Printer Friendly

Bank deposit rate deregulation and customer service levels.

Bank Deposit Rate Deregulation and Customer Service Levels

The commercial banking industry may be subject to inherent instability problems that could lead to significant disruptions in consumer transaction activity. As a result, banking is one of the more regulated industries in the United States. The services provided are considered essential, and significant concern exists about the consumer's ability to have adequate access to low-priced banking services without regulatory oversight. While the recent legislative trend is toward relaxing some of these restrictions, significant disagreement still exists over the resulting impact on consumers (Gross 1987; Kutler 1987; Thomas 1988).

The banking industry is unique because the degree of regulation aimed at protecting consumers varies significantly by state, e.g., state-determined entry and expansion restrictions and allowable organization structure. As a result, the industry is a useful testing ground in which to measure the impact of policy-induced competitive restrictions on consumer welfare. One result of these restrictions has been to increase the use of alternative forms of nonprice competition (White 1976; Lloyd-Davies 1977; Sealey 1979; Rogowski 1984; Hanweck and Rhoades 1984) and elimination of the restrictions may significantly alter industry behavior (Weisbrod 1980; Petersen 1981; Berger and Hannan 1987). For example, it has been argued that the recent elimination of bank deposit rate ceilings will result in a significant change in pricing procedures for all bank services. Significant societal effects could result as consumers are serviced by firms more accountable to the forces of the marketplace. At issue is whether or not customers will lose some of the benefits resulting from nonprice competition.

In this paper the impact of deposit rate deregulation on the influence of regulated bank entry barriers on customer service levels is evaluated. Specifically, liberal entry has been shown to result in preferred service levels to bank customers--measured as the absolute number, per capita number, or proximity of banking facilities to the consumer (Savage and Humphrey 1979; Seaver and Fraser 1979; Evanoff 1988). However, this improved service may have resulted from the existence of interest rate ceilings on deposits that have caused banks to overinvest in real physical capital. That is, without the ability to compete directly by price, a form of nonprice competition was employed in which more offices provided improved service accessibility or convenience. As deregulation occurs, banks may change pricing procedures to eliminate previously subsidized prices and to maintain profit levels in reaction to rising, unregulated interest payments. This may result in a decrease in the number of bank facilities as those opened as a nonprice means of competing may not be viable once more direct price competition is allowed. Thus, competitive deposit rates may be realized only at the expense of decreases in customer service.

A model is developed to explain the number of bank offices per service area, a service accessibility measure, and a variant of that model is used to show empirically how the service level is influenced by entry restrictions, i.e., branching restrictions. Once developed, estimates are provided for periods both prior to and after rate deregulation. If the benefits from liberal entry result mainly from nonprice competition, one would expect to see a decline in service levels in recent years due to the elimination of the need to compete on a nonprice basis. Thus, we would expect to see a shift in the relationship between the service level and its determining factors. If branching continues to lead to preferred service levels, the positive impact of branching is strongly reinforced and concerns about potential declines in customer service may be unwarranted. Our results suggest that the social benefits from liberal entry are significant, are not excessively influenced by nonprice competition, and, therefore, remain after interest rate deregulation. Thus, the social benefits of deposit price competition have not been achieved at the expense of excessive decreases in customer convenience.

The paper is organized as follows. First we briefly review the literature on the pertinent issues and then develop a theoretical model explaining one form of customer banking service--ease of accessibility. We then relax some of the assumptions of the theoretical model and empirically estimate the influence of branching laws and reaffirm previous research findings. The stability of that relationship in the period after interest rate deregulation is then evaluated with particular interest given to the role of branch restrictions. The final section summarizes the findings and offers policy recommendations.


