Equal weight vs cap weight: weighting, rebalancing and factor tilts

In depth11 min read · 15 sources

Equal weight means that every stock in the index gets the same weight, regardless of how large the company is. That sounds like more diversification (and it is), but it changes a lot more: equal-weighted indices have to sell winners and buy losers regularly, tilt towards smaller and cheaper stocks, and lag well behind in phases when a few giants carry the market. Here, an example with four stocks shows you how weights drift and what resetting them costs, which factors lie behind equal weight and what research has found out about the 1/N rule.

Two ways of weighting

With market-cap weighting (cap weight), each stock's share is its free-float market capitalisation as a proportion of the total market. With equal weighting (equal weight, 1/N), each of the N stocks gets the same share on the rebalancing date. Between two rebalancing dates, the weights change with prices:

The crucial difference lies in the last line: a cap-weighted index stays in balance by itself, because price and market capitalisation rise together. It only has to trade when stocks are added or removed or when share counts and free float change. An equal-weighted index has to reset at every rebalancing date. MSCI weights each issuer equally and rebalances quarterly as of the last business day of February, May, August and November; in between, the weights fluctuate with prices (MSCI Equal Weighted Methodology 2021). The S&P 500 Equal Weight is also reset quarterly (S&P DJI, Equal Weight FAQ 2025). For an overview of other departures from the plain index, see the article smart indexing.

Example with four stocks: drift and turnover

Illustrative figures: four stocks with market capitalisations in the ratio 50 : 25 : 15 : 10, and a portfolio of €10,000. In one quarter, A rises by 20%, B falls by 10%, C rises by 5% and D by 15%.

The return difference can be read off directly: equal weight beats cap weight when the smaller stocks do better than the large ones. In the example, Σ (0.25 − wᵢ) × rᵢ = −0.25 × 20% + 0 × (−10%) + 0.10 × 5% + 0.15 × 15% = −2.25 percentage points, exactly the difference between 7.50% and 9.75%. With equal weighting, the effective number rises from 2.90 to 4: diversification by weight increases.

What equal weighting changes in the MSCI World

Sources: MSCI World Equal Weighted Factsheet (2026); MSCI World Factsheet (2026). Sectors according to GICS. Between rebalancing dates, the weights of the equal-weighted index deviate from the average of 0.08% because they drift with prices.

Equal weighting turns an index dominated by US technology stocks into one in which industrials and financials lead and Japan weighs more than twice as much. It is therefore not just a question of diversification but a big bet against the structure of the market: ten times as much turnover and almost 5% tracking error relative to the parent index. Over the last ten years, this cost a lot of return (cumulatively +147.1% versus +250.7%), because it was precisely the largest US technology stocks that rose the most. Over three decades, the gap is around 0.8 percentage points a year. For the S&P 500, the longer track record looks different: from its launch in January 2003 to April 2023, the equal-weighted version returned 11.48% a year versus 10.29% (Edwards et al. 2023).

Factor tilts: size, value and reversal

Because small companies get the same weight as large ones, equal weight systematically favours smaller stocks, i.e. the size factor, whose premium Banz (1981) described and Fama and French (1993) included in their three-factor model (see Fama-French factors). For the S&P 500 Equal Weight, according to S&P Dow Jones Indices, the higher weight of the smaller constituents explains more than half of the historical excess return; on top of that comes a tilt towards value and against momentum (S&P DJI, Equal Weight FAQ 2025).

The tilt against momentum follows from rebalancing: if you regularly sell winners and buy losers, you profit when price movements reverse and lose when they continue. For individual stocks, there is evidence of both: one month's losers often beat the winners in the following month (Jegadeesh 1990), and over three to five years, large gains and losses partly reverse (De Bondt and Thaler 1985). This is the reversal effect. In between, over three to twelve months, trends tend to continue (Jegadeesh and Titman (1993); see momentum). Quarterly rebalancing sits right in the middle of this conflict.

Concentration phases: when equal weight lags behind

Equal weight does best when an upswing is broad and smaller stocks catch up, and worst when a few giants carry the market. This lies in the nature of stock returns: measured over their entire lifetime, a few stocks account for most of the wealth creation (Bessembinder 2018). A cap-weighted index lets these winners keep growing; an equal-weighted index cuts them back to 1/N at every rebalancing date. If such a phase lasts a long time, as in the last ten years, the shortfall adds up. If it reverses, equal weight catches up. That was the case in 2022, for example, when the S&P 500 Equal Weight beat the S&P 500 by around seven percentage points, according to S&P Dow Jones Indices (S&P DJI, Equal Weight FAQ 2025). Whether and when this happens cannot be predicted; for more on how concentrated world indices are today, see the article MSCI World vs ACWI vs FTSE All-World.

