VoxEU Column Politics and economics

Electoral sentencing cycles

In the field of criminal justice, there is wide agreement that judges should be both independent and accountable, but little consensus on how to ensure they are both. This column looks at the US state of North Carolina, where judges are elected and forced to rotate between districts. It finds that in the face of electoral incentives, judges behave like politicians in pandering to what they perceive as electors’ preferences, which can result in the unequal treatment of otherwise similar cases.  

The selection of judges faces the challenge of balancing two traits: accountability, which demands that judges resist their personal biases; and independence, which demands that judicial decisions conform to the rule of law. While the challenge is clear, little consensus exists on the best way to overcome it. Some argue in favour of appointment, others for elections, while still others prefer the random selection of judges. 

As opposed to most European countries, where judges are typically civil servants, most states in the US select judges via competitive elections with term limits. The proponents of judicial elections argue that being subject to electoral review creates incentives to reach prompt decisions and overcome personal biases. Opponents point out that elections push judges to pander to local preferences, particularly during election periods (Shugerman 2012).

Electoral sentencing cycles 

An additional cost of elections is their capacity for causing variability in sentencing, what we call ‘electoral sentencing cycles’. Several contributions in political science, law, and economics have shown that sentences tend to increase at the end of a cycle, as elections approach (Besley and Payne 2005, Helland and Tabarrok 1999, Berdejo and Yuchtman 2013). This tendency, which has been consistently interpreted as judges pandering to the electorate’s tastes for higher sentences, is also consistent with findings showing that more media pressure increases sentences, but only in districts where judges are standing in non-partisan elections (Lim et al. 2015).

But variations in sentencing during electoral periods are not necessarily driven by judges’ efforts to meet the presumed tastes of voters (what we will call the ‘pandering effect’). Instead, those variations may reflect several other confounding factors. First, judges’ increased stress due to the risk of losing their jobs could affect their sentencing behaviour (what we will call the ‘stress effect’). Second, and more importantly, variations in sentencing could reflect changes in the context in which decisions are taken (what we will call the ‘contextual effect’). This contextual effect could be due to more intense media coverage linked to the election or to an increasingly politicised working environment. It could also correspond to a change in the types of cases seen during electoral periods, for instance due to changes in police incentives. Indeed, many other elections may occur simultaneously, including elections for sheriffs. Distinguishing between these different mechanisms is essential to understanding how judges respond to elections, and assessing the desirability of electoral scrutiny.

In a recent paper (Abrams et al. 2019), we disentangle these different effects in an environment that allows us to cleanly establish the existence of sentencing electoral cycles. Our study exploits several unique features of the criminal justice system in the state of North Carolina, which is characterised by both judicial elections and forced judicial rotation. The state is divided into eight regions, in turn divided in a variable number of districts. Judges are elected in a district (i.e. their ‘home district’) but cannot stay there permanently. Every six months, they are forced to rotate across districts according to a schedule determined by the state’s chief justice. This rotation takes place within electoral cycles lasting eight years, and elections are not held at the same time across districts.

This institutional environment has key advantages: in any given term, a judge can work within or outside her home district, and the district where she is working can either have an election coming up or be outside the electoral period. These different possibilities allow us to distinguish between effects. For instance:

  • A judge working inside her home district during an election period is subject to stress, contextual effects, and pandering effects.
  • A judge working outside her home district, with an election coming up at home but not in her current district, is subject solely to the stress effect.
  • A judge working outside her home district, with no election coming up at home but one approaching in her current district, is subject solely to the contextual effect.

Using the universe of criminal decisions taken in North Carolina superior courts between

1998 and 2011, we study the evolution of sentences over the electoral cycle. As visible in Figure 1, where we only plot cases decided when a contested election is approaching at home, sentences increase an average of 25 days (equivalent to a 10% increase) before the elections. However, there is no effect when the judge is not in her home district, regardless of whether an election is held in that district. Thus, we do not find evidence of behavioural or contextual factors driving judicial decisions during pre-electoral periods. 

Figure 1 Effects of distance to elections on sentence severity

Furthermore, exploiting differences between contested and retention elections, we find no change in sentencing when a judge’s position is not contested, consistent with the idea that sentences are increased to attract voters. Finally, we show that the change in sentence severity before elections is driven by more serious crimes. The 80th and 90th percentile of sentences see a peak in sentencing two months before elections. Sentences determined in this period are up to 100 days longer. Judges appear to use the most visible crimes to signal their toughness. These results are in line with findings by Boston and Silveira (2019) that also use data from North Carolina to study how judges adapt sentencing when elections shift from state wide to the district level, showing that judges strategically adapt their sentencing behaviour to their constituencies’ political preferences in order to retain office.


In the face of electoral incentives, judges behave like politicians, strategically pandering to what they perceive as electors’ preferences. In determining the optimal system to appoint judges, the prospect that elections may cause unequal treatment of otherwise similar cases should be taken into account, prompting us to consider whether people, particularly in visible cases, should face more severe sentences when elections approach.


Abrams, D, R Galbiati, E Henry and A Philippe (2019), “Electoral Sentencing Cycles”, CEPR Discussion Paper 14049.

Berdejo, C and N Yuchtman (2013), “Crime, Punishment, and Politics: An Analysis of Political Cycles in Criminal Sentencing”, Review of Economics and Statistics 95(3): 741-756.

Besley, T and A A Payne (2013), “Implementation of Anti-Discrimination Policy: Does Judicial Selection Matter”, American Law and Economics Review 15(1): 212-251.

Boston, J and B S Silveira (2019), “The Electoral Connection in Court: How Sentencing Responds to Voter Preferences”, Mimeo: UCLA

Helland, E and A Tabarrok (1999), “Court Politics: The Political Economy of Tort Awards”, Journal of Law and Economics 42: 157-188.

Lim, C (2013), “Preferences and Incentives of Appointed and Elected Public Officials”, American Economic Review 103(4): 1360-1397.

Shugerman J H (2012), The People’s Courts: Pursuing Judicial Independence in America, Harvard University Press.

630 Reads