“Markets crash all the time. You should, at minimum, expect stocks to fall at least 10% once a year, 20% once every few years, 30% or more once or twice a decade, and 50% or more once or twice during your lifetime. Those who don’t understand this will eventually learn it the hard way.”
–Morgan Housel (The Motley Fool)
Since 1960, surprisingly few S&P 500 corrections have occurred during bull markets, and the bear markets in between appear rather infrequently¹:
1960-2014: 16 corrections (avg -14.65% & 118 days)
1985-2014: 5 corrections (avg -17.86% & 91 days)
1982-87: 1 correction
1987-2000: 2 corrections
2002-07: 1 corrections
2009-pres: 3 corrections
1960-pres: 8 bear markets (avg -37.43% & 464 days)
1985-2013: 3 bear markets (avg -46.47% & 508 days)
Now, some color for those data…
First, a lot of these corrections bottomed-out in the high teens. In fact, 4 of them ended at 19%+ drawdowns — not quite enough to be an objective bear market (i.e. 20%+). Another 4 instances ended at 9%+ drawdowns, which is not a real correction (i.e. 10%+) and hence excluded from this study.
Next, in the last 54 years captured by that data set, SPX has been stuck in a correction only ~10% of calendar days and a bear market ~19%, which means that combine, the market has been in correction or worse almost 30% of days.³
That’s far more than I had assumed, and it doesn’t even include all those drawdowns that barely missed qualification as an objective correction. Further, while we count far more (higher frequency) corrections than bear markets, the market surprisingly spends far more aggregate time in the latter state.
To emphasize age of central bank interventions and disinflation, we can look at the same data for the Great Moderation (1985-present) in isolation. Perhaps that’s a more relevant gauge for expectations, given modernity. To wit, over that time SPX has been stuck in a correction only ~4% of calendar days and a bear market ~14%, for a combine total of 19% of all days.³ Empirically, it would appear that the Great Moderation has succeeded in smoothing-out the business cycles to avert pesky corrections, but bear markets are more severe and only slightly less frequent. (Granted, the control group doesn’t include the pre-Great Depression or pre-Federal Reserve periods.)
Regardless, investors can blindly choose a day to put money to work, and they have a high probability of catching a bull trend — truthfully an overwhelming majority of the time. In addition, this study confirms the conventional wisdom that assumes bull markets progress slowly over a long period of time, and major drawdowns quickly over a shorter timeframe.
Armed with this data, I have to issue a “mea culpa” on a guiding principle of mine, formerly, “bullish optimism is my default setting because we’re in a bull market 90% of the time.” While my probability estimate was too high, the upshot remains: the market is trending bullishly a big majority of the time (70-80% of days, depending on the sampling period), and corrections occur less than once every 5 years in secular bull markets. Optimism is still a good default setting, with cash only needing to be raised once every 4-7 years.
Now, how does this help us in today’s market? I’ve lassoed a number of other quantitative studies over the past few months that may — in aggregate — help us to that end, including:
- SPX long term regression (–)
The market is currently +80% aboveitsmean (almost2sigmas) vs. prior highs of +85% before the Panic of 1907, +81% Great Depression, +150% Tech Bubble, and +89% Great Recession. While we’re approaching both 2 standard deviations and those prior high-water marks from the Great Moderation era, there’s still a little meat left on this bone — although this indicator suggests we tread lightly.
- SPX inflation-adjusted secular highs & lows (+)
The Great Recession’s bear market that bottomed in 3/2009 (-59% in real terms) was commensurate with predecessors, but today’s SPX +122% (inflation-adjusted) over 5 years is still mid-stride compared to its 5 historical secular bull market precedents, which average +424% (+396% median & 131.6% standard deviation) over 15.9 years (18.1 median & 8.8 year σ).
