## 1. Comparative Structure of Tax Revenue in Sub-Saharan Africa Countries, 1980–96

## Source details

**Canonical URL:** [1. Comparative Structure of Tax Revenue in Sub-Saharan Africa Countries, 1980–96](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04178.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04178.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04178.pdf.json)

---

### I. Introduction and research questions
- Trade liberalization typically involves reduction and unification of tariffs and relaxation of quantitative barriers, and may be accompanied by currency devaluation and domestic tax reform.
- Key research questions examined:
  - Relationship between trade liberalization and tax revenues, including taxes from international trade and other taxes, controlling for accompanying macroeconomic changes.
  - Sensitivity of the relationship to the index of liberalization and to econometric specification.
  - Relationship between exchange rate changes (devaluation/ depreciation) and tax revenues.
  - Differences between CFA franc and non-CFA franc countries in revenue response to trade liberalization.

### Theoretical considerations and mechanisms
- International trade taxes — mechanisms and elasticities:
  - Typical tax forms: import duties (usually ad valorem), export taxes (usually ad valorem or specific/ complex).
  - Revenue effects determinants: change in tariff level/coverage; price elasticity of import demand; elasticity of supply of import substitutes; starting tariff levels (possibility of tariff reductions increasing revenue if initial tariffs above revenue-maximizing levels).
  - Exchange rate effects: real depreciation raises trade tax base in domestic currency (valuation effect); if import demand inelastic short-term, valuation effect may dominate → higher import tax revenues; export-side effects typically small in short run.
  - Heterogeneity: largest tariff cuts often on goods with more elastic demand → import expansion may offset revenue losses.
- Taxes on goods and services:
  - Tariffs, excises, and VAT are typically sequenced: tariffs applied to import value; excises on base inclusive of tariffs; VAT on base inclusive of tariffs and excises.
  - Tariff cuts reduce the import-stage base for excises and VAT; offsetting effects include import value rises, decline in domestic output of import-substituting goods, and administrative constraints.
  - Exchange rate depreciation raises excise and VAT collections from imports via pass-through, conditional on demand elasticities; if excises are specific, depreciation does not automatically raise excise revenues unless excises adjusted.
- Taxes on income, profits, and capital gains:
  - Short-run trade liberalization effects via profitability changes of import-competing firms; long-run via growth and expanded tax base.
  - Exchange rate changes affect personal and corporate taxes mainly through inflation and revaluation of foreign-denominated assets/liabilities; no uniform prediction across tax systems.
  - Inflation may cause bracket creep if tax brackets not indexed.

### Data, variables, and econometric approach
- Two proxies for trade liberalization:
  - Trade liberalization 1: international trade as a share of GDP (higher → greater liberalization).
  - Trade liberalization 2: collected tariff measured by ratio of import duties to value of imports (decline → greater liberalization).
- Sample period: 1980–1996.
- Dependent variables (shares of GDP): total tax revenue; taxes on income and profits; taxes on goods and services; international trade taxes.
- Control variables (many in log form): GDP per capita (LGDP), agricultural share (AGRI), industry share (IND), real government consumption share (LGC), net transfers of aid (LNAID), terms of trade (LTOT), real effective exchange rate (LEXCHANGE, 1995=100; increase = appreciation), inflation (INFLATION), openness/trade liberalization index (OPEN), CFA dummy (CFA).
- Model: first-order dynamic panel with lagged dependent variable; country fixed effects and time effects.
- Estimation: GMM using orthogonal deviations (Arellano and Bover (1995) style); treat LGDP, INFLATION, LGC, and OPEN as endogenous and instrument with t-2 values; report (m1), (m2) and Sargan tests.

### Key empirical patterns from simple plots (1980–1996)
- Trade liberalization (openness) shows no clear pattern with major tax components.
- Collected tariff measure positively correlated with overall tax revenues and taxes on international trade (higher collected tariff = less liberal → higher revenues).
- Real effective exchange rate (appreciation) appears positively related to higher overall tax revenues and international trade taxes in simple plots.
- Regional differences: CFA countries often show positive correlation between trade liberalization measures and revenues; non-CFA countries show weaker or negative correlations. The 1994 CFA devaluation produces a visible sharp break.

