Calendrier du 21 janvier 2015
PSI-PSE (Petit Séminaire Informel de la Paris School of Economics) Seminar
Du 21/01/2015 de 17:00 à 18:30
MSE-salle B2.2
RISTINIEMI Annukka ()
A Laffer curve of sovereign debt and liquidity
WIP (Work in progress) Working Group
Du 21/01/2015 de 13:00 à 14:30
Salle de réunion du département des Sciences Sociales
DUPRAZ Yannick(University College Dublin)
VILLAR Paola(PSE)
LEROUX Mélissa()
*
La Francafrique: Does part of French Aid to Africa Finance French Electoral Campaigns?
Bayesian Model Selection and Big Data: Construction of high-frequency well-being indices from Google Trends Data
Preference for Income Unobservability and Informal Redistribution in Community and Kin Networks: A Lab-in-the-field Experiment in Senegal
Job Market Seminar
Du 21/01/2015 de 12:30 à 13:45
Campus jourdan,Bâtiment G, Rez de chaussée, Salle 8
KETZ Philipp (BROWN UNIVERSITY)
Testing near or at the Boundary of the Parameter Space
Statistical inference about a scalar parameter is often performed using the two-sided ttest. In extremum problems, where the estimator satises the restrictions on the parameter space - such as the nonnegativity of a variance parameter -, the test suers from size distortions when the true parameter vector is near or at the boundary of the parameter space. Nevertheless, the two-sided t-test continues to be used when estimates are found to be close to the boundary. This can be attributed to a lack of inference procedures that appropriately account for boundary eects on the asymptotic distribution of the estimator. To address this issue, we propose an estimator that is asymptotically normally distributed, even when the true parameter vector is near or at the boundary, and the objective function is not dened outside the parameter space. The novel estimator allows the implementation of several existing testing procedures and a new test based on the Conditional Likelihood Ratio statistic (CLR). Compared to the existing procedures, the new test is easy to implement and has good power properties. Moreover, it oers power advantages over the two-sided t-test, when the latter controls size. We also show the test to be admissible when inference is performed with respect to a scalar parameter. We apply the test to the random coecients logit model using data on the European car market and nd more evidence of heterogeneity in consumer preferences than suggested by the two-sided t-test.