An alternative to unit root tests: Bridge estimators differentiate between nonstationary versus stationary models and select optimal lag

This paper introduces a novel way of differentiating a unit root from stationary alternatives using so-called “Bridge” estimators; this estimation procedure can potentially generate exact zero estimates of parameters. We exploit this property and treat this as a model selection problem. We show that Bridge estimators can select the correct model with probability tending to 1. They estimate “zero” parameter on the lagged dependent variable as zero (nonstationarity), if this is nonzero (stationary), estimate the coefficient with standard normal limit. In this sense, we extend the statistics literature as well, since that literature only deals with model selection among only stationary variables.     

Phillips-Perron-type unit root tests in the nonlinear ESTAR framework

Summary In this paper, we propose Phillips-Perron type, semi-parametric testing procedures to distinguish a unit root process from a mean-reverting exponential smooth transition autoregressive one. The limiting nonstandard distributions are derived under very general conditions and simulation evidence shows that the tests perform better than the standard Phillips-Perron or Dickey-Fuller tests in the region of the null. We would like to thank conference participants of the Pfingsttagung 2005 in Münster for their helpful comments.     

On the asymptotic distribution of a simple unit root test for trending and breaking series

In this paper, we derive the asymptotic distribution of Popp's (2008) innovational outlier unit root test for trending series with a break. The results of Zivot and Andrews (1992) are applied to provide the limiting results of these new test statistics. We tabulate their asymptotic and finite sample critical values, and illustrate the use of the new statistics with an application to the unemployment rate series for 23 OECD countries.     

A simple unit root testing methodology that does not require knowledge regarding the presence of a break

We develop a simple methodology that allows practitioners to test for the presence of a unit root without a priori knowledge regarding the occurrence of a break under the null hypothesis. We use a pre-test that is readily available in the estimated regression used to calculate the unit root statistics, and so our methodology is very easy to implement. The t-statistic corresponding to the impulse dummy variables evaluated at break date estimator is used as a pre-test to ascertain whether a break exists under the null hypothesis. Finite sample simulations show that our methodology yields tests that maintain their size.     

The accuracy of normal approximation in a heterogeneous panel data unit root test

Tests for unit roots in panel data have become very popular. Two attractive features of panel data unit root tests are the increased power compared to time-series tests, and the often well-behaved limiting distributions of the tests. In this paper we apply Monte Carlo simulations to investigate how well the normal approximation works for a heterogeneous panel data unit root test when there are only a few cross sections in the sample. We find that the normal approximation, which should be valid for large numbers of cross-sectional units, works well, at conventional significance levels, even when the number of cross sections is as small as two. This finding is valuable for the applied researcher since critical values will be easy to obtain and p-values will be readily available.     

A new estimator for the unit root

We compare the ordinary least squares, weighted symmetric, modified weighted symmetric (MWS), maximum likelihood, and our new modification for least squares (MLS) estimator for first-order autoregressive in the case of unit root using Monte Carlo method. The Monte Carlo study sheds some light on how well the estimators and the predictors perform on different samples sizes. We found that MLS estimator is less biased and has less mean squared error (MSE) than any other estimators, and MWS predictor error performs well, in the sense of MSE, than any other predictors’ methods. The sample percentiles for the distribution of the τ statistic for the first, second, and third periods in the future, for alternative estimators, are reported to know if it agrees with those of normal distribution or not.     

Testing stability in a spatial unilateral autoregressive model

ABSTRACT

Least squares estimator of the stability parameter ? ? |α| + |β| for a spatial unilateral autoregressive process X_{k, ?} = αX_{k ? 1, ?} + βX_{k, ? ? 1} + ?_{k, ?} is investigated and asymptotic normality with a scaling factor n^5/4 is shown in the unstable case ? = 1. The result is in contrast to the unit root case of the AR(p) model X_k = α₁X_{k ? 1} + ??? + α_pX_{k ? p} + ?_k, where the limiting distribution of the least squares estimator of the unit root parameter ? ? α₁ + ??? + α_p is not normal.     

