day 2,11,12,17-24 - drivers of expected returns in im (2)
DESCRIPTION
sdfsdfasdTRANSCRIPT
-
Draft
FALL 2000 EMERGING MARKETS QUARTERLY 1
What are the drivers of averagereturns in international mar-kets? Are the sources of riskthat impact developed mar-
ket returns the same as the risk factors thataffect emerging market returns? These are thequestions explored in this study.
The work is closely related to Harvey[1995a, 1995b] and to Estrada [2000], whoexamines a number of risk measures and relatesthese measures to emerging market returns.The difference is that we expand the list of riskmetrics, and examine both developed andemerging market returns.
There are many reasons to believe thatthe sources of risk in emerging and devel-oped markets may differ. Harvey [1995a]demonstrates that the pricing relationshipsthat show some positive success in developedmarkets fail in emerging markets. Bekaert andHarvey [1995] argue that emerging and devel-oped markets cannot be treated similarly. Theypropose an asset pricing framework that allowsfor time-varying integration. Indeed, the pres-ence of barriers to international investment asdocumented in Bekaert [1995] immediatelysuggests that risk factors will differ acrossdeveloped and emerging markets.
I examine a comprehensive list of 18 riskfactors, and analyze whether these risk factorsexplain expected returns in 47 different mar-kets. In much of the analysis, I present three setsof results: developed markets, emerging mar-
kets, and all markets. The idea is to determineif the same risk factors explain expected returnsin developed and emerging markets.
I. DATA AND RISK METRICS SUMMARY ANALYSIS
Sample
I examine average returns and risk in asample of 47 countries using data from Mor-gan Stanley Capital International (MSCI). Thesample consists of 28 emerging markets and 19developed markets. Emerging market data arealso available from the International FinanceCorporation (IFC). When I replicate a num-ber of the emerging markets analyses with theIFC data, I find little difference between theIFC and MSCI samples.
The data are sampled between January1988 and December 1999. There are a num-ber of emerging markets whose data do notbegin until later, but I include only marketsthat have a history at least from January 1995.
Risk Metrics
A variety of risk metrics are used toexplain the average returns. The notation isdetailed here. For example, the raw materialsI am working with are the average total returnsin U.S. dollars for country i, which I expressas E[Ri].
Drivers of Expected Returns in International Markets
CAMPBELL R. HARVEY
CAMPBELL R.HARVEYis the J. Paul Sticht professor of interna-tional business at Duke University, anda research associate with the NationalBureau of EconomicResearch in Cam-bridge, Massachusetts.
-
One-Factor Market Model. Using a world versionof a single-factor model, where Rmt denotes the return onthe MSCI world index, I estimate the regression:
Rit rft = ai + bi[Rmt rft] + eit (1)
where rft is the U.S. 30-day Treasury bill rate, and eit is theresidual. Also, note that emt = Rmt Avg(Rmt) is used later.
SR (systematic risk) is the beta, bi in equation(1). TR (total risk) is the standard deviation of countryreturn i. IR (idiosyncratic risk) is the standard devia-tion of the residual eit. The one-factor model is studiedin Harvey [1995a, 1995b] and Bekaert and Harvey[1995].
Size. For size (market capitalization), we take the nat-ural log of average market capitalization over the relevantperiod for each country. This variable is studied in Bekaert,Erb, Harvey, and Viskanta [1997] and Estrada [2000].
Semistandard Deviations. The formula for semi-standard deviation is:
(2)
Semimean is the semistandard deviation with B =average returns for the market. Semi-rf is the semistandarddeviation with B = U.S. risk-free rate. Semi-0 is thesemistandard deviation with B = 0.
These variables are studied in Estrada [2000].Downside Beta Measures. Down-iw is the coef-
ficient from the market model using observations whencountry returns and world returns are simultaneouslynegative. Down-w is the coefficient from the marketmodel using observations when world returns are nega-tive. The first variable is studied in Estrada [2000]. Thesecond beta has not been studied before.
Value at risk. VaR is a value at risk measure. It is thesimple average of returns below the 5th percentile level.This variable is examined in Estrada [2000].
Skewness. Skew is the unconditional skewness ofreturns. It is calculated by taking the Mean(ei
3) divided bythe [Standard Deviation of (ei)]
3. Skew 5%: [(Return atthe 95th Percentile level mean return) (Return at 5thPercentile Level Mean Return)] - 1.
Coskew1 represents coskewness definition 1. It iscalculated by (sum up ei em
2)/T and divide by [squareroot of (sum of (ei
2)/T)] [(sum of (em2)/T)].Coskew2
represents coskewness definition 2. It is calculated by
(Sum up ei em2)/T and divide by [standard deviation of
(em)]3.These variables are examined in Harvey and Sid-
dique [1999, 2000a, 2000b].Spread. Kurt is the kurtosis of the return distribution.
Kurtosis is documented in Bekaert and Harvey [1997] aswell as Bekaert, Erb, Harvey, and Viskanta [1998].
Political and Country Risk. ICRGC is the log of theaverage monthly International Country Risk Guides(ICRG) country risk composite. CCR is the log of theaverage semiannual country risk rating published by Insti-tutional Investor. ICRGP is the log of the average monthlyICRG political risk ratings.
These variables are studied in Erb, Harvey, andViskanta [1996a, 1996b] as well as Bekaert, Erb, Harvey,and Viskanta [1997].
II. SUMMARY STATISTICS
Exhibit 1A (emerging markets) and Exhibit 1B(developed markets) present some summary statistics forthe risk measures over the complete sample for eachcountry. We report the mean monthly U.S. dollar returnsas well as the geometric mean return and the uncondi-tional correlation with the MSCI world in addition to theaverage values of the risk measures.
The last row of each exhibit reports the average acrossall markets. We see the usual characteristics. Emergingmarkets are more volatile and generally have more risk onthe downside (as measured by the semivariance, VaR,downside betas). In addition, the coskewness measures foremerging markets are more negative, suggesting that thesemarkets are contributing to the negative skewness of adiversified portfolio.
Exhibits 2A-2C present the correlation matrices ofthe risk measures in developed countries, emergingmarkets, and all countries. There are some interestingfindings.
Total risk and idiosyncratic risk are 99% correlatedacross emerging markets and 93% correlated in devel-oped markets. The three semivariance measures all havemore than 90% correlation with total risk in both emerg-ing and developed markets. VaR is at least -86% corre-lated with total risk in emerging and developed markets.The skewness measures are relatively little correlatedwith the other risk measures.
2 DRIVERS OF EXPECTED RETURNS IN INTERNATIONAL MARKETS FALL 2000
Semi B R B Bt- = - < (1/ T) for all Rt=1T
t( )2
-
FALL 2000 EMERGING MARKETS QUARTERLY 3
Mar
ket
E[R
]m
Gr
SRT
RIR
Mca
pSi
zeSe
mim
ean
Sem
i - r
fSe
mi-
0D
own-
b
iwD
own-
b
wV
aRsk
ewsk
ew5%
cosk
ew1
cosk
ew2
kurt
CC
RIC
RG
P
Em
ergi
ng M
arke
ts
Arg
entin
a3.
602.
180.
110.
5118
.50
18.4
317
.60
2.87
10.3
68.
678.
451.
411.
93-2
8.41
2.10
0.34
-0.2
9-1
.33
11.0
364
.57
31.2
669
.17
Jan-
88B
razi
l3.
541.
840.
301.
4018
.32
17.5
259
.22
4.08
12.2
310
.77
10.5
61.
803.
19-3
7.25
0.57
0.20
-0.3
0-1
.31
6.18
62.9
831
.76
65.8
9Ja
n-88
Chi
le2.
141.
840.
270.
527.
747.
4723
.15
3.14
5.39
4.47
4.27
1.26
1.09
-13.
390.
070.
19-0
.27
-0.5
14.
0273
.17
48.2
969
.56
Jan-
88C
hina
0.29
-0.5
30.
080.
3113
.43
13.4
71.
070.
077.
807.
857.
610.
20-0
.06
-19.
421.
180.
530.
020.
075.
1172
.83
57.4
967
.63
Jan-
93C
olom
bia
0.35
-0.0
80.
170.
439.
369.
305.
971.
796.
366.
386.
161.
181.
24-2
0.90
0.20
-0.0
6-0
.28
-0.5
54.
5063
.02
41.2
655
.43
Jan-
93C
zech
Rep
ublic
0.12
-0.2
60.
270.
608.
638.
398.
712.
166.
526.
666.
461.
501.
51-2
0.28
0.08
-0.0
7-0
.42
-0.8
04.
0780
.11
56.8
382
.28
Jan-
95
Egy
pt2.
362.
020.
180.
408.
668.
605.
201.
654.
783.
633.
410.
570.
64-1
0.41
1.49
0.45
-0.2
5-0
.49
6.02
69.9
330
.52
62.7
5Ja
n-95
Gre
ece
2.34
1.77
0.22
0.63
11.4
111
.15
14.0
42.
646.
435.
385.
170.
861.
55-1
5.91
1.92
0.16
-0.1
9-0
.54
8.74
69.2
849
.78
70.5
1Ja
n-88
Hun
gary
2.84
2.08
0.54
1.74
12.3
910
.54
6.08
1.81
8.55
7.41
7.24
1.79
2.57
-25.
461.
430.
11-0
.43
-1.0
56.
1276
.52
46.3
482
.09
Jan-
95In
dia
0.95
0.59
0.17
0.41
8.59
8.52
50.3
13.
925.
715.
375.
140.
400.
68-1
3.71
0.36
0.27
-0.0
9-0
.16
2.60
65.9
244
.42
61.2
0Ja
n-93
Indo
nesi
a2.
