Gold During Recessions: Dated Historical Performance
Gold did not rise in every U.S. recession, and a positive end-to-end change could conceal a substantial loss along the way. This study uses seven NBER peak-to-trough windows from 1973–2020, World Bank monthly gold prices and matching BLS CPI-U observations. It measures price changes, not returns earned by actual Gold IRA owners.1, 2, 3
Updated October 12, 2026. Educational analysis, not financial, legal or tax advice. Past performance does not guarantee future results. See our affiliate disclosure.
The Seven Windows and What They Measure
The NBER chronology identifies business-cycle peaks and troughs. The recession begins in the month after the peak and ends in the trough month. We use the peak month’s average price as the baseline and the trough month’s average as the endpoint. Thus “November 1973–March 1975” means comparing those two monthly observations; it is not a claim that November was the first contraction month.3
World Bank’s historical series used the London afternoon fixing before June 2025. These recession windows all precede its switch to a spot average of daily rates. Prices are nominal U.S. dollars per troy ounce, averaged within each month. The CPI series is all items, all urban consumers, U.S. city average, not seasonally adjusted. No missing month is filled with a later observation.1, 2
| Peak → trough months | Months elapsed | Gold price change | CPI change | Gold real change |
|---|---|---|---|---|
| 1973-11 → 1975-03 | 16 | +87.37% | +14.81% | +63.19% |
| 1980-01 → 1980-07 | 6 | -4.74% | +6.30% | -10.38% |
| 1981-07 → 1982-11 | 16 | +1.47% | +6.99% | -5.16% |
| 1990-07 → 1991-03 | 8 | 0.00% | +3.53% | -3.41% |
| 2001-03 → 2001-11 | 8 | +4.94% | +0.68% | +4.23% |
| 2007-12 → 2009-06 | 18 | +17.81% | +2.69% | +14.72% |
| 2020-02 → 2020-04 | 2 | +5.39% | -0.88% | +6.33% |
Calculations use the same months throughout: nominal change = end gold ÷ start gold − 1; real change = (end gold ÷ start gold) ÷ (end CPI ÷ start CPI) − 1. Figures are rounded to two decimals. In these seven windows there are five positive nominal changes, one negative and one flat. Only four show positive CPI-adjusted changes. These counts describe this measurement convention, not the chance that gold rises in the next recession.1, 2, 3
Why 1980 and 1981–1982 Tell Different Stories
January–July 1980 shows a negative nominal and real change. July 1981–November 1982 shows a small positive nominal change but a negative real change: prices rose less than the CPI basket. Combining the two recessions into one “Volcker era” interval would answer a different question and change the figures.1, 2, 3
The difference also shows why a claim that gold “beat inflation during a recession” needs both a time window and a denominator. A gain in dollars does not automatically preserve purchasing power. Equally, a flat nominal price can mean a real loss when consumer prices rise. None of these comparisons includes storage, administration, purchase premiums or sale discounts.
A Positive Endpoint Is Not Protection Along the Way
For each window, the following calculation tracks the running highest monthly average and the largest subsequent percentage drop before the trough. It is a monthly-resolution maximum drawdown, not the largest intraday or daily loss. It can understate the stress experienced by someone who had to sell during a particular trading day.1
| Window | End-to-end gold change | Largest monthly-average drawdown |
|---|---|---|
| 1973-11 → 1975-03 | +87.37% | -16.86% |
| 1981-07 → 1982-11 | +1.47% | -29.05% |
| 2007-12 → 2009-06 | +17.81% | -21.38% |
| 2020-02 → 2020-04 | +5.39% | -0.31% |
The 2007–2009 window is a useful example: gold finishes higher while its monthly-average path still contains a sizeable decline. An investor needing cash before the endpoint could have a different experience from someone measuring only December 2007 and June 2009. The monthly data do not establish the cause of the decline or promise a subsequent recovery.
Why There Is No Seven-Recession Stock Scoreboard Here
Our retrieved S&P 500 series is an official S&P Dow Jones price index, distributed through FRED with a rolling ten-year daily history. It does not cover most of these older windows and excludes dividends. We therefore do not splice unsupported old stock figures into this table or compare gold’s end-to-end move with an equity crash’s worst daily loss.4
For a like-period recent comparison, see gold versus the S&P 500 in 2026. That article aligns monthly averages and explains why even a valid price comparison is not a total-return comparison. The market analysis hub separates this historical evidence from interest-rate mechanisms and forward-looking scenarios.
What a Gold IRA Holder Can Take From the Evidence
Use the table to ask whether a planned cash need could occur during an interim loss, not simply whether gold ended higher in one earlier recession. The account’s realizable value depends on the exact metal, executable dealer bid, cash balance and charges. A statement valuation is not an offer to buy the metal from you.
Our fee definitions and cost math distinguish account charges from product pricing. The dated provider-fee comparison helps turn that distinction into an itemized quote. Neither the historical price table nor a tax wrapper supplies a future sale price.
For withdrawal pressure and custody questions rather than historical window selection, use how recession affects Gold IRA values. Selling inside an IRA and distributing money to its owner are different steps. Traditional IRA distributions generally are taxable apart from applicable basis or exclusions; qualified Roth IRA distributions are tax-free. Those rules do not insure the asset against losses.10
Frequently Asked Questions
Does this prove gold is a reliable recession hedge?
It shows mixed outcomes under a transparent definition, including negative real changes and interim losses. Seven historical episodes are too few and too different to turn the frequency count into a forecast probability. A recession label is also retrospective, not a real-time purchase signal.
Why do other recession tables report different returns?
They may use daily closes, month-end prices, annual averages, the first contraction month, or an equity-market peak instead of the NBER peak. A difference is not necessarily an error, but the conventions must be disclosed before comparisons are meaningful.
Are these net Gold IRA returns?
No. These are benchmark price-level changes before account charges, product premiums, sale discounts, taxes and investor cash flows. The amount an actual account earns cannot be inferred without its statements and transaction records.
Sources
This article is for informational purposes only and does not constitute financial advice. Gold IRA Path may receive compensation through affiliate links. Past performance does not guarantee future results. Consult a qualified financial advisor before making any investment decisions.
Editorial Team
The shared byline for Gold IRA Path research, editing, and ongoing content updates.