How do you measure the dispersion of analysts estimates

On one hand, smaller sample would have more possible outliers, which is artefactual for this strategy. On the other hand, smaller number of analysts mean that the stock is less covered and has better chance to surprise. I got to think more about the overall idea

 

I think I have seen sellside research on this. I wanna say Goldman did a study 7-8 years ago.

 

Let me search the portal, that's easy enough (or I can ask my coverage, which is even easier)

 

The paper might have been written when your coverage was in high school. It also might have been citi. They were putting out interesting stuff like this then.

 

The conclusion was higher analyst dispersion led to more volatility. But I remember how strong the signal was and a lot of the data might have been before reg SHO.

 

I have separated the data for the stocks that have earnings within 25 days. Stocks that have more than 10 analysts covering it have a median dispersion of .21 and stocks that have less than 10 analysts covering have a median dispersion of .30 so you are absolutely correct.

 

Was that the implied jump increased going into earnings or the jump was much higher than expected after earnings?

 

This makes alot of sense I have noticed that some of the analysts, have not updated there expectations for months while others have been revised. This could be very time consuming to sort through ugh..

 

I meant to say, I *don't* remember how effective the forecast was.

 

It sounds like a tricky metric to nail down, with stale analysts, number of analysts dependent on market cap etc. Could be a reasonable signal if you have other signals though