The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
FEDERAL RESERVE BANK of CLEVELAND SPRING Economic Review Spring 1984 Estimating Infrastructure Needs: M ethods and C ontroversies................... 2 In this article, economist Paul Gary Wyckoff examines the pros and cons associated with three alternative measures of capital-stock needs: technical estimates, based on the judgments of experts (such as safety engi neers and urban planners); simple compari son studies, which look at the expenditures of other, similar cities; and economic estimates, which result from an approach developed by the author in an earlier working paper. In constructing economic estimates, the author compares the expenditures of each city with those of a typical city in the sample, after accounting for the features of the city that affect the need for capital stock (for example, population, land area, and age of capital stock). The article illustrates these three alternatives, using bridge condition and highway spending data from ten midwestern urban areas. Nonbanking Operations of Bank Holding Companies .................... 11 Under current law, banking organizations are free to enter a variety of nonbanking activ ities. Proponents and opponents of this trend have identified a large number of potential benefits and costs, but there have been rela tively few empirical studies of the nonbanking operations of banking organizations. To pro vide insight on this issue, economist Gary Whalen examines the performance of the non banking subsidiaries of the 25 largest and 60 smaller regional bank holding companies over the 1981-82 interval. Economic Review is published quarterly by the Research Department of the Federal Reserve Bank of Cleveland, RO. Box 6387, Cleveland, OH 44101. Telephone: 216/579-2000. ISSN 0013-0281 Editor: Pat Wren. Assistant editor: Meredith Holmes. Design: Jamie Feldman. Typesetting: Lucy Balazek. Opinions stated in Economic Review are those of the authors and not necessarily those of the Fed eral Reserve Bank of Cleveland or of the Board of Governors of the Federal Reserve System. Material may be reprinted provided that the source is credited. Please send copies of reprinted materi als to the editor. Economist Paul Gary Wyckoff studies the public sector for the Federal Reserve Bank of Cleveland. 1. Here Choate and Walter are sum marizing the results of a government study. For details, see Harrison J. Goldin, Deteriorat ing Infrastructure in Urban and Rural Areas, Subcommit tee on Economic Growth and Stabili zation, Joint Eco nomic Committee, 96 Cong. 1 Sess. Washington: U.S. Government Print ing Office, 1979, pp. 42-53. Estimating Infrastructure Needs: Methods and Controversies by Paul Gary Wyckoff 2. For details about this rating system and for an explana tion of the terms used, see Govern ment Accounting Office (GAO) Report B-201433, Better Targeting of Federal Funds Needed to Eliminate Unsafe Bridges, August 11, 1981, Chapter 4. Federal Reserve Bank of Cleveland Americans today are reading more and more about the problems of our nation’s infrastructure—our public capital stock of roads, bridges, sewers, transit systems, and government buildings. Having been hit by back-to-back recessions and high interest rates, many state and local governments have long postponed infrastructure expenditures. Unfortunately, however, studies of the prob lem have shown that more may be needed than just catching up with a few years of deferred maintenance. For example, the Con gressional Budget Office (1983) estimated that it would cost $53 billion per year nation wide to ensure that our highways, transit systems, sewer and water facilities, and airports are, in its words, “adequate.” In a widely quoted study, Choate and Walter (1981) stated that, in the next decade, servicing the infrastructure needs of New York City alone would require $40 billion.1 At the heart of decisions about funding the nation’s infrastructure lies the elusive con cept of capital-stock needs. Lacking complete information about the desires of their con stituents and the consequences of various spending decisions, policymakers have asked researchers and other experts for guidance in determining the “proper” level of each kind of capital stock. In providing this help, the aim of the analyst is modest: to arrive at a benchmark that will enable authorities to begin debate on capital-spending plans, rather than to develop a mathematical for mula that will determine the final and optimal allocation of resources in any city. The problem of setting appropriate policy goals is perhaps most acute in the older cities of the Midwest, where an aging infra structure and changing demands for public services strain tight government budgets. Presumably, reduced population and slower rates of income growth in these cities might influence the desirable amount of capital spending, but sorting out such influences is difficult. This article explains the three different types of needs estimates that are available, 3. Because of the interest in reviving Cleveland’s econ omy, these cities were chosen to provide a basis for compar ison with Cleveland. The cities chosen bracket Cleveland in latitude, 'longitude, and population. The sample omits some important cities (for example, Indianapo lis and Columbus) because of a lack of data on the age of bridges in those cities. This infor mation is required to construct the eco nomic estimates of need below. a national bridge inventory in which bridges are rated according to structural and func tional adequacy. A bridge is considered as structurally deficient if it needs immediate repair. According to FHWA bridge inspectors, immediate repair is necessary if the super structure, substructure, or culvert is rated by bridge inspectors as being 3 or less (on a scale of 0 through 9, with 0 being the worst) or if the general appraisal or waterway adequacy I. Technical Estimates of Need ratings are 2 or less. A bridge is considered as functionally obsolete if its deck geometry, ap The most common type of capital-stock proach, roadway alignment, or load-carrying needs estimate is a technical one, involving an extensive review of the quality and quan capacity cannot safely service the road tity of existing public capital. The difference system that it’s a part of—more precisely, when bridge inspectors rate the deck geome between the actual stock and a benchmark try, underclearance, approach roadway or standard level of performance is labeled the needs gap. Although the exact benchmark alignment, overall appraisal, or waterway adequacy 3 or less.2 level used is often not explicit, the level is Of the bridges that were structurally usually based on either the analyst’s subjec tive determination of the “proper” amount of or functionally inadequate in ten large midwestern cities in 1980, the most serious public capital, or on standards rooted in the bridge problems appeared to be in Buffalo opinions of technical experts, such as civil and Rochester, where more than one-half of engineers or urban planners. Table 1 presents a typical example of such the bridges were rated as deficient.3 Struc a technical estimate of infrastructure needs. tural deficiency was a larger problem for bridges in the Midwest than functional obso To allocate federal funds, the Federal High lescence, but this fact may be inherent in way Administration (FHWA) maintains the rating process used by FHWA. When a bridge becomes structurally deficient, it auto matically is dropped from the list of func Table 1 Bridge Needs, Determined by tionally obsolete bridges. Federal Highway Administration Although technical estimates of this type Federal Bridge Inventory, 1980 are certainly necessary, especially in allo Struc Total cating budgeted funds among different pro turally Func Number defi tionally needing