## 1. US affiliates’ debt ratios and corporate taxation

## Source details

**Canonical URL:** [1. US affiliates’ debt ratios and corporate taxation](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp1412.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp1412.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp1412.pdf.json)

---

### Introduction: role of thin capitalization rules
- Interest on debt generally deductible at corporate level → incentive to finance with debt rather than equity.
- Thin capitalization rules restrict deductibility of interest above specified debt levels; regimes vary by:
  - Whether they restrict total debt (relative to assets or equity) or internal debt (debt from related parties relative to equity).
  - Treatment of excessive interest: deny deductibility for excess only, deny all interest, or requalify interest as dividends.
  - Method of application: automatic disallowance versus discretionary/arm’s-length comparisons.
- Multinational enterprises can adjust leverage across subsidiaries via international debt shifting; thin capitalization rules therefore generate cross-country variation in effective tax incentives for debt.
- Paper examines how thin capitalization rules worldwide affect capital structure of foreign affiliates of US multinationals and aggregate implications for firm leverage, interest expense, and market valuation.

### Data and coverage
- International thin capitalization dataset:
  - 54 countries over 1982-2004 (detailed debt information ends in 2004).
  - For 2004: 27 of 54 countries had explicit thin capitalization regimes.
    - 16 countries restricted total leverage.
    - 11 countries restricted internal leverage.
    - 17 countries apply rules automatically; 10 apply discretion based on arm’s-length comparisons.
- Firm-level data:
  - Confidential BEA annual survey of US Direct Investment Abroad.
  - Five benchmark years: 1982, 1989, 1994, 1999 and 2004.
  - 54,273 affiliate-year observations.
- Key variable definitions:
  - Total leverage = total US foreign affiliate debt to assets.
  - Internal leverage = internal debt owed to the US parent to equity.
  - Internal debt share = internal debt relative to total debt.
  - Country tax rate = median corporate tax rate in the affiliate host country estimated annually using affiliate-level effective tax rates.

### Descriptive statistics and trends
- Affiliate-level sample means (Panel A, Table 2):
  - Total leverage mean = 0.552.
  - Internal leverage mean = 0.154.
  - Internal debt share mean = 0.084.
- Time trends (1982–2004):
  - Average Total leverage declined from 59.4% in 1982 to 51.0% in 2004.
  - Average Internal leverage declined from 12.0% in 1982 to 2.5% in 2004.
  - Internal debt share declined from 14.9% to 6.0%.
- Thin capitalization variables (Panel A, Table 2):
  - Thin cap restriction dummy applies in 58.1% of affiliate-year observations.
  - Total leverage restriction applies in 25.2% of observations (or 43.4% of affiliates facing thin cap).
  - Mean Total leverage ratio = 0.904 (equals 1 if no restriction).
  - Mean Internal leverage ratio = 0.895 (equals 1 if no restriction).
  - Arm’s length application present in 40.6% of instances where thin cap regime is in force.
- Aggregate (consolidated firm) statistics (Panel B, Table 2):
  - On average about 15 percent of a multinational firm’s assets are subject to a thin capitalization restriction.
    - About 6 percent of assets are subject to a total leverage restriction.
    - About 9 percent of assets are subject to an internal leverage restriction.
  - Aggregate interest expense ≈ 11 percent (ratio of total interest expense to total assets).

### Univariate evidence on introductions of thin capitalization regimes
- Countries introducing total leverage restrictions (eight with sufficient data):
  - Ex post mean Total leverage lower than ex ante in all eight cases.
  - Average reduction in mean Total leverage after introduction = 4.5%.
  - Declines statistically significant in three-quarters of these countries.
  - Internal leverage mean decline = 4.5%.
- Countries introducing internal leverage restrictions (seven with sufficient data):
  - Mean Total leverage fell in all seven cases; average reduction = 4.6%.
  - Internal leverage declined in all seven instances; overall mean drop = 3.3%.

### Regression evidence — affiliate-level results
- Regression framework:
  - Dependent variables: Total leverage, Internal leverage, Internal debt share.
  - Controls: Country tax rate and affiliate and host-country controls (Net PPE/assets, EBITDA/assets, Log of sales, Growth options, Creditor rights, Political risk, Inflation), plus parent, industry, and year fixed effects.
  - Standard errors clustered by country and industry.
- Country tax rate:
  - Coefficient = 0.197, significant at 1% (higher host-country corporate tax burden associated with higher affiliate Total leverage).
  - In an analogous regression to Desai, Foley and Hines (2004) a 10% higher local tax increases total debt to assets by 2.0%.
- Thin cap restriction dummy:
  - Coefficient = -0.0214, significant at 1% (existence of a thin capitalization regime reduces affiliate Total leverage).
- Total leverage restriction:
  - On average reduces an affiliate’s Total leverage by 1.9%.
  - Stricter regimes (lower allowable total leverage ratios) associated with lower affiliate Total leverage.
  - Regression 4: Total leverage restriction coefficient = -0.0186, significant at 1%.
  - Regression 5: Total leverage restriction coefficient = -0.0445, significant at 1%; interaction with Country tax rate statistically insignificant.
  - Regressions 6 and 7 (Total leverage ratio): coefficients = 0.0307 (significant at 10%) and 0.119 (significant at 5%); interaction of Country tax rate and Total leverage ratio negative and significant in regression 7.
- Internal leverage restriction:
  - On average reduces an affiliate’s Internal leverage by 6.3%.
  - Internal leverage declines as allowable internal leverage ratio decreases.
  - Regression 4 (Table 5): Internal leverage restriction coefficient = -0.0629, significant at 1%.
  - Regression 6: Internal leverage ratio coefficient = 0.196, significant at 1%.
  - Internal restrictions reduce affiliate Total leverage by 0.8% on average (indirect effect).
- Mode of application:
  - Impact on Total leverage about twice as large when application is objective/automatic versus discretionary.
  - Impact on Total leverage from internal leverage restrictions found only when application is automatic.
- Control variable relationships (selected):
  - Net PPE/assets positively and significantly related to Total leverage (~0.0331**).
  - EBITDA/assets negatively and significantly related to Total leverage (-0.469***).
  - Log of sales positively and significantly related to Total leverage.
  - Creditor rights positively and significantly related to Total leverage.
  - Political risk negatively and significantly related to Total leverage.
  - Inflation negatively related to Total leverage.

