Introduction
Estimating the number of individuals selling sexual services is key to developing policies and services, identifying priorities, making informed funding decisions and assessing the impact of policies and services, among others. However, reliable estimates of the size of the sex industry and the characteristics of those working in it are difficult to obtain. This difficulty comes, on the one hand, from actual difficulties associated with the nature of the population (e.g. high mobility, a desire for privacy and anonymity, stigma) and on the other from what Wagenaar and Altink (Reference Wagenaar and Altink2012) call ‘morality politics’. Prostitution policy, they argue, is generally governed by ideology, emotionally charged and resistant to facts.
Yet, the existence of ample literature on the nature and prevalence of sex work and prostitution suggests that data does influence academic and policy debates. In the United Kingdom (UK), researchers have used a range of approaches to produce population estimates for those selling sex in various settings within the sex industry (e.g. Kinnell, Reference Kinnell1999; Hartworth et al., Reference Hartworth, Hartworth and Convery2012; Abramsky and Drew, Reference Abramsky and Drew2014; Brooks-Gordon et al., Reference Brooks-Gordon, Mai, Perry and Sanders2015; Sagar et al., Reference Sagar, Jones, Symons and Bowring2015; Sanders et al., Reference Sanders, Scoular, Campbell, Pitcher and Cunningham2018; Hester et al., Reference Hester, Mulvihill, Matolcsi, Lanau and Lilley-Walker2019). Ultimately, multiple estimates size of the sex industry co-exist, alongside varying estimations of the relative weight of different groups (e.g. migrant versus non-migrant workers, victims of exploitation vs independent workers) as well as multiple interpretations of such categorisations (Hester et al., Reference Hester, Mulvihill, Matolcsi, Lanau and Lilley-Walker2019). This is problematic as the effective design of policies and interventions relies on commanding reliable and comprehensive information on the reality on the ground (Shaw et al., Reference Shaw, Bloor, Cormack and Williamson1996; Walby et al., Reference Walby, Apitzsch, Armstrong, Balderston, Follis, Francis, Kelly, May-Chahal, Rashid, Shire, Towers and Tunte2016). Data limitations can also reinforce reliance on ideological positions (Wagenaar et al., Reference Wagenaar, Altink and Amesberger2017): in the absence of reliable evidence any side can claim to represent the majority view.
People working in the sex industry face enhanced risks including physical or sexual violence (Church et al., Reference Church, Henderson, Barnard and Hart2001; Shannon et al., Reference Shannon, Kerr, Strathdee, Shoveller, Montaner and Tyndall2009; Cunningham et al., Reference Cunningham, Sanders, Platt, Grenfell and Macioti2018), psychological harm and mental health risks (Hester et al., Reference Hester, Mulvihill, Matolcsi, Lanau and Lilley-Walker2019), stigma and isolation (Sanders, Reference Sanders2004; Benoit et al., Reference Benoit, Jansson, Smith and Flagg2018), and vulnerability to exploitation (Wagenaar and Altink, Reference Wagenaar and Altink2012; Hester et al., Reference Hester, Mulvihill, Matolcsi, Lanau and Lilley-Walker2019). Many also experience financial, housing and/or drug and alcohol misuse issues (Sagar et al., Reference Sagar, Jones, Jones and Clark2014; Hester et al., Reference Hester, Mulvihill, Matolcsi, Lanau and Lilley-Walker2019). As such, people selling various types of sexual services in various settings may require a range of different types of support.
In the absence of a well-developed national policy framework, prostitution policy in England and Wales has been increasingly decentralised, leaving policy and service provision to local areas, which in turn apply diverse approaches (Sagar and Croxall, Reference Sagar and Croxall2012; Niemi and Aaltonen, Reference Niemi and Aaltonen2014). This is not unique to England and Wales: in the Netherlands and Austria, for example, Wagenaar et al., (Reference Wagenaar, Altink and Amesberger2017) remark how implementation by local actors, chiefly the police and local administrations, has meant that local policy often deviates considerably from national goals. At the same time, local services as well as regulations and approaches (e.g. as related to policing) do not exist in a vacuum and will only go so far in the context of national policies and laws (Sagar and Croxall, Reference Sagar and Croxall2012; Brown and Sanders, Reference Brown and Sanders2017). Indeed, the needs, risks and experiences faced by people selling sex are shaped by the wider legal, economic and policy context. The local and the national interact in complex ways that shape the reality of commercial sex on the ground (Wagenaar et al., Reference Wagenaar, Altink and Amesberger2017). This is an issue to which we return below.
