Bayesian Econometrics Siddhartha Chib - PDF Download

(2025)

https://ebookultra.com/download/bayesian-econometrics-siddhartha-
chib/

Visit ebookultra.com today to download the complete set of

ebooks or textbooks
Here are some recommended products for you. Click the link to
download, or explore more at ebookultra.com

Bayesian Econometrics 1st Edition Gary Koop

https://ebookultra.com/download/bayesian-econometrics-1st-edition-
gary-koop/

The Oxford Handbook of Bayesian Econometrics John Geweke

(Editor)

https://ebookultra.com/download/the-oxford-handbook-of-bayesian-
econometrics-john-geweke-editor/

Bayesian Econometric Methods Gary Koop

https://ebookultra.com/download/bayesian-econometric-methods-gary-
koop/

Econometrics 1st Edition Thomas Andren

https://ebookultra.com/download/econometrics-1st-edition-thomas-
andren/
Econometrics 1st Edition Jan Tinbergen

https://ebookultra.com/download/econometrics-1st-edition-jan-
tinbergen/

Applied Econometrics 2nd Edition Dimitrios Asteriou

https://ebookultra.com/download/applied-econometrics-2nd-edition-
dimitrios-asteriou/

Bayesian reasoning and machine learning Barber D.

https://ebookultra.com/download/bayesian-reasoning-and-machine-
learning-barber-d/

Bayesian Modeling Using WinBUGS 1st Edition Ioannis

Ntzoufras

https://ebookultra.com/download/bayesian-modeling-using-winbugs-1st-
edition-ioannis-ntzoufras/

Bayesian Theory and Applications 1st Edition Paul Damien

https://ebookultra.com/download/bayesian-theory-and-applications-1st-
edition-paul-damien/
Bayesian Econometrics Siddhartha Chib Digital Instant
Download
Author(s): Siddhartha Chib, William Griffiths
ISBN(s): 9781848553088, 1848553080
Edition: None
File Details: PDF, 9.51 MB
Year: 2008
Language: english
ADVANCES IN ECONOMETRICS
Series Editors: Thomas B. Fomby and R. Carter Hill
Recent Volumes:

Volume 15: Nonstationary Panels, Panel Cointegration,

and Dynamic Panels, Edited by Badi Baltagi
Volume 16: Econometric Models in Marketing, Edited
by P. H. Franses and A. L. Montgomery
Volume 17: Maximum Likelihood Estimation of
Misspecified Models: Twenty Years Later,
Edited by Thomas B. Fomby and
R. Carter Hill
Volume 18: Spatial and Spatiotemporal Econometrics,
Edited by J. P. LeSage and R. Kelley Pace
Volume 19: Applications of Artificial Intelligence in
Finance and Economics, Edited by
J. M. Binner, G. Kendall and S. H. Chen
Volume 20A: Econometric Analysis of Financial and
Economic Time Series, Edited by
Dek Terrell and Thomas B. Fomby
Volume 20B: Econometric Analysis of Financial and
Economic Time Series, Edited by
Thomas B. Fomby and Dek Terrell
Volume 21: Modelling and Evaluating Treatment Effects
in Econometrics, Edited by Daniel L. Milli-
met, Jeffrey A. Smith and Edward J. Vytlacil
Volume 22: Econometrics and Risk Management,
Edited by Thomas B. Fomby, Knut Solna
and Jean-Pierre Fouque
 ADVANCES IN ECONOMETRICS VOLUME 23

BAYESIAN
ECONOMETRICS
EDITED BY

SIDDHARTHA CHIB
Olin Business School, Washington University

WILLIAM GRIFFITHS
Department of Economics, University of Melbourne

GARY KOOP
Department of Economics, University of Strathclyde

DEK TERRELL
Department of Economics, Louisiana State University

United Kingdom – North America – Japan

India – Malaysia – China
JAI Press is an imprint of Emerald Group Publishing Limited
Howard House, Wagon Lane, Bingley BD16 1WA, UK

First edition 2008

Copyright r 2008 Emerald Group Publishing Limited

Reprints and permission service

Contact: [email protected]

No part of this book may be reproduced, stored in a retrieval system, transmitted in any
form or by any means electronic, mechanical, photocopying, recording or otherwise
without either the prior written permission of the publisher or a licence permitting
restricted copying issued in the UK by The Copyright Licensing Agency and in the USA
by The Copyright Clearance Center. No responsibility is accepted for the accuracy of
information contained in the text, illustrations or advertisements. The opinions expressed
in these chapters are not necessarily those of the Editor or the publisher.

British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library

ISBN: 978-1-84855-308-8
ISSN: 0731-9053 (Series)

Awarded in recognition of
Emerald’s production
department’s adherence to
quality systems and processes
when preparing scholarly
journals for print
LIST OF CONTRIBUTORS

Michael K. Andersson Sveriges Riksbank, Stockholm, Sweden

Veni Arakelian Department of Economics, University of
Crete, Rethymno, Greece
Chun-man Chan Hong Kong Community College,
Kowloon, Hong Kong, China
Cathy W. S. Chen Department of Statistics, Feng Chia
University, Taiwan
Siddhartha Chib Olin Business School, Washington
University, St. Louis, MO
S. T. Boris Choy Discipline of Operations Management
and Econometrics, University of Sydney,
NSW, Australia
Michiel de Pooter Division of International Finance,
Financial Markets, Board of Governors
of the Federal Reserve System,
Washington, DC
Dipak K. Dey Department of Statistics, University of
Connecticut, Storrs, CT
Deborah Gefang Department of Economics, University of
Leicester, Leicester, UK
Richard Gerlach Discipline of Operations Management
and Econometrics, University of Sydney,
NSW, Australia
Paolo Giordani Research Department, Sveriges
Riksbank, Stockholm, Sweden
Jennifer Graves Department of Economics, University of
California, Irvine, CA
ix
x LIST OF CONTRIBUTORS

William Griffiths Department of Economics, University of

Melbourne, Vic., Australia
Ariun Ishdorj Department of Economics, Iowa State
University, Ames, IA
Liana Jacobi Department of Economics, University of
Melbourne, Vic., Australia
Ivan Jeliazkov Department of Economics, University of
California, Irvine, CA
Helen H. Jensen Department of Economics, Iowa State
University, Ames, IA
Sune Karlsson Swedish Business School, Örebo
University, Örebo, Sweden
Robert Kohn Department of Economics,
Australian School of Business,
University of New South Wales,
Sydney, Australia
Gary Koop Department of Economics, University of
Strathclyde, Glasgow, UK
Dimitris Korobilis Department of Economics, University of
Strathclyde, Glasgow, UK
Subal C. Kumbhakar Department of Economics, State
University of New York, Binghamton,
NY
Mark Kutzbach Department of Economics, University of
California, Irvine, CA
Roberto Leon-Gonzalez National Graduate Institute for Policy
Studies (GRIPS), Tokyo, Japan
Brahim Lgui Département de Sciences Économiques,
Université de Montréal, CIREQ,
Canada
Arto Luoma Department of Mathematics and
Statistics, University of Tampere,
Tampere, Finland
List of Contributors xi

Jani Luoto School of Business and Economics,

University of Jyväskylä, Jyväskylä,
Finland
William J. McCausland Département de Sciences Économiques,
Université de Montréal, CIREQ and
CIRANO, Montréal, QC, Canada
Nadine McCloud Department of Economics, The
University of the West Indies, Mona,
Kingston, Jamaica
Murat K. Munkin Department of Economics, University of
South Florida, Tampa, FL
Christopher J. O’Donnell School of Economics, University of
Queensland, Brisbane, Australia
Francesco Ravazzolo Norges Bank, Oslo, Norway
Vanessa Rayner School of Economics, University of
Queensland, Brisbane, Australia
Rene Segers Tinbergen Institute and Econometric
Institute, Erasmus University
Rotterdam, Rotterdam, The
Netherlands
Mike K. P. So Department of ISOM, Hong Kong
University of Science and Technology,
Kowloon, Hong Kong
Rodney Strachan School of Economics, The University of
Queensland, Brisbane, Australia
Sylvie Tchumtchoua Department of Statistics, University of
Connecticut, Storrs, CT
Dek Terrell Department of Economics, Louisiana
State University, Baton Rouge, LA
Justin Tobias Department of Economics, Purdue
University, West Lafayette, IN
Pravin K. Trivedi Department of Economics, Wylie Hall,
Indiana University, Bloomington, IN
xii LIST OF CONTRIBUTORS

Efthymios G. Tsionas Department of Economics, Athens

University of Economics and Business,
Athens, Greece
Herman K. van Dijk Tinbergen Institute and Econometric
Institute, Erasmus University
Rotterdam, Rotterdam, The
Netherlands
Wai-yin Wan School of Mathematics and Statistics,
University of Sydney, NSW, Australia
Arnold Zellner Graduate School of Business, University
of Chicago, Chicago, IL
BAYESIAN ECONOMETRICS: AN
INTRODUCTION

Siddhartha Chib, William Griffiths, Gary Koop and

Dek Terrell

ABSTRACT

Bayesian Econometrics is a volume in the series Advances in Econometrics

that illustrates the scope and diversity of modern Bayesian econometric
applications, reviews some recent advances in Bayesian econometrics, and
highlights many of the characteristics of Bayesian inference and
computations. This first paper in the volume is the Editors’ introduction
in which we summarize the contributions of each of the papers.

