## wp1907

## Source details

**Canonical URL:** [wp1907](https://www.imf.org/-/media/files/publications/wp/2019/wp1907.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/wp/2019/wp1907.pdf.md)
- [Structured JSON version](/-/media/files/publications/wp/2019/wp1907.pdf.json)

---

### Introduction and purpose
- Examines revenue consequences of adopting a destination-based cash-flow tax (DBCFT): a cash-flow tax with border adjustment (imports taxed, exports excluded), enabling full expensing of investment and denying interest expense deduction.
- Objective: calculate potential DBCFT revenues using an estimated tax base constructed with national accounts data for 80 countries and compare to actual CIT revenues.
- Key conceptual points:
  - DBCFT is a rent tax collected at destination.
  - Theoretical revenue impact ambiguous because existing CIT bases are eroded by tax incentives, profit shifting, and interest deductibility.
  - Empirical approach focuses on tax base and revenue effects; analyzes global adoption versus unilateral adoption.

### Estimation methodology and data
- Primary DBCFT revenue formula (equation (1)):
  - R_DBCFT = τ (Π − I + M − X)
    - Π = nonfinancial corporate gross operating surplus
    - I = corporate investment
    - M = imports
    - X = exports
    - τ = statutory CIT rate (including local/surtaxes where applicable)
- Alternative base specifications:
  - R_DBCFT = τ (VA − CE + M − X)  (equation (2))
  - R_DBCFT = τ (S − CE)  (equation (3))
- Change in revenue definition (equation (4)):
  - ΔR = R_DBCFT − R_CIT
- Correction for profit shifting (equation (5)):
  - R_DBCFT_adjusted = R_DBCFT + τ [Π_TWZ_adjusted − Π_unadjusted − ((M−X)_TWZ_adjusted − (M−X)_unadjusted)]
- Data sources and sample:
  - Gross operating surplus and investment: OECD where possible; United Nations Statistics Division to extend sample.
  - CIT revenues: World Revenue Longitudinal Database (WoRLD).
  - CIT statutory rates: IMF FAD Tax Policy Rates Database (including local/surtaxes).
  - Trade and controls: WEO (Spring 2018).
  - Resulting sample: operating surplus and investment for 80 countries; CIT revenues for 73 countries.
  - Balanced 10-year panel: 2002–2011 with 48 countries; year with most observations: 2011 with 63 countries comparable.
- Selected descriptive statistics for balanced 10-year panel (Table 1; all in percent of GDP unless otherwise noted):
  - Observations: 480
  - Imports (M) — Mean: 45.7; S.D.: 24.1; p5: 17.4; Median: 39.9; p95: 81.5
  - Exports (X) — Mean: 45.6; S.D.: 27.2; p5: 16.8; Median: 40.3; p95: 83.5
  - Investment (I) — Mean: 13.8; S.D.: 4.5; p5: 6.6; Median: 13.4; p95: 22.0
  - Gross operating surplus (Π) — Mean: 26.2; S.D.: 6.1; p5: 17.5; Median: 25.3; p95: 37.8
  - CIT revenue (R_CIT) — Mean: 3.3; S.D.: 1.7; p5: 1.4; Median: 2.9; p95: 6.3
  - Tax rate (τ) (in percent) — Mean: 26.8; S.D.: 7.3; p5: 15.0; Median: 28.0; p95: 38.9

