Did you know that if you have two arms, then you have more than the average number of arms? The Undercover Economist blog mentions this to cleverly demonstrate how statistics can deceive you. Of course, most people have two arms but the few people with one or no arm pull the mean average down below two. What does this have to do with investing? Well, most people use price multiples like the P/E ratio or P/B ratio for valuation. Like any other statistic, multiples can mislead and be misused.
In his "Little Book of Valuation," professor Aswath Damodoran describes some things to remember while using multiples. I will summarize a few points here but if you want more details, then you know where to find them. Couple that book with his free online online valuation webcasts and you have valuation treasure.
Is the multiple consistently defined?
The P/E ratio is consistent because both the numerator and the denominator refer to equity values. The price is for the equity. The earnings are for the shareholders of that equity. A company with a P/E of 10 means for every dollar of shareholder equity, the company earned its shareholders ten cents over the last year.
Price to EBITDA, however, is inconsistent. The price is for the equity. But the EBITDA is a firm value, which includes equity and debt. A price to EBITDA of 10 means that for every dollar in shareholder equity, the company earned 10 cents, of which some go toward the shareholders and some to the debt holders. Since this is an inconsistent multiple, firms with significant debt will look cheaper than ones with less debt. Price to EBITDA ratios should be higher where more money is left for the shareholder. Enterprise value to EBITDA is a better ratio because both are firm values, so debt is accounted for in the numerator. P/S is another popular inconsistent multiple.
What average are you comparing to the multiple?
It is common practice to compare a company's multiple to the average multiple of the company's sector. The average number of arms example demonstrates that the mean average is not always the right average to use. All multiples have skewed distributions. This just means it makes more sense to use the median average than the mean. For example, in January 2010 the average mean of a US stock P/E was about 30 and the median was about 15. The high outlier P/E ratios pulled the mean average up. The median meant half the P/E ratios were below 15 and half were above 15. Professor Damodoran recommends that the "standard sales pitch of a stock being cheap because it trades at a multiple less than the average for the sector should be retired in favor of one which compares the stock’s pricing to the median for the sector."
Does the multiple carry any bias?
In January 2010, only about 50% of US stocks had P/E ratios. Firms with negative earnings have meaningless negative P/E ratios so they are not reported. Multiples that leave out companies create bias. The average P/E ratio for firms in a group will have an upward bias if it does not include money losing companies. It is as biased as saying the average US algebra student has a B+ grade without including failing students. Professor Damodoran recommends various solutions to this problem. One is to adjust the average ratio down to reflect the money losing companies. Another solution is to use a ratio that does not exclude any companies. In the case of the P/E ratio, using its inverse ratio, earnings to price, prevents elimination of companies.
What variables affect multiples?
If we lived in an ideal world for multiples, we could compare apples to apples. Instead, we can only compare Apple (AAPL, Financial) multiples to those of Google (GOOG, Financial) or Hewlard Packard (HPQ, Financial) or Microsoft (MSFT, Financial). Its like comparing the prices of similar size houses with differences in neighborhoods, schools, parks, etc. Professor Damodoran argues that it is difficult to compare multiples while controlling for the differences in the variables. He lists the following multiples and the variables that would increase their value:
Price to Earnings
Higher Expected GrowthHigher PayoutLower Risk
Price to Book Ratio
Higher Expected GrowthHigher PayoutLower RiskHigher Return on Equity
Price to Sales
Higher Expected GrowthHigher PayoutLower RiskHigher Net Margin
EV to EBITDA
Higher Expected GrowthLower Reinvestment rateLower RiskHigher Return on Invested CapitalLower Tax Rate
EV to Sales
Higher Expected GrowthLower Reinvestment rateLower RiskHigher Operating Margin
Apple might have a lower trailing P/E ratio than Google or its sector while having a higher estimated growth rate. But in order to deem it cheap, you have to account for the fact that it might be riskier or have higher reinvestment needs. Claiming a stock is cheap by just comparing multiples and controlling for maybe one variable is very misleading.
