Portfoliothese (Issue 3): How to check whether a new strategy diversifies a portfolio

Three questions on the comparison benchmark, stress behaviour and return level

The previous issue of this newsletter, “What a correlation figure doesn’t tell you”, asked what a correlation says about the diversification effect of an addition. On its own, less than the figure first suggests: for the low-volatility building blocks examined, most of the risk reduction came simply from replacing a highly volatile building block with a markedly calmer one.

This issue extends the question in two directions. Instead of volatility, performance and the loss path now take centre stage. And instead of a single addition, it is about the combination of several strategies, each of which is meant to serve diversification.

Behind this lies a classic question of portfolio construction. Anyone who takes on a new strategy usually has to reduce an existing position to make room for it. If one then compares only the new portfolio with the old one, it is easy to credit the new strategy with an effect that already results from this reduction. The study yields three test questions that can be put to any new strategy: on the comparison benchmark, on stress behaviour and on the return level.

The example again comes from digital assets. Because Bitcoin fluctuates so strongly, it becomes particularly clear there how much the smaller position alone achieves.

Anyone who represents a digital asset allocation solely through Bitcoin ties it to a single price path. An investor who had put money into Bitcoin at the end of December 2021 stood 26.7% in the black at the end of June 2026. At the lowest month-end level in between, the investment was 64.2% in the red. The study examines what changes if half is shifted into other sources of return from the same investment universe. The starting point lies a few weeks after Bitcoin’s then peak in November 2021 and thus at the beginning of a severe slump. The comparison with the starting portfolio of pure Bitcoin is therefore shaped substantially by this loss phase.

The result up front: adding a building block made up of three defensive sources of return in a 50% allocation would have improved performance markedly compared with pure Bitcoin. Compared with the two reference portfolios, which hold only half the Bitcoin position and place the other half in cash or in the SOFR reference, it would have improved the maximum drawdown only marginally. And a considerable part of its lead over the SOFR reference portfolio depends on the return level of two selected individual strategies.

Performance and maximum drawdown compared

All returns are calculated in US dollars and shown as percentages. For a euro investor without currency hedging, exchange-rate effects would be added. They change all end values by the same factor and leave their ranking unchanged, but they can shift the loss paths and thus the maximum drawdown in euros. All portfolios are rebalanced once a year to their initial weights, in the main calculation at year-end, for comparison on 30 June. The observation period covers January 2022 to June 2026, that is 54 months, and is the same as in the previous issue.

The table compares pure Bitcoin with seven portfolios of 50% Bitcoin and a second asset each. Two of them serve as references: the 50% Bitcoin / 50% cash portfolio, with non-interest-bearing cash already familiar from the previous issue, referred to below as the cash reference portfolio, and the 50% Bitcoin / 50% SOFR reference portfolio, referred to below as the SOFR reference portfolio. The SOFR reference is the compounded index of the secured US money market rate and stands here for an idealised money market investment without costs. Together they are called the two reference portfolios.

As in the previous issue, three sources of return are also added individually which fluctuated comparatively little in the period examined and are referred to below as defensive: a stablecoin income strategy, which earns income on dollar-pegged tokens, a DeFi income strategy, which earns income via decentralised lending and trading protocols, and a market-neutral digital asset index made up of trading strategies without a directional bet. As a counter-check, an altcoin index excluding Bitcoin and Ethereum is added. The seventh portfolio contains the combined building block of stablecoin income, DeFi income and the market-neutral digital asset index at one third each. These three weights are also reset to one third each on the respective rebalancing date.

The altcoin row is instructive. This portfolio is worse than pure Bitcoin on both counts, performance and maximum drawdown. An addition is evidently not automatically diversification.

