Political Science

Ellis Finds Performance More Important than Partisanship in Assessing Gubernatorial Approval

Posted Jun 26, 2017

                 What impact does presidential approval have on the approval of governors of the same party? As the United States becomes more polarized, we see less and less split-ticket voting where a voter will vote for different party nominees for different political offices. Therefore, we would assume that the president, as the symbolic head of one’s party, and his approval would influence how voters view other members of one’s party. The result is that voters would hold the party accountable at election time but not hold governor’s accountable for their individual performance as governor. However, one of the bases of democratic accountability is holding individual office holders responsible for their personal performance.

                In his Research Methods paper this Spring, sophomore Charlie Ellis, with an assist from Dr. Evans, studied the impact of presidential approval on gubernatorial approval when the president and governor identify with the same party. Using multivariate regression, Ellis finds that both presidential approval and unemployment influence gubernatorial approval which supports both the idea that voters hold the president’s party and the governor responsible for their actions. In this blog post, we will discuss the various explanations for gubernatorial approval, how Ellis measures his various hypotheses, his results, and their implication for democracy.

Explanations for Gubernatorial Approval

                In the political science literature, there tend to be three major explanations for gubernatorial approval. First, the personal school tends to focus on the characteristics and attributes of individual governors and what makes them unique, such as personality and political competence, to help explain their approval (Barksdale and Green 1970; Barth and Ferguson 2002; Woodard 2013). The primary problem with these studies is the difficulty in measuring and comparing such subjective variables such as personality and competence. Under this category, we believe scandal is one variable that would be tied to an individual governor and would lower their approval. Voters hold officials, especially high ranking officials like governor, to higher standards due to their greater power. Thus, a governor involved in a scandal should have lower approval.

                Second, the economic school suggests that voters determine how well the governor is doing based on the performance of the state economy (see Hansen 1999; Orth 2001; King and Cohen 2005). If the state economy is doing well, voters believe the governor is doing a good job since they view him as the one responsible for the state economy and thus the governor has higher approval. If the state economy is not doing well, the voters blame the governor which results in lower approval.

                Third, the environmental school examines variables at the systems level that explains gubernatorial approval. The environmental school would examine factors such as presidential approval which would reflect voter assessment of the political system and would potentially lead state voters to vote against all incumbents when the president, and thus the political system, is not performing well (Crew 2002; Atkeson and Partin 1995). Conversely, a popular president would carryover and benefit all office holders. Ellis extends this line of research by identifying whether the president and governor of the same party benefit/lose based on the president’s approval or if presidential approval affects all office holders. In a more partisan time period, it would make sense that voters may hold a party accountable more than the system as a whole if the president is unpopular.

Data and Methods

                To these his hypotheses, Ellis uses linear regression to determine the impact on gubernatorial approval based on presidential approval when both the president and governor are of the same party, presidential approval overall, state and national unemployment rates, and scandal. Data on gubernatorial approval, the dependent variable, comes from Beyle’s (2002) U.S. Officials Job Approval Ratings Collection (JAR). Data on presidential approval comes from the Gallup Presidential Job Approval Center, and averaged for each year from President Reagan in 1980 to President Obama in 2010.  Unemployment percentages for states were gathered from the Iowa State University Community Indicators Program, which put together a comprehensive spreadsheet representing the percentage of unemployed public from each state for the years 1980-2015 (Annual Unemployment Rates by State). Data on national unemployment is average of monthly unemployment rate for the year and this data is available from the Department of Labor’s Bureau of Labor Statistics. Scandal is a dummy variable measured whether the governor was involved in a scandal at the time of the poll. Scandal is measured as an investigation by an official government source, either ethics commission or prosecutor. This information on scandal comes from encyclopedia entries on governors for each state. He uses data from ten states from 1980-2009.

Results

                The results in Table 1 show that governors are largely held accountable for their own performance as scandal and state unemployment explain significant portions of gubernatorial accountability. However, partisanship does matter as governors of the same party of the president see their approval and disapproval rise and drop slightly based on presidential approval. This paper also finds that there is mixed support for systemic effects of national economic performance and presidential approval overall.

Table 1

Factors that Influence Gubernatorial Approval

** statistically significant at the .01 level. * significant at the .05 level.

                The clear result is the important of state unemployment and scandal on gubernatorial approval. When the state unemployment rate increases, the governor sees as corresponding 3.38% decrease in approval and a 3.85% increase in disapproval (results not shown) for each one percent rise in the unemployment rate. In this regard, a three percent increase in unemployment in one year could lead to a 10% drop in approval. Since most governors face competitive challenges due to the high profile nature and prestige of the job, a 10% decrease is basically the difference between election and defeat. The fact that national unemployment does not influence gubernatorial approval suggests that voters hold the governor accountable for state economic performance more than national economic performance, even though national and state economic performance are correlated.

