Sample Selection Correction Research Articles

The returns to returning to school

PurposeThis work examines the returns to education for workers who pursue additional education after time out of the labor force. It compares those who remain in the labor force during additional education with those who drop out of the labor force during additional education. It compares two cohorts of the National Longitudinal Survey of Youth (NLSY).Design/methodology/approachThis work utilizes a difference equation to estimate the returns to education for workers who pursue additional education after time spent out of school and in the labor force.FindingsThe results indicate a sheepskin return of approximately 14% for those who remain in the labor force and a return of approximately 9% to years of additional education for those who drop out of the labor force. This contrasting pattern of returns is robust to sample selection correction and a variety of checks.Research limitations/implicationsThis work does not fully account for all threats to causation. Further research could pursue these and make use of data from more clearly defined periods of education.Practical implicationsThis work finds key differences between the internal labor market faced by those remaining in the labor force and the external labor market faced by those dropping out of the labor force. A policy focused on re-training workers should account for these differences.Originality/valueThis is the first work to compare workers who pursue additional education while remaining in the labor force to workers who pursue additional education and drop out of the labor force.

The impact of information and communication technology on the technical efficiency of smallholder vegetable farms in Shandong of China

Abstract Farmers have started to adopt information and communication technology (ICT), which has considerable potential to impact farm performance. This study uses data from a 2018 survey of 763 vegetable smallholder farms in China to estimate the impact of ICT on technical efficiency (TE). We adopt propensity score matching to create a balanced sample of ICT users and non-users and a stochastic frontier model with sample selection correction to compare the two groups’ TE. After accounting for self-selection bias from both observables and unobservables, the study finds a positive effect of ICT use on TE. On average, the TE score of ICT users is 0.64, whereas ICT non-users have a lower score of 0.57. A quantile regression analysis further reveals a heterogeneous impact of ICT on TE, with the largest effects among less efficient farms. These results suggest that vegetable farmers’ performance could be fostered by the widespread use of ICT.

Assessing the glass ceiling and sticky floor effects in the Argentinian tourism sector

This paper examines the gender wage gap in the Argentinian tourism sector. We use microdata from the Permanent Household Survey 2018-2019 and implement the Machado and Mata quantile decomposition method with an innovative sample selection correction. We find that the gender wage gap in tourism sector and its subsectors is not constant throughout the wage distribution. In the whole tourism sector, the gap is greater at the bottom of the wage distribution and is explained by discrimination against women. In hospitality, the gap increases over the distribution suggesting a glass ceiling effect explained by discrimination. In restaurants, however, the gap is greater at the bottom of the distribution, which provides evidence of a sticky floor effect due to differences in individuals’ endowments. Finally, the gap in travel agencies, cultural, sports and entertainment services remain relatively constant across the wage distribution.

The link between nascent entrepreneurs' role identity aspirations and their opportunity exploration and exploitation activities

AbstractAlthough nascent entrepreneurs' opportunity exploration and exploitation activities are crucial for later venture evolution and performance, it is unclear why some nascent entrepreneurs engage more in these new venture creation activities than others. Building upon role identity theory, this study advances a nascent entrepreneur's various role identity aspirations as drivers of these activities. In particular, we argue that inventor, founder, and developer role identity aspirations are linked differently to the range of opportunity exploration and exploitation activities nascent entrepreneurs undertake.We use a two‐stage Poisson model with sample selection correction on a random sample of 1582 inhabitants of Flanders, Belgium. The first stage estimates an individual's probability to engage in nascent entrepreneurship. The second stage tests the relationships between an individual's role identity aspirations and the range of exploration and exploitation activities undertaken, conditional on being a nascent entrepreneur, which holds for 162 individuals in our sample. We find that while some role identity aspirations are related to the range of opportunity exploration activities undertaken, others primarily affect the range of exploitation activities. Our findings have important implications for literature on new venture creation and entrepreneurial opportunities, for identity literature, and for nascent entrepreneurs wanting to start a venture.

