Bivariate Probit Estimation Research Articles

Choice of dialysis treatment and type of medical unit (private vs public): application of a recursive bivariate probit

ESRD patients have to deal with two choices: the first is related to the dialysis modality; the second concerns the type of dialysis unit (public vs private) where to undertake the treatment. Such a choice is related to unobservable factors, among which there might be patients' clinical factors as well as factors related to the characteristics of each unit. We employ a recursive bivariate probit estimation on a sample of ESRD Sicilian patients in order to evaluate the impact of these factors. Results can have important implications for Sicily in order to organize dialysis services: here, in fact, the number of private centres is higher than in other Italian Regions.

Genetic and environmental contributions to back pain in old age: a study of 2,108 danish twins aged 70 and older.

Self-reported 1-month prevalence of back pain in older twins assessed at intake in a population-based longitudinal survey. To determine the relative contribution of genetic and environmental factors to back pain in old age. To date, genetic contributions to back pain in old age have not been assessed, to the authors' best knowledge. Interview data given at entry into a nationwide cohort-sequential population-based survey of Danish twins aged 70 years and older in 1995, 1997, 1999, and 2001 form the basis of this analysis. Analysis of twin similarity was estimated using probandwise concordance rates, odds ratios, and tetrachoric correlations for back pain. Heritability (proportion of the population variance attributable to genetic variation) was estimated by bivariate probit estimation and adjusted for known significant environmental factors. Odds ratios for known environmental effects were estimated after controlling for age, sex, and genetic effects. Modest and nonsignificant differences between monozygotic and dizygotic twin pairs were found for probandwise concordance rates, odds ratios, and tet-rachoric correlations for both men and women. In the bivariate probit estimation, a current or previous diagnosis of osteoporosis, degenerative joint disease, or lumbar disc prolapse was found to significantly affect the risk of back pain. Additive genetic effects explained approximately one fourth of the liability to report back pain in men and none of the occurrence in women. Individual environmental effects were found to explain roughly 75% of the occurrence of back pain in men and 100% in women. Additive genetic effects are modest contributors to back pain in older men but not in women. A current or previous medical diagnosis of osteoporosis, degenerative joint disease, or lumbar disc prolapse is-strongly associated with back pain, also when genetic factors are controlled for. Because of inherent methodologic issues, this estimate of the genetic influence on back pain in old age is probably conservative.

Are Married Women in Turkey More Likely to Become Added or Discouraged Workers?

Abstract. The purpose of this study is to examine the labour market outcomes of married couples to find out which of the added and discouraged worker effects is dominant in urban Turkish families. Using household labour force survey and province‐level data from 1988 and 1994, we look for evidence regarding the dependency of the labour force participation decisions of wives and the employment status of their husbands. On yearly and pooled samples of married couples in their prime ages, bivariate probit estimates indicate that the two decisions are negatively correlated. However, the correlation is found to be statistically significant only in the economic crisis year of 1994. We interpret this finding to mean that the added worker effect dominates the discouraged worker effect.

The impact of National Health Insurance on neonatal care use and childhood vaccination in Taiwan.

The objective of this study was to assess the impact of National Health Insurance (NHI) on the utilization of neonatal care and childhood vaccination in Taiwan. Data are selected from two nationwide maternal and infant surveys undertaken in 1989 and 1996, which were funded by the Department of Health. The questionnaire was administered in all 23 administrative districts, including two metropolitan areas - the cities of Taipei and Kaohsiung. The first and second cohort consisted of 1641 and 3499 infants, respectively. This study used a bivariate probit estimation procedure to examine the factors that determine the probability of neonatal care use and vaccination by pooling the 2 years. Generally, the mothers who are older, more educated and more satisfied with birth delivery services were found to be more likely to use neonatal preventive care. The likelihood of receiving such care also tends to rise with advancing gestational age and higher probability of neonatal complication (icterus neonatorum) and being born in a hospital. The likelihood of care was also found to vary regionally with northern neonates having higher odds of receiving preventive care than non-northern neonates. In spite of having an insignificant impact on the use of care, NHI does lessen the inequality in use of these two services in various regions. However, regional variations in neonatal care use still exist. Residents of the southern area remain less likely to receive neonatal care than those of the northern area after NHI. This finding deserves serious consideration when attempting to design effective policies, such as expanding medical institutions in the aboriginal southern areas so as to increase the accessibility of such health care.

