Current accounts in the long and short run

Digitized

by

the

Internet

Archive

in

2011

with

funding

from

Boston

Library

Consortium

Member

Libraries

31

415

dewe

Massachusetts

Institute

of

Technology

Department

of

Economics

Working Paper

Series

Current

Accounts

in

the

Long

and

Short

Run

Aart

Kraay

Jaume

Ventura

Working

Paper

02-21

June

2002

Room

E52-251

50

Memorial

Drive

Cambridge,

MA

02142

This

paper

can

be

downloaded

without

charge

from

the

Social

Science

Research Network

Paper

Collection

at

MASSACHUSETTSINSTITUTE OFTECHNOLOGY

JUN

1

3

2002

'

Massachusetts

Institute

of

Technology

Department

of

Economics

Working

Paper

Series

Current

Accounts

in

the

Long

and

Short

Run

Aart

Kraay

Jaume

Ventura

Working

Paper

02-21

June

2002

Room

E52-251

50

Memorial

Drive

Cambridge,

MA

02142

This

paper

can be

downloaded

without

charge from

the

Social

Science

Research Network

Paper

Collection

at

Revised

Draft

Current

Accounts

in

the

Long and

Short

Run

Aart

Kraay*

The

World

Bank

Jaume

Ventura

MIT,

CEPR

and

NBER

June 2002

Abstract:

Faced

with

income

fluctuations, countries

smooth

their

consumption by

raising savings

when

income

is high,

and

viceversa.

How

much

ofthese savings

do

countries invest at

home

and abroad?

In otherwords,

what

arethe effects of

fluctuations in savings

on domestic

investment

and

the current

account?

Inthe long

run,

we

find thatcountries investthe marginal unit ofsavings in

domestic

and

foreign

assets inthe

same

proportions

as

intheirinitial portfolio,

so

thatthelatteris remarkably

stable. Inthe short run,

we

find thatcountries investthe marginal unitofsavings mostly

inforeign assets,

and

only gradually

do

they rebalancetheir portfolio

back

to itsoriginal

composition. This

means

that countries not onlytry to

smooth

consumption, butalso

domestic

investment.

To

achievethis, they

use

foreignassets

as

a buffer stock.

*_{Thispaperhasbeenpreparedforthe2002}

_NBER

_{MacroeconomicsAnnual.}

We

_{aregrateful toour}

discussantsFabrizioPerriandEricvanWincoopaswell astothe participantsfor theirusefulcomments.

Theopinionsexpressed hereare the authors',and do notnecessarilyreflectthoseoftheWorldBank,its

Countries are subject to transitory

income

shocks such

as

changes

inthe

terms

oftrade, fluctuations in production, policyreforms, natural disasters,

and

many

others.

There

is

ample

evidence

thatcountries

use

theirassets to bufferor

smooth

the effects ofthese

shocks

on consumption,

raising savings

when

income

is high

and

viceversa.1

The

main

goal ofthis

paper

is to

improve

our understanding ofthecombination of

assetsthatcountries

use

forthis purpose. In particular,

we

ask:

how

do

countries

allocatethe marginal unitof savings

between

domestic

and

foreign assets? Or,

equivalently,

what

are the effects of fluctuations in savings

on domestic

investment

and

thecurrent

account?

2

The

traditional

view

is thatcountries invest the marginal unitofsavings in

foreign assets. Underlying this

view

are the

assumptions

that investment risk is

weak

and

diminishing returns are strong.

The

first

assumption ensures

thatcountries invest

theirsavings only in

those

assetsthat offerthe highest

expected

return.

The second

assumption

implies that investing

any

fraction ofthe marginal unit ofsavings in

domestic

capital

would

lowerits

expected

return

below

that offoreign assets.

Hence

the marginal unit ofsavings is invested in foreign assets,justifying thetraditional rule thatfluctuations in savings leadtofluctuations in the current

account

of roughlythe

same

magnitude.

While

theoreticallycoherent, this rule

has

been

consistently rejected

by

the data.

The

top panel of Figure 1

shows

pooled

annual

observations ofthe current

account

and

savingsfor21

OECD

countries

over

thepast

30

years.

A

regression ofthe

current

account

on

savings delivers a slope coefficientthat is positive but

much

lower

than one. This is nothing butthe

famous

resultofFeldstein

and

Horioka [1980] that fluctuations in savings leadto parallelfluctuations in investment,with only

minor

effects

on

the currentaccount.

1

Forevidenceonconsumption-smoothing,see Deaton[1991]whowrites that"consumptionislessvolatile

than income,itfluctuates lessaboutitstrend,theamplitudeofitsbusinesscycle variationisless,andthe

varianceofitsgrowthrate islessthanthevarianceofthegrowthrateofincome"p.133-34.

2

Why

docountriesuseassetstosmooth consumption ratherthan simplybuyinsuranceabroad?Implicitin

thisparagraph andbasicallyinallthat followsistheassumptionthatcountriesareunableor unwillingto

selltheiridiosyncraticrisk.Thisassumptionisa centraltenetofthe intertemporalapproachtothe current

In

an

earlierpaper,

we

proposed

a

new

view

thatcountries invest the marginal

unit ofsavings as the

average

one

(Kraay

and Ventura

[2000]). Thisis

what

one

should

expect if, in contrasttothetraditionalview, investment riskis strong

and

diminishing

returns are

weak.

The

first

assumption

impliesthat countries are unwilling to

change

the composition oftheir portfolios, unless

shocks

have

large effects

on

thedistribution of asset returns.

The

second assumption ensures

thatthedistribution ofasset returns

is unaffected by the

way

countries investthe marginal unit ofsavings.

Hence,

the marginal unitofsavings isinvested as the

average

one, leading tothe

new

rulethat fluctuations insavings lead tofluctuations inthe current

account

thatare equal to

savingstimesthe

share

offoreign assets in thecountryportfolio. Thisrule notonly is

theoreticallycoherent, butit also providesasurprisingly

good

description ofthe data.

The

bottom

panel of Figure 1

shows

that a simple regression ofthe current

account

on

the interaction

between

savings

and

the shareofforeign assetsdelivers a slope

coefficientcloseto

one

and

a zero intercept. Moreover, this interaction

term by

itself

explains

around

thirtypercent ofthe

observed

variation in the currentaccount.3

Hidden

in the

bottom

panel ofFigure 1 there isa vast difference

between

the

predictive

power

ofthe

new

rule inthe long

and

the shortrun. Figure2 illustrates this

point. In thetop panel,

we

have

plottedthe

average

current

account

over a thirty-year

period against the

average

ofsavings time the

share

of foreign assets during the

same

period.

The

new

ruleexplains about

85

percentofthe long run or

average

cross-country differences in current accounts. Inthe

bottom

panel,

we

have

plotted the

(demeaned)

current

account

for

each

country

and year

againstthe

(demeaned)

interaction ofsavings

and

the initial share offoreign assetsin wealthforthe

same

country

and

year.

The

new

ruleexplains basically

none

ofthe year-to-year

within-reality.Thequestionofwhythisissoisoneofthemostintriguingpuzzlesininternationalfinance.See

Lewis[1999]forasurveyoftheliteratureonthis topic.

