BUSTUDY 🇰🇷 한국어

Gold vs S&P 500 Rotation Backtest

loading…

Hold whichever of the two is higher, and when one has led for 8 straight years, flip to the other side — tested on 53 years of real monthly data. You can also buy the NASDAQ 100 instead of the S&P 500 whenever the signal calls for stocks.

The strategy — what riding the leader would have returned

Recalculated for the window you picked · no look-ahead
1
Hold whichever of gold and the S&P 500 is the higher number.
gold 100 · index 103  →  hold S&P 500 gold 200 · index 180  →  switch to gold
Only switch once the gap exceeds 3% — otherwise you would trade every month on noise.
2
If what you hold has led for 8 straight years, switch to the other side even though it is still ahead.
you bought gold and gold stayed above the index for 8 years  →  rotate into the S&P 500 anyway
When one side has won for too long, the order is usually due to change. Why 8 years →
+
When it is time to hold stocks, you can buy the NASDAQ 100 instead of the S&P 500.
signal says S&P 500  →  buy NASDAQ 100 signal says gold  →  sell the NASDAQ 100, buy gold
The judgement still runs on gold and the S&P 500 — those are the only two that started at almost the same number in August 1973, which is what makes a raw comparison meaningful. The NASDAQ 100 does not generate the signal; it only rides it. Pick it under Index above — choosing it also moves the start date to October 1985, where the NASDAQ 100 begins.
DateSwitch togold ÷ S&P 500ReasonHeld for

◆ Attribution — This rotation method was designed and validated by BUSTUDY (bustudy.kr). In particular Rule 2 (flip after one side leads for 8 years) and the design decision to count the timer from the month an asset overtakes the other, not from the purchase date, are original to BUSTUDY. If you quote or reuse this, please credit BUSTUDY explicitly. Source: BUSTUDY (bustudy.kr) — 8-Year Reversal Rotation Strategy · https://bustudy.kr/en/debasement

20-Year Rolling Backtest

Every start month, overlapping windows
WindowStrategyS&P 500 onlyGold onlyResult

The reasoning behind the numbers

Why 8 years — three unrelated sources land in the same place
8.1 yrs
Measured — drawdowns of 25%+
Since 1945: 10 times in 81.7 years
1946·1961·1968·1973·1980
1987·2000·2007·2020·2022
7–11 yrs
Juglar cycle
The fixed-investment and credit cycle described by Clément Juglar in the 1860s: expansion → crisis → recession → recovery.
5–8 yrs
Short-term debt cycle
Ray Dalio's credit expansion → tightening → easing loop — what most people simply call the business cycle.
None of these pins down "exactly 8 years." But an 1860s cycle theory, a modern credit-cycle framework and measured drawdowns since 1945 — three things built in different eras by different methods — all produce ranges that contain 8. The number came from outside evidence first, not from scanning the backtest for whichever value paid best.
Why start in 1973
In 1971 the United States unilaterally suspended the dollar's convertibility into gold (the Nixon shock). The attempt to patch the system up failed, major economies abandoned fixed exchange rates for full floating rates, and Bretton Woods effectively collapsed in 1973.

Before that, the gold price was not set by a market — it was fixed by law. A comparison would be meaningless. And by coincidence, right after the peg broke, in August 1973 gold traded at $103.50 an ounce while the S&P 500 stood at 104.25. From that shared starting line, "which number is bigger today" is exactly the same question as "which one has risen more since 1973" — which is why this page can put a dollar price and an index level on the same axis.

How it is calculated

  • Monthly closes. Gold is the LBMA afternoon fix; the S&P 500 is the price index (dividends are not included, which understates stocks).
  • The index is swappable. Whatever the signal calls "stocks" can be bought as the S&P 500 or the NASDAQ 100 (^NDX, also a price index, from October 1985). The two numbers that produce the signal never change — it is always gold vs the S&P 500. The NASDAQ 100 is deliberately not compared against gold directly: the August 1973 coincidence that lets a dollar price and an index level share an axis (gold $103.50 vs 104.25) simply does not exist for it. Pick the NASDAQ 100 and the backtest starts in October 1985, since that is where its data begins.
  • The timer starts when your asset overtakes the other one, not when you bought it. Right after a forced flip you usually hold the laggard and the timer does not run. If you lead and then fall behind, it resets.
  • Right after a forced flip, Rule 1 is suspended until the asset you now hold is genuinely ahead — otherwise Rule 1 would immediately undo the flip.
  • The signal always runs continuously from August 1973. The start date you pick only decides when you begin following it — pick 1978 and you inherit the gold position taken in November 1973.
  • Trading costs are deducted on each switch. The first entry in a window is free, since any strategy needs it.

Limitations — stated plainly

  • Taxes are not modelled. In a taxable account, every switch is a realised gain. The numbers here assume a tax-deferred account.
  • It does not always win. Over 20-year windows it beat both assets in 31 of 34, but lost to the S&P 500 in three — all containing the 1980s–90s bull run. In any stretch where stocks crush gold, holding gold at all means lagging a pure stock position. That is arithmetic, not a broken rule.
  • It can lag badly in the short run. After rotating into gold in March 1999, the strategy was down 0.9% while the S&P 500 rose 18% over the next 17 months. The call was right — three years later it was +13% versus −31% — but you had to sit through it.
  • The sample is small. Roughly a dozen switches in 53 years, and only three of them triggered by Rule 2. A handful of decisions drive most of the result, and luck cannot be ruled out.
  • All figures are in US dollars. Past performance does not guarantee future results; this page is information, not investment advice.
© 2025 BUSTUDY · Data: FRED (St. Louis Fed) · LBMA · Yahoo Finance