Disclaimer. Don't rely on these old notes in lieu of reading the literature, but they can jog your memory. As a grad student long ago, my peers and I collaborated to write and exchange summaries of political science research. I posted them to a wiki-style website. "Wikisum" is now dead but archived here. I cannot vouch for these notes' accuracy, nor can I say who wrote them.
Kenney. 1983. The effect of state economic conditions on the vote for governor. Social Science Quarterly 64 (March): 154-162.
After examining a time series of gubernatorial election outcomes in 14 states, Kenney concludes that state economic variables (unemployment, inflation, and growth) do not significantly influence the incumbent party's share of the gubernatorial vote. In a handful of these caes unemployment or growth was significant, but generally there was no statistically significant effect.
Kenney should be applauded for being among the first to extend the economic voting debate to the state level. Nonetheless, his models are simultaneously under- and over-specified.
They are overspecified because he includes unemployment, inflation, and income simultaneously in a single regression, with only one other a variable (a dummy controlling for whether the governor belongs to the president's party). Kenney concedes that these variables are multicollinear, and he reports some tests of this problem. Still, he ought to have addressed this problem further.
They are underspecified because Kenney did not control for basic factors like whether an incumbent is running, the president's approval (his presidential dummy alone is inadequate), and so on.
Also, Kenney runs separate regressions for each of the 14 states. Thus, N for each state was between 8 and 15. With such a small number of cases and five variables being estimated, it's no wonder that none of the variables was significant. He ought to have pooled the data, perhaps including state dummies, so that he could have greater statistical leverage.
Research on similar subjects