S&P 500: Which Earnings Are Most Relevant to Its Performance? 15 comments
an article to
-
Font Size:
-
Print
- TweetThis
NIPA Corporate Profits may be a better yardstick for measuring the level of the S&P 500 index than the actual GAAP earnings of the companies involved. NIPA stands for National Income and Product Accounts, one of which is Corporate Profits. The US Bureau of Economic Analysis – BEA – includes this number as a sub account in GDP, and as such it is logically consistent with GDP.
As reported earnings of S&P 500 companies fluctuate due to the application of GAAP accounting. GAAP as presently implemented is pro-cyclical – when things are going good, little provision is permitted for losses that are accumulating, almost unnoticed. When things are going bad, write-downs bring both past and future losses together in kitchen sink quarters. NIPA Corporate Profits differ from GAAP, as will be discussed briefly later in the article.
Picturing the Relationship
click to enlarge
As the chart shows, NIPA Corporate Profits and the S&P 500 index are broadly correlated, with the linear regressions running more or less parallel. By coincidence, the units of the S&P 500 Index and NIPA Corporate profits expressed in billions of dollars are very similar. Data is quarterly. Over the 20 year period shown, there have been two periods of time where there were wider divergences.
The first is from 1998 to mid 2002, when the index surged ahead. Those were the years of the tech bubble, the early internet, the heady days of limitless possibilities. NIPA Profits headed down well before the S&P: however, the index eventually plunged to get more or less in sync and then followed profits back up in the recovery.
The second period would be the present. Profits headed down starting in the 2nd half of 2006. The S&P ran ahead for another year but eventually succumbed to economic reality, swooning to as low as 666 on an intra-day basis during the 1st quarter of this year. If profits represent the underlying economic reality, the index has seriously overshot to the downside, and over time can be expected to recover to get back in sync with longer term fundamentals.
Deriving a Formula – another way of looking at the relationship is to do a scatter chart of the S&P 500 Index as a function of NIPA Profits.
The graphing software displays the formula for the linear regression: by this line of thinking the expected value for the index profits of $1,312.6 Billion for Q1 09 would be .75 X 1,312.6 = 984.5 + 226.93 = 1,211.43, round to 1,200. That would be about 30% higher than where we are today.
What about the P/E on the S&P 500? An argument against the above rosy scenario can be made by reference to the S&P 500 P/E, currently at around 60 on a TTM basis. The thinking is, the index is way high compared to its own P/E, it could go down to some unimaginably low level on that basis. I prefer to get at the problem by viewing the S&P 500 Index in conjunction with its TTM Earnings, as shown in the following two charts. The time period runs from 2001 to 2010, with quarterly data points, other than I threw in 3/9/09 for dramatic effect.
Looking at the first chart, the horrible earnings for Q4 08 drove the index down to something of a low point, from which it has since recovered. To illustrate where this line of thinking would take the S&P, I put in projections, based on a linear regression, and using S&P “top down” estimates of future As Reported Earnings.
On the second chart, the March low is again circled as an out-lier. Using the formula for the regression, the indicated value for the S&P 500 Index would be 744.66 + (130.21 X .9) = 861.85, round to 860. That would be a disappointing plunge from where we are today.
S&P Operating Earnings – another version of S&P 500 earnings would be operating earnings, which do not include a lot of the GAAP entries which make the As Reported so horrifying. Again, I used S&P projections and a linear progression formula to develop projections out to January 2011. Again the comforting target of S&P 1,200 emerges as a distinct possibility.
What's the Difference? - A detailed discussion of the difference between NIPA Corporate Profits and the TTM As Reported Earnings on the S&P 500 is beyond the scope of this article, not to mention the skills of the author. Here is a link to a fairly detailed treatment of the subject by the BEA. Corporate Profits includes the earnings of all US corporations as defined for the purpose of the report. The publicly traded companies of the S&P 500 comprise a shifting subset of this universe, and would be about 50% by earnings, at a guess. S&P earnings are according to GAAP, while Corporate Profits are closer to taxable income.
Corporate Profits are less a result of accounting interpretation and are better for their intended purpose, which is to provide information on the state of corporate profitability consistent with GDP. Briefly, S&P 500 earnings, which include huge losses for the likes of AIG and Citigroup, not to mention mark to madness and auditor CYA losses, are not a reliable indication of the well-bing of American business as a whole nor of the desirability of investing in publicly traded US businesses. NIPA Corporate Profits are a more reliable indication along those lines, more long-term in nature.
