of committed carpoolers. With internet now available almost everywhere and for everyone, it be- comes easier to connect to a carpooling platform as an individual and to see all readily available car seats in real-time [28]. Dynamic carpooling (also called real- time carpooling) involves the use of internet and GPS tracking systems to arrange a carpool ride from anywhere possible. Dan Kirshner, researcher in this eld and maintainer of http://dynamicridesharing.org website denes dynamic carpooling as follows: A system that facilitates the ability of drivers and passengers to make one-time ride matches close to their departure time, with sucient convenience and exibility to be used on a daily basis [?]. Recently a lot of carpooling platforms have been created leading to a high number of dierent carpooling apps. Currently available are Djump, Carpool.be, CoCar, ANWB Samenrijden, BlaBlaCar, Ride Here Right Now and many others. Until now there success has been limited.

The failure of implementation is mainly caused by three reasons [28] [29]. The rst one is the critical mass. The critical mass is the minimal number of carpoolers that is required in a certain area to create a pool of users that is big enough to operate eciently. Ecient means that there are always enough users online so it is easy to ll your car as a driver, or nd a seat in a car as a passenger. The second big problem is the perception of carpooling. Carpooling, as opposed to other sustainable ways of transport (e.g. electrical cars, train), is still not considered trendy. It also lacks some important incentives like time saving or nancial advantage to become really appealing for the majority of people. The third and maybe most important problem is that valuable alternatives are available for carpooling: good organized public transport, cheap taxi services or electrical bikes.

Up to now only 2 carpool apps in Europe have succeeded in reaching the critical mass and provide a decent service to its customers at all times. These apps are BlaBlacar and carpooling.com. BlaBlacar is originally from France while Carpool- ing.com was set up in Germany. Very recently, in April 2015, BlaBlacar has taken

18 over the company Carpooling.com, thus acquiring the biggest carpooling platform ever created. With over 20 million members, and a monthly transport rate of over 2 million unique passengers, BlaBlacar is doing very well and is still expanding [30]. With Mexico being the newest country where BlaBlacar operates, the total number of countries is now 19. Due to the high amount of carpool passengers, BlaBlacar is also making a real impact environmentally. By realizing an average car occupancy rate of 2.6, compared to an average of 1.6, About 700.000 tons of CO2 have already been saved [30]. While BlaBlacar seems to be very promising for all kinds of dynamic carpooling, the platform is not yet showing any promising signs that it will also be able to tackle the commuter trac problem (which is mainly responsible for most situations of trac congestion and environmental problems). Therefore, some new ideas and governmental or company based mobility strategies are denitely worth looking at.

19 Part II Practical analysis

3 Data set analysis

To investigate the signicance of the theoretical concepts that we investigated in our literature study, we decided to test them with a real world dataset. Every third year, the Belgian government runs a survey among all Belgian companies that have at least 100 employees. The goal of this survey is to map the distribution of the daily commuter trac in Belgium for all available methods of transportation. Each involved company is asked to supply the postal code of its working oce along with the postal codes of all its employees and their daily method of transportation from home to work. These methods of transportation are subdivided in 10 separate categories that form a single vector per data entry:

[caralone, carwithothers, train, deLijn, deT ec, deMivb, collective, bicycle, motorcycle, onfoot]. The entire dataset has 263 941 entries, each consisting of the combination of a certain company (alone with the postal code of its working oces) and the postal code of one or more employees and the corresponding data vector containing their methods of transport. The dataset was provided as a raw Excel sheet containing a total of 907 311 employees across Belgium (the total active population in Belgium is about 4,5 million). Even though only about one fth of all working employees in Belgium are contained in the dataset, the information is very relevant for the current commuter trac situation in the country because the working population that is not contained in the data consists for a large part of people that do not travel large distances to get to work (for example self-employed people that work at home).

3.1 First analysis

In this section, some initial graphical representations are displayed that were generated using the raw data from Excel without any additional processing. The gures were generated with Tableau, a relatively new data visualization software tool that is becoming rapidly more popular for the analysis of large amounts of data. As students, we were able to obtain a one year free trial version of this software tool to perform our analysis. The gures below give a rst, crude impression of the infor-

20 Figure 5: The distribution of jobs across Belgium. mation contained in this dataset.

Figure 5 shows the distribution of jobs in Belgium based on the geographical coordinates of the oce where each company is located. The rst thing one notices, is that the overwhelming majority of jobs is located in Flanders, the northern part of the country. Apart from a minor cluster of cities in the Samber and Maas valley (including Luik, Namen, Charleroi and Bergen) most of the job opportunities are located above the language barrier that separates Flanders from Wallonia. Further- more, the largest portion of jobs is located in the capital of Brussels, followed by Ghent and Antwerp.

Figure 6 shows the number of employees that commute to work alone by car, displayed for every postal code of residence. We can see that especially in Brussels, Antwerp, Ghent and Leuven, only a small percentage of people commute to work alone by car (25% - 35%). However, in the southern part of the country almost everybody travels to work alone in their cars. The employees in the western part of the country also tend to do this. If we compare this view with the one from gure 5, we can see that areas where a lot of jobs are located tend to have a lower

21 fraction of commuters that go to work alone by car. The main explanation for this is that these areas are located in densely populated cities which oer more viable alternatives in the form of public transport. The southern part of the country is too sparsely populated to make intensive public transport available, and thus, most people drive to work alone by car.

Figure 6: Distribution of the employees that commute to work alone by car.

