Maximum Likelihood Probit Research Articles

Exploring the edges of clientelism: ideology, ethnicity, and partisanship in Belize

ABSTRACT Studies focusing on Belize reveal that clientelism is the glue that makes the party system stick. We explore whether ideology and ethnicity account for partisanship within that deeply clientelist setting. We use four AmericasBarometer survey waves by the Latin American Public Opinion Project (LAPOP) to study the individual-level correlates of partisan attachments with the People’s United Party (PUP) and United Democratic Party (UDP). Using binary and multinomial probit maximum likelihood estimators, we find that ethnicity correlates to partisanship with the PUP, and ideology correlates to partisanship with the UDP. The robustness checks, which focus on declared vote choices in previous national elections, corroborate our findings. The Belizean case shows how clientelist parties can develop alternative support sources and that these can vary substantially.

Stein-type control function maximum likelihood estimator for the probit model in the presence of endogeneity

A Stein-type control function maximum likelihood (CFML) estimator is suggested for the probit model in the presence of endogeneity. This novel estimator combines the probit maximum likelihood and CFML estimators. The asymptotic distribution and risk function for the new estimator is derived. It is demonstrated that, subject to certain conditions of the shrinkage parameter, the asymptotic risk of the new estimator is strictly smaller than the risk of the CFML. Monte Carlo simulations illustrate the method's superiority in finite samples. The method is also applied to analyze the impact of managerial incentives on the use of foreign-exchange derivatives.

Binary outcomes, OLS, 2SLS and IV probit

For a binary outcome Y, generated by a simple threshold crossing model with a single exogenous normally distributed explanatory variable X, the OLS estimator of the coefficient on X in a linear probability model is a consistent estimator of the average partial effect of X. Even in this very simple setting, we show that when allowing for X to be endogenously determined, the 2SLS estimator, using a normally distributed instrumental variable Z, does not identify the same causal parameter. It instead estimates the average partial effect of Z, scaled by the coefficient on Z in the linear first-stage model for X, denoted γ 1, or equivalently, it estimates the average partial effect of the population predicted value of X, These causal parameters can differ substantially as we show for the normal Probit model, which implies that care has to be taken when interpreting 2SLS estimation results in a linear probability model. Under joint normality of the error terms, IV Probit maximum likelihood estimation does identify the average partial effect of X. The two-step control function procedure of Rivers and Vuong can also estimate this causal parameter consistently, but a double averaging is needed, one over the distribution of the first-stage error V and one over the distribution of X. If instead a single averaging is performed over the joint distribution of X and V, then the same causal parameter is estimated as the one estimated by the 2SLS estimator in the linear probability model. The 2SLS estimator is a consistent estimator when the average partial effect is equal to 0, and the standard Wald test for this hypothesis has correct size under strong instrument asymptotics. We show that, in general, the standard weak instrument first-stage F-test interpretations do not apply in this setting.

Fast Cross-validation for Multi-penalty High-dimensional Ridge Regression

High-dimensional prediction with multiple data types needs to account for potentially strong differences in predictive signal. Ridge regression is a simple model for high-dimensional data that has challenged the predictive performance of many more complex models and learners, and that allows inclusion of data type-specific penalties. The largest challenge for multi-penalty ridge is to optimize these penalties efficiently in a cross-validation (CV) setting, in particular for GLM and Cox ridge regression, which require an additional estimation loop by iterative weighted least squares (IWLS). Our main contribution is a computationally very efficient formula for the multi-penalty, sample-weighted hat-matrix, as used in the IWLS algorithm. As a result, nearly all computations are in low-dimensional space, rendering a speed-up of several orders of magnitude. We developed a flexible framework that facilitates multiple types of response, unpenalized covariates, several performance criteria and repeated CV. Extensions to paired and preferential data types are included and illustrated on several cancer genomics survival prediction problems. Moreover, we present similar computational shortcuts for maximum marginal likelihood and Bayesian probit regression. The corresponding R-package, multiridge, serves as a versatile standalone tool, but also as a fast benchmark for other more complex models and multi-view learners. Supplementary materials for this article are available online.

