Monday Morning Musings: A Valuation Review

Analyst: Tobias M. Levkovich

Institutional Equity Strategy                                                  

Monday Morning Musings: A Valuation Review                                     

November 22, 2002         SUMMARY                                              

* The valuation issue has come back to the fore      

Tobias M. Levkovich         given the market rally of the past seven weeks.    

* Whereas the "Fed Model" has some flaws, we would   

argue that our earnings yield gap analyses have    

Evrard Fraise               proved quite adept at predicting inflection points 

in the market.                                     

* All of our valuation models indicate that equities 

remain relatively inexpensive versus bonds.        

* Bears take issue with operating EPS, but we would  

point out that operating EPS have a more than      

30-year relationship with business activity.       

* Technology stocks have led the way in the recent   

rally whereas Consumer Staples stocks have been the

weakest performers.                                

* We continue to think our 1,000 target for the S&P  

500 is achievable and can even be exceeded should  

                   panic buying set in.                               

OPINION

All referenced figures can only be seen in the PDF version of this document

available on GEO and FC Linx

Given the powerful rally of the past seven weeks, we expect the bear crowd to

restart its chant of an over-valued equity market, citing above average P/E

ratios and a need to refocus on reported earnings, not to mention the new

scourge of accounting, pension income.  Yet, we believe such an assault should

fall on deaf ears.  We would agree that some of the high beta stock price surges

have little, if any, fundamental support, and thus, may reverse in the next few

months.  However, that does not mean that one cannot be a tactical bull taking

advantage of trading rallies.  Moreover, those who choose to emphasize reported

earnings so as to demonstrate higher P/E ratios (and thereby support the bearish

view) should at least fess up to the fact that reported earnings are climbing

nearly 75% year over year and, thus, in a quarter or two, the reported P/E ratio

will be very similar to the operating earnings-related P/E.

Misunderstanding Valuation

While one can argue that the current dividend yield is too low relative to its

1982 or 1974 levels, we would strenuously point out that interest rates are also

much lower.  In addition, higher P/E ratios also reflect lower income

opportunities from risk-free fixed income securities vs. prior periods when

interest rates were much higher.  Therefore, it is not the absolute level of

valuation that must be considered, but rather one has to review alternative

investment returns as well.

Hence, the so-called "Fed Model" has received a great deal of attention as

people grapple with market valuations amidst the recent equity rally.  As most

of our readers are well aware, the "Fed Model" compares the 10-year Treasury

yield with an earnings yield, defined as the inverted P/E of the market, in

order to determine if equities are overvalued or undervalued relative to bonds.

Accordingly, using the current 10 Year Treasury yield of roughly 4.15% and an

S&P 500 P/E of 19.1x (trailing 4Q operating EPS), the Fed Model would suggest

that equities still are about 20% undervalued compared with their Treasury

counterpart (T Yield/Inverse PE -- 1 or 4.15% / 5.2% -1).

However, we would concede that there are some flaws with the Fed model and that

a more appropriate analysis would be to look at the measure in relation to a

five-year rolling average, similar to what we do in our earnings yield gap

analysis.  In addition, we continue to think that investors are best served by

forming valuation opinions after reviewing various methodologies.  Consequently,

we also have been running our earnings yield gap analysis versus other classes

of bond yields for some time as a means to better define the Price (or

valuation) component of our PULSE market analysis framework.

The most significant flaw with the Fed Model in our view is that it is rarely

correct in predicting inflection points.  Actually, as we have pointed in the

past, the fed model would have indicated that stocks were overvalued for 189 out

of the 204 months from 1985 through 2001 (see Figure 1).  However, given that

the markets were up tremendously during most of that period, it hardly would

have been wise to follow the model too closely.  Thus, it is more important to

look at the relationship in terms of where you are versus the past averages.  In

our work, we look at the relationship in relation to five-year rolling mean

earnings yield gap and, in particular, we focus on the extreme positions.

