Jump to ContentJump to Main Navigation
Macroeconomics and DevelopmentRoberto Frenkel and the Economics of Latin America$

Mario Damill, Martín Rapetti, and Guillermo Rozenwurcel

Print publication date: 2016

Print ISBN-13: 9780231175081

Published to Columbia Scholarship Online: September 2016

DOI: 10.7312/columbia/9780231175081.001.0001

Show Summary Details

Monetary Policy and External Shocks in a Semidollarized Economy

Monetary Policy and External Shocks in a Semidollarized Economy

Chapter:
(p.83) Chapter 5 Monetary Policy and External Shocks in a Semidollarized Economy
Source:
Macroeconomics and Development
Author(s):

Oscar Dancourt

Publisher:
Columbia University Press
DOI:10.7312/columbia/9780231175081.003.0005

Abstract and Keywords

the chapter discusses the adequate monetary policy response to management of external shocks that deteriorate the balance of payments and contracts aggregate demand. To this end the author develops an IS-LM-BP type of model adapted to the financial conditions of Peru, where the banking system operates in both domestic and foreign currency.

Keywords:   IS-LM-BP model, external shock, monetary policy, inflation targeting, sterilized intervention, Peruvian economy

The 2008–2009 crisis demonstrated that the main macroeconomic challenge facing an economy such as Peru’s is the management of external shocks that deteriorate the balance of payments and reduce aggregate demand. The aim of this chapter is to discuss what the monetary policy should be in response to these external shocks.

The arsenal of Peruvian monetary policy has several instruments linked to the credit channel and managed floating exchange rate regime.1 Since inflation targeting was implemented in 2002, the most important instrument of monetary policy has been a short-term interest rate (reference or policy rate) in domestic currency (DC). Another important tool during the last decade has been the reserve requirement ratios levied on bank deposits in both domestic and foreign currency (FC) and on the banking system’s short-term external debt. A third key instrument of monetary policy since 2002–2003 has been sterilized intervention in the foreign exchange (FX) market.2

To compare the different responses of monetary policy to external shocks, these central bank’s instruments are incorporated into an IS-LM-BP textbook model.3 This Mundell–Fleming model is adapted to the financial conditions of an economy such as Peru’s, which has a banking system that operates in both domestic and foreign currency.

The conclusion of this chapter is that a monetary policy, as suggested by Blanchard et al. (2010), which combines a Taylor rule for setting the interest rate, aimed at internal equilibrium, with a foreign exchange intervention policy of leaning against the wind, aimed at external equilibrium, can stabilize the price level and economic activity in the face of external shocks. These two rules imply that the central bank increases (decreases) the interest rate and (p.84) buys (sells) foreign currency when favorable (adverse) external shocks occur. The foreign exchange reserves of the monetary authority play an important role in the management of external shocks. This chapter also suggests a way to implement the proposal of Damill and Frenkel (2011) regarding a target for a real competitive exchange rate in an economy such as Peru’s.

The Banking System’s Interest Rates

In this model, commercial banks make loans and take deposits in both DC and FC.4 There is also a second source of loanable funds in both currencies: domestic banks have credit lines with the central bank in DC and foreign banks in FC. A local bond market does not exist. Thus, there are two lending interest rates, two deposit interest rates, and two basic interest rates: the central bank policy rate and the external interest rate, which are the cost of the banks’ credit lines in both currencies.

The first step consists of determining the lending and deposit interest rates in both currencies as a function of the interest rates of the credit lines, the reserve requirements, and the delinquency rate. These are nominal interest rates, and a zero expected inflation rate is assumed.

If commercial banks match assets and liabilities by currency and do not speculate with the future path of the exchange rate, it is as if there were two banking systems, one operating in domestic currency and the other in foreign currency. The balance sheets of both systems without equity would be as follows:

(5.1)
L=(1θ)D+U
(5.2)
L=(1θ)(D+U),
where L and L* are loans in DC and FC; θ‎ and θ‎* are reserve requirement ratios in DC and FC; D and D* are deposits in DC and FC; and U and U* are outstanding debt in DC with the central bank and outstanding debt in FC with foreign banks, respectively. It is assumed that credit lines in DC are exempt from reserve requirements and that the same reserve requirement ratio is applied to credit lines and deposits in FC.

If competition in the deposits market leads banks to set deposit interest rates that equal the cost of the two sources of funds (deposits and credit lines), we have that

(5.3)
ipD(1θ)D=iUU
(p.85)
(5.4)
ipD(1θ)D=iU(1θ)D,
where the deposit interest rates in DC and FC are ip and i p * ; the central bank policy rate is i; and the external interest rate of credit lines in FC is i*. From equations 5.3 and 5.4, we get that the deposit interest rate in DC depends directly on the central bank’s reference rate and negatively on the reserve requirement ratio in DC and that the deposit interest rate in FC is equal to the foreign interest rate, i.e.,
(5.5)
ip=(1θ)i
(5.6)
ip=i.

If competition in the loan market leads banks to set lending interest rates that generate zero profits, the interest income from the loan portfolio will be equal to the interest expense of deposits and credit lines; expected delinquency rates are taken into account (denominated as m and m* for loans in DC and FC, respectively), whereas operating costs are not, i.e.,

(5.7)
RL(1m)=ipD+iU
(5.8)
RL(1m)=ipD+iU,
where R and R* are the lending interest rates in DC and FC, respectively. Substituting equations 5.1 and 5.5 into equation 5.7 and equations 5.2 and 5.6 into equation 5.8 yields the result that the lending interest rate in DC depends directly on the reference rate and that the lending interest rate in FC depends directly on the external interest rate and the reserve requirement ratios in FC. Both lending rates also depend directly on the expected rate of delinquency,5 i.e.,
(5.9)
R=i1m
(5.10)
R=i(1θ)(1m).

If the maturity of loans were longer than that of deposits and credit lines, then the expected future reference and foreign interest rates would be additional factors that influence loan interest rates in domestic and foreign currency.6

(p.86) The is Curve

The next step is to introduce both lending interest rates in the IS curve. Production (Y) is determined by aggregate demand, which has two components: domestic demand (A), consumption plus investment, and the trade balance or net exports (Xn). That is,

(5.11)
Y=A+Xn=A(R,R,E,P,Y)+Xn(E,P,Y,Y).

