## REFERENCES

## Source details

**Canonical URL:** [REFERENCES](https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020194-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2020/english/wpiea2020194-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2020/english/wpiea2020194-print-pdf.pdf.json)

---

### Introduction and scope
- Sample: 109 developing countries; annual data for the period 2000-14.
- Dataset:
  - Monthly retail fuel prices compiled by Kpodar and Abdallah (2017) aggregated to end-of-year retail prices.
  - International gasoline prices from the US Energy Information Administration (prices expressed in USD).
- Dependent variable: pass-through coefficient (PT) defined as
  - PT_{i,t} = (Retail Fuel Price_{i,t} − Retail Fuel Price_{i,t−1}) / (Benchmark World Fuel Price_{i,t} − Benchmark World Fuel Price_{i,t−1})
- Notes:
  - Median pass-through was below 100 percent during half of the 2000-14 period.
  - Average number of data points per country: about 9.
  - Logistic transformation used to compress extreme PT values: f(PT_{i,t}) = (1 + e^{−a PT_{i,t}})^{−1} with a = 0.01 (sensitivity test with a = 0.001).

### Literature and motivation
- Gaps addressed:
  - Focus on short-term dynamic pass-through across countries and over time.
  - Explicit inclusion of macroeconomic and policy drivers (exchange rate dynamics, inflation, public deficit and debt, current account, oil price cycles, fuel price-setting mechanisms).
- Rationale: international gasoline prices used rather than crude oil prices because most developing countries import refined fuels; international gasoline price better reflects import parity for refined products.

### Model, methodology, and key variables
- Empirical model (fixed effects): f(PT_{i,t}) = α0 + A X_{i,t} + μ_i + ε_{i,t}, where f(·) is the logistic transformation.
- Key explanatory variables (lagged where appropriate):
  - Initial retail fuel price (lagged)
  - Retail gasoline price flexibility (months price changed / 12)
  - Change in international gasoline price and its volatility (standard deviation over a year using monthly data)
  - Exchange rate (local currency unit per USD)
  - Fuel price gap with neighboring countries (domestic price − average neighbor price)
  - GDP per capita (log), real GDP growth (lagged), inflation rate (log, lagged)
  - Government primary balance (percent of GDP, lagged), gross public debt (percent of GDP, log, lagged), current account balance (percent of GDP, lagged)
  - Oil rents (percent of GDP), gasoline/diesel/kerosene share in total domestic fuel consumption, CO2 emissions (kg per PPP $ of GDP, lagged)
  - Political/institutional: Polity score (first-differenced), election-year dummy, urban population share, ethnic tension score
- Stationarity and endogeneity handling:
  - Fisher panel unit root tests; non-stationary variables first-differenced (gasoline share and Polity).
  - Lagging variables by one period to mitigate simultaneity.

### Main empirical findings (gasoline, diesel, kerosene; robustness across fuels)
- General: results are consistent across gasoline, diesel, and kerosene.
- Key relationships (selected coefficients from main regressions):
  - Table 1 (Drivers of Gasoline Price Pass-Through: domestic and international oil market factors):
    - Retail gasoline price (lagged): −0.173 [0.108].
    - Retail gasoline price flexibility: 0.213 [0.052]***.
    - Change in international gasoline price: −0.122 [0.016]***.
    - Exchange rate (LCU per USD): −0.499 [0.091]***.
    - Dummy year 2008: −0.123 [0.030]***.
    - Fuel price gap with neighbors (lagged): −0.125 [0.082] (in some specifications).
    - Volatility in international gasoline price (col 5): −0.009 [0.005]*.
    - Observations and coverage: observations = 947, 785, 785, 601, 601; countries = 107, 91, 91, 72, 72; R2 = 0.14, 0.17, 0.18, 0.17, 0.17.
  - Table 2 (Macro and policy factors):
    - Change in international gasoline price: −0.497 [0.091]*** (example col 1).
    - Exchange rate: −0.124 [0.017]***.
    - Inflation rate (log, lagged): −0.163 [0.031]***.
    - Gross public debt (percent of GDP, log, lagged): 0.044 [0.025]* and 0.063 [0.024]*** (examples).
    - CO2 emissions: example col 5 = 0.147 [0.144] (not robustly significant).
    - Observations and coverage: observations = 947, 911, 896, 930, 919, 857; countries = 107, 103, 101, 104, 105, 98; R2 = 0.14–0.19.
  - Table 3 (Political and institutional factors):
    - Political/institutional factors generally show correct signs but weak statistical significance in fixed-effect estimations.
    - Pooling estimator (column 7) yields significance for election years, oil rent, and ethnic tensions; example pooling change in international gasoline price = −0.584 [0.088]***.
    - Observations and coverage: fixed-effect columns observations = 795, 921, 930, 933, 707, 631; countries = 95, 102, 105, 105, 78, 75; pooling: observations = 631; countries = 75; R2 ≈ 0.16–0.19.
- Interpretation of main effects:
  - Higher retail price flexibility → higher pass-through.
  - Large or volatile increases in international fuel prices → lower pass-through (governments cushion consumers).
  - Exchange rate depreciation → lower pass-through.
  - Higher inflation → lower pass-through.
  - Higher public debt → higher pass-through (limited fiscal buffers).
  - Price gap with neighbors: higher domestic price relative to neighbors → lower pass-through (smuggling and political pressures).
  - Political and institutional variables better explain cross-country price levels than short-term pass-through dynamics.

### Robustness checks
- Logistic transformation steepness: results stable when a = 0.001 versus a = 0.01.
- Alternative approaches:
  - Dropping the logistic transformation and excluding outliers (pass-through values exceeding sample average by one standard deviation) does not materially change findings.
  - Excluding oil-exporting countries (15 of 109): results remain materially unchanged.

### Policy implications and recommendations
- Pricing mechanism:
  - Increase retail price flexibility (automatic pricing mechanism or liberalization) to improve pass-through and contain subsidies, conditional on country circumstances and social protections.
- Timing of reforms:
  - Use periods of relatively low and less volatile international fuel prices as a window of opportunity to reform fuel subsidies.
- Macroeconomic environment:
  - Avoid undertaking subsidy reforms during periods of exchange rate depreciation and high inflation, which can undermine pass-through and raise political costs.
- Fiscal considerations:
  - Do not wait until public debt is unsustainably high to undertake subsidy reforms; limited fiscal space may force higher pass-through at politically costly times.
- Regional coordination:
  - Consider regional coordination when significant price gaps with neighboring countries drive smuggling and undermine reform incentives.
- Social sensitivity:
  - Complement pricing reforms with social measures to shield vulnerable populations from fuel price surges.