060.
720.
231.
0117
.62
17.2
115
.38
2.73
10.0
19.
158.
931.
151.
18-3
0.20
2.16
-0.0
10.
010.
0411
.61
61.3
146
.88
53.7
9Ja
n-88
Isra
el0.
890.
650.
400.
746.
856.
3316
.38
2.80
5.02
4.75
4.57
0.23
0.80
-14.
07-0
.24
-0.0
8-0
.04
-0.0
52.
9570
.41
43.1
164
.59
Jan-
93Jo
rdan
0.16
0.05
0.12
0.15
4.58
4.58
1.04
0.04
3.20
3.36
3.11
0.04
0.26
-10.
29-0
.20
0.27
-0.0
7-0
.08
4.56
63.8
729
.26
60.7
1Ja
n-88
Kor
ea1.
150.
440.
371.
1612
.47
11.6
468
.76
4.23
7.49
7.09
6.86
0.15
0.79
-23.
401.
530.
490.
080.
229.
5577
.77
66.6
172
.34
Jan-
88M
alay
sia
1.12
0.60
0.48
1.26
10.3
59.
1071
.31
4.27
7.06
6.71
6.51
1.46
1.91
-21.
360.
56-0
.10
-0.1
7-0
.38
7.02
75.0
561
.73
70.0
5Ja
n-88
Mex
ico
2.68
2.09
0.41
1.11
10.7
49.
8262
.36
4.13
7.98
6.88
6.70
1.11
1.74
-23.
89-0
.28
0.00
-0.1
4-0
.35
4.49
68.0
539
.90
68.7
4Ja
n-88
Mor
occo
1.58
1.48
-0.3
1-0
.37
4.54
4.35
4.64
1.54
3.01
2.34
2.13
1.11
-0.3
2-6
.79
0.56
-0.1
0-0
.15
-0.1
54.
3771
.39
33.6
868
.87
Jan-
93Pa
kist
an0.
22-0
.48
0.12
0.42
11.8
411
.83
5.07
1.62
8.06
8.15
7.94
-0.3
00.
56-2
6.80
0.03
0.37
0.12
0.30
4.60
58.0
228
.33
53.9
3Ja
n-93
Peru
1.30
0.79
0.41
1.15
10.1
19.
266.
941.
946.
986.
526.
351.
712.
34-2
2.71
0.58
0.01
-0.4
3-0
.84
5.48
61.3
420
.62
56.4
5Ja
n-93
Phili
ppin
es1.
210.
720.
461.
1710
.05
8.96
14.3
52.
666.
726.
326.
091.
682.
01-2
0.63
0.22
0.06
-0.1
8-0
.41
5.43
59.8
732
.16
55.8
6Ja
n-88
Pola
nd3.
622.
090.
382.
0719
.37
17.9
74.
151.
4210
.88
9.16
8.97
1.45
1.28
-29.
422.
950.
16-0
.05
-0.1
915
.62
77.8
132
.77
80.1
6Ja
n-93
Rus
sia
4.46
1.40
0.50
3.19
24.6
121
.55
19.6
12.
9816
.98
14.7
514
.54
2.68
4.76
-46.
300.
330.
020.
040.
222.
8659
.87
35.2
657
.91
Jan-
95So
uth
Afr
ica
1.37
1.02
0.54
1.24
8.30
7.00
87.8
94.
486.
125.
675.
501.
601.
89-2
0.13
-0.5
10.
11-0
.36
-0.5
35.
5272
.25
39.8
871
.26
Jan-
93Sr
ilank
a0.
09-0
.36
0.30
0.79
9.49
9.11
0.75
-0.2
96.
636.
796.
581.
502.
28-1
9.64
0.32
0.09
-0.2
6-0
.49
3.37
63.0
427
.57
57.0
9Ja
n-93
Taiw
an1.
710.
960.
300.
9512
.50
11.9
710
6.48
4.67
8.29
7.60
7.35
1.20
1.46
-25.
110.
430.
21-0
.21
-0.6
43.
9784
.29
77.3
877
.86
Jan-
88T
haila
nd1.
090.
360.
461.
4212
.12
10.7
729
.73
3.39
8.41
8.06
7.85
1.65
2.02
-26.
160.
15-0
.24
-0.1
2-0
.32
4.39
71.2
458
.81
64.7
2Ja
n-88
Tur
key
3.06
1.51
0.13
0.61
18.5
418
.44
9.82
2.28
11.3
09.
739.
500.
931.
27-2
7.82
1.02
0.42
-0.1
6-0
.72
4.98
55.7
441
.26
54.1
6Ja
n-88
Ven
ezue
la1.
630.
320.
251.
0915
.89
15.4
75.
541.
7111
.21
10.5
710
.39
1.87
3.27
-34.
80-0
.05
0.14
-0.4
2-1
.37
3.83
65.9
435
.05
64.1
3Ja
n-93
Ave
rage
1.71
0.92
0.28
0.93
12.0
411
.38
25.7
72.
537.
847.
156.
941.
151.
57-2
2.67
0.68
0.14
-0.1
8-0
.44
5.82
68.4
142
.44
65.6
8
Star
tIC
RGC
EX
HI
BI
T1
AS
um
mar
y S
tati
stic
s (m
onth
ly d
olla
r re
turn
s)
Em
ergi
ng
Mar
ket
s
Mar
ket
E[R
]m
Gr
SRT
RIR
Mca
pSi
zeSe
mi
m
ean
Sem
i- r f
Sem
i-0
Dow
n-b
iwD
own-
b
wVa
Rsk
ewsk
ew5%
cosk
ew1
cosk
ew2
kurt
ICR
GC
CC
RIC
RG
PSt
art
Dev
elop
ed M
arke
tsA
ustr
alia
1.0
20
.87
0.5
40
.75
5.4
84
.63
2.3
03
.84
3.5
23
.29
0.4
50
.60
-10
.58
0.0
80
.07
-0.1
3-0
.15
3.3
38
0.7
87
0.5
88
1.6
9Ja
n-88
8A
ustr
ia0
.73
0.4
80
.39
0.6
97
.16
6.6
22
.40
4.9
84
.82
4.6
11
.29
1.5
4-1
5.8
10
.73
0.0
1-0
.23
-0.3
84
.58
86
.29
85
.42
85
.89
Jan-
88B
elgi
um1
.39
1.2
80
.60
0.7
04
.68
3.7
62
.48
3.0
82
.60
2.3
80
.56
0.6
2-7
.84
1.2
40
.07
0.0
60
.05
6.7
88
2.0
87
9.8
47
9.4
7Ja
n-88
Can
ada
1.0
40
.93
0.6
80
.80
4.6
63
.42
2.5
63
.45
3.1
42
.95
0.7
00
.94
-9.2
7-0
.29
0.1
1-0
.22
-0.1
96
.07
83
.92
82
.98
83
.34
Jan-
88D
enm
ark
1.4
21
.29
0.5
40
.71
5.2
74
.45
2.6
43
.70
3.1
62
.95
0.3
80
.79
-9.1
80
.31
-0.0
40
.00
0.0
02
.84
85
.05
77
.15
85
.56
Jan-
88in
land
2.1
91
.83
0.5
61
.24
8.7
47
.27
2.7
15
.77
4.7
74
.53
1.0
61
.66
-14
.95
0.2
90
.41
-0.1
4-0
.25
3.3
48
3.3
67
5.5
68
6.3
2Ja
n-88
ranc
e1
.49
1.3
40
.66
0.9
35
.58
4.1
82
.77
3.9
33
.38
3.1
70
.79
1.0
0-1
0.7
4-0
.13
0.0
90
.05
0.0
53
.75
81
.29
87
.67
80
.10
Jan-
88G
erm
any
1.4
21
.25
0.6
20
.91
5.8
54
.61
2.8
34
.29
3.8
23
.63
1.1
41
.21
-13
.36
0.2
6-0
.09
-0.2
3-0
.26
4.6
38
5.1
69
1.6
38
3.0
7Ja
n-88
rela
nd1
.24
1.0
80
.67
0.9
65
.73
4.2
92
.89
3.9
53
.54
3.3
20
.69
0.9
4-1
1.2
10
.27
0.1
10
.09
0.1
04
.17
82
.18
70
.98
81
.33
Jan-
88ta
ly1
.02
0.7
70
.51
0.9
27
.19
6.2
22
.94
4.8
54
.50
4.2
70
.68
0.8
8-1
2.9
20
.50
0.1
50
.05
0.0
73
.08
78
.08
76
.42
74
.92
Jan-
882
apan
0.4
30
.18
0.7
71
.41
7.2
44
.61
3.0
04
.89
4.9
04
.64
0.8
41
.15
-13
.92
0.3
50
.25
0.2
90
.34
3.5
18
6.6
99
1.8
68
2.3
9Ja
n-88
Net
herl
ands
1.5
51
.46
0.7
40
.79
4.2
42
.87
3.0
43
.17
2.6
22
.43
0.6
50
.84
-8.0
5-0
.31
-0.0
9-0
.07
-0.0
53
.33
86
.79
88
.74
86
.93
Jan-
88N
orw
ay1
.10
0.8
50
.56
0.9
86
.94
5.7
63
.09
5.0
74
.71
4.4
91
.04
1.2
3-1
3.8
8-0
.14
0.0
3-0
.22
-0.3
13
.93
87
.56
81
.07
84
.70
Jan-
88in
gapo
re1
.36
1.1
00
.64
1.1
97
.38
5.7
03
.14
5.1
24
.68
4.4
91
.14
1.4
1-1
6.8
00
.01
0.0
8-0
.06
-0.0
85
.15
84
.97
79
.95
81
.17
Jan-
88pa
in1
.29
1.0
90
.70
1.1
26
.36
4.5
63
.18
4.4
84
.03
3.8
31
.22
1.5
1-1
3.2
30
.38
0.1
5-0
.25
-0.2
84
.31
76
.56
75
.60
73
.24
Jan-
88w
eden
1.9
11
.69
0.7
11
.19
6.6
24
.67
3.2
24
.78
4.0
13
.81
0.9
51
.35
-13
.09
-0.1
0-0
.05
-0.1
4-0
.16
3.4
98
3.1
77
7.5
48
4.2
7Ja
n-88
witz
erla
nd1
.47
1.3
40
.65
0.8
45
.12
3.8
93
.26
3.5
93
.08
2.8
80
.91
1.2
0-1
0.6
20
.27
-0.0
4-0
.05
-0.0
54
.06
90
.05
92
.85
88
.65
Jan-
88U
.K.
