8/3/2019 Capital flow-2006-05-21

1/48

A Solution to Two Paradoxes of

International Capital Flow

Jiandong Ju and Shang-Jin Wei

* Personal views, not those of the IMF

8/3/2019 Capital flow-2006-05-21

2/48

Motivation

Cross-border capital flow reached nearly $6 trillion in2004. Less than 10% goes to developing countries.

The paradox of too little capital flow: in a one-sectormodel, marginal product of capital is lower in rich

country, but the amount of capital from rich to poorcountries is too small (the Lucas Paradox)

Example: India vs the U.S. (5800% difference in MPK)

The paradox of too much capital flow: in a 2-sector,2-factor model, factor prices are equalized in a freetrade world (FPE due to Samuelson). So there is noincentive for any capital to flow.

8/3/2019 Capital flow-2006-05-21

3/48

Objectives ofthe paper

Existing explanationsofthe Lucasparadox do

notsurvive in a generalization to a 2X2 model

To build a micro-founded non-neo-classicaltheory tosolve the twoparadoxes

Tohighlight (possibly different) rolesoffinancial

development and property rightsinstitutionsin

international capitalflows

8/3/2019 Capital flow-2006-05-21

4/48

Existing explanationofthe Lucas paradox

withinaneo-classical framework

Difference in effective labor

Missing factor (e.g. human capital)

Sovereign risk (Reinhart and Rogoff)

Trade cost (Obstfeld and Rogoff) Difference in TFP (ofwhichinstitution is a special

case)

Common problem:They do notsurvive in a generalization to a neo-

classicaltwosector, twofactor model

8/3/2019 Capital flow-2006-05-21

5/48

Chain rule ofFPE

Lemma 1:Let # factors = m. Allother neo-classicalassumptions apply. For any two countries, factorprices are equalized ifthe countries can be linked by

a sequence ofcountry pairs, and ifthe countrieswithin eachpairproduce a common setofmproducts. Example: Twofactors (land and capital)

US and India may notproduce anything in common, and

may not even trade with eachother. ButFPE could hold if US-Greece (apple & apricot)

Greece-Thailand (beer and bottle)

Thailand-India (cabbage and carriage)

8/3/2019 Capital flow-2006-05-21

6/48

Ifexisting explanationsofthe Lucasparadox dontwork,

what abouttextbook reasonsthat breakthe FPEin the2X2X2 model?

Difference in technology

No eqbm in general (Panagariya)

We are NOTsaying thatFPEis realistic,

butthatitis much more difficultto escape from the

tyranny ofFPEthatthe existing literature may haverealized.

8/3/2019 Capital flow-2006-05-21

7/48

Intuitive outline ofour model

We work with a two-sector model but with two twists

To resolve the Lucas paradox, we introduce a

financial contract between entrepreneurs andinvestors: Each only gets a slice of the marginalproduct of physical capital.

To move away from FPE, we introduceheterogeneous entrepreneurs, which result insector- level DRS (despite firm-level CRS).

8/3/2019 Capital flow-2006-05-21

8/48

Re-do Lucas example: India vs the U.S.

Indias K/L ratioisonly 1/15 ofthe U.S.

Itsfinancialsystem is also muchless efficient

In the absence ofcapitalflow, the return tofinancialinvestmentislowerin India than in the U.S.

India experiences an outflowoffinancial capital Atthe same time, because Indians return to

physical capitalishigher-> InflowofFDIInflowofFDI is biggerthan itwould have been ifits

financialsystem had been more efficient Return differentialissmallerthan Lucas calculation

Muchsmallerfriction can stopthe capitalflows

8/3/2019 Capital flow-2006-05-21

9/48

Roadmap

The Model Two key parameters

Financial development

Control of expropriation risk (property rights protection)

Comparative Statics Free trade in goods

Financial capital flow

FDI World capital market equilibrium

Some very preliminary/suggestive evidence

8/3/2019 Capital flow-2006-05-21

10/48

Model Description

Within an economy (2 sectors, 2 factors)

For a given sector:

Labor

Capitalists (each endowed wone unitofcapital) Entrepreneurs + financialinvestors

Linked by financial contracts

2-period production; Liquidity shockin 2nd period

Moralhazard problem

Two country world economy

Variousscenariosofcapitalflows

8/3/2019 Capital flow-2006-05-21

11/48

Time line of the model

8/3/2019 Capital flow-2006-05-21

12/48

The Model

8/3/2019 Capital flow-2006-05-21

13/48

8/3/2019 Capital flow-2006-05-21

14/48

Financial Contract:

8/3/2019 Capital flow-2006-05-21

15/48

Solution

8/3/2019 Capital flow-2006-05-21

16/48

AllocationofCapital withinandacross Sectors

Lemma 2: The more productive entrepreneurs enter

the heterogeneous sector, while the less productiveones enter the homogeneous sector. In theheterogeneous sector, relatively more productiveentrepreneurs manage more capital.

