ArticlePDF Available

Model Specification Tests and Artificial Regression

Authors:
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
... If the endogeneity is present, the instrumental variable estimator is concluded to be consistent while the ordinary least squares estimator is inconsistent. But this test should be used cautiously with small samples due to its asymptotic validity [37]. ...
Article
Full-text available
In this article, we introduce the R package dLagM for the implementation of distributed lag models and autoregressive distributed lag (ARDL) bounds testing to explore the short and long-run relationships between dependent and independent time series. Distributed lag models constitute a large class of time series regression models including the ARDL models used for cointegration analysis. The dLagM package provides a user-friendly and flexible environment for the implementation of the finite linear, polynomial, Koyck, and ARDL models and ARDL bounds cointegration test. Particularly, in this article, a new search algorithm to specify the orders of ARDL bounds testing is proposed and implemented by the dLagM package. Main features and input/output structures of the dLagM package and use of the proposed algorithm are illustrated over the datasets included in the package. Features of dLagM package are benchmarked with some mainstream software used to implement distributed lag models and ARDLs.
... Since there are arguments for the use of both fixed and random effect models (Allison, 2009), the usual modus operandi is to consult the Hausman test to choose between the two. However, since the test does not perform well under heteroscedasticity (presence of which was confirmed by the Wald test in Tab.1), we opted to run an auxiliary regression with means of explanatory variables included in the model ( Davidson and MacKinnon, 1990;Wooldridge, 2002). Joint test that coefficients of time-averaged explanatory variables are simultaneously zero, provides proof that fixed effects estimator should be used to provide unbiased coefficient estimates. ...
Article
Full-text available
In an increasingly urbanized world, the scarcity of space is a growing problem along with land consumption and soil sealing. To achieve sustainable development and sustainable land use, society has to resolve conflicts between residential, industrial, transport, commercial and green areas while creating a balance between social, economic and ecological targets. However, coordination of sustainable land use is a challenge for policymakers. The paper examines whether the withdrawal of land from the agricultural land fund leads to development, measured both by the increase in domestic entrepreneurial activity, as well as by the increase in foreign direct investments. The results are based on the analysis of panel data on the amount of land withdrawal, newly established firms and inward flow of FDI covering 41 administrative districts of Slovak Republic over 9 years (6 years in case of the FDI, due to the availability of data). Additionally, the spatial Durbin panel model was used to examine, whether land withdrawal and its non-agricultural use generate positive spillover effects on surrounding regions in terms of increased entrepreneurial activity and flow of FDI.
Article
Corrections based on the score test for neglected heterogeneity do not appear to have been used previously in the presence of time-varying covariates. This paper uses Monte Carlo analysis to evaluate, for a grouped hazard framework constructed for an empirical application, the efficacy of such corrections.
Chapter
Full-text available
An artificial regression is a linear regression that is associated with some other econometric model, which is usually nonlinear. It can be used for a variety of purposes, in particular computing covariance matrices and calculating test statistics. The best-known artificial regression is the Gauss–Newton regression, whose key properties are shared by all artificial regressions. The chief advantage of artificial regressions is conceptual: because econometricians are very familiar with linear regression models, using them for computation reduces the chance of errors and makes the results easier to comprehend intuitively.
Chapter
Econometrics is an area of applied statistics that has developed with a very strong individual flavor although its techniques are also widely used in such disciplines as biometrics, psychometrics, and sociometrics, and, to be somewhat polemical, are applicable to a far wider statistical audience than seems aware of their need.
Chapter
The problem of statistical model selection in econometrics and statistics is reviewed. Model selection is interpreted as a decision problem through which a statistical model is selected in order to perform statistical analysis, such as estimation, testing, confidence set construction, forecasting, simulation, policy analysis, and so on. Broad approaches to model selection are described: (1) hypothesis testing procedures, including specification and diagnostic tests; (2) penalized goodness-of-fit methods, such as information criteria; (3) Bayesian approaches; (4) forecast evaluation methods. The effect of model selection on subsequent statistical inference is also discussed.
Article
Full-text available
In diagnostic testing,model specification tests play an important role. The specification of a good model is an art. Over specification yields unbiased estimates of regression coefficients,but larger variance; under specification yields biased estimates of regression coefficients and understates the variance of these estimates. In case of the left out variables or because of inclusion of variable not specified by the truth; the case of irrelevant variables. Mis-specification is usually interpreted as a case of omitted variables,and many researchers,concerned only with the bias resulting from it,the specification bias. Researchers seldom pay attention to the other aspects of mis-specification of the model. In view of the importance of these aspects of misspecification in empirical research,some major results of misspecification error tests are considered in this Article. In this research article,three types of new tests for the mis-specification of the linear regression model have been developed by using various types of residuals. © 2016,International Journal of Pharmacy and Technology. All rights reserved.
Article
This paper develops and compares non-nested hypothesis tests for linear regression models with first-order serially correlated errors. It extends the non-nested testing procedures of Pesaran, Fisher and McAleer, and Davidson and MacKinnon, and compares their performance on four conventional models of aggregate investment demand using quarterly U.S. investment data from 1951:I to 1983:IV. The data and the non-nested hypothesis tests initially indicate that none of the models are correctly specified. The tests are also intransitive in their assessments. Before rejecting these conventional models of investment demand, we investigate the small-sample properties of these different non-nested test procedures through a series of Monte Carlo studies. We find that these non-nested tests can have significant finite-sample size and power biases.
Article
This paper surveys the growing literature on diagnostic testing of models based on unit record data. We argue that while many of these tests are produced in a Lagrange multiplier framework they are often more readily derived, and more easily applied, if approached from the conditional moment testing view of Newey (1985) and Tauchen (1985). In addition we propose some new tests based on comparisons of parametric estimators with nonparametric estimators which are consistent under certain forms of misspecification. To illustrate the utility of the tests we employ them in the examination of some existing published studies. Copyright 1989 by John Wiley & Sons, Ltd.