The Asymmetric Effects of Crude Oil Prices and Exchange Rates on Diesel Prices for 27 European Countries

Abstract

Many studies have examined the asymmetric effect of US dollar-denominated crude oil prices on petroleum product prices. The ‘rockets and feathers’ argument suggests that a crude price increase raises petroleum product prices more than a corresponding decrease in crude prices lowers product prices. However, for the countries that do not use the US dollar as a medium of exchange, petroleum product prices are also affected by the exchange rates. This paper analysed the asymmetric effects of both US dollar-denominated crude oil prices and exchange rates on local currency-denominated diesel prices for 27 European countries in the short run as well as long run. The overall empirical evidence suggests that, in the short run, diesel prices react more to crude oil price increases than to a decrease, parallel to the ‘rockets and feathers’ argument. However, contrary to that argument, the long-run adjustment is the opposite. As for exchange rate shocks, again the ‘rockets and feathers’ argument holds and diesel prices respond more to exchange rate depreciation than appreciation in the short and long run.

Keywords

Rockets and feathers asymmetric effect crude oil price diesel price exchange rate

Introduction

Petroleum product prices are important because of their effects on economic performance (Kushwah & Siddiqui, 2020; Rao & Goyal, 2018). Therefore, the effects of petroleum product prices on economic performance have been investigated in various studies in the literature. Apart from several determinants such as transportation cost, rental cost and labour cost, petroleum product prices are mostly influenced by crude oil prices and exchange rates. Most of the earlier studies were associated with the ‘rockets and feathers’ hypothesis introduced by Bacon (1991), which posits that petroleum product prices react faster to rises in crude oil prices than to decreases in prices. Besides crude oil prices, another factor is the exchange rate pass-through to petroleum product prices that builds up price asymmetry due to exchange rate changes. Particularly, the recent volatility in crude oil prices attracts attention to the price asymmetry commonly seen in retail petroleum product markets. This study aimed to explore the asymmetric effects of crude oil prices and exchange rates on diesel prices for 27 European countries over the short and long run.

The ‘rockets and feathers’ hypothesis asserts that gasoline prices give a faster response to increases in crude oil prices than to decreases. It is suggested that companies, owing to their market power, are inclined to raise prices rapidly when faced with cost increases while behaving sluggishly to adjust their prices downwards in response to cost decreases to keep profit levels high. Some other possible explanations for the presence of price asymmetry include inventory adjustment costs, the level of competition at different stages of the distribution channel, consumer search costs, type of cost shock (permanent vs. transitory cost shock and long-term vs. short-term cost shock) and consumer response to changing prices (Balke et al., 1998; Berument et al., 2014; Borenstein et al., 1997; Brown & Yücel, 2000; Perdiguero-García, 2013). The existing empirical literature on asymmetric price adjustment between crude oil and gasoline prices differs by country, data set, frequency of data, the econometric model used and findings. Although the ‘rockets and feathers’ phenomenon is supported in several studies, there are also some mixed results in the literature. For instance, while Bacon (1991) found evidence for strong and rapid adjustment of UK gasoline prices to cost changes, Manning (1991) showed that price adjustment does not last long. Likewise, various results are observed in the literature, some of which confirm price asymmetry and some which conclude no asymmetry or mixed signs of asymmetry.

Several studies have referred to as proponents of price asymmetry. They include studies by Borenstein et al. (1997), Balke et al. (1998), Chen et al. (2005), Radchenko (2005), Al-Gudhea et al. (2007), Radchenko and Shapiro (2011), Pal and Mitra (2015), Polemis and Tsionas (2016, 2017) and Qin et al. (2016) for the US; Reilly and Witt (1998), Greenwood-Nimmo and Shin (2013) for the UK; Galeotti et al. (2003) and Polemis and Fotis (2014, 2015) for few European Union countries and the USA; and Pal and Mitra (2016) for India. A detailed meta-analysis done by Perdiguero-Garcia (2013) provides some evidence as to how common price asymmetry affects gasoline markets.

In contrast, various studies could not find asymmetry or report contradictory results with respect to the size and sign of price asymmetries in gasoline markets. Some of these studies include Kirchgassner and Kubler (1992) and Assane-Otoo and Schneider (2015) for Germany; Contín-Pilart et al. (2009) for Spain; Asplund et al. (2000) for Sweden; Godby et al. (2000) for Canada; Bachmeier and Griffin (2003), Honarvar (2009), Atil et al. (2014) and Bremmer and Kesselring (2016) for the USA; Grasso and Manera (2007), Meyler (2009), Clerides (2010) and Karagiannis et al. (2015) for some European countries; Bermingham and O’Brien (2011) for Ireland and the UK; Kristoufek and Lunackova (2015) and Apergis and Vouzavalis (2018) for some European countries and the USA. The existence of conflicting results shows that a consensus could not be attained concerning the rationale, size and sign of price asymmetries in the petroleum product markets.

Recently, the literature on price asymmetry in gasoline markets has drawn attention to the role of exchange rates in the price transmission mechanism due to high dependency on oil imports, particularly by European countries. For example, Ogbuabor et al. (2018) argued that exchange rate is a component of the cost of oil imports especially for oil-importing European countries, so not accounting for the effect of exchange rates in the analysis of retail energy price fluctuations might be deceptive. Moreover, Chen et al. (2018) stated that for the 27 European Union countries, the asymmetric behaviour of gasoline prices was mainly driven by the effects of exchange rate pass-through. According to statistics of the International Energy Agency, many European countries (e.g., Austria, Belgium, Czech Republic, Finland, France, Germany, Greece, Ireland, Italy, the Netherlands, Poland, Portugal, Slovakia, Spain, Sweden and Turkey) have an increasing oil import dependency ratio varying between 90% and 100% (International Energy Agency, 2014). Countries having high oil import dependency become more vulnerable to exchange rate volatility since crude oil is denominated in the US dollar. Thus, the local currency to US dollar exchange rate turns out to be a significant factor affecting the asymmetric adjustment of petroleum product prices.

Research addressing the impact of exchange rates on asymmetric price adjustments in the fuel market is rare but growing. Among them, the works of Reilly and Witt (1998), Asplund et al. (2000), Galeotti et al. (2003), Berument et al. (2014), Özmen and Akçelik (2017), Bagnai and Ospina (2015), Bagnai and Ospina (2016), Ogbuabor et al. (2018) and Chen et al. (2018) documented gasoline price asymmetry due to exchange rate changes in European countries.

This study aims to contribute to the literature on price asymmetry in different dimensions. It has been observed that the empirical literature on price asymmetry in oil markets mostly employs gasoline prices and that diesel prices are less often studied. Some recent articles including diesel price asymmetry are those by Karagiannis et al. (2015), Özmen and Akçelik (2017) and Ogbuabor et al. (2018). Therefore, this study intends to fill this gap by focusing on diesel price asymmetry in response to crude oil prices and exchange rate shocks. Another contribution is the use of a large dataset of 27 European countries; the analyses are performed separately for each country. Previously, Bagnai and Ospina (2016) and Chen et al. (2018) used a large data set from 12 Eurozone countries and 27 European countries, respectively; however, they only examined gasoline price asymmetry and did not include diesel price asymmetry in their analyses, even though diesel (or gas oil) is more widely used petroleum product in Europe. In addition, due to rapid changes in crude oil prices and exchange rates, there has been a recent rise in the frequency of data employed in various studies as seen in the weekly data of Berument et al. (2014) and Chen et al. (2018), and the daily data of Özmen and Akçelik (2017). Thus, the next contribution of our study was the use of weekly data sets, instead of monthly data. The period covered is another distinction of our study. Analysing the 2009–2018 period gives us the opportunity to investigate price asymmetry after the 2008 crisis, during which sharp rises and falls were seen in crude oil prices and exchange rates.

