The Contrarian's Favorite Indicator, and What It Actually Measures - Part 2

If you haven't read yesterday's blog post yet, please do so before continuing here.

Where sentiment has carried information, and the catch

Sentiment has historically carried signal at extremes, in the contrarian direction. J.P. Morgan Asset Management's dataset, on page 32 of the Guide to the Markets, identifies 10 sentiment peaks and 9 troughs since 1970 and annotates each with the subsequent 12-month S&P 500 price return. Following the peaks, an average of 4.8%. Following the troughs, an average of 24.1%. Michigan hit its second-lowest reading on record in late 2008 and the index rose roughly 36% over the following year. It set an all-time low in July 2022 and the market was up 16% twelve months later.

Here's the part that gets left off the chart. In that methodology, a trough is the lowest value before a series of higher highs, and a peak the highest value before a series of lower lows. Both definitions require knowing what came next. In real time a low print is just a low print. It becomes a trough retrospectively, once the recovery has already happened, which is precisely when the 24.1% has already been earned.

The current data makes the point. Michigan set a record low of 44.8 in May 2026 amid surging petrol prices. June recovered, July extended, August reversed to 51.0. Whether May was the trough gets answered in 2027, not now. And the sample is nine observations across 55 years, price-only and excluding dividends, which is worth sizing your confidence against.

The rules, and the one that's currently broken

A handful of specific measures get used alongside the peak-and-trough framing. The 80 threshold on the Conference Board Expectations Index, where a reading at or below 80 has historically been associated with a recession within the year. The spread between the expectations and present situation components. The labour market differential, jobs plentiful minus hard to get, which narrowed to +3.1 points in July. Michigan's year-ahead inflation expectations, which rose to 4.3% in August. And on the hard-data side, the retail sales control group, which strips out autos, petrol, building materials and food services and feeds directly into the GDP consumption calculation.

The 80 threshold has been in signal since February 2025. That's roughly 18 months of a recession indicator flashing without the recession arriving, which deserves scrutiny rather than repetition.

Two explanations are worth weighing. The first is partisan response distortion. Hsu noted that August sentiment fell across the political spectrum but that Republicans posted the steepest monthly decline, now 19% below readings prior to the Iran conflict and the lowest since the 2024 election. When a series swings with the news cycle and party identity, it's measuring something other than economic conditions. The second is composition. Michigan flagged notably large declines among older consumers, lower-income consumers and those without a college degree, while aggregate spending held up because upper-income cohorts kept spending. One index number averages those two experiences into a figure that describes neither.

There's a third observation that cuts against reading the surveys as a market signal at all. In the same July release, the Conference Board noted that consumers still expected higher stock prices a year out, notwithstanding recent equity volatility. Households separate how they feel about the economy from where they think markets are headed. The survey captures the former.

Come back tomorrow for the final part of this topic!

This publication is produced by Arbitrage Trade for informational and educational purposes only. It does not constitute investment advice, a recommendation to buy or sell any security, or an offer or solicitation of any kind. The analysis presented describes historical conditions and observable patterns and should not be interpreted as a forecast or prediction of future results. Past performance is not a reliable indicator of future results. Any securities, indices or instruments referenced are illustrative only. Readers should conduct their own research and consult a qualified financial professional before making any investment decision.