Abstract
In this article we study whether an elected leader has incentives to withdraw from an ongoing military conflict before elections take place. Remaining active in the conflict increases the chances of a victory, which would boost the political incumbent’s reelection probability. On the other hand, if the rival does not surrender, then the incumbent politician’s reelection probability decreases because of the costly conflict citizens have to endure. We show that when the crisis is costly to voters, the more distant the elections are, the more likely the ruling government withdraws from an ongoing war early. For low-cost conflicts, politicians never give up the fighting. We also show that confrontations in the shadow of elections are socially inefficient (too long or too short) because of the inherent misalignment of preferences between citizens and politicians.
Similar content being viewed by others
Notes
For a theoretical setup to understand their empirics, see Conconi et al. (2018).
In complete information setups, the drop-out time of players in mixed strategy equilibria follows an exponential distribution. In incomplete information setups, the distribution of the equilibrium drop-out time can be of various forms depending on the distribution of private signals (see, e.g., Menuet and Sekeris, 2022). In the present setup, we wish to keep the analysis simple and assume that the rival country’s equilibrium drop-out distribution is a function of non-modeled private characteristics of the rival and that at equilibrium, the rival drops out following a uniform distribution. This reflects a series of uncertain events possibly affecting decisions of governments or group leaders, including major exogenous shocks constraining the choices of governments (e.g., economic shocks, international wars, environmental shocks/disasters) or events modifying the very composition or survival of a government (e.g., military coups, death of a leader, snap elections). Given the uncertainty surrounding the eventual drop-out time of any decision-making entity, we can therefore always rationalize any distribution of types or drop-out times of the rival.
We are adopting a reduced form approach that greatly simplifies computations. The rationale for assuming such terminal date is that when choosing their preferred political candidate, voters will typically not base their reasoning on the likelihood of another election taking place down the line, and politicians being strategic with respect to that election and the infinitely many possible subsequent elections where war could have continued.
Observe that \(\kappa\) can be made a function of the prize \(\alpha\) without any consequence for our results.
We follow Persson and Tabellini (2000) in assuming perfect commitment of candidates to their electoral platform so as to avoid time-inconsistency issues.
Instead of considering a \(50\%\) threshold for winning elections, we could instead had considered any alternative threshold D that would be inversely related to an authoritarian regime’s strength as in Siqueira and Sekeris (2012). Such an alternative modeling assumption would not qualitatively affect our results.
Over the incumbent’s expectation, we have: \({\mathbb {P}}_I(\min \{T_I^G,T^R\}\ge T_E)={\mathbb {P}}_I(T^R\ge T_E;\;T_I^G\ge T_E)=[1-F_R(T_E)] {\textbf{1}}_{T_I^G \ge T_E}\).
References
Adam, A., & Sekeris, P. G. (2017). Self-containmnent: Achieving peace in anarchic settings. Journal of Conflict Resolution, 61(1), 173–203.
Autor, D., Dorn, D., Hanson, G., & Majlesi, K. (2020). Importing political polarization? The electoral consequences of rising trade exposure. American Economic Review, 110(10), 3139–3183.
Baliga, S., Lucca, D. O., & Sjöström, T. (2011). Domestic political survival and international conflict: Is democracy good for peace? American Economic Review, 78, 458–486.
Battaglini, M., & Harstad, B. (2020). The political economy of weak treaties. Journal of Political Economy, 128(2), 544–590.
Bjornskov, C., & Rode, M. (2020). Regime types and regime change: A new dataset on democracy, coups, and political institutions. The Review of International Organizations, 15, 531–551.
Blomberg, S. B., & Hess, G. D. (2003). Is the political business cycle for real? Journal of Public Economics,87(5–6), 1091–1121.
Bouton, L., Conconi, P., Pino, F., & Zanardi, M. (2021). The tyranny of the single minded:guns, environment, and abortion. Review of Economics and Statistics, 103, 48–59.
Bove, V., Efthyvoulou, G., & Navas, A. (2017). Political cycles in public expenditure: butter vs guns. Journal of Comparative Economics.
