Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Slides for Chapter 11

Capital Controls

Columbia University

April 15, 2021

1
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Introduction

• In Chapter 10, we saw that over the past 120 years sometimes
capital flowed fairly freely across countries and sometimes there were
significant deviations from free capital mobility.

• In the present chapter, we characterize the effects of capital con-

trols on the current account, consumption, and welfare.

• We show that capital controls can be an effective tool to reduce

current account deficits but that absent distortions or market power
they are welfare decreasing.

• We also show that in the presence of borrowing externalities or

market power capital control taxes can be desirable because they
improve welfare.

2
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

11.1 Capital Controls and Interest Rate Differentials

3
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Capital controls are restrictions imposed by governments on the flow

of financial capital into or out of a country.

The imposition of capital controls gives rise to interest rate differ-

entials that cannot be arbitraged away.

Suppose that a e country is borrowing from the rest of the world

and start with no capital controls, ie free capital mobility.

Let it be the domestic interest rate on dollar loans (the onshore

rate) and i∗t the foreign interest rate on dollar loans (the offshore
rate). As we have shown in Chapter 10, under free capital mobility
the onshore interest rate must equal the offshore interest rate,

it = i∗t .

4
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Suppose now that the government imposes a tax τt per dollar bor-
rowed internationally. The tax raises the cost of borrowing one dollar
internationally to i∗t + τt . For agents to be indifferent between off-
shore and onshore borrowing, the domestic interest rate must equal
the sum of the foreign interest rate and the capital control tax rate,

it = i∗t + τt.
The resulting onshore-offshore interest rate differential, it −i∗t , equals
the capital control tax rate, τt. The larger the capital control tax
rate is, the larger the interest rate differential will be.

This is, capital control taxes will give rise to interest rate differen-
tials. Next, we present a case study of capital control taxes imposed
in Brazil that gave rise to interest rate differentials.

5
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Capital Controls and Interest Rate Differentials in Brazil: 2009-2012

• In the wake of the global financial crisis of 2008 interest rates in the United
States and other developed countries fell to near zero. In response to such low
rates, global investors who were looking for higher yields started sending funds to
emerging market economies where interest rates were higher.

• One country that was a recipient of large inflows was Brazil. The Brazilian
authorities were concerned that these capital inflows would destabilize their econ-
omy and enacted capital control taxes on inflows. Specifically, between October
2009 and March 2012 Brazil imposed more than 10 major capital control taxes.
The measures included taxes on portfolio equity inflows, taxes on fixed income
inflows, and unremunerated reserve requirements. After March 2012 those re-
strictions were gradually removed.

• When capital inflow taxes are imposed on a specific asset or class of assets
there is always the concern that market participants can find a way to circumvent
them. One way to see if in this instance the capital control taxes were effective
is to see whether they led to non-zero covered interest rate differentials.

6
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Consider the covered interest rate differential between the Brazilian

real and the U.S. dollar.

it 360-day interest rate in Brazil on domestic currency deposits

(reais),
Et the spot exchange rate (that is, the reais price of one U.S. dollar),
Ft the 360-day forward exchange rate of U.S. dollars, and
i∗t the 360-day U.S. dollar Libor rate.

The covered interest rate differential between deposits inside Brazil

and outside of Brazil is
Et
(1 + it ) − (1 + i∗t ).
Ft
cupom
The first term of this expression is called the cupom cambial, it
cupom Et
1 + it ≡ (1 + it) .
Ft

And there is a liquid market for cupom cambials, so that we have

data on icupom
t .
7
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

We can then express the covered interest rate differential as

cupom
covered interest rate differential = it − i∗t .

Figure 11.1on the next slide plots daily data for the covered interest
rate differential for the period January 1, 2010 to December 31,
2012.

If the capital inflow controls were successful, we should see that dol-
cupom
lar interest rates inside Brazil, it , became higher than in London,
i∗t , that is, that the covered interest rate differential increased.

8
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Figure 11.1: Brazilian Real-U.S. Dollar Covered Interest Rate

Differentials: 2010-2012
4

3.5

2.5
percentage points

1.5

0.5

2010 2011 2012 2013

The figure plots daily real-dollar covered interest rate differentials computed as the spread between the cupom cambial and the U.S. dollar Libor rate for the period
January 1, 2010 to December 31, 2012. Data Source: Marcos Chamon and Márcio Garcia, ‘Capital Controls in Brazil: Effective?’, Journal of International Money and
Finance 61, 2016, 163-187. We thank the authors for sharing the data.

9
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Comments on the figure

• the covered interest rate differential was around 50 basis points until the fall
of 2010 a level also observed prior to the imposition of capital controls. This
means that the capital control measures enacted until then, which targeted mainly
portfolio equity investment, were not effective in restricting arbitrage between the
cupom cambial and the Libor rate.

• However, starting in the fall of 2010 as the government intensified capital con-
trols, the differential starts rising and reaches a peak of 4 percentage points by
April 2011, after an inflow tax of 6 percent on borrowing from abroad with ma-
turities of less than 2 years was imposed. The size of the covered interest rate
differentials suggest that the latter capital inflow taxes were indeed effective in
the sense that they prevented interest rate equalization.

• By early 2012 arbitragers seem to have found ways to bypass the capital control
tax as differentials return to normal levels of around 50 basis points.

• By June 2012 the capital control tax of 6 percent of borrowing from abroad
with maturities of less than 2 years was removed.

10
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

11.2 Macroeconomic Effects of Capital Controls

11
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Effects of Capital Controls on Consumption, Savings, and the

Current Account

Consider a two-period small open endowment economy, like the one

introduced in Chapter 3, with no initial assets (B0 = 0).

Household preferences

U (C1) + U (C2)

Period-1 budget constraint: C1 + B1 = Q1

Period-2 budget constraint: C2 = Q2 + T2 + (1 + i1)B1

T2 = lump-sum transfer
i1 = interest rate on bonds held from period 1 to period 2 (taken
as given by the household)
12
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Optimality conditions associated with the household’s problem:

U 0(C1)
0
= 1 + i1 (1)
U (C2)

C2 = Q2 + T2 + (Q1 − C1)(1 + i1) (2)

13
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Suppose that the government imposes a capital control tax on in-

ternational borrowing. Let τ1 > 0 be the capital control tax rate.
Then, if the economy is borrowing in period 1,

i1 = i∗1 + τ1,
where i∗1 is the world interest rate.

The Euler equation (1) becomes

U 0(C1) ∗
= 1 + i 1 = 1 + i 1 + τ1. (3)
U 0(C2)

This expression implies that a capital control tax on international

borrowing (τ1 > 0) distorts the intertemporal allocation of consump-
tion by increasing the interest rate faced by the household i1. The
tax creates incentives to save more and consume less in period 1.

14
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

The Government Budget Constraint

Revenue from the capital control tax: τ1(−B1 ).

