The sovereign yield curve and credit ratings in GIIPS

• Published date: 01 September 2021
• Author: Yasir Riaz,Choudhry T. Shehzad,Zaghum Umar
• Date: 01 September 2021
• DOI: http://doi.org/10.1111/irfi.12306

ORIGINAL ARTICLE

The sovereign yield curve and credit ratings

in GIIPS

Yasir Riaz

1,2

| Choudhry T. Shehzad

2

| Zaghum Umar

3

1

Department of Business Studies, Namal

Institute, Mianwali, Pakistan

2

Suleman Dawood School of Business, Lahore

University of Management Sciences, Lahore,

Pakistan

3

College of Business, Zayed University, Dubai,

United Arab Emirates

Correspondence

Yasir Riaz, Suleman Dawood School of

Business, Lahore University of Management

Sciences, Opposite Sector U, DHA., Lahore

54792, Pakistan.

Email: 14080001@lums.edu.pk

Abstract

This paper studies the impact of sovereign credit rating and

outlook changes on the shape of the sovereign yield curve

using data for five European countries, namely, Greece,

Ireland, Italy, Portugal, and Spain, known as the GIIPS for

the period of 2001–2016. We use the dynamic Nelson–

Siegel model to estimate the level, slope, and curvature of

the yield curve. Subsequently, we employ the vector auto-

regressive model to estimate the effect of sovereign rating

and outlook changes on the sovereign yield curve. We find

a significant effect of rating downgrades and an insignificant

effect of rating upgrades in all five countries; however, the

results for the effect of changes in outlook status are mixed.

Our results remain robust to various sensitivity tests.

KEYWORDS

credit ratings, dynamic Nelson–Siegel model and state-space

model, sovereign bonds, term structure of sovereign yields

JEL CLASSIFICATION

C32; E43; G12; G24; H60

1|INTRODUCTION

Following the global financial crisis, financial markets in some European countries, notably, Greece, Ireland, Italy,

Portugal, and Spain (GIIPS) were severely hit by sovereign debt crises resulting in credit crunches, recessions, and

banking sector problems in these economies (Moro, 2014). Credit rating agencies (CRAs) downgraded the sovereign

ratings of GIIPS countries and, consequently, investors had to mark GIIPS bonds as riskier which in turn resulted in

the historically high yields on bonds issued by these countries. Silvapulle, Fenech, Thomas, and Brooks (2016)

Received: 8 June 2018 Revised: 28 February 2020 Accepted: 28 March 2020

DOI: 10.1111/irfi.12306

© 2020 International Review of Finance Ltd. 2020

International Review of Finance. 2021;21:895–916. wileyonlinelibrary.com/journal/irfi 895

showed that rating downgrades were a significant factor in causing the upsurge in GIIPS yield spreads. CRAs can

affect sovereign yields through three distinct channels: firstly, CRAs perform the role of information equalizers and

their announcements have a significant effect on the sovereign bond market (Afonso, Furceri, & Gomes, 2012;

Böninghausen & Zabel, 2015; Gande & Parsley, 2005); secondly, CRAs have a direct impact through ratings-

contingent regulations (Opp, Opp, & Harris, 2013; White, 2010); and thirdly, the credit outlook and watch programs

of CRAs also affect financial markets (Alsakka & ap Gwilym, 2012; Bannier & Hirsch, 2010). Therefore, it is pertinent

to expect intense market reactions following CRAs announcements about rating and outlook upgrades or down-

grades. The objective of this paper is to examine the effect of sovereign rating and outlook changes on the shapes of

yield curves of GIIPS countries.

Earlier studies like Silvapulle et al. (2016) and Afonso et al. (2012) examined yields on bonds with specific matu-

rities; however, the reaction of yield on one maturity may not be representative of the reaction of yields on other

maturities. For example, according to the market segmentation hypothesis, the yield curves of bonds with different

maturities are unrelated and may be affected by specific demand and supply factors at each maturity date. Therefore,

yields on different maturities may react differently to the same stimulus, like a rating announcement in our case. In

this paper, we statistically estimate the magnitude and significance of CRAs' announcements on the complete yield

curve instead of estimating the elasticities of bond yields with particular maturities. This approach helps us in map-

ping the behavior of the sovereign bond market in a better way. In a nutshell, this paper explores if the sovereign

credit rating and outlook changes affect the level, slope, and/or curvature of the sovereign yield curves.

This study makes the following two contributions to literature. Firstly, to the best of our knowledge, this is the

first paper that provides estimates for the effects of four types of CRAs' announcements on the shape of the sover-

eign yield curve. Previous studies have explored the effect of these announcements on raw yields of bonds with spe-

cific maturity and not on the complete yield curve itself (Cantor & Packer, 1996; Reisen & von Maltzan, 1999;

Sy, 2002). Recent studies in this stream include Afonso et al. (2012), Böninghausen and Zabel (2015), Vu, Alsakka,

and ap Gwilym, O. (2015), and Baum, Schäfer, and Stephan (2016). Secondly, this paper introduces the dynamic

Nelson–Siegel model of Diebold and Li (2006) in the sovereign ratings literature to calculate level, slope, and curva-

ture of the yield curve and provide separate estimates for short-, medium-, and long-term effects of each type of rat-

ing announcement. Another stream of literature estimates the shape of the sovereign yield curve based on

macroeconomic factors. For example, Aguiar-Conraria, Martins, and Soares (2012), Joslin, Priebsch, and Singleton

(2014), and Ulrich (2013) find a significant effect of macroeconomic variables on the U.S. yield curve while Tam and

Yu (2008) find a similar effect on the German and the U.S. yield curves but not on the Japanese yield curve. Afonso

and Martins (2012) find a significant role of the fiscal behavior for the U.S. yield curve but not for the German yield

curve. For the Chinese market, Yan and Guo (2015) find that inflation and investment growth affect the yield curve,

while Yan and Guo (2016) propose that fiscal behavior is a significant determinant of the shape of the yield curve.

This paper reconciles the two literature streams, that is, one studying the effect of sovereign ratings and the outlook

changes on the sovereign yields of bonds with specific maturity and the other analyzing the dynamics of the shape

of the yield curve.

To estimate the proposed effect, we follow a two-step methodology. In the first step, we use the dynamic

Nelson–Siegel model to estimate the level, slope, and curvature of the sovereign yield curve and in the second

step, we estimate country-specific vector autoregressive (VAR) models and impulse response functions for each

announcement type and three latent factors. This methodology allows us to estimate the low-dimensional yield

curve factors and then study the effect of rating announcements on these factors. There are several advantages

to studying low-dimensional yield curve factors instead of high-dimensional raw yields, for example, a precise

empirical description of yield curve data, valuable compression of information, statistical tractability, good fore-

casting capabilities, and consistency with the parsimony principle (Diebold, Rudebusch, & Boragˇan Aruoba, 2006;

Gürkaynak, Sack, & Wright, 2007; Vicente & Tabak, 2008; Yu & Salyards, 2009; Zellner, 1992). Furthermore, it

allows us to segregate the effect of each type of announcement on short, medium and long-term factors of the

yield curve.

896 RIAZ ET AL.