Two-stage Probit Model Research Articles

Identifying the reciprocal causation between hit-and-run behavior and crash injury severity

Objective Hit-and-run behavior is believed to exacerbate the injury severity of traffic crashes due to the delayed emergency response for the victims. However, several previous studies indicated the opposite finding that hit-and-run crashes were associated with less severe injuries. The relevant studies mainly identified the statistical associations between hit-and-run behavior and injury severity without revealing causation between them. To this end, the study aims to explore the reciprocal causation between the two variables. Method The two-stage probit model with endogenous regressors is employed to identify the reciprocal causation between hit-and-run behavior and crash injury severity for single- and two-vehicle crashes, respectively, with the use of crash data extracted from the Crash Report Sampling System and Fatality Analysis Reporting System (2016-2019). Results The results indicate that 1) for both single- and two-vehicle crashes, the fleeing behavior can significantly increase the injury severity of the victims in the crashes while the severe injury of the victims has a negative impact on the propensity of such behavior, 2) the propensity of hit-and-run behavior is influenced by various instrumental variables such as driver age, gender, alcohol involvement, weekday, area type, and light condition, and 3) crash injury severity is significantly related to the victim age, gender, and vehicle damage. Conclusions There is a reciprocal causation between hit-and-run behavior and injury severity in traffic crashes. The analytical results can provide a reasonable explanation for the counterintuitive finding on hit-an-run crashes and help mitigate the injury severity.

Increase of catastrophic and impoverishing health expenditures in Mexico associated to policy changes and the COVID-19 pandemic.

In 2003, the Mexican Congress approved a major reform to provide health care services to the poor population through the public insurance scheme Seguro Popular. This program was dismantled in 2019 as part of a set of health system reforms and substituted with the Health Institute for Welfare (INSABI). These changes were implemented during the initial phases of the coronavirus (COVID-19) pandemic. We aimed to examine the impact of these reforms and the COVID-19 pandemic on financial risk protection in Mexico between 2018 and 2020. We performed a population-based analysis using cross-sectional data from the 2018 and 2020 rounds of the National Household Income and Expenditures Survey. We used a pooled fixed-effects multivariable two-stage probit model to determine the likelihood of catastrophic health expenditure (CHE), impoverishing health expenditure (IHE), and excessive health expenditure (EHE) among Mexican households. We also mapped the quintiles of changes in EHE in households without health insurance by state. The percentage of households without health insurance almost doubled from 8.8% (three million households) in 2018 to 16.5% (5.8 million households) in 2020. We also found large increases in the proportion of households incurring in CHE (18.4%; 95% confidence interval (CI) = 6.1, 30.7) and EHE (18.7%; 95% CI = 7.9, 29.5). Significant increases in CHE, IHE, and EHE were only observed among households without health insurance (CHE: 90.7%; 95% CI = 31.6, 149.7, EHE: 73.5%; 95% CI = 25.3, 121.8). Virtually all Mexican states (n/N = 31/32) registered an increase in EHE among households without health insurance. This increase has a systematic territorial component affecting mostly central and southern states (range = -1.0% to 194.4%). The discontinuation of the Seguro Popular Program and its substitution with INSABI during the first stages of the COVID-19 pandemic reduced the levels of health care coverage in Mexico. This reduction and the pandemic increased out-of-pocket expenditure in health and the portion of CHE and EHE in the 2018-2020 period. The effect was higher in households without health insurance and households in central and southern states of the country. Further studies are needed to determine the specific effect both of recent policy changes and of the COVID-19 pandemic on the levels of financial protection in health in Mexico.

Inequitable Financial Protection in Health for Indigenous Populations: the Mexican Case.

