A plain-English guide to the sentiment-vs-spending decoupling.

Friday's analysis piece is the live example. Michigan Consumer Sentiment Final printed at 48.1 (source: University of Michigan Surveys of Consumers), the lowest in four months, with expectations dropping 10.1 percent to 46.3 and 12-month inflation expectations climbing to 4.6 percent. But retail sales for the same month, released Sep 17, printed positively for the third consecutive month, and Friday's Durable Goods was flat. Sentiment collapsed while spending held. This piece is the framework for reading that decoupling, which has been a durable feature of the post-2022 US macro landscape.

What the decoupling is

Historically, consumer sentiment and consumer spending moved together. When sentiment fell, spending softened three to six months later; when sentiment rose, spending firmed. The correlation was tight enough that the University of Michigan's Consumer Sentiment Index was one of the specific leading indicators the Conference Board included in its Leading Economic Indicators (LEI) composite.

Post-2022, that relationship broke. Sentiment has swung between "recession-adjacent" levels and modestly-firm levels while spending has held roughly flat-to-firm on nominal terms. The specific 2022-2024 window saw the Michigan headline hit multi-decade lows (June 2022) with retail sales continuing to expand; the 2025 window saw sentiment recover partially while spending decelerated modestly. The correlation is no longer statistically significant.

The three explanations that fit the data

1. Balance-sheet resilience. The 2020-2021 fiscal transfers plus the 2022-2023 wage-catch-up moved household balance sheets to their strongest position in decades. When balance sheets are strong, sentiment moves reflect anxiety about the future rather than constraint on the present; households can be worried about tomorrow while still spending today.
2. Sentiment-as-partisan-signal. Post-2016, both Michigan and Conference Board readings show pronounced partisan effects: sentiment among the party out of power falls sharply after an election, and sentiment among the party in power rises sharply. That noise dilutes the aggregate signal.
3. Inflation-frame vs unemployment-frame. Sentiment surveys ask about "conditions" broadly; when inflation is the dominant complaint, sentiment can drop sharply even when labour markets and spending are firm. That has been the specific pattern of 2022-2026 sentiment moves.

Reading Michigan versus Conference Board

Michigan and the Conference Board Consumer Confidence Index measure related but distinct constructs. Michigan is more sensitive to inflation (asks specifically about inflation expectations and household finances); Conference Board is more sensitive to labour-market conditions (asks specifically about jobs "plentiful" versus "hard to get"). When the two diverge, the divergence itself is informative:

• Michigan weak, Conference Board firm: inflation-driven sentiment softness. The current 2026 pattern; a specific footprint of the "inflation is the main household complaint" regime.
• Michigan firm, Conference Board weak: labour-market-driven sentiment softness. The 2001, 2008-09, and 2020 patterns; typical recession footprint.
• Both weak: broad-based softening. The 1974-75 and 1980-82 patterns; a specific footprint that historically has led recessions by four to eight months.
• Both firm: expansion; the specific footprint of 2015-2019.

The four specific data points that carry the macro read

1. The Expectations sub-index level. Michigan's Expectations Index (five questions about the future) leads the Current Conditions Index (three questions about the present) by roughly three to six months. When Expectations drops sharply (as Friday's 10.1 percent single-month drop), the signal is for the following quarter's Current Conditions rather than the current one.
2. The inflation-expectations series. Michigan asks two versions: one-year-ahead and 5-to-10-year-ahead. The one-year is more volatile; the 5-to-10 is the specific "anchored" measure the Fed tracks. Friday's one-year jump to 4.6 percent is a session-mover; the 5-to-10 (typically inside 2.5 to 3.5 percent) is the more consequential trend measure.
3. The partisan split. Michigan publishes commentary on party-specific readings. A shift in the partisan split (both parties moving in the same direction, as in Friday's release) is a stronger macro signal than a shift in only one party's reading.
4. The high-income vs low-income split. Michigan tabulates by income tertile. When the top tertile joins the bottom tertile in a sentiment drop, the specific signal is that discretionary-spending resilience is at risk. That has been sporadically the case in 2025-2026; watch for a fuller convergence.

How to read a sentiment print inside a hawkish-Fed regime

Under a Warsh-led hawkish regime, a Michigan expectations collapse is a "growth-concern-rising" signal that can start to reverse the dollar's rate-differential-driven strength. That is the specific reading of Friday's tape: DXY fell 32 pips even as 10Y climbed 2.2bp. The rate-differential channel favours dollar strength; the growth-concern channel favours dollar softness or a rotation toward growth-off assets (gold, JPY on real-yield basis, CHF on flight-to-quality basis).

The framework rule is: a Michigan Expectations drop of more than 5 percent in a single month is a growth-concern signal worth logging; more than 10 percent (as on Friday) is a specific signal to watch for cross-asset confirmation in the following two weeks. If Retail Sales for the month prints negatively when it lands in October, the decoupling has broken and the framework shifts.

The three specific errors to avoid when reading sentiment

• Reading sentiment as a coincident spending indicator. Post-2022 the correlation is not statistically significant. Watch spending directly.
• Overweighting the one-year inflation expectation. Volatile and noisy; the 5-to-10-year is the trend measure the FOMC watches.
• Ignoring the partisan composition. A sentiment print that moves only because of a partisan shift is not the same signal as one that moves across parties.

Related reading

• Friday's analysis piece: the live example this framework describes.
• Inflation expectations surveys: the specific mechanics of Michigan vs New York Fed SCE vs Conference Board.
• Consumer confidence inflection signals: how to spot a regime change in the sentiment series.
• The dollar smile: the regime framework a sentiment collapse can shift.