Generating Excess Returns through Global Industry Rotation

By

Geoffrey Loudon
Macquarie University

And

John Okunev
Macquarie University

March 2006
 Abstract

Numerous studies have shown that international diversification across global equity
markets can lead to improved performance on a risk adjusted basis when compared to
investing in domestic equity markets. Early studies provided empirical evidence that
country and industry effects were important factors in explaining total returns. These
studies generally concluded country factors dominate industry factors. Recent studies
however provide evidence this view is changing in that industry effects are at least as
important as country effects. With the increasing importance of industry effects it makes
sense to investigate what investment strategies could lead to improved investment
performance in relation to global equity markets. We examine a number of commonly
used investment strategies in relation to industry rotation. Our results indicate that each of
the strategies can earn significant positive excess returns. Yet these strategies appear to
act quite differently. Each of the strategies breaks down at some point in time, but by
combining the strategies into a portfolio of strategies we are able to generate more
consistent performance with improved levels of significance.

2
Introduction

Numerous studies have shown that international diversification across global equity
markets can lead to improved performance on a risk adjusted basis when compared to
investing in domestic equity markets. Early studies by Lessard (1974) and Solnik (1974)
provided empirical evidence that country and industry effects were important factors in
explaining total returns. They concluded country factors dominate industry factors.

Following on from this early research, Grinold, Rudd and Stefek (1989) provide
additional evidence that country effects still dominate industry effects. These results are
also supported by Heston and Rouwenhorst (1994, 1995), Beckers, Connor and Curds
(1996) and Griffin and Karolyi (1998). Roll (1992) on the other hand finds that that
industry factors explain about 40% of the variation of returns. Overall, the evidence from
1970-1999 supports the notion that country factors dominate industry factors.

Recent studies however provide evidence this view is changing. Studies by Baca, Garbe
and Weiss (2000) and Cavaglia, Brightman and Aked (2000) provide evidence of the
increasing importance of industry effects. For example, Cavaglia, Brightman and Aked
(2000) show that industry effects have grown substantially from 1995-1999 and that
diversification across industries provides greater risk reduction when compared to
diversification across countries. In a recent study Phylaktis and Xia (2003) examined
country and industry effects of the Dow Jones Global country and industry indexes of 51
industry groups and 34 country indices. Using weekly data from January 1992 to
December 2001 they find that country effects still dominate industry effects but from
1999 there has been an upward shift in the industry effect. The shift in industry effects is
predominately noticed in Europe and North America whilst country effects still dominate
Asia Pacific and Latin America.

Scowcroft and Sefton (2005) provide more recent evidence in supporting the increased
importance of industry/sector effects. Their study focuses on examining momentum
effects of ten MSCI global sectors from 1992-2003. Ranking sectors by prior six months

3
return but not including the last month, and going long the top two sectors and shorting
the bottom two sectors, they find that returns to global sectors explain on average 58% of
total return variation, country factors explain on average 34% of total return variation and
stock specific factors explain on average 8% total return variation. Their research
provides the most recent evidence that industry effects are now more important than
country effects and that momentum which will be discussed in greater detail later in the
paper is largely an industry effect.

It appears that industry effects are at least as important as country effects and this raises
the question what effect does this have on global equity investment strategies. Prior to
1999 it made more sense for a global portfolio manager to take active country bets as
opposed to active industry bets. Richards (1997) and Chan, Hameed and Tong (2000)
have shown country momentum strategies can lead to significant profits. With the
increasing importance of industry effects it makes sense to investigate what investment
strategies could lead to improved investment performance in relation to global equity
markets.

In the following discussion we examine the profitability of commonly used investment

strategies in relation to industry rotation. The strategies examined are:
• Momentum
• Value
• Combination of Value and Momentum
• Growth.
Each of these strategies can generate positive excess returns and are well documented in
the literature. Momentum has been extensively researched with regards to stock specific,
country and industry strategies. Value investing has a long history of academic research.
Fama and French (1992) suggest that excess returns of value investing represent a
distress premium. Both value and momentum strategies can earn positive excess returns,
and there are diversification benefits in employing a combination strategy as the
correlation between value and momentum is generally negative. In relation to growth
strategies, there is a theoretical basis that growth in long term earnings clearly plays an

4
important role in determining the required rate of return of a stock. From the discounted
dividend model (DDM) the required rate of return is the dividend yield plus the long term
growth rate of dividends (or earnings given we know the payout ratio). There are of
course many other strategies documented in the literature. It is not the intention here to
produce an exhaustive list but rather to focus on some of the simpler strategies that are
commonly researched and implemented.

Our results indicate that each of the above mentioned strategies can earn significant
positive excess returns. Yet these strategies appear to act quite differently. Each of the
strategies breaks down at some point in time, but by combining the strategies into a
portfolio of strategies we are able to generate more consistent performance.

The format of the paper follows. Section one provides descriptive statistics of global
industry data used in this study. Section two examines the performance of industry
momentum strategies. Sections three, four and five examine the performance of the
value, combination of value and momentum and growth strategies respectively. Section
six examines the performance of an aggregate strategy and section seven examines the
effect of inversion in the US yield curve. Section eight presents the conclusions of the
paper.

1. Descriptive Statistics

Global industry data representing thirty six industries was obtained from Datastream.
These data are constructed by Datastream. For each of the industries the following series
were obtained: total return index, price index, dividend yield and earnings yield for the
period 1/1973 to 12/2005 on a monthly basis.1 Total returns and prices were calculated in
terms of US dollars. Additionally, the Tbill rate and yield on 10 year government bonds
for the US were also downloaded.

Insert Table 1

1
As industry book to price ratios were not available from Datastream, we were not able to use these as an
alternative measure of value.

5
Table 1 presents summary statistics of the annual average return , annual standard
deviation, minimum and maximum monthly return for each industry over each decade
since 1973 and for the whole sample period from 1973-2005. We have done this to
highlight the variability of the performance of industries at different times. Continuously
compounded monthly returns were calculated from the Datastream total return index.
Over the entire sample period the average annual return for all sectors was 10.8% pa.
From 1/1973-12/1979, 1/1980-12/1989, 1/1990-12/1999 and 1/2000-12/2005 the average
annual return for all sectors was 4.6% pa, 19.5% pa, 10.2% pa and 4.6% pa respectively.
It is interesting to note that the eighties produced the highest return by far and that for the
seventies and the current decade produced returns of 4.6% pa.

Insert Table 2

To highlight the variability of performance of industries Table 2 presents the top four and
bottom four industries ranked by annual return for each decade. Over the entire period the
best performing industries were: Water, Tobacco, Aerospace/Defense and Oil and Gas.
The worst performing industries were: Forestry & Paper, Media and Photography, Steel
and other Metals and Information Technology. For the seventies, not surprisingly, Oil
and Gas and Gas Distribution performed well, whilst consumer related industries
performed poorly. For the eighties Specialty Finance, Water, Tobacco and Banks
performed well, whilst Software, Aerospace/Defense, Mining and Information
Technology were the worst performing industries. This situation was reversed in the
nineties where Software and Information Technology were the best performing industries
whilst Steel, Construction, Transport and Engineering and Machinery were the worst
performing industries. For the current decade this situation reversed again where,
Tobacco, Mining, Oil and Gas and Water were the best performing industries, whilst
Information Technology, Software, Telecom Services and Media and Photography were
the worst performing industries. It is quite evident that there is sufficient variability in
returns across industries and sample periods to warrant an investigation into whether one

6
can develop an industry rotation strategy which outperforms the passive strategy of
holding industry benchmark weights.

Table 3 presents some characteristics of the industries. Panel A displays the earnings
yield, annual return, implied growth and trend growth of each industry for the whole
period.2 For each industry an earnings index was created by multiplying the earnings
yield by the price index. The implied annual growth rate of each industry was calculated
as the natural logarithm (earning end 2005/earnings start 1973)/33. If one plots the
earnings index of an industry it is apparent that earnings oscillate about a trend line. The
trend growth of an industry is the slope of regressing the earnings index against time.
Trend growth can be above/below the implied growth rate. This relation is a function of
where earnings are in relation to the economic cycle. Currently the implied growth rate is
above trend growth owing to the recent recovery in the US economy over the last two to
three years. Panel B of Table 3 presents industries ranked by their earnings yields.
Industries that have high earnings yields are Water, Tobacco, Oil and Gas,
Aerospace/Defense and Autos and industries that have low earnings yields, are
Investment companies, Information Technology, Software, Transport and
Pharmaceuticals.3 The column tagged Qnt represents the average return of each quintile
ranked by earnings yield. It is apparent that overweighting the top quintile and shorting
the bottom quintile would have produced an excess return of about 2.6% pa. The
performance of the other quintiles produced similar returns which were in the range of
10.7% to 10.0%.4 Panel C displays similar results to Panel B but in this case industries
are ranked by implied growth. In this case the results are somewhat surprising in that the
higher growth rate industries are Water, Tobacco, Health and Insurance. These would
normally be regarded as more conservative industries with the possible exception of
Health. Low growth rate industries present no surprises with Forestry/Paper, Telecom
services, Electricity and Autos experiencing the lowest growth rates. Ranking industries
by implied growth rates illustrates that the quintile with the highest growth rate has an

2
For the sake of brevity we have not included the dividend yields in Table 3.
3
We found the earnings yield/performance of Water to be quite variable and were somewhat surprised by
the performance of this industry.
4
Ranking by dividend yield produced similar results.

