Academic journal article Journal of Money, Credit & Banking

How Well Do Mexican Banks Manage Their Reserves?

Academic journal article Journal of Money, Credit & Banking

How Well Do Mexican Banks Manage Their Reserves?

Article excerpt

BANKS PLAY A major role in facilitating the way the financial sector operates. The efficiency with which banks utilize reserves is important for understanding not only operational banking behavior but also for the role that banks play in the transmission mechanism. In this paper, we analyze banks' reserve operations, and quantify the extent to which their reserve management policy is optimal--in a sense of minimizing expected reserve costs.

The reserve management problem is interesting because of the inherent asymmetry in costs incurred by the foregone interest when the bank meets its reserve requirement and the alternative, more expensive, case in which the bank incurs a penalty cost for having reserves fall below the reserve requirement. A key ingredient in assessing the probability that the reserves of a bank will fall short of its reserve requirement is the stochastic process for the bank's deposits and withdrawals. Given this process and some other relevant state variables, the bank's optimal decision can be deduced. The uncertainty about the resulting reserve position at the beginning of the next day is due to the fact that certain operations of withdrawals and deposits take place after the bank sets its overnight reserve position. Therefore, much of our empirical effort is aimed at estimating this stochastic process. Given our estimate of this process we derive the predicted optimal reserve positions. Based on the latter, the actual reserve costs are compared to the estimated costs had the banks set their reserve positions according to the model. This comparison reveals the degree to which banks follow an optimal decision process for managing their overnight reserves and the costs for not doing so.

The panel data set and the environment we study are unique. The data set contains daily observations on the reserve positions as well as other components of the balance sheet for all banks in Mexico during the years 1990-91. This data set allows us to estimate the stochastic process for deposits and withdrawals in a relatively precise way. During this period, the Mexican banking system had a very simple structure--a feature that provides a laboratory environment for our research question. In contrast to the U.S. system, the central bank in Mexico essentially served as a payment clearing house; therefore, there were no official reserve requirements, there were no significant reputation issues such as excessive use of the discount window, nor were there any complications arising from Wednesday settlements as they were done every night. Moreover, the panel structure of the data allows us to address not only the reserve behavior over time but also systematic differences in behavior across banks. Finally, reserve management is quantitatively a significant issue. For example, during 1990 the annual reserve costs of the Mexican banking industry (equivalent to 180 million dollars) were about 10% of the industry's total profit. Thus, deviations from optimal reserve positions could result in large reductions in operating profits, and consequently have an important quantitative effect on banks' overall profitability.

We find that a good forecast of the overnight withdrawals--the key variable generating overnight payment uncertainty--is based on a few types of intra-day operations. The forecasts of the overnight withdrawals predict the reserve behavior undertaken by most banks quite well. Overall, the model's reserve prediction seem to accord relatively well with observed reserve behavior. Specifically, for nine out of the 19 banks in our sample, we find insignificant differences between the model's predicted reserve behavior and the reserves actually used by the banks. For some banks, the model's predicted reserve position results in larger costs, although those are not statistically significant. These banks engage in many activities for which the bank is likely to have a better information set than the econometrician. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.