The package implements the Panel Smooth Transition Regression (PSTR) modelling.
Yukai Yang
Department of Statistics, Uppsala University
NewPSTR
initialize the modelling by creating an object of the class PSTR.
LinTest
implements the linearity tests.
WCB_LinTest
implements the wild bootstrap (WB) and the wild cluster bootstrap (WCB) linearity tests.
EstPSTR
implements the estimation of the PSTR model.
EvalTest
implements the evaluation tests.
WCB_TVTest
implements the wild bootstrap (WB) and the wild cluster bootstrap (WCB) evaluation test of no time-varying parameters.
WCB_HETest
implements the wild bootstrap (WB) and the wild cluster bootstrap (WCB) evaluation test of no remaining nonlinearity (no remaining heterogeneity).
version
shows the version number and some information of the package.
print.PSTR
prints the object of the class PSTR.
plot_transition
plots the transition function of an estimated PSTR model.
plot_response
plots curve or surfaces of the expected reponse agaist the corresponding variable.
plot_target
plots the surface of the target function for the nonlinear least square estimation.
The modelling procedure consists of three stages: Specification, Estimation and Evaluation. The package offers tools helping the package users to conduct model specification tests, to do PSTR model estimation, and to do model evaluation.
The cluster-dependency and heteroskedasticity-consistent tests are implemented in the package.
The wild bootstrap and cluster wild bootstrap tests are also implemented.
Parallel computation (as an option) is implemented in some functions, especially the bootstrap tests. Therefore, the package suits tasks running many cores on super-computation servers.
The Panel Smooth Transition Regression (PSTR) model is defined to be $$y_{it} = \mu_i + \beta_0' x_{it} + \beta_1' z_{it} g_{it} + u_{it}$$ where \(g_{it}\) is the transition function taking the logistic form with the transition variable for individual \(i\), \(x_{it}\) contains the explanatory variables in the linear part, and \(z_{it}\) contains the explanatory variables in the nonlinear part, and they can be different.
The transition function \(g_{it}\) takes the logistic form $$g(q_{it} ; \gamma, c) = \left( 1 + \exp \left( - \gamma \prod_{j=1}^{m} (q_{it} - c_j) \right) \right)^{-1}$$ with \(\gamma > 0\) and \(c_1 < c_2 < ... < c_m\). \(\gamma\) can be reparametrized as \(\gamma = \exp{\delta}\) where \(\delta\) is a real number.
Gonz<U+00E1>lez, A., Ter<U+00E4>svirta, T., van Dijk, D. and Yang, Y. (2005) "Panel Smooth Transition Regression Models", SSE/EFI Working Paper Series in Economics and Finance 604, Stockholm School of Economics, revised 11 Oct 2017.