1.2
21
.11
0.7
60
.89
4.6
83
.08
3.3
03
.20
2.7
72
.55
0.4
10
.66
-7.9
60
.49
-0.0
70
.35
0.2
73
.24
81
.78
87
.14
81
.51
Jan-
881
.58
1.5
10
.76
0.7
23
.77
2.4
63
.33
2.8
02
.26
2.0
70
.67
0.8
4-6
.69
-0.1
5-0
.09
-0.2
3-0
.14
4.3
28
3.3
59
0.1
48
1.8
6Ja
n-88
Ave
rage
1.3
11
.13
0.6
30
.93
5.9
34
.58
2.9
04
.15
3.7
03
.49
0.8
21
.07
-11
.58
0.2
10
.06
-0.0
6-0
.08
4.1
08
3.6
48
2.2
78
2.4
4
Wor
ld0
.01
0.0
10
.04
90
32
.25
9.1
1Ja
n-88
U.S
.
13
1.2
41
9.7
26
2.3
72
11.3
44
0.5
84
0.3
93
56
.85
39
6.5
51
6.7
81
54
.80
18
10
.75
21
3.9
52
2.4
55
3.1
311
3.3
21
05
.27
27
7.9
39
23
.20
39
79
.84
47
0.0
2
EX
HI
BI
T1
BS
um
mar
y S
tati
stic
s (m
onth
ly d
olla
r re
turn
s)
Dev
elop
ed M
ark
ets
MSC
I D
ata.
E[R
]: M
ean
retu
rns;
G
: Geo
met
ric
mea
n re
turn
s; :
Cor
rela
tion
with
wor
ld r
etur
ns; S
R: S
yste
mat
ic r
isk (B
eta)
; TR
: Tot
al r
isk; I
R: I
dios
yncr
atic
risk
; Mca
p: A
ver-
age
Mar
ket c
apita
lizat
ion;
Siz
e: L
og o
f ave
rage
mar
ket c
ap; S
emim
ean:
Sem
idev
iatio
n w
ith r
espe
ct to
mea
n; S
emi-
r f: S
emid
evia
tion
with
res
pect
to r
isk-f
ree
rate
; Sem
i-0:
Sem
ide-
viat
ion
with
res
pect
to 0
; Dow
n-
iw: D
owns
ide
beta
cal
cula
ted
usin
g ob
serv
atio
ns w
hen
coun
try
inde
x an
d w
orld
inde
x fa
ll sim
ulta
neou
sly; D
own-
w: D
owns
ide
beta
cal
cula
ted
usin
g ob
serv
atio
ns w
hen
wor
ld in
dex
falls
; VaR
: Val
ue a
t risk
; ske
w: S
kew
ness
; ske
w5%
: Ske
wne
ss c
alcu
late
d us
ing
form
ula
1; C
oske
w1:
Cos
kew
ness
with
wor
ld in
dex
usin
g fo
r-m
ula
2, C
oske
w-2
: Cos
kew
with
wor
ld in
dex
usin
g fo
rmul
a 3;
kur
t: K
urto
sis; I
CR
GC
: Ave
rage
ICR
G c
ompo
site
coun
try
risk
; CC
R: A
vera
ge C
CR
cou
ntry
risk
; IC
RG
P: IC
RG
Ppo
litic
al r
isk; S
tart
: Dat
a st
art
date
.
-
III. REGRESSION ANALYSIS
Bivariate Regressions
The main regression results are presented in Exhibit3. The regressions are illustrated 18 sets of graphs in Exhibit4. These regressions examine the bivariate relation betweenthe average returns and the average risk measures.
I should emphasize that we are comparing averagesto averages over the same time period. This is consistentwith some of the early tests of asset pricing models, suchas Black, Jensen, and Scholes [1972]. There is a substan-tial literature, beginning with Harvey [1991], however,that suggests that in an international context time varia-tion in the risk and returns measures is very important.Nevertheless, this unconditional analysis should proveinteresting.
The first set of regressions is the classic worldCAPM. The analysis shows a significant relation when all47 countries are included with an R2 of 27%. Interestingly,the intercept is not significantly different from zero. Thisevidence would seem to support the CAPM.
A closer inspection reveals that all of the explanatorypower is coming from emerging markets. The regressionsuggests that there is a 0% R2 for developed markets.
But here is where the graphic analysis is especiallyinsightful. The poor fit for developed markets is driven byone point; Japan has a beta of 1.41 and an average returnclose to zero. If this single observation is excluded, the fitof the developed countries dramatically improves andshows an R2 of 27%. Hence, the CAPM fares reasonablywell in this analysis.
The fit for the emerging markets might seem sur-prising, given the results in Bekaert and Harvey [1995] andHarvey [1995a], but remember that the sample is for1988-1999. During the early part of this period, many ofthe emerging markets liberalized, and the evidence inBekaert and Harvey [2000] would suggest that these mar-kets have become more correlated with the world.
The second risk measure is total risk. Asset pricingtheory says that only systematic risk, or the part of vari-ance that contributes to a well-diversified portfolios vari-ance, should be important. The regressions suggest thattotal variance can account for 52% of the variation in theemerging market returns. Variance explains practicallynone of the developed market returns. A combined anal-ysis is heavily influenced by emerging markets returns.
The results for the third risk measure, idiosyncraticrisk, are similar to total risk.
4 DRIVERS OF EXPECTED RETURNS IN INTERNATIONAL MARKETS FALL 2000
E[R
]SR
TRIR
Size
Sem
imea
nSe
mi -
r fSe
mi-0
Dow
n-b
iwD
own-
b
wVa
Rsk
ewsk
ew5%
cosk
ew1
cosk
ew2
kurt
ICR
GC
CC
RIC
RG
P
E[R
]1.
000
SR0.
060
1.00
0TR
-0.0
250.
691
1.00
0IR
-0.0
230.
389
0.93
31.
000
Size
-0.0
930.
052
-0.4
42-0
.607
1.00
0Se
mim
ean
-0.0
230.
685
0.98
70.
919
-0.4
471.
000
Sem
i - r f
-0.2
660.
651
0.95
40.
885
-0.4
020.
969
1.00
0Se
mi-0
-0.2
620.
648
0.95
30.
885
-0.4
050.
970
1.00
01.
000
Dow
n-b
iw0.
044
0.45
10.
625
0.56
7-0
.264
0.68
70.
668
0.67
81.
000
Dow
n-b
w0.
190
0.58
40.
745
0.65
6-0
.318
0.78
70.
722
0.72
90.
929
1.00
0
VaR
0.13
6-0
.624
-0.9
19-0
.859
0.44
0-0
.946
-0.9
52-0
.956
-0.8
01-0
.823
1.00
0sk
ew-0
.221
-0.1
280.
151
0.24
3-0
.218
0.04
00.
083
0.07
4-0
.023
-0.0
64-0
.061
1.00
0sk
ew5%
-0.0
470.
584
0.67
30.
580
-0.2
200.
594
0.57
00.
562
0.21
50.
372
-0.4
580.
184
1.00
0co
skew
1-0
.275
0.22
4-0
.063
-0.2
040.
273
-0.1
66-0
.102
-0.1
15-0
.523
-0.4
380.
218
0.33
00.
123
1.00
0co
skew
2-0
.264
0.17
9-0
.218
-0.3
760.
418
-0.3
12-0
.242
-0.2
54-0
.614
-0.5
490.
358
0.20
90.
069
0.95
61.
000
kurt
-0.1
01-0
.214
-0.2
21-0
.166
-0.1
13-0
.223
-0.1
78-0
.172
0.19
3-0
.004
0.06
60.
264
-0.0
06-0
.258
-0.2
091.
000
ICR
GC
-0.0
50-0
.048
-0.0
38-0
.036
-0.0
100.
009
0.03
40.
038
0.16
70.
169
-0.0
78-0
.210
-0.2
89-0
.052
-0.0
750.
039
1.00
0C
CR
-0.1
34-0
.069
-0.3
13-0
.392
0.65
2-0
.282
-0.2
24-0
.221
0.14
00.
021
0.18
0-0
.126
-0.3
970.
112
0.15
50.
096
0.56
61.
000
ICR
GP
0.22
3-0
.149
-0.0
61-0
.011
-0.1
08-0
.027
-0.0
76-0
.074
0.01
80.
127
0.03
6-0
.273
-0.2
52-0
.119
-0.1
74-0
.131
0.87
60.
