Implementing quantile selection models in Stata

In this article, we describe qregsel, a community-contributed command that implements a copula-based sample-selection correction for quantile regression recently proposed by Arellano and Bonhomme (2017, Econometrica 85: 1–28). The command allows the user to model selection in quantile regressions by using either a Gaussian or a one-dimensional Frank copula. We illustrate the use of qregsel with two examples. First, we apply the method to the fictional dataset used in the Stata Base Reference Manual for the heckman command. Second, we replicate part of the empirical application of the original article using data for the United Kingdom that cover the period 1978–2000 to compare wages of males and females at different quantiles.

Evaluation of Seven Near-Real-Time Satellite-Based Precipitation Products for Wet Seasons in the Nierji Basin, China

This study evaluated and intercompared seven near-real-time (NRT) versions of satellite-based precipitation products (SPPs) with latencies of less than one day, including GSMaP-NRT, GSMaP-Gauge-NRT, GSMaP-NOW, IMERG-Early, IMERG-Late, TMPA 3B42RT, and PERSIANN-CCS for wet seasons from 2008 to 2019 in a typical middle–high latitude temperate monsoon climate basin, namely, the Nierji Basin in China, in four aspects: flood sub-seasons, rainfall intensities, precipitation events, and hydrological utility. Our evaluation shows that the cell-scale and area-scale intercomparison ranks of NRT SPPs are similar in these four aspects. The performances of SPPs at the areal scale, at the event scale, and with light magnitude are better than those at the cell scale, at the daily scale, and with heavy magnitude, respectively. Most SPPs are similar in terms of their Pearson Correlation Coefficient (CC). The main difference between SPPs is in terms of their root-mean-square error (RMSE). The worse performances of TMPA 3B42RT are mainly caused by the poor performances during main flood seasons. The worst performances of PERSIANN-CCS are primarily reflected by the lowest CC and the underestimation of precipitation. Though GSMaP-NOW has the highest RMSE and overestimates precipitation, it can reflect the precipitation variation, as indicated by the relatively high CC. The differences among SPPs are more significant in pre-flood seasons and less significant in post-flood seasons. These results can provide valuable guidelines for the selection, correction, and application of NRT SPPs and contribute to improved insight into NRT-SPP retrieval algorithms.

arhomme: An implementation of the Arellano and Bonhomme (2017) estimator for quantile regression with selection correction

Despite constituting a major theoretical breakthrough, the quantile selection model of Arellano and Bonhomme (2017, Econometrica 85: 1–28) based on copulas has not found its way into many empirical applications. We introduce the command arhomme, which implements different variants of the estimator along with standard errors based on bootstrapping and subsampling. We illustrate the command by replicating parts of the empirical application in the original article and a related application in Arellano and Bonhomme (2018, Handbook of Quantile Regression, chap. 13).

Modelling Neighbourhood Effects in Three Dutch Cities Controlling for Selection

The non-random selection of people into neighbourhoods complicates the estimation of causal neighbourhood effects on individual outcomes. Measured neighbourhood effects could be the result of characteristics of the neighbourhood context, but they could also result from people selecting into neighbourhoods based on their preferences, income, and the availability of alternative housing. This paper examines how the neighbourhood effect on individual income is altered when geographic selection correction terms are added as controls, and how these results vary across three Dutch urban regions. We use a two-step approach in which we first model neighbourhood selection, and then include neighbourhood choice correction components in a model estimating neighbourhood effects on individual income. Using longitudinal register datasets for three major Dutch cities: Amsterdam, Utrecht and Rotterdam, and multilevel models, we analysed the effects for individuals who moved during a 5-year period. We show that in all cities, the effect of average neighbourhood income on individual income becomes much smaller after controlling for explicitly modelled neighbourhood selection. This suggests that studies that do not control for neighbourhood selection most likely overestimate the size of neighbourhood effects. For all models, the effects of neighbourhood income are strongest in Rotterdam, followed by Amsterdam and Utrecht.

College majors and wages in Turkey: OLS and quantile regression with sample selection correction

Purpose This study aims to analyze the wage differentials of the majors in college education in Turkey, which is a country implementing an ongoing expansion in college education in recent years. Design/methodology/approach The study implements Mincreian wage regression using ordinary least squares, Heckman two-step estimation and quantile regression with sample selection correction by using household labor force surveys of TurkStat from the years 2014–2017. Findings The findings indicate one of the highest heterogeneity, close to 0.50 log points, between majors in the literature. The within-heterogeneity created by majors is highest among the graduates of social-behavioral sciences, law, biology, physics, mathematics, statistics, computer, engineering and manufacturing, as shown by a 90–10 difference, which is almost 700% for some of these majors. This study shows that the natural science and technical majors that are expected to be more productive and to be paid more fall behind in the wage distribution. Research limitations/implications Estimation results show that natural science majors, except for subjects allied to medicine and engineering, are paid lower than law and service-sector-related majors. This indicates that the predictions of the skill-biased technical change hypothesis are not valid in the wage profiles in Turkey and that some majors supply more than the sectoral needs. This casts doubts on the effectiveness of the ongoing higher education expansion process of the country. Originality/value This study contributes to the literature on wage differentials of college majors, an area with limited studies. This is the first study analyzing wage differentials of the field of studies by correcting sample selection bias for the Turkish case.

Scalable Data Fusion with Selection Correction: An Application to Customer Base Analysis

A computationally scalable, statistically efficient aggregate-disaggregate data fusion method that corrects for selection bias is applied to model customer relationship dynamics at a subscription-based firm.

Gender wage gaps in Ghana: A comparison across different selection models

The wage of an individual is observed only when he/she is employed. However, getting employment requires two decisions. First, an individual has to decide to participate in the labour market, and second, an employer must decide to hire that individual. Since female labour market participation often differs from that of men, and employers’ decisions to hire may also be influenced by gender, it is appropriate to account for this double selection process. This study uses the latest household survey in Ghana to estimate gender wage gaps by correcting for this double selection process. We find that the average total gender wage gap is positive and significant irrespective of the sample selection correction method used. Our results indicate that women on average receive lower wages than men. Irrespective of the type of selection method used, our findings suggest that almost all the wage gap is a result of differences in returns, with only a small part coming from differences in observables. We find that the gender wage gap is smaller among formal wage employees and the gap decreases as education level increases. Although our findings indicate a similar trend in the wage gap across all specifications, the magnitude of the gap is sensitive to the choice of the model. This points to the need to be cautious about the choice of sample selection correction used to analyse gender wage gaps.

Informal payments in Romanian health care system. A sample selection correction

This paper draws on economic theory, sociology and political science approaches to explain informal payments in the Romanian health care system. It estimates the likelihood of paying a bribe (informal payment) using a reduced health care demand equation in a probit model with sample selection correction. Social capital, as having a relationship with doctors, and the perception of the health care system, as corrupt, are found to influence the probability of making an informal payment. The likelihood of making an informal payment in the Romanian health care system is modelled using a maximum-likelihood probit estimation with sample selection correction. In the selection equation, reduced health care demand, self-perceived health status and being afraid of diseases are used as exclusion restrictions for identifying the parameters of the econometric model.