General literacy and effective taxation system in Africa

ABSTRACT

This inquiry provides empirical evidence on the effect of general literacy on tax revenue mobilization in Africa over the period 1996–2021. This evidence is important given the low level of taxation in Africa and the lack of clarity on the nexus between a country’s literacy level and taxation effort. The paper employs a simultaneous quantile regression, and uses three different indicators of tax revenue mobilization as a percentage of GDP. The motivation for quantile regression technique is primarily based on two essential reasons: (i) the argument that the effect of literacy on taxation effectiveness is dependent on the existing level of tax revenue mobilization; (ii) tax revenue mobilization policies, to be effective, need to be adapted differently across countries with low, intermediate, and high levels of existing tax mobilization effort. Results show that improvement in general literacy appears to promote tax mobilization effort, and that the effect of general literacy on taxation is stronger for direct taxes than for indirect taxes (in terms of both significance and magnitude). Furthermore, in most of the results, the impact of general literacy is slightly stronger in the upper quantiles of the general literacy distributions. The results remain robust over alternative econometric estimation techniques.

KEYWORDS:

• General literacy
• taxation
• direct tax
• indirect tax
• Africa

1. Introduction

Does educational achievement affect countries’ taxation efforts? What is the role of education in building a culture of tax compliance and accountability? There are three main reasons for conducting this investigation in the African context. First, the very low level of tax to Gross Domestic Product (GDP) ratio in Africa (OECD/AUC/ATAF 2021), when taxation is paramount for state building and economic development. Second, the unsettled debate in the existing literature concerning the nexus of taxpayer literacy and taxation efforts. Third, the need to empirically identify the impact of taxpayer literacy on taxation efforts in order to provide more nuanced policy advice.

With regards to the first reason, the very low tax to GDP ratio for African countries suggests that these have serious problems in promoting compliance. Available evidence indicates that the average tax to GDP ratio for sub-Saharan African countries stands at around 16.6 percent (OECD/AUC/ATAF 2021). Moreover, country level variations appear to be large. In 2018, countries such as Seychelles (32.4 percent), Tunisia (32.1 percent), South Africa (29.1 percent) and Morocco (27.6 percent) had the highest tax-to-GDP ratios, while countries such as Nigeria (5.7 percent), Equatorial Guinea (5.9 percent), the Democratic Republic of the Congo (6.6 percent) and Madagascar (11.9 percent) had the lowest tax-to-GDP ratios (OECD 2020). These figures clearly indicate the existence of significant cross-country variations in tax administration effectiveness. The most effective tax administration systems generate more tax revenue relative to their national income. The average tax to GDP ratio in Africa remains low despite the robust economic growth record achieved over the last decade (Favourate Y. Sebele-Mpofu 2020; OECD/ATAF/AUC 2022). For instance, while Africa’s average economic growth has been over 3.1 percent in the 2010–2022 periods (World Bank 2023), tax to GDP ratio grew only marginally (1.5 percent). This leads into questions concerning the factors that determine the tax efforts of African countries. Given that tax is an essential instrument for economic development, the evidence for African economies warrants empirical investigation on the effectiveness of the tax administration system in the region.

The second reason is rooted in the limited amount of literature on the impact of taxpayer’s literacy on tax efforts. Chang Otto, Nichols Donald, and Schultz Joseph (1987) asserts that tax awareness increases with the level of education, and, as a consequence, compliance behaviors of taxpayers increase. Mutascu and Danuletiu (2013) argue that literacy has a strong impact on tax payments and acceptance of taxation burdens, causing a notable informational asymmetry across taxpayers. This implies that difference in tax knowledge creates information asymmetry among taxpayers, affecting compliance. The literature on the nexus between taxpayer literacy and tax effort is rather limited in African context. The very few empirical studies on African economies focused on micro-level country case studies, with the aim to generate custom made policy recommendations (Mascagni and Santoro 2018).

Previous studies have placed too much emphasis on questions such as: what are the most effective ways of increasing the tax literacy of small taxpayers? How can tax literacy programs have the greatest impact on taxpayer engagement? Focusing on such questions in a context where the general literacy rate is low is useful, but not sufficient. This paper argues that specific taxpayer literacy programs promoted to improve taxpayer engagement, though important, might not be effective in a society where the general level of literacy is poor. Therefore, it investigates the role of general literacy rate in terms of promoting tax efforts of African countries. This investigation is to the best of my knowledge the first of its kind for African countries – no other study analyzed the role of cross-country variation in taxpayer literacy in predicting taxation efforts.

With regards to the third reason, it is essential to empirically identify the relationships between cross country variations in educational outcomes and in tax efforts for policy purposes. A better understanding of the taxpayer literacy–tax effort nexus is crucial for the development of effective policies aimed at improving tax compliance, tailored differently across countries with different levels of literacy. Therefore, this paper aims to contribute to the literature by empirically assessing the impact of cross-country variations in taxpayer literacy on taxation efforts. To this end, the main research questions are: What are the regional variations across Africa in terms of taxpayer literacy rate and taxation efforts? How does a taxpayer literacy rate affect cross-country variation in taxation efforts when the existing level of tax-to- GDP ratio matters?

