Diversification Drives Higher Expected Returns, Not Just Less Risk

Published 06/11/2020

 Diversification is important. We all know that. At this point, diversification is a well-known story and it is intuitive as well. Since we cannot predict the market’s future, investing in a single company creates a lot of risk. The company could be one that goes bankrupt (for example, from 2005 to 2015, there were 303 bankruptcies of public companies or about 1% of all companies per year1), or they could be an Amazon or Facebook, which would have delivered holders incredible returns over the past 10–15 years.

The only thing we know for sure is that there is no way of telling which path a single, randomly chosen stock will take beforehand. This uncertainty of whether an individual will become a multimillionaire, lose everything or something in between, is why we diversify. Diversification reduces risk.

Diversification From a Risk Perspective

The idea that diversification is the only free lunch in investing is often attributed to Nobel Laureate Harry Markowitz. Instead of the crazy, uncertain paths in the example above, an investor can invest across the market, thereby diversifying away company-specific risk, while earning the market return. Companies go bankrupt all the time, but the entire market will not. Diversification has allowed investors to have more conviction in their future outcomes and clarity for their long-term financial plans by building more efficient portfolios. More return for each unit of risk and less risk for each unit of return increases the efficiency of a portfolio. Diversification is the cornerstone of proper investment management.

For example, the average U.S. stock has a standard deviation of about 35%.2 This means that if a stock has an expected return of 7%, the yearly return will be between –28% (7%-35%) and +42% (7%+35%) about two out of every three years. A range of –28% to +42% is a very wide band of possible outcomes. However, if a portfolio is diversified across all 3,000+ stocks that make up the U.S. market, now the standard deviation is below 20%. This means that assuming the same 7% expected return, the yearly outcomes about two out of three times will be between –13% and +27%, a tighter band than the previous example. In the second example, the company-specific risk has been diversified away, leaving only the market risk component. In the single-stock portfolio, the portfolio fully retained both types of risk: market and company-specific. One thing that Markowitz discovered is that since company-specific risk is diversifiable, it does not drive additional expected return. The arithmetic average returns are the same, but the risk of the index is much lower than the risk of the single stock. This means the diversified portfolio of U.S. stocks is much more efficient.

In addition, the diversified portfolio has an advantage in growth of wealth because of “volatility drag” on portfolios that cause geometric or compounded returns (average returns in dollar terms) to be lower than arithmetic returns the more volatile a portfolio. The simple logic is that if a portfolio falls by 20%, then it must increase in value by 25% to get back to even. Volatility drag is the formal name for the concept that a given percentage drop has a larger effect on dollar value than the subsequent same percentage gain.

Assume $100 is invested in the two portfolios from the earlier example:

1.      Single-Stock Portfolio: 7% expected arithmetic return, 35% standard deviation

2.      U.S. Market Portfolio: 7% expected arithmetic return, 20% standard deviation

They will have different geometric returns as approximated by the following formula:3

Geometric or Compounded Return = Arithmetic Return – (StDev^2)/2


After 30 years, the $100 has turned into:

Reduction of volatility has a dramatic impact on the growth of wealth, and this is the tangible benefit of diversification. There are two items worth mentioning about the analysis above. First, it ties out with known data, as over half of publicly listed stocks have negative total returns versus Treasury bills over their lifetime.4 Second, this analysis presupposes that the arithmetic returns are the same for an index versus an individual stock, however, in the next section, we will discover that this is not true either. The expected arithmetic return for a single-stock holding should be lower than for an index.

 Diversification From a Returns Perspective

The benefit of diversification has mostly been researched as a risk story, but there is a very important return benefit as well. As in the previous section, we will examine the difference between investing in a single stock versus investing in an index. If one is picking a stock at random, then the median stock return should be used as a proxy for the average stock. We looked back at the data for U.S. stocks from 1927–2017 using the CRSP U.S. stock database.5

The average yearly U.S. market return over that time period (91 years) was 12.01%. The median stock return was 8.30% on average. This means that investing in the overall index versus randomly picking a single stock would have earned on average 3.71% more per year. Again, this is a natural outcome of stock returns as stock returns are skewed, not uniform. In any given year, there are a few big winners, while the majority of stocks have returns that are less than the average. That means it should not be a surprise that the index would have higher yearly returns than the median stock, but this phenomenon has not received nearly enough attention.

Further, by looking at a single-stock’s lifetime price return versus the Russell 3000 Index, going back to 1980–2014, we know that roughly 2/3 of the universe of single stocks underperformed the Russell 3000. The median stock underperformed the index by –54%, with the extreme winners being two standard deviations over the mean as represented by the chart below.6


Combining Risk and Return

Going back to the example of compounded annual growth over 30 years, let us evaluate the combined effects of both a higher standard deviation and a lower expected return.

1.      Single-Stock Portfolio: 4.29% expected arithmetic return (7% index return – 3.71% historical spread between single stocks and index returns), 35% standard deviation

2.      U.S. Market Portfolio: 7% expected arithmetic return, 20% standard deviation


After 30 years, the $100 has turned into:

We can see that in this scenario, which considers both risk and return, the diversified approach is clearly superior to an approach concentrated in a single stock. The single stock’s increased volatility when considered cumulatively with its lower expected return actually results in a negative -1.84% compounded annual growth rate. This will be shocking to many investors, who expect single stocks to deliver returns similar to the market over the long term. In reality, most single stocks deliver much higher volatility and lower compound returns.


Most of the financial world understands the benefits of diversification from a risk perspective, but there is a benefit from a returns perspective as well. Investing in single stocks or a concentrated portfolio is significantly riskier (from a standard deviation and long-term outcome perspective) than investing in a diversified index. Additionally, single stocks or a concentrated portfolio have lower returns than index returns, which means that single-stock and concentrated portfolios are less efficient on two dimensions: risk and return. The goal of a well-constructed portfolio is to maximize returns for the lowest amount of risk. Single-stock investing accomplishes the opposite: It lowers returns for greater risk. Single-stock or concentrated-stock investors could get lucky, but the less diversified the portfolio, the more the odds tilt against these portfolios versus a diversified portfolio in the long run.


1 “Public Company Bankruptcies From 2005 to 2015.” The Wall Street Journal, Accessed August 16, 2018. 

2 “Chart 18: Today’s Ultra-Low SPX Realized Vol Is Driven by (i) Single Stock Vol at 27yr Lows and (ii) Policy Polarization Crushing Stock Correl to Levels Rarely Seen Outside of the 90s & the Tech Bubble.” Financial Times, Accessed August 16, 2018.

3 Michael Kitces, “Volatility Drag and Its Impact on (Arithmetic) Investment Returns in Monte Carlo Analysis.” Nerd’s Eye View at Kitces.com, December 27, 2017.

4 Cormac Mullen, “Lesson of the Century: Most U.S. Stocks Can’t Even Beat a T-Bill.” Bloomberg, February 1, 2017.

5 Data from Dimensional Fund Advisors and CRSP.

6 Michael Cembalest, “The Agony & the Ecstasy: The Risks and Rewards of a Concentrated Stock Position.” Eye on the Market (Special Edition), J.P. Morgan, 2014.