### Results under global adoption: main findings
- Average outcome:
  - On global average across countries and years, DBCFT revenues would be close to current CIT revenues, given prevailing statutory rates.
  - Substantial cross-country heterogeneity creates winners and losers.
- Distribution of outcomes:
  - Approximately one third of sample countries would lose substantially, one third would stay around the same level, and one third would gain substantially.
  - Some countries could end up with negative DBCFT revenue (due to refunding losses).
- Notable country-level examples (2002–2011 averages):
  - Largest estimated DBCFT revenue gains: Guatemala, Mexico, Honduras, Greece, and the United States.
  - Largest estimated DBCFT revenue losses: Luxembourg, Norway, Kazakhstan, China, and Cyprus.
- Revenue-neutral tax rates (2011):
  - For two thirds of countries the change in rate to maintain revenues would be negative or a small increase.
  - Almost 90 percent of countries could maintain revenues with rates of no more than 40 percent; some countries would require very high (unrealistic) rates or have negative bases prohibiting revenue neutrality.
- Trade balance and border adjustment:
  - Strong negative relationship between the trade balance (X−M) and revenue change: trade deficit countries tend to gain; trade surplus countries tend to lose.
  - The border-adjustment effect dominates the cash-flow component: border adjustment drives most gains/losses, with the cash-flow component having smaller and opposite-signed effects across trade positions.
- Advanced vs developing economies:
  - Developing countries on average tend to be beneficiaries of a DBCFT in the balanced sample; advanced economies tend to lose more on average.
  - Caveat: balanced sample includes fewer developing economies; wider unbalanced sample yields similar results.
- Natural resource dependence:
  - Resource-rich countries would on average lose DBCFT revenue relative to CIT revenue, primarily to the extent rents are earned by non-residents.
  - Losses concentrated in CIT receipts from the resource sector; sector-specific resource taxes could be adjusted to offset.
- Multivariate determinants (selected coefficients from Table 2; units in percent of GDP):
  - Trade balance coefficient: -0.227*** (0.028) in (1); ranges -0.171*** to -0.260*** across specifications.
  - Investment coefficient: -0.263*** (0.043) in (1); ranges -0.237*** to -0.311*** across specifications.
  - Resource rents: -0.087** (0.038) in one specification; -0.082** (0.038) in another.
  - GDP per capita (logged): -0.686* (0.359) in (3); -0.671* (0.379) in (5).
  - Observations vary by specification: e.g., Observations 1,097 (73 countries) in (1); 889 (71 countries) in (2) and (5). R-squared up to 0.445.
- Profit shifting adjustments:
  - Using Tørsløv, Wier, and Zucman (2018) adjusted series for 2015 (sample reduced to 37 countries), median difference between adjusted and unadjusted DBCFT revenue estimates is just 0.04 percent of GDP.
  - Median adjustments: trade balance adjustment median 0.17 percent of GDP; gross operating surplus adjustment median 0.13 percent of GDP; effects largely offsetting.
  - Exceptions where net adjustment exceeds 1 percent of GDP (absolute value): Luxembourg (+6.6 percent of GDP), Ireland (+3.2 percent of GDP), Cyprus (+2.4 percent of GDP). Additional countries with adjustments >¼ percent of GDP: Netherlands (+0.7 percent of GDP), Belgium (+0.3 percent of GDP).
- Trade balance reversals and NIIP risk:
  - Short-term winners with trade deficits could lose revenues in the long run if current account deficits reverse and NIIP dynamics change.
- Volatility and cyclicality:
  - DBCFT revenues are on average more volatile than CIT revenues (standard deviation of annual revenue higher for DBCFT in sample 2000–2014).
  - Fiscal cyclicality regressions (equation (6)):
    - CIT: acyclical or countercyclical depending on cyclical indicator.
    - DBCFT: more strongly aligned with cycle; procyclical using the output gap and acyclical when using GDP growth as the cycle indicator.
  - Conclusion: DBCFT is less likely to act counter-cyclically and is, on average, both more volatile and less stabilizing than the existing CIT.
- Loss-making firms and refundable credits:
  - DBCFT requires immediate refunds for losses to function; exporters may be in systematic loss positions.
  - Risks of revenue leakage from fraudulent or unsuccessful firms; loss prevalence varies across countries (Orbis data for 2005–2014 used to illustrate loss/sales ratios).

