1、International Financial Integration and Crisis Intensity,Andrew K. Rose UC Berkeley, NBER and CEPR (collaborator with Mark Spiegel),1,ASEAN Presentation, Seoul December 2010,What Caused the 2008-09 Crisis?,Defining macroeconomic event of generations A recent “Natural Experiment” Did International Fi
2、nancial Integration play any significant role? If so, multilateral or bilateral ties matter most? If bilateral, which country matters most for exposure/insurance? Big Asian Economies: China; Japan; Korea. United States,ASEAN Presentation, Seoul December 2010,2,Non-Structural Methodology,MIMIC (multi
3、ple-indicator, multiple cause) model explicitly incorporate difficulties in observing relative crisis severity Treats crisis as a latent variable Observed with error as function of observable manifestations Also do OLS, graphical analysis Sample is cross-section of 85 countries,3,ASEAN Presentation,
4、 Seoul December 2010,Crisis Performance is Latent Variable Mapping Observable Causes to Observable Manifestations,4 Manifestations from 2008-09: equity market collapse, exchange rate devaluation, economic growth, and change in creditworthiness Compare to 2006 vulnerabilities (causes) National causes
5、 International financial linkages,4,ASEAN Presentation, Seoul December 2010,Results, 1,Plausible estimates of relative severity of crisis Ex: Iceland and Latvia identified as exceptionally bad Most Asians have mild crises Use literature to find three “national” crisis causes (determinants from 2006)
6、 Real Income (negative “progressive” effect”) Financial Market Regulation (looser worse crisis) Current Account (% GDP) (deficit worse),5,ASEAN Presentation, Seoul December 2010,Results, 2,Less success linking crisis to integration variables No multilateral variables consistently enter at statistica
7、lly significant levels Most bilateral variables irrelevant too Little effect of: a) banking exposure; b) debt denomination; c) Fed swap lines Asset exposure (CPIS) matters, but not strong More exposure to US more mild crisis!,ASEAN Presentation, Seoul December 2010,6,Data,7,ASEAN Presentation, Seoul
8、 December 2010,Empirical Strategy,Exploit recent international cross section “Great Recession” a “natural experiment” Dont have to rely on time-series/panel data All information country-specific initially Then add linkages between countries National characteristics may influence vulnerability to for
9、eign shocks Medical analogy: common shocks vs. contagion,8,ASEAN Presentation, Seoul December 2010,Data Span,As crisis hit all types of countries, include both developed and EMEs All countries $10,000 per capita GDPAll countries $4,000 per capita GDP, plus population 1 million (WB: upper middle inco
10、me) Maximum of 85 countries in sample Use only publicly-available series available for reasonable span of countries,9,ASEAN Presentation, Seoul December 2010,Sample of Countries,ASEAN Presentation, Seoul December 2010,10,Differences in Crisis Severity (Manifestations/Consequences),Crisis intensity o
11、nly observed with error e.g. Berg, et al (2004) Abundant measurement error likely Initially model severity as a latent variable linked to four observable indicators for 2008-09 GDP growth (%) Change in National Equity Markets (stocks, %) Change in multilateral SDR exchange rate (%) Change in Euromon
12、ey country credit rating March 2008 - March 2010,11,ASEAN Presentation, Seoul December 2010,Table 1: Crisis Manifestations,ASEAN Presentation, Seoul December 2010,12,Methodology,13,ASEAN Presentation, Seoul December 2010,MIMIC model,MIMIC model consists of two sets of equations: (1)(2)where is crisi
13、s indicator, is an observation for potential crisis cause; is latent variable representing severity of the crisis (or lack thereof in our case), and and are well-behaved disturbances Equation (1) links observable (2008-09) manifestations of the crisis to latent variable Equation (2) links latent var
14、iable to (2006) crisis causes,14,ASEAN Presentation, Seoul December 2010,Characteristics of MIMIC Model,Substitute (2) into (1), eliminate latent variable MIMIC model is then a system of J (=4) equations with right hand sides restricted to be proportional With normalization, system is identified Nor
15、malize on equity returns (Breusch) Desirable feature of MIMIC model is ability to systematically address measurement error Estimate with STATA using GLLAMM model Rebe-Hesketh, et al (2004),15,ASEAN Presentation, Seoul December 2010,What Determinants are Associated with Crisis Incidence?,Survey liter
