Title: Some moment problems arising in financial econometrics
Authors: Christian Kleiber - Universitaet Basel (Switzerland) [presenting]
Abstract: The moment problem asks whether a distribution can be uniquely characterised by the sequence of its moments. Distributions that are not characterised by the sequence of their moments have long been known, e.g., the lognormal and certain generalised gamma distributions. We consider models from financial econometrics in which moment-indeterminate distributions may arise. Specifically, the generalised error distribution (GED) appearing in the EGARCH model is moment-indeterminate for some values of the parameters. Also, we show that one of the standard volatility models in financial econometrics, namely the stochastic volatility (SV) model, leads to return distributions that are moment-indeterminate. Perhaps somewhat unexpectedly, moment indeterminacy already arises in the classical discrete time SV model with lognormal latent volatility and independent multiplicative Gaussian noise.