. 164
( 193 .)


tal. The direction of the cumulated series is then used to discern broad market trends. Analysts
might use a moving average of cumulative breadth to gauge broad trends.

Relative strength Relative strength measures the extent to which a security has out-
relative strength
performed or underperformed either the market as a whole or its particular industry. Relative
Recent performance
strength is computed by calculating the ratio of the price of the security to a price index for the
of a given stock or
industry. For example, the relative strength of Ford versus the auto industry would be mea-
industry compared to
that of a broader sured by movements in the ratio of the price of Ford divided by the level of an auto industry
market index. index. A rising ratio implies Ford has been outperforming the rest of the industry. If relative
strength can be assumed to persist over time, then this would be a signal to buy Ford.
Similarly, the relative strength of an industry relative to the whole market can be computed
by tracking the ratio of the industry price index to the market price index.
Some evidence in support of the relative strength strategy is provided in a study by
Jegadeesh and Titman (1993). They ranked firms according to stock market performance in a
six-month base period and then examined returns in various follow-up periods ranging from 1
to 36 months. They found that the best performers in the base period continued to outperform
other stocks for several months. This pattern is consistent with the notion of persistent relative
strength. Ultimately, however, the pattern reverses, with the best base-period performers giv-
ing up their initial superior returns. Figure 19.12 illustrates this pattern. The graph shows the
Bodie’Kane’Marcus: VI. Active Investment 19. Behavioral Finance and © The McGraw’Hill
Essentials of Investments, Management Technical Analysis Companies, 2003
Fifth Edition

19 Behavioral Finance and Technical Analysis







6 11 16 21 26 31 36
“1 Month after initial ranking of stocks

F I G U R E 19.12
Cumulative difference in returns of previously best-performing and worst-performing stocks in subsequent months
Source: Jegadeesh and Titman (1993).

cumulative difference in return between the 10% of the sample of stocks with the best base-
period returns and the 10% with the worst base-period returns. Initially, the curve trends up-
ward, indicating that the best performers continue to outperform the initial laggards. After
about a year, however, the curve turns down, suggesting that abnormal returns on stocks with
momentum are ultimately reversed.

The middle two columns of the following table present data on the levels of an auto industry
index and a broad market index. Does the auto industry exhibit relative strength? That can
be determined by examining the last column, which presents the ratio of the two indexes. De-
spite the fact that the auto industry as a whole has exhibited positive returns, reflected in the Relative Strength
rising level of the industry index, the industry has not shown relative strength. The falling ra-
tio of the auto industry index to the market index shows that the auto industry has underper-
formed the broad market.

Week Auto Industry Market Index Ratio

1 165.6 447.0 0.370
2 166.7 450.1 0.370
3 168.0 455.0 0.369
4 166.9 459.9 0.363
5 170.2 459.1 0.371
6 169.2 463.0 0.365
7 171.0 469.0 0.365
8 174.1 473.2 0.368
9 173.9 478.8 0.363
10 174.2 481.0 0.362
Bodie’Kane’Marcus: VI. Active Investment 19. Behavioral Finance and © The McGraw’Hill
Essentials of Investments, Management Technical Analysis Companies, 2003
Fifth Edition

674 Part SIX Active Investment Management

Value Line is the largest investment advisory service in the world. Besides publishing the
Value Line Investment Survey, which provides information on investment fundamentals for
approximately 1,700 publicly traded companies, Value Line also ranks each of these stocks ac-
cording to their anticipated price appreciation over the next 12 months. Stocks ranked in group
1 are expected to perform the best, while those in group 5 are expected to perform the worst.
Value Line calls this “ranking for timeliness.”
Figure 19.13 shows the performance of the Value Line ranking system over the 25 years
from 1965 to March 1990. Over the total period, the different groups performed just as the
rankings predicted, and the differences were quite large. The total 25-year price appreciation
for the group 1 stocks was 3,083% (or 14.8% per year) compared to 15% (or 0.5% per year)
for group 5.
How does the Value Line ranking system work? As Bernhard (1979) explains it, the rank-
ing procedure has three components: (1) relative earnings momentum, (2) earnings surprise,
and (3) a value index. Most (though not all) of the Value Line criteria are technically oriented,
relying on either price momentum or relative strength. Points assigned for each factor deter-
mine the stock™s overall ranking.
The relative earnings momentum factor is calculated as each company™s year-to-year
change in quarterly earnings divided by the average change for all stocks.