As in Krugman (1999), domestic demand (A) is an inverse function of the debt of the private sector in local and foreign currency.7 The debt burden increases if both lending interest rates (R, R*) rise or if the exchange rate (E) rises; the debt burden is reduced if the price level (P) or economic activity (Y) rises.8 Net exports depend directly on the exchange rate (E) and external GDP (Y*) and inversely on domestic GDP (Y) and the local price level (P); the price in foreign currency of foreign goods is assumed to be 1.9

In this way, domestic demand depends negatively on the factors that determine both lending interest rates: the central bank’s reference rate, the external interest rate, the reserve requirement ratio in FC, and the delinquency rates.

As in Krugman (1999), the nominal exchange rate (and the local price level) have two opposite effects on aggregate demand. On the one hand, an increase in the exchange rate generates a higher debt burden that reduces domestic demand because banks lend in dollars to firms that sell in pesos or families who earn their income in pesos. On the other hand, an increase in the exchange rate makes domestic goods cheaper relative to foreign ones, which increases net exports. Thus, an increase in the exchange rate can be recessionary (the balance sheet effect prevails) or expansionary (the competitiveness effect prevails); we will assume that in the short run the balance sheet effect dominates over the competitiveness effect.10

In such a case, a linear IS curve could be defined as

Y=α1iα2(i+θ)α3E+α4Y+α5P.

In the IS equation, an increase in the price level rises economic activity because the balance sheet effect is bigger than the competitiveness effect. On the one hand, the debt burden decreases with a higher price level, which raises domestic demand. On the other hand, a higher price level reduces net exports because domestic goods become more expensive relative to foreign goods. If the competitiveness effect were dominant, an increase in the price (p.87) level would have the usual negative impact on economic activity.11 In the IS equation, the expected delinquency rates are assumed to be zero.

The LM Curve

The third step is to introduce both deposit interest rates into the LM curve. In this model, deposits (and loans) in domestic and foreign currency are imperfect substitutes. Deposits in domestic currency are a medium of exchange and a store of value. Deposits in foreign currency are only a store of value.

The monetary base (H) is equal to banking reserves in DC if the currency in circulation is assumed to be zero. Demand for reserve requirements in DC is equal to the reserve requirement ratio (θ‎) multiplied by deposit demand in DC, i.e., H = θ‎D. Demand for deposits in DC depends directly on economic activity (Y), the price level (P), and the deposit interest rate in DC, given by ip = (1 − θ‎)i. It is a negative function of the deposit interest rate in FC( i p * = i * ) adjusted by expected devaluation (E*/E). Thus, the equilibrium in the monetary base market is given by

(5.12)
H=θD(P,Y,i,θ,i,E/E).

As such, a lineal LM curve could be defined as

H=P+Y+α5iα7(i+EE)+α8θ,
where the net effect of the reserve requirement ratio in DC on monetary base demand is assumed to be positive.

Domestic currency monetary base and deposits are created when the central bank purchases dollars from the public and increases its foreign exchange position (PosCam) or when commercial banks increase their loans to firms and households taking more debt (U) with the central bank, i.e.,

(5.13)
EPosCam+U=H.

The central bank’s net international reserves (RIN) consist of the foreign exchange position (net external assets held by the central bank) and total bank reserves in foreign currency, θ‎* (D* + U*), which are assumed to be deposited in the central bank, i.e.,

(5.14)
RIN=PosCam+θ(D+U).

Finally, the financial wealth (RF) of companies and households is the sum of deposits minus loans in both currencies. This financial wealth is (p.88) equal to net external assets (RIN minus commercial banks’ external debt). Using equations 5.1, 5.2, 5.13, and 5.14, we obtain

(5.15)
RF=(DL)+(DL)E=(RINU)E.

The BP Curve

To complete the Mundell–Fleming model, we utilize the balance-of-payments equation with imperfect capital mobility, as in Ball (2012) or Mankiw (2010), instead of the uncovered interest rate parity equation with perfect capital mobility as in Krugman (1999) or Blanchard (2006).12

The fourth step is to introduce the lending and deposit interest rates in local and foreign currency as determinants of the capital flows in the BP curve. In this model, according to the identity of the balance of payments, the change in the central bank’s international net reserves (Δ‎RIN) equals the dollar value of net exports ( P E X n ) plus the short-term capital flow, which is defined as the change in external debt of the domestic banking system ( U * U t1 * ) minus the interest payments on foreign debt ( i t1 * U t1 * ) , i.e.,

(5.16)
ΔRIN=PEXn+(UαUt1),
where α‎ is one plus the interest rate. These foreign exchange reserves have two components: the foreign exchange position of the central bank and foreign currency banking reserves, which can vary independently of one another, i.e., Δ‎RIN = Δ‎PosCam + θ‎*(Δ‎D* + Δ‎U*), according to equation 5.14. For example, if the commercial banks take debt abroad to finance a local credit boom, loans and deposits in foreign currency will expand; thus, banking reserves in foreign currency and RIN will increase even if the central bank does not intervene in the foreign exchange market and the exchange rate is completely flexible.

Using equations 5.14 and 5.16, we can equal the result of the balance of payments with the variation in the foreign exchange position. That is,

(5.17)
PosCam=PosCamPosCamt1=PEXn+(1θ)(UαUt1)                                                           θ(DDt1).

If the central bank does not intervene in the foreign exchange market, the foreign exchange position is constant (PosCam = 0).13 External (p.89) equilibrium is defined as a balance of payments that keeps the central bank’s net asset position constant. Equation 5.17 implies a redefinition of capital flows, which now depend on the change in external debt of the banking system and in foreign currency deposits. In what follows, we discuss the factors that determine these capital flows (FK).

FK=(1θ)(UαUt1)θ(DDt1).

The local banks’ foreign debt (U*), more volatile than deposits (D*), is the factor that determines capital flows in practice. According to equation 5.2, the bank’s foreign debt depends directly on loans and inversely on deposits. That is,

U=L1θD.

Demand for loans in foreign currency depends directly on the price level (P) and economic activity (Y), as well as the lending interest rate (R) of substitute loans in local currency; it also depends inversely on the lending interest rate (R*) in foreign currency adjusted by expected depreciation (E*/E). Setting this demand for loans in foreign currency equal to the effectively loaned amount, we obtain14

(5.18)
L=L(R,R,E/E,P,Y).

Demand for FC deposits depends directly on the foreign interest rate (equal to the deposit rate in FC) adjusted by expected depreciation (E*/E); it also depends inversely on the deposit interest rate in domestic currency, which in turn depends on the policy rate and the reserve requirement ratio in domestic currency. It also depends inversely on the economic activity and price level because it is assumed that this is a reserve asset. Equating this demand with the amount actually deposited, we get

(5.19)
D=D(i,i,E/E,P,Y,θ).

