A Comparison of Bayesian and Sampling Theory Inferences in a Probit Model

HE PROBIT MODEL IS A POPULAR DEVICE for explaining binary choice decisions in econometrics. It has been used to describe choices such as labor force participation, travel mode, home ownership, and type of education. These and many more examples can be found in papers by Amemiya (1981) and Maddala (1983). Given the contribution of economics towards explaining such choices, and given the nature of data that are collected, prior information on the relationship between a choice probability and several explanatory variables frequently exists. Bayesian inference is a convenient vehicle for including such prior information. Given the increasing popularity of Bayesian inference it is useful to ask whether inferences from a probit model are sensitive to a choice between Bayesian and sampling theory techniques. Of interest is the sensitivity of inference on coefficients, probabilities, and elasticities. We consider these issues in a model designed to explain choice between fixed and variable interest rate mortgages. Two Bayesian priors are employed: a uniform prior on the coefficients, designed to be noninformative for the coefficients, and an inequality restricted prior on the signs of the coefficients. We often know, a priori, whether increasing the value of a particular explanatory variable will have a positive or negative effect on a choice probability. This knowledge can be captured by using a prior probability density function (pdf) that is truncated to be positive or negative. Thus, three sets of results are compared:those from maximum likelihood (ML) estimation, those from Bayesian estimation with an unrestricted uniform prior on the coefficients, and those from Bayesian estimation with a uniform prior truncated to accommodate inequality restrictions on the coefficients.

Children and economic development: Family size, gender preferences and human capital formation - theory and Indian cases

In the light of Gary Becker's economic theory of the family, considers how economic cost and benefit factors can influence the size of families that parents decide to have. Some support for the importance of such factors is found from results of structured interviews with wives in Kondh-dominated villages in western Orissa. These results are at variance with the hypothesis of Malthus about population growth. Factors that may alter the optimal family size as development proceeds are discussed. It is found in our sampling that, on the whole, there is a preference for daughters rather than sons although this is not as strong in the Kondh-dominated villages as in poor villages in the Santal tribal belt of West Bengal. While in the Kondh-dominated villages some discrimination in access to education in favour of boys compared to girls is present, little such or no such discrimination occurs in relation to access to food and medical attention. In the villages surveyed in the West Bengal Santal tribal belt, discrimination in favour of boys is more pronounced than in the Kondh-dominated area in Orissa. While economic considerations help to explain gender discrimination between boys and girls, we find that social and cultural factors also play a major role. Parents in a similar economic situation seem to display substantially different patterns of gender discrimination between children depending on their social and cultural content. It seems that the extent to which economic theories of the family explain family preferences and behaviour depend significantly on the social and cultural context in which they are to be applied.

Marital status, farm size and other influences on the extent of cash cropping in Kenya: A household case study

This article examines the effects of marital status, farm size and other factors on the extent of cash cropping (and allocation of land use) by means of a case study in the Nyeri district in Kenya. It was found that married women are involved in the production of a relatively greater amount of output of cash crops than unmarried women since husbands prefer to have more land under cash crops than food crops. Farmers with better quality land allocate a high proportion of it to non-food cash crops, which may expose some households to greater risks of possible famine. The proportion of land allocated to food crops declines as the farm size increases while the proportion of land allocated to non-food cash crops rises as the size of farm increases. Age is also inversely associated with subsistence. Education, though inversely associated with subsistence farming does not appear to be statistically very significant as an influence on the composition of land use and composition of farm output. With growing commercialisation, married women work more hours than unmarried ones, working not only on non-cash food crops but also on non-food cash crops. Married women seem to lose their decision-making ability with growth of agricultural commercialisation, as husbands make most decisions to do with cash crops. Married women in Kenya also have little or no power to change the way land is allocated between food and non-food cash crops.

Gender inequality, poverty and human development in Kenya: Main indicators, trends and limitations

Indicators of gender inequality, poverty and human development in Kenya are examined. Significant and rising incidence of absolute poverty occurs in Kenya and women are more likely to be in poverty than men. Female/male ratios in Kenyan decision-making institutions are highly skewed against women and they experience unfavourable enrolment ratios in primary, secondary and tertiary institutions. The share of income earned by women is much lower than men's share. General Kenyan indicators highlight declining GDP per capita, increased poverty rates especially for women, reduced life expectancy, a narrowing of the difference in female/male life expectancy rates, increased child mortality rates and an increase in the female child mortality rates. This deterioration results in an increased socio-economic burden on women, not adequately captured in the HPI, HDI, GDI and GEM. This paper advocates the use of household level gender disaggregated data because much gender inequality occurs in and emanates from the household level where culture plays a very important role in allocation of resources and decision-making. Because most human development indicators are aggregates or averages, they can be misleading. They need to be supplemented by distributional and disaggregated data as demonstrated in the Kenyan case. The importance is emphasised of studying coping mechanisms of household/families for dealing with economic hardship and other misfortunes, such AIDS.

Closing the gender gap in Bangladesh: Inequality in education, employment and earnings

Assesses the status of women in Bangladesh by analysing the dynamics of female participation in labour force and education as well as gender earnings differentials at the macro level. The study finds evidence of growing commercialisation of women’s work in Bangladesh. Although the bulk of the female labour force is engaged in self-employment activities in the rural area or in low-skilled textile and readymade garment industries in the urban area, women’s participation in high-skill and entrepreneurial jobs as well as various decision-making bodies is also on the rise. While the gender wage differentials have been considerably reduced in many industries, in general women tend to be paid less than men. There have been remarkable improvements in women’s educational attainments compared to men. Further, female access to education is found to be highly correlated with overall female labour force participation, and relative to male participation. The overall results are suggestive of an improvement in the status of women in Bangladesh.