To Pass (or Not to Pass) Through International Fuel Price Changes to Domestic Fuel Prices in Developing Countries: What Are the Drivers?
Author: Mr. Kangni R Kpodar and Patrick A. Imam

Contributor Notes

Author’s E-Mail Address: pimam@imf.org; kkpodar@imf.org
Publication Date:
25 Sep 2020
eISBN:
9781513557625
Language:
English
Keywords:
WP; price; pass-through; country; government; fuel price; pass-through coefficient; price gap; government price regulation; kerosene price; crude oil price dynamics; price differential; fuel pump price flexibility; fuel price increase; Fuel prices; Energy subsidies; Oil prices; Inflation; Consumption; Global
Download PDF (1.8 MB)

While many developing countries limit the international fuel price pass through to domestic fuel prices, others do not. Against this backdrop, we examine the factors that determine whether governments allow international fuel price changes to be passed through to domestic prices in developing countries using a dataset spanning 109 developing countries from 2000 to 2014. The paper finds that the pass-through is higher when changes in international prices are moderate and less volatile. In addition, the flexibility of the pricing mechanism allows for higher pass-through while exchange rate depreciation and lower retail fuel prices in neighboring countries inhibit it. The econometric results also underscore the fact that countries with inflation tend to experience lower pass-through, whereas those with high public debt exhibit larger pass-through. Finally, no evidence is found that political variables or environmental policies matter with regard to fuel price dynamics in the short-term. These findings, which are consistent across fuel products (gasoline, diesel and kerosene), allow us to draw important policy lessons for fuel subsidy reforms.

Abstract

While many developing countries limit the international fuel price pass through to domestic fuel prices, others do not. Against this backdrop, we examine the factors that determine whether governments allow international fuel price changes to be passed through to domestic prices in developing countries using a dataset spanning 109 developing countries from 2000 to 2014. The paper finds that the pass-through is higher when changes in international prices are moderate and less volatile. In addition, the flexibility of the pricing mechanism allows for higher pass-through while exchange rate depreciation and lower retail fuel prices in neighboring countries inhibit it. The econometric results also underscore the fact that countries with inflation tend to experience lower pass-through, whereas those with high public debt exhibit larger pass-through. Finally, no evidence is found that political variables or environmental policies matter with regard to fuel price dynamics in the short-term. These findings, which are consistent across fuel products (gasoline, diesel and kerosene), allow us to draw important policy lessons for fuel subsidy reforms.

I. Introduction

While international fuel prices are highly volatile (see Regnier, 2007 and Ftiti and Jawadi, 2020), their impact on domestic prices at the pump differ significantly from country to country. In some countries, international fuel price changes are fully passed through to retail prices with no delay. In other countries, especially developing ones, they move little or not at all, sometimes for a prolonged period, reflecting the fact that retail fuel prices are highly regulated. The lack of fuel price pass-through is the flip side of the fuel subsidies coin. The economic inefficiency implied by low fuel price pass-through and the consequent fuel subsidies are fiscally costly, inequitable and harmful for the environment (Clements et al., 2013). Despite these pitfalls, these subsidies have remained stubbornly high in many countries and proven extremely difficult to reform. Governments, especially in developing countries, have been hesitant to reduce them and to allow higher pass-through into domestic prices for fear of the political costs, including riots, given the potential impact on the poor.

Visually, Figure 1 illustrates the volatility of international gasoline prices across a number of years (mirroring crude oil price dynamics), with a sharp drop in 2008 followed by a strong recovery a year later, albeit with a certain deceleration in subsequent years until another sizeable drop is observed in 2014. The distribution of pass-through to retail gasoline prices2 indicates significant cross-country disparities. The median pass-through was below 100 percent (less than full pass-through) during half of the 2000–14 period, particularly in years recording a large swing in international gasoline prices.

Figure 1.
Figure 1.

Change in International Gasoline Price and Distribution of Pass-through to Retail Gasoline Price, 2000–14

Citation: IMF Working Papers 2020, 194; 10.5089/9781513557625.001.A001

Source: Authors’ calculationsNotes: See section III for the calculation of pass-through. A pass-through of at least 100 percent indicates that the nominal increase in international gasoline prices has been fully reflected in the prices of gasoline at the pump.

Despite a large body of literature on the economic impact of movements in fuel prices, more work is still required to fully understand the formation of domestic fuel prices. This is because studies analyzing the pass-through to domestic fuel prices do not systematically look at pass-through across countries and time – typically because of a lack of (consistent cross-sectional and long-term) data – and are instead selective in their analysis, focusing primarily on developed countries (see for instance Angelopoulou and Gibson, 2010; Bello and Contín-Pilart, 2012). Moreover, most studies tend to look at the level of fuel prices across countries rather than changes in prices over time, thereby missing important factors affecting the short-term dynamic of fuel prices (see Cheon et al., 2013 and Blair et al., 2017).

Our paper aims to fill this gap. It adds to the existing literature by systematically looking at the key factors that can impact international fuel price pass-through to domestic fuel prices both over time and across countries using a dataset spanning 109 developing countries. Unlike existing studies that focus on price levels and single countries, this paper focuses on dynamic price changes in a panel setting to better understand the role of macroeconomic and policy factors – such as exchange rate dynamics, inflation, public deficit and debt, current account deficit, oil price cycles and fuel price setting mechanisms – which other studies have largely neglected thus far. The focus is also on developing countries, where the issue of fuel subsidies is more prevalent.3 To this end, the paper calls on a unique global fuel price dataset (see Kpodar and Abdallah, 2017) that covers more countries and a longer time period than existing studies. From a policy-making perspective, understanding the mechanisms of and differences in how international fuel prices impact domestic fuel prices can shed light on why and in which circumstances countries may allow a higher pass-through, and in which circumstances they may not; thereby providing useful insights into enabling factors and barriers to fuel subsidy reforms.

The paper finds that macroeconomic factors play a critical role in the short-term dynamic of pass-through while little evidence is found for political factors, which are slow-moving and tend to be better suited to explaining cross-country differences in retail fuel price levels rather than their changes. More specifically, we find strong evidence that higher inflation discourages pass-through while high public debt is associated with higher pass-through. Large current account deficits and countries recording high carbon emissions do not necessarily exhibit larger pass-through. Furthermore, pass-through is higher during periods of moderate and less volatile change in international fuel prices and in countries with more flexible fuel price-setting mechanisms. In contrast, where fuel prices are already elevated and exchange rate depreciation occurs, pass-through tends to be lower. Finally, retail fuel price dynamics in neighboring countries matter.

The structure of the paper is as follows. Section II will provide a literature review, followed by Section III outlining the empirical strategy, including a discussion of the model, data and methodology. Section IV will discuss the main results and Section V will conclude with policy implications.

