Testing for the Null Hypothesis of Cointegration with a Structural Break

In this paper we propose residual-based tests for the null hypothesis of cointegration with a structural break against the alternative of no cointegration. The Lagrange Multiplier (LM) test is proposed and its limiting distribution is obtained for the case in which the timing of a structural break is known. Then the test statistic is extended to deal with a structural break of unknown timing. The test statistic, a plug-in version of the test statistic for known timing, replaces the true break point by the estimated one. We show the limiting properties of the test statistic under the null as well as the alternative. Critical values are calculated for the tests by simulation methods. Finite-sample simulations show that the empirical size of the test is close to the nominal one unless the regression error is very persistent and that the test rejects the null when no cointegrating relationship with a structural break is present. We provide empirical examples based on the present-value model, the term structure model, and the money-output relationship model.     

Testing for the Null Hypothesis of Cointegration with a Structural Break

In this paper we propose residual-based tests for the null hypothesis of cointegration with a structural break against the alternative of no cointegration. The Lagrange Multiplier (LM) test is proposed and its limiting distribution is obtained for the case in which the timing of a structural break is known. Then the test statistic is extended to deal with a structural break of unknown timing. The test statistic, a plug-in version of the test statistic for known timing, replaces the true break point by the estimated one. We show the limiting properties of the test statistic under the null as well as the alternative. Critical values are calculated for the tests by simulation methods. Finite-sample simulations show that the empirical size of the test is close to the nominal one unless the regression error is very persistent and that the test rejects the null when no cointegrating relationship with a structural break is present. We provide empirical examples based on the present-value model, the term structure model, and the money-output relationship model.     

Confidence Sets for Cointegrating Coefficients Based on Stationarity Tests

Standard methods for inference in cointegrating systems require all the variables to have exact unit roots and are not at all robust even to slight violations of this condition. In this article, I consider an alternative approach to inference in a cointegrating system. This involves testing the hypothesis that a cointegrating vector takes on a specified value by testing for the stationarity of the associated residual. Confidence sets for the cointegrating vector can be constructed by exploiting the equivalence between tests and confidence sets. This method has the advantage that it remains valid even if the regressors have roots that are not exactly equal to unity.     

Comment on “recent developments in bootstrapping time series”

The paper develops a general framework for identification, estimation, and hypothesis testing in cointegrated systems when the cointegrating coefficients are subject to (possibly) non-linear and cross-equation restrictions, obtained from economic theory or other relevant a priori information. It provides a proof of the consistency of the quasi maximum likelihood estimators (QMLE), establishes the relative rates of convergence of the QMLE of the short-run and the long-run parameters, and derives their asymptotic distributions; thus generalizing the results already available in the literature for the linear case. The paper also develops tests of the over-identifying (possibly) non-linear restrictions on the cointegrating vectors. The estimation and hypothesis testing procedures are applied to an Almost Ideal Demand System estimated on U.K. quarterly observations. Unlike many other studies of consumer demand this application does not treat relative prices and real per capita expenditures as exogenously given.     

LONG-RUN STRUCTURAL MODELLING

The paper develops a general framework for identification, estimation, and hypothesis testing in cointegrated systems when the cointegrating coefficients are subject to (possibly) non-linear and cross-equation restrictions, obtained from economic theory or other relevant a priori information. It provides a proof of the consistency of the quasi maximum likelihood estimators (QMLE), establishes the relative rates of convergence of the QMLE of the short-run and the long-run parameters, and derives their asymptotic distributions; thus generalizing the results already available in the literature for the linear case. The paper also develops tests of the over-identifying (possibly) non-linear restrictions on the cointegrating vectors. The estimation and hypothesis testing procedures are applied to an Almost Ideal Demand System estimated on U.K. quarterly observations. Unlike many other studies of consumer demand this application does not treat relative prices and real per capita expenditures as exogenously given.     

Small-Sample Improved Seasonal Unit Root Tests for Trending and Breaking Series

In this article three unit root tests that allow for a break in both the seasonal mean and linear trend of the data are proposed. The tests, which can be seen as small-sample corrected versions of already known asymptotic tests, are shown to perform very well in simulations, and much better than their asymptotic counterparts.     

The effect of non-normal disturbances and conditional heteroskedasticity on multiple cointegration tests

!t is well-known that Johansen's multiple cointegration tests' results and those of Johansen and Juselius' tests for restricrions on cointegrating vectors and their weights have far-reaching implications for economic modelling and analysis. Therefore, it is important to ensure that the tests have desirable finite sample properties. Although the statistics are derived under Gaussian distribution,the asympotic results are derived under a much wider class of distributions. Using simulation, this paper investigates the effect of non-normal disturbances on these tests in finite samples. Further, ARCH/GARCH type conditional heteroskedasticity is present in many economic and financial time series. This paper examines the finite properties of the tests when the error term follows ARCH/GARCH type processes. From the evidence, it appears that researchers should not be overly concerned by the possibility of small departures from non-normality when using Johansen's suggested techniques even in finite samples. ARCH and GARCH effects may be more problematic, however. In particular it becomes more important ro test whether the restriction implicit in the integrated (or near-integrated) ARCH-type Drocess actually holds in time series for the application of the cointegraiion rank tests and the test for restrictions on cointegrating weights. The tests for restrictions on cointegrating vectors apper to be robust for non-normal errors and for all ARCH and GARCH type processes considered.     