### Selected GMM regression results and diagnostics (1980–1996 full sample)
- Persistence:
  - Lagged dependent variable is positive and significant across specifications, indicating partial adjustment/persistence.
  - Examples: lag_dv = 0.563    [5.02]; lag_dv = 0.538    [5.79]; lag_dv = 0.738   [5.59]; lag_dv = 0.632   [5.35]; lag_dv = 0.472   [4.40]; lag_dv = 0.463   [3.73]; lag_dv = 0.765   [7.98]; lag_dv = 0.727   [9.14].
- Total tax revenue (selected coefficients):
  - Openness specification (column (1)): lgdp = -0.007  [-0.33]; agri = 0.137    [2.58]; ind = 0.214    [3.21]; lgc = 0.019    [2.28]; lnaid = 0.036    [1.21]; ltot = 0.029    [2.19]; lexchange = -0.011  [-0.62]; inflation = -0.023  [-4.11]; lib_index1 = 0.001    [0.11]; lib_index1cfa = 0.042    [1.38].
  - Collected tariff specification (column (2)): lgdp = -0.002  [-0.12]; agri = 0.036    [0.42]; ind = 0.133    [1.66]; lgc = 0.012    [1.01]; lnaid = 0.017    [0.52]; ltot = 0.003    [0.19]; lexchange = -0.030  [-2.07]; inflation = -0.020  [-4.88]; lib_index2 = -0.010   [-0.92]; lib_index2cfa = 0.026    [0.92].
- Income taxes (selected):
  - Column (4): lgdp = 0.027   [1.96]; agri = -0.054  [-2.05]; ind = 0.054   [1.29]; lexchange = 0.007   [0.76]; inflation = -0.003  [-0.96].
  - Interpretation: agricultural share negatively linked to income tax revenues; using collected tariff measure, per capita income positively linked and the coefficient on collected tariff suggests trade liberalization may be beneficial for income tax revenues.
- International trade taxes (Table 3 columns (5) and (6)):
  - Lagged dependent variable: (5) 0.472   [4.40]; (6) 0.463   [3.73].
  - Using openness (lib_index1): lgdp = -0.050  [-3.23]; lexchange = -0.029  [-3.75]; inflation = -0.012  [-2.21]; agri = 0.127    [1.98]; lgc = 0.015    [1.87].
  - Using collected tariff (lib_index2): lgdp = -0.043  [-2.81]; lexchange = -0.034  [-3.24]; trade liberalization not significantly linked to trade revenues with either measure.
- Taxes on goods and services (Table 3 columns (7) and (8)):
  - Lagged dependent variable: (7) 0.765   [7.98]; (8) 0.727   [9.14].
  - Using openness: ind = 0.114   [2.31]; lgc = 0.008   [1.88]; ltot = 0.015   [1.88]; inflation = -0.006 [-2.14]; lexchange and trade liberalization not significant.
  - Using collected tariff: ind = 0.137   [2.85]; ltot = 0.018   [2.15]; inflation = -0.011 [-3.98]; neither exchange rate nor trade liberalization measures significant.
- Diagnostics (selected):
  - (m1) examples: -3.481  [0.00]; -3.442  [0.00]; -3.022  [0.00]; -2.774   [0.01]; -2.838   [0.01].
  - (m2) examples: 0.841  [0.40]; -0.858  [0.39]; -0.674  [0.50]; 0.051  [0.96]; -0.195  [0.85].
  - Sargan tests examples: Sargan = 63.28   [0.90]; 69.40   [0.77]; 100.70  [0.05]; 104.3   [0.03]; 64.23  [0.89]; 78.47  [0.50]; 69.02   [0.78]; 84.85   [0.31].

### Empirical interpretation and heterogeneity
- Results sensitive to measure of trade liberalization:
  - Using openness (trade share of GDP), trade liberalization measures not significant for total tax revenue; agri, ind, lgc, and ltot exert positive effects; inflation negative.
  - Using collected tariff, lexchange and inflation are negative and significant for total tax revenue; some specifications show lib_index2 negative (example: lib_index2 = -0.012 [-1.94] in select specifications).
- Regional heterogeneity: CFA vs non-CFA differences and specification choices materially affect estimates; prior literature finds varying responses across country groups.
- Evidence of negative link of lexchange (real exchange rate appreciation) to overall and trade tax revenues in several specifications: examples include lexchange = -0.029  [-3.75] and lexchange = -0.034  [-3.24] in trade tax regressions.

