How Much Crypto Belongs in a Portfolio? Allocation, Risk and the Diversification Fallacy

Strategy
Risk Management

14 October 2026 · 17 min read

By Dominik Schneeberger, Gründer & Managing Partner

How much Bitcoin belongs in a portfolio? For a long time, the answer from major banks was: 0.0%. Many of them now name concrete ranges – most between 1 and 5% of total wealth.

The figure alone, however, only answers half the question. Three further questions are at least as important: what exactly is bought, how the exposure is managed and where the assets are held. Two portfolios with the same crypto allocation can carry completely different risks.

What banks and asset managers recommend today

An overview of published recommendations (as of October 2026):

•
BlackRock Investment Institute: 1–2% Bitcoin for investors who expect further adoption (December 2024)
•
WisdomTree: 2% Bitcoin in a global 60/40 portfolio (September 2025)
•
Morgan Stanley: 0–4% crypto, from 0% in the conservative to 4% in the opportunistic growth profile (October 2025)
•
Bank of America: 1–4% crypto for clients with a higher tolerance for volatility, implemented via spot Bitcoin ETFs (December 2025)
•
Fidelity: 2–5% Bitcoin, up to 7.5% for young investors (2024)
•
BBVA: 3–7% crypto for wealthy private banking clients, depending on risk appetite (June 2025)
•
Charles Schwab: depending on method and profile, 1.0–8.8% Bitcoin plus 0.1–2.9% Ether (April 2026)
•
VanEck: up to 6% (3% Bitcoin and 3% Ether) as the return-optimal addition to a 60/40 portfolio (November 2024)

Banks recommend allocating 0–8.8% of portfolios to crypto, depending on risk

What stands out is how defensive most figures still are. The reason lies less in doubts about the asset class than in its volatility. BlackRock calculates that just 2% Bitcoin accounts for around 5% of total portfolio risk. Morgan Stanley finds that a 6% crypto position almost doubles the volatility of a growth portfolio.

Equally important: almost all of these recommendations refer to passive, permanently held Bitcoin exposure, mostly via ETFs. They say nothing about altcoins, active strategies or the question of which platforms are used.

Volatility and drawdowns: what passive crypto exposure means

Bitcoin's volatility has fallen significantly in recent years. According to CoinDesk, it stood at around 46% annualised in 2026, compared with 84% in 2018. Our own analysis of daily price data also gives 46% for the three years to May 2026, and 54% over five years. For comparison: the MSCI World fluctuates by around 15% per year over the long term.

Beyond Bitcoin, volatility rises sharply. Ether reaches 65% over the same three-year period, the 37 altcoins we analysed 90% on average, and individual coins more than 110%.

Even more telling than volatility are drawdowns, i.e. the loss from the last peak to the trough:

•
Bitcoin: −84% (2017/18) and −77% (2021/22). Since its all-time high of October 2025, Bitcoin has at times lost up to 54% and currently trades around a third below that high.
•
Ether: −94% (2018) and −79% (2021/22). Since its all-time high of August 2025, Ether has at times lost around 70%.
•
Altcoins: Of the 37 altcoins analysed, around 80% were more than 80% below their all-time high in May 2026, and two thirds even more than 90%. The median altcoin was 95% below its all-time high, and individual coins lost almost 100%.

Volatility and drawdown: MSCI World 15%/−57%, gold 15%/−46%, Bitcoin 46%/−84%, Ether 65%/−94%, altcoins 90%/−95%

Then there is default risk: according to CoinGecko, 53% of all tokens ever recorded are now considered inactive. Most are meme and launchpad tokens – a clear indication of how large the risk is beyond established assets.

Anyone holding crypto passively has to sit through such pronounced cycles and drawdown phases in full. Losses act asymmetrically: the recovery return required rises disproportionately with every additional percentage point of drawdown.

Recovery return = 1 / (1 − drawdown) − 1 → 1 / (1 − 0.77) − 1 ≈ 335%

After a loss of 77%, a portfolio therefore needs a return of around 335% to get back to its starting value. Buy & hold is thus not a low-risk strategy but a clearly one-directional bet on future appreciation – usually without a defined invalidation.

