In How to Value a Stock Using the Income Statement, I discussed the conflict between (1) the theoretically-pure way to measure the value of a business as an ongoing concern – i.e., discounting the company’s annual free cash flows in perpetuity at an interest rate that accounts for business risk – and (2) the practical difficulties of accurately estimating either future business risk or the company’s annual cash flows over the next 10 years let alone in perpetuity.

Given the practical difficulties of accurately estimating a multitude of future data points, a short-cut method that many smart investors use to value a stock is to limit the estimation process to a single number — a snapshot multiple of a company’s current earnings (or EBITDA or free cash flow or book value).

On the surface, estimating a single number is easier than estimating multiple numbers, but the reality is that this distinction is somewhat illusory because several data “inputs” must be estimated in order to generate the single “output” multiple. Most important inputs are future earnings growth and cost of capital (i.e., business risk). Earnings growth may reasonably be estimated from past experience if the company’s business is stable, but business risk is a slippery measure that can easily be overshot or undershot.

Only Use P/E Ratios to Value Companies with “Predictable” Earnings

Because of the inherent uncertainties in estimation, I previously wrote that:

I would only consider using a P/E ratio on stable stocks with a prolonged operating history and a modicum of earnings-growth reliability.

Earnings predictability simply means that a company’s earnings have historically grown in a consistently stable pattern that is likely to continue in the future.  In contrast, a 2008 study found that corporate earnings that are very volatile over the past five years – jumping up and down haphazardly from one quarter to the next – typically signal uncertainty and unpredictable earnings for five years into the future:

Earnings volatility captures the effects of real and unavoidable economic volatility. Intuitively, firms operating in environments subject to large economic shocks are likely to have both more volatile earnings and less predictable earnings.

The volatility of reported earnings also reflects important aspects of the accounting determination of income. One such aspect is the quality of matching of expenses to revenues. Poor matching acts as noise in the economic relation between revenues and expenses, and thus the volatility of reported earnings increases in poor matching. Poor matching is also associated with poor earnings predictability because the matching noise in reported earnings obscures the underlying economic relation that governs the evolution of earnings over successive periods.

Earnings volatility not only reduces earnings predictability, but it also signals a company without a competitive advantage that has no control over pricing or a loyal customer base, but is vulnerable to the whims of a commoditized marketplace. A 2011 study found that the stocks of companies with high earnings uncertainty significantly underperform the stocks of companies with predictable earnings and that high earnings volatility “strongly predicts lower future earnings.” This helps explain “the greatest anomaly in finance”, where stocks with low betas (a measure of price volatility relative to a benchmark index) have historically outperformed high-beta stocks. Roadrunner’s six-point safety-rating system awards a safety point to low-beta stocks because of this outperformance potential.

A 10-year study by Gurufocus.com similarly found that companies with predictable earnings outperformed. In fact, based on their five-point rating system of predictability, relative outperformance increased in perfect lockstep with increases in predictability:

Companies with Predictable Earnings Outperform the Market

Source: gurufocus.com

Below is a list of ten small-cap stocks that currently enjoy the maximum five-star rating for earnings predictability:

Ten Very Predictable Small Caps

Source: Gurufocus.com

Value Line Investment Survey also ranks stocks by earnings predictability. There must be something to this!

Warren Buffett Likes Stocks with Bond-Like Cash-Flows

Warren Buffett, one of the greatest investors of all time, doesn’t like to gamble with his money and only invests in companies with predictable earnings. In the book entitled The Warren Buffett Way, Buffett was quoted telling some business school students in 1994 that:

I put a heavy weight on certainty. If you do that, the whole idea of a risk factor doesn’t make any sense to me. Risk comes from not knowing what you’re doing.

In fact, Buffett is so certain about the future earnings of companies he invests in that when he estimates the stock’s intrinsic value, he discounts projected cash flows at the same interest rate as risk-free long-term government bonds – no equity risk premium:

Buffett is firm on one point: He looks for companies whose future earnings are as predictable, as certain, as the earnings of bonds. A company’s future cash flow should take on a “coupon-like” certainty. If the company has operated with consistent earnings power and if the business is simple and understandable, Buffett believes he can determine its future earnings with a high degree of certainty. If he is unable to project with confidence what the future cash flows of a business will be, he will not attempt to value the company. He’ll simply pass.

To be clear, unlike some other prominent value investors, Buffett allegedly does attempt to value stocks via a discounted cash flow analysis, and is quoted as saying that short-hand valuations methods based on a snapshot multiple of earnings or book value “fall short.” Nevertheless, his focus on earnings predictability underscores the importance of limiting valuations based on a snapshot multiple to only a select group of companies with persistent earnings that are growing at a stable rate.