350
1.00
0
EX
HI
BI
T2
AC
ross
-Sec
tion
An
alys
is C
orre
lati
on M
atri
xD
evel
oped
Cou
ntr
ies
-
FALL 2000 EMERGING MARKETS QUARTERLY 5
EX
HI
BI
T2
BC
ross
-Sec
tion
An
alys
is C
orre
lati
on M
atri
xE
mer
gin
g M
ark
et C
oun
trie
s
E[R
]SR
TR
IRSi
zeSe
mim
ean
Sem
i - r
fSe
mi-
0D
own-
b
iwD
own-
b
wV
aRsk
ewsk
ew5%
cosk
ew1
cosk
ew2
kurt
ICR
GC
CC
RIC
RG
P
E[R
]1.
000
SR0.
562
1.00
0T
R0.
716
0.66
01.
000
IR0.
683
0.55
50.
990
1.00
0Si
ze0.
309
0.30
30.
134
0.09
21.
000
Sem
imea
n0.
678
0.74
60.
966
0.93
70.
187
1.00
0Se
mi -
rf
0.53
00.
724
0.93
20.
906
0.14
50.
982
1.00
0Se
mi-
00.
531
0.72
70.
931
0.90
40.
145
0.98
31.
000
1.00
0D
own-
b
iw0.
499
0.66
10.
439
0.36
10.
201
0.54
40.
508
0.51
11.
000
Dow
n-b
w0.
507
0.79
40.
593
0.51
80.
246
0.73
20.
724
0.72
70.
837
1.00
0
VaR
-0.5
56-0
.761
-0.9
01-0
.865
-0.2
27-0
.965
-0.9
77-0
.978
-0.5
64-0
.775
1.00
0sk
ew0.
469
0.12
60.
480
0.50
9-0
.129
0.26
50.
182
0.18
1-0
.024
-0.1
08-0
.153
1.00
0sk
ew5%
0.09
2-0
.241
0.18
50.
255
-0.1
680.
062
0.04
20.
038
-0.5
24-0
.307
0.03
40.
337
1.00
0co
skew
1-0
.054
0.04
80.
184
0.19
0-0
.020
0.11
40.
136
0.13
2-0
.536
-0.3
54-0
.056
0.16
20.
248
1.00
0co
skew
2-0
.286
0.00
8-0
.137
-0.1
59-0
.086
-0.1
66-0
.129
-0.1
32-0
.502
-0.4
340.
197
-0.0
230.
055
0.87
51.
000
kurt
0.40
10.
194
0.41
40.
429
0.03
10.
221
0.15
50.
155
0.00
9-0
.085
-0.1
970.
858
0.17
90.
155
-0.0
111.
000
ICR
GC
-0.0
940.
017
-0.2
65-0
.295
0.25
3-0
.281
-0.2
95-0
.294
-0.0
24-0
.221
0.27
30.
135
-0.0
74-0
.152
-0.0
200.
161
1.00
0C
CR
-0.1
320.
022
-0.0
21-0
.029
0.47
1-0
.029
0.00
0-0
.002
-0.1
13-0
.169
0.04
20.
050
-0.0
420.
165
0.16
1-0
.001
0.65
11.
000
ICR
GP
0.16
90.
119
-0.0
81-0
.113
0.25
9-0
.093
-0.1
41-0
.139
0.12
4-0
.041
0.10
40.
214
-0.0
24-0
.283
-0.2
440.
259
0.90
20.
499
1.00
0
E[R
]SR
TR
IRSi
zeSe
mim
ean
Sem
i - r
fSe
mi-
0D
own-
b
iwD
own-
b
wV
aRsk
ewsk
ew5%
cosk
ew1
cosk
ew2
kurt
ICR
GC
CC
RIC
RG
P
E[R
]1.
000
SR0.
519
1.00
0T
R0.
642
0.51
41.
000
IR0.
595
0.39
50.
990
1.00
0Si
ze0.
018
0.17
1-0
.394
-0.4
671.
000
Sem
imea
n0.
616
0.58
20.
980
0.96
0-0
.368
1.00
0Se
mi -
rf
0.48
80.
549
0.95
90.
945
-0.4
050.
988
1.00
0Se
mi-
00.
489
0.55
10.
959
0.94
4-0
.405
0.98
81.
000
1.00
0D
own-
b
iw0.
491
0.61
60.
515
0.45
7-0
.127
0.59
30.
566
0.56
91.
000
Dow
n-b
w0.
515
0.75
10.
624
0.55
7-0
.091
0.72
60.
710
0.71
20.
854
1.00
0
VaR
-0.5
16-0
.586
-0.9
40-0
.920
0.36
2-0
.977
-0.9
85-0
.986
-0.6
16-0
.756
1.00
0sk
ew0.
442
0.09
90.
530
0.54
9-0
.305
0.37
60.
326
0.32
50.
067
-0.0
09-0
.302
1.00
0sk
ew5%
0.11
7-0
.137
0.31
90.
359
-0.2
840.
234
0.22
90.
225
-0.3
14-0
.143
-0.1
670.
359
1.00
0co
skew
1-0
.149
0.06
7-0
.128
-0.1
680.
308
-0.1
74-0
.168
-0.1
72-0
.540
-0.3
900.
217
0.05
90.
103
1.00
0co
skew
2-0
.339
0.01
9-0
.376
-0.4
170.
310
-0.3
97-0
.380
-0.3
82-0
.568
-0.5
030.
424
-0.1
35-0
.054
0.84
61.
000
kurt
0.40
60.
157
0.48
80.
501
-0.2
210.
353
0.31
30.
314
0.11
90.
022
-0.3
420.
822
0.21
8-0
.061
-0.1
731.
000
ICR
GC
-0.2
130.
009
-0.6
08-0
.660
0.59
2-0
.608
-0.6
28-0
.629
-0.2
25-0
.333
0.60
4-0
.179
-0.2
430.
198
0.32
6-0
.170
1.00
0C
CR
-0.2
400.
010
-0.5
32-0
.586
0.72
8-0
.528
-0.5
36-0
.537
-0.2
87-0
.315
0.53
7-0
.236
-0.2
350.
360
0.44
4-0
.276
0.86
71.
000
ICR
GP
-0.0
390.
067
-0.5
11-0
.567
0.57
3-0
.510
-0.5
51-0
.551
-0.1
42-0
.227
0.51
9-0
.129
-0.2
100.
121
0.18
9-0
.111
0.95
70.
800
1.00
0
EX
HI
BI
T2
CC
ross
-Sec
tion
An
alys
is C
orre
lati
on M
atri
xA
ll C
oun
trie
s
MSC
I D
ata.
E[R
]: M
ean
retu
rns;
SR:
Syst
emat
ic r
isk (
Bet
a);
TR
: T
otal
risk
; IR
: Id
iosy
ncra
tic r
isk;
Size
: Lo
g of
ave
rage
mar
ket
cap;
Sem
imea
n: S
emid
evia
tion
with
res
pect
to
mea
n; S
emi-
r f: S
emid
evia
tion
with
res
pect
to r
isk-f
ree
rate
; Sem
i-0:
Sem
idev
iatio
n w
ith r
espe
ct to
0; D
own-
iw
: Dow
nsid
e be
ta c
alcu
late
d us
ing
obse
rvat
ions
whe
n co
untr
y in
dex
and
wor
ld in
dex
fall
simul
tane
ously
; Dow
n-
w: D
owns
ide
beta
cal
cula
ted
usin
g ob
serv
atio
ns w
hen
wor
ld in
dex
falls
; VaR
: Val
ue a
t risk
; ske
w: S
kew
ness
, ske
w5%
: Ske
wne
ss c
al-
cula
ted
usin
g fo
rmul
a 1;
Cos
kew
1: C
o-sk
ewne
ss w
ith w
orld
inde
x us
ing
form
ula
2, C
oske
w-2
: Cos
kew
with
wor
ld in
dex
usin
g fo
rmul
a 3;
kur
t: K
urto
sis; I
CR
GC
: Log
of a
ver-
age
ICR
G: c
ompo
site
coun
try
risk
; CC
R: L
og o
f ave
rage
CC
R c
ount
ry r
isk, I
CR
GP:
Log
of I
CR
P po
litic
al r
isk.
-
6 DRIVERS OF EXPECTED RETURNS IN INTERNATIONAL MARKETS FALL 2000
Risk Variable c0 p-value c1 p-value R2 Adj-R2SR 0.7777 0.091 0.1126 0.807 0.004 -0.055TR 0.9277 0.052 -0.0076 0.919 0.001 -0.058IR 0.9155 0.019 -0.0071 0.925 0.001 -0.058Size 1.0041 0.007 -0.0244 0.704 0.009 -0.050Semimean 0.9274 0.070 -0.0107 0.925 0.001 -0.058Semi - rf 1.3345 0.004 -0.1220 0.271 0.071 0.016Semi-0 1.3063 0.004 -0.1214 0.279 0.069 0.014Down-biw 0.8317 0.012 0.0624 0.857 0.002 -0.057
Down-bw 0.6351 0.066 0.2310 0.436 0.036 -0.021
VaR 1.0930 0.011 0.0182 0.580 0.018 -0.039skew 0.9309 0.000 -0.2259 0.364 0.049 -0.007skew5% 0.8909 0.000 -0.1423 0.850 0.002 -0.057coskew1 0.8474 0.000 -0.6217 0.255 0.075 0.021coskew2 0.8419 0.000 -0.5410 0.274 0.070 0.015kurt 1.0414 0.016 -0.0387 0.681 0.010 -0.048ICRGC 3.0763 0.777 -0.4956 0.840 0.003 -0.056CCR 3.5383 0.469 -0.6027 0.586 0.018 -0.040ICRGP -7.2269 0.413 1.8384 0.359 0.050 -0.006
E X H I B I T 3 ABivariate RegressionsDeveloped Countries
Risk Variable c0 p-value c1 p-value R2 Adj-R2SR 0.3699 0.272 1.0050 0.002 0.319 0.293TR -0.8754 0.059 0.1813 0.000 0.517 0.499IR -0.7778 0.107 0.1831 0.000 0.470 0.450Size 0.6242 0.202 0.2704 0.118 0.092 0.057Semimean -0.8811 0.084 0.2792 0.000 0.464 0.444Semi - rf -0.4945 0.412 0.2520 0.003 0.285 0.258Semi-0 -0.4459 0.447 0.2526 0.003 0.286 0.259Down-biw 0.2849 0.484 0.8893 0.007 0.250 0.221
Down-bw 0.4131 0.258 0.5713 0.006 0.259 0.230
VaR -0.4617 0.409 -0.0780 0.002 0.312 0.286skew 0.8635 0.003 0.6561 0.012 0.220 0.190skew5% 1.2288 0.000 0.5588 0.639 0.009 -0.030coskew1 1.2324 0.002 -0.4184 0.781 0.003 -0.035coskew2 0.9742 0.004 -0.7513 0.140 0.082 0.047kurt 0.3654 0.447 0.1618 0.035 0.160 0.128ICRGC 5.9795 0.532 -1.1071 0.625 0.009 -0.029CCR 3.3286 0.249 -0.5460 0.478 0.020 -0.018ICRGP -5.2259 0.495 1.5644 0.395 0.028 -0.009
E X H I B I T 3 BBivariate RegressionsEmerging Countries
-
The fourth risk measure is size. Size could be relatedto liquidity and the amount of information available in themarket, which are legitimate risk factors. We find thatthere is little relation between the average internationalreturns and size.
