Between January 2007 and May 2012, researchers at Texas Tech University tracked a deceptively simple number: how investors' self-reported risk tolerance moved against the S&P 500. They were using five years of data from FinaMetrica, a widely-used financial planning questionnaire, against monthly market returns.
The correlation they found during the bear market phase was 0.90.
What that means: as stock prices fell, investors consistently reported lower willingness to take risk. As prices recovered, their stated tolerance climbed back up with them. Nearly in lockstep. The investors' risk tolerance scores were following the market down and back up — which is the opposite of how rational risk-taking would behave. You should want more risk when prices are cheap and less when they're expensive. What the data shows is that investors want less risk exactly when they should want more.
Most people who have sat with a financial advisor have encountered a version of this question: "If your portfolio dropped 30% in six months, what would you do?" In the calm of an advisor's office, the honest answer for most people is: hold steady, maybe add more. The questionnaire records this as a reasonable risk tolerance. A portfolio gets built around it.
Then a real 30% drop arrives.
What the questionnaire actually measures
In 2012, Carrie H. Pan and Meir Statman published a paper in the Journal of Investment Consulting identifying five structural deficiencies in standard risk tolerance questionnaires.
The most consequential: investors' risk tolerance varies by circumstances and associated emotions, and questionnaires — administered once, in a calm setting, with nothing at stake — fail to capture this variation.
Pan and Statman aren't arguing that risk tolerance doesn't exist. They're identifying a precision problem: what questionnaires typically measure is closer to "how do you feel about risk when you have no skin in the game" — a baseline, perhaps, but not a reliable predictor of behavior when a real portfolio is falling in real time.
This distinction matters because much of the financial planning infrastructure is built on questionnaire-derived scores. The score becomes the portfolio. The portfolio becomes the expectation. And the expectation is that you will behave the way you said you would behave.
The research suggests that expectation is miscalibrated — not because anyone lied, but because risk tolerance in the abstract and risk tolerance under genuine financial duress are different cognitive events, running on different parts of the brain.
Pan and Statman identify a second deficiency worth flagging: investors often hold different risk tolerances for different goals — higher for long-horizon retirement money, lower for a near-term home purchase or an emergency fund. A single questionnaire score compresses this variation into one number, which then governs an entire portfolio.
What real fear does to risk tolerance
In 2018, Luigi Guiso, Paola Sapienza, and Luigi Zingales published a study in the Journal of Financial Economics using something rare: repeated surveys of an Italian bank's clients, combined with actual portfolio data, before and after the 2008 financial crisis.
Their central finding: after the crisis, the average investor's measured risk aversion roughly doubled. For the median investor, risk aversion increased by a factor of 3.5.
The researchers then tested four possible explanations for this shift: wealth loss, changed income expectations, updated probability estimates, and emotion-based changes in decision-making. The data pointed toward fear as the primary driver. To test this mechanism directly, they ran a lab experiment: subjects who watched a horror movie had a certainty equivalent 27% lower than those who did not — demonstrating that the emotional state of fear itself, independent of any rational update about the future, directly suppresses risk tolerance.
This is the mechanism that questionnaires cannot capture. A 30% drawdown doesn't just bring new information about company fundamentals or macroeconomic conditions. It generates an emotional response — visceral, physical, and hard to override through analysis — that rewrites, in real time, the risk tolerance number you gave an advisor six months earlier.
Paper conviction is built in calm conditions. Real conviction is tested in fear conditions. Most investors don't know which one they have until the test arrives.
The doubling of risk aversion in the Guiso-Sapienza-Zingales data was not a small fluctuation. These were investors who, like you, had filled out questionnaires and agreed to portfolios designed to reflect their stated tolerance. When the crash hit, their actual behavior diverged sharply from those stated preferences — not because they were irrational, but because fear is a physiological event that changes how the brain processes financial decisions.
The procyclical trap
The Texas Tech data reveals something more disorienting than a simple gap between what investors say and what they do. It shows that stated risk tolerance moves with the market — procyclically.