The diversification fallacy – in the equity market and in the crypto market

Anyone investing in the S&P 500 often assumes they are broadly diversified across 500 companies. In economic terms, that is only partly true. The ten largest stocks currently make up around 38 to 40% of the index – almost double the long-term average of about 24% and more than at the peak of the dotcom bubble. The IT sector alone accounts for almost 40%, and semiconductor stocks reach a record share of almost 20%.

Performance is correspondingly concentrated. By early May 2026, the ten largest stocks had delivered around 89% of the index gain. By early October, about half of the year's gain came from just four stocks: Nvidia, Apple, Micron and AMD. Buying the S&P 500 therefore largely means buying a bet on AI, tech and semiconductors.

In the crypto market, this effect is even more pronounced – not through weighting, but through correlation. Our analysis of Bitcoin, Ether and 37 altcoins based on daily returns to May 2026 shows:

•
Ether and Bitcoin correlate at 0.81 to 0.86, depending on the period.
•
Individual altcoins correlate with Bitcoin at 0.57 to 0.62 on average, and an equally weighted altcoin basket at 0.77 to 0.82.
•
A single common factor – essentially the direction of the overall market – explains 55 to 59% of all price movements of the 39 assets.
•
Mathematically, 39 crypto assets are therefore equivalent to fewer than two independent positions.
•
On days when Bitcoin lost more than 5%, an average of 96% of altcoins fell too – and by more: 9.1% on average versus 7.2% for Bitcoin.

The effective number of positions is derived as follows: for N equally weighted positions with an average pairwise correlation ρ̄, the following holds approximately:

N_eff = N / (1 + (N − 1) · ρ̄) = 39 / (1 + 38 · 0.53) ≈ 1.8

With fully independent assets (ρ̄ = 0), there would be 39 positions; with complete co-movement (ρ̄ = 1), only one. With an average correlation of 0.53 to 0.57, the crypto market is much closer to the second case.

This became particularly visible on 10 October 2025. Bitcoin closed the day at −6.5%, Ether at −11.5% and the altcoin average at −18%.

Bitcoin correlation: Ether 0.84, altcoin basket 0.82, traditional assets 0.47 or lower

Holding ten different crypto assets is therefore not diversification. It is one large position on the same risk factor – plus additional idiosyncratic risk. Altcoins can decouple from Bitcoin in the short term: in early October 2026, the 14-day correlation turned negative for the first time since July 2025, according to CryptoQuant. Such phases, however, tend to be short-lived. A genuine altcoin season like 2017 or 2021 – strong outperformance of the altcoin sector combined with low correlation to Bitcoin – has practically not occurred since.

The reasons are structural: institutional capital flows predominantly into Bitcoin via ETFs and other isolated products, and does not rotate into altcoins. Bitcoin dominance has hovered around 60% since mid-2025. Liquidity in the altcoin market, by contrast, is spread across thousands of tokens whose supply keeps growing. As a result, most altcoins behave like leveraged Bitcoin exposure: they fall with Bitcoin, only deeper.

How altcoin exposure changes the risk

Three example portfolios, rebalanced monthly to their target weights. Volatility over three and five years, maximum drawdown over five years, return over three years to May 2026:

•
Portfolio 1 – 50% Bitcoin, 30% Ether, 20% top-10 altcoins: volatility 51% / 59%, max. drawdown −76%, return +30.7% p.a.
•
Portfolio 2 – 30% Bitcoin, 20% Ether, 50% broad altcoin basket: volatility 56% / 64%, max. drawdown −78%, return +14.5% p.a.
•
Portfolio 3 – 10% Bitcoin/Ether, 90% mid- and small-cap altcoins: volatility 66% / 75%, max. drawdown −84%, return −7.6% p.a.
•
For comparison, Bitcoin only: volatility 46% / 54%, max. drawdown −77%, return +41.1% p.a.