Interest rate ceilings on commercial bank deposit services were imposed during the 1930s and regulated by the Federal Reserve System. While the degree of restrictiveness varied over time, the restrictions became binding during the 1970s as inflation pushed market rates well above allowable deposit rates. Rational behavior by bankers, aimed at avoiding the restrictions, resulted in financial innovations and the use of alternative means to compete for deposit liabilities. Alternative forms of nonprice competition developed, including the provision of free checking services, various gifts aimed at enticing new account holders, and improved customer convenience via additional bank offices.(1) The significant increase in the number of U.S. banks and bank offices during the 1970s is generally offered as evidence of this behavior (Berger and Hannan 1987; Lloyd-Davies 1977; Petersen 1981; Weisbrod 1980).

(1)The number of bank offices is not intended to be considered a comprehensive measure of service level or service accessibility. Obviously alternative means to access services are available. However, it is one important aspect of service and has been analyzed as such in previous studies of banking.

In 1980 the Depository Institution Deregulation and Monetary Control Act was introduced and allowed for the gradual phase-out of deposit interest rate ceilings starting with the introduction of interest-bearing NOW accounts in 1981. A major objective of this act was to allow the forces of the marketplace to be more operative in determining price and service levels. Specifically, it assigned the phase-out of explicit deposit rate restrictions to an interagency committee. The Garn-St. Germain Act of 1982 continued this trend by allowing Money Market Deposit Accounts, and by 1984 only minimum deposit requirements remained. While the removal of interest regulations obviously will not eliminate the use of nonprice competition, it may decrease its importance somewhat because more direct means of competing for customers will exist. In isolation, allowing banks to pay higher deposit rates should benefit consumers. The pertinent question is whether this gain may be partially or totally offset by other changes induced by rate deregulation.

One of the frequently mentioned benefits of allowing bank expansion by branching is the resulting improvement in customer service. Recent empirical studies reinforce this belief by finding that the number of bank offices--a measure of service availability--is larger when branching is allowed (Savage and Humphrey 1979; Seaver and Fraser 1979; Evanoff 1988). However, the studies analyze service levels prior to 1981, and deposit rate deregulation since that period may result in the closing of existing offices or a decrease in the introduction of new facilities.(2) This would occur if offices had been introduced solely as an alternative means to compete in a price-regulated environment. Once more direct price competition was allowed, "brick and mortar" competition would no longer be required. In this case, one of the major benefits for consumers frequently associated with branching may have been overstated. Conversely, if branching continues to lead to preferred customer service levels after rate restrictions are eased, the positive social impact is strongly reinforced. With restrictions on price competition eliminated, the number of offices present during the 1980s should be more indicative of true market forces.

(2)A substantial amount has been written recently on the possibility of numerous branches being closed (Rose 1988; Thomas 1988).


Customers choose their banks based on the total costs and benefits received. Thus explicit deposit, loan, and other service prices enter the decision process, as do implicit expenses. Convenience enters because it is a component of cost and is indirectly related to the number of offices available. As the number increases, the transportation cost declines, as does the "total cost" of obtaining services. Therefore, more services can be utilized. Actually, a branch bank has an additional beneficial characteristic, regardless of the number of offices in the service area, because the network of offices allows access at various sites. Evidence suggests this added flexibility is desired by consumers (Hanley et al. 1985). However, the following analysis ignores this additional benefit.

To model the impact of branching on service accessibility, assume that the equilibrium number of banking offices is determined by the demand for services and the cost required to provide them. Assume further that, in the absence of regulatory differences, cost conditions do not vary across bank service areas (Lanzilloti and Saving 1969). Factors influencing the level of accessibility, therefore, include those impacting service demand and the degree of regulatory stringency imposed by banking authorities. Thus

A = f(Q sub d, R), (1) where A is the availability of banking services, Q sub d is the demand for services, and R is the degree of regulatory stringency.

