3
$\begingroup$

I'm a quant working in a mainly fundamental shop. Analysts are asked to score things like management or industry trends of stocks in their "watchlist", and I am now trying to weave the results into a factor model. The model includes other factors such as earnings yield and some financial quality metrics.

The scores are simple 1 to 5 scores, where 1 is very good and 5 is very bad. A lot of thought is put into them, and they derive from team consensus (and hours of arguments).

The problem is that they are heavily skewed towards the better end of the spectrum. Stocks which would score a 5 would be considered so bad we wouldn't bother watching them (and thus are not even in our watchlist). Stocks which score a 4 are rare: only a couple of stocks, one of which has had substantial volatility.

Which statistical technique best deals with the skew in score distributions?

When I compute factor premiums with OLS, I get a very low R-squared.

$\endgroup$
6
  • $\begingroup$ given your universe, what are you filling in for missing values of the score? $\endgroup$ – phubaba Aug 21 '17 at 19:45
  • $\begingroup$ I restrict my universe to the watchlist. About a hundred stocks with score history going back two years. $\endgroup$ – lebelinoz Aug 21 '17 at 19:47
  • 1
    $\begingroup$ So in some sense, the data generating process for these "financial quality metrics" involves censoring where you have a higher probability of observing the metrics if they are good? (Because otherwise, the team wouldn't even be issuing scores.) $\endgroup$ – Matthew Gunn Aug 21 '17 at 20:39
  • $\begingroup$ Sadly, yes. Thanks for that link, it will help me explain the fallacy to my bosses. In the meantime, are these scores salvageable in a quant model? $\endgroup$ – lebelinoz Aug 21 '17 at 20:50
  • $\begingroup$ @MatthewGunn I should add that our censored watchlist is only a starting point: to create a portfolio of stocks, we choose 30 to 40 stocks from the watchlist. I'm trying to understand why our actual portfolio is underperforming the "benchmark" consisting an equal-weighted watchlist $\endgroup$ – lebelinoz Aug 21 '17 at 22:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.