There is an important gap in the literature concerning the level, inequality, and evolution of financial protection for indigenous (IH) and non-indigenous (NIH) households in low- and middle-income countries. This paper offers an assessment of the level, socioeconomic inequality and middle-term trends of catastrophic (CHE), impoverishing (IHE), and excessive (EHE) health expenditures in Mexican IHs and NIHs during the period 2008-2020. We conducted a pooled cross-sectional analysis using the last seven waves of the National Household Income and Expenditure Survey (n = 315,829 households). We assessed socioeconomic inequality in CHE, IHE, and EHE by estimating their Wagstaff concentration indices according to indigenous status. We adjusted the CHE, IHE, and EHE by estimating a maximum-likelihood two-stage probit model with robust standard errors. We observed that, during the period analyzed, CHE, IHE, and EHE were concentrated in the poorest IHs. CHE decreased from 5.4% vs. 4.7% in 2008 to 3.4% vs. 2.9% in 2014 in IHs and NIHs, respectively, and converged at 2008 levels towards 2020. IHE remained unchanged from 2008 to 2014 (1.6% for IHs vs. 1.0% for NIHs) and increased by 40% in IHs and NIHs during 2016-2020. EHE plunged in 2014 (4.6% in IHs vs. 3.8% in NIHs), then rose, and remained unchanged during 2016-2020 (6.7% in IHs and 5.6% in NIHs). In pursuit of universal health coverage, health authorities should formulate and implement effective financial protection mechanisms to address structural inequalities, especially forms of discrimination including racialization, that vulnerable social groups such as indigenous peoples have systematically faced. Doing so would contribute to closing the persistent ethnic gaps in health.

Is there an association between intimate partner violence and the prevalence of cervical cancer screening in Jordan?

Major health inequalities exist surrounding the utilisation of cervical cancer screening services globally. Jordan, a low- and middle-income country, has poor screening rates (15.8%), with barriers to accessing services, including lack of education. Emerging studies demonstrate that intimate partner violence (IPV) impacts reproductive health decisions. As a large proportion of Jordanian women have reported experiencing IPV, this study examines the association between IPV and cervical cancer screening in Jordan, the first of its kind using national-level data. Using Jordan's Demographic Health Survey 2017-18, cervical cancer screening awareness and self-reported screening were estimated in participants who answered questions on IPV (n = 6679). After applying sample weights, Heckman's two-stage probit model determined the association of awareness and utilisation of cervical cancer screening with experience of IPV, adjusting for the socio-economic factors. Of the women with privacy to answer the IPV module, 180 (3.4%) were found to be victims of sexual violence, 691 of physical violence (12.6%) and 935 (16.2%) of emotional violence. Women subjected to sexual violence were less likely to admit to having awareness of a Pap smear test; however, this did not impact screening rates. Victims of emotional violence were more likely to be screened than non-victims. No association between physical violence and cervical cancer screening was found. A significant association between cervical screening awareness and IPV demonstrates that cancer screening policies must consider IPV among women to improve screening awareness. The paper further sheds light on the paradoxical association between emotional violence and screening. It is acknowledged this situation may be far worse than reported, as women without autonomy were unlikely to answer IPV questions that may endanger them-targeted surveys on cervical cancer screening warrant further investigation.

A Gender-Based and Quasi-Experimental Study of the Catastrophic and Impoverishing Health-Care Expenditures in Mexican Households with Elderly Members, 2000-2020

ABSTRACT Latin America has experienced a rise in noncommunicable diseases (NCDs) which is having repercussions on the structuring of healthcare delivery and social protection for vulnerable populations. We examined catastrophic (CHE) and excessive (EHE, impoverishing and/or catastrophic) health care expenditures in Mexican households with and without elderly members (≥65 years), by gender of head of the households, during 2000–2020. We analyzed pooled cross-sectional data for 380,509 households from eleven rounds of the National Household Income and Expenditure Survey. Male- and female-headed households (MHHs and FHHs) were matched using propensity scores to control for gender bias in systematic differences regarding care-seeking (demand for healthcare) preferences. Adjusted probabilities of positive health expenditures, CHE and EHE were estimated using probit and two-stage probit models, respectively. Quintiles of EHE by state among FHHs with elderly members were also mapped. CHE and EHE were greater among FHHs than among MHHs (4.7% vs 3.9% and 5.5% vs 4.6%), and greater in FHHs with elderly members (5.8% vs 4.9% and 6.9% vs 5.8%). EHE in FHHs with elderly members varied geographically from 3.9% to 9.1%, being greater in less developed eastern, north-central and southeastern states. Compared with MHHs, FHHs face greater risks of CHE and EHE. This vulnerability is exacerbated in FHHs with elderly members, because of gender intersectional vulnerability. The present context, marked by a growing burden of NCDs and inequities amplified by COVID-19, makes key interlinkages across multiple Sustainable Development Goals (SDGs) apparent, and calls for urgent measures that strengthen social protection in health.