### Effects on internal debt measures and indirect impacts
- Internal leverage regressions (Table 5):
  - Country tax rate coefficients positive and significant (e.g., 0.368***).
  - Thin cap restriction dummy on Internal leverage: -0.0800***.
  - Internal leverage ratio positively associated with Internal leverage when relaxed (0.196***).
- Internal debt share regressions (Table 6):
  - Country tax rate positive and significant on Internal debt share (0.117*** to 0.130***).
  - Thin cap restriction on Internal debt share negative and significant in some columns (-0.0103*, -0.0348**).
  - Internal leverage ratio positive and significant in some specifications (0.0374*, 0.0576**).
- Indirect impact of internal leverage restrictions on Total leverage (Table 7):
  - Internal leverage restriction coefficient = -0.0082, significant at 1% — about half the impact of a total leverage restriction on Total leverage.

### Discretion (arm’s length) in application of rules
- 40.6% of regimes allow discretion based on arm’s length considerations.
- Table 8 findings:
  - Thin cap restriction coefficient = -0.0276 on Total leverage; interaction with Arm’s length = +0.0169 (both significant) → impact reduced by 61% (=0.0169/0.0276) when arm’s-length discretion applies.
  - For Internal leverage and Internal debt share regressions, many interaction terms are statistically insignificant.
  - Evidence that discretion reduces the impact of thin capitalization rules on Total leverage, but not robustly on internal leverage or internal debt share.

### Speed of adjustment to new regimes (first differences)
- First-difference regressions (Table 9; Δ denotes first differences between benchmark years):
  - Δ Thin cap restriction on Δ Total leverage: coefficient = -0.0113, significant at 10% — introduction reduces total leverage by 1.13% in first years (about half long-term estimate).
  - Δ Total leverage restriction on Δ Total leverage: coefficient = -0.0248, significant at 5% — majority of effect materializes in first years.
  - Δ Thin cap restriction on Δ Internal leverage: coefficient = -0.0452, significant at 5% — about half of corresponding level estimate.
  - Δ Internal leverage restriction on Δ Internal leverage: coefficient = -0.117, significant at 5% — larger than level estimate, suggesting rapid short-run adjustment.
  - Δ Internal debt share adjustments generally insignificant in first year.

### Consolidated firm-level effects (valuation, leverage, interest expenses)
- Tobin’s q and industry-adjusted Tobin’s q (Table 10):
  - Aggregate thin cap restriction coefficient generally insignificant on Tobin’s q.
  - Aggregate total leverage restriction negative and significant on Tobin’s q (e.g., -0.523***; -0.453** in industry-adjusted specifications).
  - Economic effect example: one standard deviation increase in aggregate total leverage restriction of 0.113 → decrease in industry-adjusted Tobin’s q of 0.051 (industry-adjusted Tobin’s q mean = -0.239; std. dev. = 1.539).
- Consolidated worldwide debt-to-assets:
  - Aggregate thin capitalization variables insignificant — suggests debt shifting across countries to keep overall leverage constant.
- Interest expenses to total assets and to debt:
  - Aggregate thin cap restriction negatively and significantly related to interest expense ratios (examples: one standard deviation increase in aggregate thin capitalization restriction variable of 0.186 translates into a decrease in aggregate interest expenses of 0.025; mean = 0.111; std. dev. = 0.199).
  - Aggregate total leverage restriction negative on Aggregate interest expense to debt: -0.126***.
  - Possible mechanisms: international debt shifting to lower-interest countries and/or choice of financial structures implying less risk to external creditors.

### Interpretation, contribution, and main conclusions
- New dataset constructed on thin capitalization rules in 54 countries for 1982-2004, recording existence, stringency, and application method.
- Thin capitalization regimes are associated with lower affiliate leverage:
  - Presence of restrictions on overall debt-to-assets reduces this leverage ratio by 1.9% on average.
  - Restrictions on borrowing from parent relative to equity reduce the targeted leverage ratio by 6.3% on average.
  - Restrictions on borrowing from parent reduce overall debt-to-assets of affiliate by 0.8% on average.
- Quantitative design parameters matter: maximum allowable ratios (Total leverage ratio, Internal leverage ratio) show significant associations with borrowing outcomes in some specifications.
- Application mode matters: automatic application yields higher impact than discretionary (arm’s length) application.
- Short-term impacts generally represent part of long-term effects, but Total and Internal leverage can respond quickly to respective restrictions.
- Thin capitalization rules reduce overall firm valuation (consistent with reduced worldwide deductibility of interest) and lower aggregate interest expense.
- Overall implication: thin capitalization rules substantially affect capital structure within multinational firms and are an important factor for taxation and corporate debt studies.

*Source: _wp1412 (IMF working paper excerpt).*

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

### References

### Listed Tables
- Table 1. Characteristics of thin capitalization rules at year-end 2004 ...............................................23
- Table 2. Descriptive statistics and correlations ..................................................................................25
- Table 3. The introductive of thin capitalization rules and borrowing ................................................28
- Table 4. Thin capitalization restrictions on total debt and total borrowing by US affiliates .............29
- Table 5. Thin capitalization restrictions on internal debt and borrowing
     from the parent relative to equity ........................................................................................30
- Table 6. Thin capitalization restrictions on the share of internal debt

and borrowing from the parent ...........................................................................................31
- Table 7. Internal leverage restrictions and the total debt and total borrowing
     by US affiliates ...................................................................................................................32
- Table 8. The application of thin capitalization rules and US affiliate financing ...............................33
- Table 9. The short-term impact of changes in thin capitalization rules on

US affiliate financing ..........................................................................................................34
- Table 10. Impact of thin capitalization rules on Tobin’s Q, debt, and interest expense at the

consolidated firm level ........................................................................................................35

### Figures
- Figure

*Source: _wp1412 - References.*

### 1. US affiliates’ debt ratios and corporate taxation ..................................................................