Costs, turnover and capacity

• Trading costs: turnover of 32.08% a year means that around a third of the index is bought and the same amount is sold. At assumed costs of 0.10% per euro traded (spread, market impact), this costs 2 × 32.08% × 0.10% = 0.064% a year, compared with only 0.006% for the MSCI World. These costs are not included in the TER but show up in the tracking difference.
• Small, less liquid stocks: because equal weight invests far more money in small companies than their market weight implies, the costs per euro traded are higher there than for the large ones.
• Capacity: the more money equal-weight funds manage, the more they are likely to move the prices of the smaller stocks on their rebalancing dates, similar to the index effect when stocks are added to an index (our assessment).
• DIY with individual stocks: if you equal-weight 20 or 50 stocks yourself, you realise gains at every rebalancing and pay the flat-rate withholding tax (Abgeltungsteuer); inside a fund, the reshuffling is tax-free for private investors. Directing new contributions into the underweight stocks reduces turnover and tax.

For advanced readers: rebalancing alpha and estimation error

The equation breaks down the difference in each period. Over many periods, a second component comes in: rebalancing itself. A portfolio that is repeatedly reset to fixed weights achieves, with the same expected returns, a higher geometric return than the fluctuating individual stocks do on average. This is the diversification return (see volatility drag and rebalancing). Constant mixes are a bet on fluctuation without a trend: they do better than buy and hold when prices move back and forth, and worse when a trend persists (Perold and Sharpe 1988).

As a rule for allocating across assets, 1/N has an advantage that has little to do with stock indices: it needs no estimates. DeMiguel et al. (2009) compared 14 optimisation models with 1/N in seven data sets; none was consistently better out of sample, measured by Sharpe ratio, certainty equivalent and turnover. According to their calculations, estimated mean-variance optimisation would need around 3,000 months of data with 25 assets to beat 1/N. For indices, this does not mean that 1/N across stocks is efficient, only that estimated “optimal” weights are often worse than no estimate at all (see covariance estimation).

What the research shows

Plyakha et al. (2014) compared equal-, value- and price-weighted portfolios of the stocks in major US indices. Rebalanced monthly, equal weight came out ahead in total return, four-factor alpha and Sharpe ratio. Part of the lead comes from higher loadings on systematic factors, but a considerable part, 42%, comes from the alpha generated by resetting to constant weights. Edwards et al. (2023) show a higher return for the S&P 500 Equal Weight than for the S&P 500 since 2003, with turnover of 22% instead of 5% a year over ten years.

The criticism of weighting by market capitalisation goes deeper: according to Arnott et al. (2005), indices weighted by sales, book value or dividends were more efficient in their tests than cap-weighted ones, because overvalued stocks automatically carry too much weight at market weight. Equal weight shares this idea without needing a metric. The opposing view: market weight is the only portfolio that all investors together can hold, and it is by far the cheapest to implement. If you deviate from it, you are betting on factor premiums that fail to appear in some decades.

Limitations and common mistakes

• Selling equal weight as “more diversification”: by weight yes, by risk not necessarily. Small companies fluctuate more, and the beta and drawdown of the MSCI World Equal Weighted were slightly above those of the parent index.
• Picking the period: from 2003 to 2023, equal weight looks good for the S&P 500; over the last ten years, it looks bad for the MSCI World. Both are samples, not rules.
• Overlooking factors: an equal-weight ETF is a combination of size, value and reversal. If you already hold small-cap or value ETFs, you double the bet.
• Underestimating costs: turnover, spreads in small stocks and a higher TER add up; compare the tracking difference against the fund's own index.
• Confusing regions with stocks: equally weighted regions (a third in the US) are something quite different from equally weighted stocks. Not investment advice.

How Portfolio Lab calculates it

In the Backtest, you can weight any selection equally and reset it with rebalancing monthly, quarterly, semi-annually, annually or by tolerance band; the average turnover per year appears in the results, and spreads and order fees can be entered as trading costs. This lets you compare a cap-weighted solution with an equally weighted selection over the same period. In Optimisation, the equally weighted portfolio, rebalanced quarterly, is the reference point on the efficient frontier and in the out-of-sample comparisons. As a market breadth indicator, the Sentiment page shows whether the equal-weighted S&P 500 is performing more strongly or more weakly than the cap-weighted one over 60 trading days. How strongly a portfolio loads on size and value is measured by the regression under Factors. Details on rebalancing and costs are in the backtest methodology.

This article explains a method; it is not investment or tax advice. Worked examples are simplified and rounded. All terms in the glossary.