Which set of earnings is used to work out the P/E of a stock

Question

So company earnings can fluctuate quite a bit, based on the market conditions, seasonal conditions, whether there was a big spend in the last year etc.

Therefore, even the last years earnings could differ quite a lot from this year.

So I was wondering like when we work out the P/E ratio, which set of earnings is used? Is it the real annual earnings from last year? Or is it the latest estimates of what the earnings will be? And who makes these estimates? Is it the market commentators or the company saying "we'd expected to make this much"?

One example I can think of is with BHP. I've been looking at their P/E ratio and it has been around 8 for the past few months.

Yesterday, they announced that profit fell 35% following the news of 'BHP Mothballs Expansion' smh.com.au

As suggested in the linked article, this news was not a surprise and even beat "Market Consensus" (which is what?) by a little bit.

So I guess in this example, as the P/E ratio did not just jump 35% on that news so that suggests to me that the p/e had already taken this news into account even before it was announced. If so, where are these numbers coming from?

Answer 1

This is a note from my broker, CMC Markets, who use Morningstar:

Morningstar calculate the P/E Ratio using a weighted average of the most recent earnings and the projected earnings for the next year. This may result in a different P/E Ratio to those based solely on past earnings as reported on some sites and other publications.

They show the P/E as being 9.93.

So obviously past earnings would usually be used but you would need to check with your source which numbers they are using. Also, as BHP's results just came out yesterday it may take a while for the most recent financial details to be updated.

Answer 2

There are two common types of P/E ratio calculations: "trailing" and "forward" (and then there are various mixes of the two).

Trailing P/E ratios are calculated as [current price] / [trailing 12-month EPS]. An alternative is the Forward P/E ratio, which is based on an estimate of earnings in the coming 12 months. The estimate used is usually called "consensus" and, to answer your question, is the average estimate of analysts who cover the stock. Any reputable organization will disclose how they calculate their financials. For example, Reuters uses a trailing ratio (indicated by "TTM") on their page for BHP.

So, the first reason a PE ratio might not jump on an announcement is it might be forward looking and therefore not very sensitive to the realized earnings.

The second reason is that if it is a trailing ratio, some of the annual EPS change is known prior to the annual announcement. For example, on 12/31 a company might report a large drop in annual earnings, but if the bulk of that loss was reported in a previous quarterly report, then the trailing EPS would account partially for it prior to the annual announcement.

In this case, I think the first reason is the culprit. The Reuters P/E of nearly 12 is a trailing ratio, so if you see 8 I'd think it must be based on a forward-looking estimate.

Answer 3

@jlowin's answer has a very good discussion of the types of PE ratio so I will just answer a very specific question from within your question:

And who makes these estimates? Is it the market commentators or the company saying "we'd expected to make this much"?

Future earnings estimates are made by professional analysts and analytical teams in the market based on a number of factors. If these analysts are within an investment company the investment company will use a frequently updated value of this estimate as the basis for their PE ratio. Some of these numbers for large or liquid firms may essentially be generated every time they want to look at the PE ratio, possibly many times a day. In my experience they take little notice of what the company says they expect to make as those are numbers that the board wants the market to see. Instead analysts use a mixture of economic data and forecasting, surveys of sentiment towards the company and its industry, and various related current events to build up an ongoing model of the company's finances. How sophisticated the model is is dependent upon how big the analytics team is and how much time resource they can devote to the company. For bigger firms with good investor relations teams and high liquidity or small, fast growing firms this can be a huge undertaking as they can see large rewards in putting the extra work in. The

At least one analytics team at a large investment bank that I worked closely with even went as far as sending analysts out onto the streets some days to "get a feeling for" some companies' and industries' growth potential.

Each analytics team or analyst only seems to make public its estimates a few times a year in spite of their being calculated internally as an ongoing process. The reason why they do this is simple; this analysis is worth a lot to their trading teams, asset managers and paying clients than the PR of releasing the data. Although these projections are "good at time of release" their value diminishes as time goes on, particularly if the firm launches new initiatives etc.. This is why weighting analyst forecasts based on this time variable makes for a better average.

Most private individual investors use an average or time weighted average (on time since release) of these analyst estimates as the basis for their forward PE.