Owning gold because its price is expected to rise is a market call. Holding gold because it changes the behaviour of the broader portfolio is an asset-allocation decision. The portfolio question is therefore more interesting than whether gold will outperform over the next six or twelve months.

Christopher Wong
Executive Director,
OCBC Group Research,
OCBC

Gold is difficult to value in the conventional sense. It has no earnings, no coupon and no stream of cash flows to discount. Yet investors have treated it as a store of wealth for centuries, while central banks continue to hold it as a reserve asset.

The puzzle is straightforward. Most financial assets derive their value from what they produce or promise to pay. Gold does neither. So, what are investors actually paying for?

Part of the answer may lie precisely in what gold lacks. Physical gold has no issuer, no maturity and no default risk. Its lack of yield is obvious. The value of not depending on somebody else’s balance sheet is harder to quantify.

Gold may not be the hedge we think it is

Gold is commonly described as an inflation hedge, a dollar hedge or a safe haven. Each description contains some truth, but none works consistently.

Over very long periods, gold has broadly preserved purchasing power. That does not make it a reliable hedge against inflation over normal investment horizons. Higher inflation can also bring tighter monetary policy, higher real yields and a stronger currency - all of which can work against gold.

Its behaviour during periods of market stress is similarly uneven. During the Russia/LTCM episode in 1998, the MSCI World Index fell around 14% from peak to trough, while gold lost about 7%. During the acute Covid sell-off in early 2020, the MSCI World Index fell by 34%, while gold declined by less than 1% and global bonds by around 3%.

In other episodes, the picture was very different. Gold gained around 16% through the Global Financial Crisis and close to 6% during the equity sell-off in late 2018.
There is no single “crisis trade”. Gold’s response depends on what is driving the stress and, just as importantly, what happens to rates, the US Dollar and liquidity.

Exhibit 1 — Gold’s defensive behaviour depends on the nature of shock
Selected MSCI World drawdowns and corresponding returns in gold and global bonds

Note: Returns are measured over the MSCI World Index’s peak-to-trough dates during selected market-stress episodes. The MSCI World and Bloomberg Global Aggregate indices refers to the USD unhedged total return Index; “Gold” is the spot price of the precious metal.

Source: Bloomberg, OCBC Group Research.

Different shocks, different defences

This distinction matters because gold does not need to be the best defensive asset in every downturn to have a role in a portfolio.

The 2022 experience is a useful example. Between the January equity peak and the October trough, the MSCI World Index fell by around 26% while global bonds fell more than 20%. Gold declined by around 7.8%.

In this case, gold hardly behaved like a textbook safe haven. Yet when both equities and bonds were under pressure and gold posted materially less losses which can matter in an investment portfolio.
That is a different way of thinking about diversification. It does not require gold to rise every time equities fall. What matters is whether its return drivers are sufficiently different to improve the behaviour of the broader portfolio when traditional diversification is under strain.

What are investors actually buying?

Gold combines a somewhat unusual set of characteristics. It is liquid, scarce, durable and free of credit risk. Unlike a bond or deposit, its value does not depend on an issuer meeting a future obligation.

That helps explain why official-sector demand remains relevant. Central banks bought more than 1,000 tonnes of gold annually from 2022 to 2024. Purchases moderated to 863 tonnes in 2025 but remained well above the 2010-2021 average of 473 tonnes. The slowdown is also instructive: strategic buyers are not completely insensitive to price.

The longer-term preference for gold nevertheless remains evident. In the World Gold Council’s 2026 survey, 89% of responding reserve managers expected global central-bank gold holdings to rise over the following 12 months, while 45% expected their own institutions to increase holdings.
This should not imply the demise of the dollar or the bond market. Rather, it reinforces a simpler point: there remains demand for an asset whose return characteristics are different from conventional financial claims.

From owning gold to building a portfolio

The portfolio question is therefore more interesting than whether gold will outperform over the next six or twelve months.

We tested this by comparing a conventional global 60/40 portfolio with a portfolio that reallocates five percentage points each from equities and bonds into gold. The first holds 60% MSCI World Index and 40% Bloomberg Global Aggregate Index. The second holds 55% MSCI World, 35% Global Aggregate and 10% gold. Both are rebalanced annually, with the exercise running from January 1990 to September 2026.

Exhibit 2 — Adding gold to a global portfolio, 1990–2026
Growth of US$100, annual rebalancing

Note: Portfolios are in USD and rebalanced annually to target weights. 60/40 comprises 60% MSCI World Index and 40% Bloomberg Global Aggregate Index; 55/35/10 comprises 55% MSCI World Index, 35% Bloomberg Global Aggregate Index and 10% gold. Initial value indexed to 100. Volatility is annualised over the full Jan 1990–Sep 2026 sample. Maximum drawdown is based on daily portfolio NAVs and deepest drawdown occurred during the GFC. Past performance is illustrative and is not indicative of future results.

Source: Bloomberg, OCBC Group Research.

Over the full period, the conventional portfolio generated a compounded annual return of around 6.9%. The portfolio with gold returned about 7.1%. The difference in return is small and should not be over-interpreted - results will vary with the sample, allocation and rebalancing convention.

The more interesting result is in the path taken to get there. Annualised volatility fell from about 10.6% to 9.9%, while the maximum drawdown narrowed from roughly 38% to 34%.

In other words, within this historical sample, introducing a modest allocation to gold improved some of the portfolio’s risk characteristics without an obvious long-run return penalty.

Exhibit 3 — The portfolio with gold generally held up better during stress periods

Note: Portfolio returns are measured over the same MSCI World Index peak-to-trough periods shown in Exhibit 1. Portfolios are rebalanced annually.

Source: Bloomberg, OCBC Group Research.

The result is not an argument that 10% is an optimal allocation. Nor does it mean gold will necessarily improve every portfolio over every period. It illustrates something narrower, that an asset can be an imperfect hedge against individual risks and still be useful at the portfolio level.

Owning gold because its price is expected to rise is a market call. Holding gold because it changes the behaviour of the broader portfolio is an asset-allocation decision.

Valuation still matters

None of this makes gold a one-way investment.

Real yields, the dollar, investor positioning and valuation still matter. Central-bank demand can slow, ETF flows can reverse and gold itself can experience sizeable drawdowns. After a powerful multi-year repricing, the price investors pay for those diversification characteristics matters too.

The portfolio case therefore does not require gold to outperform equities, bonds or cash. Nor does it require gold to rally during every market shock.

Its value lies in having a different set of return drivers - particularly during periods when the usual equity-bond relationship does not work as expected.

For long-term investors, that may be the more useful way to think about gold: less as a prediction of the next crisis, and more as a strategic diversifier within the portfolio.