The equity fall was the headline. The mechanism was the funding market. Once businesses started drawing credit lines and funds started facing redemptions, institutions needed cash immediately, and they raised it by selling whatever had a bid. That meant Treasuries, gold and investment-grade credit went down alongside equities for several sessions, which is not what any correlation table predicts.
For a retail trader the visible symptoms were narrower but sharper: spreads several times normal width, order rejections, and stop fills far from the level. The underlying cause was the same. Market makers were pricing for the risk that they could not offload the position they were about to take on.
The timeline that matters to execution
Late February saw the first serious equity fall. The week of 9 March brought the oil price war between Saudi Arabia and Russia, which knocked crude down sharply in a single session and dragged energy credit with it. Circuit breakers halted US equity trading on multiple days that month. On Sunday 15 March the Federal Reserve cut its target rate to near zero in an unscheduled announcement and restarted large-scale asset purchases. Futures still opened limit down.
The extreme point came later, on 20 April, when the expiring WTI May futures contract settled below zero because holders had no storage and had to pay to hand over the barrels. That was a contract-specific storage problem rather than a view on oil, and it caught retail traders who did not know their instrument rolled or expired.
What happened to spreads and why
A quoted spread is the market maker's compensation for holding inventory it did not choose. In March 2020 the time it took to offload that inventory rose sharply, so the compensation rose with it. Majors that normally quote well under a pip were quoting several pips at times. Exotic pairs, indices and gold widened much further. Overnight sessions were worse than London and New York, for the reasons set out in liquidity in forex.
Two practical consequences follow. A strategy whose edge is smaller than the widened spread stops being a strategy for the duration, and scalping approaches with a few pips of expected gain were simply not viable for weeks. And any stop placed a few pips from entry was being triggered by the spread rather than by the market, because the stop on a long position executes against the bid while the chart shows you something else.
Gold falling with equities in March 2020 is the cleanest available reminder that safe-haven behaviour is conditional. In a scramble for cash, an asset is sold because it is liquid, not because it is disliked.
Correlation is not a constant
Retail risk models usually assume that four positions in four different instruments are four separate bets. In a funding squeeze they are one bet, expressed four ways. Traders who had sized each position at what felt like modest risk found that a single session moved all four against them together, because the driver was liquidity rather than anything specific to any instrument.
The fix is not to abandon diversification. It is to compute worst-case aggregate exposure on the assumption that correlations go to one during stress, and to check that the resulting number is survivable. If it is not, the portfolio is too large regardless of what the historical correlation matrix says. That test belongs in the same place as the rest of your risk management rules, and it is closely related to how currency correlations behave in normal times versus stressed ones.
There is a timing detail inside the correlation point that costs people money. Correlations do not rise gradually as stress builds. They snap, usually within a single session, at the moment funding pressure appears. A portfolio that looked well spread on Friday can be one position on Monday without anything in your holdings having changed. This is why a stress test run once a quarter on historical numbers is close to useless as a live control, and why the useful version of the test is a fixed rule: cap total open risk at a number you would accept losing in one day if every position moved against you together.
What the episode changed on the firm side
Brokers learned a set of lessons that shaped the following years. Margin requirements on indices and energy were raised and in many cases never fully returned to pre-2020 levels. Several firms restricted opening new positions in specific contracts near expiry after the negative oil settlement, because a client cannot be expected to manage physical delivery mechanics. Risk desks that had been running static exposure limits moved to intraday monitoring, since the useful window for reacting had shrunk from days to minutes.
Traders can take a version of the same discipline. Know whether your instrument is a cash CFD or a dated contract, and know when it rolls. Understand that a broker raising margin mid-crisis will force a reduction if your account cannot cover it, so free margin is a buffer against the broker's rules as much as against the market. And accept that during a liquidity event, requotes and rejected orders are the system functioning as designed, not a firm acting against you, a point covered in more detail in requotes and execution.
The last lesson is the least comfortable. Between 9 March and the end of the month, the correct action for most retail traders was to reduce size and wait. Almost nobody does that, because volatility looks like opportunity when it appears on a chart and looks like something else when it arrives in your account. Leveraged trading carries a high risk of loss, and it is highest exactly when the moves look most attractive.
"Diversification is a promise made in calm markets. In a funding squeeze, everything you own is the same trade."
— Roman Onta, Executive Director, SINGUARD
Key Takeaways
- The March 2020 sell-off was a scramble for cash, so Treasuries and gold were sold alongside equities.
- Spreads widened because market makers could not offload inventory, which broke any edge smaller than the new spread.
- Correlations went to one, turning four modest positions into one large one.
- The negative WTI settlement on 20 April was a contract expiry and storage problem, not a directional view on oil.
Frequently Asked Questions
Why did gold fall during the March 2020 crash?
Institutions needed cash and sold what still had a firm bid. Gold is highly liquid, so it was among the first things sold, which pushed it down with equities before it recovered later in the year.
Why were spreads so wide in March 2020?
A spread compensates a market maker for holding a position it may not be able to offload quickly. When that offload time rose sharply, quoted spreads widened across majors, indices, gold and energy, with the worst pricing in thin overnight hours.
How did oil trade at a negative price?
The expiring WTI May contract required physical delivery at Cushing, and storage was full. Holders who could not take delivery had to pay someone to take the contract, which pushed the settlement below zero on 20 April 2020.
About the Author
Roman Onta is an Executive Director at SINGUARD. He builds the Prop Firm CRM, the Broker CRM, Scalegram and CopySignals side by side with his brother Alex Onta, and he helped on the design of eTrader, the division Alex built and leads. His ground is worldwide payment processing, AML compliance and the corporate structures brokers are built on, work the two of them carry together, shaped by executive roles in the UAE and international corporates. He lives and works in Dubai for most of the year. Meet the executive duo leading Singuard's five divisions.