VoxEU Column Labour Markets Institutions and economics

Recidivism and neighbourhood institutions

Between 30% and 50% of individuals sentenced to prison are reincarcerated in the two years after their release. Neighbourhood institutions that former inmates encounter after prison may play a role in encouraging crime desistance. This column examines the link between Evangelical church openings in Chile and reincarceration rates in the surrounding neighbourhoods. The opening of an Evangelical church in the neighbourhood significantly reduces 12-month reincarceration rates among recently released young inmates, suggesting that local institutions can provide a support network that helps former inmates cope and find work.

Between 30% and 50% of individuals sentenced to prison are reincarcerated in the two years after their release (Doleac 2020, Yukhnenko et al. 2019). These high reincarceration rates are costly for societies. Apart from the direct costs of crime, maintaining inmates in prisons is expensive. In OECD countries, for instance, the average annual expenditure per inmate is close to $70,000. 

Encouraging desistance from crime has therefore become a primary policy goal for reducing both crime and incarceration rates (Doleac 2020). Despite growing interest in understanding what factors help rehabilitate convicted criminals, we know little about the role of local institutions that former inmates encounter in their neighbourhoods after prison. There is evidence that neighbourhoods affect many outcomes, including earnings, education, marriage, and fertility as well as participation in crime (Chetty and Hendren 2018a,b, Ludwig et al. 2013, Kling et al. 2005, Sviatschi 2022), suggesting that local institutions could be important in encouraging crime desistance.

Studying the role of neighbourhood institutions in recidivism is particularly interesting in contexts where criminal offenders are geographically concentrated. This is the case for Chile – the setting we study – but also for many other countries, including the US (Card et al. 2008, Chetty et al. 2016). We give causal evidence that local institutions in the neighbourhood to which inmates return after prison matter (Barrios Fernández and Garcia-Hombrados 2022). Specifically, we show that the opening of an Evangelical church significantly reduces reincarceration rates among recently released young inmates (i.e. inmates under 30 years’ old).1 

The rise of the Evangelical church in Chile

Evangelical churches are interesting neighbourhood institutions, which in recent decades have experienced a large expansion throughout the world, particularly in disadvantaged neighbourhoods of Latin America (Costa et al. 2018, Pew Research Center 2011). In Chile, the share of the population identifying as members of an Evangelical church grew from 12% to 18% between 1992 and 2019. In contrast, the share of Catholics dropped from 77% to 45% over the same period. As illustrated in Figure 1, the number of Evangelical churches has also experienced persistent growth in the last decades, especially in disadvantaged neighbourhoods. Indeed, between 2006 and 2014 – the period we study – there were 1,659 church openings.

Figure 1 Evangelical churches opening in Santiago, 2006–2014


Notes: The figure illustrates Evangelical churches opening in Santiago – the capital city of Chile – between 2006 and 2014. Note, however, that the study discussed in this article exploits variation in church openings in the whole country.

While strongly rooted in the local communities, the Evangelical churches’ social action, proselytism, and political lobbying are transforming the social landscape of many Latin American countries (Costa et al. 2018, Fediakova 2013). According to a 2014 Pew Research Center survey of multiple Latin American countries (including Chile), Evangelicals are more likely than other religious and non-religious individuals to do charity work, visit sick people, and provide different types of support to those in need. The survey also shows that the members of Evangelical churches have a more active religious life and often participate in activities to convert and attract new people to the church.

Effect of Evangelical churches on recidivism

To identify the causal effect of Evangelical church openings on recidivism, we take advantage of rich administrative data that include the home address and exact entry and release dates of the universe of individuals under 30 years old who enter prison between 2006 and 2015. We combine these data with official records that contain the address and opening dates of all Evangelical churches that opened between 2006 and 2015 (1,659 churches). 

To overcome endogeneity concerns, we rely on a difference-in-differences strategy. Within a neighbourhood, we define a treatment and a control area. The treatment area corresponds to an inner ring immediately around the church, while the control area corresponds to an external ring slightly further away and arguably no or less affected by the church. We focus on individuals entering prison before a church opens near them and compare their reincarceration probabilities depending on whether their home is in the inner or external ring and on whether they are released from prison before or after the church opens. Our focus on individuals who enter prison before the opening of the church ensures that their first entrance to prison is not affected by the presence of the church.2 

We find that the opening of an Evangelical church reduces 12-month reincarceration rates among property-crime offenders by more than 11 percentage points, an effect that represents a drop of 18% with respect to the baseline reincarceration rates of these individuals. An important part of this drop – 7.3 percentage points – is already apparent three months after the release date. This result is consistent with the findings of Munyo and Rossi (2015), who show that a significant share of inmates re-offend soon after being released. It highlights the relevance of the conditions and support that inmates encounter immediately after leaving prison. We also find that church openings lower the number of young individuals going to prison for the first time. As in the case of reincarceration, the effect is particularly strong for property crimes.

