The impact of COVID-19 pandemic on trade balance in the East Africa Community region: evidence from a Pooled Mean Group model

Figures & data

Figure 1. Imports value for the selected EAC countries in USD million (March 2020–December 2022). Source: Central Bank of Kenya (CBK), National Bank of Rwanda (NBR), and Bank of Uganda (BOU) Statistical Databases. Notes: The value of imports is measured in USD million.

Figure 2. Exports value for the selected EAC countries in USD million (March 2020–December 2022). Source: CBK, NBR, and BOU Statistical Databases. Notes: The value of exports is measured in USD million.

Figure 3. Trade deficit for the selected EAC countries in USD million (March 2020–December 2022). Source: CBK, NBR, and BOU Statistical Databases. Notes: The values shown in Figure 3 depict a negative trade balance (trade deficit).

Table 1. Summary statistics of study variables.

Variable	Obs	Mean	Std. dev	Minimum	Maximum
Trade deficit	102	538.8225	365.9563	128.602	1498.83
Imports	102	893.7594	535.1527	201.593	2083.72
Exports	102	354.9367	183.4233	72.9909	680.931
Covid cases	102	6339.647	9128.813	39	39940
Consumer price index (CPI)	102	125.8189	16.8844	107	180.901
Exchange rate	102	1589.368	1519.2	103.74	3832.45
Broad money supply (M2)	102	14267.63	14131.44	2001.61	35806.4
Lending rate	102	15.6555	2.7774	11.75	20.9292
Crude oil price	102	70.9365	26.9332	17.66	117.72

Source: Author’s computation from STATA. Notes: For uniformity in analysis, we expressed broad money (M2) and trade deficit in USD by dividing them against each country’s monthly average exchange rate against the USD. Generally, the total value of imports exceeded that of exports for all the countries under the study period suggesting a trade deficit pattern.

Table 2. Panel unit root test results.

Variable	Testing level	W-t-bar Statistic	W-t-bar P-value	Order of integration
lnTrade Deficit	Level 1^st D	−0.6727 −3.4453	0.38340.0003	I (1)
lnImports	Level	−3.8541	0.0001	I (0)
lnExports	Level	−3.7344	0.0001	I (0)
lnCovid Cases	Level	−2.3315	0.0099	I (0)
lnConsumer Price Index	Level 1^st D	−0.0102 − 4.4944	0.49590.0000	I (1)
lnExchange Rate	Level 1^st D	−1.3383 − 4.7814	0.09040.0000	I (1)
lnM2 Money	Level 1^st D	−1.6233 − 5.1997	0.05230.0000	I (1)
Lending Rate	Level	−3.8509	0.0001	I (0)
lnCrude Oil Price	Level	−2.0649	0.0195	I (0)

Source: Compiled from STATA. Notes: D denotes difference.

Table 3. Full pooled mean group model regression results-trade deficit estimates.

Variables	(1)	(2)	(3)	(4)
D.lnTrade deficit	L.R coefficients	Kenya	Rwanda	Uganda
ECT		−1.071***	−0.997***	−1.254***
		(0.167)	(0.167)	(0.163)
D.lnCPI		−1.518	−1.311	−0.593
		(1.239)	(0.970)	(2.148)
D.lnCovid Cases		−0.0183	−0.0122	0.0337
		(0.0129)	(0.0155)	(0.0242)
D.lnExchange Rate		1.209	−0.0170	−1.230
		(1.883)	(2.662)	(1.961)
D.lnM2 Money		−1.333	1.269	−0.662
		(1.408)	(1.010)	(1.226)
D.Lending Rate		−0.120	0.0476	0.00417
		(0.235)	(0.0424)	(0.0225)
D.lnCrude Oil Price		0.0134	−0.0715	−0.0370
		(0.0987)	(0.134)	(0.152)
lnCPI	2.102***
	(0.546)
lnCovid Cases	0.0453***
	(0.0153)
lnExchange Rate	−0.284
	(0.658)
lnM2 Money	0.352
	(0.441)
Lending Rate	0.00644
	(0.0225)
lnCrude Oil Price	0.225***
	(0.0666)
Constant		−7.321	−7.237	−7.745
		(5.835)	(5.621)	(8.622)
Observations	99	99	99	99

Standard errors in parentheses. ***p < 0.01. **p < 0.05. *p < 0.1. Notes: The long-run estimates are represented in the second column and lack the difference operator ‘D’ while the third, fourth, and fifth columns present the short-run PMG estimates with the difference operator. ECT is the Error Correction Term and denotes model convergence in the long run.

Table 4. Full pooled mean group model regression results-import estimates.

Variables	(1)	(2)	(3)	(4)
D.lnImports	L.R coefficients	Kenya	Rwanda	Uganda
ECT		−1.057***	−0.888***	−0.472***
		(0.157)	(0.180)	(0.142)
D.lnCPI		−0.971	−0.229	−0.470
		(0.853)	(0.845)	(2.081)
D.lnCovid Cases		0.00121*	0.00693*	0.0574***
		(0.0102)	(0.0134)	(0.0212)
D.lnExchange Rate		−0.0667	2.188	−1.286
		(1.323)	(2.595)	(1.993)
D.lnM2 Money		−1.958*	−0.0544	−1.249
		(1.040)	(0.939)	(1.232)
D.Lending Rate		0.165	0.0870**	0.0149
		(0.183)	(0.0422)	(0.0218)
D.lnCrude Oil Price		−0.0184	−0.0858	0.353**
		(0.0778)	(0.132)	(0.145)
lnCPI	0.969*(0.513)
lnCovid Cases	0.00344(0.0161)
lnExchange Rate	−1.217*(0.716)
lnM2 Money	0.534(0.432)
Lending Rate	0.00310(0.0492)
lnCrude Oil Price	0.313***(0.0622)
Constant		1.610	3.439	2.800
		(5.424)	(5.090)	(3.135)
Observations	99	99	99	99

Standard errors in parentheses. ***p < 0.01. **p < 0.05. *p < 0.1. Notes: Imports is the dependent variable in this model.

