Post New

Predicting stock returns from charts is not usually a good idea.

But I have this theory that it depends quite a bit what the chart shows : )

Have a look at this little fella
http://www.stonewellfunds.com/SmoothedRealValue.pn...

tl;dr
The blue line is calculated without any hindsight, using only numbers available up to the date shown, and the distance of the current market price from that line is pretty predictive of forward returns. So you can get an idea of what the price might do in the next year or two, with a useful level of accuracy.

Discussion---

The graph shows grey spots showing the average real price in each calendar quarter, real book per share, and a line which is a smoothing of two different value estimates. All three are scaled arbitrarily to line up--for example, the book line is 1.37 times real book. The smoothing line looks pretty good to me: it adapts to slow spots so it's realistic, but it recognizes that over the long run value is really a pretty smooth trajectory.


A long time ago, I noted that not only is the price a volatile measure of what's going on, but even book per share is pretty volatile sometimes. So, I suggested a smoothing method for your value metric of Berkshire, intended to be just smooth enough that dips in apparent value during recessions don't show up as dips in the smoothed line representing your estimate of fair value. The idea here is that dips in apparent value are always transient, because Berkshire makes money almost every day, and losses are always made up again. This may not be 100% true, but we're Berkshire people, so let's go with it.

This was the original post on the idea, intended as a smooth metric to feed a "safe withdrawal rate" table for Berkshire stock.
http://www.datahelper.com/mi/search.phtml?nofool=y...
This is a follow up 5 years later, discussing the specific smoothing method I'm using here.
http://www.datahelper.com/mi/search.phtml?nofool=y...

The obvious value metric to use is real book per share. Or, you can use a different metric, like my "two and a half column" method. The graph here uses the simple average of the two.

I also tweak book a bit: before even doing the smoothing, I remove any drops of more than -4%. The idea is that book per share might be a slight overestimate at a cyclical top of exuberance, but not by much, so I limit any subsequent drop to 96% of the peak-to-date.

Then I smooth the numbers, using 16 quarters of value estimates, weighting them linearly. The most recent number gets a weight of 16, the second-most-recent gets a weight of 15, all the way down to a weight of 1 for the data point 16 quarters ago. This reacts to new data, but not too much. Big sudden changes a long time ago are de-emphasized.


Now the fun part: is the ratio of current price to the current point on the smoothing line predictive of forward returns? In the past, definitely. The specific exercise I did this morning looked at the ratio of price to smoothing line, and average real forward returns over the next two years. Berkshire's valuation levels have been much lower since 2008, so that makes the best predictor. It suggests that the next two years should be expected to be inflation - 1.4%/year. Using all the data from 2005, it suggests one might expect inflation - 1.5%/year. Using data starting in January 2000 (which includes some periods of higher valuation multiples) it suggests you might expect inflation + 1.8%/year.

The future may not resemble the past. And my valuation metrics, or their smoothing, may be bogus. But if (if) things are somewhat like the past, we aren't headed into a strong year or two for the stock price of Berkshire.

In case anybody cares to follow up, I used stock price $424.44 per B share, and the forecast ending price is the real average closing price 1.5 to 2.5 years from now measured in today's dollars. The new CPI figure just came out, 314.175.

As always, the forecast will be wrong, but the notion is that it's a 50/50 shot whether it will be too high or too low.

FWIW, here are the observations: dividing up the price-to-smoothed-value ratios January 2008 to January 2022 into 20 equally sized buckets of starting days, what were the average annualized real forward returns in the next two years in each bucket? With holding period endpoints 1.5 to 2.5 years later.

  19.3% (Cheapest 5% of starting dates)
  17.7%
  16.8%
  16.2%
  15.1%
  13.3%
  12.0%
  11.2%
  10.4%
   9.5%
   8.7%
   7.9%
   7.3%
   6.8%
   6.1%
   5.2%
   4.0%
   2.4%
   0.0%
  -7.7%  (Most expensive 5% of the time)

Jim

Jim

Thanks for sharing, a very interesting chart. A question. The slope of the smoothed line appears to be about the same as the slope of the smoothed line in 2008/2009. Maybe not quite the same but close. Do you have thoughts on why the growth is similar to a time period when there was a major financial crisis? Is it predictive of things to come, or a temporary slow down in growth? Or something else??

Aussi

The slope of the smoothed line appears to be about the same as the slope of the smoothed line in 2008/2009. Maybe not quite the same but close. Do you have thoughts on why the growth is similar to a time period when there was a major financial crisis?