The demand for banking services, Q sub d, is a function of particular economic and demographic factors affecting the region being evaluated. For simplicity, assume that (1) the population is uniformly distributed within the geographic unit of observation, (2) income is uniformly distributed among the population within the geographic units, (3) transportation costs are proportional to distance, (4) bank output is measured as a single index Q, and (5) competition prevails, yielding equal prices and quantities of output at each banking office. Thus, consumers maximize utility by purchasing banking services at the nearest facility. The shorter the distance traveled to obtain services, the better the service. Assume that the cost of providing services at an office can be described by

C = a - bQ + cQ raise to 2, (2) where a, b, and c are parameters. The cost function implies some indivisibility in the provision of banking offices and a typical "u" shaped average cost curve. Under competitive conditions, in the long run, offices will be established until the price of services, P sub e, equals average cost:

P sub e = AC = a/Q - b + cQ. (3)

Thus, the cost-minimizing output is dAC over dQ is equal to 0, or

Q* is equal to square root of a/c. (4)

Because the quantity of banking services is assumed proportional to customer income, the quantity of services demanded within the service area of each office is

Q = kYDR, (5) where k is a constant, Y is per capita income, D is population density, and R is the area of each banking office's service area. Thus, substituting from the above equations, square root of a/c is equal to kYDR.

R is equal to square root of a/kYD square root of c, (6) which is the optimal size of each office's service area. Given that customers desire to minimize transportation costs, the optimal number of offices, N, is given by

N = G/R (7) where G = the are of the geographic unit of observation. Collecting terms,
N = GkYD square root of c/ square root of a, and (8)
N/G = A = kYD square root of c/ square root of a, (9)

where A is a measure of service availability--offices per area--thus,

A is equal to infinity DY, (10) where infinity is equal to k square root of c/ square root of a and k, c, and a are constants.(3) Therefore, in the absence of regulation and given competitive conditions and the initial assumptions, the number of banking offices per geographic area is a multiplicative function of the population distribution, D, and per capita income, Y.

(3)The standard, although heroic, assumption of homogenous demand density within the geographic unit of observation significantly simplifies the analysis. It results in the proportionality of transportation cost assumption influencing only the consumer's choice of banking facility--the closest--and not the quantity demanded by each consumer. Dropping this assumption, the transportation cost would be incorporated by the consumer to arrive at a total price, resulting in fewer services being demanded by customers located further away from the facility. This would cause the optimal number of equal-sized institutions providing the service to be smaller. More rigorously, in determining P sub e both production and transportation cost should be included in deriving the total cost and optimal output level, Q*, associated with minimum "total" average cost. This occurs where the absolute value of the slope of the two average cost curves (production and transportation) are equal. If average transportation cost is an increasing function of output used (a realistic assumption) then the minimum of "total" cost will be at a smaller output level than minimum production cost. However, this would not impact the direction of influence of the variables discussed above. While the assumptions are obviously abstractions, we prefer to make them explicitly to allow for testable hypotheses.

To allow for empirical testing of the influence of deposit rate deregulation, adjustments to the model are required because the initial assumptions are abstractions that obviously oversimplify the true situation. Uniform income and population densities do not exist, although the smaller the unit of observation, the closer the assumptions approximately reality. Similarly, transportation costs are not proportional in all directions, and competition is unlikely to eliminate all differences in service prices and qualities. Thus, differences among geographic areas should be allowed. We also need to account for the pertinent issue being evaluated: regulation. Branching restrictions increase the costs of servicing customers if there are economies to branching. A stringent charter approval process also curtails the number of offices.

Thus, accounting for service area demographic and regulatory differences, the level of service accessibility is

A = k(P,D,Y,S,B), (11) where P is population, S is a measure of charter stringency, and B is a binary variable accounting for the ability to enter markets via branching. The remaining variables are as previously defined, and k sub P, k sub Y, k sub S, k sub B > 0, k sub D >< 0.(4) To test the hypotheses discussed earlier concerning the impact of branching on accessibility, assume that equation (11) is log linear and estimate A = 1n infinity sub 0 + infinity sub 1 1n B + infinity sub 2 1n D + infinity sub 3 1n Y + infinity sub 4 1n P + infinity sub 5 1n S.


