## _wp0740 - Appendix Tables

## Source details

**Canonical URL:** [_wp0740 - Appendix Tables](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp0740.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp0740.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp0740.pdf.json)

---

### Introduction and scope
- Focus: static effects of COMESA-related preferential trade liberalization on Uganda’s imports using commodity-level data for 1994 and 2000–2003.
- Objective: estimate impact of preferential tariff reductions on Uganda’s imports and test whether changes reflect trade creation or trade diversion.
- Data sources: COMTRADE (UN), TRAINS (UNCTAD), accessed via WITS (World Bank); commodity-level import and tariff data at 6-digit HS level; import excise taxes included in tariff rates.
- Sample years: 1994, 2000, 2001, 2002, and 2003.

### Methodology
- Theoretical framework: Armington with CES demand across origin varieties (varieties differentiated by origin).
- Empirical strategy:
  - Difference-in-difference and triple-difference estimators exploiting variation across commodities, origin countries, and time.
  - Preferred specification: regress log ratio of imports from COMESA to non-COMESA countries on log preference margin (log(1 + preferential tariff) − log(1 + MFN tariff)) with commodity-by-country fixed effects and year dummies (equation (9) in text).
- Robustness checks:
  - Adding zero-to-positive export transitions (assigning US$1).
  - Estimating elasticities by one-digit HS sectors.
  - Triple-difference using South Africa as external comparator.
  - Accounting for import excise taxes and potential tax evasion effects.

### Key empirical findings
- Elasticity estimates:
  - Elasticity of imports with respect to tariff rates: between 14 percent and 16 percent.
    - Example interpretation: a decrease of the ad valorem tariff rate by 10 percentage points implies an increase in COMESA imports of 14 percent (Table 6, column (4) interpretation).
    - If the ad valorem tariff rate decreases by 100 percent (e.g., by 10 percentage points when tariff = 10 percent), imports from COMESA countries increase by 16 percent (Table 5, column (4) interpretation).
  - Elasticity of substitution between varieties from different origin countries: approximately 1.7.
- Difference-in-difference preferred estimate:
  - Coefficient on log preference margin ≈ –1.9 (significant at the 10 percent level).
- Sectoral heterogeneity:
  - Significant elasticities predominantly in agricultural and beverage sectors (HS1, HS2, HS3); many sectors show insignificant effects.
- Trade diversion:
  - No evidence that Uganda’s imports from non-COMESA countries declined after COMESA implementation; trade effects appear associated with trade creation rather than trade diversion.
- Magnitude relative to other studies:
  - Effects for Uganda are small compared with Clausing (2001) for CUSFTA (example cited: 10 percentage point tariff decrease → 96 percent increase in U.S. imports from Canada).
  - Romalis (2005) estimates for Mexico (0.6 to 2.5) are closer to the Uganda estimates.

### Robustness and controls
- Fixed effects and dummies:
  - Commodity-by-country fixed effects control for time-invariant bilateral factors.
  - Year dummies control for common time shocks.
- Controlling for MFN liberalization:
  - Difference-in-difference estimator helps net out simultaneous MFN changes.
- Tariff measurement:
  - Tariff rates adjusted for import excise taxes.
  - Preferential tariff rates set equal to MFN when no preferential rate specified.
- Tax evasion considerations:
  - Noted that recorded imports may rise when tariffs fall due to reduced evasion.
- Triple-difference:
  - Using South Africa to net out relative border-price changes; produced imprecise elasticity estimates due to smaller sample.

### Trade policy and welfare implications
- COMESA’s preferential tariff reductions produced small positive effects on Uganda’s imports from member countries; effects concentrated in sectors where developing countries have comparative advantage.
- Lack of trade diversion implies gains from COMESA’s preferential liberalization are more likely trade-creating (welfare-improving for a small-open economy) than trade-diverting.
- Small absolute gains raise questions about the economic payoff from South-South PTAs for small low-income members and motivate consideration of noneconomic objectives (e.g., regional peace and security) and the efficient allocation of scarce administrative resources.

### Selected exact statistics and data points
- Sample years: 1994, 2000, 2001, 2002, and 2003.
- Preferential tariff rate for COMESA countries (average across tariff lines): 1994 = 11.3 percentage points; 2003 = 5.5 percentage points.
- MFN tariff rates (average across tariff lines): 1994 = 17.9 percentage points; 2003 = 10.2 percentage points.
- Imports from COMESA countries (per 6-digit HS commodity, thousand $): 1994 mean = 155.2; 2003 mean = 289.3.
- Tariff revenue in Uganda: declined significantly (by 8 percent of GDP) after the inception of COMESA.
- Estimated elasticity of substitution (σ): ≈ 1.7.
- Estimated elasticity of imports with respect to tariff rates: between 14 percent and 16 percent.
- Difference-in-difference preferred coefficient on log preference margin: –1.9 (significant at 10 percent).

### Data and tariff-rate measurement details
- Tariff rates adjusted for import excise taxes.
- Preferential tariff rate for COMESA countries in 1994 uses both PTA and MFN rates.
- Imports refer to a single 6-digit HS commodity.
- Sample restriction: commodities with data on imports from non-COMESA countries available for both 1994 and 2003; subset restricts product codes to commodities that in at least one year were imported from COMESA.
- Commodity dummies set at 6-digit HS product-code level; commodity-country dummies for pairwise combinations.