### 1. US affiliates’ debt ratios and corporate taxation

### Introduction: role of thin capitalization rules
- Interest on debt is generally deductible at the corporate level, creating incentives to finance with debt rather than equity.
- Many countries use thin capitalization rules to restrict deductibility of interest above specified debt levels; regimes vary by:
  - Whether they restrict total debt (relative to assets or equity) or internal debt (debt from related parties relative to equity).
  - The treatment of excessive interest (deny deductibility for excess only, deny all interest, or requalify interest as dividends).
  - The method of application (automatic disallowance versus discretionary/arm’s-length comparisons).
- Multinational enterprises can adjust leverage across subsidiaries via international debt shifting; thin capitalization rules are therefore an important source of cross-country variation in effective tax incentives for debt.
- This paper examines how thin capitalization rules worldwide affect capital structure of foreign affiliates of US multinationals and the aggregate implications for firm leverage, interest expense, and market valuation.

### Data and coverage
- International thin capitalization dataset:
  - 54 countries over 1982-2004 (detailed debt information ends in 2004).
  - For 2004: 27 of 54 countries had explicit thin capitalization regimes.
    - 16 countries restricted total leverage (total debt to assets/equity).
    - 11 countries restricted internal leverage (debt from related parties to equity).
    - 17 countries apply rules automatically; 10 apply discretion based on arm’s-length comparisons.
- Firm-level data:
  - Confidential BEA annual survey of US Direct Investment Abroad.
  - Five benchmark years: 1982, 1989, 1994, 1999 and 2004.
  - 54,273 affiliate-year observations.
- Key variable definitions (as used in the analysis):
  - Total leverage = total US foreign affiliate debt to assets.
  - Internal leverage = internal debt owed to the US parent to equity.
  - Internal debt share = internal debt relative to total debt.
  - Country tax rate = median corporate tax rate in the affiliate host country estimated annually using affiliate-level effective tax rates.

### Descriptive statistics and trends
- Sample means (Panel A, Table 2):
  - Total leverage mean = 0.552.
  - Internal leverage mean = 0.154.
  - Internal debt share mean = 0.084.
- Time trends (1982–2004):
  - Average Total leverage declined from 59.4% in 1982 to 51.0% in 2004.
  - Average Internal leverage declined from 12.0% in 1982 to 2.5% in 2004.
  - Internal debt share declined from 14.9% to 6.0%.
- Thin capitalization variables (Panel A, Table 2):
  - Thin cap restriction dummy applies in 58.1% of affiliate-year observations.
  - Total leverage restriction applies in 25.2% of observations (or 43.4% of affiliates facing thin cap).
  - Mean Total leverage ratio (stringency measure) = 0.904 (equals 1 if no restriction).
  - Mean Internal leverage ratio = 0.895 (equals 1 if no restriction).
  - Arm’s length application present in 40.6% of instances where thin cap regime is in force.
- Aggregate (consolidated firm) statistics (Panel B, Table 2):
  - On average about 15 percent of a multinational firm’s assets are subject to a thin capitalization restriction.
    - About 6 percent of assets are subject to a total leverage restriction.
    - About 9 percent of assets are subject to an internal leverage restriction.
  - Aggregate interest expense ≈ 11 percent (ratio of total interest expense to total assets).

### Univariate evidence on introductions of thin capitalization regimes
- Evidence from countries that introduced restrictions during sample:
  - Eight countries introduced total leverage restrictions with sufficient data; ex post mean Total leverage is lower than ex ante in all eight cases.
    - Average reduction in mean Total leverage after introduction of a total leverage restriction = 4.5%.
    - Declines in Total leverage statistically significant in three-quarters of these countries.
    - Internal leverage also declined in all eight cases; mean decline = 4.5%.
  - Seven countries introduced internal leverage restrictions; in all seven cases mean Total leverage fell.
    - Average reduction in mean Total leverage after introduction of an internal leverage restriction = 4.6%.
    - Internal leverage declined in all seven instances; overall mean drop = 3.3%.

### Regression evidence — affiliate-level results
- Regression setup:
  - Dependent variables: Total leverage, Internal leverage, Internal debt share.
  - Controls: Country tax rate and standard affiliate and host-country controls (Net PPE/assets, EBITDA/assets, Log of sales, Growth options, Creditor rights, Political risk, Inflation), plus parent, industry, and year fixed effects.
  - Standard errors clustered by country and industry.
- Key regression coefficients and findings:
  - Country tax rate:
    - Coefficient = 0.197, significant at 1% (higher host-country corporate tax burden associated with higher affiliate Total leverage).
    - In an analogous regression to Desai, Foley and Hines (2004) the paper finds a 10% higher local tax increases the total debt to assets ratio by 2.0% (consistent with prior studies reporting ~2% effects).
  - Thin cap restriction dummy:
    - Coefficient = -0.0214, significant at 1% (existence of a thin capitalization regime reduces affiliate Total leverage).
  - Existence of a total leverage restriction:
    - On average reduces an affiliate’s Total leverage by 1.9%.
    - Stricter regimes (lower allowable total leverage ratios) are associated with lower affiliate Total leverage.
  - Existence of an internal leverage restriction:
    - On average reduces an affiliate’s Internal leverage by 6.3%.
    - Internal leverage declines as allowable internal leverage ratio decreases.
    - Existence of internal leverage restrictions reduces affiliate Total leverage by 0.8% on average (internal restrictions materially affect overall leverage).
  - Mode of application:
    - The impact of thin capitalization rules on Total leverage is about twice as large when application is objective/automatic versus discretionary.
    - Impact on Total leverage from internal leverage restrictions is found only when application is automatic.
  - Control variable relationships (selected):
    - Net PPE/assets positively and significantly related to Total leverage.
    - EBITDA/assets negatively and significantly related to Total leverage.
    - Log of sales positively and significantly related to Total leverage.
    - Creditor rights positively and significantly related to Total leverage.
    - Political risk negatively and significantly related to Total leverage.
    - Inflation negatively related to Total leverage.

