Bayesian networks for imputation

Summary.  Bayesian networks are particularly useful for dealing with high dimensional statistical problems. They allow a reduction in the complexity of the phenomenon under study by representing joint relationships between a set of variables through conditional relationships between subsets of these variables. Following Thibaudeau and Winkler we use Bayesian networks for imputing missing values. This method is introduced to deal with the problem of the consistency of imputed values: preservation of statistical relationships between variables ( statistical consistency ) and preservation of logical constraints in data ( logical consistency ). We perform some experiments on a subset of anonymous individual records from the 1991 UK population census.     

An experimental investigation of the impact of ambiguity on the valuationof self-insurance and self-protection

We build two experimental markets to examine individual valuations of risk reductions with two risk-management tools: self-insurance and self-protection. We find no positive evidence that the risk-reducing mechanisms constitute a “frame.” Ambiguity in the probability on average affects valuation only weakly, and changes in the representation of ambiguity do not alter valuation. Finally, unlike the results obtained by Hogarth and Kunreuther for the case of market insurance, our findings do not provide a strong support for the “Anchoring and Adjustment” ambiguity model.     

The effects of separate taxation on labor participation of married couples. An empirical analysis using propensity score

Up to 1987 the Spanish Income Tax imposed compulsory joint filing for married couples. However, the 1988 reform allowed spouses to choose between joint and separate taxation, involving a reduction in tax rates for secondary earners. Our aim is to analyze this reform as a quasi-natural experiment, assessing the effects of tax changes on labor participation. To find out the causal effect we adopt the difference-in-differences technique. We use data from the ‘Spanish Income Tax Panel 1982–1998’. Our results show that, as a consequence of differential tax changes, married women in families more strongly affected by the fiscal reform increase their labor participation more than secondary earners from families less affected by the reform. The participation rate for secondary earners in the treatment group increases by 9.4 percentage points whereas the control group increases their participation rate by 7.8 percentage points. We define the treatment group as those secondary earners in relatively low-income families in year 1987 and the control group as those in middle-high income families, because the former experiences a stronger reduction in tax rates than the latter. As a result, we can attribute the 1.6-percentage-point-increase in participation rates to the 1988 income tax reform.     

See you on Facebook! A framework for analyzing the role of computer-mediated interaction in the evolution of social capital

Empirical studies have documented a decline in indicators of social participation in the last five decades. The responsibility of social disengagement has often been attributed to pervasive busyness and the increasing pressure on time. In this paper we argue that computer-mediated interaction, and particularly online networking, can help mitigate this downward trend. We develop a logical framework for assessing the role of the Internet in the evolution of social participation. We analyze an economy where agents can develop their social interactions through two main modes of participation, one encompassing both online networking and face to face interactions, and the other solely based on physical encounters. We study the interdependence between the increase in the pressure on time and the variation in the relative performance of the two strategies of participation.     

The Multidimensionality of Child Poverty: Evidence from Afghanistan

This paper examines multidimensional poverty among children in Afghanistan using the Alkire-Foster method. Several previous studies have underlined the need to separate children from their adult nexus when studying poverty and treat them according to their own specificities. From the capability approach, child poverty is understood to be the lack of freedom to do and to be what children themselves value and have reason to value. The case of Afghanistan is particularly relevant as years of conflict aggravated by several severe droughts, political insecurity, bad governance and ongoing violence have significantly increased poverty in the country. The paper discusses the relevant dimensions when analysing child poverty and uses data from a survey carried out by Handicap International which contains information on dimensions of children’s wellbeing that is typically missing in standard surveys. Ten dimension are considered in this paper: health, care and love, material deprivation, food security, social inclusion, education, freedom from economic and non-economic exploitation, shelter and environment, autonomy, and mobility. Our results show that younger children, those living in rural areas, girls and disabled children are the most deprived.     

The Effect of Cash Transfers on Fertility: Evidence from Argentina

In 2009 Argentina introduced a large poverty-alleviation program (AUH) that provides monthly cash transfers per child to households without workers in the formal sector. In this paper we study the potential unintended effect of this program on fertility. We apply a difference-in-difference strategy comparing the probability of having a new child among eligible and ineligible mothers both before and after the program inception. The intention to treat estimations suggest a significant positive impact on fertility in households with at least one child (around 2 percentage points), but no significant effect on childless households. Given the short time window since the implementation of the AUH, we are unable to identify whether this positive effect reflects changes in the timing of births or in the equilibrium number of children.     

Small-Sample Improved Seasonal Unit Root Tests for Trending and Breaking Series

In this article three unit root tests that allow for a break in both the seasonal mean and linear trend of the data are proposed. The tests, which can be seen as small-sample corrected versions of already known asymptotic tests, are shown to perform very well in simulations, and much better than their asymptotic counterparts.     

Parametric estimation for ARFIMA models via spectral methods

Given a fractional integrated, autoregressive, moving average,ARFIMA (p, d, q) process, the simultaneous estimation of the short and long memory parameters can be achieved by maximum likelihood estimators. In this paper, following a two-step algorithm, the coefficients are estimated combining the maximum likelihood estimators with the general orthogonal decomposition of stochastic processes. In particular, the principal component analysis of stochastic processes is exploited to estimate the short memory parameters, which are plugged into the maximum likelihood function to obtain the fractional differencingd.     

Seemingly Unrelated Ridge Regression in Semiparametric Models

This article is concerned with the problem of multicollinearity in the linear part of a seemingly unrelated semiparametric (SUS) model. It is also suspected that some additional non stochastic linear constraints hold on the whole parameter space. In the sequel, we propose semiparametric ridge and non ridge type estimators combining the restricted least squares methods in the model under study. For practical aspects, it is assumed that the covariance matrix of error terms is unknown and thus feasible estimators are proposed and their asymptotic distributional properties are derived. Also, necessary and sufficient conditions for the superiority of the ridge-type estimator over the non ridge type estimator for selecting the ridge parameter K are derived. Lastly, a Monte Carlo simulation study is conducted to estimate the parametric and nonparametric parts. In this regard, kernel smoothing and cross validation methods for estimating the nonparametric function are used.