## 2. Resource Revenue

## Source details

**Canonical URL:** [2. Resource Revenue](https://www.imf.org/-/media/files/publications/wp/2021/english/wpiea2021009-print-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2021/english/wpiea2021009-print-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2021/english/wpiea2021009-print-pdf.pdf.json)

---

### Major themes and approach
- Objective: quantify the extent of profit shifting in the extractive industries (mining and hydrocarbons) using firm-level and country-level data.
- Main data sources:
  - Country-by-Country (CBC) payment reports (European and Canadian MNEs) merged with ORBIS firm information.
  - Resource Revenue Database: unbalanced panel of 74 countries, 2000–2018.
  - Aggregate EITI reports and company-level EITI reports for several African economies.
- Estimation strategy:
  - Baseline regression: log(휋kct) = 훽(휏ct − 휏̅tk) + γ′Xctk + αc + νt + εctk, where 휋kct denotes taxable profits, and 훽 is the semi-elasticity of taxable profits with respect to international tax rate differentials.
  - Use statutory CIT rates to construct tax rate differentials (휏ct − 휏̅tk).
  - Micro- and macro-estimates combined in a seemingly unrelated regression (SUR) framework with weighted least squares and heteroscedasticity-robust standard errors.
- Controls (Xctk) include:
  - Productive capacity proxies: consolidated stock of fixed assets (ORBiS) and royalty payments (CBC and country-level).
  - Firm/group characteristics: number of entities in group, indicator (Source) to capture EITI vs CBC systematic differences.
  - Supply/demand and macro controls: imports+exports as percent of GDP (Open), log(GDP), log(Population), nominal exchange rates, inflation, commodity price indices (energy and non-energy), and year-specific fixed effects.

### Key descriptive facts and data coverage
- CBC reports cover payments by affiliates of consolidated groups in 157 countries for 2014–2019; matched to ORBIS for slightly more than 4500 observations pertaining to more than 600 consolidated groups (out of 6300 payment-level observations).
- Resource Revenue Database: 74 countries, 2000–2018 (unbalanced panel).
- Aggregate EITI reports: 53 countries, 1999–2017.
- Company-level EITI: 15 African economies, 2004–2016.
- Industry importance:
  - Among the 74 countries in the Resource Revenue Database, resource-related revenues average one third of total revenues, or 10 percent of GDP.
  - Low-income countries: natural resource revenue averaged 40 percent of total revenue between 2000 and 2012.
  - Emerging markets: average slightly more than 20 percent of total revenue before 2012.
  - Advanced economies: resource revenues average 5 percent of total revenues.
- Revenue instrument shares (until 2012):
  - Profit-based taxes (corporate income taxes and resource rent taxes) accounted for more than 40 percent of total natural resource revenue.
  - Royalties contributed around 30 percent.

### Empirical findings: semi-elasticities, heterogeneity, and drivers
- Overall semi-elasticity results:
  - A one percentage point increase in the corporate tax rate has historically reduced reported profits of mining and hydrocarbon entities by more than 3 percent (reported in the introduction).
  - Empirical range: reported profits decrease by 3 to 4 percent in response to a 1 one percentage point increase in the tax rate differential.
- Selected coefficient evidence (baseline regression results, Table 3):
  - CIT rate:
    - Model 1: -6.930*** [1.230]
    - Model 2: -2.856*** [0.616]
  - Other rate:
    - Model 1: 3.055** [1.448]
    - Model 2: 3.630** [1.722]
  - CIT rate - Other rate:
    - Model 1: -5.799*** [0.997]
    - Model 2: -2.990*** [1.253]
    - Model 4: -4.670*** [0.751]
    - Model 5: -3.738*** [0.758]
  - Log (Royalties):
    - Model 3: 0.585*** [0.033]
    - Model 6: 0.133*** [0.034]
  - Log (Total Assets):
    - Model 3: 0.389*** [0.030]
  - Log (Gdp):
    - Model 1: 0.311*** [0.067]
    - Model 2: 0.310*** [0.066]
    - Model 4: 1.246*** [0.043]
    - Model 5: 1.246*** [0.041]
    - Model 6: 1.454*** [0.200]
  - Openness and other covariates show varying signs across micro and macro models (detailed in baseline table).
- Joint estimation (SUR) and pooled estimates (Table 4):
  - Restricted coefficient on (CIT rate - Other rate):
    - Model 1: -4.081*** [0.545]
    - Model 2: -4.603*** [0.512]
    - Model 3: -3.559*** [0.641]
    - Model 4: -3.557*** [0.641]
  - Difference between macro and micro coefficient:
    - Model 1: 2.889*** [1.153]
    - Model 2: -0.302 [1.622]
    - Model 3: -0.748 [1.463]
    - Model 4: -0.798 [1.445]
  - Interpretation: pooled estimates range around semi-elasticities of -4.6 to -3.6 depending on controls; adding productive capacity controls reduces the estimated semi-elasticity.
- Non-linearities and subgroup effects:
  - Low-tax affiliate indicator = 1 for 33 percent of all groups and 60 percent of all observations.
  - Low-tax affiliates report, on average, 30 percent less profit (significant at the 1 percent level).
  - Polynomial specification yields semi-elasticity for affiliates without tax avoidance incentives estimated at 1.5; central 50 percent of observations adjust reported profits by between 1 and 3.25 percent in response to a one percentage point increase in corporate taxes.
- Sectoral heterogeneity:
  - Mining companies respond more strongly to international tax differentials than oil and gas firms; differences in measured semi-elasticities range between 1 and 2 and attain statistical significance at the 5 percent level in some specifications.
  - Oil indicator = 1 for 43 percent of observations.
- Country- and policy-level drivers (Table 7 and supporting results):
  - Transfer pricing (TP) rules: presence yields inconsistent and statistically insignificant mitigating effect in this sample.
  - Interest limitation / thin capitalization rules: sensitivity of taxable profits is reduced by half in countries with effective interest limitation rules; negative coefficient on the indicator itself suggests somewhat smaller reported profits where such rules exist.
  - Production sharing contracts: point estimates suggest reduced sensitivity but limited variation yields insignificant results.
  - Governance (control of corruption): neither level nor interaction with tax differential statistically significant.
- Random coefficients and GLS (Table 8):
  - Fixed-effect semi-elasticity estimates: around -4.045***, -3.291***, -2.307*** in first three specifications.
  - GLS estimates can be smaller in magnitude (example: -1.195 in one specification).
  - Results consistently highlight importance of interest limitation rules, low-tax affiliates, and mining vs hydrocarbon distinction.

