[Zurück]

E. Deutsch:
"Forecasting Tenure Durations from Survey Responses on Current Spells";
Vortrag: ERES 2010 17th annual conference, Bocconi University, Milan; 23.06.2010 - 26.06.2010.

Zensen und Surveys enthalten vielfach Berichte über die Verweildauern in gegenständlichen Mieten. Das erlaubt Einsichten in die Mobilität, liefert aber keine direkte Information über künftige Mobilität oder eventuellen Verbleib. Das Paper versucht, auf der Basis von Weibull-Erneuerungsprozessen die erwartbaren Verweildauern aus Spell-Informationen zu simulieren, und kontrastiert die Ergebnisse mit Weibull Survival-Schätzungen. Den empirischen Hintergrund liefern österreichische Mikrozensusdaten.

In various survey data, the tenants report the spell from past entry to the current date. Although that gives information about the mobility pattern, it does not provide any immediate forecast when the tenants will move in the future or eventually stay in probability. That sort of information would be useful for planning processes, but the data signal only the interoccurrence time. From that we can estimate the backwards recurrence time, which clearly underestimates the total stay. Inferring about the latter needs strong assumptions on the nature of the underlying stochastic processes, discussed by Feller, Cox, Karlin and Taylor, Bhat and Miller, Grimmett and Stirzacker. To approximate the basic stochastic process of durations, we perform Monte-Carlo simulations for a Weibull-process and check, whether such a procedure offers a benchmark to forecast the total spells. The outcome is contrasted with censored Weibull-survival estimates. The data used in the paper are drawn from the Austrian Census.

Tenure spell, mobility, renewal process, simulation, Weibull survivals, Austrian census