Skip to content
Menu
Finabase Learn

Strategies · Long-term investing

Equal weight vs market-cap weight: RSP against SPY, tested

How equal-weight and cap-weight indexes differ, the rules of each, and what RSP did against SPY from 2003 to 2026: level overall, far ahead to 2015 and far behind since.

On this page
  1. Two ways to weight the same companies
  2. The rules
  3. What the backtest shows
  4. Where it fails: the losing periods
  5. Compared with its neighbours
  6. How Finabase does it
  7. Frequently asked questions
  8. Summary
  9. Check your understanding

An index has to decide how much of each company to hold. The S&P 500 holds them in proportion to their market value, so the largest company can be several hundred times the weight of the smallest. An equal-weight index holds the same 500 companies at the same weight each, about 0.2%, and resets them every quarter.

Same companies, different sizes, and over two decades the two have taken turns being far ahead. This lesson states the rules of both methods, tests the equal-weight fund RSP against the cap-weight fund SPY from May 2003 to October 2026, and builds a small equal-weight portfolio by hand to show what the resetting costs. The headline is that they finished level. The path there is the lesson.

Two ways to weight the same companies

Suppose an index has three companies worth $600 billion, $300 billion and $100 billion.

Company Market value Cap weight Equal weight
A $600 billion 60% 33.3%
B $300 billion 30% 33.3%
C $100 billion 10% 33.3%

If A's share price doubles, the cap-weight fund does nothing: its holding of A doubles in value along with A's weight in the index. The equal-weight fund now has too much of A and must sell some to buy B and C. That difference in behaviour drives everything else.

Cap weight is the only weighting every investor could hold at once, since it is simply a slice of the whole market. It follows that the average dollar invested in those companies earns the cap-weight return before costs, a point William Sharpe made in 1991. It trades only when companies enter or leave the index. Its weakness is concentration: when a few companies grow very large, the fund becomes a bet on them (Market concentration: what mega caps mean for S&P 500 investors has the figures).

Equal weight treats the 500th company like the first. Compared with cap weight it holds far more of the smaller members, and because cheap, fallen stocks have small market values, more of those too. It is a tilt toward two of the classic factors, size and value, plus a mechanical rule to sell what went up.

The rules

An educational outline. The cap-weight rule first, because it is the baseline:

  1. Hold every company in the index in proportion to its market value.
  2. Trade only when the index adds or removes a company.

The equal-weight rule, as the S&P 500 Equal Weight Index and RSP apply it:

  1. Hold every company in the index at the same weight: one divided by the number of companies.
  2. Once a quarter, on the index's rebalancing date in March, June, September and December, sell part of every holding that is above that weight and add to every holding below it.
  3. When the index adds or removes a company, do the same in the fund.
  4. Between resets, let the weights drift.

A reader does not carry these rules out by hand for 500 stocks; the fund does. The rule to follow as a holder is simply to buy the fund and keep it. The second test below applies the same rule by hand to nine funds, which is small enough to see every moving part.

What the backtest shows

The test, stated in full so it can be checked:

  • Data. Month-end closing prices of RSP and SPY from May 2003, RSP's first full month, to October 2026: 282 months, about 23 and a half years.
  • Trading. Each fund is bought once and held. All the resetting happens inside RSP.
  • Costs. Both funds' fees and internal trading costs are already in their prices: 0.20% a year for RSP and about 0.09% for SPY (issuers' pages, November 2026). No cost is charged for the single purchase of either.
  • Left out. Dividends and taxes. The two funds' dividend yields have been close, so leaving them out lowers both lines by a similar amount.
RSP (equal weight) against SPY (cap weight), monthly, 2003–2026, fund prices, no dividends $1 $2 $5 Growth of $1 SPY cap weight RSP equal weight 0% −20% −40% −60% Fall from the peak 2005 2010 2015 2020 2025
RSP (equal weight) against SPY (cap weight), month-end fund prices, May 2003 to October 2026, dividends not included. Top: growth of $1 (log scale). Bottom: the fall from the highest value so far.
Measure RSP, equal weight SPY, cap weight
Total return 679.4% 679.5%
Return per year (CAGR) 9.2% 9.2%
Deepest fall from a peak −56.8% −52.2%
Longest time below a peak 65 months 64 months
Fund fee per year 0.20% About 0.09%

After 23 years the two funds had gained the same amount to within a tenth of a percentage point. RSP got there with a deeper fall in the financial crisis: it peaked in May 2007, five months before SPY, and was down 56.8% by February 2009.

The two halves of the period

Period RSP, equal weight SPY, cap weight
May 2003 – December 2014 9.5% a year 6.7% a year
December 2014 – October 2026 8.8% a year 11.6% a year

A level finish hides two long, opposite runs. For the first 11 and a half years equal weight was ahead by nearly three points a year. For the next 12 it was behind by nearly three.