### Annexes and Appendix highlights
- Annex 1 — Sample Composition: lists countries across regions (Sub-Saharan Africa, Europe and Central Asia, Latin America and the Caribbean, East Asia and Pacific, Middle East and North Africa, South Asia).
- Annex 2 — Breaking Down the Retail Fuel Price: crude oil prices, processing cost, exchange rate, international transportation costs, domestic distribution costs, local taxation, timing and price-smoothing mechanisms.
- Appendix 1 — Selected summary statistics (exact sample statistics):
  - Gasoline price pass-through (percent): observations 916; Mean 91.41; Standard deviation 504.44; Minimum -6596.09; Maximum 3240.47.
  - Diesel price pass-through (percent): observations 967; Mean 133.54; Standard deviation 703.34; Minimum -4619.34; Maximum 8001.58.
  - Kerosene price pass-through (percent): observations 692; Mean 92.35; Standard deviation 553.02; Minimum -4908.87; Maximum 7168.78.
  - Retail gasoline price level: observations 922; Mean 1.00; Standard deviation 0.42; Minimum 0.01; Maximum 2.61.
  - Gasoline price flexibility: observations 967; Mean 0.54; Standard deviation 0.40; Minimum 0.00; Maximum 1.00.
  - Change in international gasoline price: observations 967; Mean 0.12; Standard deviation 0.37; Minimum -0.58; Maximum 0.94.
  - Volatility of international gasoline price: observations 967; Mean 0.75; Standard deviation 2.64; Minimum 0.00; Maximum 33.00.
  - Log of GDP per capita: observations 967; Mean 7.42; Standard deviation 1.14; Minimum 4.66; Maximum 9.72.
  - Exchange rate (LCU per USD): observations 967; Mean 0.03; Standard deviation 0.12; Minimum -0.98; Maximum 0.98.
  - Inflation rate (log): observations 944; Mean 1.07; Standard deviation 0.06; Minimum 0.90; Maximum 1.62.
  - Public debt (percent of GDP): observations 945; Mean 3.72; Standard deviation 0.63; Minimum 1.83; Maximum 5.65.
  - Polity variable: observations 812; Mean 3.53; Standard deviation 5.50; Minimum -9.00; Maximum 10.00.
  - Election year dummy: observations 941; Mean 0.15; Standard deviation 0.36; Minimum 0.00; Maximum 1.00.
  - Oil rents (percent of GDP): observations 957; Mean 3.33; Standard deviation 8.45; Minimum 0.00; Maximum 57.57.
  - Ethnic tension score: observations 719; Mean 3.71; Standard deviation 1.21; Minimum 0.67; Maximum 6.00.
- Appendix 2 — Variable definitions and data sources: precise definitions for pass-through measures, price flexibility, price gaps, volatility measures, macro and political variables; sources include Kpodar and Abdallah (2017), US EIA, IEA, World Development Indicators, International Financial Statistics, Polity IV, Database of Political Institutions, International Country Risk Guide.
- Appendix 3 — Panel unit root test (Fisher Test using augmented Dickey-Fuller):
  - Gasoline price pass-through: Chi2 statistic 288.0 ***.
  - Retail gasoline price: Chi2 statistic 223.7 *.
  - GDP per capita (log): Chi2 statistic 267.0 **.
  - Retail gasoline price flexibility: Chi2 statistic 285.9 ***.
  - Change in international gasoline price: Chi2 statistic 332.6 ***.
  - Exchange rate (LCU per USD): Chi2 statistic 354.9 ***.
  - Volatility in international gasoline price: Chi2 statistic 227.5 *.
  - Gasoline price gap with neighbors: Chi2 statistic 503.4 ***.
  - Gasoline share in total fuel consumption: level Chi2 statistic 165.2; first difference Chi2 statistic 458.0 ***.
  - Real GDP growth rate: Chi2 statistic 578.1 ***.
  - Inflation rate (log): Chi2 statistic 605.1 ***.
  - Government primary balance: Chi2 statistic 368.1 ***.
  - Gross public debt (log): Chi2 statistic 260.5 **.
  - Current account balance: Chi2 statistic 462.1 ***.
  - CO2 emissions: Chi2 statistic 298.1 ***.
  - Polity variable: level Chi2 statistic 199.1; first difference Chi2 statistic 331.0 ***.
  - Oil rents: Chi2 statistic 120.7 *.
  - Urban population share: Chi2 statistic 405.8 ***.
  - Ethnic tension score: Chi2 statistic 363.4 ***.

### Appendix 4 — Additional regressions (selected robustness and fuel-specific results)
- Diesel (Table A1–A3): fixed-effect estimations show consistent negative coefficient on lagged retail diesel price (e.g., −0.159 [0.080]* to −0.289 [0.052]*** across specs), positive coefficient on retail diesel price flexibility (e.g., 0.186 [0.046]*** to 0.231 [0.064]***), negative on change in international diesel price (e.g., −0.080 [0.020]*** to −0.531 [0.089]*** in macro specs), negative on exchange rate (e.g., −0.440 [0.093]*** to −0.592 [0.097]***), volatility negative (−0.015 [0.005]***), and mixed significance for political/institutional variables.
- Kerosene (Table A4–A6): similar patterns — negative lagged retail price coefficients (e.g., −0.196 [0.079]** to −0.302 [0.061]***), positive price flexibility coefficients (e.g., 0.130 [0.056]** to 0.224 [0.052]***), negative change in international kerosene price (e.g., −0.075 [0.025]*** to −0.369 [0.126]***), exchange rate negative (e.g., −0.252 [0.130]* to −0.398 [0.190]**), volatility negative (−0.017 [0.002]***), limited significance of institutional variables aside from occasional ethnic tension score effects.
- Gasoline alternative logistic and non-transformed specifications (Table A7–A10):
  - Logistic with a = 0.001: coefficients remain negative for change in international gasoline price (e.g., −0.034 [0.005]*** to −0.155 [0.029]***) and negative on exchange rate (e.g., −0.128 [0.027]*** to −0.034 [0.005]*** across specs), positive on price flexibility (smaller magnitudes under transformation).
  - Without logistic transform and excluding outliers (Table A10): larger magnitudes observed — retail gasoline price (lagged): −0.787 [0.396]*; −1.453 [0.350]***; −1.467 [0.342]***; retail gasoline price flexibility: 1.174 [0.323]***; 1.431 [0.263]***; 1.715 [0.363]***; change in international gasoline price: −0.633 [0.145]***; −0.713 [0.135]***; −1.235 [0.217]***; exchange rate: −2.830 [0.629]***; −2.627 [0.572]***; −2.675 [0.641]***; inflation rate (log, lagged): −0.963 [0.178]***; gross public debt (log, lagged): 0.609 [0.146]***.

- Notes on estimation: robust standard errors clustered by country; *,**,*** denote significance at 10%, 5% and 1%, respectively. Retail fuel price flexibility proxied by number of months with price changes divided by 12. Fuel price gap with neighbors equals domestic minus average neighboring price. Volatility is annual standard deviation of monthly changes. Some variables first-differenced where non-stationary.