jects, they have two drawbacks. First, the City of bridges cient, % obsolete, % repair, % standards used are arbitrary. For example, Buffalo 4 59 136 55 why should a bridge become functionally 18 Chicago 464 7 11 obsolete when the FHWA rates it 3, rather Cincinnati 13 215 10 3 than 2 or 4? Every capital-spending needs Cleveland 279 23 0 23 estimate contains an element of subjectivity. Detroit 412 7 5 2 What the analyst is really doing is presenting 18 Louisville 217 1 17 one particular set of spending plans as being Milwaukee 769 27 5 32 better than others. Given that the goal of Minneapolis 37 291 27 10 needs estimates is to inform policymakers, Pittsburgh 207 4 39 35 the best an analyst can do is base the esti Rochester 102 57 51 6 mates on a set of values widely held by the SOURCE: Peterson et al. (1984), table 5. clientele group. The analyst then translates those values into a set of benchmark spend ing plans, quantifying how they compare with actual conditions. The problem with Economic Review • Spring 1984 including a new approach developed by the author. The strengths and weaknesses of each method are examined, and the different approaches are illustrated with examples of highway needs estimates for ten urban areas in the Midwest. The article concludes with suggestions for the best application for each method. 4. See GAO Report B-201433. 5. See Department of Transportation, Federal Highway Administration, “Design Standards for Highways: Re surfacing, Restora tion, and Rehab ilitation of Streets and Highways Other than Freeways,” Federal Register, vol. 47, no. 112 (June 10, 1982), pp. 25268-75. these technical estimates is that they fre quently embody values that may differ from those held by policymakers and hence are of little use to them. To cite another example, the federal government has developed a suf ficiency rating for each bridge in the FHWA inventory to allocate bridge rehabilitation and reconstruction grants. The structural adequacy and safety of the bridge determine 55 percent of this rating; serviceability and functional obsolescence make up 30 percent; and economic necessity accounts for 15 per cent.4 However, states might prioritize each of these objectives differently, and thus find this system of allocating funds to be of little value. A state vitally interested in economic devel opment, for instance, might weight economic necessity more heavily than 15 percent. A second disadvantage of technical esti mates is that they offer no guidelines for allo cating resources when budgets are severely constrained. Many cities, especially those in the Northeast and Midwest, simply lack enough funds to meet their needs, not only for capital spending but also for other needs. Of the nation’s 153 cities with a 1970 popu lation greater than 100,000, Bradbury (1982) found 23 cities to be suffering budgetary fiscal distress in 1977. Nineteen of these 23 cities were in the Northeast or North Central regions. Moreover, two of these cities, New York and Cleveland, moved from distress to crisis in the 1970s—that is, they were unable to meet their financial obligations. Lacking appropriate benchmarks for budgeting purposes, the danger is that communities will abandon capital investment planning altogether, pursuing instead a pay-as-you-go or, more accurately, pay-as-it-breaks strategy. Recognizing that the resources of state and local governments are limited, the federal government has recently made rule changes that, in effect, relax needs standards. For Federal Reserve Bank of Cleveland example, the FHWA has ruled that, to receive federal funds, states need not always repair highways (other than freeways) according to standards required for new construction.5 This departure allows states to minimize expenditures on projects that are less than essential. II. Average Expenditures as Measures of Need A second type of benchmark that might be used by policymakers is a simple comparison of their own capital expenditures with those of similar cities. While this approach does not clarify how to allocate funds between differ ent infrastructure projects, it does serve as a rough indicator of the appropriateness of a city’s overall capital-spending plan. Figure 1 illustrates a needs estimate in which the highway spending of each urban area in our ten-city sample is compared with the average for the entire sample. It should be empha sized that these estimates are not directly comparable with the technical estimates given in table 1. First, expenditures on the entire highway and road system, not just those on bridges, are being compared. Also, these figures represent annual averages of expenditures between 1965 and 1976 rather than conditions in 1980 (the date of the fig ures in table 1). Finally, the unit of analysis in figure 1 is not individual cities, but the aggregate expenditures of all local govern ments within an urban county. This adjust ment reflects the tendency of states to assign responsibility for highways to different levels of government; in some states, counties take more responsibility for this than in others. To maintain comparability, spending must be aggregated over all local governments within a geographic area. Large disparities exist in highway spend ing between urban counties. Milwaukee County, WI, spends almost five times as much per capita as Jefferson County (Louis roads than Milwaukee, perhaps accounting ville), KY. Being the southernmost city in this for part of the difference. Even among coun sample, Louisville may use less salt on its ties with similar climates, however, distinc Fig. 1 Highway Spending in Ten Urban Counties Annual averages, 1965-76 Average, entire sample 32.02 1972 dollars per capita SOURCES: D epartm ent of Commerce, Census Bureau, Local Government Finances in Selected Metropolitan Areas and Large Counties (annual). Economic Review • Spring 1984 6. Statistics on snowfall are from U. S. Department of Commerce, Bureau of the Census, Sta tistical Abstract of the United States (1982-83), 103rd ed., Washing ton: U.S. Govern ment Printing Office, p. 218. tions persist. Erie County spends only about two-thirds as much as Milwaukee County, despite the fact that Buffalo gets almost twice as much snow per year.6 III. Economic Measures of Need What factors are responsible for disparities of such magnitude? Should we consider them when making a needs estimate? An urban county’s land area, for example, probably influences the cost of providing highways— the more dispersed the population, the higher the cost of linking the county by highway. Indeed, the central problem with the average measure of infrastructure needs is that no two cities or counties are exactly alike, and they may differ in ways that affect capitalstock needs. What is required, then, is a method of adjusting the estimates to account for the particular circumstances faced by each jurisdiction. The result would be a third type of capital-stock-needs estimate. This approach might be called an economic esti mate of capital-stock needs, in the sense that it reflects the characteristics of each area affecting the demand for public capital stock. A Model of Infrastructure Spending The mechanics of implementing an eco nomic approach to capital-stock needs are described in Wyckoff (1984). Basically, a twostep methodology was employed. First, a pos itive model of public capital spending was developed and tested, using highway spend ing data from 1965 to 1976 for the ten urban counties listed in figure 1. Investment in public capital was modeled as a conscious choice made by public authorities, based on available resources and characteristics of the urban area. We made no particular assump tion about the nature of this public-choice mechanism. In other words, the question of what particular group controls a city—voters, a political party, a