Showing papers in "Oxford Bulletin of Economics and Statistics in 1992"
TL;DR: In this paper, the authors present an examination of such models for variables integrated at most of order one, when tests for cointegration involve statistics with non-standard asymptotic distributions.
Abstract: The recent literature on maximum likelihood cointegration theory studies Gaussian vector autoregression (VAR) models allowing for some deterministic components in the form of polynomials in time. An examination is presented of such models for variables integrated at most of order one, when tests for cointegration involve statistics with non-standard asymptotic distributions. The asymptotic distributions of these test statistics are known to be functions of the distribution of certain matrices of stochastic variables involving integrals of Brownian motions. In fact, conditional on which restrictions on the coefficients of the polynomial in time are valid, different asymptotic distributions are obtained. The cases examined exhaust the hypotheses relevant to the cointegration rank analysis of I(1) variables in models involving up to linear trends and possibly seasonal dummies. The examination solves the numerical problem in making most of the interesting quantiles of these asymptotic distributions available to the applied econometrician.
2,831 citations
TL;DR: In this paper, Monte Carlo analysis and an empirical study of U.K. money demand demonstrate that when cointegration exists, the error-correction test generally is more powerful than the Dickey-Fuller test.
Abstract: A cointegration test statistic based upon estimation of an error-correction model can be approximately normally distributed when no cointegration is present. By contrast, the equivalent Dickey-Fuller statistic applied to residuals from a static relationship has a nonstandard asymptotic distribution. When cointegration exists, the error-correction test generally is more powerful than the Dickey-Fuller test. These differences arise because the latter imposes a possibly invalid common factor restriction. The issue is general and has ramifications for system-based cointegration tests. Monte Carlo analysis and an empirical study of U.K. money demand demonstrate the differences in power. Copyright 1992 by Blackwell Publishing Ltd
1,311 citations
TL;DR: In this article, it is shown how the table in S. G. Johansen and K. Juselius (1990) can be applied to make inference on the cointegration rank.
Abstract: It is shown how the table in S. Johansen and K. Juselius (1990) can be applied to make inference on the cointegration rank. The reason that inference is difficult is that the limit distribution of the proposed likelihood ratio test statistic depends on which parameter is considered under the null. It is shown how a recent procedure for unit root testing suggested by S. G. Pantula (1989) solves the problem. The procedure is illustrated by some published econometric examples. Copyright 1992 by Blackwell Publishing Ltd
975 citations
TL;DR: In this article, a new unit root test based on an alternative parameterization which has previously been considered by Bhargava (1986) was proposed, which allows for trend under both the null and the alternative, without introducing any parameters that are irrelevant under either.
Abstract: This paper provides a new unit root test based on an alternative parameterization which has previously been considered by Bhargava (1986). This parameterization allows for trend under both the null and the alternative, without introducing any parameters that are irrelevant under either. This is not so in the Dickey-Fuller parameterizations. The new test is extracted from the score or LM principle under the assumption that the errors are iid N(0, sigma squared (epsilon)), but our asymptotics hold under more general assumptions about the errors. Two forms of the test (a coefficient test and at t-test) are derived.
825 citations
TL;DR: In this paper, the problem of representing noncausality between two sets of cointegrated variables is addressed by using the likelihood ratio tests for noncauseality, which are different depending on the distributions of the unit roots in the marginal subsystem and the autoregressive part of the conditional subsystem.
Abstract: This paper addresses the problem of representing noncausality between two sets of cointegrated variables. In this case, maximum likelihood estimates of the parameters may be derived based on Johansen's approach. Likelihood ratio tests for noncausality are derived, which are different depending on the distributions of the unit roots in the marginal subsystem and the autoregressive part of the conditional subsystem. These tests are chi-squared-distributed. A Monte Carlo experiment illustrates t he advantages of the proposed approach when compared to the usual approach based on VAR's in levels. Copyright 1992 by Blackwell Publishing Ltd
217 citations
TL;DR: In this article, the effects of dynamic specification on the size and power of three cointegration tests are investigated. But the authors focus on the residual augmented Dickey-Fuller unit root test and the likelihood ratio test in vector autoregressive models.
Abstract: textThe article discusses the use of some Monte Carlo experiments to investigate the effects of dynamic specification on the size and power of three cointegration tests. The first test, proposed by Engle and Granger (1987), is the residual augmented Dickey-Fuller unit root test. The second is a Wald test for the significance of the error correction mechanism in an autoregressive-distributed lag model, suggested by Boswijk (1989) and further developed in Boswijk (1991). The third test is a likelihood ratio test in a vector autoregressive model, proposed by Johansen (1988) and extended in Johansen and Juselius (1990).