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.

Convexity of probit weights

We demonstrate that the probit weight function is U-shaped on R, i.e., it is strictly decreasing on (−∞,0), strictly increasing on [0,∞), and strictly convex on R. Knowledge of the shape of the probit weight function can be useful in several contexts. For instance, it can resolve any confusion that may arise from a result in the classic paper of Sampford (1953). The shape of the probit weight function can also be used to justify why the computation of the probit maximum likelihood estimator (MLE) –in fact, in general, the computation of all two-step estimators whenever the first step involves estimating a probit model –may fail in the presence of outliers or unbounded parameter spaces.

Environmental consciousness and choice of bulb for lighting in a developing country

BackgroundMost countries in the world especially those in Asia and Africa have been undertaking policies meant to help promote science, technology and innovation towards meeting some provisions of the Sustainable Development Goals. However, there is still a sizable number of households who have not yet fully embraced energy-saving technologies. This study provides highlights on the economic and environmental benefits for investing in energy-saving light bulbs.MethodsUsing a survey and a multistage random sampling approach, we administered questionnaires to 1650 households in Ghana. The relevant diagnostic tests associated with cross-sectional data were undertaken. We estimated a maximum-likelihood probit model with its associated marginal effects to find out how the choice of energy-saving light bulb (behaviour) is influenced by environmental consciousness (both local knowledge and global knowledge) and other demographic factors.ResultsOur results are consistent with economic theory as well as what earlier empirical evidence found in literature. That is, environmental consciousness, education, income, etc. are very important in explaining the choice of buying energy-saving light bulbs in Ghana.ConclusionsBesides advocating for information that will make society more environmentally conscious, we further recommend the use of fiscal policies (i.e. subsidies) to support lower income brackets who are predominant in developing countries.

Application and success of R&D subsidies: what is the role of firm age?

This article evaluates the propensity of firms to apply for R&D public grants, their capacity for obtaining subsides and how these relate to the firm age. This analysis is particularly relevant in understanding why young firms may be more or less successful in their application for public subsidies. Using data from an extensive database of 19,236 firms in Catalonia, we estimate a maximum likelihood probit model with sample selection. Our results show that firm age does not exert a significant direct impact on the application propensity in R&D public call, but younger firms tend to show a larger propensity to receive an R&D subsidy. They also show that firm age exerts a moderating role through different internal resources. They highlight the need to design R&D public policies that explicitly favour applications by young firms to help overcome their obstacles to innovation.

Retirement Reconsidered: Labor Force Participation of Older Men in the United States

<p>The objectives of this article are two fold. Changes in older men’s labor force participation in the United States are first described focusing on human capital and demographic variables. A model of the labor/leisure choice and the retirement decision of older menare then estimated employing Maximum Likelihood Probit. While changes in Social Security Benefit rules are a significant factor in explaining the trend of rising retirement age among older men, the focus here is on additional factors that contribute to older men’s decision to forestall retirement. Probit coefficient estimates for three distinct age cohorts verify the effects of hypothesized determinants of the decision to retire. Specifically, the coefficient on estimated earnings is negative and hasthe largest marginal effect on the decision to retire followed by years of education and retired wife. The effects of wives’ retirement decision will likely influence and forestall the retirement decision of married men as more working women reach retirement age. The rise in labor force participation rates of older men may offset rather than reverse the decline in men’s labor force participation rates that began more than a half century ago.</p>

The Predictive Power of Financial Variables: New Evidence from Australia

Many studies have attempted to examine the predictive power of financial variables for numerous countries, but rarely does such research focus on future economic activities with respect to Australia. Financial variables are used to predict future economic events primarily because these variables are the closest indicators of the expectations and activities of investors and other economic agents. The recent global financial crisis (GFC) stemming from the subprime crisis shows that financial markets significantly influence global macroeconomic activities. In this study, we use major financial variables, such as the 90-day Treasury bill rate, 10-year Treasury bond rate, interest rate spread, and Australian stock index data. Similar to the housing prices in some other countries, those in Australia play a key role in future economic activities. In addition to financial variables, housing stock data is incorporated into our model for more realistic results, which are obtained by probit maximum likelihood estimation. We also use a general model for forecasting Australia’s GDP growth until the third quarter of 2012. The results support previous research findings, indicating that financial variables are a useful tool for forecasting future economic activities in Australia.