The gender wage gap in Myanmar: Adding insult to injury?

Using two surveys from 2017, we analyze the gender wage gap for urban workers in Myanmar. We start from a standard wage equation and condition on education, experience, health and a small set of household demographic attributes. Subsequently we control for differences in occupational choice and sector of employment. We estimate the models with sample selection correction and this leads to estimated average wage offers that are lower than the observed average wages for women. Selection into wage work results in a workforce where female wage-workers have higher levels of education compared to their male counterparts. However, average wages for female workers continue to be 29% lower than male average wages. Differences in observable attributes do not account for this gap. Instead, it is associated with a lower base wage and lower remuneration of women’s experience. Going beyond the traditional decomposition methods, we utilize our matched employer–employee survey to generate exact comparisons of female and male production workers with equal levels of education and experience, employed in the same enterprises. Even in this setting, we find a gender wage gap of 13%. Our analysis thus indicates discrimination against women in Myanmar’s labor markets.

Asset Pricing Tests of Infrequently Traded Securities: The Case of Municipal Bonds

Abstract Using a dynamic selection model, we obtain consistent and unbiased estimates of risk and returns for infrequently traded bonds and conduct the first comprehensive asset pricing test of municipal bonds using the multifactor approach. Correction for sample selection and infrequent trading problems results in substantially higher beta estimates. Besides conventional risk factors, illiquidity and taxes are important for the pricing of municipal bonds. Importantly, bond returns contain a significant liquidity risk premium. Failing to account for sample selection bias leads to erroneous inference on the magnitude of systematic risk and substantial underestimation of risk premiums. (JEL G12, G14)

Returns to formal and informal vocational education and training in India

PurposeIn the process of school-to-work transition, the role of general education and vocational education and training (VET) remains quite central. Based on the human capital theory, we estimate whether investment in VET brings additional returns for workers across the age cohorts.Design/methodology/approachThe focus of our study being the labour market in India, the data from the Periodic Labour Force Survey 2018–19, conducted by the National Statistical Office, has been used for analysis. We have applied the ordinary least square method with sample selection correction, the quasi-experimental technique of propensity score matching and heteroskedasticity based instrumental variable approach to estimate the returns with respect to no VET, formal VET and informal VET.FindingsOur study shows that workers with formal VET earn higher wages than workers with no VET or informal VET. The study finds that workers with informal VET do not earn higher wages than workers with no VET. Moreover, from the age cohort analysis, we have deduced that wage advantage of workers with formal VET persists across all age cohorts and, in fact, accentuates with an increase in age.Originality/valueWe have estimated that VET being complemented with basic general education fetches higher returns in the labour market, especially when provided through formal channels. Moreover, to the best of our knowledge, in the case of developing countries where informal VET is widely provided, this is one of the first studies that captures the return to informal VET. Lastly, complementing the existing studies on the developed countries, we have estimated the returns to VET over the life cycle of the workers.

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

PurposeThis 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/approachThe 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.FindingsThe 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/implicationsEstimation 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/valueThis 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.

Quasi-Experimental Evaluation of Alternative Sample Selection Corrections

Researchers routinely use datasets where outcomes of interest are unobserved for some cases, potentially creating a sample selection problem. Statisticians and econometricians have proposed many selection correction methods to address this challenge. We use a natural experiment to evaluate different sample selection correction methods’ performance. From 2007, the state of Michigan required that all students take a college entrance exam, increasing the exam-taking rate from 64% to 99% and largely eliminating selection into exam-taking. We apply different selection correction methods, using different sets of covariates, to the selected exam score data from before 2007. We compare the estimated coefficients from the selection-corrected models to those from OLS regressions using the complete exam score data from after 2007 as a benchmark. We find that less restrictive semiparametric correction methods typically perform better than parametric correction methods but not better than simple OLS regressions that do not correct for selection. Performance is generally worse for models that use only a few discrete covariates than for models that use more covariates with less coarse distributions.

