Prostitution and Violence: Evidence from Sweden

Perlotta Berlin, M., Immordino, G., Russo, F., & Spagnolo, G, Prostitution and Violence: Evidence from Sweden, Centre for Economic Policy Research, (August, 2020). Available at:

Findings from Advocating Opportunity’s “END DEMAND LITERATURE – DESK REVIEW” (Compiled by: Emily Dunlap, Komal Hans, and Donna Hoffman with contributions from: Kate D’Adamo and Megan K. Mattimoe):

“The Swedish law on prostitution criminalizes buyers but not sellers of sexual services and has been credited with reducing street prostitution significantly while protecting sex workers. As a consequence, this “Nordic model” is spreading rapidly around the world, but as yet there has not been any rigorous evaluation of its effects on sex market participants or on society at large. We document an increase in violence against women following the law’s introduction, ascribable mostly to domestic violence rather than violence against sex workers. We propose a model that rationalizes our findings and simulates the effects of alternative policies. The results also suggest that the lockdown on prostitution markets may have been one determinant of the surge in domestic violence observed during the Covid-19 crisis.” (Abstract, p. 3). (emphasis added).

“All in all, our evidence suggests that intimate partner violence and violence against women in general may have increased as a consequence of the Nordic model. This finding also resonates well with the observation that during the Covid-19 crisis, when sex markets practically shut down, domestic violence seems to have increased dramatically, as we discuss more extensively in the conclusions.” (pp. 3-4).

“When Canada adopted a version of the Nordic model in 2014, the Department of Justice stated that ‘overall objectives [of the reform] are to: Protect those who sell their own sexual services, protect communities, and especially children, from the harms caused by prostitution; and reduce the demand for prostitution and its incidence.’ The analysis in this paper suggests that, in Sweden, the Nordic model resulted in a smaller and probably safer prostitution market, in line with the first and third objectives. However, in relation to the second objective, it also produced a previously neglected increase in violence outside the prostitution market, in the form of domestic violence against women.” (pp. 37-38) (emphasis added).

“Our primary aim was to inform the debate on the effects of different legal regimes of prostitution. We find that the benefits of the Nordic model, namely a smaller and probably safer market for sexual services, have been accompanied by an increase in other forms of gender-based violence.” (p. 38).

“To isolate the effects of the law on the relevant outcomes, we compare the Swedish counties that are above and below average in women’s representation in police forces and among elected officials (respectively “treated” and “control” counties), two indicators that previous studies have found drive greater reporting and lower incidence of crimes against women (Iyer et al., 2012; Miller and Segal, 2018). Looking at population wide changes in rates of violence against women in Sweden before and after the 1999 prostitution law, we observe that assaults against women committed indoor by acquaintances are about 10% more numerous in treated than control counties.” (p. 3).

“This effect of the Nordic model on violence appears consistent both with the negative assessments of the policy reports cited, and with a simultaneous and independent study by Ciacci (2019) on rapes using a different methodology. But while Ciacci cannot identify the victims, and the reports are typically concerned with an increase in violence against prostitutes, the increase in violence that we measure in connection with the 1999 prostitution law affects women other than sex workers. We only find significantly more assaults indoors, perpetrated by acquaintances. Nor is there any simultaneous increase in convictions of clients in treated counties, where we observe more reports of assaults. That is, the greater violence we observe is likely not in the sex market but against non prostitutes, presumably perpetrated by frustrated former clients. In other words, we appear to have documented a negative externality for the community of criminalizing prostitution.” (p. 3).

To lend more formal support to our interpretation, we propose an equilibrium model of the Swedish prostitution market and violent behavior by clients in and outside the market under various law enforcement regimes. The model is calibrated using all information available to us on the Swedish population in order to perform comparative statics and counterfactual analysis. Specifically, we evaluate numerically the effects of various policies on equilibrium quantity, violence and violence risk inside and outside the market, reporting risk and total harm (p. 4).

“Until 1999 prostitution in Sweden was neither illegal nor regulated, but procuring sexual services and human trafficking were illegal. A ban on buying sexual services was first proposed to the Parliament on 5 February 1998, in a package with other “measures to combat violence against women, prostitution, and sexual harassment at work.” (Kvinnofrid, or Women’s Integrity, Proposition). The purchase of sexual services then became the subject of a separate law, the so called Sexköpslagen, approved on 4 April and in force as of 1 January 1999. The rest of the proposal gave rise to a host of other policies, recommendations and missions to various government agencies.” (p. 7).

“The new crime was punishable with fines or prison for up to 6 months, increased to a year in July 2011. There have been many policy assessments — by government agencies and police, feminist groups, and Swedish and international NGOs — but no rigorous study. However, the reports provide useful descriptive information on the prevalence of prostitution in general, on the enforcement effort, and on potential changes in the behavior of clients and sex workers, in particular reporting.” (p. 7).