Consumer Confidence, Inflation And Unemployment
There’s a lot of discussion regarding the negative impact of inflation on consumer sentiment. That’s definitely there – but unemployment also has a negative impact. And there is a (at least short run) tradeoff between the two. Relevantly, what would unemployment be in the absence of the American Recovery Plan, the CARES Act, and expansionary monetary policies of the Fed?
Here’s a plot of the University of Michigan consumer sentiment index, month-on-month annualized inflation rate, and the unemployment rate (the latter two rates in decimal format).
Figure 1: Top panel, University of Michigan Consumer Sentiment; Middle Panel, month-on-month CPI inflation rate annualized (log terms, decimal format); Bottom Panel, unemployment rate (decimal format). NBER defined recession dates peak-to-trough, shaded gray. Source: Univ. of Michigan, BLS, via FRED, and NBER.
What’s the relationship using a simple OLS regression? Using monthly data, 1978-2021M10:
umcsent = 1.16 – 0.86 π – 4.36 u + v
Adj. R2 = 0.42, SER = 0.095, N = 526, DW = 0.25. Bold denotes significance at 5% msl, using HAC robust standard errors.
One can report standardized coefficients (i.e., scaled by standard deviations) to convey the relative importance of each variable in the movements of the left hand side variable.
umcsent’ = – 0.26 π’ – 0.60 u’ + v’
Or the elasticity at means:
umcsent” = 1.35 – 0.034 π” – 0.315 u” + v”
So, yes, with lower inflation ceteris paribus consumer sentiment would be higher. But if that lower inflation was associated with higher unemployment, then the net effect would be ambiguous (unless one takes a stand on the tradeoff during these pandemic times).
Using quantile regression:
umcsent = 1.18 – 0.89 π – 4.62 u + v
Adj. R2 = 0.30, SER = 0.095, N = 526, Bold denotes significance at 5% msl, using Huber Sandwich robust standard errors.
Disclosure: None.