special interest group, or city workers—was left open. Our model of infrastructure spending was completed using standard tools from the eco Federal Reserve Bank of Cleveland nomic literature on investment, consumer demand, and local public finance. However, we combined these techniques in a unique way. The desired level of public capital was taken to be a function of income, the price of public capital (including interest rates), the age and value of the existing capital stock, population, area, and the amount of aid received by the community. The functional form for this relationship was taken from the Almost Ideal Demand System (AIDS), which is well known in the literature on testing the theory of the consumer. Actual spending was then related to desired stocks through a common stock adjustment function, which simply states that each city spends enough to maintain its capital stock and to eliminate some part of the gap between desired and existing capital stock. One important problem addressed in Wyckoff (1984) concerns the choice of the unit of analysis for such a study. Use of data that are aggregated over all governments within a geographic area presents a dilemma to those doing econometric research. Pre viously, researchers could never be sure that aggregate data were representative of indi vidual units. If, instead, they used data for individual units, they avoided this aggrega tion problem but risked additional error from the above-mentioned nonuniformity in the type and level of services offered by individ ual governments. Baltimore and St. Louis, for example, have integrated city and county governments, so that these governments have greater responsibilities than the city govern ments of Detroit or Cleveland. Thus, by using jurisdictions with different levels of responsibility in a cross-section estimation procedure, the researcher risks confusing expenditure differences because of vary ing levels of responsibility with additional expenditures made by one jurisdiction because of changing circumstances within that city. 7. Since writing this article, it has come to the author’s attention that this approach to esti mating expenditure needs, with some modification, has been used to esti mate intergovern mental grant needs in Europe. See Organization for Economic Co operation and Development (OECD), Mea suring Local Gov ernment Expendi ture Needs: The Copenhagen Work shop, Paris, France, 1981. Fortunately, innovation in modern demand theory has led to the development of func tional forms that fit the data well and aggre gate perfectly—that is, aggregate demand can be shown to be determined by the eco nomic conditions facing the average city in the sample. The AIDS demand functions are generally of this type. Thus, aggregate demand can be utilized without concern about its representativeness. Wyckoff (1984) shows that this property is preserved, even when noneconomic variables such as popula tion, area, and age of the capital stock are introduced into the demand function, as long as the following conditions hold: (1) across time and across counties, the intra-county distributions of city per capita income are proportional, and (2) across cities within each county, age of capital stock, area, and popu lation are independent of income. These assumptions were tested statistically and were not rejected by the data. These results allowed testing of the model with character istics of the average city in each county, rather than having to know the conditions faced by each jurisdiction in the county. For two of the independent variables, esti mates rather than actual values for the series had to be employed. The value of the capital stock was estimated from data back to 1941, based on the highway spending of the largest cities in each county. On average, the expen ditures of these cities contributed about one-half of the spending for the county as a whole. The age of the bridges in the central city of each urban county was utilized as a proxy for the age of the capital stock. The most significant determinants of highway spending were found to be popula tion, value of the existing capital stock, land area of the jurisdiction, and amount of aid received from higher levels of government. In addition, local decision-makers appeared to be more concerned with repairing and replacing existing capital stock than with building new roads and bridges to meet changed needs Economic Review • Spring 1984 in their communities. A larger population necessitated a proportional increase in high way spending, so that, other things being equal, per capita highway expenditures did not vary much across cities of different sizes. And, as expected, the larger the land area, the greater the expenditures on highways. Weaker and less consistent relationships were found between highway spending and per capita income, the price of capital goods, and the age of the capital stock. Predicted Values as Estimates of Need7 The first step of the analysis resulted in a model that explained how a typical city in the sample reacted to changes in its eco nomic characteristics. The second step com pared the actual spending of each individual city with that of a typical city under the same economic conditions. We plugged the individ ual city’s values for the independent vari ables into the equation estimated in step one, and calculated the estimate of spending that resulted, subject to an adjustment to neutral ize the effect of aid on spending. The thinking here is that income, price of capital goods, current capital-goods stock, population, area, and age of capital stock ought to be consid ered in determining highway needs and were therefore allowed to influence the needs esti mate. It might easily be argued, however, that the need for highways is simply independent of the source of financing available to the community. Does a city need more roads simply because the federal or the state gov ernment is willing to pay for them? For this reason, in determining needs estimates, each city was assigned an amount of aid equal to the average for all cities in the sample. The advantage of the resulting estimates is that they are customized for each city in the sample. That is, they reflect each area’s par ticular urban characteristics. These esti mates, then, answer a question of importance to policymakers: what would a typical city in our region do if confronted by circumstances similar to ours? I In figure 2, the average actual and needed actual and needed expenditures look small real per capita highway spending are shown on the chart, but in some cases they represent for each urban county. The gaps between significant sums of money. It turns out that Fig. 2 Real Per Capita Expenditures on Highways Annual averages, 1965-76 (predicted value of the adjusted model) Federal Reserve Bank of Cleveland the two westernmost areas in our sample— Hennepin County, MN, and Milwaukee County, WI—are farthest above their needs estimates, while two older, industrial, more eastern counties—Erie County, NY, and Cuyahoga County, OH—have the largest capital-spend ing deficits. To put these numbers in perspec tive, the Cuyahoga County deficit amounts to about 5 percent of actual expenditures, or approximately $2 million per year. The Mil waukee County surplus, on the other hand, accounts for 6 percent of actual expenditures, or approximately $3 million annually. These differences can only partly be ex plained by differences in aid (see table 2). Milwaukee and Hennepin counties do have the second and third highest aid per capita; however, Cuyahoga County receives more than the average amount of aid (sixth high est), and Erie County gets only the third low est level of aid. Clearly, some of these differ ences remain to be explained by other factors. More surprising perhaps is the wide range Table 