### Policy implications and recommendations
- Trade liberalization accompanied by appropriate monetary and exchange rate policies may preserve tax yield.
- Countries hesitant to liberalize due to revenue loss concerns should consider complementary monetary and exchange rate policies to mitigate adverse effects.
- Careful selection and validation of trade liberalization proxies is necessary because empirical results are sensitive to the measure used.
- Specific findings supportive of policy:
  - Some evidence that trade liberalization has a positive effect on income tax revenue.
  - Generally, trade liberalization is not strongly linked to total tax revenue or its components in a uniform way across specifications.

*Source: IMF working paper chapter "1. Comparative Structure of Tax Revenue in Sub-Saharan Africa Countries, 1980–96" (_wp04178).*

### 1. Comparative Structure of Tax Revenue in Sub-Saharan Africa Countries, 1980–96 ......6

### 1. Comparative Structure of Tax Revenue in Sub-Saharan Africa Countries, 1980–96

### I. Introduction
- Trade liberalization has frequently been the centerpiece of an economic development strategy in Sub-Saharan Africa.
- Trade liberalization often entails a reduction and unification of tariffs and relaxation of quantitative barriers, and may be accompanied or supported by currency devaluation and domestic tax reform.
- On devising a program of liberalization, policymakers are often hindered in forecasting tax revenues because of the uncertainty regarding the effects of trade liberalization and exchange rate changes on fiscal outcomes.
- The relationship between trade liberalization, the exchange rate, and tax revenue is therefore an issue of great practical importance.
- This paper examines this relationship in Sub-Saharan Africa.

### Research questions probed in the paper
- What is the relationship between trade liberalization and tax revenues? Does increased trade liberalization lead to a reduction in tax revenues through its effect on taxes from international trade or other taxes, controlling for accompanying macroeconomic changes?
- Is the relationship sensitive to the index of liberalization adopted? Is the relationship sensitive to the econometric specification adopted?
- What is the relationship between exchange rate changes and tax revenues? Does devaluation or currency depreciation increase or decrease tax revenue?
- Are there any differences between the CFA franc (the currency used by a group of countries in West and Central Africa) and non-CFA franc countries in the revenue response of different types of taxes to trade liberalization changes?

### Empirical and conceptual context
- The paper draws on two strands of work: one examining the relationship between trade liberalization and tax revenue and the other examining the relationship between exchange rate changes and inflation and tax revenue (or fiscal outcomes, more generally).
- Because trade liberalization is often accompanied by currency devaluation (and higher inflation), a thorough empirical investigation should consider the simultaneous relationship between trade liberalization and changes in macroeconomic variables and revenues.

*Source: IMF working paper chapter "1. Comparative Structure of Tax Revenue in Sub-Saharan Africa Countries, 1980–96".*

### Section II outlines some theoretical considerations and reviews previous empirical work in

### _wp04178 - Section II outlines some theoretical considerations and reviews previous empirical work in

### Theoretical considerations — overview
- Trade liberalization links to tax revenue mainly through international trade tax revenue; the precise relationship depends on the nature of liberalization and the response of imports and exports.
- Replacement of quantitative barriers with import duties can raise trade tax revenue depending on duty levels and import value responses; ultimate tariff reductions are likely to reduce international trade tax revenue.
- Trade liberalization effects on overall and domestic tax revenue depend on tax system structure and administrative capabilities; trade liberalization is often accompanied by introduction of a VAT or other domestic tax policy changes.
- Macroeconomic variables (including inflation and exchange rates) influence tax revenue; Tanzi (1989) posits an inverse relationship between tax revenue and the real level of the official exchange rate, with overvaluation suppressing import and export bases in domestic currency terms and reducing collections of international trade taxes and sales/excise taxes.
- Components of tax revenue interact and may reinforce or offset changes in one another; country-specific structures matter for aggregate outcomes.

