Do environmental taxes affect energy consumption and energy intensity? An empirical analysis of OECD countries

Figures & data

Figure 1. Revenue from environmental taxes in the OECD 1994–2019, as a percentage of GDP and as a percentage of total tax revenue.

Source: (OECD, 2019)

Figure 2. Composition of environmentally related tax revenue in OECD countries in 2019.

Source: OECD Environment Statistics (database) (OECD, 2021).

Figure 3. Energy consumption.

Source: Enerdata (2021),

Figure 4. Energy intensity of GDP at constant purchasing power parities (koe/$15p) (from 2000–2020).

Source: Enerdata,2021

Table 1. An a prior expectation between energy consumption and determinants of environmental taxes. Dependent variable: energy consumption and energy intensity

Explanatory variables	Description	Measurement	Expectation
Ev	Environmental tax	Total environmental tax to total tax	-
GDP	Economic growth	GDP constant 2010 US dollars	+
FD	Financial development	Index of financial development	±
$TO$	Trade openness	The ratio of import and export to GDP	+
UB	Urbanization	Total number of urban population	±

Source: Authors’ construct

Table 2. Definition and sources of variables

Label	Description	Source
Environmental tax	Environmental tax to total tax ratio/environmental tax GDP ratio	OECD database
Energy Consumption	Energy use (kg of oil equivalent)	World development indicators
Economic Growth	GDP constant 2010 US dollars	World development indicators
Financial development index	This is the ratio of Access Index, Depth Index, Efficiency Index, Institutions Index, Markets Access Index, Institutions, Markets Depth Index, Markets Efficiency Index, Markets Index	IMF data Mapper
Urbanization	Total urban population	World development indicators
Trade Openness	The ratio of export plus import to GDP	World development indicators

Source: Author’s construct

Table 3. Cross-sectional dependence test

Residual cross-section dependence test Null hypothesis: No cross-section dependence(correlation) in residuals Equation: Equation 1(1) $CD=\sqrt{\frac{2T}{N\left(N-1\right)}\left({\sum }_{i=1}^{N-1}{\sum }_{j=i+1}^N\sqrt{T}{\rho }_{ij}\right)}$(1) Periods included:20 Cross-sections included:35 Total observations: 700 Note: none zero cross-sections means detected in panel Cross-section means were removed during the computation of correlations
Test	Statistic	Df	Prob
Breusch-Pagan LM Pesaran scaled LM Pesaran CD	4704.93 119.1401 29.12867	595	0.0000 0.0000 0.0000

Note: All tests were carried at 5% significance level.

Source: Author’s compilation

Table 4. Results of unit root test

Level		LLC	IPS	ADF-Fisher	PP-Fisher
	Ev.tax%tax	−3.678*	−2.561*	109.85**	111.148**
	Ev.tax%gdp	−1.103	−0.884	89.29	103.52
	FDI	−3.551**	−0.648	100.98	111.98
	Openness	−6.510*	−3.900**	114.774**	85.87
	Urbanization	0.529	1.252	109.241**	82.394
	GDP	−2.528**	2.424	46.640	25.570
	E_COM	−0.790	2.021	69.134	76.866
	E_I	−2.008**	−3.058*	118.382*	84.841**
1^st Difference
	Ev.tax%tax	−18.409**	−14.313**	270.116**	353.62**
	Ev.tax%gdp	−12.83**	−12.157**	258.339**	361.22**
	FDI	−16.442**	−16.51**	329.17**	454.073**
	Openness	−18.090**	−14.221**	294.98**	458.91**
	Urbanization	−16.981**	−11.40**	242.401**	192.561**
	GDP	−12.205**	−8.607**	191.52**	275.58**
	Ene_use	−21.220**	−20.614**	416.766**	492.332**
	Ene_I	−11.210**	−12.238**	267.329**	393.616**

Note: Test statistic was calculated at 1% and 5% significance. P <0.01*, p <0.05**

Source: Author’s construct

Table 5. Panel co-integration results

Test	Model 1		Model 2		Model 3		Model 4
	Statistic	p-value	Statistic	p-value	Statistic	p-value	Statistic	p-value
Pedroni test
Modified Phillips-Perron	4.1807	0.0000	3.8030	0.0001	3.9990	0.0000	3.6855	0.0001
Phillips-Perron t	−4.8241	0.0000	−6.0632	0.0000	−5.2153	0.0000	−5.7784	0.0000
Augmented Dickey-Fuller	−4.7773	0.0000	−5.9967	0.0000	−5.1721	0.0000	−5.9112	0.0000
Westerlund	−2.6076	0.0046	−2.3639	0.0090	−2.5990	0.0047	−2.7776	0.0027
	H0: No cointegration among the panels. H1: All panels are cointegrated.