3

Sinceforeignassetsconstituteasmallfractionofobservedcountryportfolios, thisviewimplies that

fluctuationsinsavingsshould mostlyleadtoparallelfluctuationsininvestment,andistherefore consistent withthe Feldstein-Horiokafinding.What wefound mostsurprisingaboutthis viewinourearlierpaperis

thatithassharplydifferentimplicationsforthe currentaccountresponsetoanincreaseinsavingsin

debtorandcreditorcountries.Since debtors bydefinitionhold morethantheirwealth indomesticcapital,

theyinvestat

home

morethantheincreaseinsavings,resultingina currentaccountdeficit. Incontrast, creditorcountries investat

home

lessthanthe increaseinsavings,resulting inacurrentaccountsurplus.

country differences in current accounts.

The

contrast

between

the

two

panels indicates

adiscrepancy

between

the long

and

the short run behavior ofthecurrent account.4

Isthe short-run relationship

between

savings

and

the current

account

just noise

orare there clearpatterns behind this cloud of points?

How

do

we

reconcile the apparently

haphazard

behaviorofthe current

account

in the shortrun with its neat

behaviorinthe long run?

The

main

contribution ofthis paper,

we

think, is to provide

clear

answers

to thesequestions.

To

do

this, it is useful to start

by

pointing outthatthe

new

rule

embodies

the

view

thatthe current

account

primarily reflects portfoliogrowth,

i.e.

changes

inthe size ofthecountry portfoliowithout systematic

changes

in its

composition.

The

empirical

success

ofthe

new

rulein thetop panelof Figure

2

simply

reflectsthe observationthat thecomposition ofcountryportfolios

has

been

remarkably

stable in the long run. This is

shown

in Figure 3. If

we

want

to

understand

why

the

new

rule

performs so

poorly inthe

bottom

panelofFigure 2,

we

must

explain

how

and

why

in the short run increasesin savings lead mostlyto portfoliorebalancing, i.e. systematic

changes

inthe composition ofthecountry portfolio. Ifin addition

we

want

to reconcile the

two

panels ofFigure2,

we

must go

further

and

also explain

why

this short-run

portfolio rebalancing is

undone

inthe long run.

Our

hypothesis isthatthe pattern of portfolio rebalancing is consistentwiththe

view

that

adjustment

coststo investment are important. Ifthis is the case,

an

increase

in savingsthat raisesinvestment

reduces

the

expected

return to capital

and

induces

countries to rebalancetheir portfolios

towards

foreign assets.

Under

these

conditions,

theshort-run current

account

surplus is largerthan the

one

predicted

by

the

new

rule.

Once

savings returns tonormal, investmentdeclines,

adjustment

costs

disappear

and

thecountry portfolio returnsgraduallytoits original composition.

Throughout

this

adjustment

process, the current

account

surplus is smaller than the

one

predicted by

4

We

alsonotedthisdiscrepancyinourearlierpaper, althoughitwasmuchlesspronouncedinthesmaller

sampleof13countriesand 23years(1973-1995)that

we

usedthere.Here,

we

have been abletoextend oursampleto21 countriesand upto32years per country (1966-1997).Allthe resultsobtained inthe

the

new

rule. In the long run, the

shock does

not affectthecomposition ofthecountry

portfolio

and

the

new

rule applies.

With

thistheoretical pictureat hand,

we

go

back

tothedatato searchfor

patternsin thediscrepancies

between

the

observed

current

account

and what

the

new

rule

would

predict.

When

we

do

this, the picturethat

comes

out

from

thedata turns out

to

be

clear

and unambiguous: on

impact, countries rebalancetheir portfolios

towards

foreign assets

and

the

new

rulesystematically under-predicts the short-run effects of

increases insavings

on

the currentaccount. Intheyearsthat follow, countries

rebalancetheirportfolios

back towards

their original composition. During this period,

the

new

rulesystematically over-predictsthe currentaccount.

We

findthat the

whole

adjustment

process

lasts

about

fiveyears. Overall, theevidence isconsistent with the

view

thatadjustment coststo investment are important and, to avoid paying

them,

countries

use

foreign assets as a bufferstock to

smooth

fluctuations in investment.

The

theory presented here

can

also reconcile

two

apparentlycontradictory

observations

about

the relationship

between

the current

account

and

investment.

On

the

one

hand, the long run or cross-sectional correlation

between

investment

and

the

current

account

is

weak

(Penati

and Dooley

[1984],

Tesar

[1991]).

On

the otherhand,

theshort run or time-series correlation

between

investment

and

the current

account

is

consistently negative (Glick

and Rogoff

[1995]).

The

theory presented here predicts

that inthelong run, portfolio rebalancing issmall

and

the correlation

between

the

current

account

and

investment should

be

positive in creditorcountries

and

negative in

debtor ones.

We

show

thatthedata is consistent withthis prediction

and

thatthe

weak

cross-sectional correlation isthe resultof pooling datafrom debtor

and

creditor

countries.

The

theoryalso predicts thatin theshort run portfolio rebalancing is

important

and

this introduces a

source

ofnegative correlation

between

thecurrent

account

and

investment. This istrue in allcountries, regardlessof

whether

they are debtors orcreditors.

We

present a simple

decomposition

ofthe cross-sectional

and

time-series correlations

between

the current

account

and

investmentthat illustrates this

The

paper

isorganized

as

follows: Section I presents a stylized

model

that

encapsulatesthe

main elements

ofour portfolio-basedtheoryofthe currentaccount. Section II

uses

the

model

to study

how

countries reactto

income

shocks. Section III

examines

the empirical

evidence

and

interprets it

from

the

vantage

point ofthe theory.

Section IVinvestigatesthe relationship

between

investment

and

the currentaccount. Section

V

concludes.

I.

An

Intertemporal

Model

of

the

Current

Account

In this section,

we

presenta stylized

model

of

how

the current

account

responds

totransitory

income

shocks. Since

we

stop short of

modeling

the world

equilibrium

and

focus instead

on

a small

open economy,

these

shocks

should

be

interpreted as country-specific or idiosyncraticrisk. Following thetradition ofthe

intertemporal approach,

we

simply

assume

that countries are unableor unwilling tosell

this riskin international markets. In particular,

we

adopt

the starkest

form

ofthis

view

by

assuming

that the onlyassetthat is traded internationallyis a non-contingent bond. 5

The

model

captures

what

we

think are the essential

elements

of a

portfolio-based

theoryofthe current account. This theory is built

around

the

concept

ofcountry

portfolio

and

a simple

decomposition

ofthe current

account

that relies

on

thisconcept.

By

the country portfolio,

we

referto the

sum

ofall productive assets located within the country plus its netforeign asset position.

The

latterconsist of the

sum

ofall claims

on

domestic

assets held

by

foreigners

minus

the

sum

of all claims

on

foreign assets held

by domestic

residents. In our simple model, the only productive asset located within the country isthestock of capital,

and

the net foreign asset position is simplythestockof

5

TheintertemporalapproachwasdevelopedbySachs [1981, 1982],Obstfeld [1982],Dornbusch [1983],

Svensson andRazin [1983],Persson and Svensson[1985],andMatsuyama[1987],amongothers.

non-contingent

bonds

owned

by

the country.

