Mechanistic Over Narrative: Why Markets Are Machines, Not Stories

The analyst told a compelling story: strong consumer, resilient labour market, soft landing ahead. The narrative felt right. It was grounded in data and delivered with confidence.

The credit markets told a different story.

High-yield spreads were widening. Bank lending standards were tightening. The Federal Reserve’s Senior Loan Officer survey showed credit contracting. Defaults were starting to rise.

One was narrative. The other was mechanism. When the two disagree, I have learned to take the mechanism seriously first.

Markets do not care about your story. They are systems of transactions and credit, and they run on structure more than on sentiment.

---

The narrative trap

People are story-making machines. Give us disconnected facts and within seconds we will weave them into a plot. Daniel Kahneman called this fast, automatic, emotionally satisfying mode System 1.¹

A story feels true because it is coherent. But coherence is not causation.

Think of 2007. The narrative ran: housing is strong, subprime is contained, this is a buying opportunity. Meanwhile the ABX indices, which track subprime mortgage securities, were falling through the year.² High-yield spreads had begun widening from their mid-2007 lows. The 10-year to 3-month Treasury curve had been inverted for most of the year before. The recession would be dated from December 2007, and Lehman Brothers failed in September 2008. The stories did not survive. The credit signals had been telling the more accurate tale for more than a year.

This is mimetic desire at cultural scale. René Girard taught that we learn what to want by imitating models.³ We also learn what to believe from models: the analyst, the television guest, the investor whose record we admire. Borrowed conviction compounds until it looks like consensus, and consensus can be a long way from the mechanism underneath.

I do not mean the crowd is always wrong, or that contrarianism is a method. Reflexive opposition lets the crowd pick the question for you. The better question is: what machine is running, and what are its current inputs?

---

Dalio’s machine

Ray Dalio offers the clearest picture I know. The economy is a system of transactions and credit.⁴

1. Income enables spending.
2. Spending becomes someone else’s income.
3. The cycle repeats.

That circular flow explains trend growth but not cycles. Credit does that. When you borrow, you spend more than you earn, and your extra spending becomes someone else’s extra income. Asset prices rise, which supports more borrowing, and the loop compounds.

Then it reverses. Debt is a commitment to reduce future spending. As debt service rises, spending falls below income, and the loop runs backwards. Contraction is not simply “fear”. It is partly arithmetic.

Irving Fisher described this in 1933 as debt-deflation.⁵ Hyman Minsky formalised it as the financial instability hypothesis.⁶ Dalio describes a short-term debt cycle of roughly five to eight years and a long-term cycle of seventy-five to a hundred.⁴

The 2008 crisis fits this reading better than any story about sentiment: rising household debt, subprime expansion and securitization leverage, then debt service that could not be sustained, then defaults, contracting credit and collapsing spending. You did not need to predict when it would break. You needed to see that the mechanism was primed.

---

Cycles over lines

Most forecasting treats the future as a line: “GDP grew 2.5 percent, so we expect 2.3.” Mechanistic thinking asks where you are in a cycle, because that shapes what is likely next. One way to describe the phases:

1. Early expansion: credit easing, growth accelerating
2. Late expansion: growth strong, labour markets tight
3. Tightening: credit contracting, spreads widening
4. Slowdown or recession: credit stressed, defaults rising
5. Early easing: policy loosening, credit stabilizing

These are a way of organizing evidence, not a calendar. The National Bureau of Economic Research dates the official turning points.⁷ Since 1945, expansions have averaged about 64 months and recessions about 10, which tells you the structure repeats without telling you when the next turn comes.⁷

March 2020 shows how narrative lags mechanism. The headlines were catastrophic, and the high-yield option-adjusted spread reached 1,087 basis points on 23 March. Then policy responded on a scale few had imagined. By the end of August the spread was roughly 500 basis points.⁸ An investor who waited for the narrative to turn missed much of the recovery. The credit market had told you earlier.

---

Policy changes the machine’s payoffs

If markets are machines, policy is a set of control knobs. And the level of any one knob matters less than the combination. “The Fed funds rate is five percent” tells you very little unless you also know what the balance sheet, fiscal policy and credit channels are doing.

In 2020 and 2021 the policy rate was near zero, the Fed’s balance sheet was expanding at an unprecedented pace, fiscal support was enormous and credit was backstopped. The result was a rapid recovery, surging asset prices and, in time, inflation.

In 2022 and 2023 the Fed raised its target range from near zero to 5.25–5.50 percent, its balance sheet was shrinking, fiscal support was fading and credit was tightening after the regional bank failures. The result was slower growth, credit stress and volatile asset prices.

Same economy. Different policy settings. Different outcomes. I work through this carefully, with primary sources, in Five Lenses for Reading Monetary Policy.

---

The independence this gives you

Mimetic pressure in markets is strong. You see a respected investor buying technology and you want technology, not because you studied the mechanism but because someone you admire validated the choice.

Mechanism is a discipline against that pressure. It asks you to:

1. Define the state explicitly, with measurements.
2. Say what those measurements imply, and what would change your view.
3. Update when the state changes, not when the story changes.

It will not tell you exactly what happens next, and I would not trust anything that claimed to. It will make your reasoning visible and answerable. The way I try to combine several signals without pretending the result is a measured fact is in The State Estimator.

The analyst’s story might sound right. The credit markets do not care. I would rather watch the machine than admire the narrative.

This is general reflection, not investment advice.

---

References

Footnotes

1. Daniel Kahneman, Thinking, Fast and Slow (New York: Farrar, Straus and Giroux, 2011), pp. 19–30. The System 1 and System 2 terms come from Keith Stanovich and Richard West. ↩

2. Ingo Fender and Martin Scheicher, “The ABX: how do the markets price subprime mortgage risk?”, BIS Quarterly Review (September 2008), at bis.org. The authors describe a collapse in ABX prices since the summer of 2007. ↩

3. René Girard, Deceit, Desire, and the Novel, trans. Yvonne Freccero (Baltimore: Johns Hopkins University Press, 1965). ↩

4. Ray Dalio, How the Economic Machine Works (Bridgewater Associates, 2013), at economicprinciples.org: short-term debt cycle of about five to eight years, long-term debt cycle of about seventy-five to a hundred. ↩ ↩²

5. Irving Fisher, “The Debt-Deflation Theory of Great Depressions,” Econometrica 1, no. 4 (1933): 337–357. ↩

6. Hyman Minsky, Stabilizing an Unstable Economy (New Haven: Yale University Press, 1986). ↩

7. National Bureau of Economic Research, US Business Cycle Expansions and Contractions. For 1945–2020 the average expansion is 64.2 months and the average contraction 10.3 months. ↩ ↩²

8. ICE BofA US High Yield Index Option-Adjusted Spread, FRED series BAMLH0A0HYM2. 23 March 2020: 1,087 basis points. End of August 2020: roughly 500 basis points (508 bp in the Guggenheim Investments Q4 2020 High-Yield report; 502 bp in Concise Capital’s month-end commentary). FRED now carries only the latest three years of the series, so the 2020 levels come from secondary sources. ↩