Abstract
Thomas G. Moehrle (1995) "Nondurables
and Services Responsiveness in a Nonlinear Model: More Evidence
of Binding Borrowing Constraints."
Recent empirical analyses using the Euler Equation
approach have repeatedly rejected the stochastic implications
of the Rational Expectations-Life Cycle Hypothesis. Several
authors have argued that empirical results using aggregated
data are inconclusive because of aggregation bias.
Statistical conclusions from panel studies using food
expenditure proxies are suspect as well. This study
reexamines the stochastic implications of the hypothesis
using a broader expenditure definition constructed from the
Consumer Expenditure Survey. The study constructs sample
orthogonality conditions from the underlying nonlinear Euler
equations, and thereby relaxes the linear assumptions applied
in previous studies. If the error structure of the model is
conditionally heteroscedastic, it is shown that linearizing
can produce bias and inconsistent estimation. For this study,
hypothesis tests and parameter estimates are attained using
the numerical solution of the generalized method of moments
minimization problem. The hypothesis-testing methodology is
flexible in that it tests the "pure" Rational
Expectations Hypothesis against a borrowing constraint
hypothesis through a split sample technique. Various test
results from this study consistently show that borrowing
constraints do affect consumption growth of financially
strapped households. These conclusions are not as sharp when
using food expenditure data alone, however.
Last Modified Date: July 19, 2008
|