“Normal” Earnings Must Be Isolated to Value a Stock

Okay, now that you have narrowed down the list of companies that are suitable for a valuation based on a snapshot multiple based on their predictable earnings, the next step is to isolate a company’s “core” earnings – the “normal” earnings that are recurring and based on the company’s business operations – not financial tomfoolery or one-time events. Even if you calculate the proper earnings multiple to use (i.e., the proper P/E ratio), that multiple will not be useful if applied to a bogus “E” number. 

In other words: Both the “P/E” multiplier and the “E” base number must be accurate for the stock valuation to make sense.

Standard & Poor’s has developed a measurement of “core” earnings (Page 70) that:

focuses on a company’s after-tax earnings generated from its principal businesses. Included are employee stock option grant expenses, pension costs, restructuring charges from ongoing operations, write-downs of depreciable or amortizable operating assets, purchased research and development, M&A related expenses and unrealized gains/losses from hedging activities.

Excluded from the definition are pension gains, impairment of goodwill charges, gains or losses from asset sales, reversal of prior-year charges and provision from litigation or insurance settlements.

That definition has a lot of accounting jargon that can confuse rather than clarify. BusinessWeek explains the concept more simply:

Core earnings represent the difference between the revenue of a company’s principal, or core, business and the costs and expenses associated with deriving that revenue.

A simple example would be a chain of retail stores. The core business is running stores. Look at the revenues and the expenses from those stores, and you can find core earnings. While many retail chains may buy and sell real estate, that isn’t their main business. Neither is running a pension fund or many other things that such a company may do.

S&P believes that for an equity investor to make an investment decision based upon a company’s reasonable earnings expectation, it’s necessary to understand how that organization’s core business will perform in the future.

When analyzing a company’s income statement and balance sheets, one must extract the abnormal from the normal and isolate the core earnings.

Real-Life Example: Analyze the Financial Statements

Consider the financial statements below:

Note: Year ‘T’ corresponds to the most recent annual statement. ‘T-1’ and ‘T-2’ are the previous two annual statements.

The most recent annual earnings are $1.10 per share. A novice investor may simply decide to use this $1.10 figure as the base for calculating the stock’s value based on a snapshot earnings multiple. But that would be a mistake because the income statement includes extraordinary items.

Convert Reported Earnings to Core Earnings

Specifically, two line items entitled “one-time write-down” and “non-operating income” must be accounted for to ascertain normal earnings. Adding back the write-down and deducting the non-operating income yields a core earnings number of 10,958:

Reported net income ($7,709)

–        Non-operating income ($1,740)

+ One-time write-down ($4,989)

= $10,958

Divide $10,958 by the 6,995 shares outstanding yields core earnings per share of $1.57, which is 43 percent higher than the reported number of $1.10!

If you perform the same “core” earnings adjustments to the two previous years of financial statements, the earnings per share numbers in T-1 and T-2 become $1.22 and $1.07, respectively. So, what initially looks like a company with volatile earnings and flat growth over the past two years ($1.10 in T vs. $1.09 in T-2), actually is a company growing “core” earnings steadily at a compounded annual rate of 20% ($1.57 in T vs. $1.09 in T-2).

“Margin of Safety” May Require that Growth Rate be Discounted Before Using as P/E Ratio

According to Peter Lynch, the P/E ratio should equal the earnings growth rate and stocks growing earnings between 20-25 percent annually are his favorite (not too slow and not too fast), so arguably one could multiply $1.57 by 20 and get a stock valuation of $31.40. But our calculated growth rate is based on only two years of changes, which is too short – Lynch describes the proper earnings growth rate as “long term” or “in recent years.” Under neither of these characterizations would two years of earnings growth seem to suffice.

Furthermore, the three years of financial data in our example appear to be tumultuous ones for the company, with substantial “one-time” write-downs occurring in two of the three years. One needs to question whether write-downs truly are “one-time” in nature if they occur in a majority of the years being studied! There is also some funky recapitalization going on in year T-1 with long-term debt decreasing dramatically by 40 percent, shareholder equity increasing by 23 percent, cash balances increasing by 95 percent, and shares outstanding decreasing by 12 percent. This unusual financing activity is an additional reason to discount the company’s growth rate in my valuation.

Because of the limited financial data and numerous write-downs, the valuation principles of conservatism and “margin of safety” require that the company’s annual earnings growth rate of 20 percent be discounted by at least 20 percent, which brings the earnings growth down to a more believable 16 percent. If you use 16 as the P/E multiple for the company’s current core earnings per share of $1.57, the stock’s valuation is $25.12 – 20 percent lower than the original estimate of $31.40.