### Robustness, measurement issues, and interpretation limits
- Use of statutory CIT rates:
  - Recognized as an imperfect proxy for net tax savings because firms may face industry-specific rates, negotiated fiscal terms, tax holidays, or rent-based taxes; measurement error likely induces attenuation bias toward zero.
- CBC coverage limitation:
  - CBC reports cover European- and Canadian-based MNEs only; Resource Revenue Database and aggregate EITI reports used to assess global implications.
- Profit shifting channels emphasized:
  - (i) Debt-financing: high interest loans across entities creating tax shields.
  - (ii) Opacity in markets: underquoting intermediate product prices or over-remunerating offshore affiliates for technical services.
  - (iii) Offshore realization of capital gains from exploitation of resource wealth in low- or zero-tax jurisdictions.
- Transfer pricing challenges:
  - Determination of arm’s-length prices for natural resource products and specialized services poses notable gray areas for tax administrations.
- Comparison to prior literature:
  - Aggregate semi-elasticity of -3.5 is larger than estimates in prior studies; Beer and others (2020) report a semi-elasticity of around 1.5 for recent years.
  - Possible explanations: non-linearities, identification limits of statutory rates, and sample composition with more sensitive firms/sectors (notably extractives and mining).

### Simulation: revenue losses from profit shifting in extractives
- Semi-elasticities used for simulation (Table 9):
  - Interest Limitation = No; Mining operations = Yes: -6.105
  - Interest Limitation = Yes; Mining operations = Yes: -2.862
  - Interest Limitation = No; Mining operations = No: -4.477
  - Interest Limitation = Yes; Mining operations = No: -1.234
- Revenue loss formula:
  - Loss_c = ε_c (τ_c − τ̅_c) Base_c / τ_c
- Simulation setup:
  - Sample of 138 countries reporting resource-related tax revenues.
  - Revenue losses averaged between 2015 and 2018.
  - Distribution of product of country-specific elasticities with tax rate differentials is winsorized at the top and bottom 1 percent.
- Key simulation results (Table 10):
  - Global (138 countries):
    - Median revenue loss (percent of GDP): 0.01
    - Unweighted average (percent of GDP): 0.21
    - Weighted average (percent of GDP): 0.06
    - Revenue loss (USD billion): 43.75
  - Developing Countries (44 countries):
    - Median: 0.04
    - Unweighted average: 0.15
    - Weighted average: 0.08
    - Revenue loss (USD billion): 1.41
  - Emerging Markets (64 countries):
    - Median: 0.01
    - Unweighted average: 0.39
    - Weighted average: 0.15
    - Revenue loss (USD billion): 41.10
  - Advanced Economies (30 countries):
    - Median: 0.00
    - Unweighted average: 0.01
    - Weighted average: 0.00
    - Revenue loss (USD billion): 1.24
- Aggregate context:
  - Combined GDP of included countries is USD 76.6 trillion, on average, between 2015 and 2018.
- Interpretation:
  - Revenue losses from profit shifting in extractives amount to 0.06 percent of GDP for included countries, or around USD 44 billion.
  - Losses are more frequent and relatively larger in low income and developing countries (median country loses 0.04 percent of GDP); weighted average losses are largest in emerging markets (0.15 percent).

### Policy implications and recommendations
- Core implications:
  - International profit shifting by MNEs in the natural resource sector is a material threat to domestic resource mobilization.
  - Interest limitation rules are the most compelling empirical lever identified for constraining profit shifting.
  - Mere presence of transfer pricing laws without proactive and effective enforcement appears insufficient.
- Policy recommendations:
  - Strengthen and simplify transfer pricing rules and improve enforcement so TP laws are effective rather than merely providing “peace of mind”.
  - Limit interest deductions (interest limitation / thin capitalization rules) given evidence of halving profit shifting risks where such limits exist.
  - Consider broader fiscal regime reforms for extractive industries, including simplified administrative approaches (administrative pricing, safe harbors) and more formulaic allocation of taxable base where appropriate.
  - Improve tax treaty practices and limit tax incentives.
  - Strengthen investment negotiation practices and protections for countries imposing capital gains tax on indirect transfers.