### A. International trade taxes — mechanisms and elasticities
- Typical tax forms:
  - Import duties: usually ad valorem on import value (sometimes specific/unit basis).
  - Export taxes: usually ad valorem but sometimes specific or complex (notably export levies).
- Key determinants of revenue effects from tariff changes:
  - Change in tariff level and coverage.
  - Price elasticity of demand for imports.
  - Elasticity of supply of import substitutes.
  - Starting point for tariff changes (tariffs above revenue-maximizing levels can be reduced with revenue gains).
- Exchange rate effects:
  - Real depreciation increases the trade tax base in domestic currency terms for given import/export volumes (valuation effect).
  - If import demand is inelastic short-term, valuation effect may dominate → higher import tax revenues.
  - Real depreciation tends to increase exports (valuation and volume effects align) but export taxes are typically insignificant; short-term import effects dominate.
- Heterogeneity across goods: highest tariffs often on goods with more elastic demand; disproportionate tariff cuts on these goods may induce import expansion and offset revenue losses.

### B. Taxes on goods and services — interaction with trade liberalization and exchange rates
- Many developing countries derive a large share of goods and services tax collections from imports (in some cases one-half or more).
- Tax sequencing by convention: tariffs apply to import value; excises levied on base inclusive of tariffs; VAT levied on base inclusive of tariffs and excises.
- Trade liberalization that lowers tariffs reduces the tax base for excises and VAT at import stage; offsetting effects:
  - Import value may rise and partially or fully offset base reduction from tariff cuts.
  - Output decline of import-substituting domestic goods can reduce domestic tax bases.
  - Administrative efficiency for goods and services taxes is typically lower than for import duties, increasing uncertainty.
- Exchange rate depreciation:
  - Raises excise and VAT collections from imports via pass-through to consumer prices, but demand responses may offset increases depending on elasticities.
  - If excises are specific (per unit) rather than ad valorem, depreciation will not automatically increase excise revenues unless excises are adjusted.

### C. Taxes on income, profits, and capital gains — channels and inflation/FX effects
- Short-run effects of trade liberalization on income/profit taxes mainly via changes in profitability of import substitutes and import-competing firms; long-run effects via growth and expanded tax base.
- Exchange rate changes have limited direct effects on personal and corporate income tax collections except via revaluation of foreign-denominated assets and liabilities; inflation and resultant bracket effects matter materially.
- Personal income tax:
  - Inflation can raise nominal tax liabilities via bracket creep if brackets are not indexed.
  - Real exchange rate depreciation likely lowers real wages; if brackets are indexed, personal income tax collections may fall; if not indexed, bracket creep may still raise collections.
  - Capital income and capital gains are often weakly captured; shifts toward capital income due to depreciation unlikely to materially bolster personal income tax revenues in many developing countries.
- Corporate income tax:
  - Inflation erodes historical-cost-based depreciation allowances and may overstate profits under FIFO inventories, raising tax liabilities; conversely, rising nominal interest rates can lead to deductible interest components that understate true profits, lowering liabilities.
  - Real depreciation raises cost of imported inputs, reducing profitability for firms that rely on imports; exporters may benefit.
  - Treatment of foreign exchange gains/losses and timing of revaluation across tax systems produces no uniform prediction.

### Econometric approaches and literature review
- Cross-sectional and panel approaches relate tax revenue share to level of development, economic structure, governance, trade liberalization indices, and macro variables.
- Findings in prior studies:
  - Income level, agriculture share, and share of international trade in GDP often significant determinants of revenue ratios.
  - Persistence in tax shares suggests large country-specific components.
- Selected previous empirical results:
  - Ebrill et al (1999): panel of 27 and 105 countries (1980–1992, 1980–1995); using fixed-effects and instrumental regressions they conclude tariff reforms do not necessarily lower trade tax revenue and find depreciations significantly linked to higher trade tax revenues.
  - Ghura (1998): contrasted by finding no significant relation between exchange rate and trade tax revenue.
  - Khattry and Mohan Rao (2002): panel of 80 countries (1970–1998); fixed-effects regressions find trade liberalization negatively correlated with total tax revenue and international trade tax revenue, no significant link between exchange rate and international trade tax revenue, and countries generally below revenue-maximizing tariff—implying tariff reductions would reduce international trade tax revenues.
  - Adam, Bevan, and Chambas (2001): difference GMM dynamic panel for Sub-Saharan Africa; openness raises overall tax revenue in CFA franc countries but little effect in non-CFA countries; depreciation and removal of real exchange rate disequilibrium lower tax yield in CFA countries while the opposite occurs in non-CFA countries; effects vary by tax component.