II. A Literature Review: What Explains Differences in Pass-Through

How changes in international crude oil prices impact domestic retail fuel prices is a question that the economic literature has looked at from a number of different angles. A first set of studies focuses on statistically documenting the trends and magnitude of pass-through across countries, while a second set pays attention to the explanatory factors of cross-country variations in fuel prices.

Studies in the first group include for instance Baig et al. (2007) who use a sample of 51 emerging and developing countries for the period 2003–2006 and find evidence of limited pass-through, reflecting a rise in fuel subsidies. Similarly, in an expanded study, Coady et al. (2010) provide estimates of pass-through coefficients for gasoline, diesel, and kerosene in 155 countries between the end of 2003 and mid-2008, again confirming that pass-through levels are in most cases low, but differ significantly across countries (see also Kojima, 2012 and Coady, Flamini and Sears, 2016). More recently, Kpodar and Abdallah (2017) investigate the degree of pass-through using a novel monthly dataset of retail fuel prices in 162 countries from January 2000 to December 2014, finding evidence of positive but varying pass-through levels across regions and income groups. The persistence of the pass-through also exhibits differences across samples. There is a large body of literature examining the asymmetric response of decreases and increases in the cost of crude oil, although the results are not uniform. Borenstein et al. (1997), Balke et al. (1998), together with Al-Gudhea et al. (2007) and Blair et al. (2017) for the U.S., Bacon (1991) for the U.K., and Grasso and Manera (2007) for France, Germany, Italy, Spain and the U.K. find relatively favorable evidence for the existence of the asymmetry (see also Kpodar and Abdallah, 2017), while Norman and Shin (1991), Bachmeier and Griffin (2003) for the U.S. and Godby et al. (2000) for Canada report a symmetric response of retail gasoline prices to world crude oil price changes (see also López-Villavicencio and Pourroy, 2019). In his well-designed meta-analysis, Perdiguero-Garcia (2013) illustrates that the conflicting results reported in empirical studies may be due to various factors ranging from the estimation method used to the time-frequency of data. The reasons for the mixed evidence include differences in data span, data frequency, market structure and estimation techniques.4 These studies, while empirically interesting, are limited in that they do not systematically explain why the pass-through coefficient varies across countries.

The second group of studies attempts to answer this why by explaining differences in domestic fuel prices between countries, primarily focusing their attention on political variables. By taking a comprehensive look at economic and political drivers of fuel pricing, van Beers and Strand (2013) find that higher pass-through was associated with high income per capita, the provision of public goods (captured by health expenditure), democratic institutions and a low level of corruption. Differences in pass-throughs also reflect budgetary needs as well as energy and environmental policies as argued by Angelier and Sterner (1990), and evidenced by Rietveld and van Woudenberg (2005) in a sample of European countries. Consistent with Hammar et al. (2004), van Beers and Strand (2013) find that high gasoline consumption levels lead to more gasoline subsidies (hence lower pass-through), thereby supporting the theory of pressure groups. The complexity of the political economy is the most challenging barrier to reforming subsidies (Cheon et al., 2013 and 2015; Rentschler and Bazilian, 2016; El-Katiri and Fattouh, 2012 and 2017). According to Fattouh and El-Katiri (2012), pass-through is low when there is a lack of institutions able to redistribute wealth (through other means) or achieve other social goals. This view is echoed by Cheon et al. (2013) who stress that consumption subsidies are a common political tool for luring voters in countries with weak institutional capacity and poor governance (see also Ebeke and Lonkeng Ngouana (2015) who provide evidence of a crowding-out effect of fuel subsidies on social expenditure). Against this backdrop, policymakers have little incentive to pursue sound economic policies.

Another key factor that the literature has identified is whether a country is a net exporter or importer of the type of energy that can be subsidized or taxed. For instance, countries exporting fossil fuels may see less of a reason to resist subsidizing fuels and passing on fuel price increases, for fiscal or energy security reasons, especially when such revenues are large (Beers and Strand, 2013). Moghaddam and Wirl (2018) examine fuel subsidies in oil exporting countries, noting that the factors behind fuel subsidies in these countries include an implicit export tax embedded in world oil prices that should not be charged to domestic consumers; the redistribution of oil wealth (see also Fattouh and El-Katiri, 2012); the high political costs of raising fuel prices; and pressure from lobbying groups (agricultural sector, industries and so on). The authors find empirically that subsidy elimination hinges on the fiscal position, external buffers, and the political and social environment.

The above studies significantly contribute to understanding differences in fuel prices between countries, but also share a common shortcoming. The approach in previous studies is appropriate when looking at long-term determinants of fuel price levels, but they do not consider key macro factors influencing policymakers’ decisions to let domestic fuel prices move in the short-term where they are government-controlled. In addition, many of these studies tend to focus on a select set of advanced countries where market forces, rather than government interventions, are the most important drivers of retail fuel prices.

III. The Model, Methodology and Data Description

A. The Sample and Measurement of the Pass-Through

In this paper, we focus on a sample of 109 developing countries with annual data during the period 2000–14. The country sample (Annex 1) and the period are constrained by data availability on domestic fuel prices as provided by Kpodar and Abdallah (2017). The rationale for focusing on developing countries is that fuel prices in those countries are for the most part heavily controlled by governments as opposed to advanced economies, where domestic fuel prices respond to market forces (supply and demand). A related argument is that since the drivers of domestic fuel prices may differ between developing and advanced economies, focusing on the former group reduces sample heterogeneity (thereby improving the reliability of the results) and allows the econometric model to be tailored to this specific group of countries.

The dependent variable, which is the degree of pass-through (PT), can be calculated by comparing the price changes at retail level with the evolution of world market prices. The formula is expressed as follows:

PTi,t=Retail Fuel Pricei,t−Retail Fuel Pricei,t−1Benchmark World Fuel Pricei,t−Benchmark World Fuel Pricei,t−1(1)

  • where (i) PTi,t is the pass-through coefficient of country i in year t. For a one dollar increase in the supply cost for fuel, PT denotes how many dollars or what fraction of a dollar is passed through to the consumer.

  • (ii) the retail fuel price is the price at which gasoline is sold to final consumers at the end of year t.

  • (iii) the benchmark world fuel price is the international price of gasoline at the end of year t.5

The end-of-year retail gasoline prices are extracted from a comprehensive database of monthly retail fuel prices compiled by Kpodar and Abdallah (2017). The price data, mainly collected from national authorities, represent the prevailing retail prices per liter of gasoline all tax inclusive and expressed in local currency. The fuel prices are then converted into US dollars using the average monthly exchange rate. The data on international gasoline prices, provided by the US Energy Information Administration, represent the average spot or future contract price for gasoline delivered at specific ports.6 Prices are expressed in USD.