The distribution of bicycle commuters across Belgium is exceptionally striking. It is displayed in gure 7 along with the language barrier and the contour of the district of Brussels (both indicated with black lines). Analyzing the data, we can agree that in the Flemish part of Belgium, on average about 10%-25% of employees commute to work by bicycle. The most signicant contributions are located in Antwerp and Ghent. However, in the French part of the country (below the language barrier indicated in black) and in Brussels, almost nobody uses a bicycle to commute to work. The explanations for this striking dierence between the two regions cannot be obtained from the dataset alone, but the average live-work distance, available bicycle infrastructure (for example safe bicycle paths through dense cities and storage space at the location of employment) and the general attitude towards cycling (maintaining a healthy, young spirit) are denitely among the possible causes that should be

22 investigated.

Figure 7: Percentage of employees that commute to work by bicycle.

In contrast to the number of bicycle commuters, the percentage of employees that commute to work in a car with others is somewhat higher in Wallonia as compared to Flanders. However, as can be seen from gure 8, the fraction of carpoolers rarely exceeds 5% of the total commuting employees. Additionally, we can see that in the largest working areas (Brussels, Antwerp and Ghent) almost nobody commutes to work by car with others. One reason for this could be the intense trac density in those areas, which results in signicant extra travel time necessary to drop people o, combined with the usually better public transport alternatives that are not generally available in more rural areas.

Figure 9 shows a scatter plot containing each postal code in Belgium as a separate data point. For each postal code, the number of inhabitants is plotted on the x- axis and the percentage of carpoolers on the y-axis. We can see that there is a signicant negative correlation between the two with a statistical p-value of 0,0067 which corresponds to very strong evidence. The resulting trend line is plotted on

23 Figure 8: Percentage of employees that carpool to work. the gure as a red dotted line. We can conclude that on average, people living in large cities tend to carpool less frequently than people living in small villages. This result corresponds with the view of gure 8, where dense urban areas in large cities correspond to a low fraction of carpoolers. One possible conclusion that could be drawn from these grounds is that carpooling is often used as an alternative to public transportation when the latter is not sucient to satisfy the needs of the commuting employees, however this conclusion turns out to be false, as is pointed out by the next gure.

In contrast to gure 9 where a strong negative correlation was found between the number of inhabitants per postal code and the corresponding fraction of carpoolers, gure 10 shows that there is absolutely no such correlation when we look at the location of employment instead of the place of residence. There is thus no particular correlation between the number of employees that work in a certain city and the fraction of employees that carpool to work. If it were true that carpooling serves as a straightforward alternative to insucient public transport in sparsely populated rural areas, than there should also be a clear negative correlation between carpooling and the number of employees per postal code (because people working in small

24 Figure 9: Correlation between number of inhabitants and percentage of carpoolers per postal code. villages would often resort to carpooling to get to work whenever public transportation is lacking), however this is clearly not the case (see gure 10).

To look at this comparison in a dierent way we can think of gure 9 as an indication for the social eect of personal-based carpooling incentives: people living in smaller villages tend to carpool more often than people living in large cities. One possible explanation for this is that people from small villages have more personal contact with their neighbors and are therefore more likely to decide to start carpooling together which explains the negative correlation seen in the gure. On the other hand, gure 10 can be seen as an indication of the average eort that large companies are putting into the promotion of carpooling. By agreeing that most large companies are located in areas where a lot of people work (right side of the x-axis), it is clear that when these large companies implemented successful carpooling platforms, the percentage of carpoolers should increase in those working areas, which would result in a (obviously non-existing) positive correlation in gure 10.

The comparison of these two gures implies a rst conclusion about carpooling

25 incentives: Based on the comparison of gures 9 and 10 we can conclude that most (if not all) company based carpool incentives in Belgium are currently ineective or non-existing. The only observable geographical stimulus for carpooling seems to be the more direct social networks between possible carpooling partners, that appear to be negatively correlated with the working population of the residential area.

Figure 10: Correlation between number of employees and percentage of carpoolers per postal code.

3.2 Custom geocoding

Examining the raw data gave us some very useful initial insights in the distribution of commuter trac in Belgium, however the visualizations remain very static considering the fact that the data points always have to be grouped according to either the place of residence or the location of employment. Because of this static grouping, the actual physical displacements that correspond to the commuter trac can not be seen. To remedy this restriction we implemented a 'custom geocoding' in Tableau that allows us to visually map all the commuter trac displacements. To do this, we rst made a CSV (Comma Separated Values) le containing the GPS coordinates of every single postal code that occurs in the dataset. Next, we implemented a Matlab

26 script that generates all possible combinations of two postal codes in Belgium. These pairs correspond to all possible commuter displacements (from place of residence to the location of employment). Consequently, the Matlab script iterates over all the available entries in the dataset and stored the corresponding data vectors under the correct pair of postal codes. (Remember that the data vector for each entry contains the number of people on this commuter route (specied by the pair of postal codes) and their corresponding means of transportation.) After running the Matlab script, an output data le is generated that lists all possible daily commuter routes containing two postal codes (place of residence and location of employment) and the total number of people on this route for each possible method of transportation (see section 3). Note that displacements within the same postal code are also included as they account for a large portion of the total commuter trac (people that live and work in the same city).

Figure 11: Distribution of all daily commuter trac routes in Belgium.

The geographical distribution of all daily commuter trac is presented in gure 11. Note that in this gure, all separate transportation methods were combined to get a general view of the displacements that Belgian employees undertake on a daily

27 basis, however only displacement routes that have at least 30 daily commuters have been shown in the graph. We can see that the capital of Brussels receives by far the most signicant portion of daily commuter trac, followed by Antwerp and Ghent. Also notice the large red circles (for example in Ghent) that represent displacements within the same postal code, thus corresponding to people that live and work in the same city.

Figure 12 displays the overall distribution of the daily commuter trac represented in gure 11. We can see that more than half of all commuter trac consists of individuals that drive to work alone by car. 17% of commuters go to work by bicycle, and a total of 25% use public transport to get to work (train + buses + metro).

Figure 12: Distribution of daily commuter trac in Belgium.

