Should central banks provide reserves via repos or outright bond purchases?

In the wake of the financial crisis banks are likely to wish to hold far more highly liquid assets than before. Some of those liquid assets are likely to be held in the form of reserves at the central bank. We ask whether the central bank should provide these reserves by purchasing nominal, fixed-rate government bonds outright, or by repo-ing them in for a limited period. The key difference between these options is that with repos, the private sector retains the price risk associated with bonds, whereas this risk rests with the central bank if it purchases these bonds outright. There is a significant, practical policy issue for central banks here: should those central banks (most notably the Fed and the Bank of England) who built up a large stock of bonds during the QE operations, which were financed by creating reserves for commercial banks, expect to sell those bonds in due course or continue to hold a high proportion of them for a long period since the demand for reserves will be permanently higher? We develop and calibrate a simple OLG model in which risk-averse households hold money and bonds to insure against risk. We find that the composition of the central bank's assets should depend on how fiscal policy is conducted; but in general it has only a small impact on welfare.