Lump-sum transfers: T2

Assume government returns the revenue from the capital control

tax to households in a lump sum fashion. This implies the following
period-2 government budget constraint:

T2 = τ1(−B1 ) (4)

15
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Equilibrium

Using the government budget constraint (4) to eliminate T2 from

the household’s intertemporal budget constraint (2), yields the econ-
omy’s resource constraint

C2 = Q2 + (Q1 − C1)(1 + i∗1). (5)

Note that the economy’s intertemporal resource constraint is in-

dependent of the capital control tax rate, τ1. This is because the
government’s tax revenue is returned to households, so no resources
are lost as a consequence of the imposition of capital controls.

16
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Equilibrium (continued)

An equilibrium in the economy with capital control taxes is an allo-

cation (C1, C2) satisfying the Euler equation (3) and the economy’s
resource constraint (5), that is,
U 0(C1) ∗ +τ
= 1 + i 1 1 (3)
U 0(C2)
and
C2 = Q2 + (Q1 − C1)(1 + i∗1), (5)
given endowments Q1, Q2, the world interest rate, i∗1, and the capital
control tax, τ1.

The figure on the next slide depicts the equilibrium effects of im-
posing a capital control tax τ1 > 0.

17
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Figure 11.2: Equilibrium With and Without Capital Controls

C2

← slope = −(1 + i∗ )

Q2 A

C2∗ 0 C

C2∗ B

slope = −(1 + i∗ + τ ) →

Q1 C1∗ 0 C1∗ C1

Point A represents the endowment path, point B the equilibrium consumption

path in the absence of capital controls, and point C the equilibrium consumption
path with capital controls.
18
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Comments on the figure:

The downward sloping straight line is the economy’s intertemporal resource con-
straint, given by equation (5). The slope of this line is −(1 + i∗1).
Point A represents the endowment path, (Q1 , Q2).
Point B represents the optimal consumption path in the absence of capital con-
trols, τ1 = 0. At point B, the indifference curve is tangent to the intertemporal
resource constraint.
In the graph, when capital controls are zero, the economy runs a trade deficit
equal to C1∗ − Q1 > 0. Because B0 = 0, CA1 = T B1.
Point C in the figure represents the equilibrium when the government imposes
a capital control tax τ1 > 0. The economy’s resource constraint is unchanged.
However, households perceive an increase in the cost of borrowing from i∗1 to i∗1+τ1 .
The negative of the slope of the indifference curve now is 1 + i∗1 + τ1 > 1 + i∗1. The
imposition of the capital control tax τ1 leads to a fall in C1, and an improvement
in T B1, S1, and CA1.
Welfare is lower at point C than at point B.

Takeaway: The imposition of a capital control tax on international

borrowing discourages current consumption and causes an increase
in saving, a reduction in the trade deficit, a reduction in the current
account deficit, a reduction in the country’s net external debt, and
a reduction in welfare.
19
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Effects of Capital Controls on Investment

Consider the economy with investment of Chapter 5. To keep the

present analysis self-contained, we go over its main elements.
√
Output in period 2: Q2 = I1

I1 = the stock of capital available for production in period 2.

To build this stock of capital, firms invest in period 1. In period

1 firms borrow the amount I1 at the interest rate i1. In period 2,
firms must pay back these loans, including interest. Thus, profits in
period 2 are given by
q
period-2 profits = I1 − (1 + i1)I1 .
First-order condition for profit maximization:
1
√ = 1 + i1 .
2 I1
20
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Solving for the optimal level of investment yields

!2
1
I1 = .
2(1 + i1)

In equilibrium the domestic interest rate, i1, will be equal to the

world interest rate, i∗1 plus the capital control tax τ1
i1 = i∗1 + τ1.

Combining the above two expressions, we obtain

!2
1
I1 = ,
2(1 + i∗1 + τ1)

which shows that capital controls lower investment.

Takeaway: Capital control taxes distort not only the consumption-

saving choice of households but also the investment choice of firms
and as such are, at least in the present model, welfare decreasing.
21
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

11.3 Quantitative Restrictions on Capital Flows

22
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

• Capital controls can also take the form of quantitative restrictions

on international borrowing and lending.

• This form of capital controls, as we will show, is equivalent to

those based on taxes on international capital flows.

23
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

• Suppose the government imposes the following limit on interna-

tional borrowing

B1 ≥ −D; where D > 0.

• This implies that: C1 ≤ Q1 + D

The situation is depicted in Figure 11.3 on the next slide.

The endowment point is at point A and the optimal consumption

path in the absence of quantitative capital control restrictions is at
point B.

At point B, the economy borrows from the rest of the world in order
to finance a level of consumption, C1∗, that exceeds the period-1
endowment, Q1, Q1 − C1∗ < 0.

24
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Figure 11.3 Equilibrium under Quantitative Capital Controls

C2

← Q1 + D

← slope = −(1 + i∗ )

Q2 A

C2∗ 0 C

C2∗ B

slope = −(1 + i) →

Q1 C1∗ 0 C1∗ C1

The equilibrium under free capital mobility is at point B, where C1∗ > Q1 . Quantitative capital
controls forbid borrowing more than D, pushing households to consume Q1 + D in period 1, point
C. The domestic interest rate under quantitative capital controls (i1) is given by the slope of the
indifference curve at point C and is higher than the world interest rate i∗1 .

25
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

When quantitative capital controls are imposed, households choose

point C, and the borrowing constraint is binding, B1 = −D. In
the constrained equilibrium, in period 1 the household consumes the
endowment plus the maximum amount of borrowing allowed, D, so
0
C1∗ = Q1 + D. In period 2, the household consumes its endowment,
Q2, net of debt obligations including interest, (1 + i∗1)D, that is,
C2 = Q2 −(1 +i∗)D. So we have that in response to the quantitative
restrictions on capital inflows, current consumption falls from C1∗ to
0
C1∗ , the trade balance and the current account shrink from Q1 −C1∗ to
0 0
Q1 − C1∗ = −D, and external debt falls from C1∗ − Q1 to C1∗ − Q1 = D.

The quantitative capital control measure drive the domestic interest

rate up to i1 > i∗1.

26
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Graphically, 1 + i1 is given by the negative of the slope of the indif-

ference curve at point C .

Formally, under binding quantitative capital controls the domestic

interest rate satisfies
U 0(Q1 + D)
= 1 + i1 .
U 0(Q2 − (1 + i∗1)D)

Because Q1, Q2 , D, and i∗1 are exogenously given, this expression

represents one equation in one unknown, i1. The more stringent
capital controls are (ie, the lower D is), the higher the domestic
interest rate (i1) will be. We therefore have that the interest rate
differential, i1 − i∗1 is an increasing function of the severity of quan-
titative restrictions on capital inflows.

27
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Comparing Figures 11.2 and 11.3, it is clear that quantity-based

and tax-based capital controls give rise to the same equilibrium, in
the sense that given a capital control tax τ1 one can find a quan-
titative restriction D, such that in equilibrium, consumption, the
trade balance, the current account, the stock of external debt, and
the interest rate differential are the same under both capital-control
policies.