To get a better view on the most dominant form of commuter trac, the graphical representation of gure 11 was repeated using only the 'car alone' category from the dataset. The resulting distribution of commuters that drive to work alone in their car is visualized in gure 13, once more only routes that have at least 30 daily commuters are shown. We can see that the distribution closely resembles the

28 one in gure 11 (after all 51% of all commuters are part of this category) however, most displacements are more locally conned, meaning that employees that have to travel large distances (>60km) prefer other methods of transportation. Also notice the relatively large red circles all over the country that represent the people that drive to work alone, within their city of residence.

Figure 13: Distribution of commuter trac in Belgium consisting of employees that drive to work alone by car.

Figure 14 represents the distribution of daily commuter trac using the train to get to their location of employment. We immediately see that the capital of Brussels is the major center of daily commuter trac by train, again followed by Ghent and Antwerp.

In gure 15 we can see a detail of the representation in gure 14 that shows only train routes that have at least 150 daily commuters. We can see that the route between Ghent and Brussels is by far the most important commuter route, somewhat redening the view represented in gure 14 where Antwerp also seemed to have a large contribution. Note however that the gure was generated using commuter routes between any two exact postal codes. This means that the visible train

29 Figure 14: Distribution of commuter trac in Belgium consisting of employees that travel to work by train. route between Ghent and Brussels only incorporates commuters that actually live in Ghent (9000) and work in Brussels (1000). However, many more people take this train route everyday.

To relate the commuter trac in Brussels with the rest of the country (recall that on average about 12% of Belgian commuters goes to work by train (see gure 12)), we can compute the transport fractions from the dataset while using only commuters that live outside of Brussels, but have a job in the capital. The result is that about 50% of all employees coming from outside of the capital reach their work by train, 40% come alone by car and 10% use other methods. Only 1% of all these commuters carpool to work. We can thus conclude that because of the enormous trac congestion problem in Brussels, many people have switched to the train to reach their work while almost nobody resorts to carpooling even though still about 40% (about 60 000 people every day) drive to work alone in their car.

Considering the fact that Brussels is the most trac congested city in the world (see section 2.3.3) while 60 000 people drive to work from outside the city alone

30 in their car every single day seems almost unbelievable. It is very clear from these numbers that even though the train infrastructure seems to provide a viable alternative for many people, a lot of work still has to be done to dramatically increase car occupancy rates or nd other alternatives for people that cannot easily use the train to go to work. To shed light on some of the possible solutions in the future, the next section will discuss some of the ideas that are currently being implemented by a large and successful company located in the heart of Brussels.

Figure 15: Distribution of commuter trac in Belgium consisting of employees that travel to work by train. Only routes that have at least 150 daily commuters are shown.

4 Sustainable mobility

4.1 Case study

After combining the insights from our literature study and the analysis of the dataset, we decided it was time to test our current view on carpooling and mobility with a case study. To ensure a realistic view on the current condition of

31 company driven mobility policies, we contacted a large rm with substantial inter- ests in the promotion of sustainable mobility: Electrabel - GDF Suez. Our contact persons were general mobility manager An De Backer and Sustainable Development & Business Innovation Analyst Igor Lefebvre. This section contains several new approaches to the general mobility problem in Belgium that are based on our arranged meeting with these two people.

4.2 Carpooling

Electrabel - GDF Suez uses a large Belgian carpooling system that is the result of the merger between carpool.be and eurostop.be [31]. The platform is a cooperation between VAB and Taxistop and is a non-prot organization that aims to connect as many carpoolers as possible. The passengers only pay a small fee that helps to cover the expenses of the driver (consisting e.g. of fuel cost, insurance costs and depreciation of the vehicle). Note that neither the driver nor the organization retains any prot from the carpooling activities [32]. In the carpooling system implemented by Electrabel, employees do however obtain nancial remunerations per carpool driven kilometer. An important fact here is that the main target group for the implementation of this carpooling system within Electrabel consists of the two electricity generating nuclear facilities in Doel and Tihange. Several reasons can be found why these locations were chosen for the implementation of a carpooling system:

They are located well outside of dense urban areas that suer most severely from trac congestion, allowing carpoolers to actually get to their destination in a reasonable amount of time.

These nuclear facilities require a surprisingly large number of people to operate (each of them has about 650 permanent employees), resulting in a large concentration of people, thus increasing the exibility and benets of an implemented carpooling system.

Nuclear plants are very stable working environments, resulting in relatively xed working hours for their employees, which makes a carpooling system easier to implement as compared to other work places where exible working hours are often used.

Because they are located outside of dense urban areas, sucient parking space can be made available to accommodate all vehicles.

32 The fact that these facilities are located outside of large cities facilitates carpooling (see above), but at the same time reduces the need for such a system because most employees can just use their own cars to drive to work. Therefore extra stimulating measures are essential to motivate indecisive individuals to at least consider carpooling as a means of commuting to work. While some implementation steps are already being taken by Electrabel to promote carpooling, it became very obvious during our conversation with the mobility team that carpooling is seen as a complementary solution mainly targeted at large and easily accessible industrial parks. However for modern, crowded cities that are plagued by trac congestion, carpooling seems not to be the best option as will be explained in the next section.

4.3 Better public transportation

From the analysis of the dataset we learned that the majority of people who commute to Brussels on a daily basis reach their work using public transportation. This can be easily explained by considering the simple and yet astonishing fact that, according to the annual INRIX Trac Scoreboard [20], the daily rush hour trac jams in Brussels are the worst in the world. As if this isn't bad enough, the city of Antwerp scores second in this global ranking. Needless to say that the roads in these cities are not exactly ecient transportation routes to commute to work everyday. For this reason the main method of transportation that is promoted by Electrabel is the train. You may think your daily commute might be bad, but drivers in Brussels face the worst trac jams in the world according to the annual Trac Scorecard compiled by trac and navigation service provider INRIX in Kirkland, Wash. The second worst city in the world for trac jams Antwerp is likewise situated in Belgium, a modest-sized European kingdom that's otherwise best known for beer, chocolate, wa es and the Smurfs. Who knew? On top of the generally shorter travel time, commuting to work by train oers several other benets in comparison with the car: The train is always cheaper in the long run (when taking fuel cost, insurance, depreciation costs, ... into account).