By

thecomposition ofthe country

portfolio,

we

refer tothe

share

ofthe net foreign asset position in it.

To

interpretthe

evolution ofthe current

account

it is usefulto breakit

down

into

two

components

or

pieces:

changes

in the sizeofthecountry portfolio,

which

we

callportfolio growth;

and

changes

inthecomposition ofthecountry portfolio, orportfoliorebalancing.6

We

studya small country populated by a

continuum

ofidentical

consumers.

There

is a single

good

that

can be

used

for

consumption

and

investment.

Consumers

have access

to

two

investmentopportunities:foreign loans

and

domestic

capital.

The

instantaneous interest rate

on

foreign loansis p.

To

produce

one

unitofcapital

one

unit ofthe single

good

is required. Sincecapital is reversible

and

does

not depreciate, its

price isequalto

one

and

its return isequal tothe flow ofproduction

minus

operating

costs.

The

flowof production

generated

by

one

unit of capital is7idt+0-dco

;

where

n and

a

are non-negative constants;

and

co is a

Wiener

processes, i.e. its

changes

are

normally distributedwith E[dco]=0

and

E[dco2]=dt.

That

is, theflow ofproduction is

normally distributedwith instantaneous

mean

n and

instantaneous variance o.

The

operating costs, a, are

assumed

to

be

proportional tothe

aggregate

investment rate:

1 rik

(1)

«=*-—

(teo)

k dt

where

k isthe

aggregate

stock ofcapital at the beginningofthe period. Sincecapital

does

notdepreciate, this is also thestockofcapitalthat

was

used

in production in the previousperiod.

Note

that

we

aretreating the relationship

between

operating costs

and

investment asa congestion effectornegative externality.

One

set of

assumptions

that

justifiesthis relationship

would be

thatinvestment requiresa public input thatcosts A.

perunit of investment

and

the

government

financesthis input by raising a tax

a

on

capital.

There

might

be

alternative

and more

compelling sets of

assumptions

that

6

Implicitinthis decompositionistheassumptionthatassetpricerevaluations are small. Thismightbe a

deliverthis relationship.

The

reason

we

adopt

ithere is simply

because

it provides a

tractable

and

effective

way

tocapture the notion of

adjustment

coststo investment.7

The

representative

consumer

values

consumption

sequences

withthese preferences:

(2)

EJlnce"

5'

dt (6>0)

Given

our

assumptions about

theflow ofproduction

and

theoperating costs,

the returnto capitalis

(Tt-a)dt

+

adco; and

the representative

consumer's budget

constraint

can be

written

as

follows:

(3)

da = [((n-a)-(1-x)

+

px)a-c]-dt

+

(1-x)-a-a-dco

where

c, a

and

x

denote consumption,

wealth

and

the

share

offoreign loans in the

portfolioofthe representative

consumer.

The

budget

constraintillustratesthe standard

risk-returntrade-offunderlying investmentdecisions.

Each

extra unit ofwealth invested

in

domestic

capital ratherthan foreign loans increases the

expected

return towealth

by

n-a, atthe cost ofraising the varianceofthis return by

a

2. Finally,

we

assume

that it is

not possibleto short-sell thecapital stock, i.e. x<1

.

Solving the representative

consumer

problem,

we

findthe optimal

consumption

and

portfoliodecisions:8

(4) c

=

5•a

7

Theq-theory postulates thatinvestmentraisesthe price ofinvestmentgoodsrelative toconsumption

goods, leavingthe productivityof capitalconstant.

We

instead postulate thatinvestment lowersthe

productivityof capital, leaving therelativepriceofinvestmentand consumption goodsconstant. Itis likely

thatinrealeconomies, bothsortsofadjustmentcoststoinvestment areimportant. SeeLucas[1967]foran

earlymodelthatconsiders both typesofadjustmentcosts;andCaballero [1997]andDixitandPyndick

[1994]fortwoexcellent expositions ofexistingmodelsofadjustment costsofinvestment.

(5) x

= 1-max<{

_r~^'°

When

decidingtheirconsumption,

consumers behave

as

inthe

permanent-income

theory of

Friedman.

Equation (4)

shows

that

consumption

is a fixedfraction of

wealth

and

is

independent

ofthe

expected

return

and

volatilityofavailable assets.

When

deciding theirportfolio,

consumers behave

as

inthe

mean-variance

theory of

Markowitz

and

Tobin. Equation (5)

shows

thatthe

shares

of

each

asset in the portfolio

depend

only

on

the

mean

and

variance ofthedifferent assets

and

not

on

thelevel of wealth.

The

kink in the

demand

forforeign assets isthe result ofthe short-sale constraint

on domestic

capital, i.e. x<1.

In equilibrium,the

demand

and

supply ofcapital

must be

equal

and

this implies

that:

(6)

(1-x)a

=

k

+

dk

The

left

hand

side of Equation (6) isthe

demand

for capital. Since

we

have

assumed

that only

domestic

consumers

hold

domestic

capital, this

demand

isequal to theshare of theirwealth thatthese

consumers

want

to hold in

domestic

capital, times

wealth.

The

right

hand

side of Equation (6) isthe supplyof capital,

and

consists ofthe

capitalstock atthe beginning ofthe period plus the investment

made

during the period.

This

completes

the description ofthe model.

There

are

two

state variables: k

and

a;

and

one

shock: dco.

The

"new

rule"

model

ofour previous

paper

obtains as the

limiting

case

in

which

X—

>0. Inthis case,there are

no

adjustment costs to investment

and

the only state variable isthe levelofwealth.

Assume

that7i>p+A.(p-5). This

parameter

restriction

ensures

thatthe

economy

is productive

enough

so

thatthe

selling constraint

on

capital is

never

binding.

Then,

it is straightforwardto

use

Equations (1)-(6) toobtainthe

dynamics

forthecapital stock

and

wealth:9

(7) (8)

dk

k

da

a

=x-

1 -2 k^

ji-p-a

•

—

ay

•dt

a

T

2 a

+

p-8

dt

+

—

-adco

a

Equations

(7)-(8)provide the lawof motion ofthe

system from

any

given initial

condition

and sequence

of shocks.

Our

next goal is to

use

this

dynamical

system

to

study

how

thecurrent

account

responds

to

income

shocks.

II.

Portfolio

Growth

and

Portfolio

Rebalancing

To

illustratethe model's implications,

we

analyzethe behaviorofsavings,

investment

and

the current

account

after atransitory

income

shock.

To

do

this, it is

useful firstto establish

some

notation. Let

S

and

CA

be

savings

and

the current

1

da

1 dfx-a)

account

both

as

a

share

ofwealth, i.e.

S=

and

CA

=

. Itfollowsthat,

a

dt a dt

along

any

particular

sample

paththat

we

consider, the current

account

can be

written

as follows:

dx

(9)

CA

=

x-S

+

—

dt

Equation (9)

shows

that it is possibleto interpretthecurrent

account

as

the

sum

of

two

terms.

The

first

one

measures

the

change

inthe stockofforeign assetsthat

ToderiveEquations(7)-(8),rememberthatinthelimitofcontinuous timedkdt=0.

would

keep

constantthecompositionofthe countryportfolio

and

this is

what

we

refer

to

as

portfoliogrowth.

