What Matters More When Investing: Good Company or Good Price?

{ Euclidean Q2 2013 Letter }


Over the past five years, we have greatly expanded our technology platform and simulation capabilities.  Now, among other things, we can search for the best approaches for achieving different investment objectives, model long and short strategies, and support a wide variety of model types – ranging from non-linear models to simple ranking mechanisms.  Today, we use these tools to augment our understanding of what works in investing.  Doing so has deepened our conviction in the soundness and long-term prospects of our investment process.  Over time, through strong performance in Euclidean Fund I, we aspire to earn the right to offer you additional options for applying history’s lessons to managing your hard-earned assets. 

In the meantime, we recently had the opportunity to apply our tools to help one of our investors.  The project yielded some great insight into Joel Greenblatt’s “Magic Formula” – a simple process for buying good companies at good prices – and into how strategies for buying companies at pessimistic prices perform when price-to-earnings multiples contract.

During the past three years, as the major indexes have appreciated, public company valuations have stretched to progressively higher multiples of earnings.  These valuations have become increasingly rich [1] at the same time the average SP500 constituent company is generating unusually high earnings on each sales dollar.  Some argue that because of the relative attractiveness of the US economy and the unattractiveness of bonds, the stock market has much more room to run.  So it may.  However, the market now has two obvious ways to lose:  valuations as multiples of earnings may fall to historical norms; and companies’ earnings may compress as operating margins revert toward historical averages.  

For this reason, we feel it is worthwhile to examine how investment approaches resembling ours might have performed when price-to-earnings multiples compressed in the past.  A recent study we completed for one of our investors yields some perspective on this question. 

The “Magic Formula”

We have an investor who manages a separate set of assets with an approach that closely resembles Joel Greenblatt’s “Magic Formula.”  Joel Greenblatt is a prominent value investor who wrote the easy-to-read book, The Little Book That Beats The Market.  In his book, Greenblatt presents a “Magic Formula” for buying good companies at good prices.  He also shares evidence that one would have achieved a considerably better than market return by adhering to the formula from 1988 to 2004. 

Greenblatt’s formula is intended to buy good companies at good prices.  These concepts – good company and good price – are represented by two ratios from companies’ financial statements: Return on Invested Capital (ROIC) represents “good company” and Earnings Yield represents “good price" [2].  Although the definitions of both ratios used in our simulations are provided in the footnote, here is a quick way to think about them: 

  • Return on Invested Capital (ROIC) tells you how much cash a company generates in relation to the amount of capital tied up in its business.  As ROIC numbers increase, all else being equal, a business gets better and better.  The reason is that when you own a business, the higher your ROIC, the more money you are able to pocket every year in relation to the money you have invested in the business.
  • Earnings Yield tells you how expensive a company is in relation to the earnings the company generates.  When looking at Earnings Yield, we make certain adjustments to a company’s market capitalization to estimate what it would take to buy the entire company.  This involves penalizing companies that have a lot of debt and rewarding others that have a lot of cash.

Greenblatt’s Magic Formula gives these two ratios equal weight when selecting investments.  The Formula ranks all companies in the investable universe by Good Company (ROIC) and also by Good Price (Earnings Yield).  Then, each company’s ROIC and Earnings Yield ranks are added together.  Greenblatt’s Formula has an investor buy companies with the best combined rankings, hold each company for a year, and then move funds into new highly ranked companies.  This approach is sound and simple enough that when someone wants to learn about investing or is considering managing his own equity portfolio, we strongly encourage him to read Greenblatt’s book. 

Our Investor’s Questions

Our investor had two questions about Greenblatt’s formula that we examined with our technology and simulation capabilities. 

  1. When selecting investments, what matters more:  that you invest in good companies or that you make your investments at good (that is, low) prices.  Put another way, should Earnings Yield and Return on Invested Capital be weighted equally?  What occurs if one is given more weight in the ranking system?

  2. As Greenblatt’s study focused on the period from 1988 to 2004, how would this approach have performed across a longer period of time?  Are there environments where this approach tends to do well or poorly?