### Data and empirical methodology (as used in this study)
- Two proxies for trade liberalization:
  - Trade liberalization 1: international trade as a share of GDP (higher ratio → greater liberalization).
  - Trade liberalization 2: collected tariff measured by ratio of import duties to value of imports (decline → greater liberalization).
- Data set: augmented version of Adam et al (2001) with collected tariff and real effective exchange rates; sample period 1980-1996.
- Dependent variables (all as share of GDP): total tax revenue, taxes on income and profits, taxes on goods and services, and international trade taxes.
- Control variables included (many in log form where indicated):
  - GDP per capita (GDP)
  - Share of agriculture in GDP (AGRI)
  - Share of industry (including mining) in GDP (IND)
  - Real government consumption share in GDP (GC)
  - Net transfers of aid (NAID)
  - Terms of trade (TOT)
  - Real effective exchange rate (EXCHANGE) — index relative to 1995 (increase represents appreciation)
  - Inflation (INFLATION)
  - Openness / trade liberalization index (OPEN)
  - Dummy for CFA franc countries (CFA)
- Model form: first-order dynamic panel with lagged dependent variable; country fixed effects and time effects; potential endogeneity of per capita income, inflation, government consumption, and trade liberalization addressed via GMM instrumentation.
- Estimation method: GMM estimator using orthogonal deviations transformation (Arellano and Bover (1995) approach) with internal instruments (lagged levels and lagged explanatory variables); tests reported include 1st- and 2nd-order serial correlation and the Sargan test of over-identifying restrictions.
- Instrument strategy: treat per capita income, inflation, government consumption, and openness as endogenous; use their values dated at time t-2 as instruments.

### Key empirical patterns and diagnostics from simple plots (1980–1996)
- Simple year-country average plots:
  - Trade liberalization measure 1 (openness): no clear pattern with major tax components.
  - Trade liberalization measure 2 (collected tariff): positive correlation between higher collected tariff (less liberal environment) and overall tax revenues and taxes on international trade.
  - Real effective exchange rate: for overall tax revenues and taxes on international trade, increases in the exchange rate (appreciation) appear positively related to higher revenues.
  - Regional differences: CFA countries often show a positive correlation between trade liberalization measures and revenues; non-CFA countries show weaker or negative correlations. The 1994 CFA devaluation produces a sharp break visible in plots.

