Transactional Sex: Supply and Demand Among European Men Who have Sex with Men (MSM) in the Context of Local Laws.

Objectives: Transactional sex (TS) is generally defined as the trading of sex for material goods. Cast within the broader context of prostitution laws, we examined variations in the sociodemographic profile of men who have sex with men engaging in TS by payment direction (buying/selling).
Methods: The data were collected as part of the 38-country European Men who have sex with men Internet Survey project, conducted in 2010.
Results: About 12% of respondents reported engaging in TS in the past year. TS was associated with laws, age, education, employment, and residence. Conclusions: The striking sociodemographic differences in TS by payment direction suggest a power differential and a leading role of socioeconomic factors in TS.

INTRODUCTION

Transactional sex (TS) is generally defined as the trading (buying or selling) of sex for material benefit (i.e., exchanging money, drugs, food, shelter, or other items for sex). This includes informal bartering by individuals whose primary income is not derived from TS (Dunkle et al., 2007; Edwards, Iritani, & Hallfors, 2006; Maganja, Maman, Groves, & Mbwambo, 2007; Minichiello et al., 2000). Much has been written about financial and material motivating forces underlying women's transactional sexual relationships (e.g., Hunter, 2002). Also the more limited literature of men who trade sex with other men suggests that motivations and reasons for selling sex primarily are material-related, including trading sex for money, food, housing, and paying bills (Decker, Raj, Gupta, & Silverman, 2008; Mimiaga, Reisner, Tinsley, Mayer, & Safren, 2008; Weber et al., 2001). However, in their mixed-methods study, Mimiaga and colleagues (2008) additionally found that male street workers and Internet escorts traded sex with other men to support a drug or alcohol habit, for excitement, and because they would have sex anyway and preferred to get paid. Recent data from Australia similarly showed that men who engaged in TS were more sexually adventurous in general, reported group sex, and scored higher on measures of sexual sensation seeking (Prestage, Jin, Bavinton, & Hurley, 2014).

TS by both men and women has been linked with higher rates of HIV and sexually transmitted infections (STIs; Edwards, Halpern, & Wechsberg, 2006; Edwards, Iritani, et al., 2006). Research has shown that men who sell sex are more likely than other men who have sex with men (MSM) to engage in unprotected sex with their non-TS male and female partners (Elwood, Williams, Bell, & Richard, 1997; Estcourt et al., 2000; Koken, Parsons, Severino, & Bimbi, 2005; Prestage et al., 2007; Rietmeijer, Wolitski, Fishbein, Corby, & Cohn, 1998). However, although the potential for HIV transmission in the course of TS is indicated in studies that have revealed HIV prevalence rates of 12% to 41% among street-based samples of MSM who sell sex (Bacon et al., 2006; Belza, for the EPI-VIH Study Group, 2005; El-Bassel et al., 2000), it is not clear whether the practice of TS itself represents increased risk for HIV transmission.

Previous studies on gay and bisexual men who sell sex have not only tended to concentrate on the risk for HIV and STI transmissions, but they have been cast within paradigms that reduce such men to “vectors” of HIV and STIs (Bimbi, 2007). This has led to a limited understanding of how these men may be vulnerable in other aspects of their lives (Prestage et al., 2007; Smith & Seal, 2008). However, a small but growing body of research, largely U.S.-based, has revealed an association between selling sex and homelessness (Bobashev, Zule, Osilla, Kline, & Wechsberg, 2009; Lankenau, Clatts, Welle, Goldsamt, & Gwadz, 2005; McCarthy, Benoit, & Jansson, 2014; Newman, Rhodes, & Weiss, 2004) and psychological distress (Biello, Colby, Closson, & Mimiaga, 2014; El-Bassel et al., 2000; Friedman, Guadamuz, & Marshal, 2011; Reisner, Mimiaga, Mayer, Tinsley, & Safren, 2008; Weber et al., 2001). Weber and colleagues (2001) found that men who sold sex not only were more likely to have a high depression score and a history of residence in a psychiatric ward, but they were also 3 times more likely to be unemployed, to have less than a high school education, and to be younger compared with their counterparts who did not engage in TS. Similarly, research in Spain identified that men who sold sex had a lower educational level and the majority were immigrants (Belza, for the EPI-VIH Study Group, 2005; Belza et al., 2001;). Collectively, such research, almost exclusively on men who sell sex, suggests that understanding the relationship between TS and structural and socioeconomic status (SES) factors may be critical in understanding the broader societal factors affecting men who engage in TS with other men.

Much prior research on MSM who engage in TS has been shaped by the bias toward sampling street-based men, men residing in the United States, and men who sell sex (not those who buy sex). With this analysis, we aimed to expand previous research by examining TS within the broader context of prostitution laws, examine the variations in TS by payment direction, and identify the sociodemographic profile of European MSM engaging in TS, both on the supply and demand sides.

METHODS

Sampling and Eligibility

The data used in this analysis were collected as a part of the European Men who have sex with men Internet Survey (EMIS) project. The EMIS was a collaborative, cross-sectional study conducted simultaneously in 38 countries during the summer of 2010, with the objective of identifying prevention needs commonly unmet across diverse groups of MSM. The detailed methods of the EMIS have been reported elsewhere (Weatherburn et al., 2013). Briefly, the EMIS was an anonymous, self-administered online survey simultaneously conducted in 25 languages across Europe, including the non-European Union languages most frequently spoken in Europe: Russian, Turkish, and Ukrainian. Participants were recruited through more than 230 social media and dating Web sites for gay, bisexual, and other MSM. Residing in Europe and being an MSM and/or a man who felt attracted to men were the main eligibility criteria. Typical survey completion time was 20 min. No financial incentives were given and no IP addresses were collected. All study procedures were approved by the Research Ethics Committee of the University of Portsmouth in the United Kingdom.

