Katz and Murphy’s (1992) influential model of the college wage premium has failed when applied to out-of-sample data, making ex-post adjustments necessary. The specification searching began with Autor, Katz, and Kearney (2008) and was continued by Goldin and Katz (2008) in The Race between Education and Technology. The result is an econometric analysis of the college wage premium that depends on the undertheorized time breaks to make the model fit the data. Unit root and cointegration tests indicate that the verification of the Katz-Murphy model may have relied on spurious regressions for its validation, as Richardson (2025) has recently argued. The model should probably not be used to support a policy agenda of increasing funding for education to reduce inequality.