Sentiment and misery (and perhaps partisanship)

The College of Michigan February Client Confidence Index (preliminary) was launched at the moment. Right here is the image (reversed sequence so down is the advance) and the “distress index”, the sum of unemployment and annual CPI inflation.

Determine 1: “Distress Index” sum of unemployment and annual CPI inflation, % (blue, left scale) and the College of Michigan’s inverted client sentiment index (tan, proper scale). The NBER has outlined peak-to-trough recession dates as shaded. Supply: BLS, College of Michigan, each through FRED, NBER and creator’s calculations.

The hyperlink between the Distress Index and the Sentiment Index (FRED UMCSENT sequence) is mentioned on this article. What pursuits me is the obvious rupture of latest months. Take into account:

UMCSENT = 51.9 – 4.13MISERY – 14.8p^gasoline

Adj-R² = 0.64, SER = 7.87, DW = 0.30, Nobs = 303, pattern 1992M01-2023.01. Daring denotes significance at 10% msl, utilizing HAC sturdy normal errors.

The beta coefficient (standardized) on MISERY is 0.60 and on the log of the actual value of gasoline is 0.344, so MISERY is crucial affect on UMCSENT.

The precise, adjusted and residual values ​​are proven beneath. In December 2022, the sentiment index is overestimated by nearly 20 factors.

Determine 2: Precise (crimson) adjusted (inexperienced) and residual (blue) UMCSENT.

I’ve no specific rationalization as to why the overstatement is happening, though it’s attention-grabbing to notice that there was rising partisan divergence in perceptions of financial situations, as evidenced by, amongst different issues, Carola Binder (2023).

Supply: Binder (2023).

If the Republican flip is especially essential within the Biden period, it might sever the historic relationship between MISERY and UMCSENT.

What number of biases are there? Courtin (2022) offers barely older estimates, additionally based mostly on the Michigan survey.

curtine writes:

…the partisan hole was bigger beneath Trump and Biden than all the index’s time sequence shifts from optimism to pessimism; The Bush and Obama spreads have been about half that dimension.