8/3/2019 Capital flow-2006-05-21

17/48

Free Entry Conditions

8/3/2019 Capital flow-2006-05-21

18/48

A Stolper-Samuelson Plustheorem holds:

(Prop 1)

When p r butw

When r butw

When r but no change in w When N1 r butw

ButFPE does nothold!

8/3/2019 Capital flow-2006-05-21

19/48

DeterminationofFactor Prices

Proposition 1: An increase in N1 will decrease r but increase w.An improvement in the level of financial development willincrease r but has no effect on w. Lower expropriation riskincreases r but decrease w.

8/3/2019 Capital flow-2006-05-21

20/48

EquilibriumConditions

8/3/2019 Capital flow-2006-05-21

21/48

A RybczynskiPlustheorem holds:

(Prop 2)(under a modified non-reversaloffactorintensity)

When K (or L) N1, y1 more than y2,and p

When y1 and y2 proportionately, but

no change in p (or N1) When N1&N2 , y1& y2 proportionately,

but no change in p

8/3/2019 Capital flow-2006-05-21

22/48

Comparative Statics

Proposition 2: The increase in K will increase N1, and

decrease the relative price of good 1. The improvement inthe level of financial development, however, has no effect onoutputs and the commodity price. Lower risk expropriationdecreases N1 and N2, but has no effect on p.

8/3/2019 Capital flow-2006-05-21

23/48

Combining Propositions 1 and 2

When K/L N1 (prop 2)

r butw (prop 1)

The intuition from a one-sector modelis restored inthistwo-sector, two-factor model!

Question: Isthe LucasParadox also restored?

No! The differentialin returnsto capital dependsonc1f/(1+f), which can be very small

Evidence: Caselli and Feyrer (2005)

8/3/2019 Capital flow-2006-05-21

24/48

Moving from closed to open economy

Four-step discussion

Free trade in goods Justfinancial capitalflow (+ free trade)

JustFDI (+ free trade)

Bothtypesofcapitalflows (free trade)

8/3/2019 Capital flow-2006-05-21

25/48

Free Trade in Goods

Two countries differin factor endowments and levelsof

financial development and property rightsprotection.

Prop 3: The Heckscher-Ohlin theorem stillholds: Eachcountry exportsthe good that usesits more abundant

factorintensively.

8/3/2019 Capital flow-2006-05-21

26/48

Financial Capital Flow

Proposition 4: If the two countries have the samelevel of property rights protection and financial

development, financial capital will flow out of the

capital abundant country, and into the capital scarce

one. If the two countries have the same capital-labor

ratio, financial capital will flow out of the country

with lower financial development or poorer property

rights protection and into the other one.

8/3/2019 Capital flow-2006-05-21

27/48

ForeignDirectInvestment

Proposition 5: Suppose trade in goods is free and expropriation

risk in the two countries are the same, FDI will flow out of the

capital abundant country to the labor abundant country. If thetwo countries have the same K/L ratio, then FDI will go from the

country with poor property rights protection to the other.

8/3/2019 Capital flow-2006-05-21

28/48

Free capital mobility + free goodstrade

Ifa country haslow K/L and low , then itexperiencestwoway grossflows (outflowoffinancial capital butinflowofFDI), and a smallnetflow

e.g. China

Ifa country has a low K/L and low , thenoutflowoffinancial capital + outflowofFDI

e.g. Zimbabwe

8/3/2019 Capital flow-2006-05-21

29/48

Contrasting effects ofpoor financial development vs.

poor property rights protection

A lower level of financial development results in alower r, which generates an outflow of financialcapital. As a result, w becomes lower, which

attracts more FDI th

an oth

erwise. Worse property rights protection results in both a

lower profit, leading to less FDI, and a lower r,leading to outflow of financial capital