The train is a lot 'greener' as well, emitting about 2,5 times less carbon into the atmosphere per traveled passenger kilometer as compared to a standard personal car.

33 This lower carbon emission does not only positively aect the global climate but is also of great concern to the health of the people living in polluted cities since reoccurring smog has been shown to cause serious lung diseases that require expensive medical treatments, which are ultimately paid by the public in the form of taxes (see literature study).

Finally, people commuting to work by train have been shown to actually start their work as little as 5 minutes after arrival, while people driving to work with a car need an average of 30 minutes to recover from the stress and anxiety induced by dense, urban trac. According to a survey conducted by the mobility managers among Electrabel employees, commuting to work by car is only benecial compared to the train when the traveled distance is greater than 60 km.

When asked what the role of the government is in this issue, the mobility managers replied that: It is the duty and responsibility of large companies such as Electrabel to inuence government policies to encourage the development and construction of modern, public transport. As a consequence, Electrabels approach is simple: promote public transportation as much as possible and use cars as a nal solution for those people that have no valid alternatives. For Electrabel, promoting public transport is more vision than choice. For this reason, Electrabel chose for a very radical approach: over the course of 4 years they have moved their entire headquarters (consisting of over 3000 employees and several separate buildings) to a new, single location located merely 400 meters from one of the main train stations in Brussels. Additionally the number of parking spaces available for employees was radically reduced to provide additional incentive. The eects of this radical implementation of the mobility management strategy will be discussed in section 4.5, but for now we can conclude that carpooling is seen only as a complimentary partial solution for Electrabel that can help to transport employees from their homes to large, public transportation nodes such as bus and train stations and for rural areas outside of large cities that are not aected by heavy trac congestion. To facilitate the huge stream of employees that commute to dense urban areas on a daily basis, more fundamental and sustainable solutions need to be found that do not rely on personal vehicles as a means of transportation.

Even though a modern and satisfactory public transport seems to be the most promising solution to the mobility problem, there still remain some considerable issues with this form of transportation. First of all, delays and canceled rides are

34 very common due to the unpredictability of a large public transportation net, used by many people on a daily basis. Secondly their is still the problem of getting employees from their homes to the public transportation nodes. This problem could for example be solved by building large parking spaces for personal vehicles far outside of dense cities or by increasing the number of operating bus lines that connect living areas with transportation nodes. Finally, some hybrid forms of sustainable mobility also exist that oer a very exible solution to employees that are willing to accept some variation in their means of transport as will be briey discussed next.

4.4 Mobility packages

To further encourage employees to make the transition to a more sustainable mobility solution, several mobility packages are oered within Electrabel. These packages include dierent combinations of monthly covered car- and train kilometers, fuel cards and railcards. Some families for example have made a 'full switch', using only sustainable transport methods, leaving them stranded without a family car during the holidays. For this reason, the 'Holiday-switch' option was recently introduced, allowing them to use a fully electrical car during the year and switching to a large mono volume family vehicle during the summer months. These mobility packages are a great example of how sustainable mobility can be driven by large companies that have enough employees to oer a exible selection of alternatives.

4.5 Eectiveness of the mobility management

To investigate the result of the company based eorts to promote sustainable mobility, we can again use the dataset that was discussed in section 3. Based on the processed data matrix, we can generate a list of all postal codes where a specic company has oces, along with the corresponding number of employees and their methods of transportation. Generating this list for Electrabel leads to the result displayed in table 1 (only the 7 locations with the highest number of employees are displayed here). We can see that most of the employees either work in the headquarters in Brussels or in one of the two nuclear power plants. Moreover, we can see that for the headquarters in Brussels, more than half of all employees commute to work using public transport, indicating very clearly that the mobility management for the urban area of Brussels is indeed working (the average fraction of train commuters for Brussels is 35%, based on the dataset).

For the nuclear plants that are located outside of urban areas (and thus less acces-

35 Table 1: Distribution of all Electrabel employees over the dierent oces, along with their corresponding means of transportation. sible by public transport) most of the people commute to work by car. In Tihange, almost everybody drives to work alone, however in Doel, over 25% of employees commute to work by either using collective transportation or in a car with others. If we compare this gure with gure 8 where carpooling percentages rarely exceed 5%, we can conclude that even outside of Brussels, the mobility management is indeed working. Notice however that for the nuclear plant of Tihange, more work is needed to motivate its employees to participate in the sustainable mobility culture. To investigate the cause of this signicant dierence between Doel and Tihange, we can take a closer look at the commuter trac on those daily routes.

Figure 16 shows the geographical distribution of all cars that travel to one of the nuclear plants on a daily basis. We can see that there is a very strong correlation (as can be expected) between the location of the nuclear plant and the location of residence of the employees that work there. It is clear from the gure that the distribution of employees is very similar for both nuclear plants. Now lets look at the same graph again, but only take 'green cars' into account, which corresponds to carpoolers and collective transportation.

Figure 17 shows the geographical distribution of daily 'green cars' that commute to either Doel (blue) or Tihange (red). We can see that here, in contrast to gure 16, almost nobody in Wallonia participates in carpooling or collective transportation. Figure 17 is therefore again (also see gure 7) a very clear indication that a lot more eort needs to be spent to motivate the employees living in the southern part of the country to participate in national and sustainable mobility projects.

