The role of time‐varying rare disaster risks in predicting bond returns and volatility

• Author: Rangan Gupta,Mark E. Wohar,Tahir Suleman
• DOI: http://doi.org/10.1002/rfe.1051
• Date: 01 July 2019
• Published date: 01 July 2019

Rev Financ Econ. 2019;37:327–340. wileyonlinelibrary.com/journal/rfe

|

327

© 2018 The University of New Orleans

1

|

INTRODUCTION

Following the early work of Rietz (1988), a growing number of calibrated theoretical models have recently provided evidence

of the ability of rare disaster risks in affecting movements (returns and volatility) of asset prices (see e.g., Barro, 2006, 2009;

Barro & Jin, 2011; Barro & Ursúa, 2008, 2009, 2012; Farhi & Gabaix, 2016; Gabaix, 2012; Gourio, 2008a,b, 2012; Lewis &

Liu, 2017; Nakamura, Steinsson, Barro, & Ursúa, 2013; Wachter, 2013).

A major obstacle, however, to full- fledged empirical verification of the rare disaster models is that individual countries

rarely face actual major disasters, resulting in a small sample problem inherent in the use of actual rare disasters, which in turn,

explains the reliance of the above- mentioned papers on calibration. In this regard, Berkman, Jacobsen, and Lee (2011, 2017),

provides a solution to the small sample problem that would make empirical estimation of these models possible by focusing on

a much larger sample of potential disasters (international political crises) that are likely to cause changes in perceived rare disas-

ter probabilities. Using a detailed database of all international political crises, namely the International Crisis Behavior project

(ICB) database developed by the Center for International Development and Conflict Management, Berkman et al. (2011, 2017)

provides empirical evidence that various international crises over the period of 1918 to 2006, does indeed affect equity returns

and volatility of large number of developed and emerging economies.

Using an extended version of the ICB database, the goal of this paper is to examine the predictive power of rare- disaster

risks for the return and volatility dynamics of ten- year government bond returns of the United States over the monthly pe-

riod of 1918:01–2013:12. As a matter of comparison, we also analyze the same for long- term government bonds for another

developed country (UK) over the period of 1933:01–2013:12 and an emerging market (South Africa) covering the period

1918:01–2013:12.

Received: 27 February 2018

|

Revised: 21 August 2018

|

Accepted: 14 September 2018

DOI: 10.1002/rfe.1051

ORIGINAL ARTICLE

The role of time- varying rare disaster risks in predicting bond

returns and volatility

Rangan Gupta1,2

|

Tahir Suleman3,4

|

Mark E. Wohar5,6

1Department of Economics,University of

Pretoria, Pretoria, South Africa

2IPAG Business School, Paris, France

3School of Economics and Finance,Victoria

University of Wellington, Wellington, New

Zealand

4School of Business,Wellington Institute of

Technology, Wellington, New Zealand

5College of Business

Administration,University of Nebraska at

Omaha, Omaha, Nebraska

6School of Business and

Economics,Loughborough University,

Leicestershire, UK

Correspondence

Mark E. Wohar, College of Business

Administration, University of Nebraska at

Omaha, Omaha, NE.

Email: mwohar@unomaha.edu.

Abstract

This paper aims to provide empirical evidence to the theoretical claim that rare dis-

aster risks affect government bond market movements. Using a nonparametric

quantiles- based methodology, we show that rare disaster- risks affect only volatility,

but not returns, of 10- year government bond of the United States over the monthly

period of 1918:01 to 2013:12. In addition, the predictability of volatility holds for the

majority of the conditional distribution of the volatility, with the exception of the

extreme ends. Moreover, in general, similar results are also obtained for long- term

government bonds of an alternative developed country (UK) and an emerging market

(South Africa).

JEL CLASSIFICATION

C22, C58, G12

KEYWORDS

bond returns and volatility, nonparametric quantile causality, rare disasters