2. Fiscal capacity and its measurement

Fiscal capacity is defined as the capability of a fiscal system to raise tax revenues from a broad tax base (Besley and Persson Reference Besley and Persson2011). This concept has often been proxied in a cross-section of countries as the tax-to-GDP ratio or similar tax effort indicators. Slightly more refined measures are income taxes and nontrade taxes expressed as shares of total taxes, the income tax bias (the difference between income and trade taxes) and the formal sector share, which is inversely related with the ability of the government to raise taxes. These alternatives are based on the observation that income is more difficult to tax than goods, and therefore it requires a more structured administration.

Tax-to-GDP is readily available for many countries, hence its popularity. However, it also poses a number of problems. First, it strongly depends on the political preferences of a polity towards the size of the public sector, and the scope of redistribution, especially if we compare similar countries (Hanson and Sigman Reference Hanson and Sigman2013; Lieberman Reference Lieberman2002, Martin and Prasad Reference Martin and Prasad2014).Footnote ⁴ Second, consider two countries with the same tax-to-GDP ratio. They can afford that level in very different ways. One country could tend to expropriate its citizens, imposing a high administrative burden and giving them few or no rights to appeal; and, once revenues have been raised, it could be inefficient in transferring the money to the spending ministers who provide public services. If one country has these features that another one does not have, even if they have the same tax-to-GDP ratio, their fiscal capacity is arguably different. Other tax-effort based indicators do not provide better measures of fiscal capacity either. A higher share of income taxes in total taxes may simply reflect a culture of tax compliance (that is, a lack of tax evasion) in the country and does not tell us anything about the efficiency and effectiveness with which taxes are raised, and on the power that taxpayers have with respect to the revenue office.

More importantly, from our perspective, outcome-based measures of fiscal capacity, such as the tax-to-GDP ratio, cannot differentiate between two quite different dimensions of fiscal systems related to the exercise of taxation powers. One has to do with their effectiveness in raising tax revenues, i.e. the ability to coerce citizens to pay taxes. We call this the effectiveness dimension. The other has to do with the fairness of the exercise of taxation powers: the ability of tax systems to make the state accountable and transparent to its citizens. We call this the impartiality dimension.Footnote ⁵

In this paper, we use six indicators selected from the Public Expenditure and Financial Accountability (PEFA 2006) Program database, which provide a clear way of differentiating between the impartiality and effectiveness of tax systems. They are described below:Footnote ⁶

1. 1. Transparency of taxpayer obligations and liabilities: evaluating taxpayers’ access to information on tax liabilities and administrative procedures;

2. 2. Tax appeals: assessing the existence and functioning of a tax appeals mechanism;

3. 3. Controls in the taxpayer registration system: assessing the quality and maintenance of a taxpayer database;

4. 4. Effectiveness of penalties for non-compliance: addressing failures in registration and tax declaration obligations that assess whether penalties for all areas of non-compliance are set sufficiently high to act as a deterrence and are consistently administered;

5. 5. Quality of tax audit: assessing the planning and monitoring of tax audit and fraud investigation programmes;

6. 6. Effectiveness in collection of tax payments: looking at the frequency of complete accounts reconciliation between tax assessments, collections, arrears records and receipts by the Treasury.

The first two indicators capture the impartiality aspect of fiscal capacity, since they hinge on the relationship between the state and the public: empowering it against the taxation powers of the former or making such power clearly defined and not subject to discretion. The third one can also be seen as a measure of impartiality, because it is related to the ability of the tax system to make taxation transparent to the citizens. But its description also indicates that there is a significant element of coerciveness. So we preferred to include it in the set of effectiveness measures, to avoid criticism that such classification is ad hoc. The last three measures assess the coercive aspects of the tax system: they are all desirable features of a tax machine that aims to raise revenues.Footnote ⁷ Table 1 provides descriptive statistics for the key variables of interest. Higher scores indicate greater levels of fiscal capacity: both impartiality and effectiveness.

