The Unintended Consequences of Welfare Reforms: Universal Credit, Financial Insecurity, and Crime

Unemployment benefits (UBs) are a key component of modern welfare states. The primary motivation behind UB is to insure individuals against job losses, helping them to deal with employment shocks in the presence of credit constraints. Regardless of their importance, UB schemes are often under scrutiny. This is because they potentially increase recipients’ reservation wages, lowering search efforts and the exit rate from unemployment to jobs (e.g., Meyer 1990; Hunt 1995; Card and Levine 2000). In light of these concerns, and to reduce government spending, many countries are considering measures that increase the stringency of their benefits schemes and incentivize recipients’ transition toward employment (Andersen et al. 2019; Wickham et al. 2020). It is therefore important to broaden our knowledge on the impact of welfare policies affecting the financial position of people at the bottom of the income distribution. This can help design cost-effective regulatory frameworks and alleviate the risks associated with unintended policy responses.

This article focuses on Universal Credit (UC), an ongoing and monumental reform of the UK social security payment system. UC is a payment for people over 18 years but below State Pension age who are working on a low income or are entirely out of work. It also includes support for the cost of housing, children and childcare, financial support for people with disabilities, and carers and people too ill to work. UC was introduced in 2013 and, by the end of 2024, it will fully replace the still operative Legacy system by combining six benefits for working-age people with low or no income currently administered separately (Kennedy and Keen 2018).

UC aims to offer social protection whilst: reducing the complexity of the welfare system, encouraging recipients’ financial responsibility, decreasing welfare dependency, and incentivizing the transition toward employment. To achieve these objectives, UC introduced major changes that, we argue, have significantly affected the financial position of its recipients.

First, UC is assessed and paid in arrears with an initial minimum waiting period of 5 weeks. In addition, longer initial payment delays have often arisen due to implementation issues. As of 2017, the UK National Audit Office found 40% of claimants waited over 11 weeks from application to first payment (National Audit Office 2018). Numerous practitioners have suggested the initial waiting period and subsequent unintended payment delays have led to acute financial hardship (National Audit Office 2018; Cheetham et al. 2019). Second, UC has modified the structure of disbursements which now involve monthly payments, with the housing component of the benefits paid directly to the recipient.¹ This could affect recipients lacking budgeting skills, who may be unable to smooth consumption over a month (e.g., Foley 2011; Carr and Packham 2019). Furthermore, some claimants may use the housing component of their benefits for other expenses, thus failing to make rental payments (Hardie 2021). Finally, UC has implemented a more stringent conditionality regime, introducing severe sanctions which reduce payments to beneficiaries not adhering to its stricter guidelines, for instance not displaying sufficient effort in looking for jobs (Dwyer and Wright 2014; Webster, 2017).

This article contributes to the Economics literature by providing the first empirical evaluation of the impact of UC employing quasi-experimental methods.² In particular, we exploit the staggered rollout of the reform across parliamentary constituencies. The rollout initially targeted new claimants: single, unemployed, nonhome-owning, and without children (therefore predominantly young males).³ To conduct the analysis, we have assembled a monthly database of all 573 constituencies in England and Wales from December 2010 to February 2019. Using difference-in-differences, triple differences, event studies, and recent advances in the literature examining causal estimation with time-varying policy reforms (Goodman-Bacon 2018) we can identify credible estimates of the impact of UC.

Our analysis first documents UC has led to a significant increase in court repossession orders and eviction warrants in the social housing sector, with no effects detected on mortgages or private housing. These findings are consistent with most UC recipients renting relatively more affordable flats through social housing (Department for Work and Pensions 2019). The event study analysis reveals “well-timed” and long-lasting effects, suggesting UC has led to an increase of around 6000 repossession orders. These results provide direct evidence of the negative socio-economic effects on people at the bottom of the income distribution, also delivering a solid basis for our examination of UC’s criminogenic impacts.