Carter, D. B., & Signorino, C. S. (2010). Back to the future: Modeling time dependence in binary data. Political Analysis, 18(3), 271–292.
Carter, J., & Nordstrom, T. (2017). Term limits, leader preferences, and interstate conflict. International Studies Quarterly, 61, 721–735.
Cazals, A., & Sauquet, A. (2015). How do elections affect international cooperation? Evidence from environmental treaty participation. Public Choice, 162, 263–285.
Cederman, L.-E. (2003). Modeling the size of wars: From billiard balls to sandpiles. American Political Science Review, 97(1), 135–150.
Conconi, P., DeRemer, D. R., Kirchsteiger, G., Trimarchi, L., & Zanardi, M. (2017). Suspiciously timed trade disputes. Journal of International Economics, 105, 57–76.
Conconi, P., Sahuguet, N., & Zanardi, M. (2014). Democratic peace and electoral accountability. Journal of the European Economic Association, 12(4), 997–1028.
Cox, D. R. (1972). Regression models and life-tables. Journal of the Royal Statistical Society. Series B (Methodological), 34(2), 187–220.
Downs, G. W., & Rocke, D. M. (1994). Conflict, agency, and gambling for resurrection: The principal-agent problem goes to war. American Journal of Political Science, 38(2), 362–380.
Fearon, J. D. (1995). Rationalist explanations for war. International Organization, 49(3), 379–414.
Gartner, S. S. (2008). The multiple effects of casualties on public support for war: An experimental approach. American Political Science Review,102(1), 95–106.
Gartner, S. S., & Segura, G. M. (1998). War, casualties, and public opinion. Journal of Conflict Resolution, 42(3), 278–300.
Gartner, S. S., & Segura, G. M. (2000). Race, casualties, and opinion in the vietnam war. Journal of Politics, 62(1), 115–146.
Gartner, S. S., Segura, G. M., & Barratt, B. A. (2003). War casualties, policy positions, and the fate of legislators. Political Research Quarterly, 53(3), 467–477.
Gartzke, E. (1998). Kant we all just go along? Opportunity, willingness, and the origins of the democratic peace. American Journal of Political Science, 42(1), 1–27.
Gartzke, E. (2007). The capitalist peace. American Journal of Political Science, 51(1), 166–191.
Gaubatz, K. T. (1991). Election cycles and war. Journal of Conflict Resolution,35(2), 212–244.
Gleditsch, N. P., Wallensteen, P., Eriksson, M., Sollenberg, M., & Strand, H. (2002). Armed conflict 1946-2001: A new dataset. Journal of Peace Research, 39(5), 243–250.
Guriev, S., & Treisman, D. (2016). What makes governments popular? SSRN DP11460.
Hess, G. D., & Orphanides, A. (1995). War politics: An economic, rational-voter framework. American Economic Review, 85, 828–846.
Hess, G. D., & Orphanides, A. (2001). War and democracy. Journal of Political Economy, 109(4), 776–810.
Jackson, M. O., & Morelli, M. (2007). Political bias and war. American Economic Review, 97(4), 1353–1373.
Karol, D., & Miguel, E. (2007). The electoral cost of war: Iraq casualties and the 2004 u.s. presidential election. Journal of Politics, 69(3), 633–648.
Kleine, M., & Minaudier, C. (2019). Negotiating under political uncertainty: National elections and the dynamics of international co-operation. British Journal of Political Science, 49(1), 315–337.
Kreutz, J. (2010). How and when armed conflicts end: Introducing the ucdp conflict termination dataset. Journal of Peace Research, 47(2), 243–250.
Landau-Wells, M. (2018). High stakes and low bars: How international recognition shapes the conduct of civil wars. International Security, 43(1), 100–137.
Lindbeck, A., & Weibull, J. W. (1987). Balanced-budget redistribution as the outcome of political competition. Public Choice, 52, 273–297.
Maoz, Z., & Russett, B. (1993). Normative and structural causes of democratic peace, 1946–1986. Journal of Conflict Resolution, 87(3), 624–638.
