Welfare Cuts and the Criminal Spillovers to Neighbours of Welfare Recipients
Introduction
When Denmark’s newly elected government took office in 2001, one of its first fiscal measures was to cut the cash allowance for newly arrived refugees. The reform, known as Start-hjælp (“Start Help”), took effect on July 1, 2002 and reduced a married couple’s monthly benefit from about USD 1,370 to roughly USD 850—a one-third cut that applied only to refugees who received a residence permit on or after that date. Earlier studies show that the reform increased the criminal behavior of the affected recipients and their families (Dustmann et al., 2024a; Dustmann et al., 2024b).
Jinkins, Kuka, and Labanca (2025) show that the effects extend beyond the recipients’ household to individuals who lived in the same building as welfare recipients, referred to as “neighbors” in the paper. In doing so, the paper also quantifies how these neighborhood spillovers affect the marginal value of public funds (MVPF) associated with the reform. Although the reform reduced the welfare payments received by refugees, the MVPF should be interpreted as a comparison between the benefits and costs of maintaining higher welfare payments. Costs and benefits in the MVPF calculation are expressed in 2021 US dollars. To make results comparable across the separate samples of refugees and neighbors, the paper considers total, rather than average, changes in costs and benefits generated by the reform.
MVPF = 1.2
Net Cost
The reform generated fiscal savings for the government by reducing welfare expenditures, which the paper estimates directly for the average refugee. From these savings, the paper subtracts the forgone tax revenue on welfare transfers (which are taxable in Denmark) and add the increased tax revenue that stems from the higher labor-force participation and earnings of refugees induced by the reform (USD 1,213,306). The paper applies a tax rate of 38 percent—the bottom-bracket rate at the time of the reform—to both welfare benefits and labour income. The paper assesses these fiscal effects over a 10-year horizon and estimates a net savings of USD 43,761,460 for the 5,292 refugees in the sample.
From these savings the paper deducts the costs of increased crime. The paper considers two broad categories: costs to taxpayers and costs borne by crime victims. Following Deshpande and Mueller-Smith (2022), the paper treats the former as part of the net fiscal cost, while the latter enter the MVPF as an increase in society’s willingness to pay for welfare (see below).
Because the paper finds no significant effect on imprisonment, taxpayer costs in this setting consist solely of enforcement and prosecution expenses. The paper obtains crime-specific cost estimates from the Danish State Prosecutor’s Office and aggregate them to the property-crime and non-property-crime levels by weighting each crime’s unit cost by its sample frequency among refugees or neighbors affected by the reform. The paper computes separate cost totals for crimes committed by refugees and by non-Danish neighbors, based on the estimated reform effects on property and non-property crime counts and the unit costs described above. The paper includes only those crime categories for which the reform has a statistically significant effect: property crime for refugees and both property and non-property crime for non-Danish neighbors.
Using this approach, the paper estimates that, over the 10-year period following the reform, enforcement and prosecution costs amount to USD 55,478 for crimes committed by refugees and USD 1,914,913 for crimes committed by neighbors.
The paper uses these estimates to calculate the net cost of providing the higher benefit amount. Net costs of the greater benefits are then 43,761,460 – (55,478 + 1,914,913) = 41,791,069.
Willingness to Pay
Following the approach used in related studies (e.g., Deshpande and Mueller-Smith 2022), the paper defines a recipient’s willingness to pay (WTP) as the reduction in transfers caused by the reform—that is, the amount a recipient would pay to retain the benefits that were cut. Because welfare transfers are taxable in Denmark, the paper subtracts the tax that would have been owed on those transfers, using the 38 percent marginal rate. Over the ten years after the reform, the aggregate WTP for the 5,292 refugees in our sample is USD 42,548,153.
To obtain society’s total WTP for higher benefits, the paper adds the costs borne by crime victims. In the baseline, most conservative scenario, the paper includes only crimes that result in convictions. Unit victim-cost figures by crime type are taken from Deshpande and Mueller-Smith (2022). The paper aggregates these to the property-crime and non-property-crime levels by weighting each unit cost by the frequency of that crime among refugees or their neighbors who were affected by the reform. Separate totals are calculated for crimes committed by refugees and by non-Danish neighbors, using the reform’s estimated effects on crime counts. Only crime categories with statistically significant reform effects are included—property crime for refugees and both property and non-property crime for non-Danish neighbors. Under this approach, the paper estimates victim costs of USD 238,731 for crimes by refugees and USD 6,471,396 for crimes by non-Danish neighbors.
Total willingness to pay is then 42,548,153 + 238,731 + 6,471,396 = 49,258,280.
The paper also considers a broader scenario that accounts for crimes without convictions, which still impose losses on victims. Here the paper scale the conviction-based victim costs by the inverse of the conviction rate (i.e. the share of reported crimes that result in conviction); all other steps mirror the baseline calculation. Under this specification, the paper estimates victim costs of USD 4,334,628 for crimes by refugees and USD 26,930,515 for crimes by non-Danish neighbors, and total willingness to pay is 42,548,153 + 4,334,628 + 26,930,515 = 73,813,296.
MVPF Summary
Taking the ratio of willingness to pay to taxpayer savings net of crime costs, the paper obtains an MVPF of 1.179 for the reform. In other words, society is willing to pay USD 1.179 for each additional dollar of welfare support. This figure excludes losses borne by victims of crimes that are reported but do not lead to conviction. When those costs are included, the MVPF rises to 1.766.
The paper also provides MVPF estimates under several alternative scenarios, including the potential effects of reduced migration prompted by the reform.
References
Deshpande, Manasi and Michael Mueller-Smith (2022). “Does Welfare Prevent Crime? the Criminal Justice Outcomes of Youth Removed from SSI.” The Quarterly Journal of Economics, 137(4): 2263–2307. DOI: https://doi.org/10.1093/qje/qjac017
Dustmann, Christian, Rasmus Landersø, and Lars Højsgaard Andersen (2024a). “Refugee Benefit Cuts.” American Economic Journal: Economic Policy, 16 (2): 406-441. DOI: https://www.aeaweb.org/articles?id=10.1257/pol.20220062
Dustmann, Christian, Rasmus Landersø, and Lars Højsgaard Andersen (2024b). “Unintended Consequences of Welfare Cuts on Children and Adolescents.” American Economic Journal: Applied Economics, 16(4), 161-185. DOI: https://www.aeaweb.org/articles?id=10.1257/app.20230519
Jinkins, David, Kuka Elira, and Claudio Labanca (2025). “Welfare Programs and Crime Spillovers.” NBER Working paper 33926. https://www.nber.org/papers/w33926
Policy Features
- Category
- Criminal Justice
- Sub-Category
- Cash Transfers
- Beneficiary Type(s)
- Adults
- Country of Implementation
- Denmark
- Empirical Method
- Regression Discontinuity
- Research Type
- Primary
- Peer Reviewed
- No
- MVPF Publication Link
- www.nber.org/papers/w33926