Under the assumption that other elements involved in setting retail fuel prices are relatively stable (for instance transport costs, distribution costs and margins, see Annex 2 on the decomposition of the retail fuel price), any changes in international fuel prices that are not passed through to the end consumers would be absorbed by taxes, particularly in countries where the government regulates fuel prices. For a given change in international fuel prices, a pass-through coefficient lower than 100 percent is an indication that net fuel tax has been reduced, whereas a pass-through coefficient of more than 100 percent implies a higher net fuel tax. Similarly, a drop (respectively an increase) in the pass-through typically mirrors a loss (respectively a gain) in government tax revenue compared to the initial point.

The pass-through coefficient is a simple and straightforward metric to gauge fuel pricing policy, but there are two caveats. First, there might be a transmission lag from international gasoline prices to retail gasoline prices, reflecting fuel procurement processes, the frequency of fuel imports, and distance from markets. This implies that the denominator in equation 1 may not necessarily reflect changes in supply cost.7 But because the pass-through is measured over a one-year period in this analysis, this bias is likely to be minimal. Second, the presence of an ad valorem tax can overstate the pass-through, particularly for a large change in international fuel prices. While there is no obvious way to eliminate the effect of ad valorem taxes, the pass-through coefficient remains the best proxy for changes in fuel taxation, in line with the approach adopted in the existing literature (see for instance Kojima, 2012; and Coady, Flamini and Sears, 2016).

An initial look at the distribution of pass-through coefficients (Figure 2) points to “outliers” at both ends of the distribution.8

Figure 2.
Figure 2.

Distribution of Gasoline Price Pass-Through Coefficients, 2000–14

Citation: IMF Working Papers 2020, 194; 10.5089/9781513557625.001.A001

Source: Authors’ calculations

In a typical regression, one would exclude the extreme data points to mitigate potential biases in the regression coefficients. However, when dealing with pass-through coefficients, these data points can be quite informative. For instance, a high pass-through can occur when a country takes the opportunity of a period of stable international fuel prices to increase domestic fuel prices with a view to reducing subsidies. Rather than removing these potential outliers, we use a logistic transformation of the pass-through coefficients to “compress” extreme values into a more normal range, as follows:

f(PTi,t)=(1+e−aPti,t)−1(2)

where PTi,t is the pass-through coefficient, and a is a parameter defining the steepness of the curve. It is set at 0.01 (Figure 3) to achieve a quasi-linear transformation for about 95 percent of the distribution of pass-through coefficients (this parameter will be subject to a sensitivity test with the value of 0.001, corresponding to a quasi-linear transformation for about 99 percent of the distribution).

Figure 3.
Figure 3.

Logistic Transformation of Gasoline Pass-Through Coefficients, 2000–14

Citation: IMF Working Papers 2020, 194; 10.5089/9781513557625.001.A001

Source: Authors’ calculations

B. The Model Specification and Explanatory Variables

To explain the pass-through, the paper adopts the following model:

f(PTi,t)=α0+AXi,t+μi+ϵi,t(3)

where:

  • f(PTi,t) is the logistic transformation of the fuel price pass-through coefficient as defined above;

  • Xi,t is a set of variables including the fuel pricing mechanism, domestic and international fuel market conditions, macroeconomic and policy factors, as well as political and institutional factors that are likely to affect the pass-through.

  • μi is the country specific effect, εi,t is the error term, i represents the country and t the year.

Below we discuss how each of the control variables influences the pass-through, and the various mechanisms at play.9

Initial fuel price level and pricing mechanism. We hypothesize that countries with low initial fuel prices have more room to pass an increase in international fuel prices to consumers, and it is important to control for this in a cross-country setting since the focus of this paper is on short-term dynamics of fuel prices. The degree of pass-through is also closely linked to the price setting mechanism in place, which can range from ad hoc price changes to liberalized regimes. In the former, the government controls fuel prices, and adjusts them irregularly sometimes after a prolonged period of freeze. In the latter, fuel prices are freely determined by market forces with the role of the government restricted to setting fuel taxes and regulating the sector to ensure competition. In between lies the automatic pricing mechanism in which retail fuel prices are linked to international prices through a formula and revised at regular intervals.10 We expect more flexible price setting mechanisms (automatic pricing and liberalized regimes) to be associated with a higher pass-through. Given the lack of information and the difficulty in determining the pricing regime at a given point of time for a country, we use the frequency of price changes over a year as a proxy.11

International fuel price. The pass-through is expected to be negatively associated with changes in international fuel prices as governments would be more reluctant to pass on sizeable increases in supply costs to cushion the impact on the final consumers and the economy in general. Moreover, high volatility in international fuel prices may prompt a wait-and-see approach to fuel pricing, thus reducing the pass-through. The proponents of fuel price controls argue that government price regulation prevents this volatility from disrupting economic activities, although the benefit is not clear as the volatility is simply transferred to the budget, which in turn affects the economy. As underlined by Regnier (2007), the conventional wisdom that oil prices have been more volatile than prices of most other commodities has been used to justify price and allocation controls and energy efficiency subsidies. Taking the above reasons into consideration, the model includes the changes in international gasoline prices and the associated volatility (measured by the standard deviation over a year based on monthly data) as explanatory variables. To account for the high volatility in crude oil prices during the global financial crisis, which led to a massive drop in crude oil process (70 percent decline) in the second half of the year, after crude oil prices rallied in the first half by about 50 percent, a dummy variable is introduced.12

Price gap with neighboring countries. It is well known that large price differentials can give rise to fuel smuggling. One example is Nigeria, where heavily subsidized fuel finds its way to many neighboring countries such as Benin, Togo, Niger and Cameroon, where retail fuel prices are many times higher (smuggled fuels also stem from illegal pipeline tapping). According to Ralby (2017), fuel smuggled out of Nigeria accounts for a staggering 30 percent of all hydrocarbon products produced in the country. In advanced economies, price differentials have also encouraged cross-border “fuel tourism”, with significant tax losses for the countries with higher retail fuel prices (see Banfi, Filippini and Hunt, 2005). The challenge for the country with higher fuel prices is the limited room for higher pass-through, considering the risk of widening the price gap and thus creating more incentives for fuel tourism or for consumers to switch to the black market. The model accounts for this effect by considering the difference between the retail gasoline price in each country with the average retail gasoline price in neighboring countries.13 The higher the price gap, the lower the pass-through would be in the country with the higher gasoline price. On the other hand, in the country with the lower gasoline price, pressure to increase the pass-through may build up to curtail subsidy leakage.

Other macroeconomic and policy factors. These include the level of development, exchange rate, inflation rate, government budget balance, current account balance, business cycle, oil rent, share of fuel in total consumption basket and the level of CO2 emissions. How they can affect pass-through is discussed below.