To oversimplify just a tad:
When you have a stretch during which there is no new high in real book per share, the smoothed line will gradually start bending down towards the horizontal. The slope at any point during that time period will be relatively constant as a function of how long it has been since a fresh all time high figure, without very much regard to what book per share is actually doing during that period.

That's by design. The following assumptions are baked in:
(a) real book per share is an adequate yardstick of value. (actually the smoothing uses two different methods, but book is good enough)
(b) current real book per share is never a big exaggeration: real value is never more than 4% below its all time peak, no matter how far it drops nor for how long.

So, the similarity between the slopes in the two stretches is simply because both stretches had several quarters of seeing no fresh all time high in real observable value per share. The average keeps using "96% of peak to date" for each new quarter until a fresh high happens.

Do you have thoughts on why the growth is similar to a time period when there was a major financial crisis? Is it predictive of things to come, or a temporary slow down in growth? Or something else??

This method is an attempt to follow what has been observed, with just enough smoothing to make it smooth. It isn't a fixed slope and it has no forecast, it's just a smoothing of the recent data points. If growth slows over time, the line will go more horizontal. It could go completely horizontal if there is never again any progress in observable value. There is no attempt to estimate what will happen over the long term, it will just follow the numbers.

Of course, I do have my own expectations, unrelated to this data series. I assume the future slope will be a little gentler than the past slope. Berkshire was really doing better than expected for a few years there. But I also assume that it will continue to rise in value at a reasonable rate: I have no reason to think it will ever be worse over the long run than the "monkey with a dartboard" rate of return. I expect to see a future slope of around inflation plus 6-7%.

Jim

Nice analysis. Thanks for sharing. We see clearly that forward return depends strongly on the beginning price-to-smoothed-value ratio.

Berkshire's valuation levels have been much lower since 2008, so that makes the best predictor. It suggests that the next two years should be expected to be inflation - 1.4%/year...

I did a one-year forward model the same way. It suggests one should expect inflation - 9.6% in the next year. Ouch. That is based on the assumption taht valuation multiples will average around what they've averaged since 2008, which we might think of as "low", but it's not a terrible assumption.

Since I've gone out on a limb to share what those models say, I might as well go much further out there and forecast a complete price trajectory for the next two years. This is done totally by eyeball, but it passes through those two forecast data points. I have pencilled in inflation of 2.83%/year in the next two years (MCT inflation was 2.36% in the year to May). The result seemed extremely pessimistic, so I also did it again with just a pinch less dire fall-off, giving an expected average price range for each period.

Nominal dollars assuming constant low inflation, not today's dollars.

2024-07 $424
2024-10 $413
2025-01 $403 - 406
2025-04 $395 - 402
2025-07 $395 - 404
2025-10 $403 - 410
2026-01 $414 - 419
2026-04 $425 - 427
2026-07 $436

This should be good for a belly laugh over the next couple of years. It does suggest that there will probably still be lots of opportunities to buy at prices under $400 per B share, for those who like their veggieburgers on sale.

For whatever it's worth, this is what the equivalent real values might look like with inflation at 4.25%/year.
2024-07 $424
2024-10 $414
2025-01 $405 - 408
2025-04 $399 - 406
2025-07 $400 - 410
2025-10 $410 - 417
2026-01 $423 - 427
2026-04 $436 - 438
2026-07 $448

Jim

I have used this price pop to raise my cash allocation.

Getting ready for the apocalypse!😬

I have used this price pop to raise my cash allocation.
Getting ready for the apocalypse!

Well, the nice thing is that cash isn't so bad even if there is no apocalypse. Good investment opportunities come around pretty regularly.

Jim

I've been running my BRK at leverage = 2 with DITM LEAPs. Is it time to go back to stock and sell covered calls every week?

Jim, I am remembering your prediction for the next book value being a leap of 6% from the previous quarter. That is not going to help BRK to "remember" it has to slow down, is it?

R:)

I've been running my BRK at leverage = 2 with DITM LEAPs. Is it time to go back to stock and sell covered calls every week?

Well, there are lots of moving pieces. I don't know what else you own or the size of your cash pile, and I probably prefer not to.

But if that leverage was something put on at a time that valuations were unusually good as a way to profit maximally from the upswing, then this is decidedly not such a time, so I guess taking the leverage off makes sense. If it is just the current method you are using to hold upside to a specific number of shares over the long term for whatever reason, the answer might potentially be different. But if I had to guess your situation, taking the leverage off is more likely to be a good idea than a bad one.

Writing covered calls against that is perhaps gilding the lily. It might work and add value, but it's a more aggressive decision. As mentioned in another thread, most people find it hard to watch that lose money (even through their stock is making more money at the same time). So I'd be hesitant to recommend that.

