The effect of sample selection and initial conditions in duration models: evidence from experimental data on training

Article Abstract:

Training programs can have significant effects on the duration of participants' employment and unemployment spells. In a typical program, participants are randomly assigned to treatment and control groups. However, those who experienced being alternately employed and unemployed are not generally random subsets of the initial groups because different sorting processes were used for these groups. Traditional duration models ignore these differences, making sample selection inaccurate and resulting in misleading estimates. An estimator is proposed to resolve this problem.

Employment, Statistical sampling, Sampling (Statistics), Employee training, Ranking and selection (Statistics)

---

Continuous record asymptotics for rolling sample variance estimators

Article Abstract:

Previous research has shown changes over time of conditional covariances of asset returns. Taking account of this phenomenon, researchers have adopted strategies for adopting conditional heteroscedasticity. Such strategies include chopping the data into short blocks of time that are assumed to be homoskedastic, performing one-sided rolling regressions and performing two-sided rolling regressions. Continuous record asymptotic approximations for the measurement error in conditional variances and covariances when these methods are used are developed.

Heteroscedasticity, Asymptotes

---

Back to the future: generating moment implications for continuous-time Markov processes

Article Abstract:

Moment conditions are being derived to estimate and test continuous-time Markov models. Infinitesimal generators characterize Markov processes, in which there exists forward and reverse-time generators. The building of moment conditions with these generators suggested by stationary Markov processes are demonstrated. These moment conditions can also be used to build generalized method of moments estimators and tests. The derived econometric methods are intended to be applied to the discrete-time data of Markov processes.

Markov processes, Nonparametric statistics

Similar abstracts:

• Abstracts: The effects of general inflation and idiosyncratic cost shocks on within-commodity price dispersion: evidence from microdata
• Abstracts: Finite sample prediction and interpolation for ARMA models with missing data. Dynamic Harmonic Regression
• Abstracts: The effect of public capital in state-level production functions reconsidered. Costs and factor substitution in the provision of local fire services
• Abstracts: Short-term forecasting of industrial electricity consumption in Brazil. Modelling the impact of temperature on electricity consumption in the Eastern Province of Saudi Arabia
• Abstracts: The differential regional effects of monetary policy. Habit formation and intertemporal substitution in individual food consumption

This website is not affiliated with document authors or copyright owners. This page is provided for informational purposes only. Unintentional errors are possible.