Exponential Regression of Fractional-Response Fixed-Effects Models with an Application to Firm Capital Structure

New fixed-effects estimators are proposed for logit and complementary loglog fractional regression models. The standard specifications of these models are transformed into a form of exponential regression with multiplicative individual effects and time-variant heterogeneity, from which four alternat...

Full description

Saved in:
Bibliographic Details
Published in:Journal of econometric methods 2018-01, Vol.7 (1)
Main Authors: Ramalho, Esmeralda A., Ramalho, Joaquim J.S., Coelho, Luís M.S.
Format: Article
Language:English
Subjects:
C23
C25
Capital structure
Data models
dynamic models
endogeneity
exponential regression
fixed effects
fractional responses
heterogeneity
Longitudinal studies
Monte Carlo simulation
panel data
Variables
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:New fixed-effects estimators are proposed for logit and complementary loglog fractional regression models. The standard specifications of these models are transformed into a form of exponential regression with multiplicative individual effects and time-variant heterogeneity, from which four alternative estimators that do not require assumptions on the distribution of the unobservables are proposed. All new estimators are robust to both time-variant and time-invariant heterogeneity and can accomodate fractional responses with observations at the boundary value of zero. Additionally, some of these estimators can be applied to dynamic panel data models and can accommodate endogenous explanatory variables without requiring the specification of a reduced form model. A Monte Carlo study and an application to firm capital structure choices illustrate the usefulness of the suggested estimators.
ISSN:2156-6674
2194-6345
2156-6674
DOI:10.1515/jem-2015-0019