7
average return of 13.3%, whilst the lowest growth rate quintile experienced an average
return of 9.2%. Returns for the other quintiles were in the range of 10.2% to 10.8%.
These results suggest going long the high growth quintile and shorting the low growth
quintile would have produced an excess return of 4.1% pa.

Insert Table 3

From a macroeconomic perspective the US economy plays an important role with regards
to the global economy. Harvey (1988) has noted the shape of the US yield can have
significant effect on global equity markets. Harvey (1988) and Estrella and Hardouvelis,
(1991) have found that the spread between the yield on 5/10 year US government bonds
and the US Treasury bills has significant forecasting ability in predicting future
movements in US GDP. Maximum forecasting ability is about twelve months in the
future. They have also noted that when the US yield curve inverts this is usually
associated with a recession in twelve months time. Furthermore, Harvey (1988) noted
that when the US yield curve inverts this can have a dramatic effect on global equity
markets. We examine this issue by noting the performance of industries when the US
yield curve inverts and when the US yield curve is normal.

Table 4 depicts the performance of industries when the US yield curve is inverted and
when the US yield curve is normal. When the US yield curve is normal the top five
performing industries are Tobacco, Water, Aerospace/Defence, Information Technology
and Food & Drug Retailers and the bottom five industries are Forestry/Paper, Mining,
Steel & other metals, Diversified Industries and Chemicals. The average return of the top
five industries was 15.2 pa% whilst the average return of the bottom five industries was
10.3 pa%. This represents a difference of 4.9% pa, which indicates that when the yield
curve is normal there is potential to profit from overweighting the top five industries and
under weighting the bottom five industries. On the other hand when the yield curve is
inverted the top five performing industries are Mining, Water, Tobacco, Oil and Gas and
Forestry/Paper, and the bottom five performing industries are Information Technology,
Media & Photography, Household Goods, Autos and Software. The average return of the

8
top five industries was 14.0% pa, whilst the average return of the bottom five industries
was -17.2% pa. This represents a difference of 31.2% pa. The potential to profit here is
quite substantial and seeing that in early 2006 the US yield curve has just inverted this
may be an opportune time to rebalance industry exposures.

Insert Table 4

2. Momentum

There is an extensive body of literature that has documented the profitably of momentum
based strategies. These studies have centred on stock specific, industry and country
momentum. In the US context, studies by Jegadeesh and Titman (1993, 2001) have
shown that buying top decile stocks ranked by intermediate prior returns (three to twelve
months) and shorting the bottom decile of stocks can produce excess returns of the order
of 12% pa. These results are also supported by Chan, Jegadeesh and Lakonishok (1996)
and Lee and Swaminathan (2000). Rouwenhorst (1998) examined similar momentum
strategies for European stocks and found significant momentum profits across twelve
European markets. Country momentum was examined by Richards (1997) and Chan,
Hameed and Tong (2000) who showed that the same type of strategy is profitable for
global country indices.

There is a growing body of literature that suggests that momentum effects are essentially
driven by industry effects not stock specific effects. Moskowitz and Grinblatt (1999)
found that industry momentum subsumes industry neutral momentum for the US. These
results were also supported by O’Neal (2000) who found strong momentum effects for
US Sector funds. Grundy and Martin (2001) on the other hand provide contrary evidence
to the findings of Moskowitz and Grinblatt (1999) suggesting that industry effects are not
the primary cause of momentum, but rather stock specific return components. Swinkels
(2002) and Scowcroft and Sefton (2005) have provided additional evidence to support
industry momentum in the global context.

9
We employ a similar approach by ranking industries on prior returns. Rather than focus
on one prior return measure we examine a number prior return measures. Industries are
ranked by, one month, three month, six month, nine month, twelve month, fifteen
months, twenty four months and forty eight months prior returns. The holding period in
our analysis is for one month. For example, industries ranked by prior twelve months
return are placed into deciles at t-1 and then held for the following month.5 The process is
then repeated for each successive month. Table 5 reports the results of a long/short
strategy which goes long the top decile and simultaneously goes short the bottom decile.6
Over the whole period the highest excess return was obtained from ranking the previous
twelve month returns. This is consistent with a number of other studies for US and
European stocks. As a general observation it is apparent that the nineties was a very good
decade for momentum strategies. This decade producing significant t values for all
decades except for the forty eight month ranking.7 The results from table 5 also indicate
that momentum performed very strongly from 1980-1999. In contrast, the worst
performing decades were from 1973-1979 and 2000-2005 with excess returns not
significantly different from zero (excepting for the twelve month ranking for 1973-1979).
8

Insert Tables 5

3. Value

Research has shown value strategies can earn positive returns over the long term. For
example, Fama and French (1992) have shown that book value to market value has
predictive ability in forecasting future stocks returns in the US. In a similar vain,
Lakonishok, Shleifer and Vishny (1994) found that cash flow to price had power to
forecast stock returns in the US. These results have been supported in the global context

5
Extensions to this approach often employed in the literature is ranking on k months prior returns and
holding these positions for one, three, six, nine and twelve months.
6
Results for decile performance are available from the authors on request. Generally speaking, top deciles
produced the greatest return, whilst bottom deciles produced the lowest return.
7
Throughout the paper we consider a strategy to be significant if the t value is greater than 1.65 (5% level).
8
As noted earlier the returns to Water were somewhat erratic. We repeated the analysis without water but
obtained similar results and hence are not reported.

10
by Fama and French (1998) who examined returns to value strategies for stocks included
in the MSCI developed markets index.

Value investing typically involves selecting stocks that have high earnings yield or high
dividend yields or high book to price ratios. In the following analysis we examine the
performance of ranking industries by earnings yield and normalized earnings yield. The
results for ranking industries by dividend yield are similar to those from earnings yield
and are not reported.9 Panel B from Table 3 reports the performance of industries sorted
by earnings yield. It is apparent that if one adopts a strategy of overweighting high
earnings yield industries then one will usually over weight industries such as Water,
Tobacco, Oil and Gas, Electricity, Aerospace/Defense, Auto and Parts, Mining and
underweight, Investment companies, Information Technology Hardware, Software and
Computer Services, Transport, Pharmaceuticals. As can be seen from Table 6 this type of
strategy can lead to prolonged periods of underperformance. For example, from 1974-
2005 the long short strategy of going long the top decile and shorting the bottom decile
produced an excess return of 6.5% pa. However during the nineties this strategy realized
a loss of 4.9% pa. Much of the out performance of this strategy over the whole sample
period was due to the reversal of fortune of the value strategy in the current decade where
the average excess return was 24.9% pa. As noted earlier this was the period when
momentum performed the worst.

Insert Table 6

To overcome the criticism of always overweighting/underweighting the same types of

industries we examined a strategy of ranking industries by their normalized earnings
yield (ZEYVAL). The logic here is that when ZEYVAL is positive relative to its historic
past then the industry is attractive and when ZEYVAL is negative then earnings yields
are low relative to historic past and are considered to be unattractive. Under this scenario
growth type industries such as Information Technology and Software could be included
in the long portfolios. The results of decile ranking are displayed in Table 7. It is apparent

9
These results are available from the authors on request.

11
the returns from this approach are erratic and do not out perform the earnings yield
approach.

Insert Table 7

4. Value and momentum combination Strategy

In the previous discussion relating to value and momentum strategies it is clear that both
strategies can earn positive excess returns. Yet these strategies appear to act quite
differently. In the nineties momentum performed exceptionally well, where as value
under performed. From 2000 onwards value performed well where as momentum profits
are not significantly different from zero. Asness (1997) has examined the interaction of
value and momentum for US stocks. In his analysis, value was measured as book value to
market value and the dividend yield. Momentum was measured as the prior twelve
months momentum but not including the most recent month. His results indicated that
both value and momentum strategies generated positive excess returns, but value seems
to work best with poor momentum stocks and momentum seems to work best with
expensive stocks. We examine these issues in the following discussion in relation to
global industries.

