• Environmental and Resource Economics Review
• /
• v.12 no.4
• /
• pp.559-577
• /
• 2003

This paper analyzes household bottled water and water purifier expenditures, taking into account three important characteristics: expenditures may be censored at zero, may be interdependent across expenditure type, and may be endogenously and jointly determined. Censoring, interdependence, and endogeneity of the two expenditures are examined through simultaneous equation bivariate Tobit model. Expenditure function parameters are estimated using a 1997 household survey data collected in Seoul. The study detected interdependence between the two expenditures in the data. Moreover, the coefficient of one expenditure variable is statistically significant in the other expenditure equation. Thus, the overall results show that the simultaneous equation bivariate Tobit model employed here is appropriate for this analysis of the two expenditures. Finally estimated income and household size elasticities of the expenditures are presented.

• Environmental and Resource Economics Review
• /
• v.14 no.2
• /
• pp.447-474
• /
• 2005

Respondents, in the willingness to pay (WTP) survey, may have preference intensity about their stated WTP values. This study elicited a post-decisional intensity measure for each observed WTP answer for gathering information on the degree of preference intensity. In order to deal with the WTP data with preference intensity, this paper considers using the Type 3 Tobit model. This is usually estimated by the parametric two-stage estimation method assuming homoskedastic and bivariate normal error structure. However, if the assumptions are not satisfied, the estimates are inconsistent. The author has tested the hypotheses of homoskedasticity and normality, and could not accept them at the 1% level. The assumptions required to estimate the parametric Type 3 model are, therefore, too strong to be satisfied. As an alternative the parametric model, this study applies a semiparametric Type 3 Tobit model. The results show that the semiparametric model significantly outperforms the parametric model, and that more importantly, the mean WTP from the parametric model is significantly different from that from the semiparametric model.

• Journal of Korean Society of Forest Science
• /
• v.88 no.3
• /
• pp.400-407
• /
• 1999

The purpose of thin paper is to test difference of the two question formats, open-ended and dichotomous choice formats, in the contingent valuation method using the estimated recreational benefits. The data were collected from the visitors at the Songnisan National Park. The recreational benefit based on the equivalent variation. The two question formats, but the same content, were asked of the same individuals. In this analysis, it was used travel cost and monthly income as the exogenous variables, which assumed a linear functional form for the WTP equation. The model assumed a bivariate normal distribution on the basis of the probit and tobit model concerning the censored zero WTP. The result showed no differences in the recreational benefits from the different question formats under a same respondent. The mean benefit was estimated 25.556 Won per 5 years per visitors.