Evidence assessment of the impacts of the criminalisation of the purchase of sex: a review

Impacts Of The Criminalisation Of The Purchase Sex

The impacts of the criminalisation of the purchase of sex will be explored examining empirical evidence specifically focusing on the following issues: estimated numbers involved in prostitution (selling sex) ^[22] ; attitudinal change; application of the law; demand for the purchase of sex; impacts on individuals involved in prostitution; uptake of services to exit prostitution.

Estimated numbers involved in prostitution ^[23]

It is difficult to establish a complete picture of the numbers involved in prostitution as there are several methodological challenges to estimating accurate numbers involved in the sale of sexual services (Mujaj and Netscher, 2015). Holmström and Skilbrei (2008) note the complicating factors associated with increases in the number of foreign women involved in prostitution; cases of human trafficking for the purposes of commercial sexual exploitation and increasing online opportunities, which make it difficult to monitor the scope of prostitution. Internationally, most attention and estimates of numbers involved in prostitution focus on street-based prostitution. This means that potentially the largest forums for the sale of sex are overlooked, and also results in the circumstances of individuals involved in street- based prostitution being expanded to discussions of prostitution more generally.

According to SOU (2010) prostitution units in Stockholm and Gothenburg estimated that 50% of women engaged in street prostitution were foreign nationals. Estimates of the prevalence of street prostitution and online prostitution are available with data on indoor prostitution (in hotels etc) more difficult to estimate (Mujaj and Netscher, 2015). Despite these limitations, available evidence indicates that there has been a downward trend in street prostitution in Sweden since the 1970s.

The National Board of Health and Welfare in Sweden conducted three surveys on the impact of criminalisation of the purchase of sex focused on the perspectives of state authorities in 2000, 2003 and 2007 (Socialstyrelsen, 2004, 2008). For the third report, surveys were distributed to all police authorities in Sweden and to a sample of Swedish municipalities and to youth clinics (Socialstyrelsen, 2008). Follow-up questions were addressed to 25 municipalities and 12 police authorities that had knowledge of prostitution in their area. Responses to the follow-up questions by these authorities suggested that there had been an initial reduction in prostitution with no obvious increases up to 2006; although it was noted that prostitution arranged by phone and online had become more common. Interviews with representatives from public authorities and non-government organisations ( NGOs) suggest that initially after the Act, street-based prostitution had almost disappeared and then later returned but to a significantly lesser extent (Socialstyrelsen, 2008).

In 2014, a study was commissioned to analyse the spread, extent and forms of prostitution in Sweden (Mujaj and Netscher, 2015). The study involved a scoping survey gathering old and new knowledge on prevalence of both street and online prostitution from a range of stakeholders. Mujaj and Netscher (2015, p. 16) note that as figures are from different sources this should be borne in mind when comparing changes over time. They state: "The estimates from 2010 and onward indicate that street prostitution is relatively constant, but the basis of these findings contain many uncertainty factors since the figures from the Prostitution Units for 2010-2014 cannot be compared with previous surveys". Taking such statistical considerations into account, Mujaj and Netscher (2015) provide figures (Table 1) which indicate that estimated street prostitution has halved in Sweden since 1995.

Table 1: Estimated number of individuals engaged in street prostitution.

Year	Number of individuals
1995	650 women ( SOU, 1995)
1998/99	The introduction of the Act prohibiting the Purchase of Sexual Services meant that prostitution more or less vanished completely from the streets (National Board of Health and Welfare, 2000)
2008	300-430 women ( SOU 2010:49) [24] .
2010	200 women (contacts with the Prostitution Units).
2011-2014	200-250 women yearly (contacts with the Prostitution Units).

(Mujaj and Netscher, 2015, p. 16)

Holmstrom and Skilbrei (2008) who provided estimates of 300 women in street prostitution, and 300 women and 50 men who advertise on the internet note, it is primarily social workers who are responsible for evaluating the circumstances of the prostitution market hence most of the focus is on prostitution encountered in three of Sweden's largest cities through encounters with individuals working on the streets, or who access counselling and therapy via social work services. Holmström draws together sources from the National Board of Health and Welfare (Socialstyrelsen) and the national Criminal Investigation Department, responsible for monitoring developments in human trafficking, to assess the figures cited above ^[25] .

