This week we talked to Ascendere Associates founder J. Stephen Castellano about the financial models and consensus estimates used to help choose investments for his Systematic Long-Only model.
Covestor LIVE: Can you tell us a little bit about the financial models utilized to pick equities and ADRs for the Systematic Long Only model?
J. Stephen Castellano: The models we use with this particular strategy measure a stock's relative value within a particular sector as opposed to absolute value. In our opinion, a relative approach works very well because relative valuation and fundamental factors act as key metrics in investment decision making, which in itself leads to buy and sell activity.
The models were built assuming some basic premises of fundamental financial analysis, namely that return on capital and cash flow growth are the key drivers of an asset's valuation. We apply scores of proxies related to these factors and others, which may be summarized into four key categories: 1) Relative Value; 2) Operating Momentum; 3) Analyst Revision Momentum; and 4) Fundamental Value.
We rank these data points much in the same way that Joel Greenblatt describes in "The Little Book that Beats the Market." All the factors that we rank are equal weighted within their respective four key categories. It is a similar approach we took when we helped develop PaineWebber's Global Steel Finance product when we worked for an Institutional Investor-ranked sell side analyst in the late 1990s.
We update our models on a weekly basis to track and anticipate changes; however, we only rebalance our portfolios based on monthly ranking data. We have found that rebalancing the portfolios more frequently than a month actually detracts from performance. We believe this indicates the value of our models derive from their ability to anticipate changes in a stock's relative value as opposed to chasing the latest information -- our models cannot compete with quantitative asset management firms that can update relative valuations of stocks several times a day, but then again, they don't have to. In our opinion our models do a good job on average in finding stocks with good relative value and anticipating positive change over a few weeks or months.
The models represent the culmination of our 15 years experience in sell side and buy side equity research and related consulting projects (detailed bio). Our financial modeling approach is grounded by an intensive background working with raw data. We believe the ability to manipulate and fine-tune customized data points across thousands of stocks much in the same way a sell side analyst is able to work with a customized financial model for a single company is a major source of our advantage.
We have noted on numerous occasions that it is not uncommon for stocks that appear in our model portfolios or our weekly ranking updates to presage major ratings actions by sell side firms. See our "Nostradamus Report" for more details. Below is a non-comprehensive list of sell side actions on stocks that have been in our model portfolios since at least 10/31/2010:
11/10/2010: Williams-Sonoma (WSM) -- ThinkEquity upgraded the stock based on a recovery in home-related purchases and raised price target to $40 from $33. Oppenheimer identified WSM as a top pick in the hardlines retail sector.
11/8/2010: Albemarle Corp. (ALB) --Zacks upgrades to a Strong Buy.
11/5/2010: American Express Corp. (AXP) -- Argus upgraded the stock and raised price target to $50 due to reduced loan losses and volume growth. See also our 10/25/10 article on Seeking Alpha, "Why American Express is Finally Worth a Look."
11/5/2010: TRW Automotive Holdings Corp. (TRW) -- Deutsche Bank raised its target to $57 from $52.
11/3/2010: TRW Automotive Holdings Corp. (TRW) -- S&P Equity raised its price target raised to $62 from $50.
11/3/2010: Union Pacific Corp (UNP) -- Deutsche Bank raised price target raised to $102 from $92.
11/2/2010: Fossil Inc. (FOSL) -- BB&T Capital Markets initiated with a Buy rating.
CL: Your financial models use consensus estimates, how often do you find these estimates to be on target?
JSC: Our models for the Systematic Long-Only strategy measure analyst revision momentum -- the rate of change and the relative position of analyst revisions -- as opposed to their accuracy. We are not so much concerned by the accuracy of the estimates as we are with the general direction and relative level of the changes. Deservedly or not, stock prices in the short term do tend to fluctuate in relation to analyst revisions as opposed to any intrinsic value, so this figures into our quantitative rankings.
CL: Have there been any stocks you purchased based in part on estimates that you wish you hadn’t? Do you ever find missed estimates to work in your favor?
JSC: This past April we held Gilead (GILD) in a theoretical portfolio because it was highly ranked in every single key category we previously mentioned, including analyst revision momentum. But following the release of negative guidance with its March 2010 quarterly financial results, the stock sold off nearly 10% in a single day. This was an absurd reaction in our opinion because the reduced earnings estimates were not the result of any change in its core business (see our April 21, 2010 GILD write-up on Scribd). We thought the market reaction was wrong, but at the same time we acknowledged that GILD would be rebalanced out of our model portfolio at the end of the month due to the negative analyst revisions. As we expected, at the end of the month GILD's updated rankings were very poor for analyst revision momentum and it was subsequently removed from the portfolio.
By systematically applying our ranking criteria, we saved ourselves another 19% decline through mid July. The stock has since comeback to close to where it was last April in large part because analyst revision momentum has improved once again. Whether or not a stock’s long-term intrinsic valuation is damaged, this example clearly shows how short-term analyst revisions can play a big impact in the short term price swings of a stock, independent of the stock’s underlying fundamentals.
CL: You recently bought shares of Domtar Corp (NYSE:UFS). Can you tell us what factored into your decision to do so?
JSC: Given the recent sharp run up in Domtar Corp (UFS), up roughly 30% over the last three months (CL Note: On August 10th, UFS had a closing price of $63.08. On November 10th, UFS had a closing price of $82.00), one might assume it might be more prudent to purchase an alternative such as International Paper (IP), which has "only" run up about 14% during the same period (CL Note: On August 10th, IP had a closing price of $22.91. On November 10th, IP had a closing price of $25.96). IP might be a good choice, but we still think UFS is better. This is because among Material sector stocks in our universe of coverage, UFS ranks higher for relative value and operating momentum, and almost as well for analyst revision momentum and fundamental quality. IP looks good on relative value, but ranks less strongly on other metrics. Overall, despite the sharp run up in UFS, we think it still is the relatively more attractive stock idea. In summary, we do not factor in recent price moves in our purchase decisions; we instead are focused on relative value and relative positioning of other factors compared to other stocks in the same sector.