The fifth through seventh measures are semivariancemeasures. All are similar, and from Exhibit 2 we knowthey are all highly correlated with total variance. We findthat the semivariance measures account for a substantialpart of the variation in the emerging market returns butnot the developed market returns.
The eighth and ninth measures are downside betas.There is some relation between the downside betas andthe emerging market returnsbut little relation with thedeveloped market returns.
The tenth measure is a value at risk measure. TheVaR measures the mean return in the 5% tail of the dis-tribution. If this type of risk is priced, one would expecta highly negative VaR (high risk) to command a high aver-age return. This is exactly what we see for emergingmarketsbut not developed markets. Notice that theslope of the line in the Exhibit 4J Emerging Markets
graph is negative, implying that a more negative VaRimplies a higher expected return in emerging markets.
Harvey and Siddique [2000a] posit that investors careabout the skewness of their portfolio. Their model isbased on the work of Rubinstein [1973] and Kraus andLitzenberger [1976]. In this world, investors value the con-tribution to the skewness of the portfolio, or the coskew-ness. That is, if an asset contributes positive skewness toa diversified portfolio (high coskewness), that asset will bevaluable and will have a high price (low expected return).If the asset contributes negative skewness (negativecoskewness), to get investors to purchase this asset theprice must drop ( higher expected return).
So, in the world of the augmented CAPM, coskew-ness should count, and skewness itself should not. This isanalogous to beta and variance. In the CAPM, it is onlythe beta that is rewarded, not the total volatility. To beconsistent, only the systematic part of skewness (thecoskewness) should command a reward.
When we examine the two total skewness measures, wesee a positive relation for emerging markets but not devel-oped markets. This is similar to what we see for total volatil-
FALL 2000 EMERGING MARKETS QUARTERLY 7
Risk Variable c0 p-value c1 p-value R2 Adj-R2SR 0.2480 0.325 0.9514 0.000 0.271 0.249TR -0.1256 0.612 0.1318 0.000 0.422 0.409IR 0.0934 0.693 0.1207 0.000 0.365 0.351Size 1.1247 0.001 0.0031 0.968 0.000 0.022Semimean -0.1784 0.512 0.2070 0.000 0.389 0.376Semi - rf 0.0813 0.789 0.1832 0.000 0.247 0.230Semi-0 0.1170 0.691 0.1836 0.000 0.248 0.232Down-biw 0.2769 0.289 0.8451 0.000 0.244 0.227
Down-bw 0.3493 0.135 0.5758 0.000 0.269 0.253
VaR 0.0753 0.793 -0.0583 0.000 0.275 0.259skew 0.8478 0.000 0.5886 0.002 0.198 0.180skew5% 1.0650 0.000 0.6623 0.421 0.015 -0.007coskew1 1.0229 0.000 -0.8703 0.300 0.024 0.002coskew2 0.8964 0.000 -0.8121 0.018 0.118 0.098kurt 0.3226 0.289 0.1586 0.004 0.167 0.149ICRGC 8.2620 0.081 -1.6560 0.130 0.050 0.029CCR 3.4683 0.012 -0.5851 0.084 0.065 0.044ICRGP 2.4870 0.538 -0.3163 0.737 0.003 -0.020
E X H I B I T 3 CBivariate RegressionsAll Countries
E[Ri]: Mean return; Riski = Risk Variable; SR: Systematic risk (Beta); TR: Total risk; IR: Idiosyncratic risk; Size: Log of average market cap;Semimean: Semideviation with respect to mean; Semi-rf: Semideviation with respect to risk-free rate; Semi-0: Semideviation with respect to 0;Down-iw: Downside beta calculated using observations when country index and world index fall simultaneously; Down-w: Downside beta cal-culated using observations when world index falls; VaR: Value at risk; skew: Skewness; skew5%: Skewness calculated using formula 1; Coskew1:Co-skewness with world index using formula 2; Coskew-2: Coskew with world index using formula 3; kurt: Kurtosis; ICRGC: Log of averageICRG composite country risk; CCR: Log of average CCR country risk, ICRGP: Log of ICRP political risk.
-
8 DRIVERS OF EXPECTED RETURNS IN INTERNATIONAL MARKETS FALL 2000
EX
HIB
IT 4
A -
Sys
tem
atic
Ris
kA
ll C
ount
ries
E
mer
ging
Mar
kets
Dev
elop
ed M
arke
ts
EX
HIB
IT 4
B -
Tot
al R
isk
All
Cou
ntri
esE
mer
ging
Mar
kets
Dev
elop
ed M
arke
ts
EX
HIB
IT 4
C -
Idio
sync
rati
c R
isk
All
Cou
ntri
esE
mer
ging
Mar
kets
Dev
elop
ed M
arke
ts
y =
1.01
14x
+ 0.
2271
R2 =
0.3
101
-1012345
-1.0
00.
001.
002.
003.
004.
00Sy
stem
atic
Ris
k
Mean Returns
y =
1.00
5x +
0.3
699
R2 =
0.3
193
-1012345
-1.0
00.
001.
002.
003.
004.
00Sy
stem
atic
Ris
k
Mean Returns
y =
0.97
64x
+ 0.
046
R2 =
0.2
656
0
0.51
1.52 0
.00
0.50
1.00
1.50
Syst
emat
ic R
isk
Mean Returns
y =
0.13
18x
- 0.
1256
R2 =
0.4
221
-1012345
0.00
5.00
10.0
015
.00
20.0
025
.00
30.0
0T
otal
Ris
k
Mean Returns
y =
0.18
13x
- 0.
8754
R2 =
0.5
172
-1012345
0.00
5.00
10.0
015
.00
20.0
025
.00
30.0
0T
otal
Ris
k
Mean Returns
y =
-0.0
076x
+ 0
.927
7
R2 =
0.0
006
0
0.51
1.52 0
.00
2.00
4.00
6.00
8.00
10.0
0T
otal
Ris
k
Mean Returns
y =
0.12
07x
+ 0.
0934
R2 =
0.3
65
-1012345
0.00
5.00
10.0
015
.00
20.0
025
.00
Idio
sync
rati
c R
isk
Mean Returns
y =
0.18
32x
- 0.
7778
R2 =
0.4
699
-1012345
0.00
5.00
10.0
015
.00
20.0
025
.00
Idia
sync
rati
c R
isk
Mean Returns
y =
-0.0
071x
+ 0
.915
5
R2 =
0.0
005
0
0.51
1.52 0
.00
2.00
4.00
6.00
8.00
Idio
sync
rati
c R
isk
Mean Returns
-
FALL 2000 EMERGING MARKETS QUARTERLY 9
EX
HIB
IT 4
D -
Size
All
Cou
ntri
esE
mer
ging
Mar
kets
Dev
elop
ed M
arke
ts
EX
HIB
IT 4
E -
Sem
ivar
ianc
e to
Mea
nA
ll C
ount
ries
Em
ergi
ng M
arke
tsD
evel
oped
Mar
kets
EX
HIB
IT 4
F - S
emiv
aria
nce
to R
isk
-Fre
e In
tere
st R
ate
All
Cou
ntri
esE
mer
ging
Mar
kets
Dev
elop
ed M
arke
ts
y =
0.22
89x
+ 0.
523
R2 =
0.0
624
-1012345
-1.0
00.
001.
002.
003.
004.
005.
00Si
ze
Mean Returns
y =
0.27
04x
+ 0.
6242
R2 =
0.0
915
-1012345
-1.0
00.
001.
002.
003.
004.
005.
00Si
ze
Mean Returns
y =
0.29
39x
+ 0.
0306
R2 =
0.0
56
0
0.51
1.52 0
.00
1.00
2.00
3.00
4.00
Size
Mean Returns
y =
0.20
7x -
0.17
84
R2 =
0.3
893
-1012345
0.00
5.00
10.0
015
.00
20.0
0Se
mi -
mea
n
Mean Returns
y =
0.27
92x
- 0.8
811
R2 =
0.4
644
-1012345
0.00
5.00
10.0
015
.00
20.0
0Se
mi-m
ean
Mean Returns
y =
-0.0
107x
+ 0
.927
4
R2 =
0.0
005
0
0.51
1.52 0
.00
2.00
4.00
6.00
8.00
Sem
i-mea
n
Mean Returns
y =
0.18
32x
+ 0.
