SEED favors a value and small company bias across a diverse basket of stocks.

We favor a value and small company bias in the stock portfolios we manage. Both factors have shown over time and across markets the ability to generate additional returns to investors; although, as we cover below, that added return may be associated with added risk. Behavioral economists argue instead that both strategies add to performance by avoiding over-priced stocks. Either way, the bias has historically and across markets provided excess (i.e., above market) returns. Value and small company biases also pass two other key hurdles: they are theoretically sound and sensible, and they are simple to apply and test for.

From an academic standpoint and ours, Occam’s razor applies: the simpler the better. For pricing models, the more complex the required empirical test, the higher the risk of data mining. In other words, by running the data and adjusting the model enough times, a researcher will find good results regardless of how spurious. But just because the price-to-book ratio (to pick on one measurement) adequately captures “value” does not imply that it’s the best in practice. Even in a market efficient world, value is nothing other than a proxy for some other undefined yet undiversifiable risk factor or factors. In a mispriced asset world favored by behavioral economists, price-to-book is an even more suspect measure of value. So yes, any particular value index fund may provide better historical risk adjusted returns than the next, but SEED won’t attempt to track either one at the expense of higher taxes or costs.

In fact, we initially weight stocks more equally in a portfolio rather than weight them by their market capitalization. We understand the resulting performance will not accurately track any of the multiple “market” portfolios available. In a tax-free world, we thus won’t be “mean-variance efficient”. What we recognize is that a large basket of stocks coupled with other diverse assets can provide more than adequate diversity, arguably even better diversity when tailored to avoid risks associated with the investor’s profession or legacy assets. Because the distribution of returns across many stocks will be widely disperse, equal weightings also expand our ability to harvest tax losses (see section on Tax Loss Harvesting).

The superior historical performance of the equal versus capitalization weighted S&P 500 should not be surprising to either behaviorists who believe in simple mean reversion of stock prices or efficient market disciples who believe in the random walk of stock prices. For the mean reversion crowd, investors simply are benefitting from vacillations in popularity. For those that believe that markets are efficient, equal weighted portfolios loosely track the benefits of having a value and small cap bias. Either way, the gains we achieve in optimizing tax efficiency and tailoring the portfolio to offset the full array of an investor’s risks (namely to their own income) more then off-set the costs of failing to track one of the myriad of available indices.

We don’t want to downplay the fact that the costs of diverging from a market index can be real. The expected return of equities is already uncertain and will be only more so with an ad hoc basket of stocks. We thus prefer exchange traded funds (ETFs) or index funds for investors who have the majority of their wealth in tax-advantaged accounts. We pay attention to industry weights and prefer to utilize the strategy for only those that can buy a large (at least 50, and preferably more) basket of stocks efficiently. Otherwise, over time, any one stock can dominate the portfolio. We implement out strategy using larger companies and augment the basket with ETFs designed to more readily capture the advantages of investing in smaller companies and stocks exhibiting momentum. Explore the links below for more details on these and other model factors.