As we’re a little hurried today in preparation for travel to a 20th high school reunion, we thought we’d present a simple example to show how use of fundamental metrics, even using the simplest strategy possible, can provide outsized returns over time. There are a few catches, however, as we’ll see.
For this experiment, we used Bloomberg’s factor back-testing product to evaluate whether or not using a simple and widely used fundamental metric, Enterprise Value to Trailing 12-month EBITDA, provides excess return over time. We used the S&P 500 as our universe and grouped the results into deciles based upon EV/EBITDA, with lower valuations being better. Portfolios were rebalanced annually (year-end) in this exercise. Positions are equally allocated. For individual stocks without valid data, we assigned the stock the universe median multiple. Bloomberg in this instance will only allow us to go back to year-end 1993. Thus we considered annual results between 12/31/1993 and 12/31/2012. Ultimately, our interest lies in determining whether buying the 50 or so stocks in the first decile each year provides any benefit in terms of excess return. To make sure performance calculations remained “apples to apples,” we used Bloomberg’s universe performance calculations instead of official S&P 500 performance to benchmark the model against; in some cases, especially the results from the 1990s, less than 500 stocks were included in the database. In any case, the universe annualized returns are higher than the S&P 500 over that time.
Here are the results from the exercise. Underperformance results are highlighted.
 |
| IronHorse Capital |
Overall, simply buying top decile EV/EBITDA stocks each year through an annual rebalance produced annualized returns of approximately 2.5% per annum higher than the universe return. Behold the power of compounding as the annualized returns translate into a 927% total return for the model over those 19 years versus 576% for the total universe. Seems easy right? It’s always easy when looking a historical numbers on a page.
Here are the gotchas that trip up many attempts to implement and follow a disciplined fundamentals based strategy, whatever your favored metric.
First, and most importantly, one must overcome the tugs and pulls of human psychology. Note above that the model underperformed in 7 of 19 years and that there were a few years with dramatic underperformance. In 1998, the model underperformed by approximately 16.5%. In 1996, it was over 9 percent. Furthermore, pretend that one had actually implemented this strategy starting 12/31/1993. It would have been incredibly hard to justify sticking with it at first. The strategy underperformed the universe four out of the first six years in this example, with two of those outperformance years being among the worst. Our irrational brains would have been telling many of us to throw this strategy in the trash at various points. From another angle, many would have shied away from the strategy simply because the names picked weren’t the belles of the ball. Looking back over time, many of the low valuation stocks chosen each year weren’t necessarily the most attractive names from a press and media standpoint. Companies reach low valuations for a reason. As we’ve discussed in the past, it’s very difficult to ignore the noise surrounding these names.
Second, the strategy was more volatile than the total universe returns, not to mention there was no rhyme or reason in terms of downside protection. Standard deviation for the strategy was 23.07% vs. 19.54%. The strategy underperformed during the go-go growth-centric period from the mid to late 1990s, performed well during the dot-com crash years, hung in there pre-financial crisis, then provided little shelter during the 2008 crisis, underperforming by seven percentage points. Of course, the model made up for it in a big way in 2009. Still with a few large drawdowns and more volatile returns in general, it may have been difficult to some to stomach the ride. Nonetheless, using a simple calculation of dividing returns by standard deviation, the return per unit of risk remained slightly higher than the universe, i.e. risk adjusted returns were still slightly better even with the higher volatility.
In the end, practitioners are constantly asking why more don’t take advantage of these supposed fundamental free lunches the market provides? Simply, it’s very difficult for individuals and professionals alike to execute a strategy consistently through thick and thin. For many institutional investors, so-called career risk occupies an outsized portion of one’s thoughts. At the first sign of trouble, many abandon the plan and chase the Johnny-come lately strategy. We know from experience that there are periods where value dominates and periods where growth dominates. Or large-cap trounced small-cap and vice versa. Often, cycles go longer than even the most patient expect. It’s hard to hold the line, but those that can exhibit discipline are often rewarded over time.
