Does Competition Eliminate Discrimination. Evidence from the Commercial Sex Market in Singapore (Tables)

Table 1: Sex Worker’s Characteristics Variable Age(Sex Worker) Experience Experience in Singapore Total Monthly Income(S$) Monthly Income from Sex Days/Week Worked Hours Worked (Weekday) Hours Worked (Weekend) Customers/Day Mean 26.3 2.9 1.7 4371 3246 6.3 9.0 11.0 3.9 Col % Country China Thailand Vietnam Indonesia SgpIndian 33.5 33.0 21.0 9.1 3.4 Education Illiterate Primary Secondary High School College 21.1 50.9 20.6 5.7 1.7 MaritalStatus Single/Relationship Married Divorced Widow 67.6 11.4 19.3 1.7 Beauty(1=lowest; 5=highest) 2 3 4 5 17.6 33.5 45.5 3.4 Urban Has Child Has Pimp 35.6 22.7 47.9 N 176 * Total Monthly Income includes both sex income and non-sex income from clients Table 2: Client’s Characteristics by Sex Worker’s Country of Origin Sex Worker’s Country of Origin SingIndiana China Thailand Vietnam Indonesia Ethnicity Chinese Caucasian Bangladeshi Indian Malay Middle Eastern JapaneseKorean Black(American) 39.9 15.7 17.4 9.6 10.2 3.4 3.4 0.3 34.7 9.4 23.4 12.1 6.4 3.8 9.4 0.8 33.8 16.2 17.6 4.9 7.0 10.6 9.2 0.7 42.2 13.3 21.7 9.6 2.4 1.2 8.4 1.2 10.7 3.6 39.3 39.3 3.6 0.0 0.0 3.6 36.4 13.1 20.6 10.6 7.4 4.4 6.8 0.7 Tourist Repeat Clients Age Dressb Looks 26.6 20.5 37.2 2.3 2.0 32.3 13.9 36.5 2.2 1.8 30.7 22.5 36.9 2.8 2.4 25.3 21.7 40.7 2.5 2.1 42.9 17.9 37.0 2.3 2.0 29.5 18.8 37.3 2.4 2.0 a b SingIndian are the Singaporean Indian sex workers, who typically have dark skin tone For both Dress and Looks, 1=worst and 5=best Total Table 3: Transaction Characteristics by Client’s Ethnicity Chinese White Bangladeshi Indian Mean Rating(1=dislike, 5=like most) 4.16 3.93 3.15 2.08 All Initial Price (S$) 73.80 91.22 49.04 67.14 71.28 Initial Price (Bargain Successful) 73.88 91.13 47.45 64.62 72.13 Transaction Price 68.07 81.34 43.74 57.69 65.78 Sex Worker Approaching First 72.8 91.8 67.3 42.9 71.9 Bargain Fails 4.2 1.0 24.0 25.7 11.8 Transaction Day (Weekend=1) 58.5 60.6 54.3 64.3 58.2 Transaction Time (Night=1) 70.2 92.2 96.2 91.8 84.7 Oral Sex 76.1 89.6 33.0 62.0 67.2 Anal Sex 4.9 45.8 1.9 8.0 15.0 Duration of Service 60 85 35 40 45 2.08 2.69 1.58 1.85 2.08 Sweet Talk(1=least, 5=most) 1 Top line based on up to 174 observations. Remaining rows based on up to 677 transactions in which the sex worker was the …rst to suggest a price. Table 4: Log Initial Price by Client Ethnicity (1) (2) (3) (4) (5) (6) 0.103 (0.043) 0.127 (0.052) 0.102 (0.044) 0.100 (0.044) 0.096 (0.044) 0.143 (0.040) -0.144 (0.033) -0.186 (0.040) -0.116 (0.034) -0.108 (0.034) -0.110 (0.034) -0.135 (0.033) 0.019 (0.034) 0.015 (0.040) 0.021 (0.034) 0.030 (0.034) 0.035 (0.034) -0.035 (0.034) Sex Worker Makes 1st O¤er Sample Yes Yes Yes Yes Yes No Dummies for missing Yes No Yes Yes Yes Yes Sex Acts No No Yes Yes Yes No Sex Venue No No No Yes Yes No Service Duration No No No No Yes No 676 0.522 0.072 0.221 548 0.521 0.280 0.301 676 0.552 0.017 0.190 676 0.564 0.020 0.201 676 0.569 0.008 0.185 810 0.457 0.279 0.278 Ethnicity (Base=Chinese) White Bangladesh Indian N R2 Within R2 Between R2 Overall Standard errors in parentheses p < 0:10, p < 0:05, p < 0:01 1 2 3 Columns (1)-(5) show results with sex worker …xed e¤ects using the subsample where sex workers initiate prices. Column (6) uses the full sample All regressions include sex worker …xed e¤ects and control for client’s ethnicity (estimates for Malay, Middle Eastern, East Asian, and black American clients are not shown because the number of observations is small), client’s perceived age and age squared, dummies for repeat customer, tourist, client’s dress rating above median, client’s attractivenesss above median, client’s “sweet talk” above median. All estimates except column (2) also control for indicators for the above variables being missing. Sex act dummies: 10 combinations of sex acts plus “not reported.” Table 5: Log Final Price by Client Ethnicity (1) (2) (3) (4) (5) (6) White 0.145 (0.043) 0.182 (0.056) 0.142 (0.044) 0.142 (0.044) 0.128 (0.044) 0.181 (0.037) Bangladesh -0.125 (0.035) -0.172 (0.045) -0.081 (0.036) -0.076 (0.036) -0.075 (0.036) -0.132 (0.032) 0.014 (0.037) 0.008 (0.045) 0.029 (0.037) 0.039 (0.037) 0.032 (0.037) 0.008 (0.034) Sex Worker Makes 1st O¤er Sample Yes Yes Yes Yes Yes No Dummies for missing Yes No Yes Yes Yes Yes Sex Acts No No Yes Yes Yes No Sex Venue No No No Yes Yes No Service Duration No No No No Yes No 596 0.506 0.025 0.157 469 0.496 0.282 0.265 596 0.540 0.000 0.121 596 0.551 0.000 0.130 596 0.561 0.000 0.128 709 0.481 0.269 0.239 Ethnicity (Base=Chinese) Indian N R2 Within R2 Between R2 Overall Standard errors in parentheses p < 0:10, p < 0:05, p < 0:01 1 Table 5 repeats the regressions exercises as described in Table 4 and use log …nal price as dependent variable. Table 6: Did Sex Worker Approach Client? By Client Ethnicity (1) Fixed E¤ect Logit (2) Logit (3) Fixed E¤ect Logit (4) Logit White 1.515 (0.842) 1.244 (0.772) [0.208] 0.574 (0.632) 0.769 (0.662) [0.138] Bangladesh -0.484 (0.481) -0.447 (0.404) [-0.075] -0.423 (0.420) -0.364 (0.390) [-0.065] -1.434 (0.469) -1.271 (0.373) [-0.213] -1.621 (0.412) -1.499 (0.365) [-0.268] Yes Yes No No No Yes No Yes 483 -154.849 586 -294.156 595 -202.402 650 -345.006 Ethnicity (Base=Chinese) Indian Sex Worker Makers 1st O¤er Sample Sex Worker Characteristics N Likelihood Standard errors in parentheses p < 0:10, p < 0:05, p < 0:01 1 2 3 4 All regressions include sex worker …xed e¤ects and control for client’s ethnicity (estimates for Malay, Middle Eastern, East Asian, and black American clients are not shown because there are fewer such observations and are not the interest of this analysis), client’s perceived age and age squared, dummies for repeat customer, tourist, client’s dress rating above median, client’ attractive rating above median, as well as for indicators of above variables being missing. Sex worker characteristics: quadratic in age, quadratic in experience,four country of origin dummies, marital status, education, beauty and English skills dummies. Fixed e¤ects logit drops sex workers who always or never approached clients. Their transactions and other cases where success/failure are perfectly predicted are excluded from N. Standard errors in parentheses, clustered on sex worker in columns 2 and 4. Marginal e¤ects in brackets. Table 7: Bargaining Failed By Client Ethnicity (1) Fixed E¤ect Logit (2) Logit (3) Fixed E¤ect Logit (4) Logit White -2.073 (1.515) -1.455 (1.420) [-0.139] -1.812 (1.094) -1.753 (1.287) [-0.181] Bangladesh 1.951 (0.781) 2.114 (0.543) [0.202] 1.358 (0.632) 1.584 (0.507) [0.164] Indian 1.705 (0.743) 1.886 (0.519) [0.181] 1.265 (0.589) 1.479 (0.471) [0.153] Yes Yes No No No Yes No Yes 338 -73.253 563 -171.722 438 -101.944 627 -205.423 Ethnicity (Base=Chinese) Only Sex Worker Makes 1st O¤er Sex Worker Characteristics N Likelihood Standard errors in parentheses p < 0:10, p < 0:05, p < 0:01 1 2 3 4 All regressions include sex worker …xed e¤ects and control for client’s ethnicity (estimates for Malay, Middle Eastern, East Asian, and black American clients are not shown because there are fewer such observations and are not the interest of this analysis), client’s perceived age and age squared, dummies for repeat customer, tourist, client’s dress rating above median, client’ attractive rating above median, client’s "sweet talk" rating above median. Except column (2), all other regressions also control for indicators of above variables being missing. Sex worker characteristics: quadratic in age, quadratic in experience,four country of origin dummies, marital status, education, beauty and English skills dummies. Fixed e¤ects logit drops sex workers who always or never approached clients. Their transactions and other cases where success/failure are perfectly predicted are excluded from N. Standard errors in parentheses, clustered on sex worker in columns 2 and 4. Marginal e¤ects in brackets.