RSP divided by SPY, both starting at 1, month-end closes, 2003–2026 1.00 1.20 Equal weight relative to cap weight 2005 2010 2015 2020 2025
RSP divided by SPY, month-end prices. A rising line means equal weight is gaining on cap weight. The peak is March 2015.

The ratio of the two funds peaked in March 2015, when RSP had gained 36% more than SPY since the start. By October 2026 the whole lead was gone. An investor's opinion of equal weight depends almost entirely on which of those years they held it.

What the resetting costs: nine funds by hand

RSP's trading is hidden inside the fund. To see it, the same rule can be applied to the nine original sector funds of the S&P 500 (technology, financials, energy, health care, utilities, consumer staples, consumer discretionary, industrials and materials) at one-ninth each, from January 1999 to October 2026, with 0.10% charged on every trade.

Reset schedule Return per year Deepest fall Portfolio traded Cost paid
Every December 7.1% −50.2% 1.2 times 0.7%
Every quarter 7.1% −50.2% 2.5 times 1.3%
Every month 7.0% −50.6% 4.3 times 2.1%
Never 6.5% −49.9% Nothing 0.1%
SPY, for comparison 6.6% −52.2%

"Portfolio traded" counts the resets only, not the first purchase. Resetting did add something here: the three reset schedules returned about half a point a year more than the unreset mix, which behaved like SPY. Trading more often than once a year added nothing and cost more. Nine sector funds trade at tiny spreads; 500 individual stocks, many of them small, do not, which is part of why RSP's fee is double SPY's.

Where it fails: the losing periods

Equal weight loses to cap weight whenever the largest companies do better than the typical one.

Year RSP SPY
2020 10.2% 16.2%
2023 11.7% 24.3%
2024 11.0% 23.3%
2025 9.3% 16.4%
2026, to October 14.0% 10.8%
  • December 2022 to October 2026: RSP returned 12.0% a year and SPY 19.4%. In 2023 and 2024 the gap was more than 12 points each year. A few very large technology companies produced a large part of the index's gain, and an equal-weight fund holds each of them at 0.2%.
  • 2007 and 2008: RSP lost 0.3% and then 41.1%; SPY gained 3.2% and then lost 38.3%. Equal weight holds more in smaller and more indebted companies, which fell hardest. Its fall from peak was 4.6 points deeper.
  • 2015 and 2018: RSP lost 4.3% and 9.5%; SPY lost 0.8% and 6.3%.

It wins in the mirror-image conditions: 2009 (42.2% against 23.5%), when the most beaten-down companies rebounded most, and 2022 (−13.2% against −19.5%), when the giants fell most.

Other ways it fails or misleads:

  • It sells every winner. A company that grows from the 400th largest to the 5th is trimmed back to 0.2% four times a year on the way. Cap weight keeps all of that growth. Much of a stock market's long-run gain comes from a small number of such companies.
  • More trading, more tax inside and outside the fund. The exchange-traded fund structure shelters most of it, but turnover is still a cost.
  • The comparison depends on the start date. A study ending in 2014 called equal weight a clear winner. One starting in 2015 calls it a clear loser. Both describe the same rule.
  • It is not more diversified in the sense that matters in a crash. It fell further in the one severe bear market in the test.

Compared with its neighbours

Weighting Tilts toward Trading In this test, 2003 to 2026
Market cap (SPY) Nothing: it is the market Almost none 9.2% a year, fell 52.2%
Equal (RSP) Smaller, cheaper, recently fallen Every quarter 9.2% a year, fell 56.8%
Price (the Dow) Whichever shares have a high price Almost none Not tested here
Fundamental (by sales or profits) Cheaper companies Yearly Not tested here

Stock market indexes: S&P 500, Dow, Nasdaq and Russell explained covers price weighting and the Dow in more detail.

How Finabase does it

Finabase does not build weighted indexes, but it shows both funds and the concentration that separates them.

  • Comparison. The link above draws RSP and SPY from the same starting value. The default range is two years, which shows the recent gap and not the 23-year picture on this page (Compare stocks side by side on finabase.io).
  • Heatmap. The market heatmap sizes each company's tile by market value, so the share of the cap-weighted index taken by the largest names is visible at a glance (Market overview, heatmap and sectors on finabase.io).
  • Screener. With the scope set to the S&P 500 and the order to market cap, the screener lists the members from largest to smallest (Use the Finabase stock screener and screener builder on the web).
  • Backtest and Strategy. The Backtest tool trades indicator signals on one symbol with a fixed $10,000 per signal and no trading cost, and the Strategy tool replays weighted rules over two years, also without costs. Neither tests a weighting method over decades, and two years would have shown only the losing half of this story.