Since the correlation between long/short value and momentum strategies is -0.17 [refer
Table 11], there may be diversification benefits of trying to combine value and
momentum strategies. To address this issue we examine a combination strategy which
attempts to refine the momentum and value approaches by segregating the ranking of
each strategy into quartiles. For the period 1974-2005 we partition the rankings into:
value from Q1 (expensive) to Q4 (cheap), and momentum from Q1 (loser portfolio) and
Q4 (winner portfolios). Table 8 reports the performance of sorting value and momentum
into quartiles. It appears the greatest return of 19.5% pa was obtained from a combination
of good value and good momentum (Q4valQ4mom). The combination of poor value and
poor momentum (Q1valQ1mom) produced a return of 7.2% pa. This was not the worst
performing combination but was only marginally better than the worst performing

12
combination. A long/short strategy of Q4valQ4mom - Q1valQ1mom produced an excess
return of 12.3% pa with a T value of 3.8.

Insert Table 8

Other points of interest from Table 8 are that momentum seems to work well for the most
expensive industries. By focusing on the most expensive industries and sorting these
industries by momentum. A long/short strategy of going long the most expensive
industries (Q1val) within the winner momentum portfolio (Q4mom) and shorting the
most expensive industries with the weakest momentum (Q1val and Q1mom) yielded a
significant excess return of 9.3% pa. Significant results were also obtained when holding
the other value quartiles constant and combining this with winner/loser momentum
portfolios. From a momentum perspective there appears to be little benefit in sorting
winner momentum portfolios by value. The long/short strategy produced an insignificant
positive return of 3.1% pa. Consistent with the findings of Asness (1997) momentum
seems to work better for expensive industries and value seems to be work better by
focusing on poor momentum industries.

As noted earlier in relation to the momentum results, the nineties was an incredibly
successful decade for momentum strategies, and it is possible that the good performance
during this decade may overshadow the performance in other decades. We examine this
issue by comparing the performance of long/short strategies for each decade. Table 9
presents results of ranking industries into quartiles by value and momentum, but in this
case focusing on the performance of long/short strategies by decade. The performance of
expensive industries sorted by momentum can be quite varied across decades. For
example, the strategy of Q4momQ1val-Q1momQ1val produced insignificant excess
returns from1973-1979 and 2000-2005, marginally significant excess return of 7.9% pa
(t=1.68) for 1980-1989. During the nineties the excess return was a staggering 21% pa
(t=3.89). Overall we can conclude that momentum seems to work best for expensive
industries, but these results are driven mainly by the performance of momentum
strategies in the nineties.

13
Insert Table 9

The results are somewhat muddied for the value statement that value works best with
loser momentum portfolios. Over the whole sample period this statement is supported
however on a decade by decade basis the results are less strong. What does seem to work
consistently across differing sample periods is combining good value with good
momentum and shorting poor value and poor momentum. These results are reported in
the long/short combination column of table 12.

The long only strategy of investing in industries with good value and good momentum
out performs the passive benchmark strategy of investing in the MSCI developed market
index. From 1973- 2005 the total return of the MSCI index was 10.6% pa with standard
deviation of 14.5% pa, whilst the return of choosing industries that had a ranking greater
than 0.5 for value and momentum generated a return of 16.0% pa with a standard
deviation of 13.7% pa.10

5. Growth

From a theoretical basis growth in long term earnings clearly plays an important role in
determining the required rate of return of a stock. From the discounted dividend model
(DDM) the required rate of return is the dividend yield plus the long term growth rate of
dividends (or earnings given we know the payout ratio). Figure 1 displays a scatter plot
of implied growth against the return for each industry over the full sample period together
with a fitted regression. The adjusted R square is 46%, with a constant term of 5.2%
(t=5.0) and a slope coefficient of 0.64 (t=5.6). The results suggest that industries with
higher growth rates achieve higher returns.

Insert Figure 1

10
We chose a ranking of 0.5 because the volatility of the long only strategy was less than the volatility of
the MSCI index. Choosing a ranking of 0.75 for good value and good momentum generated a return of
19.5% pa with a standard deviation of 15.8% pa.

14
The results from Table 3 suggest that ranking industries by implied growth rates offered
potentially more profit than ranking by earnings yield. We adopt a similar approach in
this section where we rank industries by implied growth rates. We measure implied
growth by two different approaches. The first approach calculates the ex post implied
growth up to time t from the earnings index. We also calculate a rolling five year growth
rate as these numbers are often cited in the financial press. The second approach is to
calculate the ex ante implied growth rate from the DDM. This is a forward looking
estimate of future growth rates and may be a better indication of future growth. We
examine two cases, ranking industries by implied growth rates and overweighting high
growth industries and underweighting low implied growth rate industries. Secondly,
rather than just overweight industries with the highest growth rates we look for industries
that have increasing growth potential by normalizing the current implied growth rate with
respect to historic average over the previous 36 month.

Figure 2 depicts the ex ante implied growth rate calculated from DDM of the
Aerospace/Defense industry. The average implied growth rate from 1973-2005 was 8.3%
pa, and it is quite evident the implied growth can vary substantially. The implied growth
rate of Aerospace/Defense as of December 2005 is 10.1% pa from the DDM, and closely
matches the ex post implied growth of 9.1% pa estimated from earnings in Table 3.

Panel A of table 10 presents results of ranking industries by implied growth rates from
actual earnings. It is apparent there is no value in sorting industries by this measure as all
long/short results for differing sample are not significantly different zero. Also presented
in panel A is the case where industries are ranked by their previous five year growth
rates. Results in this case are also not significantly different from zero.

Insert Tables 10

Panel B of table 10 reports results similar to panel A, but in this case industry growth
rates are determined from the discounted dividend model. From 1973-1999 the long/short

15
strategy is not significantly different from zero. However from 2000-2005 low growth
substantially out performed high growth by 25%. Contrary to expectations that high
growth rates lead to higher returns we in fact obtain the reverse. Examining the strategy
by decade provides mixed results. Low growth (Dec 1) outperforms high growth (Dec10)
from 1973-1979 and 2000-2005, whilst the results are not significantly different from
zero for 1980-1989 and 1990-1999. These results indicate that overweighting industries
with high implied growth rates does not appear to lead to out performance.

Also reported in Panel B of table 10 are results of going long industries that have
increasing normalised earnings growth and shorting those industries that have decreasing
normalised earnings growth. Over the entire sample period the average excess return was
4.5% pa with a T value of 3.1. Each decade generated a significant positive excess return
excepting for the current decade where the excess return was not significantly different
from zero.

6. Aggregate Strategy

From the previous analysis it was apparent that most strategies broke down at some point
in time. In relation to long/short strategies, value produced negative excess returns from
1990-1999. Momentum produced insignificant excess returns from 1973-1979 (except for
twelve month momentum) and 2000-2005. The combination strategy of being long good
value and good momentum and shorting poor value and poor momentum produced
significant excess returns for all decades excepting from 1980-1989 (significant at the
7.5% level). Growth generated insignificant excess returns from 2000-2005. We now
examine whether there is any benefit in aggregating the strategies to form a portfolio of
strategies, as relying on anyone strategy can be problematic owing to inconsistent
performance. The strategies which are included in the aggregate strategy are: momentum
[twelve month momentum top decile (Dec10) and bottom decile (Dec1)]; value [earnings
yield top decile (Dec10) and bottom decile (Dec1)]; combination [good value and
momentum rank above 0.75 and poor value and poor momentum rank below 0.25],

16
growth [high growth (Dec10) and low growth (Dec1), based on normalised ex ante
growth].

Table 11 presents the correlation matrix of the long only strategies and the long/short
strategies. The correlations between the long only strategies are reasonably high and
there appears little potential for diversification across the strategies. What is likely
happening here is the correlations are picking up the market factor. We obtain a better
sense of diversification by examining the correlations from the long/short strategies. In
this case, it is quite evident there is much more diversification. As expected there is a
negative correlation between value and momentum. There are also low correlations
between growth and all other strategies which suggests that combining the strategies may
lead to considerable diversification.

Insert Table 11

Panel A of table 12 presents the performance of the individual long only strategies and
the equally weighted aggregation of these strategies.11 The long only aggregate strategy
provides some reduction in volatility but is able to capture most of the return. The
aggregate long only strategy substantially outperforms the MSCI index. Over the entire
sample period, the average excess return was about 6.9% pa with a lower standard
deviation than the MSCI index. In each decade the aggregate strategy out performed the
MSCI index by 8.3% pa (t=3.8), 6.8% pa (t=3.3), 3.2% pa (t=1.6) and 11.7% pa (t=3.4)
for 1973-1979, 1980-1989, 1990-1999 and 2000-2005 respectively.

Insert Table 12

Panel B of table 12 reports the performance of the long/short strategies and the equally
weighted aggregate strategy. The aggregate strategy typically has lower volatility than
the majority of the individual strategies and higher t values. The aggregate strategy

11
We have not tried to optimize the weights given to each strategy. Potentially returns could be improved
by following this approach.

