The CBOE Volatility Index (VIX) was launched within the Nineties to indicate investors expected future market risks. The Chicago Board Options Exchange’s VIX offers something unique: it uses 30-day options on the S&P 500 index to gauge traders’ volatility expectations. Essentially, it gives us a forward estimate of equity market volatility expectations.
But how accurate is that this value on a realized basis and when does it deviate from the market? We explored this query by comparing the complete range of VIX data since 1990 with the realized volatility of the S&P 500 index. We found that the market has overestimated volatility by about 4 percentage points on average. However, there have been unique times when the market has made significant miscalculations. We tell this story in a series of charts.
Figure 1 is a plot of all the time series of information. It shows that, on average, the VIX consistently exceeded realized volatility over time. And the spread was also consistent, except during peak periods (times when markets go crazy).
Exhibition 1.
In Figure 2 we summarize the information. The average realized volatility of the S&P 500 Index on a 30-day forward basis was 15.50% over the 35-year period. The average VIX (30-day forward estimate) over the identical period was 19.59%. There is a diffusion of 4.09% between the 2 values. This implies that there’s an insurance premium of 4.09 percentage points on the expected volatility to be protected against it on average.
Annex 2.
| Average (%) | Median (%) | |
| S&P Volatility (30 days ahead) | 15.50427047 | 13.12150282 |
| VIX (30-day estimate) | 19.59102883 | 17.77 |
| Difference (actual vs. estimated) | -4.086758363 | -4.648497179 |
Next, we turn to a period when there have been no major crises: from 1990 to 1996. Figure 3 shows how markets functioned during these normal times. The VIX consistently exceeded realized volatility by about five to seven percentage points.
Annex 3.
Figure 4 shows a really different time period: the worldwide financial crisis of 2008, and we see a really different story. In July 2008, realized volatility on a 30-day forward basis began to rise above the VIX. This continued until November 2008, when the VIX finally caught up and caught up with realized volatility. But then realized volatility fell again and the VIX continued to rise, exceeding realized volatility in early 2009.
Annex 4.
This appears to be a regular pattern in panics. The VIX reacts slowly to impending volatility after which overreacts once it recognizes impending volatility. This also says something about our markets: the Federal Reserve and other entities step in to dampen the VIX once things look too dangerous going forward, reducing realized volatility. We saw this dynamic again through the COVID period in Figure 5.
Annex 5.
There are two interesting insights from the charts. First, investors pay on average a 4% premium to be protected against volatility (i.e. the difference between the VIX and realized volatility). Second, the market is consistent at this premium, initially reacting slowly to large, unexpected events equivalent to the worldwide financial crisis and COVID after which overreacting.
These results illustrate the extent of premiums to be expected for insurance against extreme risks and the danger involved in paying an excessive amount of in times of market panic.