### Unilateral adoption: spillovers and strategic responses
- Mechanisms of spillovers:
  - Profit shifting channels: transfer pricing, royalties, location of debt — shift profits toward DBCFT country; DBCFT exempts exports and disallows interest deduction.
  - Real relocation: rent-earning export activity may relocate to DBCFT country; tax effects mainly harm origin CIT country.
- Modeling revenue spillovers (equation (7)):
  - dT_c / T_c = (α_c ε_K + ε_S) ω_c,D t_D
    - α assumed 1/3 (capital intensity)
    - ε_K = 2.4 (elasticity of real investment w.r.t. tax rate)
    - ε_S = 1.5 (profit-shifting elasticity)
    - ω bilateral weight proxied by bilateral FDI links
    - t_D = corporate income tax rate of DBCFT country (cut to zero in unilateral adoption)
  - Note: equation yields change in multinational profits revenue; total tax revenue impact may be overestimated because share of profits by multinationals in total profits not explicitly available.
- Illustrative simulations (2011 data):
  - United States as first mover: sizeable spillovers and revenue losses for close partners (e.g., Mexico, Canada).
  - United Kingdom first mover: smaller but still sizeable spillovers for close partners.
  - Brazil or India first mover: much smaller global spillovers.
- Incentives for policy reactions and classification under U.S. adoption (2011 simulation):
  - Policy responses considered: do nothing; reduce CIT statutory rate to zero; adopt a DBCFT.
  - Country categories:
    - Green: gain revenue by adopting DBCFT (clear incentive to follow).
    - Orange: lose revenue adopting DBCFT but would lose even more by staying with CIT (incentive to follow to minimize loss).
    - Red: lose more by adopting DBCFT than by staying with CIT (no revenue incentive to follow).
  - Second-round adoption: green and orange countries following can induce further countries to adopt due to changing spillovers; further rounds may occur but can stall.
- Overall unilateral-adoption conclusion:
  - Spillovers can be sizeable if the adopting country is large and globally integrated.
  - Spillovers can create incentives for other countries to adopt DBCFT, potentially producing a cascade of adoption among economically-linked countries.
  - Some countries would never have a revenue incentive to adopt a DBCFT (could face massive revenue losses or negative bases).

### Appendix: incidence and international revenue implications (theoretical examples)
- Setup:
  - Country DB introduces DBCFT at rate τ; rest of world (ROW) does not change tax system.
  - Currency of DB appreciates by τ / (1 − τ); wages unchanged.
  - Analysis focuses on firms by production location (DB or ROW), sales location (DB or ROW), and ownership (DB residents or ROW residents).
- Experiment 1 — location of sales and production (Table A1 summary):
  - Domestic operation (production and sales in DB), exporter (production DB, sales ROW), importer (production ROW, sales DB).
  - Tax payments:
    - Domestic operation tax payment = (p_DB − c_DB) τ
    - Exporter tax payment = −c_DB τ (refund on costs; exports untaxed)
    - Importer tax payment = p_DB τ (imports not deductible)
  - Net profit (after tax) same across firm types located in DB: (p_DB − c_DB)(1 − τ)
  - Tax borne (incidence on rents) equal for all firm types located in DB: (p_DB − c_DB) τ, provided trade balances hold.
  - Revenue collected from firms located in DB: exporters generate refunds; importers generate tax payments; with balanced trade these offset.
- Experiment 2 — ownership (Table A2 summary):
  - Tax payments/refunds are only due to firms located in DB; the location of sales determines payment/refund; ownership does not affect tax payment.
  - Incidence depends on residence of owners:
    - DB residents owning firms in DB bear the incidence.
    - DB residents owning firms in ROW bear incidence via currency appreciation reducing foreign-currency profits.
    - ROW residents owning firms in DB remit DBCFT but do not bear incidence because currency appreciation offsets burden.
    - ROW residents owning firms in ROW unaffected.
- Untaxed assets:
  - DB-residents holding assets denominated in ROW currency (bonds, land) bear DBCFT burden via appreciation (no tax remitted).
  - DB-residents holding assets denominated in DB currency do not bear burden.
  - ROW-residents holding assets denominated in DB currency gain from appreciation and face no DBCFT burden.
- Consolidated implication:
  - Government tax receipts come from firms located in DB.
  - Economic incidence of the DBCFT falls on DB-resident rent-earning investors; foreign residents are largely unaffected in incidence despite remitting payments if located in DB.
  - If DB residents own substantial foreign capital, incidence and long-run revenue implications differ (ownership structure and NIIP affect ultimate incidence).

