If you skim the list of biggest one-day percentage drops in the S&P 500, you'll see a lot of dates that look roughly right: 1987, 1929, 2020, and 2008 all dominate the list, while the rest includes some of the worst days of the Great Depression (and one day in 1940 when the market dropped on news of German progress in invading France). If you look at the biggest gainers, the years you see are—basically the same collection of years! And what you won't see are the years of raging bull markets: the decades with the best returns, the 50s, 90s, and 80s, are almost completely absent from the biggest-gains list (the one exception is a big rally two days after the crash of 1987).

So, as a general rule, the people who make money on the best few days of market performance ever are people who probably regret it a bit, since most of them were long during bigger losses just beforehand. But this raises some fun questions: why are returns so asymmetric? And why are the best days in some of the worst months?

One angle is to look at the shape of a bull market, which is part of the shape of winning trades on the long and short sides. Goldman has a nice primer on this from long ago, with the upshot being that when a long position pays off, it's usually in the form of a steady stream of good news, while a short position's payoff often takes the form of an absence of good news followed by a lot of bad news all at once.¹ And that fits in with how these theses work: if you're long a fast-growing restaurant chain, your thesis is that the next Chipotle location performs similarly to the previous one; if you own something like TransDigm, you're convinced that the next price increase they push through will be as revenue-accretive as the last one. If you're short, you generally expect the story the market believes to fall apart, which will probably happen quite fast when it does.

Some of this gets replicated at a macro scale: if the economy's growing reasonably fast, and inflation hasn't picked up, the default bet is that the next year will look similar. There will always be something to worry about, but most macro worries don't materialize, and good returns are the way equity owners get paid for the risk that these concerns do materialize.

Another way to look at it is that leverage is a pro-cyclical phenomenon: when the economy's humming, firms want to expand, and they'll borrow to do so. Consumers are flush, and will tend to buy larger houses and nicer cars than they otherwise would have (or could have), and to fund some of this with credit. All that credit expansion leads to more spending, but it also means that the current state of the economy is some mix of better fundamentals—more output per hour from more hours worked—and purely credit-induced demand. When credit creation slows down, not only does consumption reset to a lower level, but now, the economy is both more variable and more levered, which is a dangerous combination. Households and firms respond to this by deleveraging, sometimes by choice (consumers deferring spending, businesses reducing capital expenditures in order to pay down debt) and sometimes because they have no choice (foreclosure, margin calls, bankruptcy).

At the times when this is happening, the economy's higher fundamental volatility and higher leverage mean that riskiest part of the capital structure, equity, has returns magnified in both directions. Since the cause is bearish, much of that volatility will show up in sharp one-day declines, but this implies that rallies will happen, too. There's sometimes a sawtooth pattern to equity optimism, where the worse the economic news gets, the more drastic the interventions are. When the economy turns, markets start to price in that kind of response, but sometimes they underestimate the magnitude even if they get the direction right.²

The last ingredient in the volatility/downturn correlation is that the price of liquidity is higher when there's more fundamental uncertainty, and this makes the market more reactive to extreme events. To make a trade, someone needs to be willing to take the other side of the trade, and the price you have to pay them to do this is, literally, the gap between where you can buy and sell the same asset. If immediate liquidity shrinks faster than the total value of the market—a safe bet—it means that a given buy order that would have barely moved prices during a bull market can make them rip in a bear market. The financial world was falling apart in October 2008, but that was also the month of one of the biggest short squeezes of all time.

We've been covering the recent market volatility in The Diff, and have older pieces that look at the nature of liquidity as well:

• Here's more on why one misread headline sparked a temporarily multi-trillion dollar rally ($).

• Markets are always calculating new prices, but the calculation is buggier when prices move faster.

• In March of 2020, we looked at the game theory of Covid interventions ($).

• And in the same month: projects with slow feedback loops are a hedge against the psychological toll of volatility ($).

Join the discussion!

1  If that short is of the fraudier variety, not necessarily a company that completely fakes its numbers but one where the company mostly fakes its narrative, there's often a stream of good-sounding news whose credibility the market slowly discounts more and more, but still with a big drop at the end.

2  These interventions, particularly on the fiscal side, serve another purpose: if people and businesses are spending less and saving more, and are particularly sensitive to losses, they'll choose lower-risk savings vehicles, and the lowest-risk option is treasury bonds. So bigger deficits during recessions are partly about stabilizing the overall economy, but partly about responding to and taking advantage of the market's demand for more treasury bonds.