*Source: APPENDIX TABLES and Appendix sections, wpiea2020194-print-pdf (WP/20/194).*

### REFERENCES ____________________________________________________________23

### REFERENCES ____________________________________________________________23

### FIGURES
- 1. Change in International Gasoline Price and Distribution of Pass-through to Retail Gasoline Price, 2000-14 ______________________________________________________________6
- 2. Distribution of Gasoline Price Pass-Through Coefficients, 2000-14 _________________10
- 3. Logistic Transformation of Gasoline Pass-Through Coefficients, 2000-14 ____________11

### TABLES
- 1. Drivers of Gasoline Price Pass-Through: Testing for Domestic and International Oil Market Factors, 2000-2014 ___________________________________________________17
- 2. Drivers of Gasoline Price Pass-Through: Macro and Policy Factors, 2000-2014 _______19
- 3. Political and Institutional Factors: Implications for Gasoline Price Pass-Through, 2000-2014_____________________________________________________________________20

### ANNEXES
- 1. Sample Composition ______________________________________________________26
- 2. Breaking Down the Retail Fuel Price _________________________________________27

### APPENDICES
- 1. Summary Statistics and Correlation Matrix ____________________________________28
- 2. Variable Definitions and Sources ____________________________________________30
- 3. Panel unit root test _______________________________________________________32
- 4. Additional Regressions ____________________________________________________33

*wpiea2020194-print-pdf - REFERENCES ____________________________________________________________23*

### APPENDIX TABLES

### APPENDIX TABLES

### Introduction and scope
- Sample: 109 developing countries; annual data for the period 2000-14.
- Dataset: monthly retail fuel prices compiled by Kpodar and Abdallah (2017) aggregated to end-of-year retail prices; international gasoline prices from the US Energy Information Administration (prices expressed in USD).
- Dependent variable: pass-through coefficient (PT) defined as
  - PT_{i,t} = (Retail Fuel Price_{i,t} − Retail Fuel Price_{i,t−1}) / (Benchmark World Fuel Price_{i,t} − Benchmark World Fuel Price_{i,t−1})
- Notes:
  - Median pass-through was below 100 percent during half of the 2000-14 period.
  - Average number of data points per country: about 9.
  - Logistic transformation used to compress extreme PT values: f(PT_{i,t}) = (1 + e^{−a PT_{i,t}})^{−1} with a = 0.01 (sensitivity test with a = 0.001).

### Literature and motivation
- Gaps addressed:
  - Focus on short-term dynamic pass-through across countries and over time (instead of price levels or single-country studies).
  - Explicit inclusion of macroeconomic and policy drivers (exchange rate dynamics, inflation, public deficit and debt, current account, oil price cycles, fuel price-setting mechanisms).
- Rationale for using international gasoline rather than crude oil prices: most developing countries import refined fuels; international gasoline price better reflects import parity for refined products.

### Model, methodology, and key variables
- Empirical model (fixed effects): f(PT_{i,t}) = α0 + A X_{i,t} + μ_i + ε_{i,t}, where f(·) is the logistic transformation.
- Key explanatory variables (lagged where appropriate): 
  - Initial retail fuel price (lagged)
  - Retail gasoline price flexibility (proxy: number of months in a year retail prices changed relative to previous month, divided by 12)
  - Change in international gasoline price and its volatility (standard deviation over a year using monthly data)
  - Exchange rate (local currency unit per USD)
  - Fuel price gap with neighboring countries (domestic price − average neighbor price)
  - GDP per capita (log), real GDP growth (lagged), inflation rate (log, lagged), government primary balance (percent of GDP, lagged), gross public debt (percent of GDP, log, lagged), current account balance (percent of GDP, lagged)
  - Oil rents (percent of GDP), gasoline/diesel/kerosene share in total domestic fuel consumption, CO2 emissions (kg per PPP $ of GDP, lagged)
  - Political/institutional: Polity score (first-differenced), election-year dummy, urban population share, ethnic tension score
- Stationarity and endogeneity handling:
  - Fisher panel unit root tests; non-stationary variables first-differenced (gasoline share and Polity).
  - Lagging variables by one period to mitigate simultaneity.

### Main empirical findings (gasoline, diesel, kerosene; robustness across fuels)
- General: results are consistent across gasoline, diesel, and kerosene.
- Key relationships (selected coefficients from main regressions):
  - Table 1 (Drivers of Gasoline Price Pass-Through: domestic and international oil market factors):
    - Retail gasoline price (lagged): negative association; example coefficient (col 1) = −0.173 [0.108].
    - Retail gasoline price flexibility: positive association; example coefficient (col 1) = 0.213 [0.052]***.
    - Change in international gasoline price: negative association; example coefficient (col 1) = −0.122 [0.016]***.
    - Exchange rate (local currency unit per USD): negative association; example coefficient (col 1) = −0.499 [0.091]***.
    - Dummy year 2008: negative; example coefficient (col 1) = −0.123 [0.030]***.
    - Fuel price gap with neighbors (lagged): negative in some specifications; example (col 4) = −0.125 [0.082].
    - Volatility in international gasoline price (col 5): −0.009 [0.005]*.
    - Number of observations and coverage in Table 1: observations = 947, 785, 785, 601, 601; countries = 107, 91, 91, 72, 72; R2 = 0.14, 0.17, 0.18, 0.17, 0.17.
  - Table 2 (Macro and policy factors):
    - Change in international gasoline price: strong negative coefficients; example (col 1) = −0.497 [0.091]***.
    - Exchange rate: negative; example (col 1) = −0.124 [0.017]***.
    - Inflation rate (log, lagged): negative and significant; example (col 2) = −0.163 [0.031]***.
    - Gross public debt (percent of GDP, log, lagged): positive and significant in some specs; example (col 3) = 0.044 [0.025]* and (col 6) = 0.063 [0.024]***.
    - CO2 emissions: not robustly significant (example col 5 = 0.147 [0.144]).
    - Table 2 observations and coverage: observations = 947, 911, 896, 930, 919, 857; countries = 107, 103, 101, 104, 105, 98; R2 = 0.14, 0.17, 0.18, 0.17, 0.17, 0.19.
  - Table 3 (Political and institutional factors):
    - Political/institutional factors generally show correct signs but weak statistical significance in fixed-effect estimations.
    - Pooling estimator (column 7) yields significance for election years, oil rent, and ethnic tensions (example pooling change in international gasoline price = −0.584 [0.088]***).
    - Table 3 observations and coverage: fixed-effect columns observations = 795, 921, 930, 933, 707, 631; countries = 95, 102, 105, 105, 78, 75; pooling: observations = 631; countries = 75; R2 ≈ 0.16–0.19 across specifications.
- Interpretation of main effects:
  - Higher retail price flexibility → higher pass-through.
  - Large or volatile increases in international fuel prices → lower pass-through (governments cushion consumers).
  - Exchange rate depreciation → lower pass-through.
  - Higher inflation → lower pass-through.
  - Higher public debt → higher pass-through (limited fiscal buffers).
  - Price gap with neighbors: higher domestic price relative to neighbors → lower pass-through (smuggling and political pressures).
  - Political and institutional variables better explain cross-country price levels than short-term pass-through dynamics.

### Robustness checks
- Logistic transformation steepness: results stable when a = 0.001 versus a = 0.01.
- Alternative: dropping the logistic transformation and excluding outliers (pass-through values exceeding sample average by one standard deviation) does not materially change findings.
- Excluding oil-exporting countries (15 of 109): results remain materially unchanged.

