Equity Markets - Not Crazy, Just Misunderstood

Portfolio Research
Equity Markets – Not Crazy, Just Misunderstood 31 August 2006 1

Analyst
Peter Rice
Why Value is (nearly) all that matters, if you define it correctly

• Ever since the 1929 global stock market crash, the basis of valuing stocks has been dominated by the “corporate value =
equity value” principle.
• However, there is a fundamental difference in value between listed and unlisted companies: short-term realisable capital
gain (or loss). Unfortunately, no corporate valuation methodology can allow for that, by definition. As such, it becomes
necessary to redefine “value” in a traded market sense, if any expectation of capital gain or loss is to be taken into
consideration.
• By developing a valuation process specifically catering to the nature of traded stocks, we are able to show that markets
are far more value-driven than ever thought possible, with between 80% and 90% of Index price changes over the past 20
years explained for the ASX200, FTSE, S&P500 and HSIX.
• Contrary to popular belief, our work suggests that the same basic principle of value is applied consistently across
markets, sectors and stocks over time.
• The implications of this work are substantial:
• It means that markets are far more consistently predictable than ever thought possible.
• Valuation and market analysis is a far less subjective and volatile process than generally accepted.
• Dividends have a far greater role in determining market prices than generally thought, according to generally
accepted finance theory.
The Basic Principle

The value of all financial assets (whether they be stocks, bonds or property) is based on a simple principle; i.e.
Value = NPV of Expected Future Returns.
In the case of traded stocks, there are two ways to receive returns: either as:
Cash dividends; or
Capital gains.
The concept of stocks being worth the NPV of expected dividends is nothing new. The dividend discount model (DDM) was the dominant
academic principle of equity valuation from about the mid 18th century, until fairly recently.
The only problem is, it doesn’t work, if we define “work” as explaining both a large proportion of variation in market pricing, and absolute
levels of value at the same time.
If the only forms of returns available are dividends and capital gains, and dividends alone don’t explain market pricing, then, by definition, the
only possible additional contributor to price can only be some form of capital gain or loss expectation.
Again, another “basic principle” comes into play here:
Markets can only price what they can reasonably predict.
In other words, for markets to price in capital gains, there needs to be a reasonable basis for estimating those capital gains. If the market is
unable to establish a reasonable basis for estimating expected capital gain, then the expectation would be that it would price it at zero.
Adelaide Airlie Beach Geraldton Gold Coast Maroochydore Melbourne Minyama Mt. Waverley Perth Sydney
08 8407 5700 07 4946 5080 08 9964 3800 07 5631 2300 07 5409 6100 03 9242 4000 07 5478 1681 03 9831 5000 08 9268 4888 02 9247 8666
Tolhurst Noall Ltd ABN 52 003 237 536 A Participant of ASX Group. Australian Financial Service License No 238444
This information must be read in conjunction with the Analyst Certification and other important disclosures at the end of this document
© Tolhurst Noall Ltd.
Portfolio Research

Analyst
Peter Rice
Over the past 50 years, as traded equity markets have become more and more developed, the tendency for capital gains to be achieved
have been well established, with average excess returns over assumed risk-free interest rates, the established principle for estimating
market risk premiums. Given that fact, it is reasonable to expect that stock prices do reflect expected capital gains.
DCF: An Example of why Corporate Valuations Don’t Work For Traded Stocks
As soon as you get to this point in the analysis of markets, the problems with the use of corporate valuation techniques start to become
obvious. As an example of the issues, it is useful to look at discounted cash flows as a way of valuing stocks, for the simple reason that it is
the most widely used, and generally considered the most sophisticated, of corporate valuation approaches applied to stocks.
DCF is based on the principle that a company is worth the NPV of expected cash flows. This is not too far away from the basis described
above, except that it is based on the overall returns to the company, rather than to the shareholder.
The two approaches could, potentially, line up under certain circumstances. If it is assumed that cash can only have two effective uses in the
year in which it is generated, i.e. the payment of cash dividends, and/or an increase in the book value of net assets, then DCF would be
effective if market prices were equal to the net book value of the firm. In that case,
Capital appreciation = retained cash.
Unfortunately, that does not apply to any major equity market, with all major markets trading on price/book multiples of over 3 times. In
fact, the principle of stocks being worth their net asset backing had been effectively abandoned as a principle well before the 1920’s boom in
the US.
And yet, DCF remains firmly in the forefront of analyst tools in assessing stock value.
One of the other major issues with corporate valuation approaches is that different valuations apply under different conditions. DCF, for
example, only applies when buying a controlling stake in the company (and therefore, having the ability to access retained cash).
By contrast, except in very unusual cases, the valuation of traded stocks is based on an assumption of the purchase of a minority stake.
So, How Do Markets Look To Predict Capital Gain?

