How Long Did the March 2020 Stock Market Crash Last?

Key takeaways

• The S&P 500 fell 33.92% from its 19 February 2020 close of 3386.15 to 2237.40 on 23 March 2020, and it covered that ground in 23 trading days.
• Getting back took 126 trading days from the peak. The index first closed above 3386.15 on 18 August 2020, at 3389.78, which needed a gain of 51.51% off the low.
• In sterling, global equities fell 26.04% and took 195 calendar days to recover. A 60/40 of world equity and gilts fell 15.18% and was whole again in 145 days.
• Gilts gained 4.32% while equities were falling, then lost 6.34% in the eight business days to 19 March 2020 as the scramble for cash reached government bonds.
• The cushion was specific to this episode. In 2022 the S&P 500 returned -18.04% and a ten-year Treasury -17.83%, so the bond sleeve absorbed nothing at all.

The March 2020 stock market crash took 23 trading days down and 126 back up

You probably came here for a date and a duration of the March 2020 stock market crash, so here they are.

The S&P 500 closed at 3386.15 on 19 February 2020. That was the high. It closed at 2237.40 on 23 March 2020. That was the low. The fall between the two is 33.92%, and the index covered it in 23 trading days.

Climbing back took much longer. The index first closed above 3386.15 on 18 August 2020, finishing at 3389.78. That is 103 trading days after the low and 126 after the peak, so the way up took 5.5 times as many sessions as the way down. From the low, the gain required was 51.51%, because a 33.92% fall leaves you needing half as much again just to break even.

The dates are worth separating, because the two questions people type are not the same question. If you want the day the fall started, 19 February 2020 is it for US equities. If you want the day the first UK lockdown started, that is 23 March 2020. That evening the Prime Minister said "From this evening I must give the British people a very simple instruction - you must stay at home." The lockdown began on the day the index bottomed, not the day it began falling. The market had already repriced the thing the country was about to do.

Two days did most of the damage and the turn came the day after the low

The March 2020 stock market crash wasn't evenly spread across those 23 sessions. The worst session was 16 March 2020, when the S&P 500 lost 11.98%. That was the first trading day after the Federal Reserve cut on a Sunday evening. The same day, the Vix closed at 82.69. The Vix is the options market's price for the next month of S&P 500 swings. It had closed at 14.38 on 19 February 2020, so the market's own estimate of turbulence rose 5.75 times in a month.

The second-worst day was 12 March 2020, at 9.51%. Before either of them, 9 March 2020 had taken 7.60%. Those three sessions, compounded, come to a 26.4% fall on their own.

Then it turned, and it turned immediately. The best single day of the whole year came on 24 March 2020, the session straight after the low, at 9.38%. Nobody had new information about the virus that morning. What they had was the Federal Reserve's statement from the previous day. It said the committee would buy Treasury securities and agency mortgage-backed securities "in the amounts needed to support smooth market functioning."

How far each mix fell, and how long each one waited

A single index number is not what you hold. The table below takes a UK investor's portfolio on 20 February 2020, which is the day global equities peaked in sterling, and holds it unchanged through the year. The equity sleeve is the MSCI World index on a gross total return basis in pounds. The bond sleeve is a ten-year gilt, repriced every day from the Bank of England's par yield. Nothing is rebalanced, nothing is added, and every figure is measured in sterling.

Depth tracks the equity weight almost mechanically, which is what you would expect. Cut equities from 100% to 60% and the 60/40 portfolio drawdown comes in at 15.18% rather than 26.04%, so the gilt sleeve absorbed 10.86 percentage points of the fall.

The column that is less obvious is the waiting. The all-equity holder was under water for 195 days. The 60/40 holder was under water for 145. That is 50 days, seven weeks, of difference, and it comes from arithmetic rather than from skill. The shallower hole needs a smaller climb: 17.9% against 35.2%.

That result isn't the general rule, and it's worth saying so plainly. Take the 1973 to 1974 fall, measured on Aswath Damodaran's annual series and rebalanced each year end. From the end-1972 level a 100% equity portfolio lost 36.50% in nominal terms and a 60/40 lost 20.82%, and both of them were back to even in 1976. There the bond sleeve bought a shallower hole and not one year of time. The real-terms version of that episode, and the same test across every drawdown since 1928, sits in the piece on equity allocation and crash depth. The 2020 market crash recovery is one of the episodes where bonds happened to buy both things at once. One episode is not a rule.

Gilts cushioned the fall and then had a crisis of their own

The table makes the gilt sleeve look serene. It wasn't.

The ten-year gilt par yield was 0.5911% on 20 February 2020. By 9 March 2020 it was 0.1607%, and the constructed gilt sleeve had gained 4.32% while equities were collapsing. That is the hedge doing exactly what it is held for.

Then it broke. By 19 March 2020 the same yield was back up at 0.8325%, higher than where it started, and the gilt sleeve had lost 6.34% in eight business days. Investors raising cash were selling the thing that had gone up, and for a week and a half government bonds fell alongside shares. Over the whole equity drawdown window, 20 February to 16 March 2020, gilts returned just 1.11%. The cushion held on the day it was measured and very nearly didn't.

That week is the reason the Bank of England moved twice. Bank Rate went from 0.75% to 0.25% on 11 March 2020, then to 0.1% on 19 March 2020, the day the gilt selloff peaked. The second cut came alongside a decision to "increase the Bank of England's holdings of UK government bonds and sterling non-financial investment-grade corporate bonds by £200 billion to a total of £645 billion."

