What Singapore's Own Bank Says About Gold, and How That Call Has Aged
Nearly all the gold commentary that reaches Singapore is written somewhere else. New York, London, Zurich. It is worth reading what a bank headquartered here tells its own clients, and it is worth checking how that call has aged.
OCBC cut its gold forecast on 1 July. Gold has spent September trading above the near-term number in that forecast. Both halves of that sentence are useful.
What OCBC actually changed
The bank's strategists Sim Moh Siong and Christopher Wong lowered the end-2026 gold target to US$4,360 an ounce, down from US$5,100, and cut silver to US$67 from US$89.50, as reported by FXStreet and Yahoo Finance.
Four reasons were given: a sharp rise in real interest rates, renewed US dollar strength, slower ETF demand, and increasingly hawkish expectations for Federal Reserve policy.
| Forecast | Was | Now |
|---|---|---|
| End-2026 gold | US$5,100 | US$4,360 |
| End-2026 silver | US$89.50 | US$67.00 |
| Gold, September 2026 average | — | US$4,180 |
| Gold, September 2027 | — | US$4,820 |
Read the bottom two rows together and the shape of the call is clear. OCBC expected gold to sag through this year and recover through next, which is a near-term downgrade sitting inside a longer-term constructive view. The bank said as much explicitly: the revision reflects a tougher short-term backdrop, not a change of mind about where gold goes over a longer horizon.
How the call has aged
Here is the part no bank publishes about itself.
OCBC put September 2026 gold at an average of US$4,180. Gold has spent this September trading roughly between US$4,300 and US$4,500. On 1 September it was US$4,383.59, which we recorded in the September gold price update, and after the Fed's rate rise it was still around US$4,311 on 17 September.
So the near-term figure was too low, and by a reasonable margin. Meanwhile the end-2026 target of US$4,360 is approximately where gold is sitting right now, which means the bank is effectively forecasting the rest of the year to be flat.
And the reasoning? That part was right. OCBC named hawkish Fed expectations as a driver in July, and on 16 September the Fed raised rates for the first time since 2023, to 3.75% to 4.00%, with projections pointing to at least one more rise before year-end. We covered that in what the Fed hike did to gold.
The mechanism was correct and the number was wrong. That is worth understanding properly, because it is the normal outcome for forecasts, not an embarrassing one.
Why a forecast can be right about why and wrong about how much
Predicting that higher real yields pressure gold is an argument about cause and effect, and it held. Predicting the exact level gold lands on in December also requires calling the oil price, the dollar, ETF flows, central bank buying and whatever geopolitics does in the meantime.
The first is analysis. The second is closer to a guess with a decimal point on it.
Which is why a price target is the least useful part of any bank note, and the reasoning is the most useful. OCBC has been specific about what would make it more positive again: de-escalation between the US and Iran, lower oil, softer bond yields, and a more dovish Fed. That is a checklist you can actually watch, and it beats a number you cannot verify until the year is over.
Why a Singapore bank is worth reading anyway
Not because local means correct. OCBC's September number proves it does not.
It is worth reading because the bank publishes in Singapore dollars for clients who hold them, and because the currency leg matters as much as the metal leg here. A forecast that gold falls in US dollars can still mean gold rises in Singapore dollars if the exchange rate moves the other way. We went through that gap in the September update, where gold's US dollar move and its Singapore dollar move came out very differently over the same stretch.
What a forecast cut means if you are buying a chain
Almost nothing, and we would rather say that plainly than dress it up.
A bank target is a view on where an ounce of metal trades in December. It is not advice about whether to buy a bracelet in September, and OCBC is not writing for that reader. What your piece costs is set on the day:
Price = (gram weight × purity × the day's gold rate) + workmanship + 9% GST
The forecast does not enter that formula anywhere. What does enter it is gram weight, purity, workmanship and the rate that morning, and three of those four are things you choose. We ran the arithmetic on a fixed budget in what $500 buys in 916 gold.
The reframe: read the reasoning, ignore the number
Bank forecasts get revised. OCBC's have moved more than once this year, and the ones that were cut in July are already sitting below the market in September. That is not a criticism of OCBC, whose analysis of the drivers has held up well. It is the nature of putting a decimal point on a guess about December.
So use the notes for what they are good for. The list of things that would change a bank's mind is genuinely informative. The target price is the least durable sentence in the document.
We do not publish price forecasts at JJ Gold, and we are not going to start. What we do is tell you the day's rate and the exact gram weight and purity of what you are looking at, so you can check the arithmetic yourself before you commit. Every piece is real 916 or 999 gold, never plated, weighed in front of you and hallmarked, with weight and purity on the receipt. If a piece sits past what you wanted to spend in one go, it can go on Shop Now, Pay Later with $0 down and up to 12-month instalments through Atome, PayLater by Grab, or SPayLater by Shopee, and we compare what each costs in our instalment guide.
OCBC cut end-2026 gold to US$4,360 and silver to US$67 on 1 July. Its September figure of US$4,180 has already been passed, and its reasoning about a hawkish Fed was proven correct on 16 September.
Call 8950 4377 for the rate on the day before you make the trip, or come to City Plaza #01-47 (810 Geylang Road, Singapore 409286) or Lucky Plaza #02-47 (304 Orchard Road, Singapore 238863).
Read why the bank thinks it. Then go and weigh something.