Pentti Saikkonen

Asymptotically Efficient Estimation of Cointegration Regressions

An asymptotic optimality theory for the estimation of cointegration regressions is developed in this paper. The theory applies to a reasonably wide class of estimators without making any specific assumptions about the probability distribution or short-run dynamics of the data-generating process. Due to the nonstandard nature of the estimation problem, the conventional minimum variance criterion does not provide a convenient measure of asymptotic efficiency. An alternative criterion, based on the concentration or peakedness of the limiting distribution of an estimator, is therefore adopted. The limiting distribution of estimators with maximum asymptotic efficiency is characterized in the paper and used to discuss the optimality of some known estimators. A new asymptotically efficient estimator is also introduced. This estimator is obtained from the ordinary least-squares estimator by a time domain correction which is nonparametric in the sense that no assumption of a finite parameter model is required. The estimator can be computed with least squares without any initial estimations.

Comparison of unit root tests for time series with level shifts

Unit root tests are considered for time series which have a level shift at a known point in time. The shift can have a very general nonlinear form, and additional deterministic mean and trend terms are allowed for. Prior to the tests, the deterministic parts and other nuisance parameters of the data generation process are estimated in a first step. Then, the series are adjusted for these terms and unit root tests of the Dickey–Fuller type are applied to the adjusted series. The properties of previously suggested tests of this sort are analysed and modifications are proposed which take into account estimation errors in the nuisance parameters. An important result is that estimation under the null hypothesis is preferable to estimation under local alternatives. This contrasts with results obtained by other authors for time series without level shifts.

TESTING FOR A UNIT ROOT IN A TIME SERIES WITH A LEVEL SHIFT AT UNKNOWN TIME

Unit root tests for time series with level shifts of general form are considered when the timing of the shift is unknown. It is proposed to estimate the nuisance parameters of the data generation process including the shift date in a first step and apply standard unit root tests to the residuals. The estimation of the nuisance parameters is done in such a way that the unit root tests on the residuals have limiting distributions for which critical values are tabulated elsewhere in the literature. Empirical examples are discussed to illustrate the procedure.

Testing for the Cointegrating Rank of a VAR Process with Structural Shifts

Tests for the cointegrating rank of a vector autoregressive process are considered that allow for possible exogenous shifts in the mean of the data-generation process. The break points are assumed to be known a priori. It is proposed to estimate and remove the deterministic terms such as mean, linear-trend term, and a shift in a first step. Then systems cointegration tests are applied to the adjusted series. The resulting tests are shown to have known limiting null distributions that are free of nuisance parameters and do not depend on the break point. The tests are applied for analyzing the number of cointegrating relations in two German money-demand systems.

Estimation and Testing of Cointegrated Systems by an Autoregressive Approximation

This paper studies the estimation and testing of general cointegrated systems by using an autoregressive approximation. Simple estimators for both the cointegration vectors and their weight matrix in the autoregressive error correction model representation of the system are developed. Since these estimators assume that the number of cointegration vectors and their normalization are fixed in advance, convenient specification tests for checking the validity of these assumptions are also provided. The asymptotic distributions of the estimators and test statistics are derived by assuming that the order of the auto-regressive approximation increases with the sample size at a suitable rate. This generalizes some previous results derived for finite-order autoregressions as no assumption of a finite-parameter data-generating process is imposed. The estimators and tests of the paper are interpreted in terms of autoregressive spectral density estimators at the zero frequency and, in the special case of a finite-order Gaussian autoregression, their relation to maximum likelihood procedures is discussed. All estimators of the paper can be applied with simple least-squares techniques and used to construct conventional Wald tests with asymptotic chi-square distributions under the null hypothesis. The limit theory of the specification tests is nonstandard, similar to that in univariate unit root tests.

Predicting U.S. Recessions with Dynamic Binary Response Models

We develop dynamic binary probit models and apply them for predicting U.S. recessions using the interest rate spread as the driving predictor. The new models use lags of the binary response (a recession dummy) to forecast its future values and allow for the potential forecast power of lags of the underlying conditional probability. We show how multiperiod-ahead forecasts are computed iteratively using the same one-period-ahead model. Iterated forecasts that apply specific lags supported by statistical model selection procedures turn out to be more accurate than previously used direct forecasts based on horizon-specific model specifications.

TESTING FOR THE COINTEGRATING RANK OF A VAR PROCESS WITH AN INTERCEPT

Testing the cointegrating rank of a vector autoregressive process with an intercept is considered. In addition to the likelihood ratio (LR) tests developed by Johansen and Juselius and others we also consider an alternative class of tests which is based on estimating the trend parameters of the deterministic term in a different way. The asymptotic local power of these tests is derived and compared to that of the corresponding LR tests. The small sample properties are investigated by simulations. The new tests are seen to be substantially more powerful than conventional LR tests.

Maximum eigenvalue versus trace tests for the cointegrating rank of a VAR process

The properties of a range of maximum eigenvalue and trace tests for the cointegrating rank of a vector autoregressive process are compared. The tests are alilikelihood ratio type tests and operate under different assumptions regarding the deterministic part of the data generation process. The asymptotic distributions under local alternatives are given and the local power is derived. It is found that the local power of corresponding maximum eigenvalue and trace tests is very similar. A Monte Carlo comparison shows, however, that there may be slight differences in small sampies. The trace tests tend to have more distorted sizes whereas their power is in some situations superior to that of the maximum eigenvalue tests.

Test Procedures for Unit Roots in Time Series with Level Shifts at Unknown Time

Two types of unit root tests which accommodate a structural level shift at a known point in time are extended to the situation where the break date is unknown. It is shown that for any estimator for the break date the tests have the same asymptotic distribution as the corresponding tests under the known break date assumption. Different estimators of the break date are compared in a Monte Carlo experiment and a recommendation for choosing the break date in small samples is given. It is also shown that ignoring the fact that a break has occurred and applying a standard unit root test may lead to substantial size distortion and total loss of power. Example series from the Nelson-Plosser data set are used to illustrate the performance of our tests.

Trend Adjustment Prior to Testing for the Cointegrating Rank of a Vector Autoregressive Process

Testing the cointegrating rank of a vector autoregressive process that may have a deterministic linear trend is considered. Previous proposals for dealing with such a situation are either to allow for a deterministic trend term in computing a suitable test statistic or else to remove the linear trend first and then derive the test statistic from the trend‐adjusted data. In this study the latter approach is considered and a new method for trend removal is proposed that is based on estimating the trend parameters under the null hypothesis. Likelihood ratio and Lagrange multiplier type test statistics are derived on the basis of the trend‐adjusted data and their asymptotic distributions are considered under the null hypothesis and under local alternatives. A simulation comparison with other proposals is performed.