Measures and Statistical Analysis

In the structured survey, participants were asked to respond to a number of closed-ended questions, with answer options being primarily Likert scale, recency scale, and binary (e.g., yes/no). All respondents who reported any sexual contact with at least one man in the previous 12 months were asked how frequently they had “been paid by a man to have sex” and how frequently they “paid a man for having sex” with them in their country of residence (the online questionnaire auto-displayed the country name that was selected previously as the respondent's country of residence). Payment for sex (“been paid/paid”) was not defined a priori so it was left to the participants to decide whether paying noncash goods was payment. The frequency scale included the following response options: not at all, 1 to 2 times, 3 to 10 times, 11 to 50 times, and more than 50 times. In this analysis, frequently selling sex was operationalized as having been paid by a man to have sex 11 or more times in the previous 12 months. Frequently buying sex was operationalized as having paid a man to have sex 3 or more times in the previous 12 months. (To our knowledge, there is no empirical argument for a particular operationalization of frequent TS. Our categorization was largely a pragmatic one, driven also by data requirements for analyses and wishing to avoid including men who may have engaged in TS only once or twice in their lifetime.) Respondents were further asked to indicate when and in which country they last had sex abroad with a man who did not also live in the respondents’ country of residence and whether or not they paid or were paid for sex on that occasion. Single-event recall such as last sexual encounter helps minimize recall bias and has been found to be a valid representation of sexual behaviors over longer periods of time (Younge et al., 2008). Thus, the TS abroad variable serves as a data validity check in addition to offering additional information about travel-related TS.

We examined men's TS behavior in the context of prostitution laws. While researching this article, we found no other studies examining the possible influence of laws, although it is also the opinion of others (e.g., Browne & Minichiello, 1996; Scott et al., 2005) that to understand TS, examinations of TS should be done with consideration to the wider societal forces in which the behavior occurs. For the present analysis, each of the 38 EMIS countries’ legal situation concerning prostitution was assessed and assigned to one of four law categories (based on en.wikipedia.org/wiki/Prostitution_law), reflecting the legal situation in each country at the time our data were collected in 2010. The categories or groups were:

Group A, which included countries where prostitution is legal and regulated: Austria, Switzerland, Germany, Greece, Hungary, Latvia, Netherlands, and Turkey (n = 8).

Group B, which included countries where prostitution is legal but not regulated: Belgium, Bulgaria, Cyprus, Czech Republic, Denmark, Estonia, Spain, Finland, France, Ireland, Italy, Luxemburg, Malta, Poland, Portugal, Slovenia, Slovakia, and the United Kingdom (n = 18).

Group C, which included countries where prostitution is illegal and the buyers (clients) are criminally prosecuted but those who sell sex are not: Norway and Sweden (n = 2).

Group D, which included countries where prostitution is illegal and the sellers are criminally prosecuted: Belarus, Bosnia & Herzegovina, Croatia, Lithuania, Macedonia (F.Y.R.), Moldova, Romania, Serbia, Russia, and Ukraine (n = 10).

Further, with respect to independent variables, in accordance with our aim and to fill gaps in the literature, we also examined five sociodemographic characteristics: age, education, occupation, size of settlement where respondents lived, and region of origin. Education was measured according to the International Standard Classification of Education (ISCED 1997 Levels 1–6). Respondents could select one of seven occupation categories (employed full-time, employed part-time, self-employed, unemployed, student, retired, and long-time sick leave or medically retired, or other). Settlement size was dichotomized as “large cities” of at least 500,000 inhabitants versus smaller settlements. The region of origin of men engaging in TS was measured by asking for their country of birth. European countries were grouped according to a modified classification of the UN Group of Experts on Geographical Names (2006; for a description, see http://emis-project.eu/sub-regions). Countries outside of Europe were grouped in accordance with the non-European World Health Organization (WHO) regions (for a description, see http://who.int/about/regions).

All aforementioned variables were entered stepwise into two separate multivariable logistic regression models (Statistical Package for the Social Sciences Version 20; IBM Corporation, New York, NY). We constructed one model for selling sex (been paid by a man to have sex = sellers) and one for buying sex (paid a man for having sex = buyers or “clients”) to determine variables independently associated with each TS behavior. The adjusted odds ratios (AORs) and 95% confidence intervals for the variables remaining in the models are shown.

RESULTS

Of the 184,469 responses submitted, 180,988 (98.1%) met the inclusion criteria, of which 174,209 (94.4%) from 38 European countries passed the internal data validity checks and were thus included in the analyses. As Table 1 shows, about half of the men were aged 25 to 39 years old (range = 13–89 years), almost three quarters had postsecondary education (ISCED Level 4 or higher), and 54% were employed full-time. Slightly more than half of the men reported living in a large city (i.e., at least 500,000 inhabitants), and 86% were born in their current country of residence, with the two largest regions of origin for other men being West Europe and Latin America/Caribbean. Given our structural context interest, Table 1 also shows the participants’ sociodemographic profile by prostitution law group. Forty-two percent of the participants resided in Group A countries, 47% in Group B countries, 3% in Group C countries, and 8% in Group D countries.

Table 1.

