Estimating endogenous switching regression model with a flexible parametric distribution function: application to Korean housing demand

This study introduces Johnson's S_U -normal distribution which can accommodate the flexibility of true error distribution to obtain consistent estimates in an endogenous switching regression model. Simulation results indicate that the S_U -normal model outperforms the normal model for the consistency of estimators when the error distribution is nonnormal. Korean housing demand model estimated by the S_U -normal model also outperforms the normal model in terms of parameter estimates and graphical predictions.     

An application of dynamic specifications of factor demand equations to interfuel substitution in US industrial energy demand

In a recent paper, Jones (1995) [A dynamic analysis of the interfuel substitution in US industrial energy demand. J. Bus. Econ. Stat. 13 (4), 459–465] presents a dynamic analysis of interfuel substitution in US industry energy demand. The author concludes that a dynamic linear logit model is ‘superior' to a comparable dynamic translog model. The latter in fact violates concavity conditions whilst the logit formulation does not. This paper shows first of all that the dynamic formulation of the translog used in Jones (1995) is mis-specified. In fact, a parsimonious error-correction model (ECM) ‘dominates' alternative dynamic formulations, amongst which the partial adjustment mechanism used by the author. The ECM is able to generate optimal estimates of long-run and short-run elasticities, and it satisfies the concavity conditions of the cost function. Further, the theoretical framework used in this paper is the one recently proposed by Urga (1996) [On the identification problem in testing dynamic specification of factor demand equations. Econ. Lett. 52, 205–210] and Allen and Urga (1998) [Derivation and estimation of interrelated factor demands from dynamic cost function. Forthcoming in Economica]. It allows one to identify all coefficients (long-run and short-run) of the dynamic formulation via the joint estimation of the ‘effective' (short-run) cost function and the set of factor demand equations. This strategy solves, amongst other things, the parameter identification problem within the set of demand equations themselves, an issue which was originally noted by Anderson and Blundell (1982) [Estimation and hypothesis testing in dynamic singular equation systems. Econometrica, 1559–1571], re-addressed by Friesen (1992) [Testing dynamic specification of factor demand equations for US manufacturing. Rev. Econ. Stat. LXXIV (2), 240–250] and, more recently, by Urga (1996) and Allen and Urga (1998).     

A fast estimation procedure for discrete choice random coefficients demand model

We document speed-up gains of graphical processing unit (GPU) computing over central processing unit (CPU) for the estimation of discrete choice random coefficient demand model. When we use a moderate-sized GPU, the computation is six to twenty times faster, where the smallest speed-up factor, six, is obtained from a comparison with the parallel computing over sixteen CPU cores.     

The prospects of energy conservation: A different approach to the fuel demand for space heating

Zusammenfassung Die meisten empirischen Untersuchungen über das Ausmaß der jüngsten Energieeinsparungen verwendeten ökonometrische Beziehungen und Elastizitäten. Diese Arbeit jedoch geht von einer technischen Beschreibung des Services thermische Behaglichkeit aus. Basierend auf empirischem Datenmaterial über die Infrastruktur und das Klima der Stadt Wien werden sowohl die Aussichten und das Potential an Energieeinsparungen quantifiziert als auch der Effekt spezifischer Politiken. Für diesen Zweck wurde ein nichtlineares Optimierungsmodell mit disjunktiven Randbedingungen entwickelt und gelöst.

---

---

Research leading to this paper was supported by a grant from the Austria Federal Ministry of Science and Research, and that support is gretefully acknowledged. We wish to thank P. J. Jansen for discussions on this topic and also the referees for their valuable comments on this paper.     