### Policy implications and recommendations
- Pricing mechanism:
  - Increase retail price flexibility (automatic pricing mechanism or liberalization) to improve pass-through and contain subsidies, conditional on country circumstances and social protections.
- Timing of reforms:
  - Use periods of relatively low and less volatile international fuel prices as a window of opportunity to reform fuel subsidies.
- Macroeconomic environment:
  - Avoid undertaking subsidy reforms during periods of exchange rate depreciation and high inflation, which can undermine pass-through and raise political costs.
- Fiscal considerations:
  - Do not wait until public debt is unsustainably high to undertake subsidy reforms; limited fiscal space may force higher pass-through at politically costly times.
- Regional coordination:
  - Consider regional coordination when significant price gaps with neighboring countries drive smuggling and undermine reform incentives.
- Social sensitivity:
  - Complement pricing reforms with social measures to shield vulnerable populations from fuel price surges.

*Source: APPENDIX TABLES, wpiea2020194-print-pdf (WP/20/194) — data and tables as provided in the content unit.*

### REFERENCES

### REFERENCES

### Key bibliographic sources
- Angelopoulou, E. and H. Gibson, 2010, “The Determinants of Retail Petrol Prices in Greece,” Economic Modelling, 27(6), pp. 1537-42.
- 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.
- 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.
- Banfi, S., M. Filippini and L. Hunt, 2005, “Fuel Tourism in Border Regions: The case of Switzerland,” Energy Economics, 27(5), pp. 689-707.
- van Beers, C. and J. Strand, 2013, “Political Determinants of Fossil Fuel Pricing,” Policy Research Working Paper 6470, The World Bank.
- 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.
- 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.
- 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.
- 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).
- 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).
- 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.
- Ebeke C. and C. Lonkeng Ngouana, 2015, “Energy Subsidies and Public Social Spending; Theory and Evidence,” IMF Working Papers 15/101.
- El-Katiri, L. and B. Fattouh, 2012, “Energy Subsidies in the Arab World. Fossil Fuel Subsidies” European Parliament.
- 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).
- 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.
- 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.
- 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).
- 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.
- 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.
- 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.
- 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).
- 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.
- Meléndez-Jiménez, M. and A. Polanski, 2020, “Dirty neighbors — Pollution in an interlinked world” Energy Economics, Vol.86 (forthcoming).
- Moghaddam, H. and F. Wirl, 2018, “Determinants of oil price subsidies in oil and gas exporting countries” Energy Policy, Vol. 122, pp.409-420.
- 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.
- 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.
- 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.

### Annex 1 — Sample Composition (regional country lists)
- Sub-Saharan Africa: Angola; Benin; Botswana; Burkina Faso; Burundi; Cameroon; Cape Verde; Central African Republic; Chad; Comoros; Congo, Republic of; Cote d'Ivoire; Democratic Republic of the Congo; Ethiopia; Gabon; Gambia, The; Ghana; Guinea; Guinea-Bissau; Kenya; Lesotho; Liberia; Madagascar; Malawi; Mali; Mauritania; Mauritius; Mozambique; Namibia; Niger; Nigeria; Rwanda; Senegal; Seychelles; Sierra Leone; South Africa; Sudan; Swaziland; Sao Tome and Principe; Tanzania; Togo; Uganda; Zambia; Zimbabwe.
- Europe and Central Asia: Armenia; Azerbaijan; Belarus; Bulgaria; Georgia; Hungary; Kazakhstan; Kyrgyz Republic; Moldova; Romania; Serbia; Tajikistan; Turkey; Uzbekistan.
- Latin America and the Caribbean: Argentina; Bolivia; Brazil; Colombia; Costa Rica; Dominican Republic; Ecuador; El Salvador; Grenada; Guatemala; Haiti; Honduras; Jamaica; Mexico; Nicaragua; Panama; Paraguay; Peru; Venezuela.
- East Asia and Pacific: China; Fiji; Indonesia; Kiribati; Lao P.D.R.; Malaysia; Mongolia; Myanmar; Philippines; Samoa; Solomon Islands; Thailand; Timor-Leste, Dem. Rep. of; Vietnam.
- Middle East and North Africa: Algeria; Djibouti; Egypt; Iran; Iraq; Jordan; Lebanon; Libya; Morocco; Tunisia; West Bank and Gaza; Yemen.
- South Asia: Afghanistan; Bangladesh; India; Maldives; Nepal; Pakistan; Sri Lanka.

### Annex 2 — Breaking Down the Retail Fuel Price (value-chain determinants)
- Crude oil prices:
  - Crude oil price depends on worldwide supply and demand; high inelasticity of supply and demand explains large fluctuations in fuel prices.
  - Speculation and political tensions can cause sporadic fluctuations.
  - "In principle, countries all pay roughly the same price for the fuel they purchase."
- Processing cost:
  - Different crude types vary in composition, liquidity, gravity, sulfur content; processing costs vary accordingly and with refining capacity.
  - Example: 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:
  - Exchange rate fluctuations impact the price of imported oil traded in US dollars.
  - Over the last 15 years, rising fuel prices have often been exacerbated by a simultaneous depreciation against the USD; depreciation also strengthens the impact of fuel price declines in local currencies.
  - Timing lags between purchase and delivery can delay exchange-rate impacts, especially in remote areas.
- International transportation costs:
  - Location and market size affect transportation costs; small, distant markets in volatile regions incur higher insurance and transportation costs.
- Domestic distribution costs:
  - Costs include transportation, storage, distribution, marketing and retail sales.
  - Large countries with poor infrastructure face higher local distribution costs; license requirements can force providers to operate loss-making services in shallow markets offset via cross-subsidization.
  - Market size and competitiveness affect distribution costs and economies of scale.
- Local taxation:
  - Many advanced economies impose significant excise, VAT and import taxes on fuel.
  - Oil-producing countries often subsidize fuel (e.g., Venezuela).
  - European countries typically impose high taxes—often a multiple of the price of fuel.
- Other factors:
  - Timing of fuel purchases and lags contribute to price differences.
  - Price smoothing mechanisms can cause gradual transmission from international to domestic prices.