Table 1.

PEFA measures of fiscal capacity

Source: PEFA (2006), our calculations.

How correlated are the six PEFA measures of fiscal capacity with the more conventional measure of fiscal capacity – that is, tax revenues as a percentage of GDP? In Figures 1a–1f, we present scatter plots of the six measures against non-resource tax revenues/GDP for our sample of countries.Footnote ⁸ We find a positive relationship between the six PEFA measures capturing the impartiality and effectiveness of the tax system and tax revenue mobilisation. In particular, the positive correlation between the first two PEFA measures and tax revenue mobilisation suggests that how a developing country does in the impartiality dimension is a good predictor of its government's ability to raise tax revenues. Previous work by political scientists and fiscal sociologists on successful examples of tax reforms in developing countries also supports this point (see Brautigam et al. Reference Brautigam, Fjeldstad and Moore2008). In the next section, we discuss the political determinants of the impartiality and effectiveness dimensions of a taxation system.

Figure 1a.

Transparency of taxpayer obligations and liabilities and tax revenues/GDP

Data are from PEFA (2006) and ICTD (2015). R-squared = 0.15.

Figure 1b.

Quality of tax appeals system and tax revenues/GDP

Data are from PEFA (2006) and ICTD (2015). R-squared = 0.19.

Figure 1c.

Quality of the taxpayer registration system and tax revenues/GDP

Data are from PEFA (2006) and ICTD (2015). R-squared = 0.37.

Figure 1d.

Effectiveness of penalties for non-compliance and tax revenues/GDP

Data are from PEFA (2006) and ICTD (2015). R-squared = 0.38.

Figure 1e.

Quality of tax audit and tax revenues/GDP

Data are from PEFA (2006) and ICTD (2015). R2= 0.07.

Figure 1f.

Effectiveness in collection of tax payments and tax revenues/GDP

Data are from PEFA (2006) and ICTD (2015). R2 = 0.05.

3. The political determinants of fiscal capacity

In the recent literature on the political economy of development, fiscal capacity is seen as a ‘pillar’ of economic development as the expansion of the tax base allows states to invest in the public goods essential for economic development (Besley and Persson Reference Besley and Persson2011; Dincecco Reference Dincecco2017). Cohesive political institutions, seen as a stronger system of checks and balances on the executive, are believed to be one key ingredient to improve tax systems, so developing infrastructures that can raise taxes from a broad base. Where subject to effective checks and balances, incumbents will tend to promote common interests rather than using the state to retain power or redistribute to their own cronies (Besley and Persson Reference Besley and Persson2009). Thus it follows from this literature that placing limitations on the executive power is an essential condition for developing fiscally (and legally) capable states. However, this literature does not differentiate between different aspects of a taxation system, and in particular, between the impartiality and effectiveness of a taxation system: this is important to understand how constraints on the executive affect the ability to raise revenues. We argue below that the causal effect that the degree of constraints on the executive may have on fiscal capacity may differ, depending on whether the effect is on the effectiveness dimension of fiscal capacity or on the impartiality dimension of fiscal capacity. In particular, we argue that constraints on the executive are a significant determinant of the impartiality dimension of fiscal capacity and are likely to have a positive effect, while they are not a significant determinant of the effectiveness dimension of fiscal capacity.