Recently, the use of the nonlinear autoregressive distributed lags (NARDL) has been prevalent in examining petroleum product price asymmetries. NARDL modeling performs both short- and long-run analyses by estimating one single equation, which is superior to other multi-equation models. Moreover, the response of the dependent variable to positive and negative shocks in the explanatory variable can be observed through asymmetric dynamic multipliers. Additionally, contrary to other cointegration analyses, variables integrated at different orders, I(0) or I(1), can be analysed through the use of bounds testing. Thus, we followed the recent NARDL methodology to reap its advantages.

To the best of our knowledge, this study will be the first attempt to use NARDL methodology to study diesel price asymmetry for such a large data set (consisting of 27 countries). In this respect, our results coincide, to some degree, with and extend the limited literature on diesel price asymmetry. Although some previous studies (Berument et al., 2014; Karagiannis et al., 2015; Ogbuabor et al., 2018; Özmen & Akçelik, 2017) have documented diesel price asymmetry in France, Germany, Italy, Spain, Turkey and the UK, the findings regarding other European countries will be the first in the literature.

The plan of the study is as follows: the second section introduces the data and explains the methodology; the third section discusses the results of the analysis, and finally, the fourth section concludes the paper.

Methodology and Data

In this study, we have employed three variables: the weekly Brent crude oil price in US dollars/barrel, the US dollar value of the local currency (i.e., a higher exchange rate means the appreciation of the local currency) and the retail price of diesel (excluding tax) for 27 European countries (Austria, Belgium, Bulgaria, Cyprus¹ Czech Republic, Denmark, Estonia, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Latvia, Lithuania, Malta, the Netherlands, Poland, Portugal, Romania, Slovakia, Slovenia, Spain, Sweden, Turkey and the UK). We obtained the Brent crude oil price, the local currency–US dollar exchange rate and the countries’ diesel price data from the Bloomberg database. Only the diesel price data for Turkey in Turkish Lira/litre were gathered from the Turkish Energy Market Regulatory Authority. The weekly data set covers the period from 1 May 2009, to 2 November 2018, for all countries except Bulgaria and Romania. For Bulgaria and Romania, the data set covers the period from 26 June 2009, onwards. With this time span, we were able to avoid the turbulent period of the 2008 financial crisis. We used the logarithmic form of all of the data.

The objective of this study was to search specifically for

the effect of US dollar-denominated Brent crude oil prices on national diesel prices in local currency while controlling for the effect of the local currency exchange rate to the US dollar and

the effect of local exchange rates on national diesel prices while controlling for US dollar-denominated Brent crude oil prices.

For this purpose, we proposed the use of the NARDL model as developed by Shin et al. (2014) to estimate diesel prices. This model lends itself to testing the existence of the short- and long-run asymmetric effects of international crude oil prices and exchange rates on local petroleum product prices. One advantage of the NARDL model is its capability to simultaneously capture cointegration dynamics along with the symmetric (linear) and asymmetric (nonlinear) relationships among the variables (Katrakilidis & Trachanas, 2012). The other advantage is that the variables need not have the same order of integration degrees (Pesaran et al., 2001; Shin et al., 2014).

The general linear ARDL model developed by Pesaran and Shin (1999) and Pesaran et al. (2001) was advanced by Shin et al. (2014) to allow for the discrete effect of cumulative positive and negative changes in the explanatory variable on the dependent variable, which is described as the asymmetric (nonlinear) relationship between the dependent and independent variables. Thus, the equation of the NARDL model can be written as follows:

Δ y_{t} = μ + ρ y_{t - 1} + θ^{+} x_{t - 1}^{+} + θ^{-} x_{t - 1}^{-} + γ z_{t} + \sum_{i = 1}^{p - 1} α_{i} Δ y_{t - i} + \sum_{i = 0}^{q - 1} ω_{i}^{+} Δ x_{t - i}^{+} + \sum_{i = 0}^{q - 1} ω_{i}^{-} Δ x_{t - i}^{-} + \sum_{i = 0}^{q - 1} δ_{i} Δ z_{t - i} + ε_{t},

(1)

where $y_{t}$ is the dependent variable, $x_{t}$ is the independent variable, $z_{t}$ is the deterministic or control variables (such as seasonal variables, trends, other exogenous variables with fixed lags), $ε_{t}$ is iid disturbance term and

x_{t} = x_{0} + x_{t}^{+} + x_{t}^{-}

(2)

x_{t}^{+} = \sum_{i = 1}^{t} Δ x_{i}^{+} = \sum_{i = 1}^{t} m a x (Δ x_{i}, 0), x_{t}^{-} = \sum_{i = 1}^{t} Δ x_{i}^{-} = \sum_{i = 1}^{t} \min (Δ x_{i}, 0) .

(3)

The asymmetric long-run relationship between dependent and independent variables is given by

y_{t} = β^{+} x_{t}^{+} + β^{-} x_{t}^{-} + u_{t},

(4)

where the asymmetric long-run parameters related to positive and negative effects of $x_{t}$ can be denoted by $β^{+} = - ρ / θ^{+}$ and $β^{-} = - ρ / θ^{-}$ , respectively (Granger & Yoon, 2002; Schorderet, 2001, 2003).

The NARDL model given by Equation (1) is estimated by using a stepwise regression through which the lag orders p and q are determined via the general-to-specific approach, and statistically insignificant coefficients are dropped from the equation. After estimating the NARDL model, the existence of a relationship between the levels of the variables can be examined using the $t_{B D M}$ statistic, as suggested by Banerjee et al. (1998), by testing using the null hypothesis of $H_{0} : ρ = 0$ against $H_{A} : ρ < 0$ .

Additionally, the null hypothesis of no cointegration ( $H_{0} : ρ = θ^{+} = θ^{-} = γ = 0$ against $H_{A} : a t l e a s t o n e i s n o t e q u a l t o 0$ ) can be tested by standard Wald test statistics ( $F_{P S S}$ ) following Pesaran et al. (2001).

Long-run asymmetry can be tested by a standard Wald test (W_LR) under the null hypothesis of ${H_{0} : β}^{+} = β^{-}$ i.e., $H_{0} : - ρ / θ^{+} = - ρ / θ^{-}$ against $H_{A} : - ρ / θ^{+} \neq - ρ / θ^{-}$ .