While the medium- and long-term impact on the sex industry of the COVID-19 pandemic is not yet known, there is some evidence that up until that point, the number of people involved in the sex industry had been growing. The English Collective of Prostitutes (ECP, 2018), amongst others, has linked national austerity policies, and more widely poverty, to an increase in prostitution. The connection between benefit reforms (specifically the introduction of Universal Credit) and what has been termed ‘survival sex’ has since been recognised by the government’s Department for Work and Pensions (Work and Pensions Committee, 2019; see also Alston, Reference Alston2018). The growth of the sex industry has also been facilitated by new possibilities afforded by the internet: 65 per cent of respondents to a survey of individuals advertising sexual services online reported that they would not be selling sexual services if it were not for the internet (Sanders et al., Reference Sanders, Scoular, Campbell, Pitcher and Cunningham2018). The figure suggests an increase in the number of people in the industry as well as the existence of new profiles of individuals selling sexual services. Recent research (Matolcsi et al., Reference Matolcsi, Mulvihill, Lilley-Walker, Lanau and Hester2021) also identifies new and growing types of sex work.
This article assesses the current state of knowledge with regards to the prevalence of prostitution and sex work in England and Wales and concludes that no source of data allows accurate assessment of the size and characteristics of the sex industry. As a result, prostitution policy is designed without information on the nature, size and composition of target groups, or the impact of policy on the ground (Wagenaar et al., Reference Wagenaar, Altink and Amesberger2017). Governments rarely collect systematic data on prostitution and sex work (Wagenaar et al., Reference Wagenaar, Altink and Amesberger2017). Indeed, data on the sex industry is notably difficult to obtain (Shaw and Butler, Reference Shaw and Butler1998; WHO, 2010; Kingston and Smith, Reference Kingston and Smith2020). Fear of stigma, negative experiences and in some cases coercion and control mean that sex work and prostitution are often hidden (Dodsworth, Reference Dodsworth2012; Walby et al., Reference Walby, Apitzsch, Armstrong, Balderston, Follis, Francis, Kelly, May-Chahal, Rashid, Shire, Towers and Tunte2016). The sex industry is diverse, dynamic and mobile, with people moving in and out of the industry, across sectors and geographically (Whowell and Gaffney, Reference Whowell, Gaffney and Phoenix2009; Pitcher, Reference Pitcher2015; Hester et al., Reference Hester, Mulvihill, Matolcsi, Lanau and Lilley-Walker2019). Even where sex work is legal, the industry tends to operate in the shadows and individuals selling sex tend to mistrust and resist policy measures (Wagenaar et al., Reference Wagenaar, Altink and Amesberger2017). While these challenges are significant, data is essential if services are to respond to existing, emerging and changing needs. Given this, we argue for the use of aggregated local and/or regional level data to inform both service planning and wider policy development. In-depth local evidence has the potential to build a more accurate and comprehensive picture of the sex industry, that acknowledges its geographical variations. Furthermore, local mapping can directly inform service delivery through the identification of existing needs and gaps (Sagar et al., Reference Sagar, Jones, Jones and Clark2014; NSWP, 2015).
This article is structured as follows. Section 2 defines the realms of prostitution and sex work as well as their commonalities and differences, as understood and employed by this study. Section 3 provides information on the methods employed in this research. Section 4 describes the key findings with regards to the prevalence of sex work in England and Wales, the challenges associated with the rise in online markets and more generally with data collection. Section 5 introduces local mapping as a powerful yet underused tool for estimating the size and characteristics of the sex industry and discusses both its potential and limitations. To conclude we propose that if done in collaboration with local actors, local mapping can reduce uncertainty and provide key data to inform service planning and delivery, and can potentially reduce the reliance on moral judgement for policy planning.
Prostitution and sex work
Who counts and who is counted as part of the sex industry is to a large extent a political decision, and can also be influenced by the knowledge in those doing the counting of what possible categories there are to count. Individuals selling sex in the UK are diverse in terms of demographics and motivation and report a range of experiences. There is ample evidence of vulnerability, limited choice, exploitation and other harm (see e.g. Jeal and Salisbury, Reference Jeal and Salisbury2007; Coy, Reference Coy2008; Harding and Hamilton, Reference Harding and Hamilton2009), alongside research suggesting more agency and nuance in relation to harm (Sanders et al., Reference Sanders, O’Neill and Pitcher2009, Reference Sanders, Scoular, Campbell, Pitcher and Cunningham2018; Pitcher, Reference Pitcher2019) and individuals and organisations demanding to be recognised as workers and for sex work to be decriminalised (ECP, 2015; NUM, 2018). These groups have different needs, demands and power. Such diversity is a challenge when designing policy and planning services, more so, since – as will be shown – we know little about the relative composition of the current sex industry.
