A Small Open Economy Model with Currency Mismactches and a Financial Accelerator Mechanism

Abstract

We develop a two-sectors small open economy model with imperfect competition, one-period nominal price rigidities and a financial accelerator mechanism. The latter assumes an asymmetric information problem between lenders and capital good producers (entrepreneurs). Studying the zero-inflation steady state, it is shown that the model with the financial accelerator mechanism nests a fairly standard RBC model; case in which entrepreneurs “disappear" as a differentiated sector from households. It is also explained that credit market imperfections essentially reduce the aggregate supply of capital relative to the RBC case. Turning to the dynamics, we study the effects of an unanticipated and permanent increase in the level of the money supply. In this context the exchange rate jumps immediately to its new steady state level without showing any overshooting process as in Dornbusch (1976). Analysing the case without credit market imperfections but with pre-set prices, it is demonstrated that money is not neutral in the long-run, that capital adds persistence to the initial shock, and that some traditional results of the Mundell-Fleming model still hold.

Full text

Page 1

A Small Open Economy Model with Currency
Mismatches and a Financial Accelerator Mechanism
Santiago L. E. Acosta Ormaechea
University of Warwick, CV4 7AL, Coventry, UK
S.L.E.Acosta-Ormaechea@warwick.ac.uk
(preliminary version)
June 3, 2007
Abstract
We develop a two-sectors small open economy model with imperfect competition,
one-period nominal price rigidities and a ?nancial accelerator mechanism. The latter
assumes an asymmetric information problem between lenders and capital good producers
(entrepreneurs). Studying the zero-in?ation steady state, it is shown that the model
with the ?nancial accelerator mechanism nests a fairly standard RBC model; case in
which entrepreneurs \disappear" as a di?erentiated sector from households. It is also
explained that credit market imperfections essentially reduce the aggregate supply of
capital relative to the RBC case. Turning to the dynamics, we study the e?ects of an
unanticipated and permanent increase in the level of the money supply. In this context
the exchange rate jumps immediately to its new steady state level without showing any
overshooting process as in Dornbusch (1976). Analysing the case without credit market
imperfections but with pre-set prices, it is demonstrated that money is not neutral in
the long-run, that capital adds persistence to the initial shock, and that some traditional
results of the Mundell-Fleming model still hold.
JEL Classi?cation: F3, F4.
Keywords: Credit Market Imperfections; Financial Accelerator; Currency Mis-
matches; Currency Depreciation.
Acknowledgement: I would like to thank Neil Rankin for his supervision, guidance
and support. Alejo Macaya, Marcus Miller, Atsuyoshi Morozumi, Marcelo Sanchez, the
participants of the RES Easter School 2007 in Birmingham and of the 11thICMAIF
conference in the University of Crete provided helpful comments. The conventional
disclaimer applies.
1

Page 2

1Introduction
The classical Mundell-Fleming model shows that a currency devaluation has expansionary
e?ects on output through expenditure-switching e?ects. This ?nding is also obtained in the
well-known model developed by Obstfeld and Rogo? (1995). Although there is a long-standing
debate on whether this e?ect holds for developing countries or not (see Ag? enor and Montiel
1999 for a survey), the recent experiences of contractionary depreciations have revitilised
the discussion. To illustrate, Table 1 shows the negative association that existed between
currency depreciations and real GDP growth rate for a number of selected countries. It is
also noteworthy that the CPI in?ation rate has been in all these cases below the WPI in?ation
rate1.
Table 1. Selected macroeconomic indicators
Country YearNominal depreciation CPI in?ation
(Dec-yt/Dec-yt?1) -in %-
25.9
WPI in?ation Real GDP growth rate
Argentina 200267.478.3-10.9
Indonesia 199870.9 57.680.4-13.1
Korea 199832.1 7.512.2 -6.9
Malaysia 199828.35.3 10.8-7.4
Philippines1998 27.99.711.7-0.6
Thailand
Source: International Financial Statistics, IMF.
1998 24.28.1 12.2-10.5
Calvo and Reinhart (1999) have pointed out that ?nancial factors can be critical to un-
derstand contractionary depreciations due to monetary expansions. When liabilities are de-
nominated in the foreign currency while assets are denominated in the domestic currency, the
argument goes, an exchange rate depreciation increases the domestic value of the liabilities.
If the domestic value of assets does not increase pari passu with the exchange rate, indebted
agents face negative net worth e?ects. This explanation is thus a reinterpretation of the debt-
de?ation mechanism stressed by Fisher (1933), but in the context of small open economies.
Krugman (1999) has ?rstly formalised this argument in a highly stylised and static model.
He shows that a combination of currency mismatches in the private sector, imperfections in
credit markets and sudden changes in expectations could have explained what happened in
the 1997-8 South East Asian crisis. To give an insight on the importance of private sector's
foreign currency denominated liabilities, Figure 1 shows the total claims of foreign banks on
the non-bank private sector for the same countries analysed in Table 1.
1This particular feature in the evolution of in?ation rates is discussed later in this section.
2

