Testing for stationarity in heterogeneous panel data where the time dimension is finite

Abstract

and hence makes the test valid for any (T, N) combination. The asymptotic distributions of the tests are derived under the null and are shown to be normally distributed. Their moments for T fixed are derived analytically using Ghazal's (1994, Statistics and Probability letters 20, 313--319) 1 lemma 1. Finite sample size and power are considered in a Monte Carlo experiment. The proposed tests have empirical sizes that are very close to the nominal 5% level. The Monte Carlo results clearly show that the power of the test statistics increases substantially with N, T and ω (ω being the number of unit root processes under the alternative). The results indicate that the assumption that T is asymptotic rather than fixed leads to tests that are substantially oversized particularly for relatively short panels with large N. Copyright 2005 Royal Economic Society

Modelling the dynamics of market shares in a pooled data setting: econometric and empirical issues

“…The second type of tests allows for panel heterogeneity. The Im-Pesaran-Shin test (IPS test, 2003), the Fisher-type tests using ADF, and the Phillips-Perron test (Maddala and Wu, 1999;Hadri (2000), Choi (2001), Hadri and Larsson (2005)) are based on individual unit roots or coefficients i for each cross-section. 7 For simplicity, neither constants nor trends are included in the equations.…”

Time-specific disturbances in a panel stationarity test

“…The moments of Hadri and Larsson (2005) can be used to obtain a panel data test statistic that is standard normally distributed as N → ∞ for fixed T , while the moments of Jönsson (2006a) can be used to obtain an approximate standard normal distribution for the test statistic when T and N are fixed. By utilizing results from Monte Carlo simulations, Hadri and Larsson (2005) and Jönsson (2006a) show that inference can be improved by using the fixed-T moments, instead of the asymptotic moments, of the KPSS statistic.…”

“…In order to account for a finite time-series dimension, Hadri and Larsson (2005) provide finite-T moments for the KPSS statistic under the assumption that the time series under consideration are serially independent, while approximate moments for the KPSS statistic with serially dependent time series are supplied by Jönsson (2006a). The moments of Hadri and Larsson (2005) can be used to obtain a panel data test statistic that is standard normally distributed as N → ∞ for fixed T , while the moments of Jönsson (2006a) can be used to obtain an approximate standard normal distribution for the test statistic when T and N are fixed.…”

“…This was recognized e.g. by Quah (1994); Lin (1992, 1993); Levin et al (2002); Im et al (1997Im et al ( , 2003, who suggest various panel data unit root tests; by Hadri (2000), Hadri and Larsson (2005), Harris et al (2005) and Shin and Snell (2006), who suggest panel data stationarity tests; and by Pedroni (1999), Pedroni (2004), McCoskey and Kao (1998), Westerlund (2005), Larsson et al (2001) and Groen and Kleibergen (2003), who suggest panel data cointegration tests.…”

Reasearch on Regional Differences Between China's Economic Growth and Export Trade Based on the Analysis of a Random Effects Model

“…Data Science Journal, Volume 6, Supplement, 9 June 2007 (Hadri, & Larsson, 2005;Jung, Shin, & Oh, 2005;Hausman, 1978). The test formula is as follows:…”