Natural Resources, Institutions and Domestic Revenue Mobilization: From Global to Local Evidence

Abstract

The quest to mobilize sustainable domestic revenues for financing development is a challenge for many countries, especially those in the developing world. Discovering non-renewable natural resources should therefore be good news. However, the notion of a natural resource curse suggests that countries that are rich in natural resources often perform poorly on several development outcomes. A related term, fiscal resource curse, has emerged. The fiscal resource curse is the inability of countries to raise taxes from a broad base in the presence of nonrenewable natural resources. In other words, natural resource revenues displace non-resource tax effort (i.e. mobilization of tax revenues outside the natural resource sector). We take advantage of a new Government Revenue Database (GRD) to examine three key research questions on the relationship among natural resources, institutions and domestic revenue mobilization. For each question, we also examine whether context matters by examining the relationships for the full global sample as well as for developing countries. First, we examine the validity of a variant of the fiscal resource curse using a more comprehensive dataset which covers not only hydrocarbon revenues but also minerals. We employ a novel instrumental-variable strategy to estimate the causal effect of resource revenues on non-resource tax effort by exploiting the so-called “China shock”. Following her accession to the World Trade Organization in 2001, China’s engagement in the non-renewable resource trade has increased several fold, driving up commodity prices and raising resource revenues among exporting countries. China’s resource trade model with developing countries has been characterized by resource-for-infrastructure deals, which means that exporting countries also benefit from infrastructure projects rather than just liquid capital flows from the exports of natural resouces. We exploit this exogenous variation in China’s non-renewable resource trade to examine the causal effect of resource revenues on non-resource tax effect using a Two-stage Least Squares Approach. We do not find consistent evidence of a negative relationship between resource revenues and non-resource taxes. On the contrary, we find that, once we account for China’s role in the global non-renewable resource trade, a percentage point increase in resource revenues as a percentage of GDP leads to about a 0.3 percentage point increase in non-resource taxes as a percentage of GDP. China’s provisioning of energy and transport infrastructure in developing countries in lieu of export revenues might be easing the binding constraints to expanding the non-resource sector. The latter becomes the basis for increasing output in the sector and therefore increasing non-resource tax revenues. Second, the view of the New Institutionalists School is that institutions matter in explaining development outcomes. North (1990) argues that institutions shape the efficiency of markets. Others maintain the view that the quality of institutions shape fiscal policy, even in the presence of natural resources (for example, Masi, Savoia, & Sen, 2018; Botlhole, Asafu-Adjaye, & Carmignani, 2012). Evidence on the mediating role of institutions in improving tax revenues outside the natural resource sector has however been scanty. We investigate whether the type and quality of institutions can mediate the effect of natural resource rents on non-resource tax effort. We begin by proposing a simple theoretical model of social welfare maximization using optimization techniques. The model is characterized by a social planner, who allocates different levels of effort in mobilizing revenues from the natural resource sector and the non-resource sector. The planner determines the allocation of benefits to two main groups in the society: elites and non-elites. These choices are made subject to the total effort and revenue possible to maximize total welfare within the society. A key model prediction is that institutions that focus more on redistribution have a weak moderating influence on the negative effect of natural resource rents on non-resource tax effort, especially when commodity prices are high. Giving that the model is limited to looking at the role of only one type of institution, we empirically examine whether the prediction of the model is consistent by interacting natural resource rents with 12 different measures of institutions commonly used in the literature. Using Fixed-effects Estimators and a Generalized Method of Moment Estimator, we compute marginal effects of natural resource rents on non-resource tax revenues. We do not find a statistically significant impact for a majority of the institutional variables. However, we find a weak evidence of a mitigating effect for two key variables: constraints on executive and the quality of a democracy. The weak impact of these two mediating variables is premised on the fact that, while they play a mitigating role, they are unable to completely undo the negative impact of natural resource rents on non-resource tax revenues. Other covariates that we find important for determining non-resource tax revenues are the level of GDP per capita in an economy, the size of the agricultural sector and the extent to which there is control over corruption. Finally, following the literature on the differential impact of different types of natural resources on different development outcomes, we distinguish between hydrocarbon resources and nonhydrocarbon resources (metallic and non-metallic minerals) in assessing how they differentially impact on non-resource tax effort. In other words, we test the hypothesis that hydrocarbon-rich countries perform worse in non-resource tax effort than mineral-rich countries. We present a simple theoretical framework of how the exploitation of different types of natural resources yields varying incentives for non-resource tax effort by extending the theoretical framework of Knack (2009). Our theoretical model suggests that hydrocarbon-rich countries perform worse in non-resource tax effort than mineral-rich countries. Discovery of hydrocarbon resources and its exploitation is more likely to be concentrated in the hands of the elite few. Revenues are therefore prone to being expended through rent-seeking behaviour, generous tax incentives and lack of investment in tax capacity. These undermine non-resource tax effort in hydrocarbon-rich contexts compared to mineral-rich contexts where mineral occurrence is likely to be more widespread and therefore control over its exploitation is likely to be less centralized. We confirm the theoretical results using alternative data sources and alternative panel econometric techniques. In a sample of over 80 countries over the period 1980 to 2015, we find that hydrocarbon rents consistently exert a negative effect on non-resource tax effort. We do not find such consistent evidence for mineral rents nor for mineral-producing countries. This evidence is consistent at the global level and in developing countries. Our results suggest that government domestic revenue mobilization policy must take into account both local and external factors that contribute to expanding the non-resource sector, reducing informality and diversifying the economy. These factors include providing the relevant development infrastructure (for example energy and transport), improving tax administration and strengthening institutions of state. The quality of democracy, constraints on executive power and control of corruption are some key institutional factors that would be relevant for improving non-resource tax effort. Given that the risks of a lower non-resource tax effort are higher in hydrocarbon-rich countries, the factors described above should be given greater attention there.