All the factors behind Butterwire’s computation of Fundamental (ie. Base) Scores are now available in the stock snapshot of any individual equity. We use our note on 3M’s profit warning last April to illustrate this new feature in the context of Butterwire’s overall research contribution.
Helping select “Interesting” stocks to research
Butterwire does not do buy/hold/sell recommendations. But if it did, it would not follow the industry standard of assigning buys 85% of the time and holds 15% of the time, and implicitly advocate a passive investment strategy. Rather, it seeks to systematically surface the +/-15% of stocks within any given universe (whether defined by region, industry, size, value, etc.) with characteristics that make them both more liable to very high returns and less liable to very low returns. Whatever the universe selected, the number of such “interesting” stocks must not be too low (so the engine can deliver a sufficient diversity in what “interesting” is, without undue factor risk concentration) and equally not too high (so the engine can usefully contribute some “digital alpha” to the active investor’s own alpha generation process). Also the list of interesting stocks must be stable enough to be compatible with long bottom-up research and investment horizons (hence for instance the use of long history of fundamental), yet responsive enough to significant developments (hence for instance to use of complementary signals such as stock flags and alerts).
An interesting stock for Butterwire is one that has distinctive fundamental characteristics and is controversial, the logic being that the more fundamentally attractive and controversial a stock, the higher its potential to deliver abnormally high returns in the future. Interesting stocks have a higher than average chance of out-performance 12-months forward (ca. 60% vs. 50%), a higher than average chance of exceptional returns (ca. 15% vs. 5%) and lower than average chance of blowing up (ca. 5% vs. 10%). Over time, active investors are therefore more likely to add value by applying time and skill to such stocks and determine which are worthy of inclusion in their portfolio.
Looking at 3M’s stock snapshot, the current base score of 5.1 (and corresponding alpha forecast of 0%) hints at fundamentals that remain undistinctive, as indeed reinforced by all its sub-scores for fitness, value, momentum and controversy.
With all the factors underpinning the base score now available, investors can see the full detail of how a given score was set by the engine. In the case of 3M’s average Business Fitness score of 5.8, five factors score well above average (notably EVA track record, whose calculation details are available by clicking on the the “?” icon), but two are a major drag on the overall score (track record of shareholder value creation and expected growth), a negative signal reinforced by the bear arguments on the stock (i.e. great historical returns at twice cost of capital, but struggling for growth).
Note that the relative importance of each factor is dependent on macro conditions (currently downcycle, ie. low iGDP), the company’s region of risk (North America), and its meta-sector (Production). Also, the score level of a stock for a given factor is dependent on all the other stocks in the region, so that two US stocks with an EVA track record of +5% will have the same EVA score but not necessarily the same as an EM stock with a 5% EVA. Likewise, the importance of EVA in the fitness score of a US Industrial or Financial stock will not be the same as that of a Healthcare or Technology stock.
Helping build “interesting” active portfolios
The next data-heavy step that Butterwire can contribute to is the construction of “interesting” portfolios, bringing together holdings that do not unduly or unknowingly concentrate factor risks (whether sector, region, size, macro, value, momentum, etc.), whose controversies are largely stock-specific (uncorrelated), and whose overall fundamentals are more attractive than the portfolio benchmark. The portfolio outcome sought is one where the (+1-year) contributions from most stocks end up cancelling each other out (i.e. no major detractors dragging performance down), and a few stocks blow the lights out and define the year’s performance (see butterwire2017 portfolio example below, published since Jan-17 with 0 turnover since: 5 stocks out of a total of 30 accounted for all the 19% excess return to date, all of it attributable to stock selection).
Back to 3M, the April article opposed its (and that of the 6 other stocks: Schindler, Swisscom, UPS, General Dynamics, Moody's, T. Rowe Price) undistinctive fundamentals and poor controversy, to 7 “interesting” stocks with similarly attractive EVA track records (Facebook, Dollar General, Estee Lauder, Shenzhou, Titan, NIKE, PayPal). Five months on, both “portfolios” have in fact delivered similar levels of excess returns (+7.5%, where each individual stock return is calculated relative to its regional index) and tracking error (12%). What does differentiate them is that the latter group did so with much lower correlations to underlying index (0.26 vs. 0.32) and with a more favourable dispersion of returns: the best performer of the more interesting group (Dollar General, 24% excess return, IR of 1.45) did better than the best of the less interesting one (UPS, 15% excess return, IR of 0.90), likewise for the worst performers (-5% for Paypal vs. -12% for 3M).
While obviously of no statistical relevance, this anecdotal example does illustrate two important characteristics of Butterwire:
Its ability to systematically and consistently attract the investor’s attention to interesting stocks like Dollar General vs. stocks like 3M as of 1Q19, which over time compounds into significant timesaving and performance uplift
Its inability to pick Dollar General over Paypal (or UPS over 3M), in other words, its intrinsic complementarity with bottom-up research and stock-picking skill