Sociodemographic Characteristics of EMIS Participants, Total and Grouped by Prostitution Law Group Countries (Groups A–D)

	Total	Group A	Group B	Group C	Group D
Status of transactional sex:		Legal, regulated	Legal, not regulated	Illegal; clients criminally prosecuted	Illegal; sellers criminally prosecuted	Pearson's Chi-square (two-sided p value)
Countries included (by country code top-level domain):		AT, CH, DE, GR, HU, LV, NL, TK	BE, BG, CY, CZ, DK, EE, ES, FI, FR, IE, IT, LU, MT, PL, PT, SI, SK, UK	NO, SE	BA, BY, HR, LT, MD, MK, RO, RS, RU, UA
	N (%)	N (%)	N (%)	N (%)	N (%)
Sum	174,209 (100.0)	74,813 (100.0)	82,126 (100.0)	5,228 (100.0)	12,042 (100.0)
Sellers
Has frequently been paid for sex in country of residencea	1,650 (1.0)	683 (1.0)	829 (1.1)	19 (0.4)	119 (1.1)	X2 = 21.9 p < .001
Has been paid for last sex abroad	632 (1.5)	234 (1.5)	322 (1.5)	11 (0.7)	65 (3.0)	X2 = 39.2 p < .001
Buyers
Has frequently paid for sex in country of residenceb	4,910 (3.1)	2,246 (3.3)	2,286 (3.0)	40 (0.9)	338 (3.1)	X2 = 87.8 p < .001
Has paid for last sex abroad	1,970 (4.7)	810 (5.2)	943 (4.2)	81 (5.1)	136 (6.2)	X2 = 31.7 p < .001
Age
< 25 years	40,673 (23.3)	16,990 (22.7)	18,855 (23.0)	1,205 (23.0)	3,623 (30.1)	X2 = 1,965.4p < .001
25–39 years	85,193 (48.9)	35,174 (47.0)	40,609 (49.4)	2,340 (44.8)	7,070 (58.7)
40+ years	48,343 (27.8)	22,649 (30.3)	22,662 (27.6)	1,683 (32.2)	1,349 (11.2)
Educationc
ISCED 1	1,904 (1.2)	709 (1.0)	1,183 (1.5)	18 (0.3)	84 (0.7)	X2 = 18,237.8 p < .001
ISCED 2	12,006 (6.9)	5,722 (7.7)	5,730 (7.0)	254 (4.9)	300 (2.5)
ISCED 3	33,205 (19.2)	21,087 (28.4)	9,508 (11.7)	1,096 (21.1)	1,1514 (12.7)
ISCED 4	39,394 (22.8)	18,859 (25.4)	17,663 (21.7)	836 (16.1)	2,036 (17.0)
ISCED 5	44,496 (25.7)	11,351 (15.3)	23,926 (29.3)	2,393 (46.0)	6,826 (57.1)
ISCED 6	41,773 (24.2)	16,457 (22.2)	23,517 (28.8)	609 (11.7)	1,190 (10.0)
Occupation
Employed full-time	93,575 (53.7)	41,481 (55.4)	42,131 (51.3)	2,916 (55.8)	7,047 (58.5)	X2 = 1,661.7 p < .001
Employed part-time	9,671 (5.6)	4,222 (5.6)	4,192 (5.1)	357 (6.8)	900 (7.5)
Unemployed	10,457 (6.0)	3,947 (5.3)	5,614 (6.8)	260 (5.0)	636 (5.3)
Self-employed	20,196 (11.6)	8,578 (11.5)	10,285 (12.5)	397 (7.6)	936 (7.8)
Student	26,978 (15.5)	10,066 (13.5)	14,158 (17.2)	895 (17.1)	1,859 (15.4)
Retired	6,488 (3.7)	3,045 (4.1)	3,046 (3.7)	273 (5.2)	124 (1.0)
Other	6,844 (3.9)	3,474 (4.6)	2,700 (3.3)	130 (2.5)	540 (4.5)
Settlement size
< 500,000	92,363 (54.4)	41,747 (57.2)	43,680 (54.5)	2,691 (52.5)	4,245 (36.4)	X2 = 1,768.5 p < .001
≥ 500,000	77,539 (44.5)	31,208 (42.8)	36,478 (45.5)	2,430 (47.5)	7,423 (63.6)
Region of origind
Born in country of residence	146,311 (86.2)	64,798 (89.1)	66,533 (83.1)	4,321 (84.6)	10,659 (90.7)
EMIS regionse (Europe)
West Europe	4,160 (2.5)	1,096 (1.5)	2,963 (3.7)	63 (1.2)	39 (0.3)	X2 = 9,196.4 p < .001
Northwest Europe	578 (0.3)	140 (0.2)	268 (0.3)	167 (3.3)	3 (0.0)
Central Europe (West)	2,926 (1.7)	1,458 (2.0)	1,312 (1.6)	90 (1.8)	66 (0.6)
Southwest Europe	2,367 (1.4)	866 (1.2)	1,431 (1.8)	46 (0.9)	24 (0.2)
Northeast Europe	273 (0.2)	45 (0.1)	166 (0.2)	21 (0.4)	41 (0.3)
Central Europe (East)	1,623 (1.0)	794 (1.1)	745 (0.9)	53 (1.0)	31 (0.3)
Southeast Europe (European Union)g	855 (0.5)	422 (0.6)	405 (0.5)	22 (0.4)	6 (0.1)
Southeast Europe (non-European Union)h	913 (0.5)	472 (0.6)	254 (0.3)	32 (0.6)	155 (1.3)
East Europe	1,434 (0.8)	493 (0.7)	267 (0.3)	23 (0.5)	651 (5.5)
WHO regionsf (outside Europe)
United States/Canada	1,411 (0.8)	637 (0.9)	720 (0.9)	36 (0.7)	18 (0.2)
Latin America/Caribbeani	3,806 (2.2)	725 (1.0)	2,967 (3.7)	98 (1.9)	16 (0.1)
Eastern Mediterranean Regionj	597 (0.4)	178 (0.2)	378 (0.5)	33 (0.6)	8 (0.1)
African Regionk	974 (0.6)	160 (0.2)	794 (1.0)	10 (0.2)	10 (0.1)
South East Asial	412 (0.2)	166 (0.2)	205 (0.3)	36 (0.7)	5 (0.0)
Australia, New Zealand	425 (0.3)	74 (0.1)	336 (0.4)	12 (0.2)	3 (0.0)
Western Pacific Regionm	617 (0.4)	205 (0.3)	353 (0.4)	42 (0.8)	17 (0.1)

