Identifying the Early Warnings of Currency Crisis in India

Published date01 November 2019
DOI10.1177/0015732519874206
Date01 November 2019
Subject MatterArticles
Identifying the Early
Warnings of Currency
Crisis in India
Balaga Mohana Rao1
Puja Padhi1
Abstract
We empirically investigate the recent history of currency crises (stress periods)
and the factors influencing their likelihood in India. This study aims at constructing
an early warning system (EWS) to forecast the possibility of an imminent crisis in
the crisis window of 12 months by employing signal extraction methodology and
logistic regression model for the period 1986–2015. Among the 22 identified cri-
sis months (stress periods), only early episodes (1990–1991) were followed by a
devaluation and not the later periods (post 1991). Both signal extraction meth-
odology and logistic regression model indicate no relative importance to the
three generations of currency crises models. Nevertheless, the results advocate
that logistic regression model fares better than signal extraction technique. The
results also suggest that building an EWS can be an effective diagnostic technique.
JEL Codes: F31, F47, G01
Keywords
Currency crisis, early warning system, India
Introduction
The financial crises have enthused a huge theoretical and empirical debate in cur-
rent times due to its recurrent nature in the history of economics. To highlight the
importance of an economic crisis among various sections, Kaminsky, Lizondo,
Article
1 Department of Humanities and Social Sciences, Indian Institute of Technology Bombay, Powai,
Mumbai, Maharashtra, India.
Corresponding author:
Balaga Mohana Rao, Department of Humanities and Social Sciences, Indian Institute of Technology
Bombay, Powai, Mumbai, Maharashtra 400076, India.
E-mail: mohanraobalaga@gmail.com
Foreign Trade Review
54(4) 269–299, 2019
© 2019 Indian Institute of
Foreign Trade
Reprints and permissions:
in.sagepub.com/journals-permissions-india
DOI: 10.1177/0015732519874206
journals.sagepub.com/home/ftr
270 Foreign Trade Review 54(4)
and Reinhart (1998) quote Kindleberger (1978) saying that academics are inter-
ested in a crisis as they have had a history of fascination for the crises, policymak-
ers are interested because they want to prevent the crisis, and financial market
participants are interested as they can make money out of it. The financial crises
can be divided into two broad categories—currency and sudden stop crises, and
debt and banking crises. Our focus in this article is currency crisis as it has not
only swept away Argentina (2001), Brazil (1998–1999), Latin America (1980s),
Russia (1998), Southeast Asia (1997) and UK (1992) (to name a few) but it also
has caused serious economic adversities to India in the recent past (Cerra &
Saxena, 2002; Cleassens & Kose, 2013; Glick & Hutchison, 2013).
A currency crisis encompasses one of the following four features or a combi-
nation of them owing to a speculative attack—both successful and unsuccessful
attacks—on a currency: A sharp depreciation of a currency (possibly followed
by devaluation) and/or huge depletion of foreign exchange reserves or an
increase in interest rates by central bank or imposing restrictions on capital
flows (Cleassens & Kose, 2013; Edison, 2003; Glick & Hutchison, 2013). Three
generations of models1 evolved over a period of time to explain the phenome-
non of currency crisis upon the failure of previous generation model in explain-
ing the subsequent crisis episode. However, the previous literature (Cerra &
Saxena, 2002) and macroeconomic structure suggest that these three genera-
tions of models do not explain the Indian case completely despite the presence
of some of the elements of these models.
There is a plethora of empirical studies on currency crises (Berg & Pattillo,
1999; Bussiere & Fratzscher, 2006; Edison, 2003; Kaminsky & Reinhart,
1999; Kaminsky et al., 1998; Licchetta, 2011 among others) but their studies
are confined to a large sample of countries based on panel data or cross sec-
tion data whose findings or conclusions are general but not country specific as
they adopt a multicounty approach. Given the heterogeneity of the economies,
there are country-specific studies too (Alvarez-Plata & Schrooten, 2004; Ari,
2012; Cerra & Saxena, 2002; Peng & Bajona, 2008; Sachs, Tornell, & Velasco,
1996 among others); however, there is no but one study on India’s currency
crisis in 1991 (Cerra & Saxena, 2002). The current study emphasizes on
India’s exposure to currency crisis pre and post reforms, and in fixed and man-
aged float exchange rate systems, and the contributing factors of the crisis.
This study also focuses on developing an early warning system (EWS) using
‘signal extraction methodology’ (Kaminsky, Lizondo, and Reinhart [KLR])
and ‘logistic regression model’, and attempts to forecast the possibilities of a
crisis had the EWS been available. This article also tries to find the best suit-
able methodology for a successful prediction of the crisis. This article is par-
ticularly relevant in the context of rapid depreciation of rupee against the US
dollar in recent months.
The remainder of this article proceeds by providing review of literature in the
second section. Whereas the third section presents the data and selection of the
variables, the fourth section deals with the methodology, and the fifth section
discusses the empirical estimations. The final section concludes the study.

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT