Is sexual freedom a normal good?

In economics, a normal good is defined as a good where a consumer demands more of it as their income rises (in contrast, an inferior good is a good where a consumer demands less of it as their income rises). In other words, we say that normal goods have a positive income elasticity (while inferior goods have a negative income elasticity). For market goods, we can evaluate income elasticity by looking at how consumer purchases change as their income changes (controlling for other important variables, especially prices). For non-market goods such as sexual freedom, we can't look at consumer purchases but we might get some insight by looking at what happens as population-level income measures change (again controlling for other important variables, though as we are talking about non-market goods we can't control for price).

Which brings me to this recent paper (ungated earlier version here) by Feler Bose (Alma College), entitled "The determinants of sexual freedom from 1990 to 2010". To some extent it is disappointing that this paper was published in a short-form journal like Applied Economics Letters, as it doesn't give nearly enough detail on the methods. Briefly, Bose constructed a new sexual freedom index for U.S. states over the period 1990 to 2010, then regressed that index variable against median household income, as well as measures of ideology (of citizens, and of government), population density, median age, religion, and race. He finds a highly statistically significant and positive relationship between income and sexual freedom - states with higher median household incomes have higher sexual freedom. In other words, as incomes have increased people have demanded more sexual freedom, meaning that sexual freedom is a normal good.

However, I have a couple of gripes about the paper that lead me to doubt its conclusions. As mentioned above, there isn't enough detail on the new measure of sexual freedom. I have a real problem with this in particular:

This index codes sixteen variables that fall under two broad categories (marriage protection and sex crimes)... Each variable was coded on a four point scale (0, 1, 2, 3) with 3 indicating the most freedom. Further, each variable is given equal weight. To obtain the sexual freedom index value for a state during a year, each variable is added up. The higher the number, the broader the sexual freedom in that state.

I'm not sure how much validity you can assign to an index that is constructed in such a subjective manner. For instance, how were the values of 0-3 determined for each variable? Also, what is the basis for giving equal weight to "prostitution laws for pimp", "prostitution laws for prostitute", and "prostitution laws for John" (three of the variables included in the index), when these variables are likely to be closely correlated (or at least, there is no evidence provided to suggest that they aren't closely correlated). Moreover, the index used in this paper differs in construction from a similar index the author used in earlier work (described here). In that earlier paper, principal components analysis was used to construct the index (which is what I would have done, to ensure that variables that are closely correlated do not have undue influence over the index as a whole), and each variable was scored only 0 or 1 (rather than 0-3). In that earlier paper there were some odd results, with the first principal component placing a negative factor score on "age at consent" and "prostitution laws", making the overall index difficult to interpret. Perhaps that is why it wasn't used in the new paper?

It's a pity that the new index isn't available online and I can't see any follow-up work by the author. As he notes in his conclusion, some of his results (e.g. higher median age is associated with more sexual freedom) are difficult to justify. This may be simply due to a lack of robustness in the methods, which means that even though we have some results we can't say with any certainty whether sexual freedom is a normal good.