Note. EMIS = European Men who have sex with men Internet Survey; WHO = World Health Organization.

aHaving frequently been paid for sex was operationalized as having been paid by a man to have sex 11 or more times in the previous 12 months.

bHaving frequently paid for sex was operationalized as having paid a man to have sex three or more times in the previous 12 months.

cISCED = International Classification of Education (from 1997), where ISCED 1 is primary education and ISCED 6 is the second stage of tertiary education (e.g., Ph.D.).

dRegion of origin refers to those who were not born in their current country of residence (reference group) and can thus be regarded as migrants.

eEuropean countries of origin were grouped into nine European subregions (see http://www.emis-project.eu/sub-regions). Expressed by country code top-level domain, these are: West Europe = BE, FR, IE, NL, UK; Northwest Europe = DK, FI, NO, SE; Central Europe (West) = AT, CH, DE, LU; Southwest Europe = ES, GR, IT, PT; Northeast Europe = EE, LT, LV; Central Europe (East) = CZ, HU, SI, SL, PL; Southeast Europe (EU) = BG, CY, MT, RO; Southeast Europe (non-European Union) = BA, HR, MK, RS, TR; East Europe = BY, MD, RU, UA.

fNon-European countries of origin were grouped into non-European WHO regions (http://www.who.int/about/regions); the WHO region of the Americas was split into United States/Canada versus all others, and the WHO Western Pacific region was split into Australia/New Zealand versus all others.

gMigrant men who have sex with men (MSM) from Southeast Europe (European Union) were predominantly from Romania (n = 533) and Bulgaria (n = 217).

hMigrant MSM from Southeast Europe (non-European Union) were predominantly from Turkey (n = 198), Croatia (n = 177), Serbia (n = 244), and Bosnia (n = 161). Migrant MSM from East Europe were predominantly from Russia (n = 563), Ukraine (n = 259), and Kazakhstan (n = 215).

iMigrant MSM from Latin America were predominantly from Brazil (n = 1,184), Mexico (n = 435), Colombia (n = 422), Venezuela (n = 385), and Argentina (n = 379).

jMigrant MSM from the Eastern Mediterranean region were predominantly from Morocco (n = 102), Iran (n = 77), and Lebanon (n = 73).

kMigrant MSM from the African region were predominantly from South Africa (n = 358), Mozambique (n = 92), and Angola (n = 90).

lMigrant MSM from South East Asia were predominantly from India (n = 315) and Thailand (n = 80).

mMigrant MSM from the Western Pacific region were predominantly from China (n = 159), Malaysia (n = 122), and the Philippines (n = 97).

In the previous year, 12.2% of the EMIS sample engaged in TS in the country in which they resided. Buying sex was more common than selling as 11,219 men (7.0%) reported buying sex and 7,283 men (4.5%) reported selling sex at least once in the previous year. A few men reported having both bought and sold sex in the previous year (0.7%, n = 1,057). While the range for both selling and buying sex was 1 to more than 50 times, the majority of those who engaged in TS did so 1 to 2 times (51.8% selling, 60.1% buying), and only a small minority of those who engaged in TS reported it was more than 50 times in the previous year (7.5% selling, 1.3% buying). Less than 5% of the EMIS sample (4.7%) reported that they had bought sex the last time they had sex abroad and 1.5% had sold sex at this occasion.

Being Paid for Sex/Selling Sex

Frequent sellers made up 1.0% (n = 1,650) of respondents who answered the question about being paid for sex (Table 2) and 19.7% of those who had sold sex in the last year. The Hosmer-Lemeshow test showed a good fit of the multivariable logistic regression model, X (8) = 14.74, p = .65. The regression results demonstrate that within this model, all predictor variables were significantly and independently related to the criterion variable of frequent selling. Men residing in Norway and Sweden (Group C countries) were less likely than men residing in any other country group to frequently sell sex to other men (AOR = 0.42). Compared with men younger than 25 years old, the likelihood of frequently selling sex was significantly less among men older than 40 years old (AOR = 0.21) as well as men in the 25- to 39-year-old age group (AOR = 0.57). The strong relationship between TS and age is illustrated in Figure 1. Men with some or higher education were consistently less likely to frequently sell sex relative to men who had primary education only (AORs ranging from 0.63 to 0.16). Correspondingly, relative to men in full-time employment, those who were self-employed, unemployed, part-time employed, retired, or held some other non-full-time employment were consistently more likely to report that they frequently sold sex (AORs from 2.10 to 3.79). The odds of frequently selling sex was more than 2 (AOR = 2.02) for men residing in a large city relative to men residing in smaller settlements. Men who were not born in their country of residence (“immigrant” men) were significantly more likely to frequently sell sex compared with men who were born in their country of residence. This was particularly true for men originating in Latin America/Caribbean, Southeast and Central East Europe, the WHO Eastern Mediterranean Region, and the WHO African Region (AORs from 2.57 to 4.96).

Table 2.

