How Well Do Banks Manage Their Reserves?

In this paper we investigate how well banks manage their reserves. The optimal policy takes into account expected foregone interest on excess reserves and penalty costs for going below required reserves. Using a unique panel data-set on daily clearing house settlements of a cross-section of Mexican banks we estimate the deposit uncertainty banks face, and in turn their optimal reserve behavior. The most important variables for forecasting the deposit uncertainty are the interbank fund-transfers of the day, certain calendar dates, and the interest differential between the money market rate and the discount rate - a measure reflecting the bank's opportunity cost of money holdings. For most banks the model's prediction accord relatively well with the observed reserve behavior of banks. The model produces reserves costs that are significantly smaller relative to the case when reserves are set via simple rule of thumb. Furthermore, alternative motives for holding reserves (such as liquidity and reputation effects) do not seem to be the explanation for why certain banks hold relatively large reserves.