Menuet, M., & Sekeris, P. G. (2022). Engaging in wars of attrition. Annals of Economics and Statistics, 147, 79–100.
Menuet, M., & Villieu, P. (2021). Reputation and the “need for enemies’’. Economic Theory, 72(4), 1049–1089.
Morales, J. S. (2021). Legislating during war: Conflict and politics in colombia. Journal of Public Economics, 193, 104325.
Nedal, D., Stewart, M., & Weintraub, M. (2020). Urban concentration and civil war. Journal of Conflict Resolution, 64(6), 1146–1171.
Nordhaus, W. D. (1975). The political business cycle. Review of Economic Studies, 42(2), 169–190.
Persson, T., & Tabellini, G. (2000). Political Economics: Explaining Economic Policy. MIT Press.
Pervez, F. (2015). Waiting for election season: The timing of international trade disputes. Review of International Organizations, 10, 265–303.
Siqueira, K., & Sekeris, P. G. (2012). Politics and insurgencies. Economics and Politics.
Thomson, H. (2016). Rural grievances, landholding inequality, and civil conflict. International Studies Quarterly, 60(3), 511–519.
Zeigler, S., Pierskalla, J. H., & Mazumder, S. (2014). War and the reelection motive: Examining the effect of term limits. Journal of Conflict Resolution, 58(4), 658–684.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The authors have no competing interests to declare that are relevant to the content of this article.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Appendix A: Pre-electoral period
From (12), the first-order derivative is
where the marginal probability of the opponent declaring defeat, \(f_R(.)\) is given by \(1/{{\bar{T}}}\), and thus
; hence,
The first-order condition is then
The second-order derivative is
Observe that \(\Phi (T_I^G)\) admits a minimal value on \({\textbf{R}}^{+}\) at \(T_I^G=\sqrt{\frac{2\mu {\bar{T}}}{3a}}\), with \(\Phi (0)=h\kappa /{{\bar{T}}}>0\). In addition, \(\Phi (\sqrt{\frac{2\mu {\bar{T}}}{3a}}) < 0 \Leftrightarrow {{\bar{T}}}> \frac{27 \kappa ^2}{2ah\mu ^3}\); a condition that we assume throughout the paper since a fairly large time horizon is considered (a high \({{\bar{T}}}\)).
As \(\Phi (\sqrt{\frac{2\mu {\bar{T}}}{3a}})< 0\), according to the intermediate value theorem, \({\mathcal {V}}_I\) admits a local interior maximum at \({\check{T}}\in ]0,\sqrt{\frac{2\mu {\bar{T}}}{3a}}[\), which will also be the global maximum if \({\mathcal {V}}_I\left( \check{T}\right) >{\mathcal {V}}_I(T_E)\). Imposing that \(T_E\le \sqrt{\frac{2\mu {{\bar{T}}}}{3a}}\) enables us to conclude that the maximum of \({\mathcal {V}}_I\) on \([0,T_E]\) will always be such that \({\hat{T}}_I^G:=argmax_{T_I^G \in [0,T_E]} {\mathcal {V}}_I(T_I^G) =\min \left\{ {\check{T}},T_E\right\}\).
By applying the implicit function theorem to \(\Phi (T_I^G)=0\), we can then deduce that \(\partial T_I^G/\partial \kappa =-\frac{{\partial \Phi (T_I^G)}/{\partial \kappa }}{{\partial \Phi (T_I^G)}/{\partial T_I^G}}=-\frac{1/{{\bar{T}}}}{{\partial ^2{\mathcal {V}}_I}/{\partial (T_I^G)^2}}\). As \(T^I_G={\check{T}}\) is a local maximum of \({\mathcal {V}}_I\), the second-order condition implies that \(\partial {\check{T}}/\partial \kappa >0.\) In addition, looking at Eq. (15), \({\check{T}}=0<T_E\) if \(\kappa =0\), and \(\check{T}\rightarrow +\infty >T_E\) as \(\kappa \rightarrow \infty\). Hence, there is a critical level \({\underline{\kappa }}>0\), such that: \({\hat{T}}_I^G={\check{T}}<T_E\) if \(\kappa <{\underline{\kappa }}\), and \({\check{T}}\ge T_E={\hat{T}}_I^G\) if \(\kappa \ge {\underline{\kappa }}\).