The

second term

measures

the

change

in thecompositionofthe

country portfolio

and

this is

what

we

refer to

as

portfoliorebalancing.

To

develop

intuitions

about

the interplay

between

these

two

components

ofthe

current account,

we

present next a series of

examples.

In all of

them,

we

assume

the

following

sample

pathforthe production shock:

(10)

^

v

' _dt

tel-co,^)

e-o

-1 tefTpT;,) (e>0)

te[T

2,oo)

That

is, the countryexperiences a

sequence

of

unexpected

production

shocks

equalto

edt

times thecapital stockfor a finite period

and

zero afterwards.

We

refer to

the period [TLT2) asthe

shock

period

and

to (-^.T^

and

[T2 ,°°)

as

the pre-

and

post-shock

periods, respectively.

Figure

4

shows

the behavior oftheforeign asset position alongthis

sample

path.

Regardless

oftheinitial condition, duringthepre-shock period the

share

of

foreign assets

converges towards

x*

=

1

+

1 \ 1 f i A

2-X

\{2-X

X

1 (n

-

p)

-

p

+

5

+

v

7

K .

The

a

simulation behind Figure 4

assumes

that thisvalue

has

been

reached by

t=0. During

the

shock

periodthe shareofforeign assets increases steadily, albeit at adeclining

rate.

The

magnitude

ofthis increase

depends

on

X. High values of

X

implythat the

effects of increased investment

on

operating costs are large

and

provide a strong

inducement

for investorsto rebalance theirportfolios

towards

foreign assets. During

the post-shock period, investment

and

operating costs decline.

As

a result, theshare of

foreign assets slowly returns to its pre-shock level.

We

nextstudy the implications of

this behavior oftheshareofforeign assetsforthe current account.

Consider

firstthe

case

in

which

adjustment coststoinvestment are negligible,

i.e.

A—

>0. Figure

4

shows

that inthis

case

the

share

offoreign assets is constant

dx

throughout.

As

a result, there is

no

portfolio rebalancing, i.e.

—

=

;

and

the current

dt

account

is equalto portfoliogrowth; i.e.

CA

=

x•

S

.This isthe

"new

rule"

model

that

we

analyzed

in our previous paper,

and

its implicationsfora creditor

and

adebtor country are depicted in Figure 5.

The

top panel

shows

a creditor country, i.e. x*>0, whilethe

bottom

panel

shows

adebtorcountry, i.e. x*<1. Both countries raisetheirsavings during the

shock

period as a result ofthe standard

"consumption-smoothing"

motive.

Both

countries also invest

these

marginal savingsin

domestic

capital

and

foreign loans

inthe

same

proportions

as

their

average

portfolio. Since theforeignasset

share

is

small in absolutevalue,

we

findthat in both countriesthe increase in investmentis of

the

same

orderof

magnitude as

the increase in saving. But itis not exactlythe

same,

and

this leads to different current

account

responses

in debtor

and

creditorcountries.

Inthe creditorcountry investment increases

somewhat

less than savings

and

the

current

account

registers asurplus. In thedebtor country investment increases

somewhat

more

than savings

and

the current

account

registers a deficit.This is the

main

resultofour previous paper.

Consider

nextthe

case

in

which adjustment

costs toinvestment are

no

longer

negligible, i.e. X>0. Figure 6

shows

the

case

ofa countrythat is neithera debtor nor a

creditor.

By

choosing

the

case

x*=0,

we

know

thatinthe

absence

of

adjustment

costs,

the current

account

would

be

zero before, during

and

aftertheshock. Figure 6

shows

the behavior ofsavings, investment

and

the current

account

inthis case.

The

country

raises its savings during the

shock

periodforthe

same

"consumption-smoothing"

motive

as

before. But

adjustment

costs

now

discourage large

swings

in investment,

and

this affects

how

these savings aredistributed

between domestic

capital

and

foreign loans.

During the

shock

period,thecountry

uses

most

of its increase in savingsto

purchase

foreign loans, while investment increases only gradually.

Consumers

rebalancetheirportfolios

towards

foreign assets

because

the increase in investment

raisesoperating costs

and

this lowersthe

expected

return to

domestic

capital.

The

portfolio-rebalancing

component

ofthecurrent

account

is positiveand,

as

a result, the

new

rule under-predicts the current

account

surplus in the shortrun. In the post-shock

periodinvestmentfalls slowly, but

remains

higherthan

normal

forawhile. Since

productivity

has

returnedto its pre-shock level, savings return to

normal

and

the

higher-than-normalinvestment is

now

financed

by

a sale of foreign loans.

Consumers

rebalancetheir portfolios

back towards

theiroriginal composition

because

the decline

ininvestment lowers operating costs

and

this raisesthe

expected

returnto

domestic

capital.

The

portfolio-rebalancing

component

ofthe current

account

istherefore

negative and,

as

a result, the

new

ruleover-predictsthecurrent

account

surplus inthe

medium

run.

As

time passes, the country portfolio returnsto itsoriginal composition

and

the

new

rule appliesagain inthe long run.

This

example

clearly

shows

the roleofforeign loans

as

a bufferstock to

smooth

thefluctuations in investment.

Without access

toforeign loans, countries

would be

forced notonlyto invest all oftheirsavingsat

home

butalso to

do

so

contemporaneously.

Access

toforeign loans permits countries ato

spread

their

domestic

investment over time and, in this

way,

avoid paying high

adjustment

costs.

To

do

this, countriestemporarily placetheir savings inforeign loans

and

slowly convert

them

into

domestic

investment.

It is possibletodesign

more

complicated

examples

in

which

the current

account

exhibits richer

dynamics.

For instance, Figure 7

shows

the

case

of positive adjustment

costs in a creditor

and

a debtorcountry.

One

can

interpretthese

examples

as a combinationof portfoliogrowth

and

portfolio rebalancing along the lines ofthe

explanationsof Figures

5

and

6.

The

theory

developed

here thereforeequips uswith a

clear picture of the factors that

determine

how

the current

account

reacts to increases

insavings.

The

next stepisto

go

back

to actualdata

and

attemptto interpret it

from

the

vantage

point ofthetheory.

III.

The

Process

of

Current

Account

Adjustment

In the introduction,

we

argued

that in the long run

most

ofthe variation in

current

accounts

in

OECD

countries is

due

to portfolio

growth

effects, while in the short

run, current

account

fluctuations primarilyreflect

changes

in the composition ofcountry

portfolios orportfolio rebalancing.

We

made

this point

based

on

theobservationthat

thesimple interaction ofacountry's foreign asset

share

with its saving,

averaged over

thepastthirtyyears,

proved

to

be

a very

good

predictor ofthe country's

average

current account.

However,

the

same

interaction using annual data

proved

to

be

a very

poor

predictor ofyear-to-yearfluctuations in current accounts. This

was

shown

inthe

two

panelsof Figure2.10

The

theory

presented

above

has

the potential to explain these observations. In

the

presence

of

adjustment

costs to investment, thetheorypredicts that in the

short-run countries react totransitory

income

shocks by

raising savings

and

rebalancing their

portfolios

towards

foreign assets. Ifthese costs are sufficientlystrong, thetheory

can

therefore explain

why

the shortrun variation inthe current

account

is

dominated by

portfolio rebalancing,

and

notportfolio growth.