Cumulative return
Group 1: 3,083%
Group 2: 1,634%
Group 3: 717%
Group 4: 232%
Group 5: 15%





1965 70 75 80 85 90

F I G U R E 19.13
Record of Value Line ranking for timeliness (without allowing for changes in rank, 1965“1990)
Source: From Value Line Investment Survey, “Selection & Opinion,” April 20, 1990. Copyright 1994 by Value Line Publishing, Inc. Reprinted by permission:
All Rights Reserved.
Bodie’Kane’Marcus: VI. Active Investment 19. Behavioral Finance and © The McGraw’Hill
Essentials of Investments, Management Technical Analysis Companies, 2003
Fifth Edition

19 Behavioral Finance and Technical Analysis

The earnings surprise factor has to do with the difference between actual reported quarterly
earnings and Value Line™s estimate. The points assigned to each stock increase with the per-
centage difference between reported and estimated earnings.
The value index is calculated from the following regression equation
V a b1x1 b2x2 b3x3

x1 A score from 1 to 10 depending on the relative earnings momentum ranking,
compared with the company™s rank for the last 10 years;
x2 A score from 1 to 10 based on the stock™s relative price, with ratios calculated in a
similar way to the earnings ratio;
x3 The ratio of the stock™s latest 10-week average relative price (stock price divided by
the average price for all stocks) to its 52-week average relative price; and a, b 1, b 2,
and b3 are the coefficients from the regression estimated on 12 years of data.
Finally, the points for each of the three factors are added, and the stocks are classified into
five groups according to the total score.
Investing according to this system does seem to produce superior results on paper, as Figure
19.13 shows. Yet, as the nearby box points out, in practice, things are not so simple”Value
Line™s own mutual funds have not kept up even with the broad market averages. The box illus-
trates that even apparently successful trading rules can be difficult to implement in the market.

Self-Destructing Patterns
It should be abundantly clear from our presentations that most of technical analysis is based on
ideas totally at odds with the foundations of the efficient market hypothesis. The EMH follows
from the idea that rational profit-seeking investors will act on new information so quickly that
prices will nearly always reflect all publicly available information. Technical analysis, on the
other hand, posits the existence of long-lived trends that play out slowly and predictably. Such
patterns, if they exist, would violate the EMH notion of essentially unpredictable stock price
An interesting question is whether a technical rule that seems to work will continue to work
in the future once it becomes widely recognized. A clever analyst may occasionally uncover a
profitable trading rule, but the real test of efficient markets is whether the rule itself becomes
reflected in stock prices once its value is discovered.
Suppose, for example, the Dow theory predicts an upward primary trend. If the theory is
widely accepted, it follows that many investors will attempt to buy stocks immediately in an-
ticipation of the price increase; the effect would be to bid up prices sharply and immediately
rather than at the gradual, long-lived pace initially expected. The Dow theory™s predicted trend
would be replaced by a sharp jump in prices. It is in this sense that price patterns ought to be
self-destructing. When a useful technical rule (or price pattern) is discovered, it ought to be in-
validated once the mass of traders attempts to exploit it, thereby forcing prices to their “cor-
rect” levels.
Thus, the market dynamic is one of a continual search for profitable trading rules, followed
by destruction by overuse of those rules found to be successful, followed by yet another search
for yet-undiscovered rules.
Bodie’Kane’Marcus: VI. Active Investment 19. Behavioral Finance and © The McGraw’Hill
Essentials of Investments, Management Technical Analysis Companies, 2003
Fifth Edition

Paying the Piper
Performance index (12/31/83 = 100) Ratio scale
250 Group I stocks
Value Line, Inc., publishes the Value Line Investment
Survey, that handy review of 1,652 companies. Each S & P 500
week the survey rates stocks from I (best buys) to V
Value Line
(worst). Can you beat the market following these rank-
Centurion Fund
ings? Value Line tracks the performance of group I
from April 1965, when a new ranking formula went
into effect. If you bought group I then and updated
your list every week, you would have a gain of 15,391%
by June 30. That means $10,000 would have grown to
about $1.5 million, dividends excluded. The market is
up only 245% since 1965, dividends excluded. Value Line
Quite an impressive record. There is only one flaw:
It ignores transaction costs. Do transaction costs much
matter against a performance like that? What does the
investor lose in transaction costs? A percentage point a
year? Two percent?
None other than Value Line provides an answer to
12/83 12/84 12/85 12/86 6/87
this question, and the answer is almost as startling as the
paper performance. Since late 1983, Value Line has run
a mutual fund that attempts to track group I precisely. Its
What went wrong? “Inefficiencies and costs of im-
return has averaged a dismal 11 percentage points a
plementation,” says Mark Tavel, manager of the fund,
year worse than the hypothetical results in group I. The
fund hasn™t even kept up with the market (see chart). Value Line Centurion.

SUMMARY • Behavioral finance seeks to identify behavior patterns that are inconsistent with standard
economic theory and can explain observed anomalies in asset prices.
• So far, a rich set of “irrational” behavior has been documented. But with the possible


. 164
( 193 .)