I’ll add my 2 cents. I would not surprise me one bit if investors fleeing the market would choose something to chase temporarily. Energy? Berkshire? In a market sell-off of course it isn’t going to sustain but these things do happen.
Life is great if you can stand it. Remember if you sell out you’ll one day need to get back in. And when you get back in, given most here (but not all div 20!) are value investors and we tend to get back in way too early!
HA!
Great Finaancial Crisis 2008-09 BRK 50% S&P 52% identical basically.
And in terms of timescales.
It took 5 yrs in the GFC for BRK to get back to the previous peak 08-13 and 11 months in 2020.
Decline took 1 year top to bottom in the GFC and 1 month in 2020.
I’d expect a situation more like the GFC this time rather than 2020 given what Grantham says.
In both the 2008-2009 and 2020 crashes, Berkshire dropped to near book value. Depending on the severity of the next bear, that may be the bottom for Berkshire.
Yes, book value is less meaningful than it used to be due to the share buybacks and the carrying cost of assets (which are not marked to market in the calculation of book value), both of which reduce book value, compared to intrinsic value. However, if the lowering impact of these factors roughly equals the overvaluation of stocks such as Apple, then these factors may cancel out and book value would still be meaningful. Jim’s data confirm that to be the case.
"It took 5 yrs in the GFC for BRK to get back to the previous peak 08-13 and 11 months in 2020.
Decline took 1 year top to bottom in the GFC and 1 month in 2020.
I’d expect a situation more like the GFC this time rather than 2020 given what Grantham says."
That’s correct but BRK was overvalued going in to the GFC (if I recall correctly, Jim sold shares in late 2007 as the price to book was around 1.8). In contrast, it was slightly undervalued going in to the 2020 crash. That may partly explain the time it took to get back to the previous peak in the two cases.
This time, BRK is not meaningfully overvalued, which may reduce the time for it to get back to the earlier peak, compared to during the GFC.
It took 5 yrs in the GFC for BRK to get back to the previous peak 08-13 and 11 months in 2020…
That’s a good number, but it is a little dependent on what valuations were like before the plunge.
The price did spike in the early days of the crisis 2007, for example, trading briefly over 1.9 times book.
If a spike sets the high you’re waiting to hit again, a longer wait will be likely.
A alternative metric might be “how long did it stay below a reasonable valuation multiple?”
In the modern era, 1.4 times “peak to date” book per share has been about par for the course.
If that’s your test, it stayed “too cheap” for 2.1 years.
If you figured 1.45 constituted the end of the “too cheap” stretch, it was a bear of 3.28 years.
If you needed to see 1.5 to consider it a decent multiple, the dark ages lasted a full 6 years.
(The second longest stretch ever waiting for the 1.5 hurdle ended just this month, at 3.24 years).
Jim
book value is less meaningful than it used to be due to the share buybacks and the carrying cost of
assets (which are not marked to market in the calculation of book value), both of which reduce book value,
compared to intrinsic value. However, if the lowering impact of these factors roughly equals the overvaluation
of stocks such as Apple, then these factors may cancel out and book value would still be meaningful.
Another factor is how long it has been since a big acquisition.
Money spent internally on capex generally has an intrinsic value worth a multiple >1 of the cash spent.
That’s why the average company is worth a significant positive multiple of book.
Money spent on shares of a company (public or private) are worth about what was paid for the first while.
That’s why funds aren’t worth more than book (NAV).
Jim
I would say that the selloff is not just imminent; it’s begun. The S&P 500 is 9% below its peak; the NASDAQ is 16% below its peak; the Russell 2000 is 21% below its peak.
It sounds like most people on this board plan to ride BRK and the market down if they fall significantly, and then ride them back up. That’s not a bad approach, depending on one’s investment timeframe. Averaging down is also not a bad approach. Few people here, it appears, have significantly reduced their asset allocation to equities, and that’s OK, too. Timing the market is difficult, and paying capital gains tax is costly. Grantham, in his personal account, has rotated into less overvalued segments of the market, specifically emerging market value stocks, and he has shorted the NASDAQ and the Russell 2000. GMO’s approach has been to remain fully invested, but to “overlay a long-short portfolio” equal to about 30% of the base portfolio. They say that their research has show this to reduce the drawdown. All of these approaches are reasonable, but some will have better 5-year returns than others. What do you think the best approach is?
FYI: I sold some BRK and purchased some AMZN and DIS (just a small amount however).
tecmo
…
book value is less meaningful than it used to be due to the share buybacks and the carrying cost of
assets (which are not marked to market in the calculation of book value), both of which reduce book value,
compared to intrinsic value. However, if the lowering impact of these factors roughly equals the overvaluation
of stocks such as Apple, then these factors may cancel out and book value would still be meaningful.
=====
Another factor is how long it has been since a big acquisition.
Ok, and another big factor is that Berkshire is itself a big holders of securities, and these would presumably fall during a sell-off, while Berkshire’s official book value (published every 3 months) remains the same.
Berkshire’s operating company side should not suffer much in a general sell-off - it’s already at fair value, maybe about 2x book.
