An article on diversification from our investment partners:

It is generally accepted that diversification is a good thing for investors. One reason is that holding many different assets reduces the volatility (or variability of returns) of your portfolio; when one asset performs badly another may perform better. Another reason is that it can increase the probability that a portfolio will outperform the market.

There is a large dispersion of performance returns among stocks in any given period of time, and we use information in stock prices to systematically identify some of the characteristics that explain those differences in performance. With this knowledge, we seek to identify groups of stocks that we expect to have a higher return than the market as a whole, and we build portfolios that emphasise those groups of stocks.

Within these groups, the contribution to any outperformance will be unequal; some stocks may have performed well, others may have had average or poor returns. Because it is impossible to predict which stocks will do well, a portfolio that is not well-diversified across an asset class may risk missing the best-performing stocks. For this reason, we aim to have broad exposure to stocks in an asset class.

New research from Dimensional Fund Advisors considers this point by comparing characteristics and outcomes of simulated portfolios to the market. The study simulated thousands of portfolios with a weighted emphasis on those groups of stocks with higher expected returns. To analyse the impact of diversification, the portfolios varied in size with 50, 200, 500 and 1,000 stocks in each. These were compared to the Russell 1000 Index, a well-known index of the largest 1,000 US stocks where each company is weighted in proportion to its relative size.

The study finds that as the number of stocks in the simulated portfolio is increased from 50 to 1,000, there is a small decrease in its overall volatility of the portfolio and a reduction in how much the portfolio’s characteristics deviate from the market. This finding is important because not all deviations from the market are linked to out-performance, and you don’t need to make significant deviations from the market to outperform it.
That result was supported by another finding—as you increase the number of stocks in the simulated portfolio, its probability of outperforming the market increases. Over a 10-year period, a simulated portfolio of 50 stocks was likely to outperform the market 69% of the time, while a simulated portfolio of 1,000 stocks was expected to beat the market 96% of the time (see chart).
The conclusion we draw is to build diversified portfolios that continuously focus on all the stocks that are expected to deliver higher returns.

Estimated Probability of Outperforming the Weighted Russell 1000 Index over 10-Years.

Frank Russell Company is the source and owner of the trademarks, service marks, and copyrights related to the Russell Indexes

Sources: Wei Dai, “How Diversification Impacts the Reliability of Outcomes,” Dimensional Fund Advisors white paper, November 2016. To simulate US large cap portfolios with different diversification levels for each year in the sample period (1979–2016), Wei used a statistical method called bootstrapping. This enabled Wei to calculate the average volatility and correlation across 10,000 simulations. The projections or other information generated by bootstrapped samples regarding the likelihood of various investment outcomes are hypothetical in nature, do not reflect actual investment results and are not guarantees of future results. Results will vary with each use and over time.

Diversification does not eliminate the risk of market loss.