Income per capita: While admittedly an (imperfect) proxy for economic development, most studies assume that a higher income per capita is expected to lead to higher pass-through. One reason is that energy costs represent a smaller proportion of total expenses for the average consumer in higher income countries, meaning that subsidies are less important as an instrument of income support. In addition, higher-income nations generally have better alternative transfer mechanisms, reducing the incentive for politicians to “buy votes” through fossil fuel subsidies. But a counterargument could be that richer countries can afford more fuel subsidies and thus lower the pass-through.

Exchange rate. As previously discussed, exchange rate depreciation can magnify international fuel price shocks, thus prompting governments to reduce pass-through in a move to cushion the impact on consumers and productive sectors. If this mechanism dominates, we should expect the coefficient of the nominal exchange rate expressed in local currency per dollar to be negatively correlated with pass-through.

Inflation rate. By having large secondary effects through their impact on transportation costs, fuel price increases can have a significant impact on inflation. Policy makers are often concerned about high inflation eroding the purchasing power of consumers—particularly among the poor—and dampening economic growth. If inflation expectations are not well-anchored, higher inflation can feed through to wage increases and in turn push prices up further. One could, therefore, expect pass-through to be lower when inflation rises, as governments may worry that the potential inflationary impact of higher fuel prices might trigger a vicious inflationary cycle. The inflation rate is lagged by one period to ensure that it does not capture the feedback effect from the change in fuel prices.

Fiscal balance. Since the cost of an incomplete pass-through is, one way or another, ultimately borne by the budget, the ability of the government to maintain a low pass-through depends on the available fiscal space. Therefore, we expect the pass-through to be higher when the budget balance deteriorates. Public debt level is also used as an indicator of fiscal space. The fiscal balance and public debt as a ratio of GDP are lagged by one period because their contemporaneous value can be affected by the pass-through.

Current account balance. A high current account deficit is expected to be associated with a higher pass-through if it cannot be financed sustainably. The rationale is that higher fuel prices should help contain fuel consumption and therefore reduce pressure on the current account for fuel importing countries. In fuel exporting countries, lower domestic demand for fuels stemming from higher fuel prices should stimulate fuel exports and hence improve the current account balance. Taking the feedback effect from the pass-through to the current account balance into account, we incorporate the current account balance ratio (in percentage of GDP) into the model rather than the contemporaneous value.

Business cycle. Pass-through is likely to be higher during boom times when household incomes are growing, as the households can afford a higher cost of living. On the other hand, policy makers may be tempted to reduce pass-through during recessions to shield households from higher fuel prices and productive sectors from rising input costs. To test this hypothesis, we incorporate GDP growth rate into the model.

Oil rents. If a country has a fuel endowment and exports oil, the pass-through is likely to be lower. Although the opportunity costs of subsidizing domestic fuel consumption rise with the world market fuel price, so does the political cost of having higher domestic fuel prices.14 Often, the perception exists that the oil “belongs to the population” leading to a populist pressure to keep pass-through and fuel price levels low. The data on oil rents as a percent of GDP are drawn from the World Development Indicators WDI.

Share of fuel in total consumption. Pass-through may not follow the same pattern across different fuel products, even though international fuel prices co-move closely. This is often due to cross-subsidies or “targeted” subsidies. For instance, it is common in many developing countries that pass-through for diesel be lower than that of gasoline because it accounts for a higher share of domestic fuel consumption and is primarily consumed in the transport sector, the cost of which can affect food prices. Hammar, Löfgren and Sterner (2004) underline the fact that a high level of fuel consumption can lead to considerable pressure against raising fuel taxes. The importance of a fuel product is proxied by its share in total domestic fuel consumption. The data are provided by the International Energy Agency (IEA).

CO2 emissions. In light of the increasing awareness of the environmental consequences of fossil fuel consumption and the harmful impact on public health, countries with higher CO2 emissions may also exhibit higher pass-through as an attempt to curb these negative externalities (see Kaufmann, 2019; Meléndez-Jiménez and Polanski, 2020). Data on CO2 emissions stemming from the burning of fossil fuels are provided by the World Development Indicators Database and expressed in kg per PPP $ of GDP.

Political and institutional determinants. The degree to which a country is democratic can influence fuel price pass-through. Non-democratic or authoritarian regimes are typically plagued with weaker capacity, lower provision of public goods and more widespread corruption. They are averse to changes, including from volatile world oil prices, and often see energy subsidies as a substitute for providing public goods; as a result, low pass-through is likely to be prevalent in such regimes. On the other hand, in both non-democratic and democratic regimes, election years can coincide with downward pressure on fuel prices as incumbent governments seek to secure re-election. The existence of a large urban population can also deter higher pass-through, as the largest beneficiaries of fuel subsidies are in urban areas and they can exert pressure on the government to keep pass-through low. Another factor not conducive to full pass-through is political instability, reflecting concerns that higher fuel prices can further exacerbate tensions.15 The range of political and institutional factors that can potentially affect pass-through is wide, and cannot therefore be covered comprehensively. Nevertheless, focusing on a subset would make it possible to investigate in detail their importance in explaining cross-country variations of pass-through. In this paper, we use the Polity score as a measure of the level of democracy, the share of urban population in total population as a proxy of the size of pressure groups, a dummy variable for election years and finally an indicator of ethnic tensions to gauge the likelihood of political instability.16

C. Methodology

The model is estimated with a fixed-effect estimator. It allows us to take country heterogeneity into account by controlling for unobservable time-invariant country characteristics that may drive pass-through. Due to missing data, countries have an average of about 9 data points available during the period of study (2000–14). This time dimension is still relatively small compared to the cross-country dimension (109 countries), implying that a standard panel estimator, such as the fixed-effect estimator, applies. Nevertheless, it is important to address two potential issues: stationarity and endogeneity. First, we run a panel unit root test for all variables individually to ensure that they are stationary. The result of the Fisher Test (see appendix 3) suggests that the null hypothesis of unit root is rejected for all variables, except for the share of gasoline in total domestic fuel consumption and the Polity score. Consequently, the first difference of these variables, which eliminates the unit root, is included in the model. Second, it is possible that there is a feedback effect from the pass-through to some of the right-hand side variables. For instance, a higher pass-through may lead to higher inflation and reduce government subsidies, thereby translating into a lower budget deficit. Given the challenges in identifying proper instruments for the relevant variables, we mitigate the issue of simultaneous bias by lagging these variables by one period.