Maternal education and child nutritional status: evidence from Uganda

Purpose – The debate concerning the relationship between maternal education and child nutritional status is not a foregone conclusion. This paper aims to contribute to the existing debate by examining this relationship for the case of Uganda. Design/methodology/approach – Theoretically, the study was based on household models of optimization, just like in the standard consumer theory, to gain insights into household demand for the health good. Empirically, the paper employed maximum-likelihood probit models and computed marginal effects in order to obtain logically interpretable results. Findings – The paper finds that once the socio-economic factors are controlled for, the significance of maternal education, especially primary and secondary levels, in influencing child nutrition status decays but post-secondary education persists. Therefore, if mothers are exposed to the same socio-economic conditions, it is education of the mother beyond secondary level that generates a difference in the child nutrition outcomes. Practical implications – These findings suggest that efforts to improve the child health outcomes in the future need to target measures that aim to educate women beyond secondary level. The government program to extend free education at the secondary level is a good start and should be strengthened. Originality/value – Literature presents no consensus on the effect of maternal education and child nutritional status. It is often argued that maternal education is simply a proxy for the socio-economic conditions and geographical area of residence such its significance decays once these are controlled for. Yet others argue that maternal education is a single most important factor influencing child nutritional outcomes. The debate is still very hot and this study sought to contribute to this debate for the case of Uganda.

Factors Influencing Childhood Immunization in Uganda

This paper investigates the factors associated with childhood immunization in Uganda. We used nationally-representative data from Uganda Demographic and Health Survey (UDHS) of 2006. Both bivariate and multivariate approaches were employed in the analysis. The bivariate approach involved generating average percentages of children who were immunized, with analysis of pertinent background characteristics. The multivariate approach involved employing maximum likelihood probit technique and generating marginal effects to ascertain the probability of being immunized, given the same background characteristics. It revealed that slightly over 50% of children in Uganda were fully immunized. Additionally, 89%, 24%, 52%, and 64% received BCG, DPT, polio and measles vaccines respectively. Factors which have a significant association with childhood immunization are: maternal education (especially at post-secondary level), exposure to media, maternal healthcare utilization, maternal age, occupation type, immunization plan, and regional and local peculiarities. Children whose mothers had post-secondary education were twice as likely to be fully immunized compared to their counterparts whose mothers had only primary education (p < 0.01). Thus, gender parity in education enhancement efforts is crucial. There is also a need to increase media penetration, maternal healthcare utilization, and to ensure parity across localities and regions.

Management-employee relations, firm size and job satisfaction

Purpose– The purpose of this paper is to investigate the job satisfaction in relation to managerial attitudes towards employees and firm size using the linked employer-employee survey results in Britain.Design/methodology/approach– The authors first investigate the management-employee relationships and the firm size using maximum likelihood probit estimation. Next various measures of job satisfaction are related to the management-employee relations via maximum likelihood ordered probit estimates. Four measures of job satisfaction that have not been used often are considered. They are satisfaction with influence over job; satisfaction with amount of pay; satisfaction with sense of achievement and satisfaction with respect from supervisors.Findings– Main findings indicate that management-employee relationships are less satisfactory in the large firms than in the small firms. Job satisfaction levels are lower in large firms. Less satisfactory management-employee relationships in the large firms may be a major source of the observed lower level of job satisfaction in them.Practical implications– These results have important policy implications from the point of view of the firm management while achieving the aims of their organizations in particular in the large firms in the area of management-employee relationships. Improving the management-employee relations in large firms will increase employee satisfaction in many respects as well as increase productivity and reduce turnover.Originality/value– The nature of the management-employee relations with firm size and job satisfaction has not been investigated before.

