Journal of Econometrics 152 (2009) 79_80

Contents lists available at ScienceDirect

Journal of Econometrics
journal homepage: www.elsevier.com/locate/jeconom
Nonparametric and robust methods in econometrics
Luiz Renato Limaa,b,_, Marcelo Moreira c, Jack Porter d, Zhijie Xiao e
a EPGE/FGV, Brazil
b University of Tennessee, Knoxville, United States
c Department of Economics, Columbia University, United States
d Department of Economics, University of Wisconsin, Madison, United States
e Department of Economics, Boston College, United States

articleinfo
Article history:
Available online 19 April 2009
Although economic theory generally provides only loose
restrictions on the distribution of observable quantities, much
econometric work is based on tightly specified parametric models
and likelihood-based methods of inference. However, when the
parametric model is not true, the cost of imposing the strong
restrictions required for parametric estimation and testing can be
considerable.
Therefore, much effort has gone into developing procedures
that can be used in the absence of strong a priori restrictions. This
Annals Issue aims to present the latest advances in nonparametric
and robust methods in econometrics. The collection of papers
grew out of the conference ``Econometrics in Rio'' held at
Getulio Vargas Foundation, Rio de Janeiro, in July 2006. The
conference featured 6 keynote speakers, 12 invited speakers,
and 22 contributed papers (out of 70 submissions), congregating
about 50 researchers from Brazil, Canada, Europe and the United
States. The conference took place over three days, with the
presence of most important researchers in Nonparametric and
Robust Methods in Econometrics. The range of topics covered in
this Annals Issue mirrors this breadth, and includes forecasting,
functional-coefficient cointegration models, moment inequalities,
finite-sample inference for quantile regression models, hypothesis
testing against nonparametric alternatives, inference methods
for partially identified models, expectiles and splines, Parametric
Links for Binary Response Models, and instrumental variables.
Victor Chernozhukov, Christian Hansen and Michael Jansson
consider the construction of finite-sample confidence bands for
quantile regression models. The confidence bands are based on
the conditional pivotal property and can be computed using
MCMC. They illustrate the new method through the estimation
of a heterogeneous demand elasticity as well as the estimation
of heterogeneous return to schooling. They report pronounced
_ Corresponding editor.
E-mail addresses: luizr@fgv.br (L.R. Lima), mjmoreira@columbia.edu
(M. Moreira), jrporter@ssc.wisc.edu (J. Porter), xiaoz@bc.edu (Z. Xiao).
differences between confidence regions formed using the usual
asymptotics and confidence regions formed using the finite-
sample procedure in cases where the usual asymptotics are
suspect, such as inference about tail quantiles or inference when
identification is partial or weak.
Zhijie Xiao studies estimation and inference of functional-
coefficient cointegration models. It is shown that the proposed
model offers a more flexible structure of cointegration where
the value of cointegrating coefficients may be affected by
informative covariates and thus may vary over time. Kernel
and local polynomial estimators are investigated, and inference
procedures for instability of cointegrating parameters and test
for cointegration are proposed based on the functional-coefficient
estimates. Limiting distribution of the estimates and testing
statistics are also derived.
Shakeeb Khan and Elie Tamer use conditional moment inequal-
ities to conduct inference on endogenously censored regression
models. They show that, under a quantile restriction, randomly
censored regression models can be written in terms of conditional
moment inequalities. They study the identified features of these
moment inequalities with respect to the regression parameters
and show that such inequalities restrict the parameters to a set.
They also provide a simple way to convert conditional moment
inequalities into unconditional ones while preserving the informa-
tional content. This method obviates the need for nonparametric
estimation, which requires the selection of smoothing parameters
and trimming procedures. Finally, they propose a quantile mini-
mumdistance estimator which converges at the parametric rate to
the parameter vector of interest, and has an asymptotically normal
distribution. They conduct a small scale simulation study and illus-
trate their method through an application using drug relapse data.
Joel Horowitz and Sokbae Lee are concerned with inference
about a function g that is identified by a conditional quantile
regression involving instrumental variables. They propose a
test of the hypothesis that g belongs to a finite-dimensional