IV. Main Results

Table 1 presents the first set of results using gasoline prices as the dependent variable and focusing on the lagged retail gasoline price (to capture the initial level), the flexibility of the pricing mechanism, movements in international gasoline prices, the exchange rate and the price gap with neighboring countries as covariates. Overall, the results are robust with the coefficients for most variables being significant at the 1 percent level. We find that countries with higher initial gasoline prices tend to experience lower pass-through (column 1 to 5, table 1). Also, in line with expectations, the higher the frequency of changes in retail gasoline prices – a proxy of the flexibility of fuel pricing – the higher the pass-through, as evidenced by the positive and highly significant coefficient in all columns of table 1. Moreover, the results consistently support a strong negative correlation across all specifications between changes in the international gasoline price and the degree of pass-through to the retail gasoline price. This is not surprising as fuel subsidies often emerge during periods of sharp increases in international fuel prices, precisely reflecting incomplete pass-through. The coefficient for the exchange rate vis-à-vis the dollar (significant in column 1 to 5, table 1) provides evidence that countries tend to reduce pass-through in periods of currency depreciation as fuel supply costs expressed in local currency rise.

Table 1.

Drivers of Gasoline Price Pass-Through: Testing for Domestic and International Oil Market Factors, 2000–2014.

article image
Notes. Robust standard errors (clustered by country) in brackets. *,**,*** denote significance at 10%, 5% and 1%, respectively. Retail fuel price flexibility measures the flexibility of the fuel pricing mechanism, proxied by the number of months (within a year) during which retail fuel prices changed relative to the previous month, divided by 12. Fuel price gap with neighbors is the domestic fuel price minus the average fuel price in neighboring countries for the same product. Volatility in international fuel price is calculated as the standard deviation of international fuel prices over a year using monthly data. Gasoline share in total fuel consumption (lagged) is first-differenced as it is non-stationary in level.

Interestingly, the pass-through in a given country turns out to be sensitive to the price gap with neighboring countries (the coefficient is negative and significant in two out of three specifications, column 4 and 5 in table 1). A higher price gap discourages pass-through in the country with the higher retail gasoline price, confirming the pervasive negative externalities (e.g. tax revenue loss) from fuel subsidies not only in the country bearing the cost of subsidies but also in neighboring countries. As discussed above, the inability of the country with the higher gasoline price to achieve full pass-through is likely due to the increased risk of smuggling activities this move could trigger. But even in the absence of fuel smuggling, policymakers may be subject to pressure from the public with regard to why domestic fuel prices are much higher than in neighboring countries, and could therefore be tempted to limit pass-through. We further investigate the impact of the gasoline price gap on pass-through by breaking it down into positive and negative price gaps (when the gasoline price in the home country is lower than the price in neighboring countries). The results (column 3, table 1) reveal that the pressure to contain pass-through tends to increase in the country with higher gasoline prices, but in the country with lower gasoline prices, the negative price gap does not necessarily lead to an increase in pass-through. This implies that heavy subsidizers would not necessarily hike pass-through to close the gap with neighboring countries, probably because curbing smuggling through tighter border controls may be politically less costly than reforming fuel subsidies.

Higher volatility in international gasoline price tends to be associated with lower pass-through, as expected (column 5, table 1). This may reflect a wait-and-see approach to fuel pricing, but could also be a sign of fuel price smoothing. The result in column 4 (table 1) offers no evidence that the share of gasoline in total fuel consumption determines pass-through, in contrast to the findings of Hammar, Löfgren and Sterner (2004). While governments may face pressure to contain the pass-through on the most commonly consumed fuel products (and consequently the one for which an increase in the price affects the economy the most), they may also be limited by the extent to which this is affordable. As the fiscal cost of fuel subsidies is a function of the unit subsidy and the quantities consumed, a small unit subsidy can prove to be very costly considering the quantities involved.

Moving to the second set of drivers of pass-through focusing on macro-policy variables, the results in Table 2 do not suggest a meaningful difference between pass-through in periods of expansion and downturn, as the coefficient on the real GDP growth rate is positive but not significant at a conventional level (column 1 and 6). On the other hand, macroeconomic factors such as the inflation rate and fiscal space play a role in driving pass-through (column 2, 3 and 6). Indeed, in a high-inflation environment, policymakers would tend to reduce pass-through, even though this strategy can prove to be ineffective due to imported inflation stemming from higher international fuel prices. Although a large primary balance seems to be associated with lower pass-through, this correlation is not significant. The public debt level, which reflects accumulated past years’ deficits, is nevertheless positively and significantly associated with higher pass-through, reflecting limited available fiscal buffers to absorb the cost of an incomplete pass-through. There seems to be no robust link between current account balance and pass-through, suggesting that increasing pass-through to dampen domestic fuel demand (and hence fuel imports) is not as common as was thought (column 4). When current account deficit is high, some countries might choose to reduce pass-through to boost exports while others might increase it to reduce fuel imports. It also emerges that the effect of CO2 emissions on pass-through is not as strong as expected (column 5), suggesting that negative externalities from fuel consumption are not being adequately factored into the pricing of fuel products, potentially due to the reluctance of policymakers to rely on carbon taxation to fight climate change (see Kaufmann, 2019).17 Consistent with van Beers and Strand (2013), income per capita correlates positively with pass-through, although not significantly, in all specifications.

Table 2.

Drivers of Gasoline Price Pass-Through: Macro and Policy Factors, 2000–2014.

article image
Notes. Robust standard errors (clustered by country) in brackets. *,**,*** denote significance at 10%, 5% and 1%, respectively. Retail fuel price flexibility measures the flexibility of the fuel pricing mechanism, proxied by the number of months (within a year) during which retail fuel prices changed relatively to the previous month, divided by 12.

Turning to political and institutional factors, the results are very weak (table 3). All the political and institutional factors display the correct sign, but in most cases are not statistically significant. This may suggest that political and institutional factors primarily tend to affect fuel price levels (as advocated by van Beers and Strand, 2013) rather than the short-term dynamics captured by the pass-through. In addition, part of the lack of significance of the political and institutional variables could also be attributed to country fixed effects that may capture the persistence of political and institutional factors given that re-estimating the regression using a pooling estimator (table 3, column 7) yields improved results as election years, oil rent and ethnic tensions18 are significantly associated with lower pass-through.

Table 3.

Political and Institutional Factors: Implications for Gasoline Price Pass-Through, 2000–2014.

article image
Notes. Robust standard errors (clustered by country) in brackets. *,**,*** denote significance at 10%, 5% and 1%, respectively. Retail fuel price flexibility measures the flexibility of the fuel pricing mechanism, proxied by the number of months (within a year) during which retail fuel prices changed relatively to the previous month, divided by 12. The ethnic tension score lies between 0 and 6, with a lower score for countries where tensions between opposing groups are high. The Polity variable is first-differenced as it is non-stationary in level.

