On the International Transmission of Shocks: Micro-Evidence from Mutual Fund Portfolios Claudio Raddatz Sergio Schmukler Central Bank of Chile The World Bank Debt and Credit, Growth and Crises Banco de España-World Bank Conference June 18 & 19, 2012 I. Motivation Volatile capital flows, intermediated by financial intermediaries (FI) Need to understand FI and the micro aspects of their behavior In particular, do FI make capital flows more volatile and pro-cyclical? Ex-ante risk taking behavior that generate crises Ex-post propagation of shocks across markets and countries (Q and P) Intermediaries face principal agent problems that affect allocations [Underlying Investors FI Agents (Managers)] Assets Limited evidence on inner-workings of FI at international level Aggregate investments x-countries, by foreign & domestic agents, and certain intermediaries (mostly banks) I. Motivation: This Paper How do underlying investors and managers behave, react to shocks, and contribute to transmit crises across countries? Focus on international mutual funds Increasingly important drivers of capital flows Detailed micro-level portfolio data More than 1,000 equity and bond funds, monthly, starting 1996 Disentangle the behavior of Underlying investors (UI) by measuring injections/redemptions Managers (M) by changes in country-portfolio weights and cash I. Motivation: This Paper Allocations / Net Inflows Underlying investors Managers Injections / Redemptions Country 1 Country i Country N Cash I. Motivation: Specific questions What is the contribution of injections to changes in MF investment? Do managers significantly change country-portfolio weights over time? What are the determinants of investor’s injections and manager’s portfolio allocations? How do injections and allocations respond to crises? At the country level, how much of the volatility of capital flows is driven by fund managers versus investors? Are there differences between bond and equity funds? Presentation I. Motivation II. Data and Summary Statistics III. Shocks to Managers and Portfolio Reallocations IV. Behavior of Investors and Managers V. Gross and Net Country Flows VI. Conclusions II. Data: Micro-level Dataset on Mutual Funds: (EPFR ) Data coverage (monthly frequency) 1,076 funds: global and EM funds 965 equity funds: Jan 1996-Nov 2010 111 bond funds: Jul 2002-Nov 2010 7,429,000 obs./weights across funds, 124 countries, and over time Equity funds: 6,867,500 obs. Bond funds: 561,500 obs. Variables Total net assets (TNA) % of the funds’ assets allocated to each country and held in cash Investor type: active/passive Investment scopes (geographical regions) Others: fund domicile, family, main currency denomination II. Data: Additional Data Fund prices (NAV) 255,510 obs., monthly basis 90% of matches with EPFR funds: 896 equity funds, 106 bond funds Sources: Bloomberg and Datastream Used to compute returns and injections to funds Country stock and bond market indexes (U.S. dollars) 23,272 obs., monthly basis Equity markets: 86 countries, Jan 1999-Nov 2010 Bond markets: 78 countries, Jul 2002-Nov 2010 Sources: MSCI std. index, S&P BM index, local sources , JP Morgan sovereign bond index Used to compute the flows to the countries Use country-level indexes to compute returns at country level II. Data: Summary and Main Variables fijt Fijt / Ait 1 Country Level Injections/Redemptions Bloomberg, Dstream,etc. Ai , j ,t R j ,t Fi , j ,t Ai , j ,t Ai , j ,t 1 R jt Country Assets wi , j ,t & Allocations