THE SHORT ANSWER
Gold has shown calendar differences, but no month owns a dependable forecast
Grouping 798 adjacent World Bank monthly-average changes by ending calendar month produces different historical averages, medians and positive-return frequencies. The strongest median month in the full sample is January at +0.95%; the weakest is February at +0.00%. Those ranks describe one long, regime-changing sample. They do not establish a recurring trade or the cause of any return.
Strictly adjacent changes from 1960-02 through 2026-07.
Middle observation, not expected return.
Full-sample rank, not a sell signal.
Not daily month-end closes.
INTERACTIVE CALENDAR VIEW
Switch between median, average and positive frequency
Use pointer, touch or arrow keys to inspect all twelve ending-month groups. The median reduces the influence of tails; the average keeps their full effect; positive frequency counts observations above zero without measuring their size.
Source: World Bank Commodity Price Data (The Pink Sheet), monthly nominal USD per troy ounce, CC BY 4.0. GoldObserve serves the local source-preserving archive.
SUBPERIOD STABILITY LAB
Did the same calendar months stay strong across different gold-market eras?
The full-history rank can hide regime changes. This matrix repeats the same twelve-month comparison inside three completed 20-year eras and the shorter 2020-to-latest period. Read across a calendar month to test persistence, or down an era to see how its ranking changes.
Median monthly change is above zero in 2 of 4 fixed historical eras.
+2.34% · rank 3 of 12 · 7 observations
Read across a month to test persistence, then down an era to compare ranks. Exact values and sample counts are printed in every cell, so color is not the only evidence. The 2020–latest row ends 2026-07 and is not comparable in length with the three completed 20-year eras.
| Era | Sample | Average | Median | Positive frequency | Rank |
|---|---|---|---|---|---|
| 1960–1979 | 19 | +2.25% | +0.00% | 42.1% | 1 of 12 |
| 1980–1999 | 20 | +2.37% | +0.00% | 45.0% | 4 of 12 |
| 2000–2019 | 20 | +2.28% | +1.92% | 75.0% | 1 of 12 |
| 2020–latest (partial) | 7 | +3.73% | +2.34% | 100.0% | 3 of 12 |
Source: World Bank Commodity Price Data (The Pink Sheet), monthly nominal USD per troy ounce, CC BY 4.0. Method: strictly adjacent monthly averages grouped by ending month and fixed era; no interpolation.
COMPLETE MONTH TABLE
Sample size, central tendency, win frequency and actual extremes
| Ending month | Observations | Average | Median | Positive frequency | Worst observation | Best observation |
|---|---|---|---|---|---|---|
| January | 66 | +2.45% | +0.95% | 59.1% | 1982-01 / -6.34% | 1980-01 / +48.35% |
| February | 67 | +1.28% | +0.00% | 46.3% | 1981-02 / -10.23% | 1974-02 / +16.28% |
| March | 67 | +0.06% | +0.00% | 43.3% | 1980-03 / -16.69% | 1973-03 / +13.51% |
| April | 67 | +0.49% | +0.00% | 40.3% | 1980-04 / -6.68% | 2006-04 / +9.69% |
| May | 67 | +0.65% | +0.00% | 34.3% | 1974-05 / -5.23% | 1972-05 / +12.24% |
| June | 67 | +0.06% | +0.00% | 38.8% | 2006-06 / -11.70% | 1973-06 / +17.65% |
| July | 67 | +0.05% | +0.00% | 41.8% | 1981-07 / -11.28% | 1982-07 / +7.62% |
| August | 66 | +0.78% | +0.00% | 47.0% | 1973-08 / -10.83% | 2011-08 / +11.82% |
| September | 66 | +1.25% | +0.00% | 43.9% | 1975-09 / -11.66% | 1982-09 / +20.05% |
| October | 66 | +0.91% | +0.00% | 42.4% | 2011-10 / -5.98% | 1999-10 / +17.36% |
| November | 66 | -0.07% | +0.00% | 40.9% | 1978-11 / -9.25% | 1974-11 / +14.47% |
| December | 66 | +0.49% | +0.00% | 43.9% | 1980-12 / -13.92% | 1979-12 / +16.07% |
WHY SEASONALITY CAN MISLEAD
A full-sample pattern can be a tail event, regime mix or data-definition effect
RESPONSIBLE USE
Treat a calendar pattern as a question to test, not a schedule to trade
Name the price definitionMonthly average nominal USD per troy ounce.
Compare median and averageA wide gap reveals sensitivity to extreme observations.
Inspect the extreme monthsVerify whether one historical episode dominates the rank.
Check more than one eraA stable mechanism should not depend on one arbitrary start date.
Keep prediction separateDo not convert in-sample frequency into a forward probability.
Source, formula and limitsWorld Bank monthly history · CC BY 4.0
SOURCE, LICENSE & REPRODUCIBILITY
Every result uses one consistent monthly series
The source is the World Bank Commodity Price Data (The Pink Sheet), licensed CC BY 4.0. It contains 799 monthly gold averages from 1960-01 through 2026-07. The workbook was published 2026-08-04, retrieved 2026-08-07, and recorded with SHA-256 7902a77505ebdc5d202ce65f666c2ee1b04b626f042d7738ed3e6f7d112c8433.
Monthly change = (current monthly average / prior monthly average - 1) × 100. Only adjacent calendar months are used. The result is not a daily close return, intramonth high or low, dealer quote or exact investor fill.
CONTINUE THE RESEARCH
Move from calendar grouping to tails, distribution and holding periods
FREQUENTLY ASKED QUESTIONS
Questions about monthly gold-return evidence
What does gold price seasonality mean on this page?
It means grouping each month-over-month change by the calendar month in which the change ends, then comparing the average, median and positive-return frequency across those twelve groups.
Does the strongest historical month predict the next one?
No. The ranking is an in-sample description. It can be unstable across subperiods and does not account for current valuation, policy, rates, currencies or market positioning.
Why show the median as well as the average?
One extreme observation can pull an average materially. The median identifies the middle observation and makes that sensitivity visible, though it also is not a forecast.
Are these calendar-month closing returns?
No. They are percentage changes between consecutive World Bank monthly averages. A daily month-end series would produce different endpoints and extremes.
Does positive frequency mean probability of a gain?
No. It is the share of historical observations above zero in this overlapping market path, not a modeled future probability.