36 Figure 16: Total number of daily cars that commute to Doel (blue) and Tihange (red) based on the location of residence of the employees.

Figure 17: Total number of daily 'green cars' that commute to Doel (blue) and Tihange (red) based on the location of residence of the employees.

37 4.6 Future

Finally, we asked the mobility managers to give their view on the near future in terms of mobility.

Two main categories of change are being investigated right now. The rst is the modernization of public transport. In particular for Brussels, a regional express net is being built that will oer fast public transportation connections and increased fre- quency within 30km of Brussels, covering a territory inhabited by 2,5 million people. The aim of the network is to provide access to public transport, from smaller cities surrounding Brussels, that travels directly into the center of the capital. This way, employees can access the public transport more easily and the time consuming and chaotic public transfers happen outside of the densest trac areas.

On the other hand, the Belgian government plans to increase the number of active electrical cars from the current 2500 up to about 5% (= 275 000) of the total amount of daily used vehicles by the year 2020. The mobility managers judged that this is a fairly ambitious challenge, however the outcome depends signicantly on the market segments that large industrial players like Tesla will target next: Technology is improving rapidly and as soon as a reasonably priced electrical car is marketed that has a good autonomy and an acceptable recharge rate, things could go very fast.

Furthermore, cars running on natural gas (which emits a signicantly lower amount of hydrocarbons) could also prove to be a viable solution in the future. Moreover, since Electrabel is a subsidiary company of the largest french natural gas supplier (GDF Suez) it is not hard to see why this alternative is being thoroughly considered. On top of that, some clever side projects are being investigated by Electrabel that could increase the benets of a partial switch to electrical vehicles even more. One of these projects aims to reduce the power-consumption peaks that occur daily between around 11:30 - 12:30 and 16:30 - 20:00 as can be seen in gure 4.6. Both peaks are the result of the simultaneous energy consumption from large industrial machinery combined with the electricity needs of individuals. Note that the two consumption peaks correspond to the average times where people eat their daily meals and when most industrial factories still have their machinery running, resulting in a combination of industrial-, oce- and domestic demand. During the winter months, the second peak also coincides with the fall of darkness, further increasing the peak consumption of electrical power.

38 Figure 18: Electricity usage in Belgium during an average day [33].

Because electrical power cannot be stored in large quantities, it has to be produced live, i.e. the moment it is being consumed. This oers a big challenge to the power supply companies because it means that they need to provide the peak capacity to produce this maximum electrical power by building extra plants that are only used briey every day (thus greatly reducing the protability of the company). On top of that, the power plants used to capture peak demands have to be very exible in their output power because the demand uctuates vividly, resulting in very costly infrastructure and high operating expenses.

Luckily, the recent introduction of electrical vehicles might provide an unexpected solution to this problem. The latest electrical car batteries are actually capable of storing a relatively large amount of electrical power, usually more than is necessary to complete the average daily commuter distance. This means that if a signicant percentage of all employees across Belgium would be driving electrical cars, the energy still remaining in the batteries of those vehicles upon arrival (at work or at home) could be uploaded onto the national electrical grid and used to 'feed' the power consumption peak. The depleted energy would then be recharged at night, when national power consumption is signicantly lower. Such a project is of considerable importance for an energy producing company such as Electrabel because the output power supply of the nuclear plants is very dicult to modify and is preferably left unchanged for long periods of time. Therefore, the electrical batteries of large numbers of electrical cars could provide a sustainable solution

39 for the uctuating energy demand. It is exactly because of this enlarged personal interest that large companies such as Electrabel are the perfect market activators that can pave the road towards an entirely new national mobility management that is both sustainable and more ecient. After all, the world record for most trac jammed city in the world is a title we would gladly pass on.

5 Carpooling, practical implementation

To get a nal view on the distribution of carpooling trac across Belgium we can take a look at gure 19. The blue dotted curve (axis on the right side) represents the fraction of carpooling commuters for all commuter route distances. We can see that for distances around 60km almost 10% of commuters resort to carpooling, while for very small distances (<20km) and very large distances (>80km) less then 5% of commuters carpool. Now, for specic entities that want to promote carpooling (e.g. governments or private carpooling platforms) it might be interesting to focus on a specic type of commuter trac, i.e. a specic commuter route distance to maximize the eciency of invested (nancial) goods. Such entities that wish to promote carpooling must be very careful when selecting a target commuter route distance, the blue dotted curve for example can lead to some very wrong conclusions. Therefore two other curves are included with their y-axis on the left. These curves represent the fraction of total driven kilometers per commuter route distance. The red curve represents 'car alone' driven kilometers from vehicles that only have one person in them, the green curve represents 'green' kilometers, combining the categories 'car with others' and 'collective'. In contrast to the blue dotted curve, these graphs clearly show that the majority of the total driven car kilometers in Belgium are caused by short commuter routes around 10 to 20 kilometers. We can thus see that even though the fraction of carpoolers is highest for commuter distances of about 60km, the total share of these commuter routes to the total number of (carpooled) car kilometers is very small. To get an idea of the ecological impact that can be achieved by promoting carpooling we can examine gure 20.

Firstly, lets explain how gure 20 was constructed. The x-axis displays the travel distance for all commuter routes. For the construction of this gure the commuter route distances were subdivided into 11 groups with a distance dierence of 5km within each group. The rst group thus contains routes that are between [0km,5km] in length, the second group between [5km,10km] and so on. For each group we can now compute the eect of convincing a certain percentage (displayed on the y-axis)

40 Figure 19: Fraction of green cars in function of daily commuter route distance. of those commuters to start carpooling and visualize the eect of these extra green kilometers on the total amount of driven 'green' car kilometers in Belgium (see color axis on the right). The entire gure was then interpolated between all the groups to visualize more details. We can see for example that by convincing just 10% of the commuters that drive between 10km and 15km to carpool on a daily basis, we can increase the total number of 'green' car kilometers in Belgium by as much as 7%! In contrast to this, convincing 15% of the commuters that drive more than 70km to carpool on a daily basis has almost no eect on the total number of 'green' car kilometers. The gure therefore provides a very clear indication that, in contrast to the intuitive belief that carpooling initiatives should focus on long distances to maximize their positive ecological impact, they should instead primarily focus on the commuter routes between 10km and 25km.