### Summary statistics (selected)
- MFN Tariff Rate 1994 (percentage points): Obs 1,020; Mean 17.9; Std. Dev. 8.6; Min 0.0; Max 60.0.
- MFN Tariff Rate 2003 (percentage points): Obs 1,020; Mean 10.2; Std. Dev. 10.2; Min 0.0; Max 145.0.
- Imports from non-COMESA (1994, thousand $): Obs 1,020; Mean 364.6; Std. Dev. 1,289.2; Min 0.5; Max 18,223.9.
- Imports from non-COMESA (2003, thousand $): Obs 1,020; Mean 718.4; Std. Dev. 2,381.7; Min 0.5; Max 30,602.1.

### Pattern of protection (frequency distributions, selected)
- Under PTA (1994) — Total: 5,018
  - Free Trade: 334; 6.66 percent; Cum. 6.66
  - Tariffs under 5 percent: 1,172; 23.36 percent; Cum. 30.01
  - Tariffs between 5 and 10 percent: 1,558; 31.05 percent; Cum. 61.06
  - Tariffs between 10 and 25 percent: 1,367; 27.24 percent; Cum. 88.30
  - Tariffs over 25 percent: 587; 11.70 percent; Cum. 100.00
- Under COMESA (2003) — Total: 4,880
  - Free Trade: 986; 20.20 percent; Cum. 20.20
  - Tariffs under 5 percent: 2,035; 41.70 percent; Cum. 61.91
  - Tariffs between 5 and 10 percent: 1,521; 31.17 percent; Cum. 93.07
  - Tariffs between 10 and 25 percent: 305; 6.25 percent; Cum. 99.32
  - Tariffs over 25 percent: 33; 0.68 percent; Cum. 100.00
- MFN tariff rates distribution (1994) — Total: 5,018
  - Free Trade: 170; 3.39 percent; Cum. 3.39
  - Tariffs under 5 percent: 11; 0.22 percent; Cum. 3.61
  - Tariffs between 5 and 10 percent: 2,268; 45.20 percent; Cum. 48.80
  - Tariffs between 10 and 25 percent: 1,298; 25.87 percent; Cum. 74.67
  - Tariffs over 25 percent: 1,271; 25.33 percent; Cum. 100.00
- MFN tariff rates distribution (2003) — Total: 4,880
  - Free Trade: 986; 20.20 percent; Cum. 20.20
  - Tariffs under 5 percent: 12; 0.25 percent; Cum. 20.45
  - Tariffs between 5 and 10 percent: 2,021; 41.41 percent; Cum. 61.86
  - Tariffs between 10 and 25 percent: 1,813; 37.15 percent; Cum. 99.02
  - Tariffs over 25 percent: 48; 0.98 percent; Cum. 100.00

### Regression estimates (selected reported results)
- Table 5 (Dependent variable: Log of imports from COMESA)
  - Estimators: OLS, OLS, OLS, FE, FE.
  - Observations: 10,341 in all columns.
  - R-squared: 0.00, 0.00, 0.60, 0.76, 0.81.
  - Elasticity of substitution reported examples:
    - Column (3): 0.6954** 
    - Column (4): 0.8017* 
    - Column (5): 1.1924+ 
  - Significance notation: + significant at 10%; * significant at 5%; ** significant at 1%.
- Table 6 (Dependent variable: Log of imports; coefficients in levels)
  - Observations: 10,341; R-squared: same pattern as Table 5.
  - Selected coefficients:
    - Preferential tariff: 0.0069** and 0.0083+ in some specifications.
    - Preference margin: -0.0193 and 0.0096* in some specifications.
- Table 7 (Robustness: diff-in-diff and triple-diff)
  - Estimator: FE with variations.
  - Observations vary across columns: 21,500; 470; 1,090; 1,357; 317; 317.
  - R-squared examples: 0.57; 0.74; 0.77; 0.72; 0.99; 0.97.
  - Elasticity of substitution examples across columns: -0.3185; -12.0281; -8.0169; -4.4884; 1.1562; 1.1270.
- Table 8 (Estimates of Trade Diversion based on Uganda imports from non-COMESA; FE)
  - Observations: 62,302 in reported columns.
  - R-squared: 0.78 in reported columns.
  - Selected coefficients:
    - Log of MFN tariff: -0.8052; -0.7438; -0.6540; -0.6431 (with adjacent significance markers in source).
    - Log of preferential tariff: -0.1306; -0.1828; -0.3144; -0.0378.
    - Interactors: Log of COMESA imports 0.1961; 0.1750; 0.0390; 0.5702. COMESA share in imports -1.1308; 2.6690.

