Scribd
Upload
469 views

Quantile Regression

Quantile Regression

Uploaded by

HassaanAhmad
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
469 views

Quantile Regression

Quantile Regression

Uploaded by

HassaanAhmad
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

Quantile Regression

Ani Katchova

© 2013 by Ani Katchova. All rights reserved.


Quantile Regression Overview

 Quantile regression examples


 Quantile regression model
 Quantile regression coefficients and marginal effects
 Advantages of quantile regression

2
Quantile Regression

Quantile regression examples

 Consumer economics: effects of household income on food expenditures for low- and high-
expenditure households.
 Education: factors affecting student scores along their score distribution.

Quantile regression model explanations

 We use quantiles to describe the distribution of the dependent variable.


 Quantiles and percentiles are synonymous – the 0.99 quantile is the 99th percentile.
 The best-known quantile is the median. The median is the 0.50 quantile.
 The standard Ordinary Least Squares (OLS) models the relationship between one or more
independent variables x and the conditional mean of a dependent variable y.
 A quantile regression models the relationship between x and the conditional quantiles of y
rather than just the conditional mean of y.
 A quantile regression gives a more comprehensive picture of the effect of the independent
variables on the dependent variable.
 The dependent variable is continuous with no zeros or too many repeated values.

3
Quantile regression model

 The quantile regression is described by the following equation:


′

where is the vector of unknown parameters associated with the qth quantile.

 The OLS minimizes ∑ , the sum of squares of the model prediction error
 The median regression, also called least absolute-deviation regression minimizes ∑ | |
 The quantile regression minimizes ∑ | | ∑ 1 | |, a sum that gives the asymmetric
penalties | | for underprediction and 1 | | for overprediction

The qth quantile regression estimator minimizes over the objective function

∑: ′ ′
′ ∑ : ′ 1

where 0 1.

4
 In contrast to OLS and maximum likelihood, the quantile regression computational
implementation uses linear programming methods.
 We have instead of β to make clear that different choices of q estimate different values of β.

Quantile regression coefficients and marginal effects

 The standard conditional quantile is specified to be linear:

′
|

 For the jth regressor, the marginal effect is the coefficient for the qth quantile.

 A quantile regression parameter estimates the change in a specified quantile q of the


dependent variable y produced by a one unit change in the independent variable .
 The marginal effects are for infinitesimal changes in the regressor, assuming that the dependent
variable remains in the same quantile.

5
 Unlike interpretations of OLS regression, the interpretations of quantile regression results need
to specify which quantile of the dependent variable they refer to.
 Two types of significance are important for quantile regression coefficients.
o Quantile coefficients can be significantly different from zero.
o Quantile coefficients can be significantly different than the OLS coefficients, showing
different effects along the distribution of the dependent variable.

Advantages of the quantile regression

 Flexibility for modeling data with heterogeneous conditional distributions.


 Median regression is more robust to outliers than the OLS regression.
 Richer characterization and description of the data: can show different effects of the
independent variables on the dependent variable depending across the spectrum of the
dependent variable.

You might also like