The narrative fallacy: stories beat numbers

A semiconductor analyst publishes a report: "Intel is in structural decline because it lost the 7nm race to TSMC and Samsung. It will struggle to regain share. Downgrade." The report is well-written, internally coherent, and supported by evidence: foundry margins are low, capital intensity is rising, and second-sourcing is happening. Six months later, Intel announces a new architecture that performs better than TSMC's current generation. The analyst does not significantly upgrade; instead, she interprets this as "Intel is moving to a smaller disadvantage, but still declining." The evidence has changed, but the narrative has not.

This is the narrative fallacy. A human being creates a story about why something is true. The story becomes the filter through which all future evidence is interpreted. Evidence supporting the story is treated as proof; evidence contradicting the story is treated as noise or a temporary setback.

The narrative fallacy is more powerful than confirmation bias. Confirmation bias is the tendency to seek out information that confirms your belief. The narrative fallacy is the tendency to construct a coherent story and defend it against evidence to the contrary, sometimes with willful creativity.

Quick definition

The narrative fallacy is the cognitive bias toward constructing and maintaining a coherent story about causation and causality. Once a narrative is established, contradictory evidence is rationalized, ignored, or reinterpreted to fit the story. The narrative becomes more psychologically real than the data supporting it.

In equity research, the narrative is the central thesis: "Company X is improving margins," "Company Y is losing market share," "Sector Z is facing disruption." Once published, the narrative shapes how the analyst interprets all future information.

Key takeaways

• Stories organize data into meaning: Humans are not good at processing random data points; we are excellent at organizing information into stories. Analysts naturally create stories; the problem is that stories can outlive the evidence supporting them.
• Narrative consistency feels like truth: A coherent story feels more true than a disconnected set of facts. This feeling is not evidence of truth; it is evidence of narrative coherence.
• Changing your mind feels like admitting error: Once a narrative has been published and defended, changing it feels like admitting the previous analysis was wrong. This discourages intellectual flexibility.
• Cherry-picking evidence feels like research: An analyst can gather evidence supporting a narrative while ignoring contradictory evidence. The selective gathering feels like rigorous research; it is actually selective confirmation.
• Markets eventually test the narrative: If a narrative is radically wrong, the market will eventually reveal it. But the timeline is uncertain, and the analyst who held the wrong narrative for too long suffers career consequences.

Why narratives are sticky

The human brain is not a Bayesian inference machine, despite what some economists assume. Instead, it is a story-creating machine. When presented with facts (Intel's 7nm architecture failed, TSMC is ahead), the brain does not neutrally update a probability distribution. Instead, it constructs a narrative that explains the facts: "Intel is in decline because management misallocated capital and lost the foundry race."

This narrative immediately shapes perception. The analyst now interprets future data through this narrative lens. If Intel's new architecture is better than expected, the narrative reinterprets it: "This is a temporary win, but Intel's fundamental problem remains—they are late to the race and cannot catch up."

The narrative becomes self-reinforcing because the analyst is selectively attending to evidence that fits the story and filtering out or reinterpreting evidence that does not.

This is not unique to analysts. Historians call this the narrative fallacy in historical analysis: historians construct stories about why past events happened, then interpret ambiguous evidence as supporting those stories. A narrative about the decline of the Roman Empire organizes thousands of facts into a coherent whole. But the narrative shapes which facts are considered important and how they are weighted.

In equity analysis, the cost of a wrong narrative is financial loss. An analyst who develops a bearish narrative about a company will miss an important recovery if the recovery contradicts the narrative. By the time the analyst revises the thesis, substantial opportunity has been lost.

How the narrative drives forecasts

A narrative is not just a description of the past; it is a prediction about the future. "Intel is in decline" implies that Intel's market share will continue to erode, margins will remain under pressure, and capital intensity will rise. These are forecasts.

But here is the trick: once the narrative is established, the analyst is not consciously forecasting from first principles. Instead, the analyst is defending the narrative. When asking, "Will Intel's market share decline next year?" the analyst is really asking, "What evidence would contradict the decline narrative?" When no such evidence appears, the narrative feels confirmed.