Because the first date at year-end, the end of 2022, fell exactly on Bitcoin’s lowest month-end level, the two right-hand columns show the same calculation with 30 June as a counter-check. June is thereby not a better rebalancing rule, but a point of comparison. Even this difference shifts the results markedly, most strongly in the stablecoin row. The stablecoin strategy had gained in 2022 while Bitcoin lost 64.2%; year-end rebalancing therefore bought an especially large amount of Bitcoin at an especially favourable moment. The portfolio thus reaches plus 104.7%, more than either of its two components on its own; without rebalancing it would have been plus 48.1%. A considerable part of the historical performance thus depends on rebalancing and its date, and not solely on the choice of the addition. For all portfolios, 30 June lies close to the middle of what the twelve possible dates produce.

Why the comparison with pure Bitcoin is not enough

Compared with pure Bitcoin, the portfolio with the combined building block appears clearly superior at first glance: substantially higher performance and a maximum drawdown that is more than halved. The comparison with the two reference portfolios qualifies this picture.

The rows with non-interest-bearing cash and the SOFR reference show why. Even shifting half of the position brings the maximum drawdown almost to the level of the portfolio with the combined building block. On performance too, the two reference portfolios are already clearly closer to this portfolio than pure Bitcoin is. Compared with the SOFR reference portfolio, a performance gap of a good 22 percentage points remains, whereas for the maximum drawdown the gap is less than one percentage point.

The high cumulative returns, however, as the stablecoin row shows, depend heavily on rebalancing. Without rebalancing, that is, with a single 50/50 split at the end of 2021 and no later adjustment back to 50% each, the portfolios with cash and with the SOFR reference would even have been below pure Bitcoin on performance by the end of June 2026.

For the comparison between the portfolio with the combined building block and the SOFR reference portfolio, by contrast, rebalancing makes little difference: without rebalancing the performance gap is around 21 rather than a good 22 percentage points.

What the combined building block achieves

In hindsight, the portfolio with stablecoin income alone would have performed better on performance and maximum drawdown than the portfolio with the combined building block. This selection, however, is not known in advance; the combined building block spreads the weight across three sources of return and thereby reduces dependence on any single one.

On its own, that is, without Bitcoin alongside, the combined building block had a lower maximum drawdown than two of its three components. At 3.1% it was below DeFi income and the market-neutral digital asset index; stablecoin income did not close any month with a loss.

Its worst month was nevertheless, of all months, November 2022, the month of the FTX collapse. Stablecoin income was in positive territory, whereas DeFi income and the market-neutral digital asset index each lost just under 5% at the same time. The lower maximum drawdown of the combined building block therefore does not show that the three sources of return cushion one another precisely under stress.

Three reasons not to read too much into the return figures

Three points limit what can be derived from these 54 months for future portfolios.

First, a considerable part of the return advantage depends on the return level of the two income strategies. The stablecoin income strategy earned 12.4% a year over the period, compared with an annualised return of 4.1% for the SOFR reference. That is a premium of 8.3 percentage points. Such a premium on dollar-denominated balances can be understood economically as compensation for additional risks, such as counterparty, credit, custody and protocol risks. A calm return history therefore does not prove that these risks are low, only that they did not show up in the monthly returns of the period examined.

Both income strategies are selected individual strategies, not market averages. They were chosen because they represent their respective return driver, not for their historical return. Selection bias and a survivorship effect nevertheless cannot be ruled out.

What matters, therefore, is how strongly the result depends on their return level. A sensitivity calculation tests what would remain if both income strategies, with unchanged volatility, had earned only as much as the money market. Then only around a third of the combined building block’s annual excess return over the SOFR reference remains, and not even a quarter of the annual return advantage of the portfolio with the combined building block over the SOFR reference portfolio. The calculation changes only the return level; it does not isolate a diversification contribution and does not represent a stress scenario. How it is constructed is explained in the accompanying technical article for this issue.

Second, monthly data show less than one might think. For the altcoin index, for which daily values are available, the price stood more than five percent below the previous month’s closing level at some point within the month in 36 of the 54 months; ten of these months nevertheless closed positive. For the three defensive sources of return, the data used contain no daily values, and the finding for the altcoin index does not allow us to infer how they behaved within a month. For the calmest series in particular, the information that would matter most for abrupt risks is thus missing.