In order to answer these research questions, this paper carries out quantitative analysis of data from a panel of 46 African countries. The objective is to assess the impact of taxpayer literacy on tax revenue mobilization. The methodology used is based on the simultaneous quantile regression estimation technique and utilizes three different indicators of tax revenue mobilization as a percentage of GDP. Results indicate that literacy rate has a positive and statistically significant impact on tax revenue mobilization. The study provides empirical evidence for policymakers that improving literacy plays a crucial role in enhancing tax revenue mobilization. Moreover, the findings of this study stimulate further research on the relationship between literacy rates and tax revenue mobilization in the context of developing countries.

This paper is organized as follows: Section 2 provides a review of the theoretical and empirical literature. Section 3 describes the data and discusses the methodology used for the study. Section 4 analyzes, reports, and discusses the empirical results of the study. Finally, Section 5 summarizes the concluding remarks for the study.

2. The literature and conceptual framework

The theories of taxpayer compliance behavior can be categorized into five main categories: economic deterrence, fiscal exchange, social and demographic factors, comparative treatment, and political accountability (Banerjee 1992; D'Arcy 2011; Kirchler, Hoelzl, and Wahl 2008; Kunawotor et al. 2022; McKerchar and Evans 2009). This investigation is motivated by recent developments in tax compliance studies that have been based on social and psychological theories (Devos 2014). Furthermore, the literature on the determinants of countries’ tax efforts is inconclusive. Research on taxpayers’’ behavior and the tax efforts of countries has identified significant knowledge gaps that need to be addressed through systematic study.

It is widely recognized that taxpayers’ knowledge about the role of taxation in social and economic development, as well as the costs of non-compliance, are crucial factors that support tax efforts (Nichita et al. 2019; Yuliati and Fauzi 2020). Studies have shown that tax knowledge is positively correlated with tax compliance, meaning countries with more knowledgeable citizens tend to have better results in terms of tax revenue mobilization. For example, Palil, Akir, and Ahmad (2013) found that taxpayers with higher education and knowledge about taxation were more compliant. Similarly, Kirchler, Hoelzl, and Wahl (2008) argued based on empirical evidence, that taxpayers with a good understanding of taxes are more likely to comply. Eriksen and Fallan (1996) suggested that increasing taxpayer knowledge increases their consideration of compliance. However, it is worth noting that under certain circumstances, greater taxpayer knowledge may lead to non-compliance, as individuals with a better understanding of taxes and related rules may exploit loopholes.

Nevertheless, experimental evidence suggests that providing information at a low cost to taxpayers can mitigate the impact of uncertainty on compliance (James et al. 2010). There is limited research at the macro level on the role of literacy rates on the tax efforts of states. However, previous micro-level studies have documented the influence of taxpayers’ knowledge on compliance behavior (Intansari and Supramono 2022). Furthermore, Eriksen and Fallan (1996) argued that with a reasonable understanding of tax laws, people are more willing to respect the tax rules and comply with tax obligations. Additionally, individuals who are educated and knowledgeable about tax planning and tax laws are more likely to have a positive attitude towards tax compliance and reduce their tendencies for non-compliance (Bornman and Wassermann 2020; Mohdali and Pope 2014).

Tax education has been shown to improve compliance behavior by enhancing taxpayers’ understanding of tax-related matters. Several studies suggest that tax education helps taxpayers comprehend tax laws and procedures, leading to a positive attitude towards tax compliance (Clifford and Amayi 2013; Normala 2007; Reckers, Sanders and Roark 1994). Olowookere and Fasina (2013) found that tax education activities that focus on enlightening taxpayers about the socioeconomic implications of tax evasion and the transparent use of tax proceeds have a significant influence on compliance behavior. Tax education becomes particularly important when the objective is to increase tax revenue, especially in a changing environment where official tax assessments may not be sufficient (Normala 2007). Therefore, while a better understanding of tax matters promotes compliance, a lack of knowledge about the role of taxes increases non-compliance. Moreover, knowledge about taxation practices can enhance the perceived power of authorities, as awareness of their effectiveness in detecting fraud can make them appear more powerful (Kirchler, Hoelzl and Wahl 2008). Based on the debates presented in this section, the conceptual framework for this study is as follows (Figure 1).

Figure 1. Schematic design of the conceptual framework. Source: Author's construction based on the literature.

3. Research methods and the data

3.1. Methodology

The study utilizes a quantitative research approach to examine the relationship between the general literacy rate and the tax-to-GDP ratio in African countries, using secondary quantitative data. The descriptive analysis in this study focuses on comparing general literacy achievements and taxation efforts across different regions of Africa. Two techniques were utilized for this purpose. First, scatter plots were used to illustrate variations in general literacy achievements and taxation efforts at the regional and country levels. Second, country-level mean indexes of general literacy achievements and taxation efforts were utilized.

Consistent with existing literature, this study applies econometric estimation to investigate the impact of general literacy on tax revenue mobilization. The econometric model used closely follows empirical models tested in previous studies (Clifford and Amayi 2013; Devos 2014; Olowookere and Fasina 2013; Palil, Akir, and Ahmad 2013). The following model specification was used for empirical estimation: (1) $T{R}_{\mathrm{it}}=\theta +{\alpha }_1(\mathrm{SS}{E}_{\mathrm{it}})+\alpha X_{\mathrm{it}}+{\epsilon }_{\mathrm{it}}$(1) In the equations, the main dependent variable is the level of tax revenue ($T{R}_{\mathrm{it}}$), measured by the tax to GDP ratio. To enure robustness and account for the differences in the sources of taxation, we also consider direct tax revenues and indirect tax revenues as percentages of GDP. We measure general literacy achievement $(\mathrm{SS}{E}_{\mathrm{it}})$ by using the secondary school enrollment rate for each country in the sample. While the secondary school completion rate would be a better measure of literacy, we use enrollment rate due to data availability for African countries.