*Source: wpiea2021009-print-pdf - IMF working paper chapter "2. Resource Revenue" and associated sections.*

### References .............................................................................................................

### References

### Tables
- 1. Descriptive Statistics ................................................................................................................................................. 13
- 2. Descriptive Statistics, Payments Data ................................................................................................................ 13
- 3. Baseline Results .......................................................................................................................................................... 15
- 4. Combined Regression Results on Overall Effect ........................................................................................... 16
- 5. Non-Linearities and Alternative Profit Shifting Identification .................................................................. 19
- 6. Combined Regression Results on Differential Effect ................................................................................... 21
- 7. Country- and Group-Level Determinants of Profit Shifting ...................................................................... 22
- 8. Further Robustness Checks .................................................................................................................................... 25
- 9. Semi-Elasticities Used for Simulation ................................................................................................................ 26
- 10. Revenue Losses ........................................................................................................................................................ 26

### Figures
- 1. Geographical Coverage of Data ........................................................................................................................... 11

*Source: wpiea2021009-print-pdf - References (IMF).*

### 2. Resource Revenue ....................................................................................................

### 2. Resource Revenue

### Major themes and approach
- Objective: quantify the extent of profit shifting in the extractive industries (mining and hydrocarbons) using firm-level and country-level data.
- Main data sources:
  - Country-by-Country (CBC) payment reports (European and Canadian MNEs) merged with ORBIS firm information.
  - Resource Revenue Database: unbalanced panel of 74 countries, 2000–2018.
  - Aggregate EITI reports and company-level EITI reports for several African economies.
- Estimation strategy:
  - Baseline regression: log(휋kct) = 훽(휏ct − 휏̅tk) + γ′Xctk + αc + νt + εctk, where 휋kct denotes taxable profits, and 훽 is the semi-elasticity of taxable profits with respect to international tax rate differentials.
  - Use statutory CIT rates to construct tax rate differentials (휏ct − 휏̅tk).
  - Micro- and macro-estimates combined in a seemingly unrelated regression (SUR) framework with weighted least squares and heteroscedasticity-robust standard errors.
- Controls (Xctk) include:
  - Productive capacity proxies: consolidated stock of fixed assets (ORBiS) and royalty payments (CBC and country-level).
  - Firm/group characteristics: number of entities in group, indicator (Source) to capture EITI vs CBC systematic differences.
  - Supply/demand and macro controls: imports+exports as percent of GDP (Open), log(GDP), log(Population), nominal exchange rates, inflation, commodity price indices (energy and non-energy), and year-specific fixed effects.

### Key descriptive facts and data coverage
- CBC reports cover payments by affiliates of consolidated groups in 157 countries for 2014–2019; matched to ORBIS for slightly more than 4500 observations pertaining to more than 600 consolidated groups (out of 6300 payment-level observations).
- Resource Revenue Database: 74 countries, 2000–2018 (unbalanced panel).
- Aggregate EITI reports: 53 countries, 1999–2017.
- Company-level EITI: 15 African economies, 2004–2016.
- Industry importance:
  - Among the 74 countries in the Resource Revenue Database, resource-related revenues average one third of total revenues, or 10 percent of GDP.
  - Low-income countries: natural resource revenue averaged 40 percent of total revenue between 2000 and 2012.
  - Emerging markets: average slightly more than 20 percent of total revenue before 2012.
  - Advanced economies: resource revenues average 5 percent of total revenues.
- Revenue instrument shares (until 2012):
  - Profit-based taxes (corporate income taxes and resource rent taxes) accounted for more than 40 percent of total natural resource revenue.
  - Royalties contributed around 30 percent.

### Empirical findings: profit shifting elasticities and heterogeneity
- Overall semi-elasticity finding:
  - A one percentage point increase in the corporate tax rate has historically reduced reported profits of mining and hydrocarbon entities by more than 3 percent (reported in the introduction).
- Micro-data (CBC + ORBIS) results:
  - In a model without controls the measured sensitivity was around -5.8 (second model).
  - Adding controls for productive capacity (parent fixed assets and log(royalty)) reduced the measured sensitivity to -3 in the third specification.
  - Separate estimation of domestic and foreign tax rate effects: domestic tax rate negative and foreign tax rate positive; both effects statistically significant at the 5 percent level in initial specifications.
- Macro-data (Resource Revenue Database and aggregate EITI) produce effects with expected signs; measured semi-elasticities show limited variation relative to micro estimates.
- Sectoral heterogeneity:
  - Mining companies respond more strongly to international tax differentials than oil and gas firms.
  - Possible explanations include less frequent use of joint ventures/production sharing contracts in mining and greater difficulty determining arm’s-length prices for certain related-party transactions.
- Anti-abuse rules:
  - Transfer pricing rules: estimated to have a mitigating effect on the sensitivity of taxable profits, but the analysis cannot statistically reject the possibility that mere presence of transfer pricing rules is ineffective in practice.
  - Interest limitation / thin capitalization rules: profit shifting risks are reduced by half in countries that limit the amount of interest deductible for tax purposes.
- Micro vs macro comparison:
  - Using SUR and ownership information to construct representative tax rate differentials, when controlling for productive activity and ownership structures the paper is unable to reject that the semi-elasticities in micro and macro data are the same.
  - This suggests systematic differences found in previous work between micro and macro estimates may be more likely due to biased estimates (data gaps) than to adjustment-cost dynamics.