Bounded integrated processes and unit root tests

In the framework of integrated processes, the problem of testing the presence of unknown boundaries which constrain the process to move within a closed interval is considered. To analyze this problem, the concept of bounded integrated process is introduced, thus allowing to formally define boundary conditions for I(1) processes. A new class of tests, which are based on the rescaled range of the process, is introduced in order to test the null hypothesis of no boundary conditions. The limit distribution of the test statistics involved can be expressed in terms of the distribution of the range of Brownian functionals, while the power properties are obtained by deriving some asymptotic results for I(1) processes with boundary conditions. Both theoretical and simulation investigations show that range-based tests outperform standard unit root tests significantly when used to detect the presence of boundary conditions. A previous draft of the paper (Cavaliere, 2000) was presented at the 8th World Congress of the Econometric Society, Seattle, 11–16 August 2000. I wish sincerely to thank: Martin Jacobsen for his patience in discussing weak convergence to regulated Brownian motions and his valuable suggestions; the Department of Theoretical Statistics of the University of Copenhagen whose hospitality is gratefully acknowledged; Tommaso Proietti for important suggestions; Silvano Bordignon and partecipants at the CIdE seminar, University of Padua, June 2000; two anonymous referees. Partial financial support from 60% M.U.R.S.T. research grants is acknowledged.     

Robust unit root tests with autoregressive errors

ABSTRACT

This article presents a new test for unit roots based on least absolute deviation estimation specially designed to work for time series with autoregressive errors. The methodology used is a bootstrap scheme based on estimating a model and then the innovations. The resampling part is performed under the null hypothesis and, as it is customary in bootstrap procedures, is automatic and does not rely on the calculation of any nuisance parameter. The validity of the procedure is established and the asymptotic distribution of the statistic proposed is proved to converge to the correct distribution. To analyze the performance of the test for finite samples, a Monte Carlo study is conducted showing a very good behavior in many different situations.     

An efficiency Bayesian unit root test in Unobserved-ARCH models

This paper investigates the new prior distribution on the Unobserved-Autoregressive Conditional Heteroscedasticity (ARCH) unit root test. Monte Carlo simulations show that the sample size is seriously effective in efficiency of Bayesian test. To improve the performance of Bayesian test for unit root, we propose a new Bayesian test that is robust in the presence of stationary and nonstationary Unobserved-ARCH. The finite sample property of the proposed test statistic is evaluated using Monte Carlo studies. Applying the developed method, we test the policy of daily exchange rate of the German Marc with respect to the Greek Drachma.     

A comparison of alternative unit root tests

In this paper we evaluate the performance of three methods for testing the existence of a unit root in a time series, when the models under consideration in the null hypothesis do not display autocorrelation in the error term. In such cases, simple versions of the Dickey-Fuller test should be used as the most appropriate ones instead of the known augmented Dickey-Fuller or Phillips-Perron tests. Through Monte Carlo simulations we show that, apart from a few cases, testing the existence of a unit root we obtain actual type I error and power very close to their nominal levels. Additionally, when the random walk null hypothesis is true, by gradually increasing the sample size, we observe that p-values for the drift in the unrestricted model fluctuate at low levels with small variance and the Durbin-Watson (DW) statistic is approaching 2 in both the unrestricted and restricted models. If, however, the null hypothesis of a random walk is false, taking a larger sample, the DW statistic in the restricted model starts to deviate from 2 while in the unrestricted model it continues to approach 2. It is also shown that the probability not to reject that the errors are uncorrelated, when they are indeed not correlated, is higher when the DW test is applied at 1% nominal level of significance.     