1. INTRODUCTION

In 1996 two volumes of Advances in Econometrics were devoted to Bayesian

econometrics. One was on computational methods and applications and the
other on time-series applications. This was a time when Markov chain Monte
Carlo (MCMC) techniques, which have revolutionized applications of
Bayesian econometrics, had started to take hold. The adaptability of MCMC
to problems previously considered too difficult was generating a revival of
interest in the Bayesian paradigm. Now, 12 years later, it is time for another
Advances volume on Bayesian econometrics. Use of Bayesian techniques has

Bayesian Econometrics
Advances in Econometrics, Volume 23, 3–9
Copyright r 2008 by Emerald Group Publishing Limited
All rights of reproduction in any form reserved
ISSN: 0731-9053/doi:10.1016/S0731-9053(08)23021-5
3
4 SIDDHARTHA CHIB ET AL.

become widespread across all areas of empirical economics. Previously

intractable problems are being solved and more flexible models are being
introduced. The purpose of this volume is to illustrate today’s scope and
diversity of Bayesian econometric applications, to review some of the recent
advances, and to highlight various aspects of Bayesian inference and
computations.
The book is divided into three parts. In addition to this introduction, Part I
contains papers by Arnold Zellner, and by Paolo Giordani and Robert Kohn.
In his paper ‘‘Bayesian Econometrics: Past, Present, and Future,’’ Arnold
Zellner reviews problems faced by the Federal Reserve System, as described
by its former chairman, Alan Greenspan, and links these problems to a
summary of past and current Bayesian activity. Some key contributions to the
development of Bayesian econometrics are highlighted. Future research
directions are discussed with a view to improving current econometric
models, methods, and applications of them.
The other paper in Part I is a general one on a computational strategy for
improving MCMC. Under the title ‘‘Bayesian Inference using Adaptive
Sampling,’’ Paolo Giordani and Robert Kohn discuss simulation-based
Bayesian inference methods that draw on information from previous samples
to build the proposal distributions in a given family of distributions. The
article covers approaches along these lines and the intuition behind some of
the theory for proving that the procedures work. They also discuss strategies
for making adaptive sampling more effective and provide illustrations for
variable selection in the linear regression model and for time-series models
subject to interventions.

2. MICROECONOMETRIC MODELING
Part II of the book, entitled ‘‘Microeconometric Modeling’’ contains
applications that use cross-section or panel data. The paper by Murat K.
Munkin and Pravin K. Trivedi, ‘‘A Bayesian Analysis of the OPES Model
with a Nonparametric Component: An Application to Dental Insurance and
Dental Care,’’ is a good example of how Bayesian methods are increasingly
being used in important empirical work. The empirical focus is on the impact
of dental insurance on the use of dental services. Addressing this issue is
complicated by the potential endogeneity of insurance uptake and the fact
that insurance uptake may depend on explanatory variables in a nonlinear
fashion. The authors develop an appropriate model which addresses both
these issues and carry out an empirical analysis which finds strong evidence
Bayesian Econometrics: An Introduction 5

that having dental insurance encourages use of dentists, but also of adverse
selection into the insured state.
MCMC simulation techniques are particularly powerful in discrete-data
models with latent variable representations. In their paper ‘‘Fitting and
Comparison of Models for Multivariate Ordinal Outcomes,’’ Ivan Jeliazkov,
Jennifer Graves, and Mark Kutzbach review several alternative modeling
and identification schemes for ordinal data models and evaluate how each
aids or hampers estimation using MCMC. Model comparison via marginal
likelihoods and an analysis of the effects of covariates on category probabili-
ties is considered for each parameterization. The methods are applied to
examples in educational attainment, voter opinions, and consumers’ reliance
on alternative sources of medical information.
In ‘‘Intra-Household Allocation and Consumption of WIC-Approved
Foods: A Bayesian Approach,’’ Ariun Ishdorj, Helen H. Jensen, and Justin
Tobias consider the Special Supplemental Nutrition Program for Women,
Infants, and Children (WIC) that aims to provide food, nutrition education,
and other services to at-risk, low-income children and pregnant, breastfeed-
ing, and postpartum women. They assess the extent to which the WIC
program improves the nutritional outcomes of WIC families as a whole,
including the targeted and nontargeted individuals within the household.
This question is considered under the possibility that participation in the
program (which is voluntary) is endogenous. They develop an appropriate
treatment–response model and conclude that WIC participation does not
lead to increased levels of calcium intake from milk.
A second paper that illustrates the use of Bayesian techniques for analyzing
treatment–response problems is that by Siddhartha Chib and Liana Jacobi.
In their paper ‘‘Causal Effects from Panel Data in Randomized Experiments
with Partial Compliance,’’ the authors describe how to calculate the causal
impacts from a training program when noncompliance exists in the training
arm. Two primary models are considered, with one model including a
random effects specification. Prior elicitation is carefully done by simulating
from a prior predictive density on outcomes, using a hold out sample.
Estimation and model comparison are considered in detail. The methods are
employed to assess the impact of a job training program on mental health
scores.
Basic equilibrium job search models often yield wage densities that do not
accord well with empirical regularities. When extensions to basic models are
made and analyzed using kernel-smoothed nonparametric forms, it is difficult
to assess these extensions via model comparisons. In ‘‘Parametric and
Nonparametric Inference in Equilibrium Job Search Models,’’ Gary Koop
6 SIDDHARTHA CHIB ET AL.

develops Bayesian parametric and nonparametric methods that are compar-

able to those in the existing non-Bayesian literature. He then shows how
Bayesian methods can be used to compare the different parametric and
nonparametric equilibrium search models in a statistically rigorous sense.
In the paper ‘‘Do Subsidies Drive Productivity? A Cross-Country Analysis
of Nordic Dairy Farms,’’ Nadine McCloud and Subal C. Kumbhakar
develop a Bayesian hierarchical model of farm production which allows for
the calculation of input productivity, efficiency, and technical change. The
key research questions relate to whether and how these are influenced by
subsidies. Using a large panel of Nordic dairy farms, they find that subsidies
drive productivity through technical efficiency and input elasticities,
although the magnitude of these effects differs across countries.
The richness of available data and the scope for building flexible models
makes marketing a popular area for Bayesian applications. In ‘‘Semipara-
metric Bayesian Estimation of Random Coefficients Discrete Choice
Models,’’ Sylvie Tchumtchoua and Dipak K. Dey propose a semiparametric
Bayesian framework for the analysis of random coefficients discrete choice
models that can be applied to both individual as well as aggregate data.
Heterogeneity is modeled using a Dirichlet process prior which (importantly)
varies with consumer characteristics through covariates. The authors employ
a MCMC algorithm for fitting their model, and illustrate the methodology
using a household level panel dataset of peanut butter purchases, and
supermarket chain level data for 31 ready-to-eat breakfast cereals brands.
When diffuse priors are used to estimate simultaneous equation models,
the resulting posterior density can possess infinite asymptotes at points of
local nonidentification. Kleibergen and Zivot (2003) introduced a prior to
overcome this problem in the context of a restricted reduced form
specification, and investigated the relationship between the resulting
Bayesian estimators and their classical counterparts. Arto Luoma and Jani
Luoto, in their paper ‘‘Bayesian Two-Stage Regression with Parametric
Heteroscedasticity,’’ extend the analysis of Kleibergen and Zivot to a
simultaneous equation model with unequal error variances. They apply their
techniques to a cross-country Cobb–Douglas production function.

