The Window Undressing

I’m not sure when I first started reading about “window dressing” in the market. It may have been while reading Jim Cramer in the 90’s where he talked about the things managers would do to keep themselves from being fired. The basic idea is that managers would adjust their portfolios at the end of a fiscal period so it appeared they were in the best stocks and out of the worst. This way they wouldn’t lose client assets whose fees are the lifeblood of their existence. Never mind the possibility that these investors might actually check the quarterly returns and use those numbers in their decision making process…

So for example at the end of the most recent quarter a manager would liquidate any stocks that were in the portfolio that were big losers— i.e. any of the coal stocks; and then might add big winners i.e. any of the solar names that have been rock-n-rollin since last year. Then their sheep, I mean clients, would magically continue to invest with them, despite the fact that their returns were severely negative.

As a trader the only thing that matters to me regarding a certain type of market participant behavior is whether I can find a trading edge to exploit it. How would I go about doing this with respect to window dressing? I would need to model this manager behavior and see if it produced a tradeable edge in the market. There are a myriad of scenarios that one could construct and back test to look for a possible trading edge.

One of our quants tested how the 10 best and worst performers in the S&P 500 behaved in the final week of each quarter. He found something rather curious. The strongest stocks traded lower and the weakest moved higher on average. This is exactly the opposite of what one would expect under a “window dressing” scenario. Here are the results for shorting the 10 strongest and getting long the 10 weakest:

Here are the quarterly window dressing results (SPX results from last week of quarterly months 2010- buy 10 weakest and sell short 10 strongest sorted by trailing 89-day percentage change)
Total 7.1%
Avg 0.48%
Max 7.9%
Min -2.3%
Std 3%
t 0.5

Again the results run counter to expectations of people “window dressing” by buying hot stocks so that they will show up in the quarterly holdings. (results available on request)

The results here are also too volatile to trade.

This is just one scenario which in no way eliminates the possibility that some managers engage in “window dressing”. But again whether some do or some don’t is meaningless to me as a trader. I want to identify tradeable edges. The rest is for cocktail parties and CNBC.