                Second, we see that scandal can hurt a governor’s approval. A governor subject to ethical or criminal investigation sees one’s approval drop 3.26% and one’s disapproval rise 4.87%. While these results may not seem large, we need to remember that governors have the power of incumbency at their disposal that allows them to counteract any media narrative and also allows the governor to do other things to increase one’s approval such as propose bills, sign bills, dispense state aid. Considering the tools at the governor’s discretion, the drop in gubernatorial approval is important. Moreover, a scandal is also likely to increase the change that the out party will successfully challenge the governor in the legislative and next election.

                Third, we find clear, but slight indicators of partisanship having an impact. Beginning with the impact of presidential approval when the governor and president are of the same party, we see a slight positive relationship. Thus, as a same party president as the governor sees his approval increase, the governor will see a slight increase of his approval of .03 for each one percent increase in presidential approval. Naturally, it is not surprising that the governor’s disapproval numbers increase as the same party’s president’s approval increases (results not shown). However, this change is slight indicating that the impact is not great.

                The impact of party is further seen in the results for presidential approval overall. Here, we see that as presidential approval increases, gubernatorial approval decreases. Just as the same party presidential approval variable captures partisanship, the presidential approval captures partisanship in another way. This measure basically shows that when the president of the other party sees his presidential approval increase, there is a corresponding slight decrease in gubernatorial approval. These two results suggest that national party fortunes have a slight impact on gubernatorial approval.

Conclusion

                The results of this research provide evidence that citizens hold governors accountable for their actions and not those of their national party. While the president’s approval does help and hurt the governor, the impact of these variables are slight. In comparison, the impact of state unemployment and scandal are much more significant factors in influencing gubernatorial approval.

Based on these results, it is no surprise that governors spend a lot of their time on economic development. If governors can recruit industry to their state, they are likely to receive the rewards of these efforts, more so than federal officials. While tax breaks to business may not make sense in polling, the ability to use these breaks to bring in business make it worth it to the governor to use these tools as he will benefit from it.

Finally, the results suggests that we can have greater confidence in voters overall. As V. O. Key (1966) said many years ago, “the voters are not fools.” Voters may not individually show a strong knowledge of issues but they are able to examine larger trends to determine whether the office holder is doing a good job. If the economy is performing well and the governor avoids scandal, voters can use that limited information to conclude that the governor is performing well which would explain the governor’s approval. This ability to separate state and national economic performance means that governors, overall, are held accountable for their good and bad decisions.

References

 “Annual Unemployment Rates by State.” Annual Unemployment Rates by State / Iowa Community Indicators Program. Web.

Atkeson, Lonna Rae, and Randall W. Partin. “Economic and Referendum Voting: A Comparison of Gubernatorial and Senatorial Elections.” The American Political Science Review, vol. 89, no. 1, 1995, pp. 99–107. JSTOR

Barksdale, E.C., Green, Williams. Essays on Recent Southern Politics. Austin, University of Texas Press, 1970.

Barth, Jay, and Margaret R. Ferguson. “American Governors and their Constituents: The Relationship between Gubernatorial Personality and Public Approval.” State Politics & Policy Quarterly, vol. 2, no. 3, 2002, pp. 268-282. JSTOR

Beyle, Thad, et al. “Gubernatorial, Senatorial, and State-Level Presidential Job Approval: The U.S. Officials Job Approval Ratings (JAR) Collection.” State Politics & Policy Quarterly, vol. 2, no. 3, 2002, pp. 215–229. JSTOR.

Crew, Robert E., et al. “Political Events in a Model of Gubernatorial Approval.” State Politics & Policy Quarterly, vol. 2, no. 3, 2002, pp. 283–297. JSTOR.

Gallup, Inc. “Presidential Job Approval Center.” Gallup.com. 23 Jan. 2017. Web. 07 May 2017.

Hansen, Susan B. “‘Life Is Not Fair’: Governors' Job Performance Ratings and State Economies.” Political Research Quarterly, vol. 52, no. 1, 1999, pp. 167–188. JSTOR.

Key, V. O. 1966. The Responsible Electorate. Cambridge, MA: Harvard University Press.

King, James D., and Jeffrey E. Cohen. “What Determines a Governor's Popularity?” State Politics & Policy Quarterly, vol. 5, no. 3, 2005, pp. 225–247.  JSTOR.

“Labor Force Statistics from the Current Population Survey.” Department of Labor’s Bureau of Labor Statistics. https://data.bls.gov/timeseries/LNS14000000

Orth, Deborah A. “Accountability in a Federal System: The Governor, the President, and Economic Expectations.” State Politics & Policy Quarterly, vol. 1, no. 4, 2001, pp. 412–432. JSTOR.

Simon, Dennis M., et al. “The President, Referendum Voting, and Subnational Elections in the United States.” The American Political Science Review, vol. 85, no. 4, 1991, pp. 1177–1192. JSTOR.

Woodard, J. David. The New Southern Politics. Lynne Rienner Publishers, 2013.