Management-Employee Relations, Firm Size and Job Satisfaction

This paper investigates the job satisfaction in relation to managerial attitudes towards employees and firm size using the linked employer-employee survey results in Britain. We first investigate the management-employee relationships and the firm size using maximum likelihood probit estimation. Next various measures of job satisfaction are related to the management-employee relations via maximum likelihood ordered probit estimates. Four measures of job satisfaction that have not been used often are considered. They are satisfaction with influence over job; satisfaction with amount of pay; satisfaction with sense of achievement and satisfaction with respect from supervisors. Main findings indicate that management-employee relationships are less satisfactory in the large firms than in the small firms. Job satisfaction levels are lower in large firms. Less satisfactory management-employee relationships in the large firms may be a major source of the observed lower level of job satisfaction in them. These results have important policy implications from the point of view of the firm management while achieving the aims of their organizations in particular in the large firms in the area of management-employee relationships. Improving the management-employee relations in large firms will increase employee satisfaction in many respects as well as increase productivity and reduce turnover. The nature of the management-employee relations with firm size and job satisfaction has not been investigated before.

Management-Employee Relations, Firm Size and Job Satisfaction

This paper investigates the job satisfaction in relation to managerial attitudes towards employees and firm size using the linked employer-employee survey results in Britain. We first investigate the management-employee relationships and the firm size using maximum likelihood probit estimation. Next various measures of job satisfaction are related to the management-employee relations via maximum likelihood ordered probit estimates. Four measures of job satisfaction that have not been used often are considered. They are satisfaction with influence over job; satisfaction with amount of pay; satisfaction with sense of achievement and satisfaction with respect from supervisors. Main findings indicate that management-employee relationships are less satisfactory in the large firms than in the small firms. Job satisfaction levels are lower in large firms. Less satisfactory management-employee relationships in the large firms may be a major source of the observed lower level of job satisfaction in them. These results have important policy implications from the point of view of the firm management while achieving the aims of their organizations in particular in the large firms in the area of management-employee relationships. Improving the management-employee relations in large firms will increase employee satisfaction in many respects as well as increase productivity and reduce turnover. The nature of the management-employee relations with firm size and job satisfaction has not been investigated before.

WE‐A‐217A‐05: Lung Deformations and Airway Stenosis with Distal Atelectasis in Patients Treated with Stereotactic Body Radiation Therapy for Lung Tumors

Purpose: To investigate the possibility to use lung deformations to differentiate between radiation‐induced fibrosis and potential airway stenosis with distal atelectasis in patients treated with stereotactic body radiation therapy (SBRT) for lung tumors. Methods: Planning and follow‐ ups CT scans were reviewed for 46 patients who received lung SBRT treatment at our institution between 2002 and 2009, in 3 or 5 fractions, and a median total dose of 54Gy (range, 30–60). Vector lung tissue displacements were studied for patterns of parenchymal deformation only in patients with follow‐up CT scans &gt; 6 months. Planning CT lung surface contours were matched to the follow‐up scans and at least 50 anatomical landmarks inside the lung (vessel or airway branches) were identified on both scans to guide the deformation of the lung structure. Parenchyma displacements maps for the whole lung were interpolated from the landmark deformations and analyzed to identify tissue migrations towards the observed hyperdense region consistent with distal atelectasis. Results: Out of 46, only 24 patients demonstrated lung density changes with sufficient follow‐up time and image quality. Fifteen patients (62%) displayed converging deformation patterns and their average Hounsfield Unit change for the hyperdense region was 344HU and significantly higher than for the other patients (247HU, p = 0.03) supporting an alternate consolidation mechanism. The mean maximum dose to proximal airways was 38Gy for the converging pattern group versus 20Gy (p = 0.003) and a maximum likelihood Probit fit yielded TD50 = 24Gy, m = 0.56. The mean minimum dose to the hyperdense region was 15Gy versus 18Gy (p = 0.23). Conclusions: A correlation between converging deformation patterns and lung density changes was observed suggesting the possibility of an additional mechanism for lung density increase in addition to fibrosis. Correlations between converging patterns and dose to proximal airways also suggest a 24Gy dose‐threshold effect.

