Capital Markets

Value-Based Anomalies

Dr. Keith Anderson

The York Management School

1
 Learning Outcomes

• Appreciate the large apparent hole in efficient

markets theory
• Learn some more about a few of the
anomalies: the P/E, PSR and PEG ratio
• Take my PhD work into the P/E anomaly as
an example of how to advance research

2
 Value-Based Anomalies: Overview

I. Why is researching anomalies important?

II. Review of the P/E anomaly literature
III. The PSR anomaly
IV. The PEG ratio
V. My PhD research

3
Recommended Reading

 David Dreman (1998),

Contrarian Investment
Strategies: The Next
Generation
 Library G 2.63 DRE
 From £2.93 on Amazon

4
Part I: Why is Anomaly Research Important?

• CAPM: the dominant paradigm

• Proposes a straight-line relationship between a
stock’s volatility with respect to the market and
the returns expected from it
• Taught to thousands of business students every
year
• What they don’t tell you...

5
 The Academic Argument

• Many so-called ‘anomalies’ documented

• ‘Anomalies’ only for the CAPM: each seems to
show a statistic that can forecast returns
• CAPM says that beta is the only variable that
should be able to predict returns

6
 The Academic Argument

• P/E anomaly was the earliest (Nicholson (1960))

• Predated the CAPM itself (Sharpe (1964))
• Eclipsed by price-to-book-value in recent years
• Problem: No replacement model with sound
underlying theory as the CAPM has
• May move in future towards other models based
on F&F 3-factor model or behavioural finance

7
 The Practical Argument

• This is all just a dry philosophical debate for

market operators
• Want to spot shares that are going to go up
• Extra level of risk also important to them
• Several value-based quantitative funds in the US
• Only one ‘fund’ in the UK: Barclays iShare UK
Dividend Plus. See
http://uk.ishares.com/en/rc/stream/pdf/-/publish/repository/document
s/en/downloads/factsheet_ftse_uk_dividend_plus.pdf

8
 Question Sheet

• Now read the UK Dividend Plus factsheet and

try to answer the questions

9
 Part II: The Price-Earnings Anomaly

• The most popular statistic for looking at a

company’s value at its current price
• Historical P/E: Share price
Earnings in last company year

• Prospective P/E: Share price

Consensus forecast for next year
• E/P often used in academic literature – why?

10
 What is a Price-Earnings Ratio?

• Original theory (Gordon and Shapiro (1956)):

P P 11- bb

E r r- gg
E
P: current price of the stock
E: last announced company earnings
b: payout ratio
r: discount rate
g: dividend or earnings growth rate

• So higher growth rate g → smaller denominator

→ higher P/E, ceteris paribus
11
 What is a Price-Earnings Ratio?

• Standard P/E 8 to 12, or 15+ for larger growth

stocks
• Compare to tech stock boom: AOL P/E of 275,
Yahoo P/E of 1900, Amazon P/E of ∞
• Varying expectations of future earnings growth,
both cross-sectionally & through time
• Going the other way, expectations of bad news can
give extraordinarily low P/Es

12
 Early Work on the P/E Effect

• Nicholson (1960): 3-page paper with 100 stocks

• 14.7 times original investment after 20 years for
lowest P/E quintile versus 4.7 times for highest
• Glamour stocks may merit high P/Es but often
prices have grown faster than earnings
• Rarely, price action continues spectacularly
• Seen as representative of growth stocks generally
• Reality eventually makes itself felt

13
“High price-earnings multiples typically reflect investor
satisfaction with companies of high quality, or with
those which have experienced several years of
expansion and rising earnings. In such cases, prices
have often risen faster than earnings. A resultant
increase in price-earnings ratios may be justified in
individual instances, but under the impact of public
approval or even glamour, it often runs to extremes.
...
Some growth stocks appear to be exceptions, at least for
temporary periods, and in individual instances price
advances have continued spectacularly.”
Nicholson (1960) p.45
14
 Early Work on the P/E Effect