### Appendix I — Selected preferential tariff rates by 2-digit 1996 HS Codes (1994 vs. 2003; percentage-point change)
- HS 01 LIVE ANIMALS: 1994 = 4.67; 2003 = 4.00; % Point Change = -0.67
- HS 02 MEAT AND EDIBLE MEAT OFFAL: 1994 = 21.12; 2003 = 16.00; % Point Change = -5.12
- HS 03 FISH, CRUSTACEANS & AQUATIC INVERTEBRATES: 1994 = 15.34; 2003 = 6.00; % Point Change = -9.34
- HS 07 EDIBLE VEGETABLES & CERTAIN ROOTS & TUBERS: 1994 = 23.21; 2003 = 6.00; % Point Change = -17.21
- HS 08 EDIBLE FRUIT & NUTS; CITRUS FRUIT OR MELON PEEL: 1994 = 24.08; 2003 = 6.00; % Point Change = -18.08
- HS 09 COFFEE, TEA, MATE & SPICES: 1994 = 20.52; 2003 = 4.19; % Point Change = -16.33
- HS 10 CEREALS: 1994 = 11.22; 2003 = 5.13; % Point Change = -6.09
- HS 22 BEVERAGES, SPIRITS AND VINEGAR: 1994 = 55.90; 2003 = 62.67; % Point Change = 6.76
- HS 24 TOBACCO AND MANUFACTURED TOBACCO SUBSTITUTES: 1994 = 55.00; 2003 = 96.00; % Point Change = 41.00
- HS 30 PHARMACEUTICAL PRODUCTS: 1994 = 0.00; 2003 = 0.00; % Point Change = 0.00
- HS 31 FERTILIZERS: 1994 = 0.00; 2003 = 0.00; % Point Change = 0.00
- Additional selected entries (second block):
  - HS 11 MILLING PRODUCTS; MALT; STARCH; INULIN; WHT GLUTEN: 1994 = 28.82; 2003 = 11.86; % Point Change = -16.96
  - HS 41 RAW HIDES AND SKINS (NO FURSKINS) AND LEATHER: 1994 = 30.00; 2003 = 7.00; % Point Change = -23.00
  - HS 44 WOOD AND ARTICLES OF WOOD; WOOD CHARCOAL: 1994 = 28.38; 2003 = 17.29; % Point Change = -11.10
  - HS 45 CORK AND ARTICLES OF CORK: 1994 = 12.86; 2003 = 15.00; % Point Change = 2.14

### Appendix II — Selected MFN tariff rates by 2-digit 1996 HS Codes (1994 vs. 2003; percentage-point change)
- HS 46 MFR OF STRAW, ESPARTO ETC.; BASKETWARE & WICKERWRK: 1994 = 28.33; 2003 = 15.00; Change = -13.33
- HS 47 WOOD PULP ETC; RECOVD (WASTE & SCRAP) PPR & PPRBD: 1994 = 12.11; 2003 = 5.89; Change = -6.21
- HS 52 COTTON, INCLUDING YARN AND WOVEN FABRIC THEREOF: 1994 = 21.34; 2003 = 14.21; Change = -7.13
- HS 56 WADDING, FELT ETC; SP YARN; TWINE, ROPES ETC.: 1994 = 24.81; 2003 = 8.27; Change = -16.55
- HS 57 CARPETS AND OTHER TEXTILE FLOOR COVERINGS: 1994 = 30.00; 2003 = 15.00; Change = -15.00
- HS 60 KNITTED OR CROCHETED FABRICS: 1994 = 30.00; 2003 = 14.53; Change = -15.47
- HS 64 FOOTWEAR, GAITERS ETC. AND PARTS THEREOF: 1994 = 30.00; 2003 = 14.45; Change = -15.55
- HS 71 NAT ETC PEARLS, PREC ETC STONES, PR MET ETC; COIN: 1994 = 21.47; 2003 = 14.42; Change = -7.05
- HS 84 NUCLEAR REACTORS, BOILERS, MACHINERY ETC.; PARTS: 1994 = 11.31; 2003 = 0.77; Change = -10.54
- HS 85 ELECTRIC MACHINERY ETC; SOUND EQUIP; TV EQUIP; PTS: 1994 = 18.58; 2003 = 6.67; Change = -11.91
- HS 92 MUSICAL INSTRUMENTS; PARTS AND ACCESSORIES THEREOF: 1994 = 30.00; 2003 = 7.00; Change = -23.00
- HS 96 MISCELLANEOUS MANUFACTURED ARTICLES: 1994 = 27.25; 2003 = 11.08; Change = -16.17
- Categories with MFN increases (examples):
  - HS 50 SILK: 1994 = 13.00; 2003 = 15.00; Change = 2.00
  - HS 51 WOOL & ANIMAL HAIR: 1994 = 13.06; 2003 = 14.78; Change = 1.72
  - HS 53 VEG TEXT FIB NESOI; VEG FIB & PAPER YNS & WOV FAB: 1994 = 12.26; 2003 = 14.45; Change = 2.19
  - HS 54 MANMADE FILAMENTS: 1994 = 14.77; 2003 = 16.08; Change = 1.32
  - HS 97 WORKS OF ART, COLLECTORS' PIECES AND ANTIQUES: 1994 = 10.00; 2003 = 15.00; Change = 5.00

### Appendix III — Uganda's imports by country of origin (selected rows)
- Burundi: Imports 1994 = 176.68; Imports 2003 = 5,325.44; Rank = 107
- Dem.Rp.Congo: Imports 1994 = 1,140.56; Imports 2003 = 3,929; Rank = 58 (source shows "3,929 8.98 58" formatting fragment)
- Djibouti: Imports 1994 = 2,787.93; Imports 2003 = 260.00; Rank = 134
- Egypt: Imports 1994 = 951.83; Imports 2003 = 406,413.00; Rank = 30
- Kenya: Imports 1994 = 186,784.85; Imports 2003 = 1,357,194.21; Rank = 1
- Mauritius: Imports 1994 = 1,175.95; Imports 2003 = 362,479.16; Rank = 37
- Rwanda: Imports 1994 = 74.46; Imports 2003 = 6,6534.14; Rank = 53
- Zambia: Imports 1994 = 1,831.85; Imports 2003 = 3,1210.10; Rank = 63
- Zimbabwe: Imports 1994 = 2,217.92; Imports 2003 = 2,8848.67; Rank = 49
- Note: Appendix III contains an extensive country-level listing with imports for 1994 and 2003 at the commodity level.