To explain cross-country differences in the impact of general literacy on taxation, we include a range of country-level macroeconomic and institutional control variables ($X_{\mathrm{it}}$). These variables include governance quality, urban population size, trade, economic growth, aid received, fixed capital formation, foreign direct investment inflow, access to mobile telephone services, government annual expenditures, and a middle-income dummy variable. The selection of these control variables is guided by the existing literature (Bueno de Mesquita, Downs, and Smith 2017; Clifford and Amayi 2013; Mutascu and Danuletiu 2013; Ricciuti, Savoia, and Sen 2019).

The quantile regression technique is applied to estimate the model specification. Quantile regression is used for three main reasons. Firstly, taxation is persistent over time (Gang 2008), so it is important to examine the effect of general literacy on taxation using the current level of taxation. Secondly, quantile regression is a suitable technique for analyzing the relationship between variables when the interest is to find a causal link beyond the mean-to-mean effects (Cook and Manning 2009; Uribe and Guillen 2020). Thirdly, quantile regression tends to produce robust results by resisting the influence of outlier observations (Waldmann 2018). Given the persistent nature of taxation and the cross-country panel dataset used in this study, it is important to analyze how the effect of general literacy on taxation varies at different levels of taxation. This means that when the current level of taxation is different, the level of general literacy is expected to have a different impact on taxation. To achieve this, quantile regression models were estimated for the 10th, 25th, 50th, 75th, and 90th percentiles in order to identify how sensitive the effect of general literacy on taxation is to the level of existing taxation in the sample of countries.

3.2. The data

The data for this study were obtained from the World Bank, IMF, Government Revenue Data Sets (GRD) and the websites of finance ministries of African countries. The study uses a panel dataset consisting of 46 African countries from 1996 to 2021. The choice of countries and time period is based on data availability. Governance quality data is available from 1996 onwards. The observations are for annual periods, resulting in a maximum of twenty-six time series data points per country and a total of 1196 possible observations.

The lists of countries included in this study can be found in Appendix Table A1. According to the data presented in Table A1, Eswatini, South Africa, Mauritius, Tunisia, Algeria, and Cabo Verde are the top countries in terms of general literacy achievements, while Tanzania, Niger, Guinea Bissau, Uganda, Zambia, and the Central African Republic are the bottom five countries. The top countries for tax revenue mobilization (measured by the tax-to-GDP ratio) include Tunisia, Lesotho, Namibia, Solomon Islands, Seychelles, South Africa and Morocco, while Equatorial Guinea, Sudan, the Democratic Republic of the Congo, Gambia, the Congo Republic and the Central African Republic are the bottom countries. The summary statistics can be found in Appendix Table A2. The description and sources of variables are presented in Appendix Table A3. Table A4 provides the correlation matrix. From Table A4, it is apparent that the three measures of tax revenue mobilization dimensions (total tax revenue, direct tax revenue and indirect tax revenue) are highly correlated, indicating the need to conduct separate estimations for each of them.

Figure 2 presents the general literacy rate and total tax-to-GDP ratio for the sampled countries, averaged over the study period (1996–2021). The fitted line shows a strong positive relationship between the total tax revenue-to-GDP ratio and general literacy rates. Furthermore, Figure 3 illustrates a strong positive correlation between the direct tax revenue-to-GDP ratio and general literacy rates. Figure 4 indicates a very strong positive correlation between the indirect tax revenue-to-GDP ratio and general literacy rates.

Figure 2. Scatter plot of general literacy rate vs. total tax revenue- to-GDP ratio.

Figure 3. Scatter plot of general literacy rate vs. direct tax revenue-to-GDP ratio.

Figure 4. Scatter plot of general literacy rate vs. indirect tax revenue-to-GDP ratio.

4. Empirical results

Tables 1 and 2 present empirical findings on the causal relationship between different measures of tax revenue mobilization (total tax revenue to GDP ratio, direct tax to GDP ratio and indirect tax to GDP ratio) and general literacy achievements. Each table has two panels: the left panel shows the results of the generalized least square related regression, and the right panel shows the results of the quantile regression estimation. The consistent differences in the estimated coefficients of general literacy achievements among the different quantile regression results (in terms of sign, significance, and magnitude) support the importance of the empirical analysis strategy employed.

Table 1. General literacy achievement and taxation.

