How to Calculate Portfolio Beta: Formula, Excel Check and a Real 5-Stock Example

Michael Whitman · Trading since 2017, full-time since 2020 · YouTube creator, 137K subscribers · Published 2026-09-08 · Updated 2026-11-02 · Reviewed 2026-11-01

To calculate portfolio beta, multiply each holding’s beta by its share of the portfolio and add the results. Cash goes in at a beta of zero, and every beta has to be measured against the same index over the same window.

That last condition is where most of the work is. The multiplication is trivial; the betas are not. On a real five-stock portfolio below, three common ways of measuring beta gave three different answers from identical holdings on the same day.

Before you start

Each holding’s beta against a stated benchmark, since the figure is benchmark-specific. For a US stock portfolio that is usually the S&P 500, measured here with SPY. Write down the window too: one year of daily returns, two years of weekly, or five years of monthly, which is what Yahoo Finance labels “Beta (5Y Monthly)”.

The weight of each holding as a share of the total portfolio. Use today’s market value, not what you paid. A holding that doubled since you bought it carries twice the weight it had.

An understanding that beta measures one relationship and ignores everything else. It is the slope of your returns against the index’s returns. The beta page covers how it is estimated; this page is about putting the pieces together for a whole portfolio.

The steps

1. Pick one benchmark and one window for every holding

Mixing a 5-year monthly beta for one stock with a 1-year daily beta for another produces a weighted average of figures that measure different things. One index, one frequency, one end date.

2. Get each holding’s beta from one source

Either read them all from one site that states its method, or compute them yourself in Excel with =SLOPE(stock returns, index returns) on the same rows. SLOPE gives the same number as covariance divided by the index’s variance.

3. Divide each holding’s market value by the portfolio total

Include cash and anything else you hold. The weights should add to 100%.

4. Multiply each beta by its weight

A holding at 25% with a beta of 1.0727 contributes 0.2682. Cash contributes 0.

5. Add the contributions

The total is the portfolio beta. Above 1, the portfolio has moved more than the index over that window; below 1, less; below 0, it has tended to move the other way.

6. Write the window and date next to the answer

“0.73 against SPY, 60 monthly returns to 30 Oct 2026” is a result. “0.73” on its own is not, because the same holdings give a different figure on a different window.

7. Recompute it every 3 months, or after any large move

Weights drift with prices and each beta rolls forward a month at a time, so the figure changes even when you trade nothing.

A worked example

The portfolio. $100,000 on 30 Oct 2026: $25,000 Apple (AAPL), $20,000 Microsoft (MSFT), $20,000 JPMorgan Chase (JPM), $15,000 Johnson & Johnson (JNJ), $10,000 Exxon Mobil (XOM) and $10,000 in cash. The weights are 25%, 20%, 20%, 15%, 10% and 10%.

The betas. Each stock’s 5-year monthly beta against SPY, from 60 monthly dividend-adjusted returns ending 30 Oct 2026: AAPL 1.0727, MSFT 1.0472, JPM 1.0355, JNJ 0.2178, XOM 0.1647.

The arithmetic.

Read it plainly. Over those five years, when SPY moved 1% in a month, this mix moved about 0.73% on average. The two defensive names add about 0.05 between them; the three that track the market carry the rest.

Check it a second way. Build the portfolio’s own monthly return series from the same weights and regress it on SPY. The slope comes out at 0.733860, identical to the weighted sum, because a fixed-weight average of slopes is the slope of the average. That is the check a spreadsheet makes easy, and the monthly returns file has the 60 rows to run it on. Holding the cash in a Treasury bill fund (BIL, beta 0.0072) instead of at zero moves the result to 0.7346.

The original data

Same five stocks, three common windows, three answers. On one year of daily returns (251, to 30 Oct 2026) the portfolio beta is 0.39. On two years of weekly returns (104) it is 0.69. On five years of monthly returns (60) it is 0.73.

The gap comes from two stocks. On the one-year daily window Exxon Mobil’s beta is -0.72 and Johnson & Johnson’s is -0.19: over the past year both tended to fall on SPY’s up days and rise on its down days. On five years of monthly returns those same two stocks sit at 0.16 and 0.22. Apple moves too, from 0.56 on the daily window to 1.07 on the monthly one. All the figures, with their R-squared values, are in the betas file.

Yahoo’s figure is reproducible. Yahoo’s quote pages, read on 1 Nov 2026, showed Beta (5Y Monthly) of 1.07 for Apple, 2.22 for Nvidia, 0.21 for Johnson & Johnson and 0.21 for Exxon Mobil. Sixty monthly returns against SPY ending 30 Sep 2026 give 1.0686, 2.2140, 0.2138 and 0.2173, each within 0.01. Ending the window at 30 Oct instead gives 2.2046 for Nvidia and 0.1647 for Exxon Mobil, which do not round to Yahoo’s figures, so on 1 Nov the displayed figure appears to stop at the end of September.