A number of recently published academic papers have presented evidence suggesting that capital appreciation rates of traded stocks are
closely linked to reported EPS growth. These include Penman and Sougiannis (2002), Gentry et al. (1998), Subramanyam and
Venkatachalam (2001), Dechow (1994) and Sloan (1996). In the first two articles, the relationship is found to be considerably closer than
with either dividend or cash flow growth rates (again, presenting evidence against both the DDM and DCF approaches).
These findings should not be a surprise to market watchers. Approximately 80% of all written equity research focuses on earnings growth
forecasts. Stock commentaries also largely focus on earnings prospects (particularly, in the reporting season), suggesting that corporate
earnings is a critical focus of markets. In addition, P/E multiples remain the dominant benchmarks for assessing value in all major markets,
suggesting that with all other things equal, capital gain will be a function of EPS growth.
However, discounting EPS growth alone doesn’t “work” either. In reality, under GAAP rules, reported EPS and net operating cash flows are
relatively similar over time (particularly, at a market aggregate level). In most markets, the variation is in the +/-10% range.
As such, using an EPS discount model has almost identical difficulties with explaining market pricing as does a DCF-based approach.
Ironically, the work of all four authors presents what is a major indicator of the answer. If capital gain rates are closely linked to EPS
growth and stocks can produce returns in only two forms (i.e. capital gains or dividends), then value is the NPV of expected EPS PLUS
expected dividends.
BUT, this presents two major problems:
1. Dividends are generally assumed to be paid out of earnings, so by discounting streams of both expected EPS plus expected DPS,
there is an obvious double counting issue; and
08 8407 5700 07 4946 5080 08 9964 3800 07 5631 2300 07 5409 6100 03 9242 4000 07 5478 1681 03 9831 5000 08 9268 4888 02 9247 8666
Portfolio Research

Analyst
Peter Rice
2. According to Modigliani and Miller, and their Nobel Prize winning work published in 1961, dividend payout policy should be
irrelevant to market pricing. However, if this approach works, then there must be an obvious link between dividend payout policy
and market pricing.
The first question is, does this structure “work”? The second issue is, if it “works”, how do we explain it?
DPS plus EPS Explain Market Pricing

The first step in the process is to look at if EPS growth, dividend payout ratios and interest rates (as part of a discount rate) explain a
significant portion of market P/E multiples. Our analysis of the ASX200 Index month end historical P/E since 1992 produces the following
regression results:
Regression Stats
Multiple R 0.8455
R Square 0.7648
Adj. R Sq. 0.7460
Std Error 1.1268
Obs. 135
ANOVA
df SS MS F Signif. F
Regression 4 413.760 103.440 81.462 0.000
Residual 130 165.072 1.270
Total 134 578.833
Coeff. Std Error t Stat P-value Lower 95% Upper 95%
Intercept 9.402 1.258 7.472 0.000 6.912 11.891
10y Bond -55.961 6.493 -8.619 0.000 -68.807 -43.116
EPS gr 2y 20.627 5.649 3.651 0.000 9.450 31.803
DPS gr 2y 10.819 3.662 2.954 0.004 3.574 18.065
POR 17.102 1.954 8.752 0.000 13.236 20.968
The data shows that all factors are significant in explaining changes in market ratings.
There are a number of particularly relevant issues associated with these results:
• The impact of the 10 year bond yield is very significant. This is a well documented and accepted factor for the ASX200.
Various interest rates were tested, and the 10 year yield was the most relevant.
• Both EPS and DPS growth series were used, and both are statistically significant. These were sufficiently different for both
to be statistically significant: again, a fact that contradicts the Modigliani and Miller Theory.
• Dividend POR is statistically significant: the coefficient is +ve; the t statistic is relatively high, and the P value low, suggesting a
solid link with market pricing.
These numbers aren’t conclusive by themselves, but they do give a strong indication that dividends have a much greater role in determining
market prices than theory suggests.
Even more significant is the fact that both earnings and dividends have an independent influence on pricing. Generally, in finance theory, the
two are assumed to be proxies for each other. These findings appear to disprove that assumption.
The following chart gives an indication.
08 8407 5700 07 4946 5080 08 9964 3800 07 5631 2300 07 5409 6100 03 9242 4000 07 5478 1681 03 9831 5000 08 9268 4888 02 9247 8666
Portfolio Research