We define prostitution and sex work as the provision of sexual or erotic acts or sexual intimacy for payment in money or other benefit or need (Hester et al., Reference Hester, Mulvihill, Matolcsi, Lanau and Lilley-Walker2019). While the terms ‘prostitution’ and ‘sex work’ are sometimes used interchangeably, they reflect different scopes and understandings of the sex industry. The terms ‘sex work’ and ‘sex worker’ emerged to redefine the selling of sexual services as work, and to thus promote the agency of and reduce stigma against those selling sexual services (Overall, Reference Overall1992). Meanwhile, the terms ‘prostitution’ and ‘person in prostitution’ tend to be used to highlight what is seen as a lack of power and agency in selling sex, and to reject the notion of selling sex as labour. Furthermore, while ‘prostitution’ is a specific activity involving direct physical contact between the seller and buyer, ‘sex work’ (although sometimes used to refer solely to direct sex) often also includes other areas of the sex industry. This can be with or without direct physical contact (e.g. lap dance, BDSM, webcamming). Regardless of the term used, employing a wide definition as above is valuable in planning policy and services so as not to exclude groups.
Furthermore, our definition captures a wider group than a definition limited to ‘selling sexual services in exchange for money’. For example, some individuals exchanging sexual acts for money or other benefit or need may not see themselves as selling sex, such as those involved in sugar arrangements or those who sell sex ‘incidentally’ without planning to do so (Morris, Reference Morris2018). These individuals may still experience some of the support needs and risks associated with selling sex (for an overview of risks see Kinnell, Reference Kinnell2008; Sanders, Reference Sanders2004; Cunningham et al., Reference Cunningham, Sanders, Platt, Grenfell and Macioti2018). Moreover, Hester et al. (Reference Hester, Mulvihill, Matolcsi, Lanau and Lilley-Walker2019) identified a range of engagement patterns including full time as well as occasional or intermittent engagement, with sex work and prostitution often taking place in combination with other activities such as caring, studying, volunteering and paid employment. This can also cause individuals to not identify with labels or identities as a ‘sex worker’ or ‘person in prostitution’. A wide view thus also allows to better capture the diversity of pathways into, out of and within the sex industry (see also Coy, Reference Coy2008; Harding and Hamilton, Reference Harding and Hamilton2009; Dodsworth, Reference Dodsworth2012; Sagar et al., Reference Sagar, Jones, Symons and Bowring2015). Here we use both terms, prostitution and sex work, as well as ‘selling sex’, to refer to this wide range of experiences and practices. We acknowledge the ideological leanings and limitations of both terms and the existence of other terms – for example ‘transactional sex’ and ‘survival sex’ (see McMillan et al., Reference McMillan, Worth and Rawstorne2018; also Mulvihill, Reference Mulvihill2019), and that specific terms are experienced as harmful by some individuals. However, we use this range of terms here because they are the most widely-known, encompass the broader sex industry (and not only paid-for direct sex) and go some way to reflect the diversity in experiences and views on the sex industry. Recognising the range of experiences within the sex industry is, in our view, one of the starting points for a successful mapping exercise.
Methods
To assess the current state of knowledge with regards to the prevalence of prostitution and sex work in England and Wales, we proceeded in two steps: a systematic literature review and an extensive data review. The literature review aimed at identifying the current state of knowledge on the sex industry in England and Wales, unifying research published across disciplines. Between April and May 2018, the team searched Scopus, Social Services Abstracts, Web of Science, ProQuest, IBSS, OvidSP for literature published between 2000 and 2018, in English and mentioning ‘prostitut-’ or ‘sex work’ or ‘commercial sexual exploitation’. Since a substantial part of research produced by NGOs and government bodies is not published in academic outlets, the database searches were complemented by hand-searches, recommendations and citation tracking. In total, we identified 10,724 publications. A team of four researchers manually searched and coded titles and abstracts. A small proportion (236) mentioned prevalence. We narrowed the search to items referring to the United Kingdom and England and/or Wales, and prevalence including existing estimates and methodological texts. This yielded fifty-one texts which were read in full. We excluded studies that focus on specific types and settings within the sex industry. Only five publications reported primary data for the wider sex industry alongside a transparent methodology; these are discussed below. Kinnell (Reference Kinnell1999) was included despite being outside the initial time range as it has strongly influenced later studies. An updated search conducted in July 2020 for this article uncovered fifty-five further articles published between 2018 and 2020, but none fit the inclusion criteria.
In addition to the above literature searches and review, we concurrently carried out a data review informed by five sets of sources: 1) data sources used in previous research; 2) a systematic search of the UK Data Service database for studies and series from 2000 (keywords: sex, prostit*); 3) a search of the UK Data question bank for surveys containing the words prostit- or sex; 4) interviews with key stakeholders; and 5) online questionnaires distributed to selected stakeholders (five police forces, eleven service providers/NGOs, four local authorities) on existing data collection procedures. The interviews and questionnaires allowed us to assess the potential for new (including small-scale) prevalence estimates at the local or regional level. More detail is provided below.
