Stress testing: how your portfolio behaves in a crisis

Your portfolio looks great right now. Green across the board, a tidy mix of tech, an index fund, maybe a couple of individual names you feel good about. It has never lost you a night of sleep. And that is exactly the problem: a portfolio that has only lived through calm markets has never actually been tested. You know how it behaves when things are fine. You have no idea how it behaves when they are not.

That gap is what stress testing is for.

Why isn’t one volatility number enough?

Because it is abstract and treats an upside surprise the same as a crash. Most risk tools hand you a single figure — annualized volatility, standard deviation, maybe a Sharpe ratio. It is not useless, but it is abstract. “Your portfolio has 18% volatility” does not land emotionally, and it does not tell you what a bad month actually feels like in dollars. Worse, volatility treats an upside surprise and a downside crash as the same thing. You do not care about them the same way.

A stress test replaces the abstraction with something concrete and personal: if the conditions of a specific historical crisis returned tomorrow, and your current holdings behaved the way that mix of assets behaved back then, here is roughly what would happen to your balance. Not a probability. Not an average. A specific, vivid “what if the past recurred” number you can look at.

Which historical crises are worth replaying?

Three with very different shapes: 2008, the 2020 COVID crash and the 2022 rate shock. The most honest stress tests do not invent a doomsday. They replay ones that already happened. Three are worth knowing well:

• 2008 Global Financial Crisis — a slow-motion collapse where credit froze, banks failed, and broad equity indices kept falling for well over a year.
• The 2020 COVID crash — violent and fast, a steep fall within a matter of weeks, then an equally sharp recovery.
• The 2022 rate shock — different in character: stocks and bonds fell together as rates rose, punishing the classic “bonds will cushion me” assumption.

Each has a different shape — slow grind, sudden shock, correlated bleed — and your specific mix of holdings would have absorbed each one differently. That is the point. Testing against all three shows you not one worst case but a range of ways things go wrong, and which of your positions tend to lead the fall each time.

How do you estimate a portfolio loss if the market drops 30%?

Multiply the drop by your portfolio’s beta. Historical replays cover crises that happened. But you also want to ask a cleaner hypothetical: what if the whole market simply dropped 30% from here? That is where beta comes in.

Beta measures how much your portfolio tends to move relative to the market. A beta of 1.0 means you roughly track it. Above 1.0 means you amplify its moves; below 1.0 means you dampen them.

The simple version of the math looks like this:

Estimated portfolio move ≈ beta × market move

So if your portfolio has a beta of 1.3 and you model a –30% market drop:

1.3 × (–30%) ≈ –39%

On a $100,000 portfolio, that is roughly a $39,000 paper loss in the scenario — versus about $30,000 for a portfolio that merely tracked the market at beta 1.0. Same headline “market fell 30%,” very different lived experience, purely because of how much risk you were carrying relative to the index. Seeing that difference in dollars is usually more motivating than any ratio. If you want the fuller vocabulary — how VaR, beta, and drawdown fit together — this primer on portfolio risk terms is a good companion.

Why do diversified portfolios fall together in a crisis?

Because correlations between assets rise sharply when investors sell everything at once. Here is the part most risk tools quietly understate, and the reason stress testing has to be read with humility.

In normal times, your holdings are diversified. Your tech stocks, your energy name, your international fund, your bonds — they move somewhat independently. On any given Tuesday, some are up while others are down, and that offsetting is exactly what diversification buys you.

In a real crisis, that breaks. When fear takes over, investors sell everything at once to raise cash, and assets that normally have little to do with each other start falling in lockstep. Correlations that were modest in calm markets climb toward 1.0. The diversification you were counting on thins out at the precise moment you need it most.

This has a sharp practical consequence: a stress test built on normal-times correlations will tend to understate your crisis losses, because it assumes a cushioning effect that historically evaporates when it matters. Good historical replays partly capture this — 2008 and 2020 already contain the correlation spike inside the real data. But it is worth holding in mind that the calm-market version of your risk is the optimistic one.

Does a stress test predict a crash?

No. It is a fire drill, not a forecast. Be clear-eyed about the limits, or the numbers will mislead you:

• History does not repeat exactly. The next crisis will rhyme with old ones, not copy them. Your 2008 replay is an analogy, not a prophecy.
• It is not a forecast. A stress test never says a crash is coming or likely. It says: if this specific past recurred, here is the mechanical result. The probability is not part of the output.
• Your holdings today are not the holdings from back then. The model applies today’s mix to yesterday’s conditions, which is a useful approximation but still an approximation.
• The precise dollar figure is false precision. Treat the rough size as the signal; treat the exact digits as decoration.

Used this way, a stress test is less a crystal ball and more a fire drill. It does not tell you the building will burn. It tells you where the exits are and whether you can actually reach them while calm — before you find out the hard way whether the loss you modeled is one you could sit through without panic-selling at the bottom.

---

Tradune’s stress test runs the beta-scaled version on your actual holdings — a “market drops X%” shock using each position’s beta — plus an illustrative interest-rate-rise scenario by sector. It does not replay 2008, COVID or 2022 day by day; the replays above describe how that kind of test works in general. See exactly what it computes on the portfolio stress test page. This is educational analysis of how risk tools work, not a recommendation, a forecast, or advice to buy or sell anything — your own situation is yours to judge.