From equations 5.2, 5.18, and 5.19, we obtain first that demand for foreign debt (U*) depends directly on the loan and deposit interest rates in domestic currency and inversely on loan and deposit interest rates in foreign currency adjusted by expected depreciation. We then obtain that the demand for foreign debt (U*) depends directly on the price level and economic activity, i.e.,

(5.20)
U=U(i,i,E/E,P,Y,θ,θ).

(p.90) In equation 5.20, it is assumed that local banks are not subject to credit rationing in foreign markets.15

In what follows we discuss this demand for foreign debt (U*). In equation 5.20, a rise in the policy interest rate (i) causes an increase in the foreign debt of banks because it increases both lending and deposit interest rates in domestic currency, which implies that demand for loans in foreign currency is higher (becomes more attractive) and that the demand for deposits in foreign currency is lower (becomes less attractive). The external funding (U*) of local banks must increase if they want to lend more with less internal funding (D*).

In equation 5.20, a rise in the foreign interest rate (i*) causes a decrease in the foreign debt of banks because it increases both lending and deposit interest rates in foreign currency, which implies that demand for loans in foreign currency is lower (becomes less attractive) and that demand for deposits in foreign currency is higher (becomes more attractive). The external funding (U*) of local banks must decrease if they wish to lend less with more internal funding (D*).

In equation 5.20, a rise in the reserve requirement ratio in domestic currency (θ‎) causes a decrease in the foreign debt of banks because it reduces the deposit interest rate in domestic currency, which implies that demand for deposits in foreign currency is higher (becomes more attractive). The external funding of local banks (U*) must decrease if they want to lend the same with more internal funding (D*).

An increase in the reserve requirement ratio in foreign currency (θ‎*) has two opposite effects on external debt. First, an increase of θ‎* raises the lending interest rate in foreign currency, which reduces demand for loans; thus, external funding (U*) of the local banks has to diminish if they wish to lend less while their internal funding (D*) remains constant. But there is an opposite effect: to lend the same with constant deposits, more foreign debt is necessary if the reserve requirement ratio increases. In equation 5.20, it is assumed that the first effect dominates.

Last, in equation 5.20, an increase in economic activity or the price level increases the demand for foreign debt. The external funding (U*) of the local banks has to increase if they lend more with less internal funding (D*).

If we plug equations 5.19 and 5.20 into FK, we obtain an equation for the capital flows similar to the one traditionally used in IS-LM-BP models, with the exception of the presence of both reserve requirement ratios with a negative sign and of economic activity and the price level with a positive sign. Obviating the terms U t1 * , D t1 * , we have that (p.91)

(5.21)
FK=FK(i,i,E/E,P,Y,θ,θ).

From equations 5.17 and 5.21, we finally obtain the equation of the BP curve:

(5.22)
PosCam=PEXn(E,P,Y,Y)+FK(i,i,E/E,P,Y,θ,θ).

A linear BP curve could thus be defined as

PosCamPosCamt1=α9Eα10Yα11P+α12Y+α13i                                        α14(i+E)α15θα16θ,
where it is assumed that 1) an increase in the exchange rate improves the balance of payments and 2) an increase in economic activity or in the price level deteriorates the balance of payments through the trade balance, although it induces capital inflows as well.16

The AS Curve

Finally, we incorporate an aggregate supply (AS) curve to the IS-LM-BP model. The price of the domestic good (P) depends on a constant markup (z) and the unit labor cost ( W a ) , with W being the nominal wage and a the product per worker. If we set a = 1, we have that P = (1 + z)W, which implies that the real wage in terms of the domestic good is constant. If the nominal wage depends negatively on unemployment and directly on the exchange rate, because the consumer basket includes the foreign good, a linear aggregate supply curve could be

P=α17E+α18(YYP),
where it is assumed that unemployment depends negatively on the gap between effective and potential output (Y − Yp) and that α‎17 < 1, such that the real exchange rate (EP) changes in the same direction as the nominal exchange rate (E).

External Shocks and Monetary Policy

There are two basic monetary regimes in this IS-LM-BP-AS model: one with constant rates and another with constant aggregates.17 In the first case, the central bank sets the reference interest rate (i) in local currency and the exchange rate (E), as well as both reserve requirement (p.92) ratios (θ, θ‎*), and the four basic equations of the model determine economic activity (Y), the monetary base (H), the foreign exchange reserves (PosCam), and the price level (P). In the second regime, these four basic equations determine economic activity, the reference interest rate in local currency, the exchange rate, and the price level, whereas the central bank sets, in addition to both reserve requirement ratios, the monetary aggregates, i.e., the credit extended to local banks (U) and the foreign exchange position (PosCam), which implies that the central bank determines the monetary base (H).

Because it can set two of the four financial variables (PosCam, H, i, E) the central bank has two monetary policy tools. In addition, the central bank sets the reserve requirements in both currencies.18 Only if the central bank chooses not to intervene in the foreign exchange market such that Δ‎PosCam = 0, i.e., only if it opts for a freely floating exchange rate regime, will it have a unique instrument that can be the reference interest rate or the monetary base.

The basic features of this version of the Mundell–Fleming model can be presented in the framework of a conventional aggregate supply and demand model that describes a short-term equilibrium.19 If the central bank sets the interest rate and the exchange rate, aggregate demand (AD) is obtained directly from the IS equation, and aggregate supply (AS) is the same that we saw in the previous section. That is,

AD:P=α1i+α2(i+θ)+α3Eα4Yα5+1α5Y                    AS:P=α17E+α18(YYP).

The aggregate demand curve has a positive slope because the Fisher or balance sheet effect dominates the competitiveness effect. Stability in this AS-AD model requires that the slope of aggregate demand be higher than the slope of aggregate supply, i.e., that 1 > α‎5α‎18, where all α‎ coefficients are positive.

The increase in the nominal exchange rate (dE > 0) is a negative demand shock (the balance sheet effect dominates the competitiveness effect), whose strength is given by α‎3. It is also a supply shock that raises the price level for any given output gap, whose strength is given by α‎17. Here, an increase in the price level stimulates aggregate demand because it reduces the debt burden of firms and families. Hence, the indirect effect on economic activity of a devaluation through the (p.93) increase in the price level is positive, whereas the direct effect is negative. If α‎3 > α‎5α‎17, a devaluation leads to a recession, given the stability condition of 1 > α‎5α‎18. And the less steep the aggregate supply curve, the more likely it is that this devaluation leads to an increase in the price level; for example, if α‎18 = 0, the price level increases. This is what we have assumed in figure 5.1; an increase in the exchange rate shifts the DA curve to the left and the AS curve upward from point A to B; the price level rises and economic activity falls.