The test for short-run asymmetry can be in two forms, pairwise (strong) and additive (weak). The test for pairwise asymmetry includes the null hypothesis of ${H_{0} : ω}_{i}^{+} = ω_{i}^{-}$ for all $i = 0, \dots, q - 1$ against ${H_{A} : ω}_{i}^{+} \neq ω_{i}^{-}$ for all $i = 0, \dots, q - 1$ . The Wald test for additive asymmetry (W_SR) involves testing the null hypothesis of $H_{0} : \sum_{i = 0}^{q - 1} ω_{i}^{+} = \sum_{i = 0}^{q - 1} ω_{i}^{-}$ against $H_{A} : \sum_{i = 0}^{q - 1} ω_{i}^{+} \neq \sum_{i = 0}^{q - 1} ω_{i}^{-}$ .

If asymmetric behaviour is found in either the short run or long run or both, the positive and the negative dynamic multipliers are calculated to see the adjustment process of the dependent variable when $x_{t}^{+}$ and $x_{t}^{-}$ change by one unit. The respective positive and negative dynamic multipliers for $x_{t}^{+}$ and $x_{t}^{-}$ can be given as follows:

m_{h}^{+} = \sum_{i = 0}^{h} \frac{\partial y_{t + i}}{\partial x_{t}^{+,}} m_{h}^{-} = \sum_{i = 0}^{h} \frac{\partial y_{t + i}}{\partial x_{t}^{- for}} h = 0, 1, 2, \dots

(5)

where $m_{h}^{+} \to β^{+}$ and $m_{h}^{-} \to β^{-}$ as $h \to \infty$ .

Dynamic multipliers provide the insight into the asymmetric adjustment mechanism of the dependent variable in response to a shock to the system starting from the initial equilibrium until the new equilibrium state.

Analysis and Discussion

In order to apply the bounds tests following Pesaran et al. (2001), it is necessary to identify the integration order of the variables, which should not be greater than I(1). In Table 1 (Panels A and B) show the results of the Augmented Dickey–Fuller (ADF) (1979) and Phillips–Perron (1988) unit root tests for diesel prices and exchange rates, respectively.

Table 1.

Unit Root Tests for Diesel Prices, Exchange Rates and Crude Oil Price.

	ADF Tests^1,3			Phillips–Perron Tests^2,3
	Level		First Difference	Level		First Difference
Variable	Intercept	Intercept and Trend	Intercept	Intercept	Intercept and Trend	Intercept
Panel A: Diesel prices across countries
Austria diesel price	0.848 (0.994)	−2.059 (0.561)	−7.309*** (0.000)	0.576 (0.988)	−2.174 (0.497)	−7.304*** (0.000)
Belgium diesel price	−0.809 (0.811)	−3.937** (0.015)	−12.834*** (0.000)	−1.139 (0.697)	−3.772** (0.023)	−15.480*** (0.000)
Bulgaria diesel price	0.354 (0.980)	−2.215 (0.475)	−5.299*** (0.000)	0.571 (0.988)	−2.019 (0.582)	−5.238*** (0.000)
Cyprus diesel price	−0.017 (0.954)	−2.635 (0.266)	−6.040*** (0.000)	0.333 (0.979)	−2.434 (0.360)	−5.996*** (0.000)
Czech Rep. diesel price	0.348 (0.980)	−1.765 (0.713)	−6.702*** (0.000)	0.222 (0.973)	−1.883 (0.654)	−6.950*** (0.000)
Denmark diesel price	−0.251 (0.927)	−2.647 (0.261)	−8.309*** (0.000)	−0.177 (0.936)	−2.582 (0.290)	−8.298*** (0.000)
Estonia diesel price	−1.994 (0.289)	−2.606 (0.279)	−11.196*** (0.000)	−1.777 (0.389)	−2.607 (0.279)	−11.321*** (0.000)
Finland diesel price	0.617 (0.989)	−3.027 (0.131)	−11.388*** (0.000)	0.580 (0.988)	−2.843 (0.187)	−11.382*** (0.000)
France diesel price	−0.425 (0.899)	−2.854 (0.183)	−6.708*** (0.000)	0.047 (0.960)	−2.362 (0.396)	−6.611*** (0.000)
Germany diesel price	1.265 (0.998)	−1.332 (0.873)	−7.914*** (0.000)	1.195 (0.998)	−1.387 (0.858)	−7.914*** (0.000)
Greece diesel price	0.030 (0.958)	−2.583 (0.289)	−5.520*** (0.000)	0.414 (0.983)	−2.399 (0.377)	−5.484*** (0.000)
Hungary diesel price	−0.131 (0.942)	−2.388 (0.383)	−8.919*** (0.000)	−0.111 (0.944)	−2.401 (0.376)	−8.919*** (0.000)
Ireland diesel price	−0.257 (0.926)	−2.147 (0.512)	−16.524*** (0.000)	−0.931 (0.774)	−3.427* (0.055)	−16.524*** (0.000)
Italy diesel price	−0.208 (0.932)	−2.558 (0.300)	−5.140*** (0.000)	0.183 (0.970)	−2.263 (0.449)	−5.186*** (0.000)
Latvia diesel price	−0.324 (0.916)	−2.277 (0.441)	−10.491*** (0.000)	−0.227 (0.930)	−2.229 (0.467)	−10.399*** (0.000)
Lithuania diesel price	−0.030 (0.953)	−2.632 (0.268)	−7.068*** (0.000)	0.284 (0.976)	−2.219 (0.473)	−7.044*** (0.000)
Malta diesel price	−0.412 (0.901)	−1.471 (0.832)	−9.055*** (0.000)	−0.412 (0.901)	−1.471 (0.832)	−9.055*** (0.000)
Netherlands diesel price	−0.307 (0.918)	−2.631 (0.268)	−7.728*** (0.000)	−0.307 (0.918)	−2.631 (0.268)	−7.693*** (0.000)
Poland diesel price	0.161 (0.968)	−2.186 (0.491)	−8.007*** (0.000)	−0.214 (0.932)	−2.424 (0.365)	−8.219*** (0.000)
Portugal diesel price	−0.576 (0.869)	−3.066 (0.121)	−6.370*** (0.000)	−0.191 (0.935)	−2.534 (0.311)	−6.283*** (0.000)
Romania diesel price	−0.362 (0.910)	−2.441 (0.356)	−7.868*** (0.000)	−0.362 (0.910)	−2.441 (0.356)	−7.871*** (0.000)
Slovakia diesel price	0.420 (0.983)	−2.287 (0.436)	−7.476*** (0.000)	0.266 (0.975)	−2.287 (0.436)	−7.484*** (0.000)
Slovenia diesel price	0.086 (0.963)	−3.220* (0.088)	−6.488*** (0.000)	0.491 (0.986)	−2.672 (0.251)	−5.855*** (0.000)
Spain diesel price	−0.276 (0.923)	−2.797 (0.203)	−5.758*** (0.000)	0.026 (0.958)	−2.415 (0.369)	−5.732*** (0.000)
Sweden diesel price	−0.634 (0.856)	−2.074 (0.552)	−7.883*** (0.000)	−0.700 (0.840)	−2.312 (0.423)	−7.883*** (0.000)
Turkey diesel price	0.057 (0.960)	−2.817 (0.196)	−9.211*** (0.000)	0.082 (0.963)	−2.817 (0.196)	−9.213*** (0.000)
UK diesel price	0.375 (0.981)	−2.745 (0.222)	−8.632*** (0.000)	0.353 (0.980)	−2.767 (0.214)	−8.653*** (0.000)
Panel B: Exchange rates
Euro–US dollar exch. rate	−2.688* (0.080)	−1.593 (0.788)		−2.617* (0.094)	−1.620 (0.777)
Bulgarian Lev–US dollar exch. rate	−2.686* (0.081)	−1.591 (0.788)		−2.615* (0.094)	−1.618 (0.777)
Czech Koruna–US dollar exch. rate	−2.813* (0.061)	−1.511 (0.818)		−2.852* (0.056)	−1.511 (0.818)
Danish Krone–US dollar exch. rate	−2.237 (0.195)	−1.198 (0.904)	−7.073*** (0.000)	−2.216 (0.202)	−1.198 (0.904)	−7.090*** (0.000)
Hungarian Forint–US dollar exch. rate	−1.773 (0.391)	−1.728 (0.730)	−7.098*** (0.000)	−1.684 (0.436)	−1.747 (0.721)	−7.034*** (0.000)
Polish Zloty–US dollar exch. rate	−2.028 (0.275)	−1.538 (0.808)	−7.610*** (0.000)	−2.053 (0.264)	−1.535 (0.810)	−7.610*** (0.000)
Romanian Leu–US dollar exch. rate	−2.218 (0.202)	−1.722 (0.733)	−7.939*** (0.000)	−2.243 (0.193)	−1.722 (0.733)	−7.906*** (0.000)
Swedish Krona–US dollar exch. rate	−1.205 (0.669)	−1.599 (0.785)	−7.018*** (0.000)	−1.009 (0.747)	−1.667 (0.757)	−6.905*** (0.000)
Turkish Lira–US dollar exch. rate	−0.295 (0.920)	−2.174 (0.497)	−6.343*** (0.000)	−0.265 (0.925)	−2.136 (0.518)	−6.531*** (0.000)
UK Sterling–US dollar exch. rate	−2.137 (0.231)	−1.454 (0.837)	−6.907*** (0.000)	−2.042 (0.269)	−1.367 (0.863)	−6.266*** (0.000)
Brent crude oil price	−0.988 (0.754)	−2.842 (0.187)	−6.637*** (0.000)	−0.765 (0.823)	−2.272 (0.444)	−6.553*** (0.000)