S&P 500 Earnings are a good indicator as to the probable short term direction of the index, given the correlation observed. Noting the horrendous 4th quarter that has to roll through the TTM computations, S&P earnings are not going to impress until Q4 08 drops out of the computation.
Investment Implications – the accumulated fiscal and monetary stimuli applied to the economy are well in excess of anything Greenspan applied in the wake of 9/11. The fourth quarter saw a massive reset of consumer demand. There is a huge overhang of debt for both consumers and the USG. Under these circumstances it is unlikely NIPA Corporate Profits will recover rapidly: a possible scenario would call for profits to progress upward on a slow but steady path, followed by the S&P 500 index with a 6-12 month lag.
On the other hand, Q4 08 was marked by write-downs that scared the hell out of everyone. The mark to market inquisitors interrogated their victims until the expected confessions materialized. AIG alone confessed to over 60 billion of losses, and Citigroup came in second at 30 billion. Among other companies many of the write-downs reflected genuine re-structurings that will lead to leaner and more profitable businesses going forward. A certain amount of the carnage was smoke and mirrors, goodwill and other intangibles, or mark to market that will eventually revert.
Whether to look past the 2008 4th quarter – in the final analysis it's about whether investors will look past the 4th quarter 2008 as a temporary aberration, something that does not need to be accounted for on a long term basis. This decision is difficult for individual stocks – Ben Graham called it “the vexed question.” But I notice that a number of companies I own that took large write-downs in the 4th quarter are already trading as if investors are looking past them. From the standpoint of the index, AIG and Citigroup are not going to do that again, at least not real soon.
Bullish Hypothesis – as the situation is developing, I see reason to hope for S&P 1,200, based on a conjecture that NIPA Corporate Profits are indicative of stronger fundamentals for American business than are suggested by S&P 500 earnings. This can be tested every quarter when GDP comes out. The wait could be a long one, because S&P earnings on a TTM basis can't improve until Q4 08 drops out, so if the index maintains its correlation to its own earnings it won't do much for a while. In my preferred scenario, the S&P 500 trades in a nervous range for much of the 2nd half, while fundamentals improve slowly. This provides a 3-6 month window where the patient investor can accumulate positions that will yield wonderful gains when the S&P heads back up toward 1,200.
Again, there are two sets of data points that fall well off the linear regression. The cluster above the line is the same period from 1998 to 2002, when the index was well ahead of fundamentals. The second set of two points below the line would be Q4 08 and Q1 09, the meltdown caused by the financial crisis and the panic in the wake of Geithner's first major TV appearance as Secretary of Treasury.
Related Articles
|
-
The Seeking Alpha 100
See allThe top 100 stock
market authors
selected for publication
Seeking Alpha 100 » -
Fastest Climbers
Authors with the largestSee all
increase in followers
in the last 7 days
Fastest Climbers » -
Top Commenters
Top commenters,See all
as ranked by their peers
Top Commenters » -
Top Instabloggers
The 100 most active
Instabloggers.-
1
David Fry
-
2
TraderMark
-
3
Don Dion
-
4
Cliff Wachtel
-
5
Mike Havrilla
Top Instabloggers » -
1
-
Top StockTalkers
Top
StockTalkers.-
1
David Fry
-
2
Cliff Wachtel
-
3
Mike Havrilla
-
4
David White
-
5
Hedged In
Top StockTalkers » -
1
Using your derived formula, suppose earnings recover to $200 billion. then: S&P value = 1.09 X 200 + 537 = 755 !!!!!!
And it can be even worse.
Based on this longer term view, the tendancy of the S&P 500 to over-react to GAAP earnings can be regarded as a buying opportunity. GAAP, especially when modified by adding MtM, tends to run to extremes, the kitchen sink quarters.
Taken together, the two methods suggest a trajectory for the market as well as what information to monitor. So the S&P 500 will probably reach 1,200 within two years, returning 15% annually. Of course that requires NIPA profits to increase moderately quarter to quarter.
On Jul 17 01:01 PM TheFounder wrote:
> Interesting analysis. I am not clear on why are you placing a higher
> weight on the NIPA profit method than the TTM earning method. Based
> on the latter, even if the earnings recover dramatically to $300
> Billion, the corresponding S&P value will be in the mid 800's...and
> even that is optimistic.
>
> Using your derived formula, suppose earnings recover to $200 billion.
> then: S&P value = 1.09 X 200 + 537 = 755 !!!!!!
>
> And it can be even worse.
I think its time to see a type of "shadowstats" on the PE of the S&P500, just as we need to view shadowstats.com to see what CPI would look like using the same Yardstick used through the 80's.