Note however that this conclusion is partly counteracted by some practical issues that complicate the eectiveness of carpooling for short commuter distances. Firstly, short commuter distances increase the fraction of overhead time that is necessary to pick up passengers because the travel times are rather low for these short distances. And secondly, the nancial drive for individuals in term of sharing fuel costs is a lot smaller for these small distances. Nonetheless, these short commuter distances should, according to our analysis, become the primary focus of new carpooling initiatives because they have by far the largest ecological and economical (in terms of trac congestion) impact on a national scale.

41 Figure 20: Result on total driven green kilometers of convincing a certain fraction of drivers to carpool.

6 Cooperation strategies

Electrabel has clearly taken a giant leap in the practical implementation of sustainable mobility on a company based level. We must note however that Electrabel has chosen not to involve any other large companies in their mobility strategy (such as Belgacom which has their headquarters located across the street from Electrabels building). The reason for this choice was, according to the managers, that sharing a mobility strategy with another large company involves the sharing of condential information about their employees (such as home address etc..) which could violate privacy regulations. This reason seemed rather unsatisfying to us because sharing a mobility strategy with another company does not necessarily result in privacy vio- lations when handled intelligently (we will return to this issue in the next section). A more probable explanation for this choice is that a company such as Electrabel is simply large enough to implement a successful mobility program without needing other companies to increase the number of participating employees to a level that is high enough for the solution to be sustainable over time. This immediately raises the question of how smaller companies can participate in the evolution towards a sustainable mobility solution in Belgium.

42 Small companies face two major challenges in their road towards implementing sustainable mobility solutions. Firstly, they do not have sucient numbers of employees to make possible solutions accessible and exible enough. Secondly, they often do not have the money that is needed to acquire expensive oce buildings located close to public transportation nodes simply because they are often out-competed by larger companies. Moreover, if we think of the mobility packages, smaller companies will have less bargaining power with the corresponding entities, resulting in less favorable deals for their employees. Finally, on the subject of carpooling it is obvious that small companies on their own do not have enough employees to provide a sucient number of exible carpooling alternatives because the few employees are most likely spread out over many geographical areas. The next section proposes a possible solution to these problems.

7 Anonymous carpooling platform

From the arguments given in the previous section it is obvious that small companies that want to contribute to a sustainable mobility solution in Belgium have an indispensable need for cooperation with other companies. Currently there is a large number of carpooling platforms in use by many dierent companies, however all of them have a very small number of users, leading to inecient commuter routes and large excess times to pick up co-faring passengers. Moreover, many of the carpooling platforms are used for commuter routes in dense urban areas that are plagued by trac congestion. This results in negative experiences of employees that start carpooling for the rst time and end up wasting hours of travel time in trac jams anyway. This eventually causes employees to switch to other alternatives, further decreasing the public image of carpooling and prolonging the ineective eort of small companies to promote sustainable mobility.

To make progress in the evolution towards a country that is no longer plagued by the most time-consuming trac congestions in the world, we think that a major government eort will be necessary to liberate carpooling from its unfavorable image in the commuter trac scene. To do this we propose that a government initiated platform should be introduced where small companies located outside dense urban areas can register their employees based on their postal code of residence. The platform then tries to match employees across dierent companies that have similar daily commuter routes (e.g. they live and work within the same two postal codes).

43 Three factors in this proposal are worth noting:

Firstly, we suggest that the system is implemented anonymously (using only postal codes and not displaying addresses) which could convince some companies that are very keen on their employees privacy. Only when employees from a specic commuter route are matched, they can obtain each others addresses through mutual agreement of the companies own mobility management system.

Secondly, we propose that the system is initially accessible only for companies located outside of geographical areas that are frequently aected by trac congestion. This way, employees that try carpooling for the rst time can enjoy a carpooling ride without the added stress of being stuck in a trac jam. We believe that this approach could greatly benet the general image that people have about carpooling by accentuating the social and ecological aspects of a carpooling ride rather than the personal, nancial or time based incentives.

Finally, a carpooling platform can only be successful once it has acquired a certain amount of users needed for exible operation. To give such a carpooling platform a reasonable chance of success we suggest that the platform is opened for registration a considerable amount of time before it is actually initiated for use. This way, the number of involved users can steadily grow (through media campaigns on public radio and television) without losing potential users at the very start when critical mass is not yet attained.

When such a platform turns out to be successful and succeeds in persuading a signif- icant number of people to recognize the benets of carpooling, a possible expansion towards denser urban areas can be considered to try and tackle the larger problem of trac congestion. We are however a very strong advocate of the approach that Electrabel has followed in its mobility strategy, emphasizing that it is the moral duty of large and successful companies to inuence government policies to encourage the development and construction of modern public transport in the dense urban areas of our small and cosy, yet trac congested country.

8 Conclusion

To conclude this work, we can state that the huge amount of trac congestion in Belgium is mainly caused by the low car occupancy rate, especially when considering commuter trac during peak hours. Even though considerable eorts are being

44 made by the government and large companies to promote the use of public transportation methods, about half of the working population still commutes to work alone by car. Moreover we have shown that the nancial benets currently linked to carpooling have failed to motivate a suciently large portion of the commuters to try out this more sustainable form of transportation.