Characteristics of product and process innovators: some evidence from the Finnish innovation survey

The characteristics of product and process innovators among Finnish manufacturing firms with Community Innovation Survey (CIS) data are studied. Results of bivariate probit estimations suggest that, while the product and process innovations are related, they are largely driven by different factors. The ability to benefit from inward spillovers is the only variable having a symmetric effect onboth types of innovation. Cooperation with non-academic outside partners is the only other variable that becomes significant in both equations. Process innovations benefit from capital embodied technology, whereas product innovations require disembodied forms of technology.

Determinants of occupational change and mobility in rural India

With a view to redress the dearth of economic analysis on occupational dynamics in rural household economies, this study examines the nature and characteristics of occupational change and mobility in rural India. Using the unique WIDER data-set from six villages in West Bengal, India the factors causing a change of primary occupation among those who tried to change occupation over a twelve-month recall period has been analysed. Bivariate probit estimates jointly determining the probability of trying to change occupation and having been successful suggest that success in changing occupation depends crucially on socially constructed ‘status’- being older, male, from larger farming families or having higher schooling experience. They also exemplify the effects of regional diversity, levels of prosperity and different patterns of employment between agricultural and non-agricultural activities.

The institutional environment for multinational investment

This article posits that the effect of political hazards on the choice of market entry mode varies across multinational firms based on the extent to which they face expropriation hazards from their potential joint-venture partners in the host country (the level of contractual hazards). As political hazards increase, the multinational faces an increasing threat of opportunistic expropriation by the government. Partnering with host-country firms that possess a comparative advantage in interactions with the host-country government can safeguard against this hazard. However, as contractual hazards increase, the potential benefit to the joint-venture partner of manipulating the political system for it's own benefit at the expense of the multinational increases as well, thereby diminishing the hazard-mitigating benefit of forming a joint venture. A two-stage bivariate probit estimation technique is used to test these hypotheses on a sample of 3,389 overseas manufacturing operations by 461 firms in 112 countries. Copyright 2000 by Oxford University Press.

Measuring equity in access to health care

This article develops and uses methodologies to: (1) measure equity in the distribution of access to health services; and (2) measure the impact of health insurance programs on equity. The article proposes two egalitarian-based indicators for measuring equity in terms of access to health care — a concentration coefficient derived from the Gini coefficient, and the Atkinson distributional measure — and also employs a weighted Utilitarian social welfare function to measure overall levels of access. The article defines access as the use of health care by individuals with a need for care; need is measured as self-reported morbidity. The setting for the empirical application is the country of Ecuador. The Ecuador Social Security Institute runs a General Health Insurance (GHI) program, whose affiliates are primarily workers in the formal sector of the economy. The principal data source is the 1995 Ecuador Living Standards Measurement Survey. The study uses a microeconomic health care demand model and bivariate probit estimation techniques to measure the impact of insurance on health service use for each quintile of adjusted per-capita household expenditure. The study also predicts health care use and program impact for each quintile under a series of simulation scenarios corresponding to proposed expansion of eligibility for the GHI program. The GHI program increases overall access to health care, but has a negative impact on equity in the distribution of health services. The benefits of the program, calculated as its marginal impact on the probability of using of health care, have a strongly regressive distribution. Expanding eligibility to the self-employed makes the benefit more equitably distributed (but still inequitable), and increases overall social welfare considerably. Expanding eligibility to the dependents of the insured person has similar effects, although less important in magnitude.