### Short-term dynamics and consolidated firm effects
- Short-term (first-year) responses to introductions:
  - Short-term response of thin capitalization rules generally smaller than average/long-term response.
  - However, Total leverage responds quickly and fully to introduction of a total leverage restriction; Internal leverage responds quickly and fully to introduction of an internal leverage restriction.
- Consolidated firm-level effects:
  - Thin capitalization regimes reduce the multinational firm’s aggregate interest expense bill.
  - Thin capitalization regimes lower the overall valuation of the firm, consistent with reduced worldwide deductibility of interest.
  - Aggregate measures constructed:
    - Aggregate thin cap restriction = assets subject to thin cap restriction / total assets of consolidated firm.
    - Aggregate total (internal) leverage restriction = assets subject to total (internal) leverage restriction / total consolidated assets.
    - Aggregate interest expense = total interest expense / total consolidated assets.
    - Aggregate interest expense to debt = total interest expense / total debt.

### Interpretation and contribution
- Thin capitalization rules are an effective policy instrument to constrain affiliate leverage within prescribed limits and have measurable effects on:
  - Affiliate-level Total leverage and Internal leverage.
  - The mix of internal versus external debt.
  - The consolidated multinational’s interest expense and market valuation.
- The paper extends prior literature by:
  - Constructing a detailed international dataset on thin capitalization regimes for 54 countries over 1982-2004.
  - Distinguishing regimes that target total versus internal leverage and measuring stringency and application method.
  - Exploiting cross-country and time variation and within-multinational comparisons to identify effects.
  - Considering aggregate implications for firm leverage, interest expense, and valuation.

*Source: _wp1412 - 1. US affiliates’ debt ratios and corporate taxation*

### 0.272 and 0.0183, respectively, that are insignificant.

### _wp1412 - 0.272 and 0.0183, respectively, that are insignificant.

### Effects of Total Leverage Restrictions on Total Leverage
- Regression 4: Total leverage restriction variable coefficient = -0.0186, significant at 1%.
- Regression 5: Interaction of Country tax rate with Total leverage restriction; Total leverage restriction coefficient = -0.0445, significant at 1%; interaction term statistically insignificant.
- Regressions 6 and 7 (using Total leverage ratio):
  - Total leverage ratio coefficient in regression 6 = 0.0307, significant at 10%.
  - Total leverage ratio coefficient in regression 7 = 0.119, significant at 5%.
  - Interaction term of Country tax rate and Total leverage ratio in regression 7 is negative and statistically significant, suggesting tighter thin capitalization limits are less constraining in affiliates facing higher tax rates.
- Summary: Thin capitalization regimes, particularly the Total leverage restriction and Total leverage ratio variables, have a material impact on Total leverage.

### Effects of Internal Leverage Restrictions on Internal Leverage
- Regression 1 (Table 5): Country tax rate coefficient = 0.368, significant at 1%.
- Regression 2: Thin cap restriction dummy coefficient = -0.0800, significant at 1%.
- Regression 3: Interaction of Country tax rate and Thin cap restriction dummy statistically insignificant.
- Regression 4: Internal leverage restriction coefficient = -0.0629, significant at 1%.
- Regression 5: Internal leverage restriction and its interaction with Country tax rate both statistically insignificant.
- Regression 6: Internal leverage ratio coefficient = 0.196, significant at 1% (evidence that a quantitative relaxation of an internal leverage restriction increases internal leverage).
- Regression 7: Interaction of Country tax rate and Internal leverage ratio yields statistically insignificant coefficients for both the ratio and the interacted variable.
- Summary: Thin capitalization regimes that restrict internal leverage have a material impact on Internal leverage.

### Internal Debt Share and Internal Leverage Restrictions
- Regression 1 (Table 6): Country tax rate coefficient on Internal debt share = 0.117, significant at 1%.
- Regression 2: Thin cap restriction coefficient on Internal debt share negative and significant at 10%.
- Regression 6: Internal leverage ratio coefficient positive and significant (internal debt expands relative to total debt if internal leverage restriction is loosened).
- Regression 7: Internal leverage restriction and its interaction with Country tax rate estimated with significant positive and negative coefficients, respectively, suggesting internal debt to total debt ratio increases relatively little in high-tax countries as the internal leverage restriction is relaxed.

### Indirect Impact of Internal Leverage Restrictions on Total Leverage
- Table 7, regression 1: Internal leverage restriction coefficient = -0.0082, significant at 1%.
  - Impact of an internal leverage restriction on total leverage is about half the impact of a total leverage restriction on the same variable (comparison with regression 4 in Table 4).
- Table 7, regression 2: Internal leverage restriction and its interaction with Country tax rate both insignificant.
- Regressions analogous to Table 4:
  - Internal leverage ratio enters with a positive coefficient significant at 5% in one specification.
  - Internal leverage ratio and its interaction with Country tax rate are both insignificant in another specification.
- Summary: Restrictions on internal leverage appear to have a significant impact on total leverage.

### Robustness to Location Decisions
- Unreported regressions rerun for sample of parents with affiliates in the same set of countries over 1982 to 2004 yield qualitatively unaltered results.

