Evaluating the impact of standards and fiscal policies on CO 2 emissions of cars in Europe. Abstract. Introduction

Transcription

1 Evaluating the impact o standards and iscal policies on CO 2 emissions o cars in Europe Amela Ajanovic, Vienna University o Technology, Ailiation, Vienna, Austria Reinhard Haas, Vienna University o Technology, Vienna, Austria Abstract The already applied measures such as voluntary agreements between car manuactures and energy labeling or a new cars have not shown suicient reduction o greenhouse gas emissions. More rigorous options are mandatory standards and taxes. Standards should reduce the average uel intensity o a car leet. Fiscal policy measures like taxes can orce car users to change behavior to reduce travel activity, to buy smaller cars with higher energy eiciency or to switch to other transport modes i possible. The core objectives o this paper are: (i) to provide surveys on standards and iscal policies or car passenger transport in EU countries; (ii) to provide a ormal ramework that links consumers decision making processes with technical and economic aspects o service demand (this ramework allows analysis o the impact o standards and iscal policies on energy consumption and CO 2 emissions); (iii) to identiy the crucial impact parameters in this ormal ramework; and (iv) to discuss and interpret the magnitude o these impact parameters on the eectiveness o policies. The major conclusions rom this analysis are: (i) The intended introduction o standards as announced by the EU will not lead to the theoretically possible energy savings and CO 2 -reduction due to the rebound eect; (ii) The standards and uel taxes are linked via the service price elasticity. The magnitude o the service price elasticity deines which instrument is more eective; (iii) A combined tax-standard policy could lead to a win-win situation rom which the environment beneits and car drivers are not hurt; (iv) Another inding is that the introduction o a registration tax is in principle the same as the introduction o a standard and leads to the same rebound problems. Introduction Road transport which is primarily based on ossil energy is ater power generation, the second biggest source o greenhouse gas emissions (GHG) in the EU. It contributes about one-ith o the EU's total emissions o carbon dioxide (CO 2 ). Emissions rom road transport were continuously increasing over the last twenty years (EU, 2014), see Figure 1. The largest part o the total GHG emissions o road transport is caused by passenger cars. Although uel eiciency o passenger cars has been signiicantly improved in the last decade, trends toward more powerul vehicles and additional services in cars have reduced impact o uel eiciency improvements. At the same time vehicle ownership level is continuously increasing in all EU countries. In the past the emission reduction policy in the EU was mostly based on CO 2 emissions standards and iscal policy measures. However, the applied policy measures - in the irst line voluntary agreements between car manuactures and energy labeling or a new cars - have not stopped increase in GHG emissions in road transport, see Figure 1. Since targets or the emission reduction were not achieved on time more rigorous options - mandatory standards and higher taxes have been implemented in the last years. Standards could reduce the average uel intensity o a car leet. Fiscal policy measures like taxes can encourage car users to change behavior to reduce travel activity, to buy smaller cars with higher energy eiciency or to switch to other transport modes i possible.

2 Figure 1. GHG emissions (1990=1) (Eurostat, 2013) The core objectives o this paper are: (i) to provide surveys on standards and iscal policies or car passenger transport in EU countries; (ii) to provide a ormal ramework that links consumers decision making processes with technical and economic aspects o service demand (this ramework allows analyze o the impact o standards and iscal policies on energy consumption and CO 2 emissions); (iii) to identiy the crucial impact parameters in this ormal ramework; and (iv) to discuss and interpret the magnitude o these impact parameters on the eectiveness o policies. Survey on currently implemented policies For the reduction o CO 2 emissions rom passenger cars a very broad portolio o policy measures have been implemented across EU countries. Currently mostly used policy measures are standards and iscal policies. Standards or CO 2 emissions o new cars are set on the EU level and are same or all Member States, iscal policy measures are determined at the national level, and they are very dierent across the EU. Mostly used iscal measures are uel tax, car registration tax and car annual ownership tax. In this paper special ocus is put on uel taxes, registration taxes and CO 2 standards. Fiscal policy measures Due to dierent national goals, implemented iscal policy measures are very dierent across the EU. The most common taxes are uel tax, Value Added Tax (VAT), registration taxes and annual ownership taxes. However, the criteria or these taxes are very dierent. Figure 2 shows composition o gasoline prices including taxes or dierent EU countries in In the shown countries the VAT is in the range rom 19% (Germany) to 25% (Sweden and Denmark), and total tax on gasoline is in the range rom 47% (Spain) to 61% (Finland). A similar situation is with diesel, but the tax on diesel is slightly lower, in range between 42% and 57%.

