Caplan’s forthcoming The Case Against Education is a great read. It validates the signaling model applied to traditional education. It makes the case for lower federal spending on formal education as well as reduced private consumption of traditional formal higher education, such as a 4 year degree.
I love the book, although I am bugged throughout the book by his strategic conservatism. Certainly this is an intelligent approach on his part. The dominant view is pro-traditional education. By overestimating education’s return he is minimizing the capacity of this book to conflict with the status quo. Should he prove his case on such conservative grounds then using more accurate measures his case would be all the more effective. Indeed, I think he does prove his case on such conservative grounds.
While Bryan and I share a desire to alter the standard opinion in the literature, I am still bugged. I am itching for an immediate sequel which aims mainly for truthful measures of return to traditional education as well as a richer sampling of the alternatives. Two of Bryan’s children, for example, are highly successful homeschoolers. Homeschooling is not discussed in the book, but what little data I’ve seen on the subject speaks to incredibly potential for return. HSLDA produced some data. I also have anecdotal experience about math competitions where the homeschoolers took it all. I was also homeschooled for about half of a year and loved it.
On to a specific objection: Caplan creates his own list price metric and uses this as the cost to baseline the return for education. I see a ready made alternative metric which I prefer.
In Chapter 6 on page 173, footnote 27 identifies the source of Caplan’s higher education list price figure. This footnote in turn refers to Figure 5.10. The footnote is clear: List price is tuition and fees plus books and supplies. Here’s the problem: That number doesn’t measure total social cost.
Consider that federal education spending falls into two groups:
- Student subsidy
- Research grants
Category 1 may be ignored while calculating the social return to education as it can be called a transfer or a wash: There is an additional social cost to taxpayers and a gain from students. Or so goes my reading of one point made in the book on page 170. I’m not fully convinced. 3/2 does not equal 4/3, so fixed-value transfers do seem to impact the social return figure which is return/investment.
Category 2 certainly cannot be safely ignored as it doesn’t go to the student. Category 2 is non-student-augmenting education spending.
Universities share capital and labor costs between research and teaching. They thereby gain economies of scope and scale which reduce their cost of providing education services. A single professor with his computer, software, and other capital, can both teach and produce research. So the fact that universities get research grants results in their possession of a lower cost of providing education services. This means that we should include the costs of research grants as education spending and not look at student subsidies alone.
So there are two ways to measure the social cost of education and they aren’t lining up. The obvious way to me:
Total private expenditure + total public expenditure = total cost
The non-obvious way which doesn’t make sense to me, adopted by Caplan:
List price less institutional grants
I suppose to get total cost you would multiply this by the number of full time equivalent (FTE) students. Yet, I don’t think that metric would be an accurate measure of social cost. I think it would be a subset of private costs which exempts voluntary giving.
An interesting tangent – what spending counts as education spending? Appropriations to the Department of Education doesn’t cut it. The NSF funds approximately 20 percent of all federally supported basic research conducted at America’s colleges and universities. At least 26 agencies give research grants and they don’t fall under DOE. Many provide scholarships for students as well. These are generally appropriated separately from the so-called HHS bill. That bill schema was H.R.5926 in the House during 2017.
Back to Caplan. His figure 5.10 cites the College Board under S. Baum and Ma 2011. This is the article Trends in College Pricing, 2011 by Baum, Sandy and Jennifer Ma. In that piece, Figure 12A shows institutional revenue per student. Why don’t we use that number instead of list price less institutional grants? This is a ready-made metric which more or less addresses the research grant issue.
A quick point of concern about the Baum and Ma institutional revenue per student figure. Their figure includes grants as well as state and federal appropriations. Great right? Well maybe. My concern is that many university faculty members have to win grants to earn their salary. I am not sure whether grants awarded to individual faculty members are included in the institutional revenue measure, or whether they might accrue to those faculty members individually under a separate account. If the latter is the case then we have found some social expenditure which is not included. We would need to accordingly increase the 13646.67 figure I later report.
Let’s suppose the just raised concern is not the case and the College Board does include all grants to faculty, not just appropriations directly awarded to the institute. Caplan’s List Price metric from Figure 5.10 appears to derive from Baum and Ma’s Figure 1. To be noted, Baum and Ma’s Figure 1 is meant to apply to undergraduate students. Caplan’s public list price of 9412 = 8244 + 1168, which is the amount of tuition, fees, books, and supplies from Baum and Ma. These are in 2011 dollars if I’m not mistaken.
If I’m not mistaken, Caplan’s list price approach underestimates the true social expenditure per FTE student. Again, strategically this makes Caplan’s case stronger, I’m just a bit miffed because I want the true return and it seems overestimated. An alternative measure is to average the public net tuition revenue and subsidy. Looking at Baum and Ma’s Figure 13, the average of the 4 public categories for 2008-2009 is 12567.50, in 2009 dollars. If you exclude 2-year universities, since they can’t award a bachelor’s degree and that is what Caplan is assessing in Chapter 5, the FTE revenue rises to 13646.67. This is in 2009 dollars, so increase it appropriately to compare against Caplan’s list price of 9412.
This is about a 50% increase to cost, so the return to education estimate given by Caplan would be reduced by a third using this approach.