17
produced average excess returns of 11.0% pa (t=4.8), 7.5% pa (t=2.9), 8.1% pa (t=2.8)
and 11.2% pa (t=2.0) for 1973-1979, 1980-1989, 1990-1999 and 2000-2005 respectively.
Over the entire sample period the average excess return was 9.0% pa with a standard
deviation of 9.3%.

7. US Yield Curve Effect

As noted earlier from the results of table 4 movements in the US yield curve can have a
significant effect on industry returns. We explore this issue further by examining the
performance of each of the strategies when the US yield curve is normal and when the
US yield curve is inverted. Table 13 reports the performance of each strategy for
differing sample periods when the US yield curve is normal and inverted. When the curve
is normal, momentum produces the highest average excess return with results being
significant at the 1% level for every decade except from 2000-2005. The next best
performing strategy was the combination strategy which produced significant excess
returns for every decade excepting 1980-1989. Overall the worst performing strategy was
value. On the other hand when the US yield curve was inverted the worst performing
strategy was momentum. Much of this poor performance was due to under performance
from 2000-2005. From 1974-1999 the average excess return was not significantly
different from zero. Not surprisingly the best two performers were value and the
combination strategies. This seems to support the notion of a value premium which
comes into play when expectations are of a contracting economy or potential recession.

The implication of these results suggest that when the US yield curve is normal, the
weight given to the value should be reduced and the weight to momentum and
combination should be increased. Alternatively when the US yield curve is inverted the
weight of momentum and growth should be reduced and the weight to value and
combination should be increased.

18
For example, when the US yield curve is normal a potential weighting scheme could be
to reduce weights to value and growth from 25% to 15% each, and increase the weights
of momentum to 40% and combination to 30%.12 Similarly when the yield curve is
inverted both momentum and growth do not perform well where as value and
combination perform well. In case, weights for momentum and growth could be 0% and
weights to value and combination could be 50% each. In each decade the t values
increase on average by 11%, and over the whole period the t value increases from 5.5 to
6.2. Changing the weights to take into account movements in the US yield can lead to
improved performance over the equally weighted aggregate strategy, however
implementing this type of strategy assumes this phenomenon will continue in the future.

Insert Table 13

8. Conclusion

Recent evidence has established that industry effects are at least as important as country
effects in explaining global returns. In fact Scowcroft and Sefton (2005) have shown that
approximately 60% of total returns can be explained by industry effects. However these
findings pertain mostly to the last ten or so years and earlier evidence indicated that
country effects were dominant. Prior to 1999 it made more sense for a global portfolio
manager to take active country bets as opposed to active industry bets. However, with the
increasing importance of industry effects we examined a number of commonly used
investment strategies in relation to industry rotation. The strategies examined were:
momentum, value, combination of value and momentum and growth. It was found that
each of the above mentioned strategies can earn significant positive excess returns. Yet
these strategies appear to act quite differently. Each of the strategies broke down at some
point in time. For example, momentum produced significant positive excess returns from
1980-1999 and insignificant excess returns from 1973-1979 and 2000-2005. Value
produced significant positive excess returns from 1973-1979 and 2000-2005 and

12
We have not endeavored to optimize the weights, but generally speaking reducing exposure to
momentum when the US yield curve is inverted and reducing exposure to value when the US yield curve is
normal leads to higher t values than the equally weighted strategy.

19
insignificant excess returns otherwise. The combination strategy produced significant
positive excess returns for all decades excepting 1980-1989. Similarly growth produced
significant positive excess returns for each decade excepting for 2000-2005.

It was found that by combining the strategies into a portfolio of strategies we are able to
generate more consistent performance. An equally weighted aggregate long/short strategy
produced significant positive excess returns for each decade.

20
References

Asness C., (1997), “The interaction of value and momentum strategies”, Financial
Analysts Journal, 53(2), 29-36.

Baca, S., B. Garbe, and R. Weiss, (2000), “The Rise of Sector Effects in Major Equity
Markets,” Financial Analysts Journal, 56(5), 34-40.

Beckers S., G. Connor and R. Curds, (1996), “National versus global influences on equity
returns, Financial Analysts Journal, 52(2), 31-39.

Cavaglia S., C. Brightman and M. Aked, (2000), “The increasing importance of industry
factors”, Financial Analysts Journal, 56(5), 41-54.

Chan L., N. Jegadeesh and J. Lakonishok (1996), “Momentum Strategies”, Journal of

Finance, 51, 1681-1713.

Chan K., A. Hameed and W. Tong, (2000), “Profitability of momentum strategies in the
international equity markets”, Journal of Financial and Quantitative Analysis, 35, 153-
172.

Estrella A. and G. Hardouvelis, (1991), “The term structure as a predictor of real

economic activity”, Journal of Finance, 46, 555-576.

Fama E. and K. French, (1992), “The cross section of expected returns”, Journal of
Finance, 47, 427-465.

Fama E. and K. French, (1998), “Value versus Growth: The International Evidence”,
Journal of Finance 53, 1975-1999

21
Griffin J. and G. Karolyi, (1998), “Another Look at the Role of the Industrial Structure of
Markets for International Diversification Strategies”, Journal of Financial Economics, 50,
351-373.

Grinold, R. A. Rudd and D. Stefek, (1989), “Global factors: fact or fiction”, Journal of
Portfolio Management, 16(1), 79-88.

Grundy B. and J. Martin, (2001), “Understanding the nature of the risks and the source of
the rewards to momentum investing”, Review of Financial Studies, 14, 29-78.

Harvey C., (1988), “The Real Term Structure and Consumption Growth," Journal of
Financial Economics, 22, 305-333.

Heston S. and G. Rouwenhorst, (1994), Does Industrial Structure explain the benefits of
international diversification, Journal of Financial Economics, 36, 3-27.

Heston S. and G. Rouwenhorst, (1995), Industry and Country Effects in International

Stock Returns, Journal of Portfolio Management, 21, 53-58.

Jegadeesh N. and S. Titman (1993), “Returns to buying winners and selling losers:
Implications for stock market efficiency”, Journal of Finance, 48, 65-91

Jegadeesh N. and S. Titman (2001), “Profitability of momentum strategies: An evaluation

of alternative explanations”, Journal of Finance, 56, 699-720.

Lakonishok J., A. Shleifer and R. Vishny, (1994), “Contrarian Investment, extrapolation,

and risk”, Journal of Finance, 49, 1541-1578.

Lee C. and B. Swaminathan (2000), “Price momentum and trading volume”, Journal of
Finance, 55, 2017-2069.

22
Lessard D., (1974), “World, national and industry factors in equity returns”, Journal of
Finance, 29, 379-391.

Moskowitz T. and M. Grinblatt, (1999), “Do Industries explain momentum?” Journal of

Finance, 54, 1249-1290.

O’Neal E., (2000), “Industry momentum and sector mutual funds”, Financial Analysts
Journal, 56(4), 37-49.

Phylaktis K. and L. Xia, (2003), “The changing role of industry and country effects in
global equity markets”, Discussion Paper, Emerging Markets Group, Cass Business
School.

Richards A. (1997), “Winner-loser reversals in national stock market indices: Can they be
explained?” Journal of Finance, 52, 2129-2144.

Roll R. (1992), “Industrial structure and the comparative behaviour of international stock
market indices, Journal of Finance, 47, 3-41.

Rouwenhorst G. (1998), “International Momentum strategies”, Journal of Finance, 53,

267-284.

Scowcroft A. and J. Sefton, (2005), “Understanding momentum”, Financial Analysts

Journal, 61(2), 64-82.

Solnik B., (1974), “The international pricing of risk: an empirical investigation of the
world capital market structure”, Journal of Finance, 29, 365-378.

Swinkels L., (2002), “International industry momentum”, Journal of Asset Management,

3, 124-141.

