Caplan on Mathematics PhDs as a "Dead-End Job"

In an otherwise interesting post on dead-end jobs, Bryan Caplan makes the following observation:
You might think that parents would universally discourage Dead Ends.  Sometimes, they do: Few parents want their kid to tell them, "I've decided to be an actor," "I've decided to be a rock star," "I've decided to be a poet," or "I've decided to pursue a Ph.D. in philosophy."  However, there is a long list of other Dead Ends that parents enthusiastically encourage: Classical music, sports, ballet, chess, and Ph.Ds in mathematics are leading examples.
If you just did a double-take as you read the words "and Ph.Ds in mathematics," then you're not alone. In the comments stream, many people took the time to object to Caplan's conclusion. I myself dusted off my keyboard, rolled up my sleeves, and prepared to do the same.

But then I thought about it. Caplan didn't say "people who study mathematics," he said PhDs. Holding fast to this distinction, Caplan's point becomes crystal clear. Anyone who has taken the time to apply for jobs recently has likely noted the world of opportunity that awaits anyone with a degree in mathematics or statistics. But there are a lot of opportunities for holders of Bachelor degrees, and marginally fewer for those who hold Master's degrees. Finally, there are even fewer - marginally speaking - for math PhDs.

In other words, the return on investment for a math PhD is far lower than the ROI for a mathematics B.S. or M.S. Makes sense.

Maybe the phrase "Dead End" is a bit too much to describe mathematics PhDs, but I think Caplan's point is essentially a solid one.