Here's a map of economic sentiment by political affiliation, as documented by the University of Michigan Consumer Survey:
figure 1: Economic sentiment index for Michigan College Democrats/leaning Democratic (blue) and Republicans/leaning Republican (red). NBER-defined recession peak-to-trough dates appear gray. source: University of Michigan Consumer Surveyand the National Bureau of Economic Research.
While sentiment changes as control of the presidency changes, it would be interesting to examine the scale of this change and how the importance respondents place on various factors changes.
If one gets the impression that Republican/Republican-leaning respondents do cast a large negative light on Biden's presidency, then that's correct. This can be seen through regressions on unemployment, inflation, the Biden dummy, and economic confidence. SF Fed News Sentiment Indexduring the period 2017M02-2024M02 (we have a continuous sentiment time series broken down by party affiliation).
Table 1. Regression coefficients with HAC robust standard errors. Bold indicates significance at 5% msl. Sentiment, unemployment, inflation are all expressed in decimals.
Column (1) reports the results of the Michigan survey series on these variables, while column (2) reports the same results, constraining the coefficients on unemployment and inflation to be the same (i.e., regressing the “misery index” )).
Headline index inflation and unemployment are in line with expectations, while the coefficient on the Biden dummy is significantly negative. The news sentiment index also showed the expected positive signs. The F-test does not reject the null hypothesis that the unemployment and inflation coefficients are equal in size. Then “pain” has a statistically significant coefficient.
Interesting findings related to categorical responses. Democratic/Democrat-leaning respondents and independents give unemployment less weight (in absolute terms) than inflation. For Dem/lean-Dem, the F-test on whether the coefficients are the same is rejected at 5% msl. Therefore, the weight of unemployment is relatively low for this group of people. Republicans/leaning Republicans, on the other hand, place roughly equal emphasis on unemployment and inflation. Relative to Democrats/Thin Democrats, the effect of inflation is larger (2.5 vs. 2.0).
In comparison, Democrats/Lean Democrats are more likely to be affected by news sentiment than House Representatives/Lean Representatives (0.30 vs. 0.22). This shows that no matter what bias exists in news reporting, Harris and Sojourner(2024)Representative/Lean Representative respondents are less sensitive to economic information than Democratic/Lean Democratic respondents.
The most interesting finding concerns the coefficient of the Biden variable. For Dem/lean-Dem, the coefficient is +0.18, while for Rep/lean-Rep, the coefficient is -0.58, which is more than three times that (in absolute value). Another way to look at the relative impact is to look at the standard deviation normalized coefficient (sometimes called the beta coefficient). For Rep/lean-Rep, the betas for Biden and inflation (the next most influential variable) are 0.83 and 0.18, respectively. For Democrats/Democrat leaners, the ratio of Biden to news sentiment is 0.87 to 0.57.
While early data shows which party controls the presidency has a greater relative impact on House/Level House members (we don’t have that much in the 2006 to 2016 data), to the extent we can measure this The impact, overall, appears to be greater for Biden than for Democratic presidents. (Not sure if the F-test can confirm this, but the coefficient has been large (in absolute value) in the past 7 years).
This difference supports the findings of this study postal.Results not including news sentiment variable, see this postal.