You will find the insights we uncovered interesting, both regarding the ‘Magic Formula’ itself and also how strategies that focus on buying companies when they are “on sale” – something that Euclidean and the Magic Formula have in common – might perform during pessimistic periods when the average company’s valuation contracts as a multiple of earnings.

The Study and Simulation Assumptions

Our project was to examine how 10 different weightings of Good Price (Earnings Yield) and Good Company (Return on Invested Capital) might have performed across the 40-year period from 1973 to 2012.  Our goal was to find the combination that would have performed best during that period and to seek insight regarding the types of environments when different combinations might yield better results. 

In these simulations, we adopted certain assumptions that were relevant to our investor.  We assumed $100M assets under management and set the investable universe to include all companies with market capitalizations greater than $400M, with both assumptions in 2010 dollars.  As a $400M company meant something different in 2010 than in 1973, we adjusted the target portfolio assets and the minimum market capitalization by the SP500 Index value going back in time. [3] [4] 

The simulations looked at 10 different portfolios, each with 50 holdings.  The portfolios were constructed by ranking all companies in the investable universe by Good Company (Return on Invested Capital) and Good Price (Earnings Yield), and then combining the ranks based on each of 10 different weightings.  Thus, we had portfolios built with 100% Good Company / 0% Good Price on one end of the spectrum, then 90%/10%, 80%/20%, … all the way to 0% Good Company / 100% Good Price on the other side.  In the Appendix, we share other important assumptions involved in conducting these simulations. 

Simulated Results – 40 Years

You might be surprised to see the chart below.  Across the 40 years of these simulations, the best performing portfolio was the one that focused solely on buying companies at very low prices. The 100% Earnings Yield portfolio compounded at 18.6%, whereas the 50/50 and 100% ROIC portfolios returned 16.8% and 13.5%, respectively.  Any inclusion of the concept of “good company” as represented by ROIC seems to have reduced long-term returns. 

Here are several observations on these results:

  1. Meaningful differences in compounded rates of return create stunning differences over long-term periods.  Across the 40-years, $1 invested in the SP500 would have grown into $40.  This compares to $1 growing to $185 for the ROIC-only portfolio, $500 for the 50/50 portfolio, and $925 for the Earnings Yield-only portfolio.  When thinking about these numbers, note that inflation during this period averaged more than 4% and the real value of $1 in 1973 would be less than $0.20 today.
  2. This study certainly does not settle the question of when selecting investments, what matters more:  that you invest in good companies or that you make your investments at good (that is, low) prices.  This study does, however, raise the question of whether a company’s last-year Return on Invested Capital (ROIC) provides the best lens for assessing whether that company is “good” or not. 
  3. Even so, it is interesting that the portfolio focused solely on ROIC still would have done much better than the broader market.  The big lesson is perhaps that by ignoring external factors and focusing on something valuable and intrinsic to a business, you can do well across long periods even if you are not price sensitive.  Of course, the results show you do even better by being highly sensitive to price.  

Results – By Type of Environment

The results get more interesting when you look at them in smaller time slices.  For example, although the 100% earnings yield portfolio does best across the 40-year period, the Good Company (100% ROIC) portfolio wins in the 1990s. [5]  Why might this be?  Perhaps in periods like the 1990s when the stock market roars ahead without major drawdowns (such as experienced in 1973-1974, 2000-2002, and 2007-2009), you are rewarded for investing in the best of the “good companies” regardless of price.   In addition, digging into smaller time slices helps makes sense of others’ observations.  For example, Greenblatt’s book talked about the merits of a 50/50 weighting between Good Company and Good Price.  His study focused on the period 1988-2004, a period in which the 50/50 portfolio in our study also outperformed the other simulated portfolios. 

Although each slice in time can tell a slightly different story, we think a broad insight can be found by looking at the entire simulation segmented into optimistic and pessimistic periods. 

