The markets peaked [pre-coronavirus] 360 trading days ago at S&P 500 of 3386. It bottomed 23 days later at 2237 [down -34%].
Markets have also been up 30% over the pre-coronavirus peak of 360 days ago; up over 97% since this nadir 337 days ago [exactly 1⅓ years ago]. During this broader ascent since the nadir, markets have been positive on 58% of days. Been greater than +1%, on 23% of the days.
Achieving 60 all time highs en route. Suffering only a -9.6% correction, 223 days ago [almost 11 months ago]. The second worst was a -7.5% correction, which was 196 days ago [just over 9 months ago]. Not even a -5% correction since.
How do we put these reasonably awesome and new market statistics within context? Are they really unusual, and if so how? We argue here again that statistical theory and empirical data must intertwine, across a large enough horizon. We have vast periods of market drops from all-time highs, market drops while rebounding towards an all-time high, and meaningful rebounds within a market drop. In addition to combinations of the above and complexities within them, this makes contrasting and predicting market risk statistics challenging.
But never impossible to either understand, nor profit from. We’ll focus here on the first type of market drawdown statistic, from all-time highs. We are at all-time highs quite frequently after all these nowadays. This focus doesn’t overly limit as roughly ⅘ of market days are in a period of ascent from a bottom [even if below a previous watermark].
And we appreciate nearly a hundred years of data from the S&P 500, necessary to focus on many technical variables, and over a dozen cycles, and dozens of market ascents. As a result we focus on our energy there; with all data as of July 23, 2021.
Daily market changes
The issue most analysts have in studying market data is that they will process a complete batch of empirical data and deduce a theory, which is a confused mix of statistics and some sort of random personal view on markets. Nature on the other hand, has its own way. Nature is clever. Nature is neither spurious, nor trivial. And with the exception of say dinosaurs, rarely do the outcomes ever get to be existentially-limiting.
It is true that the market statistics of late appear unusual if you mine it in specific ways. On one hand there have been more [and asymmetrically so] positive days than we usually expect. We noted above that 23% of days were greater than +1%. In the 4 ascents that collectively define the previous cycle [from 2009 through the pre-coronavirus peak in 2020], we typically saw 10% or so of the days greater than +1%. And in all of the 21 full ascents prior to that [in S&P 500 history since 1928], we typically saw 12% or so of the days greater than +1%. Barring the run-up era near 1998-1999 [a total of 2 ascents out of 25], we’ve never seen ascents with 23% or more days that exceed +1%.
In the video below, we show the distribution of daily returns during all 26 ascents [and also one more for the current partial one]. Nearly 19,000 trading days in all!
Appears interesting, but we notice the seeming curiosity since the 2020 nadir, occurred “prior to” the September -9.6% correction. The 32% of greater than 1% up-days prior to the correction was nearly twice that of the history outside of it. Yet since this September -9.6% correction, the 24% days being all-time highs was nearly twice the portion of the history outside of it. And a perfect example right there, that there is not even any connection between those two popular metrics! But it also shows the appearance that this current ascent is aging.
Now it is possible that the large Federal Reserve's monetary stimulus has helped create this nice run-up in markets since early 2020. But it’s also possible that it didn’t and that something else did. It’s possible too that nothing has created this impulse, and that we are simply witnessing a rare run-up for no reason other than luck, or a continuation of the pre-coronavirus path we enjoyed for several years (see New York Times article 1 and article 2 on a period when we had only 7% of the days with greater than +1% daily changes, and 17% all-time highs).
After all, we’ve seen these statistics before as noted [large streaks of greater than +1%, and of all-time highs] in 1928-1929, in 1982ish, and in the runup of 1997-1999. Even more so than our recent 2016 or highly-valued 2019 eras. Eras with quick snap-backs and the only ones which younger investors might have only seen and are now contributing to. And we didn’t always have this kind of economic or monetary stimulus then. Large crashes "ever so" eventually and abruptly arise. So you must have a healthy dose of humility to respect that simple explanations won't bear fruit in a robust backtest. Generally there are multiple combinations of explanations at any given time, sometimes not “known” until the post-mortem narrative is penned.