Measuring the impact of health insurance with a correction for selection bias--a case study of Ecuador.

This article develops and uses methodologies to evaluate the impact of publicly-financed health insurance programmes on the use of health care. Using univariate and bivariate probit estimation techniques, the study tests and corrects for endogeneity resulting from selection bias. Potential endogeneity arises from the choice to be insured, eligibility for insurance, and differences in individuals' health status. The setting for the study is the country of Ecuador. The General Health Insurance (GHI) programme, which primarily covers workers in the formal sector of the economy, is found to have a strong positive association with the use of curative health care after correcting for selection bias, but no significant effect on the use of preventive care. Individuals with severe illnesses who are eligible for GHI have a preference for private health care, and self-select out of the GHI programme. The Seguro Campesino Social (SSC) programme, directed at farming populations, has positive but insignificant associations with both curative and preventive care.

Economic determinants of IMF financial arrangements

We analyze factors that lead to IMF approval of financial arrangements. We account for both economic variables that induce a country to seek an IMF arrangement (‘demand-side’ factors) and macroeconomic policy commitments that the IMF considers when deciding whether to approve it (‘supply-side’ factors). Using a pooled sample of annual observations for 91 developing countries over 1973–1991, we obtain maximum likelihood bivariate and univariate probit estimates of the approval of an IMF arrangement for a given country in a given year. A number of demand-side and supply-side variables are statistically significant determinants of the approval of an IMF arrangement.

Welfare Participation and Labour Market Behaviour in Canada

In this paper, bivariate probit estimation procedures and nonindependent selectivity adjustments are employed to investigate the determinants of the joint welfare participation-labor supply decisions made by single males, single females, lone fathers, and lone mothers, using data from the 1988-89 Labour Market Activity Survey. The bivariate probit results indicate that the welfare and labor-force participation decisions are not independent. Welfare program parameters, however, such as the basic allowance and the program tax rate, generally do not appear to influence the welfare participation, labor force participation, or hours worked decisions. Rather, these decisions are more strongly associated with personal characteristics and labor market conditions.

The impact of wages on job search in a transition economy - evidence with data from Eastern Germany

This study is concerned with the impact of wages on job search decisions in Eastern Germany immediately after reunification and in later periods. Several concepts measuring wage effects are used. The results of bivariate probit estimates with incomplete classification for a global and local job search show that the wage level had stronger effects on the former than on the latter job search decision directly after the collapse of the Berlin Wall. The degree of satisfaction with current income and the earnings capacities were irrelevant for the search decision. Low-paid East German workers had a higher propensity to search for a new job in the old Lander than others. This is contrary to the misleading result of isolated estimates about the job search in Western Germany. In the subsequent periods the situation has changed. The difference of earnings capacities between Eastern and Western Germany, specific East German wage premiums and intraregional wage inequality affect the decision to search for a new job in the west.

Testing for state and local government job queues

Bivariate probit estimates with partial observability identify the factors influencing the worker’s decision to seek a state or local government job and the factors influencing the government’s decision to hire particular workers. The estimates routinely confirm the presence of a local job queue but confirm the presence of a state job queue only when the role of unionization is explicitly recognized. Our results suggest that the net benefits from local and state government employment are greater than those from private sector employment. The estimates also indicate that local governments disproportionately hire racial minorities relative to the local governments’ pool of applicants but indicate no such relationship for state governments.

The Substitution between Moves and Quits

If a worker's workplace and residence locations do not maximize utility, quits and moves are substitutes; a quit or a move alone is more likely than both together. Records of a single company confirm this. Simultaneous bivariate probit estimates of move and quit behavior demonstrate that uncontrolled shocks to quits and moves are negatively correlated. Furthermore, when the company relocated from the central business district to a suburb of its metropolitan area, the occurrence of either a quit or a move virtually precluded the occurrence of the other among employees whose journeys-to-work were lengthened by the relocation. Copyright 1991 by Royal Economic Society.