Third, it remains open how independent the risks behind the three sources of return really are. For this evaluation, three events already familiar from the previous issue were defined in advance: Terra/Luna, FTX and the USDC depeg. A conspicuous joint strain was deemed to exist if at least two differently constructed building blocks lost more than twice their monthly return volatility, measured over the full period, in the same month. That applied only to FTX. This rule was, however, set with knowledge of the FTX month; that it is precisely that month which meets it is therefore not independent confirmation. The two other events did not confirm such a joint strain. And even for FTX, the return series alone cannot verify whether both building blocks were actually hit by the same source of risk. The question of common risk factors is thus neither answered nor refuted.

What remains

For the starting position examined here of 100% Bitcoin, the maximum drawdown fell, through the 50% addition of non-interest-bearing cash alone, to almost the same level as for the portfolio with the combined building block. The lead of the portfolio with the combined building block of a good 22 percentage points over the SOFR reference portfolio, by contrast, reacts sensitively to the return level of the selected strategies, and the cumulative returns depend on the rebalancing date. In the FTX month, two of the three defensive sources of return lost simultaneously; the lower maximum drawdown of the combined building block therefore does not demonstrate reliable mutual protection under stress. One of the coming issues of this newsletter will be devoted in detail to the effects of rebalancing.

The full calculation with methodology, data basis, all metrics and the event test is contained in the accompanying technical article.

Context and limitations

The study design and evaluation rules were set down in writing and frozen before the new calculations of this issue; the results of the previous study were already known at that point. Also fixed were the three events, the threshold of the event test and the rule for how the event findings are to be classified. The rebalancing convention is not part of these specifications. Analyses added subsequently are marked; in the present text this concerns the SOFR reference portfolio, also as a benchmark in the sensitivity calculation, the comparisons without rebalancing, the figures on the highest month-end level, the comparison of the rebalancing dates including the calculation as of 30 June, and the thresholds of the daily evaluation.

On the data basis: Bitcoin via the Galaxy Bitcoin Series, the altcoin index via the S&P Cryptocurrency BDA Ex-MegaCap Index, the market-neutral digital asset index via the Galaxy VisionTrack Market Neutral Index, the money market reference via the SOFR index of the Federal Reserve Bank of New York. Stablecoin income and DeFi income are two anonymised individual strategies with monthly net histories after fees. The market-neutral digital asset index is based on returns that participating managers contribute voluntarily and that are not independently verified.

On the course at the end of the period: pure Bitcoin reached its highest month-end level in July 2025 and ended the period 49% below it. The portfolio with the combined building block reached its highest month-end level in September 2025 and ended just under 25% below it. Because the 2022 decline was larger in percentage terms, this second decline does not appear in the table.

The calculations answer a question of risk and return over a completed period. They apply to the digital asset allocation on its own. Its effect on the overall portfolio depends on its weight and on the interplay with the other investments. They are not a model allocation for a client portfolio and take into account neither implementability nor liquidity or access to the strategies examined, nor the transaction costs and taxes of rebalancing. 54 monthly observations and the few available event observations do not permit statistical proof of common risk factors.

---

Boris Jurczyk is the founder of 21Nexus Capital and is therefore economically connected to the topic of digital assets. 21Nexus Capital acts as adviser to a fund that invests, among other things, in stablecoin income, DeFi income and altcoin strategies. There are currently no business relationships with the two fund companies behind the anonymised income strategies; in particular, 21Nexus is not invested there.

This article is for information purposes only and constitutes neither investment advice nor a recommendation to buy or sell financial instruments or crypto assets. Crypto assets are subject to substantial price fluctuations; a total loss of the capital invested is possible. The figures shown refer to the period January 2022 to June 2026. Past performance is not a reliable indicator of future results.