Analyst
Peter Rice
ASX200 Historical PER v Market Dividend POR

22 80%
21
75%
20
70%
19
18
65%
POR %
PER x
17
60%
16
15
55%
14
50%
13
12 45%
Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun- Dec- Jun-
92 92 93 93 94 94 95 95 96 96 97 97 98 98 99 99 00 00 01 01 02 02 03
ASX200 Hist. PER ASX200 POR
The interesting periods in this graph are:
1. The early ‘90’s, when payout ratios averaged nearly 80%, as the banks saw their earnings collapse, and attempted to hold up their
share prices by artificially holding up dividend payments, which coincided with historically high PER multiples;
2. Late ’97 and late ’99 and early ’00, where payout ratio increases coincided with increased multiples.
The reason that this analysis is interesting is that the ASX has shown the greatest variation in dividend payout ratios of any of the major
markets over the past 20 years. Payout has been much more stable in the US, UK and Asia, which possibly explains why there has been far
less focus on the role of dividend payout in market pricing.
However, the ultimate test is to consider just how much of market pricing is explained by building a valuation model, based on these
assumptions, and comparing it to actual Index prices. The result is shown below:
ASX200 Historical PER v EPS+DPS Valuations
23
21
19
17
PER x
15
13
11
9
Jun-92
Oct-92
Feb-93
Jun-93
Oct-93
Feb-94
Jun-94
Oct-94
Feb-95
Jun-95
Oct-95
Feb-96
Jun-96
Oct-96
Feb-97
Jun-97
Oct-97
Feb-98
Jun-98
Oct-98
Feb-99
Jun-99
Oct-99
Feb-00
Jun-00
Oct-00
Feb-01
Jun-01
Oct-01
Feb-02
Jun-02
Oct-02
Feb-03
Jun-03
Mkt PER Pre GW W/O Model PER
Correlation between the two series is a very high 88%.

08 8407 5700 07 4946 5080 08 9964 3800 07 5631 2300 07 5409 6100 03 9242 4000 07 5478 1681 03 9831 5000 08 9268 4888 02 9247 8666
Portfolio Research

Analyst
Peter Rice
Critically, in terms of the test as to whether or not the model “works”, the structure not only explains almost 90% of the change in P/E
multiple rating of the market over time, but also matches the absolute levels of valuation applied by the market.
Interestingly, an almost identical structure, applied to the S&P500, FTSE100 and HSIX indices produces similar results, with correlation
coefficients of >80% in all cases, over similar (if not longer) time periods.
The most serious issue now comes down to:
Why does the market appear to be double counting dividends in its pricing?
The only justifiable reason would appear to be if the market is effectively giving dividends “for free”; in other words, there is no capital loss
associated with receiving cash dividends.
If this is the case, it potentially explains why the Modigliani and Miller theory seems to not be holding, as that theory is predicated on the
assumption that the payment of a cash dividend results in a corresponding capital loss, resulting in the total return remaining the same.
Markets Don’t Appear To “Charge” a Capital Loss for Dividends Paid

In an effort to test for this phenomenon, we analysed the trading history of the top 50 stocks by market capitalisation of the ASX200 over the
period December 2000 to December 2003. Relative price performance (versus the ASX200) over 15 day trading periods from ex-dividend
date –15 days, to ex-dividend date +90 days, were investigated. Over the period, no more than 10% of the market capitalisation of the
ASX200 went ex-dividend on any one day, preventing any potential distortion of the market-relative performance. Average price
performance figures were calculated on a simple (versus market capitalisation-weighted) basis. The results of this analysis are:
Market Relative Price Performance Around Dividend Payment Dates
Trading Days Since Last Cum-Dividend Trading Day

