World, S&P and Nasdaq: three ETFs, shared drivers

Fees: low-cost vehicles, with complexity to justify

The weighted recurring cost of 0.25% per year earns B. Index vehicles keep implementation inexpensive, but paying several modest fund charges does not create several distinct return drivers. The useful comparison is whether the intended US and Nasdaq overweights justify their extra cost and rebalancing, alongside the World holding that already owns many of the same businesses.

Three holdings nevertheless require decisions about their weights and rebalancing. They can create more trades and monitoring. That work is justified if the overweights are intentional; it adds little if the goal was simply broad global equity exposure. Low fees do not reveal economic duplication.

Long-term assumptions: a conviction, not three independent bets

Adding US and Nasdaq exposure leaves the expected growth gap at -0.8 percentage points per year, for B. The model treats these sleeves largely as the same US-equity class; it does not extrapolate Nasdaq’s recent outperformance into a separate premium. Overlapping holdings therefore add little variety for the model to use when balancing growth against fluctuations.

The positive case is exposure to large, profitable companies able to invest and innovate. The counterweight is the price paid for those prospects. If profits grow less than expected, or investors demand lower valuations, a strong business can still suffer a sharp share-price decline.

This allocation expresses a larger US and technology bet than World alone. It can therefore experience extended underperformance when those exposures lag. The grade offers little support for describing the extra funds as an improvement in diversification.

Historical efficiency: a past advantage with important context

The replay returned 15.73% annually in euros from 22 October 2016 to 22 October 2026, compared with 13.66% for the reference at the same measured risk. The +2.08 percentage points per year advantage earns B. It survives the risk adjustment, so it records more than the higher raw return one might obtain simply by taking more volatility. The capture does not provide a holding-by-holding attribution of that advantage.

That is a historical strength, not an automatic justification for future weights. The window includes a favourable period for these exposures. In addition, 33.87% of the history is extended using asset-class series; not every ETF has observed prices throughout the whole period.

The reference adjusts World exposure to the measured risk, using theoretical leverage when that risk exceeds World’s. It does not reproduce every concentration risk or possible crisis path, and is not a directly purchasable product. Annualised outperformance does not show how long the allocation could remain below its previous value or behind its reference.

Crises: all three holdings transmit the same shock

The US and Nasdaq sleeves both enter the US-equity shock class, while World also has a high sensitivity to that shock. The estimated loss of -38.78% consequently approaches the US-stock reference and earns D. Reallocating between these three ETFs leaves the portfolio exposed to much the same loss mechanism.

The model uses a shared US-equity sensitivity for the Nasdaq sleeve. It applies no separate shock to chip spending or technology valuations. The result can therefore understate the consequences of a shock aimed specifically at the portfolio’s dominant theme, and should not be treated as a worst-case loss.

A larger Nasdaq weight increases the role of this shared exposure in the portfolio. A target weight or rebalancing rule describes how that exposure is maintained; neither creates a payoff that offsets equity losses. Spending needs and assets held elsewhere also affect how a decline would be experienced.

Diversification: overlapping exposures

On 30 September 2026, Nvidia accounts for 5.81% of MSCI World. On 30 September 2026, it accounts for 8.36% of Nasdaq. Apple and Microsoft also rank among the largest holdings in both indices. There is already overlap between those two funds before S&P is added. [1][2]

Technology represents 68.01% of Nasdaq under its ICB classification. This indicates substantial specialisation, but cannot be compared directly with MSCI World’s information-technology weight, which uses a different classification. A precise overlap calculation needs to match individual companies and weights at the same date. [2]

The engine groups the holdings into 1 growth source and retains no expected growth above cash when that driver stops. Its D still gives some credit to the broad-index core, but the concentrated sector mix receives less credit than World alone. Detailed S&P composition is missing and that sleeve uses its asset category. The grade supports a concentration diagnosis; the capture does not support a precise combined company-overlap percentage.

An additional holding changes diversification through its effect on the drivers of returns. Equities in other regions may reduce geographic concentration without cushioning a global crisis. A money-market reserve reduces exposed capital; other asset families bring different risks. The number of ETFs alone does not describe these differences.

Other portfolios

• World and emerging markets
• One World ETF: concentrated in the US, led by big tech
• World and money markets: cushioning losses has a cost