3. TIME-SERIES MODELING
Part III of the volume is devoted to models and applications that use time-
series data. The first paper in this part is ‘‘Bayesian Near-Boundary Analysis
in Basic Macroeconomic Time-Series Models’’ by Michiel D. de Pooter,
Bayesian Econometrics: An Introduction 7

Francesco Ravazzolo, Rene Segers, and Herman K. van Dijk. The boundary
issues considered by these authors are similar to that encountered by Arto
Luoma and Jani Luoto in their paper. There are a number of models where
the use of particular types of noninformative priors can lead to improper
posterior densities with estimation breaking down at boundary values of
parameters. The circumstances under which such problems arise, and how
the problems can be solved using regularizing or truncated priors, are
examined in detail by de Pooter et al. in the context of dynamic linear
regression models, autoregressive and error correction models, instrumental
variable models, variance component models, and state space models.
Analytical, graphical, and empirical results using U.S. macroeconomic data
are presented.
In his paper ‘‘Forecasting in Vector Autoregressions with Many
Predictors,’’ Dimitris Korobilis introduces Bayesian model selection methods
in a VAR setting, focusing on the problem of drawing inferences from a
dataset with a very large number of potential predictors. A stochastic search
variable selection algorithm is used to implement Bayesian model selection.
An empirical application using 124 potential predictors to forecast eight U.S.
macroeconomic variables is included to demonstrate the methodology.
Results indicate an improvement in forecasting accuracy over model
selection based on the Bayesian Information Criteria.
In ‘‘Bayesian Inference in a Cointegrating Panel Data Model,’’ Gary
Koop, Robert Leon-Gonzalez, and Rodney Strachan focus on cointegration
in the context of a cointegrating panel data model. Their approach allows
both short-run dynamics and the cointegrating rank to vary across cross-
sectional units. In addition to an uninformative prior, they propose an
informative prior with ‘‘soft homogeneity’’ restrictions. This informative
prior can be used to include information from economic theory that cross-
sectional units are likely to share the same cointegrating rank without forcing
that assumption on the data. Empirical applications using simulated data
and a long-run model for bilateral exchange rates are used to demonstrate
the methodology.
Cointegration is also considered by Deborah Gefang who develops tests of
purchasing power parity (PPP) within an exponential smooth transition
(ESVECM) framework. The Bayesian approach offers a substantial
methodological advantage in this application because the Gibbs sampling
scheme is not affected by the multi-mode problem created by nuisance
parameters. Results based on Bayesian model averaging and Bayesian model
selection find evidence that PPP holds between the United States and each of
the remaining G7 countries.
8 SIDDHARTHA CHIB ET AL.

‘‘Bayesian Forecast Combination for VAR Models’’ by Michael K.

Andersson and Sune Karlsson addresses the issue of how to forecast a
variable (or variables) of interest (e.g., GDP) when there is uncertainty about
the dimension of the VAR and uncertainty about which set of explanatory
variables should be used. This uncertainty leads to a huge set of models. The
authors do model averaging over the resulting high-dimensional model space
using predictive likelihoods as weights. For forecast horizons greater than
one, the predictive likelihoods will not have analytical forms and the authors
develop a simulation method for estimating them. An empirical analysis
involving U.S. GDP shows the benefits of their approach.
In ‘‘Bayesian Inference on Time-Varying Proportions,’’ William J.
McCausland and Brahim Lgui derive a highly efficient algorithm for
simulating the states in state space models where the dependent variables are
proportions. The authors argue in favor of a model which is parameterized
such that the measurement equation has the proportions (conditional on the
states) following a Dirichlet distribution, but the state equation is a standard
linear Gaussian one. The authors develop a Metropolis–Hastings algorithm
which draws states as a block from a multivariate Gaussian proposal
distribution. Extensive empirical evidence indicates that their approach
works well and, in particular, is very efficient.
Christopher J. O’Donnell and Vanessa Rayner use Bayesian methodology
to impose inequality restrictions on ARCH and GARCH models in their
paper ‘‘Imposing Stationarity Constraints on the Parameters of ARCH and
GARCH Models.’’ Bayesian model averaging is used to resolve uncertainty
with regard to model selection. The authors apply the methodology to data
from the London Metals Exchange and find that results are generally
insensitive to the imposition of inequality restrictions.
In ‘‘Bayesian Model Selection for Heteroskedastic Models,’’ Cathy W. S.
Chen, Richard Gerlach, and Mike K. P. So discuss Bayesian model selection
for a wide variety of financial volatility models that exhibit asymmetries (e.g.,
threshold GARCH models). Model selection problems are complicated by
the fact that there are many contending models and marginal likelihood
calculation can be difficult. They discuss this problem in an empirical
application involving daily data from three Asian stock markets and
calculate the empirical support for their competing models.
Using a scale mixture of uniform densities representation of the Student-t
density, S. T. Boris Choy, Wai-yin Wan, and Chun-man Chan provide a
Bayesian analysis of a Student-t stochastic volatility model in ‘‘Bayesian
Student-t Stochastic Volatility Models via Scale Mixtures.’’ They develop a
Gibbs sampler for their model and show how their approach can be extended
Bayesian Econometrics: An Introduction 9

to the important class of Student-t stochastic volatility models with leverage.

The different models are fit to returns on exchange rates of the Australian
dollar against 10 currencies.
In ‘‘Bayesian Analysis of the Consumption CAPM,’’ Veni Arakelian and
Efthymios G. Tsionas show that Labadie’s (1989) solution to the CAPM can
be applied to obtain a closed form solution and to provide a traditional
econometric interpretation. They then apply Bayesian inference to both
simulated data and the Mehra and Prescott (1985) dataset. Results generally
conform to theory, but also reveal asymmetric marginal densities for key
parameters. The asymmetry suggests that techniques such as generalized
method of moments, which rely on asymptotical approximations, may be
unreliable.

REFERENCES
Kleibergen, F., & Zivot, E. (2003). Bayesian and classical approaches to instrumental variable
regression. Journal of Econometrics, 114, 29–72.
Labadie, P. (1989). Stochastic inflation and the equity premium. Journal of Monetary
Economics, 24, 195–205.
Mehra, R., & Prescott, E. C. (1985). The equity premium: A puzzle. Journal of Monetary
Economics, 15, 145–162.
BAYESIAN ECONOMETRICS: PAST,
PRESENT, AND FUTURE

Arnold Zellner

ABSTRACT
After briefly reviewing the past history of Bayesian econometrics and Alan
Greenspan’s (2004) recent description of his use of Bayesian methods in
managing policy-making risk, some of the issues and needs that he
mentions are discussed and linked to past and present Bayesian
econometric research. Then a review of some recent Bayesian econometric
research and needs is presented. Finally, some thoughts are presented that
relate to the future of Bayesian econometrics.

1. INTRODUCTION

In the first two sentences of her paper, ‘‘Bayesian Econometrics, The First
Twenty Years,’’ Qin (1996) wrote, ‘‘Bayesian econometrics has been a
controversial area in the development of econometric methodology. Although
the Bayesian approach has been constantly dismissed by many mainstream
econometricians for its subjectivism, Bayesian methods have been adopted
widely in current econometric research’’ (p. 500). This was written more than
10 years ago. Now more mainstream econometricians and many others have
adopted the Bayesian approach and are using it to solve a broad range of

Bayesian Econometrics
Advances in Econometrics, Volume 23, 11–60
Copyright r 2008 by Emerald Group Publishing Limited
All rights of reproduction in any form reserved
ISSN: 0731-9053/doi:10.1016/S0731-9053(08)23001-X
11
12 ARNOLD ZELLNER

econometric problems in line with my forecast in Zellner (1974), ‘‘Further, it

must be recognized that the B approach is in a stage of rapid development
with work going ahead on many new problems and applications. While this is
recognized, it does not seem overly risky to conclude that the B approach,
which already has had some impact on econometric work, will have a much
more powerful influence in the next few years’’ (p. 54).
See also, Zellner (1981, 1988b, 1991, 2006) for more on the past, present,
and future of Bayesian econometrics in which it is emphasized that all
econometricians use and misuse prior information, subjectively, objectively,
or otherwise. And it has been pointed out that Bayesian econometricians
learn using an explicit model, Bayes’ Theorem that allows prior information
to be employed in a formal and reproducible manner whereas non-Bayesian
econometricians learn in an informal, subjective manner. For empirical
evidence on the rapid growth of Bayesian publications over the years in
economics and other fields that will be discussed below see Poirier (1989,
1992, 2004) and Poirier (1991) for an interesting set of Bayesian empirical
papers dealing with problems in economics and finance.
In the early 1990s, both the International Society for Bayesian Analysis
(http://www.bayesian.org) and the Section on Bayesian Statistical Science of
the American Statistical Association (http://www.amstat.org) were formed
and have been very active and successful in encouraging the growth of
Bayesian theoretical and applied research and publications. Similarly, the
NBER-NSF Seminar on Bayesian Inference in Econometrics and Statistics
(SBIES) that commenced operation in 1970, has been effective for many years
in sponsoring research meetings, publishing a number of Bayesian books and
actively supporting the creation of ISBA and SBSS in the early 1990s. In
Berry, Chaloner, and Geweke (1996), some history of the SBIES and a large
number of Bayesian research papers are presented. Also, under the current
leadership of Sid Chib, very productive meetings of this seminar in 2004 and
2005 have been held that were organized by him and John Geweke. In August
2006, the European–Japanese Bayesian Workshop held a meeting in Vienna
organized by Wolfgang Polasek that had a very interesting program. In 2005,
the Indian Bayesian Society and the Indian Bayesian Chapter of ISBA had an
international Bayesian meeting at Varanasi with many of the papers
presented that have appeared in a conference volume. In September 2006, a
Bayesian research meeting was held at the Royal Bank of Sweden, organized
by Mattias Villani that attracted leading Bayesian econometricians from all
over the world to present reports on their current work on Bayesian
econometric methodology. And now, this Advances in Econometrics volume
features additional valuable Bayesian econometric research. And last, but not
Bayesian Econometrics: Past, Present, and Future 13

least, the International Society for Bayesian Analysis has commenced

publication of an online Bayesian journal called Bayesian Analysis; see
http://www.bayesian.org for more information about this journal with
R. Kass the founding editor and listings of articles for several years that
are downloadable. These and many more Bayesian activities that have taken
place over the years attest to the growth and vitality of Bayesian analysis in
many sciences, industries, and governments worldwide.

