Financial analysts it seems are always playing games. If you ask ten of Wall Street's analysts what is the market's P/E ratio, you are likely to get ten different answers. You also will no doubt notice that most of their answers will suggest the market is currently "fair" or even "undervalued".

Which of these S&P 500 P/E ratios is correct? 18.2x, 15.5x, or 13.0x?

This is a trick question because all three are correct. But which one you choose will obviously make a massive difference in assumed valuation. Put another way: 18.2x is 42% larger than 13.0x, and is a major discrepancy and potential disaster for your portfolio.

*How it is Supposed to Work*

The P/E ratio is a valuation metric that takes the price divided by a company's earnings to come up with a ratio that can be compared across comparable assets through time. In short the P/E is used to help decide if stocks are expensive, fairly valued, or cheap.

As the ratio implies there are two inputs needed. The numerator, price, is typically straightforward and usually the current asking price of the market or stock being evaluated.

The denominator is where the issues arise. The earnings piece of the ratio it seems can be interpreted an infinite amount of ways, and therein lies the rub. Deciding which earnings assumptions to use is the crucial step in understanding the P/E ratio you are about to invest in.

*Why all the Earnings Differences?*

The earnings piece of the equation can be broken into three parts.

What kind of earnings is being used, what kind of reporting technique is being used, and finally what year of earnings is being assumed are three questions to help decipher the P/E code.

First, what kind of "earnings" is being used? There is net income earnings, operating earnings, earnings before taxes, earnings before interest and taxes; the list literally goes on and on. Most P/E ratios use the bottom line "net income".

The second item to figure out is if earnings are based on "GAAP" or "Operating" techniques.

From Standard & Poors, the go to source for S&P 500 (IVV - News) earnings analysis, "operating income excludes corporate (M&A, layoffs, financing) and unusual items". These items are typically considered non-recurring by management and are backed out of expenses. Operating earnings (also known as management earnings) is used to try to give a smoother view of earnings, attempting to exclude those items which may not always occur.

GAAP earnings on the other hand include all these costs and are based off of Generally Accepted Accounting Principles, which is what is reported to the SEC in quarterly filings. GAAP earnings are also known as "reported earnings".

The final thing to figure out is the year of earnings used. Since earnings are (wrongly) always expected to grow over time, analysts can take advantage of the rising earnings estimates to lower the P/E ratio. Simply, using forward earnings estimates provides a larger earnings number which lowers the P/E ratio. This is usually how the "earnings games" are being played.

*Why it Matters*

Tackling each of the three questions can keep things consistent and allow for better analysis. It also will help you decipher what the real P/E ratio should be. Sticking with the same kind of earnings as well as the same technique used each quarter is a great start and will provide consistent comparable P/E analysis through time.

In the following table I show two of the three earnings inconsistencies, the technique used (Management or GAAP) and year assumed, and show how it drastically changes the P/E ratio. The data is from Standard and Poor's and uses net income earnings for the S&P 500 (SSO - News) companies.

The primary driver between the two extreme P/E ratios is the date assumed. By using earnings from Dec 2014, an analyst can instantly drop the P/E ratio over 30% (from 16.2x to 12.9x). No big surprise, whichever period provides the lowest P/E is typically the P/E ratio that is quoted by the media, justifying why stocks (SDY - News) are indeed cheap (the long term average P/E ratio is around15x).

The difference between the GAAP and Management earnings makes a difference, but not near as much as the time frame assumed. By pushing out the earnings horizon, P/E ratios drop over 4 handles based on GAAP.

Think this discrepancy is extreme? The difference of the P/E on the Russell 2000 (IWM - News) dwarfs this discrepancy. The current Russell 2000 P/E based on trailing 1Q 2013 GAAP earnings is 58.7x, but the P/E expectation one year from now is only at 17.2x.

What if small cap (UWM - News) earnings don't grow next year? That will mean next year's P/E in reality is much higher than 17.2x. Don't worry though, by that time analysts will have you focusing on the expected growing 2015 earnings, again trying to justify a low P/E and time to buy.

*What to Do*

Any earnings in the future are still just estimates, but the Mar 2013 GAAP reported earnings are actual, already occurred, and real. Essentially you must believe the analysts' estimates for higher and higher earnings to justify a currently low P/E. If you want something more tangible, then using actual reported earnings from Mar 2013 is the route you should take, and that suggests P/Es are high now.

These earnings games are why as far back as September in our ETF Profit Strategy Newsletter we included in our list of 12 Mega Themes "equity valuation risk". We warned, "Cyclical earnings and margins are near all time highs=risk".

Since then earnings have flatlined as margins have indeed contracted and P/Es have gotten much larger. This continues through today, where again our Mega Theme again highlighted, "Fundamentally driven bull markets should rely more on earnings growth and less on investor's willingness to pay ever increasing multiples."

Just because analyst expectations are for continued earnings growth doesn't mean that earnings will ever get there. Did you know that March 2013's GAAP earnings are actually lower than March 2012's? This means earnings have not grown in over a year, yet analysts are still expecting significant growth through 2014.

Looking deeper into the earnings piece of the P/E ratio may help investors avoid buying a market the pundits continually insist is trading below historical averages.

In reality the market can be considered cheap and trading under 14x only if earnings indeed can grow over 30% over the next seven quarters.

The ETF Profit Strategy Newsletter uses fundamental, technical, and sentiment analysis along with common sense to see what is really going on in the stock, bond, forex, and commodity markets. We go beyond traditional analysis techniques to help investors stay on the correct side of the market.

Follow us on Twitter @ ETFguide

**More From ETFguide.com **