Published and unpublished papers

Below are some papers and other work available for downloading.

Thesis

Title: Time Varying Parameter Models for Inflation and Exchange Rates
Author: Charles S. Bos
Contents: See the more elaborate description.
Focus: Both classical and Bayesian
Date: September 2001
Published: Tinbergen Institute Research series, TI 256
Links: Introduction, WebDOC PDF document (1.5MB).

Publications and working papers

Title: A Bayesian Analysis of Unobserved Component Models using Ox
Authors: Charles S. Bos
Contents: The analysis of unobserved component models using a Bayesian approach and the Ox language is discussed in detail, including a series of example programs.
Focus: Bayesian
Date: March 2011
Published: Abstract and TI 11-048/4 (PDF, 513kb), and in the Journal of Statistical Software, Vol 41, issue 13.
Title: Relating Stochastic Volatility Estimation Methods
Authors: Charles S. Bos
Contents: A wealth of methods for estimating stochastic volatility models exists. This article tries to put them onto a similar footing, in order to compare their differences and commonalities in detail. This has the side-effect that the general methodology of estimating these models becomes clear.
Focus: Both classical and Bayesian
Date: March 2011
Published: Abstract and TI 11-049/4 (PDF, 469kb); forthcoming in the Handbook of Volatility Models and Their Applications (2012), Wiley.
Title: Models with Time-varying Mean and Variance: A Robust Analysis of U.S. Industrial Production
Authors: Charles S. Bos and Siem Jan Koopman
Contents: Investigates what components are needed for estimating U.S. industrial production, in a manner which is robust to suddenly changing situations.
Focus: Classical
Date: January 2010
Published: Abstract and TI 10-017/4 (PDF, 319kb).
Title: Spot Variance Path Estimation and its Application to High Frequency Jump Testing
Authors: Charles S. Bos, Pawel Janus and Siem Jan Koopman
Contents: Proposes a novel method to estimate variance in high frequency data, robust to microstructure noise, jumps and leverage.
Focus: Non-parametric
Date: December 2009
Published: Abstract and TI 09-110/4 (PDF, 1.2Mb), forthcoming in the Journal of Financial Econometrics.
Title: Does the Canadian Economy Suffer from Dutch Disease?
Authors: Michel Beine, Charles S. Bos, Serge Coulombe
Contents: Disentangles the evolution of Canadian currency and energy/commodity prices, to study the case of `Dutch disease' occuring in the Canadian economy due to the increased export of oil.
Focus: Bayesian
Date: January 2009, revised November 2009
Published: Abstract of discussion paper SSRN ID1336635 (PDF, 185kb), revision abstract and TI 09-096/4 (PDF, 743kb).
Title: Model-based Estimation of High Frequency Jump Diffusions with Microstructure Noise and Stochastic Volatility
Authors: Charles S. Bos
Contents: Take a high frequency financial time series, and estimate a discretisation at the 5 minute frequency of a diffusion model including jumps, microstructure and stochastic volatility with it, in a parametric approach.
Focus: Bayesian
Date: January 2008
Published: Abstract, discussion paper TI 08-011/4 (PDF, 1.1Mb) at the Tinbergen Institute.
Title: Long Memory Modelling of Inflation with Stochastic Variance and Structural Breaks
Authors: Charles S. Bos, Siem Jan Koopman and Marius Ooms
Contents: Introduce a novel method of estimating a combination of long memory in the mean and stochastic volatility, applied to postwar U.S. inflation.
Focus: Classical
Date: December 2007
Published: Abstract, discussion paper TI 07-099/4 (PDF, 340Kb) at the Tinbergen Institute.
Title: Dynamic Correlations and Optimal Hedge Ratios
Authors: Charles S. Bos and Phillip Gould
Contents: Hedge two assets using a variety of dynamic correlation models, including GARCH and SV types, and find a minimum variance hedge portfolio.
Focus: Both classical and Bayesian
Date: February 2007
Published: Abstract, discussion paper TI 07-025/4 (PDF, 355Kb) at the Tinbergen Institute.
Title: The Impact of Central Bank FX Interventions on Currency Components
Authors: Michel Beine, Charles S. Bos and Sébastien Laurent
Contents: Through extraction of currency components the impact of interventions is measured on a country- and central bank-specific basis
Focus: Bayesian
Date: January 2007
Published: Published in the Journal of Financial Econometrics, (2007), vol 5/1, pages 154-183. Older version: Abstract, discussion paper TI 05-103/4 (PDF, 499Kb) at the Tinbergen Institute.