Do diesel or kerosene prices tell a different story? This is worth investigating as different fuel products may be subject to different pricing policies, and cross-subsidization may also exit. To address this question, we rerun all the regressions in tables 1 to 3, considering diesel and kerosene as the fuel product of interest. The results, presented in tables A1 to A6 in appendix 4, are comparable to that of gasoline. They again support the importance of fuel pump price flexibility, international fuel prices, the exchange rate, the price gap with neighboring countries inflation rate and public debt as the main drivers of pass-through.

Robustness analysis.

To test the robustness of our results, we conduct a battery of tests. The above findings do not in any way depend on the logistical transformation of the pass-through variable to deal with extreme values. We demonstrate this in two ways. First, we set the value of the parameter defining the steepness of the logistic function to 0.001 instead of 0.01 as in the previous regressions (in other words, the extreme values of pass-through coefficients at both ends of the distribution are less compressed, see Figure 3). All specifications in tables 1 to 3 are rerun (see tables A7, A8 and A9 in appendix 4, respectively); the previous findings remain unchanged. Second, we drop the logistical transformation and rerun a set of specifications after excluding potential outliers defined as pass-through values that exceed the sample average by one standard deviation. Once again, this does not alter the previous findings to any significant extent (table A10 in appendix 4). Another robustness test consists in excluding oil exporting countries from the sample (15 of the109 countries). Re-running the regressions once again does not materially change the previous findings.19

V. Conclusion

Against the backdrop of the limited empirical literature on the dynamics of fuel prices, this paper attempts to shed light on the drivers of the pass-through of international fuel prices to domestic retail fuel prices. Using a dataset spanning 109 developing countries from 2000–2014, the paper finds that the pass-through is higher when changes in international prices are low, moderate and less volatile. Furthermore, the flexibility of the pricing mechanism allows for higher pass-through while exchange rate depreciation and lower retail fuel prices in neighboring countries inhibits pass-through. The results also underline the fact that countries with higher inflation tend to experience lower pass-through, whereas those with high public debt exhibit larger pass-through. Finally, we do not find evidence that political and environmental factors matters. These findings are consistent across fuel products, namely gasoline, diesel and kerosene.

The policy implications emerging from these findings can help draw important lessons for fuel subsidy reforms. Since price flexibility improves pass-through, an automatic pricing mechanism or fuel price liberalization, conditions permitting, may be a useful tool to contain fuel subsidies. Social considerations should nevertheless also be taken into account to avoid fuel price surges harming the most vulnerable part of the population. Second, the results imply that the current environment of relatively low international fuel prices provides a window of opportunity to undertake fuel subsidy reforms. Third, exchange rate depreciation and high inflation can undermine fuel subsidy reforms, considering their negative impact for pass-through. Thus, countries seeking to reform subsidies should target periods where macroeconomic conditions are relatively stable, which may provide tailwinds for this reform. While the lack of fiscal space can trigger higher pass-through, countries should not wait until public debt becomes unsustainable before undertaking fuel subsidy reforms, where this is an issue. Last, some regional coordination may be needed where the price gap with neighboring countries is a major concern.

For policymakers, the maxim that if things cannot continue forever they will stop has not been very helpful when it comes to fuel subsidies. Whereas in most countries around the world costs are visible and the policy is viewed as unsustainable, this has not led countries to implement subsidy reforms. Fuel subsidy reforms remain a major hurdle for most governments, even when they destabilize a country’s fiscal health. Our paper illustrates that, strategically speaking, there are times when subsidy reforms are easier, notably when international fuel prices are low. At present, with fuel prices and hence fuel subsidy costs being low, the time is ripe for reform, as the political costs are also low.

One possibility for extension and future research is to integrate the changes in global environmental policies taking place at the domestic and, more particularly, the global level. Following the COP-21 agreement in Paris in 2015, 197 countries have committed to limiting global temperature increases to 2°C by reducing GHG emissions caused by fossil fuels. Extending our dataset and taking the implementation of these policies in developing countries into account may help to further refine our conclusions. Such an analysis may show that environmental policies matter more and more when it comes to allowing for pass-through of international fuel prices to domestic ones.

References

  • Angelopoulou, E. and H. Gibson, 2010, “The Determinants of Retail Petrol Prices in Greece,” Economic Modelling, 27(6), pp. 1537–42.

    • Search Google Scholar
    • Export Citation

  • Angelier, J. P. and T. Sterner, 1990, “Tax harmonization for petroleum products in the EC,” Energy Policy, 18(6), pp. 500–505.

  • Bacon, R. W., 1991. Rockets and Feathers: The Asymmetric Speed of Adjustment of UK Retail Gasoline Prices to Cost Changes. Energy Economics 13, 211–218.

    • Search Google Scholar
    • Export Citation

  • Baig, T., A. Mati, D. Coady and J. Ntamatungiro, 2007, “Domestic Petroleum Product Prices and Subsidies; Recent Developments and Reform Strategies,” IMF Working Papers 07/71, International Monetary Fund.

    • Search Google Scholar
    • Export Citation

  • Banfi, S., M. Filippini and L. Hunt, 2005, “Fuel Tourism in Border Regions: The case of Switzerland,” Energy Economics, 27(5), pp. 689–707.

    • Search Google Scholar
    • Export Citation

  • van Beers, C. and J. Strand, 2013, “Political Determinants of Fossil Fuel Pricing,” Policy Research Working Paper 6470, The World Bank.

    • Search Google Scholar
    • Export Citation

  • Bello A. and I. Contín-Pilart, 2012, “Taxes, Cost and Demand Shifters as Determinants in the Regional Gasoline Price Formation Process: Evidence from Spain,” Energy Policy, 48, pp. 439–48.

    • Search Google Scholar
    • Export Citation

  • Blair, B, R. Campell and P. Mixon, 2017, “Price pass-through in US gasoline markets” Energy Economics, Vol. 65, pp.42–49

  • Chen, Shiu-Sheng, 2009, “Oil price pass-through into inflation,” Energy Economics, Vol. 31, pp.126–133

  • Cheon, A., M. Lackner and J. Urpelainen, 2015, “Instruments of political control: national oil companies, oil prices, and petroleum subsidies” Comparative Political Studies, Vol. 48, pp.370–402

    • Search Google Scholar
    • Export Citation

  • Cheon A., J. Urpelainen and M. Lackner, 2013, “Why do Governments Subsidize Gasoline Consumption? An Empirical Analysis of Global Gasoline Prices, 2002–2009,” Energy Policy, 56, pp. 382–390.

