Earnings, Analyst Estimates, and Estimating Future Prospects

This has been an interesting earnings season.  Many companies have been beating earnings estimates once certain one-time items are excluded.  This has led to criticism by market commentators alleging that earnings estimates and/or adjusted earnings are not a reliable guide for individual stocks or the market as a whole.  There’s some truth there, but let me try to give a more nuanced view.

What are we really trying to estimate?

Earnings estimates do not primarily exist for the purpose of estimating the next few quarters’ or years’ earnings.  They exist for estimating the future path of free cash flows.  Wait, what is free cash flow?  Free cash flow is the amount of money that you can take away from a business at the end of an accounting period and leave the company as well off as it was at the beginning of the accounting period.  How does that compare to earnings?  Typically, non-cash charges like depreciation and amortization get added back to earnings, and maintenance capital expenditures get deducted — what remains is free cash flow, which can be used for dividends, buybacks, and investments to build the business.  Free cash flow is similar to what I will call “run rate” earnings, though there are some differences.

In that sense, analysts are trying to estimate “run rate” earnings when they adjust for “one time” events.  Absent these abnormal occurrences which won’t happen next quarter, what would the earnings have been?  Surprises above and below the consensus estimates of the analysts provide information to investors.  Positive surprises indicate that the future run rate for earnings might be higher, and vice-versa for negative surprises.

I say “might be,” rather than “will be,” because of randomness in profitability, or, an inability to truly figure out all of the abnormalities and timing differences in a given quarter.  Typically, several positive earnings surprises must take place before estimates of the run rate begin to rise.  One is normal randomness, two is happenstance, three is a pattern, four is a change in trend.

Guiding up, guiding down

Of course, corporate management can make life easier by giving earnings guidance to analysts, and then guiding the estimates up or down as they see best.  They can also break out their estimates of operating or adjusted earnings in order that analysts can decide for themselves which adjustments are “one time,” and which aren’t.  (In my opinion, Allstate does a particularly good job with this, as does Assurant.)

But changing guidance is powerful, particularly for companies with a reputation for UPOD (underpromise, overdeliver), as opposed to companies with a reputation for OPUD (overpromise, underdeliver).  When analyzing guidance changes, one must adjust for prior earnings surprises to figure out whether th raise in guidance reflects only past successes, or forecasts greater successes in the future.

That’s a lot of “one time” events!  Why do so many of them raise operating earnings?  Shouldn’t the difference net to zero over a long enough timespan?

Alas, once we start adjusting earnings to create operating earnings, we enter a new world with a new accounting basis that superficially resembles “run rate earnings” but with a fault.  Almost every quarter has its parade of “one time” events.  On average, the adjustments raise operating earnings over ordinary GAAP earnings.  Managements are more incented to find the positive adjustments to earnings in the short run.  Obvious negative adjustments get accounted for, but non-obvious ones don’t get searched for.

But clever investors eventually adjust for this.  Here’s a way to do it.  Analyze a long period of time, say five years, or the CEO tenure, whichever is shorter.  Look at the growth in book value, and add back dividends.  Compare that to cumulative diluted earnings over the same time period.  If cumulative diluted earnings are higher, then earnings are inflated.  Here’s another way: if you have a long enough data series, add up operating and diluted earnings over a long period of time, say five years, or the CEO tenure, whichever is shorter.  If operating earnings are significantly larger, there is a company that is using operating earnings to make itself look more profitable than it actually is.

If nothing else, over a long enough period of time, this is a means of measuring the honesty of management teams where it comes to financial reporting.  In my book, honest/conservative management teams deserve higher multiples than less scrupulous management teams.  These measures document the games that management teams play in order to make their results look better than they should appear.

How are we doing versus the expectations of the market?

The investment game is one where profitability performance versus expectations determines price performance.  Prices don’t ordinarily react to backward-looking data, unless there is a big one-time charge that adds a lot to book value, or takes a lot away, perhaps impairing the future prospects of the firm.  Prices do react to the signals given through earnings relative to expectations, as it leads investors to update their views of potential future run rate earnings, or, free cash flow.

Can we make a lot of money off of this effect?

Not so much any more.  There are many investors following earnings momentum, earnings surprises, analyst estimate revisions, and price momentum, that the average investor can’t make a lot of money off of this effect.  But ordinarily it does help an investor understand what is going on when stock prices move after earnings are released.  Expectations of the future change, and that drives current stock prices.

Full disclosure: long ALL AIZ