Counting in the sex industry
Historically, the primary source of information on the nature and prevalence of sex work and prostitution in the UK has been data from support services and frontline professionals who regularly work with and support people in the industry (e.g. Kinnell, Reference Kinnell1999; Brooks-Gordon et al., Reference Brooks-Gordon, Mai, Perry and Sanders2015). While service providers hold valuable knowledge about the experiences of their clients, local circumstances and so forth, there are limitations to these estimates rooted in the numerous difficulties in identifying and accessing individuals who may need support. Those selling sex on-street/outdoors can be transient, or have difficulty making or sustaining contact with services due to issues around drug and alcohol addiction and homelessness. Meanwhile, the expansion of the internet has promoted a shift away from more visible street/outdoor prostitution towards the advertising, organising and exchange of sexual services indoors and online. Individuals in these sectors are less likely to be in contact with specialist services (Pitcher, Reference Pitcher2015; Sanders et al., Reference Sanders, Scoular, Campbell, Pitcher and Cunningham2018), and there are concerns that they are often excluded from policy development and service provision (Hubbard and Prior, Reference Hubbard and Prior2012).
Furthermore, some of those who do contact non-specialist support services (e.g. sexual health clinics) may not be willing to disclose their participation in the industry. Respondents to an online survey by Hester et al. (Reference Hester, Mulvihill, Matolcsi, Lanau and Lilley-Walker2019) described hiding their occupation from both services and their social circles and having a fear of being outed as a sex worker. Some participants who were mothers were concerned about social services questioning their ability to parent. Similarly, street workers interviewed by O’Neill et al. (Reference O’Neill, Jobe, Bilton and Stockdale2017) reported feeling uncomfortable and judged when interacting with sexual health services, which led many to avoid disclosing that they were selling sex (see also Dodsworth, Reference Dodsworth2012). Negative experiences with police, involvement in illegal activities (or those perceived as illegal) and immigration status also discourage individuals from contacting services or disclosing that they sell sex (Voices Heard Group, 2008; Changing Lives, 2016). As a result, the information held by support services and frontline professionals is unlikely to provide an accurate representation of the industry.
New research has focused on independent workers in the online sector, i.e. where sexual services are advertised, organised, and/or exchanged online (e.g. Sanders et al., Reference Sanders, Scoular, Campbell, Pitcher and Cunningham2018; Pitcher, Reference Pitcher2019). Attention to individuals working independently online sheds light on an until recently under-researched area of the industry. At the same time, a focus on the online market may result in those working on-street/outdoors and/or not working independently being excluded, thus perpetuating the issue faced by previous studies with a focus on street and indoor markets (Hester et al., Reference Hester, Mulvihill, Matolcsi, Lanau and Lilley-Walker2019). As detailed later, local mapping, while employing a wide definition as above, can help to move towards more comprehensive prevalence estimates which reflect the broader sex industry. The next sub-sections present a systematic review of the evidence on the prevalence of sex work and prostitution in England and Wales and the growing body of evidence on online sex markets.
Evidence on the prevalence of sex work and prostitution
Arguably, the most widely cited estimate for the prevalence of sex work in the UK is Kinnell’s (Reference Kinnell1999) 80,000 figure. To estimate the number of women selling sex Kinnell (Reference Kinnell1999) surveyed sixteen service providers (out of 120 thought to operate at the time) on the number of individual service users as well as an estimate of the number of sex workers in their area. The ratio between the area estimate and known service users is used as a multiplier to estimate the number of individuals selling sex. She estimated 80,000 women to work in the industry in the late nineteen-nineties. More recently, Cusick et al. (Reference Cusick, Kinnell, Brooks-Gordon and Campbell2009), Pitcher (Reference Pitcher2015) and Brooks-Gordon et al. (Reference Brooks-Gordon, Mai, Perry and Sanders2015) built on Kinnell’s approach to create updated estimates. Deriving multipliers from the average number of service users is common practice when studying hard-to-reach populations – however, the accuracy of the produced estimates is difficult to assess (WHO, 2010). For instance, individuals in contact with multiple services may be double counted. To mitigate this issue, Cusick et al. (Reference Cusick, Kinnell, Brooks-Gordon and Campbell2009) relied on recorded data on the number of unique (self-identified) sex workers that used the service in a year, while Pitcher (Reference Pitcher2015) asked services to estimate the proportion of their users who are in contact with other services. Multipliers are also affected by the characteristics of the responding organisations. For instance, larger better resourced organisations may be more likely to reply, resulting in a larger multiplier (and thus an overestimate of the population size). Furthermore, services tend to specialise in specific groups – e.g. street workers, and victims of trafficking – and may know little about other areas of the sex industry. Overall, multipliers risk introducing unknown biases in the estimates.
