Basic Analytics of IMF Lending and Surveillance ∗ Rune Jansen Hagen † August 2002 Abstract I analyse whether multilateral lending may be justified in a world of global capital markets if multilaterals have an informational advantage relative to lenders in the market for sovereign debt. I show that the adverse selection problem in this market may be solved even in the absence of lending provided the multilateral agency does not care too much about borrower country welfare. However, when lending is unconstrained the private information of the multilateral will be transferred to lenders no matter the relative weighting of welfare and lenders’ profits. In contrast, multilateral lending may not lead to lenders being able to distinguish good from bad borrowers if loan size is restricted and may in fact worsen the problem compared to a situation where the agency plays a purely informational role. 1 Introduction Private international capital markets have grown rapidly over the last 50 years, and many developing countries now have access to foreign for-profits funds in one form or another. Loans from private banks, which were the major component of commercial flows to developing countries before the debt crisis of the 1980s, are now supplemented by portfolio flows and foreign direct investment, at least in some middle-income countries. 1 The changing character of international financial markets has created a debate over the raison d’être of multilateral institutions such as the IMF and the World Bank. Given that the resources at their disposal are now dwarfed by the size of private capital markets, do they still have a role to play in international lending? Or should they instead focus on other tasks, such as distributing aid? The recurrent financial crises over the last few decades demonstrate that despite their phenomenal rate of de- velopment, private capital markets still fail, and sometimes spectacularly so. However, as economists customarily point out nowadays, the existence of market failure is not a sufficient condition for intervention since public agen- cies also fail. One must therefore demonstrate that public institutions can reasonably be expected to outperform ∗ Subject to the usual discalaimer, I wish to thank Karl Rolf Pedersen and Gaute Torsvik for commenting on previous drafts. The research reported here has been financed by the Research Council of Norway. † Department of Economics, Norwegian School of Economics and Business Administration, Helleveien 30, 5045 Bergen, Norway. E-mail: rune.hagen@nhh.no. 1 The poorest countries, in particular those in Africa, are still almost completely dependent upon public concessional flows. 1