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_____________ Economic Review ISSN 0013-0281 Winter 1981-82 Economic Review is published quarterly by the Research D epartm ent of the Federal Reserve Bank of Cleveland, P.O. Box 6387, Cleveland, Ohio 44101. Telephone: (216) 579-2000. Editor: Pat Wren. Graphics: Mike Whipkey. Typesetting: Sally Chunat. Opinions stated in the Economic Review are those of the authors and not necessarily those of the Federal Reserve Bank of Cleveland or of the Board of Governors of the Federal Reserve System. M aterial may be reprinted provided th at the source is credited. Please send copies of reprinted m aterials to the editor. W in te r 1 9 8 1 - 8 2 F e d e r a l R e s e r v e B a n k o f C lev ela n d E c o n o m ic R e v ie w C o n te n ts A B asic A nalysis of the N ew P r o te c tio n ism ...........................2 Nations engaged in free international trade experience a long-term rise in their real incomes, but often at the expense of temporary resource-adjustment and income-redistribution problems. During the 1970s the temporary adjustment problems associated with trade seemed particularly acute, and protectionist sentiments increased worldwide. Gerald H. Anderson and Owen F. Humpage illustrate the costs associated with trade restraint, focusing on devices currently used or recently proposed in the United States. Operational Policies of Multibank Holding Com panies . . . 2 0 Motivated by a belief that multibank holding company (MBHC) organizational structure and subsidiary bank performance would be systematically related, several researchers investigated the centralization policies of MBHCs over the past decade. In general, these previous studies revealed that MBHC structural centraliza tion varied widely by policy area, both within and between com panies. However, these studies were completed, for the most part, prior to the period of rapid MBHC expansion and maturation in the 1970s. Further, the samples used to generate the conclusions of these studies generally were small and unrepresentative. For these and other reasons, author Gary Whalen presents and discusses recent data concerning MBHC operational policies. These data are derived from a 1979 survey of the organizational structure of a sample of 65 holding companies in 12 states. Working Paper Review: Multibank Holding Company Organizational Structure and P erfo rm a n ce.....................................................................32 In a working paper summarized here, Gary Whalen explores whether differences in MBHC organizational centralization are sys tematically related to differences in consolidated holding company performance. The sample is cross-sectional, comprising 62 MBHCs whose management responded to a survey of operational policies in November 1979. A Basic Analysis of the N ew Protectionism by Gerald H. Anderson and Owen F. Humpage Since the early 1950s international trade has become an increasingly important component of the U.S. and world economies. In the 1970s the adjustment problems caused by increased world wide competition seemed particularly severe and heightened pressures both in the United States and abroad for protection against imports. The U.S. automobile and steel industries, for example, have advocated trade policies to reduce foreign competition in their industries. While the United States and other countries generally have balked at imposing tariffs and quotas, they have instituted an array of seemingly less bla tant trade policies often referred to as the new protectionism. These measures include both re straints on imports and incentives for exports, and in some cases the costs of these programs are greater than those of tariffs or quotas. This article describes the costs that nations incur by disrupting the free flow of international trade, focusing on the new-protectionist policies that are being substituted for the more conven tional interferences with trade, i.e., tariffs and domestically administered quotas. Such devices as gentlemen’s agreements, buy-American poli- cies, quality standards, subsidies, anti-dumping and countervailing duties, and restrictions on direct investment are described in detail. All of the techniques surveyed are currently used or have considerable support in the United States. It is not our intent to emphasize the differences among the trade restraints and incentives con sidered here, but rather to show the similarities among these trade policies. Both trade restraints and incentives transfer income among groups of individuals and impose net costs on the coun tries involved. Neither is preferable to free international trade. I. Free Trade and the Effects of Trade Barriers Since publication of Adam Sm ith’s Wealth of Nations in 1776, economists have recognized the benefits of free international trade. When one nation specializes in goods that it can produce relatively inexpensively and exchanges these goods for items that another nation produces relatively inexpensively, both nations benefit. The benefits are manifested in lower prices and a wider set of items available for consumption. Consequently, real income is higher with trade than without trade. The benefits of free inter national trade derive initially from differences in comparative costs of production. Specializa Gerald H. Anderson is an economic advisor and Owen F. Humpage is an economist, both with the Federal Reserve Bank of Cleveland. 