Joseph Hirschberg | University of Melbourne (original) (raw)

Papers by Joseph Hirschberg

Department of Economics Working Papers Series, Feb 1, 2001

This paper explores the insight from the application of cluster analysis to the results of a Data... more This paper explores the insight from the application of cluster analysis to the results of a Data Envelopment Analysis of productive behaviour.

Communications in Statistics Theory and Methods, Jan 8, 2003

This paper describes the estimation and testing of regression models that include multivariate ge... more This paper describes the estimation and testing of regression models that include multivariate generated or computed regressors in the presence of heteroskedasticity in the cross-section case. Hetero- skedasticity is often a problem in cross-section data and the usual tests for its presence cannot be applied when the heteroskedasticity is in some measure due to computed regressors. We investigate the case

Applied Financial Economics, 1992

... Fama, EF and Roll, R. (1971) Properties of symmetric stable distribution, Journal of the Amer... more ... Fama, EF and Roll, R. (1971) Properties of symmetric stable distribution, Journal of the American ... D. (1989) A selection criterion for choosing between functional forms of income distributions,Econometric Reviews, 7 ... Hirschberg, JG and Slottje, D. (1989) Remembrance of things ...

Journal of Income Distribution, Jun 6, 1997

The traditional textbook approach to avoiding the dummy trap problem is to delete a category from... more The traditional textbook approach to avoiding the dummy trap problem is to delete a category from each qualitative variable. This paper illustrates an alternative constraint which can be used to transform conventional dummy variable coefficients. This constraint serves to simplify their interpretation when the regression equation contains several qualitative variables and allows the computation of a coefficient for the deleted

This paper describes the estimation and testing of regression models that include multivariate ge... more This paper describes the estimation and testing of regression models that include multivariate generated or computed regressors in the presence of heteroskedasticity in the cross-section case. Heteroskedasticity is often a problem in cross-section data and the usual tests for its ...

Data collection and data processing procedures utilized in connection with a nationwide fire load... more Data collection and data processing procedures utilized in connection with a nationwide fire load and live load survey program are described. The techniques developed for transferring the field survey data to a form suitable for computer processing are discussed. Procedures adopted for data analysis are included. Documentation of the computer programs developed for this purpose is also presented.

The Fieller Method for the construction of confidence intervals for ratios of the expected value ... more The Fieller Method for the construction of confidence intervals for ratios of the expected value of two normally distributed random variables has been shown by a number of authors to be a superior method to the delta approximation. However, it is not widely used due in part, to the tendency to present the intervals only in a formula context. In

The estimation of regression models subject to exact linear restrictions, is a widely applied tec... more The estimation of regression models subject to exact linear restrictions, is a widely applied technique, however, aside from simple examples, the reparameterization method is rarely employed except in the case of polynomial lags. We believe this is due to the lack of a general transformation method for changing from the definition of restrictions in terms of the unrestricted parameters to the equivalent reparameterized model. In many cases the reparameterization method is computationally more efficient especially when estimation involves an iterative method. The general relationship that converts the two forms of the restricted model is derived. Examples involving systems of demand equations, polynomial lagged equations, and Splines are given in which the transformation from one form to the other are demonstrated. In addition, we demonstrate how a Wald test of the restrictions can be constructed using an augmented version of the reparameterized model. A computer program example is presented to demonstrate the equivalence. 1 We wish to acknowledge our colleagues Thomas B. Fomby and Jenny Lye for their helpful suggestions and comments on this paper, the usual caveat holds.

It is now widely believed that Foreign Direct Investment (FDI) affects positively the productivit... more It is now widely believed that Foreign Direct Investment (FDI) affects positively the productivity of individual enterprises in the host economy and this is a main argument in favour of FDI. Although China is the largest recipient of FDI after the USA, there has been little empirical research as to whether such productivity effects of FDI are present in China. In this paper we attempt to answer some of these questions using a data set constructed from the Third Industrial Census of China conducted in 1995. We employ a Data Envelopment Analysis (DEA) to estimate individual enterprise level relative productivity scores from an industry level nonparameteric production technology. These individual results are then related to the level of FDI via a Bootstrap regression which is used to insure consistent estimates of the variance covariance matrix of the regression estimates. The results are decomposed by size of enterprise and industry and provide an insight to the manner in which such spillover effects may be operating in the Chinese Economy.

The estimation of regression models subject to exact linear restrictions, is a widely applied tec... more The estimation of regression models subject to exact linear restrictions, is a widely applied technique, however, aside from simple examples, the reparameterization method is rarely employed except in the case of polynomial lags. We believe this is due to lack of a general transformation method for changing from the definition of restrictions in terms of the unrestricted parameters to the equivalent reparameterized model. In many cases the reparameterization method is computationally more efficient especially when estimation involves an iterative method. The general relationship that converts the two forms of the restricted model is derived. Examples involving systems of demand equations, polynomial lagged equations, and Splines are given in which the transformation from one form to the other are demonstrated. In addition, we demonstrate how a Wald test of the restrictions can be constructed using an augmented version of the reparameterized model. A computer program example is presented to demonstrate the equivalence.

Electronic Gaming Machine (EGM) venues. In this paper we review the literature that examines the ... more Electronic Gaming Machine (EGM) venues. In this paper we review the literature that examines the interrelationship between gambling, problem gambling and smoking in order to gauge the indirect effects of smoking bans in gaming venues. We then perform an analysis on the consequences of a smoking ban in Victoria, Australia that was instituted on September 1 st , 2002. This analysis investigates the nature of the pattern of drops in local EGM revenue and the impact on the state tax revenue.