23
 Table 1 - Performance of Global Industries
Industry 73-05 73-79 80-89 90-99 00-05
Return Stdev Min Max Return Stdev Min Max Return Stdev Min Max Return Stdev Min Max Return Stdev Min Max
AEROSPCE,DEFENCE 13.1 20.1 -5.4 4.0 19.6 23.9 -5.4 6.9 11.8 21.2 4.0 21.1 13.1 16.1 -4.0 5.5 7.8 19.9 -9.0 4.0
AUTO & PARTS 8.7 17.6 1.0 3.1 4.7 19.0 1.0 4.5 16.8 17.1 0.6 2.5 6.4 16.8 -3.3 10.3 3.8 18.3 -6.8 3.1
BANKS 11.4 18.1 4.2 9.3 4.4 16.2 4.8 9.3 23.3 19.4 4.1 6.1 6.1 18.9 -2.1 0.1 8.3 15.9 -7.6 4.2
BEVERAGES 10.9 16.3 0.0 0.8 -2.5 20.9 0.0 3.8 22.5 14.7 2.0 3.2 11.1 15.9 -6.9 -5.7 6.7 11.9 -0.3 0.8
CHEMICALS 10.1 16.5 1.9 4.0 6.1 17.7 1.9 2.9 18.2 16.6 3.2 5.2 6.6 16.0 -3.2 5.1 7.1 16.0 -4.0 4.0
CONS.&BLDG MAT. 10.4 18.2 4.3 4.7 6.9 18.8 4.1 4.7 22.3 18.2 1.8 6.1 -0.1 19.0 -2.5 4.4 11.9 15.5 -6.1 4.3
DIVERSIFIED INDS 9.3 16.8 5.3 6.6 5.0 18.4 4.3 6.6 16.7 16.5 3.6 8.3 10.3 15.1 -1.8 10.1 0.2 17.7 -1.6 5.3
ELECTR. EQUIP. 11.3 19.0 1.6 8.0 6.4 17.6 1.6 6.0 18.6 17.0 4.4 5.1 14.0 18.4 -0.6 16.5 0.3 23.8 -1.9 8.0
ELECTRICITY 10.8 14.1 0.8 2.3 5.7 15.1 0.7 2.3 20.1 15.8 -1.6 2.4 5.6 12.3 -6.6 -2.0 10.0 12.0 -1.0 0.8
ENG.&MACHINERY 8.9 18.4 -0.7 6.0 6.6 15.6 -0.7 4.8 17.4 17.7 2.8 3.4 2.4 21.1 -3.8 5.8 8.2 17.7 -4.2 6.0
FOOD PRDR./PRCR. 12.1 13.8 0.3 2.8 7.6 15.5 2.8 3.6 21.7 14.6 4.2 4.6 7.3 12.4 -4.5 -1.3 8.9 12.0 -9.4 0.3
FOOD&DRUG RETLRS 12.8 16.0 0.1 1.6 7.9 21.7 0.1 4.0 22.4 14.8 1.3 3.6 15.2 12.7 -3.6 0.8 -2.0 14.4 -8.9 1.6
FORESTRY&PAPER 8.0 20.3 -2.8 6.4 8.1 23.2 -2.8 4.6 14.8 21.3 4.0 8.3 4.5 17.2 -6.3 8.4 2.5 19.8 -9.3 6.4
GAS DISTRIBUTION 11.6 18.0 -1.9 -0.3 9.1 14.8 -1.9 7.4 21.8 22.2 2.9 3.1 6.2 15.7 -5.8 3.3 6.6 16.9 -0.3 7.4
HEALTH 11.4 17.5 -3.2 3.1 -2.0 20.7 -3.2 4.1 17.0 18.4 -1.7 3.3 15.0 17.0 -7.5 -4.1 11.5 11.1 1.0 3.1
H'HLD GDS&TEXTLS 8.9 19.3 4.8 5.1 2.8 20.5 4.8 7.1 20.1 19.0 -1.1 3.4 6.9 19.5 -0.9 17.6 0.3 17.4 -10.6 5.1
INF.TECHN.HARDW. 8.7 24.3 -0.9 7.7 0.5 19.1 -0.9 1.4 14.0 18.6 -0.4 5.2 23.3 22.0 2.0 17.3 -15.2 37.4 -2.0 7.7
INSURANCE 11.7 16.6 4.6 5.2 7.9 16.0 2.5 5.2 20.5 16.9 1.0 1.8 10.0 15.9 -5.0 4.8 4.4 17.9 -7.8 4.6
INVESTMENT COS. 10.7 16.8 2.3 4.7 5.2 20.2 2.3 3.4 16.6 17.0 5.3 6.5 11.6 13.1 -1.9 8.0 5.5 17.9 -3.6 4.7
LEIS.,ENT.&HOTEL 10.9 20.6 6.3 6.5 1.5 27.6 5.0 6.5 22.7 18.8 -0.1 9.4 10.6 16.1 -4.9 8.6 2.2 20.4 2.3 6.3
LIFE ASSURANCE 12.8 18.3 1.9 5.1 6.2 20.6 -0.6 1.9 21.5 17.8 5.8 11.2 14.0 15.4 5.5 6.8 3.7 20.2 -12.8 5.1
MEDIA&PHOTOGRA. 8.3 17.8 -1.5 1.5 -4.6 20.7 -1.5 2.8 19.9 16.8 3.9 4.0 15.0 14.0 -4.6 16.0 -7.4 19.9 -0.2 1.5
MINING 11.2 26.0 6.0 6.0 12.4 24.2 6.0 18.9 11.8 30.4 4.9 11.4 5.2 23.7 4.0 11.8 18.7 24.1 -10.4 6.0
OIL & GAS 12.9 17.6 -4.2 2.7 12.4 16.8 -4.2 -1.4 14.3 20.4 8.2 10.9 11.0 15.4 -2.5 1.7 14.5 17.3 -5.6 2.7
PER.CARE&HSHLD 10.3 15.5 -0.6 2.4 -4.0 18.1 -0.6 0.9 18.7 14.7 -1.0 3.8 16.1 15.5 -6.8 3.2 2.9 12.6 -7.8 2.4
PHARMACEUTICALS 11.8 16.2 -0.6 3.5 1.8 20.7 0.8 3.5 20.5 15.3 -2.2 0.9 16.1 14.9 -7.4 -2.6 1.4 12.9 -0.6 2.1
REAL ESTATE 9.7 21.7 2.9 16.1 3.3 27.3 5.3 16.1 18.5 22.3 -1.4 11.2 3.9 20.4 -4.5 4.1 12.0 13.8 -3.4 2.9
RETAIL, GENERAL 9.8 18.3 -2.8 4.5 -4.1 20.7 -2.8 0.7 20.9 18.1 1.1 4.6 13.8 16.1 -4.4 8.8 0.7 18.6 -15.5 4.5
SFTWR&COMP.SERV. 10.5 25.7 -2.4 6.0 -1.1 26.7 -2.4 -0.3 11.2 23.1 3.5 3.6 32.8 21.7 -1.3 24.5 -14.9 32.3 -10.0 6.0
SPC&OTH. FINANCE 11.3 23.2 2.6 5.1 1.5 18.5 2.6 3.4 28.2 27.1 1.0 1.8 5.0 24.2 -8.7 1.2 5.0 17.8 -1.4 5.1
STEEL&OTH.METALS 8.6 23.0 1.5 6.7 4.8 20.9 1.5 11.3 21.2 24.4 -2.7 19.8 -3.6 22.4 -4.5 7.6 12.5 23.6 -6.0 6.7
SUPPORT SERVICES 9.2 17.6 -5.3 4.8 -2.0 19.8 -5.3 3.4 20.5 18.3 3.8 5.9 12.9 14.8 -5.9 4.1 -3.0 17.1 -7.2 4.8
TELECOM SERVICES 9.9 17.6 -3.8 -0.3 6.5 12.4 -3.8 -1.7 19.0 17.6 -2.6 5.3 16.1 16.8 -7.9 11.7 -12.1 22.2 -5.3 -0.3
TOBACCO 16.1 20.5 -2.0 3.8 5.7 20.2 1.4 3.8 26.4 19.9 1.5 2.4 9.3 20.2 -8.2 -6.1 22.3 21.8 -6.3 -2.0
TRANSPORT 9.6 16.6 4.1 5.1 6.0 14.5 2.9 5.1 22.3 17.6 1.1 2.9 0.3 18.3 -6.5 0.8 8.0 13.0 -6.8 4.1
WATER 16.3 20.6 0.4 3.1 7.6 21.6 -0.4 0.4 27.0 23.8 4.4 24.9 13.1 18.8 -8.4 7.2 13.6 15.4 -5.6 3.1

24
 Table 2 - Top and Bottom Performing Global Industries for Differing Sample Periods
Industry 73-05 Industry 73-79 Industry 80-89 Industry 90-99 Industry 00-05
Av Return Av Return Av Return Av Return Av Return
Top 4 Water 16.3 Aerospace/Defense 19.6 SPC&Oth. Finance 28.2 Software 32.8 Tobacco 22.3
Tobacco 16.1 Oil & Gas 12.4 Water 27 Inf. Techn.Hardw. 23.3 Mining 18.7
Aerospace/Defense 13.1 Mining 12.4 Tobacco 26.4 Telecom Serv. 16.1 Oil & Gas 14.5
Oil & Gas 12.9 Gas Distribution 9.1 Banks 23.3 Pharmaceuticals 16.1 Water 13.6

Bottom 4 Forestry& Paper 8 Media & Photography -4.6 Software 11.2 Steel&Oth. Metals -3.6 Inf. Techn.Hardw. -15.2
Media & Photography 8.3 Retail,General -4.1 Aerospace/Defense 11.8 Cons&Bldg Mat. -0.1 Software -14.9
Steel&Oth. Metals 8.6 Per. Care&Hshld -4 Mining 11.8 Transport 0.3 Telecom Serv. -12.1
Inf. Techn.Hardw. 8.7 Beverages -2.5 Inf. Techn.Hardw. 14 Eng&Machinery 2.4 Media & Photography -7.4

