What is modern portfolio theory?

How can a portfolio carry less risk than its assets?

Two assets that each swing 20% a year can form a portfolio that swings 14%.

Modern portfolio theory explains how that is possible. It is the method Harry Markowitz published in 1952, in the article "Portfolio Selection" in the Journal of Finance, to evaluate each investment by the effect it has on the portfolio as a whole.

Its central idea fits in one sentence. A portfolio's risk depends less on each asset than on the way its assets move together.

The opening example is hypothetical. The full calculation is below, and any reader can repeat it in a spreadsheet.

What changed with modern portfolio theory?

The theory changed the unit of analysis: investors began to evaluate their investments as a system. Traditional investing tended to look at each asset separately, with a large dose of intuition.

Markowitz proposed measuring. Each investment gets comparable numbers, and those numbers make it possible to calculate what happens when it is combined with the others.

Measurement has an added advantage. Two portfolios built on different criteria can be placed side by side and compared using the same variables, which is impossible when each decision rests on the impression of whoever makes it.

The practical consequence is significant. An asset that is risky on its own can lower the portfolio's risk if it moves differently from what is already in it. That statement sounds contradictory until you look at the numbers.

What variables does the theory work with?

The theory works with three variables: expected return, risk and correlation. Each one answers a different question about the investment.

What does expected return mean for a portfolio?

Expected return is the average return an investor anticipates for a given period. In a portfolio, that return is the average of each asset's return, weighted by the share each one holds. A portfolio with 60% in an asset with an 8% expected return and 40% in another with 4% has an expected return of 6.4%.

How is the risk of an investment measured?

Risk is measured by volatility, the dispersion of returns around their average. A stock with 20% annual volatility strays much further from its average return than a bond with 4%.

Correlation is an indicator between -1 and 1 of how two assets move together. With a correlation of 1, both rise and fall at the same time. At 0, they move independently. At -1, when one goes up, the other goes down.

Of the three variables, correlation is the only one that exists solely between assets. An isolated asset has an expected return and a volatility, but no correlation with anything. That is why Markowitz needed to look at the whole.

Why does correlation reduce the risk of the whole?

Correlation reduces risk because one asset's declines sometimes coincide with stable or positive periods for the other, and the swings partly offset each other.

How do swings between assets offset each other?

A boat with two passengers shows the mechanism. If both lean to the same side at the same time, the boat tilts with the full force of both. If each moves independently, many of those movements cancel out and the boat rocks less.

The same happens with assets. When correlation is high, declines arrive together and add up. When it is low, a decline often coincides with an asset that holds steady, and the portfolio absorbs the blow.

The effect is gradual. A correlation of 0.5 already offsets part of the swings, though less than a correlation of 0. Every reduction in correlation takes some risk out of the whole, as long as the weights and the assets stay the same.

Return does not behave this way. A portfolio's return is always an average, while its risk can fall below the average. That asymmetry is the reason diversification exists.

What does it look like in a numerical example?

The effect is clear with two identical assets combined in equal parts. The example is hypothetical and does not correspond to any real asset.

How does changing correlation affect volatility?

Asset A and asset B each have an expected return of 8% a year and a volatility of 20% a year. A portfolio puts 50% in each.

The portfolio's expected return is 8%, regardless of correlation. Volatility does change, and it is calculated with this formula:

σ portfolio = √(w₁²σ₁² + w₂²σ₂² + 2·w₁·w₂·ρ·σ₁·σ₂)

In the formula, w is each asset's weight, σ its volatility and ρ the correlation between the two. These are the results for different correlations:

Correlation between A and B Expected return Portfolio volatility
1 8% 20.0%
0.5 8% 17.3%
0 8% 14.1%
-0.5 8% 10.0%
-1 8% 0% (theoretical case)

Scale helps put it in perspective. On $10,000, a volatility of 20% equals a standard deviation of $2,000 around the year's average result. With a correlation of 0, that deviation drops to about $1,414.

The only variable that changed from one row to the next was correlation. The assets, their returns and their weights are the same in all five rows.

The -1 row marks the mathematical limit of the effect. It helps explain the formula, but a zero-risk portfolio built from two risky assets is a textbook case. The middle rows are the ones that show the size of the effect in practice.

The formula also explains where the reduction comes from. The first two terms depend only on each asset on its own. The third, the only one that contains correlation, is the one that adds or subtracts risk depending on how the assets move together.

Where does each investment sit on the risk-return plane?

Each investment occupies a point on a plane with risk on one axis and expected return on the other. Markowitz showed that every possible investment in the market can be placed on that plane, compared with the others and combined to create portfolios.

In the example, A and B occupy the same point: 20% risk and 8% return. The portfolio with a correlation of 0 appears at the same height, but further to the left, at 14.1% risk.

Combining different assets in different proportions fills the plane with new points. Some combinations offer more return for the same risk than others. Choosing among them is the subject of the program's next classes.

The plane has immediate practical value. It shows whether a new investment moves the portfolio up, to the left or into a worse area, before committing money to it.

What are the most common mistakes when applying it?

The first mistake is measuring diversification by the number of assets. Ten stocks from the same sector, highly correlated with each other, behave almost like a single large position.

The table shows this in the first row. With a correlation of 1, combining assets does not reduce risk at all. A portfolio with many names and high correlations can carry high risk without its owner noticing.

The second mistake is treating correlation as a fixed figure. Correlations change over time, and diversification that worked with data from three years ago may have weakened today.

External shocks, such as a financial crisis or a regulatory change, can alter correlations in a short time. That is why it pays to review the portfolio and make tactical adjustments to maintain diversification, because a calculation done once grows stale.

What can readers check for themselves?

Readers can verify the mechanism with public data and a spreadsheet. These are the steps:

  1. Choose two indexes or funds with free historical price data.
  2. Download the monthly closing prices for the past three years and calculate each month's return.
  3. Calculate the correlation between the two series with the CORREL function and the volatility of each with STDEV.
  4. Create a third column with the return of a 50/50 portfolio and calculate its volatility.
  5. Repeat the exercise with a different three-year window.

How can you tell whether the prediction holds?

The prediction is specific and can be refuted. If the correlation comes out below 1, the volatility of the 50/50 portfolio will be lower than the average of the two individual volatilities. If the correlation changes between the two windows, the risk reduction will change too.

What changes when you look at the whole portfolio?

Looking at the whole portfolio changes the question worth asking about a new investment. The useful question is how that asset moves relative to what is already in the portfolio.

An attractive asset that moves like everything else adds return, but it also concentrates risk. A modest one that moves differently can do more for the stability of the whole.

The "Modern Portfolio Theory" class, part of Cronoss's Portfolio Fundamentals and Risk Management program, covers these principles with professor Rafael Gómez Gómez and shows how to analyze a portfolio as a system: Fundamentos de Portafolios de Inversión (Fundamentals of Investment Portfolios)

Educational content. Cronoss is not a broker, investment firm or financial adviser, and nothing above is an investment recommendation.