1.1. An Example of Bayesian Monetary Policy-Making

As an example of extremely important work involving the use of Bayesian

methodology and analysis, Alan Greenspan, former Chairman of the U.S.
Federal Reserve System presented an invited paper, ‘‘Risk and Uncertainty in
Monetary Policy’’ at the 2004 Meeting of the American Economic
Association that was published in the American Economic Review in 2004
along with very knowledgeable discussion by Martin Feldstein, Harvard
Professor of Economics and President of the National Bureau of Economic
Research, Mervyn King of the Bank of England, and Professor Janet L.
Yellen of the Haas School of Business, University of California, Berkeley.
The paper is notable in that it presents a comprehensive description of the
ways in which he approached and solved monetary policy problems ‘‘ . . . from
the perspective of someone who has been in the policy trenches’’ (p. 33).
Greenspan’s account should be of interest to Bayesians econometricians
and many others since he states, ‘‘In essence, the risk management approach
to policymaking is an application of Bayesian decision-making’’ (p. 37). In
addition, he writes, ‘‘Our problem is not, as is sometimes alleged, the
complexity of our policy-making process, but the far greater complexity of a
world economy whose underlying linkages appear to be continuously evol-
ving. Our response to that continuous evolution has been disciplined by the
Bayesian type of decision-making in which we have been engaged’’ (p. 39).
Feldstein (2004), after providing an excellent review of Greenspan’s
successful policy-making in the past wrote, ‘‘Chairman Greenspan empha-
sized that dealing with uncertainty is the essence of making monetary policy
(see also Feldstein, 2002). The key to what he called the risk-management
approach to monetary policy is the Bayesian theory of decision-making’’
(p. 42). After providing a brief, knowledgeable description of Bayesian
decision theory, Feldstein provides the following example to illustrate a case
of asymmetric loss in connection with a person making a decision whether to
carry an umbrella when the probability of rain is not high. ‘‘If he carries the
14 ARNOLD ZELLNER

umbrella and it does not rain, he is mildly inconvenienced. But if he does not
carry the umbrella and it rains, he will suffer getting wet. A good Bayesian
finds himself carrying an umbrella on many days when it does not rain. The
policy actions of the past year were very much in this spirit. The Fed cut the
interest rate to 1 percent to prevent the low-probability outcome of spiraling
deflation because it regarded that outcome as potentially very damaging
while the alternative possible outcome of a rise of the inflation rate from 1.5
percent to 2.5 percent was deemed less damaging and more easily reversed’’
(p. 42).
Mervyn King of the Bank of England commented knowingly about model
quality and policy-making, ‘‘Greenspan suggests that the risk-management
approach is an application of Bayesian decision-making when there is
uncertainty about the true model of the economy. Policy that is optimal in
one particular model of the economy may not be ‘robust’ across a class of
other models. In fact, it may lead to a very bad outcome should an
alternative model turn out to be true . . . Of course, although such an
approach is sensible, it is still vulnerable to policymakers giving excessive
weight to misleading models of the economy. . . . But, in the end, there is no
escaping the need to make judgments about which models are more plausible
than others’’ (pp. 42–43). These are indeed very thoughtful remarks about
problems of model uncertainty in making policy but do not recognize that
just as with Feldstein’s umbrella example above, a Bayesian analysis can
utilize posterior probabilities associated with alternative models that reflect
the quality of past performance that have been shown to be useful in
producing useful combined forecasts and probably will be helpful in dealing
with model uncertainty in policy-making.

1.2. Greenspan’s Policy-Making Problems

Below, I list and label important problems that Greenspan mentioned in

connection with his successful policy-making over the years that reveal his
deep understanding of both obvious and very subtle problems associated
with model-building, economic analyses, forecasting, and policy-making.
1. Structural changes: For example, ‘‘ . . . increased political support for
stable prices, globalization which unleashed powerful new forces of
competition, and an acceleration of productivity which at least for a time
held down cost pressures’’ (p. 33). ‘‘I believe that we at the Fed, to our
credit, did gradually come to recognize the structural economic changes
that we were living through and accordingly altered our understanding
Bayesian Econometrics: Past, Present, and Future 15

of the key parameters of the economic system and our policy stance . . . .
But as we lived through it, there was much uncertainty about the
evolving structure of the economy and about the influence of monetary
policy’’ (p. 33).
2. Forecasting: ‘‘In recognition of the lag in monetary policy’s impact on
economic activity, a preemptive response to the potential for building
inflationary pressures was made an important feature of policy. As a
consequence, this approach elevated forecasting to an even more
prominent place in policy deliberations’’ (p. 33).
3. Unintended consequences: ‘‘Perhaps the greatest irony of the past decade
is that the gradually unfolding success against inflation may well have
contributed to the stock price bubble of the latter part of the
1990s . . . The sharp rise in stock prices and their subsequent fall were,
thus, an especial challenge to the Federal Reserve’’ (p. 35).
‘‘The notion that a well-timed incremental tightening could have been
calibrated to prevent the late 1990s bubble while preserving economic
stability is almost surely an illusion. Instead of trying to contain a
putative bubble by drastic actions with largely unpredictable conse-
quences, we chose . . . to focus on policies to mitigate the fallout when it
occurs and, hopefully, ease the transition to the next expansion’’ (p. 36).
4. Uncertainty: ‘‘The Federal Reserve’s experiences over the past two
decades make it clear that uncertainty is not just a pervasive feature of
the monetary landscape; it is the defining characteristic of that
landscape. The term ‘‘uncertainty’’ is meant here to encompass both
‘Knightian uncertainty,’ in which the probability distribution of
outcomes is unknown, and ‘risk,’ in which uncertainty of outcomes is
delimited by a known probability distribution. In practice, one is never
quite sure what type of uncertainty one is dealing with in real time, and it
may be best to think of a continuum ranging from well-defined risks to
the truly unknown’’ (pp. 36–37).
5. Risk management: ‘‘As a consequence, the conduct of monetary policy in
the United States has come to involve, at its core, crucial elements of risk
management. This conceptual framework emphasizes understanding as
much as possible the many sources of risk and uncertainty that
policymakers face, quantifying those risks, when possible, and assessing
costs associated with each of the risks. In essence, the risk-management
approach to monetary policymaking is an application of Bayesian
decision-making’’ (p. 37).
6. Objectives: ‘‘This [risk management] framework also entails devising, in
light of those risks, a strategy for policy directed at maximizing the
16 ARNOLD ZELLNER