The validity of our empirical strategy relies on the parallel trends assumption. This means that in the absence of a church opening, recidivism should have followed the same trend in control and treatment areas. Figure 2 shows that at least in the six years before the church opening, there were no significant differences in 12-month recidivism rates between treated and control areas. The difference in reincarceration rates arises only after a new church opens. 

The event study concludes two years after a church opens. As discussed earlier in this section, we focus on individuals who enter prison before the church opens to ensure that their entrance to prison is not affected by the presence of the church. Since most sentences related to property crime are less than two years, we do not have the power to study what happens with individuals returning to their neighbourhood three or more years after the church opens.

Figure 2 Effect of Evangelical church openings on 12-months recidivism (property crime)


Notes: This figure illustrates how the estimated effect of Evangelical churches’ openings on recidivism evolves with time. The treated group includes individuals living 100 metres or less from the church, while the control group individuals live between 250 and 350 metres from the church. The dots represent the estimated coefficients, and the bars 95% confidence intervals. Standard errors are clustered at the neighbourhood level (i.e. inner and external rings).

We find smaller and less precise effects when focusing on individuals sentenced for drug crimes, violent crimes, and other types of crimes. It is not surprising to find a significant effect only for individuals involved in property crime. First, there are more individuals in this category, and the base level of 12-month recidivism is also higher among them. Thus, the statistical power is larger for analyses involving this specific group of inmates. 

Second, individuals that commit property and other types of crime differ in crucial traits such as psychopathy or planning measures (Boduszek et al. 2017, Seruca and Silva 2016). Consistently, individuals involved in property crime have been shown to be more responsive to the conditions they find at release and to interventions alleviating material needs (Tuttle 2019, Mallar and Thornton 1978, Berk et al. 1980). On the other hand, individuals involved in more severe types of crimes may have personal traits and links with criminal organisations that could make their rehabilitation more challenging for non-specialised institutions like Evangelical churches.

What is behind these effects?

Our findings show that Evangelical churches reduce 12-month reincarceration rates by more than 18% among property-crime offenders. Below we discuss and explore two broad classes of mechanisms that could drive our findings. We consider mechanisms related to (1) the promotion of Evangelical beliefs and values, and (2) the social support that Evangelical communities provide.

Using data from the 2002 and 2012 censuses, we show that neighbourhoods treated within this period had only one percentage point in the share of individuals who identify as Evangelicals. The modest magnitude of the coefficient indicates that the opening of these churches did not result in massive conversions in treated areas relative to control areas. In addition, we find that the effect we document is driven by individuals who already identified themselves as Evangelicals before entering prison: for them, we find a drop of 17.3 percentage points in 12-months reincarceration rates. In contrast, for individuals of no or other religion, we find the drop was not significant. These results suggest that religious conversions are not the main driver of our findings.

Our evidence is more consistent with changes in available social support that recently released individuals have when they return to their neighbourhoods. We find that the effects of Evangelical churches are particularly large in areas where the presence of the state is weaker (i.e. areas that are further away from municipality buildings and in which there are fewer public services), suggesting that Evangelical churches substitute the state in providing some social services. 

In addition, the opening of non-religious organisations in the neighbourhood to which inmates return generates similar effects to the ones we document for Evangelical churches. Relying once more on our main identification strategy, we find that the opening of organisations that promote labour insertion reduces 12-month recidivism by 11 percentage points.3 Analyses of census data indicate that the openings of Evangelical churches increased employment among Evangelical men under 30 by 2.6 percentage points.


This study provides causal evidence that local institutions of the neighbourhood to which individuals return after prison matter. We show that Evangelical churches opening in the neighbourhoods to which inmates return after prison reduces 12-month reincarceration rates among property-crime offenders by more than 11 percentage points, an effect that represents a drop of 18% with respect to the baseline reincarceration rates of these individuals.

Although we cannot perfectly identify the mechanisms behind our findings, our evidence is consistent with Evangelical churches providing a support network that helps recently released inmates cope with their immediate needs and potentially helps them enter the labour market. In contrast, conversions or changes in religiosity do not seem to play an important role in reducing recidivism. These results suggest that institutions and policy interventions giving recently released inmates access to support networks in their neighbourhoods could play an important role in encouraging desistance from crime.


Barrios Fernández, A, and J García Hombrados (2022), “Recidivism and neighborhood institutions: Evidence from the rise of the Evangelical church in Chile”, CEPR Discussion Paper 17070. 