28
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

11.4 Borrowing Externalities

and

Optimal Capital Controls

29
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Overview of Section 11.4

• Sections 11.2 and 11.3 show that capital controls can be an effec-
tive instrument to curb external imbalances, but that if the economy
is small and has well functioning markets, they are welfare decreas-
ing.

• In this section, we introduce a debt-elastic interest rate, which is a

financial friction whereby foreign lenders charge higher interest rates
for larger external debt positions.

• The debt-elastic interest rate creates an externality: individual

households, being atomistic participants in financial markets, fail
to internalize that their individual borrowing decisions collectively
determine the level of the interest rate. As a result, the economy
overborrows.

• Under these circumstances, the government has an incentive to

impose capital controls as a way to make households internalize the
fact that their borrowing drives the interest rate up. The optimal
level of capital controls is positive, eliminates overborrowing, and is
welfare increasing.
30
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

11.4.1 An Economy with a Debt-Elastic Interest Rate

31
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Suppose that the interest rate at which the small open economy can
borrow in international capital markets is increasing in the country’s
cross-sectional average of borrowing, denoted −B̄1 . If B̄1 < 0, the
country borrows, if B̄1 > 0 the country saves.

The interest rate faced by the small open economy, denoted i∗1, is
assumed to be debt elastic. Specifically,

i∗1 = I(−B̄1 )

where I(·) is a non-negative and non-decreasing function of external

debt, −B̄1.

32
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

To see how a debt-elastic interest rate works, let’s consider the

following parametric example of I(−B̄1 ):
(
i for B̄1 ≥ 0
I(−B̄1 ) = , (6)
i + δ(−B̄1 ) for B̄1 < 0

where i and δ are positive parameters.

• the country lends at the constant interest rate i

• but borrows at an interest rate that increases linearly with the level
of debt.

The figure on the next slide plots this debt-elastic interest rate
schedule.

33
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

A Debt-Elastic Interest Rate

I(−B̄1 )

i + δ(−B̄1 )

B̄1 0 −B̄1

The figure displays an interest rate schedule, I(−B̄1), which is weakly increasing
in the level of external debt, (−B̄1 ). Both i and δ are positive constants. For
B̄1 > 0, the country is a net external lender, and the interest rate is constant and
equal to i. For B̄1 < 0, the country is a net external borrower, and the interest
rate is an increasing function of the level of debt, −B̄1.

34
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Let’s now embed a debt-elastic interest rate schedule into a 2-period

small open endowment economy of the type introduced in Chapter
3. Let’s start with the household’s problem.

Households:

Preferences: U (C1) + U (C2)

Endowments: Q1 and Q2.

No initial assets or debts, B0 = 0.

Household can borrow or lend freely at the domestic interest rate i1

via a bond, denoted B1.

Budget constraint in period 1: C1 + B1 = Q1

Budget constraint in period 2: C2 = Q2 + (1 + i1)B1 .

35
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Combine budget constraints in periods 1 and 2 to obtain the house-

hold’s intertemporal budget constraint

C2 = Q2 + (1 + i1)(Q1 − C1). (7)

and use it to eliminate C2 from the lifetime utility function

The household’s maximization problem then becomes

max[U (C1) + U (Q2 + (1 + i1)(Q1 − C1))]

{C1 }
Taking derivative with respect to C1 and equating it to zero, we
obtain the optimality condition

U 0(C1) − U 0(Q2 + (1 + i)(Q1 − C1))(1 + i1) = 0.

Using the fact that Q2 + (1 + i)(Q1 − C1) = C2 and rearranging, we

obtain the usual Euler equation
U 0(C1)
0
= 1 + i1 . (8)
U (C2)
36
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

The debt-elastic country interest rate

Suppose that the country has free capital mobility. Let i∗1 be the in-
terest rate charged by foreign lenders to the country in international
capital markets. Then, in equilibrium

i1 = i∗1 (9)

But what is i∗1 in the present model? We assume that i∗1 is an

increasing function of the average external debt per household.

37
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Debt of an individual household is given by −B1 = C1 − Q1 , where

C1 is consumption of an individual household in period 1 and Q1 is
the endowment of an individual household in period 1. Let Q̄1 and
C̄1 denote cross-sectional averages of output and consumption in
period 1. Then the average debt per household in period 1 is given
by
−B̄1 = C̄1 − Q̄1. (10)
We introduce a debt-elastic interest rate by assuming that i∗1 is given
by
i∗1 = I(−B̄1 ),
where I(·) is a non-negative, non-decreasing function. Using (10)
to replace B̄1 yields i∗1 = I(C̄1 ), where Q̄1 is omitted as an argument
of I(·). This is not a problem for the present analysis, because Q̄1
is an exogenous variable, which we will keep constant throughout.

38
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Because all households are identical in preferences and endowments,

in equilibrium they all consume the same amount of goods. This
means that in equilibrium consumption per capita equals the indi-
vidual level of consumption, C̄1 = C1. So we can write the interest
rate parity condition as

i1 = i∗1 = I(C1).

39
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

An equilibrium in the economy with a debt-elastic interest rate then

are values for C1, C2, and i1 satisfying

U 0(C1)
0
= 1 + i1 (11)
U (C2)
C2 Q2
C1 + = Q1 + (12)
1 + i1 1 + i1
i1 = I(C1) (13)
given the endowments Q1 and Q2.

Note that these are the same equilibrium conditions as those asso-
ciated with the model of Chapter 3 with the only difference that in
that model the interest rate parity condition, equation (13), takes
the form i1 = i∗1, where i∗1 is an exogenous constant.

40
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

11.4.2 Competitive Equilibrium

without

Government Intervention

41
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Use the interest rate parity condition (13) to eliminate the inter-
est rate i1 from the intertemporal budget constraint (12) and the
Euler (11) to obtain

C2 = Q2 + (1 + I(C1))(Q1 − C1) (14)

and
U 0 (C1)
0
= 1 + I(C1). (15)
U (C2)

These two equations determine the equilibrium levels of consumption

in periods 1 and 2.

42
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

The economy’s resource constraint

Let’s start by analyzing the effect of the debt-elastic interest rate

on the economy’s intertemporal resource constraint, equation (14).
• The key difference with the case of a constant interest rate is the
slope. When the interest rate is constant, the slope is also constant
and equal to minus 1 + i1. When the interest rate is debt elastic,
the slope is
∂C2 h
0
i
slope of IRC = = − 1 + I(C1) + I (C1)(C1 − Q1 )
∂C1
• As in the case of a constant interest rate, the slope is negative:
increasing C1 requires sacrificing some C2.

• If the country is a borrower (C1 > Q1), the slope is greater than 1+
i1 in absolute value. Intuitively, if the country borrows an additional
unit for consumption in period 1, in period 2 it must pay not only
1 + i1 but also the increase in the interest rate, I 0 (C1), caused by
the increase in debt.

Figure 11.5 on slide 45 plots the resource constraint.

43
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Equilibrium is shown in Figure 11.5 on the next slide.