### Appendix 1 — Summary Statistics (selected variables and exact sample statistics)
- Gasoline price pass-through (percent): Number of observations 916; Mean 91.41; Standard deviation 504.44; Minimum -6596.09; Maximum 3240.47.
- Diesel price pass-through (percent): Number of observations 967; Mean 133.54; Standard deviation 703.34; Minimum -4619.34; Maximum 8001.58.
- Kerosene price pass-through (percent): Number of observations 692; Mean 92.35; Standard deviation 553.02; Minimum -4908.87; Maximum 7168.78.
- Retail gasoline price level: Number of observations 922; Mean 1.00; Standard deviation 0.42; Minimum 0.01; Maximum 2.61.
- Retail diesel price level: Number of observations 967; Mean 0.89; Standard deviation 0.41; Minimum 0.01; Maximum 2.34.
- Retail kerosene price level: Number of observations 699; Mean 0.76; Standard deviation 0.37; Minimum 0.01; Maximum 1.68.
- Gasoline price flexibility: Number of observations 967; Mean 0.54; Standard deviation 0.40; Minimum 0.00; Maximum 1.00.
- Diesel price flexibility: Number of observations 967; Mean 0.50; Standard deviation 0.39; Minimum 0.00; Maximum 1.00.
- Kerosene price flexibility: Number of observations 967; Mean 0.59; Standard deviation 0.41; Minimum 0.00; Maximum 1.00.
- Change in international gasoline price: Number of observations 967; Mean 0.12; Standard deviation 0.37; Minimum -0.58; Maximum 0.94.
- Change in international diesel price: Number of observations 967; Mean 0.10; Standard deviation 0.30; Minimum -0.45; Maximum 0.56.
- Change in international kerosene price: Number of observations 967; Mean 0.10; Standard deviation 0.28; Minimum -0.46; Maximum 0.51.
- Gasoline price gap with neighbors: Number of observations 773; Mean -0.04; Standard deviation 0.34; Minimum -1.29; Maximum 1.61.
- Diesel price gap with neighbors: Number of observations 802; Mean -0.03; Standard deviation 0.32; Minimum -1.12; Maximum 1.36.
- Kerosene price gap with neighbors: Number of observations 536; Mean -0.03; Standard deviation 0.35; Minimum -1.27; Maximum 1.46.
- Share of gasoline in fuel consumption: Number of observations 647; Mean 31.83; Standard deviation 12.15; Minimum 6.38; Maximum 73.91.
- Share of diesel in fuel consumption: Number of observations 647; Mean 54.22; Standard deviation 12.47; Minimum 4.29; Maximum 82.91.
- Share of kerosene in fuel consumption: Number of observations 647; Mean 13.95; Standard deviation 9.78; Minimum 0.72; Maximum 55.20.
- Volatility of international gasoline price: Number of observations 967; Mean 0.75; Standard deviation 2.64; Minimum 0.00; Maximum 33.00.
- Volatility of international diesel price: Number of observations 967; Mean 0.70; Standard deviation 2.51; Minimum 0.00; Maximum 34.24.
- Volatility of international kerosene price: Number of observations 967; Mean 0.69; Standard deviation 2.50; Minimum 0.00; Maximum 34.76.
- Log of GDP per capita: Number of observations 967; Mean 7.42; Standard deviation 1.14; Minimum 4.66; Maximum 9.72.
- Exchange rate (LCU per USD): Number of observations 967; Mean 0.03; Standard deviation 0.12; Minimum -0.98; Maximum 0.98.
- Real GDP growth rate: Number of observations 963; Mean 4.73; Standard deviation 5.70; Minimum -62.08; Maximum 104.49.
- Inflation rate (log): Number of observations 944; Mean 1.07; Standard deviation 0.06; Minimum 0.90; Maximum 1.62.
- Primary balance (percent of GDP): Number of observations 949; Mean -0.51; Standard deviation 6.77; Minimum -73.80; Maximum 123.49.
- Public debt (percent of GDP): Number of observations 945; Mean 3.72; Standard deviation 0.63; Minimum 1.83; Maximum 5.65.
- Current account balance (percent of GDP): Number of observations 890; Mean -4.64; Standard deviation 10.35; Minimum -80.05; Maximum 80.67.
- CO2 emissions (kg per PPP $ of GDP): Number of observations 946; Mean 0.22; Standard deviation 0.16; Minimum 0.02; Maximum 1.22.
- Polity variable: Number of observations 812; Mean 3.53; Standard deviation 5.50; Minimum -9.00; Maximum 10.00.
- Election year dummy: Number of observations 941; Mean 0.15; Standard deviation 0.36; Minimum 0.00; Maximum 1.00.
- Oil rents (percent of GDP): Number of observations 957; Mean 3.33; Standard deviation 8.45; Minimum 0.00; Maximum 57.57.
- Urban population share: Number of observations 960; Mean 46.34; Standard deviation 19.66; Minimum 8.46; Maximum 91.60.
- Ethnic tension score: Number of observations 719; Mean 3.71; Standard deviation 1.21; Minimum 0.67; Maximum 6.00.

### Appendix 2 — Variable definitions and data sources (selected)
- Gasoline price pass-through: "The ratio of the change in the retail gasoline price over the change in the international gasoline price throughout a year (as a percent)"; Sources: Authors' calculations and Kpodar and Abdallah (2017).
- Diesel price pass-through: "The ratio of the change in the retail diesel price over the change in the international diesel price throughout a year (as a percent)".
- Kerosene price pass-through: "The ratio of the change in the retail kerosene price over the change in the international kerosene price throughout a year (as a percent)".
- Gasoline price flexibility / Diesel price flexibility / Kerosene price flexibility: Number of months during which price changes in a year, divided by 12.
- Retail gasoline/diesel/kerosene price level ($USD/liter): "Price paid by consumers at the pump. The price is tax-inclusive." Source: Kpodar and Abdallah (2017).
- Change in international gasoline/diesel/kerosene price: "Percentage change in the benchmark world price" (US Energy Information Administration for gasoline).
- Gasoline/diesel/kerosene price gap with neighbors: "Difference between the retail ... price in the country and the average retail ... price in neighboring countries"; Sources: Authors' calculations and the Correlates of War Project Database (Direct Contiguity Data, 1816-2016. Version 3.2.).
- Share of gasoline/diesel/kerosene in fuel consumption: Ratios from International Energy Agency.
- Volatility of international gasoline/diesel/kerosene price: "The standard deviation of the monthly change in the world price ... over a year"; Sources: Authors' calculations and the International Energy Agency.
- Log of GDP per capita: "Logarithm of the ratio of gross domestic product (in USD) divided by the size of the population"; Source: World Development Indicators.
- Real GDP growth rate: "Percentage change in the gross domestic product in constant USD".
- CO2 emissions (kg per PPP $ of GDP): "Kg of carbon dioxide emissions divided by GDP in purchasing power parity terms".
- Oil rents (percent of GDP): "The difference between the value of crude oil production at world prices and total costs of production, expressed as a ratio to GDP".
- Exchange rate (LCU per USD): "The value in USD of one unit of local currency"; Source: International Financial Statistics.
- Inflation rate (log): "Percentage change in the consumer price index".
- Primary balance (percent of GDP): "Government net borrowing, excluding interest payments, as a percent of GDP".
- Public debt (percent of GDP): "General government public debt divided by GDP".
- Current account balance (percent of GDP): "The sum of the trade balance (exports minus imports) plus net income and net transfers, expressed as a ratio to GDP".
- Polity variable: "Polity Score is an index ranging from -10 to 10, with higher scores indicating more democratic countries"; Source: Polity IV dataset.
- Election year dummy: "Dummy variable taking 1 during election years and zero otherwise"; Source: Database of Political Institutions.
- Ethnic tension score: "An assessment of the degree of tension within a country attributable to racial, nationality, or language divisions. A low score means high tensions."; Source: International Country Risk Guide database.