(4)Coefficients on regulatory stringency measures are expected to be positive because of the manner in which they are defined, i.e., S is the approval rate, and B is one if branching is allowed, zero otherwise. The sign on population distribution is indeterminate because two opposing forces are operative. A more compact population will cause transportation costs to obtain services to be lower, perhaps, increasing the demand for services. However, a more compact population may be serviced by fewer institutions if economies of distribution exist. Previous research has generally found a negative impact.


The data used in the analysis are from four primary sources. Demographic data are from the City-County Data Files and County Business Patterns of the Bureau of the Census. Financial data are from the FDIC's Summary of Deposits. State charter approval rates are from published records of the Conference of State Bank Supervisors. Branching status data are from Amel and Keane (1986). A banking office is defined as a bank or branch at which both deposit and loan services are available.(5) County-level data were collected for the 48 contiguous states for the period 1981-1985. Use of county-level observations is considered a significant improvement over the use of data for entire states because the problems posed by variation in service area characteristics that are common to past studies should be reduced (Jacobs 1965; Lanzillotti and Saving 1969; Savage and Humphrey 1979). Analysis of still smaller units of observation is precluded by data limitations.

(5)Data adjustments were required because of the format in which census and FDIC data are presented. Independent city data for St. Louis and Baltimore were combined with the appropriate county. Virginia counties were excluded because data were not provided at the county level. Since 1980 is the last year in which detailed market level demographic data are available, estimates for the years analyzed were obtained using 1980 Census demographic data--income and population--indexed to the growth rates of County Business Pattern (CBP) payroll and employee figures. They were then deflated. Although the reported CBP figures are incomplete because of the exclusion of particular industry sectors, the included sectors are consistent across years and are assumed to proxy growth trends in each market. The percentage of the population residing in urban areas was assumed to be constant over the period analyzed. Current year (updated) data were available for the remaining variables. Similar results were found holding all demographic characteristics constant over the five-year period.

The dependent variable is measured as the number of banking offices per square mile. This is considered a preferred measure of service level, compared with the absolute or per capita number of offices, because it accounts better for the characteristics of service availability, i.e., time, distance, and cost involved in obtaining service (Mote 1974; Evanoff 1988). Explanatory variables accounting for demand and cost factors include per capita income and the percent of the population in urban areas (included to capture variation in population distribution within the unit of observation). These variables are preferable to population per square mile because the population distribution within the service area is more accurately portrayed. The state charter approval rate and state branching status binary are included to account for entry barriers.


A rather straightforward means is used to determine whether the influence of entry barriers (specifically branching) on customer service levels is as expected and whether it changed over the period when interest rates were deregulated. Equation (12) is estimated via ordinary least squares for each year over the period considered. These estimates are presented in Table 1.(6)

(6)Alternative functional forms did not prove superior in terms of t-statistics or F-statistics. To test for specification error more formally (including omitted variables), the RESET test was performed (Ramsey 1969), and the null hypothesis of no specification error could not be rejected for any year. The assumption of homoskedastic error terms could not be rejected using the nonconstructive parametric test developed by Goldfeld and Quandt (1972). Finally, casual inspection and more sophisticated tests suggested by Belsley, Kuh, and Welsch (1980) suggested that the results were not driven by influential observations and were not significantly impacted by multicollinearity. The branching impact over the period analyzed is similar to that found in previous years (Evanoff 1988).

The results for 1981 are generally consistent with expectations. The demographic factors are all significant at the five percent level and have the expected signs. The population density measure enters negatively, suggesting that some economies exist in servicing customers. Of the two regulatory variables, the charter approval rate does not have the expected positive influence on accessibility. The reasons for the insignificance are not obvious but may result from measurement problems.(7) The branching binary is highly significant, suggesting that branch restrictions do indeed reduce the number of bank offices per square mile. These findings tend to reaffirm the results of previous research and indicate that bank customers benefit from liberal branching laws.