Multivariable Regression Analyses of Having Been Paid for Sex (‘Sellers’) in the Last 12 Months and Having Been Paid for Sex at Last Sex Abroad

	Has Frequently Been Paid for Sex in Country of Residencea		Has Been Paid for Last Sex Abroad
	N (%)	AOR [95% CI]	N (%)	AOR [95% CI]
Group A: legal, regulated	683 (1.0)	Ref.	234 (1.5)	Ref.
Group B: legal, not regulated	829 (1.1)	1.01 [0.91, 1.13]	322 (1.5)	0.93 [0.78, 1.12]
Group C: paying for sex is illegal	19 (0.4)	0.42 [0.26, 0.68]	11 (0.7)	0.54 [0.29, 0.99]
Group D: being paid for sex is illegal	119 (1.1)	0.94 [0.76, 1.17]	65 (3.0)	1.59 [1.17, 2.14]
Sum	1,650 (1.0)		632 (1.5)
Age
< 25 years	639 (1.8)	Ref.	264 (4.6)	Ref.
25–39 years	818 (1.0)	0.57 [0.51, 0.65]	306 (1.4)	0.33 [0.27, 0.40]
40+ years	193 (0.4)	0.21 [0.18, 0.26]	62 (0.4)	0.09 [0.06, 0.12]
Educationb
ISCED 1	67 (3.9)	Ref.	33 (9.7)	Ref.
ISCED 2	229 (2.1)	0.63 [0.47, 0.84]	64 (3.9)	0.53 [0.32, 0.80]
ISCED 3	355 (1.2)	0.39 [0.29, 0.51]	121 (2.2)	0.31 [0.20, 0.48]
ISCED 4	408 (1.1)	0.32 [0.24, 0.43]	138 (1.9)	0.22 [0.14, 0.34]
ISCED 5	355 (0.9)	0.24 [0.18, 0.32]	153 (1.3)	0.17 [0.11, 0.26]
ISCED 6	213 (0.5)	0.16 [0.12, 0.21]	118 (0.8)	0.13 [0.09, 0.21]
Occupation
Employed full-time	492 (0.6)	Ref.	184 (0.8)	Ref.
Employed part-time	134 (1.5)	2.10 [1.71, 2.56]	44 (2.1)	2.07 [1.47, 2.92]
Unemployed	226 (2.4)	2.95 [2.49, 3.50]	68 (3.6)	3.12 [2.32, 4.21]
Self-employed	358 (1.9)	3.79 [3.28, 4.39]	146 (2.3)	3.44 [2.73, 4.32]
Student	262 (1.1)	1.12 [0.96, 1.34]	110 (2.4)	1.20 [0.91, 1.58]
Retired	45 (0.8)	2.46 [1.78, 3.40]	16 (1.1)	2.90 [1.65, 5.09]
Other	133 (2.4)	2.76 [2.22, 3.43]	64 (5.3)	4.82 [3.51, 6.63]
Settlement size
< 500,000	616 (0.7)	Ref.	246 (1.4)	Ref.
≥ 500,000	975 (1.3)	2.02 [1.81, 2.25]	364 (1.6)	1.26 [1.07, 1.50]
Region of originc/destinationd	Born in country of residence	1,189 (0.9)	Ref.	n.a.
EMIS regionse (Europe)
West Europe	27 (0.7)	1.04 [0.71, 1.53]	99 (1.3)
Northwest Europe	2 (0.4)	0.47 [0.12, 1.89]	19 (1.6)
Central Europe (West)	22 (0.8)	1.10 [0.72, 1.68]	156 (2.5)
Southwest Europe	22 (1.0)	0.98 [0.62, 1.55]	108 (1.1)
Northeast Europe	2 (0.8)	0.76 [0.19, 3.09]	1 (0.5)	Not included
Central Europe (East)	35 (2.3)	2.60 [1.84, 3.68]	18 (1.1)
Southeast Europe (European Union)	36 (4.5)	3.99 [2.80, 5.68]	8 (1.4)
Southeast Europe (non-European Union)	21 (2.4)	2.62 [1.68, 4.09]	20 (1.4)
WHO regionsf (outside Europe)
East Europe	17 (1.3)	1.40 [0.86, 2.29]	8 (1.7)
United States/Canada	12 (0.9)	1.15 [0.61, 2.17]	26 (0.9)
Latin America/Caribbean	189 (5.2)	4.96 [4.17, 5.90]	31 (1.5)
Eastern Mediterranean Region	13 (2.4)	2.69 [1.49, 4.84]	21 (2.0)
African Region	18 (2.0)	2.57 [1.57, 4.21]	10 (2.4)	Not included
Southeast Asia	4 (1.0)	1.42 [0.52, 3.83]	5 (0.3)
Australia, New Zealand	4 (1.0)	1.43 [0.53, 3.88]	6 (1.1)
Western Pacific Region	3 (0.5)	0.72 [0.23, 2.25]	6 (0.7)	<

Note. Percentages refer to the proportion of men with the respective behavior within each category/layer.

Ref. = reference group; AOR = adjusted odds ratio; CI = confidence interval; EMIS = European Men who have sex with men Internet Survey; WHO = World Health Organization. Bold values indicate statistically significant differences; n.a. = not applicable.

aHaving frequently been paid for sex was operationalized as having been paid by a man to have sex 11 or more times in the previous 12 months.

bISCED = International Classification of Education (1997), where ISCED 1 is primary education and ISCED 6 is second stage of tertiary education (e.g., Ph.D.).

cRegion of origin refers to those who were not born in their current country of residence (reference group) and can thus be regarded as migrants.

dFor having been paid for the last sex abroad, the listed regions refer to where this occurred. Bold values indicate a higher-than-average proportion.

eEuropean countries of origin were grouped into nine European subregions (see http://www.emis-project.eu/sub-regions).

fNon-European countries of origin were grouped into non-European WHO regions (http://www.who.int/about/regions); the WHO region of the Americas was split into United States/Canada versus all others, and the WHO Western Pacific region was split into Australia/New Zealand versus all others.

FIGURE 1.

Transactional sex in the last 12 months among men who have sex with men (MSM) in Europe, by age group.