Variable	Quantile levels
	GLS	Q.10	Q.25	Q.50	Q.75	Q.90
SSE	.0004714***	.000535***	.00049***	.000742***	.001372***	.001493***
	(.000132)	(.000093)	(.000085)	(.000134)	(.000099)	(.0003402)
INST	0100351***	.007133***	.005934***	.008685***	.01115***	.008464***
	(.0013958)	(.000817)	(.000833)	(.001272)	(.000819)	(.002779)
IVA	.0000303	−.000316	−.000249*	−.000260**	−.0005173	−.000798
	(.000199)	(.000238)	(.000144)	(.000133)	(.000358)	(.000592)
GDP	.0003527	.000381	.000545	0003436	.000818*	.000135
	(.000239)	(.000371)	(.000339)	(.000281)	(.000465)	(.000787)
FDI	.000315**	.000032	.0001063	−.0000379	−.000534	.000127
	(.000146)	(.000211)	(.000305)	(.0001976)	(.000686)	(.000839)
INF	−.000064	−.000575**	−.000525	−.000092	−.00007	−.0001019
	(.0000413)	(.000239)	(.000329)	(.000305)	(.000195)	(.0001764)
MOB	.0001863***	.000197 ***	.0001895***	.000265***	.000464	.0000294
	(.000035)	(.000028)	(.0000401)	(.000071)	(.000356)	(.000088)
AID	−.0004601***	−.0000236	−.0003066 *	−.000575***	−.000282	−.000181
	(.000167)	(.000155)	(.000176)	(.000218)	(.000394)	(.000718)
UPOP	.0001384	−.000705***	−.0006426***	−.000669***	.000429*	−.0009694
	(.000319)	(.000125)	(.0000777)	(.000092)	(.000237)	(.000638)
GFC	.0003859**	.000492***	.000324***	−.000279	−.000317	−.000187
	(.000169)	(.00017)	(.000115)	(.000175)	(.000258)	(.000568)
GGE	.0000231	0000313	.0000234	.0000164	.000223	.0003489
	(.0000209)	(.000163)	(.000163)	(.000149)	(.000175)	(.000945)
GNB	−.000942***	−.001052***	−.001539***	−.0004224	−.000943	−.0016022
	(.000199)	(.000405)	(.000466)	(.000679)	(.000632)	(.001052)
TRD	.0001234**	.000191 ***	.0002903***	.000335***	.000539***	.001026***
	(.000055)	(.000058)	(.0000394)	(.000078)	(.000104)	(.000233)
MID	−.026818***	.0048397	.0088312	.004446	.0048117	.012346
	(.0041295)	(.005297)	(.0059036)	(.006716)	(.005909)	(.018138)
_cons	.088539***	.069877***	.074894***	.098353***	.088159***	.094991***
	(.014039)	(.005633)	(.003955)	(.004822)	(.014015)	(.017990)
Obs.	909	909	909	909	909	909
R^2	0.4261	0.3682	0.3441	0.3436	0.3997	0.4141

Note: Dependent variable: Tax-to- GDP ratio (TR) measured by total tax revenue as a percentage of GDP. Regression results for the quartile regressions are obtained by using sqreg simultaneous quantile equations technique. ***, **, * indicates significance at ρ < 0.01, ρ < 0.05 & ρ < 0.1 respectively. Bootstrap Standard errors are presented in brackets.

Table 2. Educational achievement, direct taxation, and indirect taxation.