The implication: questionnaire scores are highest when stocks are most expensive, and lowest when stocks are cheapest. If a portfolio is periodically rebalanced to reflect "current" risk tolerance as measured by questionnaire, it will systematically shift away from equities at the best buying opportunities and toward equities at the worst ones.
The Texas Tech dataset found that although risk tolerance scores did shift during the financial crisis, the distance between the lowest and highest scores was only 7.1 percent, compared to a 51.7 percent swing in the S&P 500 during the same period.
The scores moved — but modestly, and in the wrong direction. The investors who said in 2007 they could handle a significant drop became, by early 2009, people who said they wanted less risk — exactly when prices had fallen sharply and the Guillemette-Finke dataset recorded its 51.7 percent swing in stock prices from the period's high to low.
One test of whether your risk tolerance is real: could you have added to your equity position in March 2009, when markets were at their crisis lows and financial headlines were predicting further collapse? If the honest answer is no, your questionnaire tolerance and your lived tolerance are different numbers.
Your first crash is the hardest
In 2011, Ulrike Malmendier and Stefan Nagel published a study in the Quarterly Journal of Economics using nearly five decades of Survey of Consumer Finances data. Their central finding: individuals who have experienced low stock market returns throughout their lives report lower willingness to take financial risk, are less likely to participate in the stock market, and invest a smaller fraction of their liquid assets in stocks.
The paper is titled "Depression Babies" — named for the cohort that came of age during the 1930s and carried measurably more conservative financial behavior for decades afterward, even after markets recovered. The mechanism is experiential, not rational: people aren't less risk-tolerant because they calculated that the Depression era implied lower expected returns forever. They're less risk-tolerant because they lived through it.
And crucially: younger cohorts — those with less market history — react more strongly to crashes than older investors who have experienced more full market cycles.
Your first major drawdown is the hardest because it is the first time the fear mechanism has been activated at full intensity, against real money, in a real portfolio. The investor who has lived through 1987, 2001, 2008, and 2020 carries a kind of visceral knowledge: they know, in their body, not just their spreadsheet, what recovery feels like. That experiential knowledge doesn't guarantee they'll behave perfectly, but it buffers the fear response. It provides a reference point.
The investor facing their first 30% drop has no such reference point. They have a questionnaire score, a conviction that they are a long-term investor, and an account balance that is falling in real time. The gap between the first two things and the third is what a drawdown reveals.
What paper conviction is actually worth
The research points toward a specific insight: paper conviction is a reasonable starting point and a poor finishing point.
Knowing that losses hurt roughly twice as much as equivalent gains feel good — a finding from Kahneman and Tversky's 1979 Prospect Theory — isn't just a trivia fact about behavioral economics. It's a specific prediction about how you will feel when your portfolio loses 30%. The pain will be roughly twice as large as the pleasure you felt watching it gain 30%. If you understand that in advance, you can plan for it — not eliminate it, but anticipate it.
The investors who hold through real drawdowns are not the ones who tricked themselves into believing drawdowns wouldn't hurt. They're the ones who decided, in advance, that they could survive the hurt. That's a different kind of conviction — not paper conviction, which says "I won't feel this," but real conviction, which says "I know I'll feel this and I've decided to hold anyway."
Building that conviction requires exposure to evidence before the fear arrives. What actually happened to investors who sold in March 2009 versus those who held? What has the S&P 500 done after every major correction in its history? What does the academic literature on investor timing tell you about the actual outcomes of emotional selling?
Real drawdown conviction also benefits from designing your portfolio before the crash, not during it. The investor who has pre-decided their response — who has written down exactly what they will do when the market drops 30% and why — is less susceptible to fear-driven deviation than the investor who has not.
None of this makes the fear go away. It makes the fear survivable.
The first real drawdown earns you something a questionnaire cannot: evidence about which kind of conviction you actually have.
Ready to see your full financial picture?
Try Pro free for 30 days. No bank login required. No credit card.
Create your free dashboard