Fuel consumption of European cars: The effect of standards, taxation and safety

1
Fuel consumption of European cars The effect of
standards, taxation and safety

• Theodoros Zachariadis
• Economics Research Centre, University of Cyprus
• COST 355 meeting, Madrid, May 2007

2
Contents

• The effect of standards on fuel economy
  (Clerides and Zachariadis, Are standards
  effective in improving automobile fuel economy?,
  July 2006)
• Some recent results, trying also to explain the
  share of diesel cars in each country
• Do vehicle safety requirements compromise fuel
  economy?

3
Rationale of the study on fuel economy standards

• Share of transportation in energy use and GHG
  emissions steadily rising
• It will take time for biofuels and new
  technologies (hybrids, fuel cells etc.) to be
  effective
• ? Improve fuel economy of conventional
  engines/fuels
• FE improvements may be attained through
• Higher fuel prices
• FE standards / industry voluntary commitments
• CO2-based vehicle taxation
• Autonomous technical progress
• ? How much improvement from which measure?

4
Previous similar work

• Espey (Energy Economics, 1996) Johansson
  Schipper (J. Transp. Econ. Policy, 1997)
• Greene (Energy Journal, 1990)
• Gately (Energy Journal, 1992)
• Small Van Dender (UC Irvine, 2005)
• What is new in our study
• 18 countries, 20 cross-sections, period
  1975-2004
• Period with without FE standards, with high
  low fuel prices
• FE standard is explicitly addressed as a variable

5
New-car fuel consumption and standardsin the US
and EU, 1975-2004
6
Model

• FC average sales-weighted fuel consumption of
  new cars (l/100 km)
• ? autoregressive coefficient of dependent
  variable
• a1 time trend (autonomous technical progress
  etc.)
• p gasoline price (Euros1995 per liter)
• L maximum lag length L5
• STD fuel economy standard expressed in l/100 km
• INC real GDP per capita (Euros1995)

7
Data (sample size 384)
8
Regression results
Notes Estimation carried out with the
Arellano-Bond GMM procedure. Robust t-statistics
in brackets. , and denote significance at
10, 5 and 1 level. Last column reports the
probability of the Arellano-Bond test for second
order serial correlation of residuals.
9
Policy implications 1

• Are FE standards significant for reducing
  automobile fuel consumption?
• Use data from AT, BE, FR, DE, IT, JP, SE and UK
• Split data in two periods pre-standard
  (1980-1994) and with standards (1995-2004)
• Re-estimate model without STD variable i) for
  pre-standard period ii) for entire period
• Perform a Wald test and a Chow test to examine
  stability of estimated coefficients
• Both tests reject the null of coefficient
  stability
• ? structural break, i.e. FE regulations made a
  difference

10
New-car fuel consumption in Europe and Japan,
1980-2003
11
Policy implications 2

• Given a future FE (or CO2) target to be met
  without tighter standards, how much should prices
  increase?
• In the US, tightening current CAFE standard by
  10 is equivalent to raising gasoline price by 36
  US cents2004 / gallon (result is similar with
  those of other studies)
• In Europe, stated policy target of 120 g CO2/km
  25 tighter standard retail fuel prices might
  have to double to induce similar fuel savings

12
Policy implications 3

• How might fuel consumption evolve without further
  standards and at todays fuel prices?
• Time trend coefficient a1 insignificant, near
  zero
• i.e. no autonomous improvement per year ?
• Changing consumer preferences towards more
  powerful and comfortable cars have cancelled out
  any autonomous technical progress
• European long-term models, assuming that FE will
  continue to improve at fast rates even without
  post-2010 FE regulations, may have to be revisited

13
Policy implications 4

• Are taxes always the most efficient measure?
• To tackle an externality, impose a tax and let
  the market work
• But
• Taxes less effective because of consumer myopia
• Impact of higher taxes on the whole
  economy?(e.g. sectors that use fuel as an
  intermediate good)
• Political acceptance of higher taxes
• Major externalities (accidents, congestion)
  associated with miles driven, not with fuel
  consumed

