Thirty Three Days: The Fastest Bear Market Ever Recorded

Where the Decade Started

Each strand of this series - the stimulus the meme stocks the inflation the rate hikes the regional bank bankruptcies - dates back to five weeks in early 2020. On February 19 of that year the S&P 500 closed at an all-time high. Markets were treating a new virus spreading from Wuhan as a regional story contained elsewhere. Thirty-three days later on March 23 the index had lost 34 percent.along the way it set a record that no one wanted the fastest fall from an all-time high to a bear market a 20 percent drop from a high ever measured in just 16 trading days. 1929 took longer to get there.market and there was no manual to stop the revaluation

Eight Days of Circuit Breakers

The mechanics of the crash were as historic as its speed. U.S. markets have circuit breakers automatic trading stops that activate when the S&P falls 7 percent in a single day. They were built after the 1987 crash and are rarely used. In 2020 they were activated four times in eight trading days on March 9 12 16 and 18.March 16 was the worst single session with a drop of nearly 12 percent the second-worst day in the index's history behind only 1987. That same day the VIX the volatility index that markets use as a gauge of fear closed at 82.69 the highest close ever recorded. Oil plummeted along with stocks a story that has its own entry elsewhere in this series. Credit markets closed almost entirely.Companies that needed to borrow even healthy ones couldn't get a reasonable price and some couldn't borrow at all. Four stops in eight days is not a market correction. It's a market that has temporarily stopped being able to find price

Why It Fell So Fast: The Machinery Under the Selling

Velocity like that needs a mechanical explanation and the honest answer involves plumbing that most people never think about. A significant portion of the money in modern markets is not managed by someone with an eye on the news. It is managed by a formula that targets a fixed level of volatility. Risk parity funds volatility-controlled variable annuities trend-following funds and target volatility mutual fund share classes follow some version of the same rule: hold more than one asset when it's calmholding less when there is turbulence and automatically adjusting as measured volatility changes. In a normal month that rule barely budges. In March 2020 measured volatility went from a normal calm reading to levels the market had rarely if ever recorded - the same peak behind the VIX record close mentioned above.a feedback loop not a forecast

Layer a second constraint. The banks that make stock and bond markets the dealers hold less inventory on their own balance sheets than before 2008 because post-crisis capital rules made storage risk more expensive for them. When the forced sellers arrive all at once and dealers have less excess balance to absorb the flow an order of the same size does more damage to the price than fifteen years earlier. Neither mechanism cares what the virus does next. Both help explainwhy a fall that would normally take months is compressed into three weeks

A Worked Example: What Volatility Targeting Actually Does

Let me nail down the volatility targeting mechanism with round illustrative numbers not the actual portfolio of any real fund. Suppose a fund manages $1 billion and targets 10 percent annualized volatility in its stock portfolio. A standard volatility targeting rule sizes the position as the target volatility divided by the realized volatility. If the actual volatility of the S&P 500 remains at a leisurely 12 percent the fund maintains 10 divided by12 or about 83 percent of a full notional equity position. Call it $830 million of exposure

Now push the same illustrative fund through a stress scenario loosely modeled on that week in March not on any particular fund's actual numbers. Suppose the actual realized volatility on the fund's book jumps to 60 percent a level in the same range as how extreme the previous actual VIX reading actually was. Run the same formula: 10 divided by 60 is about 0.17 or 17 percent of a full position. For our illustrative fund this means reducingequity exposure from $830 million to about $170 million a reduction of about $660 million close to an 80 percent cut in the position and the formula is not waiting for a quieter week to execute it

The issue isn't the exact numbers which I chose to be round and verifiable. It's the form: a change in measured volatility produces a large rapid mechanical change in the amount of asset a fund will hold without any human being recalculating a view of the economy in between

The Week Even the Safe Assets Broke

The detail that professionals remember most is not the fall in stocks. In mid-March US Treasuries the asset that everyone is rushing to in panic also began to fall. That was not supposed to happen and the reason it happened reveals what the accident had become. Investors everywhere funds facing redemptions foreign central banks defending their currencies companies using lines of credit needed dollars immediately and sold everything they could sell including theirsafer holdings. The phrase for this was the rush for cash. Typically dealers would step in buy the Treasuries that were being sold and hold them until a natural buyer appeared softening the price. But these were the same traders whose balance sheets had become tighter since 2008 and the sales were coming from all directions at once: risk parity funds that reduced bond exposure through the same mechanical process describedPreviously foreign accounts raising dollars leveraged funds unwinding core trades. There wasn't enough balance sheet capacity in the system to cleanly absorb it. When Treasuries break the pipes of global finance break because Treasuries are the underlying collateral for almost everything else. For a few days in March 2020 the question wasn't how far stocks would fall. It was whether the system itself would continue to function

Case Study: Bridgewater and the Risk Parity Squeeze

If you want to assign a name to the above mechanism the clearest one is Bridgewater Associates one of the world's largest hedge fund managers and the firm most closely associated with popularizing risk parity as a strategy. Ray Dalio created Bridgewater's flagship All Weather fund on a specific idea. Instead of splitting a portfolio into 60 percent stocks and 40 percent bonds by dollar value split it based on the risk each part brings and leverage the quietest assettypically bonds so that stocks and bonds contribute roughly equal amounts of volatility to the entire portfolio. In a normal environment this is a genuinely smart piece of portfolio construction because stocks and bonds typically don't fall together and the strategy relies on that trade-off