• Nicholson (1968) looked at P/E, PSR, and PBV

over holding periods up to 7 years

7-year Sales 7-year Book value 7-year

P/E
growth / price growth / price growth

< 10 131% >5 138% > 1.5 149%

10-12 87% 2-5 108% 1-1.5 112%
12-15 88% 1-2 107% 0.6-1 91%
15-20 75% 0.6-1 89% 0.3-0.6 90%
> 20 71% < 0.6 69% < 0.3 86%
15
 Early Work on the P/E Effect

• First paper by an academic Basu (1975 & 1977):

looked at NYSE industrials 1957-71
 
“The average annual rates of E/P Avg.Rtn
beta
Quintile p.a. %
return decline...as one moves
from low P/E to high P/E A (highest) 9.3 1.11
portfolios. However, contrary to A* 9.6 1.06
capital market theory, the higher B 9.3 1.04
returns on the low P/E portfolios C 11.7 0.97
were not associated with higher D 13.6 0.94
levels of systematic risk” (p666) E (lowest) 16.3 0.99

16
 More Recent Findings

• Ball (1978) re Phlogiston: see Dreman pp.151-2

• Jaffe, Keim & Westerfield (1989): January, size
& P/E effects all significant
• Fuller, Huberts & Levinson (1993): Ball’s
(1978) ‘omitted risk factors’ couldn’t account for
better returns from low P/E stocks
• Fama and French (1992 and 1993): model with
market return, size and PBV could ‘explain’ value
stock returns after the fact
• P/E effect subsumed by PBV and size effects
17
 More Recent Findings

• Lakonishok, Schleifer & Vishny (1994): value

vs. glamour stocks 1963-1990 on the basis of past
sales growth
• Differences in expected future growth rates
consistently overestimated
• Outperformance of value stocks 10-11% per year
• Value strategies aren’t fundamentally riskier –
value stocks did particularly well in ‘bad’ states of
the world

18
 LSV(1994)

• P/E did not produce as large an effect as price to

book value or price to cash flow, possibly because
“stocks with temporarily depressed earnings are
lumped together with well-performing glamour
stocks in the high expected growth/low E/P
category. These stocks with depressed earnings do
not experience the same degree of poor future
stock performance as the glamour stocks, perhaps
because they are less overpriced by the market.”
(page 1550)
19
 The P/E: Summary

• A recognised value indicator, in use since 1920s

• Efficient markets believers say it must be a proxy
for some sort of risk
• Has lost its importance in recent years due to price
to book value being preferred in the Fama and
French 3-factor model
• But still has its proponents such as Dreman (and
me)

20
 Part III: The PSR Anomaly

• A value/glamour indicator that is less easy to

manipulate than the P/E
• Particularly useful for loss-makers
• Out-of-favour firms barely making a profit could
move from the highest P/E group to a more
appropriate low PSR group
• But PSR has two important weaknesses:
– no good for financial services companies
– best to stratify by industry

21
 The PSR Anomaly
Literature on the PSR
• Nicholson (1968) first to assess PSRs
• Senchack & Martin (1987): US stocks 1975-
1984 excluding financial services firms
PSR* Quarterly
beta
Quintile return
Low PSR* 7.27% 1.113
2 5.92% 1.013
3 5.28% 0.907
4 4.47% 0.814
High PSR* 3.78% 0.892
22
 The PSR Anomaly

• Nathan, Sivakumar & Vijayakumar (2001):

important to stratify PSR by industry. Long
PSR1PE1 and short PSR5PE5 gave +37.78% p.a
• Sole UK work: Leledakis and Davison (2001)
• Found that PSR and gearing together had
significantly more predictive power than F&F’s
size and PBV together

23
 Question Sheet

• Now try the questions on the PSR

24
 Part IV: The PEG Ratio

• Popular in recent years for valuing growth stocks

• Formula is P/E ratio
earnings growth rate
• Implicit assumption: earnings growth rates tend to
be sustained over several years
• BUT Little (1962) and Rayner and Little (1966)
showed that earnings growth is indistinguishable
from a random walk

25
 The PEG Ratio

• Sort companies into past growth rate deciles:

immediately revert to average of the market
• Subsequent differences no more than would be
expected from chance alone
• Replicated using US data by Lintner and Glauber
(1967)
• More recently by Chan, Karceski and Lakonishok
(2003) looking at all US companies 1951-1997

26
 The PEG Ratio

• ‘Earnings as a random walk’ idea hasn’t dented the

popularity of Jim Slater’s Zulu Principle books
(1992 and 1996)
• Market participants still seem generally to believe
that growth rates are sustained over several years
and can be predicted. See
http://www.companynews.co.uk/star/stepbystep2.htm
• PEG ratio is also prominent in popular company
information database Company REFS which Slater
helped to design
27
 The PEG Ratio

“It ain't so much the things we don’t know that get us into
trouble. It's the things we know that just ain't so.”