Evidence from Sweden (Swedish Institute, 2010: 7) ^[26] drawing upon available data on the scale and prevalence of street prostitution indicates that street prostitution has been reduced by half since the introduction of the ban on the purchase of sexual services while in Norway and Denmark, numbers increased dramatically over the same period. Comparing the figures to Denmark where sex purchase is legal and where the population is 5.6 million compared to Sweden's 9.4 million, Holmström and Skilbrei (2008) suggest the numbers involved in prostitution in Sweden are one-tenth of that in Denmark. Holmström and Skilbrei (2008) note that it is difficult to know the extent to which successive declines in street prostitution in Sweden can be attributed to the legislation and how much to developments in communication methods such as increases in internet use. However they do note that the prostitution market has changed with increasing differentiation in practice. Writing in 2013, Skilbrei and Holmström note that the more dispersed nature of the Swedish prostitution market makes it more difficult to estimate.

Dodillet and Östergren, who are critical of the criminalisation of the purchase of sex ^[27] , (2011, p. 9) state that some authorities (citing the Swedish National Council for Crime Prevention, 2008) claim that the 'fluctuation (and therefore any claimed decrease)' of street prostitution may be a consequence of police surveillance and enforcement i.e. police priorities and practices rather than the legislation itself. However, it is likely that police priorities are influenced, if not determined, by the law. Indeed Levy (2015) notes that the Swedish model is not homogenous, with variability in how it is interpreted, enforced and resisted in policy and practice among stakeholders and across Sweden ^[28] .

Taking limitations of the available evidence into account, there does appear to be sufficient evidence to indicate that street prostitution in Sweden has declined following the introduction of legislation, with some evidence to suggest that the number of individuals involved in prostitution overall has also reduced (Joe-Cannon, 2006; Ekberg, 2004; Clausen, 2007). Taking into account limitations of data collection, Kelly et al. (2009) in a comparative analysis of nine 'prostitution regimes' ^[29] indicate that Sweden appears to have the smallest number of women selling sex, even taking population into account (see also Jakobsson and Kotsdam, 2013).

Counter-claims that prostitution has gone underground or escort agencies and brothels have increased (Moffatt, 2005) have been put forward, although it is acknowledged that this data is hard to substantiate and may be reflective of developments internationally (in Wilcox et al, 2009). It has also been suggested that displacement to Denmark has occurred although actual evidence for this is tentative and it has been argued, supports claims that criminalising has an effect of reducing street prostitution at least, within nation states, where the legislation is imposed.

In a summary report of the evaluation of the 2009 Norwegian legislation, commissioned by the Norwegian Ministry of Justice and Public Security, it is stated that despite data limitations ^[30] , there is a clear downward trend in prostitution since the law to criminalise the purchase of sex was introduced in Norway (Rasmussen et al., 2014) ^[31] . The biggest decrease has been in the Oslo street prostitution market where, based on systematic field observations, the market has stabilised at 40-65% of the market before the law. An estimate of the fall in the indoor market of 10-20% compared to before the law is also provided although with a higher degree of data uncertainty (Rasmussen et al., 2014).

Online prostitution

SOU (2010) concluded that street prostitution had decreased by half since the Act was introduced and that there were no indications that the decrease in street prostitution had led to an increase in online prostitution. ^[32] However, wider literature identifies a large international increase in the on-line promotion of sexual services (Danna, 2012; Huschke et al., 2014). Dodillet and Östergren (2011) write that access to information on internet advertisements by sex workers is limited. They cite the Malmo Prostitution Knowledge Centre, which measures internet advertising on a regular basis, and which claims that as a result of changing technology, less visible or indoor prostitution makes up four-fifths of overall prostitution prevalence compared to two-thirds prior to the ban. The figure provided by Dodillet and Östergren is based on secondary data and, it could be argued, reflects international shifts from street to indoor prostitution more generally. Dodillet and Östergren (2011) note that it is difficult to know whether displacement has occurred since the Sex Purchase Act in Sweden but that less visible indoor prostitution appears to have increased (as also found by Levy, 2015). As noted above, this seems to reflect developments internationally with an increase in indoor selling of sex and changes associated with technological developments.