The sociodemographic profile of men who sold sex abroad mirrored that of men who sold sex in their country of residence (Table 2). Men residing in Norway and Sweden were less likely to have sold sex at last sex abroad (AOR = 0.54). In contrast, men residing in a Group D country, where prostitution is illegal and the sellers are criminally prosecuted, were more likely to have sold sex at last sex abroad (AOR = 1.59). As illustrated in Figure 1, men who sold sex at last sex abroad were also less likely to be aged 25 to 39 years old (AOR = 0.33) or older than 40 years old (AOR = 0.09). Men who sold sex abroad were less likely to have higher education (AORs from 0.53 to 0.13), and were more likely not to have full-time employment (AORs from 2.07 to 4.82) and to live in a large city (AOR = 1.26). We note that selling sex abroad was more frequently taking place (above average) in the following regions: West Central Europe, Africa Region, Eastern Mediterranean Region, and Northwest Europe.

Paying for Sex/Buying Sex

In Table 3, we show the profile of the 4,910 men (3.1% of EMIS respondents and 39.9% of all men who had paid for sex in the last year) who paid a man for sex three or more times in the previous year. The Hosmer-Lemeshow test showed that the multivariable logistic regression model had a good fit, X (8) = 11.22, p = .19). As shown, with the exception of education, all of the predictor variables remained significantly associated with frequently buying sex in the context of the other variables. Buyers of sex were less likely to reside in a Group B or C country (AORs = 0.92 and 0.26, respectively) and were more likely to reside in a Group D country (AOR = 1.33). They were more likely to be older (AOR for age 40 years and older = 10.19 and AOR for age 25 to 39 years old = 3.37; Figure 1), to have full-time employment, to be self-employed or retired (AORs from 1.26 to 1.36), and to reside in a large city (AOR = 1.31). Relative to men who were born in their country of residence, frequent buyers of sex were less likely to come from East Central Europe (AOR = 0.52) and were more likely to come from the Eastern Mediterranean region (AOR = 1.97).

Table 3.

Multivariable Regression Analyses of Having Paid for Sex (‘Buyers’) in the Last 12 Months and at Last Sex Abroad

	Has Frequently Paid for Sex in Country of Residencea		Has Paid for Last Sex Abroad
	N (%)	AOR [95%-CI]	N (%)	AOR [95% CI]
Group A: legal, regulated	2,246 (3.3)	Ref.	810 (5.2)	Ref.
Group B: legal, not regulated	2,286 (3.0)	0.92 [0.86, 0.98]	943 (4.2)	0.87 [0.79; 0.96)
Group C: paying for sex is illegal	40 (0.9)	0.26 [0.18, 0.35]	81 (5.1)	0.98 [0.77; 1.26)
Group D: being paid for sex is illegal	338 (3.1)	1.33 [1.17, 1.52]	136 (6.2)	1.86 [1.52; 2.28)
Sum	4,910 (3.1)		1,970 (4.7)
Age
< 25 years	150 (0.4)	Ref.	68 (1.2)	Ref.
25–39 years	1,733 (2.2)	3.37 [2.81, 4.04]	696 (3.2)	2.26 [1.71, 2.99]
40+ years	3,027 (6.7)	10.19 [8.50, 12.23]	1,206 (8.5)	5.72 [4.32, 7.58]
Educationb
ISCED 1	54 (3.1)	Ref.	20 (5.8)	Ref.
ISCED 2	286 (2.6)	0.83 [0.61, 1.14]	116 (7.1)	1.14 [0.69, 1.88]
ISCED 3	854 (2.8)	0.95 [0.70, 1.28]	282 (5.2)	0.84 [0.52, 1.36]
ISCED 4	857 (2.4)	0.99 [0.73, 1.34]	358 (4.8)	0.97 [0.60, 1.55]
ISCED 5	1,255 (3.0)	1.01 [0.75, 1.36]	567 (4.7)	0.85 [0.53, 1.36]
ISCED 6	1,582 (4.0)	1.17 [0.87, 1.57]	618 (4.2)	0.74 [0.46, 1.18]
Occupation
Employed full-time	2,863 (3.2)	Ref.	1,133 (4.7)	Ref.
Employed part-time	200 (2.2)	0.76 [0.66, 0.89]	78 (3.6)	0.82 [0.65, 1.04]
Unemployed	174 (1.8)	0.66 [0.56, 0.78]	61 (3.2)	0.79 [0.60, 1.03]
Self-employed	1,017 (5.3)	1.36 [1.26, 1.47]	382 (6.0)	1.19 [1.05, 1.35]
Student	86 (0.4)	0.34 [0.27, 0.43]	48 (1.1)	0.59 [0.36, 0.68]
Retired	396 (6.9)	1.26 [1.12, 1.42]	195 (13.6)	1.93 [1.62, 2.31]
Other	174 (3.1)	1.08 [0.91, 1.29]	73 (6.0)	1.24 [0.95, 1.61]
Settlement size
< 500,000	2,176 (2.6)	Ref.	923 (5.3)	Ref.
≥ 500,000	2,549 (3.5)	1.31 [1.24, 1.40]	969 (4.2)	0.85 [0.77, 0.93]
Region of originc/destinationd	Born in country of residence	4,065 (3.0)	Ref.	n.a.
EMIS regionse (Europe)
West Europe	145 (3.7)	0.91 [0.76, 1.08]	177 (2.3)
Northwest Europe	13 (2.4)	0.82 [0.47, 1.44]	8 (0.7)
Central Europe (West)	90 (3.2)	0.90 [0.72, 1.12]	216 (3.4)
Southwest Europe	60 (2.6)	0.89 [0.68, 1.15]	267 (2.7)
Northeast Europe	3 (1.1)	0.54 [0.17, 1.69]	8 (3.8)	Not included
Central Europe (East)	20 (1.3)	0.52 [0.33, 0.81]	142 (8.9)
Southeast Europe (European Union)	12 (1.5)	0.72 [0.40, 1.27]	31 (5.6)
Southeast Europe (non-European Union)	16 (1.9)	0.73 [0.44, 1.21]	59 (4.1)
WHO regionsf (outside Europe)
East Europe	33 (2.5)	0.91 [0.64, 1.30]	14 (2.9)
USA/Canada	50 (3.7)	0.82 [0.62, 1.10]	51 (1.8)
Latin America / Caribbean	102 (2.8)	1.06 [0.86, 1.31]	203 (10.0)
Eastern Mediterranean Region	32 (5.8)	1.97 [1.36, 2.85]	173 (16.0)
African Region	37 (4.0)	1.03 [0.74, 1.45]	30 (7.0)	Not included
South East Asia	9 (2.3)	1.10 [0.56, 2.14]	402 (25.9)
Australia, New Zealand	8 (2.0)	0.51 [0.25, 1.03]	4 (0.7)
Western Pacific Region	8 (1.4)	0.51 [0.24, 1.09]	40 (4.3)