### Selected empirical results (GMM regression highlights, 1980–1996 full sample)
- General finding: lagged dependent variable is positive and significant — evidence of partial adjustment/ persistence in tax revenue shares.
- For total tax revenue (Table 2 columns (1) and (2)):
  - lag_dv = 0.563    [5.02] (column (1))
  - lag_dv = 0.538    [5.79] (column (2))
  - Using trade liberalization 1 (openness) specification (column (1)):
    - lgdp = -0.007  [-0.33]
    - agri = 0.137    [2.58]
    - ind = 0.214    [3.21]
    - lgc = 0.019    [2.28]
    - lnaid = 0.036    [1.21]
    - ltot = 0.029    [2.19]
    - lexchange = -0.011  [-0.62]
    - inflation = -0.023  [-4.11]
    - lib_index1 = 0.001    [0.11]
    - lib_index1cfa = 0.042    [1.38]
  - Using trade liberalization 2 (collected tariff) specification (column (2)):
    - lgdp = -0.002  [-0.12]
    - agri = 0.036    [0.42]
    - ind = 0.133    [1.66]
    - lgc = 0.012    [1.01]
    - lnaid = 0.017    [0.52]
    - ltot = 0.003    [0.19]
    - lexchange = -0.030  [-2.07]
    - inflation = -0.020  [-4.88]
    - lib_index2 = -0.010   [-0.92]
    - lib_index2cfa = 0.026    [0.92]
- For income taxes as a share of GDP (Table 2 columns (3) and (4)):
  - lag_dv = 0.738   [5.59] (column (3))
  - lag_dv = 0.632   [5.35] (column (4))
  - Selected coefficients (column (4)):
    - lgdp = 0.027   [1.96]
    - agri = -0.054  [-2.05]
    - ind = 0.054   [1.29]
    - lexchange = 0.007   [0.76]
    - inflation = -0.003  [-0.96]
- Diagnostic test statistics (selected):
  - (m1) first-order serial correlation statistic examples:
    - (m1) = -3.481  [0.00] (Table 2 column (2))
  - (m2) second-order serial correlation statistic examples:
    - (m2) = 0.841  [0.40] (Table 2 column (2))
  - Sargan test of over-identifying restrictions examples:
    - Sargan = 63.28   [0.90] (Table 2 column (1))
    - Sargan = 69.40   [0.77] (Table 2 column (2))
    - Sargan = 100.70  [0.05] (Table 2 column (3))
    - Sargan = 104.3   [0.03] (Table 2 column (4))

### Empirical interpretation and contrasts
- Results vary with the measure of trade liberalization:
  - Using openness (trade as share of GDP), trade liberalization measures are not significant for total tax revenue; agricultural share, industrial share, government consumption, and terms of trade exert positive effects; inflation is negative.
  - Using collected tariff measure, real exchange rate and inflation are negative and significant for total tax revenue (suggesting real exchange rate appreciation and higher inflation depress revenues), and lib_index2 shows negative coefficient sign (lib_index2 = -0.012 [-1.94] in some specifications).
- Regional heterogeneity (CFA vs non-CFA) and specification choices critically affect estimated relationships — consistent with literature that finds different responses across country groups and sensitivities to measurement and econometric approach.

*Source: _wp04178 - Section II of the provided IMF PDF content (excerpted text and reported regression table values).*

### 1. Year dummies are included in all specifications.

### _wp04178 - 1. Year dummies are included in all specifications.

### Regression results — income taxes (Table 2; columns (3) and (4))
- Lagged dependent variable: positive and significant.
- Using the first measure of trade liberalization:
  - Agricultural share is negatively linked to income tax revenues.
  - No other variables are significant.
- Using the second measure of trade liberalization:
  - Per capita income is positively linked to income tax revenues.
  - Agricultural share is negatively linked to income tax revenues.
  - Exchange rate and inflation variables are not significant.
  - The coefficient on the second measure of trade liberalization is negative and significant, which suggests that a higher value of the collected tariff (which is interpreted as less liberalization) is linked to lower revenues, so trade liberalization appears beneficial for income tax revenues.

### Regression results — international trade taxes (Table 3; columns (5) and (6))
- Lagged dependent variable:
  - (5) 0.472   [4.40]
  - (6) 0.463   [3.73]
- For the first measure of trade liberalization:
  - Lgdp is negatively linked: -0.050  [-3.23]
  - Lexchange (real exchange rate) is negatively linked: -0.029  [-3.75]
  - Inflation is negatively linked: -0.012  [-2.21]
  - Agri is positively linked: 0.127    [1.98]
  - Lgc (government consumption) is positively linked: 0.015    [1.87]
- For the second measure of trade liberalization:
  - Lgdp is negatively linked: -0.043  [-2.81]
  - Lexchange is negatively linked: -0.034  [-3.24]
  - Trade liberalization is not significantly linked to trade revenues using either measure.

### Regression results — taxes on goods and services (Table 3; columns (7) and (8))
- Lagged dependent variable:
  - (7) 0.765   [7.98]
  - (8) 0.727   [9.14]
- For the first measure of trade liberalization:
  - Ind (industrial share) positive: 0.114   [2.31]
  - Lgc positive: 0.008   [1.88]
  - Ltot (terms of trade) positive: 0.015   [1.88]
  - Inflation negative: -0.006 [-2.14]
  - Lexchange and trade liberalization variables are not significant.
- For the second measure of trade liberalization:
  - Ind positive: 0.137   [2.85]
  - Ltot positive: 0.018   [2.15]
  - Inflation negative: -0.011 [-3.98]
  - Neither the exchange rate nor trade liberalization measures are significant.