### Robustness, measurement issues, and interpretation limits
- Use of statutory CIT rates:
  - Acknowledged as an imperfect proxy for net tax savings in extractive industries because firms may be subject to industry-specific rates, negotiated fiscal terms, tax holidays, or additional rent-based taxes.
  - Measurement error from this proxy likely induces attenuation bias, biasing estimates toward zero.
- CBC coverage limitation:
  - CBC reports cover European- and Canadian-based MNEs only, giving a partial picture of global payments; the Resource Revenue Database and aggregate EITI reports are used to gauge global revenue implications.
- Profit shifting channels emphasized:
  - (i) Debt-financing: high interest loans across entities creating tax shields.
  - (ii) Opacity in markets: underquoting intermediate product prices or over-remunerating offshore affiliates for technical services.
  - (iii) Offshore realization of capital gains from exploitation of resource wealth in low- or zero-tax jurisdictions.
- Transfer pricing challenges:
  - Determination of arm’s-length prices for natural resource products and specialized services poses notable gray areas for tax administrations.

### Policy implications and recommendations
- For tax administration and policy design:
  - Recognize elevated profit shifting sensitivity in extractive industries and direct risk-based audits and administrative resources accordingly.
  - Strengthen profit shifting limitations, especially on interest deductibility (thin capitalization / interest limitation rules), given evidence of halving profit shifting risks where such limits exist.
  - Consider broader fiscal regime reforms for the extractive industries, including:
    - Simplified administrative approaches (administrative pricing, safe harbors) to reduce reliance on full transfer pricing analyses.
    - More formulaic approaches to allocation of taxable base for extractives where appropriate.
- For developing countries and sector-specific policy:
  - Re-examine fiscal regimes applied to extractive industries given sector-specific rent characteristics and elevated avoidance risks; options range from stronger anti-abuse rules to redesign of profit-based vs production-based instruments.
- For international tax policy debates:
  - While extractive industries are often carved out from broader corporate tax reform proposals, the paper’s findings underscore the need to tailor international and domestic reforms to sector-specific avoidance channels and revenue importance.

*Source: IMF working paper chapter "2. Resource Revenue" (excerpt provided).*

### 6. The measured semi-elasticity decreases from -4.67 in the second macro estimation (Model 5)

### 6. The measured semi-elasticity decreases from -4.67 in the second macro estimation (Model 5) to -3.7 in the third micro estimation

### Main findings on semi-elasticity and profit shifting
- The measured semi-elasticity decreases from -4.67 in the second macro estimation (Model 5) to -3.7 in the third micro estimation.
- Overall, results provide robust evidence for profit shifting: reported profits decrease by 3 to 4 percent in response to a 1 one percentage point increase in the tax rate differential.
- Covariate associations with the CIT base:
  - Positive association with the size of a country (positive effect of log of GDP).
  - Negative partial effect of population (association with per capita income level).
  - Exchange rate exerts a consistent negative effect.
  - Inflation tends to increase the returns to extractive investments.
  - Trade openness: positive impact in the macro data but negative impact in the micro data.
  - Size of representative firm: negative impact in the macro dataset; at the firm level, the size of a corporate group is positively related with the tax base.