A sufficient condition for a devaluation to improve the balance of payments, according to the BP equation, is that the indirect negative effect through the increase in the price level is smaller than the direct positive effect of the exchange rate, i.e., that α‎9 > α‎11α‎17.

Table 5.1 summarizes the positive or negative effects of different external adverse shocks of a transitory nature on the price level, economic activity, and the foreign reserves (foreign exchange position of the central bank) under a fixed or flexible exchange rate, given the interest rate and the rest of the exogenous variables.20

Monetary Policy and External Shocks in a Semidollarized Economy

Figure 5.1 A negative external shock.

(p.94)

Table 5.1 Effect of External Shocks with Fixed Interest Rate

Domestic Prices

Economic Activity

Foreign Exchange Reserves

Exchange Rate

Fixed exchange rate

International recession

0

Increase in foreign interest rate

0

Increase in expected exchange rate

0

0

0

Flexible exchange rate

International recession

±

0

+

Increase in foreign interest rate

±

0

+

Increase in expected exchange rate

±

0

+

If the central bank sets the interest rate and the exchange rate, an international recession (dY* < 0) or a capital outflow (di* > 0)are negative demand shocks that reduce economic activity (dY < 0) and the price level (dP < 0). Both shocks lead to a deterioration in the balance of payments and cause a reserve loss if the balance of payments was equilibrated in the initial situation.21 An increase in the central bank reference rate (di > 0) or in the reserve requirement ratio in foreign currency (dθ‎* > 0) are also negative demand shocks.

If the central bank wishes to stabilize both economic activity (dY = 0) and the price level (dP = 0) in the face of a transitory external shock (dY* < 0), then it must keep the exchange rate constant (dE = 0) and reduce the cost of credit in domestic currency by reducing the reference rate ( di= α 4 α 1 d Y * <0 ) to outweigh the negative effect that a drop in exports has on production and employment.22 As shown in figure 5.1, an international recession shifts the AD curve to the left; a proper cut of the interest rate shifts the AD curve to the right, returning it to its original position; and the fixed exchange rate prevents the aggregate supply curve from moving upward when the balance of payments deteriorates. Thus, we stay at point A despite the international recession. To apply this Keynesian policy, the central bank must have enough foreign currency reserves because the drop in exports and the cut in the reference rate will generate a balance-of-payments deficit (dPosCam = α‎12dY* + α‎13di < 0) if the balance of payments was equilibrated in the initial situation.23

The monetary base is reduced (the deposit interest rate in local currency falls) less than the foreign exchange reserves because domestic (p.95) currency loans of commercial banks increase (due to the cut in the reference rate). These higher local currency loans are financed by an increase in bank debt owed to the central bank.

An alternative path, keeping the exchange rate constant (dE = 0), is to make credit in foreign currency cheaper by reducing the reserve requirement ratio applied to credit lines and dollar-denominated deposits ( d θ * = α 4 α 2 d Y * <0 ) .24 This option assumes that the domestic banking system debt owed to foreign banks can increase at the same time there is an adverse external shock.25 This capital inflow could counteract the effect of a drop in exports on the balance of payments (dPosCam = α‎12dY*α‎16dθ‎* > 0).

If the central bank does not have enough foreign exchange reserves (dPosCam = 0), it has to let the exchange rate float in the face of a transitory adverse external shock. The model with a flexible exchange rate consists of three equations: the aggregate supply and demand curves already seen and the BP = 0 curve that allows us to determine the exchange rate. In the same fashion as before, the LM curve determines the monetary base.

If the central bank keeps the interest rate constant while the exchange rate floats cleanly, the effects of an international recession (dY* < 0) or capital outflow (di* < 0, dE* > 0) are similar, as shown in table 5.1.

Under this monetary policy, these external shocks constitute, just as before, a negative demand shock that tends to reduce economic activity and the price level; now, however, they also constitute a supply shock that tends to increase the price level because the exchange rate rises as the balance of payments deteriorates.

In the flexible exchange rate case, it is preferable to use a graph in the economic activity–exchange rate quadrant instead of the activity–price level quadrant and to directly observe the effect any change in the external context has on these two variables. To plot figure 5.2, the AS is inserted into both the IS and BP = 0 curves.26 The equations of the IS and BP = 0 curves are given by the following:

       IS:E=α4Yα5α18YPα1iα2(i+θ)α3α5α171α5α18α3α5α17YBP=0:E=α13i+α14(i+E)+α15θ+α16θα11α18YPα12Yα9α11α17                       +α10+α11α18α9α11α17Y.
(p.96)

Monetary Policy and External Shocks in a Semidollarized Economy

Figure 5.2. The case of capital outflows.

The IS curve has a negative slope because the depreciation of the domestic currency is contractive, i.e., dE dY = 1 α 5 α 18 α 3 α 5 α 17 <0 . The BP = 0 curve has a positive slope because an increase in the exchange rate generates a surplus in the balance of payments (α‎9 > α‎11 > α‎17) that has to be eliminated through higher economic activity, i.e., dE dY = α 9 α 11 α 17 α 10 + α 11 α 18 <0 .

In figure 5.2, a capital outflow generated by an increase in the expected exchange rate (dE* > 0) occurs, while the central bank keeps the interest rate constant and the exchange rate floats freely. This external shock shifts the BP = 0 curve to the left from point A to B; the IS curve does not move. The exchange rate increases and economic activity falls. This result requires three conditions that were previously discussed and that define the slopes of the IS and BP curves: 1 > α‎5α‎18, α‎3 > α‎5α‎17, and α‎9 > α‎11α‎17. For an increase in the exchange rate to push the price level up even in a recession, aggregate supply must be relatively flat; e.g., if α‎8 = 0, the price level increases.

(p.97) It is worth mentioning that if the exchange rate and the interest rate remain constant, this capital outflow does not affect economic activity or the price level and only reduces the foreign exchange reserves.

