Dry spells

Somebody asked about why one would buy Berkshire stock now, there being no certainty that the price will be higher any time soon.
It’s certainly true that Berkshire’s stock price, like any stock price, can go net nowhere for a long time.
If you consider inflation, those stretches are even longer.

But, the value of a share of Berkshire generally goes up over time.
So it’s pretty clear that a very long stretch of no progress in stock price means things are getting cheaper.
A flat price with a rising value means a shrinking multiple.

Since 1998 these are the four longest stretches with no net share price progress.
They all start on a fresh all time high price in real terms, as you might guess.

For the purposes of discussion let’s assume P/B is a decent rough yardstick of value, especially if you use peak-to-date book per share.

The four longest stretches are:
1998-06-19 to 2004-02-09, 5.64 years.
P/B dropped from 2.887 to 1.935 at a rate of -6.8%/year.

2007-12-10 to 2013-05-07, 5.41 years.
P/B dropped from 1.926 to 1.367 at a rate of -6.1%/year.

2004-02-25 to 2006-10-25, 2.66 years.
P/B dropped from 2.032 to 1.610 at a rate of -8.4%/year.

2014-12-18 to 2016-11-09, 1.89 years.
P/B dropped from 1.584 to 1.386 at a rate of -6.8%/year.

A few observations:

It’s probably not a good idea to buy at a high valuation multiple : )

Other than those four, you never had to wait more than a year for a fresh all time high price in real terms, which rather surprised me.

One interesting observation is that the rate of multiple compression during the long stretches was so similar.
I doubt it means that much, but it’s interesting.

The current valuation multiple of 1.27 is lower than the lowest ending valuation multiple in any of the stretches of cheapening above.
Thus one might speculate that a long net-nowhere stretch starting here would be extremely unlikely.
Too much of the maximum possible valuation compression has already happened.
The price could certainly drop a long way from here (or not), but it’s very likely to be higher than this at some point relatively soonish.



But have we also had a time with such market over valuation coming off a low interest rate base, high inflation and strongly rising rates over the medium term?

1 Like

Really interesting stuff. Another way to slice the data:

Since 1998, a randomly timed purchase had about a coin flip’s chance of being surpassed within a year: (5.64-1)+(5.41-1)+(2.66-1)+(1.89-1) = 11.6 years in which you’d wait a year or more.

If P/PeakB remains a yardstick of value, I suppose the key question is how sure we are that the central value, measured as P/B, will not continue to significantly decline as the years pass. For instance, a purchaser on December 10, 2007 may well have looked at the stretch from 1998 through 2004 and thought they were buying at a decent enough valuation multiple, given that it was below the ending valuation multiple of a pretty long stretch of no recent net highs, the 98 - 04 stretch (albeit meaningfully above the ending point on the next, shorter stretch, from 04 - 06). That would have been a wrong assumption by a fair amount, it turns out.

This is because, in recent history, the median valuation multiple has now tended to hover around 1.35 – i.e., a mark that is also below the lowest ending valuation multiple in any of the stretches identified.


I have actually been looking at something similar. I like to follow the quartiles of Price/Peak Book so I wondered how those quartiles varied during times reported book value was not a peak value.

All days from 1996 on, reported book has not been a new peak value for 32% of all trading days. I found this rather surprising as it seemed higher than I would have expected.

All days from 2012 on, reported book has not been a new peak value for 12% of all trading days. So that seems to explain why the all days number surprised me with recent history being quite different. Who knows what the future holds.

What I found interesting is if you look at the 2012 forward dates and segment out only the days reported book is not a peak value, the median price to peak bv during these days only drops to 1.32x (roughly equal to the bottom quartile of all days 2012 forward).

So let’s compare the 2012 forward Price to Peak BV for “All Days” and “Non-Peak” quartiles:

Min All Days - 0.93x
Min Non-Peak - 0.97x

Q1 All Days - 0.93x to 1.32x
Q1 Non-Peak - 0.97x to 1.12x

Q2 All Days - 1.32x to 1.37x
Q2 Non-Peak - 1.12x to 1.32x

Q3 All Days - 1.37x to 1.42x
Q3 Non-Peak - 1.32x to 1.36x

Q4 All Days - 1.42x to 1.59x
Q4 Non-Peak - 1.36x to 1.43x

So it seems if you are purchasing at valuations in the bottom quartile of peak price to BV (1.32x and ~$304 today) then you can reasonably expect to have an opportunity to sell at a profit about 50% of trading days even when book value is not setting new highs.

So to Jim’s point, it’s not a good idea to buy at a high multiple! Exercise patience, purchase at bottom quartile multiples, and stack the deck in your favor for success even in challenging environments.