What the sleeve did in your scenario
The test runs your portfolio with and without a metals sleeve through one of four named US downturns and reports the difference in ending value. Four windows is a small sample, so the readings describe what happened rather than what will.
- The sleeve helped Worth more than 2% of the starting value
- The financial crisis is the clearest case in the record, where gold rose while the S&P halved, and the 2022 inflation shock is the second. This is the deep, slow drawdown a metals sleeve is held for, and it is the event that makes the case for holding one before it is needed.
- The sleeve barely moved the result A difference under 2% either way
- The dot-com decline lands here, because gold mostly sat still while equities fell over thirty-one months. A sleeve that does nothing in a given window has still cost only its own carry, which is the trade every insurance-shaped holding makes.
- The sleeve cost you The portfolio without it ended ahead
- The first weeks of the covid crash are the example, when a liquidity shock forced investors to sell what they could sell rather than what they wanted to sell, and gold is highly liquid. It recovered within weeks, so a hedge held on a two-month view worked where one measured over a fortnight did not.
What does it mean to call gold a hedge?
The hedging claim behind most gold IRA pitches is specific: that gold rises when equities fall, so a slice of it cushions a portfolio through a downturn. It is a testable claim rather than a philosophy, and the way to test it is to run a portfolio with and without a metals sleeve through downturns that actually happened.
Across the four windows here the claim holds in two of them. Through the financial crisis gold rose while the S&P halved, which is the behaviour a hedge is held for and the clearest case of it in the record. Through the 2022 inflation shock the sleeve helped moderately. In the dot-com decline gold mostly just sat still, and in the first weeks of the covid crash it fell alongside everything else.
That last one is the useful failure. A liquidity shock forces investors to raise cash quickly and they sell what they can sell rather than what they want to sell, and gold is highly liquid. It recovered within weeks, so it hedged on a two-month view and not on a two-week one, which is a real property and not the one described on a sales call.
Why hold a metals sleeve in a retirement portfolio
A 60/40 portfolio is built on the assumption that when stocks fall, bonds hold up. That assumption held for most of four decades and it failed in 2022, when both legs lost ground together because the thing that hurt them was the same thing, and a portfolio with no third kind of holding had nowhere to go.
Gold is the most accessible holding that is neither. It is not a claim on a company's earnings or a government's coupon payments, so it is priced by a different set of forces, and that is the property a sleeve is bought for rather than any expectation that it will outperform. The World Gold Council's multi-asset research puts the strategic allocation at 5% to 10% on exactly this reasoning, measured on risk-adjusted returns.
The record across our four windows is that the sleeve does its heaviest work in a deep, slow drawdown. Through the financial crisis gold rose while the S&P halved, and through the 2022 shock it softened a decline that hit both conventional legs. In the dot-com decline it mostly sat still, and in the first weeks of the covid crash it fell with everything else as investors raised cash, recovering within weeks.
That pattern is the argument for buying a sleeve before it is needed. It costs 5% to 8% in dealer markup to establish and it is the kind of holding that looks pointless for years at a time, which is the same shape as every other form of insurance, and the two windows where it worked are the two that did the most damage to everything else.
How to run the stress test
Pick a window, set the sleeve, and read the covid result with its caveat attached.
-
Pick the window
Four are available: the dot-com crash from March 2000 to October 2002, the financial crisis from October 2007 to March 2009, the 33-day covid crash in early 2020, and the 2022 inflation shock. Each has its own causes and its own answer.
-
Set the metals sleeve
The slider runs from 0% to whatever share you want to test. Everything not in metals sits in the equity index, and the mix is set at the start of the window and left alone.
-
Compare the paired bars
Each window shows the portfolio with the sleeve against the same portfolio without it. Where the sleeve helped, the bars separate. Where it did not, they sit almost on top of each other, and the size of that gap is the whole answer.
-
Read the covid window with the caveat attached
Every figure here is measured on month-end closes, and covid is where that matters most. The S&P fell 34% intraday across 33 days in early 2020 and closed March a long way above its 23 March low, so a month-end series reports a far shallower fall than the one people lived through.
What each input means
Running all four windows tells you more than picking the one that matches your worry.
- Recession window
-
Which of the four downturns to run. The dates are month-end boundaries around the market decline rather than the NBER peak-to-trough dates, since the market and the economy did not turn on the same day in any of the four.
Where to find it Pick the one that matches the scenario you are worried about. Running all four is more useful than picking a favourite.
- Metals sleeve
-
The share of the portfolio in gold, with the rest in the equity index. Set once at the start of the window and held, so the drift you see inside the window is real rather than smoothed away by rebalancing.
Where to find it Try 0%, 10% and 20% in turn, because the differences between those three are the whole finding.
- Starting value
-
The portfolio value at the start of the window, which scales the dollar outputs and leaves every percentage alone.
Where to find it Your own balance, or leave the default.
How each window is calculated
Fixed dates, a fixed mix set at the start, and the same window run again with the sleeve turned off.
Each window is a fixed pair of month-end dates around a market decline. The portfolio is split between gold and the equity index at the start, then carried forward on monthly closes with no rebalancing, and the same window is run again with the sleeve set to zero. What is reported is the ending value, the deepest fall inside the window, and the difference the sleeve made. The equity leg reinvests dividends, which is the honest basis for a comparison across a period where dividends were being paid throughout.
ending value = (start × sleeve × gold multiple) + (start × (1 − sleeve) × equity multiple)
- sleeve
- Share of the portfolio in gold, set once at the start of the window
- multiple
- End-of-window value divided by start-of-window value for each leg
What this test leaves out
- Every account cost. The custodian fee, the storage charge, the dealer spread over spot and the tax treatment all take a real bite out of what a gold IRA returns and none of them is here.