As registered in table 5.1, under a clean floating regime, a capital outflow (induced by an increase in the foreign interest rate or the expected exchange rate) or a global recession has the same effects on economic activity and the price level. In terms of figure 5.2, an increase in the expected exchange rate only shifts the BP = 0 curve to the left, deteriorating the balance of payments, whereas the other two adverse external effects shift both the BP = 0 and IS curves to the left; i.e., they deteriorate the balance of payments and reduce aggregate demand. In all cases, the exchange rate increases, and the recession is more severe than with a fixed exchange rate because devaluation is contractionary.27

Under a clean floating regime, monetary policy cannot stabilize both economic activity and the price level if an external adverse shock occurs. If the interest rate is reduced to fight the recession, both the exchange rate and price level rise.28 And if the interest rate is raised to fight the increases in the exchange rate and price level, economic activity decreases.29 In theory, if the interest rate rises, this dilemma could be solved with an expansive fiscal policy.30 In practice, with a small government and no fiscal automatic stabilizers as in Peru, it is unlikely that the overall recessive effect of an adverse external shock and a tight monetary policy could be counterbalanced by expansionary fiscal policy or that a higher local interest rate would be able to avoid an increase in the exchange rate.31

FX Intervention and Taylor Rules

In a semidollarized economy, as we have seen, reducing the interest rate and keeping the exchange rate fixed are the adequate policy responses for stabilizing the price level and economic activity when facing external shocks that deteriorate the balance of payments and reduce aggregate demand. If the external shock only deteriorates the balance of payments, the adequate policy response for stabilizing the price level and economic activity is to keep the exchange and interest rates fixed.

However, a fixed exchange rate regime is exposed to speculative attacks. Frenkel and Rapetti (2010: 43) emphasize that, under conditions of low inflation as in Latin America since the 1990s, a limited flexibility in the exchange rate

(p.98) has been shown to be highly valuable. The lack of commitment to the level of the NER (nominal exchange rate) provides the economy flexibility to adjust to external shocks without resulting in reputational costs for the monetary authorities. The lack of commitment also eliminates the incentives of one-way bets in the FX market by speculators. In their portfolio choices between domestic and foreign assets (and liabilities), private agents have to assume the exchange rate risk. Therefore, a lower exposure to NER variations and lower financial fragility to external shocks is likely to be observed.32

If the central bank combines a Taylor rule (interest rate rises in booms and decreases in recessions) with a sterilized FX market intervention policy of leaning against the wind (buying dollars when their price falls below some target level and selling them when their price rises above this target level), it can come close to approximating this adequate policy response to external adverse shocks without being subject to the vulnerabilities inherent in a fixed exchange rate system.

According to Williamson (2010), a hybrid system such as Brazil’s, where there is limited flexibility because the central bank intervenes in the FX market, with some notion of what the “adequate” exchange rate is, works better in the real world than the two exchange policies discussed in this chapter and in textbooks (either a completely fixed or a flexible exchange rate).

This hybrid exchange policy has been used in Peru since 2002–2003. The central bank purchases dollars when its target exchange rate (EM), which is not announced, lies above the market exchange rate (E) and sells dollars when the opposite occurs. A simple intervention rule (RI) could be as follows:33

PosCamPosCamt1=α19(EME).

Plugging this equation into the BP equation, we obtain the equation of the curve BPRI, which is very similar to the BP = 0 curve except that it contains the central bank’s target exchange rate (EM) in the intercept and has a smaller slope due to the effect of the α‎19 term. In the BPRI curve, the balance-of-payments result is not equal to zero unless EM = E.34

E=α19EMα13i+α14(i+E)+α15θ+α16θα11α18YPα12Yα9+α19α11α17        +α10+α11α18α9+α19α11α17Y.

(p.99) If EM = E in the initial situation, this FX intervention rule implies that foreign exchange reserves decrease with adverse external shocks and increase with favorable external shocks. The higher α‎19 is in the RI equation, the less steep the BPRI curve will be, the higher the changes in currency reserves will be, and the more similar this managed floating regime will be to a fixed exchange rate regime; conversely, the lower α‎19 is in this equation, the steeper the BPRI curve will be and the more similar this regime will be to a clean floating regime.35

If this FX intervention rule is combined with a Taylor rule, where the policy interest rate depends directly on the output gap, it seems possible to reconcile the theory and practice of monetary policy in some emerging market economies, as proposed by Blanchard et al. (2010).36

If we incorporate the IS and BPRI curves in a Taylor rule (RT), where the interest rate is a direct function of the output gap, as in i = i0 + h (YYP), we get

                   ISRT:E=α4Y(α5α18α1h)YPα1i0α2(i+θ)α3α5α17                                    +1+α1hα5α18α3α5α17YBPRT:E=α19EMα13i0+α14(i+E)+α15θ+α16θ(α11α18α1h)YPα12Yα9+α19α11α17+α10+α11α18α13hα9+α19α11α17Y.

The new system of figure 5.2 comprises the ISRT and BPRT curves. The Taylor rule makes the IS curve steeper and the BP curve flatter; for the BPRT curve to have a positive slope in figure 5.2, α‎13 must be small. These two monetary policy rules, one that aims at internal equilibrium and the other at external equilibrium, reduce both fluctuations in the price level and in economic activity in the face of external shocks compared with a clean floating and fixed interest rate regime. As long as the α‎19 term of the FX intervention rule is large enough, no additional condition is required other than those already mentioned.

These two rules imply that the central bank increases the interest rate and buys foreign currency when favorable external shocks occur, which increase aggregate demand that leads to a boom and improves the balance (p.100) of payments, thereby pushing the exchange rate downward. Conversely, the central bank lowers the interest rate and sells foreign currency when adverse external shocks take place, which diminishes aggregate demand and deteriorates the balance of payments, thus pushing the exchange rate upward.37

Ostry et al. (2012) proposed, on the contrary, that the central bank must lower the local interest rate in the event of a capital inflow generated by a reduction in the foreign interest rate in addition to buying dollars and, symmetrically, that it must increase the domestic interest rate if there is a capital outflow in addition to selling dollars. This policy recommendation is derived from two features that do not seem to apply to an economy such as Peru’s: 1) that a decrease (increase) in the external interest rate does not affect the ISRT curve but only the BPRT curve to the left (right) of figure 5.1, and 2) that the ISRT curve has a positive slope (must be higher than the BPRT curve) in figure 5.1 because a depreciation of the national currency raises aggregate demand in the short term. These two features imply that a capital inflow causes a recession and that a capital outflow causes a boom in a flexible exchange rate regime, as in the original Mundell–Fleming model; the opposite occurs in this chapter.

With this FX intervention rule, the central bank avoids excessive appreciations and depreciations of the domestic currency relative to its target value. According to Blanchard et al. (2010: 13),

a large appreciation may squeeze the tradable sector and make it difficult for it to grow back if and when the exchange rate decreases. Also, when a significant portion of domestic contracts is denominated in foreign currency (or is somehow linked to its movements), sharp fluctuations in the exchange rate (especially depreciations) can cause severe balance sheet effects with negative consequences for financial stability, and thus, output.