Frequently asked questions

Is an equal-weight S&P 500 fund better than a regular one?

Neither was better over the full test. From May 2003 to October 2026 RSP and SPY both returned 9.2% a year at month-end prices. RSP was well ahead until 2015 and has been well behind since, and it fell further in 2008.

Why does equal weight fall behind when big technology stocks lead?

Because it holds them at 0.2% each. When a handful of the largest companies produce much of the index's gain, a fund that holds them in proportion to their size captures it and an equal-weight fund mostly does not. That was the case in 2023 and 2024.

How often does RSP rebalance?

Four times a year, following the S&P 500 Equal Weight Index. Each time, it trims the holdings that rose above an equal share and adds to those that fell below.

Does equal weight reduce risk?

It reduces dependence on a few companies. It did not reduce the fall in a bear market: RSP's deepest fall in the test was 56.8% against 52.2% for SPY.

Summary

Cap weighting holds companies in proportion to their market value and barely trades; equal weighting holds them all at the same share and resets every quarter, which tilts it toward smaller and cheaper companies and makes it sell whatever has risen. From 2003 to 2026 the equal-weight fund RSP and the cap-weight fund SPY finished level at 9.2% a year, but RSP led by almost three points a year through 2014 and trailed by almost three afterwards, and it fell further in 2008. The choice between them is a choice about concentration: equal weight avoids depending on a few giants and gives up their gains when they lead.

Check your understanding

4 questions on this lesson. Each answer comes with the reason, and nothing is scored.

  1. A company in the S&P 500 doubles in price over a quarter while the rest of the index is flat. What does each kind of fund do at the next reset?

    1. Both sell some of it
    2. The cap-weight fund does nothing; the equal-weight fund sells part of it and buys the others
    3. The cap-weight fund buys more; the equal-weight fund does nothing
    Show the answer

    B. The cap-weight fund does nothing; the equal-weight fund sells part of it and buys the others In a cap-weight fund the holding and the index weight rise together, so no trade is needed. An equal-weight fund now holds too much of the winner and trims it back to an equal share, which is why it trades every quarter and never keeps a large position in a company that keeps growing.

  2. RSP and SPY both returned 9.2% a year from 2003 to 2026. RSP led by about three points a year to 2014 and trailed by about three afterwards. What is the sound conclusion?

    1. The two methods are interchangeable
    2. Equal weight is the better method because it led first
    3. The result depends on the period: equal weight does well in broad markets and badly when a few very large companies lead
    Show the answer

    C. The result depends on the period: equal weight does well in broad markets and badly when a few very large companies lead A level finish hides two long, opposite runs with a clear cause. The funds are not interchangeable: an investor who held only the second half earned 8.8% a year in RSP against 11.6% in SPY.

  3. Why did RSP fall further than SPY in 2007 to 2009 (56.8% against 52.2%)?

    1. It charges a higher fee
    2. It holds much more in smaller and weaker companies, which fell hardest
    3. It was not rebalanced during the crisis
    Show the answer

    B. It holds much more in smaller and weaker companies, which fell hardest Equal weight gives the 500th company the same share as the first. Smaller, more indebted companies fell most in the crisis, so spreading money evenly across names did not mean less risk.

  4. In the nine-sector test, resetting every month returned 7.0% a year and resetting every December 7.1%. What does that suggest about how often to reset an equal-weight mix?

    1. More frequent resets add return
    2. Resetting more often than yearly added trading and cost without adding return
    3. Resetting never matters
    Show the answer

    B. Resetting more often than yearly added trading and cost without adding return Monthly resets traded the portfolio 4.3 times over and cost 2.1%, against 1.2 times and 0.7% for December resets, for a slightly lower return. Not resetting at all did matter: that mix returned 6.5%.

Sources and further reading

  1. S&P Dow Jones Indices — S&P 500 Equal Weight Index (each company at the same weight, reset quarterly)
  2. Invesco — S&P 500 Equal Weight ETF (RSP), fund page: expense ratio 0.20% (read November 2026)
  3. Victor DeMiguel, Lorenzo Garlappi and Raman Uppal, “Optimal Versus Naive Diversification: How Inefficient Is the 1/N Portfolio Strategy?” Review of Financial Studies 22(5), 2009, pp. 1915–1953
  4. Yuliya Plyakha, Raman Uppal and Grigory Vilkov, “Why Does an Equal-Weighted Portfolio Outperform Value- and Price-Weighted Portfolios?” working paper (2012)
  5. William F. Sharpe, “The Arithmetic of Active Management,” Financial Analysts Journal 47(1), 1991, pp. 7–9
  6. Finabase chart data — RSP, SPY and nine sector SPDR funds, month-end closes, to October 2026; results computed by Finabase