Empirical evidence: Wei 2006

8/3/2019 Capital flow-2006-05-21

30/48

Property rights protection,financial development,

and compositionofcapital flow (Wei, 2006,connectingtwo

views onfinancial globalization)IV Regression

FDI/total

foreign

liability

Portolio equity

/total foreign

liability

Portolio debt

/total foreign

liability

Loan/total

foreign liability

Institutional Quality 0.67** -0.11 0.38** -0.81*(0.29) (0.11) (0.17) (0.40)

Financialdevelopment

-0.88* 0.31* -0.40 0.65

(0.46) (0.18) (0.27) (0.66)

Resourcea

0.13 0.04 0.05 -0.15(0.13) (0.05) (0.08) (0.18)

Openness a 0.12* 0.01 -0.08* -0.23(0.07) (0.03) (0.04) (0.14)

Observations 34 34 34 33

R-squared 0.36 0.40 0.47 0.56

Measure of Institutions Average ofSix World Bank Indicators

8/3/2019 Capital flow-2006-05-21

31/48

Capital Bypass Circulation,or transfusion

Proposition 6:A unique equilibrium: Inefficient

financial system is completely bypassed.

In the transition to the eqbm, the country with a higher initialK/L always exports capital on net (i.e. running a CA deficit)

So a scenario in which the US runs a current account (CA)

deficit, China CA surplus can in principle be rationalized w/o Exchange rate policy

Mercantilist trade strategy

Fiscal deficit

Free capital mobility + free trade + free mobility ofentrepreneurs

8/3/2019 Capital flow-2006-05-21

32/48

Capital Bypass Circulation

FDIFCF

Y2

O

K

K+K*

A

B

B*

A*

L L+L*

F O*

E

CH

Y1

8/3/2019 Capital flow-2006-05-21

33/48

Capital Bypass Circulation,or transfusion

8/3/2019 Capital flow-2006-05-21

34/48

Capital MarketEquilibrium:Different

ExpropriationRisk

8/3/2019 Capital flow-2006-05-21

35/48

DifferentExpropriationRisk

In equilibrium, wage is always higher in the country with better financialinstitution or lower expropriation risk.

8/3/2019 Capital flow-2006-05-21

36/48

Prop 5:

Suppose the two countries are diversified in the

equilibrium withfree trade and free capital

mobility, then the wage rate is always (atleast

weakly)higherin the country with betterproperty

rightsprotection orwith betterfinancial

development

8/3/2019 Capital flow-2006-05-21

37/48

Conclusions

Existing explanationsofthe Lucasparadox dontsurvivein a modelwithtwosectors and twofactors. Itis difficulttosimultaneously resolve Lucasparadox and FPEin aneo-classicalframework

We build a micro-founded non-neoclassical model

Key twists:

Financial contracts

Heterogeneousfirms

The modelhighlights (potentially different) rolesoffinancial development and property rightsprotection

It generatespredictions aboutgross aswell asnetcapitalflows. It avoids boththe Lucasparadox and FPE.

8/3/2019 Capital flow-2006-05-21

38/48

Future Work (I)

Solution tootherpuzzlesin intlfinance?

Feldstein-Horioka puzzle

Shutting down risk-sharing motivation

Smallfriction to capital mobility

investment = saving

Home biasin equity holdings

Equity instead ofdirectfinancing contracts

Smallfriction to capital mobility

8/3/2019 Capital flow-2006-05-21

39/48

Future Work (II)

Empirics

Dynamics

Welfare analysis /conflictofinterest

Alternative financial contracts

Frictionsto capitalflow

8/3/2019 Capital flow-2006-05-21

40/48

Financial development, corruption,and composition

ofcapital flows: Preliminary evidence

Challenge:

measuresofinstitutions may be endogenous

Instrumental variable for government corruption:

Initial costto colonizers mortality rate ofEuropean

settlers before 1850

Acemoglu, Johnson, and Robinson (AER 2001)

Alternative: initialpopulation density in 1500

8/3/2019 Capital flow-2006-05-21

41/48

Instrumental variablesforfinancial

development:

Legalorigins: La Porta, Lopez-de-silanes,

Shleifer, and Vishny (JPE 1998)

Settler mortality

8/3/2019 Capital flow-2006-05-21

42/48

(History-based) instrumental variables

Corruption is mostly affected by settler

mortality but not by legalorigin

Financial developmentis affected by both

legalorigins and settler mortality.