2 Average Real Per Capita Aid in Ten Urban Counties, 1 9 65-7 63 County State All counties, all years New York Erie Illinois Cook Hamilton Ohio Cuyahoga Ohio Michigan Wayne Kentucky Jefferson Milwaukee Wisconsin Minnesota Hennepin Allegheny Pennsylvania New York Monroe Major city Buffalo Chicago Cincinnati Cleveland Detroit Louisville Milwaukee Minneapolis Pittsburgh Rochester Average, 1972 dollars 18.24 13.10 20.69 19.66 19.48 28.75 9.76 22.85 22.61 14.37 11.11 a. Aid includes general revenue sharing, grants for highways, and direct expenditures of state highway departm ents on local roads and streets in each urban county. SOURCES: D epartm ent of Commerce, Census Bureau, Local Govern ment Finances in Selected Metropolitan Areasand Large Conn ties (annual); and special releases from the Departm ent of Transportation, Federal Highway Adm inistration. Economic Review • Spring 1984 of per capita needs levels specified by this procedure. Since the highway spending pro cess is dominated by repair and replace ment considerations, these levels are deter mined, to a large extent, by the value of the existing capital stock that must be main tained. Thus, Jefferson County, KY, has the smallest highway .^ending need of $11.75 per person because of its low per capita income and its small, relatively new capital stock. Milwaukee County, on the other hand, despite having the smallest land area in the sample and a relatively new capital stock, has the largest capital-spending need ($46.50 per person) because of the size of the capital stock that it must maintain. This large varia tion in highway-spending needs also points up how misleading a simple average expendi ture figure can be as a measure of capitalspending needs, since it does not adjust for these differences in maintenance needs. These new estimates of capital-spending needs are not without controversy. It might be argued that the heavy emphasis in these estimates on repair and replacement indicates myopia on the part of local decision makers, a blind concern for preserving phys ical capital rather than serving the needs of their constituents. However, using this argument would necessitate showing how these politicians and administrators consis tently misjudged the amount of the public capital required. Although it is easy to argue that one city miscalculated the needs of its citizens, it is much more difficult to show that an entire region under- or over-estimated its capital-stock needs. One of the strengths of this economic approach is that the typical responses of a large group of policymakers constitute the benchmark against which actual expenditures in each area are measured. IV. Conclusion What kind of infrastructure needs estimate is best? The answer depends on the kind of information that policymakers want. In many cases, the opinion of experts is invalu able, especially in allocating funds among various projects. Increasingly, however, poli cymakers have become disillusioned with experts’ technical estimates, not only because their studies are based on arbitrary stand ards but also because these estimates fail to recognize the limitations of city budgets. If officials want a rule of thumb from which to begin budget discussions about overall capital spending, they may want to sample the opinions of their counterparts in similar cities by looking at average expen diture data. This can be very misleading, however, because every city contains special features that are not acknowledged in these comparisons. As an alternative, we have proposed economic estimates of spending needs, in which estimated needs figures are essentially the predicted values of a model that accounts for these special features. In this way, needs estimates can be individ ualized to reflect the matrix of characteristics peculiar to each local area. The economic needs approach has two advan tages over technical estimates. The cost of an economic needs estimate is only a fraction of the cost of a technical study, since no ex haustive inventory of physical units is neces sary. The results also may be of greater interest to policymakers, since arbitrary and sometimes impossible-to-attain needs stan dards are avoided. Federal Reserve Bank of Cleveland References Bradbury, Katherine L. “Fiscal Distress in Large U.S. Cities,” New England Economic Review, Federal Reserve Bank of Boston, November/December 1982, pp. 33-43. Choate, Pat, and Susan Walter. America in Ruins: Beyond the Public Works Pork Barrel, Washington, DC: The Council of State Planning Agencies, 1981. Congressional Budget Office. Public Works Infrastructure: Policy Considerations for the 1980s, Washington, DC: U.S. Government Printing Office, April 1983. Humphrey, Nan, and Peter Wilson. Capital Stock Condition in Twenty-eight Cities. Pro cessed. Washington, DC: The Urban Insti tute, February 15, 1980. Peterson, George E., et al. Guide to Bench marks of Urban Capital Condition, Wash ington, DC: The Urban Institute, in col laboration with Public Technologies, Inc., May 1984. Wyckoff, Paul Gary. “Economic Estimates of Urban Infrastructure Needs,” Federal Reserve Bank of Cleveland, Working Paper 8401, June 1984. Economist Gary Whalen researches banking issues for the Federal Reserve Bank of Cleveland. June Gates provided research assistance for this article. Nonbanking Operations of Bank Holding Companies 1. There also have been concerns voiced about possible con flicts of interest, unfair competition, and undue concen tration of resources. It is generally not possible to examine these assertions empirically. by Gary Whalen 2. Unlike banks, nonbanking sub sidiaries of bank holding companies are not generally subject to either intraor interstate geo graphic restrictions on office locations. 3. For example, see Curry (1978); Drum (1977); and Kama (1979). 11 Banking organizations first evidenced a strong desire to move into nonbanking activ ities in the early 1970s. Ever since, the appro priate types and scale of such activities and the appropriate mode of entry (that is, bank holding company subsidiary vs. bank sub sidiary vs. bank) have been hotly debated. Both proponents and opponents of banking expan sion into nonbanking areas (banking regu lators, in particular) have always been con cerned about the likely impact of nonbanking activities on the soundness of directly in volved and related banking organizations.1 In recent years, these concerns have multi plied as bankers contemplate entry into more nontraditional, presumably riskier, fields. Two diametrically opposed, extreme, ex pected impact scenarios have been depicted over the years. One view is that engaging in nonbanking activities would permit partici pating banking organizations to earn returns higher than those obtainable in traditional banking and/or diversify and so reduce their risk.2 Participating banking organizations, as well as co-affiliates, engaging in such activi ties would be sufficiently insulated from operating problems encountered when par ticular nonbanking fields were entered, as long as they did so through separately incor porated subsidiaries. The opposing view emphasizes the possi bility that nonbanking activities could turn out to be less profitable and/or more risky than expected. This could ultimately weaken related banks if their resources had to be used, either because of legal or moral obliga tions, to support troubled nonbanking com panies. Diversification benefits might not materialize if the nonbanking fields entered were closely related to banking. Empirical evidence on these issues is scant, mixed, and dated.3 In light of the recent debate about the appropriateness of bank entry into other nontraditional fields, this study explores the extent of involvement of bank holding companies in nonbanking activities and the 4. In the text, the 25 largest compa nies are referred to as large, and the 60 other companies as small or regionals. The states are Ala bama, Florida, Mas sachusetts, Michi gan, Missouri, New Jersey, Ohio, Ten nessee, Texas, Vir ginia, and Wisconsin. 