In addition, an excessive increase of the exchange rate typically raises the price level and inflation rate in an economy such as Peru’s. Presumably, these considerations should be taken into account when determining the central bank’s target exchange rate. This could be a way of implementing, in an economy such as Peru’s, the proposal of Damill and Frenkel (2011) regarding a competitive real exchange rate target.

With respect to the price level, this monetary policy has several consequences. First, if the price level depends directly on the exchange rate and the output gap, to stabilize economic activity and the exchange rate in (p.101) response to external shocks is tantamount to stabilizing the price level; in particular, this monetary policy eliminates abrupt price increases related to balance-of-payments crises generated by adverse external shocks and the lack of sufficient foreign currency reserves in the central bank.38 Second, if there is both external (EM = E) and internal equilibrium (Y = YP), the aggregate supply curve implies that the price level depends only on the target exchange rate of the FX intervention rule, i.e., P = α‎19EM. This should be the desired price level or central bank target (the equivalent to an “inflation target” in this model) for the monetary policy, comprising both a Taylor and FX intervention rule, to be consistent.39

Third, there are two types of possible anti-inflationary policies. If the price level is above the target or desired level, the central bank can raise the interest rate or sell foreign currency (i.e., reduce the intervention rule’s target exchange rate). If it increases the interest rate, both the exchange rate and economic activity decrease, and both reduce the price level. The situation changes if the central bank lowers the intervention rule’s target exchange rate. A reduction in the target exchange rate (EM) shifts the BPRI curve to the right (figure 5.2) without moving the IS curve. We move from point B to point A. The exchange rate diminishes and economic activity increases. The price level can drop if the aggregate supply curve is relatively flat. Foreign exchange reserves decrease if the balance of payments was equilibrated in the initial situation.

Concluding Remarks

The central bank’s holdings of foreign exchange reserves play a key role in the design of a monetary policy that combines a sterilized intervention FX policy of leaning against the wind with a Taylor rule for the short-term interest rate.

Keynes (1971) says in the Treatise on Money, volume VI, that “national systems develop devices and maintain large liquid reserves with the express object of having the power to maintain internal equilibrium over the short period, without too sensitive a regard for external events” (320). What is an adequate level of international reserves? It depends on the magnitude of adverse external shocks. According to Keynes (1971), to determine this level requires “a reasoned estimate of the magnitude of the drain which India might have to meet through the sudden withdrawal of foreign funds, or through a sudden drop in the value of Indian exports” (247). He adds that “this is the sort of calculation which every (p.102) central bank ought to make. The bank of a country the exports of which are highly variable in price and quantity needs a larger free reserve. The bank of a country doing a large international financial and banking business needs a larger free reserve.”

To address adverse external shock with expansionary monetary and fiscal policies, it is essential to have sufficient foreign exchange reserves. Peru’s experience with the global financial crisis of 2008–2009 suggests that it is possible to avoid major recessions and substantial increases in inflation if this necessary condition is satisfied.40

I am grateful to Gustavo Ganiko for his impeccable assistance.

(p.106) References

Bibliography references:

Armas, A. and Grippa, F. 2006. “Metas de inflación en una economía dolarizada: la experiencia del Perú.” In Dolarización financiera: la agenda de política, edited by A. Armas, E. Levy Yeyati, and A. Ize, 135–162. Lima: BCRP and FMI.

Ball, L. 2009. “Policy responses to exchange rate movements.” National Bureau of Economic Research Working Paper 15173. http://www.nber.org/papers/w15173.

Ball, L. 2012. Money, banking, and financial markets. 2nd ed. Duffield, UK: Worth Publishers.

Banco Central de Reserva del Perú (BCRP). 2009 (June). “Reporte de inflación: Panorama actual y proyecciones macroeconómicas 2009–2011.” http://www.bcrp.gob.pe/docs/Publicaciones/Reporte-Inflacion/Reporte-Inflacion-22-Junio-2009/Reporte.pdf.

Bernanke, B. S. and Gertler, M. 1995. “Inside the Black Box: The Credit Channel of Monetary Policy Transmission.” Journal of Economic Perspectives 9(4): 27–48.

Black, S. 1987. “The effect of alternative intervention policies on the variability of exchange rates: the Harrod effect.” In Exchange rate management under uncertainty, J. Bhandari, ed. Cambridge, MA: MIT Press: 73–82.

Blanchard, O. 2006. Macroeconomics. 4th ed. Upper Saddle River, NJ: Prentice Hall.

Blanchard, O., Dell’Ariccia, G., and Mauro, P. 2010. “Rethinking macroeconomic policy.” IMF Staff Position Note. http://jeromevillion.free.fr/ChroniqueSubprimes_Documents/ChroniqueSubprimes_Blanchard2010.pdf.

Clift, J. 2010 (February 12). “IMF explores contours of future macroeconomic policy. Interview with Olivier Blanchard.” IMF Survey Magazine.

Damill, M. and Frenkel, R. 2011. “Macroeconomic policies and performance in Latin America 1990–2010.” Buenos Aires: CEDES.

Dancourt, O. 2009. “Choques externos y política monetaria.” Revista Economia PUCP 32(64): 127–173

Dancourt, O. and Jiménez, R. 2010. “Perú: lecciones de la recesión de 2008–2009.” Documentos Técnicos, Iniciativa para la Transparencia Internacional. http://www.itf.org.ar/pdf/documentos/73-2010.pdf.

Dancourt, O. and Mendoza, W. 2002. Modelos macroeconómicos para una economía dolarizada. Lima: PUCP Fondo Editorial.

Dominguez, K., Hashimoto, Y., and Takatoshi, I. 2011. “International reserves and the global financial crisis.” National Bureau of Economic Research Working Paper 17362. http://www.nber.org/papers/w17362.pdf.

Dornbusch, R. and Krugman, P. 1976. “Flexible exchange rates in the short run.” Brookings Papers on Economic Activity (3): 537–584.

Dornbusch, R. 1980. Open economy macroeconomics. New York, NY: Basic Books.

Dornbusch, R. 1984. “Comments.” In Exchange rate theory and practice, edited by J. Bilson and R. Marston, 398–402. Cambridge, MA: National Bureau of Economic Research.