25
 Table 3 - Industry Characteristics in Relation to Earnings Yield and Implied Growth

Panel A Panel B Panel C

Industry EY Return Imp Gwt Trnd Gwt Industry EY Return Qnt Industry Return Imp Gwt Qnt
AEROSPCE,DEFENCE 8.2 13.1 9.1 7.1 WATER 14.50 16.3 WATER 16.3 13.5
AUTO & PARTS 8.0 8.7 5.7 5.1 TOBACCO 9.1 16.1 TOBACCO 16.1 12.0
BANKS 7.2 11.4 9.7 9.3 OIL & GAS 8.8 12.9 HEALTH 11.4 11.6
BEVERAGES 5.6 10.9 9.7 10.0 ELECTRICITY 8.4 10.8 INSURANCE 11.7 11.4
CHEMICALS 6.7 10.1 7.9 5.7 AEROSPCE,DEFENCE 8.2 13.1 LIFE ASSURANCE 12.8 11.1
CONS.&BLDG MAT. 6.5 10.4 8.2 7.8 AUTO & PARTS 8.0 8.7 FOOD&DRUG RETLRS 12.8 10.9
DIVERSIFIED INDS 7.2 9.3 7.0 6.5 MINING 7.8 11.2 12.7 PHARMACEUTICALS 11.8 10.8 13.3
ELECTR. EQUIP. 5.6 11.3 8.5 7.9 LIFE ASSURANCE 7.3 12.8 LEIS.,ENT.&HOTEL 10.9 10.4
ELECTRICITY 8.4 10.8 5.3 4.1 TELECOM SERVICES 7.3 9.9 SFTWR&COMP.SERV. 10.5 10.3
ENG.&MACHINERY 5.5 8.9 6.6 5.5 BANKS 7.2 11.4 INVESTMENT COS. 10.7 10.3
FOOD PRDR./PRCR. 7.1 12.1 8.7 8.3 DIVERSIFIED INDS 7.2 9.3 SPC&OTH. FINANCE 11.3 9.9
FOOD&DRUG RETLRS 6.0 12.8 10.9 11.0 GAS DISTRIBUTION 7.1 11.6 BEVERAGES 10.9 9.7
FORESTRY&PAPER 6.8 8.0 4.2 3.3 FOOD PRDR./PRCR. 7.1 12.1 BANKS 11.4 9.7
GAS DISTRIBUTION 7.1 11.6 8.4 6.7 FORESTRY&PAPER 6.8 8.0 10.7 RETAIL, GENERAL 9.8 9.7 10.8
HEALTH 5.2 11.4 11.6 10.1 CHEMICALS 6.7 10.1 TRANSPORT 9.6 9.6
H'HLD GDS&TEXTLS 5.0 8.9 7.5 7.3 CONS.&BLDG MAT. 6.5 10.4 STEEL&OTH.METALS 8.6 9.4
INF.TECHN.HARDW. 4.7 8.7 8.2 7.0 INSURANCE 6.3 11.7 PER.CARE&HSHLD 10.3 9.4
INSURANCE 6.3 11.7 11.4 11.1 FOOD&DRUG RETLRS 6.0 12.8 AEROSPCE,DEFENCE 13.1 9.1
INVESTMENT COS. 3.8 10.7 10.3 11.6 PER.CARE&HSHLD 5.8 10.3 OIL & GAS 12.9 8.9
LEIS.,ENT.&HOTEL 5.0 10.9 10.4 9.9 SPC&OTH. FINANCE 5.8 11.3 FOOD PRDR./PRCR. 12.1 8.7
LIFE ASSURANCE 7.3 12.8 11.1 9.5 STEEL&OTH.METALS 5.8 8.6 10.7 SUPPORT SERVICES 9.2 8.7 10.8
MEDIA&PHOTOGRA. 5.4 8.3 6.6 6.5 BEVERAGES 5.6 10.9 REAL ESTATE 9.7 8.7
MINING 7.8 11.2 7.5 3.8 ELECTR. EQUIP. 5.6 11.3 ELECTR. EQUIP. 11.3 8.5
OIL & GAS 8.8 12.9 8.9 5.4 SUPPORT SERVICES 5.6 9.2 GAS DISTRIBUTION 11.6 8.4
PER.CARE&HSHLD 5.8 10.3 9.4 8.3 ENG.&MACHINERY 5.5 8.9 INF.TECHN.HARDW. 8.7 8.2
PHARMACEUTICALS 4.9 11.8 10.8 11.3 RETAIL, GENERAL 5.5 9.8 CONS.&BLDG MAT. 10.4 8.2
REAL ESTATE 5.1 9.7 8.7 9.1 MEDIA&PHOTOGRA. 5.4 8.3 CHEMICALS 10.1 7.9
RETAIL, GENERAL 5.5 9.8 9.7 8.9 HEALTH 5.2 11.4 MINING 11.2 7.5
SFTWR&COMP.SERV. 4.8 10.5 10.3 10.1 REAL ESTATE 5.1 9.7 10.0 H'HLD GDS&TEXTLS 8.9 7.5 10.2
SPC&OTH. FINANCE 5.8 11.3 9.9 10.8 H'HLD GDS&TEXTLS 5.0 8.9 DIVERSIFIED INDS 9.3 7.0
STEEL&OTH.METALS 5.8 8.6 9.4 5.7 LEIS.,ENT.&HOTEL 5.0 10.9 MEDIA&PHOTOGRA. 8.3 6.6
SUPPORT SERVICES 5.6 9.2 8.7 8.5 PHARMACEUTICALS 4.9 11.8 ENG.&MACHINERY 8.9 6.6
TELECOM SERVICES 7.3 9.9 5.0 3.2 TRANSPORT 4.9 9.6 AUTO & PARTS 8.7 5.7
TOBACCO 9.1 16.1 12.0 12.0 SFTWR&COMP.SERV. 4.8 10.5 ELECTRICITY 10.8 5.3
TRANSPORT 4.9 9.6 9.6 8.1 INF.TECHN.HARDW. 4.7 8.7 TELECOM SERVICES 9.9 5.0
WATER 14.5 16.3 13.5 10.9 INVESTMENT COS. 3.8 10.7 10.1 FORESTRY&PAPER 8.0 4.2 9.2

26
Table 4 - Global Industry Performance Relative to US Yield Curve
Normal Yield Curve Inverted Yield Curve
Industry Return Stdev Industry Return Stdev
TOBACCO 16.3 20.4 MINING 20.4 26.6
WATER 16.0 20.5 WATER 18.5 21.4
AEROSPCE,DEFENCE 15.0 19.1 TOBACCO 14.4 21.5
ELECTR. EQUIP. 14.7 18.6 OIL & GAS 9.9 20.3
SPC&OTH. FINANCE 14.1 23.6 FORESTRY&PAPER 6.6 22.2
FOOD&DRUG RETLRS 14.3 15.2 GAS DISTRIBUTION 2.8 18.5
LIFE ASSURANCE 14.3 18.0 LIFE ASSURANCE 2.3 20.1
LEIS.,ENT.&HOTEL 14.2 19.4 INSURANCE 2.2 17.2
INF.TECHN.HARDW. 14.2 23.3 FOOD&DRUG RETLRS 2.1 20.6
SFTWR&COMP.SERV. 14.1 24.8 AEROSPCE,DEFENCE 1.6 26.4
INVESTMENT COS. 13.8 16.1 FOOD PRDR./PRCR. 1.3 17.1
FOOD PRDR./PRCR. 13.6 13.2 PHARMACEUTICALS 1.2 19.8
OIL & GAS 13.5 17.2 HEALTH -0.1 22.0
BEVERAGES 13.5 15.8 CHEMICALS -0.5 17.9
BANKS 13.4 18.1 TRANSPORT -3.5 15.2
ELECTRICITY 13.3 13.8 DIVERSIFIED INDS -3.8 17.0
PHARMACEUTICALS 13.2 15.6 PER.CARE&HSHLD -4.1 19.4
HEALTH 13.2 16.7 BANKS -4.3 17.2
GAS DISTRIBUTION 13.0 17.9 CONS.&BLDG MAT. -5.8 19.7
INSURANCE 13.0 16.6 BEVERAGES -6.2 18.9
RETAIL, GENERAL 12.9 17.9 ELECTRICITY -6.2 15.4
CONS.&BLDG MAT. 12.6 17.9 TELECOM SERVICES -6.7 18.4
TELECOM SERVICES 12.4 17.4 STEEL&OTH.METALS -7.1 22.1
PER.CARE&HSHLD 12.4 14.9 SUPPORT SERVICES -9.0 21.3
REAL ESTATE 12.3 20.9 ENG.&MACHINERY -9.9 16.6
H'HLD GDS&TEXTLS 12.1 19.0 RETAIL, GENERAL -10.1 20.4
SUPPORT SERVICES 12.1 16.9 SPC&OTH. FINANCE -10.4 19.4
AUTO & PARTS 12.1 17.2 INVESTMENT COS. -10.9 20.7
MEDIA&PHOTOGRA. 11.9 17.4 REAL ESTATE -12.0 24.7
ENG.&MACHINERY 11.7 18.6 ELECTR. EQUIP. -12.0 20.3
CHEMICALS 11.6 16.3 SFTWR&COMP.SERV. -13.5 31.0
TRANSPORT 11.3 16.7 LEIS.,ENT.&HOTEL -13.5 26.7
DIVERSIFIED INDS 11.0 16.7 AUTO & PARTS -14.1 19.8
STEEL&OTH.METALS 10.9 23.2 H'HLD GDS&TEXTLS -14.5 20.2
MINING 9.6 26.0 MEDIA&PHOTOGRA. -15.7 19.0
FORESTRY&PAPER 8.4 20.0 INF.TECHN.HARDW. -28.3 29.0