The endowment is at point B, and the equilibrium is at point C.

Because C1e > Q1, the country is a borrower.

At point C, the indifference curve has a slope equal to minus 1 + i1,

as dictated by the Euler equation (15), which is less (in absolute
value) than the slope of the resource constraint, 1 + i1 < 1 + i1 +
I 0(C1)(C1 − Q1).

Clearly, there are points on the intertemporal resource constraint

that deliver higher levels of utility than point C. Therefore the equi-
librium is inefficient.

The inefficiency originates in the fact that private households per-

ceive the cost of C1 to be 1+i1, whereas it is 1+i1+I 0 (C1)(C1−Q1) >
1 + i1

This misperception induces households to ”overborrow”, ie consume

more in period 1 than is socially optimal.
44
 Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Figure 11.5: Equilibrium in an Economy with Borrowing Externalities

C2
A
B
Q2

A
Q1 C1e C1

The locus AA represents the economy’s intertemporal resource constraint, C2 =

Q2 + (1 + I(C1 ))(Q1 − C1 ). The endowment is at point B. The competitive
equilibrium without government intervention is at point C. This equilibrium is
inefficient because there are other allocations on the resource constraint that
yield higher utility than point C.
45
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

11.4.3 The Efficient Allocation

46
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Imagine a benevolent social planner who allocates C1 and C2 to

maximize households’ utility subject to the resource constraint (14)
max U (C1) + U (C2),
{C1,C2 }
subject to
C2 = Q2 + (1 + I(C1 ))(Q1 − C1). (14)
First-order condition:
U 0(C1) 0
0
= 1 + I(C 1 ) + I (C1)(C1 − Q1). (16)
U (C2)
• At efficient allocation the slope of the indifference curve equals
the slope of the economy’s intertemporal resource constraint.

• In the competitive equilibrium, the slope of the indifference curve,

U 0(C1)/U 0(C2), is equated to the private cost of funds, 1 + I(C1),
whereas in the efficient allocation it is equated to the social cost of
funds, 1 + I(C1 ) + I 0 (C1)(C1 − Q1).

• The excess external borrowing in the competitive equilibrium is

known as overborrowing.
47
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Figure 11.6: The Efficient Allocation in an Economy with

Borrowing Externalities

C2
A
B

A
Q1 opt
C1 C1e C1

The locus AA represents the economy’s resource constraint. The endowment is at

point B. The competitive equilibrium without government intervention is at point
C. The efficient allocation is at point D, where an indifference curve is tangent to
the resource constraint. At point C the economy overborrows, ie, borrows more
than is socially optimal.
48
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

11.4.4 Optimal Capital Control Policy

49
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Can the efficient allocation be achieved in a market economy

as opposed to a centrally planned economy? The answer is
yes. The government can eliminate overborrowing and achieve the
efficient allocation by imposing a capital control tax τ1 > 0. Suppose
i1 = i∗1 + τ1, with
τ1 = I 0(C̄1)(C̄1 − Q̄1)
In equilibrium C̄1 = C1, Q̄1 = Q1, and i∗1 = I(C1 ). Given this τ1, we
have
i1 = I(C1) + I 0(C1)(C1 − Q1 ).
Using this expression, the household’s optimality condition (8) be-
comes
U 0(C1) 0
= 1 + I(C 1 ) + I (C1)(C1 − Q1),
U 0(C2)
which is identical to the optimality condition of the social planner
(16).

50
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

• Because the government rebates the revenue generated by the

capital control tax to the households in a lump-sum fashion, the
economy’s resource constraint (14) is unchanged.

• Intuitively, the capital control tax increases the effective cost of

borrowing perceived by households, which induces them to cut con-
sumption in period 1. Thus, the role of the capital control tax is to
make households internalize that the social cost of an extra unit of
consumption is not just 1 + I(C1), but 1 + I(C1) + I 0 (C1)(C1 − Q1).

• Takeaways:
– in the presence of a borrowing externality free capital mobility
ceases to be optimal.
– in the economy without govt intervention, households consume
more and borrow more in period 1 than is socially optimal.
– capital controls now are desirable as a way to eliminate overbor-
rowing and to increase welfare.
51
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Overview of Sections 11.5-11.9

The current account of a large economy affects the world inter-

est rate. Therefore, a large economy has incentives to use capital
controls to manipulate the world interest rate in its favor.

• Section 11.5 characterizes analytically eqm in a two-country model

under free capital mobility.

• Section 11.6 does the same but graphically using an Egdeworth

box.

• Section 11.7 characterizes analytically optimal capital control pol-

icy (ie optimal manipulation of the world interest rate)

• Section 11.8 does the same but graphically using an Edgeworth

box.

• Section 11.9 considers the case of optimal capital controls with

retaliation.
52
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

11.5 Capital Mobility in a Large Economy

53
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

• When a large economy like the United States, the eurozone, or

China increases its demand for international funds, the world interest
rate will in general experience upward pressure.

• Each individual household in the large economy takes the interest

rate as exogenously given. But for the large country as a whole,
the interest rate is an endogenous variable. This means that the
government of a large economy might be able to apply policies to
manipulate world interest rates in the country’s favor. For example,
if the country is running a current account deficit, the government
could impose capital controls to curb the country’s aggregate exter-
nal borrowing and induce a fall in the world interest rate.

• In this section, we characterize equilibrium in a large economy

under free capital mobility. This analysis will serve as a building
block to characterize national policies aimed at manipulating the
world interest rate, which we will take up in Section 11.7.
54
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

A two-country model

Consider a two-period world consisting of two countries, the home

country, denoted h, and the foreign country, denoted f .

The home country receives a constant endowment over time.

By contrast, the foreign country receives a lower endowment in pe-

riod 1 than in period 2.

The two economies are identical in all other respects. In particular,

both have the same preferences for consumption and start period 1
with no assets or debts.

55
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Preferences in both countries take the form

j j
ln C1 + ln C2, (17)
where C1j and C2j denote consumption in periods 1 and 2, respec-
tively, in country j = h, f .

Budget constraint of households in country j in period 1

C1j + B1j = Qj1 , (18)
Budget constraint of households in country j in period 2
j j j
C2 = Q2 + (1 + ij )B1, (19)
where ij denotes the interest rate in country j, for j = h, f . Opti-
mization implies:
j
C2 j
j
= 1 + i (20)
C1
and
j
 
j 1 j Q2
C1 = Q1 + j
. (21)
2 1+i
56
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Consider the home country

Assume constant endowments over time: Qh

1 = Qh =Q
2
Setting j = h in (21), then yields
1 Q
 
h
C1 = Q+ . (22)
2 1+i h

Trade balance
h Q ih
T B1 = h
.
21+i
Current account schedule (recall B0h = 0).

h Q ih
CA1 = h
. (23)
21+i
Net foreign asset position

h Q ih
B1 = . (24)
2 1 + ih
57
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Consider the foreign country

f f
Assume Q1 = Q/2 and Q2 = Q

Set j = f in (21) to get

f 3 + if
C1 = Q f
. (25)
4(1 + i )
f
⇒ C1 > Q/2, for interest rate below 100 percent (i.e., for any if < 1).
Trade balance
f Q if − 1
T B1 = f
,
4 (1 + i )
f
Current account schedule (recallB0 = 0)

f Q if − 1
CA1 = f
. (26)
4 (1 + i )
Net foreign asset position
f Q if − 1
B1 = . (27)
4 (1 + if )
58
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Equilibrium

Market clearing in world financial markets

f
CAh
1 + CA1 = 0. (28)

Interest rate parity: under free capital mobility,

ih = if . (29)
Let this interest rate be denoted i∗ and let’s refer to it as the world
interest rate. That is, let ih = if = i∗.