3 Figure 2. Composition o gasoline prices including taxes in 2013 (EEP, 2013) Dierent criteria or registration- and ownership tax in EU countries are shown in Table 1. A registration tax is tax paid once, by each vehicle owner, or each vehicle purchased and entered into service. It is usually based on CO 2 emissions, cylinder capacity, car price, uel consumption and power or weight o car. In the most o countries two dierent criteria are combined. However, there are also countries without registration taxes such as Germany, Sweden and United Kingdom. Annual ownership tax is paid annually, regardless o how oten the vehicle is used. In the case o passenger cars this tax is mostly based on CO 2 emissions, weight, cylinder capacity, power and uel consumption. Also this tax is not applied in all countries (e.g. France). Table 1. Criteria or a taxes on acquisition and ownership (data source: ACEA, 2013) Status: 1 January 2013 Registration tax based on: Ownership tax based on: Fuel consumption AT Fuel consumption DK Car price FI,NL Weight DK,FI,NL,SE CO 2 emissions FI,FR,NL,PT,ES CO 2 emissions FI,DE,GR,NL,PT,SE,UK Cylinder capacity GR,PT Cylinder capacity GR,PT,UK Kilowatt/weight/seats IT Kilowatt AT,IT None DE,SE,UK None FR Contrary to a uel tax which could have an impact on the short-term driving behavior (e.g. travel activity), registration tax can impact the choice o cars. Standards Important measures or the reduction o emissions in the EU are standards or CO 2 emissions rom new passenger cars. The EU Regulation (EC, 2009) on passenger cars is directly applicable in all Member States and does not need to be transposed into national law through national legal instruments. According to the Regulation average CO 2 emissions rom cars should not exceed 130

4 grams CO 2 per km by 2015 and should drop urther to 95g/km by The 130 grams target will be phased in between 2012 and 2015 (EU, 2014). The evaluation o CO 2 emissions rom new passenger cars by association is shown in Figure 3 as well as the commitments undertaken by the European (ACEA), Japanese (JAMA) and Korean (KAMA) car manuacturer associations related to the average new car emission targets. Figure 3. Evolution o CO 2 emissions rom new passenger cars by association The irst agreements with car manuactures have been voluntary agreements. Since the target o 140 gco 2 /km or 2008 was not met on time (the average or the whole car market or 2008 was g/km (Wikipedia, 2014), in 2009 the irst mandatory CO 2 emission standards or cars were adopted in the EU (de Wilde/Kroon, 2013). In practice this means that each manuacturer gets an individual annual target based on the average mass o all its new cars registered in the EU in a given year. Since only the leet average is regulated, manuacturers are able to produce cars with emissions above their indicative targets i these are oset by other cars which are below their indicative targets. Indicative emissions are established or each car according to its mass on the basis o the emissions limit value curve. This curve is set in a way that a leet average o 130 grams o CO 2 per km is achieved or the EU as a whole (EU, 2014). The limit value curve or the 2015 target is calculated using ollowing equation: CO2 _ SP M 130 ( M 0) (1) Where CO 2_SP are permitted speciic emissions, M is a mass o car in kg, M 0 is 1289 kg, and α is the slope o the LVC (0.0457). This curve is set in a way that emissions rom heavier cars have to be reduced more than those rom lighter cars. Since targets or 2015 and 2020 are mandatory, manuacturer will have to pay penalties i their average emission levels are above the target set by the limit value curve. The penalties will be based on the calculation o number o grams per kilometre (g/km) that an average vehicle registered by the manuacturer is above the target, multiplied by the number o cars registered by the manuacturer. A premium o 5 EUR per car registered will apply to the irst g/km above the target, 15 EUR or the second g/km, 25 EUR or the third g/km, and 95 EUR or each urther g/km. From 2019 every g/km o excess will cost 95 EUR (EU, 2014). Due to the implementation o mandatory CO 2 emissions standards, cars are expected to become more energy eicient to consume less uel per km driven.

