Friday, January 04, 2008

Michael Barone Isn't Making Sense

Brendan Nyhan offers a good takedown of Michael Barone's notion of a "16 year itch" in American politics.

Michael Barone's "16 year itch"

Michael Barone:

The metrically minded will see a common thread. Every 16 years--in 1976, 1992 and now in 2008--American voters have seemed less interested in experience and credentials and more interested in a new face unconnected to the current political establishment. What can explain this 16-year itch?


Chance. As Atrios would say, this has been another edition of simple answers to simple questions.

Seriously, though, Barone should know better -- political scientists spent years debating a hypothesized 36-year cycle of realigning elections that broke down under scrutiny. The same thing will inevitably happen to this even shakier hypothesis. After considering possible explanations of the supposed 16-year pattern, Barone ends up arguing that "over a period of 16 years, there is enough turnover in the electorate to stimulate an itch that produces a willingness to take a chance on something new." But why not 12 or 20 years? 16 is an entirely arbitrary number and, as Barone admits, the theory breaks down for 1944 and 1960. So in other words, the premise for the column is based on a grand total of three elections (1976, 1992, and 2008). It's absurd.

A better argument might consider the fact that the public tends to drift ideologically against the party in power, which creates openings for resurgent opposition parties, particularly when the incumbent is weak. In 1976, 1992, and 2008, the Democrats were looking for new faces to challenge the GOP after 8, 12, and 8 years of Republican rule, respectively.


I would add that the theory also overlooks 2000, when voters went with George W. Bush, a relatively inexperienced state governor, instead of Al Gore, who served 8 years as Veep, 8 years as a U.S Senator, and 8 years as a U.S. representative.

1 comment:

Mike said...

I would add that the theory also overlooks 2000, when voters went with George W. Bush, a relatively inexperienced state governor, instead of Al Gore, who served 8 years as Veep, 8 years as a U.S Senator, and 8 years as a U.S. representative.

If the theory is predicting the popular vote rather than the electoral college, then it can accommodate 2000. And I suppose that the theory isn't meant to differentiate voters in small states with more electoral votes per capita, or in swing states, from other voters. (And even if it's tracking the electoral college rather than the popular vote, it isn't clear that Bush really won even the electoral college.)