Additionally, the critical level \({\underline{\kappa }}\) is such that \({\check{T}}-T_E=0\); hence,
Appendix:B: Post-election period
After the election, the first- and second-order derivatives are
where
Hence,
Given the assumed uniform distribution, we obtain
and
There is a unique critical point at
Substituting in (18), we observe that \(v''({\hat{T}})>0\); hence ,\({\hat{T}}\) is a global minimum. Therefore, the maximum of the payoff \({\mathcal {V}}_I\) on \([T_E,{{\bar{T}}}]\) is reached at a boundary, either at \(t=T_E\) or \(t={{\bar{T}}}\).
To then find the maximum, we need to determine the sign of \({\mathcal {V}}_I({{\bar{T}}})-{\mathcal {V}}_I(T_E)\). Using Eqs. (5) and (11), we compute
where
We have \(\psi '<0\), \(\psi (+\infty )=-\infty\), and \(\psi (0)=\alpha (e^{-rT_E}-e^{-r{{\bar{T}}}})>0\).
There is a unique critical value \({\underline{v}} >0\), such that \(\psi (v)>0\Leftrightarrow {\mathcal {V}}_I({{\bar{T}}})>{\mathcal {V}}_I(T_E)\) for \(v<{\underline{v}}\), and \(\psi (v)\le 0\Leftrightarrow {\mathcal {V}}_I({{\bar{T}}})\le {\mathcal {V}}_I(T_E)\) for \(v\ge \underline{v}\). Consequently, the maximum of the payoff \({\mathcal {V}}_I\) on the post-election period is \({{\bar{T}}}\) if \(v<{\underline{v}}\), and \(T_E\), if \(v\ge {\underline{v}}\).
Appendix C: Global extremum
Proof of Proposition 4
Observe that \({\mathcal {V}}_I(\cdot )\) is a continuous mapping on \([0,{{\bar{T}}}]\). Therefore, Propositions 2 and 3 allow us to distinguish four cases, depending on the values of parameters \(\kappa\) and \(T_E\).
First: \(\kappa \ge \underline{\kappa }\) and \(v\ge {\underline{v}}\). In this case, we have \(\max _{t \in [0,T_E]} {\mathcal {V}}_I(t)={\mathcal {V}}_I(T_E)\), and \(\max _{t \in [T_E,{{\bar{T}}}]} {\mathcal {V}}_I (t)={\mathcal {V}}_I(T_E)\). Consequently, it follows that \(\max _{t \in [0,{{\bar{T}}}]} {\mathcal {V}}_I(t)={\mathcal {V}}_I(T_E)\); hence, \(T^*=T_E\).
Second: \(\kappa < \underline{\kappa }\) and \(v \ge {\underline{v}}\). In this case, we have \(\max _{t\in [0,T_E]} {\mathcal {V}}_I(t)={\mathcal {V}}_I({\check{T}})\) and \(\max _{t\in [T_E,\bar{T}]} {\mathcal {V}}_I(t)={\mathcal {V}}_I(T_E)\). Consequently, it follows that \(\max _{t\in [0,{{\bar{T}}}]} {\mathcal {V}}_I(t)={\mathcal {V}}_I(\check{T})\); hence, \(T^*={\check{T}}\).
Third: \(\kappa \ge \underline{\kappa }\) and \(v <{\underline{v}}\). In this case, we have \(\max _{t\in [0,T_E]} {\mathcal {V}}_I(t)={\mathcal {V}}_I(T_E)\) and \(\max _{t\in [T_E,{{\bar{T}}}]} {\mathcal {V}}_I (t)={\mathcal {V}}_I({{\bar{T}}})\). It follows that \(\max _{t\in [0,{{\bar{T}}}]} {\mathcal {V}}_I(t)={\mathcal {V}}_I({{\bar{T}}})\); hence, \(T^*={{\bar{T}}}\).
