Computing and Mathematical Sciences Colloquium

Monday April 17, 2017 4:00 PM

Testable Forecasts

Speaker: Assistant Professor Luciano Pomatto , Division of the Humanities and Social Sciences , California Institute of Technology
Location: Annenberg 105
Predictions about the future are typically evaluated through statistical tests. As shown by recent literature, many known tests are subject to adverse selection problems and are ineffective at discriminating between forecasters who are competent and forecasters who are uninformed but predict strategically. This paper presents necessary and sufficient conditions under which it is possible to discriminate between informed and uninformed forecasters. These conditions have a natural Bayesian interpretation. It is shown that optimal tests take the form of simple likelihood-ratio tests that compare forecasters' predictions against the predictions of a hypothetical outside observer. In addition, the paper also illustrates a novel connection between the problem of testing strategic forecasters and the classical Neyman-Pearson paradigm of hypothesis testing.
Series: Computing and Mathematical Sciences Colloquium Series
Contact: Carmen Nemer-Sirois at (626) 395-4561 carmens@cms.caltech.edu
Department of Computing + Mathematical Sciences