### Baseline regression results (Table 3)
- Data coverage:
  - Macro data covers period between 1999 and 2018.
  - Micro data covers period between 2010 and 2018.
- Models shown: Micro (Model 1, Model 2, Model 3) and Macro (Model 4, Model 5, Model 6).
- Key coefficient estimates (with standard errors in brackets):
  - CIT rate:
    - Model 1: -6.930*** [1.230]
    - Model 2: -2.856*** [0.616]
  - Other rate:
    - Model 1: 3.055** [1.448]
    - Model 2: 3.630** [1.722]
  - CIT rate - Other rate:
    - Model 1: -5.799*** [0.997]
    - Model 2: -2.990*** [1.253]
    - Model 4: -4.670*** [0.751]
    - Model 5: -3.738*** [0.758]
  - Log (Royalties):
    - Model 3: 0.585*** [0.033]
    - Model 6: 0.133*** [0.034]
  - Log (Total Assets):
    - Model 3: 0.389*** [0.030]
  - Log (Gdp):
    - Model 1: 0.311*** [0.067]
    - Model 2: 0.310*** [0.066]
    - Model 3: 0.455 [1.029]
    - Model 4: 1.246*** [0.043]
    - Model 5: 1.246*** [0.041]
    - Model 6: 1.454*** [0.200]
  - Openness:
    - Model 1: -0.005*** [0.001]
    - Model 2: 0.024** [0.011]
    - Model 3: 0.022 * [0.012]
    - Model 4: 0.019*** [0.002]
    - Model 5: 0.021*** [0.003]
    - Model 6: 0.020*** [0.003]
  - Log (Population):
    - Model 1: -0.219*** [0.084]
    - Model 2: 6.615 [6.196]
    - Model 3: 5.428 [5.681]
    - Model 4: -0.328*** [0.055]
    - Model 5: 0.284 [0.391]
    - Model 6: -0.195 [0.405]
  - Inflation:
    - Model 1: 0.037*** [0.009]
    - Model 2: 0.003 [0.019]
    - Model 3: -0.006 [0.017]
    - Model 4: 0.001 [0.002]
    - Model 5: -0.001 [0.002]
    - Model 6: -0.002 [0.002]
  - Log (Foreign exchange):
    - Model 1: -0.082*** [0.027]
    - Model 2: 1.437 [0.883]
    - Model 3: 0.686 [0.850]
    - Model 4: -0.046*** [0.017]
    - Model 5: -0.059 [0.235]
    - Model 6: 0.034 [0.228]
  - Source:
    - Model 1: -0.983*** [0.393]
    - Model 2: -1.211** [0.579]
    - Model 3: -0.849 [0.807]
    - Model 4: -1.972*** [0.232]
    - Model 5: 0.325 [0.386]
    - Model 6: 0.38 [0.372]
  - Log (Group size):
    - Model 1: 0.781*** [0.049]
    - Model 2: 0.667*** [0.055]
    - Model 3: -0.140 * [0.076]
    - Model 4: -0.362*** [0.073]
    - Model 5: -0.358*** [0.074]
  - Energy price index:
    - Model 2: 0.008 [0.007]
    - Model 6: 0.010*** [0.004]
  - Non-energy price index:
    - Model 2: -0.045 * [0.026]
    - Model 6: -0.012 * [0.007]
- Intercepts:
  - Model 1: 1.946 [1.898]
  - Model 2: -12.66 [22.527]
  - Model 3: -22.392 [23.408]
  - Model 4: -5.759*** [0.889]
  - Model 5: -5.919*** [0.586]
  - Model 6: -8.205*** [2.213]
- Fixed effects and fit:
  - Year fixed effects: Model 1 NO, Model 2 YES, Model 3 YES, Model 4 NO, Model 5 YES, Model 6 YES.
  - Country fixed effects: Model 1 NO, Model 2 YES, Model 3 YES, Model 4 NO, Model 5 NO, Model 6 YES.
  - Observations: Model 1 2564, Model 2 2564, Model 3 2165, Model 4 723, Model 5 723, Model 6 723.
  - Adjusted R2: Model 1 0.187, Model 2 0.320, Model 3 0.477, Model 4 0.683, Model 5 0.680, Model 6 0.891.
- Notes: *, **, and *** indicate statistical significance at the 1, 5 and 10 percent level respectively. Heteroscedasticity robust standard errors in brackets.

### Joint estimation and combined regression results (Table 4)
- Methodology:
  - Differences between semi-elasticities in micro- and macro data tested using a seemingly unrelated regression (SUR) framework.
  - Effects of control variables left unrestricted between macro- and micro-equations.
  - Non-uniform variance of residual errors accounted for by using the inverse variance of the data-specific residuals in a weighted least squares estimation.
- Specification steps:
  - Start with reduced specification controlling for macro variables and commodity price indices.
  - Add country-specific fixed effects and year fixed effects (Model 2).
  - Add controls for productive capacity (Model 3 and Model 4).
- Key pooled weighted least squares estimates (standard errors in brackets):
  - Restricted coefficient on (CIT rate - Other rate):
    - Model 1: -4.081*** [0.545]
    - Model 2: -4.603*** [0.512]
    - Model 3: -3.559*** [0.641]
    - Model 4: -3.557*** [0.641]
  - Difference between macro and micro coefficient:
    - Model 1: 2.889*** [1.153]
    - Model 2: -0.302 [1.622]
    - Model 3: -0.748 [1.463]
    - Model 4: -0.798 [1.445]
- Controls and fit:
  - Real controls: Model 1 no, Model 2 no, Model 3 yes, Model 4 yes.
  - Year fixed effects: Model 1 No, Model 2 Yes, Model 3 Yes, Model 4 yes.
  - Country fixed effects: Model 1 no, Model 2 yes, Model 3 yes, Model 4 yes.
  - Observations: Model 1 3287, Model 2 2888, Model 3 2888, Model 4 2915.
  - Adjusted R2: Model 1 0.654, Model 2 0.704, Model 3 0.838, Model 4 0.838.
- Interpretation:
  - In the first model, the micro-based estimate is larger, with the difference of 2.9 being statistically significant at the 1 percent level.
  - The difference lessens considerably and loses statistical significance when including country- and year-fixed effects in Model 2.
  - Restricting the coefficient on the tax rate differential between micro- and macro data yields a semi-elasticity of -4.6 (Model 2).
  - Adding controls for productive capacity (Model 3) results in the semi-elasticity decreasing to -3.6.
  - Dropping insignificant explanatory variables in Model 4 increases the sample size slightly to 2915 observations while leaving estimated effects largely unchanged.
- Notes: *, **, and *** indicate statistical significance at the 10, 5 and 1 percent level respectively. Heteroscedasticity robust standard errors in brackets. All specifications include year fixed effects. Last specification drops insignificant controls.