Sources: The authors. Bloomberg Database and the Turkish Energy Market Regulatory Authority.

Notes: ***, ** and * Indicate significance at 1%, 5% and 10%, respectively.

¹ The optimal lag order is selected by the Schwarz information criterion.

² Spectral estimation method is selected as Bartlett kernel. Bandwidth selection method is Newey-West Bandwidth.

³ Table values are t-statistics and numbers in parentheses are MacKinnon (1996) one-sided p-values.

Table 1 suggests that the series are mixed in integration, either I(0) or I(1). Thus, the NARDL model is applicable for the asymmetric adjustment analysis and the following NARDL model is estimated:

Δ {D i e s e l}_{t} = μ + ρ {d i e s e l}_{t - 1} + θ_{1}^{+} {o i l}_{t - 1}^{+} + θ_{1}^{-} {o i l}_{t - 1}^{-} + θ_{2}^{+} {e r}_{t - 1}^{+} + θ_{2}^{-} {e r}_{t - 1}^{-} + \sum_{i = 1}^{p - 1} α_{i} Δ {d i e s e l}_{t - i} + \sum_{i = 0}^{q - 1} ω_{1 i}^{+} Δ {o i l}_{t - i}^{+} + \sum_{i = 0}^{q - 1} {ω_{1 i}^{-} Δ o i l}_{t - i}^{-} + \sum_{i = 0}^{q - 1} ω_{2 i}^{+} Δ {e r}_{t - i}^{+} + \sum_{i = 0}^{q - 1} ω_{2 i}^{-} Δ {e r}_{t - i}^{-} + ε_{t}

(6)

where diesel, oil and er refer to the diesel price, crude oil price and exchange rate, respectively. The ‘+’ and ‘−’ superscripts stand for the cumulative positive and negative changes in the corresponding variables as in Equation (3).

In Table 2 (panel A) gives the test statistics on the long-run relationship among diesel prices, crude oil prices and exchange rates as gathered by estimating Equation (6).² Panel B is for the test statistics on the asymmetric effect of crude oil price changes on diesel prices, and Panel C shows that of exchange rate changes on diesel prices.

Table 2.

Asymmetric Effects of Crude Oil Price and Exchange Rate Changes on Diesel Prices.

