5%, 10% or 20% Gold? Testing Different Allocation Sizes Against the Same Portfolio
6 min read
Loading...
6 min read
"Put 5 to 10% of your portfolio in gold" is the most repeated line in personal finance content. Almost nobody shows what that 5% or 10% actually does when your stocks are falling. A gold allocation only earns its place if it visibly softens the fall, and the only way to know how much it softens is to run the numbers on a real crash, not a guess.
So we did. We took a real, sourced stress period, the IHSG's COVID-19 crash from January to March 2020, ran the same hypothetical stock-heavy portfolio through it three times with 5%, 10% and 20% gold added, and compared what each one actually cushioned.
The IHSG closed at 6,325 on 14 January 2020, its high point for the year, according to CNBC Indonesia's own retrospective on the period. By the close on 24 March 2020, it had fallen to 3,937.63, per Kontan's market report from that day, a trading session that also saw a 30-minute halt after the index dropped 5% in one afternoon.
That is a peak-to-trough decline of roughly -37.7% in about ten weeks. It is one of the sharpest, best-documented drawdowns the IHSG has ever had, which makes it a fair stress test: real dates, real closing levels, no cherry-picking a friendlier window.
Antam's 24-karat gold bar traded at Rp955.000 per gram on 14 January 2020, per Kontan and Tribunnews price reports from that day. On 24 March 2020, the same bar was priced at Rp919.000 per gram, according to Tribunnews's report that day.
Over the full ten-week window, that is a decline of about -3.8%, essentially flat next to the IHSG's -37.7%. That gap, not a positive gold return, is where the cushioning comes from.
Take a hypothetical portfolio that starts 100% in IHSG-tracking stocks, and imagine three versions of it built on 14 January 2020, identical except for how much gold sits alongside the stocks. Everything else, including that no rebalancing happens during the stretch, is held constant so the allocation size is the only variable being tested.
Portfolio return = (stock weight × stock return) + (gold weight × gold return)
| Gold allocation | Stock weight | Blended return, 14 Jan to 24 Mar 2020 | Cushioning vs. 100% stock |
|---|---|---|---|
| 0% (all stock) | 100% | -37.7% | none |
| 5% | 95% | -36.0% | 1.7 percentage points |
| 10% | 90% | -34.3% | 3.4 percentage points |
| 20% | 80% | -31.0% | 6.8 percentage points |
A bigger gold allocation gave a measurably smaller loss in this one real stress period, and the relationship was close to linear because gold's own return was small and stable relative to the size of the stock crash. That part of the math is not in dispute.
What it does not tell you is that this ratio holds in every crash. Gold's behaviour depends on why stocks are falling. In 2020, the trigger was a pandemic shock with a temporary liquidity squeeze, which is why gold dipped briefly before recovering. A crash driven by a currency crisis, a sovereign debt problem, or a shock that also hits gold demand directly could produce a very different cushioning number, smaller, larger, or briefly negative if gold gets sold for cash at the same moment stocks do, exactly as it was in mid-March 2020.
Three things worth checking before deciding on your own allocation, none of which require guessing where gold or stocks go next.
You can see the current rupiah and interest-rate backdrop that shapes moves like this on the macro context page. NetWort tracks gold, silver, platinum and palladium holdings alongside your stocks and crypto in one portfolio, so you can see your own blended drawdown during a stress period instead of assuming a number from a rule of thumb.