### Discretion (Arm’s length) in Application of Rules
- 40.6% of thin capitalization regimes allow some discretion based on arm’s length considerations.
- Table 8 (Total leverage regressions):
  - Regression 1: Thin cap restriction coefficient = -0.0276; interaction with Arm’s length = 0.0169; both significant. Interpretation: impact of thin capitalization rule on total leverage is reduced by 61% (=0.0169/0.0276) if rules are applied based on arm’s length.
  - Regression 2: Total leverage restriction coefficient = -0.0229, significant at 1%; interaction with Arm’s length = 0.0109, insignificant (point estimate suggests 48% reduction = 0.0109/0.0229 but effect not significant).
  - Regression 3: Internal leverage restriction coefficient = -0.0200 (significant at 5%); interaction with Arm’s length = 0.0244 (significant at 10%); suggests discretion fully cancels out the effect of internal leverage restrictions on total leverage.
- Table 8 (Internal leverage and Internal debt share regressions):
  - Regressions 4 and 5: interaction terms statistically insignificant.
  - Regressions 6 and 7: thin capitalization variables not statistically significant.
- Summary: Some evidence that discretion reduces the impact of thin capitalization rules on total leverage, but not on internal leverage or the internal debt share.

### Speed of Adjustment to New Regimes (First Differences)
- First-difference regressions between benchmark years (periods of five or seven years) reported in Table 9 (Δ denotes first differences).
- Δ Thin cap restriction (introduction equals 1; abolishment equals -1):
  - Regression 1 (Δ Total leverage): coefficient = -0.0113, significant at 10% — introduction reduces total leverage by 1.13% in its first years (about half of long-term estimate of -0.0214 from Table 4 regression 2).
- Δ Total leverage restriction:
  - Regression 2 (Δ Total leverage): coefficient = -0.0248, significant at 5% — similar magnitude to level regression coefficient -0.0186 in Table 4 regression 4, suggesting majority of effect materializes in first years.
- Δ Internal leverage regressions:
  - Regression 4 (Δ Internal leverage): Δ Thin cap restriction coefficient = -0.0452, significant at 5% — about half of corresponding level estimate.
  - Regression 5: Δ Internal leverage restriction coefficient = -0.117, significant at 5% — almost twice the corresponding level estimate, suggesting rapid and possibly overreactive short-run adjustment in internal leverage.
- Δ Internal debt share regressions 6 and 7: Δ Thin cap restriction and Δ Internal leverage restriction estimated with insignificant coefficients (first-years adjustment in this variable tends to be small).
- Summary: Total and internal leverage generally take longer than a few years to adjust to regime changes overall, but Total (Internal) leverage reacts relatively quickly to the introduction of a Total (Internal) leverage restriction.

### Aggregate Firm Implications (Valuation, Leverage, Interest Expenses)
- Tobin’s q (market value of overall firm’s equity plus debt to total assets):
  - Regression 1 (Table 10): Aggregate thin cap restriction coefficient insignificant.
  - Regression 2: Aggregate total leverage restriction obtains a negative and significant coefficient; Aggregate internal leverage restriction insignificant.
  - Regressions 3 and 4 (industry-adjusted Tobin’s q): similar results.
  - Economic effect: a one standard deviation increase in aggregate total leverage restriction variable of 0.113 translates into a decrease in industry-adjusted Tobin’s q of 0.051 (mean of industry-adjusted Tobin’s q = -0.239; standard deviation = 1.539).
- Consolidated worldwide debt-to-assets (regressions 5 and 6): aggregate thin capitalization variables insignificant — suggests debt shifting across countries to keep overall leverage constant.
- Interest expenses to total assets (regressions 7 and 8):
  - Regression 7: interest expense ratio negatively and significantly related to aggregate thin capitalization regime variable.
  - Regression 8: negatively and significantly related to separate aggregate total and internal leverage restriction variables.
  - Possible explanations: international debt shifting to lower-interest countries and/or choice of financial structures implying less risk to external creditors.
- Aggregate interest expenses scaled by total debt (regressions 9 and 10): qualitatively similar results to assets scaling.
  - Economic effect: one standard deviation increase in aggregate thin capitalization restriction variable of 0.186 translates into a decrease in aggregate interest expenses of 0.025 (mean = 0.111; standard deviation = 0.199).
- Summary: Thin capitalization rules affect intra-firm capital structure with aggregate implications for interest expense and firm valuation.

### Main Conclusions
- New dataset constructed on thin capitalization rules in 54 countries for 1982-2004; dataset provides existence, stringency, and application information.
- In sample, thin capitalization regimes restrict affiliate total debt to assets in about 43% of cases; other regimes restrict related-party indebtedness relative to equity.
- Average effects:
  - Presence of restrictions on overall debt-to-assets reduces this leverage ratio by 1.9%.
  - Restrictions on borrowing from parent relative to equity reduce the targeted leverage ratio by 6.3% on average.
  - Restrictions on borrowing from parent reduce overall debt-to-assets of affiliate by 0.8% on average.
- Application mode matters: automatic application yields higher impact than discretionary (arm’s length) application.
- First-year impacts generally represent part of long-term effects, but Total and Internal leverage can respond quickly to respective restrictions.
- Thin capitalization rules reduce overall firm valuation (consistent with reduced worldwide deductibility of interest).
- Overall implication: thin capitalization rules have a substantial effect on capital structure within multinational firms and are an important factor for taxation and corporate debt studies.