	Panel A: Long-run Relationship Tests		Panel B: Asymmetric Effects of Crude Oil Price Changes				Panel C: Asymmetric Effects of Exchange Rate Changes
Country	t _BDM ⁽¹⁾	F _PSS ⁽²⁾	Long-run Asymmetry W_LR⁽³⁾	Short-run Asymmetry W_SR⁽⁴⁾	L ⁺⁽⁵⁾	L ⁻⁽⁶⁾	Long-run Asymmetry W_LR⁽³⁾	Short-run Asymmetry W_SR⁽⁴⁾	L ⁺⁽⁵⁾	L ⁻⁽⁶⁾
	Panel A: Long-run Relationship Tests		Panel B: Asymmetric Effects of Crude Oil Price Changes				Panel C: Asymmetric Effects of Exchange Rate Changes
Country	t _BDM ⁽¹⁾	F _PSS ⁽²⁾	Long-run Asymmetry W_LR⁽³⁾	Short-run Asymmetry W_SR⁽⁴⁾	L ⁺⁽⁵⁾	L ⁻⁽⁶⁾	Long-run Asymmetry W_LR⁽³⁾	Short-run Asymmetry W_SR⁽⁴⁾	L ⁺⁽⁵⁾	L ⁻⁽⁶⁾
Austria (12 lag)	−8.610***	20.182***	13.076*** (0.000)	35.274*** (0.000)	0.701*** (0.000)	0.751*** (0.000)	23.689*** (0.000)	1.618 (0.204)	−0.839*** (0.000)	−1.051*** (0.000)
Belgium (6 lag)	−9.818***	29.595***	3.003* (0.084)	37.083*** (0.000)	0.815*** (0.000)	0.849*** (0.000)	2.803* (0.095)	n.a.⁽⁷⁾	−1.208*** (0.000)	−1.313*** (0.000)
Bulgaria (11 lag)	−6.870***	11.777***	62.904*** (0.000)	5.037** (0.025)	0.480*** (0.000)	0.783*** (0.000)	67.861*** (0.000)	0.132 (0.716)	−0.468** (0.048)	−1.473*** (0.000)
Cyprus (12 lag)	−10.096***	31.925***	11.135*** (0.001)	57.538*** (0.000)	0.745*** (0.000)	0.798*** (0.000)	16.134*** (0.000)	0.338 (0.561)	−0.898*** (0.000)	−1.100*** (0.000)
Czech Rep. (12 lag)	−9.138***	24.275***	0.710 (0.400)	0.001 (0.973)	0.757*** (0.000)	0.775*** (0.000)	0.454 (0.501)	7.086*** (0.008)	−1.056*** (0.000)	−1.094*** (0.000)
Denmark (11 lag)	−7.393***	13.488***	20.766*** (0.000)	14.362*** (0.000)	0.611*** (0.000)	0.693*** (0.000)	33.716*** (0.000)	0.527 (0.468)	−0.667*** (0.000)	−1.000*** (0.000)
Estonia (6 lag)	−4.323***	4.010	0.036 (0.850)	0.543 (0.462)	0.700*** (0.000)	0.689*** (0.000)	0.004 (0.949)	2.432 (0.120)	−1.112*** (0.002)	−1.124*** (0.000)
Finland (8 lag)	−9.276***	23.067***	0.120 (0.729)	12.684*** (0.000)	0.734*** (0.000)	0.725*** (0.000)	0.027 (0.871)	15.639*** (0.000)	−0.873*** (0.000)	−0.886*** (0.000)
France (11 lag)	−4.505***	4.612*	0.727 (0.394)	1.492 (0.223)	0.813*** (0.000)	0.791*** (0.000)	0.012 (0.912)	10.172*** (0.002)	−0.854*** (0.000)	−0.846*** (0.000)
Germany (12 lag)	−6.915***	11.972***	4.013** (0.046)	9.298*** (0.002)	0.781*** (0.000)	0.823*** (0.000)	3.911** (0.049)	21.158*** (0.000)	−1.209*** (0.000)	−1.342*** (0.000)
Greece (12 lag)	−10.950***	32.867***	0.441 (0.507)	27.945*** (0.000)	0.677*** (0.000)	0.687*** (0.000)	2.180 (0.141)	20.520*** (0.000)	−0.952*** (0.000)	−1.023*** (0.000)
Hungary (10 lag)	−6.578***	9.121***	14.095*** (0.000)	1.512 (0.219)	0.624*** (0.000)	0.671*** (0.000)	25.444*** (0.000)	17.797*** (0.000)	−0.643*** (0.000)	−0.802*** (0.000)
Ireland (8 lag)	−9.445***	21.158***	1.681 (0.196)	3.748* (0.054)	0.709*** (0.000)	0.740*** (0.000)	2.119 (0.146)	n.a.⁽⁷⁾	−0.913*** (0.000)	−1.026*** (0.000)
Italy (11 lag)	−10.001***	30.144***	0.023 (0.880)	47.174*** (0.000)	0.814*** (0.000)	0.817*** (0.000)	0.002 (0.969)	20.323*** (0.000)	−1.035*** (0.000)	−1.037*** (0.000)
Latvia (12 lag)	−9.299***	20.234***	5.965** (0.015)	15.300*** (0.000)	0.763*** (0.000)	0.818*** (0.000)	6.751*** (0.010)	11.247*** (0.001)	−1.020*** (0.000)	−1.206*** (0.000)
Lithuania (12 lag)	−12.926***	43.656***	39.373*** (0.000)	26.051*** (0.000)	0.711*** (0.000)	0.807*** (0.000)	38.489*** (0.000)	1.008 (0.316)	−1.003*** (0.000)	−1.305*** (0.000)
Malta (11 lag)	−3.735**	7.359***	4.020** (0.046)	2.676 (0.103)	0.549*** (0.000)	0.729*** (0.000)	2.709 (0.101)	6.519** (0.011)	−0.982* (0.057)	−1.475*** (0.000)
Netherlands (12 lag)	−5.357***	6.347**	3.013* (0.083)	0.418 (0.518)	0.727*** (0.000)	0.773*** (0.000)	5.747** (0.017)	13.555*** (0.000)	−0.817*** (0.000)	−1.017*** (0.000)
Poland (12 lag)	−9.125***	22.115***	1.913 (0.167)	20.044*** (0.000)	0.666*** (0.000)	0.689*** (0.000)	5.966** (0.015)	9.938*** (0.002)	−0.922*** (0.000)	−1.017*** (0.000)
Portugal (12 lag)	−5.045***	6.777***	0.338 (0.561)	1.488 (0.223)	0.793*** (0.000)	0.772*** (0.000)	0.665 (0.415)	1.171 (0.280)	−1.370*** (0.000)	−1.277*** (0.000)
Romania (12 lag)	−8.351***	17.047***	3.557* (0.060)	12.016*** (0.001)	0.679*** (0.000)	0.702*** (0.000)	13.393*** (0.000)	24.799*** (0.000)	−0.753*** (0.000)	−0.881*** (0.000)
Slovakia (12 lag)	−10.571***	31.851***	0.779 (0.378)	5.023** (0.026)	0.791*** (0.000)	0.809*** (0.000)	2.311 (0.129)	6.375** (0.012)	−1.075*** (0.000)	−1.175*** (0.000)
Slovenia (12 lag)	−7.594***	14.161***	0.479 (0.489)	7.730*** (0.006)	0.822*** (0.000)	0.839*** (0.000)	1.612 (0.205)	22.567*** (0.000)	−0.796*** (0.000)	−0.892*** (0.000)
Spain (12 lag)	−7.608***	13.397***	11.891*** (0.001)	21.431*** (0.000)	0.695*** (0.000)	0.741*** (0.000)	18.208*** (0.000)	8.137*** (0.005)	−0.827*** (0.000)	−1.008*** (0.000)
Sweden (10 lag)	−4.276***	3.920	2.353 (0.126)	0.830 (0.363)	0.794*** (0.000)	0.713*** (0.000)	0.088 (0.767)	1.185 (0.277)	−0.764*** (0.003)	−0.725** (0.015)
Turkey (10 lag)	−5.660***	7.594***	0.026 (0.871)	5.239** (0.023)	0.718*** (0.000)	0.723*** (0.000)	0.934 (0.334)	0.003 (0.954)	−0.706*** (0.000)	−0.782*** (0.000)
UK (12 lag)	−4.539***	6.081**	0.712 (0.399)	17.013*** (0.000)	0.793*** (0.000)	0.871*** (0.000)	1.271 (0.260)	0.103 (0.749)	−0.756*** (0.000)	−1.086*** (0.001)

Sources: The authors. Bloomberg Database and the Turkish Energy Market Regulatory Authority.

Notes: Numbers in parentheses are p-values. ***, ** and * Indicate significance at 1%, 5% and 10%, respectively. (1). BDM test statistic. (2). PSS test statistic. (3). Wald test statistic for long-run asymmetry. (4). Wald test statistic for short-run asymmetry. (5). Long-run coefficient of positive partial sum processes. (6). Long-run coefficient of negative partial sum processes. (7). Wald test for short-run asymmetry is not applicable due to estimated model parameters.

Following Shin et al. (2014), we employed t_BDM and F_PSS test statistics for the existence of a long-run relationship among the variables. The t_BDM test statistic, according to Banerjee et al. (1998), is compared to the asymptotic critical value bounds computed by Pesaran et al. (2001), with k = 2 (number of explanatory variables). The t_BDM test statistics given in panel A of Table 2 indicate the existence of a long-run relationship at the 1% significance level among diesel price, crude oil price and the exchange rate for all the countries (only Malta is at the 5% significance level). Similarly, the F_PSS test statistics of Pesaran et al. (2001) displayed in panel A of Table 2 suggest the existence of cointegration relationships at several significance levels among the variables in all countries except Estonia and Sweden. These findings are in line with those of Berument et al. (2014), Bagnai and Ospina (2015), Bagnai and Ospina (2016) and Ogbuabor et al. (2018).