EPFR Ai ,t TNA Ri ,t DStream (match) Fi ,t Ai ,t Ai ,t 1 Rit Fund Level Injections/Redemptions fit Fit / Ait 1 II. Evolution of Total Assets in Equity Funds Ai ,t i 2001-2010 800,000 90,000 80,000 70,000 60,000 50,000 40,000 30,000 20,000 10,000 0 Millions of USD 700,000 600,000 500,000 400,000 300,000 200,000 100,000 • Assets fluctuate importantly in expected manner (pro-cyclically ) • Driven by prices (returns from previous allocations) or injections? Jun. 10 Jun. 09 Jun. 08 Jun. 07 Jun. 06 Jun. 05 Jun. 04 Jun. 03 Jun. 02 Jun. 01 Jul. 00 Jan. 00 Jul. 99 Jan. 99 Jul. 98 Jan. 98 Jul. 97 Jan. 97 Jul. 96 0 Jan. 96 Millions of USD 1996-2000 II. Evolution of Portfolio Composition around the GFC Global Equity Funds Developed Europe 50% Emerging Countries 15% 14% 49% 13% 48% 12% 47% 11% 46% 10% 45% Northern Rock 44% 43% Bear Stearns 9% Lehman Brothers AIG 8% North America 23% 22% 21% 20% 19% 18% 17% 16% Average portfolio shares Sep.09 May.09 Jan.09 Sep.08 May.08 Jan.08 Sep.07 May.07 Jan.07 15% Sep.09 May.09 Jan.09 Sep.08 May.08 Jan.08 Sep.07 Jan.07 Sep.09 May.09 Jan.09 Sep.08 May.08 Jan.08 Sep.07 May.07 Jan.07 May.07 7% 42% II. Evolution of Portfolio Composition around the GFC Global Emerging Equity Funds Latin America 25% 24% 23% 22% 21% 20% 19% Average portfolio shares Sep.09 May.09 Jan.09 Sep.08 May.08 Jan.08 Sep.07 May.07 Jan.07 18% Sep.09 May.09 Jan.09 Sep.08 May.08 Jan.07 Sep.07 7% May.07 40% Sep.09 8% May.09 9% 42% Jan.09 10% 44% Sep.08 11% 46% May.08 12% 48% Jan.08 13% 50% Sep.07 14% 52% May.07 54% Jan.07 Emerging Europe 15% Jan.08 Emerging Asia 56% II. Evolution of Portfolio Composition around the GFC Global Bond Funds 25% 20% 15% 10% 5% Average portfolio shares Sep.09 May.09 Jan.09 Sep.08 May.08 Jan.08 Sep.07 May.07 Jan.07 0% Sep.09 May.09 Jan.09 Jan.07 North America 30% Sep.08 0% May.08 0% Sep.07 5% May.07 10% Sep.09 10% May.09 20% Jan.09 15% Sep.08 30% May.08 20% Jan.08 40% Sep.07 25% May.07 50% Jan.07 Emerging Countries 30% Jan.08 Developed Europe 60% II. Evolution of Portfolio Composition around the GFC Cash Weights – Global Funds Average portfolio shares Sep.09 May.09 Jan.09 Sep.09 May.09 Jan.09 Sep.08 May.08 Jan.08 Sep.07 May.07 0% Sep.08 1% May.08 2% Sep.07 3% May.07 4% Jan.07 5% Jan.07 Global Bond 20% 18% 16% 14% 12% 10% 8% 6% 4% 2% 0% Jan.08 Global Equity 6% Presentation I. Motivation II. Data and Summary Statistics III. Shocks to Managers and Portfolio Reallocations IV. Behavior of Investors and Managers V. Gross and Net Country Flows VI. Conclusions III. Variation in Assets: Decomposition of Asset Growth Aˆit rit f it Equity Funds Mean Growth Rate Returns of Assets Standard Deviation Variance Descomposition Injections/ Initial Assets Growth Rate of Assets Returns Injections/ Initial Assets All Equity Funds 2.20% 1.01% 1.15% 10.34% 47.24% 52.76% Global 1.59% 0.71% 0.88% 6.96% 54.69% 45.31% Global Emerging 2.85% 1.32% 1.46% 9.67% 49.57% 50.43% Bond Funds Mean Growth Rate Returns of Assets Standard Deviation Variance Descomposition Injections/ Initial Assets Growth Rate of Assets Returns Injections/ Initial Assets All Bond Funds 3.94% 0.69% 3.19% 8.66% 11.37% 88.63% Global 0.61% 0.31% 0.60% 7.39% 9.31% 90.69% Global Emerging 1.31% 0.43% 0.92% 10.54% 9.74% 90.26% Both injections and valuations matter III. Variance Decomposition (Tranquil vs. Crisis Times) Period All Equity Funds Global Global Emerging Period All Bond Funds All Equity Funds Global Global Emerging Equity Funds Global Financial Crisis Before Global Financial Crisis