It is imperative to highlight the political economy incentives that may face politicians in countries, such as Guyana, with accumulated revenues – whether or not these revenues are formally a “fund.” To set up institutional solutions to make it more likely that a beneficial spending path will be followed, it is necessary to understand the incentives that make it unlikely in the first place.
Why do decision-makers tend to be impatient – why are they so likely to spend too much of the money that is available instead of saving it for harder times and for future generations?
The main analysis should be concentrated on the nature of interest-group politics with power rivalry. Most of the policies involves a struggle between competing interests, and it can be shown in what follows that the competition for power among different groups can create perverse incentives to overspend in early time periods (with less left for later), even when there exists a more stable time path of spending that all groups would prefer. This main case can extend the basic analysis to discuss how overspending can be affected by the following possibilities.
- Politicians may gain political support through higher public expenditures.
- Conversely, an informed citizenry might punish politicians who imprudently overspend.
- Increased public spending may be privately valuable for the politician (independently of any political support it buys them).
- Accumulated revenues may provide an incentive to undermine the “rules of the game” embodied in the country’s governing institutions.
Interest-Group Politics with Power Rivalry
Political competition often consists of the contest between representatives of different interest groups to control the public purse and use it to the benefit of their own group. The nature of the different groups varies across different countries. The dividing lines can be based on class, ideology, ethnicity, language, religion, or other traits. The severity of the divisions, and the extent to which political competition follows interest-group lines, are also different in every country.
The degree to which politicians compete for the opportunity to favour their own group at the expense of others will affect the fiscal policies of the state. Dixie et al. (2000) show how the division of the budget in an electoral period depends on the nature of the political competition, in particular, the likelihood of a change in government in the next period.
When the sitting government decides how much of the budget to allocate to its own group, it must take into account how the opposition will behave as a result if they take power next. A similar dynamic can affect the total amount of spending of natural resource-generated wealth. A government may ideally want to spread spending evenly across all periods, but the possibility that another group takes power in the next period, and one’s own group loses out, as a result, creates an incentive to spend too much now. The incumbent government can choose its own preferred spending patterns with today’s expenditures, whereas if another group takes power tomorrow, it will allocate tomorrow’s expenditure according to its preference. This makes it preferable for an incumbent government to shift expenditure from the future to today.
Once the aggregate amount of spending for a period is set, the policymaker gives a fixed share to each group. The policy maker’s own group gets the largest share, and the share increases the more biased the policymaker is. Moreover, the more slowly diminishing marginal benefit sets in, the more unequal is the allocation. This suggests that if marginal utility diminishes more slowly in poorer societies, then poverty will be associated with deeper disagreements.
When the incumbent knows that they will be in power in both periods, they choose a time path of aggregate spending that is efficient in the sense that there is no allocation that can allow anyone group to do better without making the other group worse off. Furthermore, it is optimal for the incumbent in the sense that there is no other allocation across groups and over time that would make them better off, even at the cost of making the other policymaker worse off.
One implication of this is that the policy is time consistent – it does not matter whether the incumbent makes their decisions regarding both periods right away, or decides on the allocations one period at a time. The sequentially decided policy is the same they would choose if they could commit to the entire time path of spending allocations (i.e., for both periods) at the beginning of period 1 (first term). This is not true for all types of economic policy choices, but it is true here: with correct anticipation of their own choices in the second period, the staggering of the decision-making process has no impact on the policy maker’s choices. Some aspects of fiscal and monetary policy that rely on responses by the private sector or other actors may generate inconsistency over time (Dixie and Londregan 1995; Garfinkel and Lee 2000; Kydland and Prescott 1977).
The less stable the government – in the sense that there is a higher likelihood of an imminent change in government – the stronger the incentive for spending a lot today. Period 1 (first term) spending is higher the larger is the probability that they will be turned out of office, which shows that as the policy maker‘s power becomes more precarious, they are tempted to spend more.
The deeper is the division among the groups – the more pronounced is the tendency of politicians to favour only their own group at the expense of others – the greater will be the incentive to spend too much while one is in power.
Here it also shows that these last two effects – the instability effect and the conflict effect – reinforce each other. Instability has a more adverse effect in more divided societies.
Disagreement and instability, then, cause overspending and inefficient fiscal policy. This should not be interpreted as a claim that dictatorships are more efficient than democracies; rather, what matters is the horizon of regimes. Governments or rulers who expect to be in power for a long time have an incentive (out of their self-interest) to smooth spending over time.
This incentive disappears when power is precarious, for example, because of frequent coups or revolutions (as is in fact empirically more likely in nondemocratic systems). As can be exhibited, electoral politics can modify this finding in several ways. Moreover, if in democracies, government changes do not produce as large shifts in spending allocations, the incentive to overspend is weaker in them.
To elaborate, there are at least three reasons why Findings 3 and 4 do not imply that democracies have less efficient spending patterns than dictatorship. First, as is suggested, dictatorships may well have a higher probability that power changes. Elections are not the only thing that can threaten an incumbent’s hold on power, and dictators must typically worry more about extra-constitutional changes of power like coups d’état. Finding 3 is not about the duration of regimes, but about their expected planning horizon, where governments more secure in their power will plan for the longer term. Second, Finding 3 is based on the assumption that the probability that they will be turned out of office is fixed, and in particular, that the probability of a change in government is not affected by the incumbent’s spending decision. By relaxing this assumption below, it can be shown that when the incumbent’s political support depends negatively on overspending in the first period, they will restrain spending. If this feedback in political support is stronger in democracies, as one may plausibly think, then democracies will see more efficient spending than dictatorships. Finally, in democratic countries where the rule of law is strong, there may be less profound internal divisions and less biased policymakers, for example, because of the more predictable policy environment and the limits on what governments can do. As Finding 4 shows, less bias leads to less overspending.
Since distributive conflicts can lead to overspending of accumulated funds, we should not be surprised that the resulting front-loading of spending has distributive effects.
The distortion has distributive implications. It redistributes to the incumbent’s group from the other group, thereby partly compensating for the lower welfare the incumbent has due to the probability of losing office. Given that the incumbent prioritizes their own group within each period, the other group would prefer aggregate expenditure to be shifted toward period 2 (second-term), when the policymaker has some likelihood of deciding the allocation, and away from period 1 (first-term), when it is certain that the policymaker decides. The challenger from the other group, then, would prefer a reduction in first-period expenditure relative to the baseline efficient time path, rather than an increase. Thus, as the policymaker protects themselves against the eventuality that the other group gains power, they make the other group worse off.
An important finding here is that the loss that the incumbent inflicts on the challenger by shifting expenditure to period 1 (first term) is greater than their own gain.