*Source: _wp0740 - Appendix Tables*

### Appendix Tables

### _wp0740 - Appendix Tables

### Appendix Contents
- I. Uganda’s Preferential Tariff Rates by 2-digit 1996 HS Codes, 1994 and 2003 ..................25
- II. Uganda’s MFN Tariff Rates by 2-digit 1996 HS Codes, 1994 and 2003 ...........................27
- III. Uganda’s Imports by Country of Origin, 1994 and 2003 ..................................................29
- References................................................................................................................................33

### Figures
- 1. Uganda: Imports and Tariff Revenue (percent of GDP),1986–2003...................................18
- 2. Uganda: Imports from Developing Countries (percent of GDP),1986–2003......................18

### Tables
- 1. Summary Statistics for Uganda vis-à-vis COMESA Countries (1994–2003).....................19
- 2. Summary Statistics for Uganda vis-à-vis non-COMESA Countries (1994–2003) .............19
- 3. Pattern of Protection in Uganda in 1994 and 2003 Under the Preferential Trade Agreements (PTA in 1994 and COMESA in 2003) ................................................................20
- 4. Pattern of Protection in Uganda in 1994 and 2003 vis-à-vis the Rest of the World (MFN Tariff Rates) .............................................................................................................................20
- 5. Estimates Based on Uganda Imports in 1994, 2000, 2001, 2002, and 2003 .......................21
- 6. Estimates Based on Uganda Imports in 1994, 2000, 2001, 2002, and 2003 .......................22
- 7. Robustness Checks...............................................................................................................23

*Source: _wp0740 - Appendix Tables*

### 8. Estimates of Trade Diversion Based on Uganda Imports from Non-COMESA Countries in

### 8. Estimates of Trade Diversion Based on Uganda Imports from Non-COMESA Countries in 1994, 2000, 2001, 2002, and 2003

### Introduction and scope
- Focus: static effects of COMESA-related preferential trade liberalization on Uganda’s imports using commodity-level data for 1994 and 2000–2003.
- Objective: estimate impact of preferential tariff reductions on Uganda’s imports and test whether changes reflect trade creation or trade diversion.
- Data sources: COMTRADE (UN), TRAINS (UNCTAD), accessed via WITS (World Bank); commodity-level import and tariff data at 6-digit HS level; import excise taxes included in tariff rates.

### Methodology
- Armington framework with CES demand across origin varieties; assumes varieties of a commodity differentiated by origin.
- Empirical strategy: difference-in-difference and triple-difference estimators exploiting variation across commodities, origin countries, and time.
- Preferred specification: regress log ratio of imports from COMESA to non-COMESA countries on log preference margin (log(1 + preferential tariff) − log(1 + MFN tariff)) with commodity-by-country fixed effects and year dummies (equation (9) in text).
- Robustness checks include: adding zero-to-positive export transitions (assigning US$1), estimating elasticities by one-digit HS sectors, triple-difference using South Africa as external comparator, and accounting for import excise taxes and potential tax evasion effects.

### Key empirical findings
- Elasticity of imports with respect to tariff rates: between 14 percent and 16 percent.
  - Example interpretation from estimates: a decrease of the ad valorem tariff rate by 10 percentage points implies an increase in COMESA imports of 14 percent (Table 6, column (4) interpretation).
  - If the ad valorem tariff rate decreases by 100 percent (e.g., by 10 percentage points when tariff = 10 percent), imports from COMESA countries increase by 16 percent (Table 5, column (4) interpretation).
- Elasticity of substitution between varieties from different origin countries: approximately 1.7.
- Difference-in-difference preferred estimate: coefficient on log preference margin ≈ –1.9 (significant at the 10 percent level) — consistent with the elasticities above.
- Heterogeneity: significant elasticities predominantly in agricultural and beverage sectors (HS1, HS2, HS3); many sectors show insignificant effects, indicating cross-sector heterogeneity.
- Trade diversion test: no evidence that Uganda’s imports from non-COMESA countries declined after COMESA implementation; trade effects appear associated with trade creation rather than trade diversion.
- Magnitude relative to other studies: effects for Uganda are small compared with Clausing (2001) for CUSFTA (e.g., 10 percentage point tariff decrease → 96 percent increase in U.S. imports from Canada) and Romalis (2005) for the United States and Canada. Romalis’s estimates for Mexico (0.6 to 2.5) are closer to the Uganda estimates.

### Robustness and controls
- Commodity-by-country fixed effects control for time-invariant bilateral factors; year dummies control for common time shocks.
- Accounting for MFN tariff liberalization via the difference-in-difference estimator helps net out simultaneous MFN changes.
- Import excise taxes included in tariff measures to partially account for nontariff barriers; tax evasion effects considered (noted that recorded imports may rise when tariffs fall due to reduced evasion).
- Triple-difference estimator (using South Africa) aimed to net out relative border-price changes; this specification yielded an insignificantly different elasticity of substitution (imprecise due to smaller sample).

### Trade policy and welfare implications
- COMESA’s preferential tariff reductions produced small positive effects on Uganda’s imports from member countries; effects concentrated in sectors where developing countries have comparative advantage.
- Lack of trade diversion implies gains from COMESA’s preferential liberalization are more likely trade-creating (welfare-improving for a small-open economy) than trade-diverting.
- Small absolute gains raise questions about the economic payoff from South-South PTAs for small low-income members and motivate consideration of noneconomic objectives (e.g., regional peace and security) and the efficient allocation of scarce administrative resources.