probabilities of achieving over time our goals of price stability and the
maximum sustainable economic growth that we associate with it’’ (p. 37).
7. Expert opinion: ‘‘In designing strategies to meet our policy objectives, we
have drawn on the work of analysts, both inside and outside the Fed,
who over the past half century have devoted much effort to improving
our understanding of the economy and its monetary transmission
mechanism’’ (p. 37).
8. Model uncertainty: ‘‘A critical result [of efforts to improve our
understanding of the economy and its monetary transmission mechan-
ism] has been the identification of a relatively small set of key
relationships that, taken together, provide a useful approximation of
our economy’s dynamics. Such an approximation underlies the statistical
models that we at the Federal Reserve employ to assess the likely
influence of our policy decisions.
However, despite extensive efforts to capture and quantify what we
perceive as the key macroeconomic relationships, our knowledge about
many of the important linkages is far from complete and, in all likelihood
will always remain so. Every model, no matter how detailed or how well
designed, conceptually and empirically, is a vastly simplified representa-
tion of the world that we experience with all its intricacies on a day-to-
day basis’’ (p. 37).
9. Loss structures: ‘‘Given our inevitably incomplete knowledge about key
structural aspects of an ever-changing economy and the sometimes
asymmetric costs or benefits of particular outcomes, a central bank
needs to consider not only the most likely future path for the economy,
but also the distribution of possible outcomes about that path. The
decision-makers then need to reach a judgment about the probabilities,
costs and benefits of the various possible outcomes under alternative
choices for policy’’ (p. 37).
10. Robustness of policy: ‘‘In general, different policies will exhibit different
degrees of robustness with respect to the true underlying structure of the
economy’’ (p. 37).
11. Cost–benefit analysis: ‘‘As this episode illustrates, policy practitioners
operating under a risk-management paradigm may, at times, be led to
undertake actions intended to provide insurance against [low prob-
ability] especially adverse outcomes . . . . The product of a low-
probability event and a potentially severe outcome was judged a more
serious threat to economic performance than the higher inflation that
might ensue in the more probable scenario’’ (p. 37).
Other documents randomly have
different content
unfurnished beds and were offered unnecessary apologies by our
polished French host and Japanese hostess at dinner because of the
looting that had taken place at the time his predecessor was carried
off. There was still a certain atmosphere of suppressed dread among
the few foreign residents, for none of them was sure how soon he
might become the next victim; but mankind quickly learns to live
without discomfort under many unpleasant circumstances.
Our soldier escort changed each day, and we were entertained each
evening with the long “face-saving” process that took place before the
detail could accept the gratuity we offered them. The struggle, which
we turned over to Chang as more finely versed in Chinese etiquette
than we, was particularly arduous on that first evening, for the
commander of the detachment was a real lieutenant, and instead of
the thirty-two vociferous and violent refusals which seemed to be
required of a mere sergeant or corporal before he accepted what he
really had no intention in the world of declining, the lieutenant was
still pushing back the detested silver with fine effect when we lost
count and went inside. Three Mexican dollars distributed among ten
men for a hot and arduous thirty-mile tramp for the possible
protection of a pair of unknown foreigners might not strike one of
our own “doughboys” as anything to write home about; but for men
whose daily pay was nothing like their share of this sum, and who
draw their pay much more often in theory than in practice, the
major’s insistence that they “have a good feed on us” could not really
have sounded so immoral to them as they pretended.
The second afternoon was still fairly young when we reached the
large walled town of what its residents, at least, called Lüngbau. The
escort was to stop here, but the sergeant in command thought he
could get permission to go on with us another twenty li, or get the
next detail to start at once, if we would let him go into town and see
the commander, while we continued around the edge of it, as most
through travel does in passing crowded walled cities. Near one of the
farther gates a soldier sent by the local commandant overtook us. His
chief, he said, could not send a detail on such short notice, and he did
not think it wise for us to go on without one. Bandits had been very
active in the immediate region ahead and might even have heard of
the “important” foreigners and be looking for them.
 All this moved us little, for both the major and I knew from long
experience that it is always the next stage of the journey that is
perilous for the traveler, never the one in which he actually is.
Besides, ten straggling, poorly equipped soldiers of the Chinese type
would scarcely prevent the bandits from adding us to their collection
if they really meant to do so. But we were reckoning without our
muleteers. They had already expressed a desire to stop in Lüngbau;
the report from the commandant made them doubly anxious to stay.
We were pooh-poohing their fears and deciding to order a new start
when, following the eye of one of them, I glanced up at the city gate
close beside us. It was a picturesque little portal, but that mere fact
would not of course have drawn the attention of a Chinese muleteer.
What had aroused his interest was two frail crates, thrown hastily
together of narrow strips of wood, fastened to the face of the gate on
either side just above the arch, and each containing a human head. I
had often read of such dainty decorations on Chinese city gates, on
those indeed of our medieval ancestors; but they had always seemed
far away and long ago, something pertaining to the “good old days,”
which a prosaic modern wanderer would never have the privilege of
seeing. To come upon them, therefore, in the present year of grace
and in the full light of the ordinary, every-day life about us, tacked up
against two torn posters depicting the delights and excellencies of a
widely known brand of cigarettes, was—well, was at least a pleasant
reminder that the picturesque customs of old China had not yet all
gone into the discard, that even the modern wanderer, if he wander
long and far enough, may still once in a blue moon come upon some
of those little details linking the phonographed, sewing-machined
world of to-day with the cave-man, which he has so often envied the
travelers of bygone centuries.
These two bandits, explained the soldier messenger, prompted
now and then by the solicitous crowd that always gathers in China
about any suggestion of a controversy, or of a foreigner, had been
caught four days before in the very town where we must spend the
night, if we persisted in pushing on. I suppose the crated heads were
what any ladylike person would have called a “gruesome sight,” but I
fear they struck me merely as interesting. In China one quickly and
unconsciously gets a sense of the cheapness of human life, so that
things which would ruin a night’s sleep at home are forgotten around
the next corner. The heads each lay on one ear in the bottom of their
open-work crates, half grinning down upon passers-by. Having a
southern exposure, they had already greatly profited by the three or
four days they had been separated from their original, evidently
rather youthful, possessors to disguise their identity. They were
yellow, not the mere yellow of the Chinese, who so far north are
scarcely yellow at all, but of the yellow of a pile of crude sulphur, of a
ripe lemon; and they were in that state in which even the most
careless housewife would quickly send a cut of meat out to be buried
—deep. Moreover—and all the writers on head-adorned gates I had
ever read had never given me a hint of this little detail—they were
swarming with flies, which seemed to consider this a particularly
luscious feast.
We yielded to the reluctance of our muleteers and turned back to a
near-by inn. The sun was still high enough for a stroll through the
extramural suburb, often the most crowded part of a Chinese town,
then across Lüngbau itself, and around a half-circuit of its broad
wall, from which we could look down into many of what in other
lands would have been domestic secrets. We saw by chance, for
instance, that the big sturdy man who had followed us into the inn-
yard on his knees, because he had carelessly frozen his feet off one
night, had a big family with whom to share the remnant of a roast leg
of lamb we had given him. Somewhere among the crowded bazaars
some one succeeded in telling us that bandits were worse in this
region because it was fairly rich and they could live on the country;
but the teeming life of Chinese streets certainly flowed on its even
way in complete indifference to those heads upon the gate and to the
dangers they stood for. What was still more to the point, there was
time to take a leisurely view of the silky-brown terraced mountains
that bounded the southern horizon, and to watch the unclouded sun
sink into a fiery furnace behind them.
But for that more or less forced stop at Lüngbau we should have
ended the mule-litter stage of our journey late on the third day.
However, that might have interfered with the major’s extraordinary
success as a hunter, which was not a commonplace, vulgar matter of
quantity, but of a finesse that even a Buddhist could have applauded.
We had waded through a considerable mountain pass—at least this
wearing down of roads into cañons sometimes appreciably shortens
a climb—and had come down a steep incline to the broad flat shores
of the Yellow River. Castor oil in its native state grew head-high for
some distance along the deep sandy trail; but what roused our
genuine interest was the fact that the lowland, half a mile wide,
between us and the river, was swarming with magnificent wild ducks,
and probably geese. The major snatched the shot-gun which some
trusting sky-pilot in Peking had unwisely lent him for the journey,
and strode out into a forty-acre field literally covered with the birds.
Now and again a great flock of them rose and circled in a great
curtain across the lower sky, but this mattered little, for there were
always more where those came from; in fact, had they all risen at
once, the air could scarcely have contained them.
Nothing of course could be more reprehensible, more dastardly, in
fact, than to breathe a breath of criticism upon the marksmanship of
a host, as it were, who has risen so high in the profession in which
marksmanship is so essential; and fortunately there is not the
slightest occasion to do so. For surely the failure to make a perfect
score can honestly be accounted for by the fact that the weapon used
was already doing service long before our forefathers began to laugh
at the idiot who fancied that some day some one would invent a
“horseless carriage.” If birds will have the decency to stay where they
are until the hunter can step on their tails before firing, such a
contrivance leaves nothing to be desired. But wild ducks and geese,
even in so rarely hunted a paradise as the Yellow River valley, are not
especially cordial to strangers; one might, indeed, almost charge
them with aloofness.
However, the major did fire at last, both barrels at once, so that at
least there would not be a second recoil to embitter his
disappointment, and in spite of the fact that he had not succeeded in
getting quite near enough to his quarry to make it really worth while
to throw the weapon itself after them. Strangely enough, one of the
birds gave every evidence of having been struck, or else of having had
the scare of its life. For instead of following its myriad fellows into
the now teeming air it ran erratically along the ground, with the
major and Chang, and, I believe, two or three of the muleteers,
possibly even the cook, in hot pursuit. The most fleet-footed of this
throng—I chanced at that moment to be hovering between turning
and not turning over with my litter, and hence can give no
trustworthy testimony on the subject—at length laid hands upon the
fugitive. If it had been struck, the shot, naturally, had not penetrated
the thick feathers; perhaps it had careened off its lightly clad skull
and left it a hazy view of the situation until it was for ever too late. At
any rate, the major has the distinction of having captured in perfect
health a magnificent specimen of the wild duck family, larger than
any domestic one and beautiful as a pheasant—with a shot-gun!
One of the soldiers carried it the rest of the morning, as another
had carried his hooded falcon the day before. Our entourage
attempted to convince us that such birds were not fit to eat, but its
superiority to a Thanksgiving turkey when it appeared before us
again next day suggested that they may merely have been offering,
Chinese fashion, to throw it away for us.
 CHAPTER XX
ON TO SIAN-FU