A new unit root test against ESTAR based on a class of modified statistics

This paper proposes a new unit root test against a nonlinear exponential smooth transition autoregressive model. This model receives much attention in international macroeconomics as it has been successfully applied to a variety of financial time series. The new test is build upon the nonstandard testing approach of Abadir and Distaso (J Econom 140:695–718, 2007) who introduce a class of modified statistics for testing joint hypotheses when one of the alternatives is one-sided. The asymptotic properties of the suggested unit root test are derived. In a Monte Carlo study the popular Dickey–Fuller-type test proposed by Kapetanios et al. (J Econom 112:359–379, 2003) is compared to the new test. The results suggest that the new test is generally superior in terms of power. An application to a real effective exchange rate underlines its usefulness.     

Testing for a unit root in a nonlinear quantile autoregression framework

The nonlinear unit root test of Kapetanios, Shin, and Snell (2003 Kapetanios, G., Shin, Y., Snell, A. (2003). Testing for a unit root in the nonlinear STAR framework. Journal of Econometrics 112:359–379.[Crossref], [Web of Science ®] , [Google Scholar]) (KSS) has attracted much recent attention. However, the KSS test relies on the ordinary least squares (OLS) estimator, which is not robust to a heavy-tailed distribution and, in practice, the test suffers from a large power loss. This study develops three kinds of quantile nonlinear unit root tests: the quantile t-ratio test; the quantile Kolmogorov–Smirnov test; and the quantile Cramer–von Mises test. A Monte Carlo simulation shows that these tests have significantly better power when an innovation follows a non-normal distribution. In addition, the quantile t-ratio test can reveal the heterogeneity of the asymmetric dynamics in a time series. In our empirical studies, we investigate the unit root properties of U.S. macroeconomic time series and the real effective exchange rates for 61 countries. The results show that our proposed tests reject the unit roots more often, indicating that the series are likely to be asymmetric nonlinear reverting processes.     

New innovational outlier unit root test with a break at an unknown time

The Perron test which is based on a Dickey–Fuller test regression is a commonly employed approach to test for a unit root in the presence of a structural break of unknown timing. In the case of an innovational outlier (IO), the Perron test tends to exhibit spurious rejections in finite samples when the break occurs under the null hypothesis. In the present paper, a new Perron-type IO unit root test is developed. It is shown in Monte Carlo experiments that the new test does not over-reject the null hypothesis. Even for the case of a level and slope break for trending data, the empirical size is near its nominal level. The test distribution equals the case of a known break date. Furthermore, the test is able to identify the true break date very accurately even for small breaks. As an application serves the Nelson–Plosser data set.     

Ratio tests of a unit root

Based on the maximal invariant principle, we derive two ratio tests (locally best invariant test and point optimal test) for a unit root and compare them with previously proposed ratio tests. We also show that our ratio tests tend to have better powers than the Dickey-Fuller test and the modified Dickey-Fuller test.     

Devaluation expectations and the unit root hypothesis: The Italian Lira in the European monetary system

Summary In several studies the unit root hypothesis of EMS exchange rates is analysed within the context of devaluation expectations estimation. By means of bootstrap inference it is shown that these procedures are not compatible with standard Dickey-Fuller significance levels and may lead to a wrong rejection of the null hypothesis. In the case of the Italian Lira/Deutsche Mark exchange rate, the hypothesis of a unit root is not rejected and expectations can be modelled by means of a reflected Brownian motion. The estimated devaluation expectations are related with some macro variables which provide evidence for the structure of expectations. This research has been partially supported with 40% and 60% MURST grants. The author wishes to thank the Bank of Italy for the exchange rates and the interest rates data and Ulf S?derstr?m for providing macroeconomic indicators. Useful suggestions from Riccardo Cesari, Michele Costa and two anonymous referees are gratefully acknowledged.     

Testing for Unit Root Against Stationarity Using the Likelihood Ratio Test

In a first-order autoregressive model with drift, we derive the likelihood ratio test for a unit root against the stationary alternative. We also derive the test in a state space model with trend. Finite sample and asymptotic critical values are obtained by Monte Carlo simulations. A simulation study investigates the power performance of the likelihood ratio test and we also examine how a bias correction of the test affects the results.