Panel B of Table 2 reports the long- and short-run asymmetry properties of diesel prices in response to crude oil price changes. W_LR test statistics for long-run asymmetry, given in the first column of panel B, indicate that diesel prices have significant asymmetric responses to crude oil price changes in Austria, Belgium, Bulgaria, Cyprus, Denmark, Germany, Hungary, Latvia, Lithuania, Malta, the Netherlands, Romania and Spain in the long run. Diesel prices for the remaining countries do not show the asymmetric adjustment. Long-run asymmetry is consistent with the findings of Bagnai and Ospina (2016) for 12 Eurozone countries (Austria, Belgium, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, the Netherlands, Portugal and Spain).

The additive short-run asymmetry, denoted by the W test statistics, is reported in the second column of panel B. In the short run, diesel prices in Austria, Belgium, Bulgaria, Cyprus, Denmark, Finland, Germany, Greece, Ireland, Italy, Latvia, Lithuania, Poland, Romania, Slovakia, Slovenia, Spain, Turkey and the UK move asymmetrically against crude oil price changes at various significance levels. For the remaining countries, asymmetry is not observed in the short run. Short-run symmetry is in line with the evidence supported by Bagnai and Ospina (2016). The long-run coefficients of positive and negative changes in crude oil price, denoted by the $L^{+}$ and $L^{-}$ , are reported in the third and fourth columns of panel B of Table 2, respectively. All coefficients are found to be positive at the 1% significance level, pointing to a long-run positive relationship between diesel prices and crude oil prices. In the literature, partially parallel to our findings, Apergis and Vouzavalis (2018), Polemis and Fotis (2015), Kristoufek and Lunackova (2015), Clerides (2010), Grasso and Manera (2007) and Galeotti et al. (2003) reported mixed evidence of price asymmetry in several European countries.

As for the long- and short-run asymmetry properties of diesel prices with respect to exchange rate changes, the corresponding statistics are given in panel C of Table 2. The W_LR test statistics given in panel C document the long-run asymmetry for Austria, Belgium, Bulgaria, Cyprus, Denmark, Germany, Hungary, Latvia, Lithuania, the Netherlands, Poland, Romania and Spain in response to exchange rate shocks. In contrast, an asymmetric adjustment is not detected for the remaining countries in our sample. Likewise, parallel to our findings, Bagnai and Ospina (2016) found asymmetric long-run adjustment in 12 Eurozone countries excluding France, the Netherlands and Portugal.

The additive short-run asymmetric W_SR test statistics are reported in the second column of panel C. Accordingly, while asymmetric responses of diesel prices to exchange rate shocks were observed in the Czech Republic, Finland, France, Germany, Greece, Hungary, Italy, Latvia, Malta, the Netherlands, Poland, Romania, Slovakia, Slovenia and Spain, asymmetry could not be observed in the remaining countries. Confirming our findings, Bagnai and Ospina (2016) found evidence of symmetric short-run gasoline price adjustment in response to exchange rate shocks in 12 Eurozone countries. $L^{+}$ and $L^{-}$ statistics, the corresponding positive and negative long-run coefficients of exchange rate shocks, are given in the third and fourth columns of panel C, respectively. All coefficients are found to be significant at the 1% level, except Bulgaria, Malta and Sweden, for which significance levels are 5%, 10% and 5%, respectively. The sign of long-run coefficients indicates a negative relationship between diesel prices and the exchange rate, that is, a rise in the exchange rate (appreciation) leads to lower diesel prices, while a fall in the exchange rate (depreciation) increases diesel prices.

In order to explore the effect of changes in crude oil prices and exchange rates on diesel prices over time, we employed dynamic asymmetric (or nonlinear) multipliers, given by Equation (5). Appendices A and B depict 80-period dynamic asymmetric multipliers for changes in crude oil prices and exchange rates, respectively. The solid (dashed) black line shows the dynamic multiplier when there is an increase (decrease) in the crude oil price (Appendix A) or exchange rate (Appendix B). The bold broken red line reports the differences between the solid and dashed black lines. The upper and lower light dotted red lines are the 90% confidence intervals of the differences. If the null hypothesis that the solid broken red line is equal to zero is rejected, then we can conclude that there is a positive or negative asymmetry for the relevant period.

Appendix A reports the dynamic multipliers that allow the assessment of the short- and long-run effects of crude oil price changes on diesel prices. In the short run, the effect of a rise in the price of crude oil overrides that of a decrease, suggesting the positive asymmetric behaviour of diesel prices. In the long run, however, positive asymmetry fades away and either it is insignificant or, when statistically significant, it is negative, thereby indicating asymmetry. In short, while positive crude oil price shocks are effective in the short run, this asymmetric impact alleviates them in the long run. In other words, the ‘rockets and feathers’ hypothesis is supported in the short run but not in the long run. This evidence is in line with the findings of Kristoufek and Lunackova (2015), Polemis and Fotis (2015), Bermingham and O’Brien (2011), Clerides (2010) and Galeotti et al. (2003).

The short- and long-run effects of exchange rate changes are displayed in Appendix B. As the local currency, that is, the US dollar exchange rate increases (the local currency appreciates), the price of diesel decreases and vice versa. In general, the figures in Appendix B show that exchange rate depreciation is dominant over appreciation in the short run. In the long run, however, the effect of exchange rate depreciation loses its strength to some degree but remains statistically significant or dies out. In that sense, long-run asymmetry conforms to the findings of Bagnai and Ospina (2016); however, our short-run asymmetric adjustment argument does not agree with the short-run symmetric adjustment findings of Bagnai and Ospina (2016). Table 3 summarizes the findings concerning the behaviour of diesel prices in response to crude oil price and exchange rate shocks in the short and long run, as determined from the dynamic multipliers.

Table 3.

Effects of Crude Oil Price and Exchange Rate Shocks on Diesel Prices in the Short and Long Run

	Crude Oil Price Shocks		Exchange Rate Shocks
Country	Short Run	Long Run	Short Run	Long Run
Austria	Positive asym***	Negative asym***		Positive asym***
Belgium	Positive asym***	Negitive asym*		Positive asym*
Bulgaria	Positive asym**	Negative asym***		Positive asym***
Cyprus	Positive asym***	Negative asym***		Positive asym***
Czech Rep.			Positive asym***
Denmark	Positive asym***	Negative asym***		Positive asym***
Estonia
Finland	Positive asym***		Positive asym***
France			Positive asym***
Germany	Positive asym***	Negative asym**	Negative asym***	Positive asym**
Greece	Positive asym***		Positive asym***
Hungary		Negative asym***	Positive asym***	Positive asym***
Ireland	Negative asym*
Italy	Positive asym***		Positive asym***
Latvia	Positive asym***	Negative asym**	Positive asym***	Positive asym***
Lithuania	Positive asym***	Negative asym***		Positive asym***
Malta		Negative asym**	Negative asym**
The Netherlands		Negative asym*	Positive asym***	Positive asym**
Poland	Positive asym***		Positive asym***	Positive asym**
Portugal
Romania	Positive asym***	Negative asym*	Positive asym***	Positive asym***
Slovakia	Positive asym**		Positive asym**
Slovenia	Positive asym***		Positive asym***
Spain	Positive asym***	Negative asym***	Positive asym***	Positive asym***
Sweden
Turkey	Positive asym**
The UK	Positive asym***

Sources: The authors. Bloomberg Database and the Turkish Energy Market Regulatory Authority.

