lecture 1: empirical properties of returnsfaculty.washington.edu/ezivot/econ589/empirical... ·...
TRANSCRIPT
1
Lecture 1: Empirical Properties of Returns
Econ 589Eric Zivot
© Eric Zivot 2008
Eric ZivotSpring 2011
Updated: March 29, 2011
Daily CC Returns on MSFT
0.1
r(t)
-0.2
-0.1
0.0
© Eric Zivot 2008
1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003
-0.3
2
Daily CC Returns on S&P 500
00.
05-0
.15
-0.1
0-0
.05
0.00
© Eric Zivot 2008
1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003
-0.2
0
60
Daily CC Returns on MSFTSample Quantiles:
min 1Q median 3Q max -0.3583 -0.01282 0 0.01554 0.1787
Distribution of Daily CC Returns on MSFT
20
30
40
50
Per
cent
of T
otal
Sample Moments: mean std skewness kurtosis
0.001278 0.02539 -0.7275 16.12
Number of Observations: 4365
© Eric Zivot 2008
0
10
-0.3 -0.2 -0.1 0.0 0.1 0.2
x
3
50
60Daily CC Returns on S&P 500 Sample Quantiles:
min 1Q median 3Q max -0.229 -0.004697 0.0004677 0.00581 0.08709
Sample Moments:
Distribution of Daily CC Returns on S&P 500
10
20
30
40
Per
cent
of T
otal
pmean std skewness kurtosis
0.0003276 0.01135 -2.083 45.18
Number of Observations: 4365
© Eric Zivot 2008
0
10
-0.20 -0.15 -0.10 -0.05 0.0 0.05 0.10
x
0.2 Daily CC Returns on MSFT10/19/2000
Normal QQ-Plot
Test for Normality: Jarque-Bera
Null Hypothesis: data is normally
Test for Normality
-0.1
0.0
0.1Null Hypothesis: data is normally distributed
Test Statistics:MSFT
Test Stat 31685.64p.value 0.00
Dist. under Null: chi-square with 2 degrees of freedom
Total Observ.: 4365
© Eric Zivot 2008
-0.3
-0.2
-2 0 2
10/26/1987
10/19/1987
4
0.2
Daily CC Returns on MSFT
06/30/2003
QQ-Plot: Student-t with 4 degrees of freedom
-0.2
-0.1
0.0
0.1
03/17/1986
© Eric Zivot 2008
-0.3
-10 -5 0 5 10
03/17/1986
03/14/1986
Skew Normal Distribution
40.6
shape=5
40.6
shape=-5
-3 -2 -1 0 1 2 3
0.0
0.2
0.4
x.vals
-3 -2 -1 0 1 2 3
0.0
0.2
0.4
x.vals
0.3
0.4
shape=0
0.6
0.8
shape=1000
© Eric Zivot 2008
-3 -2 -1 0 1 2 3
0.0
0.1
0.2
x.vals
-3 -2 -1 0 1 2 3
0.0
0.2
0.4
x.vals
ξ = 0, ω = 1
5
Skew t Distribution
40.6
shape=5
StSn
40.6
shape=-5
StSn
-4 -2 0 2 4
0.0
0.2
0.4
x.vals
pdft
-4 -2 0 2 4
0.0
0.2
0.4
x.vals
pdft
30.
4
shape=0
StSn 6
0.8
shape=1000
StSn
© Eric Zivot 2008
-4 -2 0 2 4
0.0
0.1
0.2
0.3
x.vals
pdft
-4 -2 0 2 40.
