A Tobit Model with Social Interactions under Incomplete Information

Mardi | 2016-05-31
16h20-17h40 en Sully05

Yang CHAO – Lung-fei LEE – Xi QU

We build a model to analyze censored behaviors for socially associated agents under ageneral form of incomplete information. Employing the analytical tools used by Yang andLee(2014), we find that there is a unique equilibrium when parameters are within a reasonablerange. We utilize the rate of censoring and the average observed outcomes to identifymodel parameters. We suggest estimating the model by the maximum likelihood methodwith inner-loop fixed point iterations, which performs well in Monte Carlo experiments. Weapply the model to investigate municipal property tax rates in North Carolina and findsignificant competition effects.