(7)There is no reason to expect the degree of regulatory stringency in approving charters at the Federal level to differ across political boundaries. However, state differences will exist. S, the approval rate average over the three previous years, is used to capture the regulatory "mood" within the state, and there may be some concern about the consistency of S across borders. Perhaps, after a learning period, applicants in some states become aware of the requirements of the state banking authorities and file for charters only when nearly certain of acceptance. This could bias the approval rate upward and, if this level of knowledge is not constant across states, could cause it to be a poor measure of regulatory stringency. This may explain why the variable is insignificant in a number of years. The reported regressions were reestimated without the S variable, and the impact was very marginal. No deterioration of significance level, sign change, or decline in explanatory power occurred. It has also been suggested that, since the multibank holding company (MBHC) is an alternative means of expansion, it may also impact the office per area measure. While it may allow an organization to expand geographically, there is little reason to expect it to impact the number of facilities positively. MBHCs generally expand by merger or acquisition resulting in a change in ownership, but no change in the number of facilities. To the extent that MBHCs have expertise in obtaining new bank charters, they may open more new banks. However, the impact of this is probably marginal. State laws frequently disallow MBHC expansion via new bank charters, and most new banks are not affiliated with holding companies. For example, in 1980 only 32 banks or approximately 11 percent of newly chartered banks were members of BHCs. An analysis of the data supports the contention that BHCs do not positively impact the number of facilities per area. Banking markets were categorized according to the extent of MBHC activity, and t-tests were performed on the mean values of the dependent variable for various subgroups (e.g., unit banking markets with extensive MBHC activity compared with unit banking markets that do not allow MBHCs, limited branching markets with MBHCs compared to limited branching without MBHCs, etc.). Depending on how MBHC laws were interpreted, the mean value of the dependent variable either did not differ as a result of BHC status or was smaller when MBHC activity was allowed.

Estimates of equation (12) for the remaining years are also presented in Table 1 and can be compared with those for 1981 to see how the phase-out of interest rate ceilings affects the relationship. Most importantly, one can evaluate how the influence of branching changed over the period. The results show that the relationship has been remarkably stable. Coefficients for each variable carry the same sign and are of similar magnitude and statistical significance as those found in the earlier time period.(8) This occurs in spite of the phase-out of deposit rate restrictions and changes in the number of banking offices. The results indicate that over the time period analyzed the influence of branch restrictions was greatest in 1982 when office density was over 50 percent greater in areas in which branching was allowed. Although branching continues to result in superior service accessibility, the differential did decline over time. It was greatest in 1982 and declined thereafter, with the influence in the last year being significantly less than that in 1982.(9)

(8)The coefficients for the branching binary in each year over the 1981-85 period was not statistically different from that for 1981. However, the difference between 1982 and 1985 is significant at the 5 percent level utilizing a standard t-test.

(9)The methodology and technique employed to generate the reported results are rather straightforward. The findings were also robust to the technique utilized. For example, the data were pooled, and both OLS and generalized least squares estimates produced findings nearly identical to those presented here. Use of binaries for individual years also showed a significant shift in the regression in 1985--although the branching binary was of nearly identical magnitude.

Given the above findings, one can tentatively conclude that branching results in superior service accessibility for consumers even in a deregulated deposit rate environment. However, the accessibility measure is not intended to be a comprehensive measure of customer service. Rather, it is only one relatively important aspect of service, i.e., customer access to full-service offices. Alternative means of delivering bank services have been developed; the recent proliferation of automated teller machines (ATMs) being the most obvious. If the machines serve as substitutes for bank offices and are more prevalent in areas where branching is precluded (a logical assumption because banks in these areas have limited alternative means of geographic expansion) then the calculated differential in customer service may be overstated. Previous studies have excluded ATMs because of data limitations and because they have not been considered close substitutes for full-service offices (Gross 1987). Nevertheless, they are substitutes for a limited number of bank services, and data have recently become available, so their influence should be considered.(10)

(10)The author is grateful to anonymous referees for suggesting this extension of the research. Data are from ATMs Worldwide--The Nilson Report.