Note: Test statistic was calculated at 1% and 5% significance. P <0.01*, p <0.05**

Source: Author’s Construct

Table 6. Results of estimation

Model 1: dependent variable energy consumption						Model 2:dependent variable energy intensity
	FMOLS			DOLS		FMOLS		DOLS
	Coeff	t-stats		Coeff	t-stats	Coeff	t-stats	Coeff	t-stats
Ev. Tax	−0.1485*	−5.0628		−0.0887*	−1.3301	−0.213*	7.4264	−0.284*	−2.3418
Financial Dev’t	−0.122**	−1.8944		−0.0354*	−0.9415	−0.149*	−2.2459	−0.113*	−0.2851
GDP	0.3536*	8.9825		0.0740*	3.0124	0.277*	4.2441	0.187*	2.3255
Openness	0.1174*	5.1431		0.2866**	3.6484	0.208*	7.7360	0.338*	1.6652
Urbanization	0.2324*	0.9541		1.202*	6.0526	1.979*	5.5730	0.203*	0.8092
R-squared	0.3205			0.3940		0.3025		0.6045
Long-run variance	0.0004			0.0016		0.0013		0.0017

Note: Test statistic was calculated at 1% and 5% significance. P <0.01*, p <0.05**

Source: Author’s Construct

Table 7. Disaggregation of environmental tax effect (long-run coefficients)

Environmental tax	Dependent variable: Ene_Com		Dependent variable: Ene_I
	FMOLS	DOLS	FMOLS	DOLS
	Coeff	Coeff	Coeff	Coeff
Energy tax	−0.0549** (0.022)	−0.351** (0.0035)	−0.0890** (0.024)	−0.1145** (0.017)
Pollution tax	−0.0371** (0.022)	−0.0621** (0.0047)	−0.0263** (0.015)	−0.0331** (0.025)
Transport tax R-squared	−0.0142 (0.020) 0.2312	−0.389 (0.049) 0.114	−0.068 (0.031) 0.225	−0.214 (0.066) 0.182

Note: Standard errors in parenthesis, Test statistic was calculated at 1% and 5% significance. P <0.01*, p <0.05**.Dependent variables used were energy consumption (Ene_Com) and energy intensity (Ene_I).

Source Authors Construct

Table 8. Estimation results (using Share of environmental tax to GDP as independent variable)

Model 3: dependent variable energy consumption						Model 4: dependent variable energy intensity
	FMOLS			DOLS		FMOLS		DOLS
	Coeff	t-stats		Coeff	t-stats	Coeff	t-stats	Coeff	t-stats
Ev. Tax	−0.0745	−3.6532		−0.2803	−13.1973	−0.0349	−1.7096	−0.3411	−28.711
Financial Dev’t	−0.0558	−2.6739		−0.0722	−12.273	−0.0376	−1.7990	−0.0070	−0.3391
GDP	0.2782	40.257		0.1617	5.8974	0.5445	79.176	0.4443	67.398
Openness	−0.0755	−5.6948		0.0788	2.9520	0.2073	15.6201	0.1995	5.0641
Urbanization	−0.3363	−253.8		1.3011	15.4596	0.0796	60.076	0.6318	2.362
R-squared	0.9673			93.450		0.9361		0.9670
Long-run variance	0.0021			0.0002		0.0003		0.0002

Source Authors Construct

Table A1. Descriptive statistics

	E_I	ENE_COM	EV_TAX	EV_TAX_GDP	FDI	GDP	OPENNESS	URBAN
Mean	1.613755	4053.556	17.94799	2.454271	0.589968	1.11E+12	87.56817	75.46610
Median	1.558436	3614.982	7.178500	2.461500	0.601687	3.09E+11	73.09910	76.60350
Maximum	2.955845	18,178.14	7314.000	5.372000	1.000000	1.63E+13	392.8042	97.83300
Minimum	0.815568	616.6066	2.740000	0.467000	0.095064	6.62E+08	16.67948	50.62200
Std. Dev.	0.366076	2315.046	276.1805	0.807354	0.205485	2.45E+12	51.31709	11.25221
Skewness	0.742733	2.538510	26.39474	0.376237	−0.261633	4.232369	2.122545	−0.258483
Kurtosis	3.865109	13.92361	697.7881	4.096866	2.219074	22.32336	9.992648	2.366967
Jarque-Bera	86.18813	4232.122	14,160,918	51.60552	25.77321	12,980.45	1951.773	19.48290
Probability	0.000000	0.000000	0.000000	0.000000	0.000003	0.000000	0.000000	0.000059
Sum	1129.629	2,837,489.	12,563.59	1717.990	412.9779	7.78E+14	61,297.72	52,826.27
Sum Sq. Dev.	93.67435	3.75E+09	53,316,680	455.6220	29.51454	4.18E+27	1,840,777.	88,501.95