Estimation of an endogenous switching regression model with discrete dependent variables: Monte-Carlo analysis and empirical application of three estimators

The performances of alternative two-stage estimators for the endogenous switching regression model with discrete dependent variables are compared, with regard to their usefulness as starting values for maximum likelihood estimation. This is especially important in the presence of large correlation coefficients, in which case maximum likelihood procedures have difficulties to converge. Monte-Carlo simulations indicate that an estimator that corrects for conditional heteroskedasticity of the residuals is superior in almost all instances, and especially when maximum likelihood is problematic. This result is also obtained in an empirical example in which off-farm work participation equations of farm women are conditional on farm work participation status. First version received: July 1995/final version received: March 1998     

The Endogenous Sequential Probit model: An application to the demand for hospital utilization

This paper develops the Endogenous Sequential Probit model to analyze the effect of managed care on hospital utilization by adult married US population and finds that managed care gives disincentives to utilize hospital care relative to the standard indemnity plans.     

Female labour force intermittency and current earnings: switching regression model with unknown sample selection

Using the Health and Retirement Survey from the USA, this paper finds a 16% selectivity-corrected wage penalty among women who engage in intermittent labour market activity. This penalty is experienced at a low level of intermittent activity, but appears to not play an important role in a woman's decision to undertake such activity. In addition, employer preferences appear to play a larger role than human capital atrophy in the determination of the wage penalty.     

The theory of rational random behavior and its application to demand analysis

The author's theory of rational random behavior is developed, which distinguishes between the planning and the implementation stage of the decision-maker. Under the assumption of a quadratic loss function and decision variables with an unrestricted range, that theory implies normally distributed disturbances of behavioral equations with a covariance matrix that is known up to a scalar multiple. Those results have to be extended asymptotically by Barbosa for restricted decision variables and a more general loss function. The approach is applied to both the Rotterdam and the linear expenditure demand models: using U.S. postwar data for the former, the author finds that a likelihood ratio test accepts the theory at the 10 percent significance level.     

A multiregional model of China and its application

China's rapid economic transformation since 1979 has been accompanied by a stark regional disparity with the coastal provinces integrating swiftly into the world markets while inland regions lagging far behind in the industrialisation process. The growth imbalance has placed China's provinces into different stages of development and called for different policy responses to deal with different challenges they are facing. Similarly, a national policy initiated by the central government is likely to be interpreted and implemented in different ways when local governments take into consideration the different conditions in their localities. Therefore, it would be extremely valuable for policy makers and researchers to have an analytical tool that is capable of studying interactions among the provincial economies and the dynamic relationship between the centre and local governments. The multiregional econometric model of China presented in this paper is the first attempt to provide such a tool. The pioneering model consists of over 1200 equations, covering 30 provinces of China that are linked by inter-regional flows of goods and services. The model is used to study the impact of fiscal recentralisation in the 1990s. The counterfactual simulation results show that when China's central government diverted tax revenue from provincial governments, it achieved higher budgetary revenue for the central government but at the expense of lower overall budgetary revenue and slower economic growth.     

A Tobit-type estimator for the censored Poisson regression model

The Poisson counterpart to the Tobit model is presented. Formulae for the gradient and Hessian of the relevant log-likelihood function are given and incorporated into a Newton-Raphson optimization algorithm. The asymptotic covariance matrix of the estimator is detailed. As an illustration, the NR algorithm is applied to a model of individual shopping behavior.     

Technology, demand and distribution: a cumulative growth model with an application to the Dutch productivity growth slowdown

This paper argues that the case for real wage growth restraint,and the consequent restoration of profitability, which the mainstreamconsensus regards as a necessary condition for sustained outputand productivity growth, is based on weak foundations, becauseit neglects the negative impact of wage moderation on productivitygrowth. Using a general Keynesian growth model, which integratesa (wage-led or profit-led) demand regime and a productivityregime (incorporating the productivity-growth enhancing effectsof higher demand and higher real wages), the conditions areidentified under which real wage restraint fails to raise outputand productivity growth. The model is applied empirically tothe Netherlands (1960–2000).     