In a survey of online advertisements carried out in Sweden over a ten week period in 2014, 6,965 adverts were posted (Mujaj and Netscher, 2015) although the authors note several limitations to this survey including that there may be duplication of adverts. This compares with 304 online adverts in a similar survey conducted in 2006. Mujaj and Netscher (2015) note that this does not necessarily indicate an increase in the number of individuals involved in prostitution; rather that the use of online advertisments has increased. This point is also made by Fredlund et al. (2013) in relation to their study of young people and the sale of sex using a comparison of data between 2004 and 2009 in Sweden ^[33] . Fredlund et al. (2013) question whether the changes noted are the result of differences in the young people who report ever selling sex or whether it is the consequence of a change in society; for example, the internet has become the most common source for selling sex. For Mujaj and Netscher (2015), the sale of sex via the internet has increased more generally.

Change in Attitudes

In the debates and discussions which preceded the introduction of the legislation in Sweden, it was argued by those supportive of the legislation that the most important impact of the law would be on public opinion and attitudes towards prostitution (Kuosmanen, 2010: Wong, 2014). Ekberg (2004: 18 updated paper), noting the low number of arrests and convictions of men in Sweden, points out that: "It is important to remember that the main purpose of the Law is normative" and therefore has a preventative element.

Sweden's policy making has traditionally been based upon consensual and communal policy making and of introducing legislation with a normative intention i.e. where the laws are expected to support the development of new norms which will eventually result in behaviour change. Sweden claims that through the implementation of this legislation, the UN Trafficking Protocol on demand has been addressed. The 2008 Government Communication which presented an action plan against prostitution and human trafficking for sexual purposes emphasised the importance of demand, namely the purchase of sex, and underlined the link between prostitution and human trafficking for sexual purposes.

The introduction of the law criminalising the purchase of sex in Sweden was preceded by public and political debates ^[34] , which no doubt raised awareness of broader issues associated with prostitution and its introduction was accompanied by a campaign to target clients (Ekberg, 2004). Identified funds were allocated to police enforcement.

In terms of public perceptions, Kuosmanen (2010) examined the attitudes of the general public in Sweden to the Sex Purchase Act exploring whether attitudes had been affected by the law. Comparing a survey on public attitudes prior to the Sex Purchase Act (Mansson, 1996) with a survey after the Sex Purchase Act ( SIFO, 1999) in Sweden reveals that initially after the Act, there was greater support for prohibition of both the sale and buying of sex than before the Act (cited in Kuosmanen, 2010).) As Waltman (2011a) notes: "In 1996 a survey study showed that only 45% of women in Sweden and 20% of men wanted to criminalise a male sex purchaser. In 1999 81% of women and 70% of men wanted to criminalise the purchase of sex".

In Kuosmanen's (2010) report of the survey conducted in 2008, participants were asked: 'should we retain the law prohibiting the purchase of sex?' Overall, 71% were in favour of retaining the law whilst 18% were not. There were clear gender differences in attitudes with 79% of women being in favour of retaining the law in comparison to 60% of men. Comparing this to a similar question asked in a survey conducted in 2002 ^[35] , which found that 76% of respondents believed that it should be illegal to buy sex in Sweden ( SIFO, 2002 ^[36] cited in Kuosmanen, 2010). As Waltman (2011a) notes, these two surveys cannot be seen as comparable in the phrasing of this question. The 2008 survey also asked: 'should the sale of sex be prohibited by the law?' Nearly 60% of respondents believed that the sale of sex should be prohibited by law. Again, there were clear gender differences with 49% of men in comparison to 66% of women in favour of prohibiting the sale of sex. Although figures relating to the experience of purchasing of sex amongst survey respondents were very small (n=35) of those, 11 respondents (10 men and one woman) indicated that the legislation had affected their purchase of sex (five indicating they had stopped buying sex as a result of the new law and two indicating they had reduced the frequency of purchase).

As the Swedish Institute (2010) highlights, in each of the three surveys conducted since the purchase of sex was criminalised, more than 70% of respondents viewed the legislation positively although it may be that earlier figures (from the 1996 survey) can be called into question (Wong, 2014). Nevertheless, the subsequent studies indicate a high degree of acceptance of the criminal prohibition of the purchase of sexual service.