The

theoryalso predicts that in the

aftermath ofthe

shock

countries graduallyrebalance theirportfolios

back

to their

original composition. Therefore the theory

can

also explain

why

thelong-run variation

in the current

account

is

dominated

by portfoliogrowth,

and

not portfolio rebalancing.

The

theoryalso

has

very clearpredictionsforthe patterns of portfolio

rebalancingthat

we

should

observe

in the data.

The

new

rule orportfolio

growth

10

Ofcourse,onecould arguethatthisdiscrepancyinthe"between"and"within"resultsissimplydueto

muchgreatermeasurementerrorinthewithin-country variationincurrent accountsandportfoliogrowth

than inthebetween-countryvariation. Whilemeasurementerroris certainlypresent,

we

clearlydonot

thinkitisthewholestory.

One

waytoseethisistonotice that(i) measurementerrorinthe

RHS

variable

inour regressionwillbias theslopecoefficientdownwardby afactorequaltothe signal-to-noiseratio,and

(ii)measurementerrorinboththe

LHS

and

RHS

variableswillbiasthe

R

2coefficientby afactorequalto

theproductofthe signal-to-noiseratios inthetwovariables. Since

we

observeaslopecoefficientof

one-halfand an

R

2coefficientthatfallsfrom0.6 inthe"between" regressionto 0.03inthewithin regression, thisimpliesasignalto noiseratioofonly 0.5 inthe

RHS

variableand0.1 inthe

LHS

variable. Whilethere areclearlyvariousmeasurementissuesinourdata,

we

finditimplausiblethatthe dataisasnoisyasthis calculationwouldsuggest.

component

ofthecurrent

account

under-predictsthe actual current

account

during the

shock

period as countries rebalancetheir portfolios

towards

foreignassets,whilethe

new

rule over-predictsthe current

account

afterthe

shock

as countries rebalance their

portfolios

back towards

itsoriginal composition. In other words, a

contemporaneous

increasein savings should

be

associatedwith a positive portfolio-rebalancing

component

ofthe current account,while past increases in savingsshould

be

associated with negative values inthe

same

component.

Moreover, forthe

new

ruleto

apply inthe long run,these positive

and

negative

components

should

be

roughlyofthe

same

magnitude. In thissection,

we

show

that thedata is consistent with these

predictions.

We

begin

by

decomposing

observed

current

accounts

into portfoliogrowth

and

portfolio rebalancing

components.

As

in the theory, let Xct

denote

the

share

offoreign

assets in theportfolio ofcountryc atthe beginning of period t,

and

let

S

ct

and

CAc

t

denote

gross national saving

and

the current

account balance

as afraction of

GDP

during periodt.

We

measure

the portfolio-growth

component

ofthe current

account as

PG

ct

=

xct

•

S

ct, i.e. the net

purchases

offoreign assetsthat

would be observed

during

period t if a country

were

to distribute itssaving

between

domestic

and

foreign assets

in the

same

proportion

as

its existing portfolio atthe beginning ofthe period.

We

measure

the portfolio-rebalancing

component

ofthe current

account

residually

as

the

difference

between

the actual current

account

and

the portfolio-growth

component,

i.e.

PR

_ct

=CA

ct

-x

ct -Sct .

To

implement

this decomposition,

we

requiredata

on

current accounts, saving

and

theshareofforeign assets in country portfolios.

We

obtain annual data

on

current

accounts

in current

US

dollars

from

the International

Monetary

Fund's International

FinancialStatistics.

We

measure

gross national saving as the

sum

ofthe current

account

and

gross

domestic

investment in current

US

dollars,

and express

both

as a

fraction of

GDP

incurrent

US

dollars, obtaining investment

and

GDP

from

the

World

Bank's

World

Development

Indicators.

We

obtain data

on

theshare of foreign assets

in wealth

from

Kraay, Loayza,

Serven and Ventura

(2000).

We

restrictattention tothe

setof21 industrial countriesfor

which

at least

20

annual observations

on

this variable

are available

over

the period

1966-1997 covered by

this dataset.

With

data

on

saving

and

the portfolio-rebalancing

component

ofthe current

account

in hand,

we

estimate a series of

dynamic

linearregressions ofthe form:

p q /

(11)

PR

_d

=a

c

+£<|>

w

-PR

Cit_v

+£Ycv-S

Ci,_ v

+P

c

Z

ct

+u

ct

v=1 v=0

where PRc

t

and

S

ctare the portfolio-rebalancing

component

ofthe current

account

and

saving

as

described above; Zct isa vectorof control variables,

and

Ud

is awell-behaved

errorterm.

We

then

use

the point estimates ofthe coefficientsto retrievethe implied

impulse

response

function ofportfolio rebalancing in periodt+kto

an

increase insaving

dPR

c•,l

in period t, i.e. :

—

.

These

impulse

responses

provide

us

with a picture of

how

3S

ct

countries

change

thecomposition of their portfolios following

an

increase in saving.

The

results offour

such

regressions are

summarized

in

Table

1.

The

top panelof

Table

1 reportsthe estimated coefficients,while the

bottom

panel reportsthe corresponding impulse

response

functions using the21 -country

sample

of

annual

observations.

The

estimated impulse

response

functions are also plotted in thefour

panelsof Figure8.

We

begin by

assuming

that all ofthe slope coefficients are the

same

across

countries. In our simplestspecification,

we

alsoset

p=0

and

introduce

5

lags of

saving.11

The

results of thisspecification are reported in thefirst

column

of

Table

1. In this case, theimpulse

response

function simply consists oftheestimated coefficients

on

current

and

lagged saving.

We

find a strong positive

contemporaneous

correlation

between

saving

and

thecurrent account.

The

pointestimate of0.6

can be

interpreted

as the fraction of

an

increase in saving that,

on

impact,

would be

invested in foreign

assets bya countrywith zero initialforeign assets. Thisfraction

would be

slightly

higher (lower)in creditor(debtor) countries

because

ofthe portfolio-growth

component.

Sincethelatter

measures

thecurrent

account

balancethat

would keep

the composition

oftheirportfoliosconstantfollowing

an

increase in saving, it is by construction positive

in creditorcountries

and

negative in debtorones.

The

subsequent

lags ofsaving all enter with negativecoefficientsthat are decreasing in absolute value and,with theexception ofthefirst lag, are notsignificantly

different

from

zero.

These

coefficients

can be

interpreted as thefraction ofthe initial

increase in savingthat is reallocated

back towards domestic

assets in

each

ofthe

subsequent

fiveyears. Interestingly, the

sum

ofthe coefficients

on

lagged saving is

-0.09

which

is insignificantly different

from

zero. This

suggests

thatthe initial shift

toward

foreign assets is largely

undone

in the nextfive years, with the bulk of the

readjustmentoccurring in thefirstyearfollowing the increase in saving. This pattern is

consistent with the predictions ofthetheory.