*Source: Excerpt from the provided IMF working paper content.*

### REFERENCES

### _wp1412 - REFERENCES

### References
- Altshuler, R., and H. Grubert, 2003, “Taxes, repatriation strategies and multinational financial policy,” Journal of Public Economics, Vol. 87, pp. 73-107.
- Auerbach, A.J., 2002, “Taxation and corporate financial policy.” In A.J. Auerbach, and M. Feldstein, (eds.), Handbook of Public Economics, Volume III (New York: North-Holland).
- Brosens, L., 2004, “Thin capitalization rules and EU law,” EC Tax Review, Vol. 4, pp. 188-213.
- Buettner, T., M. Overesch, U. Schreiber, and G. Wamser, 2012, “The impact of thin-capitalization rules on the capital structure of multinational firms,” Journal of Public Economics, Vol. 96, pp. 930-938.
- DeAngelo, H., and R.W. Masulis, 1980, “Optimal capital structure under corporate and personal taxation,” Journal of Financial Economics, Vol. 8, pp. 3-29.
- Desai, M., F. Foley, and J. R. Hines, 2001, “Repatriation taxes and dividend distortions,” National Tax Journal, Vol. 54, pp. 859-851.
- Desai, M., F. Foley, and J. R. Hines, 2004, “A multinational perspective on capital structure choice and internal capital markets,” Journal of Finance, Vol. 59, pp. 2451-87.
- Desai, M., and J.R. Hines, 1999, “‘Basket’ cases: Tax incentives and international joint venture participation by American multinational firms,” Journal of Public Economics, Vol. 71, pp. 379–402.
- Djankov, S., C. McLiesh, and A. Shleifer, 2007, “Private credit in 129 countries,” Journal of Financial Economics, Vol. 84, pp. 299-329.
- Froot, K.A., and J.R. Hines, 1995, “Interest allocation rules, financing patterns, and the operation of U.S. multinationals,” pp. 272-307. In M. Feldstein, J.R. Hines, R.G. Hubbard, (eds.), The Effect of Taxation on Multinational Corporations, (Chicago: University of Chicago Press).
- Graham, J.R., 1996, “Debt and the marginal tax rate,” Journal of Financial Economics, Vol. 41, pp. 41-73.
- Graham, J.R., 2000, “How big are the tax benefits of debt?,” Journal of Finance, Vol. 55, pp. 1901-1941.
- Graham, J.R., 2003, “Taxes and corporate finance: A review,” Review of Financial Studies, Vol. 16, pp. 1074–1128.
- Graham, J.R., and C. Harvey, 2001, “The theory and practice of corporate finance: Evidence from the field,” Journal of Financial Economics, Vol. 60, pp. 187–243.
- Haufler, A., and M. Runkel, 2012, “Firms' financial choices and thin capitalization rules under corporate tax competition,” European Economic Review, Vol. 56, pp. 1087-1103.
- Hines, J., and E. Rice, 1994, “Fiscal paradise: Foreign tax havens and American business,” Quarterly Journal of Economics, Vol. 109, pp. 149-182.
- Huizinga, H., L. Laeven, and G. Nicodème, 2008, Capital structure and international debt shifting, Journal of Financial Economics 88, 80-118.
- MacKie-Mason, J., 1990, “Do taxes affect corporate financing decisions?,” Journal of Finance, Vol. 45, pp. 1471-1493.
- Mintz, J. and A. Weichenrieder, 2005, “Taxation and the financial structure of German outbound FDI,” CESifo Working Paper No. 1612.
- Myers, S.C., and N.S. Majluf, 1984, “Corporate financing and investment decisions when firms have information that investors do not have,” Journal of Financial Economics, Vol. 13, pp. 187–221.
- Newberry, K.J., and D.S. Dhaliwal, 2001, “Cross-jurisdictional income shifting by U.S. multinationals: Evidence from international bond offerings,” Journal of Accounting Research, Vol. 39, pp. 643-662.
- OECD, 2013, Action Plan on Base Erosion and Profit Shifting, Paris.
- Overesch, M., and G. Wamser, 2010, “Corporate tax planning and thin-capitalization rules: evidence from a quasi-experiment,” Applied Economics, Vol. 42, pp. 563-573.
- Rajan, R.G., and L. Zingales, 1995, “What do we know about capital structure? Some evidence from international data,” Journal of Finance, Vol. 50, pp. 1421-60.
- Wamser, G., 2008, “The impact of thin capitalization rules on external debt usage: a propensity score matching approach,” CESifo Working Paper No. 62.
- Weichenrieder, A., and H. Windischbauer, 2008, “Thin capitalization rules and company responses,” CESifo Working Paper No. 2456.

### Table 1 caption and variable definitions (as presented)
- Table 1. Characteristics of thin capitalization rules at year-end 2004
- This table shows key characteristics of thin capitalization rules in selected countries at year-end 2004. D denotes total debt; IID denotes single related party internal debt; TID denotes total internal debt; TIFD denotes total internal foreign debt; TFD denotes total foreign debt; E denotes total equity; IIE denotes individual internal equity; TIE denotes total internal equity; TIFE denotes total internal foreign equity; TFE denotes total foreign equity; A denotes total assets; RoE denotes return on equity.
- Data are from International Bureau for Fiscal Documentation, Brosens (2004), and national tax authorities.
- Column headings shown: (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)
- Partial row label shown: Country
- Partial row label shown: Explicit rule
- Partial row label shown: Year of first