59
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Solve for the equilibrium value of the world interest rate, i∗

Replace ih and if by i∗ in the home and foreign current account

schedules, equations (23) and (26). Then, use the resulting expres-
f
sions to eliminate CAh1 and CA 1 from (28) to obtain
Q i∗ Q i∗ − 1
+ = 0.
2 1 + i∗ 2 2(1 + i∗)
Solve for i∗, the equilibrium level of the world interest rate
1 ∗
i = ,
3
⇒ under free capital mobility the world interest rate is 33 percent.

By (26), the foreign country runs a current account deficit and

hence the domestic country a current account surplus. It follows
from (27), that the foreign country borrows internationally in period
1. In turn, if the foreign country borrows, then domestic country
must save in period 1.
60
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Solving for the equilibrium values of Cth and B1h

Setting ih = i∗ = 1/3 in equations (22) and (24) yields

7 1
C1h = Q < Q; and B1h = Q > 0.
8 8
From (19) we have:
7
C2h = Q > Q > C1h.
6

Interpretation: In spite of having a flat path of endowments, which,

if consumed, would produce a perfectly smooth path of consump-
tion, households in the home country choose to consume less than
their endowment in period 1 and to save. This is because foreign
demand for funds (discussed next) drives the world interest rate up,
inducing the home country to postpone consumption. As a result, in
period 2 the home country can enjoy a level of consumption higher
than its endowment.
61
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

f f
Solving for the equilibrium values of Ct and B1

Proceeding in an analogous fashion, we obtain the following equi-

librium values for the foreign country’s levels of consumption and
bond holdings:
f 5 1
C1 = Q > Q,
8 2
f 5
C2 = Q < Q,
6
and
1
B1f = − Q < 0.
8
Intuitively, facing an upward sloping path of endowments, the for-
eign country borrows in period 1 to smooth consumption over time.
So it consumes above its endowment in period 1, below its endow-
ment in period 2, and maintains a short bond position in period
1.
62
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

The Level of Welfare under Free Capital Mobility

Welfare under free capital mobility can be found by evaluating the

utility function (17) at the respective equilibrium consumption levels.
This yields
49
 
ln C1h + ln C2h = ln Q2 (30)
48
for the home country, and
25 2
 
f f
ln C1 + ln C2 = ln Q ,
48
for the foreign country.

Exercise 11.9 asks you to show that both countries are better off
under free capital mobility than under financial autarky. An implica-
tion of this result is that it does not pay for either country to impose
capital controls so high that all intertemporal trade stops.

But is there is a capital control policy that induces an equilibrium

in which the level of welfare is higher (for the country imposing the
controls) than in the eqm with free capital mobility?
63
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

11.6

Graphical Analysis of Equilibrium

under Free Capital Mobility

in a Large Economy

64
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Thus far we analyzed the equilibrium under free capital mobility in a

two-country world algebraically for a specific functional form of the
utility function, U (C) = ln C.

We will now repeat the analysis but instead of following an algebraic

approach we will use two powerful graphical objects in general equi-
librium analysis, the offer curve and the Edgeworth box, both created
by the Irish economist Francis Ysidro Edgeworth (1845-1926).

65
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

The Offer Curve

Given endowments, Q1 and Q2, we wish to characterize the optimal consumption

path (C1, C2) for different values of the interest rate.

Let i0 denote the interest rate such that C1 = Q1 and C2 = Q2. Next find the
optimal (C1 , C2) for other values of i1.

Starting from i1 = i0 consider a decline in the interest rate. This leads to C1 ↑

and C2 ↓.∗

Starting from i1 = i0, consider higher values of i1, by the substitution effect C1 ↓
and C2 ↑.

The figure on the next slide shows how (C1 , C2 ) change with the interest rate.

It follows that the optimal consumption path as the interest rate declines describes
a downward sloping locus in the space (C1 , C2) that crosses the endowment point.
This locus is the offer curve.

The figure on the slide after the next plots the offer curve.

∗
If the income effect associated with a decline in interest rates dominates the substitution effect, it is possible that C2 ↑. In what follows we assume that the substitution
effect always dominates.

66
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Figure 11.7: Optimal Intertemporal Consumption Choice at

Different Interest Rates
C2j

I2 I1 I0

Qj2 A0

A1
slope = −(1 + i 0 )
A2
slope = −(1 + i 1 )

slope = −(1 + i 2 )

I0 I1 I2

C1j
Qj1

The figure displays the optimal consumption choice for three different values of the interest rate, i0 , i1 , and i2 , satisfying i0 > i1 > i2 . Each interest rate is associated
with a different intertemporal budget constraint. The higher the interest rate is, the steeper the intertemporal budget constraint will be. The intertemporal budget
constraint I 0 I 0 is induced by the highest of the three interest rates, and the intertemporal budget constraint I 2 I 2 by the lowest. The associated optimal consumption
path induced by the interest rate associated with budget constraint I 0 I 0 is the endowment point, A0 . The intertemporal budget constraints I 1 I 1 and I 2 I 2 produce
optimal consumption choices given by points A1 and A2 , respectively. The offer curve (not shown) connects points A0 , A1 , A2 .

67
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Figure 11.8: The Offer Curve

j
C2
J

A0
j
Q2
A1
A2
J

j j
Q1 C1

The offer curve is the locus JJ, which connects all optimal consumption allocations at different
interest rates. The offer curve crosses the endowment point A0 . The figure also shows the
indifference curve that crosses the endowment point. At the endowment point this indifference
curve is tangent to the offer curve. All points on the offer curve other than the endowment point
are preferred to the endowment point itself.
68
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

The Edgeworth Box

We now construct the Edgeworth Box, which is shown on the next slide.

• The length of the horizontal side of the box is the global endowment of goods
in period 1, Qh1 + Qf1.

• The height of the box is the global endowment in period 2, Qh2 + Qf2.

• The southwest corner of the box is the origin of the foreign country and is
indicated by the symbol O f . For the foreign country, consumption and the en-
dowment in period 1 are measured on the horizontal axis from the origin O f to
the right, and consumption and the endowment in period 2 are measured on the
vertical axis from O f upward. Welfare of households in the foreign country rises
as the allocation (C1f , C2f ) moves northeast in the box.

