Below some buyback levels by WEB

If you're interested in keeping an eye on Mr Buffett's buyback strategies, here's a geeky number for you.

A fit for how many A-share equivalents are being bought per working day as calculated by Excel:
=MAX(0, -1700 * PB + 2540)

where PB is ratio of average price P/B in the date range, using the highest-to-date known book per share on each day.

It's not a reliable formula, but it gives a quick idea.

It implicitly gives no buybacks above 1.5 times book.
Using the average P/B in a reported date interval, to generalize just a hair, there have been buybacks above 1.49 times peak-to-date book, but none above 1.5, so something like 1.5 seems to be the "too expensive" cutoff.

The buybacks in Jan and Feb 2022 were outliers in the sense of lots of buybacks at higher average multiples in the period. That could be a bit of a quirk of the data though: the price moved a huge amount in March, and the buybacks were loaded to the low end at the first part of the month. The average price paid was 4.5% lower than the average market price in the month for both A and B shares.

Jim