February 12, 2009


More from old Fortune "Keeping Up" columns by the late Daniel Seligman. Here's one where using 1980s hardware and software, and inputting data by hand, he did something that hardly any professional journalists could imagine doing today.

December 4, 1989

Can it really be? The liberals now draw their support more from the rich than the poor? The Schumpeter effect is already in place? Here we allude to a certain famous proposition originally (1942) put forward in Joseph A. Schumpeter's monumental work Capitalism, Socialism, and Democracy: that market economies contain the seeds of their own destruction. In the course of creating material abundance, they also create masses of wealthy, educated, critically minded people who judge mere wealth production to be beneath them and gravitate naturally to antibusiness attitudes. The capitalist class itself assimilates these attitudes. "It absorbs the slogans of current radicalism," Schumpeter wrote, "and seems quite willing to undergo a process of conversion to a creed hostile to its very existence."

... That question can best be answered by a spreadsheet. So, nervily courting ophthalmia, we keyed in (a) 1980 census data on median family income for each of our country's 435 congressional districts and (b) the 1988 ADA rating for each district's Congressman. On command, Lotus 1-2-3 then proceeded to quantify the relationships between these two sets of data for the entire country and for certain regions.

Blearily eyeballing the data, we conclude that something has indeed changed -- and in the direction forecast by Schumpeter and Ladd. In some regions, there now is a discernible positive correlation between income and liberalism. In New England, for example, the correlation is 0.18. (Positive correlations range from 0 to 1.0.) For Massachusetts alone, it is 0.29.

In some other areas, the old relationships are lingering in place. In the South, the correlation between income and liberalism is still distinctly negative. For the states of the old Confederacy, the coefficient is -0.32, a figure that tells us Southern folks continue to vote their pocketbooks. The same may be said of California, whose 45 districts collectively weighed in at -0.29.

It turns out that these divergent patterns neatly offset each other. Taking the country as a whole, income predicts virtually nothing about the strength of liberal attitudes; the nationwide correlation coefficient is positive but, at 0.05, barely visible. Especially to our own aching orbs.

My published articles are archived at iSteve.com -- Steve Sailer


Half Sigma said...

I bet this has changed a lot since 1989, and using the same statistical technique today there would be quite a strong correlation between wealth and liberalism.

Anonymous said...

I imagine the fact that Seligman was a (half-) Jew scares a lot of isteve readers off items like this one. Doesn't fit their template.

Audacious Epigone said...


Andrew Gelman's Red State Blue State book takes your assertion on. In the Northeast, that is the case, but it's ambiguous in the Midwest and not at all the case in the South. This is apparent enough from the fact that ~85% of whites voted for McCain in Alabama and Mississippi. Race accounts for half of the wealth gap in the South.