Narratives vs Data

The Chicken or the Egg

One of the oldest debates in economics takes on a new form through the lens of Narrative Economics. Does data create narratives, or do narratives create data?

The intuitive answer seems obvious. Of course data comes first. Inflation goes up, the “inflation panic” narrative follows. Employment numbers disappoint, the “recession” narrative spreads. Cause (data) leads to effect (narrative). Simple, right?

Shiller’s answer is considerably more complicated. The relationship is bidirectional, and narratives lead data more often than most people realize.

When Narratives Move First

Think back to early 2022. By the time U.S. CPI hit the 7% range, the “inflation panic” narrative had already been dominating markets for months. From the second half of 2021, warnings about inflation had been spreading through social media and financial outlets, and there is substantial analysis suggesting that the narrative itself accelerated corporate pricing decisions.

The mechanism works like this. When the expectation “prices will keep rising” becomes widespread, businesses conclude “we need to raise prices now” and act preemptively. Consumers think “I should buy before things get more expensive” and pull purchases forward. These behavioral responses actually push prices higher. The narrative created the very reality it predicted.

The Economics of Self-Fulfilling Prophecy

The self-fulfilling prophecy is a core mechanism of Narrative Economics. When enough people believe in a particular future, their belief-driven behavior literally makes that future happen.

The textbook example is a bank run. A rumor (narrative) spreads that a bank might fail. Depositors rush to withdraw their money. The mass withdrawal drains the bank’s liquidity and actually causes it to fail. A story that may have been false at the outset becomes true through the behavior it triggered.

Recessions can work through the same mechanism. When the “recession is coming” narrative reaches sufficient intensity, consumers cut spending, businesses delay investment, and hiring freezes spread. These individually rational defensive actions, aggregated across an economy, actually slow growth — and the recession becomes real. The narrative manufactured the future it forecast.

Not every narrative is self-fulfilling, of course. The narrative “an asteroid will hit Earth tomorrow” will not summon an asteroid regardless of how many people believe it. Self-fulfilling prophecies are particularly powerful in domains where human behavior directly shapes outcomes — economics, finance, politics.

The Framing Effect: Same Data, Different Stories

Another critical mechanism in the narrative-data relationship is the framing effect. Identical data can have opposite market impacts depending on which narrative frame is applied.

Consider a jobs report. Nonfarm payrolls come in 300,000 higher than expected. The same number supports two entirely different narratives:

Frame A — “The Fed can’t cut rates”: Employment is too strong, so the Fed will delay rate cuts. Higher rates are bad for stocks. Market falls.

Frame B — “Soft landing confirmed”: Strong employment without recession proves the economy is healthy. Corporate earnings should follow. Market rises.

The exact same data point triggers opposite market reactions. The difference is not the data itself but the dominant narrative through which the data is interpreted. In 2022, Frame A dominated, so strong jobs data pulled stocks down (“good news is bad news”). By 2023, Frame B gained strength, and similar data pushed stocks higher.

Oil Prices: Where Narrative Runs Ahead of Fundamentals

The crude oil market provides some of the clearest examples of narratives leading fundamentals. Oil prices frequently move on narrative well before any actual change in supply or demand occurs.

When geopolitical tensions escalate in the Middle East, the “supply disruption” narrative pushes oil prices higher even when no actual supply has been lost. Traders are not betting on the probability of a supply cut — they are betting on how other traders will respond to the narrative. It is Keynes’s beauty contest in its purest form.

The same dynamic plays out before OPEC meetings. “Production cut likely” or “deal will collapse” narratives move prices days before the actual decision. Sometimes the pre-meeting narrative-driven price movement is larger than the reaction to the actual outcome.

Recognizing the Narrative Filter

The practical takeaway from all of this is straightforward. We believe we are looking at data “as it is,” but we are always viewing it through the filter of the dominant narrative.

Data feels objective. Numbers are numbers. But our judgment about which numbers matter and what those numbers mean is already shaped by the prevailing narrative. When “inflation panic” dominates, all eyes are on CPI day. When “AI revolution” dominates, Big Tech earnings consume all the oxygen. Important data released during the same period goes largely ignored.

Why does this matter? Recognizing the narrative filter does not allow you to remove it. We are human, and we are all operating under the influence of the dominant narrative. But at minimum, we can ask ourselves: “Am I looking at data right now, or am I looking at a story?” That question is the first step in separating signal from noise.

So, Which Comes First?

Shiller’s conclusion is clear. Data and narratives form a continuous feedback loop, each reinforcing the other. Sometimes data creates the narrative. Sometimes the narrative creates the data. The important question is not “which comes first?” but “which direction is the feedback loop spinning right now?”

That is what RETIC tracks every day. Which narratives are today’s data reinforcing, and which narratives are shaping tomorrow’s data. And as always, our predictions based on this tracking are humbly prepared to be wrong. The feedback loop spins fast, and even when you can see it clearly, predicting where it stops is another matter entirely.

What we can offer is a better understanding of the spin. And in markets where most participants do not even realize the loop exists, awareness of the mechanism — even without predictive precision — is a genuine advantage. A small one, perhaps. But we will take what we can get.