5 Formal ramework To analyse the impact o various policies on overall CO 2 emissions we create in the ollowing a ormal ramework. This ramework is based on technical relations between CO 2 emissions, energy and service demand. It builds on works provided by Wirl (1992), Walker/Wirl (1993) and Haas (2009). It urthermore considers consumers decision making processes and their sensitivity to technical standards and taxes. How service is produced basic technical ramework The basic presumption is that consumers do not demand energy or technology per se, but energy services (S), in this case mobility vehicle kilometres driven (vkm). Total energy or mobility used, which is mostly dependent on uel intensity (FI) o cars and number o kilometres driven per year (vkm), can be written as: E vkm FI (2) Emissions rom passenger car depend in principle on total energy used in cars as well as on the type o energy used. Dierent uels (e.g. gasoline, diesel, biouels, etc.) have dierent speciic CO 2 emissions coeicient. The average speciic CO 2 emissions coeicient ( CO2 ) can be improved with better quality o ossil uels, higher share o biouels with better ecological perormance, more electricity rom renewable energy sources, etc. CO CO 2 emissions (CO 2 ) rom passenger cars can be described as: E vkm FI vkm CO (3) 2 CO CO2 2 with CO 2.total CO 2 emissions [ton CO 2 /yr] CO2.CO 2 emission actor o uel [kg CO 2 /litre] FI...uel intensity [litre/100 km] CO 2_SP speciic CO 2 emissions [kg CO 2 /km] Consumers decision-making model 2 _ SP To ind out how policies work we have to know how consumers make decisions. In principle consumers try to maximize the beneits they enjoy rom the consumption o a speciic service or a good and to minimize the monetary and other eorts related to the consumption o this good or service, see Wirl (1992) and Walker/Wirl (1993). I we look at the service vkm consumers will maximize the utility they obtain rom consuming this service minus the disutilities, (see Wirl (1992)). Disutilities in this context are uel costs (P E), the capital costs (ρ I(η)), transaction costs and the possible damage or the environment. In the ollowing we neglect the last two parts and obtain: max u( vkm) p E( ) I ( ) (4) E, P uel price including tax ρ..annuity actor I(η) Investment costs The irst-order condition with respect to energy leads to:

6 ' u P u P FI E ( T) (5) In principle, better eiciency η(t) lowers the marginal costs o a particular service and thus raises the demand or this service. For example, a more uel eicient car implies that one can aord higher mobility. Figure 4 provides a graphical representation o this eect or two dierent levels o marginal service costs P S. Figure 4. Choice o service level or vkm driven or dierent uel intensities o a car (based on Wirl 1992) The two marginal cost curves correspond to two dierent equipments, FI 1 <FI 0, where the technology 1 is more eicient. The downward sloping curve represents marginal willingness to pay or the service, u (vkm). The optimum is given where marginal beneits equal marginal costs. Thus, the cars with FI 0 implies the service demand vkm 0 while the more eicient cars with FI 1 implies a higher service level vkm 1. Consumer decision can be aected with policy measures, e.g. uel tax (τ ), eiciency standards (η ) and registration tax (τ R ). Including this policies in Eq. (4) we obtain: maxu( vkm) ( p ) E( ) ( I( ) ) (6) E, R Modelling the demand or service Based on the above derived model or consumer decision making we summarize that consumers strive or service demand, and service demand depends on energy service price (Ps), income (Y), investment costs (IC) and quality (q) rom other possible attributes (x). However, in the short-term service demand in this case vkm - can be written as: vkm ( P, Y) (7) With P s P FI S This service price can be changed with policies: - Rising the uel tax τ which increases P and consequently Ps or - Reducing uel intensity by a standard (FI ) which decreases Ps (8)