2 Federal Reserve Bank of Cleveland tion and trade augment these benefits by in creasing the degree of worldwide competition, encouraging international diffusion of new tech nology, and promoting economies of scale in production. The greater the difference between trading partners’ pre-trade price patterns, the greater the gains from free trade. If an economy is to reap the benefits of free international trade, it also must incur the resource-adjustment costs and income-redistribution problems associated with specialization and trade. Trade changes relative prices and forces a reallocation of resources. Over time, a nation engaged in trade experiences further changes in relative costs, technology, and tastes that alter the composition of its exports, im ports, and domestic production. The adjustment does not occur instantaneously, and during the transition some resources—including people— may be unemployed. After the transition, the benefits of specialization and trade are not dis tributed evenly throughout the economy. Trade theory suggests that the factor used intensively in the production of a nation’s import commod ity may experience a decrease in its relative wage as international trade expands. Conse quently, while total real income rises for a nation expanding its trade, some groups within a nation may suffer a reduction in real income. Nations often impose trade barriers in an at tempt to reap the benefits of trade without in curring the associated transition costs and income-redistribution problems. Trade barriers are costly non-solutions, often involving effects not anticipated by their proponents. They arbi trarily transfer income from consumers to do mestic producers, governments, and sometimes foreign exporters and introduce net losses to society resulting from production and consump tion inefficiencies. The income-redistribution effects and net costs of tariffs and quotas, classical protectionist de vices, are shown in figure 1. With the imposition of a per-unit tariff (/), the price of the imported good rises by the amount of the tariff (t = P i - Po). Consumers, who purchase less at the higher price, suffer a loss in purchasing power shown in panel B by the lightly shaded area Po P i f h. 3 Not all of this income loss, however, is a net loss to the home country. Part (area Po Pi e b) is transferred to producers of the protected com modity through the higher prices they charge, and part (area e f g h ) accrues to the domestic government as tariff revenues. In addition to these income transfers, the home country suf fers tariff-induced net losses stemming from greater inefficiency in production and foregone consumption opportunities. These net losses are shown by the darkly shaded areas b e g and d f h, respectively, in figure 1. Quotas have effects similar to tariffs, with one important difference involving area e f g h. In the case of a tariff, the domestic government captures this income in the form of tariff reve nues. In the case of a quota, the amount captured by the home country depends on how import licenses are sold. If import-license fees are set at Pi minus Po dollars per unit, the home country captures the entire area e f g h . On the other hand, if import-license fees are set at less than Pi minus Po per unit, part of the income lost by consumers will accrue to foreign producers and exporters rather than to the home country’s government. Throughout this article it is as sumed that governments always capture these revenues through the sale of import licenses when imposing a quota. Sometimes the consumer of a protected good also is a producer, using the protected item as an input to his own manufacturing process. A tariff or quota on an input raises the cost of producing with that input. A tariff on imported steel, for example, increases the cost of producing domes tic cars and consequently may force domestic car producers to seek protection against foreign car imports. In this way protectionist measures become catalysts of further trade restraint, espe cially along lines of production. II. The N ew Protectionism Realizing that the world is better off when trade flourishes, developed countries have made great advances toward lowering tariffs and lim iting quotas since World War II. Most of the gains were made under the auspices of the Gen- 4 Economic Review □ Winter 1981-82 Fig. 1 Free Trade, Tariffs, and Quotas Panel A Effects of Free Trade Panel B Effects of Tariff or Quota P rice P rice Sw Im p o rts Q u a n tity Free Trade. A simple model of the market for a single traded good is shown in panel A, where line Dd represents the domestic (or home country’s) demand schedule for a traded good, and line Sd represents its supply curve. In the absence of international trade, the home coun try would produce and consume 5 units of the good in question at a price of P2 . The horizontal line Swrepresents the world-supply schedule. A horizontal world-supply