Mechanism
The great vibecession was real but surveys missed the distribution
The standard story goes: unemployment stayed low, wages rose, inflation cooled, yet consumer confidence tanked. Broken measurement. People are irrationally gloomy. But I think we're misreading what the indices actually captured.
The Conference Board and UMich surveys ask about personal finances and six-month outlook. Those questions do something weird when you have high inequality, concentrated wage gains, and sectoral shock. A software engineer getting a 15% bump in real compensation feels *fine*. A service worker whose nominal wage stayed flat while housing doubled feels *terrible*. Both show up in aggregate data. The median person was actually squeezed—rising food and rent costs, childcare pricing that never deflated, credit card debt climbing back up. The aggregate statistics were correct. The sentiment was correct. But sentiment indices average across those two experiences and call it "consumer confidence," which flattens distributional reality into a single scalar.
What's harder to defend is that *nothing changed*. Labor force participation among prime-age women dipped in 2023. Delinquencies on auto loans went up faster than unemployment would predict. Revolving credit growth accelerated, which is usually a sign people are extending themselves. Nominal wage growth outpaced real wage growth for most workers once you account for actual consumption baskets (CPI is a Laspeyres index with old weights). The vibecession wasn't invented. The surveys just couldn't isolate whose vibecession it was, and that ambiguity got misread as inaccuracy rather than informativeness about dispersion.
1 comment
Log in to comment.
When you say the Laspeyres weighting problem meant nominal wages outpaced real for most workers—are you talking about something like the BLS experimental CPI-U-RS, or more just the basket mismatch between what CPI actually tracks versus what people on the lower end of that wage distribution were actually spending on? Because I ask because I spent five years watching our ridership collapse during the 2008 contraction in ways that had nothing to do with unemployment rates looking okay on paper, and everything to do with people dropping from "takes the bus sometimes" to "walks or stops making the trip." But we could never quite pin down whether our models were missing a distributional shock or just measuring the wrong people in the first place.
The labor force participation dip for prime-age women in 2023 is what actually makes me skeptical here though. That's not ambiguous sentiment data—that's people making a binary choice. If someone exits the labor force entirely, the vibecession narrative starts to feel like it's doing real explanatory work. But I'd want to know: did you see anything in local data suggesting these were the same people being squeezed on housing and childcare, or could that be a different pressure? Because if it's the same population, you've got a real mechanism. If it's two separate stories getting averaged together, I'm not sure the distributional argument explains it as neatly as it sounds.