In Conclusion
You can slap a nice looking multiple to most stocks and make them look pretty. XYZ Widgets sells for a forward price to EBITDA of only 4.5! What a steal! Any company looks like a bargain when framed with a low attractive multiple. This means you better know the nature of multiples very well if you use them to value companies.
In his "Little Book of Valuation," professor Aswath Damodoran describes some things to remember while using multiples. I will summarize a few points here but if you want more details, then you know where to find them. Couple that book with his free online online valuation webcasts and you have valuation treasure.
Is the multiple consistently defined?
The P/E ratio is consistent because both the numerator and the denominator refer to equity values. The price is for the equity. The earnings are for the shareholders of that equity. A company with a P/E of 10 means for every dollar of shareholder equity, the company earned its shareholders ten cents over the last year.
Price to EBITDA, however, is inconsistent. The price is for the equity. But the EBITDA is a firm value, which includes equity and debt. A price to EBITDA of 10 means that for every dollar in shareholder equity, the company earned 10 cents, of which some go toward the shareholders and some to the debt holders. Since this is an inconsistent multiple, firms with significant debt will look cheaper than ones with less debt. Price to EBITDA ratios should be higher where more money is left for the shareholder. Enterprise value to EBITDA is a better ratio because both are firm values, so debt is accounted for in the numerator. P/S is another popular inconsistent multiple.
What average are you comparing to the multiple?
It is common practice to compare a company's multiple to the average multiple of the company's sector. The average number of arms example demonstrates that the mean average is not always the right average to use. All multiples have skewed distributions. This just means it makes more sense to use the median average than the mean. For example, in January 2010 the average mean of a US stock P/E was about 30 and the median was about 15. The high outlier P/E ratios pulled the mean average up. The median meant half the P/E ratios were below 15 and half were above 15. Professor Damodoran recommends that the "standard sales pitch of a stock being cheap because it trades at a multiple less than the average for the sector should be retired in favor of one which compares the stock’s pricing to the median for the sector."
Does the multiple carry any bias?
In January 2010, only about 50% of US stocks had P/E ratios. Firms with negative earnings have meaningless negative P/E ratios so they are not reported. Multiples that leave out companies create bias. The average P/E ratio for firms in a group will have an upward bias if it does not include money losing companies. It is as biased as saying the average US algebra student has a B+ grade without including failing students. Professor Damodoran recommends various solutions to this problem. One is to adjust the average ratio down to reflect the money losing companies. Another solution is to use a ratio that does not exclude any companies. In the case of the P/E ratio, using its inverse ratio, earnings to price, prevents elimination of companies.
What variables affect multiples?
If we lived in an ideal world for multiples, we could compare apples to apples. Instead, we can only compare Apple (AAPL, Financial) multiples to those of Google (GOOG, Financial) or Hewlard Packard (HPQ, Financial) or Microsoft (MSFT, Financial). Its like comparing the prices of similar size houses with differences in neighborhoods, schools, parks, etc. Professor Damodoran argues that it is difficult to compare multiples while controlling for the differences in the variables. He lists the following multiples and the variables that would increase their value:
Price to Earnings
Higher Expected GrowthHigher PayoutLower Risk
Price to Book Ratio
Higher Expected GrowthHigher PayoutLower RiskHigher Return on Equity
Price to Sales
Higher Expected GrowthHigher PayoutLower RiskHigher Net Margin
EV to EBITDA
Higher Expected GrowthLower Reinvestment rateLower RiskHigher Return on Invested CapitalLower Tax Rate
EV to Sales
Higher Expected GrowthLower Reinvestment rateLower RiskHigher Operating Margin
Apple might have a lower trailing P/E ratio than Google or its sector while having a higher estimated growth rate. But in order to deem it cheap, you have to account for the fact that it might be riskier or have higher reinvestment needs. Claiming a stock is cheap by just comparing multiples and controlling for maybe one variable is very misleading.
In Conclusion
You can slap a nice looking multiple to most stocks and make them look pretty. XYZ Widgets sells for a forward price to EBITDA of only 4.5! What a steal! Any company looks like a bargain when framed with a low attractive multiple. This means you better know the nature of multiples very well if you use them to value companies.