-15 +15 +30 +45 +60 +75 +90 +105
Max 20.29% 21.80% 15.51% 17.74% 20.35% 22.52% 19.33% 14.86%
Min -25.97% -35.70% -37.27% -24.74% -25.67% -41.73% -20.96% -36.25%
Avge. 0.06% 0.68% -0.36% 0.60% -0.26% -0.05% 0.40% 0.31%
StDev 5.8% 5.5% 5.5% 5.3% 6.1% 6.4% 5.0% 5.5%
The first observation to make is what is not evident here. The ASX200 traded on an average dividend yield of approximately 3.8% during the
period. Given that cash dividends are generally paid twice each year, the expected average adjustment to prices relative to market
as a result of those dividend payments would be in the region of 1.5-2.0%. This would obviously be expected in the first 15 trading days
after the last cum-dividend trading day. The results above show that this does not, in fact take place. To the contrary, the average
relative price performance of stocks in this period is positive, and does, in fact, show the highest average relative price
performance of any of the periods looked at, with no greater standard deviation in those results. This would appear to be a significant
result.
Given the fact that numerical averages have been used in the calculations to highlight the prevalence of the phenomenon, as against the
percentage of the market to which this applies, the actual magnitude of the relative price performance has little real relevance, except in
comparison to the figures achieved for the other periods.
Further, the performance in the first period, being the 15 days immediately prior to the last cum-dividend trading day, shows little evidence to
suggest that there is any significant short-term, dividend capture related buying of stocks in the sample, which may have caused some
distortion of the results. However, any short-term unwinding of such trading positions is likely to have worked against the phenomenon
identified if it did exist.
The second test that was carried out was to investigate the number of observations where there had been a clear return of relative stock
prices to cum-dividend levels in the immediate post ex-dividend trading period. A period of 15 trading days was arbitrarily chosen. The
08 8407 5700 07 4946 5080 08 9964 3800 07 5631 2300 07 5409 6100 03 9242 4000 07 5478 1681 03 9831 5000 08 9268 4888 02 9247 8666
Portfolio Research

Analyst
Peter Rice
comparative cum-dividend relative (to the ASX200) price was calculated as the average of the relative price for the 5 trading days ending 6
days before the dividend payment date. This was done to remove any distortion from a potential run-up in price in the last 5 days before the
ex-dividend date, but had allowed sufficient time for any post-earnings result price adjustment.
The maximum relative price achieved in the 15 trading days after the stock went ex-dividend was recorded and compared to the cum-
dividend relative price comparison calculation. If the maximum price achieved was greater than the comparison figure, this was recorded as
a “full retracement”. If the maximum achieved relative price was less than the comparative, then the proportion of the maximum relative price
from the theoretical ex-dividend price was compared to the calculated relative dividend payment (cash dividend paid / ASX200 Index closing
on the last cum-dividend day).
Results for the sample used showed that 388 observations out of a total sample of 548 observations (71%) traded at a maximum relative
price equal to or greater than the comparative cum-dividend pricing. A further 20 observations, or 3.7%, retraced between 75% and 100% of
the dividend adjustment.
An additional phenomenon was observed. Stocks that had announced results significantly below market expectations suffered relative
capital depreciation for a period after the announcement of such results, which went well beyond the corresponding ex-dividend date. This
adjustment to negative results announcements clearly distorts the results associated purely with the examination of the post dividend
payment phenomenon being investigated. We identified 56 instances in our sample where this was evident.
If these observations are removed from the sample, the percentage of full retracement observations rises to 79%, while the
percentage of stocks achieving 75% retracement or better increases to 83%.
These results appear to be consistent with the results achieved in the first test, and significantly reinforce the existence of what appears to
be a wide-spread practice for markets to return stock prices to cum-dividend levels relatively quickly after the payment of cash dividends.
Similar results were achieved on a similar basis, in the US, the UK, and Hong Kong.
On a prima facie basis, this evidence seems to suggest that the market is, in effect, giving investors cash dividends “for free”, and
as such, confirms the findings of Gentry et al. (2002) and Penman and Sougiannis (1998). In addition, the findings are consistent with
the long-popular practice of “dividend stripping” that was widely used as a short-term trading strategy to boost returns prior to the
introduction of the 1986 US Tax Reform Act.
It had been argued since the publication of the work of Miller and Modigliani (1961) that the practice was theoretically unprofitable, yet it
persisted to such an extent that the law was changed in an attempt to eradicate it as a practice over 20 years later.
DISCLAIMER
Disclosure of Interest
Tolhurst Noall Ltd (‘Tolhurst Noall’) and/or entities and persons connected with it may have an interest in the securities the subject of the
recommendations set out in this report. In addition, Tolhurst Noall and/or its agents will receive brokerage on any transaction involving the
relevant securities.
TNL may seek from the companies subject to this Research Report and/or their shareholders, advisory mandates or mandates for dealings in
securities, and therefore may receive commissions or fees from the companies, and/or their shareholders, at some time in the future.
Disclaimer
The information and opinions contained in this report have been obtained from sources Tolhurst Noall believed to be reliable, but no
representation or warranty, express or implied, is made that such information is accurate or complete and it should not be relied upon as such.
Information and opinions contained in the report are published for the assistance of recipients, but are not relied upon as authoritative and may
be subject to change without notice. Except to the extent that liability cannot be excluded, no Tolhurst Noall Group company accepts any liability
for any direct or consequential loss arising from any use of material contained in this report.
General Advice Warning
This report is intended to provide general advice. In preparing this advice, Tolhurst Noall did not take into account the investment objective, the
financial situation and particular needs of any particular person. Before making an investment decision on the basis of this advice, you need to
consider, with or without the assistance of a securities adviser, whether the advice is appropriate in light of your particular investment needs,
objectives and financial circumstances.
Explanation of Tolhurst Noall's Recommendation and Risk Rating system:
Recommendations are assessments of each Tolhurst Noall Analyst's view of potential total returns over a 1 year period relative to the
performance of the All Ordinaries Accumulation Index.
Expected total Return is measured as (capital gain (or loss) + dividend)/purchase price
We have divided our recommendations into five main categories:
Strong Buy: Expected Total Return in excess of 25% over a 1 year period relative to All Ordinaries Accumulation Index
Buy: Expected Total Return between 15% and 25% over a 1 year period relative to All Ordinaries Accumulation Index
Accumulate: Expected Total Return between 5% - 15% over a 1 year period relative to All Ordinaries Accumulation Index
08 8407 5700 07 4946 5080 08 9964 3800 07 5631 2300 07 5409 6100 03 9242 4000 07 5478 1681 03 9831 5000 08 9268 4888 02 9247 8666
Portfolio Research