0813
R2 =
0.2
469
-1012345
0.00
5.00
10.0
015
.00
20.0
0Se
mi-r
f
Mean Returns
y =
0.25
2x -
0.49
45
R2 =
0.2
851
-1012345
0.00
5.00
10.0
015
.00
20.0
0Se
mi-r
f
Mean Returns
y =
-0.1
22x
+ 1.
3345
R2 =
0.0
709
0
0.51
1.52 0
.00
1.00
2.00
3.00
4.00
5.00
6.00
Sem
i-rf
Mean Returns
-
10 DRIVERS OF EXPECTED RETURNS IN INTERNATIONAL MARKETS FALL 2000
EX
HIB
IT 4
G -
Sem
ivar
ianc
e to
Zer
oA
ll C
ount
ries
Em
ergi
ng M
arke
tsD
evel
oped
Mar
kets
EX
HIB
IT 4
H -
Dow
nsid
e B
eta,
Wor
ld a
nd L
ocal
Mar
ket D
own
All
Cou
ntri
esE
mer
ging
Mar
kets
Dev
elop
ed M
arke
ts
EX
HIB
IT 4
I - D
owns
ide
Bet
a, W
orld
Dow
nA
ll C
ount
ries
Em
ergi
ng M
arke
tsD
evel
oped
Mar
kets
y =
0.18
37x
+ 0.
117
R2 =
0.2
482
-1012345
0.00
5.00
10.0
015
.00
20.0
0Se
mi-0
Mean Returns
y =
0.25
26x
- 0.4
459
R2 =
0.2
863
-1012345
0.00
5.00
10.0
015
.00
20.0
0Se
mi-0
Mean Returns
y =
-0.1
214x
+ 1
.306
3
R2 =
0.0
685
0
0.51
1.52 0
.00
1.00
2.00
3.00
4.00
5.00
Sem
i-0
Mean Returns
y =
0.84
51x
+ 0.
2769
R2 =
0.2
439
-1012345
-1.0
00.
001.
002.
003.
00D
own
Bet
a - i
w
Mean Returns
y =
0.88
93x
+ 0.
2849
R2 =
0.2
502
-1012345
-1.0
00.
001.
002.
003.
00D
own
Bet
a - i
w
Mean Returns
y =
0.06
24x
+ 0.
8317
R2 =
0.0
02
0
0.51
1.52 0
.00
0.50
1.00
1.50
Dow
n B
eta
- iw
Mean Returns
y =
0.57
58x
+ 0.
3493
R2 =
0.2
691
-1012345
-1.0
00.
001.
002.
003.
004.
005.
006.
00D
own
Bet
a -w
Mean Returns
y =
0.57
13x
+ 0.
4131
R2 =
0.2
589
-1012345
-1.0
00.
001.
002.
003.
004.
005.
006.
00D
own
Bet
a - w
Mean Returns
y =
0.23
1x +
0.6
351
R2 =
0.0
361
0
0.51
1.52 0
.00
0.50
1.00
1.50
2.00
Dow
n B
eta
- w
Mean Returns
-
FALL 2000 EMERGING MARKETS QUARTERLY 11
EX
HIB
IT 4
J - V
alue
at R
isk
All
Cou
ntri
esE
mer
ging
Mar
kets
Dev
elop
ed M
arke
ts
EX
HIB
IT 4
K -
Unc
ondi
tiona
l Ske
wne
ssA
ll C
ount
ries
Em
ergi
ng M
arke
tsD
evel
oped
Mar
kets
EX
HIB
IT 4
L - S
kew
5%
Tai
lA
ll C
ount
ries
Em
ergi
ng M
arke
tsD
evel
oped
Mar
kets
y =
-0.0
583x
+ 0
.075
3
R2 =
0.2
746
-1012345
-50.
00-4
0.00
-30.
00-2
0.00
-10.
000.
00V
alue
At R
isk
Mean Returns
y =
-0.0
78x
- 0.4
618
R2 =
0.3
122
-1012345
-50.
00-4
0.00
-30.
00-2
0.00
-10.
000.
00V
alue
At R
isk
Mean Returns
y =
0.01
82x
+ 1.
0932
R2 =
0.0
184
0
0.51
1.52
-20.
00-1
5.00
-10.
00-5
.00
0.00
Val
ue A
t Ris
k
Mean Returns
y =
0.58
87x
+ 0.
8478
R2 =
0.1
978
-1012345
-1.0
00.
001.
002.
003.
004.
00Sk
ew
Mean Returns
y =
0.65
61x
+ 0.
8635
R2 =
0.2
202
-1012345
-1.0
00.
001.
002.
003.
004.
00Sk
ew
Mean Returns
y =
-0.2
259x
+ 0
.930
9
R2 =
0.0
486
0
0.51
1.52
-0.5
00.
000.
501.
001.
50Sk
ew
Mean Returns
y =
0.66
23x
+ 1.
0651
R2 =
0.0
145
-1012345
-0.4
0-0
.20
0.00
0.20
0.40
0.60
Skew
5%
Mean Returns
y =
0.55
88x
+ 1.
2288
R2 =
0.0
086
-1012345
-0.4
0-0
.20
0.00
0.20
0.40
0.60
Skew
5%
Mean Returns
y =
-0.1
423x
+ 0
.890
9
R2 =
0.0
022
0
0.51
1.52
-0.2
0-0
.10
0.00
0.10
0.20
0.30
0.40
0.50
Skew
5%
Mean Returns
-
12 DRIVERS OF EXPECTED RETURNS IN INTERNATIONAL MARKETS FALL 2000
All
Cou
ntri
esE
mer
ging
Mar
kets
Dev
elop
ed M
arke
ts
EX
HIB
IT 4
N -
Cos
kew
ness
Mea
sure
2A
ll C
ount
ries
Em
ergi
ng M
arke
tsD
evel
oped
Mar
kets
EX
HIB
IT 4
O -
Kur
tosi
sA
ll C
ount
ries
Em
ergi
ng M
arke
tsD
evel
oped
Mar
kets
y =
-0.8
703x
+ 1
.022
9
R2 =
0.0
238
-1012345
-0.6
0-0
.40
-0.2
00.
000.
200.
40C
oske
w1
Mean Returns
y =
-0.4
184x
+ 1
.232
4
R2 =
0.0
03
-1012345
-0.6
0-0
.40
-0.2
00.
000.
20C
oske
w1
Mean Returns
y =
-0.6
217x
+ 0
.847
4
R2 =
0.0
754
0
0.51
1.52
-0.4
0-0
.20
0.00
0.20
0.40
Cos
kew
1
Mean Returns
y =
-0.8
121x
+ 0
.896
4
R2 =
0.1
18
-1012345
-1.5
0-1
.00
-0.5
00.
000.
50C
oske
w2
Mean Returns
y =
-0.7
513x
+ 0
.974
2
R2 =
0.0
818
-1012345
-1.5
0-1
.00
-0.5
00.
000.
50C
oske
w2
Mean Returns
y =
-0.5
41x
+ 0.
8419
R2 =
0.0
698
0
0.51
1.52
-0.6
0-0
.40
-0.2
00.
000.
200.
40C
oske
w2
Mean Returns
y =
0.15
86x
+ 0.
3226
R2 =
0.1
673
-1012345
0.00
5.00
10.0
015
.00
20.0
0K
urto
sis
Mean Returns
y =
0.16
18x
+ 0.
3654
R2 =
0.1
604
-1012345
0.00
5.00
10.0
015
.00
20.0
0K
urto
sis
Mean Returns
y =
-0.0
387x
+ 1
.041
4
R2 =
0.0
102
0
0.51
1.52 0
.00
2.00
4.00
6.00
8.00
Kur
tosi
s
Mean Returns
EX
HIB
IT 4
M-C
oske
wne
ss M
easu
re 1
-
FALL 2000 EMERGING MARKETS QUARTERLY 13
EX
HIB
IT 4
P - I
CR
GC
Com
posi
te C
ount
ry R
isk
Em
ergi
ng M
arke
tsD
evel
oped
Mar
kets
All
Cou
ntri
es
EX
HIB
IT 4
Q -
Inst
itutio
nal I
nves
tor
Cou
ntry
Cre
dit R
atin
gA
ll C
ount
ries
Em
ergi
ng M
arke
tsD
evel
oped
Mar
kets
EX
HIB
IT 4
R -
ICR
G P
oliti
cal R
isk
All
Cou
ntri
esD
evel
oped
Mar
kets
y =
-1.1
071x
+ 5
.979
5
R2 =
0.0
093
-1012345
4.00
4.10
4.20
4.30
4.40
4.50
ICR
G C
ount
ry R
isk
Com
posi
te
Mean Returns
y =
-1.6
56x
+ 8.
262
R2 =
0.0
501
-1012345
3.90
4.00
4.10
4.20
4.30
4.40
4.50
4.60
ICR
G C
ount
ry R
isk
Com
posi
te
Mean Returns
y =
-0.4
956x
+ 3
.076
3
R2 =
0.0
025
0
0.51
1.52 4
.30
4.35
4.40
4.45
4.50
4.55
ICR
G C
ount
ry R
isk
Com
posi
te
Mean Returns
y =
-0.5
851x
+ 3
.468
3
R2 =
0.0
647
-1012345
0.00
1.00
2.00
3.00
4.00
5.00
CC
R C
ount
ry R
atin
g
Mean Returns
y =
-0.3
163x
+ 2
.487
R2 =
0.0
025
-1012345
3.90
4.00
4.10
4.20
4.30
4.40
4.50
4.60
ICR
G P
oliti
cal R
isk
Mean Returns
y =
-0.5
46x
+ 3.