Implementing Quantile Selection Models in Stata

This article describes qregsel, a Stata module to implement a copula-based sample selection correction for quantile regression recently proposed by Arellano and Bonhomme (2017, Econometrica 85(1): 1-28). The command allows the user to model selection in quantile regressions using either a Gaussian or an one-dimensional Frank copula. We illustrate the use of qregsel with two examples. First, we apply the method to the fictional data set employed in the Stata base reference manual for the heckman command. Second, we replicate part of the empirical application of the original paper using data for the UK that covers the period 1978-2000 to compare wages of males and females at different quantiles.

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.

Private and Social Returns to Investment in Education: the Case of Turkey with Alternative Methods

ABSTRACT This paper estimates private and social returns to investment in education in Turkey, using the 2017 Household Labour Force Survey (latest available at the time of writing) and alternative methodologies. The analysis uses the 1997 education reform of increasing compulsory education by three years as an instrument. This results in a private rate of return on the order of 16% for higher education and a social return of 10%. Using the number of children younger than age 15 in the household as an exclusion restriction, sample selection correction is applied, and it shows that the returns to education for females are higher than those for males. Contrary to many findings in other countries, private returns to those working in the public sector are higher than those in the private sector, and private returns to those who followed the vocational track in secondary education are higher than those in the general academic track. The paper discusses the policy implications of the findings.

District or distributed space heating in rural residential sector? Empirical evidence from a discrete choice experiment in South China

Rural residential space heating is strongly linked to policy considerations related to the clean energy supply and rural sustainable development. As there has been a hot debate on whether to promote district heating in South China, this paper aims to investigate household heating choices with a particular focus on rural residential sector, using the survey data from Sichuan Province. In order to investigate the main influencing factors of household decision-making behaviors, a multinomial logit (MNL) model with sample selection correction and an alternative-specific conditional logit (ASCL) model were used to respectively analyze the actual and stated choices. The estimation results show that energy-specific attributes such as safety level and smoky level have statistically significant effects on household stated preferences for heating systems. Households prefer to adopt lower-cost heating system with high quality energy sources. Among household-specific characteristics, income level, educational level of the decision-maker, household demographic structure, and household location are important determinants of heating fuel choices. These findings suggest that future energy policy should pay more attention to its combined effect on both cost and quality of heating system as well as its different regional effects. Besides, enhancing household socio-economic status should also be attached importance in policy design.

The Association Between Overcrowded Households and Adolescent Pregnancy

Literature suggests that socioeconomic environment influences the probility of teenage pregnancy. Specifically, this paper reports an investigation about the association between overcrowded households and teenage pregnancy in a developing country. Using a probit model with sample selection correction, which provides a robust estimate, we analyze data of 10,245 Ecuadorian women between 12 and 19 years old. Data was collected from official surveys conducted in 2006 and 2014. Empirical evidence shows that an overcrowded household is an influential factor on the probability of adolescent pregnancy in Ecuador. The involvement of policy makers, parents, schools, and NGOs is necessary to design preventive and corrective policies regarding adolescent pregnancy. In particular, governments should implement policies that ensure the rights of adolescents to access mechanisms for preventing and reporting sexual violence whether it comes from their households or anywhere else.

Inequality in tax evasion: the case of the Spanish income tax

Purpose The purpose of this paper is to estimate tax evasion and its impact on progressivity, redistribution and the measurement of inequality, using microdata from the Spanish income tax for 2001-2004. Design/methodology/approach The approach follows Feldman and Slemrod (2007) by exploiting the relation of charitable donations with the composition of income but introduces two methodological innovations, which could be useful for further studies: correction for sample selection with a Heckman two-step setting and the calculation of different evasion rates for top incomes with an interaction term. Findings Evasion in capital incomes was significant throughout these years. Financial incomes were reported at around 50-70 per cent of their real value, with the lowest estimates corresponding to the top decile. Revenues from fixed capital display similarly low compliance rates for the top 10 per cent. Tax evasion in self-employment incomes (direct assessment) is estimated at 20 per cent for 2001. Mostly because of a composition effect, this means that fraud was higher at the top of the income distribution, thus having a regressive impact. Inequality statistics and top income concentration estimates should, therefore, be revised upwards. Originality/value This is the first paper to estimate the distributive impacts of tax evasion in Spain, and one of very few internationally.

