The hedge that stops hedging

Gold is sold as a safe haven in normal stress. In a liquidity crisis, it is sold for a completely different reason: it is liquid, it has unrealised gains, and it is the easiest thing on the book to turn into cash.

This is what happened in 2008 and again in March 2020. In both episodes gold fell hard alongside equities in the first phase, then rallied sharply once central banks backstopped the system. Investors who had sized gold as a portfolio hedge found it moving with their losses precisely when they needed it most.

Why it happens

The mechanism is not sentiment, it is margin. When volatility spikes across the board, margin requirements rise, and leveraged participants must raise cash. They sell what they can, not what they want to. Gold, which is deep and usually profitable, is near the top of that list. On top of that, a dollar funding squeeze strengthens the dollar, which pressures gold through the channel from Part 1.

The shape of the episode

The pattern is remarkably consistent:

Phase 1 — liquidation (days to weeks). Correlations converge toward one. Gold, equities, and industrial commodities all fall. The dollar rips.
Phase 2 — policy response. Central banks cut, backstop, and expand balance sheets. Real yields collapse.
Phase 3 — gold reprices. With real yields crushed, the dominant driver reasserts itself and gold makes its move.

So gold is a hedge — but on a lag, and only after the policy response. It is not a hedge for the first two weeks.

What to do about it

Do not size a gold long as if it is uncorrelated. In the scenario where you most need it, it is correlated.
Reduce gross exposure into known stress events rather than relying on diversification.
If you are holding gold as a hedge, hold it for phase 3 — and have the liquidity to survive phase 1.
Watch the dollar funding channel. A spiking dollar in a risk-off move is the tell that the liquidation phase is running.

How to check this yourself

Pick the last market stress event you remember — a banking scare, a flash crash, a policy shock — and pull the gold and equity charts for that window. Note the correlation. Then do the same for a calm quarter. The two numbers will differ, sometimes in sign. That gap is the whole lesson: the diversification you measured in calm markets is not the diversification you get in a crunch.