Salut Fabrice, While I fully agree with you that we don't know enough about the cost of discrimination towards LGBT people, it is to my view less of an issue of interest than an issue of measurement. It is easier to obtain information (especially since it is often self-reported) on ethnicity, race, caste and gender than sexual orientation in survey context. Suppose that a question on sexual orientation was introduced in household surveys. Given the social stigma, it is very likely that the ones who suffer the least from discrimination would be the ones who would truthfully answer the question, leading to a dramatic underestimation of the cost of LGBT discrimination; you could even end up finding better outcomes for (self-reported) LGBT! It is therefore in my view methodological shortcoming rather than lack of interest that is driving our ignorance on this issue.