Narrow Window (Jan. 2003-Feb. 2007) (Mar. 2008-Dec. 2009) Injections/ Injections/ Returns Returns Initial Assets Initial Assets 36.74% 63.26% 67.01% 32.99% 37.06% 62.94% 65.40% 34.60% 33.54% 66.46% 70.15% 29.85% Global Financial Crisis Wide Window (Mar. 2007-Oct. 2010) Injections/ Returns Initial Assets 57.65% 42.35% 60.44% 39.56% 64.71% 35.29% Bond Funds Global Financial Crisis Narrow Window (Mar. 2008-Dec. 2009) Injections/ Returns Initial Assets 18.78% 81.22% 2.66% 97.34% 26.23% 73.77% Global Financial Crisis Wide Window (Mar. 2007-Oct. 2010) Injections/ Returns Initial Assets 11.82% 88.18% 4.45% 95.55% 20.59% 79.41% Before Global Financial Crisis (Jan. 2003-Feb. 2007) Injections/ Returns Initial Assets 12.36% 87.64% 5.18% 94.82% 12.90% 87.10% Not driven by a common time component III. Significant Variation in Country Weights: Coefficients of Variation CV j (w j ) wj Equity Funds Across Funds Within Target Region 1.57 Within Funds 0.07 NonTarget Region 0.66 0.61 0.05 0.41 0.72 Across Funds 0.66 0.04 1.56 0.85 Within Funds 0.53 0.03 0.68 0.93 Number of Funds Global 155 Global Emerging 187 Target Region Cash 1.15 Bond Funds Global 30 Global Emerging 81 Across Funds 2.09 0.19 0.56 1.22 Within Funds 0.66 0.08 0.22 0.46 Across Funds 1.23 0.14 1.35 1.78 Within Funds 0.44 0.05 0.35 1.21 Manager’s decision on how to allocate flows may play an important role Presentation I. Motivation II. Data and Summary Statistics III. Shocks to Managers and Portfolio Reallocations IV. Behavior of Investors and Managers V. Gross and Net Country Flows VI. Conclusions IV. Behavior of Investors: Injections to Equity Funds A. Equity Funds Variables Country Crisis Injections/Average Assets -0.048* -0.003 -0.009 -0.013 (0.014) (0.012) (0.010) (0.011) Global Crisis -0.018* -0.008† (0.001) (0.004) Lagged Fund Returns 0.161* 0.119* (0.024) Country of Origin Returns (0.023) 0.261* 0.222* 0.171* (0.033) 0.178* (0.039) 0.135* (0.024) (0.023) (0.028) Time Fixed Effects No No No No No Yes No Country of Origin-Time Fixed Effects No No No No No No Yes 41,232 41,232 40,492 39,479 38,764 38,764 40,492 0.035 0.016 0.036 0.017 0.047 0.028 0.050 0.031 0.065 0.046 0.114 0.092 0.174 0.090 No. of Observations R-squared Adj.R-sq Pro-cyclical and subject to wealth effects *=1%, †=5%, ~=10% IV. Behavior of Investors: Injections to Bond Funds B. Bond Funds Variables Country Crisis Injections/Average Assets -0.081* -0.070* -0.018 -0.031 (0.021) (0.018) (0.016) (0.023) Global Crisis -0.038* -0.028* (0.006) (0.008) Lagged Fund Returns 0.229† 0.205† (0.111) Country of Origin Returns (0.102) 0.464* 0.468* 0.126~ (0.070) 0.107 (0.067) 0.337* Time Fixed Effects No No No (0.148) No Country of Origin-Month Fixed Effects No No No No No No Yes No. of Observations 3,520 3,520 3,445 3,261 3,196 3,196 3,445 R-squared 0.061 0.065 0.073 0.068 0.092 0.156 0.266 Adj. R-sq 0.038 0.041 0.051 0.044 0.069 0.107 0.087 Even more than equity funds (0.127) No (0.121) Yes No *=1%, †=5%, ~=10% IV. Behavior of Managers: Framework Behavior of (log) weights ijt ijt 1 rjt rit Crisis jt ij it ijt Comes from log-linearization of identity wijt wijt1 ( Rijt f ijt ) ( Rit f it ) ijt ijt 1 (rjt rit ) ( fijt fit ) it ijt buy and hold Plus an implicit flow equation ( fijt fit ) ijt 1 rjt rit Crisis jt ij it ijt IV. Behavior of Managers: Framework ijt ijt 1 rjt rit Crisis jt ij it ijt α and β = 1: buy-and-hold strategy α and β ≠ 1: cyclicality of flows from managers to countries β < 1 counter-cyclical relative flows ϒ response