The

restof

Table

1 reports avariety of robustness

checks on

this basic result.

We

begin by introducing lagged values ofthe portfolio-rebalancing

component

ofthe current account,

and

findthat thefirst

and second

lags are stronglysignificant,while

third (and higher) lagsare not.12

Although

this slightly altersthe point estimatesofthe

coefficients

on

current

and

lagged saving,

we

find thatthe

shape

of impulse

response

function isverysimilar tothat reported in thefirst regression.

The

main

differenceis

thattheinitial shifttoward foreign assets is slightlysmallerthan before, at

50

percent of

the increase in saving.

11

In unreportedresults,

we

findthatfifthandhigherlagsofsavingareinsignificantlydifferentfrom zeroin

mostspecifications,andaddinghigher lags haslittleeffectonthe pointestimatesofthe coefficientsonthe

firstfivelags.

12

We

areassumingherethatthetimedimensionofour panelissufficientlylargethat

we

canobtain consistentestimatesofthe coefficientson thelaggeddependentvariableinthepresenceof fixed effects relyingon large-Tasymptotics.

Remember

alsothatsavingisconstructedas investmentplus the current

account,andthelatteris highlycorrelated withthedependentvariableinEquation(11). Tothe extentthat

the portfolio-rebalancing componentofthe currentaccountismeasuredwitherrorsthatare persistentover

time,thiscould introduceacorrelation betweenthe residualsandcurrentandlaggedsaving. Inthe specifications withlagsofthedependentvariable,wetest foranddonotrejectthenullofnoserial

dependenceinthe residuals,andsowe canruleoutthispotentialsourceofbiasinourestimated impulse responses.

Inthe next

column

we

augment

the specification ofthe previous

column

with several additional control variables.

To

the extent thatthere are other

shocks

to returns that

change

the desired

composition

of country portfolios,

and

totheextentthat these are correlated with saving, thiswill bias our results in directions

which

depend

on

the signsofthese correlations.

For

example,

ifthere areglobal

shocks which

raise saving

and

investmentin all countries (such

as

changes

inworld interest rates),

we

will

be

underestimating the size ofthe initial shifttowardforeignassets

when

saving

increases. Similarly, ifin countries

and

yearsin

which

saving is high, factorsthat

increase the desired rateof

investment

(such

as

population or productivitygrowth),

we

may

again

be

underestimating the shift

toward

foreign assets.

To

controlforthese

factors,

we

introduce

year

dummies

tocapture global shocks, population growth,

and

Solow

residuals as a proxyforproductivitygrowth.13

The

third

column

of

Table

1

reports the results of this

augmented

regression. Population

growth

and Solow

residualsenter significantlywith the

expected

negative signs,

and

we

find a largershift

toward

foreign assetsthan before, with

75

percent oftheinitial increasein saving

allocated

toward

foreign assets.

However,

the

subsequent

pattern ofadjustment isthe

same

as

before, with the initial shift

toward

foreign assets being reversed in next

few

years.

Inthefinal

column,

we

relaxthe

assumption

thatthe slope coefficients in

Equation (11) arethe

same

across

countries,

and

instead estimate this equation separatelyfor

each

country.

Because

ofthefairly shorttime-series available for

each

country,

we

adopt

a

more

parsimonious

lag structure, introducing only

two

lags ofthe

dependent

variable

and

of saving,

as

well as population

growth

and

Solow

residuals.

We

reportthe

average

and

standard deviation across countries oftheestimated

coefficients in the last

column

of

Table

1. 14

Notsurprisingly,

we

find thatthe

country-13

We

constructSolowresidualsasthegrowthin

GDP

atconstantprices lessgrowthinemploymenttimes

the period averageshareoflabourin

GDP,

drawingthelattertwovariablesfromthe

OECD

LabourForce

StatisticsandNationalAccounts.

Inthepresenceofparameter heterogeneityacrosscountries, thepooled estimates reportedinthe

previoustwocolumnswill not deliverconsistentestimatesofaverage (acrosscountries)ofthese

parameterswhenthereisalaggeddependentvariable(PesaranandSmith (1995)). However,the

averageacrosscountriesoftheestimatedcoefficientswillprovideaconsistentestimateoftheaverage

by-country

parameters

are

much

less preciselyestimated,

and

the dispersion across

countries in the pointestimates is large. Nevertheless,

we

find results that are

qualitatively

and

quantitativelyquite similar tothoseinthe previous columns.

On

average, thefraction of

an

increasein savingthat is allocatedtoforeign assets is0.7,

and

this initial shift

toward

foreign assets isquickly

undone

in

subsequent

periods.

One

drawback

oftheannual data

on which

we

have

relied

so

faristhat it is not

informative

about

theintra-year

dynamics

of saving

and

the current account. For 12 of

the countries in our

sample,

we

were

abletoobtain quarterlyobservations

on

the

currentaccount, investment

and

GDP

beginning in

1980

orearlier

from

the

International Financial Statistics

and

the

OECD

Quarterly National Accounts. Forthese

countries,

we

linearly interpolatethe annual data

on

theforeignasset share

and use

it

to construct quarterly portfoliogrowth

and

rebalancing

components.

We

then

re-estimate Equation (11) using quarterly data, introducing eight lags of the portfolio

rebalancing

component

ofthe current account,

and

eight lags of saving.

We

do

not

have

quarterlydata

on

population or

employment

growth required to introducethe

same

control variables

as

inthe previous regressionswith annual data (Regressions 3

and

4 in

Table

1).

We

therefore include only a setof period

dummies

and

real

GDP

growth

as

controls.

As

before,

we

summarize

the results ofthese country-by-country regressions

by

computing

the

mean

and

standarddeviation acrosscountries ofthe estimated impulse responses.

As shown

in the top panel of Figure9,

we

find that

on

impact, just

over

60

percentof

an

increase in saving thatlasts

one

quarteris invested abroad. Beginning immediately inthe next quarter, this initialshift

toward

foreign assets begins

to

be

reversed

as

countries run current

account

deficits. If

we

considera

shock

to

saving that lastsfour quarters,the pattern that

emerges

is verysimilar to that

we saw

in the annual data. This is

shown

inthe

bottom

panel of Figure 9. During the

shock

period, countries run positive but declining current

account

surpluses

as

they

use

response.

We

findresults thatarequantitativelyquitesimilaracross allspecificationsdespitethis potential

sourceofbiasinthe estimates whichimpose parameter homogeneityacrosscountries.

foreign assets

as

a bufferstockto

smooth

investment. In

subsequent

years, countries run current

account

deficits inorderto restoretheir original pre-shock portfolios.

To

sum

up,while portfolio growth explains

much

ofthe long-run variation in

current accounts, portfolio rebalancing

dominates

in the short run. In all ofour

specifications,

we

find thatthe portfolio-rebalancing

component

ofthe current

account

follows a

remarkably

clearpattern.

On

impact,

up

tothree-quartersof a

shock

to saving

is invested

abroad as

countries

use

foreign assets

as

a bufferstockto

smooth

investment intheface of adjustmentcosts. In

subsequent

periods, the initial increase

in saving

produces

current

account

deficits

as

countries shifttheir portfolios

back

to

their original composition.

IV.