3286
R2 =
0.0
195
-1012345
0.00
1.00
2.00
3.00
4.00
5.00
CC
R C
ount
ry R
atin
g
Mean Returns
y =
-0.6
027x
+ 3
.538
4
R2 =
0.0
178
0
0.51
1.52 4
.20
4.30
4.40
4.50
4.60
CC
R C
ount
ry R
atin
g
Mean Returns
y =
1.56
44x
- 5.2
269
R2 =
0.0
279
-1012345
3.90
4.00
4.10
4.20
4.30
4.40
4.50
ICR
G P
oliti
cal R
isk
Mean Returns
y =
1.83
85x
- 7.2
269
R2 =
0.0
497
0
0.51
1.52 4
.25
4.30
4.35
4.40
4.45
4.50
ICR
G P
oliti
cal R
isk
Mean Returns
Em
ergi
ng M
arke
ts
-
ity, and is consistent with the notion that the emerging mar-kets are not fully integrated into world capital markets.
For the coskewness measures, the results are similaracross the two methods. In both developed and emerg-ing markets, there is a negative relation, suggesting thatmore negative coskewness gets a higher expected returnwhich is what the theory would suggest.
From an asset pricing perspective again, it is the con-tribution to kurtosis of a well-diversified portfolio (thecokurtosis) that should be priced. We do not attempt tomeasure the cokurtosis. We find a positive relationbetween kurtosis and returns in emerging markets but notin developed markets.
The last measures are country risk ratings. Erb, Har-vey, and Viskanta [1996a, 1996b] show that there is a rela-tion between country risk ratings and subsequent returns.The analysis in Exhibit 3 is different. We are comparingaverage country risk to average returns during the sameperiod. Nevertheless, most of the results are consistent withthose of Erb, Harvey, and Viskanta.
We find for the ICRG Composite and the Institu-tional Investor Country Credit Rating that there is a neg-ative relation in both developed and emerging markets.The relation is weak, however. The political risk measurein exhibit 4R has a positive slope, suggesting that higherpolitical risk is associated with lower expected returnswhich is counter-intuitive.
14 DRIVERS OF EXPECTED RETURNS IN INTERNATIONAL MARKETS FALL 2000
Risk1 / Risk2 c0 p-value c1 p-value c2 p-value R2
SR / TR 0.8522 0.1000 0.2774 0.6730 -0.0385 0.7160 0.012SR / IR 0.8177 0.1130 0.1526 0.7670 -0.0169 0.8420 0.006SR / Size 0.8946 0.1150 0.1220 0.7970 -0.0253 0.7020 0.013SR / Semimean 0.8671 0.1100 0.2681 0.6810 -0.0565 0.7280 0.011SR / Semi - rf 1.0727 0.0270 0.7604 0.1970 -0.2431 0.0970 0.166SR / Semi-0 1.0284 0.0310 0.7421 0.2070 -0.2403 0.1040 0.159SR / Down-biw 0.7699 0.1130 0.0943 0.8600 0.0304 0.9390 0.004
SR / Down-bw 0.7115 0.1310 -0.1446 0.8010 0.2857 0.4470 0.040
SR / VaR 0.9074 0.0650 0.4448 0.4570 0.0380 0.3760 0.053SR / skew 0.8732 0.0700 0.0607 0.8970 -0.2217 0.3920 0.050SR / skew5% 0.6725 0.2100 0.2483 0.6710 -0.3788 0.6910 0.014SR / coskew1 0.6200 0.1830 0.2397 0.6080 -0.6864 0.2330 0.091SR / coskew2 0.6446 0.1660 0.2081 0.6550 -0.5818 0.2610 0.082SR / kurt 0.9574 0.1700 0.0755 0.8760 -0.0354 0.7220 0.012SR / ICRGC 2.8540 0.7990 0.1084 0.2000 -0.4683 0.8530 0.006SR / CCR 3.3791 0.5070 0.0959 0.8390 -0.5868 0.6070 0.020SR / ICRGP -7.9100 0.3930 0.1787 0.7030 1.9555 0.3480 0.059TR / IR 0.9280 0.0790 -0.0079 0.9700 0.0003 0.9990 0.001TR / Size 1.1992 0.1230 -0.0249 0.7700 -0.0339 0.6450 0.014TR / Semimean 0.9171 0.1000 -0.0264 0.9560 0.0295 0.9680 0.001TR / Semi - r * 0.8827 0.0050 0.7752 0.0000 -1.2421 0.0000 0.658TR / Semi-0* 0.6921 0.0300 0.7319 0.0000 -0.1819 0.0000 0.611TR / Down-biw 0.9245 0.0600 -0.0262 0.7900 0.1385 0.7620 0.007
TR / Down-bw 0.9436 0.0460 -0.1133 0.3080 0.5699 0.2060 0.099
TR / VaR 0.8558 0.0720 0.1935 0.3080 0.0967 0.2510 0.082TR / skew 0.9160 0.0580 0.0026 0.9730 -0.2272 0.3820 0.049TR / skew5% 0.8716 0.1500 0.0035 0.9730 -0.1659 0.8740 0.002TR / coskew1 0.9235 0.0520 -0.0129 0.8620 -0.6278 0.2660 0.072TR / coskew2 0.9947 0.0400 -0.0263 0.7290 -0.5798 0.2670 0.077TR / kurt 1.1478 0.1100 -0.0150 0.8480 -0.0429 0.6660 0.013TR / ICRGC 3.1693 0.7780 -0.0081 0.9160 -0.5057 0.8450 0.003TR / CCR 4.1312 0.4510 -0.0224 0.7800 -0.7070 0.5560 0.023TR / ICRGP -7.1810 0.4330 -0.0035 0.0498 1.8327 0.3760 0.050*Extreme multicollinearity.
f
E X H I B I T 5 AMultiple RegressionsDeveloped Markets
-
Multivariate Regressions
The multivariate regressions are presented in Exhibit5. These regressions use two risk factors. We concentrateon regressions with beta and one other risk factor andregressions with total risk and one other risk factor.
If the world behaves according to the augmentedCAPM with skewness, we should have both beta andcoskewness in the regression. It turns out that this type ofregression is reasonably successful. In developed markets,the market beta has a positive coefficient, and the coskew-ness has a negative coefficient. Neither coefficient is sig-nificant at conventional levels in the regression with only19 observations.
In the emerging markets sample, this regression ismore successful. The coefficient on the market beta is sig-nificant, and the coefficient on the coskewness measureis negative and significant, at the 10% level using the sec-ond definition of coskewness. This regression is able toexplain 40% of the emerging market returns. There are28 observations in the regression.
In the full sample, the regression with coskewnessdefinition 2 is quite successful. Both coefficients are signif-icantly different from zero and are the right sign. The regres-sion explains 40% of the variation in the average returns.
The alternative model is one with total variance andtotal skewness. This model does not fare well in developedmarkets; it does much better in emerging markets. The
FALL 2000 EMERGING MARKETS QUARTERLY 15
Risk1 / Risk2 c0 p-value c1 p-value c2 p-value R2
SR / TR -0.8203 0.0800 0.2851 0.3870 0.1546 0.0020 0.532SR / IR -0.7703 0.1010 0.4749 0.1220 0.1436 0.0030 0.519SR / Size 0.1156 0.7980 0.9271 0.0050 0.1295 0.4040 0.338SR / Semimean -0.7869 0.1450 0.2272 0.5640 0.2401 0.0130 0.472SR / Semi - rf -0.2003 0.7430 0.6672 0.1210 0.1238 0.2720 0.352SR / Semi-0 -0.1763 0.7660 0.6641 0.1240 0.1245 0.2710 0.352SR / Down-biw 0.1567 0.6890 0.7405 0.0640 0.4000 0.3060 0.348
SR / Down-bw 0.2985 0.4070 0.7753 0.1190 0.1825 0.5520 0.329
SR / VaR -0.2127 0.7130 0.5912 0.1900 -0.0427 0.2260 0.359SR / skew 0.0722 0.8190 0.9142 0.0020 0.5653 0.0100 0.480SR / skew5% 0.0677 0.8600 1.1087 0.0010 1.4635 0.1490 0.375SR / coskew1 0.2511 0.5460 1.0120 0.0020 -0.6276 0.6200 0.326SR / coskew2 0.0283 0.9390 1.0088 0.0010 -0.7625 0.0720 0.404SR / kurt -0.2439 0.5900 0.9008 0.0030 0.1222 0.0660 0.407SR / ICRGC 5.5166 0.4930 1.0083 0.0020 -1.2202 0.5220 0.331SR / CCR 2.5628 0.2890 1.0108 0.0020 -0.5938 0.3580 0.342SR / ICRGP -3.5800 0.5820 0.9835 0.0030 0.9505 0.5440 0.330TR / IR -0.8258 0.0700 0.5232 0.0400 -0.3656 0.1650 0.554TR / Size -1.2669 0.0180 0.1743 0.0000 0.1879 0.1290 0.561TR / Semimean -0.8183 0.0980 0.2314 0.1010 -0.0842 0.7060 0.520TR / Semi - r -0.3113 0.4650 0.4239 0.0000 -0.4872 0.0040 0.658TR / Semi-0 -0.4206 0.3150 0.4182 0.0000 -0.4763 0.0040 0.653TR / Down-b -1.0380 0.0280 0.1560 0.0000 0.4062 0.1350 0.559
TR / Down-b -0.8699 0.0630 0.1624 0.0010 0.1424 0.4640 0.528
TR / VaR -0.6711 0.1500 0.2888 0.0010 0.0661 0.1350 0.559TR / skew -0.7931 0.0900 0.1618 0.0000 0.2254 0.3080 0.537TR / skew5% -0.8635 0.0690 0.1833 0.0000 -0.2512 0.7700 0.519TR / coskew1 -1.2476 0.0210 0.1903 0.0000 -1.4755 0.1660 0.554TR / coskew2 -1.0189 0.0310 0.1748 0.0000 -0.5026 0.1690 0.553TR / kurt -1.0111 0.0440 0.1684 0.0000 0.0501 0.4180 0.530TR / ICRGC -5.8392 0.4140 0.1881 0.0000 1.1570 0.4860 0.527TR / CCR 0.9363 0.6500 0.1807 0.0000 -0.4873 0.3710 0.533TR / ICRGP -9.7990 0.0710 0.1860 0.0000 2.1230 0.0980 0.568
f
iw
w
E X H I B I T 5 BMultiple RegressionsEmerging Markets
-
loading on total variance is positive, and the loading onskewness is negative which is what one would expect ifthe markets are completely segmented. The two factorscan explain about 52% of the variation in the averagereturns. In the model with all countries, the total varianceremains important but the skewness measures are nolonger significant.