As it turns out, the three years of financial statements used in this example are from the real-life 1988-1990 financial statements of spirits manufacturer Brown-Forman (NYSE: BF-B). The actual stock price at the time the 1990 financial statements were released in August of that year was $23.25, which is 7.4 percent below my fair-value estimate of $25.12. But, based on a 12 percent discount rate and the stock’s actual dividend-adjusted stock performance over the 20 years from 1990 to 2010, the stock’s present value in August 1990 was worth $26.11, 3.9 percent higher than my $25.12 estimate, so my valuation estimate is narrowly situated in between the stock’s actual market price (+7.4 percent) and the stock’s actual intrinsic value (-3.9 percent). Not bad.

Anchoring Bias is the Enemy of Accurate Stock Valuation

To be fair, I calculated the stock valuation after already knowing what Brown-Forman’s stock price was in August 1990, so my calculation is tainted by “anchoring bias.” Would I have decided to discount the company’s 20-percent earnings growth rate by 20 percent if I hadn’t known that a 20 P/E would overvalue the company? I like to think so, but we’ll never know.

A financial blogger conducted an experiment using the Brown-Forman financial statements above, whereby he provided the same financial data to two groups of stock analysts, except that he provided one group of analysts with no information about the stock’s current market price while providing the other group with false information that the stock was currently trading at $87.12.  The analyst group provided with the false market-price information estimated the stock’s value (on average) at $34, significantly higher (68%) than the average value estimate of $20.21 in the other group.

Conclusion: anchoring bias is real and explains why so many retail investors are willing to buy overvalued stocks – their valuation of a stock is tainted by its current market price. The Internet bubble of 1999-2000 proved that the stock market is not efficiently priced and market prices cannot be relied upon to reflect true value. Remember the wise saying of value investor Benjamin Graham:  “Price is what you pay, value is what you get.”

Intrinsic value and market price often diverge and the way to make money is to buy stocks that have a market price below intrinsic value and sell stocks that have a market price above intrinsic value. Using market prices to determine intrinsic value defeats the entire purpose of value investing.

Around the Roadrunner Portfolios

HomeAway. We were less than thrilled to learn that giant travel site Expedia.com, which owns Booking.com, was jumping into HomeAway’s space with both feet. And mounting competition to our small cap momentum play HomeAway is one reason we’re selling HomeAway.

HomeAway is still a great business and remains the world’s leading online marketplace of vacation rentals, but the price momentum of its stock has been tripped up by competition and work needed on its Internet platforms. With more and larger competitors entering its market, it’s become apparent that HomeAway has to fight off the invaders with more marketing expense while simultaneously investing more cash into an upgraded e-commerce system infrastructure.

All of these added expenses are affecting profitability. A recent J.P. Morgan report cut estimated 2014 EBITDA to $111.8 million from $119.1 million. Consensus analyst estimates for full-year 2014 average 61 cents a share for a P/E of 56 at its current price of $34.50, which could be justified if earnings were growing more quickly than they actually are. The market sees this, and has taken HomeAway off the fast track. We’re selling HomeAway.

PriceSmart is the Costco of Latin America, but unlike Costco, PriceSmart’s comparable sales growth is weakening. Monthly comps have actually been declining from double digits seen in 2012, to the high single digits last year to the mid and low single digits this year. This softening has been reflected in the stock price, which peaked above $120 late last year, but has settled down below $90 now. The current price may be smart as a long-term value play, but not as a momentum play. We’re selling PriceSmart.  

SolarWinds has had some nice earnings surprises since we recommended it last fall and the price peaked in February at $49.95 a share. But since that time the price has settled down in a sub-$40 range.  We think problems digesting acquisitions, salesforce hiring and marketing expenses involved in peddling new products has taken the wind out of SolarWinds’ price momentum. We’re selling SolarWinds.

Valmont Industries cut its earnings outlook and said three of its four main business segments are weak. The company warned its second quarter profits will be in the $2.35 to $2.40 range, versus analyst expectations of $2.79 a share. And 2014 doesn’t look good overall for the Omaha, Neb.-based metal products maker. It now thinks full-year earnings will drop to $9.35 to $9.65 per share, versus the $10 to $10.50 it had signaled earlier. We’ve selling Valmont Industries.

The Oil Patch may be booming, and Western Refining is in a perfect location to leverage WTI crude at its refineries, but it’s geographic edge isn’t translating to the bottom line. Also, new rulings that may allow broad oil exports from the US after a 40-year ban could erode margins at refiners, including Western Refining.

After rising from about $25 a share around $40 in January, Western Refining’s earnings and its stock momentum have stalled – it’s been waffling around $40 since. And the possible sea change in oil export policy would only hurt this refiner struggling to improve profitability. We’re selling Western Refining.