Statistica Neerlandica special issue on Benchmarking, Temporal Disaggregation and Reconciliation of Systems of Time Series just published: An interview with co-editor Baoline Chen

Features

  • Author: Statistics Views
  • Date: 01 Nov 2018
  • Copyright: Photograph appears courtesy of Baoline Chen

Statistica Neerlandica has published a special issue on Benchmarking, Temporal Disaggregation and Reconciliation of Systems of Time Series. The special issue includes ten articles, all addressing different aspects of the practical challenge in compiling systems of economic time series, which were brought together by co-editors Baoline Chen, Tommaso Di Fonzo and Nino Mushkudiani. 

Baoline Chen has a doctoral degree in economics from Indiana University, and started her career in academia. She spent seven years teaching economics at Rutgers University, New Jersey, where she earned her tenured associate professorship. During those years, her research focused on applied time series macroeconometrics, productivity analysis, and numerical techniques for solving dynamic economic models. After joining the US Bureau of Economic Analysis (BEA) in 2001, her research gradually transitioned to measurement issues in the compilation of official statistics, as described below.

Upon the publication of the recent special issue, StatisticsViews.com talks to Baoline Chen about the inspiration behind the special issue, her work on economic time series and her career at the US Bureau of Economic Analysis.

thumbnail image: Statistica Neerlandica special issue on Benchmarking, Temporal Disaggregation and Reconciliation of Systems of Time Series just published: An interview with co-editor Baoline Chen

1. What was the motivation behind putting together this special issue?

The idea of compiling a special issue on the subjects of benchmarking, temporal disaggregation and reconciliation of system of time series came from the editorial board of Statistica Neerlandica. The motivation was to bring together a collection of the most recent research that addresses a practical challenge in the compilation of systems of economic time series, especially at national accounts, official statistical agencies, and central banks, and to stimulate new research studies in this area.

2. How did the other two guest editors, Tommaso Di Fonzo and Nino Mushkudiani, become involved with the project?

Dr Nino Mushkudiani, who is a researcher at the Department of Methodology at Statistics Netherlands (CBS), was the first one invited in early January 2017 to be a guest editor for this special issue. Subsequently later that month, per her recommendation, Professor Tommaso Di Fonzo and I were invited to be co-editors. For several years Dr Mushkudiani has been working on the development and implementation of an optimization-based macro-integration technique to reconcile systems of time series data at CBS. Professor Di Fonzo, chairman of the Statistics Department at University of Padova, has been conducting innovative research on these topics, both in theories and applications, for more than three decades. I have known Dr Mushkudiani and Professor Di Fonzo for quite some time through our common interests in this area of research. It has been a very rewarding co-editorial experience to compile this special issue.

3. Can you tell us a little bit about the special issue and what it is about this topic that is of particular interest now?

This special issue brings together ten of the most recent contributions in the research on benchmarking, temporal disaggregation, and reconciliation of economic time series. It offers a mixture of both theoretical and applied studies that cover three broad topics: 1) temporal benchmarking of a single time series; 2) indirect estimation through temporal disaggregation; and 3) reconciliation of system of time series. A sequential benchmarking and an entropy-based benchmarking method are proposed for temporal benchmarking of a single time series. A dynamic regression model and a data driven approach are described and illustrated with national accounts data for temporal disaggregation; and an overview of practical feasibility, ease of use, and availability of software for alternative procedures for temporal disaggregation is also presented.

On reconciliation of time series with contemporaneous constraints, a maximum-likelihood procedure is proposed to tackle the problem of balancing a system of accounts. To account for contemporaneous constraints observed in time, a simultaneous least-squares procedure is demonstrated to show that reconciliation of a system of accounts in consecutive years can be done at very disaggregated level of detail; and an in-depth description of the practical experiences in the implementation of a macro-integration technique at a Netherlands statistical agency is presented. Moreover, some technical and practical issues in the reconciliation of systems of seasonally adjusted time series are addressed.

A more detailed overview of these studies can be found in the Preface for the special issue, which is free to read online.

4. Who should be reading this special issue and why?

Statisticians and practitioners whose work involves compilation of systems of economic time series using data with noise or data from different sources could potentially benefit from reading this special issue. I would think that statisticians and practitioners from official statistical agencies, national accounts and central banks will find these studies especially helpful, because they provide some potentially useful solutions to different aspects of the challenge in benchmarking, temporal disaggregation and reconciliation of systems of economic time series; and they also provide some helpful practical experiences in the implementation of a variety of statistical procedures.

5. What has been the most exciting development that you have worked on during your career at the US Bureau of Economic Analysis?

During the 17 years working at BEA, I have had the opportunity to work on a variety of interesting projects related to measurement issues in the compilation of official statistics. My research includes the introduction of an adaptive expectation framework to measure property and casualty insurance services in the national accounts; implementation of optimization frameworks for temporal disaggregation of sub-annual estimates in the US national accounts when only annual data and sub-annual indicators are available; a generalized least squares procedure for balancing the system of input-output accounts according to the estimated reliabilities of initial estimates; a simultaneous reconciliation procedure to reconcile a large system of time series accounts at various levels of disaggregation; estimation of capital and productivity in the manufacturing industries, and a state space framework for real-time estimation of future revision of GDP growth. Recently I have been involved in the research on now-casting early estimates of detailed GDP components in the national accounts and detecting and removing residual seasonality in the indirectly seasonally adjusted GDP estimates.

6. Are there people or events that have been particularly influential in your career?

There have been several people who have positively impacted my research at different stages of my career. I would like to mention three of them here. In the early years of my career, I benefited greatly from collaboration with Peter A. Zadrozny, who is a senior researcher at the US Bureau of Labor Statistics. As an experienced time series macro-econometrician, he helped me broadened the scope of my research from theoretical modelling and numerical analysis to applied macroeconomic time series analysis. Examples of our work include state space time series modelling for macroeconomic applications, multiple-step perturbation method for solving dynamic economic models, and estimation of vector autoregressive models with mixed-frequency data. These research experiences paved the way for my later research on measurement issues in official statistics at BEA.

For my research on the topics of benchmarking, temporal disaggregation and reconciliation, I feel very grateful to Professor Estela Bee Dagum and her former research collaborator Pierre A. Cholette, who was a senior methodologist at Statistics Canada. I benefited not only from reading their research that spans over several decades but also from numerous communications with Dr Cholette to answer my questions and to share their research experiences during my early research in this area. Their book Benchmarking, Temporal Distribution, and Reconciliation Methods for Time Series (2006, Springer Lecture Notes in Statistics, 186) helped me gain a comprehensive understanding of the subject matter.

Statistica Neerlandica: Benchmarking, Temporal Disaggregation, and Reconciliation of Systems of Time Series is available to read online. 

Related Topics

Related Publications

Related Content

Site Footer

Address:

This website is provided by John Wiley & Sons Limited, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ (Company No: 00641132, VAT No: 376766987)

Published features on StatisticsViews.com are checked for statistical accuracy by a panel from the European Network for Business and Industrial Statistics (ENBIS)   to whom Wiley and StatisticsViews.com express their gratitude. This panel are: Ron Kenett, David Steinberg, Shirley Coleman, Irena Ograjenšek, Fabrizio Ruggeri, Rainer Göb, Philippe Castagliola, Xavier Tort-Martorell, Bart De Ketelaere, Antonio Pievatolo, Martina Vandebroek, Lance Mitchell, Gilbert Saporta, Helmut Waldl and Stelios Psarakis.