Inferring Zambia’s HIV prevalence from a selected sample

ABSTRACT This paper estimates HIV prevalence in Zambia from survey data that are subject to sample selection: some surveyed individuals do not consent to take an HIV test. We introduce semiparametric estimators that incorporate recent developments in machine learning. The semiparametric estimators perform well in Monte Carlo experiments and obtain narrower confidence intervals than a fully parametric estimator when the model is misspecified. Our semiparametric estimates of the HIV rate are roughly equal to the rate in the selected sample. In contrast, recent parametric estimates find a higher rate – implying that some form of sample-selection correction is warranted. Further, parametric estimates find a 14% decline in Zambia’s HIV prevalence from 2007 to 2014, whereas semiparametric estimates find that the HIV rate was relatively stable over this time period.

Adoption of E-Commerce by Individuals and Digital Divide: Evidence from Spain

E-commerce penetration rates are distant among those groups of individuals with the lowest and the highest levels of online shopping adoption. This is an indicator of digital divide, having negative effects in terms of untapped opportunities for people, companies and the whole economy. Key socioeconomic and demographic determinants of adoption of e-commerce are explored, analyzing a dataset of 174,776 observations for the period 2008-2017 in Spain. The empirical analysis is based on a standard neoclassical utility maximization framework. Linear probability model, logistic regression, and Heckman’s sample selection correction model have been used. The results suggest that e-commerce adoption is positively related with being male, having higher levels of education, income and digital skills, being Spanish, and being employed; while being female, older and belonging to a household of two or more members have negative effects. An interaction between digital skills and age has been introduced in the model, where high digital skills seem to have a positive influence, partly counteracting the lower odds for some age groups. Policy recommendations related to demand and supply measures are suggested to foster the adoption of e-commerce.

Monte Carlo Evidence on the Estimation Method for Industry Dynamics

Abstract This study presents a structural estimation method for nonlinear stochastic dynamic models of heterogeneous firms. I perform a Monte Carlo experiment to evaluate the performance of the estimators for the AR(1) dynamic panel data subject to sample selection without exogenous regressors. The results suggest a strong need to correct the sample selection and that the proposed structural estimation method works well. These results are important for practical situations where the assumptions of the standard sample selection correction methods are not satisfied.

The Link Between Role Identity Aspirations And Opportunity Exploration And Exploitation

This study puts forward a nascent entrepreneur’s different role identity aspirations as drivers of specific new venture creation activities. In particular, we argue that nascent entrepreneurs’ inventor, founder, and developer role identity aspirations are linked to the degree to which they undertake opportunity exploration and opportunity exploitation activities to start their venture. A Poisson model with sample selection correction, performed on a random sample of 716 inhabitants of Flanders, the Dutch-speaking part of Belgium, largely confirms our hypotheses. As a result, this study has important implications for the literature on new venture creation, the literature on entrepreneurial opportunities and the identity literature, as well as for nascent entrepreneurs wanting to start a venture.

Efficiency of correction for sample selection in QUAIDS models: an example for the fish demand in Germany

This paper investigates the performance of the Quadratic Almost Ideal Demand System (QUAIDS) with and without accounting for the consistent two-step estimation for censoring of the dependent variable as well as the use of quality-adjusted prices. The demand systems are estimated for a commodity that has a relatively low consumption level in the German market, fish. Thus, the zero observations are quite high and the need of a correction procedure is more urgent. The results show differences in the applied methods. It turns out that ignoring consistent two-step estimation and quality-adjusted prices leads to less elastic demand estimates.