Table B1. Quantile regression estimates (dependent variable: energy consumption)

Variables	Q25	Q50	Q75	Q90
Environmental taxes	−0.1705** (0.050)	−0.0656 (0.055)	−0.1984** (0.073)	−0.0362 (0.023)
Financial Development	0.0793 (0.174)	−0.0034 (0.005)	0.1871** (0.098)	0.0389 (0.339)
Economic Growth	0.1135** (0.044)	0.0453** (0.013)	0.0103 (0.013)	−0.0094 (0.043)
Trade openness	0.3852** (0.075)	0.3315** (0.045)	0.2183** (0.036)	0.1120 (0.074)
Urbanization	0.8226** (0.108)	1.1176** (0.105)	1.3619** (0.122)	1.8167** (0.314)
Constant	−0.3501 (1.550)	0.8716 (0.569)	1.3142** (0.703)	0.7373 (2.669)
Obs.	700	700	700	700
Pseudo R-Squared	0.1923	0.1657	0.2010	0.2278

Table C1. Quantile regression estimates (dependent variable: energy intensity)

Variables	Q25	Q50	Q75	Q90
Environmental taxes	−0.1367** (0.104)	−0.204** (0.144)	−0.0137** (0.019)	−0.045** (0.013)
Financial Development	−0.088* (0.268)	−0.1493 (0.039)	−0.1905* (0.090)	−0.0771 (0.086)
Economic Growth	−0.0743 (0.054)	−0.0752 (0.035)	−0.1270 (0.011)	−0.1484 (0.012)
Trade openness	−0.1129 (0.086)	−0.1434 (0.1825)	−0.2858 (0.089)	−0.3863 (0.037)
Urbanization	0.0874* (0.092)	0.1474* (0.265)	0.7372* (0.088)	0.4446* (0.154)
Constant	3.6699 (1.784)	3.891 (1.997)	3.1708 (0.639)	5.7324 (0.822)
Obs.	700	700	700	700
Pseudo R-Squared	0.1651	0.083	0.1846	0.2502

Table C2. Panel generalized methods of moments Dependent variable: Energy Intensity

Variable	Coefficient	Std Error	t-Statistic	prob
L_EV_TAX	−0.117841	0.154021	0.765118	0.0445
L_FDI	0.411750	0.239903	1.716318	0.0866
L_GDP	−0.011094	0.041541	−0.267061	0.7895
L_T O	0.345392	0.064034	5.393850	0.0000
L_U	1.082194	0.187505	5.771540	0.0000
C	2.294591	1.761204	1.302853	0.1931
Root MSE	0.617049	R-squared	0.331374
Mean dependent var	8.174653	Adjusted R-squared	0.341475
S.D. dependent var	0.535176	S.E. of regression	0.619852
Sum squared resid	253.1985	Durbin-Watson stat	0.971289
J-statistic	8.38E-19	Instrument rank	6

Table C3. Panel generalized methods of moments. dependent variable: energy consumption

Variable	Coefficient	Std Error	t-Statistic	prob
L_EV_TAX	−0.086091	0.081433	−1.057201	0.0308
L_FDI	−0.058956	0.126840	−0.464805	0.6422
L_GDP	−0.105914	0.021963	−4.822321	0.0000
L_TO	−0.185840	0.033856	−5.489145	0.0000
L_U	0.359689	0.099137	3.628214	0.0003
C	3.791231	0.931174	4.071450	0.0001
Root MSE	0.617049	R-squared	0.331374
Mean dependent var	8.174653	Adjusted R-squared	0.341475
S.D. dependent var	0.535176	S.E. of regression	0.619852
Sum squared resid	253.1985	Durbin-Watson stat	0.971289

OECD. 2019. Taxing energy use for sustainable development opportunities for energy tax and subsidy reform in selected developing and emerging economies. https://www.oecd.org/tax/tax-policy/taxing-energy-use-for-sustainable-development.pdf

 Google Scholar