Saturday, November 27, 2010

A Woman. A Prostitute. A Slave.

New York Times article "A Woman. A Prostitute. A Slave." (November 27th, 2010)states that, when it comes to ending forced prostitution "There are no silver bullets, but the critical step is for the police and prosecutors to focus more on customers (to reduce demand) and, above all, on pimps."Corrections has a silver bullet: legalize prostitution. Essentially reiterating an earlier point ("Enabling prostitution" from September 3rd, 2010), it seems rather obvious that forced prostitution is an inferior input into prostitution services. Presumably, upon the legalization of prostitution, willing prostitute supply will shift out dramatically more than forced prostitutes.  This will cause forced prostitute quantity to shift down. We depict the silver bullet to severely decrease forced prostitution in the United States graphically below (click to enlarge).
The responsibility for the larger amount of unwilling prostitutes is on those that oppose legalized prostitution.


  1. except that's not what actually happens:

  2. Hi Hannah,

    Thanks for reading! While your link touches on relevant subjects, the paper doesn't actually address the causal links that are the heart of this blog. The paper's flaw is that it looks only at correlation, and indeed one might suspect that the same countries that legalize prostitution may have higher human trafficking, all things not being equal. After all, there are reasons they have these laws! Insofar as these are correlated with non-capitalist, non-democtatic, non-Western countries, we might expect higher rates of both human trafficking and legalized prostitution. But not because legalized prostitution causes human trafficking, but because something in these countries causes both.

    The use of partial controls in linear regressions like the ones the authors run are as likely to cause bias and inconsistency as running a regression without them, if you follow the logic of multiple regression. This can be a difficult point to understand, and indeed most of sociology seems to ignore it.

    Please, feel free to peruse our other articles to see examples of why papers like this may show correlation, but do not show causation, and why finding correlation through linear regression can so often fail to find even valid correlations.