March 2020 was close to the worst-case scenario for that assumption. Stocks and bonds sold together for stretches of the month which is exactly the correlation breakout risk that the peg is least prepared for and volatility spiked in both legs of the portfolio at once triggering the deleveraging math from the worked example above on the stock side and the bond side simultaneously. I don't have the exact Bridgewater fund level numbers in front of me andI'm not going to pretend I have them. What is well documented is the direction. The funds that run this playbook and there are many beyond Bridgewater went through one of the toughest times on record in the first quarter of 2020 and the mechanism above explains why regardless of the exact performance of each company

The Response That Drew the Floor

The Fed acted faster than ever in its history. An emergency half-point rate cut on March 3 essentially did nothing the market continued to fall. On Sunday March 15 the Fed cut rates to zero and announced $700 billion in bond purchases and the market fell 12 percent the next trading day as investors read panic into the magnitude and timing of the response rather than calm. The bottom finally came on March 23.March when the Federal Reserve announced that its bond purchases would be effectively unlimited and crossing a line it had avoided throughout its history said it would buy corporate bonds directly supporting the credit of American businesses with the backing of Treasury money from the CARES Act a package this series covers in its next entry. March 23 the only day of maximum monetary strength was the exact day the market bottomed. I don't think that's a coincidence. It might be the most important mechanical fact of this entire series:Sales stopped once the forced sellers described above finally had a large enough backer to make continuing selling futile

The Rebound Nobody Traded Well

What followed confused almost everyone including myself. The S&P 500 regained its February high on August 18 2020 less than five months after hitting bottom the fastest full recovery from a bear market on record while the real economy was still mired in the worst recession since the Depression. The gap between a booming market and roughly 10 percent unemployment seemed obscene to many people and became a political grievance.defining.But the market never priced in the present.What the bailout had all but guaranteed would survive the shutdown were the prices of zero interest rates trillions of dollars in stimulus and corporate cash flows.Investors who sold at the bottom and waited for the economy to visibly improve before buying back missed out on one of the great rallies in market history.It's a costly lesson about the gap between the economy and the market and one that a whole generation of new investors including this author learned in real time ininstead of doing it in a textbook

The Threshold Problem: Why Fastest Oversells It

Now let me argue against my own headline because I think this particular record is more fragile than it seems. The bear market has a specific almost arbitrary definition: about 20 percent below the most recent closing high. Twenty is a round number that market historians established not a threshold with any real theoretical basis behind it. The fastest bear market record is actually a record for the fastest time to cross a specific conventionally chosen line and that's a more limited statement than it seems.the fastest fall

Think about what the puck actually rewards. A drop that falls in a nearly straight line with few relief spikes along the way crosses a fixed threshold faster than a drop of similar or even larger total size that zigzags its way down. Both 1929 and 2008 produced total losses much larger than the 34 percent COVID drop but they took a winding path to get there with spikes along the way that delayed the time whenthe index actually closed 20 percent below its previous peak. COVID's path down was unusually straight which is truly remarkable but straightness and severity are not the same and the headline figure mixes them up without saying so. Let's also consider the year 1987. Black Monday alone took about 20 percent off the index in a single session arguably a more violent shock in the strictest sense than anything in March 2020 however due tothat the market stabilized afterward rather than continuing downward does not hold the record for the fastest bear market by this definition

I still think the COVID crash was extraordinary. I just don't think the fastest bear market ever is doing as much explanatory work as the headline suggests. It's about measuring the shape of a crash against an arbitrary line not measuring the total damage and it's worth knowing before repeating the statistic in a room that will ask you to defend it

How I Actually Think About Crash Speed Now

This is what really changed in the way I read a sell-off after analyzing all of this and it has nothing to do with predicting the next drop which I don't think I or anyone else can do reliably

My read is that the most useful thing to look at during a rapid decline is not how much stocks are down. It's whether historically uncorrelated assets are moving together. If stocks fall and bonds rise like they're supposed to that looks like an ordinary revaluation: investors move from risk to safety in a way that makes sense on their own terms. If stocks and bonds fall together that's something else entirely and March 2020 is a clear example of that. It usually means thatsomeone or many at once needs cash immediately and is selling whatever is liquid rather than what they actually have a negative opinion on. This is closer to a plumbing problem than an opinion about the future and plumbing problems tend to need political backing before they are over not just the passage of time

The way I would actually use this and I want to be clear that this is a way to read a crisis as it's happening not investment advice is to treat a Treasury sell-off during a stock sell-off as a signal to weight central bank and Treasury statements on economic data because what resolves a liquidity event is a support not a better jobs report. The first time I studied this crisis I put the emphasis the other way around focusing on the virus curve rather than the underlying market pipelines.plumbing is what really explains the bottom line

The Bottom Line

The COVID crisis compressed an entire market cycle into five weeks a plumbing failure that reached the Treasury market itself and a policy response that redefined what central banks are willing to do. Part of that speed was real and structural. Volatility directed at funds and risk parity strategies deleveraging at the same time and bank operators with less room for maneuver on their balance sheets than before 2008 turned an already strong sell-off into something faster than they expected.The markets have ever recorded. Some of the fastest bear market labels are an artifact of timing to an arbitrary 20 percent line rather than measuring the total damage. Both are true at the same time and I think you need both to really understand what happened instead of just repeating the headline. Their fingerprints are on every other entry in this series: the stimulus that fueled the mania and inflation of 2021 the zero rates that mispriced everything.that 2022 had to revalue and the precedent that markets now expect an unlimited bailout every time the fund seems scary enough. Thirty-three days and the rest of the decade was spent paying the bill