    • Search Google Scholar
    • Export Citation

  • Clements, B., D. Coady, S. Fabrizio, S. Gupta, T. Alleyne, and C. Sdralevich, eds., 2013, Energy Subsidy Reform: Lessons and Implications (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

  • Coady, D., J. Arze del Granado; L. Eyraud, H. Jin, V. Thakoor, A. Tuladhar and L. Nemeth, 2013, “Automatic Fuel Pricing Mechanisms with Price Smoothing : Design, Implementation, and Fiscal Implications” Technical Notes and Manuals No.2012/03 (Washington DC: IMF)

    • Search Google Scholar
    • Export Citation

  • De Gregorio, J., O. Landerretche, C. Neilson, C. Broda and R. Rigobon, 2007, “Another Pass-through Bites the Dust? Oil Prices and Inflation” Economía Vol. 7, pp. 155–208

    • Search Google Scholar
    • Export Citation

  • Ebeke C. and C. Lonkeng Ngouana, 2015, “Energy Subsidies and Public Social Spending; Theory and Evidence,” IMF Working Papers 15/101.

    • Search Google Scholar
    • Export Citation

  • El-Katiri, L. and B. Fattouh, 2012, “Energy Subsidies in the Arab World. Fossil Fuel Subsidies” European Parliament http://www.europarl.europa.eu/RegData/etudes/IDAN/2017/595372/IPOL_IDA(2017)595372_EN.pdf

    • Search Google Scholar
    • Export Citation

  • El-Katiri, L. and B. Fattouh, 2017 “A Brief Political Economy of Energy Subsidies in the Middle East and North Africa” pp.58–87 in G. Luciani, (Ed.), Combining Economic and Political Development: The Experience of MENA. (Brill: Leiden)

    • Search Google Scholar
    • Export Citation

  • Ftiti, Z. and F. Jawadi, 2020. “On the Oil Price Uncertainty,” The Energy Journal, (forthcoming).

  • Gangopadhyay, S., B. Ramaswami and W. Wadhwa, 2005 “Reducing subsidies on household fuels in India: how will it affect the poor?” Energy Policy, Vol. 33, pp. 2326–2336

    • Search Google Scholar
    • Export Citation

  • Gao, L, H. Kim, R. Saba, 2014, “How do oil price shocks affect consumer prices?” Energy Economics, Vol. 45, pp. 313–323

  • Gelos, G. and Y. Ustyugova, 2017, “Inflation responses to commodity price shocks – How and why do countries differ?” Journal of International Money and Finance, Vol. 72, pp. 28–47.

    • Search Google Scholar
    • Export Citation

  • Gillingham, R. and M. Keen, 2012, “Mitigation and Fuel Pricing in Developing Economies,” Chapter 6 in Fiscal Policy to Mitigate Climate Change. A Guide for Policymakers, edited by Ian W.H. Parry, Ruud de Mooij and Michael Keen, pp. 103–132 (Washington DC: IMF).

    • Search Google Scholar
    • Export Citation

  • Hammar, H., A. Löfgren and T. Sterner, 2004, “Political Economy Obstacles to Fuel Taxation,” Energy Journal, 25(3), pp. 1–17.

  • Karl, T., 1997, “The Paradox of Plenty: Oil Booms and Petro-states” (California: University of California Press)

  • Kaufmann, Robert, 2019, Pass-through of motor gasoline taxes: Efficiency and efficacy of environmental taxes, Energy Policy, Vol.125, pp. 207–215

    • Search Google Scholar
    • Export Citation

  • Kojima, M., 2012, “Oil Price Risks and Pump Price Adjustments,” Policy Research Working Paper No. 6227 (Washington: World Bank).

  • Kpodar, K. and C. Abdallah, 2017, “Dynamic Fuel Price Pass-Through: Evidence from a New Global Retail Fuel Price Database” Energy Economics, 66, pp.303–312.

    • Search Google Scholar
    • Export Citation

  • Labandeira, X., J. M. Labeaga and X. López-Otero, 2017, “A Meta-Analysis on the Price Elasticity of Energy Demand,” Energy Policy, Vol. 102, pp. 549–568

    • Search Google Scholar
    • Export Citation

  • LeBlanc, Michael and Chinn, Menzie David, 2004, “Do High Oil Prices Presage Inflation? The Evidence from G-5 Countries” UC Santa Cruz Economics Working Paper No. 561; (SCCIE Working Paper No. 04–04).

    • Search Google Scholar
    • Export Citation

  • López-Villavicencio, A. and M. Pourroy, 2019, “Inflation target and (a)symmetries in the oil price pass-through to inflation”, Energy Economics, Vol.80, pp. 860–875

    • Search Google Scholar
    • Export Citation

  • Meléndez-Jiménez, M. and A. Polanski, 2020, “Dirty neighbors — Pollution in an interlinked world” Energy Economics, Vol.86 (forthcoming).

    • Search Google Scholar
    • Export Citation

  • Moghaddam, H. and F. Wirl, 2018, “Determinants of oil price subsidies in oil and gas exporting countries” Energy Policy, Vol. 122, pp.409–420

    • Search Google Scholar
    • Export Citation

  • Ralby, I., 2017, “Downstream Oil Theft: Global Modalities, Trends, and Remedies”, Atlantic Council, Global Energy Center.

  • Rentschler, J and M. Bazilian, 2016, “Reforming fossil fuel subsidies: drivers, barriers and the state of progress” Climate Policy, Vol.3062, pp. 1–24

    • Search Google Scholar
    • Export Citation

  • Regnier, Eva, 2007, “Oil and Energy Price Volatility” Energy Economics, Vol. 29, pp.405–27

  • Rietveld, P. and S. van Woudenberg, 2005, “Why Fuel Prices Differ,” Energy Economics, 27(1), pp. 79–92.

  • Siddig, K., A. Aguiar, H. Grethe, P. Minor and T. Walmsley, 2014, “Impacts of removing fuel import subsidies in Nigeria on poverty” Energy Policy, Vol. 69, pp. 165–178

    • Search Google Scholar
    • Export Citation

  • Stinnett, D, J. Tir, P. Schafer, P. Diehl and C. Gochman, 2002, “The Correlates of War Project Direct Contiguity Data, Version 3.” Conflict Management and Peace Science, 19 (2), pp.58–66.

    • Search Google Scholar
    • Export Citation

Annex 1. Sample Composition

article image

Annex 2. Breaking Down the Retail Fuel Price

At the micro-level, the determinants of fuel pass-through by country can be broken down by looking at the value chain:

  • Crude oil prices: Crude oil is an important factor in explaining price movements. The price depends on the worldwide supply of crude oil, and the worldwide demand for it. The high inelasticity of supply and demand explains the large fluctuation in fuel prices—small changes on either side have a high impact on prices. Speculation and political tensions can also cause fuel prices to fluctuate sporadically. In principle, countries all pay roughly the same price for the fuel they purchase.

  • Processing cost: There are different types of crude oil in the world, and they vary greatly in composition, liquidity, gravity, sulfur content and other factors. The cost of processing crudes, therefore, varies based on these factors as well as on the refining capacities available. The diesel prices from US Gulf Coast are for ultra-low-sulfur diesel while in many Latin American markets, the sale is still of high-sulfur diesel.