14
Conclusions of the study on FE standards

• If there were no standards in force, car fuel
  economy would not have improved considerably
• Very high fuel price increases required in Europe
  if fuel economy to be improved without standards
• Absent technological breakthroughs or an economic
  recession, FE will only improve further with
  tighter standards
• Raising fuel taxes is not an option for Europe,
  could be considered in the US together with
  stricter standards (modified CAFE rules)

15
Recent extensions

• Focus on European countries only
• Fuel consumption may also depend on
• total vehicle taxes (registration, circulation,
  insurance etc.)
• urbanisation and population density
• ratio of retail gasoline/diesel price
• Except for gasoline/diesel ratio, other variables
  not available as a time series but only as a
  country-specific figure for a given year (i.e.
  fixed effect)
• Efficient estimation of dynamic panel models
  wipes out fixed effects, therefore adding these
  as explanatory variables is not possible
• Feedback requested are national data on vehicle
  taxation available for several years?

16
Effect of gasoline/diesel price ratio

• Price ratio was constructed from retail fuel
  prices (source IEA)
• To avoid endogeneity/collinearity
• Gasoline price is the average of the previous
  three years
• Gasoline/diesel ratio is the current years price
  ratio
• Using both price variables improves estimation

17
Is there a safety fuel economy trade-off?

• Car manufacturers dont respect their CO2
  commitment legislation to cut CO2 emissions
  from cars to come soon
• EU Environment Commissioner, 03/11/2006
• Decrease in CO2 emissions has recently slowed.
  This is due to strong customer demand for larger
  and safer vehicles and disappointing consumer
  acceptance of extremely fuel-efficient cars
• European car industry (ACEA), 05/11/2006
• Better car safety does not jeopardise emission
  reduction the added weight due to safety
  interventions is negligible
• European Transport Safety Council, 13/11/2006

18
Safety vs. fuel economy

• Two questions
• Does safety affect vehicle mass?
• Does safety affect fuel consumption / CO2
  emissions?
• US studies analyse relationship between traffic
  fatalities and attributes of vehicles involved in
  accidents see Ahmad and Greene, Transp. Res.
  Record 1941(2005) 1-7
• Earlier results showed that lower fuel
  consumption leads to less safety ? more
  fatalities
• Recent evidence is inconclusive

19
Safety vs. fuel economy Empirical analysis

• Car safety data obtained from EuroNCAP website
  for 193 cars of model years 2000-2007(www.euronca
  p.com)
• EuroNCAP provides consumers with independent
  information about a cars safety
• Ratings for three tests are provided Adult
  occupant test, pedestrian test, child protection
  test
• Score is provided in integer numbers (e.g. 0-30)
  and then codified in stars (excellent is 5
  stars for adult children tests, 4 stars for
  pedestrian test)
• For each model tested, EuroNCAP provides exact
  model description (e.g. Peugeot 207cc, 1.6 sport
  1), kerb weight and model year

20
Safety vs. fuel economy Empirical analysis (2)

• For each one of the 193 EuroNCAP car models, fuel
  consumption CO2 data were retrieved from the
  2001-2006 databases of the German Federal Motor
  Transport Authority (KBA) (purchased on CD-ROMs)
• Data for 2007 models were obtained from online
  databases of the UK Vehicle Certification Agency
  (VCA) (www.vcacarfueldata.org.uk) of portal
  www.carpages.co.uk
• Linear regressions
• massi f(safetyi, engine_sizei, dsl_dummy,
  year_dummy)
• CO2i f(safetyi, engine_sizei, dsl_dummy,
  year_dummy)

21
Results (1) Safety effect on vehicle mass is
very small
22
Results (2) Safety effect on CO2 is marginally
significant, small and negative!
23
Safety vs. fuel economy tentative conclusion

• Better car safety does not jeopardise emission
  reduction the added weight due to safety
  interventions is negligible
• European Transport Safety Council, 13/11/2006
• ETSC is probably right !
• Results are similar if we observe subsets of the
  whole sample (e.g. if we exclude SUVs and/or
  superminis, observe family cars and/or MPVs only)
• Results are similar if safety variable includes
  both adultpedestrian test ratings
• Results are consistent with Ahmad and Greene
  (2005) who used fatalities as dependent variable
• Please comment!