Furthermore, the profiles of those accessing services partially reflect the existence of appropriate and accessible services (Hartworth et al., Reference Hartworth, Hartworth and Convery2012). Visibility of and support for specific groups also varies by area. For instance, there are no male specialist services listed in the UKNSWP directory for Wales, two in London and three in Yorkshire. Equally, a reduction in the number of services (e.g. as a result of austerity) would result in reduced population estimates regardless of actual changes in the population. For instance, if participation in the sex industry is increasing following a recession and/or benefit cuts, while simultaneously services are being reduced, estimates from service providers are likely to result in an under-estimation and misrepresentation of the sex industry.
Some studies have sought to overcome data limitations by combining information from multiple sources. Abramsky and Drew (Reference Abramsky and Drew2014) sought to estimate the size of the indoor and street prostitution markets in the UK by combining data gathered on the indoor market (Dickson, Reference Dickson2004) with London Metropolitan Police data on street prostitution. Their study has been heavily criticised chiefly for their reliance on London data to build national estimates, and the use of police dataFootnote 1 (see Brooks-Gordon, et al., Reference Brooks-Gordon, Mai, Perry and Sanders2015, for a detailed discussion).
Brooks-Gordon et al. (Reference Brooks-Gordon, Mai, Perry and Sanders2015) complemented data from services with existing research on under-represented populations such as men, and checked their estimates with people directly involved in the sex industry. The assumption is that people in the industry often know each other and can estimate how many people operate in their geographical area. While the contrast of multiple sources of data allows a degree of triangulation which can help evaluate the observed estimates, such an assumption can be problematic, particularly when respondents are asked to provide estimates for people in other areas of the industry (WHO, 2010).
So how many people are in the sex industry in the UK? With the different methods and scopes noted above, it is unsurprising that the estimates vary substantially. Brooks-Gordon et al.’s (Reference Brooks-Gordon, Mai, Perry and Sanders2015) estimate for 2009-13 is 72,816 individuals, very close to the numbers obtained by Kinnell (Reference Kinnell1999), and Pitcher (Reference Pitcher2015) who estimated 85,714 people selling sex in the UK. Cusick et al. (Reference Cusick, Kinnell, Brooks-Gordon and Campbell2009) report a ‘conservative estimate’ of 35,882 but note that the actual figure may be up to three times as high, i.e. over 100,000 (Cusick et al., Reference Cusick, Kinnell, Brooks-Gordon and Campbell2009: 714). Abramsky and Drew (Reference Abramsky and Drew2014) estimated around 58,000 women to be selling sex in 2004. The lack of consensus in the estimates reflect differences in methodology and population of interest (e.g. different settings/types, genders) and are also a testament to the difficulty of producing reliable national estimates. The expansion of the internet has changed the sex industry and added new challenges.
Online sex markets, new sources and new challenges
The expansion of the internet has substantially altered the sex industry and added complexity to prevalence estimation. A range of actors including independent sex workers, traffickers, brothels and others advertise sexual services online. Advertising takes place through both sex industry-specific platforms and classifieds sites. Online advertising platforms are an obvious but also problematic source of prevalence data (Sanders et al., Reference Sanders, Scoular, Campbell, Pitcher and Cunningham2018; Hester et al., Reference Hester, Mulvihill, Matolcsi, Lanau and Lilley-Walker2019). Counting profiles on online advertising platforms can provide a quick and rough estimate of the size of the online market.
However, it is also likely to lead to an over-estimation of the number of people selling sex. Some individuals will have multiple profiles, either on the same or different platforms; while profiles that are no longer active may still appear in platforms (Sanders et al., Reference Sanders, Scoular, Campbell, Pitcher and Cunningham2018; Hester et al., Reference Hester, Mulvihill, Matolcsi, Lanau and Lilley-Walker2019). In their analysis of online platforms, Sanders et al. (Reference Sanders, Scoular, Campbell, Pitcher and Cunningham2018) found up to 32 per cent of profiles to be inactive, while Hester et al. (Reference Hester, Mulvihill, Matolcsi, Lanau and Lilley-Walker2019) report that around half of the profiles on two of the most-used sites are incomplete. An additional risk lies in analysing only the most widely used platforms, which would lead to an under-representation of groups more likely to use specialist platforms such as ethnic minorities or men selling sex to men. Finally, the information published by websites relies on self-reported data. However, descriptions of personal characteristics such as age, nationality or sexual orientation do not necessarily reflect those in the industry, as advertisers tend to modify profiles to include characteristics that are perceived as more desirable (Hester et al., Reference Hester, Mulvihill, Matolcsi, Lanau and Lilley-Walker2019). All in all, while using online data sources is essential to understanding the sex industry today, relying on online sources alone risks mis-estimation, overlooking practices that are part of the sex industry, and obtaining an inaccurate picture of the characteristics of individuals selling sexual services.