### Nonlinearities and comparison to prior literature
- The aggregate semi-elasticity of -3.5 is larger than estimates reported in previous studies.
- Using a meta-analysis approach, Beer and others (2020) report a semi-elasticity of around 1.5 for recent years.

*Source: wpiea2021009-print-pdf - 6. The measured semi-elasticity decreases from -4.67 in the second macro estimation (Model 5) to -3.7 in the third micro estimation*

### 1.5 and 3.2.

### 1.5 and 3.2.

### Factors explaining differences with previous work
- At least three factors could explain differences between the paper’s findings and prior studies:
  - Non-linearities in the response of taxable profits to tax rate differentials, possibly due to fixed costs of tax avoidance. Dowd and others (2017) report that US multinationals with tax haven affiliates are more sensitive to taxation, with the semi-elasticity of taxable profits being increased by between 3.4 and 7.2 among this group.
  - Tax rates may be a suboptimal identification device because actual firm-level tax rates are unobserved, and statutory tax rate changes are infrequent and episodic. Tax rate differentials increase variation and control options, but do not fully resolve endogeneity concerns.
  - Profit shifting incentives vary across countries, industries, and firms; the present study may include firms more sensitive to tax differences. Differences in industry structure and profit drivers between petroleum and mining could imply increased sensitivity in the mining sector.

### Non-Linearities
- Low-tax affiliate indicator in sample:
  - Indicator = 1 for 33 percent of all groups and 60 percent of all observations.
  - Interaction of indicator with tax rate differential yields a negative coefficient (increased sensitivity for groups with low-tax affiliates) but the interaction coefficient is not statistically significant.
  - The indicator (not interacted) suggests low-tax affiliates report, on average, 30 percent less profit, significant at the 1 percent level.
- Polynomial specification:
  - Replacing the low-tax indicator with a 3rd degree polynomial in the tax rate differential yields:
    - Semi-elasticity of affiliates without tax avoidance incentives (company facing no tax rate differential) estimated at 1.5.
    - Coefficients on both second and third orders are large and statistically significant.
    - Evaluated using observed distribution, the central 50 percent of observations would adjust reported profits by between 1 and 3.25 percent in response to a one percentage point increase in corporate taxes.
- Interpretation:
  - Strong presence of extractive affiliates in low-tax jurisdictions helps explain increased sensitivity of taxable profits.

### Alternative identification devices (earnings shocks)
- Strategy: Compare earnings response to shocks in high-tax countries with response in low-tax countries; shocks used include commodity prices and group-level profitability.
- Key empirical findings (Table 5 highlights):
  - Positive direct effect of earning shock variable on reported profits in all specifications.
  - Negative coefficients on the interaction between earning shock and a dummy for high-tax locations.
  - Using oil price index yields large standard errors impeding definite conclusions (Model 3).
  - Interaction significant at the 10 percent level when using mineral prices (Model 4): reported profits increase as mineral prices increase, but significantly less so among firms/countries facing an above average tax rate differential.
  - Micro-level (parent-level profitability) specifications (Models 5 and 6): subsidiary-level recorded profit responds positively to increases in parent-level profitability, but impact is significantly reduced in high-tax jurisdictions.
  - Replacing consolidated profit with consolidated profit minus taxable profits at the subsidiary level (Model 6) leaves effects unchanged, providing additional indirect evidence for profit shifting.
- Table 5 summary statistics:
  - Observations: 2915 (Models 1–4); 2148 (Models 5–6).
  - Adj R2: 0.837, 0.838, 0.828, 0.828, 0.407, 0.407 (Models 1–6 respectively).

### Heterogeneity in profit shifting: industry and country drivers
- Rationale:
  - Profit shifting ability varies by industry, firm, and country-level factors (e.g., tax administrative capacity, anti-abuse measures).
- Differential effects between petroleum and mining:
  - Three reasons mining may be more sensitive to tax differentials:
    - Less transparent pricing and more vertically integrated production in mining, increasing scope for transfer pricing;
    - Petroleum joint ventures and third-party traders can constrain cost inflation opportunities;
    - Production sharing contracts in petroleum may mandate higher standards of auditing/reporting that attenuate profit shifting.
  - Oil indicator in sample:
    - Oil = 1 for 43 percent of observations.
  - Table 6 findings (differential effect between hydrocarbon and mining):
    - Overall tentative evidence that mining operations are more sensitive to tax rate differentials than hydrocarbon entities.
    - Difference in measured semi-elasticities ranges between 1 and 2 and attains statistical significance at the 5 percent level in some specifications (Model 3).
    - When including country-fixed effects, difference lessens to 1.2 and confidence bands increase (Model 4).
  - Table 6 summary statistics:
    - Observations: 3298 (Models 1–2), 2899 (Models 3–4), 2926 (Model 5).
    - Adjusted R2: 0.667, 0.669, 0.71, 0.817, 0.817.
- Country-level drivers (Table 7):
  - Transfer pricing (TP) rules:
    - TP rules indicator takes value 1 when country requires TP documentation and published guidance.
    - Existence of TP rules tends to reduce sensitivity in prior work, but in this sample the effect appears inconsistent and coefficient is insignificant.
  - Interest limitation (thin capitalization / EBITDA or debt-to-equity rules):
    - Indicator = 1 in countries/years that restrict deductibility of interest.
    - Statistically significant interaction suggests sensitivity of taxable profits is reduced by half in countries with effective interest limitation rules.
    - Negative coefficient on the indicator itself suggests reported profits are somewhat smaller in countries with interest limitation rules, possibly due to increased compliance burden.
  - Production sharing contracts:
    - Dummy indicates countries/years with production sharing frameworks.
    - Point estimates suggest reduced sensitivity, but limited variation leads to insignificant results.
  - Governance (control of corruption):
    - Neither the control of corruption variable nor its interaction with tax rate differential is statistically significant.
  - Table 7 summary statistics:
    - Observations: 2904, 2902, 2915, 2892 (Models 1–4).
    - Adj R2: 0.836 across models.