7 The level o short-term service demand with respect to kilometre driven depends on available income (Y) 1 and the service price o a km driven (P S ). In other words we assume that once a car is purchased there are no other attributes or driving more or less than lower or higher energy service prices. The impacts on vkm driven is analysed in Ajanovic/Haas (2012) by applying a cointegration approach to the ollowing model: Where: C vkm Ps Y vkm C P (9) sy S Intercept Demand or service, vehicle km driven in year t in a country Weighted average price o service vkm driven (calculated by means o weighted uel prices) Real private inal consumption expenditures The most interesting numbers o this analysis are the energy service price elasticities α S because they contain inormation or both - price and eiciency impact. It is the crucial parameter or our urther analysis. Empirical analyses o the magnitude o price elasticities have been conducted in many papers, e.g. Dahl (2012), Dargay (1993) and Goodwin et al. (2004). Investigations conducted by the authors o this paper see e.g. Ajanovic/Haas 2010, 2011 and 2012 have resulted in long-term service price elasticities o about (-0.4) to (-0.45). Related to the above relections this leads to the ollowing interpretations: the increase in uel prices due to a uel tax o 1% would result in energy savings o about 0.4% to 0.45%. With a standard which decreases CO 2 emissions by 1% the savings would be between 0.55% and 0.6%. About the same eect would result rom a CO 2 -dependent registration tax. Interaction o policies The principle o how changes in eiciency due to standards and in prices due to taxes aect energy consumption is depicted in Figure 5. For a uel tax the reduction in energy consumption ΔE results rom higher service price Ps τ remaining on the same curve η 0. When a standard is implemented we switch rom η 0 to η 1 leading to a reduction ΔE o energy consumption. However, due to a lower service price Ps η this saving eect is lower than the theoretical eect which is ΔE η. Figure 5. How a tax vs a standard works 1 Note, that urther on in the empirical analyses o this paper we use Private Consumption Expenditures (PCE) as a proxy or income

8 In this section we analyze the impacts o changes in uel intensity due to standards vs changes in uel prices due to taxes on energy consumption. This is important to derive conclusions with respect to the eect o the implementation o standards or uel intensity vs the eect o the introduction o uel taxes leading to increasing uel prices. The rebound eect One o the most critically discussed issues with respect to the implementation o standards or uel intensity or corresponding CO 2 emissions is the rebound eect. The service cost savings or car users usually lead to the change in driving behavior. The behavioral response to the introduction o a new more eicient technology or other measures implemented to reduce energy use is called direct rebound eect. The basic principle o a rebound eect is shown in Figure 6. Point 1 shows the initial situation (E 1 energy consumption, 1 uel eiciency, S 1 service, in this case vehicle kilometer driven vkm 1 ). With the increasing energy eiciency rom 1 to 2 theoretically energy consumption could be reduced rom E 1 to E 2 th. Due to the higher eiciency, service price is lower, which causes increase in service demand. Due to the increase in energy eiciency rom 1 to 2 and the resulting rebound eect energy consumption will be reduced to E 2 pr instead to E 2 th (Ajanovic/Haas, 2010a). Figure 6. The rebound eect (Ajanovic/Haas, 2010a) Using the deinition o the service price elasticity, the dierence in vkm driven caused by the rebound eect is calculated as: ( P FI) vkmreb vkm, P vkm s 1 (10) P FI 1 Where α vkm,ps is the elasticity o vehicle kilometres driven with respect to service price P S. Using previous equations and the undamental deinition described in Greene (1997), the elasticity o energy consumption with respect to a change in uel intensity is derived, (or detail see Ajanovic/Haas, 2012):, 1 (11) E FI vkm, P S From the Eq. 11 it can be seen that the elasticity o energy consumption with respect to a change in uel intensity (γ E,FI ) is one plus the elasticity o energy service (in our case vkm) with respect to service price (Ps).