E arly on the fourth day we climbed up out of a great road cañon to

a mammoth stone archway that marks the boundary between
Honan and Shensi provinces, and immediately pitched down again
into another chasm of equal depth. Nor was there any improvement
in the fragile soil, in the endless lines of coolies going and coming, or
in the mangy beggars who squatted, loudly lamenting, in the dust
here and there along the sunken road all the way into strongly walled
Tungkwan. This important outpost of Shensi Province lies just over
the Honan border, on the Hoang Ho, yellow river indeed here at this
shallow season, across which one may see the loess hills of the
province of Shansi, just then suffering acutely from drouth. The
world had worn away from about the massive wall that surrounds the
town, as it does from about even a mud shrine in the loess country,
so that we had to climb again, rather stiffly, to reach the imposing
city gate that admitted us.
In strict duty, no doubt, the soldiers straggling about it should
have demanded our passports, which the Wai-chiao-pu in Peking,
whose privilege it is to look after “outside-country people,” had
smeared with half an acre of red-ink-stamped characters purporting
to be permission to visit five specified provinces, after which they
must, officially at least, be returned to Peking for further desecration.
But all the soldiers said was “Pien-tze,” and the Chinese visiting-
cards we produced in answer to that laconic request were evidently
all they wanted as proof of our identity. Since the major’s name
chances to begin with Ph, forcing him also to pass as Mr. Fei in
Chinese, we were at once taken for brothers, even in the face of
decided facial proof to the contrary, and passed on our way
unquestioned.
 The native pastor of the Fu-ying-tong, as the Chinese call a
Protestant mission, was not in town. But in the interior of China any
Caucasian passes at face value, at least until he has definitely been
proved a counterfeit, and we were soon installed in several dusty,
slightly furnished rooms of the rambling, temple-like compound,
while Chang and the cook explored the kitchen with the caretaker.
Had we arrived an hour earlier we might perhaps have gone on at
once and reached Sian-fu that same night. For, strange as it sounded,
there was a motor-bus line running more or less daily half-way
across “Hidden Shensi,” from Tungkwan to the capital. But the buses
started early in the morning; moreover, with all our dunnage we
should probably need a special car, and there was just then none in
town. If we really wished to go on next day, it would be best, they
told us, if the major in his official capacity should wire the Tuchun at
Sian-fu, to whom this little venture in less sluggish transportation
personally belonged. Meanwhile, there was the matter of settling
with our muleteers, and deciding how much cumshaw—without
which no transaction in China is considered properly closed—we
cared to give them. Tungkwan, too, was large and interesting
enough, with a wall which clambered for a long way along the crest of
a ridge high above us; but there is much sameness to most Chinese
cities, and this one seemed to offer nothing unique. But at least there
was something of that quality in a leisurely half-day for the ablutions,
razor-wieldings, resorting, and repose of which we were in arrears.
It did indeed require a special car for all our expedition, and even
at that I was forced to banish to the running-board the chauffeur’s
assistant, who habitually fills out the front seat of any public
conveyance of this sort in China. His duties seemed to be to crank the
car, to attend the wants of a perpetually parched radiator, to tinker
with the engine whenever there was the slightest chance to do so,
and in general to help the imported chauffeur to reduce the exiled
vehicle from a movable to an immovable object as soon as possible.
The driver had been brought all the way from Tientsin to grace
Shensi’s new enterprise, having been chosen evidently because of
what he did not know about automobile engines and their proper
manipulation, and therefore sure to be free from prejudice. If we
understood rightly, the conveyance had been carried piecemeal
through the loess cañons on mule-back, and no doubt some of the
parts had been assigned tasks for which they had never been trained.
But it is axiomatic that nothing short of total dissolution will prevent
a Ford truck from functioning, and less than two hours after this one
had been requested to start we were staggering in spasmodic jerks
out through the western city gate.
It is 290 li from Tungkwan to Sian-fu, almost exactly the same
distance as we had made in mule-litters in more than three days; so
that though we never attained breathless speed the journey felt rapid
by comparison. Once through the massive stone archway that
separated city from country, the going did not at first seem to be
appreciably better than the alleged road behind us; one gasped at the
temerity of any one, especially the timid Chinese, actually setting out
on so ideal a route for an obstacle race with the expectation of really
reaching a destination nearly a hundred miles away. But in time we
came to realize that it was what the Chinese consider an unusually
fine road. Loess had for the most part given way to a somewhat more
cohesive soil, and there were no real cañons. When he was Tuchun of
the province, the “Christian general” had built, mainly with soldier
labor—the two words seem incompatible in China—this raised
highway beside the old haphazard route all the way from the frontier
to the capital. His intentions had been excellent; but his funds were
limited, the soil available contains not a hint of stone or gravel, and
public coöperation was of course wholly lacking. The general had
done his best to replace this last un-Chinese asset by board signs set
up at frequent intervals along the way, with a warning that the
highroad was reserved for automobiles only, and that any other use
of it would be severely punished. His successor had evidently tried to
keep in force this unprecedented interference with Chinese freedom
of individual action, and his authority was certainly considerable, as
witness the fact that only here and there had the sign-boards even yet
been turned into fuel. But the Tuchun could scarcely be expected to
patrol the famous highway personally, and even at that he could not
have kept an eye on all parts of it at once. Therefore it was much
more densely thronged than the typical Chinese road down below it.
Donkeys, mules, pack-cattle, rickshaws—these often run the eighty-
seven miles in less than two and a half days, and make the round trip
in five, at a cost to the passenger of about two American dollars—
innumerable wheelbarrows, especially coolies in never ending
procession, prefer to ignore the sign-boards, if indeed even the slight
minority who can read them consider the prohibition as really
meant. Worst of all, whole regiments of the Tuchun’s own soldiers
were moving eastward, evidently in order to be more immediately
available to their real commander-in-chief, Wu Pei-fu, and more
than half the carts that carried these and their helter-skelter
paraphernalia were themselves frankly disobeying the placarded
order. These sharp-tired, two-wheeled contrivances are
magnificently designed for ruining a road, particularly one built
merely of earth, in the shortest possible time, and the result of the
trespassing of even the few thousand we passed during the one day
can readily be imagined. Then there were many spots where the
Chinese genius for never repairing anything until repair is absolutely
unavoidable manifested itself, and here and there some farmer had
frankly chopped the highway in two to make a passage for his
irrigation-ditch, a privilege as time-honored as China’s written
language.
 An old tablet in the
compound of the chief
mosque at Sian-fu,
purely Chinese in form,
except that the base has
lost its likeness to a
turtle and the writing is
in Arabic

This famous old portrait

of Confucius, cut on
black stone, in Sian-fu,
is said to be the most
 authentic one in
existence

A large town of cave-dwellers in the loess

country, and the terraced fields which support it

Samson and Delilah? This blind boy, grinding

grain all day long, marches round and round his
stone mill with the same high lifted feet and
bobbing head of the late Caruso in the opera of
that name