Unit root econometrics and economic nonlinearities

This paper concentrates on some shortcomings of contemporary unit root econometric methodology (testing for cointegration, common roots and stationarity) where the dynamics of an economy are described by a nonlinear process. It is shown that, in such circumstances, traditionally applied unit root econometrics may not lead to interpretable or statistically significant results. Two cases of such nonlinearities are discussed: (i) a stochastically nonlinear data generating process and (ii) a time-varying parameters cointegrating relation, typical of an economic reform process. It is shown that case (i) consists of a wide family of economic processes and in most such cases the results of standard unit root tests are not directly interpretable. Case (ii) does not result in a (conventionally understood) error-correction representation of a cointegrated process. Some Monte Carlo experiments evaluate the validity of cointegration tests in situations where there is a change in the cointegration parameter and from cointegration regime to noncointegration and vice versa. A simple method of estimation through simulation is proposed and its finite-sample properties examined.     

Structural Instability and the Production-Smoothing Model of Inventories

The production-smoothing model of inventories implies that inventories, labor inputs, sales, and factor input prices are cointegrated if sales and factor prices are I(1) with one cointegrating vector for each state variable held. These propositions are tested in six nondurable-goods industries. All industries provide evidence of cointegration. Fewer quasi-fixed factors are found than previous research often assumed. Estimates of cointegrating vectors provide implausible parameter estimates. Rank stability tests, with fixed or seqentially chosen breakpoints, indicate that the cointegrating matrix has unstable rank. Parameter estimates of cointegrating vectors do not provide much support for the production-smoothing model of inventories.     

A Bayesian Approach to Modeling Time-Varying Cointegration and Cointegrating Rank

A multivariate model that allows for both a time-varying cointegrating matrix and time-varying cointegrating rank is presented. The model addresses the issue that, in real data, the validity of a constant cointegrating relationship may be questionable. The model nests the submodels implied by alternative cointegrating matrix ranks and allows for transitions between stationarity and nonstationarity, and cointegrating and noncointegrating relationships in accordance with the observed behavior of the data. A Bayesian test of cointegration is also developed. The model is used to assess the validity of the Fisher effect and is also applied to equity market data.     

Aliternative methods for estimating long-run responses with applications to australian import demand

This paper compares and contrasts a number of single-equation and systems estimators of long-run responses with application to a three-variable import demand model. Two variants of Box and Tiao's (1977) canonical estimator are developed and associated tests for the number of cointegrating vectors are introduced. A simulation study indicates that, while both Box-Tiao estimators have empirical distributions with fatter tails than the normal, there is evidence that the incidence of extreme values is even greater with Johansen's (1988) ML procedure.     

Aliternative methods for estimating long-run responses with applications to australian import demand

This paper compares and contrasts a number of single-equation and systems estimators of long-run responses with application to a three-variable import demand model. Two variants of Box and Tiao's (1977) canonical estimator are developed and associated tests for the number of cointegrating vectors are introduced. A simulation study indicates that, while both Box-Tiao estimators have empirical distributions with fatter tails than the normal, there is evidence that the incidence of extreme values is even greater with Johansen's (1988) ML procedure.     

Cointegration testing under structural change: reducing size distortions and improving power of residual based tests

This paper investigates how standard residual based tests for cointegration—under structural change in the long run relationship—can be modified in order to reduce size distortions and improve power, by following the same ideas used in the unit root context. This is a natural strategy given that these tests are unit root statistics applied to estimated residuals from a cointegrating regression. In order to assess the finite sample performance of the alternative tests, a Monte Carlo experiment will be implemented to analyze size and power. Critical values for the tests constructed with GLS detrended data, proposed by Elliot et al. (Econometrica 64:813–836, 1996), are obtained by simulation.     

Inference on cointegrating ranks using lr and lm tests based on pseudo-likelihoods

This paper considers Lagrange Multiplier (LM) and Likelihood Ratio (LR) tests for determining the cointegrating rank of a vector autoregressive system. n order to deal with outliers and possible fat-tailedness of the error process, non-Gaussian likelihoods are used to carry out the estimation. The limiting distributions of the tests based on these non-Gaussian pseudo-)likelihoods are derived. These distributions depend on nuisance parameters. An operational procedure is proposed to perform inference. It appears that the tests based on non-Gaussian pseudo-likelihoods are much more powerful than their Gaussian counterparts if the errors are fat-tailed. Moreover, the operational LM-type test has a better overall performance than the LR-type test. Copyright O 1998 by Marcel Dekker, Inc.     