Comparing the attitudes of Norway and Sweden, a longitudinal comparative study, using a difference-in-differences methodology ^[37] , aimed to explore the effect of the Norwegian criminalisation of the purchase of sex in 2009 on attitudes towards prostitution (Jakobsson and Kotsadam, 2010). Jakobsson and Kotsadam (2010) hypothesised that the criminalisation of buying sex in Norway would affect attitudes towards prostitution, that there would be a greater effect of the law on younger people and people living in greater proximity to prostitution in Oslo, and, that people who trust politicians are more inclined to change their attitudes in accordance with a legal change. Jakobsson and Kotsadam (2010) posit that a focus on specific groups of young people and Oslo residents enabled them to investigate the role of context in law reform.

Jakobsson and Kotsdam's (2010) internet survey was conducted with a random sample of the public, aged 15-65, over two sweeps in August 2008 and August 2009 in Norway and Sweden. ^[38] Using a control group, the premise was to compare the changes in attitudes among individuals in a country where there has been a change in the law (Norway) to the changes in attitudes among individuals in a similar country without such a change during the period (Sweden). They note variations in attitudes towards criminalisation of buying sex. In Oslo, changes in attitudes were evident and Jakobsson and Kotsadam (2010) suggest that this might be due to prostitution being more visible in Oslo and the effects of increased proximity. Also, younger people were more likely to change their position toward thinking that buying sex should be illegal than older people. However, the authors noted that the study was unable to distinguish between any 'direct' effect of the law and the effect attained via the media debate, which started before the first wave of the survey, although this was related to the legal reform. The appropriateness of Sweden as a control group was also questioned given the introduction of similar legislation ten years previously.

Findings from the Gender Equality Barometer in Finland ^[39] , a survey of public attitudes, experiences and opinions relating to gender equality, show that one in three women and two in three men approve of men and women buying sex from an individual involved in prostitution (Finish Ministry of Social Affairs and Health, 2013). However, the law in Finland differs significantly from the Swedish model and is unlikely to have the same normative effect. Specifically examining the impact of the partial criminalisation of purchase legislation in Finland, which prohibits the purchasing of sex from a victim of human trafficking or procuring, on attitudes towards the purchase of sex, the survey findings suggest that there has been consistency in attitudes after the law was introduced across the majority of the sample. For example, of those who had indicated that, before the enactment of the legislation, they had neutral opinions about men who purchase sex, the majority (89%) maintained neutral views after the law. However, of those who held negative opinions about men who purchase sex prior to the law, just over two thirds maintained their views whilst a third became more negative. In interviews with experts from various units of the police, the prosecution service, the Border Guard, the social services and NGOs, the partial criminalisation of purchase prohibiting the purchasing of sex from a victim of human trafficking or procuring in Finland, was not thought to have had an obvious normative effect on behaviour or attitudes (Niemi and Aaltonen, 2014) ^[40] .

Alongside ongoing support for the criminalisation of the purchase of sex legislation however, as noted above (p32), there also appears to have been an increase in the proportion of respondents who would like to see the sellers of sexual services/individuals involved in prostitution penalised in Sweden and Norway (Rasmussen et al., 2014; Amnesty International, 2016).

Impacts on Demand

The overall picture appears to be one of continued, but decreased demand for prostitution in countries where the purchase of sex has been criminalised. The Swedish Institute (2010) indicates that men state they are less likely to purchase sex in Sweden as a result of the legislation although noting that the number of prosecutions under the legislation is largely dependent on the priorities set by the police and the resources they have available. In Finland, where the purchase of sexual services or offering sexual services against payment is prohibited in a public place and where a partial sex purchase ban was enacted in 2006 prohibiting the purchasing of sex from a victim of human trafficking or procuring (Niemi and Aaltonen, 2014), a survey conducted in 2007 found that purchasing of sex had slightly decreased among young men but had dropped by as much as half among middle-aged men since 1999 (2007 Finsex survey cited in Niemi and Aaltonen, 2014).

The deterrent effect of the law may be anticipated, but is difficult to evidence empirically.

Professionals working with clients in the Swedish context who participated in a study conducted by Levy (2015) who is critical of the legislation, reported that clients do not tend to cite the law as being the reason why they have attended services to access support and notes there is limited evidence as to how the Swedish Sex Purchase Act may have impacted on clients' attitudes towards purchasing sex. Based on their review of research and reports on the criminalisation of the purchase of sex, Dodillet and Östergren (2011) write that some clients were not particularly concerned about the prospect of detection.