Differential Probability in Unmet Healthcare Needs Among Migrants in Four European Countries

Migrants and refugees try to reach Europe to seek protection and a better life. The responsiveness and stewardship of the European countries health system have an impact on the ability to access healthcare. This study aims to investigate the differential probability of healthcare unmet needs among migrants living in four European countries. We used a 2019 cross-sectional data from the European Union Income and Living Conditions survey. We performed a two-stage probit model with sample selection, first to identify the respondents with need for care, then those who need it but have not received it. We analysed reasons for unmet needs through accessibility, availability and acceptability. We then performed country studies assessing the national health systems, financing mechanisms and migration policies. Bringing together data on financial hardship and unmet needs reveals that migrants living in Europe have a higher risk of facing unmet healthcare needs compared to native citizens, and affordability of care remains a substantial barrier. Our results showed the country heterogeneity in the differential migrants’ unmet needs according to the place where they live, and this disparity seems attributed to the health system and policies applied. Given the diversity of socioeconomic conditions throughout the European countries, the health of migrants depends to a large degree on the integration and health policies in place. We believe that EU policies should apply further efforts to respect core health and protection ethics and to acknowledge, among others, principles of ‘do-no-harm’, equity and the right to health.

Does exposure to health information through mobile phones increase immunisation knowledge, completeness and timeliness in rural India?

IntroductionImmunisation plays a vital role in reducing child mortality and morbidity against preventable diseases. As part of a randomised controlled trial in rural Madhya Pradesh, India to assess the impact...

Family background and curricular tracking among Taiwanese high school students

Prior research studying the relationship between family background and curricular tracking estimates conditional logistic regressions without considering the sample selection, and finds small or non-robust effects. Given that sample selection bias may underestimate of the effects of family background on later schooling, we analyze Taiwan Education Panel Survey (TEPS) and compare models with and without adjusting the selection bias. We find that the effects of family background on curricular tracking become stronger in the Heckman two-stage probit model for selection bias than in the simple logistic regression. These results imply family background may play a more important role in students’ curricular choice than documented in the literature.

Does quality of electricity matter? Household-level evidence from the Philippines

The Philippines is a country that faces many development challenges, including providing reliable and good quality electricity. Access to good quality electricity connection matters because it affects many aspects that increase productivity (e.g. education, health and business) and can make lives easier and more comfortable. While access has improved over time, many households still suffer from poor quality of their connection brought about by incidents of power outages, fluctuations and low voltage. This article attempts to examine whether better electricity quality improves household welfare. We apply a two-stage probit-ordered probit model to overcome endogeneity caused by reverse causation between electricity quality and household income. We find that households experiencing better electricity quality decrease the probability of remaining in the lowest income category by 23%.

Wealth and Health in South Africa

In this paper, we investigate the impact of wealth on health in South Africa using the National Income Dynamics Study (NIDS). We estimate a two-stage probit model with inheritance as an instrumental variable for wealth. We find no significant effect of wealth on health at the individual level, consistent with most of the results found for developed countries. Alternative specifications to the health outcomes (self-reported health versus reported diseases) as well as the introduction of gifts as an additional instrumental variable delivers similar results. In addition, we decompose wealth into liquid and illiquid wealth. Despite the health effect being higher for liquid than for non-liquid wealth, none of these measures involve substantial or significant effects on health.

Investigating consumer preferences and willingness to pay for Orange-fleshed Sweet potato (OFSP) juice in Rwanda

Abstract While micronutrient malnutrition has become a major public health problem in developing countries, consumption of biofortified foods such as orange fleshed sweet potato has proved to be cost effective means of addressing vitamin A deficiency. This paper describes consumer preferences and willingness to pay for Orange fleshed Sweet potato (OFSP) juice based on a structured survey and taste tests administered to 946 participants (384 female and 562 male) chosen at random in Rwanda. Four juices types were tested: two popular brands of 100% pineapple juice, one 100%-OFSP juice, and one 80% OFSP- 20% pineapple juice blend. Consumers ranked different attributes of each juice such as aroma, taste, color, “right” amount of sugar, and aftertaste using a Likert scale (1 to 5). Heckman two-stage probit model was used to analyze willingness to pay and a multinomial logit model was used to analyze the determinants of juice choice. Gender of the consumer, juice buying frequency, aroma, the taste of the juice, and vitamin A knowledge were positively associated with willingness to pay and juice choice. Without nutritional information on OFSP juice, the willingness to pay for the standard juices were statistically higher; but with nutritional information, the willingness to pay and choice of OFSP juice was significantly improved. It is concluded that nutrition information is an important factor influencing the acceptance and willingness to pay for OFSP juice.

