Most research on what pushes people toward crime focuses on familiar suspects: unemployment, poverty, family breakdown, or exposure to violence. Serious illness rarely makes the list. Yet a diagnosis of a life-threatening disease can upend a person’s finances, shorten their time horizon, and shake their mental health all at once. Could something like a cancer diagnosis actually change whether someone breaks the law?
That is the question at the heart of an article in the American Economic Journal: Applied Economics. Drawing on decades of Danish administrative records linking health, income, and criminal histories, the authors find that a cancer diagnosis is associated with a measurable and lasting rise in the probability that a person commits a crime.
The puzzle of health and crime
Steffen Andersen of Copenhagen Business School, together with Elin Colmsjö of the University of Copenhagen and Gianpaolo Parise and Kim Peijnenburg of EDHEC Business School, wanted to test an old idea from economics. Gary Becker and Isaac Ehrlich argued decades ago that people weigh the benefits of a crime (money, for instance) against its costs (punishment, and the time horizon over which that punishment matters). Health problems, in principle, could shift every part of that equation: income from legal work might fall, the cost of a future prison sentence might matter less if survival is uncertain, and risk tolerance or mental state might change as well.
Testing this empirically is hard. People in poor health differ from healthy people in many ways that also predict crime, from childhood environment to lifestyle. To get around the problem, the authors focused on one specific event: a first cancer diagnosis. About 40% of Danes develop cancer at some point, across all demographic groups, and for any given individual, the exact year of diagnosis is largely unpredictable.
How the study was built
The team combined several Danish national registries covering the full population. They pulled health records from the National Patient Registry, criminal charges and convictions from the Central Crime Registry, and detailed information on income, employment, wealth, and family status from tax and demographic registries. The final sample included roughly 368,000 people diagnosed with cancer between 1980 and 2018, observed in a window of ten years before and after diagnosis. The analysis was limited to people aged 18 to 62 to avoid confusion with retirement effects.
The key methodological move was to compare people diagnosed in a given year with people of the same age who would eventually get cancer but had not yet been diagnosed. This comparison, combined with fixed effects for each individual and each age-by-year combination, isolated the effect of the diagnosis itself from background traits and aging.
What the data showed
In the year of the diagnosis, criminal activity actually dipped. Treatment is physically demanding and often involves extended hospital stays, and many patients also have savings that cushion the initial blow. But starting roughly two years after diagnosis, the probability of a criminal conviction rose, climbed for several years, and then stayed elevated for more than a decade.
On average, a cancer diagnosis was linked to a 14% increase in the probability of a conviction in any given year, compared with a baseline annual rate of 0.69%. The effect appeared both among people with no prior criminal record (a 14% rise in first-time offenses) and among those with a record (a 14% rise in re-offenses).
Three pathways from diagnosis to crime
The authors then looked at why this pattern emerges. They found evidence for several overlapping channels.
The first is economic. A cancer diagnosis was associated with a drop in salary income of about 4.8%, a decline in total income of 1.5%, and a small but persistent drop in employment. The rise in crime was largest for groups most exposed to these financial hits: people who do not own a home, single individuals, men, and those with less education. Most of the additional crimes were property offenses like theft and fraud, which fit an economic motive. Notably, the authors found that people did not turn to the most serious crimes. Homicide, aggravated robbery, and similar offenses did not rise. The additional offenses were smaller-stakes crimes: shoplifting, minor assault, vandalism, fraud.
The second pathway involves survival. If a person expects to live a shorter life, the threat of future punishment carries less weight in today’s decisions. To test this, the researchers estimated each patient’s expected decline in five-year survival probability based on cancer type, age, gender, and marital status. People whose survival probabilities dropped more sharply were more likely to commit crimes afterward, though the very sickest patients committed fewer crimes, likely because severe illness limits what someone can physically do.
The third pathway involves mental health. Cancer patients were more likely to seek psychological help after diagnosis, and the cancer-crime link was about 2.5 times stronger among those who did. The authors found no evidence that cancer changed people’s underlying risk preferences, based on a smaller sample of patients who had participated in risk-attitude experiments.
Interestingly, while non-economic crimes like violence started from a much lower base, they rose by a larger percentage (38%) than economic crimes (14%). The authors interpret this as a sign that financial motives alone cannot explain the pattern.
Do welfare policies change the picture?
The final part of the study examined whether government support could soften the effect. In 2007, Denmark consolidated 271 municipalities into 98, shifting decision-making authority over social policies. Some municipalities became more generous to cancer patients after the reform, others less so. The authors measured these changes and compared crime responses in “stingy” versus “generous” municipalities.
The impact of a cancer diagnosis on crime was substantially larger in places where social support was cut. The additional effect of living in a stingy municipality was roughly 0.08 percentage points, on top of the baseline rise. The authors interpret this as evidence that the economic channel is real and that welfare programs can partially offset it.
Caveats
A few limitations are worth keeping in mind. The study measures convictions, not all crimes committed, so some of the rise could in principle reflect changes in how likely patients are to be caught. The authors run a series of checks against this possibility, including examining the time between offense and apprehension and showing that less generous welfare policies (which should affect actual criminal incentives, not detection) amplify the effect. They argue these results point to a genuine rise in offending rather than an artifact of enforcement.
The findings also come from Denmark, a country with universal health insurance and an extensive safety net. In settings where patients face large out-of-pocket medical bills, the mechanisms and magnitudes could look quite different. The researchers note that an out-of-pocket medical expense channel is unlikely to drive their results, precisely because Danish patients do not face such bills.
The average age of a cancer patient in the sample is 52, considerably older than the ages at which events like marriage, childbirth, or job loss tend to occur. That makes health shocks a distinctive trigger for crime later in life, a period that other research on “turning points” has largely left unexplored.