This is where the narrative fallacy becomes financially harmful. The analyst is not updating their forecast based on new evidence; they are selectively gathering evidence that supports the forecast already embedded in the narrative.

Consider an alternative example. A growth-stock analyst publishes a report: "SaaS Company X is the leader in its category and will grow 40% for the next decade." The narrative is "winner takes all; X will dominate." Now, the analyst encounters data: Company X's growth is slowing from 50% to 40% due to competitive pressure. In a narrative-free analysis, this would cause a forecast revision. But in the narrative framework, the analyst reinterprets: "Growth is normalizing, but at a still-impressive 40%, which maintains the dominance thesis." The narrative survives the evidence update.

The role of publication and reputation

Once an analyst publishes a thesis, changing it becomes emotionally and professionally costly. Changing a rating from Buy to Sell means admitting the previous analysis was wrong. This creates a bias toward defending the original narrative even when evidence shifts.

This is exacerbated in sell-side research, where an analyst's rating is public and tracked. A portfolio manager who changes their mind has only their own opportunity cost to bear. A sell-side analyst who changes their mind has their reputation and client relationships at stake. Clients who bought based on a Buy rating become angry if the analyst downgrades. The analyst faces career risk.

As a result, sell-side analysts tend to revise ratings slowly, even when evidence shifts. The narrative persists longer than the data supporting it.

Institutional factors also play a role. An analyst who publishes a bearish thesis on a large company may find the company becomes less willing to grant access and management meetings. The company's investor relations may direct other companies away from the analyst. These career pressures discourage analysts from developing bearish narratives, or if they do, from defending them when challenged.

Narrative and market cycles

The narrative fallacy interacts dangerously with market cycles. In bull markets, optimistic narratives become self-reinforcing. A narrative that "software is the future and all software companies will earn premium multiples" becomes stronger as software multiples expand. The narrative feels more and more true as evidence accumulates (in the form of rising stock prices).

Then the market turns. Software multiples compress. The narrative should be updated. But the narrative is now deeply embedded in how the analyst thinks. Instead of revising, the analyst narrates a new story: "The selloff is temporary; software is still the future, just in a different form." The original narrative persists, disguised in new language.

By the time the analyst has revised the narrative, the opportunity to sell at the top has long passed. The analyst is describing a market that no longer exists.

Real-world examples

Financial stocks 2007–2009: Analysts developed a narrative in the early 2000s: "Financial innovation and lending have reduced risk. Banks are safer than they were a decade ago." As housing prices rose, this narrative felt more true. Quarterly earnings beat expectations, validating the narrative. When evidence of deteriorating credit quality emerged in 2007–2008, analysts initially reinterpreted it as temporary. The narrative only collapsed when the market forced a reckoning.

Amazon's profitability: For fifteen years, Amazon's narrative was: "We are sacrificing short-term profits for long-term market dominance." This narrative explained the company's low profitability and justified high valuations. The narrative was so strong that when Amazon began earning substantial profits in 2015–2016, many analysts reinterpreted it: "Profits are temporary; growth is what matters." The narrative persisted even when the underlying fact (low profits) changed.

Tesla's production challenges: The Tesla narrative for years was: "Elon Musk and Tesla will revolutionize car manufacturing; the Model 3 will be produced at the promised volumes." Each time production fell short, the narrative survived reinterpretation: "Production is ramping faster than expected, given the complexity." The narrative outlasted the evidence supporting it by years.

Value investing 2010–2020: A narrative emerged in the 2000s that "value investing has died because growth dominates." As value underperformed, the narrative felt more true. When value rallied 30% in 2016, the narrative was updated to "value is bouncing but will eventually lose again." When value rolled over in 2017–2018, the narrative felt vindicated. The narrative shaped how analysts interpreted a dataset that actually showed value was cyclical, not dead.

Common mistakes

Mistake 1: Falling in love with your narrative instead of testing it. Once you have written down a detailed thesis, you naturally become invested in its truth. Separate the narrative (your story) from the data (the evidence). Actively seek data that contradicts the narrative, not just data that supports it.

Mistake 2: Treating narrative coherence as evidence of truth. A story can be internally consistent and completely wrong. Coherence is a property of the story, not the underlying reality. A narrative about "Intel's decline" is coherent and elegant; that does not mean Intel is in decline.