*Source: IMF Working Paper wp1907 — Appendix Tables (selected excerpts).*

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

### References

### Tables
- 1. Descriptive Statistics for Balanced 10-year Panel ......................................................................................... 11
- 2. Change in Revenue and Explanatory Variables .............................................................................................. 18
- 3. The Cyclicality of the CIT and DBCFT ................................................................................................................. 23

### Figures
- 1. DBCFT and CIT Revenues, (Averages over 2002-2011) ............................................................................... 12
- 2. Revenue-Neutral Tax Rates in 2011 ................................................................................................................... 13
- 3. Trade balance and Revenue Change, Excluding Resource-Rich Economies ...................................... 14
- 4. Border Adjustment versus Cash Flow Component of Revenue Change .............................................. 14
- 5. Revenues by Income Group and Level .............................................................................................................. 15
- 6. Revenues by Resource Dependence .................................................................................................................. 17
- 7. The Impact of Adjusting for Profit Shifting ...................................................................................................... 19
- 8. Trade Balance and Net International Investment Position, 2014 ............................................................ 21
- 9. Standard Deviation of Annual Revenue, 2000-2014 .................................................................................... 22
- 10. Average Ratios of Losses to Sales, 2005-14 (percent) .............................................................................. 24
- 11. Revenue Spillovers from Unilateral DBCFT Adoption, 2011 ................................................................... 27
- 12. Revenue Changes under DBCFT and CIT Following Hypothetical U.S. 
 Adoption of DCBCFT in 2011 ................................................................................................................ 28
- 13. Second-Round Spillovers ..................................................................................................................................... 29

### Appendix
- Appendix 
  - International Revenue and Incidence Implications of a DBCFT .................................................................... 32

_4_

### Appendix Tables

### wp1907 - Appendix Tables

### Introduction and purpose
- The paper examines revenue consequences of adopting a destination-based cash-flow tax (DBCFT), a cash-flow tax with border adjustment (imports taxed, exports excluded) that enables full expensing of investment and denies interest expense deduction.
- Objective: calculate potential DBCFT revenues using an estimated tax base constructed with national accounts data for 80 countries and compare to actual CIT revenues.
- Key conceptual points:
  - DBCFT is a rent tax collected at destination.
  - Theoretical revenue impact ambiguous because existing CIT bases are eroded by tax incentives, profit shifting, and interest deductibility.
  - Empirical approach focuses on tax base and revenue effects; global adoption versus unilateral adoption are analyzed.

### Estimation methodology and data
- Primary DBCFT revenue formula (equation (1)):
  - R_DBCFT = τ (Π − I + M − X)
    - Π = nonfinancial corporate gross operating surplus
    - I = corporate investment
    - M = imports
    - X = exports
    - τ = statutory CIT rate (including local/surtaxes where applicable)
- Alternative base specifications:
  - R_DBCFT = τ (VA − CE + M − X)  (equation (2))
  - R_DBCFT = τ (S − CE)  (equation (3))
- Change in revenue definition (equation (4)):
  - ΔR = R_DBCFT − R_CIT
    - Uses actual CIT revenues to reflect real-world complications and compliance.
- Correction for profit shifting using Tørsløv, Wier, and Zucman (2018) (equation (5)):
  - R_DBCFT_adjusted = R_DBCFT + τ [Π_TWZ_adjusted − Π_unadjusted − ((M−X)_TWZ_adjusted − (M−X)_unadjusted)]
- Data sources and sample:
  - Gross operating surplus and investment: OECD where possible; United Nations Statistics Division to extend sample.
  - CIT revenues: World Revenue Longitudinal Database (WoRLD).
  - CIT statutory rates: IMF FAD Tax Policy Rates Database (including local/surtaxes).
  - Trade and controls: WEO (Spring 2018).
  - Resulting sample: operating surplus and investment for 80 countries; CIT revenues for 73 countries.
  - Balanced 10-year panel: 2002–2011 with 48 countries; year with most observations: 2011 with 63 countries comparable.
- Selected descriptive statistics for balanced 10-year panel (Table 1; all in percent of GDP unless otherwise noted):
  - Observations: 480
  - Imports (M) — Mean: 45.7; S.D.: 24.1; p5: 17.4; Median: 39.9; p95: 81.5
  - Exports (X) — Mean: 45.6; S.D.: 27.2; p5: 16.8; Median: 40.3; p95: 83.5
  - Investment (I) — Mean: 13.8; S.D.: 4.5; p5: 6.6; Median: 13.4; p95: 22.0
  - Gross operating surplus (Π) — Mean: 26.2; S.D.: 6.1; p5: 17.5; Median: 25.3; p95: 37.8
  - CIT revenue (R_CIT) — Mean: 3.3; S.D.: 1.7; p5: 1.4; Median: 2.9; p95: 6.3
  - Tax rate (τ) (in percent) — Mean: 26.8; S.D.: 7.3; p5: 15.0; Median: 28.0; p95: 38.9