00.2
0.4
0.6
x.vals
pdft
ξ = 0, ω = 1, ν = 5
QQ-Plot: MLE of Skew-t for MSFTlocation = -0.004, scale = 0.020, shape = 0.298, df = 4.973
0.2
-0.1
0.0
0.1
msft
© Eric Zivot 2008
-0.1 0.0 0.1 0.2
-0.3
-0.2
st quantiles
mle computed with R package sn, qqPlot() from R package car
6
0 05
Daily CC Returns on S&P 50010/21/1987
Test for Normality: Jarque-Bera
Null Hypothesis: data is normally
Normal QQ-Plot Test for Normality
-0.15
-0.10
-0.05
0.0
0.05
10/26/1987
Null Hypothesis: data is normally distributed
Test Statistics:SP500
Test Stat 326705.3p.value 0.0
Dist. under Null: chi-square with 2 degrees of freedom
© Eric Zivot 2008
-0.20
-2 0 2
10/19/1987
Total Observ.: 4365
0.1Daily CC Returns on S&P 500
06/30/2003
QQ-Plot: Student-t with 4 degrees of freedom
-0.1
0.0
03/17/1986
© Eric Zivot 2008
-0.2
-10 -5 0 5 10
03/14/1986
7
QQ-Plot: MLE of Skew-t for SP500location = 0.001, scale = 0.007, shape = -0.099, df = 3.329
0.10
-0.10
-0.05
0.00
0.05
sp50
0
© Eric Zivot 2008
-0.10 -0.05 0.00 0.05 0.10
-0.20
-0.15
-
st quantiles
mle computed with R package sn, qqPlot() from R package car
Monthly CC Returns on MSFT
0.3
0.4
r(t)
3-0
.2-0
.10.
00.
10.
2
© Eric Zivot 2008
1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003
-0.4
-0.3
8
Monthly CC Returns on S&P500
.05
0.10
-0.1
5-0
.10
-0.0
50.
000.
© Eric Zivot 2008
1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003
-0.2
0
30
Monthly CC Returns on MSFTSample Quantiles:
min 1Q median 3Q max -0.3861 -0.03587 0.0336 0.09736 0.4384
Distribution for Monthly Returns on MSFT
10
20
Per
cent
of T
otal
Sample Moments: mean std skewness kurtosis
0.03357 0.1145 0.1845 4.004
Number of Observations: 208
© Eric Zivot 2008
0
-0.4 -0.2 0.0 0.2 0.4
x
9
30
Monthly CC Returns on S&P500Sample Quantiles:
min 1Q median 3Q max
Distribution for Monthly Returns on S&P 500
10
15
20
25
Perc
ent o
f Tot
al
max -0.2066 -0.01921 0.0122 0.03875
0.125
Sample Moments: mean std skewness kurtosis
0.008219 0.0459 -0.8377 5.186
Number of Observations: 208
© Eric Zivot 2008
0
5
-0.2 -0.1 0.0 0.1
x
0.401/01/2001
01/01/1987
Test for Normality: Shapiro-Wilks
Test Statistics:MSFT
Normal QQ-Plot Tests for Normality
0.0
0.2
Test Stat 0.9906p.value 0.9558
Dist. under Null: normalTotal Observ.: 208
Test for Normality: Jarque-Bera
Test Statistics:MSFT
Test Stat 9.9223
© Eric Zivot 2008
-0.4
-0.2
-3 -2 -1 0 1 2 3
04/01/2000
p.value 0.0070
Dist. under Null: chi-square with 2 degrees of freedom
Total Observ.: 208
10
QQ-Plot: Student’s t with 10 df
0.405/01/2003
06/01/2003
0.0
0.2
05/01/2003
© Eric Zivot 2008
-0.4
-0.2
-2 0 2
03/01/1986
0.1
Monthly CC Returns on S&P500
01/01/1987
Test for Normality: Shapiro-Wilks
Test Statistics:SP500
Normal QQ-Plot Tests for Normality
-0.1
0.0
Test Stat 0.9699p.value 0.0154
Dist. under Null: normalTotal Observ.: 208
Test for Normality: Jarque-Bera