Variable	Direct taxation						Indirect taxation
	GLS	Q.10	Q.25	Q.50	Q.75	Q.90	GLS	Q.10	Q.25	Q.50	Q.75	Q.90
SSE	.000274***	.000169**	.00021***	.000431***	.001005***	.00109***	.000213**	.000345***	.000409***	.000354***	.000192**	.000116
	(.000068)	(.000069)	(.000043)	(.0000824)	(.000065)	(.000072)	(.000099)	(.000622)	(.000068)	(.000088)	(.000086)	(.00017)
INST	.003403***	.00379***	.00306***	.00416***	.005775***	.00207***	.007396***	.004012***	.004132***	.004192***	.004647***	.00465***
	(.000734)	(.000432)	(.000568)	(.000616)	(.000576)	(.000892)	(.001009)	(.000828)	(.001063)	(.000799)	(.000926)	(.000926)
IVA	.0000547	.00031**	.000301**	.000525***	.000558***	.0002616	−.000201	−.000788**	−.000907**	−.001026**	−.000679**	−.0005326
	(.000103)	(.000127)	(.00015)	(.000101)	(.000109)	(.000221)	(.000149)	(.000055)	(.000064)	(.000184)	(.000269)	(.000387)
GDP	−.000169	.0000557	.0000698	−.000198	−.000089	.000394	.000519***	.000676***	.000862***	.0003426	.000407	−.000168
	(.000119)	(.000111)	(.000157)	(.000308)	(.000423)	(.000315)	(.00019)	(.000176)	(.000203)	(.000381)	(.00042)	(.00059)
FDI	.000248***	.000192**	.0000976	.0001334	−.0000983	.000708**	.000077	−.000026	−.0000614	−.000169	−.000103	−.000499*
	(.000073)	(.000095)	(.000131)	(.000095)	(.000333)	(.000326)	(.000115)	(.000155)	(.000135)	(.000189)	(.000307)	(.000252)
INF	.000753	−.000033	.000112	−.0000119	−.0000126	−.000038	−.000062*	−.000915***	−.000445	−.000181	−.000078	−.000087
	(.000021)	(.000064)	(.000067)	(.0000712)	(.000036)	(.000083)	(.000033)	(.000241)	(.000348)	(.000239)	(.000143)	(.000056)
MOB	.000127***	.000105***	.000136***	.000161***	.0000371*	.0000285	.000048*	.000117**	.0000485	.0000173	.0000275	.0000464
	(.0000182)	(.000023)	(.000028)	(.000034)	(.0000206)	(.000044)	(.000027)	(.000046)	(.000042)	(.0000234)	(.00003)	(.000047)
AID	−.000253***	.000896	−.0000722	−.000219*	.000093	−.000094	−.000221*	.0000602	−.000053	−.000336*	−.000372	−.000357
	(.000084)	(.000058)	(.000086)	(.000119)	(.000153)	(.000145)	(.00013)	(.000112)	(.000156)	(.000196)	(.000253)	(.000315)
UPOP	.0000293	−.000089*	−.00038***	−.000653***	−.00032***	−.000357**	.000161	−.00028***	−.000172**	−.000026	−.000143	−.000328
	(.000188)	(.000047)	(.000074)	(.000089)	(.000084)	(.000172)	(.000208)	(.000099)	(.000081)	(.000134)	(.000203)	(.000297)
GFC	.000352***	.0001038	−.0000499	−.000215*	−.00026***	−.000019	.000089	.000471***	.000303***	.000231	−.000048	−.000264
	(.000087)	(.000099)	(.000083)	(.000118)	(.000100)	(.000284)	(.000132)	(.000115)	(.000095)	(.000142)	(.000174)	(.000251)
GGE	.000565	.000665	.000965	.0000255	.000191	.000269	.0000115	.000021	.000472	−.0000176	.0000891	.000653
	(.000011)	(.000058)	(.000066)	(.000102)	(.000194)	(.000336)	(.000017)	(.000076)	(.000154)	(.000132)	(.000328)	(.000444)
GNB	−.00041***	−.00069** ***	−.001095***	−.0007922**	−.000559	−.000466	−.00054**	.0001318	.0002862*	.0000567	−.000501	−.000873
	(.000099)	(.000275)	(.000299)	(.0004)	(.000345)	(.000317)	(.000157)	(.000171)	(.000164)	(.000248)	(.000387)	(.000559)
TRD	.000156***	−.000027	.0000993**	.000176***	.000099***	.000207**	−.000018	.000073	.0000504	.000228***	.000503***	.000538***
	(.000028)	(.000035)	(.0000392)	(.0000458)	(.000039)	(.000083)	(.000042)	(.000048)	(.000039)	(.000041)	(.000077)	(.000099)
MID	−.001867	−.00636*	.0056034	.009428**	.010283***	.02937***	−.02864***	−.010348**	.0021701	−.001927	−.0005902	.004777
	(.002076)	(.003512)	(.0051675)	(.003804)	(.003479)	(.007510)	(.00321)	(.004667)	(.002639)	(.003218)	(.003577)	(.004287)
_cons	.01433***	.00827**	.0168535***	.024378***	.012586**	.008991	.073947***	.054518***	.064896***	.072631***	.073363***	.08556***
	(.008604)	(.003226)	(.004426)	(.004914)	(.005812)	(.010654)	(.00882)	(.005439)	(.004164)	(.00625)	(.00889)	(.011554)
Obs.	901	901	901	901	901	901	901	901	901	901	901	901
R2	0.4371	.2261	0.2478	0.3052	0.4276	0.4796	0.2286	0.3046	0.2415	0.2551	0.3038	0.3408

Note: Dependent variable: Tax-to- GDP ratio (TR) measured by total tax revenue as a percentage of GDP. Regression results for the quartile regressions are obtained by using sqreg simultaneous quantile equations technique. ***, **, * indicates significance at ρ < 0.01, ρ < 0.05 & ρ < 0.1 respectively. Bootstrap Standard errors are presented in brackets.

The general findings presented in Table 1 regarding the causal association between general literacy achievements and cross-country variations in taxation efforts in African economies can be summarized as follows: In the analyzed African countries, there is a progressive positive improvement in general literacy, which has a significant and positive impact on cross-country variations in tax revenue mobilization. This is supported by the high statistical significance of the coefficients in all the models used to assess the effect of general literacy achievements on tax revenue mobilization. These results align with the idea in the literature that better taxpayer knowledge improves compliance behavior and contributes to improved tax revenue mobilization. Mutascu and Danuletiu (2013) provided strong evidence that literacy has a significant impact on tax revenue mobilization. The results of the quantile regressions clearly show that the positive impact of the general literacy rate is greater at higher quantiles. This means that the coefficients of general literacy achievements are higher at higher quantiles, which aligns with the hypothesized claims. These results also provide empirical evidence for the argument in the literature that tax revenue is persistent over time (Gang 2008; Mascagni 2016). Overall, the findings suggest that countries with better general literacy achievements are able to mobilize more tax revenue.

Besides, most of the control variables that are statistically significant appear to have coefficients with the expected sign. For example, governance quality consistently shows a positive and statistically significant impact on revenue mobilization efforts. This finding aligns with previous research suggesting that countries with better rule of law, strong control of corruption, high regulatory quality, and strong institutional accountability are more successful in mobilizing tax revenues (Dickerson 2014; Favourate Y. Sebele-Mpofu 2020; Ricciuti, Savoia, and Sen 2019). Similarly, access to mobile telephone services has a positive and statistically significant effect on tax revenue mobilization. This is because access to mobile telephone services improves taxpayers’ access to tax-related information, thereby enhancing taxpayer compliance (Eissa and Zeitlin 2014; Okunogbe and Santoro 2023). Additionally, external trade has a significant positive effect on tax revenue mobilization, which is consistent with previous empirical results (Baunsgaard and Keen 2005; Nwosa, Saibu, and Fakunle 2012).