### Results under global adoption: main findings
- Average outcome:
  - On global average across countries and years, DBCFT revenues would be close to current CIT revenues, given prevailing statutory rates.
  - Substantial cross-country heterogeneity creates winners and losers.
- Distribution of outcomes:
  - Approximately one third of sample countries would lose substantially, one third would stay around the same level, and one third would gain substantially.
  - Some countries could end up with negative DBCFT revenue (due to refunding losses).
- Notable country-level examples (2002–2011 averages):
  - Largest estimated DBCFT revenue gains: Guatemala, Mexico, Honduras, Greece, and the United States.
  - Largest estimated DBCFT revenue losses: Luxembourg, Norway, Kazakhstan, China, and Cyprus.
- Revenue-neutral tax rates (2011):
  - For two thirds of countries the change in rate to maintain revenues would be negative or a small increase.
  - Almost 90 percent of countries could maintain revenues with rates of no more than 40 percent; some countries would require very high (unrealistic) rates or have negative bases prohibiting revenue neutrality.
- Trade balance and border adjustment:
  - Strong negative relationship between the trade balance (X−M) and revenue change: trade deficit countries tend to gain; trade surplus countries tend to lose.
  - The border-adjustment effect dominates the cash-flow component: border adjustment drives most gains/losses, with the cash-flow component having smaller and opposite-signed effects across trade positions.
- Advanced vs developing economies:
  - Developing countries on average tend to be beneficiaries of a DBCFT in the balanced sample; advanced economies tend to lose more on average.
  - Caveat: balanced sample includes fewer developing economies; wider unbalanced sample yields similar results.
- Natural resource dependence:
  - Resource-rich countries would on average lose DBCFT revenue relative to CIT revenue, primarily to the extent rents are earned by non-residents.
  - Losses concentrated in CIT receipts from the resource sector; sector-specific resource taxes could be adjusted to offset.
- Multivariate determinants (Table 2: change in revenue and explanatory variables; coefficients in percent of GDP; GDP per capita logged):
  - Trade balance coefficient: -0.227*** (0.028) in (1); ranges -0.171*** to -0.260*** across specifications — negative and statistically significant.
  - Investment coefficient: -0.263*** (0.043) in (1); ranges -0.237*** to -0.311*** — negative and statistically significant.
  - Resource rents: -0.087** (0.038) in specification showing it, and -0.082** (0.038) in another — negative and statistically significant.
  - GDP per capita (logged): -0.686* (0.359) in (3); -0.671* (0.379) in (5) — higher income associated with lower revenue gains.
  - Interaction terms (Investment*Advanced, Trade balance*Advanced) not statistically significant in reported specifications.
  - Observations vary by specification: e.g., Observations 1,097 (73 countries) in (1); 889 (71 countries) in (2) and (5). R-squared up to 0.445.
- Profit shifting adjustments:
  - Using Tørsløv, Wier, and Zucman (2018) adjusted series for 2015 (sample reduced to 37 countries), median difference between adjusted and unadjusted DBCFT revenue estimates is just 0.04 percent of GDP.
  - Median adjustments: trade balance adjustment median 0.17 percent of GDP; gross operating surplus adjustment median 0.13 percent of GDP; effects largely offsetting.
  - Exceptions where net adjustment exceeds 1 percent of GDP (absolute value): Luxembourg (+6.6 percent of GDP), Ireland (+3.2 percent of GDP), Cyprus (+2.4 percent of GDP). Additional countries with adjustments >¼ percent of GDP: Netherlands (+0.7 percent of GDP), Belgium (+0.3 percent of GDP).
- Trade balance reversals and NIIP risk:
  - Short-term winners with trade deficits (gaining from DBCFT) could lose revenues in the long run if current account deficits reverse and NIIP dynamics change.
  - Figure 8 analysis: countries split by trade balance and net international investment position (NIIP) to indicate short-term vs future revenue risk.
- Volatility:
  - DBCFT revenues are on average more volatile than CIT revenues (standard deviation of annual revenue higher for DBCFT in sample 2000–2014), though not universally so across all countries.
- Cyclicality (econometric assessment using standard fiscal cyclicality regression, equation (6)):
  - CIT: found to be acyclical or countercyclical depending on cyclical indicator (output gap or growth rate).
  - DBCFT: more strongly aligned with cycle; procyclical in regressions using the output gap and acyclical when using GDP growth as the cycle indicator.
  - Conclusion: DBCFT is less likely to act counter-cyclically and is, on average, both more volatile and less stabilizing than the existing CIT.
- Loss-making firms and refundable credits:
  - DBCFT requires immediate refunds for losses to function (some firms, notably exporters, may be in systematic loss positions).
  - This raises risks of revenue leakage from fraudulent or unsuccessful firms; loss prevalence varies across countries (loss/sales ratios illustrated using Orbis data for 2005–2014).

