High Return, Long Road Back: Half Bitcoin, Half Gold Compared with Equities

A fixed mix of 50 per cent Bitcoin and 50 per cent gold returned 28.6 per cent per year in euros since the end of 2017, compared with 12.7 per cent for a globally diversified equity investment (MSCI World). However, the mix needed 756 days to regain its previous high, against 323 days for the MSCI World.

This article examines what lies behind this contrast, whether a money market cushion (a euro overnight ETF) resolves it, and what remains of the effects when small allocations are added to an equity portfolio. It is a historical comparison without a recommendation. No statement is made about suitability as an equity substitute in any specific mandate.

Test design

Four portfolios are compared in euros on the basis of daily closing prices from 29 December 2017 to 30 September 2026. Weights are reset to their target values once a year. Costs, taxes and product fees for Bitcoin and gold are not taken into account. The comparison starts at the end of 2017; the months of Bitcoin euro prices available since 05/2017 are not used. Further start dates follow at the end of the article.

High return, deep loss, long recovery

The comparison starts with the loss side. The maximum drawdown measures the largest percentage distance from a previously reached peak to a subsequent trough. Peak-to-recovery time is measured in calendar days from the peak preceding the maximum drawdown until that peak is first regained.

The largest decline of the MSCI World occurred in the Covid crash (19 February 2020 to 23 March 2020). For the Bitcoin/Gold portfolio, the decline ran from 8 November 2021 to 19 December 2022, and for the mix with money market from 26 June 2019 to 16 March 2020.

The annual reallocations had a substantial effect on the result (table). With rebalancing, the return of the Bitcoin/Gold portfolio was above that of Bitcoin (22.1 per cent) and gold (14.9 per cent) on their own. The low correlation of daily returns between Bitcoin and gold (0.10) alone does not explain this return advantage; return paths, differences in volatility and the reallocations themselves play a role. A decomposition was not the subject of the study.

Who bears the risk?

A portfolio weight of 50 per cent does not mean 50 per cent of the risk. Measured as contribution to portfolio volatility, Bitcoin accounted for around 93 per cent of the measured volatility risk in the mix with gold, and for around 94 per cent in the mix with 50 per cent money market, with a portfolio weight of only 25 per cent. A check using weekly returns, which reduces the time offset between market closes, gives around 94 per cent in each case. Gold contributed the remainder, money market practically nothing. The risk measure captures only volatility, not all risks.

With 19.6 per cent volatility per year, the mix of money market, gold and Bitcoin fluctuated considerably less than the Bitcoin/Gold portfolio (34.9 per cent), but more than the MSCI World (15.7 per cent). It is therefore not a risk-equivalent counterpart to the equity index, but a different risk with a lower weight on Bitcoin. Daily returns of Bitcoin correlated with gold at 0.10 and with the MSCI World at 0.19.

The cross-check: a simpler defensive equity investment

The cross-check shows whether the mix of money market, gold and Bitcoin offered additional benefit compared with a simple, more defensive equity investment. The money market cushion reduced the maximum drawdown relative to the Bitcoin/Gold portfolio from 42.9 per cent to 26.7 per cent and the recovery period from 756 to 497 days; the return fell to 17.2 per cent per year. A mix of 50 per cent money market and 50 per cent MSCI World serves as the comparison. It returned 7.1 per cent per year with a maximum drawdown of 17.5 per cent, volatility of 7.8 per cent and a recovery period of 324 days. Its maximum drawdown was thus smaller than for the two alternative portfolios with Bitcoin examined, but the return was clearly lower.

More informative is a look at the four phases in which the MSCI World stood more than 10 per cent below its previous high. The loss is measured from the equity high to the equity low, here for the portfolios run continuously since the end of 2017.

The mix of money market, gold and Bitcoin lost less than the MSCI World in all four phases and less than the mix of money market and MSCI World in three of four phases; it did worse in the phase 01/2022 to 06/2022. If the portfolios are instead reset to target weights at each equity high (event test), it was better than the simple money market/equity mix in only two of four phases. The Bitcoin/Gold portfolio, run continuously, lost less than the MSCI World in three of four phases, and in two in the event test.

Over the full period, the other part of the picture emerges. Compared with the money market/equity mix, the three-asset mix had a deeper maximum drawdown of its own (26.7 per cent versus 17.5 per cent), higher volatility (19.6 per cent versus 7.8 per cent) and a longer recovery (497 versus 324 days); in return, it achieved a higher return (17.2 versus 7.1 per cent per year). Less loss during an equity crisis therefore does not mean lower portfolio risk.

Return and entry date

Because investors enter at different times, an investment at target weights was simulated for each month-end, and performance was measured over the following 12, 24 and 36 months.