Hence, when our earnings yield gap studies indicate that we are more than two

standard deviations from the five-year mean (which they do currently), we know

that we are in territory that only occurs 4.5% of the time as there is a 95.5%

chance of being within two standard deviations of that mean.  Therefore, either

the bond yield has to go up or the market P/E needs to increase in order to

correct the extreme relationship.

Figure 1:

Source:  Salomon Smith Barney

Moreover, unlike the traditional Fed Model, our earnings yield gap analyses have

been very good historically at predicting inflection points in the market.  As

Figure 2 indicates, the S&P 500 has been up an average of 20.3% after our model

has been more than two standard deviations below the mean and the market has

been down an average of 6.7% when the model has been more than two standard

deviations above the mean.  In addition we would point out that of the 19 times

that the model has been more than two standard deviations below the mean, only

two of those periods have resulted in negative market performance 12 months

later.

Figure 2:

Source:  Salomon Smith Barney

However, many people have raised concerns that the 10-year Treasury note is no

longer as valuable in valuation work for three key reasons:

*   There is an increased demand for 10-Year Treasuries ever since the U.S.

government ceased issuing 30-Year bonds;

*   Equity investors plowed into "risk free" Treasuries due to all the

corporate malfeasance issues, and;

*   Corporate bond investors got "burned" buying into what were considered

cheap utility and telecom issues and then reverted back to seemingly less

risky Treasury notes.

These factors created a lot of buying pressure on the 10 Year yield making it

less desirable for comparison purposes.  We did not include the "convexity

trade" factor in the various Treasury pressures since it appears to have eased

back of late.  Nonetheless, more than a year ago, we started publishing our

earnings yield gap model with Moody's Aaa and Baa yields in place of the 10-

year.  Nonetheless, as Figures 3, 4 and 5 suggest, we are still more than two

standard deviations from the mean unless one looks at Baa yields (Figure 5), but

we are a bit reluctant to use the investment grade index for our valuation

purposes since it tends to confuse market and corporate risk.  Indeed,

investment grade spreads and the stock market have become strongly correlated

and thus one is mixing related data to arrive at what should be independently

derived conclusions.

Figure 3

Source: Salomon Smith Barney

Figure 4

Source: Salomon Smith Barney

Figure 5

Source: Salomon Smith Barney

One big difference between the 10-year yield study and the ones that employ

corporate bond yields is that that the latter did not signal a bottom in

September 2001. Actually, we would argue that the aforementioned pressures that

were placed on the 10 Year Treasury yield accounted for the majority of the

difference during that period in 2001.  However, all three models have been

suggesting for over a month that equities should be bought relative to their

bond counterparts.  In fact, at one point, every model was more than three

standard deviations from the mean.

Given the extreme levels of the models, we thought it would be interesting to

stress test them to see what would have to occur to either yields or equity

market levels in order for the models to revert to three standard deviations

below the mean or increase back toward neutral territory.  As can be seen in

Figure 6, the results are remarkable.  Even in the Baa model, the most

conservative of the three, equities would need to increase roughly 40% in order

for equities to reach parity with its historical average relationship if yields

were to remain constant.  Additionally, if we hold S&P 500 levels constant, Baa

yields would have to increase to 9.1% in order for parity to be achieved.   We

agree that such a scenario would not occur, as yields would not stay constant if

equities were to increase in such an aggressive nature and equities would not

stay constant if yields were to increase so dramatically.  However, it is still

very useful to measure what it would take in order for certain scenarios to

transpire, in our opinion.

Figure 6

Source:  Salomon Smith Barney

Another earnings yield gap model that we find useful from time to time employs a

median P/E multiple (see Figure 7).  The median multiple allows us to look at an

earnings yield that is not heavily biased by any given sector weight at any

given time as we are always looking at the middle of the rung company.  Hence,

one could argue that issues that can affect a weighted average multiple such as

the tech boom or corporate malfeasance would have less impact on the median

company of such a large basket of stocks.  Nevertheless, this model also

suggests that we are more than two standard deviations below the five-year mean.