8/3/2019 Capital flow-2006-05-21

43/48

The basic specification:

(1) Composition(j) = 1 Corruption(j)

+ 2 FinDev(j) + Z(j) + e(j)

Zj is a vectorofcontrol variables,

1, 2, and are parameters

ej is a random error.

8/3/2019 Capital flow-2006-05-21

44/48

Table 7:Addingmore control variables (IVRegressions)

FDI/total foreign

liability

Portolio

equity/totalforeign liability

Portolio debt

/total foreignliability

Loan/total

foreign liability

(1) (2) (3) (4) (5) (6) (7) (8)

Corruption(GCR/WDR) -0.55** -0.42* 0.09 0.17* -0.34** -0.27* 0.69* 0.29

(0.24) (0.25) (0.10) (0.09) (0.14) (0.13) (0.34) (0.28)

Financial development -0.87* -0.76 0.31* 0.38** -0.48* -0.42* 0.72 0.28

(0.48) (0.46) (0.19) (0.16) (0.26) (0.25) (0.68) (0.54)

Resourcea

0.13 0.12 0.04 0.04 0.05 0.05 -0.15 -0.16

(0.13) (0.13) (0.05) (0.04) (0.07) (0.07) (0.18) (0.14)

Opennessa

0.13* 0.17** 0.01 0.04 -0.09** -0.07* -0.22 -0.39**

(0.07) (0.07) (0.03) (0.02) (0.04) (0.04) (0.15) (0.12)

FDI restrition Dummy -0.01 -0.06 -0.00 -0.03* 0.05* 0.02 -0.04 0.09

(0.05) (0.06) (0.02) (0.02) (0.03) (0.03) (0.07) (0.06)

Log(GDP) 0.04* 0.03** 0.02* -0.10**

(0.02) (0.01) (0.01) (0.02)

Observations 34 34 34 34 34 34 33 33

R-squared 0.36 0.43 0.40 0.58 0.53 0.59 0.57 0.74

8/3/2019 Capital flow-2006-05-21

45/48

Explainingthe RatioofFDI/Total Foreign Liabilities in 2003

IV regressions

Corruption(GCR/WDR) -0.10** -0.65** -0.56**

(0.04) (0.23) (0.24)

Financial development 0.17* -1.07** -0.88*

(0.09) (0.44) (0.46)

Resourcea

0.13(0.13)

Openness a 0.12*

(0.07)

Observations 40 34 34 34

R-squared 0.15 0.09 0.28 0.36

8/3/2019 Capital flow-2006-05-21

46/48

FirstStage Regressions:

Using Histories toInstrumentModern-day Institutions

Corruption(GCR/WDR) Financial development Institutional Quality

(1) (2) (3) (4) (5) (6) (7) (8) (9)

Log(settler

mortality)

0.46** 0.31** -0.21** -0.38** -0.29**

(0.08) (0.08) (0.03) (0.07) (0.07)

Log(Population 0.27** 0.10 -0.07**

density in 1500) (0.07) (0.08) (0.03)

Legal origin 0.37 0.62** -0.18** -0.14* -0.18** -0.06

(French) (0.23) (0.22) (0.08) (0.08) (0.08) (0.17)

Legal origin 0.00 0.00 0.74* 0.00 0.00 0.00

(German) (0.00) (0.00) (0.38) (0.00) (0.00) (0.00)

Legal origin 0.00 0.00 0.70* 0.00 0.00 0.00

(Scandivanian) (0.00) (0.00) (0.38) (0.00) (0.00) (0.00)

Legal origin 0.71 0.79 -0.25** -0.29 -0.14 -0.98**

(Socialist) (0.66) (0.72) (0.10) (0.21) (0.25) (0.45)

Observations 44 48 40 44 120 60 73 70 61

R-squared 0.44 0.24 0.36 0.20 0.14 0.47 0.14 0.33 0.29

8/3/2019 Capital flow-2006-05-21

47/48

Evidence isstillpreliminary, butintriguing

More needsto be done

8/3/2019 Capital flow-2006-05-21

48/48

Welfare Impacts: Financial Capital flow

The welfare effect of financial capital outflow isdetermined by the trade off between investors' gain

and entrepreneurs' loss. If the later dominates theformer, welfare is reduced at home due to financialcapital outflow.