Search results

Long-run changes in living standards occupy an important place in development and growth economics, as well as in economic history. An extensive literature uses heights to study historical living standards. Most historical heights data, however, come from selected subpopulations such as volunteer soldiers, raising concerns about the role of selection bias in these results. Variations in sample mean heights can reflect selection rather than changes in population heights. A Roy-style model of the decision to join the military formalizes the selection problem. Simulations show that even modest differential rewards to the civilian sector produce a military heights sample that is significantly shorter than the cohort from which it is drawn. Monte Carlos show that diagnostics based on departure from the normal distribution have little power to detect selection. To detect height-related selection, we develop a simple, robust diagnostic based on differential selection by age at recruitment. A companion paper (H. Bodenhorn, T. Guinnane, and T. Mroz, 2017) uses this diagnostic to show that the selection problems affect important results in the historical heights literature.

The purpose of this study is to examine the effects of health on wages of Australian workers, with a focus on gender differences and the role of macroeconomic conditions in the effects.

The first 15 waves of the Household, Income and Labour Dynamics in Australia survey are used to estimate a wage model that accounts for the endogeneity of health, unobserved heterogeneity and sample selection bias.

The results show that, after accounting for the endogeneity of health, unobserved heterogeneity and sample selection bias, better health increases wages for Australian male workers, but not for female workers. The results also show that accounting for the endogeneity of health, unobserved heterogeneity and potential sample selection bias is important in estimating the effects of health on wages. In particular, a simple ordinary least squares estimator would underestimate the effect of health on wages for males, while overestimate it for females, and simply addressing the endogeneity of health using instrumental variables could overestimate the effect for both genders. It is also found that the effects of health on wages fall under depressed macroeconomic conditions, perhaps due to reduced job mobility and increased presentism during a recession.

This study adds to the international literature on the effects of health on wages by providing empirical evidence from Australia. The model applied to estimate the effects takes advantage of a panel dataset to address the bias resulting potentially from all the sources of the endogeneity of health, unobserved heterogeneity and sample selection. The results indeed show that failing to address these issues would substantially bias the estimated effects of health on wages.

This chapter proposes an approach toward the estimation of cross-sectional sample selection models, where the shocks on the units of observation feature some interdependence through spatial or network autocorrelation. In particular, this chapter improves on prior Bayesian work on this subject by proposing a modified approach toward sampling the multivariate-truncated, cross-sectionally dependent latent variable of the selection equation. This chapter outlines the model and implementation approach and provides simulation results documenting the better performance of the proposed approach relative to existing ones.

This chapter studies a snowball sampling method for social networks with endogenous peer selection. Snowball sampling is a sampling design which preserves the dependence structure of the network. It sequentially collects the information of vertices linked to the vertices collected in the previous iteration. The snowball samples suffer from a sample selection problem because of the endogenous peer selection. The author proposes a new estimation method that uses the relationship between samples in different iterations to correct selection. The author uses the snowball samples collected from Facebook to estimate the proportion of users who support the Umbrella Movement in Hong Kong.

Prior analyses of racial bias in the New York City’s Stop-and-Frisk program implicitly assumed that potential bias of police officers did not vary by crime type and that their decision of which type of crime to report as the basis for the stop did not exhibit any bias. In this paper, we first extend the hit rates model to consider crime type heterogeneity in racial bias and police officer decisions of reported crime type. Second, we reevaluate the program while accounting for heterogeneity in bias along crime types and for the sample selection which may arise from conditioning on crime type. We present evidence that differences in biases across crime types are substantial and specification tests support incorporating corrections for selective crime reporting. However, the main findings on racial bias do not differ sharply once accounting for this choice-based selection.

This paper analyzes the age pay gap in Italy (22%), particularly as it is of interest in an aging society and as it may affect social cohesion. Instead of the traditional approach for model selection, we use a machine-learning approach (post double robust Least Absolute Shrinkage Operator [LASSO]). This approach allows us to reduce Omitted Variable Bias (OVB), given data restrictions, and to obtain a robust estimate of the conditional age pay gap. We then decompose the conditional gap and analyze the impact of four further potential sources of heterogeneity (workers', sectors', and occupations' permanent heterogeneity as well as sample selection bias). The results suggest that age discrimination in pay is only perceived but not real in Italy for both men and women.

The purpose of this paper is to examine if and how entrepreneurial assistance programs, through guided preparation, affect start-up success.

–This study uses Heckman's two-stage sample selection model to predict the effect of contact and interactions with entrepreneurial support programs on start-up outcomes while taking into account the entrepreneur's self-selection into obtaining support from these programs.

The results indicate that, after controlling for individual characteristics, activities undertaken during the start-up process, organizational characteristics and external factors, guided preparation contributes to a greater likelihood of achieving positive start-up outcome. This finding holds even after controlling for the entrepreneur's self-selection into contacting and using outside assistance.

Results suggest that self-selection bias remains a concern when studying the impact of assistance programs on start-up outcomes. Future research should make sure to address self-selection in their analysis.

The study's results have implications for the design of start-up programs. It highlights the importance of delivery structures that are fluid, flexible, interactive, experiential, and tailored to the individual entrepreneur's needs.

This study focusses on assistance programs broadly defined (includes many different types of programs) and provides an empirical analysis that addresses self-selection.

We show how to use a smoothed empirical likelihood approach to conduct efficient semiparametric inference in models characterized as conditional moment equalities when data are collected by variable probability sampling. Results from a simulation experiment suggest that the smoothed empirical likelihood based estimator can estimate the model parameters very well in small to moderately sized stratified samples.

Without controlling for selection bias and the potential endogeneity of the treatment by using proper methods, the estimation of treatment effect could lead to biased or incorrect conclusions. However, these issues are not addressed adequately and properly in higher education research. This study reviews the essence of self-selection bias, treatment assignment endogeneity, and treatment effect estimation. We introduce three treatment effect estimators – propensity score matching analysis, doubly robust estimation (augmented inverse probability weighted approach), and endogenous treatment estimator (control-function approach) – and examine literature that applies these methods to research in higher education. We then use the three methods in a case study that estimates the effects of transfer student pre-enrollment debt on persistence and first year grades. The final discussion provides guidelines and recommendations for causal inference research studies that use such quasi-experimental methods.

We examine the sample self-selection and the use of LIFO or FIFO inventory method. For this purpose, we apply the Heckman-Lee’s two-stage regression to the 1973–1981 data, a period of relatively high inflation, during which the incentive to adopt the LIFO inventory valuation method was most pronounced. The predicted coefficients based on the reduced-form probit (inventory choice model) and the tax functions are used to derive predicted tax savings in the structured probit. Specifically, the predicted tax savings are computed by comparing the actual LIFO (FIFO) taxes vs. predicted FIFO (LIFO) taxes. Thereafter, we estimate the dollar amount of tax savings under different regimes. The two-stage approach enables us to address not only the managerial choice of the inventory method but also the tax effect of this decision. Previous studies do not jointly consider the inventory choice decision and the tax effect of that decision. Hence, the approach we use is a contribution to the literature. Our results show that self-selection bias is present in our sample of LIFO and FIFO firms and correcting for the self-selection bias shows that the LIFO firms, on average, had $282 million of tax savings, which explains why a large number of firms adopted the LIFO inventory method during the seventies.