Towards the end of my debate with Steve Pearlstein, he posed an intriguing question.  My paraphrase: Suppose half of higher education really is pure waste.  What’s the efficient government response?  Should government should cut its subsidy by 50%?  Or what?

Steve’s asking a complex question, so let’s start with a simple case: perfectly elastic (i.e. horizontal) supply of education.  In this scenario, subsidies have no effect on price.  To cut Q by 50%, then, you have to cut total spending by 50%, too.  Since total spending=private spending + subsidy, you simply have to ask two questions:

1. Was the subsidy initially more than 50% of total spending?  If so, the
efficient subsidy is (.5*initial total spending-private spending). 

Numerical example: Initial spending=$500B, optimal spending=$250B, and initial subsidy=$350B.  Upshot: Cut the subsidy to $100B.  That way, private spending + subsidy=$250B, the efficient level.

2. Was the subsidy less than or equal to 50% of total spending?  If so, the efficient subsidy is 0

Numerical example: Initial spending=$500B, optimal spending=$250B, and
initial subsidy=$200B.  Upshot: Cut the subsidy to $0B.  Private
spending ($300B) still exceeds the efficient level of $250B, but unless the government can give negative subsidies (i.e. taxes), there’s further better it can do.

Knowing Steve, he’ll protest the unrealism of my key simplification.  After all, the supply of education isn’t perfectly elastic in the real world, so cutting subsidies automatically cuts prices.  However, this plain fact just amplifies my austere conclusions. 

To see why, recall that total spending=price*quantity.  If we cut subsidies enough to halve quantity, and prices fall too, total spending falls by more than 50%.  And given the 50% waste premise, a greater than 50% fall in total spending is the efficient outcome!