Test Statistics:SP500
Test Stat 65.7551
© Eric Zivot 2008
-0.2
-3 -2 -1 0 1 2 3
08/01/1998
10/01/1987
p.value 0.0000
Dist. under Null: chi-square with 2 degrees of freedom
Total Observ.: 208
11
Monthly CC Returns on S&P500
06/01/2003
QQ-Plot: Student’s t with 7 df
-0.1
0.0
0.1
© Eric Zivot 2008
-0.2
-4 -2 0 2 4
04/01/1986
03/01/1986
Testing for Autocorrelation
AC
F.4
0.6
0.8
1.0
Daily CC Returns on MSFT
Test for Autocorrelation: Ljung-Box
Null Hypothesis: no autocorrelation
Test Statistics:SP500
Test Stat 35.1892p.value 0.0191
Lag0 10 20 30
0.0
0.2
0.6
0.8
1.0
Daily CC Returns on S&P 500
Dist. under Null: chi-square with 20 degrees of freedom
Total Observ.: 4365
Test for Autocorrelation: Ljung-Box
Null Hypothesis: no autocorrelation
Test Statistics:MSFT
© Eric Zivot 2008
Lag
AC
F
0 10 20 30
0.0
0.2
0.4
0 MSFT Test Stat 43.2323p.value 0.0019
Dist. under Null: chi-square with 20 degrees of freedom
Total Observ.: 4365
12
Stylized Facts of Daily Asset Returns
0.30
0.10
Microsoft Returns
0.20
0.05
S & P 500 Returns
1986 1990 1994 1998 2002
-0
1986 1990 1994 1998 2002
-0
1986 1990 1994 1998 2002
0.00
0.08
Microsoft Squared Returns
1986 1990 1994 1998 2002
0.00
00.
040
S & P 500 Squared ReturnsVolatility clustering
1986 1990 1994 1998 2002
0.00
0.30
Microsoft Absolute Returns
1986 1990 1994 1998 2002
0.00
0.30
S & P 500 Absolute Returns
Sample Autocorrelations of Daily Returns
L
AC
F
0 5 10 15 20
0.0
0.6
Microsoft Returns
L
AC
F
0 5 10 15 20
0.0
0.6
S&P 500 Returns
Lag Lag
Lag
AC
F
0 5 10 15 20
0.0
0.6
Microsoft Squared Returns
Lag
AC
F
0 5 10 15 20
0.0
0.6
S&P 500 Squared Returns
Microsoft Absolute Returns Microsoft Absolute Returns
Dependence in volatility
Lag
AC
F
0 5 10 15 20
0.0
0.6
Lag
AC
F
0 5 10 15 20
0.0
0.6
13
Stylized Facts for Monthly Asset Returns
1986 1990 1994 1998 2002
-0.3
0.4
Microsoft Returns
1986 1990 1994 1998 2002
-0.2
00.
10
S&P 500 Returns
1986 1990 1994 1998 2002
0.02
0.18
Microsoft Squared Returns
1986 1990 1994 1998 2002
0.00
5
S&P 500 Squared Returns
Microsoft Squared Returns S&P 500 Squared Returns
Less volatility clustering
Less volatility dependence
Lag
AC
F
0 5 10 15 20
0.0
0.6
q
Lag
AC
F
0 5 10 15 20
0.0
0.6
q
MSFT and S&P 500 Daily Returns
0.0
SP500
0.0
0.1
MSFT
-0.2
-0.1 Sample covariance matrix
MSFT SP500 MSFT 0.0006380948 0.0001700987
SP500 0.0001700987 0.0001262902
Sample correlation matrixMSFT SP500 MSFT 1.0000000 0.5992023SP500 0.5992023 1.0000000
-0.3
-0.2
-0.1
1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003
14
EWMA Volatilities and Correlations
EWMA Conditional Correlation
SP500
EWMA Conditional Volatilities
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Correlations
60.
08
MSFT
0.01
0.02
0.03
0.04
SP500
0.1
0.2
0.3
1986 1988 1990 1992 1994 1996 1998 2000 20020.02
0.04
0.06
198619871988198919901991199219931994199519961997199819992000200120022003