The Effect of Protest Votes on the Estimates of WTP for Use Values of Recreational Sites

Contingent valuation studies are often characterized by a considerable number of protest responses, which may cause selectivity bias on the final estimates for WTP. Sample selection models can detect and – if necessary – correct selectivity bias. In economic applications where the relevant dependent variable is continuous, sample selection models are generally estimated using Heckman's 2-step method rather than the FIML estimator. Either method has its own drawback: computational complexity for the FIML method, susceptibility to collinearity problems for the 2-step method. Using data on valuation of forest resources for recreational use, we analyse the performance of the two estimators. In this application, given the presence of some collinearity, the FIML is preferred to the 2-step method. A procedure is outlined to deal with selectivity problems in similar settings.     

The Impact of Computer Use On Earnings in the UK

The effect of new technology on relative demands for workers has been the subject of much research in economics. Krueger (1993) and others have studied the impact of computers on earnings in the US and elsewhere. Such studies have been criticised for ignoring the possibility of bias due to unobserved heterogeneity between computer users and non‐users, resulting in computer users not being a random sub‐sample of all workers. As well as looking at the effects of computers on earnings in the UK, this paper extends previous analyses by using a sample selection framework to deal with the bias problem. Results indicate not only that returns to computer use are positive but that it is important to correct for the sample selection bias.     

Quantile regression with sample selection: Estimating women's return to education in the U.S.

This study uses quantile regression techniques to analyze changes in the returns to education for women. The data used is the March Current Population Survey for the years 1968, 1973, 1979, 1986 and 1990. The first step in estimating the single (linear) index selection equation uses Ichimura's (1993) semiparametric procedure. To correct for an unknown form of a sample selection bias in the quantile regression, the second step incorporates a nonparametric method, using an idea similar to one developed by Heckman (1980) and Newey (1991) for mean regression, and Buchinsky (1998) for quantile regression.  The results show that: (a) the returns to education increased enormously for the younger cohorts, but very little for the older cohorts; (b) in general the returns are higher at the lower quantiles in the beginning of the sample period and higher at the higher quantiles by the end of the sample period; (c) there is a significant sample selection bias for all age groups at almost all quantiles; (d) toward the end of the sample period there is a significant convergence of the returns at the various quantiles, especially for the younger cohorts and age groups; and (e) the semiparametric estimates of the selection equation are considerably different from those obtained for a parametric probit model.     

Asymmetries in Household Consumption and Liquidity Constraints — A Switching Regression Approach

Most of the empirical work in the Euler equation tradition in the consumption literature, that addressed the issue of liquidity constraints, has tested the hypotheses of Excess sensitivity and Liquidity constraints, with the sample selection based on a single exogenous criterion. Such a method will give misleading inferences, if the sample selection criterion is correlated with the Euler equation errors. In this paper, an attempt is made to correct for the sample selection bias and estimate the probability of a household being constrained simultaneously with the Euler equation parameters. Using the Additional Rural Incomes Survey (ARIS) data on rural Indian households, I find that consumption is excessively sensitive to income changes for constrained than for the unconstrained. In addition, I also find that consumption responds asymmetrically to positive and negative changes in income indicating a preference asymmetry for the unconstrained and binding liquidity constraints for the constrained group.

     

Female labour supply in Australia: Another example of a backward-bending labour supply curve

Hours worked by employed married females in Australia are analysed using techniques to correct for sample selection bias. Consistent with the conclusions of the study of female labour supply in the U.S. and Canada by Nakamura and Nakamura (1981), working wives work fewer hours per week when paid higher hourly wage rates.     

IMPACT OF PRICING STRUCTURE SELECTIVITY ON URBAN WATER DEMAND

Interest in demand management policies has intensified as residential water demand has grown in the United States. Using data from the 1984 American Water Works Association (AWWA) survey, the study here provides an empirical analysis of the differences in price elasticities of demand across water pricing block structures and examines these structures' "conservation-orientedness." However, a potential sample selection bias exists. That is, in cities where people are more interested in conservation, utility managers may be more likely to select a rate structure that they believe is conservation-oriented–an increasing block structure, for example. Managers' selectivity bias may cause research results either to understate or to overstate a particular block structure's impact on water conservation. The analysis here corrects for this selectivity bias in estimating water demand and tests whether consumers respond to average prices or to marginal prices. Correcting for selectivity bias involves an explicit analysis of the factors that influence utility managers' selection of rate structures. Estimating water demand under increasing and decreasing block structures suggests that sample selection bias remains a problem worthy of further investigation.     