Sterling absorbed a chunk of the fall for a UK investor

The same index fell by different amounts depending on the currency you count in. The MSCI World index lost 33.99% in dollars and 26.04% in pounds. The gap is sterling, which fell hard in the same fortnight, and a weaker pound lifts the sterling value of overseas holdings.

So a UK investor holding unhedged global equities took 7.95 percentage points less of the covid stock market crash than a dollar investor did. That is not a free lunch. Currency works the other way round just as often. Over 2020 as a whole the same index returned 16.50% in dollars and 12.90% in pounds, so the pound gave back most of what it had lent. What it did in March was change the shape of the fall while it was happening.

Why 2022 is the reason not to count on the cushion

Here is the strongest objection to everything above, and it is a serious one. The bond sleeve worked in 2020 because yields had room to fall and because the shock was a growth shock. Neither condition is permanent.

In 2022 the S&P 500 returned -18.04% on a total return basis and a ten-year US Treasury returned -17.83%. The two legs fell together, by almost the same amount, because the shock was inflation rather than growth. A 60/40 built for March 2020 would have given you nothing in 2022. What actually held up across six separate equity falls, and how much the answer varies by the kind of shock, is the subject of the piece on diversification in a crisis.

There is a second objection, and it is about what "the market" means. The headline recovery belongs to a global index dominated by US technology. A UK investor who held only UK shares had a different 2020 entirely: the FTSE All-Share returned -9.56% for the calendar year, so that investor was still down when the year ended. The speed of the reversal was a property of one index, not of equities as a class.

The reversal was priced off two central banks, not off the virus

The sequence matters, because it is the part a reader can check. The World Health Organization made the assessment that covid-19 "can be characterized as a pandemic" on 11 March 2020. Equities in sterling bottomed five days later. The Federal Reserve cut its target range to "0 to 1/4 percent" on 15 March 2020 and committed to buying "at least $500 billion" of Treasury securities and "at least $200 billion" of agency mortgage-backed securities. The S&P 500 bottomed eight days after that, on the day the Federal Reserve dropped the cap on those purchases.

The economy didn't recover on that schedule. The National Bureau of Economic Research dates the US contraction from February 2020 to April 2020, a contraction of 2 months against an average of 17.0 months across its cycles from 1854 to 2020. That is the shortest contraction in the whole table. The next shortest ran 6 months, from January to July 1980. Markets reversed before the recession was over, which is the ordinary case rather than the strange one. The same pattern shows up in the data on bear market recovery time across earlier cycles.

The series behind these numbers, and what they cannot tell you

Everything above is checkable, so here is exactly what these March 2020 stock market crash figures are built from, and where they are weakest.

The US index figures are daily closes of the S&P 500 price index, published by the Federal Reserve Bank of St Louis. They exclude dividends, which over a six-month window is worth a fraction of a percent, and they are closing prices, so the intraday lows were deeper than anything quoted here.

The sterling equity figures are the MSCI World index, gross total return, in pounds. Gross means dividends are reinvested with no withholding tax and no fund charges, so a real holding would have done slightly worse than the index on every line of the table.

The gilt sleeve is constructed, and this is the softest part. Each day a par bond yielding yesterday's rate is repriced at today's rate, with one day of coupon accrued. That proxy returned 5.85% over 2020, against 8.50% for the FTSE Actuaries UK Conventional Gilts All Stocks index, which holds longer-dated bonds and therefore gained more as yields fell. Read the gilt column as the behaviour of a ten-year bond rather than of a gilt fund you could have bought.

Nothing is rebalanced. A portfolio topped back up to target during the fall bought equities near the low. It would have recovered faster than the table shows, and that's a separate question with its own evidence.

Finally, a month-end statement would have told you a different story. Checking only month ends, the sterling global equity holder saw -10.73% at the end of February and -20.15% at the end of March 2020. The actual worst close was -26.04%. Lower-frequency data always flatters a drawdown, because the true bottom rarely lands on the last day of a month. Tracking the real path is what a tool like LedgerTouch is for, and it is the difference between a 20% fall and a 26% one on the same portfolio.

What would change the conclusion

Three things would.

The first is the starting yield. The gilt sleeve could rally 4.32% in February 2020 because a 0.59% yield still had somewhere to go. From a higher starting yield there is more room; from a yield near zero there is almost none, and 2022 is what that looks like.

The second is the speed of the policy response. Bank Rate reached 0.1% within 28 days of the sterling peak and the Federal Reserve removed the limit on its purchases on the day US equities bottomed. A slower response, or a shock that monetary policy cannot address, breaks the link between the 23 trading days down and the 126 back up.

The third is the kind of shock. March 2020 was a growth shock with an identifiable end, which is why the recovery could be priced long before the virus was. An inflation shock, a solvency shock or a war doesn't have that property. The separate question of whether people who sell stocks in a market crash ended up better off has its own evidence, and it is not flattering.

The number worth carrying out of this is not the 33.92%. It is the 50 days. That is what 40% in gilts bought a UK investor in 2020, and in 2022 the same 40% bought nothing. The cushion is not a property of the allocation. It is a property of the crisis, and you find out which kind you are in afterwards.