Mistake 3: Reinterpreting contradictory evidence instead of revising the narrative. When data contradicts the narrative, the first instinct is to reinterpret. Resist it. Ask: what would it take for me to abandon this narrative entirely? If the answer is "nothing," the narrative is not a hypothesis; it is a belief.

Mistake 4: Defending the narrative publicly before updating it privately. Sell-side analysts publish research, then defend it in client calls. This commitment makes it harder to change privately. Separate publication from belief. A published report is a report; your true belief might differ, and evolve.

Mistake 5: Confusing narrative simplicity with truth. A simple, elegant story is more memorable and feels more true. But the world is often complex and messy. A narrative that "Company X is a winner" is simpler than "Company X has competitive advantages in segment A, competitive disadvantages in segment B, and uncertain outcomes in segment C." But the second narrative is more likely to be true.

FAQ

Q: Is it impossible to have a narrative without falling into the narrative fallacy?

A: No. Humans need narratives to organize information. The key is to treat the narrative as a hypothesis to test, not as truth to defend. Write down the narrative, then ask: "What evidence would contradict this? What am I NOT looking for?" If you can articulate the conditions under which the narrative is false, you have created a testable hypothesis rather than a fixed belief.

Q: How do I know if my analysis is driven by narrative fallacy?

A: Ask yourself: (1) Have I actively sought evidence that contradicts my thesis? (2) Have I changed my mind in the last six months based on new information? (3) Can I articulate the conditions under which my thesis is wrong? (4) Am I interpreting ambiguous evidence as supporting my thesis? If you answer no to (1)–(3) or yes to (4), narrative fallacy is likely present.

Q: Should I avoid creating narratives altogether and stick to pure numbers?

A: Pure numbers without narrative are hard to interpret. A P/E ratio of 18 is meaningful only in a narrative context (narrative: "this company is growing faster than peers, justifying a premium multiple" vs. "this company is in decline and the market is overpaying"). The solution is not to avoid narratives but to treat them as working hypotheses, not fixed truths.

Q: How do I write research without falling into narrative fallacy?

A: Write the narrative, then write the critique of the narrative. For every reason you think the thesis is true, write down reasons it might be false. Present both. Tell readers the conditions under which your view is wrong. This forces you to confront the narrative fallacy while still communicating your thesis.

Q: Do successful investors use narratives or avoid them?

A: Successful investors typically use narratives but hold them lightly. They tell themselves a story about why an investment makes sense, but they monitor constantly for evidence that the story is wrong. When the evidence arrives, they change the narrative or exit the position. The key is intellectual flexibility, not the absence of narrative.

Q: Is the narrative fallacy the same as confirmation bias?

A: No. Confirmation bias is seeking out information that supports your belief. Narrative fallacy is constructing and defending a story to explain information, even when the story changes as evidence shifts. Confirmation bias is active (you seek supporting evidence); narrative fallacy is passive (the story organizes whatever evidence you encounter).

Related concepts

• Confirmation bias: The tendency to seek and interpret information as supporting your existing beliefs.
• Belief perseverance: The tendency to maintain a belief even when presented with contradictory evidence.
• The sunk-cost fallacy: Continuing to invest in a belief because you have already invested time and money defending it.
• Hindsight bias: The tendency to see past events as more predictable than they were, creating false narratives about past analysis.
• Path dependence: How initial conditions and choices constrain future outcomes, making some narratives more likely to persist.

Summary

A narrative is a story that explains why events happened and predicts why future events will happen. Narratives are how humans organize information and make sense of complexity. But narratives also create a dangerous bias: once established, they filter evidence and persist even when contradicted.

The analyst who publishes a thesis creates a narrative. The narrative then shapes how all future evidence is interpreted. Evidence supporting the narrative feels like confirmation; evidence contradicting it is rationalized or ignored. The narrative becomes more real than the data.

The antidote is intellectual flexibility: treat your narrative as a hypothesis to test, actively seek contradictory evidence, and revise when the evidence demands it. The analyst or investor who changes their mind based on new information will be more successful over time than one who defends a narrative against reality.

Next

Read the next article: Herding among sell-side analysts.