Note. Percentages refer to the proportion of men with the respective behavior within each category/layer. Ref. = reference group; AOR = adjusted odds ratio; CI = confidence interval; EMIS = European Men who have sex with men Internet Survey; WHO = World Health Organization. Bold values indicate statistically significant differences; n.a. = not applicable.

aHaving frequently paid for sex was operationalized as having been paid by a man to have sex three or more times in the previous 12 months.

bISCED = International Classification of Education (1997), where ISCED 1 is primary education and ISCED 6 is the second stage of tertiary education (e.g., Ph.D.).

cRegion of origin refers to those who were not born in their current country of residence (reference group) and can thus be regarded as migrants.

dFor having paid for the last sex abroad, the listed regions refer to where this occurred. Bold values indicate a higher-than-average proportion.

eEuropean countries of origin were grouped into nine European subregions (see http://www.emis-project.eu/sub-regions).

fNon-European countries of origin were grouped into non-European WHO regions (http://www.who.int/about/regions); the WHO region of the Americas was split into United States/Canada versus all others, and the WHO Western Pacific region was split into Australia/New Zealand versus all others.

Concerning the men who reported that they had paid for sex the last time they had sex abroad, these men were less likely to reside in a Group B country (AOR = 0.87) and were more likely to reside in a Group D country (AOR = 1.86). We note that no difference could be seen for men from Group C countries. Additionally, the men who reported buying sex the last time they had sex abroad were older (AORs = 2.26 and 5.72) rather than younger than age 25 years (Figure 1), were more likely to be self-employed (AOR = 1.19) or retired (AOR = 1.93), and were less likely to be a student (AOR = 0.59). However, they were less likely to reside in a large city (AOR = 0.85). The behavior of buying sex abroad more frequently took place (above average) in the following regions: South East Asia, Eastern Mediterranean Region, Latin America/Caribbean, East Central Europe, African Region, and Southeast Europe European Union countries.

DISCUSSION

We examined the sociodemographic characteristics of MSM engaged in the supply and demand sides of TS within the broader context of prostitution laws. Our results taken from a large population of gay, bisexual, and other MSM in Europe document that a segment of the general population of these men engages in TS. In the previous year, 12.2% of the EMIS sample engaged in TS in their country of residence, but for the great majority, this was an infrequent behavior, suggesting that TS among European MSM primarily occurs as a matter of opportunity. As expected, clients formed the largest group of people involved in TS. Seven percent reported buying sex and 4.5% reported selling sex at least once in the previous year in their country of residence (0.7% reported having both bought and sold sex). To our knowledge, there are no reliable estimates of TS among MSM, and the lack of similarity across studies in samples, operationalizations, and measurements complicates comparisons of rates across studies (for a discussion of this issue, see, e.g., Minichiello & Scott, 2014). However, studies conducted on MSM in high-income countries have suggested that up to 37% have ever been paid or have ever paid for sex with another man (Koken et al., 2005; Prestage et al., 2007; Weber et al., 2001). A recent study methodologically similar to ours, conducted among 2,306 Australian gay and bisexual men, showed that 16.7% reported ever selling sex to another man and about 25% reported ever buying sex from another man (Prestage et al., 2014). Studies that have been conducted on the general population of mainly heterosexual men suggest that the proportion who report buying sex from a sex worker (almost exclusively female) in the previous year varies by country and European region and ranges from a low of less than 1% to 11% (Carael, Slaymaker, Lyerla, & Sarkar, 2006).

We used a broad definition of TS that included informal trading outside the realm of commercial sex work. Thus, with this article, we neither aim to nor have an empirical basis to address the regulation of sex work. In most countries, prostitution laws have been structured with a female sex worker in mind, and unlike female sex work, male sex work is quite invisible—perhaps increasingly so with the emergence of new technologies (see, e.g., Baral et al., 2014; Grov et al., 2014; Minichiello & Scott, 2014; Walby, 2012). Nonetheless, it is important in our analysis to consider the potential influence of the continuum of legal approaches to TS across Europe on TS behavior among MSM. Although prostitution is legal in almost all European Union countries, it is illegal in most non-European Union countries in Europe, particularly in the Nordic countries and successor states of Yugoslavia and the Soviet Union. Our results suggest that TS laws may be associated with MSM's TS behavior. Interestingly, although criminalization of selling sex seemed unrelated to MSM selling sex, it seemed to increase the odds of MSM buying sex when only the seller is criminalized. Men residing in countries where buyers are criminally prosecuted (Norway and Sweden) had lower odds of both frequently selling and buying sex. However, given the small number of countries in this group and their homogeneity, the results for Norway and Sweden may be a reflection of broader cultural factors, such as nationals’ high income and norms about sex. It bears mentioning that Sweden in 1989 was the first country in Europe to criminalize the purchase of sexual services, with the perspective that prostitution is by definition coercive. Norway (and Iceland) followed suit in 2009. Of note, among MSM who resided in a country where sellers are prosecuted, the rate of TS abroad was considerably higher, both with regard to selling and buying sex, relative to MSM in the other country groups. Rates of buying sex were also considerably higher abroad than in one's country of residence among MSM in both Group C and Group D countries, as compared with MSM in countries where TS is legal. This may suggest that in countries where TS is illegal, the behavior is being relocated to countries where it is not. Stated differently, the legal situation of TS in one country might influence men to shift their TS encounters to countries with supportive TS legislation. According to Baral and colleagues (2014), men who sell sex in Eastern Europe—where TS is illegal—are known to migrate to countries in Central and Western Europe, such as Germany and Switzerland—where TS is legal. The extralegal status of TS, in that it complicates sellers’ efforts when dealing with potential and actual clients, makes up one aspect of structural factors that are interwoven with TS. For example, in previous studies, perceived neighborhood disorder and poverty have been associated with TS (Harcourt, Egger, Donovan, 2005; Latkin, Curry, Hua, & Davey, 2007). Additional structural and social issues associated with TS should be examined in future studies.