• The northeast corner of the box is the origin of the home country and is indicated
by the symbol O h . For this country, consumption and the endowment in period
1 are measured on the horizontal axis from O h to the left, and consumption and
the endowment in period 2 are measured on the vertical axis from O h downward.
Welfare of households in the home country rises as the allocation (C1h , C2h ) moves
southwest in the box.

• The endowments of the two countries are given by point A0.

• Any point in the box represents an allocation of consumption across time and
countries that can be achieved with the existing global endowments.
69
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

The Edgeworth Box — before adding offer curves

Qh1
Oh
Qf2 + Qh2

0
Qh2
A

Qf2

Qf1 + Qh1
Of
Qf1

The width of the box is equal to the world endowment of goods in period 1, Qf1 + Qh1. The heights
of the box is equal to the world endowment of goods in period 2, Qf2 + Qh2 . The origin of the
foreign country is O f and the origin of the home country is O h . The endowment point is A0.

70
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

• Next, we add the offer curves to the Edgeworth box. This is shown
in the figure on the next slide.

• The offer curve of the foreign country is the locus F F .

• The offer curve of the home country is the locus HH.

• Both offer curves must cross the endowment point A0.

71
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

The Edgeworth Box — after adding the offer curves

Qh1
Oh
Qf2 + Qh2
F Qh2
A0
H

Qf2

Qf1 + Qh1
Of
Qf1

The offer curve of the foreign country is the locus F F , and the offer curve of the home country is
the locus HH. Both offer curves must cross the endowment point, A0. For a given country, any
point on the country’s offer curve is preferred to the endowment point. The slope of a straight
line from a point on the offer curve, of country f say, to A0 is equal to −(1 + if ). For points to
the right of A0 the interest rate if is lower than the autarky interest rate and for points on the
offer curve FF to the left of A0 the interest rate supporting that allocation, if , is higher.

72
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Equilibrium Under Free Capital Mobility

The figure on the next slide presents the equilibrium allocation.

• The equilibrium under free capital mobility is given by point B, where the two
offer curves intersect for a second time.

• In equilibrium, the foreign country, which has a relatively low endowment in

period 1, borrows from the home country.

• The equilibrium world interest rate, i∗, is determined by the slope of the line
that connects points A0 and B. This line is the intertemporal budget constraint
faced by the domestic and foreign households at the equilibrium world interest
rate i∗ . This interest rate is lower than the domestic interest rate in the foreign
country under financial autarky, which is determined by the slope of the foreign
household’s indifference curve at the endowment point A0. By the same logic,
we have that the equilibrium interest rate under free capital mobility, i∗, is higher
than the domestic interest rate in the home country under financial autarky.

• We have therefore established that allowing for free capital mobility eliminates
interest rate differentials.

• Because in equilibrium both countries are on their respective offer curves, they
are both better off than under autarky. Thus free capital mobility is welfare
improving.

73
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Figure 11.9: Equilibrium Under Free Capital Mobility in a Two-

Country Model
Qh1
Oh
Qf2 + Qh2
F Qh2
A0
H

Qf2
B
F

H
← slope= −(1 + i∗ )

Qf1 + Qh1
Of
Qf1

The equilibrium under free capital mobility is point B. The slope of the line that connects points
A0 and point B is −(1 + i∗ ), where i∗ is the equilibrium world interest rate under free capital
mobility.

74
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

The Allocation Under Free Capital Mobility is Pareto Optimal

• At point B the indifference curves of the home and foreign house-

holds both have a slope equal to −(1 + i∗).

• Thus, at point B the indifference curves of the home and foreign

households are tangent to each other.

• This implies that at no point inside the Edgeworth box can both
countries be better off than at point B. In other words, any other
attainable consumption allocation makes at least one country worse
off relative to the allocation associated with the equilibrium under
free capital mobility.

• When an equilibrium has this property, we say that it is Pareto

optimal.

75
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

11.7 Optimal Capital Controls in a Large Economy

• A large country that is borrowing has monopsony power in inter-

national funds markets. Unlike the country as a whole, individual
households do not have market power in financial markets as they
are atomistically small. Thus, exploiting the country’s market power
can only be achieved via government intervention.

• For a large country that is a borrower, capital controls have two

opposing macroeconomic effects: (1) they distort the intertemporal
allocation, which is welfare decreasing. (2) they lower the world
interest rate, which is welfare increasing. We will now present an
economy in which this tradeoff is resolved in favor of (2).

• Let’s go back to the algebraic example introduced in Section 11.5.

Recall the foreign country was the borrower and the home country
the lender.
76
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

• Assume that in response to capital controls imposed by the foreign

country, the home country does not retaliate by imposing its own
capital controls.

⇒ demand for international funds by home country still is

Qi∗
B1h = ∗
. (31)
21+i

• We turn to the setting of the optimal capital control tax, τ , in

country f next.

77
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

In Section 11.1 we analyzed an economy with capital control taxes

and here we will repeat some of these results to keep the presentation
self-contained.

• The capital control tax τ creates a wedge between the world in-
terest rate and the interest rate in the foreign country,

if = i∗ + τ. (32)

• Government rebates capital control tax revenue via lump sum

transfers, T
f
T = −τ B1 . (33)

• Budget constraints of the household:

f fQ
C1 + B1 = (34)
2
f f
C2 = Q + T + (1 + if )B1 (35)

78
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Combining (32), (33), and (35), in eqm we have that

f f
C2 = Q + (1 + i∗)B1 , (36)
which implies that country f, in eqm, does not loose any resources
due to the capital control tax.

The government of country f internalizes that:

f
(1) in eqm world capital market clears: B1 + B1h = 0
(2) demand by country h for funds is: B1h = Q i∗
2 1+i∗
f f
(3) in eqm total period 2 resources are: C2 = Q + (1 + i∗)B1 , and
(4) C1f + B1f = Q
2.

f f
From here we can express C1 and C2 as functions of the world
interest rate i∗
f Q 1 + 2i∗
C1 = (37)
2 1 + i∗
and
f Q
C2 = (2 − i∗). (38)
2
79
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

f f
Now use these two expressions to eliminate C1 and C2 from the
utility function of the foreign household to obtain

Q 1 + 2i∗
!
Q
 
ln + ln (2 − i∗) .
2 1 + i∗ 2

This is the indirect lifetime utility function because it is expressed in

terms of a price, the world interest rate, i∗, instead of consumption,
f f
C1 and C2 .

80
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

The government of country f solves:

Q 1 + 2i∗
" ! #
Q

max ln + ln (2 − i∗)
{i∗ } 2 1 + i∗ 2

Taking the derivative of the indirect utility function with respect to

i∗ and setting it to zero yields
∗ 2 ∗ 1
i + 2i − = 0,
2
q
The two solutions are: i∗ = −1± 3 . The only economically sensible
2
(ie i∗ > −1) solution is
s
3
= 0.22, i∗ = −1 +
2
or 22 percent. Under optimal capital controls the world interest rate
is lower than under free capital mobility (22 versus 33 percent).

