External volatility from sharp movements in terms of trade and reversals in gross financial flows is a pressing concern for developing and emerging countries, and has been a key driver of aggregate fluctuations in the recent financial crisis. Unfettered, it disturbs exchange rates, interest rates and domestic economic activity, generates boom-bust cycles in commodities and induces inflationary and deflationary pressures which in turn make monetary and macroeconomic management extremely difficult. What should developing countries do to protect themselves? We study external insurance from the perspective of a global equilibrium model, exploring the conditions under which self-insurance by developing countries can be self-defeating in the aggregate. This can occur for two reasons. First, precautionary saving can depress world interest rates sufficiently to trigger speculation on commodities. This increases external volatility and accentuates the precautionary saving motive. Second, low world interest rates can have additional adverse macroeconomic effects by increasing financial fragility in developed economies, increasing external volatility in developing countries. The second strand of our project concerns the public component of the precautionary saving: the accumulation of international reserves. Developing countries accumulated large amounts of reserves, reflecting their desire to mitigate the consequence of sharp capital flow reversals, sudden stops, or internal fiscal crises. These reserves are mostly held in the form of low-yield, non-contingent U.S. assets, Treasuries, which makes this form of self-insurance costly. We argue that this can be rationalised in a world of severely incomplete markets where country-tailored macroeconomic insurance is unavailable or too expensive. We develop the analysis by distinguishing between country-specific and global crises. In the case of country-specific crises, self-insurance arrangements lead to over-accumulation of reserves and low yields on dollar assets which increase the cost of self-insurance. These arrangements are dominated by pooling schemes where developing countries would issue contingent drawing rights on a common pool of international reserves. We study potentially better institutional arrangements where developing countries pay an insurance premium to a Lender of Last Resort (LLR) in tranquil times, in exchange for a liquidity payout from the LLR in times of crisis. Such a system would relieve the downward pressures on U.S. interest rates and tame the incentives for leverage. If the purpose is to insure against global crises, the analysis is different. The dollar appreciated during the worst phase of the crisis, as investors retreated to the safe-haven of U.S. Treasuries. This makes U.S. Treasuries attractive insurance tools against global crises. But one might legitimately be concerned that if many countries liquidate their dollar assets at the same time, the value of the dollar would plummet, creating a form of dollar systemic risk. We describe some international cooperation mechanisms to take care of this problem ex-ante. These two theoretical projects will provide a framework with which ICG teams will be able to advise developing countries on the advisability of their self-insurance policy and on the potential benefits of alternative policies negotiated at a global level.