Title: Inference for Adaptive Time Series Models: Stochastic Volatility and Conditionally Gaussian State Space form
Authors: Charles S. Bos and Neil Shephard
Contents: This paper enlarges the class of State Space models to allow for stochastic volatility in each of the disturbances. The existing methodology of sampling from the posterior for such models is improved upon, such that the estimation becomes viable.
Focus: Bayesian
Date: January 2004
Published: In Econometric Reviews, (2006), vol 25/2-3, pages 219-244. Older version of abstract, discussion paper TI 04-015/4 (PDF, 954Kb) at the Tinbergen Institute.
Title: On Model Selection Criteria as a Starting Point for Sequential Detection of Non-linearity
Authors: Charles S. Bos and Ana Justel
Contents: This article discusses the paper `Detecting Non Linearity in Time Series by Model Selection Criteria' by D. Peña and J. Rodriguez.
Focus: Classical
Date: October 2005
Published: In the International Journal of Forecasting, (2005) vol 21/4, pages 749-754, with underlying programs.
Title: Time Series Modelling using TSMod 3.24
Authors: Charles S. Bos
Contents: Review of the TSMod package by James Davidson
Focus: Classical
Date: November 2003
Published: International Journal of Forecasting (2004), vol 20/3, pages 512-522 old version as a discussion paper at the Tinbergen Institute, with underlying datasets and programs
Title: Adaptive Radial-based Direction Sampling: Some Flexible and Robust Monte Carlo Integration Methods
Old title: Adaptive Polar Sampling with an application to a Bayes measure of Value-at-Risk
Authors: Luc Bauwens, Charles S. Bos, Herman K. van Dijk and Rutger D. van Oest
Contents: A class of more robust sampling method, using adaptive radial-based schemes, is developed.
Focus: Bayesian
Date: September 2003
Published: Journal of Econometrics (2004) vol 123/2, pages 201-225, previous version as EI 2003-22/a
Title: State Space Models with a Common Stochastic Variance
Authors: Siem Jan Koopman and Charles S. Bos
Contents: State space models including a common stochastic variance component are estimated using (simulated) maximum likelihood, with application.
Focus: Classical
Date: September 2003
Published: Journal of Business & Economic Statistics (2004), vol 22/3, pages 346-357. Previous version as TI 02-113/4.
Links: Abstract, software, TI 02-113/4 (400 kb, old version), JBES pdf.
Title: Inflation interval forecasting by long memory regression models
Authors: Charles S. Bos, Philip Hans Franses and Marius Ooms
Prepared for: 20th International Symposium on Forecasting (ISF), June 21-24, 2000, Lisbon, Portugal
Contents: For monthly US inflation rates, the out-of-sample performance over horizons between 1 and 24 months of short memory ARMA(1,1), ARIMA(1,1,1) and long memory ARFIMA(0,d,0) models are compared. Explanatory variables are introduced. Forecast performance of the long memory model is better when the coverage of the forecast intervals is considered.
Focus: Classical
Date: September 2002
Published: Published in the International Journal of Forecasting (2002), vol 18/2, pages 243-262.
Links: Abstract, TI-01-029/4 (Older version, PDF 400kb)
Title: Daily Exchange Rate Behaviour and Hedging of Currency Risk
Authors: Charles S. Bos, Ronald J. Mahieu and Herman K. van Dijk
Prepared for: Conference on Inference & Decision Making, Rotterdam
Contents: A set of 7 different but well-known alternative models are used for evaluating the decision of an investor who is considering to hedge his currency risk. The final returns of several strategies, based on a utility function or on the Value-at-Risk or the Sharpe ratio, are compared. Underlying data is the US Dollar/German D-Mark daily exchange rate, with corresponding interest rates.
Focus: Bayesian
Date: November 2000
Published: Older version as Econometric Institute report EI2000-25/a, published in the Journal of Applied Econometrics, (2000), Vol. 15/6.
Links: Abstract, EI2000-25/a (PDF, 539 KB), JAE 15/6, p 671-696 (PDF, 327 KB)
Title: Long Memory and Level Shifts: Re-Analyzing Inflation Rates
Authors: Charles S. Bos, Philip Hans Franses and Marius Ooms
Contents: The relation between the occurrence of level shifts in a univariate time series and the estimated degree of fractional integration is investigated. Inflation rates for the G7 countries are modelled both including and excluding the possibility of breaks, to check the possible importance of shifts in the model
Focus: Classical
Date: August 1999
Published: Empirical Economics, Vol 24, pp. 427-449, 1999.