We first discuss the relationship between the nature of constraints on the executive and the effectiveness of taxation systems. To start with, consider two types of autocrats: the first is the autocrat as a ‘stationary bandit (that) has an encompassing interest in the territory he controls and accordingly provides domestic order and other public goods’ (Olson Reference Olson1993: 569). The second type of autocrat is a ‘roving bandit’ with a sufficiently short time horizon, so it would be ‘[in] his interest to confiscate the property of his subjects, to abrogate any contracts he has signed in borrowing money from them, and generally to ignore the long-run economic consequences of his choices’ (Olson Reference Olson1993: 572). The first autocrat, with a long time horizon, is a leader with a long-term vision and a commitment to enact institutional reforms and policies that are likely to be growth enhancing (such as Deng Xiaoping in China). The second type of leader has a short-term vision (perhaps because he is in an unstable political environment where he may lose power), and engages in high levels of predation (such as Mohammed Farah Aidid and Ali Mahdi Mohamed in Somalia in the 1990s). The autocrat with a long time horizon – the ‘stationary bandit’ – who is interested in staying in power, as well as maximizing long-term income to mobilise tax revenues both to provide public goods to his own citizens and to extract some of the revenues for his personal benefit, has a strong incentive to invest in the effectiveness of the taxation system. By doing so, the autocrat can maximise tax revenues for a given tax rate. On the other hand, the autocrat with a short time horizon – the ‘roving bandit’ – has little interest in investing in building up the effectiveness of the taxation system, when the returns to the investment are possibly far into the future, and not within the timespan that he considers to be of relevance. Similarly, in democracies, the median voter hypothesis suggests that there is an incentive for rulers to invest in effective tax systems, so that the party in power can provide the public goods necessary for re-election. At the same time, in democracies, ‘the tenure, terms and time horizons of democratic political leaders are even shorter than those of the typical autocrat, and the democracies lose a great deal of efficiency because of this’ (Olson Reference Olson1993: 572).

This suggests that when it comes to the effectiveness dimension of fiscal capacity, there is no reason to expect that executives with limited checks on their authority will behave differently from executives with significant constraints on their power with respect to making taxation systems effective in collecting more revenues for the state. Therefore, the degree of constraints on the executive is not a significant determinant of tax system effectiveness, with the relationship between constraints on the executive and the effectiveness dimension of fiscal capacity likely to go either way.

What about the relationship between the degree of constraints on the executive and the impartiality dimension of taxation systems? Here, we may expect authoritarian regimes to behave differently from democratic systems. Fairness in taxation systems may be seen as part of a ‘fiscal contract’ between the state and its citizens (Brautigam et al. Reference Brautigam, Fjeldstad and Moore2008). Transparency and accountability of taxation systems are about state–society relations, involving an exchange of tax revenues for services.Footnote ⁹ Creating mechanisms of accountability and placing constraints on rulers facilitate the existence of a fiscal bargain, at the heart of the relationship between citizens and rulers. According to Levi (Reference Levi1988), it should reduce the transaction costs of taxes by making compliance ‘quasi voluntary’ and by building ‘tax morale’ (Doerrenberg and Peichl Reference Doerrenberg and Peichl2013; Luttmer and Singhal Reference Luttmer and Singhal2014). Citizens should be more willing to enter into a fiscal contract with the state, as they have more control over its actions and greater belief in its legitimacy (Bates and Lien Reference Bates and Lien Da-Hsiang1985). Accountability and responsibility processes in tax systems ‘engage taxpayer-citizens collectively in politics and leads them to make claims on government for reciprocity, either through short-term conflict or long-term increases in political engagement’ (Prichard Reference Prichard2010: 13). Such processes are more likely to emerge in cohesive political systems where there are significant constraints on the power of the executive, and where politicians have an incentive to signal the legitimacy of the state by making the tax system more transparent and non-discriminatory (Cheibub Reference Cheibub1998). Furthermore, transparency in taxation systems is more likely to emerge in regimes where political elites are more constrained in their powers to evade taxes or bend tax rules in their favour (while in regimes with limited checks on the executive, elites face little constraints in avoiding taxes or in devising a non-transparent tax system that discriminates in their favour).