Analyst
Peter Rice
Hold: Expected Total Return between -5% and 5% over a 1 year period relative to All Ordinaries Accumulation Index
Sell: Expected Total Return less than -5% over a 1 year period relative to All Ordinaries Accumulation Index
Risk Ratings:
Risk is a subjective assessment of overall risk within a company including price volatility and earnings variability, external liquidity, and size.
We divide our risk into three categories:
High: Company typically has high price volatility and earnings variability, low external liquidity and has a small market capitalisation.
Medium: Company typically has moderate price volatility and earnings variability, external liquidity and a medium size market capitalisation.
Low: Company typically has low price volatility and earnings variability, high external liquidity and is a large size market capitalisation.
Analyst verification
I verify that I, Peter Rice, have prepared this research report accurately and that any financial forecasts and recommendations that are
expressed are solely my own personal opinions. In addition, I certify that no part of my compensation is or will be directly or indirectly tied to the
specific recommendation or financial forecasts expressed in this report.
08 8407 5700 07 4946 5080 08 9964 3800 07 5631 2300 07 5409 6100 03 9242 4000 07 5478 1681 03 9831 5000 08 9268 4888 02 9247 8666

Equity Markets - Not Crazy, Just Misunderstood

Hochgeladen von

Dokumentinformationen

Originalbeschreibung:

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Equity Markets - Not Crazy, Just Misunderstood

Hochgeladen von

Copyright:

Verfügbare Formate

Portfolio Research

Equity Markets – Not Crazy, Just Misunderstood 31 August 2006 1

Why Value is (nearly) all that matters, if you define it correctly

• The implications of this work are substantial:

The Basic Principle

Value = NPV of Expected Future Returns.

Again, another “basic principle” comes into play here:

Markets can only price what they can reasonably predict.

Equity Markets – Not Crazy, Just Misunderstood 31 August 2006 2

Capital appreciation = retained cash.

So, How Do Markets Look To Predict Capital Gain?

BUT, this presents two major problems:

Equity Markets – Not Crazy, Just Misunderstood 31 August 2006 3

DPS plus EPS Explain Market Pricing

The following chart gives an indication.

Equity Markets – Not Crazy, Just Misunderstood 31 August 2006 4

ASX200 Historical PER v Market Dividend POR

ASX200 Hist. PER ASX200 POR

The interesting periods in this graph are:

Mkt PER Pre GW W/O Model PER

Correlation between the two series is a very high 88%.

Equity Markets – Not Crazy, Just Misunderstood 31 August 2006 5

The most serious issue now comes down to:

Markets Don’t Appear To “Charge” a Capital Loss for Dividends Paid

Market Relative Price Performance Around Dividend Payment Dates

Trading Days Since Last Cum-Dividend Trading Day

Equity Markets – Not Crazy, Just Misunderstood 31 August 2006 6

Equity Markets – Not Crazy, Just Misunderstood 31 August 2006 7

Das könnte Ihnen auch gefallen