Notes: ***, ** and * Indicate significance at 1%, 5% and 10%, respectively.

Despite having some mixed results, our findings overall support the ‘rockets and feathers’ phenomenon to the degree that diesel prices in the short run give a greater response to increases than decreases in crude oil prices. In the long run, either the impact of the increase in crude oil price dies out, or the influence gets reversed. Compared to the evidence provided by Bagnai and Ospina (2016) for gasoline price adjustment in 12 Eurozone countries, our results contrast both in the short- and long-run analysis.

As for exchange rate shocks, depreciation of exchange rates is more dominant in the short run, suggesting escalating diesel prices, which is in line with the findings of Reilly and Witt (1998), Asplund et al. (2000), Galeotti et al. (2003), Berument et al. (2014) and Özmen and Akçelik (2017). In the long run, either the effect of depreciation remained significant with some degree of reduction, as Bagnai and Ospina (2016) found for 12 Eurozone countries, or it faded away. Only Germany and Malta presented negative short-run asymmetry in response to exchange rate shocks, as opposed to typical movements in other countries, meaning that in the short run, exchange rate appreciation has more influence over diesel prices than depreciation in Germany and Malta.

Conclusion

In this study, we explored the short- and long-run asymmetric adjustment properties of pre-tax diesel prices to crude oil price and exchange rate changes in 27 European countries using a NARDL model. Despite having mixed results, we found evidence supporting the ‘rockets and feathers’ phenomenon so that, in the short run, diesel prices respond more to crude oil price increases than to their decreases. In the long run, however, either this effect vanishes or the effect of decreases in crude oil prices on diesel prices exceeds that of the increases. As for exchange rate shocks, depreciation of the local currency against the US dollar has more effect on diesel prices than appreciation in the short and long run. Thus, although we have some mixed results, it can be generalized that diesel prices adjust asymmetrically in response to changes in crude oil prices and exchange rates in the short and long-run. As regards the size of the adjustment, the average long-run coefficients of positive and negative charges of the crude oil price are 0.721 and 0.762, whereas those of exchange rate are −0.908 and −1.073, respectively. Put differently, a 1% positive (negative) shock in crude oil price generates an almost 0.72% increase or 0.76% decrease in diesel prices, prevailing negative asymmetry in the long run. As for the exchange rate, pass-through to diesel prices, a 1% appreciation (depreciation) in exchange rate yields a 0.91% (1.07%) decrease (increase) in diesel prices, indicating positive asymmetry in the long run. Moreover, diesel prices are more sensitive to exchange rate changes than to crude oil price changes, and this result is the same across all countries considered. Thus, the long-term effects of exchange rate shocks might be more severe than those of crude oil shocks with regard to diesel price adjustments in 27 European countries.

Countries have more power to influence their exchange rates through their monetary and fiscal policies than the world crude oil prices. Since exchange rate asymmetry has a higher level of influence on pre-tax diesel prices than the crude oil asymmetries, exchange rate stability is an important factor in the asymmetric responses of diesel prices. Policy makers may adopt different policies to promote exchange rate stability. Moreover, the asymmetric effects of crude oil prices on pre-tax diesel prices are short-lived; either the asymmetry disappears or gets reversed in the long run. Thus, policymakers may not like to adjust their tax codes to eliminate the adverse effects of crude oil price volatilities.

Appendix A. Dynamic Multipliers for Crude Oil Price Changes.

Sources: The authors. Bloomberg Database and the Turkish Energy Market Regulatory Authority.

Appendix B. Dynamic Multipliers for Exchange Rate Changes.

Sources: The authors. Bloomberg Database and the Turkish Energy Market Regulatory Authority.

Footnotes

Acknowledgements

The authors are grateful to Anita Akkaş, Tara Sylvestre and the anonymous referees of the journal for their extremely useful suggestions to improve the quality of the article. The usual disclaimers apply.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The authors received no financial support for the research, authorship and/or publication of this article.

Notes

References

Al-Gudhea

, Kenc

, & Dibooglu

(2007). Do retail gasoline prices rise more readily than they fall? A threshold cointegration approach. Journal of Economics and Business , 59, 560–574.

Apergis

, & Vouzavalis

(2018). Asymmetric pass-through of oil prices to gasoline prices: Evidence from a new country sample. Energy Policy , 114, 519–528.

Asplund

, Eriksson

, & Friberg

(2000). Price adjustment by a gasoline retail chain. Scandinavian Journal of Economics , 102, 101–121.

Assane-Otoo

, & Schneider

(2015). Retail fuel price adjustment in Germany: A threshold cointegration approach. Energy Policy , 78, 1–10.

Atil

, Lahiani

, & Nguyen

(2014). Asymmetric and nonlinear pass-through of crude oil prices to gasoline and natural gas prices. Energy Policy , 65, 567–573.

Bachmeier

, & Griffin

(2003). New evidence on asymmetric gasoline price responses. Review of Economics and Statistics , 85, 772–776.

Bacon

(1991). Rockets and feathers: The asymmetric speed of adjustment of UK retail gasoline prices to cost changes. Energy Economics , 13(3), 211–218.

Bagnai

, & Ospina

(2015). Long and short-run asymmetries and hysteresis in the Italian gasoline market. Energy Policy , 78, 41–50.

Bagnai

, & Ospina

(2016). “Asymmetric asymmetries” in Eurozone markets gasoline pricing. Journal of Economic Asymmetries , 13, 89–99.

10.

Balke

, Brown

, & Yücel

(1998). Crude oil and gasoline prices: An asymmetric relationship? Economic and Finance Policy Review , 1, 2–11.

11.

Banerjee

, Dolado

, & Mestre

(1998). Error-correction mechanism tests for cointegration in a single-equation framework. Journal of Time Series Analysis , 19, 267–283.

12.

Bermingham

, & O’Brien

(2011). Testing for asymmetric pricing behavior in Irish and UK petrol and diesel markets. Energy Journal , 32(3), 1–27.

13.

Berument

M. H.

, Sahin

, & Sahin

(2014). The relative effects of crude oil price and exchange rate on petroleum product prices: Evidence from a set of Northern Mediterranean countries. Economic Models , 42, 243–249.

14.

Borenstein

, Cameron

, & Gilbert

(1997). Do gasoline prices respond asymmetrically to crude oil price changes? Quarterly Journal of Economics , 112, 305–339.

15.

Bremmer

, & Kesselring

(2016). The relationship between U.S. retail gasoline and crude oil prices during the Great Recession: ‘Rockets and feathers’ or ‘balloons and rocks’ behavior? Energy Economics , 55, 200–210.

16.

Brown

, & Yücel

(2000). Gasoline and crude oil prices: Why the asymmetry? Economic and Financial Review, 3rd Q , 23–29.