Comments on specification of household expenditure functions and equivalence scales by nonparametric regression

This paper considers Lagrange Multiplier (LM) and Likelihood Ratio (LR) tests for determining the cointegrating rank of a vector autoregressive system. n order to deal with outliers and possible fat-tailedness of the error process, non-Gaussian likelihoods are used to carry out the estimation. The limiting distributions of the tests based on these non-Gaussian pseudo-)likelihoods are derived. These distributions depend on nuisance parameters. An operational procedure is proposed to perform inference. It appears that the tests based on non-Gaussian pseudo-likelihoods are much more powerful than their Gaussian counterparts if the errors are fat-tailed. Moreover, the operational LM-type test has a better overall performance than the LR-type test. Copyright O 1998 by Marcel Dekker, Inc.     

The selection of zero-non-zero patterned cointegrating vectors in error-correction modelling

An effective and efficient search algorithm has been developed to select from an 1(1) system zero-non-zero patterned cointegrating and loading vectors in a subset VECM, B^q(l)y(t-1) + B^q-1 (L)Ay(t) = ε( t ) , where the long term impact matrix Bq(l) contains zero entries. The algorithm can be applied to higher order integrated systems. The Finnish money-output model presented by Johansen and Juselius (1990) and the United States balanced growth model presented by King, Plosser, Stock and Watson (1991) are used to demonstrate the usefulness of this algorithm in examining the cointegrating relationships in vector time series.     

SEPARATION, WEAK EXOGENEITY, AND P-T DECOMPOSITION IN COINTEGRATED VAR SYSTEMS WITH COMMON FEATURES

The aim of this paper is to study the concept of separability in multiple nonstationary time series displaying both common stochastic trends and common stochastic cycles. When modeling the dynamics of multiple time series for a panel of several entities such as countries, sectors, firms, imposing some form of separability and commonalities is often required to restrict the dimension of the parameter space. For this purpose we introduce the concept of common feature separation and investigate the relationships between separation in cointegration and separation in serial correlation common features. Loosely speaking we investigate whether a set of time series can be partitioned into subsets such that there are serial correlation common features within the sub-groups only. The paper investigates three issues. First, it provides conditions for separating joint cointegrating vectors into marginal cointegrating vectors as well as separating joint short-term dynamics into marginal short-term dynamics. Second, conditions for making permanent-transitory decompositions based on marginal systems are given. Third, issues of weak exogeneity are considered. Likelihood ratio type tests for the different hypotheses under study are proposed. An empirical analysis of the link between economic fluctuations in the United States and Canada shows the practical relevance of the approach proposed in this paper.     

Asymptotic Inference on Cointegrating Rank in Partial Systems

The likelihood ratio test for cointegrating rank is analyzed for partial (or conditional) systems in the vector autoregressive error-correction model. Under the assumption of weak exogeneity for the cointegrating parameters, the asymptotic distributions are given and tables of critical values are provided. A discussion is given of some of the assumptions of the model, why they are needed, and how they are tested.     

The selection of zero-non-zero patterned cointegrating vectors in error-correction modelling

An effective and efficient search algorithm has been developed to select from an 1(1) system zero-non-zero patterned cointegrating and loading vectors in a subset VECM, B ^q (l)y(t-1) + B ^q-1 (L)Ay(t) = ?( t ) , where the long term impact matrix Bq(l) contains zero entries. The algorithm can be applied to higher order integrated systems. The Finnish money-output model presented by Johansen and Juselius (1990) and the United States balanced growth model presented by King, Plosser, Stock and Watson (1991) are used to demonstrate the usefulness of this algorithm in examining the cointegrating relationships in vector time series.     

Conditionally Efficient Estimation of Long-Run Relationships Using Mixed-Frequency Time Series

I analyze efficient estimation of a cointegrating vector when the regressand and regressor are observed at different frequencies. Previous authors have examined the effects of specific temporal aggregation or sampling schemes, finding conventionally efficient techniques to be efficient only when both the regressand and the regressors are average sampled. Using an alternative method for analyzing aggregation under more general weighting schemes, I derive an efficiency bound that is conditional on the type of aggregation used on the low-frequency series and differs from the unconditional bound defined by the full-information high-frequency data-generating process, which is infeasible due to aggregation of at least one series. I modify a conventional estimator, canonical cointegrating regression (CCR), to accommodate cases in which the aggregation weights are known. The correlation structure may be utilized to offset the potential information loss from aggregation, resulting in a conditionally efficient estimator. In the case of unknown weights, the correlation structure of the error term generally confounds identification of conditionally efficient weights. Efficiency is illustrated using a simulation study and an application to estimating a gasoline demand equation.