All of the approaches discussed above highlight and provide valuable information on some sectors of the industry, but leave other areas and groups obscured. Studies relying on data from service providers will likely under-estimate victims of trafficking and exploitation, migrants as well as escorts and those linked to the online industry, who are less likely to be in contact with specialist support services. Research relying on online sources will likely under-estimate groups such as sexual and ethnic minorities, who tend to use specialist channels and sites, as well as groups who are less likely to access (whether because of resources, safety or other reasons) the internet to participate in online research.
Awareness of the trade-offs associated with each methodological approach is key for the development of social policy responses as it will affect conclusions about the relative weight of different groups. Gender breakdowns, for example, vary widely. Using data from specialist services, Pitcher (Reference Pitcher2015) estimates men to be 7 per cent of the industry, while an analysis of the main online platform reports 42 per cent of advertisers to be men (Import.io, 2014). Reliance on incomplete data therefore can muddle not only the overall size but also the characteristics of the industry. This is critical when defining priorities for planning or estimating unmet needs for services. Next, we examine other sources of data available for estimating the prevalence of prostitution and sex work in England and Wales.
Mapping local data
Our data review identified no national datasets that allow for estimating the prevalence of prostitution and sex work in the UK, England or Wales. The National Survey of Sexual Attitudes and Lifestyles currently only collects information on whether respondents have paid for sex. The proportion of men who report having paid for sex has increased over time, from 2 per cent in 1990 to 4 per cent in 2000, and 11 per cent in the latest survey (Ward et al., Reference Ward, Mercer, Wellings, Fenton, Erens, Copas and Johnson2005; Jones et al., Reference Jones, Johnson, Wellings, Sonnenberg, Field, Tanton, Erens, Clifton, Datta, Mitchell, Prah and Mercer2015) – however, the trend may reflect increased likelihood to report rather than increase in demand, supply or number of exchanges. Thus, currently in the UK, no source of data allows for assessing changes in the prevalence of sex work and prostitution.
Research using a range of sources has evidenced geographical variations in the size and nature of the industry (Hartworth et al., Reference Hartworth, Hartworth and Convery2012; Brooks-Gordon et al., Reference Brooks-Gordon, Mai, Perry and Sanders2015; Changing Lives, 2016), which imply different needs and priorities. Sex markets vary across locations in terms of the size of the street scene, modalities of sex work and prostitution, nationality and gender of those involved in the trade, amongst other characteristics. Some areas in Northern England have a small or non-existent street scene, while in larger cities, street areas tend to be larger and clearly defined (Hester et al., Reference Hester, Mulvihill, Matolcsi, Lanau and Lilley-Walker2019). In some regions, escorts are mainly organised by agencies, while in others independent escort online adverts are more common. The London scene is largely international and multi-ethnic (Brooks-Gordon et al., Reference Brooks-Gordon, Mai, Perry and Sanders2015), while a study in Cardiff targeting BAME female indoor workers could only identify three individuals (Sagar et al., Reference Sagar, Jones, Croxall, Tyrie and Chimba2011). Similarly, while the male escort scene is well-organised in cities such as London and Manchester, other areas do not appear to have a significant male scene (Hester et al., Reference Hester, Mulvihill, Matolcsi, Lanau and Lilley-Walker2019).
Such geographical diversity highlights the importance of local-level data to identify policy priorities and planning service provision. Mapping exercises describe the physical and online spaces where the sex industry operates, identify networks and can provide a rough estimate of the population size, its characteristics and patterns of activity (WHO, 2010). Generally, the process involves defining a geographic scope, contacting a broad range of specialised and non-specialised services, and gathering quantitative and qualitative information on the nature of the sex industry, the locations where it takes place, the individuals involved and their needs. When combined with methods to capture the online industry, such as online surveys, mapping has strong potential to uncover unmet needs and facilitate service planning (Hartworth et al., Reference Hartworth, Hartworth and Convery2012). Indeed, mapping is used for planning in a range of intervention areas, from supporting bereaved children (Childhood Bereavement Network n.d.) to enhancing rural communities (Creative Rural Communities, 2017).