Table 5. Full pooled mean group model regression results-export estimates.

Variables	(1)	(2)	(3)	(4)
D.lnExports	L.R coefficients	Kenya	Rwanda	Uganda
ECT		−0.862***	−0.260*	−0.584***
		(0.157)	(0.143)	(0.140)
D.lnCPI		−0.285	2.004	1.016
		(0.947)	(1.578)	(1.923)
D.lnCovid Cases		−0.00266	0.0163	0.0131
		(0.00925)	(0.0216)	(0.0202)
D.lnExchange Rate		0.0717	−2.522	−2.147
		(1.314)	(4.262)	(1.965)
D.lnM2 Money		1.276	0.476	−0.0141
		(0.962)	(1.643)	(1.122)
D.Lending Rate		0.0115	0.140**	0.00106
		(0.152)	(0.0678)	(0.0211)
D.lnCrude Oil Price		−0.0413	0.0619	0.166
		(0.0973)	(0.238)	(0.154)
lnCPI	−1.683(1.570)
lnCovid Cases	−0.0000986(0.0176)
lnExchange Rate	0.657(1.723)
lnM2 Money	−2.155**(1.018)
Lending Rate	−0.00450(0.0454)
lnCrude Oil Price	0.500***(0.101)
Constant		27.31**	6.108	14.80
		(12.10)	(4.797)	(9.478)
Observations	99	99	99	99

Standard errors in parentheses. ***p < 0.01. **p < 0.05. *p < 0.1. Notes: Exports is the dependent variable in this model.

Table A1. Different summary statistics per panel.

Variable		Mean	Std. dev	Min	Max	Obs
Trade deficit	Overall	538.8225	365.9563	128.602	1498.83	N = 102
	Between		414.2238	216.6756	1006.078	n = 3
			135.6607	114.6465	1031.575	T = 34
Imports	Overall	893.7594	535.1527	201.593	2083.72	N = 102
	Between		616.7851	355.3099	1566.689	n = 3
	Within		173.9581	351.36	1410.79	T = 34
Exports	Overall	354.9367	183.4233	72.9909	680.931	N = 102
	Between		211.1891	138.6342	560.6111	n = 3
	Within		60.1318	196.4219	496.8659	T = 34
Covid	Overall	6339.647	9128.813	39	39,940	N = 102
Cases	Between		3281.591	3913.618	10,073.5	n = 3
	Within		8722.667	−3674.853	36,206.15	T = 34
CPI	Overall	125.8189	16.8844	107	180.901	N = 102
	Between		17.0504	115.2488	145.4887	n = 3
	Within		9.4527	114.8672	161.2312	T = 34
Exchange	Overall	1589.368	1519.2	103.74	3832.45	N = 102
Rate	Between		1849.966	111.9273	3664.221	n = 3
	Within		61.6184	1439.657	1757.597	T = 34
M2 money	Overall	14,267.63	14,131.44	2001.61	35,806.4	N = 102
	Between		17,207.61	2391.137	34,001.29	n = 3
	Within		584.7597	12,605.43	16,072.73	T = 34
Lending	Overall	15.6555	2.7774	11.75	20.9292	N = 102
Rate	Between		3.2631	12.1312	18.5718	n = 3
	Within		0.7385	12.615	18.0129	T = 34
Crude	Overall	70.9365	26.9332	17.66	117.72	N = 102
Oil	Between		0	70.9365	70.9365	n = 3
Price	Within		26.9332	17.66	117.72	T = 34

Notes: Table A1 provides comprehensive summary statistics across the variables and countries forming the panel.

Table A2. Correlation matrix.

	Covid cases	CPI	Exchange rate	M2 money	Lending rate	Crude oil price
Covid cases	1
CPI	−0.2809	1
Exchange rate	−0.1744	−0.2579	1
M2 money	0.3049	−0.4580	−0.6018	1
Lending rate	−0.2886	0.1005	0.8676	−0.8557	1
Crude oil price	−0.0793	0.3565	0.0035	0.0244	−0.0119	1

Source: Compiled from STATA. Note. Generally, from the results presented in Table A2, multicollinearity was deemed not to be a problem in this study. Further, to control for any potential correlation, we logarithmically transformed the variables in our estimated model(s).

Table A3. Pedroni cointegration test results.

Pedroni’s cointegration tests
Model 1 Regressand = Trade Deficit			Model 2 Regressand = Imports			Model 3 Regressand = Exports
Test statistics	Panel	Group	Test statistics	Panel	Group	Test statistics	Panel	Group
v	−0.8174	.	v	−0.2111	.	v	0.3275	.
rho	−1.102	−0.6066	rho	0.131	0.75	rho	0.2469	0.8036
t	−5.765	−6.981	t	−2.793	−3.054	t	−1.85	−1.769
adf	−4.868	−5.941	adf	−0.6536	0.06772	adf	−2.411	−2.446
Ho: No Cointegration

Source: Author’s Computation from STATA.

Data availability statement

The data used for this study and the supporting Stata analyses do-files can be made available upon reasonable request from the author. The data is freely and readily accessible and was sourced from the country-specific Central Banks, the National Bureau of Statistics, Ministries of Health, and OPEC databases.