Periods of Optimism (increasing Earnings Multiples) and Pessimism (Decreasing Earnings Multiples)

Professor Robert Shiller of Yale University invented the Shiller P/E (also known as the Cyclically Adjusted PE or CAPE) to measure the market's valuation.  The Shiller P/E looks at the SP500 index level in relation to the long-term (trailing 10-year) earning power of the index’s constituent companies.  We used the Shiller P/E to segment the simulation period into optimistic and pessimistic periods. 

Our first step was to divide the study into 10 four-year segments.  We grouped those four-year segments into pessimistic and optimistic periods.  Pessimistic periods were those when the Shiller P/E fell across the four-years.  Optimistic periods were those when the Shiller P/E was rising.   In the following chart, you can see how the Shiller P/E evolved over the simulation period. 

Next, we looked at the returns of the SP500 and three simulated portfolios – 100/0, 50/50, and 0/100 weightings of Good Company (Return on Invested Capital) and Good Price (Earnings Yield) – grouped into pessimistic and optimistic periods.  

As you review this chart, you will notice that the best weighting of Good Company (Return on Invested Capital) and Good Price (Earnings Yield) depends on the type of environment you find yourself in.  During optimistic times, you may do better by not being price sensitive.  Then again, pretty much all approaches – including being indexed to the market – tend to do well during optimistic times.  However, in pessimistic periods when price-to-earnings multiples compress, there is an overwhelming advantage to being invested in the least-expensive companies (e.g., those with the highest earnings yields).  

Relating This Study to Euclidean

We are always amazed by how well one can do across long periods by committing to a simple approach for selecting investments based on companies’ intrinsic qualities.  Whenever we explore variations on this theme, our conviction in Euclidean’s investment process deepens. 

Euclidean’s approach is similar to the simple models shown in this study in that we apply a systematic approach to investing in good companies at good prices.  Our approach differs, however, in important ways regarding how we think about what makes for a good company and how we think about price. 

For example, we have found that there are better ways of assessing a company’s true character – that is, the “goodness” of a company – than simple measures such as Return on Invested Capital.  We have found that one can do better by developing a nuanced understanding of a company through the relationships between its long-term operating history, consistency of operations, riskiness of its balance sheet, and rate of growth.  

We have also found that concepts like return on capital and earnings yield are better examined over longer periods of time than just one year.  It makes sense that we found this to be so.  First, companies can make certain accounting decisions, such as how they value inventory, that impact the timing of earnings.  Thus, two companies that are the same in every way except for a different accounting choice can show different quarterly and annual results.  These differences tend to average out when companies are examined over long periods.  In addition, consider what you really care about as an investor.  Do you care that a company had a good year and is cheap on last-year earnings?  Or do you care that you are buying a company that has a consistent track record of generating results and is inexpensive based on its long-term, demonstrated earning power? 

Of course, in any given year or period, the approach that works best depends on the circumstances of the day.  For example, Euclidean would have done better in 2011 and 2012 if we had operated with a larger number of positions and constrained our investment universe to larger companies.  However, across the past 40 years, we would have greatly benefited by looking for investments in a larger pool of opportunities and concentrating our funds.  This is why we do not alter our approach.  We believe we have the weight of history on our side, and we expect that our investors will benefit as we adhere to our process during the decades ahead.  We accept that across intermediate periods we will often look more or less smart than we really are. 

With this in mind, we note that we have done well during our first five years in a period characterized by generally increasing valuation multiples.  As we look ahead, we feel it is inevitable that multiples will someday compress.  When they do, we believe our style of investing is likely to perform even better. 

*****

We greatly value the privilege of managing a portion of your assets and want you to be an informed Euclidean Investor.  We are available to discuss the content shared here, individual positions in our portfolio, or any other questions you might have.  Please call us at any time.  We enjoy hearing from you. 

Best Regards,

John & Mike

[1]  Professor Robert Shiller of Yale University invented the Schiller P/E to measure the market's valuation.  The Schiller P/E looks at the SP500 index level in relation to the 10-year average inflation-adjusted, annual earnings for the SP500 constituent companies.  The current Shiller P/E ratio of 23.3 is approximately 40% higher than its historical average (from 1880 through today) of 16.5.