*Source: _wp1412 - REFERENCES*

### introduction

### _wp1412 - introduction

### Thin capitalization rules: cross-country coverage and design features
- Many countries report the presence or absence of thin capitalization rules, with details on:
  - Definition numerator, definition denominator, ratio, substantial shareholding, direct (Dir) or indirect (Ind) shareholding, application of Automatic (Auto) or arm's length (AL) considerations, remedy (non-deductibility (Nd) or reclassification as dividend (Div)), applicability for all or exceeding debt, and whether rules apply to interest of interest of interest (I) or net interest (NI).
- Selected country design details (as presented):
  - Argentina: Yes 1999; Dir; Auto; Div; Exc; I; Rules apply to Foreign.
  - Australia: Yes 1987; Ind; AL; Nd; Exc; I; 15%; Rules apply to Foreign.
  - Canada: Yes 1972; Ind; Auto; Nd; Exc; I; 25%; Rules apply to Foreign.
  - France: Yes 1979; IID; E; 1.5; 50% Dir; Auto; Nd; Exc; I; All.
  - Germany: Yes 1994; IID; IIE; 1.5; 25% Ind; AL; Div; Exc; I; All (note: since 2008, Germany applies an earnings stripping rule).
  - Italy: Yes 2004; IID; IIE; 5; 25% Ind; AL; Div; Exc; I; All (note: repealed in 2008 and replaced by earnings stripping rule for some firms).
  - Japan: Yes 1992; TIFD TIFE; 3; 50% Ind; Auto; Nd; Exc; I; Foreign (ratio condition noted: total debt to third parties to total equity should be over 3).
  - Netherlands: Yes 2004; D E; 3; 33% Ind; Auto; Nd; Exc; NI; All.
  - United Kingdom: Yes 1988; D E; 1; 75% Ind; AL; Nd; Exc; I; All.
  - United States: Yes 1989; D E; 1.5; 50% Ind; AL; Nd; Exc; I; Foreign (presented for information only; safe harbor provision; fact and circumstances approach).
- Notes highlight legal and operational nuances (examples):
  - Note 2/: EU members cannot discriminate between national and other EU companies following Lankhorst-Hohorst (2003); many thin cap rules were extended to domestic companies or repealed for EU companies from 2004.
  - Remedies range from non-deductibility to reclassification as dividend; some countries apply market/arm’s length rate considerations.

### Descriptive statistics (affiliate and consolidated firm levels)
- Panel A. Affiliate-level (N denotes observations):
  - Total leverage: N 54,269; Mean 0.552; Median 0.502; Std. dev. 0.420.
  - Internal leverage: N 52,092; Mean 0.154; Median 0.000; Std. dev. 1.884.
  - Internal debt share: N 51,524; Mean 0.084; Median 0.003; Std. dev. 0.352.
  - Country tax rate: N 54,273; Mean 0.299; Median 0.314; Std. dev. 0.110.
  - Thin cap restriction (dummy): N 54,273; Mean 0.581; Median 1.000; Std. dev. 0.493.
  - Arm’s length (application dummy): N 54,273; Mean 0.236; Median 0.000; Std. dev. 0.424.
  - Net PPE/assets: N 54,273; Mean 0.202; Median 0.115; Std. dev. 0.230.
  - EBITDA/assets: N 54,273; Mean 0.122; Median 0.101; Std. dev. 0.231.
  - Log of sales: N 54,273; Mean 9.884; Median 10.247; Std. dev. 2.735.
- Panel B. Aggregate firm-level:
  - Tobin’s q: N 4,558; Mean 1.349; Median 1.015; Std. dev. 1.257.
  - Industry-adjusted Tobin’s q: N 4,558; Mean -0.236; Median -0.253; Std. dev. 1.222.
  - Aggregate debt: N 4,558; Mean 0.253; Median 0.232; Std. dev. 0.190.
  - Aggregate interest expense: N 9,705; Mean 0.023; Median 0.017; Std. dev. 0.022.
  - Aggregate thin cap restriction (asset-weighted): N 9,705; Mean 0.156; Median 0.097; Std. dev. 0.186.

### Correlations (affiliate-level highlights from Panel C)
- Selected pairwise correlations (exact values as presented):
  - Total leverage with Internal debt share: 0.238.
  - Total leverage with Country tax rate: 0.068.
  - Internal leverage with Internal debt share: 0.416.
  - Thin cap restriction with Total leverage restriction: 0.569.
  - Total leverage restriction with Internal leverage restriction: -0.343.
  - EBITDA/assets with Total leverage: -0.262.
  - Log of sales with Total leverage: 0.072.
- Bolded (italicized) correlations in the source indicate significance at the 1% (5%) level.