And one month changes it. Roll that window forward to 30 Oct 2026 and Exxon Mobil’s beta drops from 0.2173 to 0.1647. October 2026 is the reason: Exxon rose 15.05% while SPY fell 0.91%, and the month that dropped out (October 2021) had both rising, by 9.61% and 7.02%.

How much each beta explains. R-squared is the share of a stock’s monthly swings that SPY accounts for. On the five-year window it is 0.45 for Apple, 0.38 for Microsoft, 0.44 for JPMorgan, 0.04 for Johnson & Johnson and 0.01 for Exxon Mobil. For the portfolio as a whole it is 0.72. A beta built on an R-squared of 0.01 is a slope through a cloud, which is why it can swing by 0.05 in one month.

The check. Nasdaq’s closing prices, which carry no dividends, give one-year daily betas of 0.5613 for Apple and -0.7083 for Exxon Mobil (Yahoo-based: 0.5648 and -0.7223) and a five-year monthly beta of 0.2019 for Johnson & Johnson against 0.2178. Different feeds and the missing dividends move the third decimal, not the story.

Few videos cover this. Of the 24,971 unique videos in this site’s search study, 3 put beta in the title, at a median of 45 views across 3 channels.

How to tell it worked

The weights add to 100%, cash included.

Every beta comes from 1 benchmark and 1 window, and both are written next to the answer.

The weighted sum matches a regression of the portfolio’s own monthly returns to at least 4 decimal places, if you have the return series.

And the figure is dated within the last 3 months.

What beta leaves out

Concentration. Beta averages away exactly what a concentrated portfolio is exposed to. In the example, Apple, Microsoft and JPMorgan each have a beta near 1.05, yet their monthly returns swung with a standard deviation of 7.12%, 7.59% and 6.96% over the five years, against 4.46% for SPY. Microsoft’s worst month lost 17.15% (June 2026); SPY’s worst lost 9.24% (September 2022). A beta of about 1 did not mean index-like risk for any of them. See concentration risk for how to measure that directly.

What the stocks do with each other. A correlation grid shows whether your holdings are the same bet. Beta only shows how each one has related to the index.

The benchmark you did not pick. Regressed on the Nasdaq-100 fund QQQ instead of SPY, the same portfolio has a beta of 0.43, with an R-squared of 0.45. Neither number is wrong; they answer different questions.

Why the figure drifts

Prices change the weights. If the example portfolio had been bought on 29 Oct 2021 in the same dollar amounts, with the cash held in the Treasury bill fund BIL, and left alone, it would be worth $212,230 on 30 Oct 2026, with Apple at 27.78% of it and Exxon Mobil at 16.31% instead of 10%.

Using today’s betas, that drifted mix (the cash still counted at 0) has a beta of 0.7356 against 0.7339 for the original weights. Here the winners sat on both sides of 1 and nearly offset each other; in a portfolio where the high-beta names did all the rising, the same drift pushes the figure up. A rebalance resets the weights.

The betas roll. Each month a new return enters the window and an old one leaves, as the Exxon Mobil example shows.

When it fails

Treating it as a risk summary. A beta near 1 reads as market-like risk. The single stocks above show it can sit on top of swings far larger than the market’s.

Mixing windows or benchmarks. The example spans 0.39 to 0.73 on the same holdings; a mix of sources can land anywhere in between and mean nothing.

Leaving out cash. Ten percent in cash at a beta of zero is what pulls the example below the 0.82 the five stocks alone would give.

Using purchase weights. Today’s market values set the weights.

Reading a low-R-squared beta as a hedge. Exxon Mobil’s negative one-year beta came with an R-squared of 0.12. It describes a year in which oil and stocks often moved apart; it is not a promise that the stock will rise the next time the index falls.

Sizing a hedge from the wrong window. If you use beta to size an index hedge, the window you choose sets the hedge size, and on this portfolio that choice nearly doubles it.

Questions people ask

What is the formula for portfolio beta?

Portfolio beta = w1 x b1 + w2 x b2 + … + wn x bn, where each w is a holding’s share of the portfolio and each b is its beta against the same index. Cash has a beta of 0. For a single stock, beta = covariance(stock, index) / variance(index).

How do you calculate portfolio beta in Excel?

Put each holding’s weight and beta in two columns and use =SUMPRODUCT(weights, betas). To get the betas themselves, put monthly returns for each stock and the index in columns and use =SLOPE(stock, index). Running SLOPE on the portfolio’s own return column should match the SUMPRODUCT.

What is a good beta for a portfolio?

There is no good or bad value, only a match or a mismatch with the market exposure you meant to have. A reading of 0.73 is roughly three quarters of the index’s monthly moves, measured over the stated window.

Why does Yahoo show a different beta from another site?

Different windows, frequencies and end dates. Yahoo uses five years of monthly returns; others use one or three years, daily or weekly. On the example, those choices span 0.39 to 0.73.

Related

Beta explains how a single stock’s figure is estimated and what the slope means. Correlation is the figure that shows what beta cannot: whether your holdings move together. And the Sharpe ratio is the other common portfolio statistic, scoring return against total volatility rather than against the index.

This page is educational, not financial advice. Test every idea on your own charts before risking money.