Altcoin portfolios show higher volatility and lower returns than Bitcoin

In this period, more altcoins did not mean more diversification, but higher volatility, deeper drawdowns and weaker returns. In pronounced altcoin phases such as 2021, the picture can temporarily reverse. Note also that the analysis only covers coins with a complete data history. Total losses such as Terra/Luna are not included – the actual risk of an altcoin portfolio is likely to be higher.

Crypto in the total portfolio: low correlation with traditional assets

While crypto assets are highly correlated with each other, their relationship with traditional asset classes is much weaker. Correlation of Bitcoin's monthly returns over three, five and ten years to April 2026:

•
Global equities (MSCI World): 0.37 / 0.47 / 0.33
•
Gold: −0.27 / −0.08 / −0.12
•
US bonds: −0.06 / 0.19 / 0.16
•
Listed real estate (REITs): 0.16 / 0.33 / 0.27
•
US residential property (Case-Shiller): 0.01 / −0.12 / 0.05
•
Money supply M2 (US, monthly change): 0.05 / −0.05 / 0.09

These values fluctuate considerably over time. The rolling 12-month correlation with the MSCI World has ranged from −0.55 to 0.76 since 2016. Short-term measurements currently show a correlation with the S&P 500 close to zero and with gold of around 0.5, the highest since 2020.

The often-cited link with the money supply shows up mainly over the medium term. On an annual basis, the growth rates of Bitcoin and US M2 correlate at 0.25 (ten years) to 0.32 (five years). For global money supply, Lyn Alden found that Bitcoin moved in the same direction in 83% of all 12-month periods. Bitcoin is therefore more a barometer of global liquidity than of corporate earnings.

For portfolio construction, this is the decisive point: a crypto allocation can diversify the total portfolio – but hardly any diversification arises within the crypto allocation itself. Spreading capital across multiple coins merely increases volatility and drawdowns – and in the vast majority of cases it does not deliver a higher long-term return either. However, the figures above only apply to passive buy & hold (long-only). Active strategies behave differently: when exposure is partially reduced or short positions are taken selectively, dependence on market direction falls – and with it the correlation with the rest of the portfolio.

Passive or active: two different risks

Buy & hold means full, permanent exposure in one direction. The return depends solely on the market rising over the investment horizon. Interim book losses of 50 to 80% are not an exception but part of every crypto cycle so far.

An actively managed, market-adaptive approach addresses precisely this: with defined invalidation points, position sizes based on a risk budget, reduced exposure in weak market phases and – where the mandate allows – short positions. The aim is significantly shallower drawdowns with controlled risk. This is not guaranteed: active strategies can also be wrong at times and go through loss phases. For a detailed comparison, see our article on Crypto Asset Management vs. Buy & Hold.

The underestimated risk: counterparties

A considerable part of crypto risk has nothing to do with the market. It lies with the exchanges, custodians and lenders through which investments are made. Some of the largest losses of recent years were caused not by falling prices but by the failure of counterparties:

•
Mt. Gox (2014): Around 850,000 bitcoins disappeared; only about 200,000 were recovered. Repayments to creditors are still ongoing.
•
BitMart (2021): Hack causing losses of around USD 196 million.
•
Terra/Luna (May 2022): Collapse of an algorithmic stablecoin, wiping out around USD 40 billion in market value.
•
Three Arrows Capital (June 2022): Liquidation of the crypto hedge fund with claims of around USD 3.5 billion.
•
Celsius (June 2022): Withdrawal freeze with customer liabilities of around USD 4.7 billion.
•
FTX (November 2022): Insolvency with a shortfall of up to USD 8 billion. Around two and a half years later, customers were compensated at November 2022 prices – and missed the price rally in between.
•
Bybit (February 2025): Theft of around USD 1.5 billion, attributed by the FBI to North Korean hackers.
•
10 October 2025: Liquidations of more than USD 19 billion within a single day. On several platforms, auto-deleveraging forcibly closed even hedged positions, and the stablecoin USDe temporarily fell to around USD 0.65 on Binance (FTI Consulting).
•
Bitget (September 2026): Theft of more than USD 350 million from hot wallets (TechCrunch); withdrawals were suspended for several days. Bitget's USD 464 million user protection fund covered the loss (24/7 Wall St.).