With a 36-month holding period, the Bitcoin/Gold portfolio was ahead on return in all 70 start dates but had the smaller maximum drawdown in only 17 per cent of start dates. The mix of money market and MSCI World had the smaller drawdown for all start dates but the higher return in none.

Gold or Bitcoin? Small allocations

For asset managers, the practical question is usually not a full alternative to equities but a small allocation. For this purpose, 10 percentage points of the MSCI World allocation were replaced by gold, by Bitcoin, or by 5 percentage points of each. This makes it possible to see whether an effect comes from Bitcoin, from gold or from the combination.

In the equity portfolios examined, adding Bitcoin mainly raised the return, while adding gold improved the loss metrics more clearly: gold reduced the maximum drawdown and volatility and made the drawdown smaller for all start dates in the periods shown, while Bitcoin raised the return with an unchanged maximum drawdown. The combination lay between the two individual variants on return and drawdown. Whether a separate combination effect exists beyond that cannot be assessed with this comparison. On the recovery from the respective maximum drawdown, the allocations with Bitcoin (265 and 264 days) were below the MSCI World (323 days), while the gold allocation, at 322 days, was at a similar level.

What this means for the portfolio decision

First: protection during an equity crisis and a portfolio's own loss risk are two different properties. A mix can perform comparatively well in equity crises and still have deeper declines of its own and longer recoveries. Both belong in the assessment.

Second: with small allocations to an equity portfolio, gold mainly improved the loss metrics and Bitcoin mainly the return. For Bitcoin, the contribution to measured volatility risk, in addition to the portfolio weight, is informative, also with a view to a mandate's risk budget. The study provides no suitable limit.

Third: for loss-bearing capacity, the time to recovery matters alongside the size of the loss. The two main portfolios with Bitcoin needed 497 and 756 days to regain the high before their maximum drawdown, the MSCI World 323 days. With small allocations, this difference was not observed.

Anyone seeking loss dampening finds the benchmark in a simpler defensive equity investment. When seeking additional return opportunities, the loss phases and risk contributions taken on for them have to be assessed. No particular allocation follows from this.

The comparison describes a period of exceptional Bitcoin appreciation. The results are neither upper limits for future losses nor expectations for future returns.

Methodology and robustness

The result depends on the start date. The two more recent periods were added after the results had been reviewed.

Since the end of 2017, the maximum drawdown of the MSCI World stems from the Covid crash (33.6 per cent); since the end of 2020, from spring 2025 (21.6 per cent). For the money market/Gold/Bitcoin portfolio, the maximum drawdown occurred during the Covid crash in the sample starting at the end of 2017, and during the Bitcoin bear market from 11/2021 to 12/2022 in the sample starting at the end of 2020. Consequently, the three-asset mix has a smaller maximum drawdown than the MSCI World for the 2017 start (26.7 per cent versus 33.6 per cent), but a larger one for the 2020 start (25.3 per cent versus 21.6 per cent).

Quarterly instead of annual rebalancing gives the Bitcoin/Gold portfolio a return of 29.2 per cent per year and a maximum drawdown of 48.3 per cent. Volatility and risk contribution were calculated uniformly from simple daily returns. The rolling periods each run from month-end to month-end.

---

Data and limitations: Daily closing prices in euros, own calculation. MSCI World: iShares Core MSCI World UCITS ETF USD (Acc), ISIN IE00B4L5Y983, Euronext Paris (ticker SWDA), accumulating; the ETF's costs are included in the price. Money market: Amundi EUR Overnight Return UCITS ETF (Acc), ISIN FR0010510800, Euronext Paris (ticker CSH), accumulating; four individual price outliers (7 October 2015, 24 March 2026, 10 July 2026, 24 July 2026) were replaced by the mean of the neighbouring days; the results change by less than 0.1 percentage points as a result. Gold: spot price XAU/EUR, excluding storage and product costs. Bitcoin: exchange price BTC/EUR, daily value at 00:00 UTC, excluding product costs. Common trading days of gold and the MSCI World ETF; on 48 of these days no price of the money market ETF is available, and the last price is carried forward there (return, maximum drawdown and recovery period of the main period remain unchanged if restricted to days with actual common observations, volatilities change slightly, for Bitcoin/Gold from 34.9 to 35.1 per cent); Bitcoin movements at weekends are included in the return up to the next common trading day. The closing times of the series differ, which can slightly distort correlations and risk contributions. A trading cost assumption of 0.2 per cent on reallocated volume changed returns by less than 0.1 percentage points; purchase costs on entry and product fees are not taken into account. The analysis describes the past and contains no forecast. Calculation as of 30 September 2026.

No investment advice. Market observation only.

An article by 21Nexus. The content is provided for general information and context. It constitutes neither investment advice nor a recommendation to buy or sell financial instruments and does not replace individual advice.