### Selected exact statistics and data points (as reported)
- Sample years: 1994, 2000, 2001, 2002, and 2003.
- Preferential tariff rate for COMESA countries (average across tariff lines): 1994 = 11.3 percentage points; 2003 = 5.5 percentage points.
- MFN tariff rates (average across tariff lines): 1994 = 17.9 percentage points; 2003 = 10.2 percentage points.
- Imports from COMESA countries (per 6-digit HS commodity, thousand $): 1994 mean = 155.2; 2003 mean = 289.3.
- Tariff revenue in Uganda: declined significantly (by 8 percent of GDP) after the inception of COMESA.
- Estimated elasticity of substitution (σ): ≈ 1.7.
- Estimated elasticity of imports with respect to tariff rates: between 14 percent and 16 percent.
- Difference-in-difference preferred coefficient on log preference margin: –1.9 (significant at 10 percent).
- Sectors with largest reductions in preferential tariff rates between 1994 and 2003 include: “edible fruit and nuts...,” “vegetable plaiting materials...,” “essential oils, etc.; perfumery, cosmetic...,” “prep feathers, down etc..,” and “musical instruments...”; sectors with large MFN reductions include “coffee, tea, mate and spices,” “vegetable plaiting materials,” and “raw hides and skins and leather.”

*Source: _wp0740 - 8. Estimates of Trade Diversion Based on Uganda Imports from Non-COMESA Countries in 1994, 2000, 2001, 2002, and 2003 (extract from IMF working paper).*

### 03. The tariff rates are

### 03. The tariff rates are

### Data and Methodology
- Tariff rates are adjusted for the existence of import excise taxes.
- The tariff rate for COMESA countries in 1994 is the average tariff rate faced by COMESA countries in 1994, which uses both PTA and MFN rates.
- Imports refer to a single 6-digit HS commodity.
- Preferential tariff rates are set equal to MFN tariff rates when no preferential rate is specified.
- Sample restriction: commodities for which data on imports from non-COMESA countries is available for both 1994 and 2003; subset restricts product codes to commodities that in at least one year were imported from COMESA.
- Commodity dummy variables are set at the 6-digit HS product-code level. Commodity-country dummy variables are for pairwise combinations of commodities and import-origin countries.
- Broad HS codes are defined in Appendix I.

### Summary Statistics (Uganda vis-à-vis COMESA and non-COMESA, 1994–2003)
- MFN Tariff Rate 1994 (percentage points): Obs 1,020; Mean 17.9; Std. Dev. 8.6; Min 0.0; Max 60.0.
- MFN Tariff Rate 2003 (percentage points): Obs 1,020; Mean 10.2; Std. Dev. 10.2; Min 0.0; Max 145.0.
- Imports from non-COMESA (1994, thousand $): Obs 1,020; Mean 364.6; Std. Dev. 1,289.2; Min 0.5; Max 18,223.9.
- Imports from non-COMESA (2003, thousand $): Obs 1,020; Mean 718.4; Std. Dev. 2,381.7; Min 0.5; Max 30,602.1.

### Pattern of Protection in Uganda (1994 and 2003)
- Under PTA (1994) — frequency distribution (Freq., Percent, Cum.):
  - Free Trade: 334; 6.66; 6.66
  - Tariffs under 5 percent: 1,172; 23.36; 30.01
  - Tariffs between 5 and 10 percent: 1,558; 31.05; 61.06
  - Tariffs between 10 and 25 percent: 1,367; 27.24; 88.30
  - Tariffs over 25 percent: 587; 11.70; 100.00
  - Total: 5,018; 100.00
- Under COMESA (2003) — frequency distribution (Freq., Percent, Cum.):
  - Free Trade: 986; 20.20; 20.20
  - Tariffs under 5 percent: 2,035; 41.70; 61.91
  - Tariffs between 5 and 10 percent: 1,521; 31.17; 93.07
  - Tariffs between 10 and 25 percent: 305; 6.25; 99.32
  - Tariffs over 25 percent: 33; 0.68; 100.00
  - Total: 4,880; 100.00
- MFN tariff rates distribution (1994):
  - Free Trade: 170; 3.39; 3.39
  - Tariffs under 5 percent: 11; 0.22; 3.61
  - Tariffs between 5 and 10 percent: 2,268; 45.20; 48.80
  - Tariffs between 10 and 25 percent: 1,298; 25.87; 74.67
  - Tariffs over 25 percent: 1,271; 25.33; 100.00
  - Total: 5,018; 100.00
- MFN tariff rates distribution (2003):
  - Free Trade: 986; 20.20; 20.20
  - Tariffs under 5 percent: 12; 0.25; 20.45
  - Tariffs between 5 and 10 percent: 2,021; 41.41; 61.86
  - Tariffs between 10 and 25 percent: 1,813; 37.15; 99.02
  - Tariffs over 25 percent: 48; 0.98; 100.00
  - Total: 4,880; 100.00

