The Theory of Reserve Accumulation, Revisited

Uncertainty about a government willingness to repay its outstanding liabilities upon auctioning new debt creates vulnerability to belief-driven hikes in borrowing costs. We show that optimizing policymakers will eliminate such vulnerability by accumulating reserves up to ensuring post-auction debt repayment in all (off-equilibrium) circumstances. The model helps explaining why governments hold significant amounts of reserves and appear reluctant to use them to smooth fundamental shocks. Quantitatively, the model explains reserve holdings up to 3% of GDP if debt is short term, 2.4% with long-term debt—as long bond maturities mitigate vulnerability to belief-driven crises.