Transcript The Saliency of the BC Carbon tax

PRICE AND CARBON TAX EFFECTS ON
GASOLINE AND DIESEL DEMAND
By
Jean-Thomas Bernard1
Grant Guenther2
and
Maral Kichian3
October 23, 2015
1.
Economics Department, University of Ottawa, [email protected]
2.
Transport Canada, [email protected]
3.
Economics Department and Graduate School of Public and
International Affairs, University of Ottawa, [email protected]
ORDER OF THE PRESENTATION
o
Introduction
o
The B.C. Carbon Tax
o
Literature Review
o
Data
o
Models and Estimation Results
o
Conclusion
INTRODUCTION
Research Objectives:
o
Do consumers and firms respond in the same
way to carbon taxes ?
o
Are the effects of changes in the different
components that make up total energy prices
the same ? That is, do tax effects differ from
pure price effects ?
o
Also, do excise taxes have the same effects as
carbon taxes ?
We use the data related to the introduction of the
B.C. carbon tax in 2008 to test this hypothesis.
THE B.C. CARBON TAX
o
Broad Base
o
Phase in:
o
$10/tonne of CO2 on July 1st, 2008.
o
$5 added on July 1st every year until 2012 to reach
$30.
o
This is equivalent to a tax of 6.67¢/litre of gasoline
and 7.68¢/ litre of diesel.
o
Revenue Neutral.
LITERATURE REVIEW
o
o
Traditional Specification: no difference between oil price
and tax effects.
Dahl (2012)
Hughes, Knittel and Sperling (2008)
Price and tax effects on gasoline demand.
Davis and Kilian (2011)
Li, Linn and Muehlegger (AEJ: EP, 2014)
Rivers and Shaufele (JEEM, 2015)
Coglianese, Davis, Kilian and Stock (NBER, 2015)
DATA
Monthly domestic sales of gasoline in litre
Statistics Canada, The Supply and
Disposition of Refined Petroleum
Products in Canada, Catalogue no.
45-004-X.
Monthly domestic sales of diesel in litre
ibid.
Monthly regular gasoline average final price and total taxes in
Vancouver, (¢/litre)
Natural Resources Canada,
http://www2.nrcan.gc.ca/eneene/
sources/pripri/prices_bycity_e.cfm
Monthly diesel average final price and total taxes in Vancouver,
(¢/litre)
ibid.
C.P.I. index (2007=100.0)
Statistics Canada. Table 326-0020 Consumer Price Index (CPI), 2011
basket, monthly (2002=100).
Real Gross Domestic Product at base price in 2007 $
Quarterly data provided by
Conference Board of Canada.
B.C. monthly population
Statistics Canada. Table 051-0005 Estimates of population, Canada,
provinces and territories, quarterly
(persons).
Monthly Rainfall at Vancouver International Airport in mm
http://climate.weather.gc.ca
Real effective exchange rate monthly average, C.P.I. based
(2010=100.0)
Bank of Industrial Settlement
(BIS).
Gasoline sales
Diesel sales
Jan-13
Jul-12
Jan-12
Jul-11
Jan-11
Jul-10
Jan-10
Jul-09
Jan-09
Jul-08
Jan-08
Jul-07
Jan-07
Jul-06
Jan-06
Jul-05
Jan-05
Jul-04
Jan-04
Jul-03
Jan-03
Jul-02
Jan-02
Jul-01
Jan-01
Jul-00
Jan-00
Jul-99
Jan-99
Jul-98
Jan-98
Jul-97
Jan-97
Monthly Gasoline and Diesel Sales
(litres per capita )
120.00
110.00
100.00
90.00
80.00
70.00
60.00
50.00
40.00
30.00
Real Price net of carbon & excise tax
Real excise tax
Real carbon tax
Jan-13
Jul-12
Jan-12
Jul-11
Jan-11
Jul-10
Jan-10
Jul-09
Jan-09
Jul-08
Jan-08
Jul-07
Jan-07
Jul-06
Jan-06
Jul-05
Jan-05
Jul-04
Jan-04
Jul-03
Jan-03
Jul-02
Jan-02
Jul-01
Jan-01
Jul-00
Jan-00
Jul-99
Jan-99
Jul-98
Jan-98
Jul-97
Jan-97
Real Price of Regular Gasoline
(2007 cents/litre)
120.00
100.00
80.00
60.00
40.00
20.00
0.00
Jan-13
Jul-12
Jan-12
Jul-11
Jan-11
Jul-10
Jan-10
Jul-09
Jan-09
Jul-08
Jan-08
Jul-07
Jan-07
Jul-06
Jan-06
Jul-05
Jan-05
Jul-04
Jan-04
Jul-03
Jan-03
Jul-02
Jan-02
Jul-01
Jan-01
Jul-00
Jan-00
Jul-99
Jan-99
Jul-98
Jan-98
Jul-97
Jan-97
Real GDP per capita
(2007$ in thousand)
45.00
40.00
35.00
30.00
25.00
MODELS AND ESTIMATION RESULTS
Modelling Concerns (1):
July 2008:
Three effects coincide (US recession, seasonality, carbon tax)
July 2009, 2010, 2011, 2012:
Two effects coincide (seasonality, carbon tax)
• In particular, the seasonal effect is very important
• Need to make sure it does not mask the carbon tax effect in the model
- We assume the July seasonal effect is the same over the whole sample
- We regress sales series on monthly dummies over 1997:1-2007:12
- We generate a new series (sales_xjuly) by subtracting the July
coefficient from the sales series over 2008:1-2014:12
- Sales_xjuly already incorporates the July effect, over 2nd subsample
- We use this series as the dependent variable in our models
MODELS AND ESTIMATION RESULTS
Modelling Concerns (2):
Final Energy Prices include up to four components:
- pure price effects
- excise tax effects
- carbon tax effects (whenever applicable)
- other tax effects (e.g., GST)
• Evidence suggests that price effects (inclusive of excise taxes) are
different than carbon tax effects; see Rivers and Schaufele (2015)
• But price effects in these studies include excise taxes !