Make or buy: the case of harvesting mechanization in coffee crop in Brazil

This paper investigates the make-or-buy choice on the harvesting mechanization transaction in Brazilian coffee production. It empirically tests this organization choice following the transaction attributes of asset specificity, uncertainty and path dependence. Using a two-stage probit model conditioned by credit availability in a sample of 105 coffee growers in Brazil, our study evidences that the governance decision follows the efficient alignment argument of the Transaction Cost Economics. The research contributes to the understanding of institutional arrangements in place at coffee grower perspective. An additional descriptive field investigation presented the existence of three different types of arrangement as a solution for coffee harvesting mechanization: total vertical integration (make), outsourcing contracts (buy) and tapped vertical integration. Those results demonstrate a shift from a spot transaction pattern to more sophisticated contractual tools. We present managerial implications by illustrating the determinants of each decision, either make or buy, in the harvesting mechanization transaction in coffee production which can provide support to efficient strategies elaboration by farmers or service providers.

Pharmaceutical Start-ups’ Technology and Financing Strategy

It is necessary for pharmaceutical start-ups to access external financing sources, such as angel, independent venture capital (IVC), and corporate venture capital (CVC), in order to proceed their drug development processes. Given that technology is often the most important resource for pharmaceutical start-ups, it is important to understand how pharmaceutical start-ups’ technological characteristics serve to determine their financing sources. Using a dataset that contains 1730 fundraising activities of 529 pharmaceutical start-ups, this study provides systematic evidence that technological progress and uncertainty significantly impact start-ups’ decisions on the type of funding source. These findings are supported by several empirical approaches, including multinomial logit, sequential logit, and two-stage probit models. Along with the complexity and length of product development cycle in the pharmaceutical sector, these findings suggest that start-ups should strategically arrange financing sources as the drug development processes are proceeded.

The effect of early childhood malnutrition on child labor and schooling in rural ethiopia

The long term effects of malnutrition in early childhood on the child’s physical stature have been widely investigated over the last few decades. A number of empirical studies show that malnutrition shocks in early childhood lead to long-lasting deficiencies in the physical fitness of the person, especially in terms of height. Other studies also show that better physical stature is strongly associated with higher cognitive ability and better educational attainment. However, previous studies have paid little attention to the role physical fitness may play in influencing the choice parents make in selecting their child for child labor activities and how this association may influence the accuracy of the estimated relationships between schooling and physical stature. This paper examines how physical stature of a child measured in terms of age standardized height influences his/her selection for family labor activities versus schooling in rural Ethiopia. Two-stage bivariate probit models are estimated for child labor and schooling using malnutrition caused by exposure to weather shocks in early childhood as a source of identification for the child’s physical stature. The first stage results show that exposure to significant weather shocks during the first three years of the child’s life generally have a lasting negative effect on his/her age standardized heights measured later in life. The effect is particularly strong when the child is exposed to the weather shock during his/her second year of life. Estimation results also indicate that physically more robust children are more likely to combine child labor and schooling than those children who are physically less fit. On the other hand, I find no evidence that better physical stature of the child leads to higher probability of his/her selection for fulltime child labor activities. The findings are consistent across two different cohorts of children and two different identification strategies. Estimation results for both the older and younger cohort of children in my sample show that better access to early childhood nutrition can improve the child’s prospects for attending schooling but may also put the child in additional pressure to participate in family labor activities which may be reflected in poor performance in schooling. These findings imply that policies intended to promote schooling through nutrition support programs could be more effective if they are accompanied by programs (such as income support schemes in the form of conditional cash transfers) that could mitigate the parents’ need for child labor.

A Two-Stage Probit Model for Predicting Recovery Rates

We propose a two-stage probit model (TPM) to predict recovery rates. By the ordinal nature of the three categories of recovery rates: total loss, total recovery, and lying between the two extremes, we first use the ordered probit model to predict the category that a given debt belongs to among the three ones. Then, for the debt that is classified as lying between the two extremes, we use the probit transformation regression to predict its recovery rate. We use real data sets to support TPM. Our empirical results show that macroeconomic-, debt-, firm-, and industry-specific variables are all important in determining recovery rates. Using an expanding rolling window approach, our empirical results confirm that TPM has better and more robust out-of-sample performance than its alternatives, in the sense of yielding more accurate predicted recovery rates.