Feldstein, M. 1999. “Self-protection for emerging market economies.” National Bureau of Economic Research Working Paper 6907. http://www.nber.org/papers/w6907.pdf. (p.107)

Frenkel, R. and Rapetti, M. 2010. “A concise history of exchange rate regimes in Latin America.” Working Paper 2010–01. University of Massachusetts, Amherst.

Henderson, D. 1984. “Exchange market intervention operations: their role in financial policy and their effects.” In Exchange rate theory and practice, edited by J. Bilson and R. Marston, 359–406. Cambridge, MA: National Bureau of Economic Research.

International Monetary Fund (IMF) 2010. “How did emerging markets cope in the crisis?” https://www.imf.org/external/np/pp/eng/2010/061510.pdf.

Keynes, J. M. 1971. A treatise on money: collected writings. vols. V and VI. London: The Royal Economic Society.

Krugman, P. 1999. “Analytical afterthoughts on the Asian crisis.” web.mit.edu/krugman/www/MINICRIS.htm.

Krugman, P. 2000. “Crises: the price of globalization?” Paper for the Jackson Hole Conference Global Economic Integration: Opportunities and Challenges. Federal Reserve of Kansas City, MO.

Krugman, P. and Obstfeld, M. 2001. Economía internacional: teoría y política. 5th ed. Madrid: Addison-Wesley.

Mankiw, N. G. 2010. Macroeconomics. 7th ed. Duffield, UK: Worth Publishers.

Ostry, J., Ghosh, A., and Chamon, M. 2012. “Two targets, two instruments: monetary policy and exchange rate policies in emerging market economies.” IMF Staff Discussion Note. https://www.imf.org/external/pubs/ft/sdn/2012/sdn1201.pdf.

Rossini, R. and Vega, M. 2007. “El mecanismo de transmisión de la política monetaria en un entorno de dolarización financiera: el caso del Perú entre 1996 y 2006.” Documento de Trabajo 2007–017. http://www.bcrp.gob.pe/docs/Publicaciones/Revista-Estudios-Economicos/14/Estudios-Economicos-14-1.pdf.

Rossini, R., Quispe, Z., and Rodríguez, D. 2011. “Capital flows, monetary policy and forex intervention in Peru.” Banco Central de Reserva del Perú, Documento de Trabajo 2011–08. http://www.bis.org/publ/bppdf/bispap57r.pdf.

Taylor, J. B. 1993. “Discretion versus policy rules in practice. Carnegie-Rochester Conference Series on Public Policy 39: 195–214.

Tobin, J. 1980. Asset accumulation and economic activity. Chicago, IL: University of Chicago Press.

Tobin, J. and Braga de Macedo, J. 1980. “The short-run macroeconomics of floating exchange rates: an exposition.” In Flexible exchange rates and balance of payments: essays in honor of Egon Sohmen, edited by J. Chipman and C. Kindleberger, 5–28. Amsterdam: North-Holland.

Williamson, J. 2010. “Exchange rate policy in Brazil.” Peterson Institute for International Economics, Working Paper 10–16. http://www.iie.com/publications/wp/wp10-16.pdf. (p.108)

Notes:

(1.) The credit channel in this chapter refers to the effect that monetary policy instruments have on loan volume and the interest rates charged by banks. See Bernanke and Gertler (1995).

(2.) For a description of the use of these diverse instruments in Peru, see Rossini et al. (2011).

(3.) Regarding the motives that justify this option, see Ball (2009).

(5.) If short-term external debt is exempt from reserve requirements, as was the case in the Peruvian economy between 1992 and 2004, the lending interest rate in FC would not depend on the FC reserve requirement ratio. The remuneration on FC banking reserves can also be included as a determinant of the lending interest rate in FC. See Dancourt and Mendoza (2002).

(6.) The pass-through coefficients that link changes in the reference interest rate and changes in the banking system’s interest rates in DC have been increasing over the last decade, as documented by BCRP (2009); these coefficients are higher for the bank interest rates of shorter maturities.

(7.) To simplify, we have not considered fiscal policy and public debt. If there is no local bond market, the fiscal deficit would be equal to the increase in external public debt.

(8.) The exchange rate is defined as the peso value of a dollar.

(9.) It is assumed that the Marshall–Lerner condition holds.

(11.) This IS curve implies that the aggregate demand curve has a positive slope on the economic activity–price level quadrant given the interest rate; if the competitiveness effect dominates, the aggregate demand curve has the usual negative slope given the interest rate. See Tobin (1980).

(12.) The uncovered interest rate parity equation implies that the exchange rate is only a function of the domestic and foreign interest rates and of the exchange rate expected by the bondholders. If the home and foreign central banks set the domestic and foreign interest rates, and if the expected exchange rate is also an exogenous variable, it is clear that a sterilized purchase or sale of foreign currency by the central bank cannot (p.103) change the exchange rate. Alternatively, this interest rate parity equation implies that the central bank can only fix either the interest rate or the exchange rate. For a sterilized intervention to change the price of the foreign currency in the Mundell–Fleming model given these three exogenous variables, it is necessary to use a balance-of-payments equation as in Ball (2012) or a modified uncovered interest rate parity with a risk premium as in Krugman and Obstfeld (2001). Both versions of the Mundell–Fleming model suppose that the domestic and foreign assets are imperfect substitutes and thus introduce an additional factor, the relative supply of domestic and foreign assets, that also influences the exchange rate and that the central bank can manipulate with sterilized intervention; in both cases, the central bank can fix both the interest and exchange rates. See Dornbusch and Krugman (1976) and Dornbusch (1984).

(13.) It is assumed that the foreign exchange position neither earns interest income nor suffers changes in valuation.

(14.) A similar equation exists for the market of loans in DC. For the demands of loans and deposits in both currencies to be consistent with one other given financial wealth, it is assumed that 1) the reserve requirement ratio in DC influences only both types of deposits, and 2) the reserve requirement ratio in FC does not directly influence deposits or loans but rather only through the lending rate in FC. The loan market in DC remains in the shadow in the same way that the bond market does in the original IS-LM-BP model.

(15.) Banks cannot always obtain all the external debt they want at the current interest rate. During the crises of 2008–2009 and 1998–2000, Peruvian banks suffered an abrupt cut in their short-term external credit lines, a similar phenomenon to a depositor’s run. These circumstances can be represented imperfectly in this framework as an increase in the foreign interest rate.

(16.) An increase in the exchange rate improves the trade balance in dollars if the Marshall–Lerner condition holds and if the coefficient of the exchange rate on aggregate supply is less than unity, assuming that the balance of payments was equilibrated in the initial situation. The increase in the exchange also has a positive effect on the capital account.