Market drawdowns
Now we also have the somewhat complementary issue of why we have not seen compelling market crashes, which are deeper and without bounceback. We might have valid reasons. Or we might simply be guessing. The factors most relevant above, we see have large errors in mapping to the recurring timeliness and severities of a market correction. And as with the rest of this article, the history of nature gets deeper with each passing day. Everything “unusual” becomes a normal part of a historical range. Many times in life you might not expect something, until it occurs.
209 days since the last market correction:
Periods of market up-moves and all-time highs will be interrupted somehow:
Additionally every asset class goes through it’s own daily pattern. Just like people respond a little differently to the same weather, in different driving locations. Not every asset class has been around as long as the U.S. stock market.
In theory, over a long time however [as in still a couple more years or so] they should be "much more" in harmony. But the same could have always been said of any point in history, making this period not necessarily special. Even Japan in some ways has finally recovered from its own major 1980s bubble.
Be patient with what appears to be -but may not be- an anomaly. Invest with caution. Know that a severe corrections can and will ultimately unfold, as always. If you are leveraged in exotic assets, you certainly learned the wrong lessons from this good period. Your investments may break far more that you might assume at the moment.
Larger market corrections [e.g. -20% or so bear markets] will always be part of one of the next market corrections. They weren’t in early 2018. But they can't be avoided and always rear their head inopportunely [sometimes right after a -10% or so correction]. We saw this later in 2018 [in the absence of a recession too].
We didn't see many back-to-back corrections so far since the 2020 nadir. But given our unusual upside activity, we could likely at some point. And they don't always simply snap-back. The ongoing ascents and rapid snap-backs makes this a [somewhat] unusual post- global financial crisis era. For example, in limited cases so far they never snapped back [e.g., some popular domestic and foreign ETNs earlier in 2018, on the smaller correction, or some cryptocurrencies in the post-coronavirus era].
Summary
All patterns have to come, and must be replaced with similar outcomes in the figurative casino shoe. Else you don’t have a market! Happy periods come and linger, but they don't last forever. Unhappy periods also come and go over periods "generally longer" than we've seen! We've been in an "unnaturally" happy period now, and as noted it has been like this for some time...
Daily market changes [albeit fat-tailed] are more random in the normal way people are accustomed to digesting statistics. In terms of their variance, ongoing streaks, and hundreds times more data points to explore. Versus episodes [and then severities] of market crashes are a different form of statistical modeling, and a summary that is much more difficult to model. We are fooled only through some temporary shifts in parameters.
We had “reasons” for the market to crash or to rally along the way They did when I was running the analytics team for Timothy Geithner, in TARP. During that time we also faced some cross-currents of similar stimulus and government shutdowns. And Taper Tantrum. Mild inflation. Debt-downgrades and global civil unrest. Testing of a new President.
And all this occurred after the NBER declared the recession over, and the recovery began. How do we know this time is much different? We can’t fully know. We are likely guessing for the moment. And the current stimulus we know will taper at some point in the near future.
Invest like you have less sense than you like about what "can" happen next. The video above showed we have barely seen the larger pendulum up and especially down currents we can see [and that’s during good times!]
We may continue in the current form for a while longer [regardless of how long the previous streak occurred], but will abruptly change as well at some point. Comfort is your enemy. Not repairing the roof of your portfolio, when the sun is out, could be your own [but not your neighbor's] self-imposed risk. Most do not see a market depression, just as they didn't in the late 1920s, late 1960s, and late 1990s.
The theories and practical lessons here have all been seen in some various format before. In addition to the article linked above, we have in this article here the lessons [from missing data, to aspects of life that emulate a simulation] in my new book, Colors and Numbers. One can also see our previous website articles on degree of freedom adjustments in market changes, elevated valuations, intra-tier risks and our initial article on recherché streaks.
When we fool ourselves with assumptions of a new natural environment, we make many imprudent decisions, and all at the wrong time. We see driver seatbelts, sports helmets, novel vaccines, etc. And assume the world is now much safer, “forever”.
So indeed, stay the course. But to internalize this means more than knowing that the environment is unusual and some aspects of it must eventually regress. It means not going above and beyond on risk. Acting just the same and just as prudently even though it seems everything is safer [just because you "said so", but only in the rear-view mirror].
Our nature is always developing new irksome qualities. And borrowing from 20th-century literary poet, T.S. Elliot, we will, "... arrive where we started. And know the place for the first time."
Supplemental
August 3a
August 3b
August 5a
August 6a
No comments:
Post a Comment