Is There a Global Safe Haven?

• DOI: http://doi.org/10.1111/infi.12078
• Published date: 01 December 2015
• Author: Livio Stracca,Maurizio Michael Habib
• Date: 01 December 2015

Is There a Global Safe Haven?

Maurizio Michael Habib and Livio Stracca

European Central Bank.

Abstract

In this paper we search for global safe-haven assets by analysing the

impact of crisis episodes and risk shocks on the behaviour of foreign

investors and selected asset prices and yields. We regress de-trended net

purchases by foreign investors of debt instruments issued in safe and

liquid financial markets on estimated global risk-aversion shocks and

crisis episodes identified by both quantitative and qualitative means.

On the price side, we also look at changes in long-term government

bond yields and the appreciation of the effective exchange rate condi-

tional on risk shocks. We find that US short-term debt –in particular

government debt –is the best, thoug h still imperfect, example of a

global safe-h aven asset. In the major ity of cases we find that foreign

investors head for the exit.

I. Introduction

The behaviour of foreign investors vis-

a-vis risk shocks is lar gely an empirical

matter, as it is not straightforward to derive unambiguous implications from

existing theoretical models. While an extensive literature on the home bias

We would like to thank Giulia Filippeschi and Letiz ia Montinari for excellent research assis tance. We

are grateful to Dimitri os Vagias, Frank Warnock and three anonymous referees for useful suggest ions

and comments. The views ex pressed in this paper belong to the authors a nd are not necessarily

shared by the European Central Ban k.

International Finance 18:3, 2015: pp. 281–297

DOI: 10.1111/infi.12078

© 2015 John Wiley & Sons Ltd

contributed to our understandi ng of the role of distance as a fac tor in portfolio

allocation from a stat ic perspective (see C oeurdacier and Rey 201 3 for a review),

relatively little is kn own on how foreign investors react to news in a dyna mic setting.

In particular, not all models point to retrenchment a s an optimal be haviour of

foreign investors in reaction to glob al or local risk shocks. For in stance, foreign

investors may be less well informed tha n domestic investors (Gehri g 1993; Tille and

van Wincoop 2010) and/or less attentive or familiar with domestic assets (Mondria

et al. 2010; Van Nieuwerburgh and Veldkamp 2010), or have less legal protection

against default (Broner et al. 2010) or a different exposure to home and foreign

shocks (Devereux and Sutherland 2011). However, a rise in global risk and global

risk aversion has, for example, unclear effects on the degree of information

asymmetry between domest ic and foreign investors (infor mational rents by dom es-

tic investors may go down when risk goes up). An in crease in global ri sk aversion

should lead to furt her portfolio d iversification to reduce t he overall level of risk

and hence to larg er foreign portfoli o investment, but if th at were true, we would

observe portfolio inflows by foreign invest ors as well as por tfolio outflows by

domestic investors. Instea d, the existing literature suggests exac tly the opposite:

Broner et al. (2013) convinci ngly show that g ross capital flows tend to re-trench in

crisis times.

The original contribution of this paper to the literature on capital flows is threefold.

First and foremost, we look in greater detail at the assets, debt securities –in particular

those issued by the government –and countries that are most likely to serve as safe

havens in crisis times. We provide a finer partition of cross border investment than, for

example, Broner et al. (2013). In particular, we seek to understand whether there are

assets and countries that represent the exception to the general rule of portfolio

retrenchment as a reaction to rising uncertainty. Second, we focus on the external

liabilities in balance of payments statistics in order to gauge foreign demand for

domestic securities. Similar to Forbes and Warnock (2012), we attempt to identify

‘surges’(i.e. sharp increases) and ‘stops’(i.e. sharp decreases) in capital inflows that are

driven by foreigners. Compared with Forbes and Warnock (2012), we look at the

demand by foreign investors conditional on the materialization of a certain structural

shock and not in general, and we look at a longer sample and not just the most recent

crisis. We consider two measures of risk: crisis episodes as well as estimated global

risk-aversion shocks. Third, we collect the capital flow data from a variety of national

sources at the highest available (monthly) frequency.

1

1

Moreover, in the working paper version of this study, we control for net issuance of securi ties by

domestic residents for the e uro area and the euro-area countries for whic h high-quality data on ne t

issuance are available . Net issuance is i ncluded in order to check to what extent foreigners are

absorbing a larger or smalle r supply of domestic securities th an domestic investors. Althoug h the

statistical significance, and in some case s also the size, of the coeffici ents associated with cris is

dummies or the globa l risk shock is slightly reduced, qu alitatively the main resu lts are the same.

282 Maurizio Michael Habib and Livio Stracca

© 2015 John Wiley & Sons Ltd