Intuitively, because the foreign country is a borrower it benefits

from a lower world interest rate. (By the same token, the home
country, a lender, does not benefit.)
81
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Country f welfare: ln( 25.2122

48 Q2 ) under optimal capital controls (ver-
25 2
sus ln( 48 Q ) under free capital mobility)

Country h welfare? Find consumption from:

f f f f
C1h = Qh
1 + Q1 − C1 ; and C2h = Qh
2 + Q2 − C2

Country h welfare: ln(1.0103Q2) under optimal capital controls (of

country f) (versus ln(1.0208Q2) under free capital mobility).

Optimal capital controls raise welfare in country f and lower it in

country h.

82
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

The domestic interest rate in country f under optimal capital

controls

Find if : From Euler equation:

f
1+i = C2 = 1.5
f
C1
That is, the domestic interest rate increases from if = 0.33 under
free capital mobility to if = 0.5. Why? Because the gov’t must
incentivate private households to borrow less.

Find τ :
τ = if − i∗ = 0.5 − 0.22 = 0.28

The tax rate on capital inflows is a hefty 28 percent.

83
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Because the capital controls raise the domestic interest rate (if ),
the current account deficit of the foreign economy falls from

f
B1 = −0.1250Q under free capital mobility to

f
B1 = −0.0918Q under optimal capital controls.

84
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

11.8 Graphical Analysis of Optimal Capital Controls in a Large

Economy

85
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Let’s now use the Edgeworth box to analyze optimal capital controls
in the foreign country, see Figure 11.11 on the next slide.
• In setting capital controls, the objective of the foreign country is to attain a
point on the home country’s offer curve, HH, that maximizes the foreign country’s
utility. The foreign country is constrained to pick a point on the home country’s
offer curve because, by construction, only the allocations on the offer curve can
be obtained as a market outcome, that is, by an appropriate choice of the world
interest rate. Consequently, the allocation associated with the optimal capital
control policy is one at which an indifference curve of the foreign country is
tangent to the offer curve of the home country.
• This allocation is point C in Figure 11.11. The indifference curve attained by
the foreign country under optimal capital controls is U U .
0
• The world interest rate under optimal capital controls (i∗ ), is defined by the
slope of the line that connects points A0 and C. Clearly, this line is flatter than
the one connecting points A0 and B. This means that the imposition of optimal
capital controls causes the world interest rate to fall.
• Also, the optimal capital control policy in the foreign country makes the foreign
country better off at the expense of the home country, whose welfare goes down.
Recall that the equilibrium under free capital mobility is Pareto optimal, so the
improvement in the foreign country’s welfare must be welfare decreasing for the
home country.
• These results echo those obtained algebraically in Section 11.7.
86
 Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Figure 11.11 Optimal Capital Controls in a Large Economy

Qh1
Oh
Qf2 + Qh2
F Qh2
A0 U
H

Qf2 U
B
F 0
slope = −(1 + i∗ )

H slope = −(1 + i∗ )

Qf1 + Qh1
Of
Qf1
The offer curve of the home country is HH and that of the foreign country is F F . The endowment
is at point A0. The equilibrium under free capital mobility is at point B, and the equilibrium under
optimal capital controls in the foreign country is at0 point C. The interest rate under free capital
mobility is i∗ and under optimal capital controls i∗ < i∗ . The indifference curve attained by the
foreign country under optimal capital controls is U U . The fact that this indifference curve is not
tangent to the intertemporal budget constraint that crosses point C implies that the equilibrium
with capital controls is Pareto inefficient.
87
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

The Allocation Under Optimal Capital Controls is Inefficient

• The equilibrium under optimal capital controls in the foreign coun-

try fails to be efficient. Let’s show why.

• At point C, by construction the indifference curve of the home

country (not shown) must be tangent to the intertemporal budget
0
constraint associated with the interest rate i∗ .

•The indifference curve of the foreign country (the locus UU), on

the other hand, is tangent to the offer curve of country h (the locus
HH).

• Since the budget constraint and the offer curve (the locus HH)
intersect at C, the slopes of the home and foreign indifference curves
at point C are not the same.

• This implies that the equilibrium allocation under capital controls

is inefficient in the sense that the home country could be made
better off without making the foreign country worse off.
88
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

11.9 Retaliation

89
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

• Thus far, we have assumed that, as the foreign country imposes

controls on capital inflows, the home country does not retaliate by
imposing its own restrictions on capital flows.

• Because the home country is a large economy, it will in general

have an incentive to retaliate. In turn, the foreign country would
have an incentive to readjust its capital control policy in response
to the retaliation of the home country.

• The equilibrium that will emerge under this strategic interaction

depends on what type of game the two countries play in setting
capital controls. We will focus on one type of game known as Nash
equilibrium.

• Essentially, in a Nash equilibrium each country sets its own capital

control tax optimally taking as given the capital tax rate of the other
country. An equilibrium is reached when the capital control tax that
each country takes as given is indeed the tax rate that is optimal
for the other country.
90
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Equilibrium Requirements

As shown earlier, the intertemporal resource constraint with capital

controls is the same as in the case without capital controls, that is,
for j = h, f , we have
j j
j C2 j Q2
C1 + ∗
= Q1 + ∗
. (39)
1+i 1+i

And the Euler equations continue to be requirements of equilibrium:

j
C2 j
j
= 1 + i . (40)
C1

91
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Consumption in the Foreign Country

The foreign country imposes controls on capital inflows:

if = i∗ + τ f ,

Setting j = f in (39) and (40) and solving for C1f and C2f , yields
f
f Q2
f f ∗ f
Q1 + 1+i∗
C1 = K (i , τ ) ≡
1+i∗+τ f
1 + 1+i∗
and
f Qf2
f f ∗ f ∗ f
Q1 + 1+i∗
C2 = L (i , τ ) ≡ (1 + i + τ )
1+i∗ +τ f
1 + 1+i∗

where i∗ and τ f are endogenous variables.

92
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Consumption in the Home Country

The home country imposes controls on capital outflows.

ih = i∗ − τ h .
Setting j = h in (39) and (40) and solving for C1h and C2h, yields

h Qh2
h h ∗ h
Q1 + 1+i ∗
C1 = K (i , τ ) ≡
1+i∗ −τ h
1 + 1+i∗
and
h Qh
2
h h ∗ h ∗ h
Q1 + 1+i ∗
C2 = L (i , τ ) ≡ (1 + i − τ ) ,
1+i∗−τ h
1 + 1+i∗

where i∗ and τ h are endogenous variables to be determined in equi-

librium.
93
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Market clearing in the goods market in period 1 requires that

global consumption equal the global endowment,
f
K f (i∗, τ f ) + K h(i∗, τ h) − Q1 − Qh
1 = 0.
This equation expresses the world interest rate as an implicit function
of the tax rates in the home and foreign countries. We then write

i∗ = I(τ f , τ h).
Using this relation to eliminate i∗ from consumption in both periods
in the home and foreign countries we can write
f
C1 = K̃ f (τ f , τ h) ≡ K f (I(τ f , τ h), τ f ),

f
C2 = L̃f (τ f , τ h) ≡ Lf (I(τ f , τ h), τ f ),

C1h = K̃ h(τ f , τ h) ≡ K h(I(τ f , τ h), τ h),

C2h = L̃h(τ f , τ h) ≡ Lh(I(τ f , τ h), τ h).