- Average secular bull market (+)
Since the Great Depression, there have been 12 bull markets, but only 3 have survived into a 6th consecutive year (not including today’s yet) and 1 failed (FY1987 -20% starting from Black Monday). All 4 of those extended bull markets had an average 6th year return of +10% with a 23.6% standard deviation — high due to a small sample size. Today’s market shows a clear correlation to prior secular bull runs, especially the 1950s bull market (1949-56) — perhaps the best analogue to today’s, considering the geopolitical environment, interest rates, and demographics.
- Momentum (+)
Contrary to conventional memes, buying at alltime highs like today’s has historically been a good strategy, because the frequency of new closing highs since 2013 is a signal of a secular bull market a la 1952-65 and 1983-99.
- Mean reversion (PUSH)
Despite a lot of debate citing overfitted data, SPX typically mean reverts after big years like 2013. After >25% CY performance, SPX’s subsequent mean return is only 5.47 vs 11.5% historical average. While that’s underperforming the average, it’s still a healthy return — although it’s hard to place much faith in any expected returns using this data, because their 18.27% standard deviation is indicative of high variability.
- Investor surveys (+)
- Gallup poll of Americans’ view of the stock market shows “irrational non-exuberance,” with only 46% thinking equity investment is a good idea and 50% calling it a bad idea — both nowhere bear historical extremes
- Retail sentiment’s bull/bear ratio @ 1.1x is far from the 1.8x average observed in prior, exuberant extremes.
- Retail equity allocations @ 66.9% are above their long term average of 60%, but not quite up to prior extremes of 70% in 2006 or 77% in 2000.
- Strategists’ equity allocations @ 54.1% are below their long term average of 60.3% and prior extremes of 67% in 2007 or 71% in 2002.
In sum (4+/1PUSH/1–), these suggest we should, objectively, not fear a big, bad, bear market today.
Now, I have a few reactions regarding the opening quote, including a takeaway within the context of all the quantitative data presented herein…
First, the theory of portfolio management is a lot different than its practice. Hindsight bias is one reason for that perception/reality gap. Another is loss aversion. For example, we humans worry more about getting caught in a bear market than missing a bull — at least until the latter stages of bull rallies, when our friends all brag about their stock market riches at cocktail parties. During a correction, people extrapolate the trend and step-out of the market… but it’s real hard for them to step back in. After all, we don’t know it’s going to be a mere correction until it’s over.
Second, corrections and bears don’t occur like clockwork. They don’t fall at set intervals — not even within months, years, or half-decade of some schedule. Everyone’s time horizon is different, but real investors shouldn’t trifle with corrections. Throughout the Great Moderation, SPX’s average annual drawdown has been -14.5% with a 10.2% standard deviation. You can see there have been a lot of little drawdowns in the single digits, and a few large drawdowns amounting to bear markets. However, compare that ephemeral risk with the reward of SPX’s average annual return of +10% with a 16.6% standard deviation — inclusive of those drawdowns. Investors with long time horizons should focus on the opportunity cost of missing a bull market ( e.g. +100%) juxtaposed against the risk of a hiccup (-15%). In fact, the upside you sacrifice in a balanced portfolio helps you withstand brutal downside in garden-variety bear markets
Beyond that, we’re left to fear the big ones — the generational crashes. Sometimes, in hopes of frontrunning such crises, you can try your best to control for your own biases and watch out for the psychological excesses of others. Most of the time, nobody sees it coming — a black swan is an unknown unknown by definition.
So, unless my multidisciplinary process renders a high conviction sell signal from a number of indicators that all say a secular crash is coming, I’m best left to trust the discipline of prudent asset allocation and diversification.
Just because I’ve determined that such a generational crash isn’t coming here and now, that doesn’t mean that corrections won’t occur as ever; it just means that we shouldn’t fear them.
–Romeo (hattip Barry Ritholtz)
²Correction= 10% drawdown; Bear market= 20%
³1960 – 2014 = 54 years * 365 days = ~19,710 days
1985 – 2014 = 29 years * 365 days = ~10,585 days