There was a certain satisfaction in seeing Nemesis, in the form of

our staggering, stuttering truck and the regular bus we sometimes
passed, sometimes dropped behind, overtake these lawbreakers
whom neither authority nor public opinion was able to curb. There
are few automobiles in Shensi Province, probably never more than
ten, and few of the throng along even its most nearly modern road
are in a frame of mind to meet one without what the “movie” world
calls “registering astonishment.” Most of them register a very
exaggerated form of it, which not only affects all the muscles of the
body but often manifests itself even in their domestic animals. With
their creaking wheelbarrows and a heavy head wind to hamper their
hearing, many permitted us almost to step on their heels before they
showed any inclination to give us the right of way; but this selfish
attitude was more than offset by the alacrity with which they did so
when once their minds were made up. At times the road immediately
ahead was so crowded with coolies and mule-drivers fleeing wild-
eyed at cross-purposes that we were forced to pause and even to halt
until the atmosphere had cleared itself sufficiently to make out the
ruts again. The conventional line of action was to abandon
wheelbarrow, animals, or pole-slung burdens at once and to go, quite
irrespective of destination. The road being from six to ten feet above
the surrounding country, barely wide enough in most places for one
car to run comfortably, with sheer sides and often a deep trench on
either hand, the punishment which overtook many of the trespassers
almost fitted the crime.
The coolies invariably grinned broadly or laughed aloud at their
own discomfiture, with that quick and genuine sense of humor which
transforms their rude, comfortless lives into a kind of perpetual
game and makes them, for all their many less agreeable qualities,
almost lovable. The few travelers of the haughtier classes, however,
strove to preserve the dignified deportment due their high standing,
even in the face of this ridiculous contrivance of inhuman speed from
the barbarous outside world. But they did not always succeed in
upholding all the precepts of Confucius. Among scores, probably
hundreds, who performed extraordinary feats of agility for our
beguilement during that day, the prize should be awarded to a man
we passed less than two hours out of Tungkwan. He was unusually
well dressed, as if of the wealthier merchant class, and was also
bound westward, seated high above his stout mule on the pile of
bedding and baggage in cloth saddle-bags which the well-to-do
Chinese long-distance traveler carries between himself and his
saddle. The mule under him was jogging comfortably along on the
edge of his own side of the road—which in most of China is the left—
though not on his own road, leaving us room to pass without more
than the hazard to which the brink-loving chauffeur habitually put
us. The animal showed every evidence of self-control and the ability
to handle the situation without mishap, but he reckoned without his
merely human master. We were perhaps ten yards behind them
when the man’s ears and brain coördinated and he looked around.
His first impulse was evidently to snatch the reins and attempt to
better the already perfect behavior of his mount, but the un-
Confucian speed with which we were lessening the already slight
distance between us confirmed him in the impression that it would
be safer to dismount with all seemly haste and leave the animal to its
own fate. Without losing an iota of his poise or dignity, or even his
position for that matter, the haughty gentleman calmly slipped off
his high seat on the ostensibly safe side, still in the right-angled
attitude of a sitting person—and admirably maintained that pose
until he disappeared, seat first, into a cross between a swamp and a
lake which unfortunately bordered the road at that particular place.
The chauffeur and I had the exclusive benefit of this portion of the
performance; the rest was reserved for those bouncing on our
baggage in the truck itself. When the major first became aware of the
existence of the haughty trespasser, it was in the form of a mere
head, topped by a dripping Chinese skull-cap, protruding from the
body of water alongside, and his last view of him as he receded into
the horizon was of a water-gushing figure clinging to the edge of the
road and shaking his open hand after the disappearing truck in the
gesture which the Chinese substitute for shaking the fist, while the
mule stood just where he had been abandoned, patiently awaiting
the good will of his temperamental master.

With the end of October it had turned distinctly colder, which was
fortunate; for the heat of Honan would have made the exertions
often required of us much less of a pastime than they were. Though it
had been smilingly new when it reached the province three months
before, our poor old truck resembled some maltreated, ill fed donkey
which even its heartless Chinese owner must soon turn out to die, yet
which faithfully toiled on to the very best of its ability. So long as it
hobbled along beneath him, the alleged chauffeur had not a worry in
the world; but whenever the slightest hill or sand a bit deeper than
usual brought us to a halt he was as helpless as a Hottentot with an
airplane. Having roared the engine almost out from under its hood,
as the only antidote suggesting itself to him, he sat supinely back in
his seat, at the end of his resources, and waited for some one else to
do something about it. Luckily there are always plenty of coolies
within call on any important route in China; but their natural
timidity increased in the presence of the strange snorting monster
that most of them had only seen hastily from a distance, and it
required the force of example to get them to approach and exert
themselves. Thus it came about that, though we had paid rather
generously for the transporting of our expedition from the boundary
to the capital of the province, we furnished the motive-power
ourselves for a considerable fraction of the journey.
For one short distance there were a few rocks and trees; but we
were soon in swirling loess again, dust so thick that it covered our
faces as with a white mask. Now and again we passed a high-walled
town, usually through the inevitable extramural suburb, a long line
of ramshackle mud huts, with men crowded together under the
thatch awnings, eating all manner of strange and unsavory-looking
native dishes. Even in the rare cases when we entered the city itself
there was nothing much more imposing. All morning long Hwa-
shan, second only to Tai-shan among the five sacred mountains of
China, walled off the southern horizon with its series of jagged
ranges, shaped not unlike a mammoth sleeping elephant, their
sunless northern slopes like a great perpendicular wall of beautiful
blue-gray color, topped by a wonderfully fantastic sky-line. About
2200 B.C. an early emperor of what was China in those days, with this
region as a nucleus, used to go to Hwa-shan to offer sacrifices and to
give audiences to his subjects, and the range has been sacred in
Chinese eyes ever since.
One might have fancied that a world war was on again, so often
were we held up by endless east-bound trains of soldiers, most of
them lounging in straw-roofed carts of two wheels, red banners with
white characters flying. It was noticeable that no one but the soldiers
had horses, of which most of China has been drained by her
swarming, autonomous militarists. Companies, even battalions, were
busily drilling here and there; two or three times we passed large
military camps in tents of wigwam shape, with a modernity about
them that looked incongruous against such backgrounds as a great
medieval, anachronistic city wall blackened by the centuries. Twice
we passed mule-carts laboring east or westward with the mails; all
day long a distorted line of telegraph-poles bearing a sagging wire or
two stretched haphazard into the distance.
The country grew a bit more rolling, with even less suggestion of
loess, as we neared Sian-fu. For miles the way was lined with
countless graves, ranging from dilapidated little cones of mud to
immense mounds. Bygone glories lay all over the landscape,
monument upon turtle-borne monument, so much more important
from the Chinese point of view than passable roads. At length the
great east gate of Shensi’s capital rose above the horizon, like some
huge isolated apartment-house, and just as the last daylight of
October flickered out we roared our jerky way up its broad main
street to our destination.

To say that I was disappointed in Sian-fu would be somewhat

overstating the case. But as nearly as I can recall the preconceived
picture, always so swiftly melting away in the glaring sunshine of
reality, I expected something more “wild and woolly,” something a
bit less like an abridged edition of Peking. Surely the city that was for
centuries the chief Manchu stronghold of the west, almost their
second capital, which had welcomed the cantankerous old dowager
fleeing before the justifiable wrath of the Western world, which had
seen such cruel and unnecessary bloodshed during the birth of the
republic, which had so often been the outpost on the edge of a great
Mohammedan rebellion, might at least have had some faint thrill,
some little hint of hidden danger, left to cheer up the jaded
wanderer. Instead, there was the same flat, placid city partly within
and partly without a mighty stone wall, swarming with the harmless
pullulations of petty traders, cheerfully enduring all the time-
honored discomforts of China, quite like those which lie scattered
like unto the sands of the sea in number over all the vast land that so
long gave Peking its undivided allegiance.
One stepped out of the big post-office compound where most
English-speaking foreigners find hospitality, upon that surprisingly
broad main street, to find it paved with something that has long
since lost the smoothness essential to comfortable rickshaw riding,
and lined for much of its length with houses unusual in northern
China, being of two stories. Along this one may come upon wood-
turners quite like those of Damascus in their methods—a little
shallow, frontless shop, a kind of Indian bow with a loose string for
lathe, a sometimes toe-supported chisel. Perhaps a householder
would find more interesting the long rows of wheelbarrows, filled
with huge chunks of that splendid anthracite so abundant and so
cheap in northwestern China, backed up against the curb and
patiently awaiting purchasers. But at the big bell-tower marking the
center of the city this broad street contracts to squeeze its way
through the resounding, dungeon-like arch, and never again regains
its lost breath. Here the paving is of big flagstones, worn so convex
that riding is not merely uncomfortable but well nigh impossible,
except to those who are inured by generations of such experiences, or
to whom the loss of “face” would be fatal. Others, at least new-
comers, may rather welcome this unspoken invitation to dismount
and stroll. For though there may be nothing in it not to be seen in a
hundred other places in China, “sights” are as compact in this busiest
street of Sian-fu as if they had purposely been gathered together here
as into a museum.
This main thoroughfare, and the one crossing it at right angles
beneath the bell-tower, cut the Shensi capital into its definite
quarters. The one on the right hand, as one comes in from the east,
is, or rather was, the Manchu city, given over now largely to great
open spaces; for here hundreds of the then ruling class jumped into
wells or otherwise violently did away with themselves, or were
violently done away with, to a number popularly estimated at more
than five thousand, when China last threw off an alien yoke and
announced itself a republic. Mere mud walls, with the brick or stone
facings gone to serve in some other capacity, mark most of the
compounds of what were perhaps for centuries Manchu palaces. Of
the palaces themselves there are few traces; dust and bare earth are
much more in evidence, though trees have survived to an extent
almost suggestive of Peking. Beyond this, filling the northwest
quarter, is the Mohammedan section, much more crowded and with
few open spaces—with none, perhaps, except they be public or
private courtyards. There are towns in western China where
Moslems must live outside the walls; but Sian-fu has been more
charitable toward her unabsorbable minority, and even during the
great rebellion they retained their intramural quarter, suffering little
more than constant surveillance, and no doubt occasional reviling.
Whether or not they would be driven back into it again if the
worshipers of Allah chose to live in some other part of town matters
not, for custom is as strong a bond with them as with their fellow-
Chinese, and whatever is Moslem about Sian-fu will be found in this
quarter, at least when bedtime comes. Here are all the mosques; here
are women who have scarcely stepped outside their compounds in a
generation, not even with covered faces; from here set forth each
morning the water-carriers, the muleteers, the common porters who
profess the faith of Medina. Outwardly the stroller through this
quarter may find it scarcely at all different from that Chinese half of
the city which lies to the south of its main thoroughfare. He may note
that the skullcaps of men and boys are more likely to be white than
black, that he sees only the most poverty-stricken class of women,
and not many of those, that many of the passers-by have liquid black
eyes and a very trifle more self-assertion, a slightly less lamb-like
expression than the common run of Chinese. Possibly it will occur to
him, too, that more of the little mutton-shop restaurants wide-
opening on the pulsating main street are on the north side of it, and
that the men who tend and patronize them also favor white skullcaps
and have something intangibly redolent of the Near East in features
and manners. But his eye is likely to be caught by more conspicuous
things along the stone-hard thoroughfare,—big whitish loaves of
bread nearly two feet in diameter and only two or three inches thick,
the splashes of color of myriad heaps of ripe persimmons, an
occasional woman with natural feet, relics not of Mohammedan but
of Manchu custom. There live half a million people within the city
walls and as many more in the environs, say unofficial guessers, and
about one in ten of these are Moslems and a bare two thousand
Manchus, the latter now mainly servants and recognizable to the
others by their Peking dialect and the somewhat different dress of
the women.

