Click here for a link to ORIGINAL article:
By John P. Hussman, Ph.D. | 2 August 2010
All rights reserved and actively enforced.
| It is impossible to properly estimate long-term cash flows based on a single year of earnings, regardless of whether one uses actual net earnings or projected operating earnings. It is impossible to properly value the stock market based on a single year of earnings, regardless of whether one uses actual net earnings or projected operating earnings. |
Writing each of these sentences only once is woefully inadequate. If I had my way, investors would have to write them over and over five days a week. Wall Street analysts would have to write them a hundred times a day, immediately upon arriving to work.
In recent weeks, I've seen "valuation" arguments that literally treat future estimated operating earnings as if they are a pure, immediately distributable dividend that will grow indefinitely without the need for capital investment, while sustaining current record profit margins forever. I've heard analysts say, with a straight face, that stocks are 'cheaper here' than they were at the 2009 lows, because the ratio of the S&P 500 to the current forward operating earnings estimate is lower today than it was 16 months ago. I've seen analysts presume to "capitalize" earnings into some sort of market valuation by doing nothing more than dividing estimated operating earnings by corporate bond yields that are presently nearly indistinguishable from Treasury yields.
The primary question investors need to ask is whether these analysts have actually examined the historical record of these approaches— not just whether they have an anecdote about some extreme such as 2000 or 1987— but whether they have done a robust, long-term evaluation. Unless a "valuation" methodology is accompanied by long-term, decade-by-decade evidence showing that the valuation method is actually correlated with realized, subsequent market returns (particularly over a horizon of say, 7-10 years), then you are not looking at the sound valuation work of an investment professional. You are either looking at a random guess or a sales pitch.
Don't get me wrong. There are many thoughtful, well-disciplined financial planners and asset managers— usually far away from Wall Street— who are excellent stewards of their customers' investments. My difficulty is not with those professionals, but with the careless and inept reasoning that passes for analysis hour after hour on the financial news. If you take away one thing from this week's comment, it is that stocks are a claim to a long-term stream of cash flows that will actually be distributed to investors over time, and that this stream of cash flows cannot be estimated from a single year's earnings number.
The main reason for this is that profit margins vary from year-to-year over the business cycle, and tend to mean-revert over the long-term. Earnings (net and operating) tend to be depressed during periods of economic strain, but when they reflect compressed profit margins, they are strongly associated with above-average rates of subsequent growth over the following 7-10 years. In contrast, earnings that reflect elevated profit margins are strongly associated with poor rates of subsequent growth.
When analysts take earnings figures at face value, and presume to "capitalize" them simply by dividing by interest rates, they demonstrate a Kindergartener's grasp of securities valuation. Case in point is the treatment of forward operating earnings. The first problem is that analysts tend to treat these as if they are distributable cash flows. Unfortunately, operating earnings exclude a whole range of charges that may not occur on an annual basis, but are legitimate costs and losses incurred as part of the ordinary course of business.
Meanwhile, operating earnings often include a benefit from those very same "extraordinary" sources— provided they make positive contributions (witness the large boost to the operating earnings of major banks this quarter, resulting from the reduction in reserves for future loan losses [[potentially simply postponing those losses to a future reckoning: normxxx]]). Forward operating earnings take these hypothetical earnings to the next level, and are based on the year-ahead forecasts of Wall Street analysts. As long-term readers of these comments know, I am terribly concerned about the increasingly careless use of operating earnings as a measure of stock valuation, because I have yet to see an operating earnings model that is not ignorant, devious, misleading, lacking in historical evidence, repeatedly catastrophic, or all of the above. Not least of these concerns is that the commonly quoted "norm" of 15 for the P/E ratio properly applies to the ratio of the S&P 500 to trailing 12-month net earnings, which are invariably much lower than forward operating earnings.
Operating earnings are not even defined under Generally Accepted Accounting Principles (GAAP). They were spawned by Wall Street in the early 1980's, so there is (conveniently) no long-term history for this measure, meaning that the valuation bubble between the late 1990's and 2007 represents a significant chunk of the observable record. Still, the increasingly common use of this earnings measure requires us to somehow deal with it constructively. As it turns out, the lack of history prior to 1980 is not particularly difficult to overcome.
We can very accurately explain the relationship between forward operating earnings and standard earnings measures using variables that have been observable throughout history. We can form good estimates prior to 1980 on that basis (see August 20, 2007 Long Term Evidence on the Fed Model and Forward Operating P/E Ratios). It is then straightforward to calculate objects such as the Fed Model (the ratio of the forward operating earnings yield to 10-year Treasury yields), and to demonstrate that it has zero correlation with subsequent market returns.
