For the record, I love Christine Benz at Morningstar. She writes great articles that are accessible and informative, and often with a non-traditional take on things. That’s why I’m so disappointed with her recent article on asset allocation for retirement. In fact, articles like this make me crazy.
Selecting a stock and bond mix is just a way to try to target volatility. (Even asset class volatilities can vary over time, so it’s not a perfect solution. But volatilities tend to be reasonably stable, so I can buy into the idea of volatility buckets.) But traditional asset allocationists often make much broader claims. Here are all of the things that typical strategic asset allocation cannot do:
1) It can’t help you predict what your returns will be. It can tell you what your returns would have been in the past, but that has no effect on what returns will look like in the future. Most asset allocations simply assume that equity returns will always be positive and somewhere around the historical norm. That’s a crucial problem because most asset allocations count on the equity returns to drive overall growth.
2) It can’t help you predict your risk level. Volatility might be somewhat predictable, but risk is a different animal. You can’t eat low volatility if it turns out you did not invest in assets with good returns.
3) It cannot make the investing process predictable. Everyone wants certainty. As long as historical norms continue, it seems like the process is fairly predictable. If there is a paradigm shift, you will quickly realize it was not. Markets are not predictable. The primary function of strategic asset allocation seems to be to generate really nice-looking pie charts.
Yet many, many articles contain these sorts of homilies about asset allocation:
Most experts agree that your retirement portfolio’s asset allocation–its mixture of stocks, bonds, and cash–will have the biggest impact on how much it grows, as well as its risk level. Unfortunately, retirement savers seeking guidance on what an appropriate asset allocation may have a hard time knowing where to look. Sure, you could certainly do worse than adopting Jack Bogle’s simple formula: 100 minus your age equals how much you should hold in stocks. But what if you want to come at the problem with a greater sense of precision?
Morningstar’s Asset Allocator tool provides another, goal-based view of asset allocation, harnessing your own portfolio information if you’ve saved one on Morningstar.com. The tool calculates how likely you are to meet financial goals based on your current portfolio value, monthly investments, time horizon, and asset mix, and is useful for fine-tuning your allocation.
I’m not trying to pick on Morningstar. Their strategic asset allocation tool, I’m sure, is as good as anyone else’s. The point is that all of the tools are flawed because their fundamental premise is flawed: they rely completely on historic return streams being repeated in one fashion or another. If you were a Japanese investor in 1990 working with 40 years of data (1950 – 1990) to construct a strategic asset allocation for your retirement, it would no doubt include a large equity component because historic equity returns had been both large and positive for a long period of time. Asset allocation steered you directly into a disaster. Since we know this has already happened in other markets, why do we continue to court disaster here? Do you really believe that other markets can go down, but not the US? Why would you continue to use a tool that you know will eventually fail?
After all, it’s not as if alternatives do not exist. There are various kinds of dynamic asset allocation to choose from. Tactical asset allocation using relative strength is just one form; other analysts try to forecast returns based on asset valuation or to rebalance across asset classes when sentiment gets too one-sided. How well something works is often a function of how well it is implemented, but no one can make a failed process work regardless of how cleverly it is implemented. I urge you to rethink your asset allocation methodology before it bites you in the ass. If your asset allocation is not responsive to actual price changes, it is pretty much useless.
A Regretful Strategic Asset Allocator
[...] What can be done to minimize the problems associated with being too bullish at the top and too bearish at the lows? For starters, it makes zero sense to rely on an investor’s expressed risk tolerance when constructing an asset allocation–as pointed out above by Dan Ariely. Furthermore, reliance on an investor’s age also presents another set of problems (bonds can also have horrific drawdowns.) Also popular is the practice of reliance on long-term historical inputs, such as variance, expected return, and correlation, in order to construct the “optimal” portfolio. We’ve frequently dealt with the fallacies of modern portfolio theory. [...]