The rapid increase in the cost of college in recent decades — and the associated explosion in student debt, which now totals nearly $1.3 trillion nationally — is all too familiar to many Americans. But few understand what has caused the tuition boom, particularly at the public institutions that enroll roughly two-thirds of all students at four-year colleges. Many commenters, particularly in the popular press, focus on ballooning administrative budgets and extravagant student amenities. Those elements have played a role, to be sure, but by far the single biggest driver of rising tuitions for public colleges has been declining state funding for higher education.
It is true that today’s students enjoy better amenities .... It is also true that today’s universities employ far more administrators and staff who don’t have any direct role in either research or instruction. ...
And it is also true that professors (at least those on the tenure track) are paid better than the people who held those same jobs years ago. ...
All of those trends add to the cost of college, but not by that much. At most, about a quarter of the increase in college tuition since 2000 can be attributed to rising faculty salaries, improved amenities and administrative bloat. By comparison, the decline in state support accounts for about three-quarters of the rising cost of college.
Hat tip: PAG
Here is the data for the states with the biggest funding decrease per student (all of the states are in the article):
The "three-quarters" number is close to the unconditional mean of the share of tuition hike explained by cuts (mean = 82) which is the logic that I think is being used by 538. The share of tuition hike column is the negative of state funding per student divided by the increase in tuition.
Another way to think about the data is with a linear regression (my SPSS output is below). A model of the increase in tuition (dT) as a function of the change in state funding per student (dS) holding starting tuition constant (T0 = current tuition - increase in tuition) is:
dT = 1.357 - 0.248*dS + 0.313*T0
This model tells us that the increase in tuition rises $248 for each $1000 dollar decrease in state funding per student and states with higher tuition before 2000 tend to raise tuition more than states with lower tuition. The R-squared is 0.47 which tells us that 47% of the variation in the tuition increase is due to variation in changes in state funding per student and starting tuition. A regression model with only the change in state funding per student included has an R-square of 0.304.
So, I think that decreases in state funding explain about 25% (using the regression coefficient) or 30% (using the R-squared) of the increase in tuition instead of "three-fourths."