Crystal ball prices

[mungofitch]

It has been an interesting ride so far this year–high prices, low prices, all kinds of prices in between.

I have a model which attempts to predict the rate of return from Berkshire’s stock in the next three years based on valuation levels, and the recent rate of change of value.
Unsurprisingly, the stock gets a slightly better market multiple when values have been rising strongly lately.
Technically it’s a model to predict the average real market price 1-5 years later, but that’s pretty much the same as a three year prediction.

As the price-to-value ratio changes, the prediction changes.
When the price was pretty full 2022-03-28, the prediction was that the next three years would get you inflation + 5.41%.
When the price was cheap 2022-06-23, the prediction was that the next three years would get you inflation + 11.06%.
With the recent partial rebound, the current prediction is three years of inflation + 9.08%/year.
That works out to an expected price “around” three years from now of $587,055…measured in today’s dollars.

The model isn’t to bad, if all you want is a prudently sensible expectation rather than actual true predictions.
For a spot check:
Three years ago it predicted a price now of $395,397 measured in August 2019 dollars.
That equates to a price now of $457,055 in today’s dollars.
The actual price is $452,697, so the prediction was too high by just under 1%.
It’s not usually that good : )

Another time it was pretty good:
Its highest ever prediction was at the market bottom in 2009.
It predicted three years of inflation + 17.22%/year.
The actual result was inflation + 19.8%/year.
Give the chaos at t he time, I’d count that a big win.

During the pandemic bear the three year forecast peaked at inflation + 15.49%/year.
We still have to wait a while to see how well it did.
So far we have 2.4 years at inflation + 22.4%/year.

None of this should be taken too seriously, as it includes two big assumptions which are far from sure things:

• That future valuation multiples will roughly resemble past ones, and
• That future growth in value will be roughly similar to past rates

Still, if you want a guess, it’s probably better than reading Reddit.

Jim

[rrr12345]

"it includes two big assumptions which are far from sure things:

• That future valuation multiples will roughly resemble past ones, and
• That future growth in value will be roughly similar to past rates"

Reasonable assumptions, supported by history.

*P/B reversion: Historically P/B has tended to revert to the mean in about three years, more or less. I usually assume ten years to revert, as it adds less to the estimated return. If P/B reverts from today’s ratio of 1.44 to a mean of 1.48 in three years, it adds 0.9 percentage points to the annualized return; if P/B reverts from today’s ratio of 1.44 to a mean of 1.48 in ten years, it adds 0.3 percentage points to the annualized return. Call it 0.5%.

*Value growth rate: Since mid 1998 BV per share has risen 10.5%/year, least squares, with an r^2 of 0.985. My estimated IV/share has also risen 10.5%/year, nominal, with an r^2 of 0.985 (not surprising since my IV estimate is based on BV).

Although I’m sure my analysis is much cruder than your analysis, I might therefore estimate a nominal, annualized total return of 10.5% + 0.5% = 11%, but then lower it a bit from that figure by the loss of Warren Buffett as CEO and chief investment officer. This estimate seems reasonable to me since Warren has described Berkshire’s operating companies as “slighter better” than the S&P, and since historically Berkshire’s stock portfolio has outperformed the S&P by a few percentage points, and also since Berkshire uses leverage in the form of float. Further support for this estimate is that BRK.A stock has historically outperformed the S&P by a few percentage points, more when the return of the S&P is low, as it might be expected to be over the next ten years starting from today’s market P/E. Just FWIW.

Thank you for sharing your forecast, Jim. Your forecast is higher than mine, but I’m sure that your analysis is more detailed.

[palmersq]

It will be a regime change when Buffett passes. I just can’t tell which direction - a relief rally upon the eventual passing of the old guard vs a lower-confidence vote on the new guard. Old model probably gets little predictive power around or even pass the juncture. It’s not going to be business as normal.