### Accounting for additional heterogeneity (random coefficients and GLS)
- Formalization:
  - Semi-elasticities modeled as β = β + δ′ z_{ckt} + η_k, allowing deterministic drivers z and stochastic group-level η.
- Estimation approach:
  - Estimate group-specific semi-elasticities via dummy-variable regressions; compute weighted average (fixed-effect estimate) using inverse standard error weights.
  - Use estimated group-level semi-elasticities to construct efficient GLS estimates allowing block-diagonal covariance (macro vs micro).
- Table 8 key results:
  - Fixed-effect semi-elasticity estimates: around -4.045***, -3.291***, -2.307*** in the first three specifications (standard errors in brackets).
  - GLS estimates are smaller in magnitude in some specifications (e.g., semi-elasticity of -1.195 in Model 3).
  - Group-specific fixed effects included in all reported models; country-specific fixed effects added in some columns.
  - Results reinforce importance of interest limitation rules, low-tax affiliates, and mining vs hydrocarbon distinction.
  - Observations across models: 2912, 2607, 2631, 2631, 2631, 2631, 2583, 2625.
  - R2 ranges: 0.494 to 0.684 across models.

### Simulation: revenue losses from profit shifting in extractives
- Semi-elasticities used for simulation (Table 9):
  - Interest Limitation = No; Mining operations = Yes: -6.105
  - Interest Limitation = Yes; Mining operations = Yes: -2.862
  - Interest Limitation = No; Mining operations = No: -4.477
  - Interest Limitation = Yes; Mining operations = No: -1.234
- Revenue loss formula (following Hines and Rice (1994)):
  - Loss_c = ε_c (τ_c − τ̅_c) Base_c / τ_c
- Simulation setup:
  - Sample of 138 countries reporting resource-related tax revenues.
  - Revenue losses averaged between 2015 and 2018.
  - Distribution of product of country-specific elasticities with tax rate differentials is winsorized at the top and bottom 1 percent.
- Simulation results (Table 10):
  - Global (138 countries):
    - Median revenue loss (percent of GDP): 0.01
    - Unweighted average (percent of GDP): 0.21
    - Weighted average (percent of GDP): 0.06
    - Revenue loss (USD billion): 43.75
  - Developing Countries (44 countries):
    - Median: 0.04
    - Unweighted average: 0.15
    - Weighted average: 0.08
    - Revenue loss (USD billion): 1.41
  - Emerging Markets (64 countries):
    - Median: 0.01
    - Unweighted average: 0.39
    - Weighted average: 0.15
    - Revenue loss (USD billion): 41.10
  - Advanced Economies (30 countries):
    - Median: 0.00
    - Unweighted average: 0.01
    - Weighted average: 0.00
    - Revenue loss (USD billion): 1.24
- Aggregate context note:
  - The combined GDP of included countries is USD 76.6 trillion, on average, between 2015 and 2018.
- Interpretation:
  - Revenue losses from profit shifting in extractives amount to 0.06 percent of GDP for included countries, or around USD 44 billion.
  - Losses are more frequent and relatively larger in low income and developing countries (median country loses 0.04 percent of GDP); weighted average losses are largest in emerging markets (0.15 percent).

### Conclusion and policy implications
- Main findings:
  - International profit shifting by MNEs in the natural resource sector is a material threat to domestic resource mobilization.
  - The most compelling empirical result is the impact of interest limitation rules in constraining profit shifting.
  - Weak evidence for efficacy of transfer pricing rules in constraining international profit shifting in this sample; mere presence of TP laws without proactive and effective enforcement appears insufficient.
  - Mining firms show increased sensitivity of taxable profits relative to hydrocarbon operations in some specifications, but the statistical significance is limited.
  - No statistically significant effect of production sharing arrangements on profit shifting behavior, though point estimates indicate slightly reduced semi-elasticity.
- Policy recommendations:
  - Concerted effort needed to close current profit shifting channels by:
    - Strengthening and simplifying transfer pricing rules (and improving enforcement to make TP laws effective rather than merely providing “peace of mind”).
    - Limiting interest deductions (interest limitation rules shown to reduce sensitivity).
    - Improving tax treaty practices.
    - Limiting tax incentives.
    - Strengthening investment negotiation practices.
    - For countries imposing capital gains tax on indirect transfers, strengthening protections.

