Matching Goals with Strategies
Portfolio goals can come in different forms. We could have an absolute return goal of, say, 10% per year. Alternatively, we could have a relative return goal of, say, outperforming the S&P 500 by 2% per year. These two are certainly different, as we could, for example, outperform the S&P 500 by 2% with only an 8% return when the S&P 500 returned just 6%.
While distinguishing between these types of goals is not always top of mind, when we think about how to achieve them, I think that distinguishing between the two types is quite important. To see this, we will consider, below, the simplest approaches for achieving each type of goal. As we will see, the best approaches for the two types are quite different. Both strategies we will discuss below are familiar, but they pair up best with different goals.
As a result, if we want to accurately assess the performance of a portfolio manager, it is critically important to make sure that we assess them using the type of goal they can reasonably promise to achieve, not the type that they cannot. Matching up clients with one type of goal to managers using a strategy best suited to a different goal will lead to distress for both parties. Similarly, if we are managing our own portfolio, using a strategy that does not pair up best with our return goal is likely to lead to disappointment.
The Simplest Absolute Return Strategy
Approach: calculate a reverse DCF for some assets. If the implied rate of return at the current price of the asset is above our goal (e.g., 10%), with a margin of safety, then buy some.
If we are able to find 10–20 assets to buy, then our diversified portfolio should achieve the goal. Our margin of safety gives us room for a some of these assets to underperform due to unforeseen events (though likely not a recession that would impact all of them simultaneously).
It is important to note that we do not have to value every available asset. We can put many if not most into the “too-hard pile”. As long as we are correct in our assessment of the assets we bought, we can feel confident about achieving the required absolute return.
On the other hand, this strategy makes few promises about relative returns. In order to estimate those, we would need know the market return, which depends on the returns of all the assets, even the ones in our “too hard” pile. If we, for example, ignored the entire tech sector and that sector ended up including some of best businesses in history, which ended up becoming a large part of the index, then we would expect to underperform.
Over a long enough period of time, we might be willing to make an assumption about what market returns will be, but even over a decade, there are considerable variations in what these can be. Hence, while the strategy described above is straightforward to understand in terms of absolute returns, there is substantial uncertainty about its relative returns.
Even if we are willing to commit to the strategy for a long period of time and are willing to make strong assumptions about market returns over that period, I would argue that this is not the simplest way to achieve a outperformance.
The Simplest Relative Return Strategy
Approach: identify a factor that is under-appreciated / under-priced by the market. Construct a portfolio that tilts toward assets that with more of that factor and away from those assets with less.
This portfolio may own all available assets, but it will have a higher weighting on those assets we believe are under-priced and a lower weighting on those we believe to be over-priced. As long as those in the former group outperform those in the latter group, we will outperform the market, regardless of the overall market return.
If we have an explicit goal for outperformance, then we will need to do more. In particular, we need an expectation (likely based on past returns) for how much the over-weight stocks will outperform the under-weight ones. If that number is less our relative return goal, then we will need to identify additional under-appreciated factors. We can add those into the portfolio as well simply by adjusting the weights. Once we have identified a collection of largely-independent factors whose total expected outperformance exceeds our goal, with some margin of safety, we can feel confident about achieving the desired relative return.
On the other hand, this strategy can make few promises about the absolute level of returns. Those depend on the overall market return, which is highly variable year-to-year.
As above, if we are willing to commit to the strategy for a long period of time and are willing to make assumptions about the market return over that period, then we can estimate the absolute return as well. However, while the connection to relative performance is direct, there is substantial uncertainty about the absolute return of this strategy. Certainly, it is safe to say that the first strategy is a simpler way to achieve an absolute level of return.
Lessons for Discretionary Value Investors
Either strategy above can be used by “value investors”. The first strategy seems to be the standard approach for discretionary value investors, but the latter will also work if you tilt toward factors that identify value stocks. This has been harder said than done in recent years, as the flaws in metrics that have traditionally identified under-valued stocks (e.g., P/B) have become more prominent, but more refined factors can still do the job.
In principle, both strategies can achieve both types of goals if executed properly for a long period of time. However, over any small period of time (say, less than a decade), the two strategies make different promises to us about their returns.
As we discussed, relative returns depend not only on the returns of the assets we buy but also on the returns of the assets we do not buy. Since performing an accurate reverse DCF for every available asset would be incredibly difficult — even Warren Buffett has a “too hard” pile — this strategy is not the best choice for achieving a relative return goal.
For that, it is much simpler, I think, to focus instead on finding under-appreciated factors. That can be uncomfortable for discretionary investors, though, as they could be required to buy assets without knowing what they are worth. Indeed, the first factor tilt to include should probably not be value but rather momentum, as it has a better recent track record, and that would mean, in many cases, buying companies that look too expensive, something a discretionary value investor would rather short than buy.
I think the best answer for discretionary value investors may not be to change strategies but instead to change goals: they should feel more confident about their ability to provide an absolute return than a relative one. Barring a recession that depresses the market for years, their strategy should be able to predict returns over the period of a few years with decent accuracy. Indeed, achieving an absolute return is the goal for which their strategy for is ideally suited.