Page 3

0
5
10
15
20
25
30
Argentina / end-'01Indonesia / end-'97Korea / end-'97Malaysia / end-'97Philippines / end-'97 Thailand / end-'97
Source: BIS and IMF
As % of GDP
Figure 1. Claims of foreign banks on non-bank private sector
This rather imperfect measure of ?nancial \dollarisation"2,3puts forward the idea that
foreign currency denominated debt has been an important element in the forefront of the
currency depreciations in these countries.
Aghion et al (2000, 2001) follow the same line of reasoning as Krugman (1999), but provide
a higher degree of formalisation. In particular, they assume the production of a single tradable
good that faces one-period nominal price rigidities, in a context where the private sector has
liabilities denominated in the foreign currency. An exchange rate depreciation induced by
a monetary expansion thus generates negative net worth e?ects that reduce investment and
output (i.e., balance-sheet e?ects).
Besides the lack of microfoundations present in their approach4, the assumption that
the tradable good sector faces nominal price rigidities seems to be an important drawback
of their model. In this regard, Burstein et al (2005) show, analysing 5 recent episodes of
large devaluations5, that the main source of changes in the real exchange rate has been the
slow adjustment in the prices of nontradable goods. This can provide an explanation of the
relatively lower increase in the CPI in?ation rate (since this index is highly in?uenced by
nontradable goods) vis a vis the WPI in?ation rate observed in Table 1.
2Financial dollarisation is a widely used expression to indicate that the liabilities of certain sectors in a
country are denominated in the foreign currency. Notice, however, that this foreign currency is not necessarily
the US dollar.
3In particular, owing to lack of information we are not able to discriminate the currency of denomination
of these liabilities. However, being the creditors foreign banks, it seems to be very likely that these loans were
denominated in foreign currencies. Notice also that we are explicitly excluding currency mismatches in the
public sector, which has been a critical feature specially in the Argentine crisis in 2001/2.
4For example, they directly postulate the existence and the form of credit constraints without deriving it
from primitive assumptions.
5They consider the cases of Argentina (2002), Brazil (1999), Korea (1997), Mexico (1994) and Thailand
(1997).
3

Page 4

The objective of this paper is thus to provide a rigorous though realistic framework in which
to analyse why currency depreciations, owing to a monetary expansion, can be contractionary
in the short-run. With this aim, we develop a well microfounded dynamic monetary general
equilibrium model for a small open economy that considers a tradable and a nontradable sec-
tor, imperfections in credit and goods markets, currency mismatches and one-period nominal
price rigidities in the intermediate nontradable good.
Following Obstfeld and Rogo? (1996, Ch. 10.2) the output of the tradable sector will be
assumed exogenous. Since we want to concentrate our attention in the nontradable sector, this
assumption highly simpli?es the analysis without a?ecting the main objectives of the paper.
The nontradable sector is composed of a ?nal producer ?rm, which is perfectly competitive,
and a continuum of intermediate ?rms that face monopolistic competition as in Blanchard
and Kiyotaki (1987).
The production of capital is modelled as in Carlstrom and Fuerst (1997). This capital
is afterwards utilised by intermediate ?rms. There is a continuum of entrepreneurs, each
one producing capital with only one input, which is part of the ?nal nontradable good. The
production function of each entrepreneur has an idiosyncratic and stochastic element. To
determine the amount of investment placed in production, entrepreneurs utilise their net
worth in conjunction with external funding. This funding is, however, subject to frictions due
to the presence of an asymmetric information problem between lenders and entrepreneurs.
All the borrowing that entrepreneurs obtain is assumed to be denominated in the foreign
currency.
Cespedes et al (2004) develop a similar model but with only one sector of production
(tradable) and sticky wages. In the present paper we consider two sectors and fully ?exible
wages. Choi and Cook (2004), Cook (2004) and Devereux et al (2006) are probably the
closest references. Essentially, all these models build on variants of the ?nancial accelerator
mechanism developed in Carlstrom and Fuerst (1997) and Bernanke et al (1999). However,
besides some important di?erences in the speci?cation of the models, they are only interested
in numerical solutions to evaluate di?erent exchange rate and monetary policies. Hence, there
are relevant results and interactions that are hidden in the \black box" typically associated
with calibration methods.In contrast, the present paper's objective is to work through
the analytics of the model so as to provide, whenever possible, an analytical solution that
highlights in a transparent way the mechanisms by which monetary and exchange rate policies
a?ect the economy.
4

Page 5

Although this paper is still work in progress, there are a number of intermediate results
worth emphasizing. We ?rstly studied the properties of the model in the zero-in?ation steady
state. We show that those variables associated with the ?nancial accelerator mechanism yield
simple steady state solutions; depending only on the subjective discount factor, monitoring
costs and the fraction of expected pro?ts that entrepreneurs devote to consumption. Compar-
ing the cases with and without credit market imperfections we show that the latter converges
to a fairly standard RBC model in which entrepreneurs \disappear" as a di?erentiated class
from households. It is also shown that credit market imperfections essentially reduce the
supply of capital relative to the RBC case.
Turning to the dynamics, we study an unexpected and permanent increase in the level
of the money supply under a ?oating exchange rate regime. It is shown that the nominal
exchange rate immediately jumps to its new steady state level, therefore not showing any
overshooting process (contrary to Dornbusch (1976)). As Fender and Rankin (2003) point
out, this particular feature is a direct consequence of the household's logarithmic preferences
assumed in the model. Without credit market imperfections and zero initial net foreign assets
the monetary expansion with pre-set prices improves the short-run trade balance surplus,
giving place to an accumulation of net-foreign assets. Owing to this e?ect money is not
neutral in the long-run. It is thus possible to show that the ?nal nontradable output is
positively a?ected in the short- and long-run. It is also explained that the long-run neutrality
of money is recovered eliminating capital from the model.
The remainder of the paper is organised as follows. Section 2 develops the main elements
of the model with the exception of the production of the capital good. Section 3 explains
how capital is produced in the economy and develops the ?nancial accelerator mechanism.
Section 4 deals with aggregation and de?nes the equilibrium conditions of the model. Section
5 analyses the steady state. Section 6 deals with the dynamics of the model. Section 7
presents concluding remarks.
2The model
We consider a small open economy model with two sectors: one tradable and one nontrad-
able. The economy is composed of ?rms, households, the government and entrepreneurs that
mutually interact within a monetary framework. The remainder of this section describes in
detail the characteristics of each sector.
5