Judging by R2, the best-fitting regression is the onewith total risk and a measure of semivariance, althoughthis is misleading because of the extreme collinearitybetween the semivariance measures and the total risk.
The message of this analysis is as follows. A worldCAPM with coskewness appears to have some ability toexplain average returns in both developed and emerging
markets. In emerging markets, the systematic risks (betaand coskewness) are not complete measures of risk. Thatis, extra variation can be explained by adding total risksto the regression. Even so, the results of Bekaert andHarvey [2000] suggest that these total risks may be becom-ing less important with capital market liberalizations.
IV. SUMMARY
I have examined 18 measures of risk in 47 interna-tional markets. Particular attention is paid to a measure ofdownside implied by asset pricing theory: coskewness.This risk measure captures the contribution that an assetmakes to a well-diversified portfolios total skewness. A
16 DRIVERS OF EXPECTED RETURNS IN INTERNATIONAL MARKETS FALL 2000
Risk1 / Risk2 c0 p-value c1 p-value c2 p-value R2
SR / TR -0.3051 0.2370 0.4638 0.0540 0.1054 0.0000 0.470SR / IR -0.2491 0.3280 0.6110 0.0080 0.0944 0.0000 0.459SR / Size 0.3817 0.2400 0.9780 0.0000 -0.0451 0.5150 0.278SR / Semimean -0.2923 0.2900 0.4350 0.0970 0.1610 0.0010 0.427SR / Semi - rf -0.1101 0.7130 0.6483 0.0200 0.1114 0.0460 0.335SR / Semi-0 -0.0878 0.7640 0.6460 0.0210 0.1119 0.0450 0.335SR / Down-biw 0.0555 0.8360 0.6375 0.0320 0.4776 0.0840 0.320
SR / Down-bw 0.1791 0.4770 0.5493 0.1210 0.3252 0.1300 0.309
SR / VaR -0.0937 0.7440 0.5945 0.0360 -0.0371 0.0320 0.344SR / skew 0.0580 0.8020 0.8795 0.0000 0.5256 0.0010 0.427SR / skew5% 0.0881 0.7430 1.0000 0.0000 1.0737 0.1310 0.308SR / coskew1 0.0873 0.7460 0.9747 0.0000 -1.0720 0.1380 0.307SR / coskew2 -0.0099 0.9680 0.9638 0.0000 -0.8360 0.0040 0.396SR / kurt -0.3296 0.2950 0.8553 0.0000 0.1302 0.0080 0.381SR / ICRGC 7.5150 0.0640 0.9550 0.0000 -1.6895 0.0720 0.323SR / CCR 2.6242 0.0270 0.9561 0.0000 -0.5971 0.0400 0.339SR / ICRGP 2.5330 0.4660 0.9618 0.0000 -0.5372 0.5100 0.278TR / IR -0.4942 0.0790 0.5095 0.0020 -0.3758 0.0150 0.495TR / Size -0.9385 0.0180 0.1566 0.0000 0.1638 0.0100 0.503TR / Semimean -0.0596 0.8290 0.1963 0.0990 -0.1075 0.5750 0.426TR / Semi - r 0.2858 0.1970 0.4402 0.0000 -0.5842 0.0000 0.622TR / Semi-0 0.1582 0.4580 0.4334 0.0000 -0.5715 0.0000 0.616TR / Down-b -0.2859 0.2760 0.1092 0.0000 0.3710 0.1020 0.457
TR / Down-b -0.1824 0.4660 0.1083 0.0010 0.2064 0.2030 0.443
TR / VAR 0.0218 0.9290 0.2729 0.0000 0.0823 0.0240 0.486TR / skew -0.0723 0.7740 0.1168 0.0000 0.1853 0.3000 0.436TR / skew5% -0.1286 0.6050 0.1381 0.0000 -0.5315 0.4260 0.431TR / coskew1 -0.1604 0.5300 0.1299 0.0000 -0.4071 0.5340 0.427TR / coskew2 0.1218 0.6230 0.1230 0.0000 -0.2730 0.3510 0.434TR / kurt -0.2510 0.3760 0.1199 0.0000 0.0467 0.3590 0.433TR / ICRGC -9.1055 0.0530 0.1654 0.0000 2.0120 0.0560 0.469TR / CCR -1.4252 0.3100 0.1456 0.0000 0.2930 0.3470 0.434TR / ICRGP -10.7762 0.0030 0.1714 0.0000 2.4047 0.0030 0.530
f
f
iw
w
E X H I B I T 5 CMultiple RegressionsAll Markets
-
negative coskewness would imply that adding an asset tothe portfolio will decrease the skewness of the portfolio.Given that investors like positive not negative skewness,this asset would have to have a high expected return to getinvestors to purchase it.
Risk measures implied by asset pricing theory, inparticular world beta and coskewness, work reasonablywell in capturing the cross-section of average returns inworld markets. Yet, consistent with the evidence inBekaert and Harvey [1995], many emerging marketsappear to be impacted by total risk measures like varianceand skewness. This is consistent with their less-than-complete integration with world capital markets.
ENDNOTES
I wish to thank Lakshman Easwaran for his excellentresearch assistance and comments.
REFERENCES
Bekaert, Geert. Market Integration and Investment Barriers inEmerging Equity Markets. World Bank Economic Review, 9(1995), pp. 75-107.
Bekaert, Geert, Claude B. Erb, Campbell R. Harvey, andTadas E. Viskanta. The Cross-Sectional Determinants ofEmerging Equity Market Returns. in Peter Carman, ed.,Quantitative Investing of the Global Markets: Strategies, Tactics,and Advanced Analytical Techniques. Chicago: Glenlake Pub-lishing, 1997, pp. 221-272.
. Distributional Characteristics of Emerging MarketReturns and Asset Allocation. The Journal of Portfolio Manage-ment, Vol. 24, No. 2 (Winter 1998), pp. 102-116.
Bekaert, Geert and Campbell R. Harvey. Emerging EquityMarket Volatility. Journal of Financial Economics, Vol. 43, No.1 (1997), pp. 27-77.
. Foreign Speculators and Emerging Equity Markets.Journal of Finance, 55 (2000), pp. 565-613.
. Time-Varying World Market Integration. Journal ofFinance, Vol. 50, No. 2 (1995), pp. 403-444.
Black, Fischer, Michael Jensen, and Myron Scholes. TheCapital Asset Pricing Model: Some Empirical Tests. MichaelJensen, editor. Studies in the Theory of Capital Markets. NewYork: Praeger, 1972.
Erb, Claude B., Campbell R. Harvey, and Tadas E. Viskanta.Expected Returns and Volatility in 135 Countries. The Jour-nal of Portfolio Management, Vol. 22, No. 3 (1996a), pp. 46-58.
. Political Risk, Financial Risk and Economic Risk.Financial Analysts Journal Vol. 52, No. 6 (November/Decem-ber 1996b), pp. 28-46.
Estrada, Javier. The Cost of Equity in Emerging Markets: ADownside Risk Approach. Emerging Markets Quarterly, Vol. 4,No. 3 (Fall 2000), pp PAGES TO FOLLOW ONCEJOURNAL IS PAGINATED
Harvey, Campbell R. Predictable Risk and Returns in Emerg-ing Markets. Review of Financial Studies, Vol. 8. No. 3 (1995a),pp. 773-816.
Harvey, Campbell R. The Risk Exposure of Emerging EquityMarkets. World Bank Economic Review, Vol. 9, No. 1 (1995),pp. 19-50.
. The World Price of Covariance Risk. Journal of Finance,46 (1991), pp. 111-157.
Harvey, Campbell R. and Akhtar Siddique. AutoregressiveConditional Skewness. Journal of Financial and QuantitativeAnalysis, Vol. 34, No. 4 (1999), pp. 465-488.
. Conditional Skewness in Asset Pricing Tests. Journalof Finance, 55 (2000a), pp. 1263-1295.
. The Cross-Section of Expected Risk Exposure. Work-ing paper, Duke University, 2000b.
Kraus, Alan, and Robert Litzenberger. Skewness Preference andthe Valuation of Risk Assets. Journal of Finance, 31 (1976), pp.1085-1100.
Rubinstein, Mark. The Fundamental Theory of Parameter-preference Security Valuation. Journal of Financial and Quan-titative Analysis, 8 (1973), pp. 61-69.
FALL 2000 EMERGING MARKETS QUARTERLY 17