of flows to a crisis (on top of what is captured by β) Test for persistence of weights and response to returns and crises Similar results if reallocations (relative to buy-and-hold) studied Econometric considerations discussed in the paper Dynamic panel, UR, Endogeneity IV. Behavior of Managers: Country Weights A. Equity Funds Log Lagged Weights Relative Returns Country Crisis Log Country Weights Monthly 0.986* 0.982* 0.983* 0.901* 0.901* 0.899 * (0.001) (0.001) (0.001) (0.002) (0.002) (0.002) 0.622* 0.647* 0.993* 0.959* 0.956* 0.598 * (0.051) (0.057) (0.013) (0.049) (0.013) (0.013) -0.020* (0.003) No Yes No No No No No Yes No No No No Fund Fixed Effects Date Fixed Effects Fund-Date Fixed No No Yes Effects Country of DestinyNo No No Fund Fixed Effects Log Lagged Weights=Relative Returns 0.000* 0.000* 0.446 No. of Observations 458,458 458,458 458,458 Semi Annual Annual 0.568* 0.307* (0.012) (0.026) 0.857* 0.567* (0.032) (0.035) -0.069* -0.118* (0.017) (0.026) No No No No No Yes Yes Yes Yes Yes Yes Yes Yes Yes 0.000* 0.000* 0.000* 458,458 458,458 458,458 0.000* 62,949 0.000* 26,018 *=1%, †=5%, ~=10% IV. Behavior of Managers: Country Weights (Bonds) B. Bond Funds Log Country Weights Monthly Semi Annual Annual Log Lagged Weights 0.974* 0.969* 0.970* 0.868* 0.866* 0.866* 0.448* 0.102~ (0.002) (0.003) (0.003) (0.008) (0.009) (0.009) (0.037) (0.059) Relative Returns 0.237* 0.238* 0.638* 0.219* 0.608* 0.611* 0.296* 0.310* (0.091) (0.091) (0.079) (0.084) (0.073) (0.073) (0.101) (0.100) Country Crisis -0.016 -0.017 -0.026 (0.011) (0.050) (0.084) Fund Fixed Effects No Yes No No No No No No Date Fixed Effects No Yes No No No No No No Fund-Date Fixed Effects No No Yes No Yes Yes Yes Yes Country of DestinyNo No No Yes Yes Yes Yes Yes Fund Fixed Effects Log Lagged Weights=Relative Returns 0.000* 0.000* 0.000* 0.000* 0.000* 0.000* 0.127 0.023† No. of Observations 39,183 39,183 39,183 39,183 39,183 39,183 5,035 1,959 *=1%, †=5%, ~=10% Relative flows less pro-cyclical than in equity: contagion, precautionary savings? IV. Behavior of Managers: Cash Weights A. Equity Funds Log Cash Weights Variables Log Lagged Weights Monthly 0.587* (0.006) Relative Returns (-Fund return) 0.729* (0.083) 0.389* (0.008) 0.700* (0.102) 0.360* (0.008) 0.169~ (0.088) Country Crisis 0.377* (0.009) 0.494* (0.099) 0.096~ (0.051) Global Crisis 0.158* Semi Annual Annual 0.112* -0.083 (0.024) 0.188* (0.071) 0.116 (0.158) 0.116† (0.050) -0.181 (0.138) 0.498~ (0.284) 0.111 (0.018) (0.049) (0.101) Origin Returns -0.168 -0.437* -0.034 Fund Fixed Effects No Yes Yes (0.116) Yes (0.097) Yes (0.119) Yes Time Fixed Effects No No Yes No No No 0.087~ 0.002* 0.029† 0.237 0.288 0.479 Log Lagged Weights=Relative Returns *=1%, †=5%, ~=10% Counter-cyclical cash positions IV. Behavior of Managers: Cash Weights (Bonds) B. Bond Funds Log Cash Weights Variables Log Lagged Weights Relative Returns Country Crisis Global Crisis Origin Returns Fund Fixed Effects Time Fixed Effects Log Lagged Weights=Relative Returns Monthly 0.654* 0.449* 0.446* 0.433* (0.022) (0.029) (0.029) (0.030) -0.459~ -0.422 -0.682 -0.381 (0.264) (0.303) (0.456) (0.298) -0.537* (0.172) -0.028 (0.047) 0.261 (0.520) No Yes Yes Yes No No Yes No 0.000 * 0.004* 0.015 † 0.007* Semi Annual 0.119 (0.078) 0.166 (0.257) -1.175~ (0.670) -0.039 (0.138) 0.991 (0.949) Yes No 0.867 Annual -0.380† (0.176) 0.510~ (0.295) -1.923~ (1.057) 0.371~ (0.186) -0.362 (0.930) Yes No 0.010 * *=1%, †=5%, ~=10% Pro-cyclical cash positions: precautionary savings? (levels