A globally flexible,quadratic almost ideal demand system: an application to demand for meats and fish in Taiwan

A new demand system, called a globally flexible, quadratic almost ideal demand system (GF–QUAIDS), integrates both the quadratic Engel curve and the Fourier series. This novel GF–QUAIDS model is applied to study the demand for meats and fish in Taiwan. Empirical results show that the GF–QUAIDS fits the data well and that, by using the likelihood ratio test, its nested models are rejected, including the locally flexible functional form and the linear Engel curve. Nevertheless, other empirical model-selection measures indicate that the GF–QUAIDS can be as good as its famous nested models such as the AIDS, QUAIDS and GF–AIDS.     

Multi-country tests for the oscillator model with slowly varying coefficients

This paper applies Ramsey's (1992) forced oscillator model discovered with US data to Canada, Japan, UK, Belgium, France, Germany, and Italy. We also examine the aggregate production index for consumer goods in the US. Thus, the model is applied to and the results checked against data that were not involved in the discovery of the model in the original paper. The adequacy of the model is evaluated and international comparisons are made. The similarity of the estimates for all of the data sets examined suggest that the model may be widely applicable for consumer goods production indices, although France and Italy provide partial exceptions to this statement.     

Solar energy and the national energy dilemma : A model for policy evaluation

This study presents solar-enhanced energy scenarios for the future together with the policy considerations they imply. The methodology known as system dynamics is employed, and the work reported here builds on an earlier model developed by Sasser.Policy-related implications include the following. First, after-tax costs of solar technology must be competitive with the costs for coal as well as those for oil and gas to achieve a substantial market penetration. Second, the collapse predicted by the original Sasser model used in this study can be postponed by solar technology alone but not avoided. If solar technology is allowed to supply nearly 100% of the energy needs of each energy consumption sector, then each sector is allowed 100 years more growth before the inevitable collapse occurs. Third, the amount of imported crude oil that can be displayed by solar energy in the next two decades of this century is modest.     

A new energy model to capture the behavior of energy price processes

Owing to the vague fluctuation of energy prices from time to time, a new energy model, which considers both the mean-reverting behavior and the long memory property, is proposed in this paper. Since the problem of estimating parameters, in discrete time for this model, plays a central role in forecast inference, the problem of estimating the unknown parameters has been dealt with for the fractional Ornstein–Uhlenbeck process observed discretely. The asymptotic properties of these estimates are also provided. The numerical simulation results confirm the theoretical analysis and show that our method is effective. To show how to apply our approach in realistic contexts, an empirical study of energy in China, namely Daqing crude oil, is presented. The empirical results seem reasonable when compared to the real data.     

A diffusion model with application to evolving medical technologies

This paper reports on a system dynamics simulation model which was developed to study the diffusion of medical technologies but which should be applicable to other sorts of technologies as well. The model addresses both the adoption and the changing extent of use of an evolving, product-based technology and also endogenously accounts for changes in actual and perceived performance. Two specific medical case studies provide background to the model and are used for testing the model's ability to reproduce various aspects of historical behavior. The model's relatively complex structure is presented in the form of influence diagrams showing major flows and relationships. The paper concludes with a discussion of contributions and possible future uses and modifications of the model.     

A model of advertising with application to the German automobile industry

This paper develops a dynamic duopolistic model of advertising and price competition. Advertising accumulates a stock of goodwill which enters directly into the demand functions and both the cooperative and predatory effects of advertising are considered. It is shown that firms invest more in advertising the higher the degree of cooperativeness of advertising and some comparative static results are provided. In the second, empirical part of the article, estimations of the demand equations are presented using data from the German automobile industry. The main conclusion drawn is that advertising plays an important role in this market and has a predatory nature.     

日本企业IT应用面面观

90年代以来迅速发展的IT革命,在人类历史上第一次形成了全球性的信息通讯网络,这不仅有力地促进了信息化社会的发展,也给企业经营战略带来了很大的影响。日本企业在IT应用方面亦有了很大的进展并形成了自己的特点,在增加信息化投资、实现企业信息化的同时,从新产品开发、零部件