Lesotho and Rwanda are an interesting comparison among African countries. They are similar in many respects (both are small and landlocked, have analogous per capita GDP levels and comparable functional shares of GDP, and both have arranged semi-autonomous revenue authorities), but are somehow different in their openness, with Lesotho much more reliant than Rwanda on international trade. They differ from a political point of view: according to the Freedom in the World Report by Freedom House, Lesotho has been ‘partly free’ since 1992, with periods during which it was rated ‘free’ (e.g. the years 2002–8), whereas Rwanda has been consistently ‘not-free’. The two fiscal systems are quite different: Rwanda relies on direct and indirect taxes, whereas Lesotho mainly exploits trade taxes. Rwanda stresses effectiveness in revenue collection, while Lesotho does not attempt to chase the large unregistered economy, because voters can revolt in the ballot box against what is perceived as excessive taxation on those who live at the subsistence level. Yet the more consensual Lesotho has a larger tax effort ratio than Rwanda (D'Arcy Reference D'Arcy2012).Footnote ¹⁰ Historically, the example of the introduction of direct income tax in the United Kingdom in 1799 provides a clear case of political institutions in determining the impartiality dimension of fiscal capacity, particularly around the provision of tax information and a more accountable tax system. As North and Weingast (Reference North and Weingast1989) argue, after the Glorious Revolution of 1688, there was a shift in political power from the Crown to the Parliament where there was a significant increase in the constraints on the executive – as they note, ‘it was now the King in Parliament, not the King alone’ (p. 816). As Levi (Reference Levi1988: 144) notes:

This example shows that the greater constraints on the executive in post-Restoration British history played a causal role in the increase in the impartiality dimension of fiscal capacity, which in turn allowed for the passage of the income tax in 1799. Thus, the Besley–Persson argument on the role of cohesive political systems in building the state's fiscal capacity applies more to the impartiality dimension of fiscal capacity rather than to its effectiveness dimension. We can restate our argument on the differing role that political institutions play in augmenting the various dimensions of fiscal capacity, by means of two propositions:

In the next section, we propose an empirical strategy to test the above hypotheses using cross-national, cross-sectional data.

4. Econometric methods and identification

Since the objective of the paper is to look at the structural conditions under which countries develop capable states, regressions based on cross-section averages are a suitable approach as they test relationships whose mechanisms have long-run characteristics. This is an established methodological choice in econometric research on institutions (prominent examples are Acemoglu et al. Reference Acemoglu, Johnson and Robinson2001; Acemoglu et al. Reference Acemoglu, Johnson, Robinson and Thaicharoen2003) and implies explaining the dependent variable using long-term averages of the explanatory variables to capture structural cross-country variation.Footnote ¹¹ Hence, the regression specification takes the form:

(1) $$\begin{equation} F{C_{i,T,T - 1}} = {\beta _0} + \beta \cdot {W_{i,t,t - 1}} + {X'_{i,t,t - 1}} \cdot \phi + {\varepsilon _{i,t,t - 1}} \end{equation}$$

where, FC_{i,T,T −1} captures the quality of current fiscal institutions as the average of a given dimension of fiscal capacity of interest for country i between the end of the sample period, T, and T-1, captured here by the six PEFA indicators. Besley and Persson (Reference Besley and Persson2011) suggest that fiscal and legal capacities have common determinants and that investing in one dimension of state capacity simultaneously reinforces the other, i.e. there are complementarities. By extension, we apply this hypothesis to the different dimensions of fiscal capacity.

On the right-hand side, W_{i,t,t −1} is the determinant of fiscal capacity of interest, averaged between times t and t-1, with t < T-1, and β represents its long-run effect on fiscal capacity. It is measured as the average value of Constraints on the executive from the Polity IV dataset from 1965 (or independence year, if later) up to 2004 (Marshall et al. Reference Marshall, Jaggers and Gurr2011), since we are interested in the long-run component of these constraints (and not in the annual fluctuations). But taking this variable as a 2000–4 average, as an alternative measure, delivers similar results (available on request). This variable measures the extent of constitutional limits on the exercise of arbitrary power by the executive, i.e. whether the executive power is subject to institutionalised checks and balances (on a scale from 1 to 7, where 1 indicates unlimited authority of the chief executive and 7 indicates executive parity or subordination, with intermediate values indicating moderate to substantial power limitations). Similarly, X_{i,t,t −1} is a set of controls (described and discussed in the results section). Finally, ε_{i,t,t −1} is the error, capturing all other omitted factors.Footnote ¹²

Figure 2 provides some preliminary evidence on the relationship between the degree of Constraints on the executive and the different dimensions of fiscal capacity. There seems to be a positive correlation of constraints on the executive with both impartiality measures and effectiveness measures. However, while useful to illustrate the behaviour of key variables, one should not yet be tempted to give any causal interpretation to such correlations.