### Regression Estimates — Uganda Imports (1994, 2000, 2001, 2002, 2003)
- Table 5: Dependent variable: Log of imports (Uganda's imports from COMESA countries)
  - Columns (1)–(5) specification summary:
    - Estimators: OLS, OLS, OLS, FE, FE.
    - Observations: 10,341 in all columns.
    - R-squared: 0.00, 0.00, 0.60, 0.76, 0.81 respectively.
    - Elasticity of substitution estimates (selected reported coefficients with significance):
      - Column (3): 0.6954** (robust SE in parentheses in source)
      - Column (4): 0.8017* 
      - Column (5): 1.1924+ 
    - Notes: Robust standard errors; + significant at 10%; * significant at 5%; ** significant at 1%.
  - Variable definitions:
    - Log of preferential tariff = log(100 plus) Uganda's tariff rate (1994, 2000, 2001, 2002, 2003).
    - Log preference margin = log(100 plus) Uganda's preferential tariff rate for COMESA countries minus log(100 plus) Uganda's customs-duty rate for non-COMESA countries.
    - Year dummy variables, commodity dummies, commodity-country dummies applied depending on specification.
- Table 6: Same sample with dependent variable Log of imports; coefficients reported in levels:
  - Estimators: OLS, OLS, OLS, FE, FE.
  - Observations: 10,341 in all columns.
  - R-squared: 0.00, 0.00, 0.60, 0.76, 0.81.
  - Selected coefficient significance:
    - Preferential tariff: 0.0069** and 0.0083+ in some specifications.
    - Preference margin: -0.0193 and 0.0096* in some specifications.
  - Constant terms shown across columns (e.g., 1.9879; 1.9888; 2.2772; 1.8901; -1.1295).
  - Notes: Robust standard errors; + significant at 10%; * significant at 5%; ** significant at 1%.

### Robustness Checks and Trade Diversion Estimates
- Table 7: Robustness checks — difference-in-differences and triple-difference specifications
  - Columns (1)–(6) summary:
    - Estimator: FE with variations; Observations vary: 21,500; 470; 1,090; 1,357; 317; 317.
    - R-squared: 0.57; 0.74; 0.77; 0.72; 0.99; 0.97.
    - Elasticity of substitution (examples):
      - Column (1): -0.3185
      - Column (2): -12.0281
      - Column (3): -8.0169
      - Column (4): -4.4884
      - Column (5): 1.1562
      - Column (6): 1.1270
    - Selected significance: 6.7830+; 3.0652**; 2.1271*; others shown in source.
  - Definitions:
    - Diff-in-diff: COMESA vs. non-COMESA, Uganda vs. South Africa uses log of relative imports (1994, 2001).
    - Ratio of preference margin: Uganda vs. South Africa = log difference between preference margin in Uganda and in South Africa.
- Table 8: Estimates of Trade Diversion based on Uganda imports from non-COMESA countries (1994, 2000, 2001, 2002, 2003)
  - Columns (1)–(4) (FE estimator):
    - Observations: 62,302 in columns (1)–(3); 62,302 in column (4) (source presents 62,302 repeatedly).
    - R-squared: 0.78 in all reported columns.
    - Coefficient examples:
      - Log of MFN tariff: -0.8052; -0.7438; -0.6540; -0.6431 with significance markers (e.g., 0.2869**; 0.3440*; 0.3444+; 0.3458+ shown in source in adjacent formatting).
      - Log of preferential tariff: -0.1306; -0.1828; -0.3144; -0.0378 (other coefficient fragments shown in source).
      - Log of COMESA imports and COMESA share in imports included as interactors in some specifications (coefficients reported in source: e.g., Log of COMESA imports 0.1961; 0.1750; 0.0390; 0.5702; COMESA share in imports -1.1308; 2.6690).
    - Number of differences: One in all presented columns.
  - Notes: Robust standard errors in parentheses; + significant at 10%; * significant at 5%; ** significant at 1%.

### Appendix I — Uganda's Preferential Tariff Rates by 2-digit 1996 HS Codes (1994 and 2003)
- Selected entries (Preferential Tariff Rates, 1994; Preferential Tariff Rates, 2003; % Point Change):
  - HS 01 LIVE ANIMALS: 4.67; 4.00; -0.67 (first occurrence in appendix listing).
  - HS 02 MEAT AND EDIBLE MEAT OFFAL: 21.12; 16.00; -5.12
  - HS 03 FISH, CRUSTACEANS & AQUATIC INVERTEBRATES: 15.34; 6.00; -9.34
  - HS 04 DAIRY PRODS; BIRDS EGGS; HONEY; ED ANIMAL PR NESOI: 14.60; 13.20; -1.40
  - HS 07 EDIBLE VEGETABLES & CERTAIN ROOTS & TUBERS: 23.21; 6.00; -17.21
  - HS 08 EDIBLE FRUIT & NUTS; CITRUS FRUIT OR MELON PEEL: 24.08; 6.00; -18.08
  - HS 09 COFFEE, TEA, MATE & SPICES: 20.52; 4.19; -16.33
  - HS 10 CEREALS: 11.22; 5.13; -6.09
  - HS 22 BEVERAGES, SPIRITS AND VINEGAR: 55.90; 62.67; 6.76
  - HS 24 TOBACCO AND MANUFACTURED TOBACCO SUBSTITUTES: 55.00; 96.00; 41.00
  - HS 30 PHARMACEUTICAL PRODUCTS: 0.00; 0.00; 0.00
  - HS 31 FERTILIZERS: 0.00; 0.00; 0.00
- Appendix continued (selected 1994 vs. 2003 comparisons, second block):
  - HS 01 LIVE ANIMALS: 6.67; 7.00; 0.33
  - HS 02 MEAT AND EDIBLE MEAT OFFAL: 30.00; 25.00; -5.00
  - HS 03 FISH, CRUSTACEANS & AQUATIC INVERTEBRATES: 30.00; 15.00; -15.00
  - HS 07 EDIBLE VEGETABLES & CERTAIN ROOTS & TUBERS: 30.00; 15.00; -15.00
  - HS 09 COFFEE, TEA, MATE & SPICES: 25.15; 7.09; -18.06
  - HS 11 MILLING PRODUCTS; MALT; STARCH; INULIN; WHT GLUTEN: 28.82; 11.86; -16.96
  - HS 16 EDIBLE PREPARATIONS OF MEAT, FISH, CRUSTACEANS ETC: 30.00; 15.00; -15.00
  - HS 22 BEVERAGES, SPIRITS AND VINEGAR: 29.05; 71.67; 42.62
  - HS 24 TOBACCO AND MANUFACTURED TOBACCO SUBSTITUTES: 50.00; 105.00; 55.00
  - HS 30 PHARMACEUTICAL PRODUCTS: 0.00; 0.00; 0.00
  - HS 31 FERTILIZERS: 0.00; 0.00; 0.00
  - HS 41 RAW HIDES AND SKINS (NO FURSKINS) AND LEATHER: 30.00; 7.00; -23.00
  - HS 44 WOOD AND ARTICLES OF WOOD; WOOD CHARCOAL: 28.38; 17.29; -11.10
  - HS 45 CORK AND ARTICLES OF CORK: 12.86; 15.00; 2.14
- Appendix I provides full 2-digit HS code listing with preferential tariff rates in 1994 and 2003 and the percentage-point change for each code (complete table entries appear in the source).