Berkshire’s stock holdings, however, constitute almost half the value of Berkshire, and presumably are worth about their market value, or slightly less, given the future capital gains taxes owed on these. Almost half of Berkshire’s value is in this collection of stocks, half of this being Apple but the other half is arguably overvalued too. If the stock holdings sold off by 40%, but this had not yet been updated in a quarterly report, Berkshire’s share price might go to very low multiples of the most recently published book value, even if the multiples (1 and 2) on these two parts of the company had not changed.
If someone is following Berkshire and looking for a good price to buy back in, I think it makes a lot of sense to separate out these two components. Mr Market may be schizophrenic, but he’s not stupid.
dtb
The amazing thing to me is how rigidly investment managers of foundations, target funds and the like have stuck to their strategic asset allocations during the stock and bond bubbles, earning nothing in bonds and risking significant loss in US equities. One foundation that I talked to said that the Uniform Prudent Investment Act (UPIA) required them to use the asset allocation that they were using. This is somewhat true. The UPIA does require that managers use Modern Portfolio Theory to allocate assets, but it does not require them to use silly assumptions about the near term returns of those asset classes. Bonds and US equities face poor, near term returns. All asset allocations need not be based on 30 year forecasts.
“He [Grantham] suggests that he thinks the on-trend normal level of the S&P 500 would be about 2500 today.”
There are many ways to to calculate the trend. The basic approach is to plot (log-log) the S&P index against something upon which it logically depends, such as revenues or earnings (not time). I tried plotting it against GDP. As in Jim’s analysis, the trendline as of 12/21 depends on the start date of the plot. If I start the plot in 1929, the trendline as of 12/21 is 1,915. If I start the plot in 1945, the trendline as of 12/21 is 2,186. If I start the plot in 1985, the trendline as of 12/21 is 3,413. It’s good to include in the plot a number of periods when the S&P was clearly undervalued, such as 1981, and clearly overvalued, such as 1999.
He [Grantham] suggests that he thinks the on-trend normal level of the S&P 500 would be about 2500 today.
I have seen research where they are talking about SP500 23 earnings pegged at $250. Economy is still doing good, it can take a hit after rate increase, etc, but now doing good. If it continues and we actually have $250, are you expecting sp500 will trade at 2500 level?
The chances of wide eyed bull succeeding is far higher than a strong bear. I can understand being cautious but not that negative.
I have seen research where they are talking about SP500 23 earnings pegged at $250.
Economy is still doing good, it can take a hit after rate increase, etc, but now doing good.
If it continues and we actually have $250, are you expecting sp500 will trade at 2500 level?
The chances of wide eyed bull succeeding is far higher than a strong bear. I can understand being cautious but not that negative.
Earnings are cyclical.
The 2023 earnings could be high or low, but the 2023 cyclically adjusted earnings will be pretty predictably between the two.
Extrapolating GDP and estimating public market profits as a likely percentage of GDP makes a lot of sense. GDP growth is pretty predictable over fairly long timeframes.
A simpler method is just to extrapolate the real net earnings of the index over time, which is pretty close to the same thing.
Then it depends what sort of time frame you consider. Bullish optimists take shorter time frames because the tax cuts make a nice steep slope to the trend line.
The trend line of real S&P 500 earnings (and cap weight predecessors) using any start date 2010-2014
or anything 1999-2005 gives a pretty cheery figure in the $141-146 range in today’s money for calendar 2023.
Any start date for the trend prior to that gives a lower extrapolated figure for 2023.
For example, a trend line starting 1995 would extrapolate to 2023 earnings of about $135 in today’s money.
So, sure, $250 is entirely possible for that particular year, but it would be about 75% above trend, give or take.
Most of the value of any equity is far in the future, so no single year matters much.
Jim
I have seen research where they are talking about SP500 23 earnings pegged at $250.
Economy is still doing good, it can take a hit after rate increase, etc, but now doing good.
If it continues and we actually have $250, are you expecting sp500 will trade at 2500 level?
…
The trend line of real S&P 500 earnings (and cap weight predecessors) using any start date 2010-2014
or anything 1999-2005 gives a pretty cheery figure in the $141-146 range in today’s money for calendar 2023.
PS
Let’s say the 2023 S&P cyclically adjusted earnings come in at the midpoint of that extrapolation in today’s dollars, $143.50.
In that case, Mr Grantham’s target of S&P 2500 would be 17.4 times cyclically adjusted earnings.
My very rough expectation of something more like 3000 would be 20.9 times cyclically adjusted earnings.
Those don’t sound like crazily pessimistic figures.
Jim
Earnings are cyclical
There are many folks argue the board is not a perma-bear… here is your prime example. 
The cape wearing, mean revisionists have predicted the world is going to go back to buggy whips because if you take 300 years of industrial age…
Sometimes things change and such change don’t mean revert, or you can rewind the clock. So I consider cyclically adjusted earnings only for cyclical industries like commodities. Not everything is going to mean revert, if it does, I would like my age and hair to mean revert.