Figure 7:

Source: Salomon Smith Barney

In addition we have recently introduced two new renditions of the earnings yield

gap model.  The first, is a model in which the P/E of the market is calculated

using trend line EPS (since 1905) and the second is a model in which we use real

10 Year Treasury yields (Yield less inflation) to see the effect on the

relationship.  Both provide interesting results.

We would use trend EPS in an earnings yield gap analysis (see Figure 8) in order

to distance ourselves from the cyclical nature of quarterly operating EPS,

thereby placing a greater emphasis on the bond yield and equity market price

over time as we are assuming EPS grow at the constant rate that they have grown

over the last 100 years.  The result is that the earnings yield gap analysis has

fallen below two standard deviations from the mean for the first time since the

1990-91 recession.

Figure 8:

Source:  Salomon Smith Barney

In Figure 9 we attempted to examine the earnings yield gap using a real yield in

the analysis, since some investors have argued that we are comparing nominal

trends in earnings against a more constant real Treasury yield over time.  While

we might be willing to debate the concept, even using a real yield gap analysis,

we still find the gap at more than two standard deviations below the mean.

Hence, even on this metric, stocks look cheap vs. bonds.

Figure 9:

Source:  Salomon Smith Barney

Reported Versus Operating EPS

Since all of our work does tie to operating earnings numbers, we are certain

that bears will revert to using reported numbers to make their point on

valuation.  Yet, as can be seen below, operating earnings have a more than 30-

year relationship with business activity (Figure 10), which is hard to deny.

Moreover, the simple fact that a few companies can take one-time goodwill

write-offs and basically wipe out the clean earnings of the bulk of the S&P 500

does not mean that the aggregated reported numbers make a whole lot of sense

either, in our opinion.

Figure 10:

Source:  Federal Reserve and Salomon Smith Barney

Finally, we thought a review of the rally thus far would be appropriate.  In

this vein, we thought it would be useful to look at the major indices, various

growth and value indices as well as the 10 S&P sectors to see which groups have

performed the best since October 9t h .  As one can see in Figure 11, the Nasdaq

has outperformed its nearest major index by more than 10% since the rally began.

This is largely due to the large tech presence within the index as tech is also

the strongest performing sector within the S&P.  We would note that the only

sector that is down since October 9t h  is Consumer Staples, a sector we

downgraded 10 days ago due to pricing concerns.  While it is not shocking that

the most beaten-up sectors have rallied the hardest and could move higher as

"beta bets" continue to be placed (given that there are only five weeks left in

the trading year), we do harbor doubts over the underlying fundamentals.

Nevertheless, upward momentum and trading volumes suggest that the current

rally still has legs based on the aforementioned valuation work.  Furthermore,

our proprietary sentiment work clearly does not show any reason to back away

yet from equities, although sentiment has gone neutral in our "Other PE" index.

Thus, we think that we can see our 1000 target for the S&P 500 exceeded as a

panic buy can set in.  Many money managers seem to be trailing their benchmark

indices and may try to swing for the fences with just over a month left in the

year.

Figure 11:

Through November 21, 2002

Source:  FactSet and Salomon Smith Barney

Ogni lettore deve considerarsi responsabile per i rischi dei propri investimenti e per l’uso che fa delle informazioni contenute in queste pagine. Lo studio che propongo ha come unico scopo quello di fornire informazioni. Non e’ quindi un’offerta o un invito a comprare o a vendere titoli. Ogni decisione di investimento/disinvestimento è di esclusiva competenza dell'investitore che riceve i consigli e le raccomandazioni, il quale può decidere di darvi o meno esecuzione.

The information contained herein, including any expression of opinion, has been obtained from, or is based upon, sources believed by us to be reliable, but is not guaranteed as to accuracy or completeness. This is not intended to be an offer to buy or sell or a solicitation of an offer to buy or sell, the securities or commodities, if any, referred to herein. There is risk of loss in all trading.