Prior to the criminalisation of the purchase of sex in Northern Ireland ( NI), a multi-method study including an online survey and face to face interviews with sex workers and clients as well as phone interviews with nine representatives across NI councils, was conducted to explore the potential impacts of the criminalisation of the purchase of sex (Huschke et al., 2014). A key finding from the survey was that more than one third of Northern Ireland-based clients did not know what the legal context of prostitution was. From face-to-face interviews with clients (N = 10), none had established the legal status of prostitution before paying for sex ^[41] .

Mujaj and Netscher (2015) compared population-based surveys to specifically explore the purchase of sexual services in Sweden (i.e. whether men had bought sex during the past 12 months, rather than whether sex had ever been purchased), in order to measure how many people were 'active' buyers. In 2014, approximately 0.8 percent of the men surveyed had bought sexual services during the past year (Mujaj and Netscher, 2015). This compared with 1.2 percent of the men surveyed in 2011; 1.8 percent in 2008; and, 1.3 percent in 1996 (Mansson, 1998; Kuosmanen, 2008; Priebe and Svedin, 2010 cited in Mujaj and Netscher, 2015).

Supporting the evidence of a reduction in the purchase of sex provided by Mujaj and Netscher (2015), the Nordic Gender Institute (2008) (cited in Claude, 2010: 15) also claim that the number of sex buyers in Sweden has declined since the introduction of the sex purchase law. A poll, taken to determine whether the law had influenced individual patterns of behaviour, compared results with those of a similar poll taken in 1996. The findings revealed that the number of male sex buyers in these studies had decreased from 13.6 percent to 7.9 percent, where each poll questioned 2,500 individuals between 18 and 74 years of age ^[42] .

Wilcox et al. (2009) highlighted studies which showed that men are more likely to buy sex when they are abroad than at home. Similarly, Kuosmanen's (2010) survey research on public attitudes of men who said they had purchased sex (based on a small sub-sample of 41 men), noted that 71% reported that their most recent purchase of sex took place abroad. Studies in Finland, which has partially criminalised the purchase of sex, have similarly found that sex is primarily purchased abroad (Lammi-Taskula, 1999 and Marttila, 2008 cited in Kuosmanen, 2010). Kuosmanen (2010: 258) concludes that the context in which sex is purchased is "substantially dependent both on the domestic market and the extent to which it is regulated by legislation and other norms in the area".

Kulick (2003 cited in Levy, 2015) provides some evidence of sex workers using the legislation to blackmail or extort money from clients, who are unlikely to contact the police given that the purchase of sex is prohibited, however this appears to be an issue which happens in prostitution generally as clients may be broadly reluctant to contact the police and to disclose their purchase of sex (also noted by the Norwegian Ministry of Justice and Police Affairs (2004)).

Ekberg (2004), the Norwegian Ministry of Justice and Police Affairs (2004) and Waltman (2011a and b) note the considerable emphasis that is placed on reducing exploitation and addressing those who exploit and profit from the sale of sex as third parties. SOU (2010) claims that the law has limited the growth of the sex market and subsequently, the involvement of organised criminals in prostitution in Sweden.

Jakobsson and Kotsdam (2013) using data on 46 European countries from UNODC and the International Labour Organisation ( ILO), while noting the limitations of available data, draw a positive correlation between prostitution legislation and rates of trafficking, with restrictive legislation associated with lower rates of human trafficking. ^[43] Although outwith the remit of this review and heavily caveated, other data appears to support this claim, for example Cho et al. (2013). ^[44]

Application and enforcement of the law

Danna (2012) notes that while no special funds were allocated to social services alongside the introduction of legislation to criminalise the purchase of sex, significant additional funds were assigned to the police to address human trafficking and to respond to prostitution in public spaces and on-line advertising on the internet. Given that both human trafficking and prostitution are typically low in terms of 'reported' crimes, the subsequent emphasis on surveillance that is required to enforce the legislation is important. Police efforts to combat human trafficking also involve responding to prostitution (Claude, 2010), leading to claims that the sex purchase legislation thus has a deterrent effect on human traffickers (see also Ekberg, 2015).

