quantregCF 1.0.0

The control function approach for quantile regression models

QuantregCF

This Python package is an implementation of the control function approach for quantile regression models proposed by Sokbae Lee in "Endogeneity in Quantile Regression Models: A Control Function Approach," Journal of Econometrics, 141: 1131-1158, 2007.

Installation

This project can be installed using pip:

pip install quantregCF

Usage

from quantregCF import quantregCF

beta, se = quantregCF(option, degree, tau_first_stage, tau_second_stage, data)

quantregCF returns a list of estimated coefficients and a list of standard errors.

Parameters

The main function is

quantregCF(option, degree, tau_first_stage, tau_second_stage, data)

option = 0 if the second-stage quantile regression uses polynomial series.

option = 1 if the second-stage quantile regression uses B-spline.

degree is the degree of the polynomial/B-spline. (integer)

tau_first_stage is the value of tau for the first-stage quantile regression. (between 0 and 1)

tau_second_stage is the value of tau for the second-stage quantile regression. (between 0 and 1)

data is a list of length two that contains information on the dataset.

data = [`dataframe`, `var_lst`]

Each element in data is defined as followed:

dataframe is the dataset in the pandas DataFrame format.

var_lst is a list of length four that contains the name of variables of interest. The first two elements of var_lst are strings and the last two elements are lists of strings.

`var_lst` = [`dep_var`, `endog_var`, `exog_var_lst`, `iv_var_lst`]

Each element in var_lst is defined as followed:

dep_var and endog_var are the names of dependent variable and endogenous right-hand side variable.

exog_var_lst and iv_var_lst are the lists of names of exogenous included variables and instrumental variables.

Example

The file fishdata.py illustrates how to use quantregCF on the well-known Graddy’s Fulton fish market data. To run the code, simply download fishdata.py from the tests folder and run python fishdata.py. Below is a code snippet showing how to load the data and use quantregCF.

# load data into DataFrame `df` from https://www.kathryngraddy.org/research#pubdata
data_source = "https://uploads-ssl.webflow.com/629e460595fdd36617348189/62a0fd19b6742078eed59f47_fish.out.txt"
df = pd.read_csv(data_source, sep="\t")
var_lst = ['qty', 'price', ["day1", "day2", "day3", "day4"], ["stormy", "mixed"]]
data_lst = [df, var_lst]

# regressions using B-splines in the second-stage
beta, se = quantregCF(option=1, degree=3, tau_first_stage=0.5, tau_second_stage=0.5, data=data_lst)

# calculate the 95% confidence interval
ci_lb = beta[0] - 1.96 * se[0]
ci_ub = beta[0] + 1.96 * se[0]

Dependencies

• NumPy
• Pandas
• SciPy
• CVXPY
• Urllib