Figure 2a.

Transparency of taxpayer obligations and liabilities and Constraints on the executive

Source: PEFA (2006) and Polity IV (Marshall et al. 2011).

Figure 2b.

Quality of tax appeals mechanism and Constraints on the executive

Source: PEFA (2006) and Polity IV (Marshall et al. 2011).

Figure 2c.

Quality of the taxpayer registration system and Constraints on the executive

Source: PEFA (2006) and Polity IV (Marshall et al. 2011).

Figure 2d.

Effectiveness of penalties for non-compliance and Constraints on the executive

Source: PEFA (2006) and Polity IV (Marshall et al. 2011).

Figure 2e.

Quality of tax audit and Constraints on the executive

Source: PEFA (2006) and Polity IV (Marshall et al. 2011).

Figure 2f.

Effectiveness in collection of tax payments and Constraints on the executive

Source: PEFA (2006) and Polity IV (Marshall et al. 2011).

Before estimating (1), we should first discuss whether estimating the impact of political institutions is subject to identification problems. Although there are good reasons to expect a causal relationship between rulers’ accountability and fiscal capacity development, OLS estimates are insufficient to document such a relationship. Building a political system is clearly an endogenous process, driven by a variety of social forces, including state actors. One such possibility is that taxation may lead to increased political representation (e.g. Moore Reference Moore2007; Timmons Reference Timmons2010). When estimating the relationship from the data, the effect of constraints on the executive could then be affected by reverse causality, hence subject to bias in OLS regressions. A concern is also that the effect of political systems may be endogenous in the statistical sense, namely correlated with the regression disturbances because of measurement error. Therefore, one might expect the coefficients on Constraints on the executive to be biased both away from zero and towards zero. The magnitude of the two types of bias, and their combined effect, is an open question, but here we attempt to address the problem using an instrumental variable approach, presenting estimates from different methods.

Our instrument has a prominent place in the literature: historical settler mortality, as captured by the (log of) mortality rate due to the disease environment at the time of colonisation. Acemoglu et al. (Reference Acemoglu, Johnson and Robinson2001) documented that such a variable picks the exogenous variation in the type of institutions built in the former European colonies. Where European colonisers settled en masse, life was organised around inclusive institutions, i.e. subjecting the ruling elite to binding limitations to their power. Where they could not settle, due to adverse sanitary conditions, institutions were extractive, i.e. subject to little or no constraint on the rulers. Compared to its alternatives, this instrument had a more plausible justification (see Acemoglu Reference Acemoglu2005). Perhaps for this reason, it has survived criticism, which has been made on the grounds of data quality and accuracy (e.g. Albouy Reference Albouy2012). Since it was proposed, it has been exploited successfully to identify the effect of the Constraints on the executive variable (e.g. Acemoglu et al. Reference Acemoglu, Johnson and Robinson2001; Acemoglu et al. Reference Acemoglu, Johnson, Robinson and Thaicharoen2003). Of course, the exclusion restriction that the instrument does not affect the second stage left-hand-side is always one of the most vulnerable parts of any IV identification strategy. So, while we rely on Acemoglu et al.’s (Reference Acemoglu, Johnson and Robinson2001) intuition on the plausibility that settlers’ mortality does not directly affect level of fiscal capacity (other than through its effect on Constraints on the executive), we also address exclusion restriction concerns through econometric testing.