However, it remains under-used for the study of sex work and prostitution in the UK context. Our systematic review uncovered just two local mapping studies in England and one regional study in Wales. Barefoot Research and Evaluation (2010) examined the extent and dynamics of the sex market and sexual exploitation in County Durham and Darlington (North England) using a qualitative network mapping approach consisting of structured interviews with a range of professionals. They identified 109 adults to be involved in the local sex industry. While the use of qualitative data to generate estimates has limitations, the approach allowed identifying populations previously not known to services, including ten children being sexually exploited. Prompted by a scarcity of specialist services in their area Hartworth et al. (Reference Hartworth, Hartworth and Convery2012) used a mixed approach to conduct an extensive mapping of sex work and prostitution in North East England through interviews with 200 professionals and practitioners from specialist and non-specialist services, who identified service users, other individuals in the industry, as well as locations where paid sex is known or thought to happen. The team then visited those locations to verify the information obtained through observation and interviews and to request further leads. Their iterative process revealed a heterogeneous sex market, with a diverse range of services, locations and prices charged. Moreover, Hartworth et al. (Reference Hartworth, Hartworth and Convery2012) uncovered segments of the sex industry population that were previously unidentified, including exploitation, behind the ‘small visible sex market’ in the region (p. 87). They estimated around 1000 individuals to be involved in sex work in the area at any given time, and conclude that this invisibility means agencies often fail to realise the range of needs faced by those in the sex industry. Finally, Sagar and colleagues (Reference Sagar, Jones, Jones and Clark2014) conducted a mapping exercise for Wales where they engaged 200 agencies and organisations, 100+ sex workers, residents and businesses through interviews, questionnaires and engagement events. The research reported sex work as taking place in all local authorities in Wales. All three mapping exercises identified forms of sex work previously unacknowledged in the area: a strength of the approach.
Complementing traditional mapping exercises, as in the above examples, with surveys can be a powerful approach to access groups not known to services. Online surveys can, at a relatively limited cost, be accessed by potentially thousands of users covering a wide geographical range (Hine, Reference Hine, Gilbert and Stoneman2016). Enhanced anonymity means that people are more likely to disclose their activities. The downside is that verification can be difficult and researchers’ control over the data is reduced, potentially opening the door to misuse (e.g. people other than the target population responding). Close collaboration and information sharing with specialist and non-specialist services and online platforms can help mitigate these issues. Equally, it is crucial to consider inequality in access to the internet, and trust: vulnerable groups (such as those who are dealing with homelessness, substance misuse and/or who are exploited), as well as those with language and other barriers, are less likely to complete online surveys (Hester et al., Reference Hester, Mulvihill, Matolcsi, Lanau and Lilley-Walker2019). Gathering data through multiple mechanisms and from multiple sources is fundamental to build a comprehensive understanding of the industry.
Mapping can be seen as both costly and potentially putting sex workers at risk. Indeed, the UKNSWP notes that
In many cases, programmers concentrate on commissioning mapping and population size estimates merely to accumulate more data without plans to provide services. This uses up precious resources which could be used to provide services (NSWP, 2015: 1).
The identification of sex workers can also put them at risk. Data collection efforts may have direct effects (e.g. questions may re-traumatise people who have experienced exploitation or violence), but also potential unintended consequences (Sanders et al., Reference Sanders, Scoular, Campbell, Pitcher and Cunningham2018). For example, the identification of locations where individuals offer sexual services can be useful for service delivery planning but could also put those selling sex at risk of being ‘outed’ or targeted for violence or theft. Including input from those currently and formerly selling sex and support organisations, can both ease the collection of data and assist in the prioritisation of safety and ethical procedures.
Indeed, ethical considerations are a crucial aspect of both research and social intervention, even more so when working with potentially vulnerable groups such as those involved in the sex industry. Some sex workers experience harm as a result of police enforcement (SWARM, 2018). They may also be vulnerable to ‘outing’ and/or harassment (NSWP, 2015). Mapping exercises should prioritise the needs and safety of people offering sexual services (Harding and Hamilton, Reference Harding and Hamilton2009; WHO, 2010; NSWP, 2015). This includes ensuring that the mapping is carried out purely to inform local support and service provision, that any data collected is kept secure and that none of the published information can be used for enforcement purposes or to harass or otherwise harm those selling sex (Kingston and Smith, Reference Kingston and Smith2020). Researchers and institutions aiming to estimate the prevalence of prostitution and sex work must use approaches that reduce and control bias, as well as balance the potential risks for those in the industry (Walby et al., Reference Walby, Apitzsch, Armstrong, Balderston, Follis, Francis, Kelly, May-Chahal, Rashid, Shire, Towers and Tunte2016). Concerns about confidentiality and the protection of vulnerable service users make data sharing challenging for services, and data collection should as far as possible accommodate this, as well as conform to ethical considerations in terms of the privacy of individuals. Involving paid peer researchers – with lived experience in the sex industry – in the mapping exercise, from design through to data collection and analysis to writing up and dissemination, is crucial and can help reduce risks and improve trust. Any local mapping being carried out must thoroughly consider such ethical implications of their research.
When conducted well, the advantages of mapping outweigh its limitations. Gathering data at the local level allows policy and programme planning to be grounded on in-depth knowledge of the population. While no single source has a view of the industry as a whole, a wide range of organisations hold partial data. This includes, as noted, charities and NGOs working directly with and/or carrying out advocacy for people involved in sex work and prostitution, and police, but also other organisations such as health services, other criminal justice agencies such as courts and probation services, local authorities (which in some jurisdictions license venues where sexual services are sold or otherwise monitor the sector), social services, and tax authorities. While all these sources have their limitations, when brought together, they can provide more robust and comprehensive estimates than using one or two sources alone.