Marginal Effects for binary response variables are nonlinear functions of the parameter estimates... more Marginal Effects for binary response variables are nonlinear functions of the parameter estimates. Many of the applications of these models also contain nonlinear regressors. We adopt a simple linear transformation of the parameter estimates that dramatically simplifies the calculation of marginal effects and corresponding standard errors for these models. We also use this approach to present an empirical example in which marginal effects and the corresponding confidence interval are graphically illustrated and interpreted.

A number of papers have reported a link between the consumption of alcohol and wages earned. This... more A number of papers have reported a link between the consumption of alcohol and wages earned. This relationship has been found to be positive for low to moderate levels of alcohol but then turns to negative as the level of consumption reaches a certain level of consumption although. Although this relationship differs between men and women and it also appears to be mitigated by other activities such as smoking very little of this research attempts to make inferences concerning the location of the point of inflection - the level of consumption where the negative aspects begin to dominate the relationship. Although this inflection point as been defined and in fewer cases a confidence interval has been constructed (for example MacDonald and Shields, 2001). In this paper we use the published evidence to make inferences concerning the turning points found in these studies. The aim of this study to determine whether there are some overarching conclusions as to the nature of the turning poin...

The testing of incremental cost-effectiveness ratios for competing interventions has been examine... more The testing of incremental cost-effectiveness ratios for competing interventions has been examined using a number of techniques for the construction of confidence bounds. Both the Fieller method and the bootstrap have been proposed as alternatives to the traditional delta method for the construction of confidence intervals, however it has not been previously recognized that a form of Fieller's method could be adapted directly to the bootstrap. The traditional application of the bootstrap to ratios always results in the definition of finite intervals even when the distribution of denominator may be located near zero. Using the technique applied here we can benefit from the Fieller approach while relaxing the assumption of a specific distribution such as the normal or the F-distribution.

This chapter presents the econometric methods that are used in health economics to model individu... more This chapter presents the econometric methods that are used in health economics to model individuals health care costs. These methods are used for prediction, projection and forecasting, in the context of risk adjustment, resource allocation, technology assessment and policy evaluation. The chapter reviews the literature on the comparative performance of the methods, especially in the context of forecasting individual health care costs, and concludes with an empirical case study.

The good health of an individual is a combination of uncontrollable factors that includes genetic... more The good health of an individual is a combination of uncontrollable factors that includes genetics and random events and controllable factors through the regulation of activities such as smoking, drinking, eating, exercise and other informed choices. Since the work of Grossman (1972) a significant relationship between health and earnings is predicted. This paper investigates this relationship using four indicators of healthy activity in the context of Australia. Using the 1995 Australian National Health Survey we simultaneously examine the effects of drinking, smoking, eating and exercising on wages. Special attention is given to nonlinearities and interaction of these effects between each other and with age as is suggested in the medical literature. To model the interaction of smoking with these other effects separate models are fit for smokers and nonsmokers which account for the potentional for selectivity bias. Results are given separately for men and women.

Regression specifications frequently employ regressors that are defined as the product of two oth... more Regression specifications frequently employ regressors that are defined as the product of two other regressors to form an interaction. Unfortunately, these models have a number of potential difficulties when it comes to interpretation. In this paper, we discuss two common aspects of these specifications that can lead to difficulties. The first is the change in parameter estimates and tstatistics when different definitions of dummy variables are used. The second involves an application of Fieller’s theorem to draw appropriate inferences from a regression with interaction variables.

Department of Economics Working Papers Series, Feb 1, 2001

Communications in Statistics Theory and Methods, Jan 8, 2003

Applied Financial Economics, 1992

Journal of Income Distribution, Jun 6, 1997

This paper describes the estimation and testing of regression models that include multivariate ge... more This paper describes the estimation and testing of regression models that include multivariate generated or computed regressors in the presence of heteroskedasticity in the cross-section case. Heteroskedasticity is often a problem in cross-section data and the usual tests for its ...

The estimation of regression models subject to exact linear restrictions, is a widely applied tec... more The estimation of regression models subject to exact linear restrictions, is a widely applied technique, however, aside from simple examples, the reparameterization method is rarely employed except in the case of polynomial lags. We believe this is due to lack of a general transformation method for changing from the definition of restrictions in terms of the unrestricted parameters to the equivalent reparameterized model. In many cases the reparameterization method is computationally more efficient especially when estimation involves an iterative method. The general relationship that converts the two forms of the restricted model is derived. Examples involving systems of demand equations, polynomial lagged equations, and Splines are given in which the transformation from one form to the other are demonstrated. In addition, we demonstrate how a Wald test of the restrictions can be constructed using an augmented version of the reparameterized model. A computer program example is presented to demonstrate the equivalence.

This paper compares the performance of three related systems of input demand equations in their a... more This paper compares the performance of three related systems of input demand equations in their ability to model time-of-day electricity demand. They are derived from the Translog, Generalized Leontief, and the QuadraticSquareroot
flexible cost function specifications. The data used in the
estimation are firm-level monthly observations of seven time differentiated electricity inputs. Linear demand equations are specified in order to facilitate the firm-level estimation. The assumption of symmetric separability so often employed in demand studies of this nature has been relaxed in favor of non-symmetric separability which allows the substitution
of non-electric inputs for electricity inputs. Comparisons of these specifications are made on the basis of the log-likelihood, the Akaike information criterion, in-sample prediction, and the precision of the price elasticity estimates. The results are reported in terms of individual firm model comparisons