Crew and Weiher: Gubernatorial popularity in three states

Crew and Weiher. 1996. Gubernatorial popularity in three states: A preliminary model. Social Science Journal 33: 39-54.

In Brief

Though there have been numerous studies of presidential popularity, there have been few studies of gubernatorial popularity. Those studies that have been performed have focused on gubernatorial elections, as there is a dearth of good data on gubernatorial approval. The authors use time series data in California, and less complete data for Minnesota and Iowa, to take a deeper look at gubernatorial approval. Their findings largely confirm what earlier studies have suggested: Gubernatorial approval is driven by national factors (presidential approval and national economic conditions) more than anything else.

Variables

For the California study, the authors use quarterly data spanning just over a decade. The data is not perfect, and they fill in some missing values with averages. Since their analyses of Minnesota and Iowa use similar variables, I will focus on the California case. Variables:

• Lagged popularity
• California (quarterly) unemployment
• US (quarterly) unemployment
• Presidential approval level
• Dummy: Whether the president and governor belong to the same party
• Interaction of this dummy with the two national-level variables
• "Honeymoon": The number of quarters since the governor's inauguration

Findings

Like previous studies, the authors find that state-level economic conditions have no effect. Voters punish the governor for national trends, and there is a distinct honeymoon effect.

Comments and Criticism

Peltzman (1987) observes that we should use change variables, not objective values. Specifically, why measure the unemployment rate rather than the 'change' in the unemployment rate?