As astounding and unbelievable it is to imagine we would ever have a PE of 100+, that is the PE on the S&P 500 today, about PE 134. Any of who have read, I'll use Sy Harding's "Riding a Bear", or any source of material that goes into Historical Averages, Highs, Lows, etc., would know the 10-15-20 PE range meanings.
Keep in mind that the current 134 PE includes the highly fudged Q1/09 Earnings, assisted by a retroactive Accounting Rule Change, the Mark-to-Makeup numbers used. Once Q2/09 Earnings are in the PE ratio on the S&P 500 will be in the 1,900 range. So I'm sure people are working on another new Yardstick. But for those who want to use what has always been used to determine the PE with the standard Yardstick, you can simply go the official S&P Website. I'll leave the link, and a paste of the PE:
www2.standardandpoors....
[ same link made Tiny ]
tinyurl.com/5g8k8z
S&P 500 Statistics
As of June 30, 2009
Total Market Value ($ Billion) 8,045
Mean Market Value ($ Million) 16,090
Median Market Value ($ Million) 6,532
Weighted Ave. Market Value ($ Million) 68,624
Largest Cos. Market Value ($ Million) 341,141
Smallest Cos. Market Value ($ Million) 643
Median Share Price ($) 27.875
P/E Ratio* 134.01 <<<<<&l...
*Based on As Reported Earnings.
======================...
[ they use TMT instead of TTM, Twelve Months Trailing, instead of the more commonly used Trailing Twelve Months, same thing though. I believe most who read seekingalpha will understand why the media does not mention these stats. It is truly astounding to see ]
Estimate based on Q2/09
www.decisionpoint.com/...
DECISION POINT
OVERVIEW OF MARKET FUNDAMENTALS
Wednesday 7/1/2009
TMT P/E Ratio (GAAP)
2009 Q1 / PE 134.20
Estimate Q2/09 / PE 1923.60
One comment/criticism. You are using twenty years of data for the correlation, and in that period credit became easier and tax policy better for the investor and corporations, with leverage also increasing.
Going forward that will certainly not be the case. If you can find longer history which includes more adverse conditions (credit, taxes, etc) and the correlation still holds - then you have a winner!
My gut says that the more adverse credit and tax situations should slow corporate profits and a market unwillingness to pay the multiples seen during that 20 year period you studied. Hence, the correlation should break down. But thanks for the heads up; I'll pay more attention to NIPA corporate profits in the future.
On Jul 18 05:24 AM Tom Armistead wrote:
> My reason to prefer the NIPA method would be that NIPA has signaled
> the last two major moves 1 year in advance. By watching this you
> could know a move is coming and have a pretty good idea of the magnitude.
>
>
> Based on this longer term view, the tendancy of the S&P 500 to
> over-react to GAAP earnings can be regarded as a buying opportunity.
> GAAP, especially when modified by adding MtM, tends to run to extremes,
> the kitchen sink quarters.
>
> Taken together, the two methods suggest a trajectory for the market
> as well as what information to monitor. So the S&P 500 will probably
> reach 1,200 within two years, returning 15% annually. Of course that
> requires NIPA profits to increase moderately quarter to quarter.
>
Dancing diva, good points, there is very little reason to expect the present adminsration to create a climate as favorable for business as what prevailed previously. On the other hand, interest rates can be expected to stay low, which should push P/E multiples up.
I suspect that there was a loss of confidence that the government (from either administration) was able to control the economy and prevent disaster, also a growing fear of the fallout from excessive financialism, tending to make cash more attractive than any other investment regardless of yield.
For a future article I may experiment with introducing other factors such as prevailing interest rates which are directly relevant to P/E, or other time periods, the 20 years is arbitrary, I used in on natural gas and just got used to it.
untrusting, you have a point on the cash flow. I don't know if S&P aggregates this number to make it easily accessible.
Tom
> Why not use cash flow vs. S&P index and then calulate your regression
> and correlation. Cash flow and S&P index are both objective numbers.
> Corporate profits and NIPA are both subject to bias depending upon
> how the entities choose to "manipulate" the numbers.
operating cash flow? operating cash flow minus maintenance capex? free cash flow?
Are S&P earnings weighted the way their stock prices are weighted?
If not, the PE would be meaningless.
Thanks
www2.standardandpoors....
I thouight of working a discussion of that issue into my article but passed on the grounds it would make the whole discussion unwieldy. But all of this shows that we should not let anyone tell us stocks and expensive or cheap based on the P/E of the S&P because there are so many versions of earnings and then you get the complication of Siegel's position.