Now, with my ranting on mean-reversion out of the way, I don’t know what is going to happen on 2023, whether Fed will actually do as many rate hikes as market is expecting, whether Fed will follow through on their words in doing QT, i.e., they are really going to shrink FED balance sheet, whether US congress now got comfortable with uncontrolled money printing will be weaned away, etc. Many of these have positive and negative effects on economy. The $250 spy earnings assumes 8% earnings increase in 2022, which may be more attainable with supply chain easing and cap-ex cycle starting, will 2023 see 10% earnings growth, which is baked in to that assumption, I don’t know.
What I do know is, future is unknown, unpredictable, there are variety of outcomes possible and in most of those scenarios SP500 earnings is going to be far higher than pandemic low of $140. What that is going to be? IDK, but I dearly hope it is far better than 2020 economy.
The cape wearing, mean revisionists have predicted the world is going to go back to buggy whips because if you take 300 years of industrial age…
Sometimes things change and such change don’t mean revert, or you can rewind the clock. So I consider cyclically adjusted earnings only for cyclical industries like commodities. Not everything is going to mean revert, if it does, I would like my age and hair to mean revert.
More strawman arguments. Many on the board have argued that profitability (margins) might be permanently elevated; but the argument that they will continue to rise “forever” if of course impossible. There have been lots of discussions trying to explain the elevated margins - some of the explanations even make sense!
tecmo
…
Earnings are cyclical
…
There are many folks argue the board is not a perma-bear… here is your prime example.
So, you’re saying…earnings aren’t cyclical?
I dare say, a few hundred million business people and stock investors will be very startled to learn this.
What I do know is, future is unknown, unpredictable, there are variety of outcomes possible and
in most of those scenarios SP500 earnings is going to be far higher than pandemic low of $140.
FWIW, the pandemic low (rolling four quarters) was $94.13, not $140.
So the figure of S&P 500 earnings in the low $140s in 2023 suggested by extrapolating the real earnings trend would be 52% up from the pandemic lows, not equal to the lows.
I have no idea whether S&P 500 earnings will be $250 in 2023.
But if they manage that, it sure will be a lollapalooza of a result. (achieving the wildly improbable?)
$250 S&P 500 earnings in 2023 would correspond to total US corporate earnings being just over 16% of GDP.
(Using the fairly reasonable approximation that S&P 500 earnings growth is about the same as total US corporate earnings growth in this stretch)
That’s up about 37% from the current figures which represent the all time high in the Fed database back to the 1940s.
Have a look at what corporate profits at 16% of GDP looks like in context https://fred.stlouisfed.org/graph/?g=cSh
Let’s just say that 16% wouldn’t be my central expectation.
Jim
So, you’re saying…earnings aren’t cyclical?
You are arguing against something that I didn’t say. Why?
Economy goes through cycles is something everyone understands. But even your cyclical lows can and are higher than past cycle lows’.
At the moment, economy is doing fine and 101 econ says inflation results in economic growth, or GDP growth. What we are suffering is high inflation, thus higher GDP. Predicting cyclically low earnings is not aligned.
I linked the numbers I use and you seems to be using different set of numbers. There are no 4 quarter rolling period where I see $95 earnings starting Q1 2020.
The way I see is 8% increase in 2022 from 2021 and another 10% increase in 2023 from 2022. We can apply that to the number you are using.
Perhaps you can provide your numbers reference. Without that, we are arguing apples and oranges.
So, you’re saying…earnings aren’t cyclical?
…
You are arguing against something that I didn’t say. Why?
Well, because you pretty much said it?
You said that the assertion that earnings are cyclical was a “prime example” of being a permabear.
OK.
So, either you don’t believe earnings are cyclical, or you believe being a permabear is a sound position to take.
Being a permabear is clearly an incorrect/nonsensical stance (we’re on the same page there), so by extension, I can only assume you don’t believe earnings are cyclical.
Or were you 'fessing up to thinking the permabears are right?
Either way–
No, believing earnings are cyclical isn’t being a permabear, it’s understanding how the world works.
In some years they’re really high, and in some they’re really low.
The main thing to understand is that neither one matters–only the trend level of earnings matters for the broad market.
That’s because only aggregate future earnings matter, and that future, though squiggly at short time frames, will have an average level which the trend line follows.
It’s certainly possible that 2023 S&P 500 earnings might come in at $250, however unlikely it seems to anyone but sell-side bankers.
But the bigger point is that it’s irrelevant–that’s definitely not going to be a point of the trend line, nor an indication of its future.
The average of the subsequent decade is going to be at least 20-25% lower than that in real terms even on optimistic assumptions.
(e.g., earnings continuing to rise faster than GDP, past their current record percentage)
There are no 4 quarter rolling period where I see $95 earnings starting Q1 2020.
Use the figures from Standard and Poors. That’s the actual figure for the four quarters to end 2020.
Google “sp-500-eps-est.xlsx”, check cell J142.
It’s the same figure as is reported elsewhere, e.g. Pinnacle, who reports based on a different methodology.
(Dow Jones data services used by Pinnacle reports the sum of known and published bottom-up trailing
four quarters as reported, whereas S&P’s figure is “profits earned during the 12 months ending on that date”.
The S&P figure is correct, but can only be known well after the date in question. The DJ method is “as known at the time”)
Jim