Mental Health Service Needs in the Prison Boom: The Case of Children of Incarcerated Mothers

This study identifies the factors related to mental health service use among children of incarcerated mothers. Data on 700 children collected from a diverse sample of mothers in Arizona are used, and a two-stage probit model with sample selection is estimated to assess the various child, mother, and caregiver characteristics associated with children’s use of mental health services. Results indicate that children involved in child protective services (CPS) and children cared for by grandparents are more likely to receive mental health services, whereas children of Native American mothers and children who have been exposed to violence are less likely to receive services for their mental health needs. These findings have important implications for correctional policy regarding the intake screening of female inmates and suggest that criminal justice agencies communicate more closely with CPS and community-based services to ensure children’s mental health needs are addressed while their mothers are in prison.

Patterns of earnings differentials across three conservative European welfare regimes with alternative education systems

The aim of this paper is to investigate, from a generational perspective, the effect of human capital on individual earnings and earnings differences in Germany, France and Italy, three developed countries in Western Europe with similar conservative welfare regimes but with important differences in their education systems. Income inequalities between and within education levels are explored using a two-stage probit model with quantile regressions in the second stage. More precisely, drawing upon 2005 EU-SILC data, returns on schooling and experience are estimated separately for employees and self-employed full-time workers by means of Mincerian earnings equations with sample selection; the sample selection correction accounts for the potential individual self-selection into the two labour force types. Although some determinants appear to be relatively similar across countries, state-specific differentials are drawn in light of the institutional features of each national context. The study reveals how each dimension of human capital differently affects individuals’ earnings and earnings inequality and, most of all, how their impacts differ along the conditional earnings distribution and across countries. In the comparative perspective, the country's leading position in terms of the highest rewards on education also depends on which earnings distribution (employee vs. self-employed) is analysed.

International Migration and School Enrollment of the Left-Behinds in Albania: A Note

We study the impact of international migration on enrollment of the left-behinds in Albania using the 2008 and 2012 Living Standard Measurement Survey. We employ a two-stage Probit model controlling for endogeneity with an instrumental-variable approach. We find that migration has a negative and significant impact on enrollment in both waves. This result holds across different sub-samples and alternative ways to proxy within-household migration. Our findings confirm similar results for 2005 and cast doubts on the effectiveness of recent educational reforms and migration-oriented policies in Albania.

Risky remote purchasing and identity theft victimization among older Internet users

The low self-control/risky lifestyles perspective posits that people deficient in self-control engage in certain risky behaviors that increase their exposure to motivated offenders in the absence of capable guardianship, which in turn elevates their risk of victimization. Using survey data from telephone interviews conducted in Florida and Arizona with individuals aged 60 and over, the current study tests whether this theoretical framework partially explains risky remote purchasing and identity theft victimization among older Internet users. Results from the two-stage probit models conform to expectations: Individuals with lower levels of self-control have a significantly higher probability of making a purchase after receiving an unsolicited email from a vendor with whom they have not previously done business. What is more, making a risky remote purchase significantly increases the probability of identity theft victimization. The findings not only speak to the generality of the low self-control/risky lifestyles perspective, but also indicate that older Internet users can reduce their victimization risk by taking specific precautions.

Relocation and Social Support Among Older Adults in Rural China.

China's economic reforms have led to millions of citizens being relocated to support infrastructure development, reduce poverty, and address ecological, disaster-related and other concerns. This study expands on previous research on the implications of relocation in China by examining the impact of rural elders' relocation on the perceived availability of emotional, instrumental, and financial support. Data were drawn from the Ankang Study of Aging and Health conducted with a representative sample of 1,062 rural residents aged 60 and over living in Ankang, China. Two-stage probit and least squares regression models assessed the impact of relocation on familial and nonfamilial emotional, instrumental, and financial support. Relocation was negatively associated with the number of social support resources that older adults perceived as being available. Although this was the case with regard to both familial and nonfamilial support, it was particularly evident with regard to family support and, within families, with regard to instrumental rather than financial or emotional support. Relocation has negative implications for the number of social support resources perceived to be available by older adults in rural China. China will need to come to terms with how to provide for the instrumental support needs of an aging society.