7th jacques polak annual research ...discussion of: financial integration, macroeconomic volatility...

The views expressed in this paper are those of the author(s) only, and the presence of them, or of links to them, on the IMF website does not imply that the IMF, its Executive Board, or its management endorses or shares the views expressed in the paper.

77TTHH JJAACCQQUUEESS PPOOLLAAKK AANNNNUUAALL RREESSEEAARRCCHH CCOONNFFEERREENNCCEE NNOOVVEEMMBBEERR 99--1100,, 22000066

Discussion of: Financial Integration,Macroeconomic Volatility and Welfare

by Martin Evans and Victoria Hnatkovska

Fabrizio PerriUniversity of Minnesota, Minneapolis FED and NBER

IMF Seventh Jacques Polak Annual Research Conference,Washington, November 2006

The Issue

What are the consequences of changes internationalfinancial integration?

Answers from EH modelOn co-movements: LargeOn volatility: Small and ambiguousOn welfare: Tiny

What can we learn about financial integration and our modelsabout financial integration?

The Issue

What are the consequences of changes internationalfinancial integration?

Answers from EH modelOn co-movements: LargeOn volatility: Small and ambiguousOn welfare: Tiny

What can we learn about financial integration and our modelsabout financial integration?

The Issue

What are the consequences of changes internationalfinancial integration?

Answers from EH modelOn co-movements: LargeOn volatility: Small and ambiguousOn welfare: Tiny

What can we learn about financial integration and our modelsabout financial integration?

Financial Integration and co-movement: theory

More financial integration implies more equalization ofmarginal valuations (MRS) across states and dates

ExamplesComplete Markets. Complete equalization of MRS acrossall dates and states, regardless of shock processFinancial Autarky. MRS driven solely by shocks process

In EH model MRS ' C, thus more integration leads to higherconsumption co-movement (from 0 to 0.73), but not necessarilyoutput co-movement

Financial Integration and co-movement: theory

More financial integration implies more equalization ofmarginal valuations (MRS) across states and dates

ExamplesComplete Markets. Complete equalization of MRS acrossall dates and states, regardless of shock process

Financial Autarky. MRS driven solely by shocks processIn EH model MRS ' C, thus more integration leads to higherconsumption co-movement (from 0 to 0.73), but not necessarilyoutput co-movement

Financial Integration and co-movement: theory

More financial integration implies more equalization ofmarginal valuations (MRS) across states and dates

ExamplesComplete Markets. Complete equalization of MRS acrossall dates and states, regardless of shock processFinancial Autarky. MRS driven solely by shocks process

In EH model MRS ' C, thus more integration leads to higherconsumption co-movement (from 0 to 0.73), but not necessarilyoutput co-movement

Financial Integration and co-movement: theory

More financial integration implies more equalization ofmarginal valuations (MRS) across states and dates

ExamplesComplete Markets. Complete equalization of MRS acrossall dates and states, regardless of shock processFinancial Autarky. MRS driven solely by shocks process

In EH model MRS ' C, thus more integration leads to higherconsumption co-movement (from 0 to 0.73), but not necessarilyoutput co-movement

Financial Integration and co-movement: evidence

Financial Integration and co-movement: lessons

The EH model (as many others) predicts a strong relationbetween financial integration and consumptionco-movement.

Support for this implication is limitedMain problem is the strong link between MRS andconsumption (also not supported by pricing data)Next: look at the evidence using more sophisticated MRS

Financial Integration and co-movement: lessons

The EH model (as many others) predicts a strong relationbetween financial integration and consumptionco-movement.Support for this implication is limited

Main problem is the strong link between MRS andconsumption (also not supported by pricing data)Next: look at the evidence using more sophisticated MRS

Financial Integration and co-movement: lessons

The EH model (as many others) predicts a strong relationbetween financial integration and consumptionco-movement.Support for this implication is limitedMain problem is the strong link between MRS andconsumption (also not supported by pricing data)

Next: look at the evidence using more sophisticated MRS

Financial Integration and co-movement: lessons

The EH model (as many others) predicts a strong relationbetween financial integration and consumptionco-movement.Support for this implication is limitedMain problem is the strong link between MRS andconsumption (also not supported by pricing data)Next: look at the evidence using more sophisticated MRS

Financial Integration and Volatitilty

Macro Volatility = Volatility of Underlying Shocks + Amplification

In EH model (as in many others) shocks do most of thework.Small role of financial integration (Volatility of output goesfrom 0.77% to 0.84%Evidence from financial crises suggest a link betweenfinancial integration and shocks itself (in Mexico andArgentina large drops in TFP)

Next: explore models with the volatility of shocks candepend on integration

Financial Integration and Volatitilty

Macro Volatility = Volatility of Underlying Shocks + Amplification

In EH model (as in many others) shocks do most of thework.Small role of financial integration (Volatility of output goesfrom 0.77% to 0.84%Evidence from financial crises suggest a link betweenfinancial integration and shocks itself (in Mexico andArgentina large drops in TFP)Next: explore models with the volatility of shocks candepend on integration

Financial Integration and Welfare

In EH the gains of going from FA to BE are 0.006% oflifetime consumptionFrom BE to HI even smaller

Why?

Shocks to non tradable cannot (by assumption) be sharedShocks to tradables are trend stationary (AC = 0.7)Business Cycle Shocks are small (Lucas)

Financial Integration and Welfare

In EH the gains of going from FA to BE are 0.006% oflifetime consumptionFrom BE to HI even smaller

Why?

Shocks to non tradable cannot (by assumption) be sharedShocks to tradables are trend stationary (AC = 0.7)Business Cycle Shocks are small (Lucas)

Financial Integration and Welfare

In EH the gains of going from FA to BE are 0.006% oflifetime consumptionFrom BE to HI even smaller

Why?

Shocks to non tradable cannot (by assumption) be sharedShocks to tradables are trend stationary (AC = 0.7)Business Cycle Shocks are small (Lucas)

Is Financial Integration irrelevant?

Not if it affects trend growth (Obstfeld, 1994)Not if offers individuals (as opposed to countries) betterinsurance opportunities (Davis, Nalewaik and Willen, 2001)Not if shocks are not trend stationary (Aguiar andGopinath, 2005)

Gains from holding a world portfolio

yit = ρyit−1 + εit

u(yit) =1

1− σ(yit)1−σ

σ = 2, σε = 0.02, β = 0.99

200 countries, 200 periods

If ρ is close to 1 shocks cannot be insured with a simplebondWelfare impact of stock market financial integration can bevery large!

Gains from holding a world portfolio v/s a countryportfolio

0.9 0.92 0.94 0.96 0.98 1 1.020

2

4

6

8

10

12

ρ (persistence of shocks)

Per

cent

of l

ifetim

e co

nsum

ptio

n

What next

If gains are so large is then very important to study optimalportfolio diversification (EH, 2005)

Next: why does not capital flow from poor (and unstable)countries to rich (and stable) countries?

What next

If gains are so large is then very important to study optimalportfolio diversification (EH, 2005)Next: why does not capital flow from poor (and unstable)countries to rich (and stable) countries?