Other publications

Title: Ox programming and Bayes
Authors: Charles S. Bos
Contents: A short review of Ox and its usage in Bayesian research
Focus: Bayesian
Date: March 2005
Published: Published in the ISBA Bulletin (2005, Vol 12 no 1), available in PDF
Title: Explaining Adaptive Radial-based Direction Sampling
Authors: Luc Bauwens, Charles S. Bos, Herman K. van Dijk and Rutger D. van Oest
Contents: To-the-point explanation of the ARDS method, see also the main article on Adaptive Radial-based Direction Sampling
Focus: Bayesian
Date: October 2003
Published: In the Proceedings of the American Statistical Association
Title: A Comparison of Marginal Likelihood Computation Methods
Authors: Charles S. Bos
Prepared for: Computational Statistics 2002, August 24-28, Berlin, Germany
Contents: An overview of methods for computing the marginal likelihood is given, and compared for their accuracy
Focus: Bayesian
Date: August 2002
Published: As TI 02-084/4 and in the COMPSTAT 2002 - Proceedings in Computational Statistics, pages 111-117.
Links: Abstract, TI 02-084/4 (162 kb)
Title: On the Variation of Hedging Decisions in Daily Currency Risk Management
Authors: Charles S. Bos, Ronald J. Mahieu and Herman K. van Dijk
Prepared for: 6th World Meeting of the International Society for Bayesian Analysis, May 2000, Crete
Contents: This paper is an extension to Daily Exchange Rate Behaviour and Hedging of Currency Risk, focussing on the hedging decisions taken for the different models intended at hedging currency risk.
Focus: Bayesian
Date: May 2000
Published: In Bayesian Methods with Applications to Science, Policy, and Official Statistics: Proceedings of ISBA 2000, and published as Econometric Institute Report 2000-20/a.
Links: Abstract, EI2000-20/a (PDF, 294 kB), ISBA (PDF, 161 kB)
Title: Adaptive Polar Sampling with an application to a Bayes measure of Value-at-Risk
Old title: Adaptive Polar Sampling: A New MC Technique for the Analysis of Ill-behaved Surfaces
Authors: Luc Bauwens, Charles S. Bos and Herman K. van Dijk
Prepared for: 6th Valencia International Meeting on Bayesian Statistics, Valencia and Econometric Society European Meeting '99, Santiago de Compostela
Contents: A more robust sampling method, as an alternative to Metropolis-Hastings sampling and Gibbs sampling, is developed.
Focus: Mainly Bayesian, slightly classical
Date: 1999-2001
Published: Previous version as TI 98-071/4, present version as EI-9940/a, short note in `Proceedings in Computational Statistics, 2000'
Links: Abstract, EI-9940/a (PS, 535 KB)

[Main page] [Top] [Next: Software]
Last change: 20/10/2011