Table 2 presents the findings regarding the impact of general literacy on cross-country variations in direct tax revenue mobilization and indirect tax revenue mobilization respectively. The general findings that can be established from the results presented in Table 2 are as follows:

The results clearly show that general literacy has a positive and statistically significant effect on the direct tax revenue mobilization in African countries. This means that differences in the development of general literacy achievements can explain the cross-country variations in direct tax revenue mobilization. This finding is in line with the literature, which suggests that taxpayers’ literacy has a positive impact on tax revenue mobilization. For example, Adam and Webly (2012) provides empirical evidence that tax knowledge is crucial for voluntary tax compliance, especially in determining an accurate tax liability. Additionally, it is argued that taxpayer knowledge is one of the most influential factors in determining compliance behavior under the self-assessment system (Wadesango and Mwandambira 2018). Furthermore, the empirical findings support the argument in the literature that the effect of literacy on direct tax revenue mobilization is slightly higher at higher quantiles of the literacy distribution. This finding provides empirical support for the existing argument in the literature that tax revenue is persistent over time (Gang 2008; Mascagni 2016).

However, general literacy has a positive and statistically highly significant impact on indirect taxation only at the lower quantiles and median quartile levels. The effect of general literacy on indirect taxation appears to be insignificant at higher quantiles. It is important to note that literacy affects direct and indirect taxes differently. Direct taxes are expected to be paid directly by the taxpayer, who has the obligation to pay taxes on their economic activities as defined in tax laws. This means that this type of tax can be evaded or avoided by the ultimate taxpayer in cases of tax fraud. On the other hand, indirect taxes are collected by third parties or businesses on the sale of goods and services. In this case, the actual taxpayer, the consumer, cannot evade the tax on their own. This is why direct taxes are referred to as taxes on honesty.

Furthermore, most of the statistically significant control variables have the expected signs. For instance, government quality appears to have a significant positive impact on the mobilization of direct tax revenue. This means that these results provide empirical evidence for the argument in the literature that countries with better governance infrastructures can generate more tax revenue (Dickerson 2014; Favourate Y. Sebele-Mpofu 2020; Ricciuti, Savoia, and Sen 2019). Similarly access to mobile phone services appears to have a significant positive effect on direct tax revenue mobilization. Access to mobile services improves access to tax-related information and communication platforms thereby enhancing taxpayer compliance (Okunogbe and Santoro 2023). Furthermore, trade openness is found to have a significant positive impact on the mobilization of direct tax revenue. However, the effect of government budget deficit is found to be statistically significant and negative.

5. Conclusion and policy recommendations

This study aims to contribute to the literature on the association between general literacy achievements and taxation efforts by addressing the following research questions: What are the regional variations across Africa in terms of taxpayer literacy rate and taxation efforts? How does the literacy rate of taxpayers affect cross-country variation in taxation efforts when the existing level of tax-to- GDP ratio matters? The study used the gross enrollment level in secondary schools to assess the effect of general literacy on taxation efforts in African countries. The empirical evidence is based on a panel data set of 46 countries in Africa from 1996 to 2021. The econometric models employed quantile regression techniques and the generalized least square approach. I chose quantile regression estimation techniques because they are useful for identifying the effects of general literacy achievements when the existing level of taxation effort is relevant to the causal association. Based on the empirical findings, this investigation suggests that general literacy achievements influence the variation in taxation levels across countries. The empirical results strongly support the positive and significant effects of general literacy achievements on cross-country variations in taxation efforts. Therefore, it is important to understand how these findings contribute to the literature on the relationship between taxpayer knowledge and the taxation efforts of countries.

In previous literature, the positive association between taxpayer knowledge and taxation has lacked empirical validity due to the persistent nature of tax revenue. This study aims to provide empirical validity by adopting an estimation strategy that considers the persistent nature of taxation. Previous literature has primarily focused on micro-level taxpayer knowledge to assess the relationship between taxation efforts and literacy (Olowookere and Fasina 2013). However, this study’s findings demonstrate that the positive association between general literacy achievements and taxation remains true even when considering the existing level of taxation.

Consistent with previous empirical findings, this investigation contributes to the literature by examining the effects of general literacy on taxation using different indicators of taxation efforts in Africa. The findings further advance the literature by showing that the general literacy of taxpayers explains cross-country differences in taxation (Kirchler, Hoelzl, and Wahl 2008; Palil, Akir, and Ahmad 2013). The findings imply that countries with higher levels of general literacy perform better in terms of tax revenue mobilization compared to countries with lower levels of general literacy. This is because a higher level of literacy indicates that taxpayers can understand and analyze the legal and social aspects of taxation (Clifford and Amayi 2013; Normala 2007; Reckers, Sanders and Roark 1994). Additionally, higher levels of literacy suggest that taxpayers have a better understanding of taxes and are more likely to comply with tax regulations (Kirchler, Hoelzl, and Wahl 2008). In terms of policy implications, these findings are important for tax adjustments and revenue mobilization. Countries with better economic performance can increase their tax revenue by improving literacy levels. This means that African countries, many of which are poor, need to simultaneously focus on achieving quality economic growth and improving literacy.