curve results from the assumption that the home country’s demand for, and the supply of, the traded good is too small relative to world supply to influence the world price (Po). Under free international trade the home country would consume 9 units of the traded good at the world price of Po. Domestic Im p o rts Q u a n tity producers would supply 1 unit of the good at price P o, and imports would equal 8 units. The benefits of trade accruing to the home country relate to greater efficiency in produc tion and improved opportunity for consump tion. Consumers, for example, are willing to pay the prices given along their demand curve (Dd) for various quantities of the good; with free trade, however, they pay only Po per unit. The price difference for each additional unit pur chased represents a net gain to consumers. The triangular area under the demand curve, bounded by a' c' d \ equals the total tradeinduced net gain to the home country resulting from the improved opportunity for consump tion. Similarly, domestic producers are willing Federal Reserve Bank of Cleveland to supply quantities of the good at prices given by the supply curve (S«0, but a foreign producer can supply the good at a lower price (Po). The price difference for each unit represents a net gain to the home country from improved effi ciency. The triangular area under the supply curve bounded by a' b' c' equals the total tradeinduced net gain to the home country asso ciated with increased efficiency in production. The total area a' b'd', therefore, represents the net gains accruing to the home country from free international trade. In contrast, the area Po P2 a 'b ' does not represent a trade-induced net gain or loss to the home country. This area represents a realloca tion of revenues from domestic producers of the traded good to consumers, resulting from the lower price of the traded good. Consumers use this revenue to finance additional pur chases of the traded good, as well as purchases of other items. The amount spent on additional imports eventually finds its way back to the home country through exports or foreign direct investments. Consequently, resources left un employed as a result of trade in time find employment in other industries, although probably only at lower relative wages (see, for example, Luttrell 1981). Tariffs and Quotas. If the home-country imposes a per-unit tariff of P i minus Po on the traded-good imports, the domestic price would rise to Pi and domestic consumption would fall to 7 units, as shown in panel B. The quantity supplied domestically would rise to 3 units, and imports would fall to 4 units. Consumers now buy less and pay more, but part of their realincome loss is transferred to other sectors of the domestic economy and does not represent a net loss to the home country. The lightly shaded area Po Pi f h represents the total de cline in consumers’ purchasing power. Part of this real-income loss, given by the area Po P i b e, accrues to domestic producers, who now charge P i . This price exceeds the marginal cost of production, given by the domesticsupply schedule, for each unit up to the third unit of the traded good produced. Another por tion of the consumer-income loss, given by area e f g h , goes to the domestic government as tariff revenues. Like other tax receipts, tariff revenues may be used to benefit consumers. Tariffs, nevertheless, impose net losses on the home country (and the world) from in creased inefficiencies in production and fore gone consumption opportunities. The part of the net loss attributable to resources wasted by shifting production to the less efficient domes tic producers is given by the darkly shaded triangular area beg. Foreign producers can produce the second and third units of the traded good for less than domestic producers. The part of the net loss attributable to foregone consumption opportunities is given by the darkly shaded triangular area d f h . Under the free-trade price of Po per unit, consumers would have purchased an eighth and ninth unit of the traded good. Notice, too, that the more price-elastic the home country’s demand and supply, the greater the net losses asso ciated with the tariff. Imposing a quota instead of a tariff alters the conclusions in one important respect. Assume that the government imposes a quota, limiting imports to only 4 units. Again, the free-trade situation depicted in panel A is altered to that shown in panel B, as the domestic market bids up the price of the traded good to P i. The decline in consumers’ purchasing power again is given by area Po P i f h , and triangles b e g and d f h indicate the net losses to the world. The major difference between the tariff and quota involve the area e f g h. In the case of a tariff, this amount accrues to the domestic government as tariff revenues. In the case of a quota, the share of area e f g h obtained by the home country depends on how import licenses are issued. If, for example, the home country sells import licenses at a price of P i minus Po dollars per unit, it would capture the entire area e f g h. Alternatively, if import licenses are sold at a lower price or are given to foreign governments, producers, or exporters, those groups would receive part or all of the income lost by consumers. 