Gross flows estimation from the Survey of Income and Program Participation

摘要: The paper focuses on one of the main longitudinal surveys conducted by US Bureau Census, Survey Income and Program Participation (SIPP); specifically it regards problem estimating gross flows in labour market when data are affected classification errors. Some analysis presented show that typical sources errors data, like time-in-sample effect, do not affect SIPP observed dramatically. so-called seam effect dominates other phenomena potentially bias means surveys. On basis this evidence, a strategy to correct force with from is proposed. an ad hoc specification latent class analysis. Gross among states major tool analyse dynamics market. In most industrialised countries central statistical office conducts order collect information participation. Longitudinal follow units over time; they may be distinguished between (i) panel surveys, where same respondents collected at successive moments time (ii) retrospective cross-sectional including questions; more often, practical situations, mixture two typologies. To dynamics, usually individuals classified into three mutually exclusive states: Employed (E), Unemployed (U) Not (N). With net estimated. regard changes individual level. whether repeated or retrospective, always measurement reference estimation market, these termed errors, since cause erroneously their position Even if sometimes cancel out flows, can ignored flows. Sources for reviewed [5]. Panel attrition, which essentially unit nonresponse, either because leave sample before end