Background
Asset allocation is the primary determinant of both risk and return in many portfolios. Numerous studies have concluded that the percentage distribution of financial assets (cash, stocks, bonds, international, real estate, venture capital, and other investments) has accounted for much of the variability of a portfolio's return, while market timing and security selection typically account for a smaller percentage, although there is much debate about these studies and their exact meaning (see Research below). Asset classes can be classified in broad terms like stocks, bonds, and cash or they can subdivided further into large-cap and small-cap, value and growth, international, and combinations of each. Bonds can be subdivided into short, intermediate, and long term, tax-free, high-yield, convertible, and international classifications.
Asset allocation should be consistent with an investors goals, constraints and time horizon. The goal of asset allocation is to achieve the highest return for the acceptable level of risk, or alternatively the lowest risk for a needed rate of return. By combining assets with different characteristics in a portfolio, an investor can achieve higher returns with lower risk over the long term. Adding high risk asset classes and investments to a portfolio may seem risky, but its likely the net effect will be to both increase returns and lower the risk of the portfolio. As an example, while international stocks may be riskier than US stocks to an American investor, adding international equities to a portfolio of US stocks actually lowers the risk of the portfolio because the assets have low correlation. That is, sometimes when US stocks go up, international stocks go down and vice versa. The net effect is a portfolio that has less risk because it fluctuates in value less.
A problem with mutual funds and managers that shift asset classes within their funds, is that they alter the asset allocation for any investor in the fund. This is a reason many choose to use index funds (index funds do not shift asset allocation). Specific target allocations (i.e., 60/40 stocks/bonds) can be viewed as a starting point. It can be useful to establish asset allocation ranges since over time, fluctuations in value and income received will alter asset allocations which creates a need to rebalance and/or review goals and constraints.
There are many "Asset Allocation" funds that intend to eliminate the need for investors to choose their own asset allocation. These funds typically attempt to provide investors with better returns through actively purchasing undervalued assets classes and under weighting overvalued asset classes. Most of these funds combine active market timing and asset allocation. David Dreman (False Prophets, Forbes 1/13/97) points out that according to Morningstar data if you expect to outperform the market using these funds you are likely to be disappointed. The S&P on average has significantly outperformed equity funds which have in turn outperformed asset allocation funds.
Trivia: What unique asset class outperformed the S&P 500, commercial real estate, international stocks, and bonds for both ten and twenty year periods ending 1995? Answer.
Research
Many researchers have studied funds to determine what accounts for variations in returns. A common conclusion of many of these studies is that a large percentage of the variability is attributable to asset allocation. The most frequently cited study was by Gary P. Brinson, Brian D. Singer, and Gilbert L. Beebower and was titled "Determinants of Portfolio Performance II: An Update." 1 They assessed the impact of passive (benchmark) and active asset allocations and security selection on 82 large pension plans over the 1977-87 period and found that on average, benchmark asset allocation (allocation policy) explained 91.5% of the variation in quarterly returns. In other words, the decision to invest in asset classes (stocks, bonds, etc.) was more important than the selection of individual securities in this sample. The issue was discussed in this interview with William F. Sharpe about Style Analysis.
The predecessor study was titled "Determinants of Portfolio Performance." 2 That study looked at the returns of 91 large pension funds from 1974 to 1983 and found that on average 93.6% of the total variation in actual plan results could be attributed to investment policy. Less than 5% of the returns were determined by security selection. However, the articles came under attack in the last few years from William Jahnke of Financial Design. The Financial Design web site's resources includes a wealth of information on the subject for those interested in learning more about the debate, including Jahnke's article that started the debate titled "The Asset Allocation Hoax" and numerous follow-ups by Jahnke.
Another article on the topic was published in the January/February 2000 issue of the Financial Analyst Journal and can be read on the Ibbotson web site. See
Does Asset Allocation Policy Explain 40%, 90%, or 100% of Performance? by Roger G. Ibbotson and Paul D. Kaplan. A follow up letter to the Editor of the Financial Analysts Journal appeared in the May/June issue by John Nuttall. For much more on the debate see The Importance of Asset Allocations. See also The Grinch Who Stole Asset Allocation (RR) from Dow Jones Investment Advisor (Dec 97).
See also
1. Gary P. Brinson, Brian D. Singer, and Gilbert L. Beebower, "Determinants of Portfolio Performance II: An Update," Financial Analysts Journal, May/June 1991.
2. Gary P. Brinson, L. Randolph Hood, and Gilbert L. Beebower, "Determinants of Portfolio Performance," Financial Analysts Journal, July/August 1986.