Fourth: \(\kappa <{\underline{\kappa }}\), and \(v<{\underline{v}}\). In this case, we have \(\max _{t\in [0,T_E]} {\mathcal {V}}_I(t)={\mathcal {V}}_I({\check{T}})\) and \(\max _{t\in [T_E,\bar{T}]} {\mathcal {V}}_I(t)={\mathcal {V}}_I({{\bar{T}}})\). Since \(\left. {\mathcal {V}}_I({\check{T}})\right| _{\kappa< \underline{\kappa }}<\left. {\mathcal {V}}_I(T_E)\right| _{\kappa \ge \underline{\kappa }}\), it follows that \(\max _{t\in [0,{{\bar{T}}}]} {\mathcal {V}}_I(t)={\mathcal {V}}_I({{\bar{T}}})\); hence, \(T^*={{\bar{T}}}\). \(\square\)
Proof of Corollary 1
We first establish the relationship between \({\underline{v}}\) and \(T_E\) by implicitly differentiating (20):
Since \(\psi '(\cdot )<0\), the sign of \(\partial {\underline{v}}/ \bar{T}_E\) is given by the sign of the numerator, i.e., by the sign of \(\partial \psi ({\underline{v}})/\partial T_E\).
From Eq. (20), we obtain
where
Observe that \(\Lambda _{{\underline{v}}}(T_E)\ge 0 \Leftrightarrow T_E\le {\check{T}}_E:= {{\bar{T}}} - (\alpha r - 2{\underline{v}})/\underline{v}r\). Since \(\alpha r >2{\underline{v}}\), if \(\Lambda _{\underline{v}}(0)>0\), then \({\check{T}}_E\in ]0,{{\bar{T}}}[\). Consequently, we deduce:
In words, We thus deduce that the relationship between \(\underline{v}\) and \(T_E\) is described by an inverted U-shaped curve if \(\Lambda _{{\underline{v}}}(0)>0\), and \({\underline{v}}\) monotonically decreases in \(T_E\) otherwise.
Let us now assume that values of \(T_E\) are very high (formally, \(T_E>{\check{T}}_E\)). In this case, we have \(\partial {\underline{\kappa }} /\partial T_E >0\) (as shown in Appendix A), and \(\partial {\underline{v}} /\partial T_E < 0\). Namely, if \(T_E\) is high enough, \({\underline{\kappa }}\) is also high but \({\underline{v}}\) is low. Hence, as \(T_E\) increases it is increasingly likely that \(\kappa <{\underline{\kappa }}\) and \(v\ge {\underline{v}}\), i.e., that the equilibrium is described by case 2 in Fig. 1. Under this scenario, the incumbent drops out before elections since \(T^*\in (0,T_E)\).
Assume next that the value of \(T_E\) declines. Since \(T_E\) is initially high, \({\underline{v}}\) will consequently increase, while \({\underline{\kappa }}\) decreases. Hence, as \(T_E\) decreases it is increasingly likely that \(\kappa \ge {\underline{\kappa }}\) and \(v <{\underline{v}}\), i.e., that the equilibrium is described by case 3 in Fig. 1. Under this scenario, the incumbent drops out after elections since \(T^*={{\bar{T}}}\).
Further decreases of \(T_E\) imply that \(T_E\) ultimately falls below the threshold \({\check{T}}_E\), and the relationship between \(\underline{v}\) and \(T_E\) then reverses (i.e., \(\partial {\underline{v}}/\partial T_E>0\)). Consequently both bounds \({\underline{\kappa }}\) and \({\underline{v}}\) become smaller and it becomes increasingly likely that \(\kappa \ge {\underline{\kappa }}\) and \(v\ge {\underline{v}}\), i.e., that the equilibrium is described by case 1 in Fig. 1. Under this scenario, the incumbent drops out at the election date \(T^*=T_E\).