16、ature (Rose and Spiegel 2009a, 2009b, 2010), find three determinants Use these as national controls throughout Real Income (negative effect) Credit Market Regulation (looser is worse) Current Account, % GDP (higher is better),ASEAN Presentation, Seoul December 2010,16,Sensitivity Analysis,Drop Excha
17、nge Rate Consequences Restrict to Asian/Pacific Countries (and USA) OLS with growth as dependent variable OLS with stock market change as regressand,ASEAN Presentation, Seoul December 2010,17,MIMIC Estimates with Controls,ASEAN Presentation, Seoul December 2010,18,Findings consistent with Literature
18、,Real GDP per capita has negative, significant effect (the “progressive” crisis) More Credit Market Regulation associated with milder crisis (Giannone et al), but rarely significant Current Account surpluses associated with milder crises,ASEAN Presentation, Seoul December 2010,19,Crisis was Progress
19、ive,ASEAN Presentation, Seoul December 2010,20,Role of Current Account,ASEAN Presentation, Seoul December 2010,21,International Financial Integration,22,ASEAN Presentation, Seoul December 2010,Multilateral Financial Linkages,Examine (6) imperfect multilateral proxies for 2006 financial integration,
20、a)-e) as %GDP a) net foreign assets; b) external debt; c) short-term external debt; d) financing via international capital markets; e) international reserves. f) currency union dummy,23,ASEAN Presentation, Seoul December 2010,Unchanged Methodology,Always include (3) national causes Same sensitivity
21、analysis,ASEAN Presentation, Seoul December 2010,24,Role of Multilateral Financial Linkages,ASEAN Presentation, Seoul December 2010,25,Poor Multilateral Results,Little effect of international financial integration, after taking into account (3) domestic factors 30 coefficients; one significantly dif
22、ferent from zero at 1%; two more at 5%. Short-Term External Debt/GDP consistently negative effect (Blanchard et al) Significant 2/5 times,ASEAN Presentation, Seoul December 2010,26,Proceed on to Bilateral Linkages,Exposure to Individual Countries Different Sectors: All External Assets (CPIS) Debt, L
23、ong-Term Debt as well Bank Loans (BIS) Debt Denomination (WB) Fed swap-lines,ASEAN Presentation, Seoul December 2010,27,(17) Measures of Bilateral Financial Linkages,Share of overseas assets held in USA (as proportion of all overseas wealth) Also consider China, Japan, Korea Also consider debt, long
24、-term debt BIS Consolidated banking data Available for Japan, USA PPG debt denominated in yen/$ Finally, (endogenous) existence of Fed swapline,28,ASEAN Presentation, Seoul December 2010,CPIS: American Exposure Helps!,29,ASEAN Presentation, Seoul December 2010,Special Role of US,Crisis originated in
25、 US Provider of International Reserves Monetary Anchor (especially in Asia) Dollar Appreciation in 2008 But exposure to US assets associated with more mild crises,ASEAN Presentation, Seoul December 2010,30,Robust to Exact Asset Considered,31,ASEAN Presentation, Seoul December 2010,Graphical Evidence
26、: US Exposure,ASEAN Presentation, Seoul December 2010,32,Japanese Exposure (note x-scales),ASEAN Presentation, Seoul December 2010,33,Korea (note x-scales),ASEAN Presentation, Seoul December 2010,34,China (note x-scales),ASEAN Presentation, Seoul December 2010,35,BIS Consolidated Banking Claims, Deb
27、t Denomination, Fed Swaplines,36,ASEAN Presentation, Seoul December 2010,Outliers Important for Japanese BIS consolidated banking data,ASEAN Presentation, Seoul December 2010,37,American Analogue,ASEAN Presentation, Seoul December 2010,38,Some Sensitivity, Bigger Mystery,Why does exposure to America
28、 (most likely epi-center of 2008 crisis) seem to help? Special Advantage of “Exorbitant Privilege”? Interesting Future Research topic Still, results not very strong Most measures insignificant No evidence that Asian links helped in crisis,39,ASEAN Presentation, Seoul December 2010,Summary and Conclu
29、sion,40,ASEAN Presentation, Seoul December 2010,Examine causes and consequences of 2008-09 Financial Crisis,MIMIC Methodology explicitly confronts fact that “crisis severity” observed with error Account for national crisis causes that work in literature Cross-sectional focus: only trying to explain
30、relative incidence, not timing,41,ASEAN Presentation, Seoul December 2010,Message,Financial Integration across countries not very important in understanding crisis incidence Seems to provide little insurance (though may also limit contagion) Multilateral measures especially weak Bilateral American Financial Influence: Enduring Impact,ASEAN Presentation, Seoul December 2010,42,