also larger) Presentation I. Motivation II. Data and Summary Statistics III. Shocks to Managers and Portfolio Reallocations IV. Behavior of Investors and Managers V. Gross and Net Country Flows VI. Conclusions V. Gross and Net Country Capital Flows: Two Measures Gross Flows (includes valuation effects) Aˆ jt sijt1 wˆ ijt sijt1 Aˆ it i i Growth rate of weights Growth rate of fund assets Net Flows f jt sijt1 wˆ ijt rjt rit sijt1 f it i i Return-adjusted growth rate of weights Injections V. Net Flows Shares (% of Country Growth Rate) Return-Adjusted Growth Rate Injections of Weights All Countries Type Active Passive Frequency Monthly SemiAnnual Annual Variance Decomposition (% of Variance of Country Growth Rate) Return-Adjusted Growth Rate Injections of Weights 88.4% 11.6% 84.8% 15.2% 87.4% 15.0% 12.6% 85.0% 86.8% 30.9% 13.2% 69.1% 88.4% 11.6% 84.8% 15.2% 83.3% 16.7% 78.9% 21.1% 80.6% 19.4% 73.0% 27.0% Manager’s behavior explains most of MF net capital flows to countries Larger for active funds V. Gross and Net Country Flows: Quantitative Effects 10% decline in lagged fund returns reduces injections in 1 pp If all funds investing in a country experience such decline, gross flows will decline in 1 pp This is close to the median gross flow across countries (2%) 10% decline in country of origin returns reduces injections in 2 pp 10% decline in relative returns (holding fund returns constant) induces a similar decline in gross flows A country crisis leads to a 2% decline in gross flows 10% decline in relative returns yields a 1 pp decline in relative flows Similar to the unweighted average growth in net flows in the sample (-1.5%) • If this is accompanied by a low fund performance or low returns in the country of origin, that can induce large redemptions (4 pp decline) V. Gross and Net Country Flows: Quantitative Effects For a shock to injections to have no effect on a country’s net flows we need relative flows to compensate in the same amount Only countries that are doing relatively well, would not be seriously affected by shocks to the injections by underlying investors Even in this case, contagion may be an important source of capital flows Presentation I. Motivation II. Data and Summary Statistics III. Shocks to Managers and Portfolio Reallocations IV. Behavior of Investors and Managers V. Gross and Net Country Flows VI. Conclusions VI. Conclusions: Main Results MF assets fluctuate substantially over time, pro-cyclically Particularly pronounced effects during crises Large reallocations during global crisis, consistent with retrenchment Both investors and managers behind these movements, changing their investments substantially over time … and shaping fluctuations in capital flows Neither managers nor investors exploit potential arbitrage opportunities by being contrarian during crises No stabilizing role: amplify crises & transmit shocks x countries Important policy lesson: Runs even among equity-type investors VI. Conclusions: Main Results Not the case that bad shocks at home country propel more investments abroad, to the contrary Underlying investors do not act either as deep-pocket international investors buying assets abroad at fire sale prices Their behavior exerts pressure on managers, who need to react to these shocks Evidence not consistent with constant country weights, which change substantially over time Managers move away from countries experiencing crises Cash positions actively used, differently for equity and bonds Thank you!