Download presentation
Presentation is loading. Please wait.
Published byHilary Campbell Modified over 6 years ago
1
Factor Analysis, HAR-RV Regressions, and F-Tests
Zed Lamba 4/9/08
2
ECON 201FS Recap from Monday Established through Factor Analysis that Realized Variation amongst 10 tech stocks in S&P 100 had high communality. Average Uniqueness was only , when analysis was done on ln(RV) Question to be settled: how does introducing the market into the mix affect things? Can we establish whether the communality results from an industry effect or a market effect?
3
S&P 500 Log Levels 04/97 – 10/07 Critical to note scale on Y-axis
ECON 201FS S&P 500 Log Levels 04/97 – 10/07 Critical to note scale on Y-axis
4
Factor Analysis on ln(RV)
ECON 201FS Factor Analysis on ln(RV) As before, there is 1 significant common factor Goes against expectation that introducing the market would result in 2 common factors
5
ECON 201FS Factor Analysis contd. Average Uniqueness (ignoring S&P): (slightly lower than before) Average Uniqueness (including S&P):
6
Factor Analysis Analysis
ECON 201FS Factor Analysis Analysis S&P 500 Realized Variation (Uniqueness = ) essentially not well-explained by common factor (0.6 is considered threshold) Possible explanations: Communality amongst stocks results from industry effect S&P 500 contains numerous tech stocks that have not been examined here, so results are biased
7
ECON 201FS HAR-RV Regressions MSFT RV regressed against all stocks (including MSFT) and S&P All regressors are lagged by 1 day, 5-day lag average, and 22-day lag average Regressors’ Explanatory Power (Adjusted R2) = (was without S&P)
8
Testing 1-day lags Regressors Tested F-Statistic P-Value
ECON 201FS Testing 1-day lags Regressors Tested F-Statistic P-Value All stocks’ 1-day lag other than MSFT 5.97 0.0000 S&P day lag added to above 6.29 MSFT 1-day lag added to above 9.28
9
Testing 5-day average lags
ECON 201FS Testing 5-day average lags Regressors Tested F-Statistic P-Value All stocks’ 5-day average lags other than MSFT 9.18 0.0000 S&P day average lag added to above 9.33 (reduction in significance?) MSFT 5-day average lag added to above 14.93
10
Testing 22-day average lags
ECON 201FS Testing 22-day average lags Regressors Tested F-Statistic P-Value All stocks’ 22-day average lags other than MSFT 1.62 0.1034 S&P day average lag added to above 1.46 0.1483 MSFT 22-day average lag added to above 1.47 0.1368
11
ECON 201FS Testing CSCO All 3 lags (1-day, 5-day average, and 22-day average) tested F-Statistic = 4.03 P-Value = Conclusion: Mildly significant
12
ECON 201FS Testing DELL All 3 lags (1-day, 5-day average, and 22-day average) tested F-Statistic = 13.56 P-Value = Conclusion: Strongly significant
13
ECON 201FS Testing EMC All 3 lags (1-day, 5-day average, and 22-day average) tested F-Statistic = 0.13 P-Value = Conclusion: Completely insignificant
14
ECON 201FS Testing HPQ All 3 lags (1-day, 5-day average, and 22-day average) tested F-Statistic = 0.86 P-Value = Conclusion: Completely insignificant; surprising given DELL result. Perhaps Compaq merger can explain some unrelated RV?
15
ECON 201FS Testing IBM All 3 lags (1-day, 5-day average, and 22-day average) tested F-Statistic = 1.05 P-Value = Conclusion: Insignificant, goes with EMC result
16
ECON 201FS Testing INTC All 3 lags (1-day, 5-day average, and 22-day average) tested F-Statistic = 7.63 P-Value = Conclusion: Strongly significant
17
ECON 201FS Testing MSFT All 3 lags (1-day, 5-day average, and 22-day average) tested F-Statistic = 13.1 P-Value = Conclusion: Very strongly significant, as expected
18
ECON 201FS Testing ORCL All 3 lags (1-day, 5-day average, and 22-day average) tested F-Statistic = 2.34 P-Value = Conclusion: Almost significant, not quite
19
ECON 201FS Testing TXN All 3 lags (1-day, 5-day average, and 22-day average) tested F-Statistic = 0.15 P-Value = Conclusion: Completely insignificant
20
ECON 201FS Testing XRX All 3 lags (1-day, 5-day average, and 22-day average) tested F-Statistic = 1.95 P-Value = Conclusion: Insignificant
21
ECON 201FS Testing S&P 500 All 3 lags (1-day, 5-day average, and 22-day average) tested F-Statistic = 1.90 P-Value = Conclusion: Insignificant (recall Possible Explanations from Factor Analysis section)
22
Testing Significant Stocks Together
23
Regressing against Significant Stocks
ECON 201FS Regressing against Significant Stocks Originally, 33 regressors combined for R2 of 12 regressors here have combined R2 of Thus, 12 regressors account for 99.13% of R2 Should have reduced this to 9 regressors by removing 22-day average lags
24
Extensions for Final Presentation
ECON 201FS Extensions for Final Presentation Check S&P 500 handling Now that long-term industry trends have been established, see where GOOG fits in Extract common factor indicated in Factor Analysis and use as regressor, as per Prof. Bollerslev’s suggestion
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.