5 6 Economic Review □ Winter 1981-82 eral Agreement on Tariffs and Trade (GATT), initially negotiated in 1947 (see box). This period of increasing trade liberalization lasted until the early 1970s, when the Organization of Petro leum Exporting Countries’ price increases and the harsh worldwide recession of 1974-75 created balance-of-payments and employment problems more severe than the typical cyclical patterns. The ensuing climate of slow real growth, high unemployment, and reduced productivity gains created a more difficult environment in which trade-induced resource adjustments could be made. Since the early 1970s, there has been a noticeable increase in pressures for trade protection, as Balassa (1978) and Tumlir (1979) have observed.1 Recent protectionist measures, however, have distinguished themselves from those of previous periods, consequently earning the label new pro tectionism. The most obvious distinguishing characteristic of the new protectionism is its re jection of the tariff and, to a lesser extent, the quota. A GATT inventory completed in 1981 lists more than 600 separate non-tariff barriers that affect industrial-products trade.2 In part, the growing importance of non-tariff trade bar riers can be attributed to the success of GATT. The first six GATT negotiations dealt almost ex clusively with tariff reductions; it was not until the Tokyo Round, concluded in 1979, that GATT began considering non-tariff barriers. The growth of the new protectionism may reflect an attempt by countries to institute trade restraints without destroying GATT as a multinational forum for discussing trade problems. Even with the Tokyo Round, there is ample room for countries to im pose trade barriers. Nevertheless, the prolifer ation of non-tariff barriers is attributable to more than just the relative demise of tariffs and the desire to maintain GATT. 1. K rauss (1978) argues th a t increased protectionism is a natural extension of the growing role of governm ent in the economy. 2. GA T T Focus, February 1982, p. 2. GATT is preparing a corresponding inventory for agricultural products. See also Inventory of Industrial N T B s, U.S. D epartm ent of Com merce, W ashington, D.C., 1973. Under floating-exchange rates, non-tariff trade barriers have the added advantage of providing a more constant degree of restraint than ad valorem tariffs, which are imposed at a fixed percentage of the value of the goods. An exchange-rate change that lowers the home currency price of imports, for example, proportionately reduces tariff duties and the degree of protection the tariff affords. Quantity restraints are insensitive to such ex change-rate movements and consequently have become relatively more attractive than tariffs. Another major reason for the relative growth of the “nontraditional” forms of protectionism is that they are, in some cases, more palatable to for eign producers and their governments than ta r iffs or quotas administered by the home country. Trade-policy negotiations often involve a quid pro quo\ many new-protectionist measures enable for eigners to capture part or all of the economic rents created by trade restraints. As explained in the discussion of gentlemen’s agreements, the quid pro quo often can be directed more closely to affected parties. Another reason for the proliferation of the new protectionism is that the techniques often are not blatantly anti-trade. Krauss (1978) argues that al most anything governments do to intervene in the private sector has consequences for international trade. Loan guarantees, such as those extended to Chrysler Corporation, are designed as domestic employment policies; nevertheless, they increase a firm ’s ability to compete against foreign pro ducers. New-protectionist techniques, such as quality standards, often do not appear as overtly anti-trade as tariffs or quotas, but their effects often are similar. In fact, many of the new-protectionist devices inflict greater net losses on the home country than tariffs or quotas, because they allow for eigners to acquire the economic rents that the home country would capture as tariff revenue or import-license fees. Moreover, many of the newprotectionist measures function as quotas, re stricting the market mechanism more severely than a tariff. Although tariffs distort market out comes, they nevertheless allow shifts in demand (e.g., through changing tastes) and supply (e.g., through efficiency gains) to influence the quan- Federal Reserve Bank of Cleveland 7 General A greem ent on Tariffs and Trade The General Agreement on Tariffs and Trade (GATT) is a multilateral treaty for reducing existing trade barriers and defining rules for conducting international trade. The agree ment also provides a forum for discussing international trade disputes. The principles underlying GATT are widely accepted; rough ly four-fifths of world trade currently is con ducted by nations supporting the agreement. GATT grew out of efforts by the United States during World War II to seek multi national tariff reductions. Concluded in 1947, the first negotiating session, or