### Overall findings and interpretation
- Strong persistence over time in total tax revenues and all components of revenues, evidenced by positive and significant coefficients on the lagged dependent variable across specifications.
- Some evidence that trade liberalization has a positive effect on income tax revenue, but otherwise trade liberalization is not strongly linked to total tax revenue or its components.
- Results are not uniform across specifications, and no significant difference between CFA and non-CFA countries is found.
- Sensitivity of results to the measure of trade liberalization indicates the need for careful consideration of the best proxy for liberalization.
- Some evidence of a negative link of real exchange rate appreciation to:
  - Overall tax revenues
  - Trade tax revenues
  - Specific estimates: lexchange coefficients include -0.029  [-3.75], -0.034  [-3.24], 0.009   [0.66], 0.011   [1.33] across columns (5)–(8).
- For the second measure of liberalization, with the exception of income taxes and trade taxes, inflation is negatively linked to revenues.
- Policy implication: trade liberalization accompanied by an appropriate monetary and exchange rate policy does not have a significant adverse effect on overall tax revenue though it may affect income tax revenue.

### Selected GMM estimation diagnostics and statistics (Table 3)
- Coefficients and robust t-ratios (selected):
  - lgdp: -0.050  [-3.23]; -0.043  [-2.81]; 0.004   [0.18]; 0.007   [0.34]
  - agri: 0.127    [1.98]; 0.057    [0.83]; 0.001   [0.01]; 0.045   [0.88]
  - ind: 0.045    [0.80]; 0.059    [0.80]; 0.114   [2.31]; 0.137   [2.85]
  - lnaid: -0.004  [-0.12]; 0.013  [0.32]; 0.002   [0.11]; 0.005   [0.26]
  - lib_index1: 0.002     [0.23]; -0.004 [-0.59]
  - lib_index2: 0.010     [1.12]; -0.010 [-1.25]
- Serial correlation tests in first-differenced residuals:
  - (m1) -3.442  [0.00]; -3.022  [0.00]; -2.774   [0.01]; -2.838   [0.01]
  - (m2) -0.858  [0.39]; -0.674  [0.50]; 0.051  [0.96]; -0.195  [0.85]
- Sargan test of over-identifying restrictions:
  - Sargan 64.23  [0.89]; 78.47  [0.50]; 69.02   [0.78]; 84.85   [0.31]

### Policy recommendations and implications (from conclusions and interpretation)
- Trade liberalization, when accompanied by appropriately supportive monetary and exchange rate policies, may preserve tax yield.
- Countries reluctant to undertake trade liberalization for fear of revenue loss should consider complementary monetary and exchange rate policies to mitigate adverse revenue effects.
- Careful selection and validation of trade liberalization measures (proxies) is necessary due to sensitivity of empirical results to the measure used.

### Data definitions (Appendix)
- Lag_dv: lag of dependent variable.
- Lgdp: (log) real per capita income
- Agri: agricultural share in GDP
- Ind: industry (including mining) share in GDP
- Lgc: (log) government consumption as a share of GDP
- Lib_index1: (log) openness: international trade as a share of GDP
- Lib_index2: (log) openness: ratio of import duties to value of imports
- Lnaid: (log) net transfers of aid.
- Inflation: annual inflation
- Ltot: (log) terms of trade.
- Lexchange: (log) real effective exchange rate, 1995=100. Increase indicates appreciation.

### Country classification (Appendix)
- Non-CFA: Burundi; The Gambia; Ghana; Kenya; Madagascar; Mauritania; Malawi; Mauritius; Rwanda; Sierra Leone; Tanzania; Uganda; Zambia; and Zimbabwe.
- CFA: Benin; Burkina Faso; Central African Republic; Cote d’Ivoire; Mali; Niger; Senegal; and Togo.

*Source: _wp04178 - 1. Year dummies are included in all specifications.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2004/_wp04178.pdf_