We acknowledge that the experiences of those in the sex industry are influenced both by the local context and by the wider legal and social policy frameworks (Brown and Sanders, Reference Brown and Sanders2017). Coordinated mapping initiatives would allow comparison within and across regions on the prevalence and nature of prostitution and sex work, as well as the welfare and needs of those working in the industry, providing key evidence to inform regional and national policies, and interventions. If resources allow, conducting update mappings every few years could help identifying changes in the industry. Updates could also be informed by support organisations who are often aware of emerging or changing needs among their users. Local mapping can focus on particular settings and types of sexual services. However, given that individuals can be involved in various settings simultaneously or over time (Hester et al., Reference Hester, Mulvihill, Matolcsi, Lanau and Lilley-Walker2019), mapping exercises that seek to look at the wider sex industry can help capture information relevant to the wider experiences and needs of individuals, not to mention a wider range of individuals.
Looking ahead
The (UK) sex industry is thought to be growing as a result of both worsening material conditions for individuals (Alston, Reference Alston2018) and the expansion of the online industry (Sanders et al., Reference Sanders, Scoular, Campbell, Pitcher and Cunningham2018). The characteristics and experiences of those working in the sex industry are also changing: meaning new policy responses are required to respond to new needs and modes of work. We know the sex industry to be widely diverse, dynamic and mobile (Whowell and Gaffney, Reference Whowell, Gaffney and Phoenix2009; Wagenaar and Altink, Reference Wagenaar and Altink2012; Pitcher, Reference Pitcher2015; Hester et al., Reference Hester, Mulvihill, Matolcsi, Lanau and Lilley-Walker2019). People in the industry have different trajectories, degrees of control and constraints. Some are victims of trafficking and exploitation. For others, sex work can be empowering, or simply a way to pay the bills (Mulvihill, Reference Mulvihill2019; Pitcher, Reference Pitcher2019). In this context, the selection of sources used to create estimates and breakdowns is not neutral: it determines which voices are amplified and which are left unheard. Relying on partial estimates and incomplete sources is likely to result in biased conclusions that misrepresent segments of the industry and can obscure others (e.g. male or disabled individuals).
This article proposes a wider use of local mapping in England and Wales as a strategy to inform both prostitution policy and service provision. In a context in which the governance of sex work largely happens at the local level, local mapping can inform service planning and delivery so that it addresses the specific needs of the local population. Indeed, without reliable data on the size and characteristics of the sex industry and those involved in it, it is hard to make policy. Developing local data and information is particularly relevant in light of evidence that the known sex markets may be a small part of the sex industry in the area. Local (and regional) information is key to inform service provision and good policy practice.
Equally, the sex industry is shaped by wider structural and material conditions (e.g. economic and labour market context, the benefits system, national prostitution legislation, gender inequalities and so on) which are beyond the control of local governments and can limit the impact of local initiatives (Brown and Sanders, Reference Brown and Sanders2017). However, local mapping has the potential to inform and enrich wider debates on prostitution policy by highlighting the diversity of the industry both within and between locations. In doing so, it has the potential to help practitioners and policymakers move towards a more comprehensive and multifaceted understanding of sex work and prostitution.
Prostitution policy is often informed by moral judgement with little attention to data (Wagenaar and Altink, Reference Wagenaar and Altink2012). Data by itself is unlikely to solve the difficulties of designing prostitution policy. Doing so requires some political will and the development of networks of trust. However, the absence of data that reflects the diversity of the sex industry hampers the development of such policies. Thus, the collection of reliable evidence is necessary, if not sufficient for the development of evidence informed prostitution policy. Mapping can and has been used to understand the needs of and inform planning for other vulnerable populations such as homeless youth and drug usersFootnote 2 . This approach does not erase the difficulties of researching the sex industry (Wagenaar and Altink, Reference Wagenaar and Altink2012). No source of data is perfect and there is no single solution. Furthermore, data collection involves risks, and thus it is key that activities prioritise the safety of those involved in the sex industry. We propose that if done in collaboration with local actors, enhanced use of mapping can reduce uncertainty and provide key data to inform service planning decisions. While the results of single mapping exercises cannot necessarily be extrapolated to other areas, or indeed to the national or country level, they can be pivotal to improving service provision and reducing reliance on moral judgement when planning policy. With those aims in mind, counting should come alongside a thoughtful, open and respectful debate on the nature of prostitution and sex work.
Funding statement
The research that inspired this article was funded by the UK Home Office and the Office of the Police and Crime Commissioner for South Wales. This paper has also received funding from the postdoctoral fellowship programme Beatriu de Pinós, funded by the Secretary of Universities and Research (Government of Catalonia) and by the H2020 programme Marie Sklodovska-Curie grant agreement nº 801370.