Gold closed at a record $5,405 an ounce on January 29, 2026, touching $5,595.47 intraday. Five months later, on June 25, it settled at $4,001.80 — roughly 25% below the peak and down about 7% for the year at the midpoint. This week it has come back: gold traded near $4,627 on August 25, a three-month high, about 16% above the June low and still roughly 14% under January's record close.
Most coverage has framed that round trip as a story about direction. The more useful number for a retirement portfolio is a different one.
The Drawdown Was Ordinary. The Volatility Wasn't.
The World Gold Council's own history is blunt about the decline. Since 1971, there have been eight episodes in which gold fell more than 20% after setting a record high. The average drawdown was 36%; the median was 29%. A 25% fall from a record is not an anomaly in gold — it is close to the middle of the distribution.
What changed in 2026 was the ride, not the destination. Realized volatility peaked above 50% during the first half of the year, against a 20-year average of about 17%. It has since cooled to below 30% on a 30-day basis — still well above normal. Measured across the full series since 1971, gold's volatility this year has sat in the top few percentiles.
This is the part that gets missed. Diversification value depends on two things: how an asset moves relative to your other holdings, and how much it moves at all. A 10% gold sleeve running at 50% volatility contributes roughly three times the portfolio risk of the same 10% sleeve at 17%. The allocation you wrote down did not change. The risk it carries did.
Central Banks Kept Buying Anyway
Official-sector demand ran in the opposite direction of price. Central banks bought a net 288.9 tonnes in Q2 2026 — the strongest second quarter in the World Gold Council's series — and 533 tonnes across the first half, led by Poland at 51 tonnes and the People's Bank of China at 33 tonnes, its largest quarterly addition since late 2023.
That is worth understanding correctly. Central banks buy on multi-year reserve mandates and are largely price-insensitive. Their demand is a slow structural support, not a signal about the next two quarters — which is exactly why it kept rising while the price fell 25%.
Practical Takeaways
- Size the sleeve to the volatility, not the thesis. Research generally supports a 5–15% allocation for diversification. At 2026's volatility levels, the same percentage is doing considerably more work in your risk budget than the backtest assumed.
- Set rebalancing bands before you need them. Mechanical bands — trim at +5 points over target, add at −5 — turned this year's swing into a source of returns. Discretion mostly turned it into regret.
- Expect the drawdown you have already been shown. A median 29% decline after a record high is the base rate. If a repeat would force you to change your retirement date or spending, the position is too big today.
- Account for gold IRA friction. Dealer spreads and depository procedures make metals slower and costlier to rebalance than an ETF or fund. A wider band is often more realistic than a tight one you cannot execute.
- Treat forecasts as ranges. The Council's own half-year scenarios span +5% to +20%, −5% to +5%, and −5% to −15%, with a baseline of roughly ±5% around $4,100/oz. Analyst year-end targets run from about $4,360 to $4,900. That spread is the honest picture.
Gold did its job as a long-horizon diversifier this year. It just did it while behaving like a risk asset — and for anyone within a few years of drawing income, that distinction is the one worth planning around.
Sources: World Gold Council – Gold Mid-Year Outlook 2026: Point break; World Gold Council – Gold Demand Trends Q2 2026; World Gold Council – Gold Price Volatility data series; USAGOLD – Daily Precious Metals Market Report (August 25, 2026); GoldSilver – Gold Price Outlook August 2026