The question then becomes— is there any way that forward operating earnings (FOE) can be employed as a useful measure of market valuation? The answer is actually yes.
Valuing The S&P 500 Using Forward Operating Earnings
I should emphasize from the outset that the proper valuation of a stream of future cash flows requires one actually to model the stream of cash flows earned, delivered as dividends, reinvested, and obtained as terminal, liquidation or buyout payments. Methods that rely on multiples such as P/E, price/revenue, price/dividend and so forth are approximations at best. They perform better if the fundamental metric being used is relatively smooth, and varying rates of growth are explicitly taken into account.
That said, discounted cash flow models can always be mapped to "multiple-based" models. The issue there is that you'd better know what assumptions are baked in, and whether they're reasonable. For example, in an ideal world, a stock that earns E, pays a proportion d of that out in dividends, reinvests the rest to grow at a perfectly constant rate g, and is expected to stay in business into the indefinite future, should have a P/E ratio of d/(k-g) where k is the desired long term rate of return (say 0.10 or 10%) that the stock should be priced to deliver. For more on discounted dividends, see the September 12, 2005 comment, The S&P 500 as a Stream of Payments.
Unfortunately, all of this stuff gets abused. I remember in the market bubble of the late 1990's and into 2000, a bunch of Wall Street analysts were saying that stocks were still cheap based on the 'dividend discount model'. The problem was that they were assuming growth rates (g) of 10% or higher, when the peak-to-peak growth of S&P earnings and dividends had never exceeded 6-7% over periods of a decade or more. The Dow 36,000 guys basically tried to justify a P/E of 100 for the Dow by assuming that earnings were dividends, and then picking a "g" that was so close to "k" that the denominator of the above model was 0.01, or 1%. I wondered why they didn't go all the way and set k=g so they could publish "Dow Infinity."
But I digress.
The two main failures of standard FOE analysis are that 1) analysts assume a long-term norm for the P/E ratio that properly applies to trailing net, not forward operating earnings, and; 2) analysts fail to model the variation in prospective earnings growth induced by changes in the level of profit margins, and therefore wildly over- or underestimate long-term cash flows that are relevant to proper valuation. By dealing directly with those two issues, we can obtain useful implications about market valuation. As I have frequently noted, it is not theory, but simple algebra, that the long-term annual total return for the S&P 500 over any horizon T can be written as:
Long term total return = ((1+g)(future PE / current PE)^(1/T)- 1) + (dividend yield(current PE / future PE + 1) / 2)
The first term is just the annualized capital gain, while the second term reasonably approximates the average dividend yield over the holding period. For the future P/E, one can apply a variety of historically observed P/E ratios in order to obtain a range of reasonable projections, but the most likely outcome turns out to be somewhere between the historical mean and median. You have to get two things right: the "normal" future P/E and the prospective long-term earnings growth rate g. 'Standard' FOE analysis misses on both counts.
Very simply, looking out over a 7-10 year horizon, the proper historical norm for price-to-forward operating earnings is approximately 12.7. Moreover, one cannot simply apply the long-term operating earnings growth rate of 6.3% (0.063) as an unchanging measure of g. Rather, an accurate growth rate for the model has to reflect the level of profit margins at any point in time, since the current P/E multiple may reflect either depressed or elevated earnings. For a 10-year investment horizon, the proper value of g should take into account the gradual normalization of margins. Historically, the best estimate is approximately:
g = 1.063 x (0.072 / (FOE/S&P 500 Revenues))^(1/10) - 1
The chart below presents the historical projections obtained using this analysis, along with the actual 10-year total returns achieved by the S&P 500.
Currently, the forward operating earnings model above suggests an average annual 10-year total return for the S&P 500 of 5.5%, while indicating that the S&P 500 was briefly and moderately undervalued at the 2009 market low. Not surprisingly, the 1988-1991 model projections underestimated the market's subsequent 10-year total return, as the period a decade later— between 1998 and 2001— represented the extremes of the subsequent valuation bubble. Overall, the correspondence between projected and actual 10-year market returns is very close. Taken together, my impression is that this is a record that a reasonable and informative valuation model ought to produce.
Last week, I noted that our standard model (which is based on peak-to-peak earnings) suggested a somewhat higher 10-year total return near 6.7% annually. The difference between that model and the forward operating earnings model above is that the peak-to-peak model calculates a growth rate based on the position of earnings within a long-term growth channel, and does not formally take profit margins into account. The chart below provides a good range of what our most reliable valuation models project for S&P 500 total returns over the coming decade, and presents the complete post-war history.