The

Current

Account

and

Investment

Over

the past

20

years considerable empirical effort

has

been

devoted

to

documenting

the correlations

between

investment

and

thecurrent account.

Two

stylized facts

have emerged.

First, cross-countrycorrelations

between

investment

and

the current

account

are

weak

(Penati

and Dooley

[1984],

Tesar

[1991]).

Second,

within

countriesthetime-series correlation

between

investment

and

thecurrent

account

is

consistently negative (Glick

and

Rogoff[1995]).

We

document

thatthese

two

stylized

facts hold in our

sample

ofcountries in Figure 10. In the toppanel

we

plot long-run

averages

ofthe current

account

as afraction of

GDP

(on thevertical axis)against long-run investment rates (onthe horizontal axis) forthe21 industrial countriesin our sample.

Across

countries,

we

find a very

weak

negative correlation

between

thetwo,

with a coefficientof-0.036. Inthe

bottom

panel,

we

plotthe

same

two

variables

expressed as

deviations

from

country

means,

pooling all available annual

observations. Withincountries, the correlation

between

investment

and

thecurrent

account

is strongly negative, witha coefficient of-0.329.15

Thisdifference

between

the correlations

between

the current

account

and

investment inthe long

and

shortrun are consistent with the

view

ofthe current

account

proposed

in this paper.

To

see

this, it is useful towritethecurrent

account

and

investment

as

follows:

(12)

CA

ct

=x

ct -Sct

+PR

ct

(13) l

ct

=(1-x

el

)-S

et

-PR

el

These

equations

decompose

the current

account

and

investment into their

portfolio-growth

and

portfolio-rebalancing

components.

The

key

observation toexplain

thepattern ofcorrelations

between

the current

account

and

investment is thatthe

long-run relationship

between

thesevariables is

dominated by

their portfolio-growth

components,

while the short-run relationship is

dominated by

the portfolio-rebalancing

components.

To

make

this statement precise,

we

decompose

thecoefficientof a

regressionofthe current

account

on

investment into thecontributions ofportfolio

growth

and

portfolio rebalancing. Let p

be

this regression coefficient

and

define PG

_

Cov(xS,(1-x)l)

_gnd

PR

_

Cov(CA,l)

Cov(x-S,(1-x)l)

_Sjnce

_b=3

pg

+b

pr

Var(l) P Var(l) Var(l) ' P P P

'

we

interpret (3

PG

and

(3

PR

as

the contributions of portfoliogrowth

and

portfolio

rebalancingtothe relationship

between

the current

account

and

investment.

When

we

perform this

decomposition

on

the

between

estimatorin the top panel

of Figure 10,

we

find that (3

PG

=-0.041

and

(3

PR

=0.005. Consistentwith thetheory,

portfolio rebalancing plays

no

role inthe long run

and

the relationship

between

the

current

account

and

investment reflects only portfolio growth. Moreover, thetheory

predicts thatthecorrelation

between

thecurrent

account

and

investment should

be

Thisisalmostexactlythesameastheaverageofcountry-by-countryestimates reportedinGlickand

Rogoff[1995],

negative in debtor countries

(where

x>1)

and

positive in creditorcountries

(where

x<1).

The

intuition is simple

and

follows immediately

from

the

new

rule: in debtorcountries

increasesin saving

generate

even

greater increases in investment, leading to current

account

deficits, while in creditorcountriesthe increase in investment is less than that

ofsaving, leading tocurrent

account

surpluses. Since our

sample

ofcountries consists

ofa mixture of

15

debtor

and

6creditorcountries,

we

should expectto find a negative

butnot especially strong correlation

between

investment

and

the current

account

in a

cross-section thatpools all countries together. This is exactly

what

we

found

inthe top

panel ofFigure 10. But

when

we

divide

our

sample

into debtors

and

creditors

and

compute

thecorrelations separately inthe

two

groups,

we

shouldfind a negative

correlation

among

debtors

and

a positive correlation

among

creditors. Figure 1 1

shows

thatthis isthe case.

Of

course,

we

have

onlya very small

sample

ofcreditors

and

debtors,

and so these

differences in slopes should

be

takenwith a grain of salt.

Nevertheless,

we

notethatthey are consistentwiththe theory.

When

we

perform the

same

decomposition on

thewithinestimator inthe

bottom

panel of Figure 10,

we

findthat (3

PG

=-0.014

and

_pPR=-0.315. Consistentwith the theory, portfolio rebalancing is importantin theshort run

and

this introduces asource

of negative correlation

between

thecurrent

account and

investment. In the

presence

of

adjustment

costs, a

shock

to

income

in a given period triggers

an adjustment process

thatlastsfor

many

periods. In particular,

a

positive

shock

to

income

raises saving

contemporaneously and

isfollowed

by

several periods of portfolio rebalancing,

as

countries

have

higherthan

normal

investment financed by current

account

deficits in

orderto restoretheir pre-shock portfolios.

The

oppositeoccurs

when

thereis a

negative shock.

Thus

positive

shocks

triggera "ripple"effectof

subsequent

higher investment

and

lower current accounts,

and

vice versafor negativeshocks. This

"ripple" effect is a

source

of negative correlation

between

investment

and

the current

account

within countries.

V.

Concluding

Remarks

By

reconciling long

and

short run data, this

paper

completes

the

view

ofthe

cyclical behaviorofsavings, investment

and

the current

account

in industrialcountries

that

we

first

proposed

in

Kraay

and Ventura

[2000].

Faced

with

income

shocks,

countries

smooth

consumption by

raising savings

when

income

is high

and

vice versa.

In the short run, countries invest

most

oftheirsavings inforeign assets, only to

rebalance their portfolios

back

to their original composition in the nextfourto five years. Inthe long run, countryportfoliosare remarkably stable, the

new

ruleapplies

and

fluctuations in savings lead'tofluctuationsin the current

account

thatare equal to

savings times the

share

offoreign assets in thecountry portfolio.

By

usingforeign

assets

as

abuffer stock, countries

smooth

investment inorderto

save on

adjustment

costs.

An

interesting implication ofthis

view

of international capital flows isthatthe

stock offoreign assets

and

the current

account

are

more

volatilethan

consumption,

investment

and domestic

capital stock. Butthis

does

not

mean

that internationalcapital

flowsare afactor thatcontributesto

making

macroeconomic

aggregates

more

volatile

orunstable.

To

the contrary, the

view

presented here

suggests

that theabilityto

purchase

and

sell foreign assets allowscountries not onlyto

smooth

their

consumption,

but alsotheir investment. Foreign assets

and

the current

account

absorb

partofthevolatilityofthese other

macroeconomic

aggregates.

References

Caballero, Ricardo,

"Aggregate

Investment," In

John

B.Taylor

and

Michael

Woodford,

Eds.,

Handbook

of

Macroeconomics,

Vol. 1B.

Amsterdam,

The

Netherlands:

Elsevier, 1999.

Deaton,

Angus.

Understanding Consumption.

Oxford:

Clarendon

Press, 1992.

Dixit,

Avinash

K.,

and Robert

S. Pindyck,

Investment

Under

Uncertainty. Princeton,

NJ: Princeton University Press, 1994.

