Be careful with those trailing earnings
A lot people, including me, rely heavily on P/E ratios to evaluate securities or the market as a whole. One of the problems with the P/E ratio is the denominator, earnings. There are times when one should be extremly careful with the earnings figure. Now is one such time.
For the first time since 1933, S&P 500 earnings for a quarter (the 4th quarter of 2008) is negative. The loss is largely due to massive losses reported at financial companies. It's not clear how much of those losses are real, since earnings that depend on write-offs can reverse if the underlying assets gain value. In any case, the negative earnings causes the P/E ratio to be negative for the 4th quarter of 2008, or, if you are looking at the full trailing year, a very large P/E (since that negative quarter plus the other three positive quarters ends up in a tiny earnings numbers.)
Impact On P/E Ratios
To see what I mean, consider the current valuations of some indexes. The S&P 500 has a trailing 12 month P/E of 70.85. The Russell 2000, which represent small-cap stocks, has a negative or undefined P/E. And the Dow Jones Transportation Average has a P/E of 4514.31! The DJTA P/E might actually be the highest P/E for a major average in the history of America. I certainly have never seen any major index with a P/E above 4000 :)
Clearly those numbers cannot be interpreted the normal way you would look at P/E ratios. Unless you think the losses are going to continue forever, one shouldn't rely on these P/E ratios. To overcome this extreme data point, you may want to look at forward P/Es or P/Es based on operating income, or P/Es excluding the financial sector*. In terms of forward P/Es, the indexes are posting typical P/Es between 15 and 20 (Russell 2000 and DJTA still have very high forward P/Es of 47.49 and 53.53, respectively.) I wouldn't blindly follow analyst forecasts; use this as a rough guide.
(* You should never exclude a sector because it is posting losses. Then you end up overstating the market results. However, my view is that the current situation is unique. We are facing a situation where a lot of financials posted large one-time losses. It is reasonable to conclude that these companies will not post similar losses in the future. )
Impact On Dividends
The negative earnings in 4Q08 for the S&P 500 also impacts dividend analysis. I think it mostly impacts anyone trying to gauge market valuation based on dividend payouts; I don't think what I am about to say is an issue for dividend-oriented investors.
In the July 2009 Elliott Wave Theorist, a newsletter for traders that I came across (the July one is freely available if you register), Robert Prechter has the following chart illustrating the dividend payout ratio. Prechter is making a bearish case based on the notion that the dividend payout ratio is 300%. In other words, companies are paying out way more than they earn and this is unsustainable.
.png)
I looked at dividend payouts and annoucements in a post a few months ago and share Robert Prechter's sentiment. However, it is not as bearish as Prechter implies. Prechter implies that the companies must cut dividends if earnings don't triple. I don't think we can draw much from the above-100% payout ratio because of the abnormal negative earnings quarter I mentioned above.
First of all, some companies will pay out more than they earn because, well, executives can mask the problem while shafting the shareholders. This involves taking on more debt or issuing shares and paying out dividends. Financial companies have been notorious in the last two years years. I have beaten to death the issue of how AIG not only paid out dividends but also increased it while issuing shares to pay for them. I have also cited the Canadian banks for their shareholder-wealth-destroying strategy of issuing a ton of shares in the last year in order to, largely, keep paying out dividends at a constant level. This tactic is not sustainable but it can go on for a long time.
Furthermore—this is really the main point—it is not that inconceivable to see earnings triple. Prechter is correct in suggesting that this is an extreme scenario but we are in an unusual, once-in-50-years, scenario. It can happen because of the extremely low earnings in the last year. Indeed, analysts are forecasting earnings rise slightly more than 3x (you can tell quickly by noticing how the trailing p/e drops from around 70 to a forward p/e around 18.)
To sum up, we need to be careful with the earnings numbers because of massive losses posted in the last year. The important question is not the abnormaly low earnings in the past year, but what happens in the future. The question to think about, if you like thinking about the big picture, is how far earnings will rise. Analysts have the S&P 500 P/E dropping from 70 to 18 but is that too optimistic? Or is not optimistic enough? If you like thinking about dividends rather than P/Es, the question is where the payout ratio will end up. If it ends up at, say, an 80% payout ratio, it's probably not very bullish (since companies will end up paying out almost all their earnings as dividends); but if it falls to 60%, it means they still have lots of profits beyond their current dividend payments.
