1989issue C071-6

A precious-metal price as a changing intermarket equation

Stepwise multiple linear regression estimated gold, platinum, and silver from inflation gauges and foreign-exchange rates over stated 1975-1989 samples. The coefficient of determination, the unexplained residual, and the gold-platinum ratio then showed how tightly each metal still tracked those intermarket inputs.

Stepwise multiple linear regression estimated a metal’s mean price from related series such as inflation gauges and, when they still helped, foreign-exchange rates.
For 1975-1989, consumer and producer price indices associated about 77% of gold-price variation; adding exchange rates raised that share to 85% and left a 15% unexplained residual.
In the June 1982-January 1989 platinum sample, stepwise selection discarded both inflation indices and attributed 76% of price variance to exchange-rate changes.
From January 1975 the gold-platinum ratio averaged near 0.92 and later appeared to settle below 0.8, while silver showed the weakest inflation-and-currency linkage of the three metals.

Entries in this reading3 entries

Current trackCorrelation analysis52 readings Linear regression112 readings Intermarket analysis136 readings

Estimate the metal from inflation and currency first

Intermarket analysis here means reading one metal against inflation gauges and foreign-exchange rates rather than against its own price history alone. The working tool is linear regression: a fitted linear equation that maps ordered inflation, currency, or related observations into an estimated metal price over a stated sample.

Correlation analysis then measures how tightly the metal’s path tracks those inflation and currency inputs across a defined lookback. The archive workflow used stepwise multiple linear regression to estimate a metal’s mean price from related series such as inflation gauges. Stepwise selection keeps only the intermarket variables that help explain the dependent metal price and can drop the rest.

What the gold samples associated with those inputs

For 1975-1989, consumer and producer price indices together produced a coefficient of determination of 0.7664, or about 77% of gold-price variation. The coefficient of determination is the share of price variation associated with the fitted intermarket equation over a stated sample.

Adding foreign-exchange rates associated 85% of gold-price variation with the three intermarket inputs and left 15% to other influences. That leftover share is the unexplained residual: the portion of price variation left after the selected inflation and currency variables are accounted for. Over June 1982-January 1989 the same gold specification explained 72% of price change, leaving a 28% residual.

Gold monthly average versus inflation-and-FX regression, 1975–1988

Gold’s monthly average (jagged) and the stepwise fit from CPI, PPI and foreign-exchange rates (smoother) from January 1975 through 1988. The two lines stay close through the 1980 spike and the mid-1980s slide; the leftover gap is the unexplained residual the article puts at about 15 percent. Values were read off the ratio-scale Gold plot (the second attached figure), not copied from the printed graphic.Gold · monthly · 1975-01-01T00:00:00.000Z to 1988-01-01T00:00:00.000Z

Y values are approximate readings from a ratio-scale raster; the source quotes R² = 0.85 once foreign-exchange rates join CPI and PPI, and 0.7664 for the two inflation gauges alone. Sample in the article runs through January 1989; the printed x-axis ticks end at 88-1-1.

When platinum dropped the inflation gauges

Platinum from January 1975 to January 1989 had about 80% of price variance associated with domestic inflation and exchange rates. In the June 1982-January 1989 platinum sample, stepwise selection discarded both inflation indices and attributed 76% of price variance to exchange-rate changes, alongside import dependence above 99%.

Silver and the gold-platinum ratio

Silver’s inflation-and-FX regressions explained about 57% of variance from January 1975 to January 1989 and about 66% from June 1982 to January 1989, the weakest linkage of the three metals.

From January 1975 the gold-to-platinum price ratio averaged near 0.92 and later appeared to settle below 0.8. The gold-platinum ratio is the relative price of gold versus platinum, used as a spread-style marker of whether the pair still occupies its usual range.

How the historical analysis used the equations

The historical analysis treated the fitted equations as a way to flag unusually high or low metal values, while noting that the type and weight of explanatory factors change over time.

Educational research material, not investment advice. Historical source context does not establish present-day performance.

2 of 37 in the Correlation analysis track

19901-10 pp.Next on Correlation analysisTwo clocks for copper: a factor regime, a regression baseline, and leftover moving-average timingMonthly averages were the regime clock: copper was paired with inflation, rates, the dollar, and inventories and scored from exact inverse through no measured link to exact match.

All readings on this track · 37 readings

By fieldIntermarket and spreads148 readings By year198983 readings

All 52 readings tagged Correlation analysis

Also on Correlation analysis5 readings

1990Earnings yield, rate correlation and regression for equity value 2020Bond spreads as an equity regime lens 1990Two clocks for copper: a factor regime, a regression baseline, and leftover moving-average timing 1991Constructing a two-market linear correlation check 1991Name the window, then combine leaders