Comments on: Don’t Get Trapped by a Crowd in Stocks

By: Kamuran

Kamuran — Thu, 20 Dec 2012 10:35:54 +0000

Hi thanks a lot for a discerning post, I really found your blog by mistake while searching on Google for something else closely related. In any case before I ramble on too much i would just like to say how much I loved your post, I have bookmarked your site and also taken your RSS feed, Once more thank you very much for the blog post keep up the good work.

By: Vadim Zlotnikov

Vadim Zlotnikov — Wed, 05 Dec 2012 12:34:10 +0000

Yes. This is what I meant by the development of “option or volatility” strategies. I think your suggestion of focusing on the left tail is exactly right.

By: Joseph Hogue

Joseph Hogue — Tue, 04 Dec 2012 23:42:51 +0000

While the ex-post volatility is higher, presumably after the trade has fallen apart and everyone rushed to the exit, would the ex-ante volatility be lower than normal due to unidirectional complacency? I would think this would lend itself well to buying puts as portfolio insurance.

By: Vadim Zlotnikov

Vadim Zlotnikov — Tue, 04 Dec 2012 12:49:06 +0000

You make a very good point. The nature of crowding is loosely tied to the business cycle, with cyclicals becoming crowded as we approach a cyclical peak. There is also some mild correlation between 12-month price momentum (or relative strength) and the degree of crowding. Likewise, there is some relationship between valuation and crowding. However, the real value of a crowding signal is less “directional” and more tied to specifics. In other words, which of the in-favor cyclicals or high momentum stocks/industries are most crowded. Furthermore, I would not use a crowding signal as a source of alpha, which would require one to correctly time the outflows from crowded trades (it is doable using persistence measure of order imbalances, but it tends to reverse quickly). I think the real value of understanding specific crowded positions comes down to risk management. In a crowded trade, the ex-post volatility and negative skew is much higher than would be indicated by the historical relationships. Crowding is not a measure of future return, but always a measure of future risk. Hence, I would recommend sizing the exposure of a crowded trade at a lower level than a similarly attractive uncrowded trade. One can also develop volatility or option trades around the notion of crowding.

By: Joseph Hogue

Joseph Hogue — Tue, 04 Dec 2012 04:00:03 +0000

I like the idea and agree with the need for a more defined qualifier for “crowed” but it seems the research results in the typical distinction between cyclical and defensive sectors. I wonder what the last graph would look like around the market peak. Would it be reversed with cyclicals being ”overcrowded”? In this case it boils down to the billion dollar question of calling the market trough and peak, which has been tried a few times.