Becker’s application of rational utility models to criminal choices made by individuals (1968) predicts a reduction in legitimate income opportunities, such as the one experienced by recipients registered on a more stringent welfare regime, will incentivize criminal behavior motivated by economic gain. In line with this prediction, our empirical analysis shows UC has led to a significant rise in acquisitive crimes. Effects are driven by an increase in burglaries and vehicle crimes (i.e., theft of and from a vehicle). Our findings are corroborated by placebo tests randomly permuting the date of UC adoption across constituencies. They also hold against numerous robustness checks such as the inclusion of an extensive set of socio-economic covariates, constituency-specific trends, constituency-by-month fixed effects, and changes to the sample of analysis and functional forms used. Moreover, an event study analysis reduces concerns pre-trends may be confounding the causal interpretation of the results or inflating our estimates. The event study analysis also shows the criminogenic effects of the reform emerge on impact and last for almost 3 years, until the end of the sample period. Our estimates suggest UC has caused around 35,000 burglaries and 25,000 vehicle crimes, at a societal cost of £465 million.

Our findings are of timely importance for at least two reasons. First, during the period of our analysis 1.5 million first-time recipients claimed UC, but the program will directly affect the lives of more than eight million people, who will transition from the Legacy system onto UC, by the time the latter is fully implemented in 2024 (Kennedy and Keen 2018). Second, other countries are currently evaluating reforms mirroring UC’s practices (Wickham et al. 2020). The need for such reforms has recently intensified because of the augmented pressure welfare systems experienced during to the coronavirus pandemic and the following cost of living crisis. Thus, our study of a specific UK measure may have broader applicability informing policymakers worldwide regarding the potential unintended consequences of welfare interventions affecting the incentives to offend for low-income individuals at the margins of crime.

Our work documents the criminogenic effects of a large-scale UK welfare reform by exploiting implementation features (e.g., the staggered roll-out) and newly assembled granular data covering 573 constituencies observed monthly for a period of around 8 years. Our article is closely related to the work of Machin and Marie (2006) who study the impact of the Jobseeker’s Allowance (JSA) introduction in the United Kingdom in 1996. Analogously, JSA simplified the UK employment benefit system and brought a more stringent welfare regime. Using a mix of quasi-experimental designs and qualitative evidence on 45 areas in England and Wales observed quarterly, the researchers show crime rose more in areas where a greater proportion of claimants were affected by the JSA introduction, a result that is consistent with our findings and is strongly suggestive of the link between crime, benefit cuts, and sanctions.

The literature employing quasi-experimental designs to study the criminogenic impacts of changes in the structure of welfare payments and austerity measures offers an important backdrop to our findings. In our study, we show evidence suggesting stricter sanctions and the payment of the housing benefits to the claimant, rather than the landlord, are channels through which the welfare reform may affect crime. Other studies have focused on aspects such as changes in payments frequency (Foley 2011; Carr and Packham 2019), duration of temporary benefits (Bindler 2016), benefit cuts (Andersen et al. 2019; Fetzer et al. 2019; Giulietti and McConnell, 2021; Melander and Miotto 2021), food stamps ban (Tuttle 2019), and the receipt of universal basic income (Watson et al. 2020). Ultimately, our work provides a different context to previous studies, thus extending the external validity of the literature identifying criminogenic impacts of welfare interventions affecting the financial position of people at the bottom of the income distribution.⁴

Finally, our study documents a stark increase in the number of evictions from social housing, contributing to the literature examining the impact of welfare reforms on housing security and other socio-economic outcomes. A closely related paper is Fetzer et al. (2019) who examine the impacts of a cut to rent subsidies for low-income families in the United Kingdom in April 2011. They show the policy generated an increase in evictions, leading to a negative spiral of individual bankruptcies, statutory homelessness, rough sleeping, and—consistent with our results—property crimes. An extensive literature has documented how evictions have long-lasting negative effects on consumption and access to credit (Humphries et al. 2019), mental and physical health (Burgard et al. 2012; Fowler et al. 2015), achievement of children (Chyn 2018), and labor markets outcomes (Desmond et al. 2016). In light of this evidence, we believe...