Gold scenario simulator
What has gold tended to do when real interest rates and the dollar move? A transparent two-factor model on free public data. Illustrative, not advice.
Your scenario
In plain English: in past months when US real yields moved +0.50 points and the broad dollar moved 0.0%, gold moved by about −3.7% on average, but individual months were often far from that. Illustrative only; not a forecast.
Quick grid: modelled monthly gold move
| Real yield change | Dollar −5% | Dollar unchanged | Dollar +5% |
|---|---|---|---|
| −0.50 points | +9.5% | +3.7% | −1.9% |
| Unchanged | +5.8% | 0.0% | −5.6% |
| +0.50 points | +2.2% | −3.7% | −9.2% |
The model, in full
Monthly change in gold (log %, COMEX gold futures front month, month-end close) = a + b × change in the US 10-year real yield (percentage points) + c × change in the broad US dollar (log %). Ordinary least squares, one month at a time.
| Term | Estimate | Standard error | t-statistic |
|---|---|---|---|
| Constant (a): average month | +1.00% | 0.32 | 3.1 |
| Real yield (b): per +1 point | −7.32% | 1.45 | -5.1 |
| Dollar (c): per +1% | −1.14% | 0.20 | -5.8 |
Sample: Jan 2007 to Oct 2026, 171 month-on-month changes (months missing from the free gold series are skipped, not filled in). R² = 0.36: these two measures explain about 36% of gold's monthly moves. Typical miss (residual standard error): 4.2% a month. The scenario leaves out the constant, so it shows only the part linked to rates and the dollar.
Sources: gold, COMEX gold futures via Yahoo Finance, monthly, delayed; real yield: 10-year real yield (TIPS), FRED; dollar: broad trade-weighted US dollar index, FRED. Model rebuilt Mon 16 Nov, 17:45 UK.
Limits: past relationships shift, the sample includes crisis years, and a two-factor model leaves out central-bank buying, inflation surprises and geopolitics. Illustrative, not advice.