5. Under section 4(c)8 of the Bank Holding Company Act, the Federal Re serve Board specifies the types of non banking activities in which bank holding companies may be come involved. (The Comptroller of the Currency has sim ilar authority for national banks.) In general, the Federal Reserve permits bank holding com panies to engage in activities “closely related to banking” that are expected to result in net bene fits for the public. The approved activ ities constitute the so-called laundry list; see Whitehead (1983), pp. 10-11. 6. The reasons that holding companies choose to structure their nonbank sub sidiaries as parent vs. bank subsidiar ies are not clear, nor is a clear pattern of organization or trend in organiza tion evident. 12 discernible impacts of this involvement on the parent holding companies and their sub sidiary banks. The nonrandom sample of hold ing companies examined consists of the 25 largest bank holding companies in the United States (as of year-end 1982) and 60 smaller regional companies in 11 different states with an average consolidated asset size of $3.8 billion.4 I. Involvement in Nonbanking Activities We can determine the general types of non banking activities that bank holding compa nies engage in by examining their annual reports or other published materials. These sources show that the sample companies most frequently were involved in consumer and commercial finance (including indus trial banks), mortgage banking, leasing, and insurance sales related to extensions of credit.5 Fewer, generally larger holding com panies own factoring operations, small bus iness investment companies, or discount brokers; operate futures subsidiaries or troubled S&Ls; or engage in other autho rized nonbanking activities. However, it is much more difficult to obtain precise quantitative estimates of holding com pany involvement (and the returns earned) in each type of nonbanking activity. Hold ing companies can engage in most nonbank activities, either through a subsidiary of the parent company or through a subsidiary of a bank affiliate, or both.6 Banks that own non banking subsidiaries are not required to pro vide balance-sheet and income-statement data for each subsidiary individually. Banks may even engage in certain nonbanking activities directly (for example, operate a leasing divi sion). Thus, their nonbanking operations become an unidentifiable part of their own consolidated reports of income and condition. Further, a nonbanking subsidiary of the parent may engage in a variety of nonbank ing activities and is not required to report disaggregated performance data. The best Federal Reserve Bank of Cleveland available quantitative information on the nonbanking operations of bank holding com panies is aggregated data for all the nonbank ing subsidiaries of the parent holding com pany contained in Y-9 reports filed with the Federal Reserve. (Y-9 reports contain consoli dated and parent company balance sheets and income statements.) This is the source of the data presented in the tables in this article. Accordingly, the extent of bank holding com pany involvement in nonbanking activities is understated to some unknown extent. The data in table 1 indicate that bank holding companies generally are not heavily involved in nonbanking activities. The large companies are more actively engaged in these activities than the regionals, but consider able variation exists even within each group. Equity investments in nonbanking subsid iaries averaged only 4.9 percent of parent company total assets at the 25 largest com panies at year-end 1982; this figure was just 2.2 percent for the small companies. Because of the wide variation in this ratio within each group, even the relatively low mean ratios exaggerate the typical level of holding com pany involvement in nonbanking activities somewhat. The substantial within-group variation is indicated by the relatively large size of the standard deviation of the ratio of nonbank equity/parent company total assets relative to the mean and the wide range in this ratio for each group: 14.7 percentage points for the large companies and 19.1 per centage points for the regionals. Under such circumstances the median value of this ratio is a superior indicator of the typical extent of holding company involve ment in nonbanking activities. It is just 3.8 percent for large companies and 1.3 per cent for regionals. An additional 19.7 percent of the assets of large parent companies, on average, consisted of advances to (that is, debt of) nonbanking subsidiaries. At smaller companies the mean 7. For perspective, the mean ratio of equity investment in bank subsidiaries to parent company total assets was 49.6 percent and 77.2 percent at larger and small compa nies, respectively. The mean ratios of advances to bank subsidiaries relative to parent company total assets were 4.1 percent and 2 .7 percent, respec tively. ratio was just 4.2 percent. Once again, mea sures of dispersion indicate that this ratio differs considerably across companies, both between and within groups. The ratio is gen erally higher and less variable for the large companies, suggesting that they typically assist their nonbank affiliates in raising funds. Fifteen of the regionals made no ad vances at all to their nonbank subsidiaries.7 Alternatively, equity in nonbank subsid iaries constituted 9.8 percent of large com pany equity investments in all subsidiaries, on average, in 1982. Total parent company investment in nonbank subsidiaries (equity plus advances) averaged 29.7 percent of total large parent company investment in all subsidiaries in the same year. Comparable figures for the regional companies are 2.8 per cent and 7.2 percent, respectively. Again because of the skewed nature of the data, the median values of these ratios are slightly below the respective mean for both large and small companies. Interestingly, comparison of the 1982 ratio of equity investment in nonbanking activi ties to equity investment in all subsidiaries with its level in 1978 indicates that many regional companies are less involved in nonbank activities now than they were in the past. This ratio actually declined at 27 of these companies over the period. The mean change in the 1982 ratio minus the 1978 ratio was 5.0 percentage points for the large companies and a negative 0.4 percentage point for the smaller ones. II. Performance of Nonbanking Subsidiaries The impact of nonbank subsidiaries on the holding company and related banks depends not only on the scale of nonbank operations but also on the performance of these sub sidiaries over time. One indicator of nonbank performance is net income earned by these subsidiaries relative to the equity investment Table 1 Measures of Holding Company Involvement in Nonbanking Activities Figures in percentage points; 1982 year-end data Ratio Equity investment in nonbank subsidiaries/ parent company total assets Advances to nonbank subsidiaries/ parent company total assets Equity investment in nonbank subsidiaries/ equity investment in all subsidiaries Parent company investment in nonbank subsidiaries/ parent company investment in all subsidiaries Large holding companies Standard Mean Median deviation Range Regional holding companies Standard Mean Median deviation Range 4.90 3.75 4.02 14.67 2.20 1.30 3.13 19.10 19.74 15.48 18.18 58.22 4.22 0.85 7.87 40.39 9.77 8.60 8.22 31.57 2.81 1.53 3.70 21.52 29.68 25.45 21.06 65.70 7.24 3.46 9.50 44.77 Economic Review • Spring 1984 8. Calculation of nonbank net income presumes that non bank subsidiaries are 100 percent owned by their parent holding companies. of the