94
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

The government of the foreign country picks τ f to maximize the

utility of the foreign household taking as given the tax rate in the
home country, τ h.
h i
f f h f f h
max ln K̃ (τ , τ ) + ln L̃ (τ , τ )
{τ f }
The first-order condition is
f f
K̃1 (τ f , τ h) L̃1(τ f , τ h)
f f h
+ f f h = 0,
K̃ (τ , τ ) L̃ (τ , τ )
f f
where K̃1 (τ f , τ h) and L̃1(τ f , τ h) denote, respectively, the partial
derivatives of K̃ f (τ f , τ h) and L̃f (τ f , τ h) with respect to the first
argument, τ f . This optimality condition implicitly defines the tax
rate in the foreign country, τ f , as a function of the tax rate in the
home country, τ h. We write the solution for τ f as
τ f = Rf (τ h).
This relationship is called the reaction function of the foreign coun-
try. It represents the optimal tax response of the foreign country as
a function of the tax rate in the home country.
95
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Likewise, the government of the home country chooses τ h to

h f h h f h
h i
max ln K̃ (τ , τ ) + ln L̃ (τ , τ ) ,
{τ h}

taking as given τ f .

The associated first-order condition is

K̃1h(τ f , τ h) L̃h
1 (τ f , τ h)
h f h
+ h f h = 0.
K̃ (τ , τ ) L̃ (τ , τ )

Solving this expression for τ h, we can write

τ h = Rh(τ f ),
which is the reaction function of the home country.

96
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Figure 11.12: on the next slide displays the reaction functions of the
home and foreign countries in the space (τ h, τ f ) for the algebraic
example studied in Section 11.5.
• The Nash equilibrium is at point A, where the two reaction functions intersect.

• In the Nash equilibrium τ f = 0.18 and τ h = 0.30.

• The cross country interest rate differential, if − ih = τ f + τ h , widens from 28

(=28+0) percent when the home country is passive to 48(=18+30) percent when
the home country retaliates.

• Point B in the figure corresponds to the case in which the foreign country
behaves strategically and the home is passive, which is the case we studied in
Section . Comparing points A and B, we see that retaliation by the home country
makes the foreign country lower its capital control tax rate (28 versus 18 percent).

• Point C in the figure corresponds to the case in which the home country behaves
strategically and the foreign country is passive.

• Regardless of whether a country is borrowing or lending, retaliation by the other

country lowers its own capital control taxes relative to the situation in which the
other country is passive.

97
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Figure 11.12: Capital Control Reaction Functions of the Home

and Foreign Governments
0.5
Rf (τ h )
Rh (τ f )
0.45

0.4

0.35

0.3
B
τf

0.25

0.2
A

0.15

0.1

0.05

C
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
τh

The function Rf (τ h) is the reaction function of the foreign country. It expresses the optimal
capital control tax rate of the foreign country as a function of the tax rate of the home country.
Similarly, Rh(τ f ) is the reaction function of the home country, representing the optimal tax rate
in the home country as a function of the foreign country’s tax rate. The intersection of the
two reaction functions gives the Nash equilibrium capital control tax rates in the two countries.
Replication file: tauf_tauh_num.m in two_country.zip.
98
 Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Table 11.1: Comparison of Equilibria Under Alternative Capital Control Policies

Welfare
Policy i ∗ τ f τ h CA f f h
Autarky – – – 0 ln(0.5000Q ) ln(1.0000Q2 )
2

Free Capital Mobility 0.33 0 0 -0.125Q ln(0.5208Q2) ln(1.0208Q2 )

Home Country Passive 0.22 0.28 0 -0.092Q ln(0.5253Q2) ln(1.0103Q2 )
Foreign Country Passive 0.55 0 0.35 -0.073Q ln(0.5082Q2) ln(1.0318Q2 )
Retaliation—Nash Eqm 0.45 0.18 0.30 -0.060Q ln(0.5111Q2) ln(1.0219Q2 )

Comments:
• Not surprisingly, intertemporal trade, as measured by the absolute size of the current account,
is the largest under free capital mobility and the smallest under optimal capital controls with Nash
retaliation.
• Also not surprisingly, a country’s welfare is the highest when it imposes optimal capital controls
and the other country is passive.
• It is somewhat surprising, however, that the home country is better off under optimal capital
controls with retaliation than under free capital mobility. Thus, it is optimal for the home country
to impose capital controls regardless of whether this triggers a capital control war or not. This is
not the case for the foreign country, which prefers free capital mobility to a capital control war.
• An interesting question is whether it pays for the foreign country to compensate the home
country for abiding to free capital mobility.
• Finally, if one country imposes optimal capital controls unilaterally, it is in the interest of the other
country to retaliate. To see this, note that in both countries welfare is higher under retaliation
than in the absence thereof.

99
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Summing Up

• Capital controls drive a wedge between the domestic interest rate

and the world interest rate.

• If under free capital mobility, a small country borrows from the rest
of the world, then the imposition of capital controls drives domestic
interest rates up, depresses current consumption, and improves the
current account.

• If under free capital mobility, a small country lends to the rest

of the world, then the imposition of capital outflow controls lowers
domestic interest rates, increases current consumption, and worsens
the current account.

• In a small open economy without distortions, capital controls are

always welfare decreasing.
100
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Summing Up (continued)

• In the presence of borrowing externalities, capital controls can be

welfare increasing, as they can be effective in eliminating overbor-
rowing.

• In a two-country world, free capital mobility is in general preferred

to financial autarky.

• In a two-country world, free capital mobility results in a Pareto

optimal allocation, that is, any other feasible allocation makes at
least one country worse off.

• For an economy that has market power in global financial markets

it might be welfare improving to impose capital controls to move
the world interest rate in its favor.

101
Schmitt-Grohé, Uribe, Woodford “International Macroeconomics” Slides for Chapter 11: Capital Controls

Summing Up (concluded)

• A large economy that runs a current account deficit benefits from

imposing controls on capital inflows that drive the world interest
rate down provided the rest of the world does not retaliate.

• A large economy that runs a current account surplus benefits from

imposing controls on capital outflows that drive the world interest
rate up provided the rest of the world does not retaliate.

• In a two-country world, the allocation under optimal capital con-

trols fails to be Pareto efficient, that is, there is a feasible realloca-
tion of resources that would make at least one country better off
without making the other country worse off.

• In a two-country world, if one country imposes optimal capital

controls unilaterally, it is in the interest of the other country to
retaliate.

• In a two-country world, it may be welfare improving for one country

to initiate a capital control war.
102