*Source: wpiea2021009-print-pdf - 1.5 and 3.2.*

### REFERENCES

### REFERENCES

### Cited works
- Beer, Sebastian, Ruud de Mooij, and Li Liu, 2020, International Corporate Tax Avoidance: a review of the channels, magnitudes, and blind spots, Journal of economic surveys, 34: 660-688.  
- Beer, Sebastian, and Jan Loeprick, 2015, Profit shifting: Drivers of Transfer (mis)pricing and the Potential of Countermeasures, International Tax and Public Finance, 22, issue 3, p. 426-451.  
- _________, 2017, Taxing Income in the Oil and Gas Sector — Challenges of International and Domestic Profit Shifting, Energy Economics, 61, issue C, p. 186- 198.  
- Blouin, J. L., Harry Huizinga, Luc Laeven, and Gaëtan Nicodème, 2014, "Thin Capitalization Rules and Multinational Firm Capital Structure," CESifo Working Paper Series 4695, CESifo Group Munich .  
- Boadway, Robin and Michael Keen, 2014, Rent Taxes and Royalties in Designing Fiscal Regimes for Non-Renewable Resources, CESifo Working Paper Series No. 4568  
- Buettner, Tiess and Georg Wamser. 2013. Internal Debt and Multinational Profit Shifting: Empirical Evidence from Firm-Level Panel Data, National Tax Journal, National Tax Association, vol. 66(1), pages 63-95, March.  
- _________, Michael Overesch, and Ulrich Schreiber, 2012,"The Impact of Thin-Capitalization Rules on the Capital Structure of Multinational Firms," Journal of Public Economics, Elsevier, vol. 96(11), pages 930-938.  
- Cooper, Graeme and Vann, Richard, 2016, “Transfer Pricing Money – the Chevron Case”, Legal Studies Research Paper 16/72, University of Sydney, Sydney  
- Durst, Michael, 2016, “Improving the Performance of Natural Resource Taxation in Developing Countries”, ICTD Working Paper 60, Institute of Development Studies, Brighton.   
- Fatica, Serena and Gregori, Wildmer Daniel, 2020, How much profit shifting do European banks do?, Economic Modelling, 90, issue C, p. 536-551. 
- Hines, James and Eric M. Rice, 1994, Fiscal Paradise: Foreign Tax Havens and American Business, Quarterly Journal of Economics, 109, 149-82.  
- Saunders-Scott, Molly J., 2015, Substitution Across Methods of Profit Shifting, National Tax Journal, 68, issue 4, p. 1099-1120.  
- Grubert, Harry, 2003, Intangible Income, Intercompany Transactions, Income Shifting, and the Choice of Location, National Tax Journal, 56(1), 221–242.  
- Dischinger Matthias and Nadine Riedel, 2011, Corporate Taxes and the Location of Intangible Assets within Multinational Firms, Journal of Public Economics, Volume 95, Issues 7–8, Pages 691-707.  
- Hebous, Shafik and Niels Johannesen, 2015, “At Your Service! The Role of Tax Havens in International Trade with Services,” CESifo Working Paper Series 5414, CESifo Group Munich .  
- Heckmeyer, Jost and Michael Overesch, 2017, Multinationals’ Profit Response to Tax Differentials: Effect Size and Shifting Channels, Canadian Journal of Economics, forthcoming.  
- International Monetary Fund, 2019, Corporate taxation in the global economy, Washington DC.  
- International Monetary Fund, 2020, Mining and Multinationals in Sub-Saharan Africa, Washington DC.  
- Joshi, P., Outslay, E. and Persson, A., 2020, Does Public Country‐by‐Country Reporting Deter Tax Avoidance and Income Shifting? Evidence from the European Banking Industry. Contemp Account Res. Accepted Author Manuscript. doi:10.1111/1911-3846.12601  
- Crivelli, E., Ruud De Mooij, and Michael Keen, 2016, Base Erosion, Profit Shifting and Developing Countries, FinanzArchiv: Public Finance Analysis, 2016, vol. 72, issue 3, 268-301  
- Griliches, Zvi, and Jerry A. Hausmann, 1986, “Errors in Variables in Panel Data” Journal of Econometrics 31, 93-118  
- King, Mervyn A., 1974, “Taxation and the Cost of Capital,” The Review of Economic Studies, Volume 41, Issue 1, January 1974, Pages 21–35, 
- Lohse, Theresa, and Nadine Riedel, 2013, "Do Transfer Pricing Laws Limit International Income Shifting? Evidence from European Multinationals," CESifo Working Paper Series 4404, CESifo. 
- Merz, Julia and Michael Overesch, 2017, Profit shifting and tax response of multinational banks, Journal of Banking & Finance, Volume 78, Pages 57-68.  
- Overesch, Michael, and Hubertus Wolff, 2019, Financial Transparency to the Rescue: Effects of Country-by-Country Reporting in the EU Banking Sector on Tax Avoidance (Available at SSRN: https://ssrn.com/abstract=3075784 or http://dx.doi.org/10.2139/ssrn.3075784  
- Overesch, Michael, and Georg Wamser, 2010, Corporate Tax Planning and Thin- Capitalization Rules: Evidence from a Quasi-Experiment, Applied Economics, 42, issue 5, p. 563-573.

*Source: wpiea2021009-print-pdf - REFERENCES*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2021/english/wpiea2021009-print-pdf.pdf_