Simulation-based inference: A survey with special reference to panel data models

doi:10.1016/0304-4076(93)90037-6

Journal of Econometrics

Volume 59, Issues 1–2, September 1993, Pages 5-33

https://doi.org/10.1016/0304-4076(93)90037-6 Get rights and content

Abstract

In this article we study the recent developments of inference methods based on simulations. In particular, we discuss the Simulated Generalized Method of Moments, the Simulated Maximum Likelihood Method, and the Simulated Pseudo Maximum Likelihood Methods. The asymptotic properties of the estimators are described when the number of observations n goes to infinity, and we distinguish the case where the number H of simulations per observation is fixed from the case where this number also goes to infinity. In the former case, the possible asymptotic bias is evaluated and, in the latter case, we carefully examine the consequences of the assumptions on the relative divergence rates of n and H. We also show how these methods apply to various contexts, in particular to the case of panel data models.

References (30)

C. Gourieroux et al.
Simulated residuals
Journal of Econometrics
(1987)
L. Bauwens
Bayesian full information analysis of simultaneous equation models using integration by Monte Carlo
(1984)
H. Bierings et al.
Pseudo maximum likelihood techniques in a simple rationing model of the Dutch labour market
(1989)
A. Börsch-Supan et al.
Smooth unbiased multivariate probability simulators for maximum likelihood estimation of limited dependent variable models
B. Brown
Simulation based semi-parametric estimation and prediction in nonlinear systems
(1990)
B.W. Brown et al.
Residual-based procedure for prediction and estimation in nonlinear simultaneous systems
Econometrica
(1984)
J.S. Batler et al.
A computationally efficient quadrature procedure for one-factor multinomial probit model
Econometrica
(1982)
D. Duffie et al.
Simulated moments estimation of Markov models of asset prices
(1989)
J. Geweke
Efficient simulation from the multivariate normal distribution subject to linear inequality contraints and the evaluation of constraint probabilities
(1989)
C. Gourieroux et al.
Additive log-differenced probability models for count data

C. Gourieroux et al.

Simulation based econometrics in models with heterogeneity

Annales d'Economie et de Statistique no. 20/21

(1991)

C. Gourieroux et al.

Pseudo maximum likelihood methods: Theory

Econometrica

(1984)

V.A. Hajivassiliou et al.

Country heterogeneity and external debt crises: Estimation by the method of simulated moments

(1989)

V. Hajivassiliou et al.

Simulation of multivariate normal rectangle probabilities: Methods and programs

(1991)

L.P. Hansen

Large sample properties of generalized method of moments estimators

Econometrica

(1982)

Cited by (99)

