قطب بندی، عدم اطمینان و شکست سرمایه گذاری عمومی

In a political economy model, the effect of political polarisation on a government's intertemporal choice between redistribution and public investment is shown to be similar to the effect of political uncertainty. Moreover, polarisation and uncertainty reinforce one another in their impact on public underinvestment and may ultimately lead to no investment at all.

Political polarisation captures the phenomenon that societies are not homogeneous, but made up of different groups, which are often competing for political influence. Such heterogeneity can be based on ethnic, linguistic, religious diversity, income distribution and/or other social characteristics. Abstracting from the specific type of heterogeneity I model polarisation as the degree, to which the government favours one group over another, i.e. the degree, to which the heterogeneity actually matters for decision making. In a political economy model I can show that polarisation has similar effects as political uncertainty (government's chance of losing power) in producing public underinvestment. Probably the most important result is that polarisation and political uncertainty reinforce one another in their effect on underinvestment. Ultimately, this may lead to total public investment failure. This finding helps explain, for instance, that many developing countries, especially in sub-Saharan Africa, suffer from poor public services in health, education, infrastructure, etc. despite large public expenditure (Pradhan, 1996). The effects of political uncertainty (sometimes called political instability) are well documented in the model-theoretic literature. Most papers are two-period models. The chance of a change of decision maker who might take different, less desirable, decisions in the second period produces a negative spill-over onto the incumbent decision maker in the first period. Future outcomes and the effects of today's actions onto the future are more heavily discounted. In Cukierman et al. (1992), for instance, the result is higher seigniorage revenue today instead of economic reform effective tomorrow, in Devereux and Wen (1998), it is higher public activity and lower growth. Tabellini and Alesina's (1990) median voter typically chooses to borrow from the future to pay for higher public goods spending today.1 However, none of these models disentangles the effects of polarisation from those of political uncertainty, even though agents have preferences over two types of public goods. In contrast to political uncertainty, polarisation has only been scrutinised under specific circumstances. On the one hand, there is a lot of empirical evidence documenting the key role of political polarisation for growth collapse and development failures.2Easterly and Levine (1997) find that, empirically, ethnic diversity is the single most important cause of slow growth in Africa.3Collier and Gunning (1999) argue that ethnic diversity promotes ethnic favouritism, i.e. diverting public spending to ethnic groups instead of creating better conditions for the whole of society.4 On the other hand, there is the literature on clientelism.5Robinson and Verdier (2002) analyse clientelism in a game-theoretic framework with an asymmetric political structure. It is optimal for patrons to invest too little in public services while overproviding public employment, which they demonstrate to be an “incentive-compatible way for patrons to control clients.”6 This paper acknowledges that political uncertainty (chance of losing power) and polarisation (heterogeneity) are distinct empirical phenomena relevant for both developing and more developed countries. The political economy model I present abstracts from specific forms of polarisation or political uncertainty and captures directly the trade-off between redistribution and public investment. In an intertemporal public finance model with two rivaling groups, a government can choose between (efficient) public investment and spending on public goods, one of which is only beneficial to one group, the other is only useful to the other group. There is more or less redistribution depending on how much the government prefers one public good over the other (polarisation). However, the incumbent faces an exogenous chance of losing power to the opposition with different preferences (political uncertainty).7 The new insight I derive is that polarisation and uncertainty are mutually dependent in their effect on public investment. The reason is that it is perfectly rational for a government to invest less in period 1 when beneficial effects of its policies may not fully accrue to it in the future. The incumbent government is effectively discounting the future more heavily in the presence of polarisation and uncertainty. If the alternative government is known to favour the other group (polarisation), political uncertainty will cause underinvestment because public investment will be used for achieving other objectives in case of government change. However, in the extreme case of no polarisation at all, political uncertainty has no effect because the other government would pursue identical policies in the future anyway. Similarly, if the government stays in power with certainty, it is irrelevant, if there is a lot of polarisation or not. Furthermore, it is shown that polarisation and uncertainty reinforce one another in their effect on public underinvestment. Typically, there is a threshold, beyond which the government does not want to invest at all. Reducing polarisation and political uncertainty above the threshold and thereby effectively discounting the future less severely leads to a strong increase in public investment at first. However, the additional investment increments (for less and less polarisation and uncertainty) become smaller because marginal investment profitability goes down. Section 2 presents the intertemporal framework of the theoretical model. In Section 3, the government's maximisation problem is simplified by aggregating polarisation and uncertainty into a composite measure of political instability. Section 4 presents the interior solution for aggregate political instability. Section 5 discusses the danger of public investment failure by emphasising how polarisation and uncertainty reinforce one another and by presenting the corner solution of no public investment at all. Section 6 concludes.

This paper captures a government's decision problem between efficient public investment and redistribution in a parsimonious model of political instability. The chance of another government being in power and taking undesirable decisions in the future produces a negative spill-over onto those in power today. In this paper, it is actually optimal for the current government to totally refrain from spending on public investment, if there is heavy discounting produced by political instability. As political stability increases (and discounting decreases), a threshold is reached, beyond which it is optimal to increase investment with marginal investment being strong at first. Then, however, the additional investment increments become smaller, because marginal investment profitability goes down. The theoretical model captures the impact of two distinct causes of political instability as they appear in reality: polarisation (rivaling groups in society) and political uncertainty about the future government. Probably the most important result is that their effect on a government's effective discounting is multiplicative. Underinvestment in infrastructure, anti-corruption measures, health, or education is particularly severe in a country with a legacy of group-specific policies and numerous government changes. Therefore, the paper offers an explanation for appalling levels of efficient public investment in sub-Saharan African countries, which are troubled by ethnic strife combined with a history of coup d'états or revolutions. But the model may also help explain inefficient government sectors in unstable Western democracies such as France (during the Forth Republic) or Italy (before the break-up of the traditional party system in 1992).Nonetheless, for several reasons the modeling approach seems to be particularly relevant for certain low-income countries. First, political uncertainty in some of these countries is inherent to the political structure of the country rather than caused by electoral uncertainty as in Western democracies. Capturing political uncertainty by an exogenous parameter may, therefore, be deemed more appropriate for such countries. Second, ethnic diversity is a real additional problem in many low-income countries. This means that polarisation is a much more clear-cut and severe problem than in Western democracies where the electoral process entices the parties to compete for the votes in the centre. Third, the disregard for private sector decisions on labour, consumption and investment would certainly not be suitable simplifications for industrialised countries, but may be seen as a first approximation in some low-income countries, where there is either no economic growth or it depends on external factors (like foreign direct investment). Relaxing some of the assumptions of this paper may offer scope for future research. It may be worth while exploring if there are any trade-off effects when other sources of government revenue are included. Another extension would be to model the effect of public investment on growth, when the private sector optimises its investment and consumption decisions. A further extension might be to capture the interdependence between growth and political instability. So far, political instability was endogenised in the literature by modeling voting10 or by modeling insurgencies either in response to the relative military strengths of government and rebels (Grossman, 1999) or dependent upon “rent-seeking economic reforms” (McBride, 2005). However and particularly relevant for capturing the government decision problem in low-income countries: better economic conditions directly affect the chance of a coup d'état or revolution. For these countries, it seems appropriate, therefore, to capture the feedback of government actions on growth and model its interdependence with political instability. It would be interesting to explore the link to the literature on endogenous growth and rent-seeking such as Ehrlich and Lui (1999) or Park et al. (2005).