17.

Chen

, Finney

, & Lai

(2005). A threshold cointegration analysis of asymmetric price transmission from crude oil to gasoline prices. Economics Letters , 89, 233–239.

18.

Chen

, Polemis

, & Stengos

(2018). Re-examining the asymmetric gasoline pricing mechanism in EU: A panel threshold analysis . MPRA Paper No. 89575.

19.

Clerides

(2010). Retail fuel price response to oil price shocks in EU countries. Cyprus Economic Policy Review , 4(1), 25–45.

20.

Contín-Pilart

, Correljé

, & Palacios

(2009). Competition, regulation, and pricing behavior in the Spanish retail gasoline market. Energy Policy , 37, 219–228.

21.

Dickey

, & Fuller

(1979). Distribution of the estimators for autoregressive time series with a unit root. Journal of the American Statistical Association , 74(366), 427–431.

22.

Galeotti

, Lanza

, & Manera

(2003). Rockets and feathers revisited: An international comparison on European gasoline markets. Energy Economics , 25, 175–190.

23.

Godby

, Lintner

, Stengos

, & Wandschneider

(2000). Testing for asymmetric pricing in the Canadian retail gasoline markets. Energy Economics , 22, 349–368.

24.

Granger

, & Yoon

(2002). Hidden cointegration . San Diego: University of California.

25.

Grasso

, & Manera

(2007). Asymmetric error correction models for the oil–gasoline price relationship. Energy Policy , 35, 156–177.

26.

Greenwood-Nimmo

, & Shin

(2013). Taxation and the asymmetric adjustment of selected retail energy prices in the UK. Economics Letters , 121, 411–416.

27.

Honarvar

(2009). Asymmetry in retail gasoline and crude oil price movements in the United States: An application of hidden cointegration technique. Energy Economics , 31, 395–402.

28.

International Energy Agency. (2014). Energy supply security emergency response of IEA Countries 2014 . https://www.iea.org.

29.

Karagiannis

, Panagopoulos

, & Vlamis

(2015). Are unleaded gasoline and diesel price adjustments symmetric? A comparison of the four largest EU retail fuel markets. Economic Modelling , 48, 281–291.

30.

Katrakilidis

, & Trachanas

(2012). What drives housing price dynamics in Greece: New evidence from asymmetric ARDL cointegration. Economic Modelling , 29(4), 1064–1069.

31.

Kirchgassner

, & Kubler

(1992). Symmetric or asymmetric price adjustment in the oil market: An empirical analysis of the relations between international and domestic prices in the Federal Republic of Germany 1972–1989. Energy Economics , 14, 171–185.

32.

Kristoufek

, & Lunackova

(2015). Rockets and feathers meet Joseph: Reinvestigating the oil–gasoline asymmetry on the international markets. Energy Economics , 49, 1–8.

33.

Kushwah

S. V.

, & Siddiqui

A. A.

(2020). Relationship between oil price movements and stock returns of oil firms in oil importing economies. Global Business Review , 1–17. https://dx-doi-org-s.web.bisu.edu.cn/10.1177/0972150920917018.

34.

MacKinnon

(1996). Numerical distribution functions for unit root and cointegration tests. Journal of Applied Econometrics , 11(6), 601–618.

35.

Manning

(1991). Petrol prices, oil price rises and oil price falls: Some evidence for the UK since 1972. Applied Economics , 23, 1535–1541.

36.

Meyler

(2009). The pass-through of oil prices into euro area consumer liquid fuel prices in an environment of high and volatile oil prices. Energy Economics , 31(6), 867–881.

37.

Ogbuabor

, Orji

, & Anthony-Orji

(2018). Asymmetric adjustment of retail energy prices in Europe: Analysis of structural change and exchange rate dynamics. Journal of Renewable and Sustainable Energy , 10, 015904.

38.

Özmen

, & Akçelik

(2017). Asymmetric exchange rate and oil price pass-through in motor fuel market: A microeconometric approach. The Journal of Economic Asymmetries , 15, 64–75.

39.

Pal

, & Mitra

(2015). Asymmetric impact of crude price on oil product pricing in the United States: An application of multiple threshold nonlinear autoregressive distributed lag model. Economic Modelling , 51, 436–443.

40.

Pal

, & Mitra

(2016). Asymmetric oil product pricing in India: Evidence from a multiple threshold nonlinear ARDL model. Economic Modelling , 59, 314–328.

41.

Perdiguero-Garcia

(2013). Symmetry or asymmetry oil prices? A meta-analysis approach. Energy Policy , 57, 389–397.

42.

Pesaran

, & Shin

(1999). An autoregressive distributed lag modeling approach to cointegration analysis . In S. Strom, Econometrics and Economic Theory in the 20th Century: The Ragnar Frisch Centennial Symposium. Cambridge: Cambridge University.

43.

Pesaran

, Shin

, & Smith

(2001). Bounds testing approaches to the analysis of level relationships. Journal of Applied Econometrics , 16(3), 289–326.

44.

Phillips

, & Perron

(1988). Testing for unit roots in time series regression. Biometrika , 75, 335–346.

45.

Polemis

, & Fotis

(2014). The taxation effect on gasoline price asymmetry nexus: Evidence from both sides of the Atlantic. Energy Policy , 73, 225–233.

46.

Polemis

, & Fotis

(2015). Rent-seeking oligopolistic behaviour in European gasoline markets. Economics Bulletin , 35(1), 827–833.

47.

Polemis

, & Tsionas

(2016). An alternative semiparametric approach to the modelling of asymmetric gasoline price adjustment. Energy Economics , 56, 384–388.

48.

Polemis

, & Tsionas

(2017). Asymmetric price adjustment in the U.S gasoline industry: Evidence from Bayesian threshold dynamic panel data models. International Journal of Economics and Business Research , 24(1), 91–128.

49.

Qin

, Zhou

, & Wu

(2016). Revisiting asymmetric price transmission in the U.S. oil-gasoline markets: A multiple threshold error-correction analysis. Economic Modelling , 52, 583–591.

50.

Radchenko

(2005). Oil price volatility and the asymmetric response of gasoline prices to oil price increases and decreases. Energy Economics , 27, 708–730.

51.

Radchenko

, & Shapiro

(2011). Anticipated and unanticipated effects of crude oil prices and gasoline inventory changes on gasoline prices. Energy Economics , 33, 758–769.

52.

Rao

T. D.

, & Goyal

(2018). Global commodity and oil price movements, macroeconomic performance and challenges for an emerging economy: The Indian experience. Global Business Review , 19(3), 650–674.

53.

Reilly

, & Witt

(1998). Petrol price asymmetries revisited. Energy Economics , 20, 297–308.

54.

Schorderet

(2001). Revisiting Okun's law: An hysteretic perspective . San Diego: University of California.

55.

Schorderet

(2003). Asymmetric cointegration . Geneva: University of Geneva.

56.

Shin

, Yu

, & Greenwood-Nimmo

(2014). Modelling asymmetric cointegration and dynamic multipliers in a nonlinear ARDL framework. In Horrace

, Sickles

(Eds.), Festschrift in Honor of Peter Schmidt: Econometric methods and applications (pp. 281–314). New York: Springer.