Joel Dickson Interview

November 17, 2003

Vanguard tax guru and fund co-manager discusses Vanguard's quantitative approach to managing its active stock funds.

This article originally appeared November 17, 2003

Joel M. Dickson, a Vanguard principal, teams up with Gus Sauter as co-portfolio manager of the actively managed mutual fund assets that the Quantitative Equity Group oversees. Those funds include Vanguard Strategic Equity Fund and portions of the Morgan Growth, Explorer, Windsor II, and Equity Income funds. Mr. Dickson earned an A.B. degree at Washington University in St. Louis and a Ph.D. at Stanford University, and has worked in investment management for Vanguard since 1996.

Q: How do you pick stocks?

A: We're quantitative in both our stock selection and risk-taking processes.  We use models to select stocks, but we maintain a risk posture similar to the benchmark. 

Tracking error is always a consideration, but I wouldn't refer to what we do as 'enhanced indexing.'  We definitely have an active strategy.  The question for us is: How much do you want to loosen up your risk profile [relative to the benchmark] to let in an active strategy?

Q: How do active fund managers fare against relevant stock indexes?

A: Two-thirds of active mangers are going to underperform the benchmark over time - that's just simple math.  There are certain natural laws that you can't do anything about, and math is one of those.

Is there an ability to outperform consistently over time?  That's the age-old question. 

When you ask active managers why they think they can outperform consistently, you get a lot of different answers, but it's essentially the same answer every time: We're smarter than everyone else.  We have the best research and we know more about the companies than anyone else. 

To us, that's not a good justification.  The market in aggregate ends up being smarter than everyone else. 

Instead, we tend to rely on a behavioral finance justification for why an active manager might outperform.  There are consistent predictable biases among investors and analysts.  Can those biases be exploited?  Our models look for those biases and determine if they can or cannot be exploited.

Our models are, in essence, a behavioral finance justification for selecting stocks. 

The big risk here is false positives.  Or you may form your own opinions about which strategy may work, when in fact the overall goal is to eliminate such biases.

So we think this is a better justification for why active management might work.  Does it guarantee we will outperform?  No, of course not.  Does it tilt the odds in our favor?  We hope so. 

We certainly have to keep costs low or we're not going to outperform no matter what processes we use.  We have to consider not only expenses, but also transaction costs, and tax costs. 

Q: Why is portfolio diversification so important?

A: Portfolio diversification is key because you want to expose yourself ONLY to the risks your models say you should.  You don't want stock specific risk or other unintended risk creeping into the portfolio.  Most active managers fail because their portfolios are too concentrated, in our opinion.

For example, there are more mid-cap stocks out there than large-cap stocks.  So if you made equal bets on all the stocks you view as attractive, you'd end up with a small-cap bias.  However, we want to reflect the risk of the index on a CAP-WEIGHTED basis.  This helps diversify away stock-specific risk.  We don't want to end up with an unintended small-cap bias.

Q: Does studying indexing (or the index) make you a better active manager?

A: I think you can be a better manager by understanding the principles that underlie the benefits of indexing.  Such as the value of broad representation or diversification.  Also, a focus on minimizing costs as the single most important factor, whether you're using an active OR passive strategy. 

Q: Are there certain markets where active mangers have the upper hand?

A: People say active managers are more successful in so-called less efficient markets like small-cap and international.  However, there are also more costs in those areas.

With our own models we have had more success in small-cap stocks than in large-cap stocks.  But to me this isn't really a market efficiency issue.  We attempt to find the best stocks in any particular market segment.  You simply have more mid- and small-cap stocks, so you're able to make more meaningful comparisons among companies.  In large caps, you just can't make as many peer comparisons. 

Obviously, if your model if successful at differentiating among stocks, it will be more successful when your sample size is larger.  For example, stock specific risk [which will dominate all other risks] is different when your sample size is 3, instead of say 40 stocks.  In other words, stock-specific risk will screw up your models every time.


Find your next ETF

Reset All