15. Q&A

Today I'll respond to a few questions that readers have asked me since I started the blog. Got more questions or feedback? Add them as comments below.

Q: How do you pick a mutual fund?

A: Let's start with how not to pick a mutual fund: looking at past performance. Morningstar and other publications have a rating system that takes into account past performance over short and long term, and gives funds high ratings if they beat their peers in the same investment class.

What they don't tell you is that the best funds of one year are often below average the following year. In fact, on average, 5-star funds will underperform in following years! This amazing result, confirmed again and again in research, is due to multiple reasons:

• Fund managers who bet on a sector and win (say, Internet stocks in 1999 or Real Estate in 2002), will stay with this sector when it loses steam and comes crashing back to Earth
• Since on average all the stock funds do the same (a bit less than the index), the return-to-the-mean phenomenon will mean that good years are often followed by less-than spectacular year
• Only a tiny fraction of lucky managers (0.1%) manage to beat the index consistently over long periods. Everyone else will have to balance their good years with bad years. 

So how do I choose my funds? I first zoom in on the sector I want to cover (say, US market or Muni bonds or International stocks). I look for funds with high diversification. I then narrow the selection to low-cost index funds, and look for low turnover. In some cases (such as Muni bonds) this is more difficult due to tax issues, but in general I end up with a few options that have more or less the same performance. Sites like Fidelity and Morningstar make this sorting easy. 

Q: When is a good time to buy a mutual fund?

A: The answer to this question is always the same - "Now". You can't predict the market. Maybe in hindsight, you'll enter right before a crash. Or maybe you'll enter right before a boom. But your best bet is not to enter at all - just be in the market all the time, keeping your preferred stock/bond ratio, and investing in small amounts any savings you have. I do not recommend sitting on a pile of cash and trying to time the market - this is way to risky.


Q: How to split the pie?

A: This depends on your level of risk tolerance, your investment horizon and how well you sleep at night. I'm a pretty aggressive/adventurous guy, and my split is 80/20 (stock/bonds). Others go with a more balanced 70/30 or even 60/40. I wouldn't recommend less than 60% in stocks for people under the age of 60. A good test of your risk tolerance is to ask yourself how you'd sleep at night if the stock portion of your portfolio drops by 25%. Note that with a 60/40 split, this means only 15% drop in your total portfolio, while with a 80/20 split this amounts to 20%. Try to run the numbers in your head - if you're going to lose sleep on a big drop, increase the bonds ratio. If you can see the longer time horizon and stick with the plan, go with a more aggressive approach. The one thing you don't want to do is to change the pie allocation in response to market events.


Q: What are your main DOs and DON'Ts?

A: There are so many of them... check my earlier posts for the essentials. You can start with Dilbert's advice.


Q: What is the mathematical background for these recommendations?

A: This is a little too much to cover here, I'll try to cover it in more details in a few weeks. But the basis is the following: while stocks have given, on average, the highest returns, they also carry with them the highest risks. It turns out that for any given return you wish to achieve, there's a certain mix of stock and bonds that can be predicted to achieve this return (based on historical averages) while minimizing the risk. For example, a 50/50 stock/bond portfolio with no rebalancing has gained 8.6% over the last 60 years, with a risk (standard deviation) of 12.2%. The same portfolio, with annual rebalancing, has gained 9% annually - or 25% more over the life of the investment. Moreover, the risk for the latter portfolio, as measured in standard deviation, is 25% lower (9.3%). In other words, more rewards, less risk. Achieving the same gain without rebalancing would have required you to increase the stock ratio and take even more risk.

The reason is that a balanced portfolio enables you (or rather, forces you) to sell stocks when they increase in value and buy more of them when they lose value - "sell high, buy low." Ferri's "All about asset allocation" gives many more details and graphs. In particular, the numbers I'm quoting here are from table 3-3.