1. Final Presentation
Mingwei Lei
Econ 201

2. Research Idea
Past research have shown evidence of high asset correlations in the period of heightened market volatility:
Campbell, Koedijk, and Kofman- 2002
Butler Joaquin
This phenomenon is also well known in the business industry
Empirical exploration of the relationship between asset returns correlation and market (SPY) volatility

3. The Process Pair up stocks to be analyzed along with SPY
Match up data of stocks and SPY
Partition data into periods (1-day, 5-days, 20 days) to be analyzed
Find the optimal sampling frequency to calculate returns correlation for each partition
Plot correlation against market standard deviation
Perform transformations (log, Fisher) to attain a more linear relationship
Perform regression analysis

4. Correlation Signature (Period- 1 day)

5. Correlation Signature (Period- 5 days)

6. Correlation Signature (Period- 20 days)

7. BAC & GS Correlation vs. Market Standard Deviation (Period – 1 day)

8. BAC & GS Correlation vs. Ln(MktStd) (Period – 1 day)

9. BAC & GS Fisher Transformed Correlation vs. MktStd (Period – 1 day)

10. BAC & GS Fisher Transformed Corr vs. Ln(MktStd) (Period – 1 day)

11. BAC & GS Correlation vs. Market Standard Deviation (Period – 5 days)

12. BAC & GS Correlation vs. Ln(MktStd) (Period – 5 days)

13. BAC & GS Fisher Transformed Correlation vs. MktStd (Period – 5 days)

14. BAC & GS Fisher Transformed Corr vs. Ln(MktStd) (Period – 5 days)

15. BAC & GS Correlation vs. Market Standard Deviation (Period – 20 days)

16. BAC & GS Correlation vs. Ln(MktStd) (Period – 20 days)

17. BAC & GS Fisher Transformed Correlation vs. MktStd (Period – 20 days)

18. BAC & GS Fisher Transformed Corr vs. Ln(MktStd) (Period – 20 days)

19. Regression Results (Period- 1 day)
Regressand
Regressor β1
t-stat β0 R2
BAC and GS
Corr
ln(MktStd)
0.148
17.55
1.140
28.39
0.0955
Fisher Corr
MktStd
15.51
14.05
0.362
28.40
0.1083
0.229
18.11
1.600
26.20
0.1200
JPM & GS
0.134
15.77
1.098
27.36
0.0835
14.15
13.66
.416
34.43
0.0908
0.214
16.51
1.572
25.06
0.1058

20. Regression Results Cont. (Period- 1 day)
Regressand
Regressor β1
t-stat β0 R2
WMT and JPM
Corr
ln(MktStd)
0.1563
18.68
1.086
27.76
0.1019
Fisher Corr
MktStd
13.94
14.27
0.255
21.67
0.1028
0.2070
18.54
1.373
25.92
0.1151
WMT and KO
0.1475
15.29
0.983
21.59
0.0851
12.96
12.11
0.189
15.32
0.0915
0.1877
15.55
1.206
21.02
0.0975
WMT and VZ
0.1954
22.66
1.243
30.42
0.1634
14.39
0.192
15.05
0.1496
0.2469
21.97
1.53
28.30
0.1768

21. Regression Results (Period- 5 days)
Regressand
Regressor β1
t-stat β0 R2
BAC and GS
Corr
ln(MktStd)
0.1290
9.50
0.956
17.79
0.1131
Fisher Corr
MktStd
5.507
7.73
0.391
20.23
0.1150
0.1878
10.01
1.248
16.52
0.1348
JPM & GS
0.1004
8.42
0.881
18.44
0.0860
4.420
7.89
0.462
28.85
0.0829
0.1537
8.85
1.161
10.45
0.1014

22. Regression Results Cont. (Period- 5 days)
Regressand
Regressor β1
t-stat β0 R2
WMT and JPM
Corr
ln(MktStd)
0.1233
10.17
0.832
17.82
0.1336
Fisher Corr
MktStd
4.688
8.82
0.275
18.17
0.1341
0.1531
10.31
0.976
16.84
0.1431
WMT and KO
0.1242
8.78
0.768
13.97
0.1298
4.822
6.36
0.194
10.11
0.1463
0.1504
8.65
0.889
13.02
0.1439
WMT and VZ
0.1725
13.14
0.987
18.91
0.2462
6.185
10.57
0.195
12.70
0.2426
0.210
12.88
1.156
17.68
0.2679

23. Regression Results (Period- 20 days)
Regressand
Regressor β1
t-stat Β0 R2
BAC and GS
Corr
ln(MktStd)
0.1314
5.74
0.869
11.64
0.1338
Fisher Corr
MktStd
2.763
5.21
0.380
12.35
0.1343
0.186
6.10
1.098
10.80
0.1562
JPM & GS
0.0843
4.60
0.763
11.98
0.0858
1.980
5.44
0.470
22.13
0.0910
0.1316
5.03
0.979
10.67
0.1030

24. Regression Results Cont. (Period- 20 days)
Regressand
Regressor β1
t-stat β0 R2
WMT and JPM
Corr
ln(MktStd)
0.1229
6.69
0.735
12.59
0.2021
Fisher Corr
MktStd
2.266
6.81
0.264
13.15
0.1908
0.1470
6.84
0.836
11.99
0.2084
WMT and KO
0.1240
5.90
0.672
9.92
0.2040
2.426
5.95
0.180
8.73
0.2299
0.1463
5.80
0.756
9.20
0.2149
WMT and VZ
0.1618
8.12
0.828
12.39
0.3351
3.156
11.44
0.183
10.92
0.3666
0.1947
8.06
0.951
11.70
0.3593

25. Conclusions
The results definitely suggest that there exists a negative relationship between asset correlations and market volatility
Results imply that diversification works the least when it is needed the most
Portfolio managers and risk management practices must allow for time variant asset correlations and understand how asset correlations change with the market