Hopefully the main point didn’t get lost in the other interesting tidbits:
Suppose we want to do something like Jim’s analysis, but use IV and not Book, in the hopes of maybe even getting a tighter relation.
That’s in a different thread, so to rephrase the issue in terms of @rrr12345 graph of "Price/Trendline Oct 1964-Jun2022):
when you’re way below 1 (undervalued) do you tend to get higher forward average returns and when you’re way above do you tend to get lower?
My understanding of @rrr12345 work is that IV on a date would be the TrendlinePrice for that date in a plot of log(Price ) vs log(Book).
Is the relation between IV and future return tighter than Book and forward return?
But for the analysis suggested above, you can’t take IV from the trend line determined from all the data i.e.including data in the future of the date for which you’re determining IV, because that requires a crystal ball. You need to calculate the trendline using all data to the past of each date of interest. There’d be a ‘burn-in’ period to establish a gross trend, and then you start calculating trendlines and hence IV for each date after the burn-in period in a walk-forward process (lots of trendlines).
Having had my attention drawn to trend lines by the above thought, I then started noticing these breaks I referenced in the trendlines, even before doing much of any walkforward analysis. And things would get even more complicated when you start determining trendlines in a walk forward manner.
So I thought I’d post it and see what people might think.
I agree, it’s probably no co-incidence that BRK and the SP500 share a similar break.