Monitoring International Borrowers The IMF Role in Bank

Monitoring International Borrowers: The IMF Role in Bank and Bond Markets By Eichengreen, Kletzer, and Mody Discussant Athanasios Vamvakidis

Empirical question • Impact of IMF programs on market access and cost of funds • 6, 700 loan transactions between emerging market borrowers and banks, 3, 500 new bond issues • 1991 -2002

• Banks act as delegated monitors on behalf of depositors • Dispersed bondholders lack the capacity or/and the incentives to incur the large costs of monitoring • Therefore, the IMF’s public monitoring role should be more valuable to bondholders • The IMF could signal valuable information to the private sector • The IMF program can be a commitment mechanism for good policy plans/intensions

Main Results • IMF programs reduce spreads more in markets dominated by bonds than in those dominated by bank loans • This result is stronger for countries with intermediate levels of external debt/GDP ratios • This result is stronger for programs that turn precautionary

Discussion • The theoretical framework discussed in the paper helps put the empirical tests in perspective • The empirical results are consistent with theoretical predictions

• Global variables: US industrial growth (results are consistent with Arora and Vamvakidis (2004 a), log of swap rate, EMBI volatility How about: Growth in trading partners (Arora and Vamvakidis (2004 b), Regional growth Sector shocks

• Increased IMF transparency after 1997 Most program information/country reports, on policies, projections, and risks, are now public information Emphasis on transparency by the IMF for its members (Fiscal and data ROSC, FSAP and other similar facilities) Has the ability of an IMF program to provide a positive signal increased as a result of more transparency? Results for before and after 1997 Results for programs that publish the LOI and staff reports versus programs that don’t

• IMF programs (positive signals): All IMF programs Precautionary Programs turned precautionary Negative program related signals? Program negotiations that failed, indicating the need for a program, but the absence of good policy plans, of IMF monitoring, and of policy commitment (can use staff reports to get the data) Precautionary arrangements that turn non-precautionary Programs that collapsed: IMF pulling the plug

• Causality The authors claim that the use of data on transactions reduces causality concerns. However, the paper needs to do more to explain this claim and address remaining causality concerns

• The paper tests if the results differ by issuer (bank lending versus bond issues), by the level of external debt/GDP ratio, and by the size of IMF lending Other possible dimensions: Region (both for lenders and borrowers) Level of GDP per capita History of crises Openness of capital markets Level of development of the financial sector