Best Investment Ideas
Randolph B. Cohen
Harvard Business School – Finance Unit
London School of Economics
Goldman Sachs Group, Inc.
March 15, 2010
When asked to talk about his portfolio, the typical investment manager will identify a position therein and proceed to describe the opportunity and the investment thesis with tremendous conviction and enthusiasm. Frequently the listener is overwhelmed by the persuasiveness of the passionate presentation. This leads to a natural follow-up question: how many investments make up the portfolio. Informed that the answer is, e.g., 150, the questioner will often wonder how anyone could possess such depth of knowledge and passion for so many disparate companies. Pressed to answer, investment managers have been known to sheepishly confess that their portfolio contains a few core high-conviction positions “the “best ideas” ?and then a large number of additional positions which may have less expected excess return but which serve to “round out” the portfolio.
This paper attempts to identify ex ante which of the investments in managers? portfolios were their best ideas and to evaluate the performance of those investments. We ?find that best ideas not only generate statistically and economically signifi?cant risk-adjusted returns over time but they also systematically outperform the rest of the positions in managers?portfolios. We ?find this result across all combinations of speci?fications: different benchmarks, different risk models, different de?finitions of best ideas. The level of outperformance varies depending on the specifi?cation, but for our primary tests falls in the range one to four percent per quarter.
These ?ndings have powerful implications for our understanding of stock market efficiency. Previous research has generally found that money managers do not outperform benchmarks net of fees. Mark Rubenstein referred to this fact as the e¢ cient-markets faction?s “nuclear bomb” against the “puny ri?es” of those who argue risk-adjusted returns are forecastable. Subsequent work has shown quite modest outperformance of around 1 percent per year for the stocks selected by managers (ignoring all fees and costs). We believe this paper is the ?rst to show evidence that the typical active manager can select stocks that deliver economically large risk-adjusted returns.
The paper also has strong implications for the optimal behavior of investors in managed funds. Our fi?ndings suggest that while the typical manager has a small number of good investment ideas that provide positive alpha in expectation, the remaining ideas in the typical managed portfolio add no alpha at all. Managers have understandable incentives to include these zero-alpha positions. Without them, the portfolio would contain only a few names, leading to increased volatility, price impact, illiquidity, and regulatory/litgation risk. Adding additional stocks to the portfolio can not only reduce volatility but also increase portfolio Sharpe ratio. Perhaps most important, adding names enables the manager to take in more assets, and thus draw greater management fees. But while the manager gains from diversifying the portfolio, it is likely that typical investors are made worse o¤. We suggest that investors who put only a modest fraction of their assets into each managed fund can have substantial gains if managers choose less-diversifi?ed portfolios.
The rest of the paper is structured as follows. In section 2 we brie?y discuss related literature. In section 3 we provide motivation and our methodology. In section 4 we
summarize the dataset. In section 5 we describe the results and their implications.
Section 6 concludes.
Best Investment Ideas 2. Related literature
There are several plausible reasons why examining total portfolio performance may be misleading concerning stock-picking skills. First, manager compensation is often tied to the size of the fund?s holdings. As a consequence, managers may have incentives to continue investing fund capital after their supply of alpha-generating ideas has run out. This tension has been the subject of recent analysis, highlighted by the work of Berk and Green (2004).