A look at the critiques of the asset-allocation theory and at some funds that don't follow it blindly.

Two bear markets in one decade have shaken investors' faith in the tenets of Modern Portfolio Theory and the asset-allocation strategies the theory spawned. Critics say that blind faith in MPT led to lopsided asset allocation. The theory did not properly account for systemic risk, which caused investors to allocate too heavily in stocks, and now their portfolios will need years to get even.

**A Quick Primer**

It's no exaggeration to say that MPT is the foundation of all modern finance. Before Harry Markowitz introduced MPT in the 1950s and William Sharpe took the theory to its logical conclusion by formulating the capital asset pricing model in the following decade, there really was no coherent account of how assets ought to be priced and in what proportion investors ought to hold them in their portfolios. MPT incorporates various measures of volatility and return such as standard deviation, R-squared, and beta.

MPT suggests how investors can strike a balance between risk and return. Markowitz showed that it makes sense for investors to hold securities or assets only in such proportion that the combined portfolio either achieves maximum return for a given level of risk, or minimizes the risk for a given level of return. The theory gave rise to rules that could recommend an asset-allocation formula for any investor. You'd just tell the model what asset classes you were considering, input expected return, variance, and correlation data for those assets, and out came a list of which assets you should choose and in what proportion. This portfolio would have the highest Sharpe ratio, which is another key MPT statistic that measures how much return a portfolio generates for every unit of risk (as defined by standard deviation).

These portfolio-construction ideas are generally applicable even if you don't use specific asset-allocation software. For example, by reducing your portfolio beta, you can make it less vulnerable to a slide in the overall market. The trick is to diversify into funds or asset classes that have low beta. If you're looking for signs of true active management, screen for funds with low R-squared. And you can look at a fund's Sharpe ratio to assess its risk/reward potential: the higher the ratio, the more return it has squeezed in the past from a given level of risk.

**The Problem**

Recent market shocks have shown that past experience is only a weak predictor of future results. So, how can you know what average return and variance to expect from an asset in the future and how the various assets under consideration would correlate? Typically, the models use historical data that serve as estimates of the expected future values. But correlations can be unstable because, at times of market stress, assets that were previously found to be uncorrelated can suddenly move in lock step. For example, domestic and foreign equities and commodities all tanked in the fourth quarter of last year.

One key challenge is that the theory uses standard deviation to measure variability or risk, but this is valid only if returns follow a normal distribution. The symmetric, bell-shaped normal curve allows for many convenient and elegant results but does a poor job of predicting extreme outcomes, and thus it isn't all that realistic. Standard MPT models often put very low, 1-in-200 year probabilities on market slides that actually seem to happen every few years. Thus, the theory underestimates risk, causing investors to load up on risky assets like equities.

**Useful but Not Guaranteed**

Some investors leaned too heavily on MPT models as though they were all they needed, but those models still serve a purpose. Diversification is still a great way to reduce risk and earn a higher level of return in the long run. Studies show that asset allocation is still of tremendous importance, even after last year's meltdown.

In fact, most constructive ideas about improving asset allocation retain the basic framework of MPT. These ideas suggest practical tweaks to the theory, rather than any radical remedies. For example, one idea is to improve risk measurement. This means less reliance on the normal distribution and more on other distributions or approaches that entertain the possibility of extreme losses. A combination of lower allocation to especially risky assets and hedging tools can then be used to protect the portfolio. Another suggestion is to engage more in tactical or dynamic asset allocation. Rather than stay with a static allocation to equities, say, this approach involves shifting the mix based on macro views or valuation analysis.