VoxEU Column Politics and economics

Democracy and accountability: The perverse effects of term limits

The democratic peace – the regularity that democracies do not go to war with each other – is one of the most robust findings in political science. This column presents recent research showing that democratic leaders unable to seek another term in office behave like autocrats. Accountability to voters in the next election lies at the heart of the democratic peace.

Economists have studied in depth how alternative forms of government can lead to systematic differences in outcomes (see Persson and Tabellini, 2004). Autocracies differ from democracies in their economic and political performances. In democracies, political leaders are accountable to the electorate, since periodic elections make it possible to reward or punish them for their actions at the voting booth. In autocracies, leaders escape this general scrutiny and usually rely on a military apparatus or on the support from key groups to sustain their rule.

However, this classification is too broad and does not do justice to the complexity and diversity of political regimes. As Tim Besley and Masa Kudamatsu argued last year on Vox, autocracies in which the leaders are subject to the control of key supporting groups have better economic performances than regimes in which such discipline is missing. This observation underlines the importance of checks and balances in providing incentives to guide the behaviour of political leaders. Conversely, if the fundamental principle common to all democracies is that the power resides in the people, there are a wide variety of democratic regimes. From presidential to parliamentary systems, with all possible shades in between, each democracy has its particularities that lead to large differences in leaders’ accountability.

Many democracies with presidential or semi-presidential political systems impose restrictions on the tenure of their executives. Term limits may reduce the disciplining effect of electoral accountability, as politicians who cannot be re-elected have little to lose from displeasing voters and may thus behave in a more self-interested way. Term limits come in different varieties. Many countries impose “strong” term limits, which rule out re-election after a fixed number of terms. For example, since 1917 Mexico has had one-term limits, ruling out the possibility of re-election of the President altogether, while since 1951 the United States has had two-term limits, which allow re-election only once. Other countries impose “weak” term limits, which only restrict the number of consecutive terms a person can serve. For instance, since 1994 Panama has made re-election of the President possible after skipping two terms. Figure 1 shows that a significant number of countries impose strong term limits on their leaders, making it relevant to understand how this institutional feature affects policy choices.

Figure 1. Strong term limits in 2001

Source: Conconi, Sahuguet and Zanardi, 2008

Democratic peace and term limits

In a recent paper (Conconi, Sahuguet and Zanardi, 2008), we examine the importance of electoral accountability for international peace and cooperation, focusing on the impact of executive term limits on inter-state conflicts. One of the few stylised facts in international relations is that democratic states are much less likely to fight one another than other pairs of states. According to an often-quoted statement by Jack Levy (1988), the democratic peace phenomenon is “as close as anything we have to an empirical law in international relations.” The idea that democracies do not fight each other can be traced back to Immanuel Kant’s 1795 essay “Perpetual Peace”. Kant’s argument was that policymakers in non-democratic states are more likely to engage in conflicts because they are not constrained by electoral accountability.

What is the link between electoral accountability and the likelihood of inter-state conflicts? And if electoral accountability is the explanation for the democratic peace, is there any effect of term limits on the likelihood of conflicts? To address these questions, we describe a simple theoretical model of self-enforcing international peace. Our analysis is based on the fundamental observation that, without a supranational authority with direct powers to punish violations, governments will only refrain from aggressive behaviour if they perceive that doing so is in their interest. The use of military force is beneficial in the short-run but has long-term detrimental consequences. Each country can gain by launching an assault on another country to obtain a portion of its wealth and resources; however, once the attacked country responds by defending itself, being in a conflict is often costly for all countries involved compared to peace.

The incentives of leaders to maintain a cooperative attitude with their foreign partners differ with the type of political regime. In democracies, leaders who want to stay in power need to behave in the interest of the voters. This observation leads to predictions about the likelihood of conflict in different dyads (between autocracies, between democracies, and between democracies and autocracies). Electoral incentives create accountability and discipline policymakers. This explains the lower probability of conflict observed between democracies: the threat of losing office reduces politicians’ willingness to break peaceful relations with other countries. From this perspective, term limits, which restrict the number of mandates a politician can serve in office, should hinder cooperation, since term limits reduce and can even eliminate the incumbent’s benefits from future periods in office, which reduces voters’ ability to punish leaders who engage in costly conflicts.

Empirical evidence

These results are intuitive and their policy implications possibly far reaching. Is there any empirical evidence to support them? To examine the impact of re-election motives on the likelihood of conflicts, we have collected information about the different types of executive term limits adopted by countries over time. Combining this information with a large dataset of inter-state conflicts over the period 1816-2001 for a total of 177 countries, we assess the validity of our theoretical predictions by comparing the conflict patterns of democracies with no term limits to those of democracies with one-term or two-term limits.

Our analysis of the determinants of inter-state conflicts provides strong support for the accountability argument. In line with the existing empirical literature on the democratic peace, we show that democratic dyads are significantly less likely to be in conflict than mixed or autocratic dyads. Crucially, however, this result does not hold for democracies in which the executive is in the last possible mandate, which are as likely to be involved in conflicts as autocracies. Thus the presence of binding term limits invalidates the democratic peace phenomenon.

There is another implication of term limits for the accountability of political leaders. Democratic leaders are accountable as long as they face election in the future. This means that in democracies with two-term limits, there should be more accountability when leaders are in their first mandate than when they are in their second mandate. We indeed find that the likelihood of conflicts in democracies with two-term limits depends on whether the executive is in the penultimate or in the last possible mandate.


Our analysis of the impact of term limits on inter-state conflicts confirms that domestic political institutions can have a crucial impact on economic and political outcomes. In democracies without term limits, periodic elections provide the means to hold opportunistic political leaders accountable for their foreign policy decisions. In autocracies and democracies with term limits, in which there is no need for “contract renewal”, politicians can adopt unpopular policies with no repercussion on whether or not they are able to stay in power. Some caution is clearly warranted in interpreting these results. Though our analysis shows that political systems in which the leaders are subject to re-election are good for peace, this should not be taken to imply that democratisation of dictatorships will necessarily lead to peace. The take-home message, as pointed out by Daron Acemoglu, Davide Ticchi and Andrea Vindigni recently on Vox, is that policymakers should carefully consider the complexity of the political environment when trying to shape or guide the transition to democracy.


Acemoglu, D., D. Ticchi, and A. Vindigni (2008): “A Theory of Military Dictatorships,” VoxEU.org.

Besely, T., and M. Kudamatsu (2007): “What Can We Learn from Successful Autocracies?,” VoxEU.org.

Conconi, P., N. Sahuguet, and M. Zanardi (2008): “Democratic Peace and Electoral Accountability,” CEPR Discussion Paper 6908.

Levy, J. S. (1988): “Domestic Politics and War,” Journal of Interdisciplinary History 18, 653-673.

Persson, T., and G. Tabellini (2004): “Constitutions and Economic Policy,” Journal of Economic Perspectives 18, 75-98.

7,244 Reads