Ten Stocks Are Enough For Diversification In Asia


Ten Stocks are Enough in Asia

Andrew Stotz

University of Science and Technology of China (USTC) – School of Management

Wei Lu

Where Are All The Big Tech Stocks On The 13Fs For Q3?

technological 1666892351Most of the time, it's easy to spot trends within hedge funds' quarterly 13F filings, but things were different in the third quarter. One of the easiest trends to spot has been tech stocks, but this time around, there was relatively little movement in tech names. Rising And Falling Stars In Tech Of course, that Read More

University of Science and Technology of China (USTC) – School of Management

July 17, 2014


This paper looks across 13,000 stocks in Asia excluding Japan over 10 years to determine the optimum number an active manager should hold to reduce unsystematic risk. We randomly select stocks for inclusion in equally weighted portfolios that are held for one year and then are reselected based on the new year’s universe. We find that 10 stocks removed 64% of unsystematic risk and, after this, the marginal impact reduces significantly. An additional 10 stocks will only take this number to 74%, but will drive the active fund’s performance closer to that of a passive fund.

Ten Stocks are Enough in Asia – Introduction

This study attempts to identify the optimum number of stocks that an active fund manager should hold in a portfolio of stocks in Asia (we only consider Asia excluding Japan). The unique aspects of this study are its inclusion of all Asian markets, annual reselection, annual rebalancing, and its dynamic data set.

All markets in Asia – This research builds portfolios of companies across Asia from China H shares to shares in India.

Annual reselection – Most prior research on diversification randomly selects portfolios of two, three, four and more stocks, holds those stocks for the period (e.g., five, 10 or more years) of the study and then measures the volatility of those portfolios. The weakness of such studies is that now very few fund managers hold a portfolio for such a length of time, in fact, recent research shows that turnover of the average portfolio in the US is now reaching 100% per year. To make the results of this research more realistic, we reselect stocks annually; meaning, a portfolio of 10 stocks, reselected annually for 10 years would mean the investor owns 100 stocks over the period.

Annual rebalancing – When we reselect stocks each year we apply equal weighting to the new stocks in the portfolio.

Dynamic universe – Our list of stocks available for reselection at each year’s end were actually investable at that time. The two main elements that make a stock investable are that it is large and liquid enough to which to allocate money. Each year, we remove all those which do not fit this criteria, which makes the results more robust and applicable to real world investing. In addition, this methodology allows us to include new listings which, given the boom times in Asian markets over the past 10 years, would be a serious omission.

We start with the assumption that the active fund manager is guided by four competing objectives: to reduce uncompensated risk, to reduce complexity, to reduce costs, and to maximize return. Of these four objectives, only the first is helped by increasing the number of stocks in a portfolio; all others are better achieved by having less stocks in a portfolio. Fewer stocks means less complexity, less work and, a chance at higher than market return. The main costs come from finding that new stock, keeping track of changes happening in that company, and making decisions about when to buy more or when to sell.

Besides overcoming the performance drag of the above costs, an active fund manager is also expected to “beat the market”, but the more stocks he holds, the more likely it is that his performance will mimic that of a passively managed fund.

So, the object is not only to minimize diversifiable risk, but rather to find the optimum number of stocks after which adding the next stocks fails to bring significant benefit to the investor’s risk and return.

This paper starts with a literature review, follows with a description of the data-set construction, and then introduces our methodology. This is followed by our results and analysis and ends with our conclusion.

See full article by SSRN

No posts to display