### Appendix 3 — Panel unit root test (Fisher Test using augmented Dickey-Fuller)
- Test 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."
- Variables with Chi2 statistics reported (levels include significance markers where provided):
  - Gasoline price pass-through: Chi2 statistic 288.0 ***.
  - Retail gasoline price: Chi2 statistic 223.7 *.
  - GDP per capita (log): Chi2 statistic 267.0 **.
  - Retail gasoline price flexibility: Chi2 statistic 285.9 ***.
  - Change in international gasoline price: Chi2 statistic 332.6 ***.
  - Exchange rate (local currency unit per USD): Chi2 statistic 354.9 ***.
  - Volatility in international gasoline price: Chi2 statistic 227.5 *.
  - Gasoline price gap with neighbors: Chi2 statistic 503.4 ***.
  - Gasoline share in total fuel consumption: Variable in level Chi2 statistic 165.2; Variable in first difference Chi2 statistic 458.0 ***.
  - Real GDP growth rate: Chi2 statistic 578.1 ***.
  - Inflation rate (log): Chi2 statistic 605.1 ***.
  - Government primary balance (percent of GDP): Chi2 statistic 368.1 ***.
  - Gross public debt (percent of GDP, log): Chi2 statistic 260.5 **.
  - Current account balance (percent of GDP): Chi2 statistic 462.1 ***.
  - CO2 emissions (kg per PPP $ of GDP): Chi2 statistic 298.1 ***.
  - Polity variable: Variable in level Chi2 statistic 199.1; Variable in first difference Chi2 statistic 331.0 ***.
  - Oil rents (percent of GDP): Chi2 statistic 120.7 *.
  - Urban population share: Chi2 statistic 405.8 ***.
  - Ethnic tension score: Chi2 statistic 363.4 ***.

*Content derived from the PDF chapter/section titled "REFERENCES" (wpiea2020194-print-pdf).*

### Appendix 4. Additional Regressions

### Appendix 4. Additional Regressions

### Diesel Price Pass-Through: Domestic and International Oil Market Factors (Table A1)
- Estimation type: Fixed-effect estimations; Number of observations and countries vary across specifications:
  - (1) 967 observations, 109 countries, R2 0.12
  - (2) 789 observations, 92 countries, R2 0.15
  - (3) 789 observations, 92 countries, R2 0.18
  - (4) 623 observations, 72 countries, R2 0.16
  - (5) 623 observations, 72 countries, R2 0.17
- Robust significant coefficients (standard errors in brackets; *,**,*** denote significance at 10%, 5% and 1%):
  - Retail diesel price (lagged): -0.159 [0.080]*; -0.191 [0.053]***; -0.169 [0.051]***; -0.165 [0.054]***; -0.167 [0.054]***
  - Retail diesel price flexibility: 0.186 [0.046]***; 0.192 [0.049]***; 0.190 [0.045]***; 0.221 [0.065]***; 0.231 [0.064]***
  - Change in international diesel price: -0.080 [0.020]***; -0.092 [0.023]***; -0.110 [0.023]***; -0.094 [0.026]***; -0.099 [0.026]***
  - Exchange rate (local currency unit per USD): -0.440 [0.093]***; -0.541 [0.093]***; -0.529 [0.090]***; -0.590 [0.096]***; -0.592 [0.097]***
  - Dummy year 2008: -0.099 [0.025]***; -0.083 [0.029]***; -0.102 [0.029]***; -0.071 [0.033]**; -0.057 [0.033]*
  - Fuel price gap with neighbors (lagged): -0.180 [0.099]*; -0.200 [0.106]*; -0.208 [0.108]*
  - Positive gap (local diesel prices higher, lagged): -0.665 [0.177]***
  - Negative gap (lagged): 0.257 [0.114]**
  - Volatility in international diesel price: -0.015 [0.005]***
- Non-significant or small coefficients: GDP per capita (log) and diesel share in total fuel consumption (lagged) generally not significant.

### Diesel Price Pass-Through: Macro and Policy Factors (Table A2)
- Estimation type: Fixed-effect estimations; Number of observations and countries vary:
  - (1) 961 obs, 109 countries, R2 0.12; (2) 943 obs, 106 countries, R2 0.15; (3) 925 obs, 104 countries, R2 0.15; (4) 957 obs, 106 countries, R2 0.14; (5) 946 obs, 107 countries, R2 0.14; (6) 886 obs, 101 countries, R2 0.17
- Robust significant coefficients:
  - Retail diesel price (lagged): -0.158 [0.081]*; -0.251 [0.048]***; -0.269 [0.057]***; -0.259 [0.047]***; -0.260 [0.049]***; -0.255 [0.055]***
  - Retail diesel price flexibility: 0.186 [0.046]***; 0.184 [0.046]***; 0.187 [0.050]***; 0.181 [0.046]***; 0.179 [0.047]***; 0.181 [0.050]***
  - GDP per capita (log): 0.025 [0.075]; 0.102 [0.051]**; 0.144 [0.062]**; 0.111 [0.048]**; 0.113 [0.050]**; 0.146 [0.062]**
  - Change in international diesel price: -0.440 [0.093]***; -0.517 [0.086]***; -0.531 [0.089]***; -0.499 [0.089]***; -0.506 [0.089]***; -0.542 [0.086]***
  - Exchange rate (local currency unit per USD): -0.080 [0.020]***; -0.079 [0.021]***; -0.092 [0.023]***; -0.088 [0.021]***; -0.086 [0.021]***; -0.082 [0.024]***
  - Dummy year 2008: -0.100 [0.026]***; -0.086 [0.026]***; -0.086 [0.028]***; -0.092 [0.027]***; -0.083 [0.025]***; -0.078 [0.029]***
  - Inflation rate (log, lagged): -0.180 [0.029]***; -0.198 [0.028]***
  - Gross public debt (percent of GDP, log, lagged): 0.051 [0.029]*; 0.064 [0.027]**
- Insignificant: Real GDP growth rate (lagged), Government primary balance, Current account balance, CO2 emissions (large standard errors).

### Diesel Price Pass-Through: Political and Institutional Factors (Table A3)
- Estimation type: Fixed-effect estimations; observations and countries vary: (1) 810 obs, 96 countries, R2 0.14; (2) 941 obs, 104 countries, R2 0.14; (3) 957 obs, 107 countries, R2 0.14; (4) 960 obs, 107 countries, R2 0.14; (5) 719 obs, 79 countries, R2 0.15; (6) 643 obs, 76 countries, R2 0.16
- Robust significant coefficients across specs:
  - Retail diesel price (lagged): -0.289 [0.052]***; -0.250 [0.047]***; -0.253 [0.048]***; -0.257 [0.048]***; -0.228 [0.047]***; -0.247 [0.050]***
  - Retail diesel price flexibility: 0.199 [0.052]***; 0.173 [0.046]***; 0.189 [0.046]***; 0.183 [0.046]***; 0.200 [0.054]***; 0.222 [0.057]***
  - GDP per capita (log): 0.148 [0.055]***; 0.104 [0.050]**; 0.103 [0.050]**; 0.118 [0.052]**; 0.090 [0.050]*; 0.151 [0.055]***
  - Change in international diesel price: -0.483 [0.099]***; -0.496 [0.088]***; -0.500 [0.088]***; -0.504 [0.088]***; -0.480 [0.095]***; -0.474 [0.100]***
  - Exchange rate (local currency unit per USD): -0.084 [0.030]***; -0.081 [0.020]***; -0.088 [0.020]***; -0.089 [0.021]***; -0.071 [0.022]***; -0.095 [0.033]***
  - Dummy year 2008: -0.086 [0.035]**; -0.088 [0.026]***; -0.096 [0.026]***; -0.093 [0.026]***; -0.076 [0.026]***; -0.101 [0.038]**
- Political/institutional variables generally not significant: Polity variable (first difference), Election year dummy, Oil rents, Urban population share, Ethnic tension score show limited or no robust significance.