That was one reason I liked using the total earnings in dollars rather than the EPS as provided by S&P, then working with ratios rather than a P/E...
On Jul 31 02:07 PM jimmy46 wrote:
> Hello Tom
> Are S&P earnings weighted the way their stock prices are weighted?
>
> If not, the PE would be meaningless.
> Thanks
There are other twists. For one, if you go to the S&P spreadsheet (links are in other comments above), the earnings posted on the page where they compute the P/E ratio are way different (much smaller) than if you look on the previous page of the spreadsheet, which is labeled "ISSUE LEVEL DATA". The adjustment from the raw earnings data to the number they use is based on their "divisor," which (as near as I can figure it out) is an all-purpose adjustment number they use to both weight the index and to make sure the index doesn't change when they drop a company from the 500 and replace it with another. That happens more often than you think (there have been about 32 replacements from Q4 last year to now).
Yet another problem is that when S&P replaces a company, they don't go back and change prior values of earnings from the old company to the new. That is, they don't restate the prior earnings to reflect the new composition of the index. So earnings from prior quarters are frozen once posted, but companies change later on that may have had very different earnings.
I was collecting data from S&P and elsewhere to write an article about the S&P's "true" P/E ratio, but I haven't had the time to complete it. Right now, I think Siegel was right, but I'm not positive.
On top of all that, of course, is how useful is the TTM P/E ratio when you are trying to look forward and make decisions for the future?
Obviously the S&P was too highly valued in the 2000 (1999-2002?) period relative to NIPA profits and that shifts the line upward. You have these dates circled on the graph. What if you ran the correlation again using a dummy variable for the obvious outlyers (including those two below the line in graph 2 as well)? That would shift the regression equation line downward by ?. Just eyeballing it that could reduce the S&P fair value by at least 100 points, perhaps more.
Hopefully you will see this comment and reply. Thank you again for a great article.
Two important factors not included in the analysis are interest rates and unemployment. I spent some time this weekend adding them in, they both have pretty good correlation. Higher interest lowers the S&P, as does higher unemployment. For interest I used Moody's corporate Baaa rate, available from the Fed weekly.
Using values as of 6/30/09, the new analysis suggested S&P 1098. Since then, corporate interest rates have been plunging while unemployment remains at recored highs. New info is due out next week, very interesting to see what comes of it.
The rapid drop in the interest rates has pushed predicted S&P up to over 1,300, suspect in my opinion with unemployment where it is. On the other hand, the very high interest rates during the credit crisis were rapidly reflected in th S&P going down, so maybe low interest rates will have the opposite effect.
What I couldn't get past was the unemployment, there is no history to compare the huge increases to and predict the effect on the indexes. It is over 3 standard deviations from its long term average and once you get into that area these methods and formulas are of limited use, all they can tell you is you are in uncharted territory.
At this point I am investing on the basis that the S&P will make 1,200 by December next year, guessing that there will be a sharp correction somewhere between now and then. I will be more optimistic if unemployment decreases rapidly, or if Washington extends benefits long enough to mitigate the shock to the economy.
Thanks for you interest in my article.
On Aug 30 10:11 AM dancing diva wrote:
> I calculated the fair value of the S&P using your formula and
> the latest release of NIPA profits and came up with an S&P level
> of 1165. However, I noticed something rereading the article that
> I missed earlier. It related to your second graph and the regression.
>
>
> Obviously the S&P was too highly valued in the 2000 (1999-2002?)
> period relative to NIPA profits and that shifts the line upward.
> You have these dates circled on the graph. What if you ran the correlation
> again using a dummy variable for the obvious outlyers (including
> those two below the line in graph 2 as well)? That would shift the
> regression equation line downward by ?. Just eyeballing it that could
> reduce the S&P fair value by at least 100 points, perhaps more.
>
>
> Hopefully you will see this comment and reply. Thank you again for
> a great article.
Thanks for your reply.
While I would normally agree lower interest rates would significantly boost the S&P, as they have done historically, I really do think this time it's different. That's because banks are still not lending freely and credit is tight, despite the low interest rates. In the past the low level of interest rates ( which implies easy credit conditions) are bullish for the stock market because it allows busineses to expand freely. That's certainly not the case this time.
The reason I suggested the line should be shifted downward (and a dummy variable used for the clearly outlying data points such as near the 2000 stock market peak) that at least in the next 10-20 years it's doubtful that p/e's will get that stupidly high. Those data points shift the regression line up considerably.
Using the unemployment rates makes sense to me. Typically p/e's are highest during trough earnings, and trough earnings correlate well (with some time adjustment) to the highest rates of unemployment.