### Empirical evidence: impact of thin capitalization and other restrictions on affiliate borrowing
- Table 3 — Mean changes around introduction of restrictions:
  - Panel A (restrictions on total leverage): average Total leverage before 0.582; after 0.537. Average Internal leverage before 0.106; after 0.061.
  - Panel B (restrictions on internal leverage): average Total leverage before 0.624; after 0.578. Average Internal leverage before 0.090; after 0.057.
  - Individual country examples show statistically significant declines in Total leverage and Internal leverage in many cases (stars in the table indicate significance at 10%, 5%, 1% levels).
- Table 4 — Regressions: Thin capitalization restrictions and total debt (dependent variable: Total leverage; observations 54,269):
  - Country tax rate coefficients: 0.197*** (col 1), 0.223*** (col 2), 0.217*** (col 3), 0.187*** (col 4), 0.174*** (col 5), 0.193*** (col 6), 0.493*** (col 7).
  - Thin cap restriction: -0.0214*** (col 1); -0.0272 (col 2).
  - Total leverage restriction: -0.0186*** (col 3); -0.0445** (col 6).
  - Total leverage ratio (maximum allowable): 0.0307* (col 5); 0.119** (col 7).
  - Net PPE/assets positive: ~0.0331** across columns.
  - EBITDA/assets negative: -0.469*** across columns.
  - Observations 54,269; R-squared 0.097.
- Table 5 — Regressions: Thin cap restrictions and internal debt relative to equity (dependent variable: Internal leverage; observations 52,092):
  - Country tax rate positive and significant: 0.368*** (col 1), 0.468*** (cols 2–6).
  - Thin cap restriction: -0.0800*** (col 1); -0.0806 (col 2).
  - Internal leverage ratio: 0.196*** (col 5).
  - EBITDA/assets positive though small: ~0.0943***.
  - Log of sales strongly positive: ~1.871*** to 1.952***.
  - Observations 52,092; R-squared 0.003.
- Table 6 — Regressions: Thin cap restrictions on share of internal debt (dependent variable: Internal debt share; observations 51,524):
  - Country tax rate positive and significant: 0.117*** to 0.130*** (columns 1–6).
  - Thin cap restriction: -0.0103* (col 1); -0.0348** (col 2).
  - Internal leverage ratio: 0.0374* (col 5); 0.0576** (col 6).
  - Net PPE/assets strongly positive: 0.107***.
  - EBITDA/assets negative: -0.121***.
  - Observations 51,524; R-squared 0.026.
- Table 7 — Internal leverage restrictions and total debt (dependent variable: Total leverage; observations 54,269):
  - Country tax rate positive and significant: 0.211*** (col 1), 0.224*** (col 2), 0.194*** (col 3).
  - Internal leverage restriction: -0.0082*** (col 1).
  - Internal leverage ratio: 0.00463** (col 2).
  - Observations 54,269; R-squared 0.097.
- Table 8 — Application of thin capitalization rules and US affiliate financing (various dependent variables; observations vary):
  - Country tax rate coefficients positive and significant across outcomes (e.g., Total leverage 0.224*** (col 1); Internal leverage 0.463*** (col 4); Internal debt share 0.128*** (col 6)).
  - Thin cap restriction negative on Total leverage: -0.0276*** (col 1).
  - Thin cap restriction interaction with Arm’s length: +0.0169** for Total leverage (col 2).
  - Internal leverage restriction negative on Internal leverage: -0.0200** (col 4) and -0.0575** (col 5).
  - Net PPE/assets: positive for Total leverage (0.0354**), negative for Internal leverage (-0.0751), positive for Internal debt share (0.108***).
  - EBITDA/assets negative for Total leverage (-0.487***) and for Internal debt share (-0.129***), but positive for Internal leverage (0.112***).
  - Observations: Total leverage regressions 54,262 (R-squared 0.099); Internal leverage regressions 52,092 (R-squared 0.003); Internal debt share regressions 51,524 (R-squared 0.027).
- Table 9 — Short-term (one-period change) effects of changes in thin capitalization rules (panel of changes; observations 15,202 or fewer depending on outcome):
  - Δ Country tax rate associated with Δ Total leverage: 0.167** (col 1), 0.175** (col 2), 0.171** (col 3).
  - Δ Thin cap restriction associated with Δ Total leverage: -0.0113* (col 1); Δ Thin cap restriction associated with Δ Internal leverage: -0.0452** (col 3).
  - Δ Total leverage restriction: -0.0248** (col 2).
  - Δ Net PPE/assets positive and significant for changes in leverage (0.160***).
  - Δ EBITDA/assets negative and significant for changes in leverage (-0.339***).
  - Observations 15,202 for change regressions; R-squared ~0.061 for change in total leverage.
- Table 10 — Consolidated firm-level impacts (observations vary by dependent variable):
  - Aggregate thin cap restriction: negative but generally not significant on Tobin’s q and firm-level aggregates as presented (examples: -0.182 (Tobin’s q col 1); -0.00338 (Aggregate debt col 5) with standard errors reported).
  - Aggregate total leverage restriction: significantly negative on Tobin’s q: -0.523*** (col 1) and -0.453** (col 2).
  - Aggregate total leverage restriction negative on Aggregate interest expense to debt: -0.126*** (col 5).
  - Net PPE/Assets large negative coefficients on Tobin’s q: -0.773*** (col 1) and -0.778*** (col 2).
  - EBITDA/Assets strongly positive for Tobin’s q: 5.621*** (col 1) and 5.610*** (col 2); negative for Aggregate debt and Aggregate interest expense measures.
  - Aggregate observations and R-squared: Tobin’s q regressions N 4,558 (R-squared 0.279); Aggregate debt regressions N 9,705 (R-squared ~0.110–0.111); Aggregate interest expense regressions N 9,705 (R-squared ~0.116); Aggregate interest expense to debt regressions N 9,705 (R-squared ~0.135–0.136).

### Key empirical interpretations drawn from the reported results
- Thin capitalization rules are associated with lower affiliate leverage:
  - Cross-sectional and change regressions show negative coefficients on thin cap restriction variables for Total leverage (e.g., -0.0214*** in Table 4; -0.0276*** in Table 8).
  - Internal leverage and internal debt share respond negatively to the presence of thin cap restrictions (e.g., Internal leverage: -0.0800*** in Table 5; Internal debt share: -0.0103* in Table 6).
- Country tax rate is positively associated with affiliate leverage measures:
  - Country tax rate coefficients are consistently positive and significant across Total leverage and Internal leverage regressions (examples: 0.197*** in Table 4; 0.368*** in Table 5; 0.224*** in Table 8).
- Quantitative design parameters matter:
  - The maximum allowable ratios (Total leverage ratio, Internal leverage ratio) and explicit quantitative restrictions show significant associations with borrowing outcomes in some specifications (e.g., Total leverage ratio 0.119** in Table 4 col 7; Internal leverage ratio 0.196*** in Table 5 col 6).
- Remedial approaches and enforcement modalities affect responses:
  - The interaction Thin cap restriction × Arm’s length is positive for Total leverage (0.0169** in Table 8 col 2), indicating how application of arm’s length considerations can moderate the association between thin cap rules and affiliate leverage.
- Short-term changes and dynamic responses:
  - One-period changes in thin cap restrictions have measurable short-term effects on affiliate leverage (Δ Thin cap restriction -0.0113* on Δ Total leverage in Table 9).
- Consolidated firm-level effects:
  - Aggregate measures of restrictions have detectable associations with firm-level outcomes (e.g., aggregate total leverage restriction is negatively associated with Tobin’s q: -0.523*** in Table 10 col 1).
  - Firm fundamentals (Net PPE/assets and EBITDA/assets) are strong predictors of Tobin’s q, aggregate debt, and interest expense measures.

*Source: _wp1412 - introduction (IMF working paper content provided in the source PDF unit).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2014/_wp1412.pdf_