parent company in nonbank activities.8 Various summary statistics for this perfor mance measure defined over various time intervals appear in table 2. Again, it should be noted that performance of nonbank sub sidiaries of holding company banks is not reflected in the Y-9 data used to construct these measures. Examination of the mean and median values of this ratio reveals that the nonbanking sub sidiaries of larger companies generally have been more profitable than those of smaller companies. The exception is 1982, when the median return on nonbank equity of regionals was slightly above that of large companies. However, we cannot reject the hypothesis that the mean returns of the two groups of com panies are equal for any time period exam ined using formal statistical tests. None of the largest companies realized losses on their nonbanking operations in 1982, down from four in the previous year. The figures for the regional companies are 10 and 17, respec tively. The 1978-82 mean nonbank returns of two large and thirteen smaller companies were negative. The rate-of-return data indicate that 1981 was a particularly difficult year for the non bank subsidiaries of virtually all holding companies. Market interest rates attained Table 2 M easures of Nonbank Performance Figures in percentage points Ratio Mean Large companies Standard Median deviation Range Mean Regional companies Standard Median deviation Range 1. Nonbank subsidiary net income/equity investment in nonbank subsidiaries Time period 11.34 12.97 24.92 -66.4 - 66.3 1982 15.26 12.23 13.56 4.51- 64.02 8.93 9.49 64.10 -239.3 -364.1 1981 10.26 12.76 11.82 -17.14- 29.30 10.14 13.52 38.83 -152.52-184.72 Average 1981-82 12.81 13.98 -1.24- 34.10 7.83 -9.82-119.81 11.63 10.42 87.81 -238.14-532.6 Average 1978-82 14.85 11.20 23.51 2. Nonbank subsidiary net income/equity investment in nonbank subsidiaries minus bank subsidiary net income/equity investment in bank subsidiaries Time period -80.71- 52.23 0.66 24.90 1982 3.10 -0.60 12.03 -1.59 -8.38- 45.14 -4.30 -2.72 64.90 -252.92-360.41 1981 -2.82 1.45 12.92 -38.14- 15.23 -2.95 -0.59 39.13 -166.20-176.32 Average 1981-82 0.14 2.27 7.20 -13.97- 16.14 -1.73 -2.49 88.13 -248.68-524.25 Average 1978-82 1.76 -0.86 23.76 -21.38-108.42 3. Equity investment in nonbank subsidiaries/equity investment in all subsidiaries times ratio 2 above Time period -2.69- 2.73 0.01 0.75 -0.02 1982 -0.04 -0.03 0.74 -2.42- 1.45 1.12 -2.22- 7.07 -0.09 -0.01 1981 -0.30 0.04 -2.97- 0.92 0.98 -2.22- 2.97 -0.09 0.001 0.70 Average 1981-82 -0.15 -0.05 0.80 -2.51- 1.13 0.62 -2.58- 1.34 -0.21 -0.01 Average 1978-82 -0.25 -0.03 0.57 -2.08- 0.34 4. Revenues paid by nonbank subsidiaries to parent/parent total operating income 3.47 10.71 0- 41.17 8.20 0- 76.95 Average 1981-82 26.23 21.32 22.89 5. Dividends paid by nonbank subsidiaries to parent/parent total operating income 0- 23.84 0.30 3.30 1.36 2.64 0- 11.15 Average 1981-82 1.63 0.62 6. Dividends paid by nonbank subsidiaries to parent/parent equity investment in nonbank subsidiaries 0- 52.26 0.95 11.40 6.60 Average 1981-82 3.78 5.74 0- 22.53 1.35 - 14 Federal Reserve Bank of Cleveland 9. These last two figures are the mean of the following ratio for the companies in each group: stan dard deviation of rate of return on nonbank equity over 1978-82, divided by the mean rate of return on nonbank equity over 1978-82. unprecedented cyclical peaks in this year. The lackluster performance of nonbank sub sidiaries may have stemmed from heavy reli ance on short-term funds or the existence of usury ceilings, or both. The analysis also reveals that holding com panies’ returns on their equity investment in nonbank activities have varied considerably across companies and over time. The varia tion is particularly notable at the regional companies. The 1982 coefficient of variation of nonbank return on equity was 88.9 percent at large companies and 219.8 percent at the regional companies. The ranges are 59.5 and 132.7 percentage points, respectively. The mean coefficient of variation of return on non bank equity for large companies defined over the 1978-82 interval was 105.7 percent; for the smaller companies, it was 249.0 percent.9 Nonbank returns are considerably more variable than returns earned by holding com panies on their equity investment in bank subsidiaries. The coefficient of variation of return on bank equity was approximately the same for large and small companies—roughly 20 percent in 1982 and 30 percent over the 1978-82 interval. Insight into the relative profitability of nonbank activities can be obtained by com paring holding company rates of return on nonbank equity to a similar measure defined for bank subsidiaries. Summary statistics of such a measure defined over several time periods also appear in table 2. Given the vari ability of nonbank returns, it may be more informative to focus on the 2-year and 5-year average differences in returns reported in the table. The mean and median figures sug gest that larger companies generally have earned somewhat higher, but not markedly higher, returns on their nonbank activi ties than they have in banking. The opposite is generally true for smaller companies. How ever, none of the means is significantly dif ferent from zero for either group. Because of the year-to-year variability in nonbank returns, these summary measures obscure some interesting patterns in the disaggregated data for large and small com Economic Review • Spring 1984 panies. The rate of return on nonbank equity exceeded that earned on bank equity at 10 of the large companies and 20 of the regional companies in three of the five years over the 1978-82 interval. However, the nonbank operations of just two large companies were more profitable than their banking operations in four of these years, and not one large com pany managed to earn a higher return on its nonbank equity than it did on its equity investment in banking in every year during this period. Fifteen regionals managed to accomplish this feat in four years, and nine of these achieved it in all five years. Differences in nonbank subsidiary per formance across companies could result from any number of factors. It may be explained by differences in the degree of involvement of individual companies in particular types of permissible nonbank activities. For example, some companies may have elected to become involved in activities with high expected returns and risks. Alternatively, companies may simply differ in the organization of their nonbank activities. For example, certain nonbank activities might be grouped into a nonbank subsidiary of the lead bank by some companies but not by others. The length of time companies have engaged in particu lar types of activities also might influence reported performance. Presumably, experi ence is an advantage. Nonbank performance might depend on whether the subsidiaries acquired their nonbank subsidiaries or started them de novo (that is, from scratch). Gener ally, de novo operations are unprofitable for some period after start-up. Differences in the size and timing of nonbank acquisitions and/or the method of acquisition accounting employed might also influence reported nonbank performance. Differences in nonbank performance might be explained by the absolute level or rate of growth of their nonbank operations, or the size of their nonbank operations relative 10. Leverage is a measure of the extent to which a firm uses debt rather than equity to finance its assets. Rate of return on equity, the perfor mance measure used here, is the product of rate of return on nonbank assets and nonbank leverage (nonbank assets divided by nonbank equity). 