Showing papers by "Neil Shephard published in 2015"

Comparing sample survey measures of English earnings of graduates with administrative data during the Great Recession

Abstract: This paper compares survey based labour earnings data for English graduates, taken from the UK’s Labour Force Survey (LFS), with the UK Government administrative sources of o?cial individual level earnings data. This type of administrative data has few sample selection issues, is substantially longitudinal and its large samples mean the earnings of subpopulations can be potentially studied (e.g. those who study a speci?c subject at a speci?c university and graduate in a speci?c year). We ?nd that very broadly the LFS and administrative data show a similar distribution of graduates’ earnings. However, the administrative data has considerably less gender disparity, higher high quantiles and more time series persistence. We also report on how the distribution of graduate and non-graduate earnings fell during each year of the Great Recession.

Unobserved Components and Time Series Econometrics

Abstract: This volume presents original and up-to-date studies in unobserved components (UC) time series models from both theoretical and methodological perspectives. It also presents empirical studies where the UC time series methodology is adopted. Drawing on the intellectual influence of Andrew Harvey, the work covers three main topics: the theory and methodology for unobserved components time series models; applications of unobserved components time series models; and time series econometrics and estimation and testing. These types of time series models have seen wide application in economics, statistics, finance, climate change, engineering, biostatistics, and sports statistics. The volume effectively provides a key review into relevant research directions for UC time series econometrics and will be of interest to econometricians, time series statisticians, and practitioners (government, central banks, business) in time series analysis and forecasting, as well to researchers and graduate students in statistics, econometrics, and engineering. Contributors to this volume - Craig Ansley Fabio Busetti, Bank of Italy Piet de Jong, Macquarie University Francis X. Diebold, University of Pennsylvania Giuliano De Rossi, Macquarie Securities Gabriele Fiorentini, University of Florence Simon J. Godsill, University of Cambridge Andrew Harvey, University of Cambridge Jouni Helske, University of Jyvaskyla Siem Jan Koopman, VU University Amsterdam Tatjana Lemke, University of Cambridge Alessandra Luati, University of Bologna Jun Ma, University of Alabama Geert Mesters, Universitat Pompeu Fabra Charles R. Nelson, University of Washington Jukka Nyblom, University of Jyvaskyla Pilar Poncela, Universidad Autonoma de Madrid Tommaso Proietti, University of Tor Vergata Esther Ruiz, Universidad Carlos III de Madrid Enrique Sentana Neil Shephard, Harvard University Andrea Stella, Federal Reserve System James H. Stock, Harvard University Kamil Yilmaz, Koc University

Moment conditions and Bayesian nonparametrics

Abstract: Models phrased though moment conditions are central to much of modern inference. Here these moment conditions are embedded within a nonparametric Bayesian setup. Handling such a model is not probabilistically straightforward as the posterior has support on a manifold. We solve the relevant issues, building new probability and computational tools using Hausdorff measures to analyze them on real and simulated data. These new methods which involve simulating on a manifold can be applied widely, including providing Bayesian analysis of quasi-likelihoods, linear and nonlinear regression, missing data and hierarchical models.

Likelihood inference for exponential-trawl processes

Abstract: Integer-valued trawl processes are a class of serially correlated, stationary and infinitely divisible processes that Ole E. Barndorff-Nielsen has been working on in recent years. In this Chapter, we provide the first analysis of likelihood inference for trawl processes by focusing on the so-called exponential-trawl process, which is also a continuous time hidden Markov process with countable state space. The core ideas include prediction decomposition, filtering and smoothing, complete-data analysis and EM algorithm. These can be easily scaled up to adapt to more general trawl processes but with increasing computation efforts.