To increase the pay o from governmental eorts we propose that a national carpooling platform is initiated where commuters working outside of trac congested areas can be anonymously matched with other participants. Furthermore, we recognize that the national mobility problem can not be solved by merely promoting carpooling, large governmental investments are also necessary to modernize the existing public transportation methods and increase the accessibility of the public transportation nodes from residential areas. Finally, we are of the opinion that many companies should seriously consider the necessity of having oce buildings located in dense urban areas. Population density is not likely to decrease in the near future, therefore if we want to pave the road towards a sustainable future and a growing economy, some further creative and innovative ideas might be necessary. After all, we think we can safely assume that nobody likes to be stuck in yet another trac jam.

References

[1] Wikipedia. Carpooling. http://en.wikipedia.org/wiki/Carpool, April 2015.

[2] Laurent Van Malderen Thomas Vanoutrive and Bart Jourquin. Mobility management measures by employers: Overview and exploratory analysis for belgium. EJTIR, pages pg 121141, June 2010.

[3] Department of Civil and Environmental Engineering. Rideshare history and statistics. http://ridesharechoices.scripts.mit.edu/home/histstats/, January 2009.

[4] Fréderic Mazella Robin Chase and Jean-Louis Jourdan. Carsharing. http://p2pfoundation.net/Carsharing, Februari 2015.

[5] Geschiedenis van de auto (1885-1904), 2015.

[6] Adam Hodges. Roping the wild jitney: The jitney bus craze and the rise of urban autobus systems. Planning Perspectives, 21(3):253276, 06 2006.

45 [7] Ross D. Eckert and George W. Hilton. The jitneys. Journal of Law and Eco- nomics, 15(2):293325, 10 1972.

[8] Cinzia Pastorello (European Environment Agency). Occupancy rates of passenger vehicles. June 2010.

[9] FOD Financien. Carpool, 2014.

[10] What price do i pay for carpooling? http://www.carpool.be/brussels/particulieren/info, 2015.

[11] Het Laatste Nieuws Redactie. Eerste veroordeling voor brusselse uber-chaueur: auto verbeurd verklaard. http://www.hln.be/hln/nl/957/Binnenland/article/detail/2308281/2015/05/04, May 2015.

[12] Taxistop + Carpool.be. Fiscale handleidiing voor de carpooler, 2015.

[13] FOD Financien. Fiscaal stelsel carpooling, 2015.

[14] FOD Financien. Forfait en werkelijke kosten.

[15] European Environment Agency. New cars co2 emissions well below europe 2015 target. http://www.eea.europa.eu/highlights/new-cars2019-co2-emissions-well, April 2015.

[16] Natuur en Energie Vlaanderen Departement Leefmilieu. Co2 uitstoot door wagens, August 2009.

[17] Jos Dings. How clean are europe's cars? an analysis of carmaker progress towards eu co2 targets in 2011, December 2012.

[18] VMM + Hoge Gezondheidsraad. Luchtvervuilende stoen en hun eect op de gezondheid, December 2013.

[19] Jim Gorzelany. The world's most trac-congested cities. http://www.forbes.com/sites/jimgorzelany/2013/04/25/the-worlds-most- trac-congested-cities/, April 2013.

[20] Inrix. Trac congestion scorecard. http://www.forbes.com/pictures/fk45meg/1- brussels-belgium/.

[21] INRIX. Inrix scorecard. http://inrix.com/scorecard/, July 2014.

46 [22] Lieven Desmet. Oeso geeft belgie leadvies. http://www.ademloos.be/nieuws/oeso-geeft-belgi2013.

[23] Stokols J. Stokols D., Novaco R. and Campell J. Trac congestion: Type a behavior and stress. Journal of applied Psychology, 63:pg 467480, 1978.

[24] Singer J. & Baum A. Schaeer M., Street S. Eects of control on the stress reaction by commuters. Journal of applied social Psychology, (18):944957, 1988.

[25] Kurt Meers. Files kosten ons 2 miljoen per werkdag. http://www.madeinlimburg.be/nieuws/les-kosten-ons-2-miljoen-per- werkdag/, March 2014.

[26] OECD. Oecd economic surveys: Belgium 2013. better use of infrastructures to reduce environmental and congestion costs. OECD publishing, 2013.

[27] Carpool incentive programs: Implementing commuter benets as one of the nation's best workplaces for commuters.

[28] Daniel Graziotin. An analysis of issues against the adoption of dynamic carpooling. 2010.

[29] Guilaume verhaeghe. Dynamic carpooling doesn't exist. blog.djump.in/post/58232155553/dynamic-carpooling-doesnt-exist, August 2014.

[30] BlaBlacar. Blablacar in een notendop. https://www.blablacar.nl/blog/over- ons, April 2015.

[31] Smove.be. Taxistop lanceerde carpool-app in samenwerking met vab. http://smove.be/aanbod/taxistop-lanceerde-carpool-app-samenwerking-met- vab.

[32] Carpool.be. Kost van carpoolen. http://www.carpool.be/vlaanderen/particulieren/kost/index.

[33] Elia. Belgian grid load data. http://www.elia.be/en/grid-data/data-download, 2014.