11. Specifically, these were relation ships where the cor relation coefficient was significant at the 10 percent level. to their banking activities. It might be diffi cult to manage large and/or rapidly growing nonbank operations; or, rapid growth might reflect a preference for volume at the expense of profits. Nonbank performance might also depend on how highly the holding company leverages its nonbank operations.10 Or, perfor mance might be influenced by the size of the parent company. A large company may be able to realize and share various econo mies with its nonbank subsidiaries or pos sess superior management. Nonbank subsid iary performance could even be related to the performance of a company’s banking subsidiaries. Companies with highly profit able banking subsidiaries could afford to sacrifice nonbank profits for growth. It should be noted that, in most of these cases, arguments could be made to sup port the opposite type of expected impact. For example, the profitability of bank and nonbank subsidiaries might be positively correlated because of common superior parent company management, for example, or shared organizational economies. Thus, the expected relationship between nonbank profitability and each of these factors is generally ambiguous. Unfortunately, available data allowed only a few of these possibilities to be inves tigated. Specifically, measures of nonbank subsidiary profitability were correlated with the following: total parent company equity investment in nonbank activities; nonbank equity investment relative to equity invest ment in all subsidiaries; parent company advances to nonbank subsidiaries as a per centage of equity invested in such subsid iaries; parent company total assets; rate of return on bank equity; and various other measures of nonbank and parent company growth over the 1978-82 interval. Very few significant correlations were detected for either large or small companies.11 A significant negative relationship (correla tion coefficient = -0.374) was found between Federal Reserve Bank of Cleveland the rate of return on nonbank equity and the proportion of total subsidiary equity invest ment accounted for by such subsidiaries for the large holding companies only. A signifi cant positive relationship (correlation coeffi cient = 0.625) was detected between the measure of nonbank leverage (advances to nonbank subsidiaries divided by equity invest ment in such subsidiaries) and nonbank prof itability for large companies. Exactly the opposite result was found for the regional companies (correlation coefficient = -0.474). The rates of return on bank and nonbank equity were found to be significantly posi tively related (correlation coefficient = 0.416) but only for the large companies. These admittedly limited findings simply do not allow any definitive statements to be made about the causes of the observed var iation in nonbank performance. Further empirical research on this issue is necessary to answer this question and is beyond the scope of this study. The net impact of nonbank operations on the level of holding company returns depends on the interaction of two factors: 1. the proportion of equity investment in all subsidiaries accounted for by nonbank operations; 2. the difference in the rate of return earned on equity investment in nonbank and bank operations. Summary statistics for products of these two factors, again for several different time periods, appear in table 2. The mean and median net impacts are negative but slight for both large and regional companies in all time periods examined. However, the sample data do allow formal rejection of the hypothesis that the 5-year average mean nonbank net impact is zero for both large and small companies (5 percent level, 2-tail test). A look at the disaggregated data revealed that the average marginal net profitability impact of nonbank operations was positive at 15 of the 25 large companies and 28 of the 60 regionals over the 1981-82 interval. For the 5-year period, the figures drop to 8 and 25, respectively. 12. The variability of holding company returns depends on the variability of bank and nonbank returns, and the proportion of hold ing company invest ment in each, as well as on the cor relation of returns. 13. Seven of the correlation coeffi cients for large banks and seven for small banks were significant (10 p er cent level, 1-tail test). 14. Bank subsid iaries may also make a limited number of loans to nonbank co-affiliates under section 23A of the Bank Holding Company Act. The amount of such loans is not reported on Y-9 forms. The impact of the variability of nonbank returns on the variability of consolidated holding company returns depends crucially on the correlation between the rate of return earned on bank and nonbank activities.12 If bank and nonbank returns are negatively correlated, involvement in nonbank activi ties could moderate the variability of holding company returns. This is the basis of the alleged advantage of diversifying into non bank fields. Correlation of bank and nonbank return data over the 1978-82 interval revealed that negative correlations existed for 14 of the large companies and 32 of the smaller com panies.13 These findings suggest that slightly more than half of the sample companies de rived some measure of diversification ben efits from involvement in nonbank activities. Given these findings, it is not surprising that little correlation was discovered between various measures of the extent of holding company involvement in nonbank activities and the variance of consolidated holding company returns on equity calculated over the 1978-82 interval. Nonbank rate of return figures paint a some what incomplete picture of the contribution of nonbank subsidiaries to the holding com pany organizaton. For example, some net earnings are generally retained by nonbank subsidiaries and thus are not available for parent company use. Revenues actually upstreamed (that is, transferred) by nonbanking subsidiaries to the parent company averaged 26.2 percent of parent company gross income over the 1981-82 interval vs. just 8.2 per cent at the regionals (see table 2). Bank sub sidiaries provided the lion’s share of parent company operating income at both large and small companies—59.0 percent at the former, 85.3 percent at the latter. However, the rela tively large standard deviations and maximum values of the nonbank ratios indicate that nonbank operations contribute materially to the income of a number organizations. However, part of the revenue paid by non bank subsidiaries to the parent is compen sation for services rendered or funds advanced. A critical measure of the contribution made Economic Review • Spring 1984 by nonbank subsidiaries to the holding com pany organization is dividends paid to the parent. At larger companies, dividends paid by nonbank subsidiaries averaged 3.8 percent of equity investment in such activities and 1.6 percent of parent company gross income over the 1981-82 interval; at the regional companies, these ratios were 6.6 percent and 1.4 percent, respectively. The mean ratios obscure the fact that no dividends were paid in 1982 by the nonbank subsidiaries of 11 of the large companies and 33 of the regionals. The nonbank subsidiaries of 7 large and 31 regionals paid no dividends over the entire 1981-82 interval. Such performance is reflected in the considerably lower median values of these ratios. Bank dividends, on the other hand, averaged 35.4 percent of parent company operating income and 5.3 percent of the equity investment in banking operations at large companies over the 1981-82 interval. Comparable figures for the regionals were 59.5 percent and 7.0 percent, respectively. The importance of bank dividends to the consoli dated organization is also reflected in the fact that bank dividends averaged 119.2 percent of parent company dividends at large com panies and 158.2 percent at regional compa nies over the 1981-82 interval. The figures suggest the possibility that