"... In PAGE 14: ...8 Mean educational outcomes are lower in the stable blended family subsamples than for all siblings. Table2 reports the descriptive statistics for the NLSY-Child sample and our stable blended family subsample. There are 4,320 siblings in the sample, of whom 418 individuals 6 Following the census definition, we say that a blended family is one that must include at least one stepparent, stepsibling and/or half-sibling.... ..."

"... In PAGE 14: ... Table1 reports the means and standard deviations of the variables used in the NLSY siblings sample and stable blended family subsample. Average years of schooling in the sibling sample is 13.... In PAGE 14: ...5 Mean educational outcomes are lower in the stable blended family subsample than for all siblings. In addition to means and standard deviations, Table1 reports the standard deviation within families and the percent of the variance of the educational outcome and family structure variables accounted for by within family variation. In the sibling sample, over half of the variation in family structure and educational outcomes is accounted for between families.... ..."

"... In PAGE 18: ... We begin with simple tests of differences in mean schooling outcomes. The top panel of Table5 tests the null hypothesis of no difference in mean schooling outcomes between siblings in stable blended families and siblings from traditional families in the NLSY sample. For all four schooling outcomes we reject the null hypothesis of no difference in schooling outcomes.... In PAGE 19: ... Biological children in blended families are necessarily younger than their half-siblings who are stepchildren, and the relatively poor outcomes for biological children in blended families could be the result of birth order. The bottom panel of Table5 compares the educational outcomes of the youngest children in traditional families with the biological children in blended families. In three of the four outcomes biological children from blended families have significantly lower educational attainment.... ..."

"... In PAGE 22: ... They found that AFQT scores are a significant determinant of wages and explain 63% of the black-white wage differential after controlling for education.19 Table2 presents analogous results for our sample. In column 1 of table 2 we report the overall black-white wage differential to be explained using log wages.... ..."

"... In PAGE 13: ... There is no evidence that the presence of an incidence effect explains the non-linearity.12 To contrast the effects of training on wages implied by the different functional forms, the last column in Table3 shows the predicted effect of training at the median of the distribution of positive hours of training. The implied effect of the median hours of training differs by more than a factor of 12 between the different specifications.... In PAGE 24: ... Rate of Return Results. Table 6 shows results for the specifications considered in Table3 , with terms for lagged and lead training added. (Wage observations for the year 2000 were omitted, as lead training is not observed.... ..."

"... In PAGE 24: ...ection 2.1.3 this value is made up of three terms: (i) net unemployment benefits, (ii) the value of home production, production in the informal sector or any non-pecuniary value, less (iii) search costs. Table5 reports the decomposition of b F N into these components according to the estimates of Table 2. As in Tables 3 and 4, each entry reports the sample mean and standard deviation of the relevant variable using the point estimates of the parameters in Table 2.... In PAGE 25: ...S. data: Eckstein and Wolpin (1995), using NLSY data, report (see their Table5 , p.282) reservation wages defined as rJU at the steady state.... ..."

"... In PAGE 25: ... This gives separate estimation of the time path of each combination of the strata variables, while the remaining variables are specified as proportional effects with a single set of coefficients for all strata. Table3 shows the number of spells in each stratum, and indicates that there are sufficient data for precise estimation of the baseline hazards. === Insert Table 3 About Here === Jobs crossing NLSY interview dates have multiple records in the data with different values of covariates.... In PAGE 25: ... Table 3 shows the number of spells in each stratum, and indicates that there are sufficient data for precise estimation of the baseline hazards. === Insert Table3 About Here === Jobs crossing NLSY interview dates have multiple records in the data with different values of covariates. This results in a multi-spell structure of the data, and since a respondent can change jobs more than once, the data also have a repeated-event structure.... ..."

"... In PAGE 16: ...ikelihood to obtain estimates of the behavioral parameters. We set $ =.95, T = 40, and M = 2. We show below that our main results are not very sensitive to these parameter settings. Our basic results are shown in Table2 . We find that differences in expected income are a significant determinant of migration decisions for this population.... In PAGE 16: ... This is an annual migration rate of 3.69%, and the first column in Table2 matches this rate by setting the probability of moving to each of J-1 locations to a constant value, namely , with J = 51.18 The next columns show that population size, distance, home and previous locations and age all have highly significant effects on migration.... In PAGE 21: ... A natural interpretation of this is mover-stayer heterogeneity: some people are more likely to move than others, and these people account for more than their share of the observed moves. We simulated the corresponding statistics for the model by starting 1000 replicas of the NLSY individuals in the observed initial locations, and using the model (with the estimated parameters shown in Table2 ) to generate a history for each replica, covering the number of periods observed for this individual. In the case of the homogeneous model, the fit is good, although both the proportion of people who never move, and the proportion of movers who move more than once are a bit low, relative to the... In PAGE 25: ...he data. Our model includes age as a state variable, to capture the finite horizon effect just discussed. The model also allows for the possibility that age has a direct effect on the cost of migration; this can be regarded as a catch-all for whatever is missing from the human capital explanation.24 The results in Table2 show that this direct effect is large and significant. 5.... In PAGE 35: ...Appendix B: Validation of ML Estimates The parameter estimates from Table2 were used to generate replicas of each NLSY observation, starting from the actual value in the NLSY data, and allowing the model to choose the sequence of locations. Table 10 gives results for 10, 100, and 1,000 replicas of each NLSY observation.... ..."