I picked up a man of standing in the Moslem faith one morning

and strolled out to the chief mosque. Outwardly there was nothing to
distinguish it from any Chinese compound, enclosing perhaps a
temple, to judge by the typical tile roofs and the tree-tops rising
above it. Indeed, the courtyard itself, beautiful with its old trees and
buildings, filled with the twitter of birds, which seemed to make it a
kind of sanctuary, restful and peace-loving in atmosphere, would not
easily have been recognized as containing anything but the usual
promiscuous mixture of the Buddhist, Taoist, and Confucian beliefs.
There were the same wooden tablets bearing two or three big
Chinese characters leaning out from under the eaves; the same
curious little figures adorned the upturned gables; there had been a
genuinely Chinese indifference about cleaning up after the birds. But
closer inspection brought out the underlying Mohammedanism. Not
far from the entrance stood a big stone tablet, purely Chinese in
form, even to the top-heavy dragon carvings; but the text that
covered it was not Chinese but Arabic. Here and there were other
stone-cut bits of that same tongue; “Yalabi” the not inhospitable
group that had gathered about me called it, though one or two
murmured something sounding like “Toorkee.” The beautiful little
three-story tower of pillar-borne roofs turned out to be the minaret
from which a Chinese muezzin singsongs the faithful to prayer.
Certainly it was leaving Chinese custom behind to be required,
however courteously, to leave my shoes at the door of the mosque
itself before I could step through the cloth-hung opening of a
building which up to that moment might have been anywhere in
China. But inside we had at last left China entirely behind. Not a
suggestion was to be seen of those myriad fantastic and demoniacal
figures which clutter up the interior of Chinese temples; the Koran’s
prohibition of graven images had been obeyed to the letter, and the
final sanctuary itself, where the men of Sian-fu’s northwest quarter
gather each Friday to turn their faces westward toward Mecca and
pray, was as severely beautiful in its Arabic style as if it had been
directly copied from the Alhambra.
The Islamites of China, or at least of Sian-fu, seem to have lost that
fierce inhospitality toward the unbeliever which makes it impossible
for those not of the faith even to enter many a famous mosque
farther west. Centuries of dwelling among them has given even the
intolerant Mussulmans much of the tolerance, or at least of the easy-
going, almost indifferent attitude, toward their religious
paraphernalia, which is so characteristic of the Chinese. There was
no objection, so long as I removed my shoes, to my wandering at will
in every part of the mosque, to stepping within the niche in the west
wall which takes on much of the sanctity of Mecca, not even to my
photographing it. The Chinese Moslems, indeed, seem never to have
heard of the Prophet’s implied injunction against permitting one’s
likeness to be transferred to paper; any refusal to stand before my
kodak among the group that trailed me about the compound was
probably due to mere Chinese superstitions, coupled with that dread
of giving their fellow-men the faintest opening for ridicule which is
one of the strongest traits in the Chinese character. For these fellows
were essentially Chinese, for all their religion, their swarthier
complexions and more Semitic noses; even the few among them
whose features would not have been conspicuous in a throng of
Turks or Arabs had all the little mannerisms, and to all appearances
the identical point of view, except in their alien faith, of their fellow-
countrymen.
Though there is no intermarriage between the Chinese
Mohammedans and their neighbors, the blood that runs in their
veins is largely the same. When the militant faith of Islam swept in
upon China from the west, at the time when it was spreading in all
directions, and was halted in our own only by the activity of Charles
Martel in France, the surest way of escaping the sword was to
embrace the new faith; and no one moves more quickly under the
inspiration of fear than the Chinese. Then, too, the conquerors
needed wives, or at least women, and took them from among the
conquered. Perhaps its greatest gains were during the inflow of trade
following the victories of Kublai Khan. For a long time it was, and
probably still is, the custom to adopt Chinese children into
Mohammedan homes. Thus the Turkish or Arabic features of the
invaders have been greatly modified, and even the few who have a
trace of these left seem to be greatly outnumbered by the purely
Chinese descendants of those who embraced the faith under
compulsion, so that even within a mosque compound it is often only
by inference, or the catching of some slight detail of custom or
costume, that the stranger can recognize a “Hwei-Hwei.” Foreigners
resident where the Mohammedans are numerous claim to be able to
tell one at sight, if only by a faintly more stiff-necked attitude toward
the rest of the world, a drawing of the line, beyond which he refuses
to be imposed upon, just a trifle closer to his own rights than do his
pacific Chinese fellows. Step into a temple at any time, and you will
receive nothing but profound courtesies from the Chinese, however
unwelcome you may be at that moment, say these experienced
Westerners; enter a mosque when a service is in progress, however,
and while the customary outward politenesses may not be lacking,
the atmosphere will be charged with something that says as distinctly
as a placard, “This is not the time to call.” I had a little hint of this
myself just before taking my departure. A high dignitary, what we
might call a bishop, wearing a strange blue costume and supported
as he tottered along by two lesser officials, issued from an inner court
on his way to perform some ceremony in a private family. My request
to photograph him was declined, not discourteously, but very
definitely and very promptly, as if, being a hadji who had made the
pilgrimage to Mecca, he was well aware of the ban which the Prophet
put on the making of likenesses, whatever might be the general
ignorance of it about him; and something gave me the feeling that if I
had attempted to act contrary to his wishes the smiling group of his
coreligionists about me would have found some unviolent Chinese
way of preventing me.
The non-believers among whom they live have, of course, other
terms than “Hwei-Hwei” for the Moslem minority, some of them so
far from complimentary as to be out of usage in any but the lowest
society. One of the less unkindly ones is “Pu-chih-jew-roe-ren,” the
don’t-eat-meat people. The Mohammedans have a name or two for
themselves and their religion so respectful and self-complacent that
their fellows decline to use them, so that the middle ground of
“Hwei-Hwei” is the one on which the two sections of the community
commonly meet. This term means something roughly corresponding
to “the associated people,” the single character for hwei meaning,
approximately, “association.” The Y. M. C. A. which functioned—
under a boyhood friend of the major, from Maine, it turned out—in
the quarter of Sian-fu opposite to that of the mosques was known as
the “Ching-Nien-Hwei,” which is quite the same as our own
abbreviation, except that our third letter, with all that it stands for, is
left out. This does not of course mean that the religious element is
lacking in the organization as it exists in Sian-fu—quite the contrary
seemed to be the case; but to stroll into the purely Chinese
compound, with its Chinese buildings, its board placards covered
with only Chinese characters, was also not to realize at once that one
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

ebookultra.com