We study patterns of firms' advertised preferences for gender, age, beauty and height in a sample of ads from a Chinese internet job board, and interpret these patterns using a simple employer search model. Overall, about 10 percent of job ads expressed a gender preference; these percentages were 25%, 8% and 2% respectively for age, beauty and height. The share of firms who placed at least one ad
... [Show full abstract] listing each of these preferences during our five-month sampling period was 34, 47, 29 and 10 percent respectively. Our model suggests that firms posting such ads must have strong underlying preferences for these characteristics. Consistent with our model, firms are more likely to express preferences for "US-prohibited" employee attributes when education and experience requirements for the job are low; thus our sample of ads for relatively skilled positions likely understates the frequency of advertised 'discriminatory' preferences in China. Cross-sectional patterns also suggest some role for customer discrimination, product market competition, and corporate culture. Using the recent collapse of China's labor market as a natural experiment, we find that firms' advertised wages, and their education and experience requirements, respond to changing labor market conditions in the direction predicted by our search model, while firms' advertised preferences for age, gender, height and beauty do not.