The net of my long and short positions is the smallest net long I've had in a very long time. (though that's hard to calculate, as it depends on how much, if any, of my stock is called away from the calls I've written)

Jim, I am remembering your prediction for the next book value being a leap of 6% from the previous quarter. That is not going to help BRK to "remember" it has to slow down, is it?

Honestly I have little idea how the market will react to the Q2 book figures. Markets can be pretty perverse. It seems pretty certain that book will look good, largely due to the very large increase in the market price of Apple shares in Q2. But yes, in general, the market reacts to current book value more than it does to "on trend" value, both in unusually high and unusually low quarters, so even if Q2 book is "unusually" high, that may not be taken into account.

Jim

"I've been running my BRK at leverage = 2 with DITM LEAPs. Is it time to go back to stock and sell covered calls every week?"

Well, there are lots of moving pieces...
But if that leverage was something put on at a time that valuations were unusually good as a way to profit maximally from the upswing, then this is decidedly not such a time, so I guess taking the leverage off makes sense.

Ken Fisher likes to call the market "The Great Humiliator -- its aim is to humiliate as many people as it can for as long as it can for as many dollars as it can."

Too true!

I closed out my BRK leaps earlier this year, when the P/B got to 1.6. At 383.

I recently sold a 425 covered call after BRK went up big several days in a row. "Surely it is being over-bought and will fall back."

BRK-B is currently 438. Blew right past the CC breakeven point.

Me: Humiliated both ways.
Market: Success!!

In addition to the historic step change in BRK’s valuation in 1998…

… the year 2018 saw another momentous event in the price discovery scheme. That is when Buffett announced the “I’ll buy when I feel like it’s a good value” regime. It replaced the 1.1x and later 1.2x BV regime. That was in place for 6 years and the market never offered them a chance to buy any share; the bump up to 1.2x happened to accommodate the single private transaction with the Ueltschi estate. To underline the regime change to “watch us”, Buffett simultaneously dropped reporting the BV multiple column in the Letter.

The upshot of this regime change in 2018 was that Mr. Market had to look in the mirror to see “what they bought”. And of course Berkshire being able to buy $70 Billion and counting since.

For me, nothing underscored the regime change more than Greg Warren’s repeated questions to Buffett on the buyback mechanism; one prominent one was “Do you use the yet-to-be-announced intra-quarter book value to buyback?”. Besides the oxymoronic nature of that question (Berkshire doesn’t waste time or energy on such financial gimmickry), Buffett said “no” to Gregg but that didn’t stop him reporting that Buffett confirmed that they did. No matter what he or Mr. Market thinks, this was a master stroke from Buffett.

So, we’ll wait until November 15 to see “what they did”.

So, we’ll wait until November 15 to see “what they did”.

Indeed. But still, I doubt (and hope) they are not buying today.

There does seem to be a correlation between valuation multiples and quantity of shares repurchased. It's very rough and ready, but if you fit a straight line through the observations and squint a bit, you get a "no buybacks" point around 1.5 times book.

Jim

<<<There does seem to be a correlation between valuation multiples and quantity of shares repurchased. It's very rough and ready, but if you fit a straight line through the observations and squint a bit, you get a "no buybacks" point around 1.5 times book.
>>>

Agreed. It does show that the willingness threshold has moved up from 1.2 to 1.5x.

As always, Mr. Buffett thinks decades out and this new regime was no reactionary move in 2018. It’s meant for the ages, long past his actuarial or real life expectancy.

We’re in great hands

Just a bit of a follow up on the notion of this smoothed value line being predictive of where the stock price might go next---

Since the line has so darned few squiggles, you can extrapolate it a tiny bit. As a result, you get an extremely good idea of where the line would be today, even though this quarter's figures aren't built into the calculation yet. Nor even last quarter's, at the moment.

Since we know the level of the line at each date in the past, and we know the level of the price at each date in the past, we know the average ratio between the two. Looking at the average ratio since 2005 and applying it to today's level of that line, the price today "should be" $370.22 per B share in today's dollars. Thus, today's price of $441.26 is about 19% above what you would expect at the usual multiple since then. If the future value trends and future valuation multiples are not very much different from those typical of the past, we have already received the benefit of price rises we might reasonably have expected over the next couple of years. Stated in a much simpler way, the price has gone up more than the value has lately.