27
 Table 5 Performance Long/Short Industry Momentum D10 - D1

1 mth 3 mth 6 mth 9 mth 12 mth 15 mth 24 mth 48 mth

74-79 Av Ret 7.89 3.47 -3.41 5.17 10.37 2.38 0.43 0.60
Std 12.52 13.43 14.51 13.36 11.79 14.39 13.85 12.63
T val 1.54 0.63 -0.58 0.95 2.15* 0.41 0.08 0.12

80-89 Av Ret 9.21 1.40 6.17 10.48 11.95 8.23 3.00 -2.26
Std 12.96 15.35 16.26 16.03 17.36 16.44 15.88 17.26
T val 2.27* 0.29 1.22 2.09* 2.20* 1.60 0.60 -0.42

90-99 Av Ret 7.74 16.08 10.86 16.46 21.86 18.27 14.47 6.46
Std 15.58 16.90 16.23 16.33 18.20 17.78 17.82 17.17
T val 1.59 3.05** 2.14* 3.22** 3.84* 3.29** 2.60** 1.20

00-05 Av Ret -3.89 0.54 3.59 3.03 3.74 1.37 -4.02 -3.16
Std 19.98 21.08 23.52 23.73 23.73 23.35 22.52 20.52
T val -0.48 0.06 0.37 0.31 0.39 0.14 -0.44 -0.38

74-05 Av Ret 6.07 6.23 5.36 9.98 13.23 9.00 4.81 0.84
Std 15.24 16.77 17.52 17.37 18.15 18.02 17.61 17.11
T val 2.27* 2.12* 1.74* 3.27** 4.16** 2.85** 1.56 0.28

74/99 Av Ret 8.34 7.53 5.76 11.56 15.40 10.74 6.82 1.76
Std 13.89 15.64 15.89 15.58 16.61 16.58 16.28 16.26
T val 3.12** 2.50** 1.89* 3.86** 4.82** 3.37** 2.18* 0.56
* significant at 5%
** significant at 1%

28
 Table 6 - Performance Earnings Yield EYVAL

DEC1 DEC2 DEC3 DEC4 DEC5 DEC6 DEC7 DEC8 DEC9 DEC10 L/S

74-79 Av Ret 6.82 9.89 1.97 4.39 10.30 11.16 12.32 9.22 17.47 17.50 10.69
Std 16.44 17.44 19.38 18.71 16.68 17.49 16.77 15.66 15.19 14.93 10.37
T value 2.53**

80-89 Av Ret 20.23 23.97 22.77 20.97 20.65 18.51 18.76 22.60 21.81 24.54 4.32
Std 17.96 19.34 14.97 16.23 17.21 15.97 14.73 14.84 13.84 14.33 13.23
T value 1.04

90-99 Av Ret 15.88 11.31 14.34 14.27 11.85 7.06 13.68 8.30 9.14 10.99 -4.89
Std 18.93 15.63 14.84 14.42 14.29 14.62 14.62 14.18 14.01 12.18 16.31
T value -0.96

00-05 Av Ret -6.31 0.16 -1.42 10.62 8.03 7.62 7.08 11.58 9.09 18.59 24.91
Std 21.35 15.34 13.48 14.14 13.49 14.87 16.11 13.68 15.18 15.21 18.85
T value 3.1**

74-05 Av Ret 11.42 12.94 11.73 13.83 13.61 11.52 13.79 13.56 14.66 17.87 6.45
Std 18.81 17.27 15.78 15.85 15.59 15.67 15.34 14.64 14.44 14.01 15.22
T value 2.42**

74/99 Av Ret 15.46 15.85 14.73 14.57 14.88 12.41 15.32 14.01 15.93 17.71 1.51
Std 18.01 17.59 16.16 16.23 16.02 15.85 15.16 14.87 14.27 13.75 19.97
T value 0.39
significant at 1%

29
 Table 7 - Performance Normalised Earnings Yield ZEYVAL

DEC1 DEC2 DEC3 DEC4 DEC5 DEC6 DEC7 DEC8 DEC9 DEC10 L/S

74-79 Av Ret 10.78 5.97 14.09 9.62 8.79 11.97 5.56 10.15 11.97 14.09 3.32
Std 16.91 16.28 16.56 15.68 17.73 16.85 16.33 16.18 19.94 17.61 13.18
T value 0.62

80-89 Av Ret 21.18 21.54 18.81 22.10 22.77 20.69 22.53 24.25 22.58 18.78 -2.40
Std 15.81 15.99 16.84 14.88 15.36 17.25 14.49 15.91 15.61 15.37 12.21
T value -0.63

90-99 Av Ret 14.06 17.67 12.21 14.39 12.73 12.31 15.79 7.89 6.90 4.24 -9.82
Std 14.21 14.09 13.25 14.46 14.20 13.65 15.51 16.13 16.25 16.18 11.74
T value -2.68**

00-05 Av Ret 6.95 3.87 6.72 4.15 6.03 4.86 7.09 7.00 7.07 11.03 4.08
Std 15.68 15.65 15.93 14.24 15.46 15.65 16.29 13.60 14.93 15.70 12.20
T value 0.78

74-05 Av Ret 14.36 14.13 13.61 14.01 13.89 13.49 14.37 13.28 12.80 11.90 -2.45
Std 15.53 15.49 15.56 14.85 15.55 15.84 15.57 15.72 16.64 16.15 12.31
T value -1.13

74/99 Av Ret 16.04 16.46 15.18 16.25 15.68 15.45 16.02 14.70 14.10 12.10 -6.63
Std 15.48 15.40 15.46 14.93 15.56 15.86 15.39 16.15 17.01 16.27 12.39
T value 2.78**

** significant at 1%

30
 Table 8 - Performance of Sorting Value and Momentum into Quartiles 1974-2005

VALUE

Q1 Q2 Q3 Q4 Return
Expensive Cheap Difference

Q1 (Loser quartile) Average Return 7.20 8.60 9.90 12.90 5.80

Std 18.80 16.90 17.30 15.70 15.70
T value 2.10**

Q2 Average Return 6.80 11.90 11.10 13.40 6.60

Std 17.40 16.90 15.00 14.70 14.70
T value 2.57**
MOMENTUM
Q3 Average Return 10.60 14.90 12.10 15.00 4.50
Std 16.20 15.20 14.40 15.10 12.50
T value 2.05*

Q4 (winner quartile) Average Return 16.40 15.40 13.90 19.50 3.10

Std 16.80 16.10 15.60 16.90 17.40
T value 1.00

Return Difference 9.30 6.90 4.00 6.60 12.30

Std 17.40 15.30 15.60 16.20 15.50
T value 3.05** 2.6** 1.50 2.33** 4.8**
* significant at 5%
** significant at 1%

Table 9 Long/Short Strategy of Combination Value and Momentum

Sorted by Quartiles
Q1 VAL Q2 VAL Q3 VAL Q4 VAL Q1 MOM Q2 MOM Q3 MOM Q4 MOM
(Q4-Q1)MOM (Q4-Q1)MOM (Q4-Q1)MOM (Q4-Q1)MOM (Q4-Q1)VAL (Q4-Q1)VAL (Q4-Q1)VAL (Q4-Q1)VAL

74-79 Av Ret -1.14 2.99 -1.99 4.73 8.45 9.06 7.88 14.32
Std 15.92 16.04 14.41 13.43 15.24 13.72 12.64 15.84
T value -0.18 0.46 -0.34 0.86 1.36 1.62 1.53 2.21**