Berk, R A, K J Lenihan and P H Rossi (1980), “Crime and poverty: Some experimental evidence from ex-offenders”, American Sociological Review 45(5): 766–86.

Boduszek, D, A Debowska and D Willmott (2017), “Latent profile analysis of psychopathic traits among homicide, general violent, property, and white-collar offenders”, Journal of Criminal Justice 51: 17–23.

Card, D, A Mas and J Rothstein (2008), “Tipping and the dynamics of segregation”, Quarterly Journal of Economics 123(1): 177–218.

Chetty, R, N Hendren, F Lin, J Majerovitz and B Scuderi (2016), “Childhood environment and gender gaps in adulthood”, American Economic Review 106(5): 282–88.

Chetty, R, and N Hendren (2018a), “The impacts of neighborhoods on intergenerational mobility I: Childhood exposure effects”, Quarterly Journal of Economics 133(3): 1107–62.

Chetty, R, and N Hendren (2018b), “The impacts of neighborhoods on intergenerational mobility II: County-level estimates”, Quarterly Journal of Economics 133(3): 1163–228.

Costa, F, A Marcantonio Junior and R Castro (2018), “Stop suffering! Economic downturns and Pentecostal upsurge”, FGV EPGE Economics Working Papers (Ensaios Econômicos da EPGE) 804, EPGE Brazilian School of Economics and Finance – FGV EPGE (Brazil).

Doleac, J (2020), “Encouraging desistance from crime”, Journal of Economic Perspectives.

Fediakova, E (2013), Evangélicos, política y sociedad en Chile: dejando ‘el refugio de las masas’, 1990–2010, Concepción, Chile: Centro Evangélico de Estudios Pentecostales, CEEP; Providencia, Santiago, Chile: Instituto de Estudios Avanzados, Universidad de Santiago de Chile.

Goodman-Bacon, A (2021), “Difference-in-differences with variation in treatment timing”, Journal of Econometrics 225(2): 254–77.

Kling, J R, J Ludwig, and L F Katz (2005, 02), "Neighborhood effects on crime for female and male youth: Evidence from a randomized housing voucher experiment", Quarterly Journal of Economics 120 (1), 87–130.

Ludwig, J, G J Duncan, L A Gennetian, L F Katz, R C Kessler, J R Kling, and L Sanbon-matsu (2013, May), Long-term neighborhood effects on low-income families: Evidence from moving to opportunity, American Economic Review 103 (3), 226–231.

Mallar, C D and C V D Thornton (1978), "Transitional aid for released prisoners: Evidence from the Life Experiment", Journal of Human Resources 13 (2), 208–236.

McCall, P, K Land, C Dollar, and K Parker (2013), "The age structure-crime rate relationship: Solving a long-standing puzzle", Journal of Quantitative Criminology 29, 167–190.

Munyo, I and M Rossi (2015), "First-day criminal recidivism", Journal of Public Economics 124 (C), 81–90.

Seruca, T and C F Silva (2016), "Executive functioning in criminal behavior: Differentiating between types of crime and exploring the relation between shifting, inhibition, and anger", International Journal of Forensic Mental Health 15 (3), 235–246.

Sviatschi, M M (2022), "Making a narco: Childhood exposure to illegal labor markets and criminal life paths", Econometrica, forthcoming.

Pew Research Center (2011), Global Christianity: A report on the size and distribution of the world’s Christian population, Washington, DC: Pew Research Center.

Tuttle, C (2019), “Snapping back: Food stamp bans and criminal recidivism”, American Economic Journal: Economic Policy 11(2): 301–27.

Ulmer, J, and D Steffensmeier (2014), “The age and crime relationship: Social variation, social explanations”, in The Nurture Versus Biosocial Debate in Criminology: On the Origins of Criminal Behavior and Criminality, SAGE Publications.

Yukhnenko, D, S Sridhar and S Fazel (2019), “A systematic review of criminal recidivism rates worldwide: 3-year update”, Wellcome Open Research 4(28).


1 Understanding how to encourage young offenders to desist crime is particularly relevant because crime participation significantly decays with age (Doleac 2020). Indeed, individuals under 30 years old have the highest risk of committing a crime (McCall et al. 2013, Ulmer and Steffensmeier 2014).

2 A nice feature of our research design is that we do not need to exploit variation on staggered church openings to identify the effects of interest. Thus, we can abstract from the challenges highlighted by Goodman-Bacon (2021) in the context of two-way fixed effect (TWFE) specifications.

3 The effect of organisations that provide support with alcohol and drug abuse rehabilitation is not statistically significant. There are fewer of such organisations in our sample, which reduces the statistical power of our analyses.