Asymmetric adjustment of clean energy demand in the OECD countries

ABSTRACT

This paper investigates the asymmetric demand for clean energy for a panel of OECD countries. We construct a price index for clean energy based on information on the energy composition of electricity generation, and focus on the possibility of non-symmetric clean and dirty clean consumption due to rising and falling energy prices, and to positive income growth and negative income growth. Using an Autoregressive Distributed Lag (ARDL) model that encapsulates both long-run equilibrium among energy demands, energy prices and income, and short-run dynamics. We find evidence that clean energy demand adjusts more aggressively to a price decline than a price rise. In addition, we find that clean energy consumption is more responsive to negative income changes than to positive income changes. A climate policy implication is that policy should account for reversal effect of clean energy consumption during periods of economic downturn.

KEYWORDS:

• Asymmetric responses
• consumer demand
• ARDL

JEL CLASSIFICATION:

• C23
• Q41
• C13

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ This link,https://public.tableau.com/views/IRENARETimeSeries/Charts?:embed=y :showVizHome=no publish=yes :toolbar=no, provides trend of the world electricity generation by renewable from 2000 onwards.

² See the online supplemental appendix of Papageorgiou, Saam, and Schulte (2017).

³ See, for example, Guo et al. (2021).

⁴ See, among others, Dargay and Gately (1995).

⁵ The periods of oil crises are 1973, 1979, 1990. These episodes saw substantial increases in oil prices.

⁶ The three-way decomposition for a series $x_t$ is as follows.

$x_t=x_t^{max}+x_t^{+}+x_t^{-}$

where $x_t^{max}=\mathit{max}(x_0,\dots ,x_t)$, $x_t^{+}={\sum }_{i=0}^t\mathit{max}(0,(x_{i-1}^{max}-x_{\mathit{i}-1})-(x_i^{max}-x_i))$, and $x_t^{-}={\sum }_{i=0}^t\mathit{min}(0,(x_{i-1}^{max}-x_{\mathit{i}-1})-(x_i^{max}-x_i))$.

⁷ See the online supplemental appendix of Papageorgiou, Saam, and Schulte (2017).

⁸ Our treatment of the electricity price as a weighted average of the clean energy price and the dirty energy price can be interpreted as a solution to a problem where an electricity producer aims to maximize its profits subject to a Cobb-Douglas production function which combines clean and dirty energy to generate electricity. This problem leads to a standard solution where revenues from the sale of electricity equal the total cost incurred from the use of clean and dirty energy.

⁹ We follow Greenwood-Nimmo and Shin (2013) to define the cumulative dynamic multiplier effects of $x_{it}^{+}$ and $x_{it}^{-}$ on $E_{i,t}^j$ as follows: $m_h^{x+}={\sum }_{v=0}^h\frac{\mathrm{\partial }{E}_{i,t+v}^j}{\mathrm{\partial }{x}_{it}^{+}}$ and $m_h^{x-}={\sum }_{v=0}^h\frac{\mathrm{\partial }{E}_{i,t+v}^j}{\mathrm{\partial }{x}_{it}^{-}}$, where $x_{\mathit{it}}\in \left\{y_{\mathit{it}},p_{it}^c,p_{it}^d\right\}$. Notice that when h increases to infinity, the cumulative dynamic multiplier converges to the long-run elasticities.

¹⁰ Since the industrial and electricity generating sectors could have wholesale prices while the household sector often faces higher retail prices. For instance, the mean nominal price of steam coal over the period of 1978–2015 is 2761.068 US$\mathrm{\$}$ per TJ in the industrial sector, 2202.147 US$\mathrm{\$}$ per TJ in the electricity-generating sector, 7089.235 US$\mathrm{\$}$ per TJ in the household sector. Although the prices are cheaper in the electricity-generating sector, it is more appropriate to replace the missing values than the residential price level.

¹¹ We compared the growth rate of energy of a specific European country to the average growth rate of that energy in OECD Europe and find most European countries have a similar growth rate of specific energy. The reason why we use this interpolation approach is that although the price level of specific energy in each country is different due to taxes or transport cost, the growth rate of energy should be similar to capture the same upward and downward price movement in the international energy market.

¹² Note here the nominal dirty price $\mathit{n}{p}_{it}^d$ is re-constructed by weights of shares of electricity generation from three dis-aggregated dirty energies: $\mathit{n}{p}_{it}^d=\frac{s_{it}^{oil}}{s_{it}^d}\times \mathit{n}{p}_{it}^{oil}+\frac{s_{it}^{c\mathrm{oal}}}{s_{it}^d}\times \mathit{n}{p}_{it}^{\mathit{coal}}+\frac{s_{it}^{n\mathrm{gas}}}{s_{it}^d}\times \mathit{n}{p}_{it}^{\mathit{ngas}}$ where $s_{it}^d=s_{it}^{oil}+s_{it}^{\mathit{coal}}+s_{it}^{\mathit{ngas}}$.