Determinants of diarrhoea and acute respiratory infection among under-fives in Uganda

Diarrhoea and acute respiratory infection (ARI) are leading causes of mortality and morbidity in children under the age of five in developing countries. On the African continent, pneumonia (14%) and diarrhoea (17%) cause more child deaths than Malaria (16%), HIV/AIDS (4%), and measles (1%) combined. This paper set out to investigate the factors associated with the occurrence of diarrhoea and ARI incidence for children under five years in Uganda. We used a nationally representative Uganda Demographic and Health Survey (UDHS) (2006). Sampling was done in two stages. In the first stage 321 clusters were selected from among a list of clusters sampled in the 2005/06 Uganda National Household Survey (UNHS), 17 clusters from the 2002 Census frame from Karamoja, and 30 internally displaced camps (IDPs). In the second stage, households in each cluster were selected as per UNHS listing. In addition 20 households were randomly selected in each cluster. Questionnaires were used during data collection. During the analysis, a maximum likelihood probit model was used in order to ascertain the probability of occurrence of diseases. On average, 32% and 48% of children in the survey suffered from diarrhoea and ARI in the two weeks prior to the survey date. The occurrence was concentrated amongst children aged 0-24 months. Mother's education, especially at postsecondary level, reduced the probability of diarrhoea occurrence but had no effect on ARI occurrence. First hour initiation and exclusive breastfeeding reduced the probability occurrence of both diarrhoea and ARI. Other significant factors associated with the occurrence of both diseases include: regional and location differentials, wealth status, type of dwelling, mother's occupation, child age, and child nutritional status. Policy interventions should target female education, eliminate location and regional disadvantages, and educate the population to adopt breastfeeding practices recommended by the World Health Organization (WHO). The government should also ensure proper dwelling places for the population that are associated with favourable health outcomes. Other proper feeding practices together with breastfeeding (after six months), should be made known to the masses so as to reduce the number of children that are malnourished and growth retarded.

Maternal education and childbirth care in Uganda.

Globally, over 500,000 females die of complications related to pregnancy and childbirth each year, and of these, over 99% of deaths occur in developing countries such as Uganda. Utilisation of modern and professional care during delivery is important in lowering maternal mortality. This paper sets out to investigate the factors associated with the utilisation of modern and professional childbirth care so as to inform policy makers on the pertinent factors that need to be influenced by policy. A nationally representative Uganda Demographic and Health Survey (UDHS) (2006) was used. Sampling was done in two stages. In the first stage 321 clusters were selected from a list of clusters sampled in the 2005/06 Uganda National Household Survey (UNHS), 17 clusters from the 2002 Census frame from Karamoja, and 30 internally displaced camps (IDPs). In the second stage, the households in each cluster were selected as per the UNHS listing. In addition an additional 20 households were randomly selected in each cluster. Questionnaires were used during data collection. During the analysis, a maximum likelihood probit technique was employed. Prior to this, a bivariate approach was used to generate average percentages of mothers using the childbirth care services by background characteristics. It is found that maternal education is the strongest predictor, especially at post-secondary level (highest marginal effect of 33% and P<0.01), associated with the utilisation of childbirth care. Whereas partner's education at all levels is important, maternal education is observed to exert a much stronger association. Other factors significantly associated with the utilisation of professional childbirth care include community infrastructure, occupation, location, and regional differences, wealth status, religion, and age cohorts. These findings suggest that whereas all levels of education are important, the effects of post-secondary education are more pronounced. Therefore, efforts to improve professional childbirth care utilisation need to focus on female education beyond secondary level. In addition, measures are needed to improve agricultural productivity which might improve earnings and childbirth care utilisation. Government should also undertake to improve community infrastructure across all regions and locations. Both government and donors can ensure universal access to professional childbirth care irrespective of the ability to pay.