### Kerosene Price Pass-Through: Domestic and International Oil Market Factors (Table A4)
- Estimation type: Fixed-effect estimations; (1) 693 obs, 81 countries, R2 0.12; (2) 515 obs, 65 countries, R2 0.12; (3) 515 obs, 65 countries, R2 0.13; (4) 371 obs, 46 countries, R2 0.13; (5) 371 obs, 46 countries, R2 0.14
- Robust significant coefficients:
  - Retail kerosene price (lagged): -0.196 [0.079]**; -0.226 [0.063]***; -0.205 [0.061]***; -0.166 [0.059]***; -0.171 [0.060]***
  - Retail kerosene price flexibility: 0.185 [0.053]***; 0.133 [0.058]**; 0.130 [0.056]**; 0.179 [0.077]**; 0.205 [0.073]***
  - GDP per capita (log): 0.085 [0.070]; 0.123 [0.061]**; 0.141 [0.057]**; 0.129 [0.051]**; 0.128 [0.053]**
  - Change in international kerosene price: -0.075 [0.025]***; -0.080 [0.033]**; -0.083 [0.033]**; -0.073 [0.036]*; -0.083 [0.036]**
  - Exchange rate (local currency unit per USD): -0.252 [0.130]*; -0.305 [0.151]**; -0.308 [0.151]**; -0.382 [0.191]*; -0.398 [0.190]**
  - Dummy year 2008: -0.106 [0.029]***; -0.104 [0.035]***; -0.107 [0.035]***; -0.090 [0.035]**; -0.074 [0.035]**
  - Positive gap (local kerosene prices higher, lagged): -0.383 [0.148]**
  - Negative gap (lagged): 0.136 [0.080]*
  - Kerosene share in total fuel consumption (lagged): 0.005 [0.003]* (in some specs)
  - Volatility in international kerosene price: -0.017 [0.002]***

### Kerosene Price Pass-Through: Macro and Policy Factors (Table A5)
- Estimation type: Fixed-effect estimations; (1) 690 obs, 81 countries, R2 0.12; (2) 678 obs, 79 countries, R2 0.15; (3) 661 obs, 77 countries, R2 0.15; (4) 683 obs, 78 countries, R2 0.15; (5) 672 obs, 79 countries, R2 0.15; (6) 634 obs, 75 countries, R2 0.18
- Robust significant coefficients:
  - Retail kerosene price (lagged): -0.197 [0.078]**; -0.278 [0.056]***; -0.287 [0.063]***; -0.286 [0.056]***; -0.274 [0.056]***; -0.253 [0.059]***
  - Retail kerosene price flexibility: 0.187 [0.053]***; 0.199 [0.053]***; 0.191 [0.059]***; 0.180 [0.055]***; 0.215 [0.053]***; 0.224 [0.052]***
  - GDP per capita (log): 0.085 [0.069]; 0.157 [0.054]***; 0.175 [0.063]***; 0.160 [0.051]***; 0.174 [0.056]***; 0.202 [0.065]***
  - Change in international kerosene price: -0.253 [0.130]*; -0.333 [0.124]***; -0.357 [0.132]***; -0.317 [0.128]**; -0.341 [0.126]***; -0.369 [0.126]***
  - Exchange rate (local currency unit per USD): -0.079 [0.024]***; -0.071 [0.026]***; -0.081 [0.028]***; -0.088 [0.025]***; -0.081 [0.026]***; -0.073 [0.028]**
  - Dummy year 2008: -0.110 [0.029]***; -0.091 [0.030]***; -0.091 [0.031]***; -0.103 [0.030]***; -0.090 [0.030]***; -0.090 [0.031]***
  - Inflation rate (log, lagged): -0.111 [0.026]***; -0.122 [0.030]***
  - Gross public debt (percent of GDP, log, lagged): 0.024 [0.024]; 0.053 [0.023]**
  - Current account balance (percent of GDP, lagged): 0.003 [0.002]; 0.004 [0.002]**
  - CO2 emissions (kg per PPP $ of GDP, lagged): 0.266 [0.220]; 0.447 [0.193]**

### Kerosene Price Pass-Through: Political and Institutional Factors (Table A6)
- Estimation type: Fixed-effect estimations; observations and countries vary: (1) 589 obs, 71 countries, R2 0.14; (2) 672 obs, 77 countries, R2 0.14; (3) 683 obs, 79 countries, R2 0.15; (4) 686 obs, 79 countries, R2 0.15; (5) 501 obs, 57 countries, R2 0.15; (6) 451 obs, 56 countries, R2 0.15
- Robust significant coefficients:
  - Retail kerosene price (lagged): -0.302 [0.061]***; -0.282 [0.057]***; -0.286 [0.056]***; -0.289 [0.059]***; -0.280 [0.065]***; -0.290 [0.073]***
  - Retail kerosene price flexibility: 0.183 [0.060]***; 0.181 [0.055]***; 0.189 [0.056]***; 0.187 [0.055]***; 0.153 [0.058]**; 0.148 [0.063]**
  - GDP per capita (log): 0.190 [0.059]***; 0.156 [0.054]***; 0.158 [0.052]***; 0.164 [0.052]***; 0.144 [0.058]**; 0.201 [0.066]***
  - Change in international kerosene price: -0.244 [0.135]*; -0.309 [0.128]**; -0.317 [0.127]**; -0.327 [0.127]**; -0.268 [0.140]*; -0.203 [0.146]
  - Exchange rate (local currency unit per USD): -0.100 [0.037]***; -0.077 [0.026]***; -0.086 [0.026]***; -0.084 [0.026]***; -0.047 [0.027]*; -0.081 [0.039]**
  - Dummy year 2008: -0.120 [0.043]***; -0.094 [0.030]***; -0.110 [0.032]***; -0.098 [0.028]***; -0.068 [0.034]**; -0.117 [0.048]**
  - Ethnic tension score: 0.061 [0.024]**; 0.065 [0.026]**
- Other institutional variables (Polity first difference, Election year dummy, Oil rents, Urban population share) generally not robustly significant.

