<p>Fama will win in due time. EMT forms the basis of literally all of modern finance as an ideal type. Even if markets are not always efficient as a practical matter (although they mostly are in the sense that Fama means it as opposed to the lay understanding), the very fact that we use the EMT framework to consider deviations on the margin and how to correct them speaks volumes to the models overall influence. In reality, I think his work is far more important than say Beckers or Fogels, since we can actually utilize its outcomes on a day-to-day basis. </p>
<p>As for Levitt, the critics have largely been repudiated on the abortion article (see below). The real battle over whether his work matters, that is, do we really care about cheating algorithms and the like, is one which I believe the profession is mostly on his side about as part of a larger strain of economic intellectual imperialism. </p>
<p>An unusually good summary from Wikipedia:</p>
<p>In November 2005, two Federal Reserve Bank of Boston economists published a working paper (Foote and Goetz 2005 [3]) which argued that the results in Donohue and Levitt’s abortion and crime paper were due to statistical errors by the authors - in particular the omission of certain statistical controls that Donohue and Levitt had claimed to have used and using the total number of arrests and not the arrest rate in explaining changes in the crime rate. The Economist remarked on the news of the errors that “for someone of Mr Levitt’s iconoclasm and ingenuity, technical ineptitude is a much graver charge than moral turpitude. To be politically incorrect is one thing; to be simply incorrect quite another.”[4] However, subsequent correction for the errors pointed out by the critique turned out to not reverse or eliminate the link originally expounded by Levitt and Donahue, thus removing speculation that the original model was purposefully distortionary. Theodore Joyce had previously criticised the results in 2003 finding that no such link existed (“Did Legalized Abortion Lower Crime?” Journal of Human Resources, 2003, 38(1), pp. 1 -37.); however, Joyce’s criticism has been refuted by the authors and others as profoundly flawed, a result of omitted variables bias. Responses to both criticisms can be found here (Foote/Goetz reply) and here (Joyce reply). See Legalized abortion and crime effect.</p>