The findings are consistent with existing evidence on literacy and tax revenue mobilization in Africa. The evidence shows that Africa has made progress in literacy in recent years, although there are significant variations at the country and regional levels. According to the Global Education Report (UNESCO, 2021), countries such as Seychelles, Algeria, Mauritius, Egypt, Kenya, Tunisia, Eswatini, Ghana, Morocco, South Africa, and Cape Verde – mostly in western, northern, and southern Africa – have achieved high educational achievements. However, countries like Niger, Burkina Faso, Mali, the Central African Republic, Ethiopia, Eritrea, Guinea, and Zambia – mostly in eastern and central Africa – have lower educational achievements. This indicates that literacy achievements are not uniform across Africa, and there are significant differences at the country level.

Similarly, there are significant regional variations in tax revenue mobilization in Africa. Southern, northern and western African countries show better tax revenue mobilization compared to the rest of the African regions. For example, the top countries in terms of tax revenue mobilization (tax-to-GDP ratio) include Tunisia, Lesotho, Namibia, Solomon Islands, Seychelles, South Africa and Morocco – mostly in the southern, northern and western African regions. On the other hand, Equatorial Guinea, Sudan, the Democratic Republic of the Congo, Gambia, Congo Republic, and the Central African Republic are at the bottom- mostly in the eastern and central African regions.

Acknowledgements

The author gratefully acknowledges the financial support from the Ethiopian Civil Service University. Furthermore, the author is indebted to the two anonymous reviewers as well as the editors of Development Studies Research for the valuable comments received on the earlier draft of the manuscript. The usual caveats apply

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by Ethiopian Civil Service University [grant number 026/2022].

Appendices

Table A1. List of countries in the sample.

Country	Country	SSE	TTR	DTR	ITR
BFA	Burkina Faso	24.40863	0.137678	0.050631	0.087047
BWA	Botswana	80.89293	0.151651	0.101498	0.050153
CAF	C. Afr. Republic	19.46103	0.073104	0.017492	0.055467
CIV	Cote D’voire	31.8965	0.121567	0.044447	0.07712
COG	Congo Rep	43.75706	0.071929	0.02276	0.049169
CPV	Cabo Verde	83.97667	0.190855	0.06287	0.127985
EGY	Egypt	81.34586	0.136084	0.061533	0.075041
ETH	Ethiopia	27.75354	0.107218	0.038475	0.068742
GAB	Gabon	39.87807	0.114113	0.035654	0.078459
GHN	Ghana	54.81711	0.112276	0.047715	0.064562
GMB	Gambia	25.39805	0.086707	0.024634	0.062057
GNQ	Equ. Guinea	44.33362	0.088004	0.077154	0.01085
LSO	Lesotho	46.89806	0.236044	0.098214	0.137828
MAR	Morocco	62.97293	0.267174	0.136445	0.13073
MDG	Madagascar	28.34643	0.097277	0.025081	0.072196
MLI	Mali	31.17339	0.139038	0.044679	0.094359
MOZ	Mozambique	21.14795	0.147732	0.058717	0.089631
MUS	Mauritius	87.41929	0.182666	0.04872	0.133946
MWI	Malawi	35.70737	0.103255	0.046673	0.056582
NAM	Namibia	62.18743	0.1993	0.125216	0.074084
SDN	Sudan	39.09811	0.053359	0.00822	0.027402
SLB	Sol. Islands	35.17613	0.21274	0.059771	0.152969
SYC	Seychelles	79.74223	0.310291	0.110031	0.200257
TGO	Togo	39.66869	0.121512	0.026673	0.094839
TUN	Tunisia	84.28659	0.260824	0.143477	0.117346
TZA	Tanzania	12.72542	0.095427	0.031853	0.064053
ZAF	South Africa	93.631	0.249538	0.148326	0.101212
ZMB	Zambia	17.79453	0.151764	0.071591	0.080173
ZWE	Zimbabwe	43.42735	0.159283	0.071538	0.085874
UGA	Uganda	10.15183	0.086412	0.023868	0.062544
SWZ	Eswatini	97.98835	0.13679	0.08889	0.047899
SEN	Senegal	32.69275	0.153895	0.047891	0.106004
RAW	Rwanda	26.39011	0.126975	0.052149	0.074826
NER	Niger	14.97176	0.098149	0.030088	0.068061
LBY	Libya	65.18843	0.037413	0.013728	0.026101
KEN	Kenya	41.49088	0.123792	0.053201	0.070592
CMR	Cameroon	41.50842	0.12119	0.045224	0.075965
COD	Congo dem. Republic	34.80838	0.068909	0.024557	0.044353
GIN	Guinea	31.32913	0.103223	0.038957	0.064393
GNB	Guinea Bissau	17.40202	0.062255	0.016471	0.043617
LBR	Liberia	32.22412	0.110288	0.03383	0.07621
MRT	Mauritania	24.66189	0.097858	0.034921	0.06286
BDI	Burundi	28.88612	0.139045	0.038312	0.100789
BEN	Benin	41.08649	0.103514	0.020877	0.083863
DZA	Algeria	83.52301	0.11316	0.041188	0.071972

Source: Own computation based on data from world development indicators and the international revenue statistics.

Table A2. Summary statistics (1996–2021).