round, led to the GATT framework as well as significant tariff reductions. Twenty-seven nations par ticipated at that time. In the sixth or Kennedy Round, which was concluded in 1967,53 partic ipants reduced industrial tariffs by two-fifths on average. The seventh or Tokyo Round was concluded in 1979 and marked an important advance in GATT negotiations. For the first time participants specifically addressed non tariff trade barriers. The most important principle underlying GATT is that nations should be nondiscriminatory in conducting trade. Embodied in the “most-favored-nation” clause (Art. I), this principle holds that trade privileges offered to one country should be extended immediately and unconditionally to all nations. The non discrimination principle is primarily respon tity of imports. In this sense a tariff does not totally eliminate the market mechanism. Quotas, as Balassa (1978) observes, do not allow shifts in supply and demand to increase the amount imported. Gentlemen's Agreements An increasingly common form of new protec tionism that is particularly favored in the United States is the gentlemen ’s agreement, whereby an exporting country “voluntarily” restricts ship ments of specific commodities to an importing country. A gentlemen’s agreement essentially is sible for the success of GATT in reducing trade barriers and encouraging international trade. A second important principle holds that trade barriers that are necessary should be in the form of tariffs. Quotas are largely prohibited. Tariffs discourage competition and efficiency somewhat less than quantitative restrictions. A third principle deals with consultation and negotiation: countries should meet under GATT auspices to discuss trade problems, and re straints should be negotiated. Despite GATT, trade restraints do exist. Some restrictive trade practices in existence before GATT were continued under a grand father clause, thought necessary for the agree ment’s successful negotiation. Customs unions and free-trade areas can discriminate against nonmembers. Participants apparently believe that the trade-creating effects of these organi zations exceed the trade-distorting effects. Quantitative restrictions are permissible in cases of balance-of-payments difficulties, and “safeguard” provisions allow nations to insti tute import restrictions on goods that threaten serious injury to domestic producers. Moreover, “voluntary” restraints generally are viewed as consistent with GATT. Import restraints also reflect, to a limited extent, the fact that GATT is not international law. It is a contract ulti mately enforced only by national honor, inter national prestige, and the threat of retaliation. a quota, but one that allows the foreign ex porting country to capture much of the economic rents associated with trade restraint. Conse quently, the costs associated with gentlemen’s agreements typically exceed the costs of an equally restrictive tariff or quota. The rising use of gentlemen’s agreements partially reflects the relative ease with which they can be enacted and the flexibility with which they can be administered. Whereas quotas result from specific legislation and are subject to congressional review, gentlemen’s agreements usually are established by a specific admini stration under its ability to conduct for 8 Economic Review □ Winter 1981-82 eign policy.3Bergsten (1975) suggests that being adm inistrative in origin leaves gentlemen’s agreements open to abuse; others argue that many secret gentlemen’s agreements have re sulted from their not being subject to legislative review (see Morkre and T arr 1980, p. 36). Gen tlemen’s agreements usually are negotiated bi laterally for specific commodities, whereas quo tas are determined unilaterally for protected commodities from all nations. In 1981, for ex ample, the Reagan administration negotiated a gentlemen’s agreement with Japan to restrict automobile imports, but did not seek similar re straints against German, British, French, or Italian car imports. Because of the bilateral aspect of a gentlemen’s agreement, the foreign exporting country has more influence on the agreement terms than under a quota. Often, for example, gentlemen’s agreements allow exporters to borrow in any given year against future limitations or against past, unused limitations. Because gentlemen’s agreements are discrim inatory, they violate the spirit of GATT’s mostfavored-nation clause. Nevertheless, negotiations currently are under way to sanction the use of gentlemen’s agreements under GATT’s “safe guard” provisions. Proponents of this protec tionist device argue that, because gentlemen’s agreements are voluntary, they conform with the GATT philosophy. Exporting countries, how ever, frequently must negotiate gentlemen’s agreements under threat of tariffs or quotas that is less desirable from their perspective. Export ing countries usually do not view gentlemen’s agreements as voluntary. Bergsten (1975) suggests that gentlemen’s agreements may be less restrictive than tariffs or quotas, in part because the exporting country organizes the market and because the incentives to ensure strict compliance are smaller for the 3. Strictly