Federal Reserve Chairman Ben Bernanke has set an inflation target that may turn into a bull’s eye on the Republican party. Bernanke wants to cut the current "core" inflation rate of 2.9 percent to 2 percent. Over the next two years that would increase the unemployment rate from 4.8 percent to 5.5 percent, throwing two to three million more Americans out of work. Last week, Bernanke "paused" the Fed’s two-year policy of raising interest rates, but to contain inflation, rates will have to go up. Unemployment will go up with them. If Bernanke doesn’t raise rates enough to increase unemployment, inflation, driven by surging oil prices and rising wages, will escalate. Robert J. Gordon, a professor of economics at Northwestern who studies the trade-off between unemployment and inflation, contends that over the next two years both will increase simultaneously. "I think the Fed is facing an absolutely classic case of stagflation," Gordon told Edmund L. Andrews of The New York Times, "a situation in which they cannot win."

Gordon’s "they" might just as easily apply to the incumbent party. Right now security issues—"cut and run" vs. "stay what course?"—dominate the political foreground. These issues might be salient in November, and possibly in 2008, though whether to the credit or debit of the GOP remains to be seen. But American politics knows no more certain a predicative metric than that increasing unemployment or rising prices in an election year defeat the incumbent party. The phenomenon is called "economic retrospective voting." In the long sweep of political history, it appears that, more perhaps than any other factor, the answer to Ronald Reagan’s question in his first debate with Jimmy Carter in 1980—"Are you better off now than you were four years ago?"—decides elections.

In a path-breaking 1971 study, "Short-Term Fluctuations in U.S. Voting Behavior, 1896-1964," Gerald H. Kramer, a Yale political scientist, correlated economic conditions with the fortunes of incumbent parties. The key variable, Kramer found, was real personal income—that is, income adjusted for inflation. In "off-years," when there were no presidential elections, falling real incomes predicted defeat for the incumbent party in statewide races for the U.S. House. For example, a 10 percent decrease in per capita income translated into a loss of 40 House seats. In their 1998 paper, "Twenty-five Years After Kramer: An Assessment of Economic Retrospective Voting Based Upon Improved estimates of Income and Unemployment," D. Roderick Kiewiet and Michael Udell substantiated Kramer’s basic thesis. Using estimates of pre-World War II unemployment levels that had been unavailable to Kramer, they found that unemployment, not real personal income, is "the single variable that most powerfully affects congressional elections."