A Multivariate Unit Root Test Based on the Modified Weighted Symmetric Estimator for VAR(p)

Multivariate unit root tests for the VAR model have been commonly used in time series analysis. Several unit root tests were developed. Most of the estimators of coefficient matrices developed in the VAR model are obtained using ordinary least squares estimators. In this paper, we suggest a multivariate unit root test based on a modified weighted symmetric estimator. Using a limited Monte Carlo simulation, we compare the powers of the new test statistic and the test statistic suggested in Fuller (1996).     

The Quality of Life in Korea: Comparative and Dynamic Perspectives

Social Indicators Research -     

Treatment Effects With Unobserved Heterogeneity: A Set Identification Approach

We propose the sharp identifiable bounds of the potential outcome distributions using panel data. We allow for the possibility that statistical randomization of treatment assignments is not achieved until unobserved heterogeneity is properly controlled for. We use certain stationarity assumptions to obtain the sharp bounds. Our approach allows for dynamic treatment decisions, where the current treatment decisions may depend on the past treatments or the past observed outcomes. As an empirical illustration, we study the effect of smoking during pregnancy on infant birthweight. We find that for the group of switchers the infant birthweight of a smoking mother is first-order stochastically dominated by that of a nonsmoking mother.     

Testing for a unit root in an arima(p,1,0) signal observed with ma(q) noise

An ARIMA(p,1,0) signal disturbed by MA(q) noise is an ARIMA(p,1, p+q+1) process restricted by nonlinear constraints on parameters. For this model with a unit root the restricted maximum likelihood estimator (RMLE) of the unit root is strongly consistent and it has the same limiting distribution as the ordinary least squares estimator of the unit root in an AR(1) model tabulated by Dickey and Fuller (1979). A modified RMLE is proposed which has the same limiting properties as the RMLE and is computationally much simpler. Simulation results show that our unit root tests based on the modified RMLE perform very well for small samples and compare favorably with the Said and Dickey (1985) tests with respect to both sizes and powers. An illustrative example from sample survey is given.     

Group sequential designs using both type I and type II error probability spending functions

We propose flexible group sequential designs using type I and type II error probability spending functions. The proposed designs preserve the overall significance level and power and allow the repeated testing to be perloimed at a flexible schedule. Computational methods are described. An example on a mega clinical trial is provided.     

Gender differences in client–provider relationship as active ingredient in substance abuse treatment

The client–provider relationship is increasingly evaluated as an active ingredient in the delivery of substance abuse treatment services. This study examines gender differences in client–provider relationship as an important treatment ingredient affecting retention in treatment and reduced post-treatment substance use. The study uses data collected for the National Treatment Improvement Evaluation Study (NTIES), a prospective, cohort study of U.S. substance abuse treatment programs and their clients. Data on individual characteristics were collected at the pre-treatment interview; on client–provider relationship and services received at treatment exit; and on post-treatment drug use at 12 months post-treatment. The analytic sample consists of 3027 clients from 59 service delivery units (1922 men and 1105 women). Structural equation modeling (SEM) was used to assess the structural relations and causal connections between relationship and service variables and treatment outcome variables. Results indicate that a positive client–provider relationship is related directly to longer duration and reduced post-treatment drug use for the total sample and for men analyzed separately. For women, a positive client–provider relationship was related directly to treatment duration and only indirectly to reduced post-treatment drug use. The findings point to the significance of including client–provider relationship in service delivery models – both as a therapeutic element as well as an element facilitative of matching services to specific client needs.     

The Economic Value of Bilingualism for Asians and Hispanics1

This study examines how bilingualism affects the wages of Asian and Hispanic workers using 2000 Census data. In contradiction to the general belief that bilingualism can provide a competitive advantage in the labor market, we find no evidence that 1.5‐generation and U.S.‐born Asian and Hispanic bilingual workers generally have higher wages than their English monolingual co‐ethnics; in some cases, in fact, their wages are significantly lower. In search of specific circumstances under which bilingualism might provide an economic advantage, we also examine interactions of language with such variables as education, employment in the public rather than the private sector, and the size of the population of mother‐tongue speakers. With limited exceptions, we find no sign of greater economic returns to bilingualism. Since bilingualism requires considerable effort to maintain across generations in the United States, we conclude that the virtual absence of economic rewards for it creates pressure for linguistic assimilation.     

Normal tests for unit roots based on instrumental variable estimators

For estimating unit roots of autoregressive processes, we introduce a new instrumental variable (IV) method which discounts large values of regressors corresponding to the unit roots. Based on the IV estimator, we propose new unit root tests whose limiting null distributions are standard normal. Observation at time t is adjusted for mean recursively by the sample mean of observations up to the time t. The powers of the proposed tests are better than those of the Dickey–Fuller tests and are comparable to those of the tests based on the weighted symmetric estimator, which are known to have the best power against stationary alternatives.     

Bayes least squares linear estimation of densities

Linear, least squares statistical methods in which the "parameters" are interpreted as random variables were introduced by Whittle, and further developed by Hartigan and others. They are applied here to the problem of estimating the coefficients in an orthogonal expansion of a multivariate density, given a simple random sample.