It's A Question Of Quadrix

1/2/2017


We've been using our Quadrix stock-rating system to pick stocks since 2000. Over the years we've figured out a lot about how and why the quantitative system works. But even longtime subscribers still ask us questions about Quadrix, and sometimes we ask questions ourselves as part of the never-ending quest to best make use of the system.

Here are a few of those questions.

Q How has the Overall score performed over the last few years?

A Quadrix hasn't worked as well in recent years as it has over the long haul. However, we designed it to tap into many different operating metrics precisely to keep the score relevant even when some of the individual factors aren't working.

Since 1992, Dow Jones U.S. Index stocks in the top quintile (top one-fifth) based on Quadrix Overall scores averaged 12-month returns 2.4% higher than those of the average stock in the index. We calculated this average using rolling 12-month periods (January through December, February through January, etc.) starting in January 1992, a total of 287 periods.

Of course, Quadrix performed better during some of those periods than others, as shown below.

Stocks with high Overall scores have averaged returns below the typical stock in the index in the last 12 rolling periods. However, the Overall score has outperformed the market in 61% of the last 287 rolling 12-month periods, and on average top scorers have outperformed the typical index stock over the last three, five, and 10 years of rolling periods. We expect Quadrix to bounce back strong after its period of underperformance, just as it has after rough patches in the past.

Q Which individual Quadrix statistics work best?

A We ask ourselves this question often because the answer changes frequently. Quadrix relies on roughly 100 statistics to generate the Overall score, and the effectiveness of those individual metrics varies.

Over long periods, valuation ratios have showed the most predictive power. In fact, of the 10 individual Quadrix factors that have worked best since 1992, seven are part of the Value score. Returns over the last five years tell a different story, with the top 10 factors including three valuation ratios, our earnings-predictability score, and six measurements of estimate revisions, four from the Earnings Estimates score and two from the Momentum score.

The forces driving stock returns over the last five years are not the same as those that have proven most effective over the last 20-plus years. However, there was some overlap. Four individual factors were among the 10 most effective during both periods — valuation ratios price/free cash flow and price/sales, and estimate revisions for the current quarter and the next quarter.

Given the consistent effectiveness of both valuation ratios and profit-estimate trends, investors should always look at a stock from those angles before making a buy. We certainly do.

Q What works better at finding stocks that outperform, absolute valuation ratios like price/sales or relative valuation ratios like price/sales compared to the five-year average ratio?

A In terms of raw outperformance, absolute ratios are the clear winners, with the highest-scoring (cheapest) stocks delivering the strongest outperformance over both long and short periods. We saw some recurring themes, with four valuation ratios (price/free cash flow, price/sales, price/earnings, and enterprise value/EBITDA) among the seven most effective individual Value factors since 1992, and over the last five years.

Of course, we also measure effectiveness using winning percentage, or the percentage of periods in which top scorers outstripped the average stock. By this measuring stick, absolute valuation ratios look even better. For example, stocks with the top price/sales ratios exceeded the typical stock's return in 62% of periods since 1992. In contrast, price/sales to five-year average outperformed 46% of the time, and price/sales to three-year average 36%. We saw similar trends with absolute versus historical ratios for price/earnings and price/book.

These findings inspire another question. If absolute valuation metrics work so much better than relative metrics, why use the historical comparisons at all? The answer lies in part with the nature of the relative valuation ratios themselves, and in part with the design of Quadrix.

While relative valuation ratios don't work as often as absolute metrics do, over long periods of time top scorers have averaged index-beating returns. This suggests that when relative metrics do work, they tend to work very well. If we ignored these ratios, we'd miss out on the high returns they sometimes generate. In addition, part of the power of Quadrix lies in its ability to capture so many aspects of a company's fundamentals. Some individual factors don't work well on their own, but they bring their own spice. The stew wouldn't be as savory without them.

Q Which Quadrix scores work best together?

A We start our analysis with the Overall score, which is based on a weighted average of the six category scores — Momentum, Value, Quality, Financial Strength, Earnings Estimates, and Performance. The diversity of this approach provides Quadrix with much of its predictive mojo. However, investors relate more to some categories than others.

Value investors prefer to own stocks with low valuation ratios, while other investors focus on operating growth or gravitate toward stocks with strong share-price action. The Overall score captures data points from these and many other investment approaches (see www.DowTheory.com/Go/Quadrix for more on how Quadrix works). However, investors who prefer to lean their selections in one direction or another can also make use of our system.

In some cases a combination of high scores in Overall and one of the six categories has created portfolios with greater outperformance than stocks with just high Overall scores. The Overall/Value and Overall/ Performance combos stand out, as stocks that score above 80 in both scores have historically averaged 12-month returns at least 3.9% higher than the index average.

Keep in mind that while companies earning high Overall scores tend to earn solid scores in several of the categories, requiring at least 80 in multiple scores limits the number of choices available. Rather than fixating on a particular category score, we advise readers to focus on stocks with high Overall scores and let the category chips fall where they may. If you have a particular taste for Value stocks, rather than limiting yourself to the small group of stocks with very high scores for both Overall and Value, seek out stocks with high Overall ranks and rule out those with low Value scores — say, below 50.

Q In which sectors does Quadrix work best?

A The answer to this question depends on the time period.

Since 1992, top Overall scorers in eight of the 10 market sectors outperformed the average stock in the sector, with five sectors (energy, materials, consumer discretionary, financials, and technology) averaging at least 1.1% outperformance. However, over the last five years only seven sectors outperformed, with the Overall score providing by far the biggest boost in consumer staples and health care, sectors with defensive characteristics and where the Overall score hasn't been especially effective over the long haul.

Unfortunately, predicting when Quadrix will become more effective in a particular sector is difficult. We continue to advise readers to maintain a diversified portfolio, confident that Quadrix will work in most sectors, most of the time.


Current Hotline

Stock Spotlight

Individual Stock Reports

ISRs make stock research easy!

Perhaps the most valuable two page reports available anywhere.

All the data you would normally have to plow through years of 10-K filings, earnings reports, and reams of market data to assemble — yours all in one concise report.

ISRs contain our proprietary Quadrix scores — find out how we rate all the stocks in the S&P 500.

Visit us at individualstockreports.com