The role of branching in the above analysis may be overstated if ATMs are concentrated in unit banking areas. However, analysis of mean values for the absolute number or per capita number of ATMs for 1984-85 indicates that they are actually more prevalent in branching states. Correlations between these two measures of ATM presence and the branching binary variable also indicate that their presence is positively related to the ability to branch.(11) Thus, not only do customers in branching areas have more full-service facilities, but also they apparently are provided more ATMs. While no casual relationship is assumed, the evidence suggests that the presence of ATMs does not bias upward the service differential found in branching areas. Thus, even accounting for ATMs, consumers are likely to benefit from price competition between banks with no appreciable offset in the availability of service. 11 The correlation coefficient between the binary and the number of ATMs = .06 and with ATM per capita = .19. The correlation with the per capita value, the more meaningful of the two measures, was significantly different from zero at the 1 percent level. Values for individual years were similar.


The U.S. banking industry has long been subject to significant regulation to ensure consumers access to a safe, reliable flow of financial services. One form of this regulation--limiting geographic expansion--has been shown to result in inferior consumer service. However, deposit rate restrictions may have resulted in the proliferation of bank offices in areas where expansion was allowed--a means of nonprice competition--and are the major reason for this service differential. Thus, consumer benefits attributed to a bank's ability to provide services through multiple offices may not be realized without deposit rate ceilings. Consumer benefits gained from price competition may simply be offset by lower service levels. With the recent deregulation of interest payments on bank deposits, this argument can be empirically evaluated. This paper has presented a model of bank office density, a customer service level measure, and then used that model to evaluate changes during the deregulated period.

Results indicate that improved service levels result from branching per se and are not significantly affected by deposit rate deregulation. While the differential between the number of bank offices per square mile in branch and unit banking states has declined slightly in the last few years, it is still substantial. The obvious policy implications are that consumers should welcome and encourage the liberalization of branching laws. While one could argue that some reservations may be in order because service accessibility is only one aspect of the branching-consumer welfare issue, the benefits found here serve to complement those generally found when evaluating branching and other aspects of consumer well-being (Baer et al. 1988).

As with most empirical studies, a few qualifying statements appear to be in order. First, the results do not imply that deposit rate regulation has had no impact on the industry or, more importantly, on customer service levels. It has had obvious positive effects on the consumer's ability to obtain competitive rates on deposits and potential effects on bank profitability. Rather, the results suggest that the social benefits of branching are realized even as banks are allowed to offer competitive rates explicitly. That is, the benefits of deposit rate competition have not been achieved at the expense of customer convenience. Second, data limitations require some simplifying assumptions, although they are not thought to impact the results appreciably. Finally, although substantial turnover has occurred in the industry during the time period studied, it is possible that future trends may show that the differential in customer service levels between areas with different branching laws will continue to narrow. However, the differential has been persistent over time, and the changes required to eliminate it totally would be substantial. The improved customer service realized with branching is expected to continue, although future research efforts will be needed for verification.

Table : 1 Bank Office Density--1981-85
COPYRIGHT 1989 American Council on Consumer Interests
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1989 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Evanoff, Douglas D.
Publication:Journal of Consumer Affairs
Date:Jun 22, 1989
Previous Article:Rental-purchase agreements: a preliminary investigation of consumer attitudes and behaviors.
Next Article:A content analysis of problem-resolution appeals in television commercials.

Related Articles
Banking plus: with competition and deregulation, banks in Indiana are offering more non-traditional financial products.
The benefits of branching deregulation.
The challenges facing community banks: in their own words.
The real effects of U.S. banking deregulation.
Banking on Africa.

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters