The US Consumer “Confidence Game”

By in
The US Consumer “Confidence Game”

When it comes to determining the health of the US consumer economy, there are really just 2 key issues. The first is the “ability to spend” (employment levels and wage growth). The second is “desire to spend” (consumer confidence).

We know US labor markets look good just now, and consumer confidence seems to be similarly strong. But a long run historical look at the University of Michigan Consumer Sentiment tells a more nuanced story.

Consider a graph of the Michigan data back to 1978, when this survey first went to monthly readings:

Here is what we see:

#1: The current cycle (2009 – present) started with the longest period of below-average consumer confidence since at least the 1970s.

  • The long run (1978 – 2019) average sentiment reading is 86.2.
  • It took almost 6 years to hit those levels (first reached in October 2014) after the November 2009 lows. But it only took 3 months post the 2000 recession to hit average confidence, 5 months post the 1990 recession, and 11 months after the 1982 recession.

#2: This cycle’s peak reading of 101.4 is well below the tops of the last 2 cycles (Jan 2004, 103.8, Jan 2000, 112.0). We have thought about that disparity a lot over the years and see many possible causes:

  • An aging population, more cautious than in their earlier lives, is one explanation.
  • Then there is the slower pace of wage growth to consider, something that could weigh on consumers’ economic confidence.
  • High levels of student loan debt could certainly hurt confidence among younger demographics.
  • No matter the cause, sentiment readings that remain well below prior peaks (and not far from cross-cycle averages) as an important issue to watch.

#3: The current reading of 95.7 (out on Friday) shows better consumer confidence than the last 2 months (September’s 93.2 in the chart above, October’s 95.5). Going into Holiday spending season that is good news, especially since this year’s Black Friday – Christmas period is shorter than usual.

Now, as for the relationship between “ability to spend” (labor market conditions) and “desire to spend” (confidence), we want to highlight that these two drivers of the US consumer economy do not always move in tandem. Just consider that despite record low levels of American unemployment just now, “peak sentiment” was almost 20 years ago. What’s going on here?

To answer that question we looked at the trailing 2-year correlations of changes in US unemployment relative to changes in consumer sentiment from 1978 to 2019. There is a chart below with that data, but here are our 2 takeaways:

  • High correlation levels (+0.6) between falling unemployment and rising consumer sentiment generally occur early in an economic cycle. This occurred in 1980, 1994, 2009 and 2011. The only exceptions were in 1987 and 1989.
  • Past those periods, falling unemployment has no noticeable impact on consumer sentiment. In fact, the average 2-year correlation between changes in employment and sentiment from 1978 – present is actually negative 0.2.

Here is the chart showing these historical correlations:

Summing up: at this late point in the current US economic cycle there is no causal link between falling unemployment (ability to spend) and rising consumer sentiment (desire to spend). That’s one important reason why the Federal Reserve is so focused on wage growth and stable financial asset prices, which drive both “ability/desire to spend”. Their recently completed mid-cycle course correction seems to have had its desired effect on the latter, at least for now. A US-China trade deal is the remaining piece of the puzzle to see consumer sentiment and dovetails neatly with what we described in “Markets” today.