[2] ROIC = (Earnings Before Interest & Taxes + Depreciation – CapEx) / (Net Working Capital + Net Fixed Assets) Earnings Yield = (Earnings Before Interest & Taxes + Depreciation – CapEx) / Enterprise Value (Market Value + Debt – Cash)

[3] For example, as the SP500 Index was approximately 1130 in January 2010 and just above 620 in January 1996, the $400M minimum market capitalization threshold for 2010 would have been approximately $220M at the start of 1996.  

[4] This assumption creates results that differ from Greenblatt’s.  In Greenblatt’s book, his investable universe is the 3,500 largest domestic companies.  By limiting our simulations to companies with market capitalizations of $400M or more in 2010 dollars, there is an average of 1,500 companies in our investable universe across the period 1973-2012.  As you might expect, by investing in a smaller investment universe and thus having access to fewer opportunities, our simulated returns are considerably lower than what Greenblatt shared in his book.  

[5] To keep this letter at a reasonable length, we have not included the returns by decade detail.  We are happy to share this information with you.  Just call or email us anytime.  


We share these numbers because they are easy-to-communicate measures that show the results of our systematic process for buying shares in historically sound companies when their earnings are on sale. [6] [7]

It is important to note that Euclidean uses similar concepts but different measures to assess individual companies as potential investments.  Our models look at certain metrics over longer periods and seek to understand their volatility and rate of growth.  Our process also makes a series of adjustments to company financial statements that our research has found to more accurately assess results, makes complex trade-offs between measures, and so on.  These numbers should, however, give you a sense of what you own as a Euclidean Investor.  In general, higher numbers for these measures are more attractive.  The key measures are:

  1. Earnings Yield – This measures how inexpensive a company is in relation to its demonstrated ability to generate cash for its owners. A company with twice the earnings yield as another is half as expensive; therefore, all else being equal, we seek companies with very high Earnings Yields.  Earnings Yield reflects a company’s past four-year average earnings before interest and tax, divided by its current enterprise value (enterprise value = market value + debt – cash).
  2. Return on Capital – This measures how well a company has historically generated cash for its owners in relation to how much capital has been invested (equity and long-term debt) in the business. At its highest level, this measure reflects two important things.  First, it is an indicator of whether a company’s business is efficient at deploying capital in a way that generates additional income for its shareholders.  Second, it indicates whether management has good discipline in deciding what to do with the cash it generates.  For example, all else being equal, companies that overpay for acquisitions, or retain more capital than they can productively deploy, will show lower returns on capital than businesses that do the opposite.  Return on Capital reflects a company’s four-year average earnings before interest and tax, divided by its current equity + long-term debt. 
  3. Equity / Assets – This measures how much of a company’s assets can be claimed by its common shareholders versus being claimed by others.  High numbers here imply that the company owns a large portion of its figurative “house” and, all else being equal, indicates a better readiness to weather tough times.
  4. Revenue Growth Rate – This is the annualized rate a company has grown over the past four years. 

[6] All Euclidean measures are formed by summing the values of Euclidean’s pro-rata share of each portfolio company’s financials.  That is, if Euclidean owns 1% of a company’s shares, it first calculates 1% of that company’s market value, revenue, debt, assets, earnings, and so on.  Then, it sums those numbers with its pro-rata share of all other portfolio companies.  This provides the total revenue, assets, earnings, etc. across the portfolio that are used to calculate the portfolio’s aggregate measures presented here. 

[7] The S&P 500 measures are calculated in a similar way as described above.  The market values, revenue, debt, assets, earnings, etc., for each company in the S&P 500 are added together.  Those aggregate numbers are then used to calculate the metrics above.  For example, the earnings yield of the S&P 500 is calculated as the total average four-year earnings before interest and taxes across all 500 companies divided by those companies’ collective enterprise values (all 500 companies’ market values + cash – debt).


