Elections, economic development and debt servicing in Africa

Figures & data

Figure 1. Number of presidential elections by countries, international foundation for electoral systems data, 1985–2015.

Figure 2. Government expenditure in pre-election and election years, world development indicators data, 1985–2015.

Figure 3. Government expenditure and GDP per capita, world development indicators data data, 1985–2015.

Figure 4. The distribution of debt servicing on election by country, international financial statistics data, 1985–2015.

Table 1. Description of variables and data sources

Variable	Definition	Expected Sign	Data source
Debt servicing	Sum of principal repayments and interest actually paid in currency, goods, or services on long-term debt, interest paid on short-term debt, and repayments (repurchases and charges) to the IMF.		IFS
Election	Election year/season is a dummy, with 1 indicating election year and 0 if non-election year.	±
GDP per capita	GDP per capita is the overall GDP of a country divided by the population in 2017 constant dollars	+	WDI
Inflation	Inflation is proxied by the consumer price index and it is captured as the price of a weighted average market basket of consumer goods and services purchased by households.	±	WDI
Unemployment	Unemployment rate is the proportion of unemployed as a per centage of labour force	+	WDI
Governance System	System of governance is a dummy, with 1 indicating simple majority and 0 if parliamentary system of governance	±
Age Dependency Ratio	Dependency ratio is the proportion of people aged less than 16 and greater than 65 in the population	+	WDI
Election System	Electoral system is a dummy, with 1 indicating presidential system of government and 0 if parliamentary system of governance.	±

Note: WDI is World Development Indicators (Accessed: https://databank.worldbank.org/source/world-development-indicators#) and IFS is IMF’s International Financial Statistics (Accessed: https://data.imf.org/regular.aspx?key=61545853)

Table 2. Distribution of means of some selected variables across years of election

Variable	Obs.	Election	No election	Difference	Combined	t-statistics
Government Expenditure	1333	18.519 (0.774)	15.758 (0.229)	2.761 (0.631)	16.178 (0.228)	4.377***
Unemployment	1333	8.294 (0.486)	6.668 (0.228)	1.626 (0.544)	6.951 (0.207)	2.987***
Dependency ratio	1333	89.136 (0.811)	88.929 (0.397)	0.207 (0.934)	88.965 (0.356)	0.222
GDP per capita	1333	6.424 (0.078)	6.441 (0.033)	0.016 (0.085)	6.438 (0.031)	0.182
Inflation	1333	65.214 (0.087)	59.743 (0.044)	5.471 (0.103)	61.990 (0.041)	7.234***

Robust standard errors in parentheses;

***p <0.01, **p <0.05, *p <0.1

Table 3. Bivariate estimates for effect of election variables on debt servicing (dependent variable: debt servicing)

	(1)	(2)	(3)
Variable	OLS	OLS	OLS
Election	−0.039***
	(0.0093)
Governance system		0.041*
		(0.0234)
Electoral system			0.012
			(0.0086)
Constant	0.141***	0.053	0.113***
	(0.0044)	(0.0463)	(0.0152)
Observations	1,333	1,333	1,333
R-squared	0.0120	0.0031	0.0020

Robust standard errors in parentheses

*** p <0.01, ** p <0.05, * p <0.1

Table 4. Effect of electoral on government debt servicing (dependent variable: debt servicing)

	1	2	3	4	5
Variables	OLS	RE1	RE2	GMM1	GMM2
Elections	−0.1882***	−0.0102***	0.0001	−0.0114***	−0.0130***
	(0.0520)	(0.0030)	(0.0005)	(0.0021)	(0.0011)
Unemployment	−0.0015***	−0.0268	0.0039	−0.4562***	−0.2646***
	(0.0005)	(0.0178)	(0.0044)	(0.0765)	(0.0772)
Government system	0.0476***	0.4534***	0.0176	0.0055	0.0180
	(0.0150)	(0.1202)	(0.0244)	(0.0481)	(0.0296)
Dependency ratio	0.0005	0.0003	0.0002**	−0.0018***	−0.0012***
	(0.0003)	(0.0002)	(0.0001)	(0.0003)	(0.0004)
GDP per capita	−0.0420***	0.0020	0.0002	−0.0099***	−0.0120***
	(0.0058)	(0.0053)	(0.0005)	(0.0026)	(0.0027)
Election $\times$ GDP per capita	0.0241***	–	−0.0001*	–	0.0126*
	(0.0075)	–	(0.0000)	–	(0.0069)
Electoral system	0.0083	0.0061	−0.0086**	0.0059***	0.0104***
	(0.0080)	(0.0110)	(0.0038)	(0.0010)	(0.0015)
Inflation	0.0370***	0.0046	−0.0011	0.1826**	0.2202***
	(0.0105)	(0.0056)	(0.0015)	(0.0815)	(0.0539)
Debt servicing (−1)	–	–	0.9904***	−0.0150***	−0.0145***
	–	–	(0.0034)	(0.0022)	(0.0027)
Constant	0.1354*	6.2294***	−0.0138	–	–
	(0.0727)	(0.4540)	(0.0813)	–	–
Observations	1,133	1,133	1,333	1,169	1,169
R-squared	0.1941	-	-	–	–
Countries	42	42	42	40	40
Instruments	-	-	-	33	37
Hausmann Test	na	-	-	-	-
LM Test	0.000 [1.000]	-	na	na	na