### Unilateral adoption: spillovers and strategic responses
- Mechanisms of spillovers:
  - Profit shifting channels: transfer pricing, royalties, location of debt — these increase measured exports of DBCFT country and imports of CIT country; DBCFT exempts exports and disallows interest deduction so DBCFT country faces no direct tax on shifted profits.
  - Real relocation: rent-earning export activity has incentive to relocate to DBCFT country; tax effects mainly harm the origin CIT country.
- Modeling revenue spillovers (Beer, Klemm, and Matheson (2018) style; equation (7)):
  - dT_c / T_c = (α_c ε_K + ε_S) ω_c,D t_D
    - α assumed 1/3 (capital intensity)
    - ε_K (elasticity of real investment w.r.t. tax rate) = 2.4 (De Mooij and Ederveen, 2008 meta-study)
    - ε_S (profit-shifting elasticity) = 1.5 (Beer, De Mooij, and Liu, 2018 meta-study)
    - ω bilateral weight proxied by bilateral FDI links (FDI_In + FDI_Out relative to world)
    - t_D = corporate income tax rate of DBCFT country (cut to zero in unilateral adoption)
  - Note: equation yields change in multinational profits revenue; total tax revenue impact may be overestimated because share of profits by multinationals in total profits not explicitly available.
- Illustrative simulations for first mover choices (2011 data):
  - United States as first mover: sizeable spillovers and revenue losses for close partners (e.g., Mexico, Canada).
  - United Kingdom first mover: smaller but still sizeable spillovers for close partners.
  - Brazil or India first mover: much smaller global spillovers.
- Incentives for policy reactions:
  - Countries facing revenue losses from unilateral DBCFT adoption by a partner have three broad responses considered here:
    - Do nothing (stay with CIT and suffer spillovers).
    - Reduce their own CIT statutory rate to zero (extreme and costly).
    - Adopt a DBCFT themselves (follow).
  - Simulation (U.S. adoption in 2011): countries are categorized:
    - Green: gain revenue by adopting DBCFT (clear incentive to follow).
    - Orange: lose revenue adopting DBCFT but would lose even more by staying with CIT (incentive to follow to minimize loss).
    - Red: lose more by adopting DBCFT than by staying with CIT (no revenue incentive to follow).
  - Second-round adoption: if green and orange countries follow, this can induce further countries to adopt due to changing spillovers; further rounds may occur but can stall if remaining countries are small or face large revenue costs.
- Overall unilateral-adoption conclusion:
  - Spillovers can be sizeable if the adopting country is large and globally integrated.
  - Spillovers can create incentives for other countries to adopt DBCFT (first- and second-round effects), potentially producing a cascade of adoption among economically-linked countries.
  - Some countries would never have a revenue incentive to adopt a DBCFT (could face massive revenue losses or negative bases).