- Variously attributed to Josh Billings,

Artemus Ward and Mark Twain

28
 Part V: My PhD Work

•Academic: widens the gap in returns

between high and low P/E stocks that needs to
be explained by any valuation theory
•Practitioners: obvious practical applications
for portfolio managers
Two areas of results covered here:
1. Long-term P/Es
2. Deconstructing the P/E ratio

29
 Data Sources

• All UK shares 1975-2003 including AIM

• LSPD for company names, then Datastream
thereafter
• Excluded financial sector companies, including
investment trusts
• For companies that went bust, set RI manually to 0

30
 Part VA: The Long-Term P/E

•All papers up to now use the one-year

historical P/E for assigning companies to deciles
or quintiles
•Why should only one year of earnings be
relevant for predicting future returns?
“Average earnings...should cover a period of not less
than five years, and preferably seven to ten years”
(Graham and Dodd (1934, p.452))
•Remember LSV (1994) p.1550 quote
•Assign companies to deciles n using EP1, the
 EPSij
traditional E/P, up to EP8: EPn  j 1 31
i
nPi
 Long-Term P/E Decile Returns
EP1 EP2 EP3 EP4 EP5 EP6 EP7 EP8

Highest P/E 18.28% 18.20% 18.62% 16.65% 17.84% 17.83% 18.15% 16.26%

Decile 2 19.25% 19.36% 16.41% 17.98% 16.94% 17.42% 16.16% 16.71%

Decile 3 18.38% 17.32% 18.92% 18.68% 17.78% 17.51% 17.05% 16.43%

Decile 4 16.44% 18.96% 19.45% 18.42% 19.49% 17.81% 18.61% 18.42%

Decile 5 17.96% 18.06% 17.73% 18.58% 17.62% 19.11% 18.34% 19.54%

Decile 6 18.53% 18.73% 19.32% 18.98% 19.97% 19.69% 19.81% 19.81%

Decile 7 21.59% 19.53% 19.86% 20.77% 19.61% 20.18% 19.86% 19.39%

Decile 8 20.86% 20.55% 21.33% 22.11% 21.81% 20.42% 20.58% 21.11%

Decile 9 22.47% 21.75% 22.00% 22.08% 22.48% 22.88% 22.48% 23.05%

Lowest P/E 24.26% 22.82% 21.89% 22.18% 24.27% 25.51% 27.57% 27.87%

D10 – D1 5.98% 4.62% 3.28% 5.52% 6.44% 7.67% 9.42% 11.62%

32
Long-Term P/E Decile Returns

28%
26%
24%
22%
20%
18%
16%
14%
12%
10%
EP8
EP1

33
 Weighting Past Years of Earnings
Increasing into the Increasing into the
future past
Equal Inverse Inverse Linear
Linear Linear
weights square square Regression
Highest P/E 16.26% 16.89% 17.84% 15.81% 15.19% 16.68%
Decile 2 16.71% 16.30% 16.44% 16.05% 16.79% 14.83%
Decile 3 16.43% 18.10% 18.03% 16.95% 16.52% 17.40%
Decile 4 18.42% 17.24% 17.48% 18.35% 19.05% 17.60%
Decile 5 19.54% 18.93% 17.96% 19.14% 19.00% 19.77%
Decile 6 19.81% 19.66% 20.03% 20.69% 19.42% 18.28%
Decile 7 19.39% 20.11% 20.76% 19.92% 20.80% 20.36%
Decile 8 21.11% 20.79% 21.50% 20.27% 21.00% 22.33%
Decile 9 23.05% 23.83% 20.77% 23.48% 23.84% 22.89%
Lowest P/E 27.87% 26.68% 27.75% 27.98% 27.01% 28.45%
D10 – D1 11.62% 9.79% 9.91% 12.17% 11.82% 11.77%

