IV Quantile Regression for Group‐Level Treatments,With an Application to the Distributional Effects of Trade

We present a methodology for estimating the distributional effects of an endogenous treatment that varies at the group level when there are group‐level unobservables, a quantile extension of Hausman and Taylor, 1981. Because of the presence of group‐level unobservables, standard quantile regression techniques are inconsistent in our setting even if the treatment is independent of unobservables. In contrast, our estimation technique is consistent as well as computationally simple, consisting of group‐by‐group quantile regression followed by two‐stage least squares. Using the Bahadur representation of quantile estimators, we derive weak conditions on the growth of the number of observations per group that are sufficient for consistency and asymptotic zero‐mean normality of our estimator. As in Hausman and Taylor, 1981, micro‐level covariates can be used as internal instruments for the endogenous group‐level treatment if they satisfy relevance and exogeneity conditions. Our approach applies to a broad range of settings including labor, public finance, industrial organization, urban economics, and development; we illustrate its usefulness with several such examples. Finally, an empirical application of our estimator finds that low‐wage earners in the United States from 1990 to 2007 were significantly more affected by increased Chinese import competition than high‐wage earners.     

Method Madness

I caution you not to expect a cookbook solution to your difficulty. I cannot tell you what to do in ten easy steps. Precious few successful career searches are conducted in so mechanical a fashion. You must decide the best sequence of events for yourself. —Howard Figler, The Complete Job Search Handbook     

A Mathematical Approach to Categorization and Labeling of Qualitative Data: the Latent Categorization Method

As text databases increasingly become available to researchers, the limits to human cognition are rapidly reached. Focusing on examining objective realities, this paper introduces the latent categorization method, a novel new research method for analysis of large and midsize data sets. This method clusters text artifacts and extracts the words that were most important in creating the clusters. Further, it demonstrates a set of techniques for extracting knowledge from a representative data set involving 6135 abstracts from a variety of business‐related journals .     

Hutterite fecundability by age and parity: Strategies for frailty modeling of event histories

Effective fecundability declines with age and parity. Furthermore, women differ in their effective fecundability: some women have persistently low or high monthly chances of live-birth conception. Estimates are presented concerning the magnitude of these effects in a natural-fertility population: 406 Hutterite women in North America who had 3,206 births, largely in the 1940s and 1950s. The estimates are based on models that incorporate the effects of persistent heterogeneity and that use the full information provided by multiple-spell duration data. In addition, hazards rather than probabilities are modeled, piecewise linear hazard functions are used, and age and parity effects are decomposed systematically. These methods permit the development of more elaborate models of changing fecundability and of heterogeneity in postpartum amenorrhea.     

Factors affecting adoption in China, 1950-87

Though adoption in China is known to have increased as more girls were abandoned and became available for adoption following the introduction of the country's one-child policy in the 1980s, little is known about Chinese adoption practices. This paper investigates the factors affecting adoption in China in the period 1950-87 using data from the National Two-Per-Thousand Sample Survey on Fertility and Contraception of 1988. The results show that the national adoption rate was 2.3 per cent. Childless women were more likely than women with children to adopt and did not show a sex preference. Women with children by birth used adoption to secure a child of the 'missing' sex. Women who had experienced the death of a child were more likely to adopt than those who had not. Women with children may have used adoption as a strategy to circumvent the strict family planning policies.     

Structural unemployment. An analysis of recruitment and selection mechanisms based on panel data among Danish long-term unemployed

The perception of structural unemployment – summarised in the notion of 'Eurosclerosis'– became almost hegemonic during the 1990s. Policy makers all over Europe tried, by means of supply-side policies, to counteract the lack of incentives in the developed European welfare states, the lack of qualification on the post-industrial labour markets and the personal decay due to long-term unemployment. However, based on the critical case of Denmark, this article challenges the perception of structural unemployment and suggests an alternative business cycle/barrier perception. At the macro level it is difficult to explain the Danish decline in unemployment from 1994 to 2000 within the structure perception. The lack of explanatory power of the structure perception is further highlighted in micro-level analyses conducted on a panel study of long-term unemployed. Based on the unemployed's own assessments, we find no indications of supply-side problems. These results are supported by analyses of actual labour market integration of the long-term unemployed in the period between 1994 and 1999, which show that education level and previous unemployment had no noteworthy influence on labour market integration, whereas age had a decisive influence. These surprising results further undermine the perception of structural unemployment and the supply-side policies rooted in this 'mistaken' problem definition.     

Maximum Likelihood Inference for Multivariate Frailty Models Using an Automated Monte Carlo EM Algorithm

We present a maximum likelihood estimation procedure for the multivariate frailty model. The estimation is based on a Monte Carlo EM algorithm. The expectation step is approximated by averaging over random samples drawn from the posterior distribution of the frailties using rejection sampling. The maximization step reduces to a standard partial likelihood maximization. We also propose a simple rule based on the relative change in the parameter estimates to decide on sample size in each iteration and a stopping time for the algorithm. An important new concept is acquiring absolute convergence of the algorithm through sample size determination and an efficient sampling technique. The method is illustrated using a rat carcinogenesis dataset and data on vase lifetimes of cut roses. The estimation results are compared with approximate inference based on penalized partial likelihood using these two examples. Unlike the penalized partial likelihood estimation, the proposed full maximum likelihood estimation method accounts for all the uncertainty while estimating standard errors for the parameters.     

Behavioral health service use and costs among children in foster care

This article compares behavioral health service use and cost for foster care versus nonfoster care children; children before, during, and after foster care placement; and successfully reunified versus nonsuccessfully reunified foster care children. Behavioral health service costs for children in foster care were higher than for children not in foster care. Children in foster care used more services during their foster care placement than before placement and after discharge. Nonsuccessfully reunified children received a significantly larger quantity of services than those successfully reunified.