Though each of our metrics is slightly different, all concur that the market was moderately, but not historically undervalued, at the 2009 low, which briefly approached the level of valuation that was observed at the 1970 low, but was nowhere close to the valuations seen at points such as 1950, 1974 and 1982. I clearly underestimated the willingness of investors to drive stocks back to strenuous overvaluation so quickly. Earlier this year, the market was more overvalued than at any point prior to the late-1990's bubble, and is currently near the same level of overvaluation as the 1972 and 1987 market peaks. At present, all concur that the S&P 500 is most likely priced to deliver a 10-year total return of roughly 6%, albeit with the likelihood of significant interim volatility.
Stocks are emphatically not cheap on a historical basis. Analysts who encourage investors to take a different view on valuations should at least be expected to present similarly extensive historical evidence supporting those perspectives.
| [Important Notes— A spreadsheet including forward operating earnings estimates (bottom-up are most common) and the most recent four quarters of index revenues can be downloaded at no charge by registering with www.standardandpoors.com. To compute the current 12-month FOE from 2010 and 2011 estimates, take the weighted average. For July (month 7), the 2010 estimate was 82.15, and the 2011 estimate was 94.60, producing a weighted average of (12-7)/12 * 82.15 + 7/12 * 94.6 = 89.41, with 12-month trailing revenues of 918.98 for the index. The calculation of g in the model above uses trailing 12-month revenues and is calibrated on that basis. The implied 10-year rate of total return can be compared with potential risks and alternative rates of return as one chooses. Of course, it's up to individual investors to decide what level of potential return is acceptable, but keep in mind that historically, investors have not generally tolerated low or mid single-digit implied equity returns for long, so lower projected returns are also associated with dramatically higher risk of intermediate-term loss. Investors who believe that a 10-year expected return in stocks of about 6% is an outstanding prospect relative to historical and probable market volatility, and believe that other investors, in aggregate, will embrace that same belief indefinitely, are welcome to classify stocks as "fairly valued" here. We are not in that category.] |
One final note to our colleagues in finance— we read a lot of research here— articles, blogs, news clips, subscription services, academic papers, and so forth. You'll see comments and links all over our research referencing various sources. We're happy to see our own work picked up, and are particularly pleased when professors use it in their classes. But come on, guys— when you use someone's content, or agree with someone's point, acknowledge the person and then add your own work. Don't just lift models or material.
Bill and I do our best to always include attribution when we use other people's work (send us a note if we miss one!)— except that when we criticize someone's approach, we may omit names unless the person appears to be damaging or misleading investors. I know this is the age of the internet, but common courtesy still applies. None of us is 13 anymore.
Economic Risks Continue
Last week, new claims for unemployment remained within their recent range at 457,000, GDP growth came in at a disappointing 2.4% for the second quarter, nearly half of which represented inventory accumulation, the ECRI Weekly Leading Index deteriorated [still further] to a -10.9% growth rate, and the markets were cheered somewhat by a Chicago Purchasing Managers Index above 60. Clearly, there are cross-currents to the data here. For our part, we are closely focused on the leading components of the data, rather than measures that provide less timely information.
Our own Recession Warning Composite is a primary basis for concern. Notably, once a recession signal emerges, the composite often tips back and forth due to variations among individual components. Those signals tend to become solid as an economic downturn takes hold, but in any case, just one signal is sufficient. With other leading measures also clearly deteriorating, the primary fact that would change our views would be uniform improvement in those leading measures, not simply a rally in the S&P 500 sufficient to take it a bit over its level 6 months ago.
The ECRI Weekly Leading Index continues to receive attention, and in our view, for good reason. As I've noted before, the WLI growth rate is tightly correlated with the Purchasing Managers Index, with the strongest correlation being at a lead-time of 13 weeks. Notably, the WLI growth rate 13 weeks ago was still a positive 12.7, so on that basis, we're still not quite at the point where we would expect to see a wholesale deterioration in the PMI. A 12.7% read on the WLI growth rate is consistent with a PMI just over 60, so the Chicago PMI number was actually right in the pocket, and does not contradict expectations of softer economic activity, at least not yet.
The national PMI indices come out this week, but again, the July figures may be a mixed bag. In contrast, one would expect the August and September figures to show material weakness. To the extent that we do not observe that, and particularly if other leading measures improve, our concern about probable economic weakness will abate. We're not at that point at present.
New claims for unemployment will also be important early evidence of economic conditions over the next few months. Below, I've inverted the ECRI Weekly Leading Index, plotting it against the amount by which weekly new claims for unemployment (4-week average) exceed their 5-year norm. While WLI downturns (blue spikes on the chart) are not always associated with a spike in new claims (see 1987, when the weak WLI was driven almost exclusively due to stock price weakness), the leading tendency of the WLI is strong enough that we should not ignore the potential for increased layoffs.
