Portfolio Theory, Transaction Costs, and the Demand for Time Deposits
Gess, Alan C., Journal of Money, Credit & Banking
1. THE ISSUE
MANY BANKS REVIEW WEEKLY the interest rates they pay on their small time deposits. If the U.S. Treasury yield curve changes significantly, banks change their deposit rates to reflect the new market interest rates. Large rate changes occurred during 1991 and 1992 when the decline in the Treasury bill rate induced banks to reduce the spread between rates on flexible and fixed rate deposits by 200 to 300 basis points. In spite of these large changes in relative rates, few households rebalanced their deposit portfolios. This paper addresses two questions about this absence of deposit-rebalancing trades. First, why did households not rebalance their holdings of small time deposits when relative rates of return changed? Second, how much potential interest did households forego by not rebalancing? Households might not rebalance because they view different types of deposits as being poor substitutes and even large rate changes are too small to change optimal holdings. This is an optimal-portfolio explanation. Alternatively, they may not rebalance because the costs of rebalancing eliminate the gains from rebalancing. This is a transaction-cost explanation.
The theory of banking posits that through economies of specialization banks reduce intermediation costs by issuing large-denomination loans with maturities that best serve borrowers, and small-denomination time deposits with maturities that best serve savers (Benston and Smith 1976). While banks reduce transaction costs, they cannot drive them to zero. What effects do transaction costs have on depositors? Are they significant or insignificant? Are depositors' portfolios far from or close to the efficient frontier? Are transaction costs sufficiently high that depositors seldom rebalance? Or, are they so small that depositors continually rebalance? In short, how close is the small time-deposit market to the perfect market ideal of zero transaction costs?
The main news of this paper is that it is not transaction costs that keep depositors from rebalancing. Instead, depositors did not rebalance because there were insignificant gains from doing so. This is because the banks changed relative rates among deposits that are poor substitutes by large amounts, and among deposits that are close substitutes by much smaller amounts. Some deposits are such close substitutes that optimal rebalancings among them involve large trades even in response to small changes in relative yields. Optimal trades often exceed the average balances in the accounts. But, here is the rub: The gain from rebalancing is not different from zero since the households are essentially indifferent to their holdings of deposits that are close substitutes. Optimal-rebalancing trading volume among close substitutes is very large, but the gain is not different from zero. Optimal-rebalancing trades among poor substitutes are essentially zero. In short, households did not rebalance because it was not optimal for them to do so even if transaction costs were virtually zero.
2. PREVIOUS RESEARCH USED IN THIS STUDY
A critical feature of this research is that it disentangles the effects of asset substitution and transaction costs by applying portfolio theory to the demand for interest paying deposits. Banks offer depositors a wide array of deposits: passbook accounts, NOW accounts, money market accounts, and certificates of deposits of different maturities and yields. Because the yields on these accounts are not perfectly positively correlated, depositors can increase their risk-adjusted expected returns by holding a portfolio of deposits instead of a single deposit. That is, depositors can apply portfolio theory to their deposit holdings. To make such a risk-diversification strategy effective, the household must determine the efficient frontier, select the amount of risk it is willing to bear, choose the initial portfolio, and reallocate money among assets as the efficient frontier changes in response to changes in relative expected rates of return. …