9 Figure 7 depicts the eect o a uel tax versus a standard depending on the service price elasticity. For example, i a tax in the magnitude o 100% is introduced and the price elasticity is (- 0.3) then the energy saving eect is 30%. I a standard in the magnitude o 100% is introduced and the price elasticity is e.g. (-0.3) then the energy saving eect is 70% and the rebound eect due to more km driven is 30%. Figure 7. Eect o a tax versus standard depending on service price elasticity A CO 2 emission standard should lead to increasing uel eiciency o new cars. This could reduce energy service price or car drivers and lead to change o the behaviour, e.g. cars are used more requently and/or on longer distances. A irst major conclusion o this paper is that the preerence or taxes versus standards depends solely on the magnitude o the service price elasticity α vkm. I this α S is small (e.g. -0.1) the uel tax eect is almost neglectable. In this case a standard is clearly preerable with an empirical saving eect o about 90% o theoretically calculated 100%. I at other hands α vkm is high it is vice versa and the tax is preerable. As mentioned beore α vkm is in the range o -0.4 to This leads to an ambiguous situation and it is likely that a combined introduction o standards and uel taxes depending on the service price elasticity will lead to the most beneicial solution or society, see Figure 8. Figure 8. How taxes and standards interact and how a win-win situation is derived or society

10 How registration taxes work As shown in Table 1 in some EU countries registration tax is based on the speciic CO 2 emissions (CO 2_SP ) o a car. The relation between speciic CO 2 emissions and uel intensity is given in Eq. (3). Corresponding to this equation, speciic CO 2 emissions set by standards ( ) are: _ where FI is to the standard related uel intensity. A CO 2 based registration tax leads to purchase o cars with lower speciic CO 2 emissions per km driven. Figure 9 depicts the relation between a registration tax based on CO 2 emissions and the speciic CO 2 emissions o cars chosen. The higher the registration tax is the lower are the speciic CO 2 emissions o the average sold cars. For every required emission standard (CO 2_SP ) a corresponding registration tax (τ R_CO2 ) could be implemented to meet this standard and vice versa, see example in Figure 9. _ _ (13) Figure 9. Relation between a registration tax (τ R_CO2 ) and the speciic CO 2 emissions (CO 2_SP ) o cars It should be noted that the relationship between a registration tax and the speciic CO 2 emissions will also depend on the elasticity o investment costs (γ). A high registration tax will lead to the purchase o cars with lower speciic CO 2 emissions. Cars with lower speciic CO 2 emissions have lower uel intensity. I we introduce speciic CO 2 emissions the corresponding uel intensity (FI ) will aect service price Ps: P S ( P ) FI Where FI is a unction o speciic CO 2 emissions: FI =(CO 2_SP ) Since FI =(CO 2_SP ) and CO 2_SP =(τ R_CO2 ) this leads to the ollowing relections: rom Eq. (13) ollows P (, ) (15) S P S R _ CO 2 This means that the service price does not just depend on the uel tax τ but also rom registration tax τ R_CO2. Finally, rom Eq. (7) we obtain: vkm ( P S ) (, ) (16) R _ CO 2 (14)