holding companies extensively involved in nonbanking activities may attempt to draw more heavily on the resources of their subsid iary banks to support their nonbanking oper ations. Specifically, higher dividends and/or management fees may be imposed on their bank affiliates.14 It is also possible that heavy involvement in nonbank activities may result in the parent company being operated in a more risky manner. To obtain insight on these issues, various ratios were constructed to reflect bank fee and dividend burdens and were correlated with the measures of parent company involve ment in nonbanking activities identical or similar to those defined in table 1. All ratios were 1981-82 averages. The ratios and a sum mary of the correlation results appear in table 3. No association was detected between was evident between fees paid by subsidiary nonbank involvement and bank fees and divi banks to the parent and one measure of dends for large holding companies. At regional the scale of its nonbank operations. companies, a positive significant correlation Additional correlation results suggest Table 3 Correlations: Nonbank Involvement, Bank Burdens, and Parent Risk Bank dividend and fee burden ratios 1. Subsidiary bank fees/ subsidiary bank net income 2. Bank fees/equity invest ment in bank subsidiaries 3. Bank fees/bank income paid to parent 4. Bank dividends/ bank net income 5. Bank dividends/equity investment in bank subsidiaries 6. Ratio 1 plus ratio 4 7. Ratio 2 plus ratio 5 M easures of involvement in nonbank activities Investment Equity investm ent in in nonbanking Equity in nonbanks subsidiaries nonbank subsidiaries Equity in all Parent total Parent total subsidiaries assets assets Large Regional Large Regional Large Regional 0 0 0 + 0 0 0 0 0 0 0 + 0 0 0 0 0 0 0 + 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 + 0 0 0 0 0 0 0 0 + 0 + + + + + 0 + 0 + + + + + 0 + 0 0 0 + 0 0 0 + 0 + + + 0 + 0 Parent risk ratios 1. Parent short-term debt/ parent equity 2. Parent total debt/ parent equity 3. Parent double leverage ratio (equity investment in subsidiaries/parent equity) 4. Parent total interest expense/gross income 5. Consolidated net income/parent total interest expense 6. Amount of double leverage/consolidated net income 0 0 + 0 ' ■ 0 0 KEY: + = positive significant correlation (10 percent level, 2-tail test). - = negative significant correlation (10 percent level, 2-tail test). 0 = insignificant correlation. Federal Reserve Bank of Cleveland Investment in nonbanks Investment in all subsidiaries Large Regional 0 + 0 0 0 that, in general, a direct relationship exists between measures of parent company risk and the degree of involvement in nonbank activities at both large and small companies. The greater a parent company’s involve ment in nonbanking operations, the higher is the company’s reliance on debt and interest expenses and the lower is its debt coverage. ment in nonbank activities. Since holding companies generally are engaged in the same types of activities, the variation in nonbank impacts across companies suggests that man agement quality is a critical determinant of nonbank subsidiary performance. This, in turn, suggests that a “typical” impact of a particular type of nonbank activity on any holding company (or holding companies) in general is difficult to predict. The implication III. Conclusion is that regulatory alterations in the number The data deficiencies noted in this article and/or authorized scale of permissible non suggest that the results reported should be banking activities of holding companies viewed with caution. With this in mind, the should be gradual rather than abrupt in either results generally suggest that commercial direction (that is, whether liberalizing or re banking remains the core business of the typ stricting holding company nonbanking activi ical holding company. Involvement in non ties). In particular, entry into more nontrabank activities appears to be relatively limited, ditional fields should be carefully considered. particularly at regional companies. Nonbank profitability varies widely across companies References for reasons that are largely unclear. Large Cates, David. “Analyzing the Parent of a Bank size, however, does not appear to guarantee Holding Company,” Magazine of Bank superior nonbank subsidiary performance. Administration, November 1976. Nonbank profitability also appears to vary considerably over time. Curry, Timothy J. “The Performance of Bank The net impact of these activities on Holding Companies,” in The Bank Hold the level and variability of holding company ing Company Movement to 1978: A Compen returns generally seems to be negligible. These dium , Board of Governors of the Federal findings result from the fact that nonbank Reserve System, 1978. involvement is typically small and that non bank and bank profitability are not markedly Drum, Dale. “The Nonbanking Activities of Bank Holding Companies,” Business Con different and are negatively correlated at roughly half of the companies. The latter find ditions, Federal Reserve Bank of Chicago, March-April 1977. ing does indicate that some companies have obtained some measure of diversification ben Karna, Adi. “Bank Holding Company Prof efits from engaging in nonbank operations. itability: Nonbanking Subsidiaries and Involvement in nonbank operations does not Financial Leverage,” Journal of Bank generally appear to be strongly related to Research, Spring 1979. bank subsidiary fees and dividend burdens but does appear to be positively correlated Wall, Larry D., and Robert A. Eisenbeis. “Risk with parent company leverage. Considerations in Deregulating Bank Activ ities,” Economic Review, Federal Reserve Perhaps the most noteworthy findings of Bank of Atlanta, May 1984. this study are the evident wide variations in the extent of holding company involvement in Whitehead, David D. “Interstate Banking: these activities and the contribution of non Inventory,” Economic Review, bank operations to the holding company organ Taking Federal Reserve Bank of Atlanta, ization. Disaggregated data indicate that a May 1983. considerable number of, but not all, holding companies have derived benefits from involve Economic Review • Spring 1984 The Federal Reserve Bank of Cleveland publishes an infor mative research periodical called Economic Commen tary. Following are the titles published since January 1984. If you are inter ested in receiving this publication, either future or back issues, please contact our Public Information Cen ter, Federal Reserve Bank of Cleveland, P.O. Box6387, Cleve land, OH 44101, (216) 579-2048. Economic Commentary Reorganizing the U.S. Banking Regulatory Structure Sandra Pianalto 4/9/84 The International Debt Situation Owen E Humpage 1/3/84 Deregulation and D eposit Pricing Paul R. Watro 4/23/84 Commercial Bank Holdings of Treasury Debt Gary Whalen 1/16/84 The Economy in 1984: Industry P erspectives Robert H. Schnorbus 5/7/84 Closely Watched Banks Paul R. Watro 1/30/84 Nominal Income Targeting John B. Carlson 5/21/84 Collective Bargaining and Disinflation Mark S. Sniderman and Daniel A. Littman 2/13/84 Seeking a Stable Economic Environment Karen N. Horn 6/4/84 Monetary Policy in the 1980s Karen N. Horn 2/27/84 Regional Interstate Banking Gerald H. Anderson, Thomas M. Buynak, and James J. Balazsy, Jr. 6/18/84 Banking without Interstate Barriers Thomas M. Buynak, Gerald H. Anderson, and James J. Balazsy, Jr. 3/12/84 The Japanese Cost Advantage in Automobile Production Susan A. Loos 7/2/84 The Monetary Targets in 1984 William T. Gavin 3/26/84 Rate Deregulation and D eposit Shifting Paul R. Watro 7/16/84 The Costs of a Protectionist Cure Michael F. Bryan and Owen F. Humpage 7/30/84 Small-Issue IDBs—Tax Policy in Search of a Focus Paul Gary Wyckoff 8/13/84 20 Federal Reserve Bank of Cleveland
@FedFRASER