Cliff J. Huang

Semiparametric Smooth Coefficient Models

In this article, we propose a semiparametric smooth coefficient model as a useful yet flexible specification for studying a general regression relationship with varying coefficients. The article proposes a local least squares method with a kernel weight function to estimate the smooth coefficient function. The consistency of the estimator and its asymptotic normality are established. A simple statistic for testing a parametric model versus the semiparametric smooth coefficient model is proposed. An empirical application of the proposed method is presented with an estimation of the production function of the nonmetal mineral industry in China. The empirical findings show that the intermediate production and management expense has played a vital role and is an unbalanced determinant of the labor and capital elasticities of output in production.

On stationary tests in the presence of structural breaks

Abstract We examine the effect of a structural break on stationarity tests. It has been shown that stationarity tests suffer from size distortion problems if a structural break exists but is ignored. This problem parallels the power loss problem of unit root tests ignoring an existing break. Nonetheless, the distributions of stationarity tests are asymptotically invariant to the exclusion of the existing break under the alternative hypothesis of a unit root. Our results thus clarify the nature of the effects of a structural break on inference regarding integrated time series.

On the Testing of Regression Disturbances for Normality

Abstract OLS and BLUS residuals in the classical linear regression model are compared with respect to testing the normality of regression disturbances. It is shown that in theory both types of residuals suffer from the common problem of lack of independence under the alternative hypothesis of nonnormal disturbances. Further, Monte Carlo experiments appear to show the superiority of OLS over BLUS residuals, as well as the superiority of the Shapiro-Wilk test over other tests studied, when one is testing for normality.

Estimation of Seemingly Unrelated Tobit Regressions via the EM Algorithm

An expectation-maximum (EM) likelihood algorithm is used to estimate two seemingly unrelated Tobit regressions in which the dependent variables are truncated normal. An illustrative example on the determination of the life-health insurance and pension benefits is also given.

Evidence on nonlinear error correction in money demand: the case of Taiwan

This paper proposes a nonlinear error-correction model based upon smooth transition regression methodology. The model is specified such that the short-run adjustment toward long-run equilibrium is nonlinear and that the error correction is a smooth function of long-run deviation. Empirical results obtained from estimating M2 money demand in Taiwan support the hypothesis of a nonlinear error-correction process and provide better interpretation of change in the demand for money.

Predicting Bank Failures: The Role of Structure in Affecting

The acceleration in the number of US bank failures during recent years provides valuable data for developing bank failure prediction models. A probit models which incorporates various bank structure variables as well as traditional financial ratios is used to explain bank failures during the 1982—3 period. The empirical findings suggest that important information regarding a banks likelihood of failure is contained in its balance sheet and income accounting data. It is also found that banks which affiliate with holding companies or are larger in size have a significantly lower probability of failure. This suggests that states which impose unit-banking rules or block holding company formation may be adding to failure risks.

Hypothesis testing in path models

Abstract The three prior criteria for linear recursive causal (path) model evaluation are shown to be equivalent to a more general technique. Hotellings T2 is introduced as a means of evaluating general hypotheses for the entire model, and some consequences of treating the model as a whole are discussed. The “twice standard error” rule is shown to be misleading for a number of reasons. An example of the application of the procedure and the computational technique is given, and some additional applications are suggested.

E-commerce, R&D, and productivity: Firm-level evidence from Taiwan

This study investigates the impact of e-commerce and R&D on productivity, using a unique panel dataset obtained from Taiwanese manufacturing firms for the period from 1999 to 2002. We specifically consider the network externalities of e-commerce and employ the system generalized method of moment (GMM) technique to deal with the endogenous problem of e-commerce adoption. The empirical results show that both e-commerce and R&D capital have a positive influence on productivity, while R&D exhibits a larger productivity-enhancing effect. We also find a complementary relationship between e-commerce and R&D on enhancing productivity. Crucially, the inter-industry network externality of e-commerce significantly contributes to productivity.

Optimal allocation of faculty time under uncertainty in production

Abstract A model of expected utility maximization is used to study the optimal allocation of a professors time between research and teaching in a situation where both production activities are subject to uncertainty. It was shown that, in general, a professors optimal allocation of time under uncertain production departs from the certainty strategy. An explicit model was developed to evaluate the effects of time effort in research and teaching with respect to changes in the degree of uncertainty, the Pratt-Arrow index of risk aversion, and the unit output rewards of research and teaching (for example, “merit pay”).

Uncertainty and total factor productivity in the Taiwanese banking industry

In this article, we formulate a behavioural model under uncertainty to estimate Total Factor Productivity (TFP) in the Taiwan banking industry. In particular, the article provides a model based on the safety-first rule under uncertainty to measure the risk premium in banking operations that are subject to loan default and other investment risks. With panel data of 40 banks in 1981–1996, a translog cost function and the associated share equations are used to estimate the dual rate of Total Cost Diminution (TCD), the dual Returns To Scale (RTS) and the derived primal rate of TFP. A constant elasticity of transformation output function is employed to construct an aggregated output index of loan and investment activities. The empirical results indicate zero productivity growth and a highly risk-averse banking industry. Government-owned banks are generally more risk-averse than privately owned banks. As expected, the Taiwan banking industry became more risk-venturesome after the deregulation and liberalization of the industry and during the stock market boom of the late 1980s.