The Swedish Institute (2010) notes that following some initial uncertainty, police officers and prosecutors now consider that the provisions of the law are working effectively, with offences of sexual purchase relatively easy to investigate and to process, although attempted purchase is more difficult to prove. In 2010, uncertainties remained around the status of the individual involved in prostitution (i.e. as either witness or injured party) during court proceedings (see also Waltman, 2011b). Initial difficulties applying the Sex Purchase Act in Sweden related to evidence for corroboration for attempts to purchase sex while actual purchase was easier to prove with police officers apprehending individuals following sex ^[45] .

The number of clients prosecuted in Sweden increased from 46 reported crimes of buying sexual services in 1999 to 392 in 2010 in Stockholm (communication by BRA cited in Danna, 2012). 1,300 men were cautioned between 1999 and 2005 under the law and one male had received a prison sentence during that period (Scoular et al., 2008; Danna, 2012). Kelly et al. (2009: 39) note that there were 163 charges for the purchase of sex in 2006 with 108 convictions, while 38 trafficking offences were recorded in 2006 with 27 relating to sexual exploitation and 11 convictions overall.

There has been a steady increase in the number of reported crimes, almost exclusively resulting from police investigations. Wong (2014: 192-3) quoting statistical data from the Bra database notes that some years are characterised by a notable increase in crimes reported and determines that this can be attributed to seizure of client records on a single occasion from an organised prostitution ring (also referred to by Waltman, 2011b).

Ekberg (2015) using data from the National Council for Crime Prevention notes that the legislation has resulted in increased numbers of men being apprehended for attempting to purchase or having purchased a sexual service. Numbers of reported offences have also increased from 94 in 1999 to 601 in 2014. Numbers of convictions have increased, rising from 11 in 1999 to 391 in 2013.

In Finland, Niemi and Aaltonen (2014) identify issues with the application of the law partially criminalising purchase: where there is denial of a sex purchase; where sex purchase is not proven; or where a person admits sex purchase but denies awareness that the individual involved in prostitution was a victim of procuring or human trafficking. For example, an identified concern was that sex workers would be unlikely to cooperate in the provision of evidence against clients. In research which hypothetically explored the potential effects of the criminalisation of the purchase of sex in Northern Ireland, including 18 interviews with service providers and experts, police officers interviewed perceived that criminalisation of purchase would be difficult to police in practice (Huschke et al., 2014). However, Wong (2014) notes that there has been a steady increase in the number of persons actually convicted with the majority of suspects confessing and accepting a penal order, or being caught in the act, whereby evidence tends to be sufficient to obtain a conviction.

Experiences with the police

Whilst the selling of sexual services is not criminalised in Sweden, Levy's (2015) small qualitative study with individuals involved in prostitution suggests that police attempts to eradicate prostitution has included informing landlords of tenants selling sex and reporting individuals to hotels and venues so they will be banned from returning, as well as police officers visiting individuals involved in prostitution at their place of residence (Danna, 2012; Levy, 2015). In Levy's research, concerns were raised in interviews with sex workers and representatives from agencies, regarding the treatment of migrant sex workers, (as opposed to victims of human trafficking), by the police, with examples of cases where individuals involved in prostitution were deported from Sweden. For Levy, based on his interviews, this was perceived to be an act of 'destabilising sex work in Sweden' (Levy, 2015, p. 200). In Sweden, Levy (2015) reports that sex workers' experiences with the police are variable highlighting that experiences are more favourable where police are specially trained in dealing with prostitution. Whilst some of the sex workers interviewed identified encounters with police which were favourable; there were also reports by sex workers and service providers of police unprofessionalism and harassment. This reflects similar findings from Amnesty International (2016) on the situation in Norway where they report evidence that individuals involved in prostitution were subject to intrusive and targeted policing. Many of the women interviewed in the Amnesty International report claimed to be reluctant to contact the police stating that increased enforcement of the law had resulted in the eviction of individuals involved in prostitution from their place of work and/or home.

Ostergren (2004), who is critical of the legislation, claims that due to the drop in demand, individuals involved in prostitution in Sweden have expressed a reticence to report clients to the police when they are the victims of a crime. Dodillet and Östergren (2011) note that it appears that clients are less willing to assist as witnesses in cases in which profiteers who exploit the sexual labour of others are prosecuted, however Mujaj and Neitscher (2015) note that relations with the police and individuals involved in prostitution have improved in this respect as sellers of sex are not themselves subject to prosecution or criminalisation.