Standard errors in parentheses; *** p <0.01, ** p <0.05, * p <0.1

Table A1. Unit root test results for the variables at levels

Variables	Constant		Constant and	Trend
	CADF (Z-t-bar)	CIPS	CADF (Z-t-bar)	CIPS
Debt servicing	−0.679	−3.293**	1.508	−2.526
GDP per capita	−6.232***	−5.280***	−7.141***	−3.524***
Inflation	5.184	−1.510	3.908	−2.111
Dependency ratio	0.487	−1.367	0.814	−2.571
Government expenditure	−1.353***	−3.363**	−2.347**	−3.771**
Unemployment	−1.521***	−4.243***	−4.044***	−5.065***

A single asterisk (*) denotes significance at 10% level, two asterisks (**) at 5% level, and three asterisks (***) at 1% level. Both CADF & CIPS test the $H_0$: All panels contain unit root against $H_1$: Some panels are stationary. CADF means Cross-sectionally Augmented Dickey–Fuller while CIPS refers to Cross-sectionally Augmented Im Pesaran Shin

Unit root test results for the variables at first difference

Variables	Constant		Constant and	Trend
	CADF (Z-t-bar)	CIPS	CADF (Z-t-bar)	CIPS
Debt servicing	−11.153***	−5.337***	−6.834***	−9.513***
GDP per capita	−13.889***	−6.666***	−13.610	−7.388***
Inflation	−8.510***	−4.931***	−5.164***	−8.534***
Dependency ratio	−13.324***	−5.176***	−11.515***	−6.863***
Government expenditure	−16.551***	−5.823***	−13.096***	−7.525***
Unemployment	−12.876***	−5.365***	−15.952***	−5.845***

A single asterisk (*) denotes significance at 10% level, two asterisks (**) at 5% level, and three asterisks (***) at 1% level. Both CADF & CIPS test the $H_0$: all panels contain unit root against $H_1$: Some panels are stationary. CADF means Cross-sectionally Augmented Dickey–Fuller while CIPS refers to Cross-sectionally Augmented Im Pesaran Shin.

Variables included in the debt servicing model

Variable	Obs.	Mean	Std. Dev.	Min	Max
Election year	1,169	0.1523	0.3594	0.0000	1.0000
Government expenditure	1,169	16.1783	7.8087	2.0471	69.5428
Unemployment	1,169	7.4617	8.0207	0.0000	39.3000
Governance system	1,169	1.9752	0.1556	1.0000	2.0000
Dependency ratio	1,169	88.0333	13.1466	40.7957	112.6829
GDP per capita	1,169	6.4648	1.1082	4.6121	10.0583
Electoral system	1,169	1.6775	0.4676	1.0000	2.0000
Inflation	1,169	4.7366	0.3609	4.1614	10.1069
Debt servicing	1,169	13.4462	13.0059	0.0000	101.8891

Source: Authors’ construct using Stata V.15

Post-estimation test for the GMM estimations

	Debt servicing
Description of test	Chi2 value	z	Prob.
Arellano-Bond test for AR(1) in first differences		16.6	0.0006
Arellano-Bond test for AR(2) in first differences		1.05	0.251
Sargan test of over-identification restrictions	232.17		0.219
Hansen test of over-identification restrictions	8.97		0.372
Difference-in-Hansen tests of exogeneity of instrument subsets: GMM instruments for levels
Hansen test excluding group	7.63		0.287
Difference (null H = exogenous)	0.97		0.173
Hansen test excluding group	15.72		0.293
Difference (null H = exogenous)	0.81		0.114

Source: Author’s construct using Stata V.15

List of countries in the study

Angola	Eritrea	Madagascar	Senegal
Benin	Ethiopia	Malawi	Seychelles
Botswana	Gabon	Mali	Sierra Leone
Burkina Faso	Gambia	Mauritius	South Africa
Cameroon	Ghana	Mauritania	Sudan
Cape Verde	Guinea	Mozambique	Tanzania
Central African Republic (CAR)	Guinea Bissau	Namibia	Togo
Comoros	Ivory Coast	Niger	Uganda
Congo (Brazzaville)	Kenya	Nigeria	Zambia
Democratic Republic of Congo	Lesotho	Rwanda	Zimbabwe
Equatorial Guinea	Liberia	Sao Tome and Principe

Source: Authors’ construct, 2021.