DETERMINATION OF PARTICIPATION DECISION, HIRING DECISION, AND WAGES IN A DOUBLE SELECTION FRAMEWORK: MALE-FEMALE WAGE DIFFERENTIALS IN THE U.S. LABOR MARKET REVISITED

The magnitude of the male-female wage differential is known to be highly sensitive to the specification of the wage equations used. An important source of misspecification is the failure to correct the sample selection bias that results from estimating the wage equation obtained through two sequential decisions: the worker's decision to participate in the labor market and the employer's decision to hire. Estimation of the wage equation ignoring this double selection process leads to biased estimates, and consequently the resulting male-female wage differentials are likely to be misleading. Following a double selection approach and using a sample from the National Longitudinal Survey of Youth, this article examines the determinants of not only the wage equation but also the worker's participation and the employer's hiring decisions in both male and female samples. The study further demonstrates that the unexplained male-female wage differential remains underestimated when the roles of both selection decisions are ignored in the estimation of wage equations.     

Evaluating Social Policy by Experimental and Nonexperimental Methods

Although it is important to establish causal relationships in social policy evaluation, the effects are difficult to observe due to sample selection. To evaluate the performance of estimators designed to handle sample selection bias, we analyse data from a Norwegian rehabilitation project with a randomised experimental design. The data permit us to compare the performance of different nonexperimental estimators with the experimental results. In our case study we find that nonexperimental evaluation based on sample selection estimators with selection terms that fail to meet conventional levels of statistical significance is highly unreliable. The difference in difference estimator and propensity score matching estimators perform better in our context.
JEL classification : C 51; J 24; H 43; I 12     

EVIDENCE ON GENDER WAGE DISCRIMINATION IN PORTUGAL: PARAMETRIC AND SEMI‐PARAMETRIC APPROACHES

In this paper we consider several alternative approaches to analyze gender wage discrimination. Besides the traditional OLS estimator, we use two other approaches to control for sample selection bias problems: the parametric procedure suggested by Vella and Wooldridge, and the Li and Wooldridge semi‐parametric estimator. We study the case of Portugal, employing data from the European Community Household Panel. The results reveal that the discrimination estimates are sensitive to the different econometric approaches. In fact, when sample selection bias is taken into account, the discrimination values are reduced and are typically not significant.     

Assessing Changes in Capital and Investment as a Result of Fishing Capacity Limitation Programs

This paper provides a framework to evaluate the effectiveness of the European Multiannual Guidance Programmes on capacity limitation based on the evolution of the prices of used fishing vessels and on the measures of capital and investment derived from the estimation of a hedonic price function related to vessels characteristics. Attention is paid not only to the correct specification of the numerical variables but also, to testing for sample selection bias, to checking for structural change, to analysing the convenience of using alternative estimation methods and to discussing the potential bias as a result of asymmetric information in the used markets. The application of this framework to the Basque fleets evidences clear and coincident signs of successful adjustment after MAGP₃, but also a substantial investment in the next periods as a result of the modernisation of the fleets.     

Female income and the divorce decision: evidence from micro data

Escalating divorce rates during the 1960s and 1970s led to large numbers of academic investigations into the causes of divorce. Most of these studies concentrated on a significant increase in female income that resulted from rising female labour force participation rates. The difficulty with quantifying these arguments is that it is possible to observe the income of married females or it is possible to observe the income of divorced females, but it is not possible to observe both outcomes, simultaneously. This research attempts to resolve these difficulties by using sample selection techniques to correct for possible bias from simple observation of the income of married and divorced females.     

Wage Effects of Drinking in Australia

This article analyses the wage effects of drinking using the Australian Twin Registry data. A multinomial logit framework is employed to explain the allocation of workers across various drinking states, and to correct for selection bias in the wage equations. It is found that there is a significant positive wage premium for moderate drinking. A favourable family background (during childhood) is positively related to the measure of moderate drinking. Greater genetic endowments are also associated with moderate drinking.     

On bias correction in the multivariate sample-selection model

Heien and Wessells' two-step estimator for the multivariate sample-selection model has been used extensively during the past 15 years. A modified version of it, with slightly different selectivity regressors, has also appeared in the empirical literature. Both estimators are unfounded and generally do not correct for the sample selectivity bias as intended but have continued to gain popularity in empirical applications. The properties of the modified Heien–Wessells procedure are investigated in both the bivariate and multivariate contexts, and the conditions under which this estimator fails to correct for sample selectivity are examined. The theoretical properties are demonstrated with a simulated random sample.     