80-89 Av Ret 7.92 6.06 0.95 8.00 4.19 -2.70 4.80 4.27
Std 14.91 14.76 15.02 18.61 15.93 16.05 13.86 19.17
T value 1.68* 1.30 0.20 1.36 0.83 -0.53 1.09 0.70

90-99 Av Ret 21.05 12.87 12.37 8.65 3.00 7.66 -0.99 -9.40
Std 17.12 15.00 16.13 16.18 15.25 13.67 11.76 16.80
T value 3.89** 2.71** 2.43** 1.69* 0.62 1.77 -0.27 -1.77

00-05 Av Ret 2.07 2.05 1.16 2.46 10.37 18.02 9.75 10.75
Std 22.00 15.89 16.51 14.66 16.91 14.64 11.23 16.19
T value 0.23 0.32 0.17 0.41 1.50 3.02** 2.13* 1.63

74-05 Av Ret 9.25 6.87 4.02 6.56 5.76 6.60 4.48 3.08
Std 17.39 15.29 15.58 16.21 15.74 14.73 12.53 17.44
T value 3.05** 2.58** 1.48 2.33** 2.10* 2.57** 2.05* 1.01

74/99 Av Ret 10.88 7.97 4.67 7.50 4.71 4.00 3.28 1.33
Std 16.17 15.15 15.38 16.55 15.48 14.68 12.80 17.69
T value 3.50** 2.73** 1.58 2.35** 1.58 1.42 1.33 0.39
* significant at 5%
** significant at 1%

31
 Table 10 - Performance of Long/Short Implied Growth Strategies

Ex Post Earnings Ex Ante Earnings

Panel A Panel B
Implied Gwth 5 year Gwth Implied Gwth Z Imp Gwth

74-79 Av Ret 1.45 -1.40 -8.65 6.05

Std 12.65 11.53 10.73 6.31
T Val 0.28 -0.30 1.97* 2.35**

80-89 Av Ret -2.17 2.11 -4.32 5.88

Std 14.06 12.62 13.23 7.23
T Val -0.49 0.54 -1.04 2.60**

90-99 Av Ret 2.06 1.56 4.89 4.14

Std 13.03 13.96 16.31 7.59
T Val 0.51 0.36 0.96 1.74*

00-05 Av Ret 5.14 1.44 -24.91 1.39

Std 11.30 15.64 18.85 11.55
T Val 1.11 0.23 -3.24 0.30

74-05 Av Ret 1.18 1.24 -5.98 4.46

Std 12.98 13.47 15.38 8.22
T Val 0.52 0.52 -2.20 3.1**

74/99 Av Ret 0.24 1.19 -1.50 5.20

Std 13.35 12.94 14.17 7.20
T Val 0.09 0.48 -0.55 3.75**

* significant at 5%
** significant at 1%

32
 Table 11 Correlation Matrix 1974-2005

Long Only Long/Short

Mom Val Combo Gwth Mom Val Combo Gwth

Mom 1.00 0.65 0.71 0.77 Mom 1.00 -0.17 0.48 0.23
Val 1.00 0.76 0.81 Val 1.00 0.42 0.09
Combo 1.00 0.73 Combo 1.00 0.30
Gwth 1.00 Gwth 1.00

Table 12 - Aggregate Strategy

Panel A LONG ONLY L/S Panel B Long/Short

Mom Val Combo Gwth Agglong MSCI Agglong-MSCI Mom Val Combo Gwth AggLS
Dec10 Dec10 D10-D1 D10-D1
WGT 0.25 0.25 0.25 0.25 0.25 0.25 0.25 0.25
74-79 Av Ret 16.50 17.50 21.69 16.14 16.95 8.68 8.27 10.37 10.69 18.36 6.14 10.98
Std 15.37 14.93 16.24 12.78 13.07 14.36 5.30 11.79 10.37 9.59 6.37 5.59
T val 3.82** 2.15* 2.53** 4.69** 2.36** 4.81**

80-89 Av Ret 28.29 24.54 26.39 25.29 26.13 19.38 6.75 11.95 4.32 7.97 5.88 7.53
Std 19.13 14.33 17.47 14.44 14.83 14.65 6.48 17.36 13.23 16.53 7.23 8.38
T val 3.33** 2.2* 1.04 1.54 2.6** 2.87**

90-99 Av Ret 23.61 10.99 14.39 13.30 15.57 12.33 3.24 21.86 -4.89 11.20 4.14 8.08
Std 16.17 12.18 13.87 13.44 11.91 13.98 6.61 18.20 16.31 15.66 7.59 9.23
T val 1.57 3.84** -0.96 2.29* 1.74 2.80**

00-05 Av Ret 8.55 18.59 13.85 8.08 12.30 0.94 11.73 3.74 24.91 14.64 1.39 11.17
Std 17.56 15.21 15.16 13.79 13.76 14.63 8.34 23.73 18.85 18.09 11.55 13.39
T val 3.44** 0.39 3.24** 1.98* 0.30 2.04*

74-05 Av Ret 20.92 17.87 19.43 16.64 18.53 11.71 6.86 13.23 6.45 12.17 4.47 9.03
Std 17.33 14.01 15.76 13.82 13.48 14.45 6.74 18.15 15.22 15.49 8.23 9.34
T val 5.8** 4.16** 2.42** 4.48** 3.09** 5.51**

74/99 Av Ret 23.77 17.71 20.69 18.71 19.95 14.20 5.75 15.40 2.25 11.61 5.21 8.54
Std 17.20 13.75 15.90 13.79 13.40 14.34 6.29 16.61 14.02 14.86 7.21 8.17
T val 4.76** 4.82** 0.83 4.06** 3.76** 5.43**
* significant at 5%
** significant at 1%

33
 Table 13 -Performance of Global Industry Strategies in relation to the US Yield Curve

Panel A CURVE INVERTED Panel B CURVE NORMAL

Mom Val Combo Gwth Mom Val Combo Gwth
74-79 Av Ret 5.02 12.86 30.53 1.56 12.15 9.97 14.30 5.51
Std 15.54 8.89 10.93 4.20 10.38 10.88 8.91 5.89
T val 0.79 3.54** 6.84** 0.91 2.87** 2.24* 3.93** 2.29*

80-89 Av Ret -6.37 11.05 13.59 10.22 16.06 2.81 6.71 4.90
Std 14.53 12.53 16.91 6.52 17.78 13.41 16.51 7.38
T val -1.40 2.82** 2.57** 5.02** 2.21* 0.51 1.00 1.63

90-99 Av Ret NA NA NA NA 21.86 -4.89 11.20 4.14

Std NA NA NA NA 18.20 16.31 15.66 7.59
T val NA NA NA NA 3.84** -0.96 2.29* 1.74*

00-05 Av Ret -41.04 66.84 28.36 -0.40 8.64 20.32 13.14 1.59
Std 32.03 25.46 22.86 14.14 22.52 17.75 17.67 11.36
T val -3.14** 6.43** 3.04** -0.07 0.94 2.80** 1.82* 0.34

74-05 Av Ret -7.17 20.05 22.28 5.32 16.09 4.55 10.76 4.09
Std 18.41 14.77 15.76 7.34 17.99 15.22 15.43 8.13
T val -2.22* 7.74** 8.06** 4.14** 5.10** 1.70 3.98** 2.87**

74/99 Av Ret -1.24 11.86 21.21 6.32 17.84 0.83 10.20 4.68
Std 14.89 10.91 14.56 5.67 16.76 14.39 14.88 7.18
T val -0.43 5.65** 7.57** 5.80** 5.50** 0.30 3.56** 3.39**
* significant at 5%
** significant at 1%

34
 Figure 1
Implied Growth vs Return
Return(t) = a + b * Implied Growth(t) + e(t)
a = 5.94 (t=5.0), b= 0.54 (t=5.6), R^2 = 0.46
18.0

16.0

14.0

12.0
Return %

10.0

8.0

6.0

4.0

2.0

0.0
0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0
Implied Growth%

35
 Implied Growth %
Ja
n-
7

0.0
2.0
4.0
6.0
8.0
10.0
12.0

Ja 5
n-
7
Ja 6
n-
7
Ja 7
n-
7
Ja 8
n-
7
Ja 9
n-
8
Ja 0
n-
81
Ja
n-
8
Ja 2
n-
8
Ja 3
n-
8
Ja 4
n-
85
Ja
n-
8
Ja 6
n-
87
Ja
n-
8
Ja 8
n-
8
Ja 9
n-
90
Ja
n-
9

Date
Ja 1
n-
9
Ja 2
n-
9
Ja 3
n-
9
Ja 4
n-
9
Ja 5
n-
96
Ja
n-
9
Ja 7
Figure 2 Ex Ante Implied Growth Aerospace/Defense

n-
9
Ja 8
n-
9
Ja 9
n-
0
Ja 0
n-
0
Ja 1
n-
0
Ja 2
n-
03
Ja
n-
0
Ja 4
n-
05

36