What’s the matter with college graduates?

 Traditionally, more educated people were more likely to vote for conservative parties, even if you adjusted for income, but this relationship has weakened or reversed in many nations, including the United States.  The usual explanation for this change is more educated people are more conservative on economic issues, but more liberal on “social issues," so changes in voting pattern reflect changes in the relative importance of economic and social issues.   To quote David Leonhardt, “people vote based less on their income and more on their cultural attitudes.”  

 But there’s another possible explanation of the shift that hasn’t gotten much attention:  maybe the relationship between education and economic opinions is changing.  I considered this possibility in a paper that I presented last week at the World Congress of Sociology using data from a large number of nations included in the World Values Survey.  I considered four questions—whether incomes should be made more equal, whether government ownership of business should be increased, whether the government should take more responsibility for caring for people, and whether competition is bad because it brings out the worst in people—and computed the difference between the opinions of university graduates and others, adjusting for income (and a few other variables).  The figure shows the association of higher education with the sum of opinions on the four questions:

The horizontal red line indicates no difference:  below that line means more educated people are more conservative than less educated people, above means they are to the left of less educated people.  The relationship seems to change with per-capita GDP:  in poorer nations, university graduates are to the right of other people on these issues, but in more affluent nations, they are generally to the left.   (This relationship does not hold in the formerly socialist nations). 

Of course, the affluent nations differ from the poorer nations in a number of ways:  most obviously, they are generally in Western Europe or were settled by people from Western Europe.  So a cross-sectional relationship between per-capita GDP and the direction of opinion could reflect enduring cultural differences rather than affluence.  Fortunately, the WVS has included these questions since 1990, so it’s possible to see if the relationship has changed over the last 30 years.  Here are the changes, adjusting for differences in the nations included in different waves.  The zero point is the differences in 1990–positive values mean more leftward (or less rightward) differences relative to 1990.

For three of the four questions, the “education gap” has moved to the left--for competition, there hasn't been any clear change.  The move is pretty steady, although there seems to be some short-term variation (between the first and second waves, there was no change, or maybe even a small rightward change).  

So it seems that there’s something about “modernization” or economic development that leads to a change in the relationship between education and economic views.  What might that be?  One possibility is that affluence does increase the importance of "values" relative to immediate self-interest, but that economic opinions reflect mix of self-interest and values rather than immediate self-interest.  That is, educated people become more willing to help the poor even if that might increase their taxes, and less educated people become more willing to refuse a "handout" even if they would benefit.   Robert Lane proposed that freedom and equality were core principles of modern society, and although almost everyone accepts them in a general way, more educated people are more likely to apply them consistently--as a result, people in "the professional class" were more likely to be consistent egalitarians than people in the working class.  I think he may have been on to something.