Coups d'état and the foreign exchange market

• Published date: 01 July 2020
• Author: Hippolyte Weneyam Balima
• Date: 01 July 2020
• DOI: http://doi.org/10.1111/twec.12905

1928

|

wileyonlinelibrary.com/journal/twec World Econ. 2020;43:1928–1950.

© 2019 John Wiley & Sons Ltd

Received: 15 September 2018

|

Revised: 19 September 2019

|

Accepted: 5 November 2019

DOI: 10.1111/twec.12905

ORIGINAL ARTICLE

Coups d'état and the foreign exchange market

Hippolyte WeneyamBalima

International Monetary Fund, Washington, DC, USA

KEYWORDS

coups d'état, currency crises, foreign exchange market

Chaotic scenes of rebel soldiers taking to the streets of Istanbul and Ankara on Friday

caused thelira to plummet 5 percent—its biggest fall since 2008.

The Financial Times (17July 2016)

1

|

INTRODUCTION

On 15 July 2016, a coup d'état attempt by a fraction within the Turkish's Armed Forces took place in

Turkey against President Recep Tayyip Erdogan. The fraction tried to control Ankara and Istanbul—

the country's two largest cities—but failed to do so after forces loyal to the President defeated them.

While the coup d'état failed, the value of the lira—Turkish's national currency—depreciated by 5%

vis-à-vis the US dollar on the day following the coup. The Turkish experience is not an isolated case.

Almost three decades after the end of the cold war, coups d'état continue to be common around the

world and often lead to changes in leaderships and institutions (Leon, 2014).

Previous experiences of currency crises in many countries, including the crash of the Mexican

peso in 1994, the collapses of the Turkish lira in 1994 and 2001 and the collapse of the currency

board in Argentina in 2002, highlighted the importance of exchange rate behaviours for economic

stability and attracted academics and policymaker's interest in exchange rate dynamics (Alsakka & ap

Gwilym, 2012; Beine, Janssen, & Lecourt, 2009; Bollerslev & Melvin, 1994; Daude, Levy Yeyati, &

Nagengast, 2016; Frankel & Froot, 1990; Loria, Sanchez, & Salgado, 2010; Melvin & Taylor, 2009;

Sarno & Taylor, 2001). However, while the drivers of a country's foreign exchange market have long

been an important policy question (Benigno, Chen, Otrok, & Rebucci, 2016; Edwards, 1988; Evans

& Lyons, 2002; Hooper & Morton, 1982; Krueger, 1983), the relation between coups d'état and the

foreign exchange market has yet to be studied formally in the academic literature.

This paper uses a different set of identification strategies to analyse the effect of coups d'état on the

foreign exchange market. The maintained assumption in the paper, deeply discussed in the following

section, is that the occurrence of a coup d'état can affect the foreign exchange market through, at least,

its nested effect on the real economy, uncertainty about a new government's policies and on economic

agents' productive decisions, institutional quality and development aid. In particular, I stress that when

a country experiences a coup d'état, some uncertainty may rise about the future developments of its

|

1929

BALIMA

economic policies including fiscal, monetary, exchange rate or capital control policies. For instance,

the occurrence of a coup d'état may elevate transfer and convertibility risk—the risk that the govern-

ment will impose capital controls or some restrictions on current account transactions that would im-

pede the private sector's ability to convert the local currency into foreign exchange. Such uncertainties

may lead to capital outflows and therefore a depreciation of the coup d'état country's currency against

currencies of its trading partners. Capital outflows may also arise from uncertainty that could be chan-

nelled through downward revisions of economic projections given that coups are generally followed

by sharp economic downturns.1 Depreciation can also stem from market participants' irrational exu-

berance as the result of unexpected speculative upsurges surrounding the foreign exchange market

(Schiller, 2000).

Building on these intuitions, I empirically examine the effect of coups d'état on the foreign ex-

change market using a monthly panel dataset covering 150 countries over the period 1980-M01

to 2015-M12. Econometrically, I rely on two different identification strategies to assess the effect

of coups d'état on the foreign exchange market. First, in the benchmark analysis, I employ the en-

tropy balancing methodology,a generalisation of the conventional matching method proposed by

Hainmueller (2012). This method has been used by Neuenkirch and Neumeier (2016) to study the

effect of economic sanction on poverty. Entropy balancing allows to achieve a high degree of covariate

balance by using a reweighting scheme that directly incorporates covariate balance into a weight func-

tion that is applied to the sample units (Hainmueller, 2012). Practically, entropy balancing imposes

a number of balance constraints that imply that the distributions of the matching covariates of coup

d'état and non-coup d'état groups in the preprocessed data match exactly on all prespecified moments.

Another crucial advantage of this methodology is that it allows theproper control for country and time

dummies to roll out unobservable factors at country and time levels. To sum up, using the entropy

balancing allows to identify the effect of coups d'état by comparing coup d'état and non-coup d'état

countries that are as similar as possible in terms of pre-coup d'état observable characteristics, after

purging for country- and time-specific factors. Second, I follow Jones and Olken (2009) and Masaki

(2016) and focus exclusively on country-month observations where coups d'état took place. I stress

that conditionally on the occurrence of coups d'état, the fact that a coup succeeds in unseating the

government is exogenous to the foreign exchange market. I exploit this exogenous component in the

success of coups d'état to estimate the impact of coups on the foreign exchange market using standard

panel fixed-effect regressions.

My empirical analysis shows that the occurrence of coups d'état negatively affects the foreign ex-

change market in coup d'état countries, relatively to non-coup d'état countries. On average, a country

nominal effective exchange rate—the value of its national currency against its main trading partners'

weighted basket of currencies—depreciates by 0.35 per cent following the occurrence of a coup d'état.

This estimated depreciation is economically meaningful as it corresponds to almost twice the average

exchange rate depreciation in our sample on non-null country-month observations. I strongly demon-

strate that this finding is robust to a large set of alternative specifications of the entropy balancing

methodology as well as the use of an alternative identification strategy. I also find that the occurrence

of a coup d'état creates spillover effects by generating downside risks to neighbouring countries.

Moreover, I show that the effect of coups varies depending on countries' financial buffers (i.e. the

level of international reserves) and the exchange rate regimes. Regarding the former, the magnitude of

coups' effect is particularly higher when the level of financial buffers is relatively low. This suggests

that a comfortable stock of international reserves provides immunity against exchange rate movement

1 For instance, previous research suggests that the shock of a coup d'état reduces income by about 3.5% in the year of the coup

and by a cumulative total of 7% (Collier, 2009), and leads to international sanctions in the form of a reduction of aid

disbursements (Masaki, 2016).