When Fear Spikes, Should You Buy?

By Victor Haghani and James White¹
ESTIMATED READING TIME: 5 min.

In a recent FT article, the head of investment strategy at a major US bank identifies a VIX above 30% as a powerful buy signal, noting that such entry points have yielded positive returns 70% to 83% of the time (depending on horizon) with average six-month gains of 12.4%. David Giroux, who manages the $66 billion T. Rowe Price Capital Appreciation Fund (PRWCX), shares the tactical view: “Historically, when the VIX is above 25% or 30%, the forward 12-month returns for the S&P 500 have been extremely attractive-well above the long-term average.”²

The pattern they describe is real, but it doesn’t necessarily follow that investors should increase their equity exposure when VIX is elevated. In fact, there’s a stronger case in theory and in practice for doing the opposite. The central issue is that a high expectation of a positive return, or even an above-average positive return, isn’t by itself a sensible investing goal. Rather, investors should seek to maximize their risk-adjusted expected return, so that’s how the quality of an investment strategy should be evaluated.

In that spirit, we’ll use Sharpe Ratio as a reasonable proxy for risk-adjusted return.³ We pulled VIX and S&P 500 data going back to January 1990 and compared several asset allocation approaches. A static stock/T-bill mix produced a Sharpe ratio of 0.50. A strategy that increased equity exposure when the VIX exceeded 30% – the FT article prescription – came in slightly lower at 0.47.⁴

However, strategies that reduced exposure during high-volatility periods did modestly better: sizing inversely to VIX, or simply cutting equity allocation from 100% to 70% when the VIX exceeded the median level of 17.5%, both produced Sharpe ratios of 0.54. A simple one-year momentum strategy, which cuts exposure when the market is down, which tends to be when VIX is elevated, generated the highest Sharpe ratio of the bunch at 0.59.⁵

None of these differences are statistically significant. With 36 years of data, the standard error on the Sharpe ratio is at least 0.17.⁶ So this history in and of itself doesn’t prove much, though it certainly runs counter to the thesis that buying when VIX is above 30% has been a good thing to do.

When data can’t give us the answer, we have to think from first principles. If your objective is to maximize the risk-adjusted return of your savings, finance theory and common sense both suggest that higher risk, all else equal, calls for cutting exposure, not adding to it. And, when expected returns are higher, all else equal, you should want to add to your exposure.⁷ By calling on VIX to do double duty – answering both questions at once – investors wind up with a muddled strategy that doesn’t work in theory and is likely to disappoint in practice.

1. We thank our partner Jerry Bell for his helpful comments. Of course, there is only so much he can do to help us, and all remaining short-comings are solely our own.
   
2. David Giroux, T. Rowe Price “Insights” (Market Volatility Outlook), October 2022.
   
3. Historical Sharpe Ratio is calculated as the arithmetic annual return over a period divided by the annualized volatility over the period. It doesn’t map directly to risk-adjusted return – but, for a properly scaled strategy maximizing Sharpe Ratio, it will also maximize risk-adjusted return, given a variety of standard assumptions.
   
4. We also tested a slower moving strategy that would add a fixed amount of stock exposure every day VIX was above 30% and reduce exposure every day it was below 30%, keeping exposure within a maximum and minimum band. We tested a range of daily increments/decrements from 0.1% to 0.5%, which resulted in a de minimis increase in average Sharpe ratio from 0.50 to 0.51.
   
5. For a deeper dive, see “Hold the Dip” by AQR (2025):
   “If investors want to add value through timing market entry and exit points, history shows it’s better to align with momentum than to buy the dip…”
   
6. The standard deviation of the Sharpe ratio (\(\text{SR}\)) from a sample \(N\) years long is \(\sqrt{\frac{\frac{1 + \text{SR}^2}{2}}{N}}\), assuming the returns are drawn from the same unchanging distribution. A rule of thumb for a long time period and not very high \(\text{SR}\) is \(\sqrt{\frac{1}{N}}\).
   
7. Using the Merton share formula – \(\frac{\mu – r_f}{\gamma \sigma^2}\) – a doubling of volatility (\(\sigma\)) from 15% to 30% requires the risk premium (\(\mu – r_f\)) to quadruple just to maintain your current position. To justify doubling your exposure when the VIX doubles, the premium would need to increase eight-fold, an expectation that strains credulity.