Appendix D: Social welfare
By differentiating (14) with respect to \(T_S\), we have:
Given the assumed uniform distribution, we obtain
, and
Observe that \(u^{'}(0)<0\) and that \(u^{'}({\bar{T}})>0\). Moreover, when \(u^{'}(T_S)=0\), \(u^{''}(T_S)>0\) which implies that function u(.) will only change sign once and that it is accordingly U-shaped on \([0,{{\bar{T}}}]\) with a minimum at \({{\bar{T}}}_S=\bar{T}-\alpha /v\). Consequently, \(\max u(t)\in \{u(0),u({{\bar{T}}})\}\).
We know, first, that \(u(0)=0\) since if the planner were to drop out in \(T=0\), in which case the WoA’s duration would be zero and the focal country would lose with certainty and obtain no gains.
Second, \(u({{\bar{T}}})=\int _{x=0}^{x={{\bar{T}}}} \left\{ \int _{0}^x -ve^{-rt}\;\text {d}t+\alpha e^{-rx} \right\} f_R(x)\;\text {d}x\). Indeed, if the planner drops out at \({{\bar{T}}}\), the country wins for sure and the representative citizen obtains \(\alpha\). The cost of this certain victory is given by the period cost of the WoA, weighted by the number of periods the planner expects the rival to be actively fighting the WoA. This cost is given by \(\int _{0}^x -ve^{-rt}\;\text {d}t\).
Computing this payoff, we obtain:
with,
Hence, the social optimum hinges on the sign of \(\Psi (v)\) since \(\Psi (v)> 0\Rightarrow T_S={{\bar{T}}}\), and \(\Psi (v)\le 0\Rightarrow T_S=0\). Following the same reasoning as in Appendix B, we can apply the intermediate value theorem on function \(\Psi\) and deduce that there is a unique level \({\hat{v}}>0\), such that: \(v < {\hat{v}}\Leftrightarrow \Psi (v) > 0 \Leftrightarrow T_S^*= {{\bar{T}}}\), and \(v \ge {\hat{v}}\Leftrightarrow \Psi (v)\le 0 \Leftrightarrow T_S^*=0\).
In Appendix C we show that \(v< {\underline{v}} \Leftrightarrow \psi (v)> 0\Leftrightarrow T^*={{\bar{T}}}\), and that otherwise \(T^*=T_E\) or \(T^*={\check{T}}\in (0,T_E)\).
We next show that \({\underline{v}} <{\hat{v}}\). To establish this result, from Eq. (20), we obtain:
It is clear that \(\Lambda '(T_E)\ge 0\) under the condition \(\bar{T}\le T_E+(r\alpha -2v)/rv\) that we assume to be true. In this case, \(\Psi (v)-\psi (v)\ge \Lambda (0)= v{{\bar{T}}}\left( 1-e^{-r\bar{T}}\right) >0\), so that \({\underline{v}} <{\hat{v}}\). Hence, there are three cases:
-
(i)
If \(v<{\underline{v}} <{\hat{v}}\). In this case, the incumbent and the social planner both decide to drop out at \({{\bar{T}}}\), i.e., \(T^*=T_S^*={{\bar{T}}}\). The political equilibrium is socially optimal.
-
(ii)
If \({\underline{v}} \le v < {\hat{v}}\). In this case, according to Proposition 4, the incumbent chooses either \(T^*=T_E\) or \(T^*=\check{T} \in (0,T_E)\), while the social planner decides \(T_S^*={{\bar{T}}}\). The political equilibrium is socially suboptimal.
-
(iii)
If \(v \ge {\hat{v}}\). In this case, according to Proposition 4, the incumbent chooses either \(T^*=T_E\) or \(T^*={\check{T}} \in (0,T_E)\), while the social planner decides \(T_S^*= 0\). The political equilibrium is socially suboptimal.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Adam, A., Menuet, M. & Sekeris, P.G. Conflict under the shadow of elections. Public Choice (2024). https://doi.org/10.1007/s11127-024-01148-2
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s11127-024-01148-2