Simulation of multidimensional diffusions with sticky boundaries via Markov chain approximation
2023, European Journal of Operational Research
Citation Excerpt :
For multidimensional diffusions, we can consider using our simulation method in a simulation-based estimation approach, such as those discussed in Gourieroux & Monfort (1993) and Gallant & Tauchen (1996).
In this study, we develop a novel method to simulate multidimensional diffusions with sticky boundaries for which the Euler scheme fails. We approximate the sticky diffusion process by a multidimensional continuous-time Markov chain (CTMC) that can be easily simulated. We provide two approaches to construct the CTMC. In the first approach, we approximate the infinitesimal generator of the sticky diffusion by finite difference using standard coordinate directions. In the second one, we match the local moments using the drift and the eigenvectors of the covariance matrix as transition directions. The first approach does not always guarantee a valid Markov chain, whereas the second one can. We show that both construction methods yield a first-order simulation scheme, which can capture the sticky behavior and is free from the curse of dimensionality. As applications, we use our method to simulate the limit of a queuing system with exceptional service policy and a multi-factor short rate model for low interest rate environment.
Prediction of crop biophysical variables with panel data techniques and radar remote sensing imagery
2021, Biosystems Engineering
Citation Excerpt :
Finally, the panel data methodology for time series was applied to study the possible relationships between field and satellite data, allowing the modelling of unobservable heterogeneity. Panel data techniques for short time series are usually applied in economics ((Arellano, 2003; Baltagi, 2008; Engel & Reinecke, 2012; Gourieroux & Monfort, 1993; Hsiao, 2007; Mátyás & Sevestre, 2013) and others) but have not previously been used in remote sensing applications. Typically, remote sensing observations should be validated through in situ observations (the so-called “ground truth”) that are both time consuming and spatially scarce.
Since the late 1970s, remote sensing techniques have been proven to be suitable for characterizing and monitoring plants and crops. In particular, synthetic aperture radar (SAR) missions contribute considerably to this prediction effort. However, the main issue when using SAR image series together with field observations is the scarcity of data due to the difficulty of acquiring field measurements. This research aimed to contribute to solving this problem with an alternative statistical model that can overcome the lack of a long, robust series of field-based ground truth observations. The main novelty of this research is the evaluation of the potential of a panel data approach to radar remote sensing imagery for predicting crop biophysical variables. For this purpose, RADARSAT-2 imagery was acquired over the study area in central Spain. Simultaneously, a field campaign was deployed to estimate crop parameters in the same area and to validate the results of the modelling. The analysis of the influence of the crop type on the incidence angle and the polarimetric parameters showed a strong influence of the co-polar channels (HH, VV), the entropy (H) and the coherence between the co-polar channels (γ_HHVV), with the differences being higher at 25°. The panel data analysis method demonstrated that good predictions, with R² greater than 0.78, were achieved for all biophysical variables analysed in this study. Overall, this novel statistical approach with remote sensing data showed great applicability for the prediction of crop variables, even with a short series of observations.
Direct comparison of agent-based models of herding in financial markets
2016, Journal of Economic Dynamics and Control
Citation Excerpt :
This process is often complicated by the presence of nonlinearity and/or emergence of complexity from simple rules, which makes the inference of parameter values with traditional statistical methods difficult. Existing solutions to this problem rely on simulation methods such as simulated maximum likelihood (SML) or the method of simulated moments (MSM), both reviewed in Gouriéroux and Monfort (1993). More recently, Bianchi et al. (2007) advocate the use of the indirect inference approach of Gouriéroux and Monfort (1996), which generalises MSM by using a binding function rather than directly selecting the moments that need matching.
The present paper tests a new model comparison methodology by comparing multiple calibrations of three agent-based models of financial markets on the daily returns of 24 stock market indices and exchange rate series. The models chosen for this empirical application are the herding model of Gilli and Winker (2003), its asymmetric version by Alfarano et al. (2005) and the more recent model by Franke and Westerhoff (2011), which all share a common lineage to the herding model introduced by Kirman (1993). In addition, standard ARCH processes are included for each financial series to provide a benchmark for the explanatory power of the models. The methodology provides a consistent and statistically significant ranking of the three models. More importantly, it also reveals that the best performing model, Franke and Westerhoff, is generally not distinguishable from an ARCH-type process, suggesting their explanatory power on the data is similar.
On the estimation of zero-inefficiency stochastic frontier models with endogenous regressors
2016, Economics Letters
In this paper, we investigate endogeneity issues in the zero-inefficiency stochastic frontier (ZISF) models by mean of simultaneous equation setting. Specifically, we allow for one or more regressors to be correlated with the statistical noise. A modified limited information maximum likelihood (LIML) approach is used to estimate the parameters of the model. Moreover, the firm specific inefficiency score is also provided. Limited Monte Carlo simulations show that the proposed estimators perform well in finite sample.
Mixed logit with a flexible mixing distribution
2016, Journal of Choice Modelling
This paper presents a flexible procedure for representing the distribution of random parameters in mixed logit models. A logit formula is specified for the mixing distribution, in addition to its use for the choice probabilities. The properties of logit assure positivity and provide the normalizing constant for the mixing distribution. Any mixing distribution can be approximated to any degree of accuracy by this specification. The researcher defines variables to describe the shape of the mixing distribution, using flexible forms such as polynomials, splines, and step functions. The gradient of the log-likelihood is easy to calculate, which facilitates estimation. The procedure is illustrated with data on consumers' choice among video streaming services.
Endogeneity in stochastic frontier models
2016, Journal of Econometrics
Stochastic frontier models are typically estimated by maximum likelihood (MLE) or corrected ordinary least squares. The consistency of either estimator depends on exogeneity of the explanatory variables (inputs, in the production frontier setting). We will investigate the case that one or more of the inputs is endogenous, in the simultaneous equation sense of endogeneity. That is, we worry that there is correlation between the inputs and statistical noise or inefficiency.
In a standard regression setting, simultaneity is handled by a number of procedures that are numerically or asymptotically equivalent. These include 2SLS; using the residual from the reduced form equations for the endogenous variables as a control function; and MLE of the system that contains the equation of interest plus the unrestricted reduced form equations for the endogenous variables (LIML). We will consider modifications of these standard procedures for the stochastic frontier setting.
The paper is mostly a survey and combination of existing results from the stochastic frontier literature and the classic simultaneous equations literature, but it also contains some new results.

View all citing articles on Scopus

View full text

Simulation-based inference: A survey with special reference to panel data models

Abstract

Journal of Econometrics

Bayesian full information analysis of simultaneous equation models using integration by Monte Carlo

Pseudo maximum likelihood techniques in a simple rationing model of the Dutch labour market

Smooth unbiased multivariate probability simulators for maximum likelihood estimation of limited dependent variable models

Simulation based semi-parametric estimation and prediction in nonlinear systems

Residual-based procedure for prediction and estimation in nonlinear simultaneous systems

Econometrica

A computationally efficient quadrature procedure for one-factor multinomial probit model

Econometrica

Simulated moments estimation of Markov models of asset prices

Efficient simulation from the multivariate normal distribution subject to linear inequality contraints and the evaluation of constraint probabilities

Additive log-differenced probability models for count data

Simulation based econometrics in models with heterogeneity

Annales d'Economie et de Statistique no. 20/21

Pseudo maximum likelihood methods: Theory

Econometrica

Country heterogeneity and external debt crises: Estimation by the method of simulated moments

Simulation of multivariate normal rectangle probabilities: Methods and programs

Large sample properties of generalized method of moments estimators

Econometrica