parametric family against a nonparametric alternative, and show
0304-4076/$ _ see front matter ' 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.jeconom.2009.04.001
80 Guest editorial / Journal of Econometrics 152 (2009) 79_80
that such a test is not subject to the ill-posed inverse problem
of nonparametric instrumental variables estimation. Asymptotic
results are provided and a Monte Carlo simulation is conducted to
illustrate the finite-sample performance of the test.
Giuliano De Rossi and Andrew Harvey show that time-varying
quantiles satisfy the defining property of fixed quantiles in having
the appropriate number of observations above and below. They
consider time-varying expectiles and show how they can be
estimated by a state space signal extraction algorithm. Because
the state space form can handle irregularly spaced observations,
the proposed algorithms can be adapted to provide a viable means
of computing spline-based non-parametric quantile and expectile
regressions.
Marcelo Moreira applies classical exponential-family statistical
theory to develop a unifying framework for testing structural
parameters in the simultaneous equations model under the
assumption of normal errors with known reduced-form variance
matrix. In the limited-information model, he characterizes the
entire class of similar tests in a model with only one endogenous
explanatory variable. In the full-information framework, he
proposes a family of similar tests for subsets of endogenous
variables' coefficients. When the model is just-identified, the
Anderson_Rubin, score, and (pseudo) conditional likelihood ratio
tests are optimal. When the model is over-identified, the (pseudo)
conditional likelihood ratio test has power close to the power
envelope when identification is strong.
Joao Issler and Luiz Renato Lima propose a novel approach
to econometric forecasting of stationary and ergodic time series
within a panel-data framework. They consider an equal-weighted
forecasting device coupled with bias correction and show that
it is potentially superior to other techniques in several contexts.
In particular, it is asymptotically equivalent to the conditional
expectation, i.e., has an optimal limiting mean-squared error. They
also develop a zero-mean test for the average bias and discuss the
forecast-combination puzzle in small and large samples. Monte
Carlo simulations are conducted to evaluate the performance of
the bias-corrected average forecast in finite samples. An empirical
exercise, based upon data from a well-known survey is also
presented. Overall, these results show promise for the feasible
bias-corrected average forecast.
Roger Koenker and Jungmo Yoon show that although the
familiar logit and probit models provide convenient settings
for many binary response applications, a larger class of link
functions may be occasionally desirable. They consider two
parametric families of link functions: the Gosset link based
on the Student t latent variable model with the degrees of
freedom parameter controlling the tail behavior, and the Pregibon
link based on the (generalized) Tukey family with two shape
parameters controlling skewness and tail behavior. Both Bayesian
and maximum likelihood methods for estimation and inference are
explored. Implementations of the methods are illustrated in the R
environment.
Alfred Galichon and Marc Henry propose a test of the validity
of a set of theoretical restrictions on the relationship between
economic variables, which do not necessarily identify the data
generating process. The restrictions can be derived from any model
of interactions, allowing censoring and multiple equilibria. When
the restrictions are parameterized, the test can be inverted to yield
confidence regions for partially identified parameters.
Antonio Galvao extends unit root tests based on quantile re-
gression proposed by Koenker and Xiao (2004) to allow station-
ary covariates and a linear time trend. A simulation experiment
is described, illustrating the finite-sample performance of the unit
root test for several types of distributions. The test based on quan-
tile autoregression turns out to be especially advantageous when
innovations are heavy-tailed. An application to the CPI-based real
exchange rates using four different countries suggests that real ex-
change rates are not constant unit root processes.
Acknowledgements
Weare indebted to Cristina Igreja and Tania Macario for helping
us organize the conference. We also wish to thank the following
sponsors for financial support: Getulio Vargas Foundation, CAPES,
CNPq and Banco Itau. Finally, we are grateful to Cheng Hsiao and
Sena H. Schlessinger for their support and guidance in setting up
this volume as well as to all referees for their insightful comments,
which greatly improved the quality of the contributions to this
volume.