*Source: _wp0740 - 03. The tariff rates are (IMF PDF chapter content provided).*

### Appendix II: Uganda's MFN Tariff Rates by 2-digit 1996 HS Codes, 1994 and 2003

### Appendix II: Uganda's MFN Tariff Rates by 2-digit 1996 HS Codes, 1994 and 2003

### MFN tariff rates by 2-digit HS codes (1994, 2003, Percentage point change)
- 46  MFR OF STRAW, ESPARTO ETC.; BASKETWARE & WICKERWRK — 28.33 15.00 -13.33
- 47  WOOD PULP ETC; RECOVD (WASTE & SCRAP) PPR & PPRBD — 12.11 5.89 -6.21
- 48  PAPER & PAPERBOARD & ARTICLES (INC PAPR PULP ARTL) — 15.05 9.21 -5.83
- 49  PRINTED BOOKS, NEWSPAPERS ETC; MANUSCRIPTS ETC — 8.42 4.05 -4.37
- 50  SILK, INCLUDING YARNS AND WOVEN FABRIC THEREOF — 13.00 15.00 2.00
- 51  WOOL & ANIMAL HAIR, INCLUDING YARN & WOVEN FABRIC — 13.06 14.78 1.72
- 52  COTTON, INCLUDING YARN AND WOVEN FABRIC THEREOF — 21.34 14.21 -7.13
- 53  VEG TEXT FIB NESOI; VEG FIB & PAPER YNS & WOV FAB — 12.26 14.45 2.19
- 54  MANMADE FILAMENTS, INCLUDING YARNS & WOVEN FABRICS — 14.77 16.08 1.32
- 55  MANMADE STAPLE FIBERS, INCL YARNS & WOVEN FABRICS — 15.91 10.13 -5.78
- 56  WADDING, FELT ETC; SP YARN; TWINE, ROPES ETC. — 24.81 8.27 -16.55
- 57  CARPETS AND OTHER TEXTILE FLOOR COVERINGS — 30.00 15.00 -15.00
- 58  SPEC WOV FABRICS; TUFTED FAB; LACE; TAPESTRIES ETC — 20.37 15.00 -5.37
- 59  IMPREGNATED ETC TEXT FABRICS; TEX ART FOR INDUSTRY — 17.20 14.00 -3.20
- 60  KNITTED OR CROCHETED FABRICS — 30.00 14.53 -15.47
- 61  APPAREL ARTICLES AND ACCESSORIES, KNIT OR CROCHET — 20.00 15.00 -5.00
- 62  APPAREL ARTICLES AND ACCESSORIES, NOT KNIT ETC. — 20.00 15.00 -5.00
- 63  TEXTILE ART NESOI; NEEDLECRAFT SETS; WORN TEXT ART — 19.48 15.04 -4.45
- 64  FOOTWEAR, GAITERS ETC. AND PARTS THEREOF — 30.00 14.45 -15.55
- 65  HEADGEAR AND PARTS THEREOF — 30.00 15.00 -15.00
- 66  UMBRELLAS, WALKING-STICKS, RIDING-CROPS ETC, PARTS — 30.00 15.00 -15.00
- 67  PREP FEATHERS, DOWN ETC; ARTIF FLOWERS; H HAIR ART — 30.00 15.50 -14.50
- 68  ART OF STONE, PLASTER, CEMENT, ASBESTOS, MICA ETC. — 21.79 14.33 -7.46
- 69  CERAMIC PRODUCTS — 22.50 13.93 -8.57
- 70  GLASS AND GLASSWARE — 17.98 12.95 -5.04
- 71  NAT ETC PEARLS, PREC ETC STONES, PR MET ETC; COIN — 21.47 14.42 -7.05
- 72  IRON AND STEEL — 11.96 7.07 -4.89
- 73  ARTICLES OF IRON OR STEEL — 17.64 6.65 -10.98
- 74  COPPER AND ARTICLES THEREOF — 12.07 7.00 -5.07
- 75  NICKEL AND ARTICLES THEREOF — 10.31 9.00 -1.31
- 76  ALUMINUM AND ARTICLES THEREOF — 15.88 7.65 -8.24
- 78  LEAD AND ARTICLES THEREOF — 12.00 6.30 -5.70
- 79  ZINC AND ARTICLES THEREOF — 12.73 5.60 -7.13
- 80  TIN AND ARTICLES THEREOF — 12.22 5.25 -6.97
- 81  BASE METALS NESOI; CERMETS; ARTICLES THEREOF — 10.00 6.94 -3.06
- 82  TOOLS, CUTLERY ETC. OF BASE METAL & PARTS THEREOF — 20.62 7.00 -13.62
- 83  MISCELLANEOUS ARTICLES OF BASE METAL — 20.00 12.11 -7.89
- 84  NUCLEAR REACTORS, BOILERS, MACHINERY ETC.; PARTS — 11.31 0.77 -10.54
- 85  ELECTRIC MACHINERY ETC; SOUND EQUIP; TV EQUIP; PTS — 18.58 6.67 -11.91
- 86  RAILWAY OR TRAMWAY STOCK ETC; TRAFFIC SIGNAL EQUIP — 10.00 0.00 -10.00
- 87  VEHICLES, EXCEPT RAILWAY OR TRAMWAY, AND PARTS ETC — 15.83 10.24 -5.60
- 88  AIRCRAFT, SPACECRAFT, AND PARTS THEREOF — 10.00 0.93 -9.07
- 89  SHIPS, BOATS AND FLOATING STRUCTURES — 17.35 3.06 -14.29
- 90  OPTIC, PHOTO ETC, MEDIC OR SURGICAL INSTRMENTS ETC — 12.14 2.39 -9.74
- 91  CLOCKS AND WATCHES AND PARTS THEREOF — 29.45 15.00 -14.45
- 92  MUSICAL INSTRUMENTS; PARTS AND ACCESSORIES THEREOF — 30.00 7.00 -23.00
- 93  ARMS AND AMMUNITION; PARTS AND ACCESSORIES THEREOF — 9.41 7.00 -2.41
- 94  FURNITURE; BEDDING ETC; LAMPS NESOI ETC; PREFAB BD — 24.86 15.17 -9.70
- 95  TOYS, GAMES & SPORT EQUIPMENT; PARTS & ACCESSORIES — 20.23 14.07 -6.16
- 96  MISCELLANEOUS MANUFACTURED ARTICLES — 27.25 11.08 -16.17
- 97  WORKS OF ART, COLLECTORS' PIECES AND ANTIQUES — 10.00 15.00 5.00