In our results, the influence of age on TS was especially striking. As in Prestage and colleagues’ recent study in Australia (2014), among both sellers and buyers of sex, TS was strongly associated with age, but with one important differential. Selling sex steadily decreased with age and was most likely among men younger than age 25, while buying sex steadily increased with age, with the odds of buying being 10.2 for men aged 40 years and older. This clearly reflects the traditional age dynamic of older individuals paying younger men for sex. It also reflects previous research on male escorts (Fipaza, Karlyn, Tun, Mbizvo, & Manzini 2011; Koken et al., 2005; Mimiaga et al., 2008), underscoring that youth is a valuable commodity in TS. In fact, in his qualitative study, Walby (2012) highlights that older escorts recognized the limits of their aging bodies and deliberately strategized to stay in the escort business.

Like Prestage and colleagues (2014) and Weber and colleagues (2001), we found that in addition to age, sociodemographic factors associated with selling sex were low education and not having full-time employment. This association between selling sex and low SES suggests an economic imperative to which TS is a response. Consistent with the empowerment paradigm explaining that selling sex is driven by lower levels of human capital (McCarthy et al., 2014), in our study, selling sex seems to be at least partially motivated by financial need, where the TS income may not be a principal source of income but at any rate may supplement other sources of revenue.

The possible economic disempowerment link between low SES and selling sex is strengthened by the finding that MSM who frequently sold sex also were more likely not to be born in their country of residence. That is, they were more likely to be ‘migrants,’ especially from Latin America/Caribbean, Southeast Europe, and Africa. Men selling sex not only seemed to be economically vulnerable, but also socially vulnerable. This finding is documented elsewhere (e.g., McCarthy et al., 2014). For example, Belza and colleagues (2001; Belza, for the EPI-VIH Study Group, 2005) discussed the issue of socioeconomic marginalization among MSM engaging in TS in relation to their similar research findings, which showed that large proportions of male sex workers in Spain came from another country, primarily in the regions of Latin America and North Africa. The researchers explained that because migrants historically experience greater challenges in establishing social networks and face barriers to accessing services and financial and material resources, TS among MSM may reflect social disadvantage. Moreover, given our finding of striking SES differences in TS by payment direction, in these TS encounters, the relationships reflected a power differential and social hierarchy between two men having sex. Contrary to MSM who sold sex, MSM who paid for sex were considerably older and more likely to have steady employment. These associations between TS and SES held also for last sexual encounter abroad—these patterns are highlighted by our finding that buying sex abroad was particularly prevalent within economically disadvantaged regions such as Latin America/Caribbean, Northern Africa, and Southeast Asia.

Our study comes with limitations. Although large and diverse, this nonrandom sample is not necessarily representative of the larger population of MSM in Europe and the generalizability of our results is uncertain. All data were self-reported and are subject to the common limitations of cross-sectional survey research, including recall error and social desirability bias, which might particularly concern illegal behavior. In this analysis and related EMIS analyses concerning illegal behavior (e.g., drug use), however, we have detected no indications of biased reporting. We cannot examine causation in the analysis. TS is a highly heterogeneous activity, with transactions taking place in various settings (e.g., saunas, public toilets, clubs, on the street) and arranged in various ways (e.g., advertisements, in-person contact). Not only were such variations unmeasured in our study (except for TS abroad, which was primarily arranged via online escort profiles), but there are likely also a spectrum of experiences and health risks among men who trade sex with other men that were unexamined in our analysis. A different operationalization of frequent TS could produce different results but would also on the one hand run the risk of including infrequent TS and, on the other hand, run the risk of analyses being underpowered. Also, a different, and possibly more nuanced, categorization of prostitution laws could produce different results. However, the four-group categorization used in our analysis was the only one we identified that accurately covered all countries and was up-to-date.

Limitations notwithstanding, our study results of TS among European MSM in the context of prostitution laws reveal that laws may be associated with MSM's TS behavior and there are important differentials between men on the supply side and those on the demand side of TS. First, the age dynamic of older individuals paying younger ones for sex highlights that youth is a main tradable commodity. Second, our results affirm the importance of SES to TS among MSM. The striking SES differences in TS by payment direction suggest both a power differential and a leading role of socioeconomic factors in TS. The characteristics of men who sell sex to other men suggest this is an economically and socially vulnerable group. Our results argue for a continued social and cultural focus on MSM who engage in TS. However, although more research is needed that pays attention to MSM as a whole engaged in TS within a structural and individual socioeconomic context, it would also be important to examine TS within a broader health context.