Given that the 5-year norm for new unemployment claims is about 400,000 weekly, it's possible we could be looking at new claims well above 500,000 weekly by September. Again, the relationship is not tight enough to form the basis for strong predictions, but it is clear that investors should not easily discard caution on the basis of one positive economic figure or another. Importantly, the WLI growth rate was 14.9% 23 weeks ago, so we are still in the period where the recent deterioration in leading indicators may not be confirmed by coincident data.
With regard to mortgage losses, the housing market remains in a fascinating period of "extend and pretend," where it is clear that mortgage conditions are worsening, but we have not yet seen an explosion in foreclosures, or a movement of the growing inventory of foreclosed homes onto the open market. The main exception is Fannie Mae and Freddie Mac, which have been accelerating foreclosures in recent months, accompanied by a regular influx of funds from the U.S. Treasury (which already exceeds $145 billion) to bail out losses on what would otherwise be insolvent GSE debt. The Lender Processing Services (LPS) June Mortgage Monitor provided the most recent report last week, noting that
|
As James Saccacio, the CEO of RealtyTrac observed a week ago,
|
On Sunday's edition of Meet the Press, Alan Greenspan remarked that there is a huge number of homes that would go "underwater" if home prices were to slip by another 5-7%, and that such an event could potentially trigger a large wave of additional foreclosures. I am not familiar enough with the price distribution of existing mortgages to contribute much insight here, but the remark does help to explain why banks appear so reluctant to bring the massive inventory of delinquent and foreclosed properties onto the market. It is not clear precisely at what point the burgeoning inventory of foreclosures and delinquent mortgages will impact the markets, but it is clear that conditions are not improving, and that fresh economic weakness would tend to destabilize an already fragile situation.
At the same time, it is not out of the question that we may be quietly allowing U.S. banks to go insolvent without disclosure, [[we've done it in the past: normxxx]] covering the losses over time out of wide interest spreads on existing loans, and that we may be able to avoid outward evidence of mortgage deterioration simply by allowing the Treasury to go further and further into deficit on behalf of the GSEs. Undoubtedly, all of this will produce future strains in the form of inflation risk, longer-term commodity price pressures, fiscal instability, stagnant lending activity, continued failure of smaller institutions, further loan writedowns, and other events. The losses are inevitable, and to some extent even quantifiable.
The real question is who will actually bear those losses and when. The official policy is clearly for the public to do so, massively, preferably quietly, and over a very long period of time. Still, my impression is that fresh economic weakness could prove to be a tipping point, and that both investors and the public should understand that they are likely to pay terribly for the current abundance of apparently 'free' lunches.
Market Climate
As of last week, the Market Climate for stocks was mixed— valuations remain unfavorable, technical action was mixed but tenuous, with various indices flirting with widely observed levels of support and resistance (e.g. the 1100 level on the S&P 500), while leading measures of economic activity remain decidedly unfavorable. As I noted last week, the average historical outcome of similar combinations has been negative, largely because the deterioration in economic measures [eventually] tends to trump technical action even when it has been more favorable than it is at present. Still, there is a clear speculative element in day-to-day market action here, as trend-following investors remain heavily focused on very specific price levels, which can trigger short-term bursts of buying and selling pressure.
Overall, my impression is that the near-term dynamics of the market are likely to be dominated by this sort of speculative trend following activity— primarily because it will probably still take another 4-8 weeks until sensitive coincident economic measures (such as ISM figures and new claims for unemployment) begin to predictably reflect the deterioration we've seen in various composites of leading indicators. Again, if we do not [actually] observe that deterioration, and particularly if the leading measures broadly improve, our concerns about the economy would tend to abate. For now, we remain defensive.
In bonds, the Market Climate remained characterized last week by moderately unfavorable yield levels and favorable yield pressures. The Strategic Total Return Fund continues to carry a duration of just under 4 years, largely in intermediate-term Treasury securities. I am not at all convinced that Treasuries with a maturity much past 5 years will provide adequate yields to maturity, or even positive real returns, as we move through this decade.
But inflation concerns are clearly a longer-term issue. We are likely to observe strikingly larger budget deficits ahead for a while, but at the same time, the eagerness of investors to hold default-free paper is likely to be quite strong. In that environment, you can print a lot of government liabilities with seemingly no consequence.
It's in the back half of this decade where the eagerness to hold those liabilities will probably abate, with unfortunately no ability to reduce the supply since the dough will already have been spent. That will likely push us into an inflationary experience much like the 1970's, which followed the rapid increase in government spending and the move to persistent deficits after the introduction of the Great Society programs of the late-1960's.
M O R E. . .
No comments:
Post a Comment