47 Attachments

48 FISCAAL STELSEL CARPOOLING – INKOMSTENJAAR 2014, AANSLAGJAAR 2015

WERKNEMER-PASSAGIER

EIGEN WAGEN VAN DE WERKNEMER-CHAUFFEUR - VERGOEDING WOON-WERKVERKEER BETAALD DOOR WERKGEVER FIRMAWAGEN TER BESCHIKKING GESTELD DOOR DE WERKGEVER - GEEN VERGOEDING WOON-WERKVERKEER

BETAALD DOOR WERKGEVER GEORGANISEERD GEMEENSCHAPPELIJK VERVOER (GGV)?

iue2:Fsa apoigsse nBelgium. in system carpooling Fiscal 21: Figure Voorwaarde: GEORGANISEERD GEMEENSCHAPPELIJK VERVOER (GGV)?

de werkgever moet de regels en de modaliteiten voor het gebruik van het gemeenschappelijk vervoer vastleggen en toezien op het gebruik ervan door de werknemers. Bovendien moeten de regels voor het gebruik van het georganiseerd gemeenschappelijk vervoer, evenals de voorwaarden Voorwaarde: van de vergoedingen, opgenomen zijn in een collectieve overeenkomst of in individuele schriftelijke overeenkomsten die zijn gesloten tussen de werkgever of de groep van werkgevers en elke werknemer; de werkgever moet de regels en de modaliteiten voor het gebruik van het gemeenschappelijk vervoer vastleggen en toezien op het gebruik ervan door de werknemers. Bovendien moeten de regels voor het gebruik van het georganiseerd gemeenschappelijk vervoer opgenomen zijn in een collectieve overeenkomst of in individuele schriftelijke overeenkomsten die zijn gesloten tussen de werkgever of de groep van werkgevers en elke werknemer;

JA NEE

JA NEE VRAAGT TOEPASSING VAN ZIJN WERKELIJKE VRAAGT TOEPASSINGNEE VAN ZIJN WERKELIJKE BEROEPSKOSTEN? BEROEPSKOSTEN?

JA NEE JA Het voordeel dat voor de werknemer-passagier Het voordeel dat voor de werknemer-passagier voortvloeit uit het NEE voortvloeit uit het gebruik van het GGV voor de woon- carpoolen met de werknemer-chauffeur die de beschikking heeft werkverplaatsingen moet worden aangemerkt als een over een firmawagen, maar dat niet georganiseerd wordt door de ste Werkelijke beroepskosten: vrijgesteld sociaal voordeel (art. 38, § 1, 1 lid, 11°, werkgever valt buiten het toepassingsgebied van de Werkelijke beroepskosten: WIB 92). inkomstenbelastingen en geeft derhalve geen aanleiding tot een 49 belastbaar feit. - forfait van 0,15 euro per Door de werkgever * - forfait van 0,15 euro per afgelegde kilometer m.b.t. toegekende * * * afgelegde kilometer m.b.t. de verplaatsing tussen de vergoeding als * * de verplaatsing tussen de woonplaats en de vaste Door de werkgever terugbetaling of woonplaats en de vaste plaats van tewerkstelling Werkelijke beroepskosten: betaalde vergoeding betaling van plaats van tewerkstelling zonder dat de in Werkelijke beroepskosten: voor GGV: reiskosten van de zonder dat de in aanmerking genomen - forfait van 0,15 euro per afgelegde kilometer m.b.t. VRIJSTELLING voor woonplaats naar de aanmerking genomen afstand tussen de de verplaatsing tussen de woonplaats en de vaste - forfait van 0,15 euro per afgelegde kilometer m.b.t. de een bedrag dat plaats van afstand tussen de woonplaats en de plaats plaats van tewerkstelling zonder dat de in aanmerking verplaatsing tussen de woonplaats en de vaste plaats van maximaal gelijk is tewerkstelling: woonplaats en de plaats van tewerkstelling hoger genomen afstand tussen de woonplaats en de plaats tewerkstelling zonder dat de in aanmerking genomen afstand aan de prijs van een VRIJSTELLING voor van tewerkstelling hoger dan 100 kilometer mag zijn van tewerkstelling hoger dan 100 kilometer mag zijn tussen de woonplaats en de plaats van tewerkstelling hoger dan treinabonnement een bedrag ten dan 100 kilometer mag zijn (art. 66bis, WIB 92). (art. 66bis, WIB 92). 100 kilometer mag zijn (art. 66bis, WIB 92). eerste klasse voor belope van (art. 66bis, WIB 92). die afstand (art. 38 , maximum 380 euro § 1, 1ste lid, 9°, b, OF voor AJ 2015 (art. OF OF ste OF WIB 92). 38 , § 1, 1 lid, 9°, - zijn werkelijke bijdragen* in c, WIB 92). - zijn werkelijke bijdragen* in de carpooling ten belope - zijn werkelijke bijdragen* in de carpooling ten belope van 75% - zijn werkelijke bijdragen* in de carpooling ten belope van 75% voor zover hij uiteraard het bewijs levert van voor zover hij uiteraard het bewijs levert van de daadwerkelijke de carpooling ten belope van 75% voor zover hij de daadwerkelijke betaling (art. 66, § 1, WIB 92). betaling (art. 66, § 1, WIB 92). van 75% voor zover hij uiteraard het bewijs levert uiteraard het bewijs levert van de daadwerkelijke van de daadwerkelijke betaling (art. 66, § 1, WIB betaling (art. 66, § 1, WIB 92). 92). * * * * * * Door de werkgever Door de werkgever betaalde toegekende vergoeding als *Deze bijdragen zijn bij de werknemer-chauffeur niet belastbaar wanneer de samengetelde bijdragen van de verschillende passagiers vergoeding voor GGV: terugbetaling of betaling van VOLLEDIG BELASTBAAR, dus reiskosten van de woonplaats betaald voor de door de werknemer-chauffeur afgelegde carpoolafstand niet meer bedragen dan de kilometervergoeding die de Staat GEEN VRIJSTELLING (art. 38 , naar de plaats van aan haar personeel toekent voor dienstverplaatsingen (0,3461 euro/km van 1.7.2013 tot 30.6.2014 - 0,3468 euro/km van 1.7.2014 tot § 1, 1ste lid, 9°, WIB 92). tewerkstelling: VOLLEDIG 30.6.2015). BELASTBAAR, dus GEEN VRIJSTELLING (art. 38 , § 1, 1ste lid, 9°, WIB 92).