  • Exchange rate: Fluctuations in the exchange rate impact the price of imported oil, which is traded in US dollars. Over the last 15 years, the adverse impact of rising fuel prices has often been exacerbated by a simultaneous depreciation against the USD strengthening the impact of fuel price declines in local currencies. Note that sometimes, particularly in remote areas given the lagged effect between purchase of fuel and delivery, the exchange rate impact may not be felt immediately.

  • International transportation costs: Depending on the location of the market, as well as the size of the market, there can be a large gap between import and export prices. A small, distant market in a volatile region (requiring higher insurance costs) will have higher transportation costs than a large market located close to the oil source that benefits from economies of scale in terms of transportation.

  • Domestic distribution costs: Once the fuel has entered the country, often via a port, it must be distributed across the territory. Costs include transportation, storage, distribution, marketing and retail sales. In a large country with poor infrastructure, local distribution costs should be higher. In addition, as part of the license, many countries require that the oil company provide a public service – force them to locate in areas where the market is shallow and make losses – allowing them to recoup the loss through cross-subsidizing from other activities, for instance. Other factors that can impact distribution costs include the size of the market – are there economies of scale in distribution – as well the level of competitiveness of the market.

  • Local taxation: In many AEs, such as Belgium for instance, it is common to impose significant excise, VAT and import taxes, among others, on the fuel consumed. In other countries, often producers of fuel such as Venezuela, prices are subsidized and below those on the international market. Much of the pricing differences paid by local consumers is dependent on national pricing policies. European countries have typically imposed high taxes – often a multiple of the price of fuel – while other countries, such as oil producing countries, often subsidize the fuel.

  • Other factors. The timing of fuel purchases is an additional factor. This is especially the case in small markets, where there are often lags between the purchase of fuel on the world market and the final retail price. Other countries have smoothing mechanisms in place, meaning that the transmission from international to domestic taxes occurs gradually over time.

Appendix 1. Summary Statistics and Correlation Matrix

Summary Statistics

article image

Correlation Matrix

article image
Notes. Grayed coefficients are significant at 1, 5 or 10% level.

Appendix 2. Variable Definitions and Sources

article image
article image

Appendix 3. Panel unit root test

article image
Notes. Fisher Test for panel unit root using an augmented Dickey-Fuller test (Ho: unit root); *,**,*** denote significance at 10%, 5% and 1%, respectively. A significant Chi2 statistic indicates that Ho is rejected, hence the variable is stationary. Variable in level means variable as defined in the left column.

Appendix 4. Additional Regressions

Table A1.

Drivers of Diesel Price Pass-Through: The Role of Domestic and International Oil Market Factors, 2000–2014

article image
Notes. Robust standard errors (clustered by country) in brackets. *,**,*** denote significance at 10%, 5% and 1%, respectively. Retail fuel price flexibility measures the flexibility of the fuel pricing mechanism, proxied by the number of months (within a year) during which retail fuel prices changed relative to the previous month, divided by 12. Fuel price gap with neighbors is the domestic fuel price minus the average fuel price in neighboring countries for the same product. The volatility in international fuel price is calculated as the standard deviation of international fuel prices over a year using monthly data.
Table A2.

Drivers of Diesel Price Pass-Through: Do Macro and Policy Factors Matter? 2000–2014

article image
Notes. Robust standard errors (clustered by country) in brackets. *,**,*** denote significance at 10%, 5% and 1%, respectively. Retail fuel price flexibility measures the flexibility of the fuel pricing mechanism, proxied by the number of months (within a year) during which retail fuel prices changed relative to the previous month, divided by 12.
Table A3.

Political and Institutional Factors: Implications for Diesel Price Pass-Through, 2000–2014

article image
Notes. Robust standard errors (clustered by country) in brackets. *,**,*** denote significance at 10%, 5% and 1%, respectively. Retail fuel price flexibility measures the flexibility of the fuel pricing mechanism, proxied by the number of months (within a year) during which retail fuel prices changed relative to the previous month, divided by 12. The ethnic tension score lies between 0 and 6, with a lower score for countries where tensions between opposing groups are high. Polity variable is first-differenced as it is non-stationary in level.
Table A4.

Domestic and International Oil Market Factors and their Impact on Kerosene Price Pass-Through, 2000–2014

article image
Notes. Robust standard errors (clustered by country) in brackets. *,**,*** denote significance at 10%, 5% and 1%, respectively. Retail fuel price flexibility measures the flexibility of the fuel pricing mechanism, proxied by the number of months (within a year) during which retail fuel prices changed relative to the previous month, divided by 12. Fuel price gap with neighbors is the domestic fuel price minus the average fuel price in neighboring countries for the same product. The volatility in international fuel price is calculated as the standard deviation of international fuel prices over a year using monthly data.
Table A5.

Impact of Macro and Policy Factors on Kerosene Price Pass-Through, 2000–2014

article image
Notes. Robust standard errors (clustered by country) in brackets. *,**,*** denote significance at 10%, 5% and 1%, respectively. Retail fuel price flexibility measures the flexibility of the fuel pricing mechanism, proxied by the number of months (within a year) during which retail fuel prices changed relative to the previous month, divided by 12.
Table A6.

Political and Institutional Factors and their Impact on Kerosene Price Pass-Through, 2000–2014

article image
Notes. Robust standard errors (clustered by country) in brackets. *,**,*** denote significance at 10%, 5% and 1%, respectively. Retail fuel price flexibility measures the flexibility of the fuel pricing mechanism, proxied by the number of months (within a year) during which retail fuel prices changed relative to the previous month, divided by 12. The ethnic tension score lies between 0 and 6, with a lower score for countries where tensions between opposing groups are high. Polity variable is first-differenced as it is non-stationary in level.
Table A7.

Drivers of Gasoline Price Pass-Through: Testing for Domestic and International Oil Market Factors and Alternative Logistic Function, 2000–2014

article image
Notes. Robust standard errors (clustered by country) in brackets. *,**,*** denote significance at 10%, 5% and 1%, respectively. In the above regressions, the logistic transformation of the dependent variable uses the value of 0.001 (compared to 0.01 for the main results) for the parameter a defining the steepness of the logistic function curve. Retail fuel price flexibility measures the flexibility of the fuel pricing mechanism, proxied by the number of months (within a year) during which retail fuel prices changed relative to the previous month, divided by 12. Fuel price gap with neighbors is the domestic fuel price minus the average fuel price in neighboring countries for the same product. The volatility in international fuel price is calculated as the standard deviation of international fuel prices over a year using monthly data. Gasoline share in total fuel consumption (lagged) is first-differenced as it is non-stationary in level.