• Therefore, we separate total energy prices into net prices, excise taxes,
carbon taxes
• We allow these three components to have possibly different impacts
MODEL: GASOLINE
∆𝑔𝑡 =
12
𝑗=1 𝑎𝑗 𝑚𝑗
+ 𝛼𝑝 ∆𝑝𝑡 + 𝛼𝑒 ∆𝑒𝑡 + 𝛼𝑐 ∆𝑐𝑡 + 𝛽𝑦 ∆𝑦𝑡
+𝛽𝑟 ∆𝑟𝑡 +
-
𝜌𝑖 𝑔𝑡−𝑖 + 𝑣𝑡
𝑔𝑡 is log of real per capita gasoline sales
𝑝𝑡 is log of real net prices (net of excise and carbon taxes)
𝑒𝑡 is log of real excise taxes
𝑐𝑡 is log of real carbon taxes
𝑦𝑡 is log of real per capita Canadian GDP
𝑚𝑗 are monthly dummies
𝑟𝑡 is the log of rainfall
ESTIMATION RESULTS: GASOLINE
Variable
Coefficient
Std. dev.
P-value
Net price
-0.057
0.025
0.0231
Excise tax
-0.614
0.177
0.0006
Carbon tax
-0.176
0.035
0.0000
Canadian GDP
1.266
0.595
0.0347
Rainfall
-0.003
0.002
0.1906
Sales Lag 1
-0.598
0.061
0.0000
Sales Lag 2
-0.347
0.060
0.0347
Adjusted 𝑅2
0.80
Durbin-Watson
2.12
LM serial correlation 𝜒 2 test (4 lags) P-value
0.1404
ARCH(1) test 𝜒 2 test P-value
0.1191
ARCH(4) test 𝜒 2 test P-value
0.6238
WALD TEST RESULTS: GASOLINE
Null Hypothesis
𝜒 2 test P-value
Net Price = Excise Tax
0.0015
Net Price = Carbon Tax
0.0056
Excise Tax = Carbon Tax
0.0148
For Gasoline:
- Net price effects are significantly different than tax effects
- Excise tax effects are significantly different than carbon tax effects
- All three effects influence gasoline sales
- Excise taxes have the biggest effect, followed by carbon taxes, then by
net price changes
MODEL: DIESEL
∆𝑑𝑡 =
12
𝑗=1 𝑎𝑗 𝑚𝑗
+ 𝛼𝑝 ∆𝑝𝑡 + 𝛼𝑒 ∆𝑒𝑡 + 𝛼𝑐0 ∆𝑐𝑡 + 𝛼𝑐1 ∆𝑐𝑡+1
+ 𝛽𝑦 ∆𝑦𝑡 + 𝛽𝑌0 ∆𝑌𝑡 + 𝛽𝑌1 ∆𝑌𝑡−1 +
-
2
𝑘=1 𝜌𝑘
𝑑𝑢𝑚𝑘𝑡 + 𝑣𝑡 + θ𝑣𝑡−1
𝑑𝑡 is log of real per capita diesel sales
𝑝𝑡 is log of real net prices (net of excise and carbon taxes)
𝑒𝑡 is log of real excise taxes
𝑐𝑡 is log of real carbon taxes
𝑦𝑡 is log of real per capita Canadian GDP
𝑌𝑡 is log of real per capita US industrial production (manufacturing)
𝑚𝑗 are monthly dummies
𝑑𝑢𝑚𝑘𝑡 is a dummy for the recession period (2008Q3; 2009Q1)
ESTIMATION RESULTS: DIESEL
Variable
Coefficient
Std. dev.
P-value
Net price
-0.011
0.055
0.8419
Excise tax
0.0354
0.080
0.6567
Carbon tax
-0.125
0.060
0.0396
Carbon tax lead 1
0.125
0.058
0.0322
Canadian GDP
2.282
0.917
0.0138
US GDP
-2.572
0.643
0.0001
US GDP Lag 1
1.434
0.626
0.0227
Dummy 1
-0.380
0.058
0.0000
Dummy 2
0.354
0.075
0.0000
Adjusted 𝑅2
0.74
Durbin-Watson
2.03
LM serial correlation 𝜒 2 test (4 lags) P-value
0.6238
ARCH(1) test 𝜒 2 test P-value
0.6026
ARCH(4) test 𝜒 2 test P-value
0.0916
WALD TEST RESULTS: DIESEL
Null Hypothesis
𝜒 2 test P-value
Net Price = Excise Tax
N.A.
Net Price = Carbon Tax
N.A.
Excise Tax = Carbon Tax
N.A.
For Diesel:
- Net price effects are not significant for diesel sales
- Excise tax effects are not significant for diesel sales
- Only carbon taxes significantly affect diesel sales (at t and t+1)
- However, cumulative effects over the two periods are nil.
CONCLUSION
• Carbon taxes influence consumers (gasoline market) differently than
firms (diesel market)
• Within each market, the different components making up the total
price have different effects on sales in that market
• Therefore, it is important to calculate emission reductions on the basis
of the correct elasticities; i.e. we must account for different price and
tax effects.
• B.C. neutrality calculations were done on the basis of total price
elasticities ! The observed deficit is likely due to the underestimation of
the consumer response to the carbon tax.