Euclidean’s Largest Holdings as of May 15, 2013

In 2013, we established a practice of sharing our 10 largest positions.  We are sharing this information because a growing number of you have expressed an interest in talking through individual positions as a means of better understanding how our investment process seeks value.  Also, Euclidean aspires to grow such that next year our size would require us to disclose, via 13F filings, our positions 45 days after the end of each quarter.   We feel that adopting a practice this year of disclosing our largest positions with a similar lag will lead us nicely into our new requirements in the future. 

We are available to discuss these holdings with you at your convenience.  We are happy to explain both why our models have found these companies to be attractive as well as our sense of why the market has been pessimistic about their future prospects. 

Euclidean’s 10 largest positions as of May 15, 2013 (in alphabetical order)

  1. Almost Family - AFAM
  2. Deckers Outdoor - DECK
  3. DeVry - DV
  4. GameStop - GME
  5. G-III Apparel Group - GIII
  6. ITT Educational Services - ESI
  7. Kirkland's - KIRK
  8. Lincoln Educational Services - LINC
  9. True Religion Apparel - TRLG
  10. USANA Health Sciences - USNA


Appendix – Simulation Assumptions

  1. The historical simulation results herein do not represent the results of actual trading and may not reflect the impact that material economic and market factors may have had on the adviser’s decision making if the adviser were actually managing client money.  The simulated results were achieved through the retroactive application of a model designed with the benefit of hindsight.  No investment strategy or risk management technique can guarantee return or eliminate risk in any market environment.
  2. In the simulation, Standard & Poor’s Compustat database was used as a source for all information about companies and securities for the entire simulated time period.  From 1987 to 2012, Computstat’s point in time database was used such that the simulation processed financial data concurrent with when it would have been accessible in the past.  Prior to 1987, the period when the historical record is less clear on when financial data would have been received, the simulations assume that financial data was not available to investors until 90 days following the end of the applicable fiscal quarter. 
  3. The simulations were restricted to non-financial companies listed on the NYSE, NASDAQ, and AMEX stock exchanges.
  4. The simulations assumed $100M assets under management and assumed that the investable universe includes all companies with market capitalizations greater than $400M, with both assumptions in 2010 dollars.  As a $400M company meant something different in 2010 than in 1973, the simulations adjusted the target portfolio assets and the minimum market capitalization by the SP500 Index value going back in time.  For an example, as the SP500 Index was around 1130 in January 2010 and just under 620 in January 1996, the $400M minimum market capitalization threshold for 2010 would have been approximately $220M at the start of 1996. 
  5. Companies in the investable universe were ranked by Earnings Yield and Return on Invested Capital (ROIC).  In each of the ten simulations, these ranks were combined by a specific weighting Earnings Yield and ROIC. For example, one simulation’s weighting was 30% Earnings Yield and 70% Return on Invested Capital, whereas another simulation’s weighting was 80% Earnings Yield and 20% Return on Invested Capital.  Each simulation invested in the 50 companies with best combined ranks of that simulation’s respective weighting of Earnings Yield and ROIC.
  6. To minimize the potential impact, positive or negative, of market timing and to show how an equally-weighted, 50-position portfolio might have performed at each point in time, the portfolios were rebalanced monthly to equally weight the 50 securities in each portfolio.
  7. The purchase and sale price for a security was the volume-weighted average closing price for the security over the first 10 trading days of each month.  The simulations assumed a trading cost of $0.02 per share. The simulations also assumed a maximum participation of 15% of a target holding’s daily volume over the 10-day trading window. 
  8. For the model and the S&P 500, the annual returns presented are calculated by compounding the monthly returns (including dividends) between Dec 31st of the prior year and Dec 31st of the current year.
  9. The simulation performance does not reflect the deduction of any investment advisory fees.
  10. Back-tested performance results have certain inherent limitations.  No representation is being made that any model or model mix will achieve performance similar to that shown.  Simulated performance and actual prior performance provide no guarantee of future performance.