Poverty and childhood undernutrition in developing countries: A multi-national cohort study

The importance of reducing childhood undernutrition has been enshrined in the United Nations' Millennium Development Goals. This study explores the relationship between alternative indicators of poverty and childhood undernutrition in developing countries within the context of a multi-national cohort study (Young Lives). Approximately 2000 children in each of four countries - Ethiopia, India (Andhra Pradesh), Peru and Vietnam - had their heights measured and were weighed when they were aged between 6 and 17 months (survey one) and again between 4.5 and 5.5 years (survey two). The anthropometric outcomes of stunted, underweight and wasted were calculated using World Health Organization 2006 reference standards. Maximum-likelihood probit estimation was employed to model the relationship within each country and survey between alternative measures of living standards (principally a wealth index developed using principal components analysis) and each anthropometric outcome. An extensive set of covariates was incorporated into the models to remove as much individual heterogeneity as possible. The fully adjusted models revealed a negative and statistically significant coefficient on wealth for all outcomes in all countries, with the exception of the outcome of wasted in India (Andhra Pradesh) and Vietnam (survey one) and the outcome of underweight in Vietnam (surveys one and two). In survey one, the partial effects of wealth on the probabilities of stunting, being underweight and wasting was to reduce them by between 1.4 and 5.1 percentage points, 1.0 and 6.4 percentage points, and 0.3 and 4.5 percentage points, respectively, with each unit (10%) increase in wealth. The partial effects of wealth on the probabilities of anthropometric outcomes were larger in the survey two models. In both surveys, children residing in the lowest wealth quintile households had significantly increased probabilities of being stunted in all four study countries and of being underweight in Ethiopia, India (Andhra Pradesh) and Peru in comparison to children residing in the highest wealth quintile households. Random effects probit models confirmed the statistical significance of increased wealth in reducing the probability of being stunted and underweight across all four study countries. We conclude that, although multi-faceted, childhood undernutrition in developing countries is strongly rooted in poverty.

Factors influencing the uptake of screening services for breast and cervical cancer in Taiwan

Cervical cancer and breast cancer have had the first and second highest occurrence rates among women in Taiwan since 1981 and have also been two of the leading causes of cancer death. In order to more effectively promote preventive medical care programs, it is important to identify the key determinants of women's behavior regarding their decisions to engage in screenings. This study aims to identify the major factors that affect the utilization of breast and cervical cancer screening among women in Taiwan, who are covered by universal National Health Insurance. Data were obtained from the 2001 National Health Interview Survey for the Taiwan area and targeted women between 30 and 75 years old. The sample size consisted of 5,611 individuals for Papanicolaou (Pap) tests and breast examinations by a doctor. For breast ultrasounds and mammograms, 3,875 individuals were included. The propensity among women in Taiwan to use the two types of screening services, i.e., Pap test and breast examinations, were estimated by maximum likelihood probit regressions. The findings in this study show that the likelihood of a woman receiving a Pap test or a clinical breast examination depends on a variety of factors such as age, marital status, income level, education, and health status. Women with lower socioeconomic status were found to be much less likely to undergo the freely available cancer screening services. In addition, healthy behavior, such as not smoking and engaging in exercise, had a positive effect on the uptake of screening. The results of this study enable us to understand more thoroughly the characteristics of women who undergo a Pap test and breast examination, as well as the factors that influence them in Taiwan. The findings can help formulate related policies that are directed at removing the barriers to accessing medical care and targeting those at-risk groups. This analysis provides new evidence of the factors affecting the utilization of preventive care among women in a developing country, which are comparable to those of other countries, and may shed further light on the issue of promoting cancer screening and women's health.