34
 The Long-Term P/E Ratio - Summary

• EP8 resolution 12% versus 6% for traditional EP1

• Any fund manager who takes a longer-term view
of earnings will gain a competitive advantage
• Individual past years of earnings aren’t equally
valuable for predicting returns: earnings from five
to eight years ago are better at predicting returns
than the traditional P/E!
• Not much success with clever weighting systems.
Remaining work sticks to equally-weighted EP8

35
 Part VB: Decomposing the P/E Ratio

• Several regular influences on a company’s P/E,

including:
– The year
– The sector
– Size
– Idiosyncratic effects
• Question: do these different effects have different
values in predicting returns?

36
 Influences on the P/E
Year
Low P/E

High returns

Sector
High P/E

P/E
High returns
Low P/E

Size
High returns
Low P/E

High returns

Idiosyncratic
Low P/E

High returns
37
 Market Average P/Es by Year

30

25

20
Average
15
P/E

10

38
 Sector P/Es

• Average P/E calculated for each sector with ten or

more company/year returns
• Averaged over all 29 years
• 132 LSPD G17 categories
• P/E varies from 28.7 (Oil and gas exploration and
production) to 6.4 (Steel)
• Conclusion: Slightly better returns for high P/E
sectors!

39
 Size P/Es

• Group companies into 20 categories based on

market capitalisation on portfolio formation date
• E/P averaged for each category over all years
• Category limits vary from year to year
• Average P/Es vary consistently according to
market capitalisation
• Large fund managers naturally tend to buy large
companies to avoid liquidity problems

40
 Average P/Es by Size Category
20
18
16
14
12
Average
10
P/E
8
6
4
2
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Market Value Category
41
 Calculating the Idiosyncratic E/P

• Define the idiosyncratic E/P as that part of a

company’s E/P that is not explained by its year,
sector or size:
ActualEPi YearEPi MVEPi G17 EPi IdioEPi
   
AverageEP AverageEP AverageEP AverageEP AverageEP

• Note this is not a regression and there is no error

term! Rearranging,
ActualEPi  AverageEP 3
IdioEPi 
YearEPi  MVEPi  G17 EPi
42
 E/P Decomposition
Linear Regression Model
• Linear regression model is
Rtn01i   0  1YearEPi   2 MVEPi   3G17 EPi   4 IdioEPi  ui

• Coefficients are

Rtn01 = 0.736 + 0.479 YearEP + 0.274 MVEP

-0.093 G17EP + 0.091 IdioEP

• All significant at 1% level

43
 E/P Decomposition
Linear Regr Model

• Year E/P is much better at predicting returns, but

we can’t choose which year to invest in!
• MVEP three times as useful as G17EP or IdioEP
• Effect of G17 E/P is the opposite way round to the
others: a higher sector P/E means better returns

44
Decile Returns using the Decomposed E/P
Decomposed
EP1 EP8
E/P
Highest P/E 18.28% 16.26% 11.93%
Decile 2 19.25% 16.71% 14.66%
Decile 3 18.38% 16.43% 15.84%
Decile 4 16.44% 18.42% 16.94%
Decile 5 17.96% 19.54% 20.49%
Decile 6 18.53% 19.81% 19.64%
Decile 7 21.59% 19.39% 21.27%
Decile 8 20.86% 21.11% 23.84%
Decile 9 22.47% 23.05% 23.72%
Lowest P/E 24.26% 27.87% 30.23%
D10 – D1 5.98% 11.62% 18.30%
45
 Decomposed E/P Portfolio Example
£10,000,000

£1,000,000

£100,000

£10,000

£1,000
83
75

77

79

81

85

87

89

91

93

95

97

99

01

03
19

19

19

19

19

19

20
19

19

19

19

19

19

19

20
LinRegr Value LinRegr Glamour EP8 Value EP8 Glamour

46
 E/P Decomposition Summary

• Four influences (year, sector, size and

idiosyncratic components) identified
• Sector works in the opposite direction to the rest
• Decomposed the influences and put them back
together again in a weighted P/E
• Sector influence reversed
• D10-D1 difference in returns: 6% for traditional
P/E → 12% for EP8 → 18% for decomposed E/P
• Value decile now returns 30% per annum