We can also look at the past to see what usually happened next. Historically this distance above the "norm" was followed on average by meaningfully negative real returns in the next year (-13% ish), and slightly negative in the next two years (-3%/year ish). Though of course the future may be different: this "what happened before" view has the flaw of potentially overweighting the observation that following the few periods of similarly high valuation multiples the stock just happened to go on to become undervalued, not merely averagely-valued, a year or so later. It would make more sense to expect an average valuation multiple as your base line assumption, not a bad one.

These are not predictions per se, just following a line of reasoning. IF certain trends and relationships are similar, THEN that's what the implication is for the stock price.

Jim

“Historically this distance above the "norm" was followed on average by meaningfully negative real returns in the next year (-13% ish), and slightly negative in the next two years (-3%/year ish).”

Thank you for sharing Jim. This is meaningful data for those of us that live off investments. For now, at least, I am still finding it is easier to trim other names than BRK to pay for life expenses but something to noodle for sure as the price rises disproportionately.

Mungofitch:
We can also look at the past to see what usually happened next. Historically this distance above the "norm" was followed on average by meaningfully negative real returns in the next year (-13% ish), and slightly negative in the next two years (-3%/year ish). Though of course the future may be different: this "what happened before" view has the flaw of potentially overweighting the observation that following the few periods of similarly high valuation multiples the stock just happened to go on to become undervalued, not merely averagely-valued, a year or so later. It would make more sense to expect an average valuation multiple as your base line assumption, not a bad one.

May I ask what you will be doing about this? Do you keep a core holding of BRK even believing the stock price is that much too high?

On the one hand, I often wish I had a retirment port which was 1/3 SPY, 1/3 BRKB and 1/3 MRFOX, and that I didn't trade them and just lived with the 15-20% return. On the other hand it seems nuts to hang on to something that "should" go down for two years. On the third hand, I'm still glad I didn't dump out of my BRK yet since it seems to still be creeping up a bit. On the fourth hand, I am quickly running out of hands.

R:)

We can also look at the past to see what usually happened next. Historically this distance above the "norm" was followed on average by meaningfully negative real returns in the next year
...
May I ask what you will be doing about this? Do you keep a core holding of BRK even believing the stock price is that much too high?

Usually I do, but not this time. I have almost no net position in Berkshire at the moment.

I'm cash heavy because I have some cash from something else from earlier in the year that I haven't deployed yet, which is unrelated. But I have a negligible amount of "net" Berkshire position as I've written high strike calls against essentially all of my remaining long positions, having sold the rest of my long positions. The last big "net long" block was dumped July 18 with B shares at $448.79 (check the chart, I'm quite proud of the luck of that timing!)

Oddly, since most of my current long term long position is held as deep in the money calls, I am now in a "bull call spread": long and short call options on the same thing but at different strike prices. Losing money slowly on time erosion of the low-strike calls, and making money rather more quickly on the time erosion of the high strike calls--if it stays in a range I make a little money every day. But, other than that, it's close to a neutral position, depending on what the stock price does and when: I might have no position at all soon if the stock price goes even higher and the calls I've written are exercised, or I'll keep the portion I haven't sold if the stock price retreats.

I have (just) enough confidence in my valuation methodology to believe that there isn't a decent chance of making a decent return in the next year, so I'll wait for a more attractive valuation before pouncing. Ideally Berkshire, maybe something else. There's always another sale sooner or later. And I don't pay capital gains tax so lightening up or switching positions doesn't have that drag. And, unusually, you can get meaningfully positive real interest rates on cash at the moment. I've done very well in the first half of the year (we've all seen that Berkshire is up 23.3% year to date), so even a little bit of interest income will be more than enough to satisfy me for a while.

Jim

PS
You might wonder why I still have any deep in the money calls rather than plain stock, since calls make most sense only when you want leverage because the near term outlook is particularly good--which is isn't. The reason is that the amount of time value that I could realize on selling them (near the bid price, far from the ask price) is not great, and corresponds to a low interest rate on the "borrow". If I sold them and switched that same size of stock allocation to plain stock I would be saving costs from the eroding time value in the calls, but eating into a cash pile that is currently earning a higher rate sitting in T-bills, so I'd be worse off. If I could trade at the midpoint of bid and ask it would make sense, but I can't.

“You might wonder why I still have any deep in the money calls rather than plain stock, since calls make most sense only when you want leverage because the near term outlook is particularly good--which is isn't.”

Thanks for your full explanation. I’ve got some LTCG DITM calls Jan. 2025 (150 & 160 Strikes) & am strongly considering selling them relatively soon & locking the gains as they’ve served us well, cautious outlook over next 6 months & we have no plans or need to exchange those funds for any more stock at present.

Post New

Berkshire Hathaway FAQ

---

Contact Shrewd'm
Contact the developer of these message boards.