The Turkey Problem in Finance: How Safe Assets Blow Up

Before 2008, mortgage-backed securities had returned positive yields in every single year since their creation.

Not one down year. Not one loss. Just steady, reliable returns funded by homeowners paying mortgages. The data was impeccable. The trend line was smooth. The rating agencies gave AAA ratings. The banks were confident. The investors were confident.

This is the Turkey Problem in its most literal financial form.

The Data Looked Perfect

The MBS (mortgage-backed security) was introduced as a financial innovation — a way to distribute risk and provide liquidity to the mortgage market. In practice, it worked smoothly for years. The mortgages paid. The securities returned. The system seemed to have solved the fundamental problem of lending: how to make a risky individual loan into a safe collective asset.

Every year that passed added more data confirming the safety of the instrument. 1990: positive return. 1995: positive return. 2000: positive return. By 2005, with 15 years of uninterrupted positive returns, the statistical confidence was extremely high. The banks built increasingly complex models around it. The leveraging increased. The exposure grew.

All of this was rational given the data available.

The problem wasn't the calculation of returns or the statistical rigor of the models. The problem was that the entire dataset was generated under a single regime: a period in which home prices had never declined nationally.

That assumption had never been tested. It wasn't in the historical data because the thing had never happened.

The Structural Break

The regime change came in 2006-2007. Home prices stopped rising and started falling nationally for the first time in the post-war era.

This single change invalidated the entire foundation of the MBS valuation. Once home prices were falling, defaults spiked (homeowners underwater had no incentive to pay). Once defaults spiked, the securities backed by those mortgages lost value. Once the securities lost value, the leverage that the banks had taken became unmanageable.

The banks weren't dumb. The rating agencies weren't corrupt (at least not primarily). The investors weren't reckless (at least not intentionally). They were all turkeys on day 1000, looking at solid data.

What the Data Hid

The historical data on mortgages hid several things:

1. The assumption embedded in the prices. Every year that prices went up, the market was implicitly assuming that prices would continue going up. The models called this assumption "conservative," but it was actually a bet. A structural break in that bet would break the model completely.

2. The correlation risk. When everything is going well, defaults are isolated and regional. But the MBS system is built on the assumption that defaults stay isolated. If defaults become correlated — if falling prices create incentives to default across regions simultaneously — the entire distribution assumption changes. The data from the boom period didn't contain examples of high correlation, so the models couldn't price it.

3. The concentration risk. The banks didn't just hold mortgages; they held leveraged exposure to mortgages. A system that returns 5% per year is interesting. A system that returns 5% per year and allows you to borrow 10x the capital and keep the 50% return? That's intoxicating. The leverage was rational given the perceived safety. But it meant that when the assumption broke, the losses were catastrophic.

The Turkey's Perspective vs. the Butcher's

A turkey investor in 2005 looked at 15 years of data and saw safety.

A butcher — someone who understood the structural assumptions built into the system — would have asked: what happens if the assumption breaks?

The butcher's question isn't about probability or prediction. It's about fragility. What if home prices fall? Can the system survive it? Are there leverage limits that would protect against it?

The answers were: the system cannot survive it, because the entire valuation depends on the assumption holding. And the leverage was unlimited because the risk was assumed to be non-existent.

The Non-Turkey Financial Position

After understanding the Turkey Problem, how do you position yourself in financial markets?

Not by trying to predict the break — you can't know when home prices will fall, or when a new financial instrument will blow up, or when the next structural assumption will be revealed as false.

Instead, you ask:

1. What structural assumptions is my position making? If I own this asset, what has to remain true for it to keep working?

2. What would break if the assumption reversed? If the assumption inverts — prices fall instead of rise, defaults spike instead of staying low — what happens to my position?

3. Can I survive the break? Do I have enough capital, enough diversification, enough flexibility to survive if the assumption breaks?

If the answer to #3 is no, then you have a Turkey Problem. You're safe as long as the assumption holds, and catastrophic if it breaks.

The banks in 2008 couldn't survive the break because they had leveraged themselves to the limit of the assumption. They had optimized for a single regime.

The non-turkey position: optimize for multiple regimes. Hold enough cash and capital that you can survive when assumptions break. Avoid leverage that makes you dependent on a single scenario playing out.

The Lesson for Your Own Investments

The deeper lesson: be skeptical of smooth data.

Smooth returns are not evidence of safety — they're evidence that you might be in a single regime. Regime changes happen. Assumptions get invalidated. Structural breaks arrive.

The more stable something looks, the more important it is to ask: what assumption is this stability depending on? And what would happen if that assumption were wrong?

That's non-turkey thinking in finance.