47
 Summary

I hope you now

• Understand more about some of the anomalies
facing efficient markets theory and how to test
them
• Particularly about the P/E ratio and why it has
dropped out of favour recently
• Appreciate the sort of work that could advance
this research

48
 References
• Ball, R. 1978. Anomalies in Relationships between
Securities' Yields and Yield-Surrogates. Journal of
Financial Economics, 6(2/3): 103-26.
• Basu, S. 1975. The Information Content of Price-
Earnings Ratios. Financial Management, 4(2): 53-64.
• Basu, S. 1977. The Investment Performance of Common
Stocks in relation to their Price-Earnings Ratios. The
Journal of Finance, 32(3): 663-82.
• Chan, L.C.K., Karceski, J. & Lakonishok, J. 2003. The
Level and Persistence of Growth Rates. The Journal of
Finance, 58(2): 643-84.
• Dreman, D.N. 1998. Contrarian Investment Strategies:
The Next Generation. New York: Simon & Schuster.

49
 References
• Fama, E.F. & French, K.R. 1992. The Cross-Section of
Expected Stock Returns. The Journal of Finance, 47(2):
427-65.
• Fama, E.F. & French, K.R. 1993. Common Risk Factors
in the Returns on Stocks and Bonds. Journal of Financial
Economics, 33(1): 3-56.
• Fuller, R.J., Huberts, L.C. & Levinson, M.J. 1993.
Returns to E/P Strategies, Higgeldy Piggeldy Growth,
Analysts' Forecast Errors, and Omitted Risk Factors.
Journal of Portfolio Management, 1993(Winter): 13-24.
• Gordon, M. & Shapiro, E. 1956. Capital Equipment
Analysis: The Required Rate of Profit. Management
Science, 3: 102-10.
• Graham, B. & Dodd, D. 1934. Security Analysis. New
50
York: McGraw-Hill.
 References
• Jaffe, J., Keim, D.B., & Westerfield, R. 1989. Earnings
Yields, Market Values, and Stock Returns. The Journal
of Finance, 44(1): 135-48.
• Lakonishok, J., Schleifer, A. & Vishny, R. 1994.
Contrarian Investment, Extrapolation, and Risk. The
Journal of Finance, 49(5): 1541-78.
• Leledakis, G. & Davidson, I. 2001. Are Two Factors
Enough? The U.K. Evidence. Financial Analysts
Journal, 57(6): 96-105.
• Linter, J. & Glauber, R. 1967. Higgledy Piggledy Growth
in America. In Lorie, J. and Brealey, R., editor, Modern
Developments in Investment Management. Hinsdale, IL:
Dryden Press.

51
 References

• Little, I.M.D. 1962. Higgledy Piggledy Growth. Journal

of the Oxford University Institute of Statistics, 24(4):
387-412.
• Miller, W. and McGarry, P.C. 1966. 452 ways to beat the
market. Financial Analysts Journal, 23(4): 96-105.
• Nathan, S., Sivakumar, K., & Vijayakumar, J. 2001.
Returns to Trading Strategies Based on Price-to-Earnings
and Price-to-Sales Ratios. Journal of Investing, 10(2):
17-28.
• Nicholson, S.F. 1960. Price-Earnings Ratios. Financial
Analysts Journal, 16(4): 43-45.
• Nicholson, S.F. 1968. Price-Earnings Ratios in relation to
Investment Results. Financial Analysts Journal, 24(1):
105-09. 52
 References
• Rayner, A.C. & Little, I.M.D. 1966. Higgledy Piggledy
Growth Again. Oxford: Basil Blackwell.
• Senchack, A.J. & Martin, J.D. 1987. The Relative
Performance of the PSR and PER Investment Strategies.
Financial Analysts Journal, 43(2): 46-56.
• Slater, J. 1992. The Zulu Principle. London: Texere
Publishing.
• Slater, J. 1996. Beyond the Zulu Principle. London:
Texere Publishing.

53