- The premium and the buy-back discount. A real account holds coins bought above spot and sold below it, so the round trip is paid twice before any window opens.
- Anything inside a month, and the covid window is where that matters most, since the S&P fell 34% intraday and month-end closes cannot see it.
- Bonds, cash and everything else a real portfolio holds, since this runs two assets against each other and nothing more.
- Any forecast, because four windows with four different causes describe what happened rather than what comes next.
Three downturns, with and without gold
The window where the hedge worked, the window where it half worked, and the window where it did not.
The financial crisis, 20% sleeve
October 2007 to March 2009, the deepest equity decline in the set.
- Window
- Global financial crisis
- Sleeve
- 20% gold
- Start
- $100,000
The sleeve reduced the loss materially, and gold ended the window up.
This is the window the whole hedging argument rests on, and it holds. Gold rose while equities fell by half, which is the deep drawdown a metals sleeve is bought for and the clearest case of it in the record.
The dot-com crash, 20% sleeve
March 2000 to October 2002, a slower decline over two and a half years.
- Window
- Dot-com crash
- Sleeve
- 20% gold
- Start
- $100,000
The sleeve helped, mostly by not falling rather than by rising much.
Gold was coming off a twenty-year low here and did little in absolute terms. The sleeve's contribution was that a fifth of the portfolio sat out the decline, which is a weaker claim than the one usually made for it and it is still a real one.
The covid crash, 20% sleeve
February to March 2020, 33 days.
- Window
- COVID 33-day crash
- Sleeve
- 20% gold
- Start
- $100,000
Almost no separation between the two bars on month-end data.
Gold sold off alongside everything else in the middle of March as investors raised cash, which is what it usually does in the first days of a liquidity shock before it recovers. A hedge that arrives a fortnight late is a real hedge and it is not the one described on the call.
The four windows, and what they contain
Four figures cover the sample size, the measurement limit and the cost a hedge has to clear before it starts helping.
- 4 Windows in the sample
Dot-com, the financial crisis, covid and the 2022 inflation shock. Four is a small sample and each had its own causes, so the results describe what happened rather than what will.
- 34% The S&P's intraday covid fall
Across 33 days in February and March 2020. A month-end series reports far less, because the index closed March a long way above its 23 March low.
- 5% to 8% What the sleeve costs to buy
The dealer markup on IRA-approved metal, paid before any of these windows begins. A hedge has to clear its own entry cost before it starts helping.
- 2 of 4 Windows where the sleeve clearly helped
The financial crisis strongly and the 2022 inflation shock moderately. The dot-com window was mild and covid was close to neutral on this data.
What people get wrong about gold as a hedge
The first is the claim itself, and the rest are about what a four-window sample can carry.
Does gold always rise when stocks fall?
No, and the exceptions land at the worst moments. It rose through the financial crisis, and in the first weeks of the covid crash it fell alongside everything else, because a liquidity shock makes investors sell whatever they can sell, and gold is easy to sell. The correlation is not fixed and it moves most at exactly the moments a hedge is supposed to be working.
Is four downturns enough to prove the hedging case?
No, it is enough to test a claim and not enough to settle it. Each of the four had different causes, from a valuation unwind to a banking collapse to a pandemic to an inflation shock, and a hedge that worked in two of them has a record rather than a property.
Does the stress test show what a gold IRA would have done?
No, it prices metal at spot and stops there. A real gold IRA pays a setup fee, an annual custodian charge, depository storage and a dealer spread over spot, and it holds coins bought above spot and sold back below it, so a rollover pays that spread twice before any window begins.
Is a bigger metals sleeve a better hedge?
No, not automatically, because a bigger sleeve moves the portfolio further in both directions, and outside a downturn that has a cost. Sizing the sleeve against your whole position is a separate question, which the allocation calculator handles against the 5% to 10% band the diversification research lands on.
Gold hedging questions
Does gold protect against a stock market crash?
Sometimes, and in two of the four windows here it clearly did, most strongly through the financial crisis when gold rose while the S&P halved. In the first weeks of the covid crash it fell alongside everything else as investors raised cash, and in the dot-com decline it mostly just sat still. The record supports a sleeve that does its heaviest work in a deep equity drawdown, which is the event it is held for, rather than one that softens every decline.
Why did gold fall during the covid crash?
A liquidity shock forces investors to raise cash quickly, and they sell what they can sell rather than what they want to sell. Gold is highly liquid, which makes it one of the first things sold, and it recovered within weeks. A hedge that works on a two-month view and not a two-week one is still doing its job, provided you are holding it for the decade rather than for the fortnight.
What is the right size for a gold sleeve?
The diversification research most often cited, including the World Gold Council's own multi-asset work, lands on 5% to 10%. This calculator will run any size you set and does not recommend one, and the allocation calculator handles the sizing question against your whole portfolio rather than against a single window.
Why only four recessions?
Because the gold price only floats from 1968 and the windows before 2000 either predate reliable monthly equity total-return data on the same basis or were not equity-market declines. Four is a small sample and we would rather name it as four than pad the set with windows that do not answer the question.
Does the equity side include dividends here?
Yes. A recession window measured without dividends would overstate the equity decline, and since the whole point of this tool is testing whether the sleeve helped, the leg it is being compared against has to be measured properly. Our gold vs stocks backtest reinvests them too.
Would a gold IRA have delivered these results?
No, and the gap is not small. Before a single window opens, a rollover pays 5% to 8% in dealer markup, and it pays a further spread on the way out. Add the setup fee, the annual administration and the storage charge and a ten-year hold takes roughly 9% to 17% of the balance, which comes off every figure on this page.