While crime is traditionally considered a young man’s game, the number of crimes committed by older adults has climbed disproportionately in recent decades.1 The main emphasis of the crime economics literature to date has been on factors with an early but long-lasting impact on criminal paths, such as education, family background, and opportunities upon entering the labour market (e.g. Cullen et al. 2006, Heckman et al. 2013). However, the rising share of crimes committed by older adults, often with a previously clean record, calls for a better understanding of late-in-life determinants. In Andersen et al. (2021), we document the effects of one of the most impactful and widespread adverse events over the life cycle – namely, cancer. To that end, we leverage rich administrative data from Denmark that allow us to link health and criminal records at the individual level and empirically explore whether (and why) affected individuals ‘break bad’.
Cancer causes crime
We explore the impact of health shocks on criminal activity in a large database covering the entire Danish population. The setting provides us with an ideal testing ground to conduct this analysis. First, detailed information at the individual level allows us to follow the medical history and criminal record of each person over the course of his or her life. Second, thanks to universal health insurance, the medical costs of cancer treatment are fully covered for all Danes. This allows us to exclude an almost mechanical channel through which a person is forced by personal bankruptcy to resort to illegal activities. Our focus, instead, is on how changes in expectations about future income streams and life expectancy alter the economic incentive to commit crime.
In Figure 1, we show that the probability of committing a crime increases on average by 13% following a cancer diagnosis. This effect is subdued in the immediate years after diagnosis but intensifies over time and persists for more than ten years. Furthermore, we show that cancer not only drives repeat offenders to increase the number of violations they commit, but also leads individuals without a criminal record to violate the law for the first time. Furthermore, we document an increase in the crime propensity of (healthy) spouses of cancer patients. We rule out the possibility that these effects are driven by the impact of cancer on the probability of being arrested. To identify a causal effect relation, we compare in each year the probability of breaking the law for individuals diagnosed with cancer with that of similar individuals (same gender, age, and year of observation) who are not yet diagnosed with cancer but will be in the future.
Figure 1 The effect of cancer on crime
Why does cancer increase crime?
To explore the potential channels, we rely on Becker (1968) and Ehrlich (1973) rational theories of crime. One prediction of these theories is that the decision to commit a crime depends on an array of factors that include the difference between the remuneration of legal and illegal activities, the perceived probability of punishment, and the personal attitude towards risk. As health shocks affect each of these dimensions, we consider them one at a time.
First, we uncover evidence for an economic motive. In particular, most of the crimes that follow a cancer diagnosis are of an economic nature (e.g. stealing, fraud, dealing) and, to a significantly lesser extent, of a violent nature (e.g. murder, assault). We collect additional evidence for an economic motive by exploring the heterogeneity in the impact of cancer on crime for different subgroups of the population. We find that people who experience an income decline after cancer tend to turn to illegal activities. Relatedly, those more vulnerable to the economic repercussions of cancer – i.e. individuals who lack pre-existing insurance in the form of financial wealth, home-equity, education, or marriage – are more likely to ‘break bad’.
Second, our analysis identifies the existence of a survival probabilities channel: Individuals for whom cancer induces a larger decrease in survival probabilities – and who are therefore less likely to face punishment for their criminal offenses – are more likely to violate the law (Figure 2). By contrast, we do not find support for a preference channel. Specifically, we rely on elicited measures of risk preferences from two experiments conducted between 2003 and 2010, which we match to a subset of the individuals in our sample. We find that, if anything, cancer increases risk aversion rather than decreasing it.
Figure 2 Survival probability channel
Who breaks bad?
We find the strongest response to health shocks for men, singles, and people with less education (the latter finding is in line with Bell et al. 2018). In particular, the propensity to commit crime after cancer is five times larger for men than for women. The heterogeneity based on education and marital status is also striking, as married people or people with above median education are less likely to start committing crimes following a severe health shock. Furthermore, individuals who have family members with a criminal record are more likely to break the law themselves after a cancer diagnosis. In general, this latter result confirms previous findings in the literature that individuals with criminal peers are more likely to commit crime (Bilings et al. 2016).
Can policy interventions mitigate the negative impact of cancer on crime?
We find that welfare policies can alleviate the negative externality induced by health shocks. The evidence for this result is based on an administrative reform that reallocated decisional authority on social policies across Danish municipalities as an exogenous source of variation in welfare support. We document that a decrease in the generosity of social security fosters an increase in the sensitivity of crime to health shocks. Individuals who experience the largest reduction in economic subsidies due to the reform increase their crime rates by roughly twice as much following a cancer diagnosis. This shows the importance of social policies in mitigating the relationship between illegal activities and health shocks, even when those policies are not designed to fight crime or improve healthcare.
In sum, we document an important and surprising driver of criminal activity: adverse health events. This finding adds to a growing literature on the determinants of crime (Dix-Carneiro et al. 2017, Foley 2008, Machin and Meghir 2004, Pinotti 2020). In the context of the Covid health crisis, it is of paramount importance to develop a better understanding of the implications of health shocks. While most countries have so far experienced an overall contraction in crime – as lockdowns restricted the ability of individuals to commit offences – our results suggest that this trend may be reversed in the near future. As people are left to cope with the effects of the virus and welfare support expires, the more vulnerable segments of the population may find themselves drawn to illegal activities if social support is not in place.
Andersen, S, G Parise and K Peijnenburg (2021), “Breaking Bad: How Health Shocks Prompt Crime”, CEPR Discussion Paper No. 15899.
Becker, G S (1968), “Crime and Punishment: An Economic Approach”, Journal of Political Economy 76: 169–217.
Bell, B, R Costa and S Machin (2018), “Why Education reduces Crime”, VoxEU.org, 14 October.
Billings, S, D Deming and S Ross (2016), “Neighbourhood Spillovers in Youth Crime: Social Interactions Matter”, VoxEU.org, 11 July.
Cullen, J B, B A Jacob and S Levitt (2006), “The effect of school choice on participants: Evidence from randomized lotteries”, Econometrica 74(5): 1191–1230.
Dix-Carneiro, R, R Soares and G Ulyssea (2017), “Economic Shocks and Homicides in Brazil”, VoxEU.org, 31 August.
Ehrlich, I (1973), “Participation in Illegitimate Activities: A Theoretical and Empirical Investigation”, Journal of Political Economy 81: 521–565.
Foley, F (2008), “Welfare payments, liquidity constraints, and crime”, VoxEU.org, 5 August.
Heckman, J, R Pinto and P Savelyev (2013), “Understanding the mechanisms through which an influential early childhood program boosted adult outcomes”, American Economic Review 103(6): 2052–86.
Machin, S and C Meghir (2004), “Crime and economic incentives,” Journal of Human Resources 39: 958–979.
Pinotti, P (2020), “Burden of Proof: Measuring and Understanding Crime”, VoxEU.org, 1 August.