For the first time since 1933, S&P 500 earnings for a quarter (the 4th quarter of 2008) is negative. The loss is largely due to massive losses reported at financial companies. It's not clear how much of those losses are real, since earnings that depend on write-offs can reverse if the underlying assets gain value. In any case, the negative earnings causes the P/E ratio to be negative for the 4th quarter of 2008, or, if you are looking at the full trailing year, a very large P/E (since that negative quarter plus the other three positive quarters ends up in a tiny earnings numbers.)
Impact On P/E Ratios
To see what I mean, consider the current valuations of some indexes. The S&P 500 has a trailing 12 month P/E of 70.85. The Russell 2000, which represent small-cap stocks, has a negative or undefined P/E. And the Dow Jones Transportation Average has a P/E of 4514.31! The DJTA P/E might actually be the highest P/E for a major average in the history of America. I certainly have never seen any major index with a P/E above 4000 :)
Clearly those numbers cannot be interpreted the normal way you would look at P/E ratios. Unless you think the losses are going to continue forever, one shouldn't rely on these P/E ratios. To overcome this extreme data point, you may want to look at forward P/Es or P/Es based on operating income, or P/Es excluding the financial sector*. In terms of forward P/Es, the indexes are posting typical P/Es between 15 and 20 (Russell 2000 and DJTA still have very high forward P/Es of 47.49 and 53.53, respectively.) I wouldn't blindly follow analyst forecasts; use this as a rough guide.
(* You should never exclude a sector because it is posting losses. Then you end up overstating the market results. However, my view is that the current situation is unique. We are facing a situation where a lot of financials posted large one-time losses. It is reasonable to conclude that these companies will not post similar losses in the future. )
Impact On Dividends
The negative earnings in 4Q08 for the S&P 500 also impacts dividend analysis. I think it mostly impacts anyone trying to gauge market valuation based on dividend payouts; I don't think what I am about to say is an issue for dividend-oriented investors.
In the July 2009 Elliott Wave Theorist, a newsletter for traders that I came across (the July one is freely available if you register), Robert Prechter has the following chart illustrating the dividend payout ratio. Prechter is making a bearish case based on the notion that the dividend payout ratio is 300%. In other words, companies are paying out way more than they earn and this is unsustainable.
.png)
I looked at dividend payouts and annoucements in a post a few months ago and share Robert Prechter's sentiment. However, it is not as bearish as Prechter implies. Prechter implies that the companies must cut dividends if earnings don't triple. I don't think we can draw much from the above-100% payout ratio because of the abnormal negative earnings quarter I mentioned above.
First of all, some companies will pay out more than they earn because, well, executives can mask the problem while shafting the shareholders. This involves taking on more debt or issuing shares and paying out dividends. Financial companies have been notorious in the last two years years. I have beaten to death the issue of how AIG not only paid out dividends but also increased it while issuing shares to pay for them. I have also cited the Canadian banks for their shareholder-wealth-destroying strategy of issuing a ton of shares in the last year in order to, largely, keep paying out dividends at a constant level. This tactic is not sustainable but it can go on for a long time.
Furthermore—this is really the main point—it is not that inconceivable to see earnings triple. Prechter is correct in suggesting that this is an extreme scenario but we are in an unusual, once-in-50-years, scenario. It can happen because of the extremely low earnings in the last year. Indeed, analysts are forecasting earnings rise slightly more than 3x (you can tell quickly by noticing how the trailing p/e drops from around 70 to a forward p/e around 18.)
To sum up, we need to be careful with the earnings numbers because of massive losses posted in the last year. The important question is not the abnormaly low earnings in the past year, but what happens in the future. The question to think about, if you like thinking about the big picture, is how far earnings will rise. Analysts have the S&P 500 P/E dropping from 70 to 18 but is that too optimistic? Or is not optimistic enough? If you like thinking about dividends rather than P/Es, the question is where the payout ratio will end up. If it ends up at, say, an 80% payout ratio, it's probably not very bullish (since companies will end up paying out almost all their earnings as dividends); but if it falls to 60%, it means they still have lots of profits beyond their current dividend payments.
From what I've seen of American companies, they tend to pay out a regular dividend for a year after earnings have cratered. If the EPS collapse continues for more than a year, then the dividend gets cut or passed - but there's typically a year's worth of lead time.
ReplyDeleteBased upon this rule of thumb, I'd say that there's no real immediate threat to the dividend level of the S&P 500 as a whole - and there won't be unless we're in the early stages of Great Depression II.
By the way, if you're interested: traditional corporate law permitted the paying out of dividends up to the point where retained earnings shrunk to zero.