Getting tired of financial engineering? How about finding some companies that are creating shareholder value by innovation, as reflected in R&D? This article presents a screen that selects companies that are trading at a low price-to-R&D ratio (market capital divided by R&D expense, abbreviated in this article as PRD).
(Click charts to expand)
A 10-Year Backtest
The backtest draws from stocks in the S&P 500 (SPY) and selects a portfolio of no more than 20 stocks, with each position limited to 5%. The portfolio is rebalanced every three months, allowing for 0.5% slippage on the bid/ask, and its performance is compared to the index.
Performance is very promising: the portfolio with quarterly rebalancing would have turned $100 into $353.20 between 3/31/2001 and 11/22/2011, for an annualized return of 12.6% per year.
- R&D expense between 3% and 15% of revenue
- and Interest Coverage > 5 or Dividend Yield between 1% and 7%
- Sorted by PRD, lower is better, and the best 20 stocks selected
R&D as a % of revenue simply limits possible selections to companies that have a meaningful expense for the item. Interest Coverage or Dividend Yield are alternate criteria to test for financial strength: not too restrictive; to be sure enough prospects pass the screen.
The 20 prospects with the lowest PRD are selected. The thinking is, the investor is buying R&D, and the less he pays for what he wants, the better. All R&D is not created equal. It's difficult to judge quality, and simply requiring that the company be financially strong may eliminate prospects where R&D is a waste of money.
Validating a Screen
The first concern is possible data mining. It's possible to select and tweak variables until the desired result is obtained, when the variables may have little relation to the desired outcome. Usually this manifests as an inability to replicate results when using a different set of data. This screen performs extremely well on the S&P 500, as shown, and very nearly as well on the Russell 1,000. Performance on the Russell 2,000 works out to 14.7% annualized for the same approximate 10-year period, and performance on the Russell 3,000 is similar. The criteria selected are relevant, and the screen works on multiple sets of data, to include large and small cap.
Too frequent rebalancing can increase reported performance, at the expense of ignoring the labor, slippage and commissions involved in frequent trading. Results are not as impressive if rebalancing is done once per year, but the screen still outperforms the index by a meaningful amount. A buy and homework investor selecting stocks along the lines suggested would monitor the quarterly performance as earnings came out and adjust the portfolio as needed.
Survivorship bias refers to the error of selecting from among current members of an index, so that purported past performance doesn't include companies that were eliminated during the period of time being tested - the likes of WaMu or Lehman Brothers. The screens were created at StockScreen123, and the indexes as made available for backtesting have been constituted to eliminate survivorship bias.
The screen doesn't produce any prospects from the Financial sector. In spite of their extensive innovation, they don't have R&D budgets, and don't make the cut.
Here are the 20 stocks this screen would select today, together with their valuations according to the primary metrics, as well as PRD:
Xerox (XRX) has recovered from a patch of very weak performance and with the acquisition of ACS (American Computer Services) is now a factor in data services as well as office imaging. The company spends 4% of its revenue on R&D. The stock is cheap on a cash flow basis, at a P/CF of 5.03. I wrote a bullish article back in May, and I still like the prospects, primarily because of the use of strong free cash flow to buy back shares.
Pfizer (PFE) spends 14% of its revenue on R&D. In common with many Big Pharma names, the company has a blockbuster coming off patent - Lipitor. The company cut its dividend in half in 2009, but has continued to increase it since, by over 10% in both 2010 and 2011. It now yields 4.2%, and the investor will get paid to wait until new drugs under development come to fruition. R&D expense was over $9 billion during the past year, and Jack Holland hears some gurgling in the pipeline.
Cisco (CSCO) CEO John Chambers has come in for heavy criticism as growth slowed and competitors impinged on the company's territory. A heavy round of cost-cutting and the realization that Cisco is still a heavyweight with a dominant position have initiated a recovery in the share price. I hold it in my portfolio, and completed an analysis to develop a price target of $24, although I haven't written an article on it yet. The cult of personality cuts both ways, and the extremely harsh attitudes toward the company and its CEO were most likely overdone.
Western Digital (WDC) has taken a hit from flooding in Thailand, where much of its hard disk production is located. The company had insurance, and should be able to sell its production at reasonable markups, as the market will be supply constrained. Competitor Seagate Technology (STX) suffered less damage from the flooding, and may gain market share. The company is well run and financially strong, but because of questions about the future of HDD (hard disk drives) vs. SSD (solid state drives) it may become a perpetual value candidate. I have reservations about continuing my small position here.
Lexmark (LXK), faced with competition and dwindling growth in its printer business, has been acquring software companies to extend its reach in the information management business. Competitors such as Hewlett Packard (HPQ) and Xerox have been doing the same. With a PRD of 6.84, the investor is buying a substantial amount of R&D at a discount. The question is, with 9% of revenue devoted to R&D, where's the beef? The shares are cheap based on P/E or Cash Flow, and the dividend yields 3% at today's prices. Management should either monetize the value purportedly created by R&D, or reduce the expense and deploy the funds for the benefit of shareholders.
Based on the results of this backtesting, low PRD is viable as a valuation metric to select stocks that will outperform over time. While financial engineering in the form of borrowing for buybacks or acquisitions may create value, it is far more appealing when a company creates something innovative, or extends existing technology in ways both they and their customers know are possible.
This way of looking at stocks also develops some questions investors should be asking. Is the R&D budget well spent? Does the company maintain its research during down cycles, so as to be able to perform strongly when conditions improve, or does it slash the expense as a short-sighted cost saving measure? Is reported R&D expense a residue of research in progress arising from acquisitions?
While R&D is frequently associated with tech stocks that don't pay dividends, the inclusion of Dividend Yield as a measure of financial strength brought in a number of attractive prospects, and improved the overall performance of the screen.