15
Return Forecasting by Quantile Regression QWAFAFEW December 2010 Larry Pohlman and Lingji

Return Forecasting by Quantile Regression QWAFAFEW December 20101 Larry Pohlman and Lingjie Ma

Embed Size (px)

Citation preview

Page 1: Return Forecasting by Quantile Regression QWAFAFEW December 20101 Larry Pohlman and Lingjie Ma

Return Forecasting by Quantile Regression

QWAFAFEWDecember 20101

Larry Pohlman and Lingjie Ma

Page 2: Return Forecasting by Quantile Regression QWAFAFEW December 20101 Larry Pohlman and Lingjie Ma

2

Outline

• The Math• Examples• Multivariate Model• Results

Page 3: Return Forecasting by Quantile Regression QWAFAFEW December 20101 Larry Pohlman and Lingjie Ma

3

The Math and Code

• Model

• OLS Estimation

• QR Estimation

• R, S+, Stat, SAS

uXby

2)(minˆ bxy TiibOLS

Tii

Tiixy bxy

Tii

TiibQR bxybxy

))(1()(minˆ

Page 4: Return Forecasting by Quantile Regression QWAFAFEW December 20101 Larry Pohlman and Lingjie Ma

4

What does QR do?

T

uT

y

uTT

y

TT

xFxxQ

FxxxF

uδxβxy

))((

)(

)(

1

11

Sample Quantile:

Conditional Quantile:

1,0,:

,yYProb

y

y

FyfQ

F

Page 5: Return Forecasting by Quantile Regression QWAFAFEW December 20101 Larry Pohlman and Lingjie Ma

5

Example: Book to Price

Page 6: Return Forecasting by Quantile Regression QWAFAFEW December 20101 Larry Pohlman and Lingjie Ma

6

Why QR?

A natural question is under what conditions will QR be “better” than OLS?

1. Full picture view: heterogeneity

If there is heterogeneity, then QR will provide a more complete view of the relationship between variables through the effects of independent variables across quantiles of the response distribution.

2. Robustness: fat-tail distribution

If the conditional return distribution is not Gaussian but fat-tailed, the QR estimates will be more robust and efficient than the conditional mean estimates

Page 7: Return Forecasting by Quantile Regression QWAFAFEW December 20101 Larry Pohlman and Lingjie Ma

7

Example Price Momentum

Page 8: Return Forecasting by Quantile Regression QWAFAFEW December 20101 Larry Pohlman and Lingjie Ma

8

Example Return on Equity

Page 9: Return Forecasting by Quantile Regression QWAFAFEW December 20101 Larry Pohlman and Lingjie Ma

9

Multivariate Model

• Book to Price• Earnings to Price• Cashflow to Enterprise Value• Balance sheet Accruals• Return on Equity• Price Momentum (9 months)• Earnings Momentum (9 months)

Page 10: Return Forecasting by Quantile Regression QWAFAFEW December 20101 Larry Pohlman and Lingjie Ma

10

Model Variable Plots

Page 11: Return Forecasting by Quantile Regression QWAFAFEW December 20101 Larry Pohlman and Lingjie Ma

11

Results: Equal Weight Quintiles

Page 12: Return Forecasting by Quantile Regression QWAFAFEW December 20101 Larry Pohlman and Lingjie Ma

12

Results: Cap Weighted Qunitiles

Page 13: Return Forecasting by Quantile Regression QWAFAFEW December 20101 Larry Pohlman and Lingjie Ma

13

Optimized Portfolios TE=3%

Page 14: Return Forecasting by Quantile Regression QWAFAFEW December 20101 Larry Pohlman and Lingjie Ma

14

Optimized Portfolios TE=6%

Page 15: Return Forecasting by Quantile Regression QWAFAFEW December 20101 Larry Pohlman and Lingjie Ma

15

Conclusion

■ Conditional mean method is still attractive

■QR provides a full-picture distributional view

■ Link between distribution estimates and point portfolio.