The assumption of linearity is tested using five statistical tests for the US and the Canadian unemployment rates. An AR(p) model was used to remove any linear structure from the series. Strong evidence in favour of non-linearity was found in the case of Canada. The result for the US is not so...

We study the effects of ARCH errors on the performance of the commonly used lag length selection criteria. The most important finding of this study is that SIC, FPE, HQC and BIC perform considerably well in estimating the true autoregressive lag length, even in the presence of ARCH errors. Thus,...

It is well known that, in a multinomial probit, only the covariance matrix of the location and scale normalized utilities are identified. In this note, we explore the relation between these identifiable parameters and the original elements of the covariance matrix, to find out what can be learnt...

When testing for a change in mean of a time series, the null hypothesis is no change in mean. However, a change in mean causes a bias in the estimation of serial correlation parameters. This bias can cause nonmonotonic power to the point that if the change is big enough, power can go to zero. In...

In this paper we test the stationarity properties of the consumption-income ratio for a sample of 14 European Union countries over the period 1960-1999 utilizing recent advances in panel unit root and asymmetric unit root tests. We find that a failure to take account of asymmetries, would imply...

Most of the existing autoregressive models presume that the observations are perfectly measured. In empirical studies, the variable of interest is unavoidably measured with various kinds of errors. Thus, misleading conclusions may be yielded due to the inconsistency of the parameter estimates...

Estimating the lag length of autoregressive process for a time series is a crucial econometric exercise in most economic studies. This study attempts to provide helpfully guidelines regarding the use of lag length selection criteria in determining the autoregressive lag length. The most...

This paper studies the estimation of time series regression when both regressors and disturbances have long memory. In contrast with the frequency domain estimation as in Robinson and Hidalgo (1997), we propose to estimate the same regression model with discrete wavelet transform (DWT) of the...

This paper deals with the issue of the Permanent Income Hypothesis (PIH) and we show that consumption and income may be fractionally cointegrated. We use a semiparametric frequency domain procedure of Robinson (1995a), and the results show that the UK and the Japanese consumption and income are...

In this note, we derive the finite sample bias of the modified ordinary least squares (MOLS) estimator, which was suggested by Wansbeek and Knaap (1999) and reconsidered by Hayakawa (2006a,b). From the formula for the finite sample bias, we find that the bias of the MOLS estimator becomes small...