### Observations from Appendix II data
- The table reports MFN tariff rates by 2-digit 1996 HS codes with three columns: MFN Tariff Rates, 1994; MFN Tariff Rates, 2003; and Percentage Point Change MFN Tariff Rates.
- Largest percentage point reductions (examples from the table):
  - HS 92 MUSICAL INSTRUMENTS: -23.00 percentage points (30.00 to 7.00)
  - HS 56 WADDING, FELT ETC; SP YARN; TWINE, ROPES ETC.: -16.55 percentage points (24.81 to 8.27)
  - HS 96 MISCELLANEOUS MANUFACTURED ARTICLES: -16.17 percentage points (27.25 to 11.08)
- Some HS categories show tariff increases between 1994 and 2003:
  - HS 50 SILK: +2.00 percentage points (13.00 to 15.00)
  - HS 51 WOOL & ANIMAL HAIR: +1.72 percentage points (13.06 to 14.78)
  - HS 53 VEG TEXT FIB NESOI: +2.19 percentage points (12.26 to 14.45)
  - HS 54 MANMADE FILAMENTS: +1.32 percentage points (14.77 to 16.08)
  - HS 97 WORKS OF ART, COLLECTORS' PIECES AND ANTIQUES: +5.00 percentage points (10.00 to 15.00)

### Appendix III (as it appears in the source)
- Appendix III presents "Uganda's Imports by Country of Origin, 1994 and 2003" with columns "Imports, 1994 (US$ thousand)" and "Imports, 2003 (US$ thousand)" based on data at the commodity level.
- Representative rows included in the source (country, imports 1994, imports 2003, Rank fields as shown):
  - Burundi 176.68 5,325.44 107
  - Dem.Rp.Congo 1,140.56 3,929 8.98 58
  - Djibouti 2,787.93 260.00 134
  - Egypt 951.83 406,413.00 30
  - Eritrea 0.00 1,3684.62 84
  - Ethiopia 521.70 4,453.58 95
  - Kenya 186,784.85 1,357,194.21 1
  - Madagascar 0.00 1,361.99 129
  - Malawi 890.89 4,1300.13 57
  - Mauritius 1,175.95 362,479.16 37
  - Namibia 0.00 1,36200.91 64
  - Rwanda 74.46 6,6534.14 53
  - Sudan 28.25 8510.08 117
  - Swaziland 0.00 1,369,999.18 23
  - Zambia 1,831.85 3,1210.10 63
  - Zimbabwe 2,217.92 2,8848.67 49
- The source continues with an extensive country-level listing (numerous additional country rows for 1994 and 2003 imports are presented in the PDF).

*Source: Appendix II and accompanying appendices as provided in the supplied content.*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2007/_wp0740.pdf_