134 citations
TL;DR: In this paper, the authors exposits Wiener distribution theory for I(1) time series as an overview to a special issue on testing integration and cointegration, and applies it to an autoregressive process, a bivariate regression, and the similarity and power properties of two single-equation tests for co-integration.
Abstract: The paper exposits Wiener distribution theory for I(1) time series as an overview to a special issue on testing integration and cointegration. The behavior of an I(1) series is related to a Wiener process to derive the limiting distribution of its sample mean. Other Wiener processes are related to functions of the normal distribution. The analysis is applied to an autoregressive process, a bivariate regression, and the similarity and power properties of two single-equation tests for cointegration. Systems analyses of cointegration based on the Johansen approach are derived by successive concentration of the likelihood function. An empirical model for Norwegian consumption expenditure is examined. Copyright 1992 by Blackwell Publishing Ltd
83 citations
68 citations
TL;DR: In this paper, the authors analyzed the determinants of Norwegian households' consumption expenditure, using quarterly data for the period 1966(1)-1989(4) and found that consumption, income and a broad measure of households wealth appear to form a cointegrating relationship.
Abstract: The authors analyze the determinants of Norwegian households' consumption expenditure, using quarterly data for the period 1966(1)-1989(4). Expenditure, income and a broad measure of households wealth appear to form a cointegrating relationship. Likelihood ratio tests do not disprove that the cointegrating vector can be estimated efficiently from a single equation, i.e. a "consumption function." On this basis, a dynamic model of private expenditure is developed. Despite structural breaks in both the income and wealth processes, the consumption function displays considerable parameter constancy. Thus, the model can be attributed at least some degree of autonomy and structural invariance. On the basis of the parameter invariance tests, the "Lucas-critique" is in fact refuted. Consequently, the rational expectations approach yields (surprisingly) little insight into the causes of "breakdown" in pre-existing models. Instead their results suggest a more prosaic "left out variables" explanation, namely long-run wealth effects. Copyright 1992 by Blackwell Publishing Ltd
66 citations
49 citations
TL;DR: In this paper, the authors developed tests for a unit root based on the Durbin-Hausman principle using the ordinary least squares estimator and the pseudo instrumental variables estimator using the current variable as an instrument to formulate test statistics.
Abstract: The author develops tests for a unit root based on the Durbin-Hausman principle. The ordinary least squares estimator and the pseudo instrumental variables estimator using the current variable as an instrument are employed to formulate test statistics. The limit distributions of these tests are expressed as inverses of the squared Brownian functionals that have not been used previously in testing for a unit root. Finite sample and asymptotic distributions of Durbin-Hausman tests are tabulated by simulations. It is shown by simulations that Durbin-Hausman tests are more powerful than Dickey-Fuller tests when there is an intercept or a linear time trend in the model. Copyright 1992 by Blackwell Publishing Ltd
TL;DR: A unified framework to derive the distribution of conventional statistics under a unit root is presented in this paper, which is based on formulae which can generate (analytically as well as numerically) the densities and distributions of statistics such as the t ratio, the normalized autocorrelation coefficient, and many more.
Abstract: A unified framework to derive the distribution of conventional statistics under a unit root is presented. It is based on formulae which can generate (analytically as well as numerically) the densities and distributions of statistics such as the t ratio, the normalized autocorrelation coefficient, and many more. The name density (or distribution) generating equation is given to these formulae. As a practical example, the numerical and analytical aspects of the distribution of the t ratio under a unit root are discussed. Suggestions for further applications and extensions of these formulae are also given. Copyright 1992 by Blackwell Publishing Ltd
TL;DR: In this paper, the authors discuss some aspects of the error correction model for the Norwegian consumption function proposed in Brodin and Nymoen (1991) and discuss the pursued simplification process since the simplified model contains an error correcting variable that includes a contemporaneous variable and further includes combinations of variables which may be hard to interpret.
Abstract: textThe article discusses some aspects of the error correction model for the Norwegian consumption function proposed in Brodin and Nymoen (1991). The main focus is on the pursued simplification process since the simplified model contains an error correcting variable that includes a contemporaneous variable, and it further includes combinations of variables which may be hard to interpret. The economic implications of the alternative model for the Norwegian consumption, are that, apart from deterministics, consumption is effected by current income and wealth, and that the dynamics are established by an average of past changes in consumption/income ratios.