11 From this equation is clear that a registration tax based on CO 2 emissions will also have an impact on service demand - vehicle kilometer driven. In principle a CO 2 based registration tax works like a standard and leads to the same eect. With a higher CO 2 based registration tax demand or cheaper and more energy eicient vehicles rise. Due to the better eiciency o cars, service price decreases and leads to a direct rebound eect caused by more km driven, see Figure 10. Figure 10. Relation between speciic CO 2 emissions and the vehicle km driven Conclusions The major conclusions rom this analysis are: 1. The intended introduction o standards as announced by the EU will not lead to the theoretically possible energy savings and CO 2 -reduction. The major reason is that the rebound eect due to more km driven and larger cars will eat up a part o the theoretically calculated savings. 2. Despite the act that there are a numerous parameters that constitute a ormal ramework, to explain energy consumption and CO 2 emissions, there is only one important parameter which inluences the inal CO 2 emissions and this is the service price elasticity. 3. In addition standards and uel taxes are linked via the service price elasticity and are not independent! The magnitude o the service price elasticity deines which instrument is more eective the tax or standard. Another inding is that the introduction o a registration tax is in principle the same as the introduction o a standard and leads to the same rebound problems. The registration taxes may cover a speciic part o the standards component. 4. Our results or the service price elasticity are about -0.4 to This result leads to the situation that a mix o policies is preerable. A simultaneously introduced tax will compensate or the rebound eect without hurting car drivers due to service vkm driven remaining at same service price. A combined tax-standard policy will lead to a win-win situation or the environment and car drivers. Figure 8 describes how the policy designs should look like in principle. The uel tax should compensate the standard to an extent so that inally the service price beore and ater policy introduction remains the same. Reerences ACEA Tax Guide Ajanovic A., R. Haas, The role o eiciency improvements vs. price eects or modeling passenger car transport demand and energy demand Lessons rom European countries. Energy Policy

12 Ajanovic A, R. Haas On the eectiveness o standards vs. taxes or reducing CO2 emissions in passenger car transport Europe. Proceedings, Risø International Energy Conerence (RIEC). Ajanovic A., R. Haas The impact o iscal policies on passenger car transport energy demand in OECD countries. Proceedings, 29 th USAEE/IAEE Annual North American Conerence in Calgary. Canada. Ajanovic A., R. Haas. 2010a. On extracting the impact o consumer behaviour and the rebound eect in passenger car transport in selected EU countries. Proceedings,, 8 th BIEE Academic Conerence, Oxord. Dahl C.A Measuring global gasoline and diesel price and income elasticities. Energy Policy 41. Dargay, J.M., Are price and income elasticities o demand constant? The UK Experience EE16. Oxord Institute or Energy Studies. EC European Commission. Regulation (EC) No 443/2009 o the European Parliament and o the Council o 23 April 2009 setting emission perormance standards or new passenger cars as part o the Community s integrated approach to reduce CO2 emissions rom light-duty vehicles. June EEP Europe s Energy Portal. accessed on EU website on Climate action in the transport sector: accessed on Eurostat accessed on Goodwin, P.B., Dargay, J.M., Hanly, M., Elasticities o road traic and uel consumption with respect to price and income: a review. Transport Reviews 24 (3), Greene D.L Theory and Empirical Estimates o the Rebound Eect or the U.S. Transportation Sector. ORNL. Haas, R., Nakicenovic, N., Ajanovic, A., Faber, T., Kranzl, L., Mueller, A., Resch, G., Towards sustainability o energy systems: a primer on how to apply the concept o energy services to identiy necessary trends and policies. Energy Policy 36, Walker, I.O., Wirl, F., Irreversible price-induced eiciency improvements: theory and empirical application to road transportation. The Energy Journal 14 (4), Wikipedia accessed on de Wilde H.P.J., P. Kroon Policy options to reduce passenger cars CO2 emissions ater 2020 ECN-E Wirl F The impact o conservation, technology and the new registration tax on automobiles on energy consumption. Empirica Austrian economic paper. Vol.19, No. 1.