As short-term volatility leaks lower and lower and more and more talking heads use this 'risk-index' as reason to be longer and longer stocks, we thought it might be useful to get some context on recent movements in volatility-related factors. Whether its seasonality, volatility term structure, or the high-yield credit market, VIX looks low (underpricing short-term risk). Perhaps it is a plethora of new-year-new-book covered call euphoria or just belief in the LTRO firewall fixing tail-risk (or US decoupling shifting us to moderation), short-term options are sending some different messages to the risk-is-on-like-donkey-kong 'broadcast' that the contemporaneous (and in no way leading) VIX is being mis-understood as indicating. We present six perspectives that should be considered before more nickels are picked up in front of the micro (earnings) and macro (you name it) steamroller.
First, VIX is considerably seasonal. The chart below shows the average drop is short-term volatility has been very impressive into year-end and indeed for the first week or two fo the new year there is little to no pick up from this seasonal. By week 3, seasonals show that VIX picks up notably. Is all this VIX compression, simply a December/January-effect on VOL? inferring the seasonals would mean a higher VIX is about to begin?
Second, demand for index protection remains much stronger than it seems underlying single-names suggest as implied correlation diverges from VIX. Whether this is covered call writing en masse or simply the market makers and professionals more than happy to carry the crash risk (knowing something we mere mortals do not?), implied correlation (which measures the relative demand for index volatility versus the volatility of the underlying names in the index) has risen consistently in the last week or so even as Vol has dropped notably. This divergence tends to revert with Vol moving to implied correlation (in this case VIX higher).
Third, compared to the high-yield credit market, vol is considerably too low. It would be hard to argue that the high-yield credit market did not better reflect real-money investor's greater-than-short-term risk-appetite for well, risk, given the huge skew to getting-it-wrong in credit. The tight relationship between HY credit and volatility (which is inherently non-linear and dynamic but for the sake of simplicity we track in this chart) has broken in the last few months and short-term volatility is not at all pricing similar risks to the HY market. Arguing that perhaps HY should revert lower is a stretch also when one reflects on the market participants' attitudes, real-money, size, and perspective. Perhaps HY is not pricing the impact of an inevitable QE3 as much as vol? We suspect the truth is somewhere in between and vol will rise.
Fourth, the term structure of short-term vol to medium-term vol is very steep which has tended to infer over-optimism and a rapid reversion to higher short-term vol levels. The relationship between 1m and 3m vol for stocks can provide some insight into just how over-excited all those premia sellers are. It seems the natural momo players from stocks have found a new home and chase vol lower and lower, selling more and more premium, until of course, crack. It would appear that both the steepness (i.e. short-term vol is very low compared to medium-term) and absolute levels of vol suggest that maybe those picking up nickels in front of the steam-roller need their knuckles rapped once again.
Fifth and Sixth are related to the differences between a normal distribution and the real distribution of returns implied by options prices for SPY (the S&P 500 ETF). If we create a more complex model for valuing options (away from the Black-Scholes model) then we can 'calibrate' this model to understand how skewed (left or right skewed - i.e. how much upside/downside risk is expected) and how kurtotic (a big word for fat-tailed or how much extreme risk is expected).
The implied kurtosis (which really reflects what the options prices are saying aboutthe risk of extreme moves) is showing that both long- and short-term indications are for a rising expectation of extreme moves. This rising expectation is perhaps what is helping price the index vol (implied correlation above) higher than the underlying micro vols...the point being that under the surface, extreme risk expectations are rising (which is not what VIX is reflecting).
The implied skewness (which is different from the skew in vol that many are used to thinking about) is calibrated by adjusting the underlying distribution of returns to meet options prices. In the case of SPY, we can see longer-term expectations of notable downside skew rising significantly in the last week or two (in fact the highest (lowest on chart) since the US downgrade). Short-term options do indeed reflect the lower concern of downside as they have risen recently but remain around the median for risk and are also notably divergent from longer-term options. After profit-taking on skewness positions as we crashed in Q3 2011, short-term options skewness does not reflect the move in stocks and longer-term skewness is much more concerned at downside.
All-in-all, it would appear that when we scratch at the surface of VIX, we find some notable references for why it has dropped (seasonals, HY compression, profit-taking on complex options, covered call writing?) but perhaps more importantly the clear picture is one of an instrument (short-term vol as VIX is untradable directly) that has been pushed too far and risk-reward on the risk-index seems skewed to significant upside from here.
Charts: Bloomberg