In addition to Rick Garnett’s post on yet another of Geoff Stone’s ill-founded speculations, I’ll highlight these points made by Eugene Kontorovich on the Volokh Conspiracy:
The relatively high rate of liberal justices recruiting from Republican-appointed judges could simply mean the latter are simply perceived as better judges, mentors, etc. than their Democratic-appointed counterparts.… There has been significantly greater proportion of Republican-appointed judges on the courts of appeals over most of the past two decades. This is the period during which current judges have built their reputation, and Republicans have had a quantitative edge. While this difference has closed in recent years, I assume newly-appointed judges will on the whole do worse in fee[d]ing clerks to the Supreme Court.
Update: Orin Kerr also weighs in with excellent observations, including:
Stone fails to consider supply and demand. As most readers know, liberal students heavily outnumber conservative students at the law schools that tend to generate the pool of circuit clerks. At the same time, the federal judiciary is roughly evenly divided between GOP and Dem appointed circuit judges. This creates an imbalance in law clerk hiring. Clerks often want to work for bosses with similar views, and the judges often feel the same way. But judges also want the best clerks. With the supply of conservative clerks relatively low, GOP-appointed circuit judges regularly hire liberal clerks, while Dem-appointed circuit judges only rarely hire conservative clerks. I think this explains the numbers Stone identifies. Justices who want to choose liberal clerks can choose clerks who worked for Democratic appointed circuit judges or Republican appointed circuit judges. On the other hand, Justices who want to choose conservative clerks will choose almost entirely from former clerks of GOP appointed circuit judges.
See also Kerr’s discussion of Justice Ginsburg’s hiring. As he observes: “[I]f [Stone’s] theory fails when you look at applications, it suggests there may be a problem with the theory.”