In 2025 alone, crypto assets worth more than USD 3.4 billion were stolen, according to Chainalysis.

Our research paper on counterparty risk management comes to a clear conclusion: there is no single perfect counterparty. What matters is redundancy. If, for example, exposure is distributed evenly across four independent trading venues, an isolated total failure limits the immediate damage mathematically to 25% instead of 100% – provided the platforms do not share common weak points in the areas of IT security, (algorithmic) stablecoins, market makers or tech & server infrastructure. Anyone holding their entire crypto allocation on a single platform, by contrast, carries a concentration risk that does not appear in any volatility metric. More on this under Security & Custody.

Using several uncorrelated exchanges/wallets also reduces the probability of larger losses. Assuming, for illustration, an annual default probability of p = 0.5% per exchange and full independence, the following applies to the simultaneous failure of two specific exchanges:

P(A ∩ B) = p² = 0.5% × 0.5% = 0.0025%

→ A simultaneous failure of two uncorrelated exchanges is therefore 200 times less likely than a single failure.

The expected loss, however, stays the same: EL = Σ pᵢ · LGDᵢ · wᵢ is 0.5% both for one exchange and for four exchanges of equal quality. Redundancy is therefore tail management, not a reduction of expected loss. And it only works to the extent that the exchanges are genuinely independent: if the probability that exchange B fails once exchange A has failed is 25% instead of 0.5%, the factor shrinks from 200 to 4.

Redundancy keeps expected loss at 0.5% while limiting single-failure damage to 25%

Our allocation recommendation

From these considerations, we derive our ranges. They depend not only on the risk profile, but above all on how the exposure is managed.

Passive buy & hold, spread across at least two independent platforms:

•
Low risk: 3–5% of total wealth
•
Medium risk: 5–10%
•
High risk: 10–20%

Active management with market-adaptive long/short strategies, spread across several platforms:

•
Low risk: 5–10% of total wealth
•
Medium risk: 10–20%
•
High risk: 20–35%

This deliberately puts us above the recommendations of most banks. These, however, refer to passive Bitcoin exposure in standard portfolios, not to actively managed mandates. The higher ranges for active management are only justifiable if risk per position is limited, loss limits are defined and there is no permanent, full market exposure. Without these conditions, the buy & hold ranges apply.

What this means for a total portfolio

Our starting point is a typically structured portfolio: 30% equity ETFs, 30% real estate, 20% bonds, 10% gold and commodities and 10% crypto held as buy & hold. We compare it with a portfolio without crypto, a higher buy & hold allocation and two actively managed variants. The traditional assets are adjusted proportionally in each case.

Assumptions in the medium scenario (return / volatility p.a.):

•
Equities (MSCI World): 7% / 15%
•
Real estate, held directly: 5% / 8%
•
Gold and commodities: 5.5% / 15%
•
Bonds: 3.5% / 6%
•
Crypto buy & hold, Bitcoin-heavy: 15% / 50%
•
Active crypto, based on the target parameters of our DEUS Alpha Quant strategy: 20% / 25%, correlation with equities 0.15 and with crypto buy & hold 0.45

The assumptions for traditional assets are based on long-term capital market assumptions, including those of J.P. Morgan and MSCI. We simulated 20,000 ten-year paths with monthly rebalancing and realistically fat tails. Shown are the expected return, volatility, the range of one standard deviation down and up, and the maximum drawdown over ten years in the median scenario (in brackets: the worst 5% of scenarios).

Model note: Portfolio volatility follows from the weights w, volatilities and correlations of the assets (covariance matrix Σ). The risk contribution of an asset i shows what share of total fluctuation it accounts for; all contributions add up to 100%. Under a normal distribution, the range of ±1 standard deviation covers around two thirds of all years. Because of the fat tails, extreme years occur more often in the model than under a normal distribution. The simulation is not a forecast but a comparison of scenarios under fixed assumptions.