(18.) The central bank has two monetary policy tools in the models of Henderson (1984), Dornbusch (1984), Krugman and Obstfeld (2001), Tobin and Braga de Macedo (1980), and Ball (2012).

(19.) With a fixed exchange rate, these equilibria are not necessarily lasting because they do not exclude, for example, a balance-of-payments deficit and a sustained loss of reserves by the central bank. With a completely flexible exchange rate, they are lasting. See Tobin (1980).

(20.) The classic external shock in the Peruvian economy is a drop in the international prices of export commodities, which deteriorates the balance of payments and reduces aggregate demand; it is equivalent to a global recession or an increase in external interest rates, two of the adverse shocks shown in table 5.1. See Ball (2009) and Dancourt (2009).

(21.) It is assumed that the direct effect of both adverse external shocks that deteriorates the balance of payments dominates the indirect effect that improves the balance of payments through a decline in economic activity and the price level. (p.104)

(22.) This result is obtained by totally differentiating the AS-AD system; it does not change if the devaluation is expansionary.

(23.) The Peruvian authorities responded to the external shock of 2008–2009 with a similar Keynesian policy. It was the first adverse external shock in the last half-century that was dealt with monetary and fiscal expansionary policies. From 1998 to 2000, after a similar external shock, both the interest and exchange rates were raised; the recession was much worse than the one that occurred in 2008–2009, and a banking crisis ensued. In the first case, there were enough foreign exchange reserves available; in the second, there were not. See Dancourt and Jiménez (2010).

(24.) These two types of expansionary monetary policies have differentiated effects on credit in DC and FC. A cut in the reference rate increases loans in DC, but it can lead to a contraction in loans in FC. The reduction in the reserve requirement ratios in FC increase loans in FC but can lead to a reduction in loans in DC.

(25.) During the external shocks of 1998–2000 and 2008–2009, the banking system’s short-term foreign debt and domestic loans in foreign currency dropped in the Peruvian economy despite the reductions in the FC reserve requirement ratio. See Dancourt and Jiménez (2010). A capital inflow generated by a reduction in the foreign interest rate can certainly be neutralized with an increase in the reserve requirement ratio in FC. Economic activity and the price level will not change if everything else remains constant.

(26.) This is similar to Krugman and Obstfeld (2001) and Mankiw (2010).

(27.) As stated in endnote 20, it is assumed that the direct effect of these adverse external effects, which leads to a deterioration of the balance of payments, is larger than its indirect effect. If the expected exchange rate changes, there is no indirect effect.

(28.) A sufficient condition for a higher (lower) interest rate to cause a decrease (increase) in economic activity is that the impact of the interest rate on aggregate demand is larger than the impact on capital flows (α‎1 > α‎13) and that the impact of the exchange rate on the balance of payments is larger than its impact on aggregate demand (α‎9α‎11α‎17 > α‎3α‎5α‎17).

(29.) The aggregate supply curve should be relatively flat. If α‎18 = 0, the exchange rate must remain constant so that the price level does not change. The interest rate must increase to counteract the upward pressure on the exchange rate from the adverse external effect; this leads to a recession of the economy.

(30.) A reduction of the reserve requirement ratio in FC is ineffective when facing a capital outflow. See endnote 23.

(31.) Both the interest and exchange rates rose significantly during the 1998−2000 crisis in the Peruvian economy. See Dancourt and Jiménez (2010). The argumentation supposes that the α‎13 coefficient of the local interest rate in the BP curve would be a lot smaller than the α‎14 term of the external interest rate or the expected exchange rate.

(32.) A limited flexibility in the exchange rate can also contribute to the de-dollarization of the banking system. See Armas and Grippa (2006) and Krugman (2000).

(33.) In Peru, these FX interventions are sterilized so that the short-term interest rate does not vary; the central bank operates with a target zone but does not announce the limits of the exchange rate band, which Williamson (2010) criticizes. It is clear, (p.105) however, that the strength and frequency of the interventions indicate the market operators for the approximate location of this exchange rate band.

(34.) Dornbusch (1980) and Black (1987) combine an FX intervention rule and a balance-of-payments equation.

(35.) The Peruvian central bank’s foreign exchange position, which represented 19 percent of GDP in early 2008, was reduced by 27 percent between April 2008 and April 2009, while the exchange rate increased 20 percent between the minimum and peak of this period.

(36.) Blanchard et al. (2010) ask: “Isn’t it time to reconcile practice with theory, and to think of monetary policy more broadly, as the joint use of the interest rate and sterilized intervention, to protect inflation targets while reducing the costs associated with excessive exchange rate volatility?” According to Blanchard et al. (2010), “imperfect capital mobility endows central banks with a second instrument in the form of reserve accumulation and sterilized intervention. This tool can help control the external target while domestic objectives are left to the policy rate.”

(37.) During the boom, the central bank can also raise the reserve requirement ratios on FC deposits and external credit lines or reduce the remuneration on reserve requirements in FC to raise the lending interest rate in FC. During a recession the central bank can do the opposite. Another option is to keep the reserve requirement ratio in FC at a high level and the remuneration at a low level during both booms and recessions to gradually de-dollarize the banking system and thus reduce the risk of a financial crisis. The extreme volatility of external funding, which led to the banking crisis of 1998−2000, could have precipitated another crisis in 2008–2009 had it not been for the intervention of the Peruvian central bank. See Dancourt and Jiménez (2010).

(38.) From 1950 to 2010, there were seven large recessions in the Peruvian economy, one in each decade with the exception of the 1980s, when there were two. In five of these recessions, inflation rose. In the last two recessions, 1998–2000 and 2008–2009, inflation declined.

(39.) In this chapter, the central bank’s price target would not be the consumer price index, which depends on the exchange rate (E) and the domestic price level (P), according to the weights that the imported and domestic goods have on the consumer basket. Instead, the target of the central bank would be the domestic price level (P). In both cases, however, the central bank must stabilize the exchange rate and the output gap to stabilize the price level.

(40.) Damill and Frenkel (2011) and IMF (2010) found that emerging economies with larger foreign exchange reserves suffered milder recessions during the external crisis of 2008–2009, taking into account other factors such as the change in exports, short-term foreign debt, or the growth rate before the crisis. Dominguez et al. (2011) found that emerging economies with larger foreign currency reserves experienced higher growth after the 2008–2009 crisis, taking into account other factors such as the change in terms of trade or the growth rate before the crisis. See also Feldstein (1999).