Variable	Obs	Mean	Std. Dev.	Min	Max
TTR	1,105	0.137749	0.069993	0.005733	0.43977
DTR	1,071	0.053431	0.037806	0.001927	0.196883
ITR	1,079	0.080634	0.040607	0.003744	0.366344
SSE	1,143	43.33583	26.48407	2.72863	114.3809
INST	1,194	−.000936	2.173423	−5.96862	5.237885
IVA	1,117	26.91627	13.36686	3.243096	86.66956
GDP	1,188	4.493104	7.820707	−62.0759	149.973
FDI	1,187	4.327417	9.153414	−9.5977	161.8238
INF	1,184	12.82991	123.9003	−27.0487	4145.106
MOB	1,190	47.80609	46.83113	0	184.298
AID	1,166	8.218221	9.328793	−0.25092	92.14146
UPOP	1,193	42.31967	18.14672	7.412	89.37
GFC	1,087	21.3179	8.949663	0	61.46902
GGE	1,170	39.81442	92.99952	0	1018.327
TRD	1,189	75.24271	36.23873	17.85861	311.3541
GNB	1,184	−2.48252	7.844425	−130.796	32.502
MID	1,193	0.413803	0.492531	0	1

Table A3. Definitions of variables.

Variable	Sign	Definitions of variables (measurement)	Source
Taxation effort	TTR	Total tax revenue– to- GDP ratio (% of GDP)	World Bank
Direct Tax effort	DTR	Total direct tax revenue- to- GDP ratio (% of GDP)	GRS
Indirect Tax effort	ITR	Total Indirect tax revenue– to- GDP ratio (% of GDP)	GRS
Educational Achievement	SSE	Secondary School Enrollment (% of Gross).	World Bank
Industry Value Added	IVA	Industry (including construction), value added (% of GDP)
Governance Quality	INST	First Principal Component of Government Effectiveness, corruption control, rule of law, regulatory quality, voice and accountability, and political stability. The data for these variables are obtained from the world-wide- governance indicators	PCA
Capital formation	GFC	Measured by investment in to fixed capital facility (% of GDP)	World Bank
Foreign direct investment	FDI	Net inflow of foreign direct investment (% of GDP).	World Bank
MOBILE PHONES	MOB	Mobile phone subscriptions (per 100 people)	World Bank
Economic Growth	GDP	Rate of annual growth in domestics physical product	World Bank
Inflation	INF	Annual Consumer Price Inflation	World Bank
Foreign aid	AID	Total Development Assistance (% of GDP).	World Bank
Urban Population	UPOP	Percentage of the total population living in urban areas	World Bank
Trade	TRD	Measured by the ratio of the total trade (i.e. import plus export) (% of GDP).	World Bank
Government Expenditure	GGE	Measure as annual government expenditure (% of GDP).	IMF
Budget Deficit	GNB	Government net difference between annual revenues and annual expenditures (% of GDP).	IMF
Middle income	MID	Middle income dummy which assumes a value of 1 if the country is in the middle category as classified by the world bank, otherwise 0 value	World Bank

Table A4. Pairwise correlation matrix.

	TTR	DTR	ITR	SSE	INST	IVA	GDP	FDI	INF	MOB	AID	UPOP	GFC	GGE	TRD	GNB	MID
TTR	1
DTR	0.847*	1
ITR	0.853*	0.461*	1
SSE	0.425*	0.598*	0.268*	1
INST	0.485*	0.503*	0.496*	0.412*	1
IVA	0.009	0.076*	−0.234*	0.212*	−0.146*	1
GDP	−0.002	−0.047	0.029	−0.081*	0.009	0.003	1
FDI	0.077*	0.044	0.069*	−0.031	−0.034	0.022	0.216*	1
INF	0.105*	−0.100*	0.193	−0.117*	−0.092*	0.107*	0.021	0.001	1
MOB	0.312*	0.459*	0.166*	0.635*	0.242*	0.122*	−0.067*	0.034	−0.061*	1
AID	−0.202*	−0.287*	−0.039	−0.444*	−0.132*	−0.377*	0.040	0.159*	0.001	−0.329*	1
UPOP	0.149*	0.287*	−0.052	0.597*	0.087*	0.556*	−0.064*	0.071*	−0.012	0.474*	−0.446*	1
GFC	0.169*	0.207*	0.099*	0.209*	0.259*	0.273*	0.173*	0.232*	−0.094*	0.288*	−0.173*	0.242*	1
GGE	0.091*	−0.022	−0.001	−0.061*	−0.181*	0.089*	0.037	0.158*	−0.001	−0.060*	−0.067*	0.109*	−0.009	1
TRD	0.362*	0.324*	0.292*	0.357*	0.227*	0.333*	0.035	0.329*	0.053	0.274*	−0.028	0.411*	0.289*	0.101*	1
GNB	−0.013	−0.050	0.0083	−0.144*	0.018	0.189*	0.155*	0.049	0.032	−0.165*	0.006	−0.005	0.023	−0.015	0.059*	1
MID	0.415*	0.502*	0.227*	0.685*	0.411*	0.282*	−0.066*	−0.081*	−0.045	0.464*	−0.384*	0.545*	0.222*	0.035	0.291*	−0.068*	1

Note: * indicates significance at ρ < 0.05. Bootstrap Standard errors are presented in brackets.