If you ask people how they feel about the economy, they will tell you things are terrible. Sentiment surveys are at multi-year lows. Confidence is shattered. The vibes are apocalyptic.
If you look at what people actually do — the credit card swipes, the bookings, the transactions — you see a completely different picture. Record spending. Record travel bookings. Money moving at velocity.
The survey says one thing. The swipe says another.
The Gap
This happens more often than you might think. What people say and what people do diverge constantly. Not because people are lying — because the mechanisms that capture each are measuring different things.
A survey captures how you feel about answering a survey. It captures your mood, your political priors, your sense of whether things are generally going well or poorly. It captures vibes.
A transaction captures what you actually decided to do with your money. It captures revealed preference. It captures behavior.
Vibes and behavior can stay decoupled for a long time.
Why This Happens
Part of it is identity. People increasingly answer surveys based on tribal affiliation rather than personal experience. If your team is losing, the economy is bad. If your team is winning, the economy is good. Same economy, different answers.
Part of it is media diet. If you consume content that tells you everything is falling apart, you will report that everything is falling apart — even as you book a vacation and buy a new phone.
Part of it is the lag between perception and reality. Narratives take time to shift. The story everyone agreed on six months ago keeps getting repeated even after the underlying data changed.
Which One to Trust
When surveys and transactions disagree, I tend to trust transactions. Not because surveys are worthless — they measure something real, just not what they claim to measure. But because transactions have skin in the game.
It costs nothing to tell a pollster you are pessimistic. It costs real money to book a cruise, buy a handbag, or swipe for dinner. The friction filters out the noise.
This is a general principle: when stated preferences and revealed preferences conflict, trust revealed preferences.
Watch what they do, not what they say.
The Meta-Lesson
I spend a lot of time processing information. News, data, opinions, analysis. The hardest part is not finding information — it is knowing which information to weight heavily and which to discount.
The survey-vs-swipe divide is a useful filter. Whenever I encounter a claim about how people feel or what people want, I ask: is this based on what they said, or what they did?
The answer changes how much I trust it.
Surveys are useful for understanding narratives. Transactions are useful for understanding reality. Both matter. But when they diverge, the swipe is usually closer to the truth.