Vol smirk is almost always a mechanical artifact of a falling price (implied vol rises as price falls). Additionally, changes in E/P (d/p, b/p) are primarily a consequence of changes in P, as the numerators tend to be much less volatile. So, the best indicators reduce to “flavors” of negative price momentum.
One thinks that, but look carefully at the data. The smirk in the options price insures vol goes up when prices fall. In fact, the intra day correlation is nearly 100% (Joanne hill shows us that) . So as markets fall, implied vol goes up. You can see it in VIX charts - all the major vol spikes correlate with selloffs. It is not a mark of informed investors (as one might hope).
You're talking about AFTER the correction already started, which is true. I'm talking about BEFORE the correction starts when the market is still in an uptrend. In some instances, vol skew was up b/c of portfolio insurance.
I think it’s an empirical question. Intraday changes in S&P implied vol, for example, are almost 100% directly correlated with movements in the index as changes in the index simply traces different levels of vol implicit in the smirk. Specifically, implied Vol rises when prices rise, or fall, as they index moves along the volatility surface. On extreme days, the smirk itself moves, but not much in most cases and almost never on tick to tick timing. The key, however, is that the vol surface is a smirk and not a smile. So, on average, implied vol goes up much more when the market falls than when it rises. This is why most of the big spikes in daily implied vol are associated with down markets. It isn’t hedging (that’s already implied in the smirk). So, again, the bigger changes in implied vol correspond with falling prices - ie negative momentum. Of course, although the smirk is mostly stationary intra-day, the smirk moves over time reflecting a myriad of things. The key point here is that the larger changes in implied vol are associated with down markets in daily data, but the relationship can be fuzzier for monthly data. However, having looked at it a lot, I am pretty confident that the dominant data points in a statistical analysis will correspond to large down markets. This isn’t driven by investor activity as much as it is an artifact of the shape of the vol curve (which IS influenced by investor sentiment to a degree).
That’s my point. As prices fall, smirk causes implied vol to rise. A statistical study will correlate increased vol with future falling prices. But, that is equivalent to saying that falling prices (which, via smirk causes the vol increase) predict future falling prices. But that is the same as saying momentum is predictive (which it is). We looked hard at this and the smirk entanglement was enlightening. I’ve refereed a few papers on this topic and it usually boils down to momentum.
Thanks so much Doc, Larry Connors is truly brilliant.
Glad it helped
One could also use Log-Periodic Power Law Singularity (LPPLS):
• This mathematical model identifies bubbles by detecting faster-than-exponential growth in asset prices.
• It uses patterns of price oscillations (log-periodic oscillations) as the system approaches a critical point.
Vol smirk is almost always a mechanical artifact of a falling price (implied vol rises as price falls). Additionally, changes in E/P (d/p, b/p) are primarily a consequence of changes in P, as the numerators tend to be much less volatile. So, the best indicators reduce to “flavors” of negative price momentum.
Vol skew can increase before a correction when investors bid up OTM puts, buying them for protection.
One thinks that, but look carefully at the data. The smirk in the options price insures vol goes up when prices fall. In fact, the intra day correlation is nearly 100% (Joanne hill shows us that) . So as markets fall, implied vol goes up. You can see it in VIX charts - all the major vol spikes correlate with selloffs. It is not a mark of informed investors (as one might hope).
You're talking about AFTER the correction already started, which is true. I'm talking about BEFORE the correction starts when the market is still in an uptrend. In some instances, vol skew was up b/c of portfolio insurance.
I think it’s an empirical question. Intraday changes in S&P implied vol, for example, are almost 100% directly correlated with movements in the index as changes in the index simply traces different levels of vol implicit in the smirk. Specifically, implied Vol rises when prices rise, or fall, as they index moves along the volatility surface. On extreme days, the smirk itself moves, but not much in most cases and almost never on tick to tick timing. The key, however, is that the vol surface is a smirk and not a smile. So, on average, implied vol goes up much more when the market falls than when it rises. This is why most of the big spikes in daily implied vol are associated with down markets. It isn’t hedging (that’s already implied in the smirk). So, again, the bigger changes in implied vol correspond with falling prices - ie negative momentum. Of course, although the smirk is mostly stationary intra-day, the smirk moves over time reflecting a myriad of things. The key point here is that the larger changes in implied vol are associated with down markets in daily data, but the relationship can be fuzzier for monthly data. However, having looked at it a lot, I am pretty confident that the dominant data points in a statistical analysis will correspond to large down markets. This isn’t driven by investor activity as much as it is an artifact of the shape of the vol curve (which IS influenced by investor sentiment to a degree).
That’s my point. As prices fall, smirk causes implied vol to rise. A statistical study will correlate increased vol with future falling prices. But, that is equivalent to saying that falling prices (which, via smirk causes the vol increase) predict future falling prices. But that is the same as saying momentum is predictive (which it is). We looked hard at this and the smirk entanglement was enlightening. I’ve refereed a few papers on this topic and it usually boils down to momentum.
You have a point, and I don't disagree with it.
However, I disagree with "the vol smirk is almost always a mechanical artifact of a falling price".
There are instances when the market goes UP (and not down), and the vol skew still increases due to hedging activity.
A recent example is early December 2024, as well as the fall of 2017 and many others.
I hope this time I'm being clearer and my message is coming across.