### Appendix: incidence and international revenue implications (theoretical examples)
- Assumptions:
  - Country DB introduces DBCFT at rate τ; ROW does not change tax system.
  - Currency of DB appreciates by τ / (1 − τ) in textbook fashion; wages unchanged.
  - Focus on firms by production location (DB or ROW), sales location (DB or ROW), and ownership (DB residents or ROW residents).
- First experiment: location of sales and production (Table A1 summary)
  - Domestic operation (production and sales in DB), exporter (production DB, sales ROW), importer (production ROW, sales DB).
  - Tax payments:
    - Domestic operation tax payment = (p_DB − c_DB) τ
    - Exporter receives refund on costs: tax payment = −c_DB τ (exports untaxed)
    - Importer pays tax on sales: tax payment = p_DB τ (imports not deductible)
  - Net profit (after tax) same across firm types located in DB: (p_DB − c_DB)(1 − τ)
  - Tax borne (incidence on rents) equal for all firm types located in DB: (p_DB − c_DB) τ, provided trade balances hold.
  - Revenue collected from firms located in DB: exporters generate refunds, importers generate high tax payments; with balanced trade these offset.
- Second experiment: ownership (Table A2 summary)
  - Tax payments/refunds are only due to firms located in DB; the location of sales determines payment/refund; ownership does not affect tax payment.
  - Incidence depends on residence of owners:
    - DB residents owning firms in DB bear the incidence.
    - DB residents owning firms in ROW bear incidence via currency appreciation reducing foreign-currency profits (no DBCFT remittance).
    - ROW residents owning firms in DB remit DBCFT but do not bear incidence because currency appreciation offsets burden.
    - ROW residents owning firms in ROW unaffected.
- Untaxed assets:
  - DB-residents holding assets denominated in ROW currency (bonds, land) bear DBCFT burden via appreciation (no tax remitted).
  - DB-residents holding assets denominated in DB currency do not bear burden.
  - ROW-residents holding assets denominated in DB currency gain from appreciation and face no DBCFT burden.
- Consolidated implication:
  - Government tax receipts come from firms located in DB.
  - Economic incidence of the DBCFT falls on DB-resident rent-earning investors; foreign residents are largely unaffected in incidence despite remitting payments if located in DB.
  - If DB residents own substantial foreign capital, incidence effects differ (discussion continues beyond provided excerpt).

*Source: IMF Working Paper wp1907 — Appendix Tables (selected excerpts).*

### introduction this means that they bear a high burden. In the long term this translates into high

### wp1907 - introduction this means that they bear a high burden. In the long term this translates into high

### Effect on revenues and trade balance
- Owning foreign capital entitles them to income streams from abroad, which can finance a trade deficit, which boosts revenues.
- Hence countries with a strong IIP will ultimately benefit in revenue terms from a DBCFT.
- Even though this is collected on imports, the burden is still on residents.

### Incidence of the DBCFT under location-specific rents
- Suppose there is a large location-specific rent earned by exporters in DB (e.g., a natural resource, or network benefit from Silicon Valley).
  - If the asset is domestically owned, the burden of the DBCFT will be borne by owners even though no tax is collected from the exporting firm.
  - The owners will (ultimately) use their income from exports to finance imports, which are taxed.
  - If the asset is foreign owned, however, there is no tax burden. There are revenue losses: the profits of nonresidents weaken the income balance, requiring trade surpluses.

### Notes and cross-references
- 35
- 25
- 25 Auerbach (2017b) also points out that it will depend on the rates of return on assets. E.g., a negative IIP need not require future trade surpluses, if income earned on the smaller foreign assets is higher than the financing cost of the larger foreign liabilities.

*wp1907 - introduction this means that they bear a high burden. In the long term this translates into high*

---


_Source: https://www.imf.org/-/media/files/publications/wp/2019/wp1907.pdf_