Filtering Returns for Unspecified Biases in Priors when Testing Asset Pricing Theory

Procedures are presented that allow the empiricist to estimate and test asset pricing models on limited-liability securities without the assumption that the historical payoff distribution provides a consistent estimate of the market's prior beliefs. The procedures effectively filter return data for unspecified historical biases in the market's priors. They do not involve explicit estimation of the market's priors, and hence, economize on parameters. The procedures derive from a new but simple property of Bayesian learning, namely: if the correct likelihood is used, the inverse posterior at the true parameter value forms a martingale process relative to the learner's information filtration augmented with the true parameter value. Application of this central result to tests of asset pricing models requires a deliberate selection bias. Hence, as a by-product, the article establishes that biased samples contain information with which to falsify an asset pricing model or estimate its parameters. These include samples subject to, e.g. survivorship bias or Peso problems.     

Hedonic Price Analysis and Selectivity Bias

Hedonic valuation of quality attributes can be misleading when theassumption that these attributes are exogenous to sample selection isviolated. This paper considers the simultaneity between hedonic valuationand sample selection in the context of a model of producer behavior andinvestigates empirically the case where land is demanded for use as an inputeither in agricultural production or in touristic development. The empiricalanalysis suggests that failing to correct for sample selection results in abiased valuation of the effect of water salinity on agricultural land.     

Dynamics of self-rated health and selective mortality

Self-rated health status (SRHS) is one of the most frequently used health measures in empirical health economics. This article analyzes the first seven waves of the Health and Retirement Study (HRS) and finds that (1) all available lags have decreasing but significant predictive power for current SRHS and (2) SRHS and future mortality are strongly related which leads to a specific selection problem known as survivorship bias. A parsimonious joint model with an autocorrelated latent health component in both the SRHS and the mortality equation is suggested. It is better able to capture the empirical facts than commonly used models including random effects and/or state dependence and better able to correct the survivorship bias than commonly used strategies such as inverse probability weighting.     

Complex ballot propositions,individual voting behavior,and status quo bias

One concern about direct democracy is that citizens may not be sufficiently competent to decide about complex policies. This may lead to exaggerated conservatism in the voting decision (status quo bias). To investigate how complexity affects individual voting behavior, we develop a novel measure of proposition complexity (using official pre-referendum booklets) and combine it with post-referendum survey data from Switzerland. Using Heckman selection estimations to account for endogenous variation in participation rates, we find that an increase in proposition complexity from the 10th to the 90th percentile would decrease voters' approval by 5.6 ppts, which is often decisive: an additional 12% of the propositions in our sample would be rejected.     

Health expenditure models: A comparison using panel data

In this article, we focus on the estimation of outpatient expenditures with panel data. We model the logarithm of expenditures and consider five different models. The first two are two-part and sample selection cross-section models. Two-part panel data models turn out to be inappropriate for dealing with expenditures. We thus estimate sample selection models with panel data: one without a lagged dependent variable and two with a lagged dependent variable. These two latter models differ in their assumptions on the variance of the residuals. Modelling heteroscedasticity may indeed be important to avoid the bias due to the retransformation problem. We show that lagged dependent variables are important factors for heteroscedasticity. For the models with state dependence, we provide a new solution to the initial conditions problem by controlling for generalised residuals. We establish that panel data models highly improve the correlation explained by the model in the time-series dimension without damaging the fit in the cross-section dimension. For all indicators of fit, the model with state dependence and heteroscedasticity seems to dominate the others.     

Non-random selection into entrepreneurship in the realm of government decentralization and corruption

For the past few decades, the extant literature on corruption has primarily relied on firm-level survey measures – for example, those from the World Bank – to explore the relevant empirical determinants of this illegal practice. However, these studies have potentially overlooked an underlying econometric problem – namely, non-random selection into entrepreneurship – that may bias all the estimated determinants to date if ignored in the analysis. Here, I assess this possibility by applying the traditional Heckman (1979) correction procedure in a novel way: using two different samples. I use my proposed solution in the context of government decentralization and firm-level corruption as a plausible application. Specifically, I revisit the question of the causal impact of government decentralization on firm-level corruption when the underlying sample selection issue is addressed. Results are worth noting. I find reasonable evidence of selection bias. On controlling for this, fiscal decentralization substantially decreases firm-level bribery, in general. This finding is in contradiction to the results reported by naive estimation strategies where the sample selection issue is completely ignored.     

The determinants of female labour supply in Hungary

The aim of this paper is to closely examine the effect on a married woman's labour supply decision of non-labour income and of her own wage rate in a transition economy. The paper first shows that labour force participation rates for Hungarian females, though decreasing, were still extremely high and comparable to male economic activity in western countries. It then estimates labour supply parameters using a model that controls for sample selection issues associated with the labour supply of married women. Wage elasticity is estimated to be high and positive (1.81) while the income effect is consistent with leisure being a normal good. No sample selection bias is found.