Dombusch,

Rudiger, "Real Interest Rates,

Home

Goods,

and

Optimal

External

Borrowing,"JournalofPolitical

Economy,

XCI

(1983), 141-153.

Feldstein, Martin,

and Charles

Horioka,

"Domestic

Savings

and

International Capital Flows,"

Economic

Journal,

XC

(1980), 314-329.

Glick,

Reuven

and Kenneth

Rogoff, "Global versus Country-Specific Productivity

Shocks

and

the Current Account," Journalof

Monetary

Economics," 35

(1995),

159-192.

Kraay, Aart,

and

Jaume

Ventura, "Current

Accounts

in

Debtor

and

Creditor Countries,"

QuarterlyJournal of

Economics,

XCV

(2000), 1137-1166.

Kraay, Aart,

Norman

Loayza, Luis Serven,

and

Jaume

Ventura, "Country Portfolios,"

NBER

Working

Papers, No. 7795, July2000.

Lewis, Karen, "Trying to Explain

Home

Bias in Equities

and Consumption,"

Journal of

Economic

Literature,

XXXVII

(1999),

571-608.

Lucas, Robert E. "Adjustment

Costs

and

the

Theory

ofSupply," JournalofPolitical

Economy,

(1967)75, 321-334.

Matsuyama,

Kiminori, "Current

Account

Dynamics

in a Finite Horizon Model," Journal

ofInternational

Economics,

XXIII (1987), 299-313.

Meilon, Robert,

"Optimum Consumption

and

Portfolio

Rules

in a

Continuous-Time

Model," Journal of

Economic

Theory, III (1971), 373-413.

Obstfeld, Maurice,

"Aggregate

Spending

and

the

Terms

ofTrade: Is

There

a

Laursen-Metzler Effect?" QuarterlyJournal of

Economics,

XC

(1982), 251-270.

Obstfeld Maurice,

and Kenneth

Rogoff,

"The

Intertemporal

Approach

tothe Current Account." In

Gene Grossman

and Kenneth

Rogoff, eds.,

Handbook

of

International

Economics.

Amsterdam,

The

Netherlands: Elsevier, 1995.

Penati, A.

and

M. Dooley, 1984. Current

account

imbalances

and

capital formation in

industrial countries, 1949-81,

IMF

Staff

Papers

31, 1-24.

Persson, Thorsten,

and

Lars

Svensson,

"Current

Account

Dynamics

and

the

Terms

of

Trade: Harberger-Laursen-Metzler

Two

Generations

Later," Journal of Political

Economy,

XCIII (1985), 43-65.

Pesaran,

Hashem

and

Ron

Smith, (1995). "Estimating

Long-run

Relationships

from

Dynamic

Heterogenous

Panels". Journal of Econometrics.

68

(1995) 79-113. Sachs,Jeffrey,

"The

Current

Account

and

Macroeconomic

Adjustment

in the 1970s,"

Brookings

Papers

on

Economic

Activity, 1 (1981), 201-268.

,

"The

Current

Account

in the

Macroeconomic

Adjustment

Process," .

Scandinavian

Journal of

Economics,

LXXXIV

(1982), 147-159.

Svensson,

Lars,

and Assaf

Razin,

"The

Terms

of

Trade and

the CurrentAccount:

The

Harberger-Laursen-Metzler Effect,"Journal ofPolitical

Economy,

XCI

(1983),

97-125.

Tesar, Linda, "Savings, Investment

and

International Capital Flows," Journal of

International

Economics,

XXXI

(1991), 55-78.

Ventura,

Jaume,

"A Portfolio

View

ofthe U.S. Current

Account

Deficits". Brookings

Papers on

Economic

Activity. (2001), July.

Tabid:

Portfolio

Rebalancing

and Saving

(An nual

Data

for 21

Countries)

Reqression 1 Rearession 2 Reqression 3 Reqression 4

Mean

SDof

Coef S.E Coef S.E Coef S.E Coef Coefs

CoefficientEstimates sy 0.598 0.096 0.504 0.080 0.746 0.079 0.691 0.286 sy(-1) -0.281 0.133 -0.611 0.102 -0.824 0.104 -0.767 0.383 sy(-2) -0.120 0.106 0.112 0.077 0.109 0.070 0.123 0.167 sy(-3) -0.120 0.095 -0.043 0.073 0.040 0.067 sy(-4) -0.102 0.103 -0.031 0.065 -0.063 0.061 sy(-5) -0.060 0.078 0.020 0.058 0.019 0.057 pr(-1) 0.754 0.056 0.845 0.057 0.837 0.216 pr(-2) -0.114 0.069 -0.081 0.066 -0.152 0.186 pr(-3) -0.031 0.049 -0.076 0.047 dq * -0.375 0.050 -0.390 0.198 dpop -0.684 0.188 -0.267 1.293 CountryEffects

Y

Y

Y

YearEffects

N

N

Y

ImpulseResponses t 0.598 0.096 0.504 0.054 0.746 0.059 0.691 0.286 t-1 -0.281 0.133 -0.231 0.096 -0.193 0.095 -0.179 0.222 t-2 -0.120 0.106 -0.119 0.058 -0.114 0.056 -0.111 0.142 t-3 -0.120 0.095 -0.122 0.060 -0.098 0.054 -0.088 0.106 t-4 -0.102 0.103 -0.102 0.042 -0.122 0.047 -0.059 0.076 t-5 -0.060 0.078 -0.039 0.028 -0.068 0.040 -0.038 0.063 t-6 -0.014 0.024 -0.040 0.037 -0.024 0.057 t-7 -0.003 0.020 -0.019 0.035 -0.018 0.048 t-8 0.000 0.016 -0.008 0.033 -0.014 0.041 t-9 0.001 0.012 -0.002 0.030 -0.013 0.036 t-10 0.001 0.009 0.001 0.027 -0.011 0.032

Note: Thistable reportsthe results ofestimating Equation (11)inthepaper. Thefirstthreecolumnsassumeslope

coefficients are he

same

acrosscountries. The lastcolumnreportsthe

mean

andstandarddeviation across

countriesofcountry-by-countryestimates Standard errors fortheimpulseresponsesincclumns(2)and(3)are

simulated using500 drawsfromtheestimateddistributionof coefficients.

Figure

1:

The

Traditional

Rule

and

the

New

Rule

Traditional

Rule

nationalsaving :rvations foran 0.1 -0.05 -Q.

Q

O

0-*3 § ( g -0.05

-<

£ -0.1 -3

o

-0.15 --0.2

y=0.231x-

0.062 R2= 0.124

*

*

Saving/GDP

New

Rule

0.1 -y =0.939x-_0.002

R

2 =0.301 Ji.Q5,

W"

*

0.05 5alth) § -0.15 *0.1

*

_^

_-1W5

*WM

1

^^^

**

£

^

+

U

-<%-5

-0.15 --0.2

(Saving/GDP)x (Foreign

Assets/W

Note: Thetopi

(gross national

unbalancedpar

bottom) panelplotsthe currentaccount balance as a shareof

GDP

againstgross

saving interacted with the foreignassetposition),poolingallavailableannual obs«

leiof21

OECD

countriesoverthe period 1966-1997.