σ_P = √(wᵀ · Σ · w) RC_i = w_i · (Σ · w)_i / σ_P²

•
No crypto: return 5.4%, volatility 7.2%, range −1.8% to +12.6%, max. drawdown −11% (−22%)
•
10% crypto buy & hold (medium risk): return 6.4%, volatility 9.4%, range −3.1% to +15.8%, max. drawdown −15% (−29%); crypto accounts for 40% of total risk
•
20% crypto buy & hold (high risk): return 7.3%, volatility 13.1%, range −5.8% to +20.4%, max. drawdown −23% (−42%); crypto accounts for 69% of total risk
•
20% crypto active (medium risk): return 8.3%, volatility 8.1%, range +0.2% to +16.4%, max. drawdown −11% (−22%)
•
35% crypto active (high risk): return 10.5%, volatility 10.5%, range 0.0% to +21.0%, max. drawdown −14% (−28%)

Portfolio simulation compares expected returns and 10-year drawdowns across five crypto allocations

The result shows the core of this article. Even 10% passive crypto exposure accounts for around 40% of total portfolio fluctuation; at 20%, it is almost 70%. An actively managed 20% allocation, by contrast, achieves volatility and drawdown in this scenario that are almost equal to those of the portfolio without crypto – at a higher expected return.

The difference becomes even clearer under stress. For this, we assume the losses from the crypto bear market of late 2021/2022: crypto buy & hold −77%, equities −25%, bonds −17%, real estate −10%, gold +5%. In this scenario, the portfolio with 20% crypto buy & hold would lose around 27%, the portfolio without crypto around 15%. With 20% actively managed crypto, the loss would be around 18%, provided the active strategy limits its maximum loss to 30%.

Importantly, though, the active variant is only as good as the strategy behind it. The assumptions in this model calculation for our proprietary DEUS Alpha Quant trading strategy are based on target values and a backtest (April 2023 to April 2026, maximum drawdown −29.1%). For comparison, a deliberately pessimistic scenario: if the active strategy achieves only half its target return (10% instead of 20%) with considerably higher volatility (35% instead of 25%), the picture reverses. 35% active exposure would then result in a total portfolio volatility of 13.7% and a median drawdown of −24% – comparable to 20% crypto buy & hold. Choosing the manager and the quality of their risk management are therefore at least as important as the allocation size itself.

Conclusion

The question 'How much crypto?' falls short. Two portfolios with a 10% crypto allocation can differ many times over in risk – depending on whether the exposure is in Bitcoin or small-cap altcoins, whether it is held passively or actively managed, and whether it sits on a single platform or is spread across several independent counterparties.

What matters is not only how much crypto is in the portfolio, but how that exposure is managed, allocated and held.

Would you like to know which crypto allocation suits your portfolio and your goals? In a personal initial consultation, we assess your situation together – from the right risk profile to the structure of a segregated managed account.

Methodology: crypto price data from Coin Metrics (daily, to 23 May 2026; Bitcoin, Ether and 37 altcoins with available history). Volatility: standard deviation of daily returns × √365. Correlations: Pearson coefficient, crypto on a daily basis, traditional assets on a monthly basis. Top-10 altcoins: XRP, BNB, DOGE, ADA, TRX, LINK, BCH, LTC, XLM, DOT; broad basket: all 37 altcoins; mid/small caps: the remaining 27. Baskets equally weighted, rebalanced monthly. Traditional assets: monthly data for MSCI World (URTH), US Aggregate Bonds (AGG) and REITs (VNQ) including distributions, FRED (M2SL, Case-Shiller), gold price as monthly average. Simulation: 20,000 ten-year paths with monthly returns from a multivariate Student's t-distribution (4 degrees of freedom), monthly rebalancing, own calculation. Counterparty risk: model calculation from our research paper 'Counterparty Risk Management for Crypto Assets'. Past results and simulations do not guarantee future performance. This article is for information purposes only and does not constitute investment advice.