### Gasoline Price Pass-Through: Domestic and International Oil Market Factors and Alternative Logistic Function (Table A7)
- Estimation type: Fixed-effect estimations; (1) 947 obs, 107 countries, R2 0.07; (2) 785 obs, 91 countries, R2 0.09; (3) 785 obs, 91 countries, R2 0.09; (4) 601 obs, 72 countries, R2 0.10; (5) 601 obs, 72 countries, R2 0.10
- Logistic transformation parameter a used: 0.001 (alternative to main results' 0.01)
- Robust significant coefficients:
  - Retail gasoline price (lagged): -0.071 [0.024]***; -0.071 [0.024]*** (in some specs)
  - Retail gasoline price flexibility: 0.031 [0.017]*; 0.038 [0.018]**; 0.036 [0.018]**; 0.047 [0.022]**; 0.048 [0.022]**
  - Change in international gasoline price: -0.034 [0.005]***; -0.030 [0.005]***; -0.030 [0.005]***; -0.027 [0.006]***; -0.028 [0.006]***
  - Exchange rate (local currency unit per USD): -0.128 [0.027]***; -0.131 [0.031]***; -0.129 [0.031]***; -0.143 [0.037]***; -0.144 [0.037]***
  - Dummy year 2008: -0.030 [0.009]***; -0.024 [0.009]**; -0.024 [0.009]**; -0.023 [0.011]**; -0.021 [0.011]**
  - Fuel price gap with neighbors (lagged): -0.056 [0.026]**; -0.083 [0.029]***; -0.084 [0.029]***
  - Volatility in international gasoline price: -0.002 [0.001]**
  - Positive gap (local fuel prices higher, lagged): -0.096 [0.036]***
- Gasoline share in total fuel consumption (lagged, and first-differenced) not significant in most specs.

### Gasoline Price Pass-Through: Macro and Policy Factors with Alternative Logistic Function (Table A8)
- Estimation type: Fixed-effect estimations; (1) 947 obs, 107 countries, R2 0.07; (2) 911 obs, 103 countries, R2 0.09; (3) 896 obs, 101 countries, R2 0.09; (4) 930 obs, 104 countries, R2 0.09; (5) 919 obs, 105 countries, R2 0.09; (6) 857 obs, 98 countries, R2 0.10
- Robust significant coefficients:
  - Retail gasoline price (lagged): -0.052 [0.031]*; -0.093 [0.020]***; -0.099 [0.022]***; -0.093 [0.020]***; -0.094 [0.021]***; -0.092 [0.021]***
  - Retail gasoline price flexibility: 0.031 [0.017]* (other specs smaller and not significant)
  - GDP per capita (log): 0.051 [0.016]***; 0.059 [0.018]***; 0.050 [0.016]***; 0.050 [0.018]***; 0.057 [0.020]***
  - Change in international gasoline price: -0.127 [0.027]***; -0.154 [0.026]***; -0.139 [0.031]***; -0.139 [0.028]***; -0.140 [0.028]***; -0.155 [0.029]***
  - Exchange rate (local currency unit per USD): -0.034 [0.005]*** across specs
  - Dummy year 2008: -0.031 [0.009]***; -0.022 [0.008]***; -0.022 [0.008]**; -0.023 [0.008]***; -0.022 [0.008]***; -0.020 [0.008]**
  - Inflation rate (log, lagged): -0.022 [0.013]*; -0.024 [0.012]**
  - Gross public debt (percent of GDP, log, lagged): 0.012 [0.007]*; 0.014 [0.007]**
- Other macro variables show modest or mixed significance.

### Gasoline Price Pass-Through: Political and Institutional Factors with Alternative Logistic Function (Table A9)
- Estimation type: Fixed-effects; observations and countries vary: (1) 795 obs, 95 countries, R2 0.09; (2) 921 obs, 102 countries, R2 0.09; (3) 930 obs, 105 countries, R2 0.09; (4) 933 obs, 105 countries, R2 0.09; (5) 707 obs, 78 countries, R2 0.10; (6) 631 obs, 75 countries, R2 0.13
- Robust significant coefficients:
  - Retail gasoline price (lagged): -0.097 [0.022]***; -0.090 [0.021]***; -0.096 [0.021]***; -0.091 [0.020]***; -0.067 [0.020]***; -0.080 [0.023]***
  - Retail gasoline price flexibility: 0.029 [0.017]*; 0.024 [0.017]; 0.025 [0.017]; 0.040 [0.018]**; 0.043 [0.022]*
  - GDP per capita (log): 0.059 [0.019]***; 0.047 [0.017]***; 0.051 [0.017]***; 0.055 [0.017]***; 0.038 [0.017]**; 0.060 [0.022]***
  - Change in international gasoline price: -0.134 [0.031]***; -0.137 [0.029]***; -0.141 [0.029]***; -0.141 [0.028]***; -0.146 [0.032]***; -0.145 [0.033]***
  - Exchange rate (local currency unit per USD): -0.043 [0.008]***; -0.031 [0.005]***; -0.034 [0.005]***; -0.034 [0.005]***; -0.036 [0.005]***; -0.058 [0.009]***
  - Dummy year 2008: -0.033 [0.013]**; -0.022 [0.008]***; -0.021 [0.008]***; -0.025 [0.008]***; -0.032 [0.009]***; -0.050 [0.013]***
  - Oil rents (percent of GDP): -0.001 [0.001]; -0.003 [0.001]** (in some specs)
  - Other institutional variables (Polity first difference, Election year dummy, Urban population share, Ethnic tension score) largely not robustly significant.

### Regressions without Logistic Transformation and Excluding Potential Outliers (Table A10)
- Estimation type: OLS-style fixed-effect variants without logistic transform; potential outliers excluded (pass-through values exceeding sample average by one standard deviation).
- Three specifications reported:
  - (1) 543 observations, 72 countries, R2 0.13
  - (2) 763 observations, 96 countries, R2 0.17
  - (3) 569 observations, 74 countries, R2 0.18
- Robust significant coefficients (selected):
  - Retail gasoline price (lagged): -0.787 [0.396]*; -1.453 [0.350]***; -1.467 [0.342]***
  - Retail gasoline price flexibility: 1.174 [0.323]***; 1.431 [0.263]***; 1.715 [0.363]***
  - Change in international gasoline price: -0.633 [0.145]***; -0.713 [0.135]***; -1.235 [0.217]***
  - Exchange rate (local currency unit per USD): -2.830 [0.629]***; -2.627 [0.572]***; -2.675 [0.641]***
  - Fuel price gap with neighbors (lagged): -0.593 [0.185]***; -0.796 [0.173]***; -1.396 [0.286]***
  - Inflation rate (log, lagged): -0.963 [0.178]***
  - Gross public debt (percent of GDP, log, lagged): 0.609 [0.146]***
  - CO2 emissions (kg per PPP $ of GDP, lagged): 1.953 [1.147]* (specification (3))
- Non-significant or large-variance coefficients reported for volatility, real GDP growth rate, government primary balance, current account balance, polity, election year, oil rents, urban population share, ethnic tension score in various specs.

Notes (from source tables):
- Robust standard errors clustered by country; *,**,*** denote significance at 10%, 5% and 1%, respectively.
- Retail fuel price flexibility is 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 equals domestic fuel price minus the average fuel price in neighboring countries for the same product.
- Volatility in international fuel price is the standard deviation of international fuel prices over a year using monthly data.
- Some variables are first-differenced where non-stationary in level (e.g., gasoline share in total fuel consumption, Polity variable).

*Appendix 4. Additional Regressions (tables and notes as provided).*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2020/english/wpiea2020194-print-pdf.pdf_