One argument for more concentrated portfolios is the concept of active share as a metric for active management. 
In a paper published in 2009 by Martijn Cremers and Antti Petajisto, then of the Yale School of Management, the concept of active share – which measures the percent of a portfolio’s holdings that differ from its benchmark – was introduced.
From their analysis, the authors concluded that active share is a better measure of active management than tracking error alone, as well as a better indicator of future outperformance.
According to the study, as active share measures differentiation from the benchmark, it is a proxy for stock selection. Tracking error, on the other hand, measures the volatility of portfolio returns in excess of the benchmark, and thus is a proxy for systemic factor risk. Active share ranges from 0 per cent (a pure index fund) to 100 per cent (fully active and completely different than the benchmark).
The study considers a US manager with an active share of 60 per cent or greater as truly active, and concludes that managers with high active share have historically outperformed.
Portfolios with higher active share tend to outperform
Source: Lazard Asset Management
As reported in Petajisto’s 2013 update, Active Share and Mutual Fund Performance, the above chart shows the relative performance of US all-equity funds from 1990 to 2009, segmented by five assigned active share and tracking error categories defined by the authors.
One should note that in this context the ‘Concentrated’ category does not imply a small number of holdings as thought of commonly, but rather this definition is for portfolios that combine high active share and high tracking error.
In the common use definition of concentrated portfolios, which we consider both the Stock Pickers and Concentrated categories, the results reflect strong outperformance for high active share portfolios. It is also important to realise that concentrated portfolios engender high active share because they hold a limited number of ideas and therefore only have a small overlap with an index.
Rise in close indexing
Source: Lazard Asset Management
High active share portfolios outperform across market caps and the performance of these funds held up well through the financial crisis.
As noted by the author: “I found that the most active stock pickers have been able to add value to their investors, beating their benchmark indices by about 1.26 per cent per year after all fees and expenses. Factor bets have destroyed value after fees. Closet indexers have essentially just matched their benchmark index performance before fees, which has produced consistent underperformance after fees. The results are similar during the 2008–09 financial crisis, and they also hold separately within large-cap and small-cap funds.”
Another key argument in the paper is that the ‘average’ active manager appears to underperform because of the inclusion of closet indexers in typical sample sets. Closet indexers, in our opinion the perfect example of over-diversifiers, typically maintain positions that overlap closely with the benchmark to ensure that their performance does not deviate significantly from the index, while many times claiming to be active managers.
We believe closet indexers may prove to be significantly expensive managers, as their strategies are unlikely to provide meaningful outperformance, particularly net of fees. It is also noted in the study that the amount of assets managed by closet indexers exploded in the mid-1990s.
In 2009, these managers represented about a third of the total assets, an increase from approximately 1 per cent in 1980, as illustrated in above diagram. The rise of closet indexers and their classification as active managers helps to explain the common perception that most active managers tend to underperform.
How many stocks make a diversified portfolio?
We have discussed various studies that examine the outperformance of concentrated managers; yet, some observers may continue to believe that a portfolio with a concentrated number of holdings is too risky.
However, it should be noted that most diversification benefits are realised after relatively few securities are added to a portfolio. As such, the practice of adding too many securities for diversification purposes leads to marginal risk-reduction results.
The below is the standard deviation pattern of a hypothetical equal weighted portfolio is displayed. As illustrated in the exhibit, the steepest drop in risk reduction occurs before the addition of the tenth security.
Risk reduction rate slows with more stocks
Source: Lazard Asset Management
The curve flattens thereafter showing a slower rate of reduction in total portfolio risk as additional stocks are added.
In Elton and Gruber’s book Modern Portfolio Theory and Investment Analysis, the authors concluded that the average standard deviation of a portfolio of one stock was 49.2 per cent, and that increasing the number of stocks in the portfolio to 1,000 could reduce its standard deviation to a limit of 19.2 per cent.
They also concluded that with a portfolio of 20 stocks the risk was reduced to approximately 20 per cent. Therefore, while the first 20 stocks reduced the portfolio’s risk by 29.2 percentage points, the additional stocks between 20 and 1,000 only reduced the portfolio’s risk by about 0.8 percentage points. We display an approximation of these results above as we feel that many investors do not realize how few securities are actually needed to realize the benefits of diversification (via standard deviation).
While we admit that concentrated portfolios will generally show more volatility than diversified portfolios, we believe that there are other ways to introduce diversification into a portfolio rather than simply holding more securities.
For example, in the Lazard US Equity Concentrated strategy, one of the main tenets of the portfolio construction process is blending diversified cash flow streams. By this, we mean that we seek to invest in approximately 20 companies that each have different business operations and objectives. We do not just combine the top ideas of our analysts, but the best combination of ideas from a risk perspective. We arrive at what we believe to be the optimal combination of stocks by understanding the revenue, earnings, cash flow, and balance sheet contribution of each individual company. We also consider how each company will interact with every other name in the strategy, as well as how it will affect the financial productivity (return on equity, free cash flow yield), valuation, and leverage of the strategy as a whole.
An offshoot of this process is that we typically do not invest in diversified businesses such as financials (particularly, investment, commercial, or regional banks) or diversified industrials, as the business models of the companies are diversified within themselves. We believe our process helps ensure that the strategy is not focused on any one theme or factor.
In our view, diversifying cash flows in a concentrated mandate is especially important as it can help mitigate much of the risk inherent in a concentrated portfolio. For example, we have found that not only has the Lazard US Equity Concentrated strategy outperformed most major developed-market indices over time, it has also provided lower levels of volatility.
By design, concentrated strategies facilitate investing in the highest conviction ideas and therefore limit overlap with an index – leading to high active share, which in turn, is linked to potential outperformance.
In our view, both theory and evidence support the notion that concentrated portfolios are well-positioned to generate alpha. We feel that by adding a concentrated strategy investors will capture the returns of the foremost investment ideas without diluting performance with over-diversification. In conclusion, we believe investors should focus on concentrated portfolios, where fundamental analysis shines.
Martin Flood is a director, portfolio manager and analyst at Lazard Asset Management. The views expressed above are his own and should not be taken as investment advice.