speaking, there are two types of gentlem en’s agreem ents: orderly m arketing agreem ents (OMAs) and voluntary export restrain ts (VERs). OMAs are formalized VERs, authorized under the 1974 T rade Act, subject to spe cific procedural stages (petition, hearing, review) and some tim es to congressional reversal. VERs are established under adm inistrative powers and are not subject to congressional or public scrutiny. See Morkre and T a rr (1980), pp. 35-7. exporting country than for the importing coun try. The enforcement efforts of a foreign govern ment strain its ties with its own business com munity and often require the development of a costly bureaucracy to administer the program. More importantly, however, gentlemen’s agree ments may be less restrictive than tariffs or quotas, as they do not affect all foreign ex porters. Actually, gentlemen’s agreements may stimulate exports from non-restrained foreign producers. In 1977, the Carter administration entered into a gentlemen’s agreement with Japan to restrict U.S. imports of Japanese color tele visions. Morkre and T arr (1980) argue that the restrictions on Japanese televisions contributed to such a large increase in color-television im ports from South Korea, Taiwan, Mexico, Singa pore, and Canada that the total volume of such imports into the United States in 1978 exceeded that of 1977. Restrained exporters sometimes invest in other foreign countries to circumvent protectionist measures; or they sometimes ship goods produced in restrained countries through non-restrained nations to disguise their origin. The substitution phenomenon often results in pressures to negotiate additional gentlemen’s agreements with other producing countries. Bergsten (1975) argues that these non-restrained countries, fearing future gentlemen’s agree ments, may encourage exportation of commodi ties to build a larger export base against which to negotiate. Similarly, restrained countries con tinuously may strive to meet their quotas to prevent future cuts if existing gentlemen’s agree ments are renegotiated. Nations bent on trade restraint theoretically can negotiate and police a network of gentle men’s agreements as restrictive as any quota. In such cases the importing country might incur costs greater than those of an equally restrictive quota. In the case of a quota administered so that the importing country’s government captures all of the export-license fees, the net cost to society associated with restraint is shown in figure 1, panel B, to equal the darkly shaded areas b e ^and d f h . In the case of a gentlemen’s agreement, however, the home country incurs an additional net cost equal to the foregone Federal Reserve Bank of Cleveland import-license fees shown by the lightly shaded area e f g h. These foregone revenues accrue to foreign exporters in the form of economic rents associated with the price increase or to the foreign governments that enter into a gentlemen’s agreement as export-license fees. Although the economic costs of gentlemen’s agreements exceed those of equally restrictive quotas, they offer certain advantages for inter national diplomacy not found in the more tradi tional protectionist measures. A nation institut ing protectionist measures often provides some form of quid pro quo to reduce the likelihood of foreign retaliation. This may take the form of eliminating tariffs or quotas on one commodity while imposing them on another. In the case of gentlemen’s agreements, however, the quid pro quo is automatic, taking the form of the eco nomic rents transferred to foreigners. Through its issuing of export licenses, the exporting government often can direct these rents to the firms most severely affected by the gentlemen’s agreement and simultaneously may gain greater political influence over the industries affected by the gentlemen’s agreement. Small foreign firms unaffected by the restraints naturally find gentlemen’s agreements attractive, but large firms also may find them attractive if the issu ance of export licenses stabilizes competition. The middle-sized foreign firms, growing rapidly in the export market, would be the most severely harmed by gentlemen’s agreements. Buy-American Policies The United States, like many other countries, has a history of procurement policies that dis criminate against foreign suppliers. Govern ments usually adopt such policies to promote self-sufficiency, particularly in the production of military goods or high-technology items, or to offer implicit subsidies to specific producers. The U.S. government, in sharp contrast to most of its foreign counterparts whose discrimination is less open, has stated explicitly its preference for domestic producers and has followed a policy of price favoritism. A domestic price must exceed a foreign price by a specific margin before pro 9 curement officers may buy the foreign item. Until recently, domestic producers were allowed a margin of 6 percent over foreign prices in most circumstances, 12 percent in the case of small businesses, and 50 percent for de fense procurement.4 Many governments follow a more subtle form of discriminatory procurement, called general favoritism. As the name implies, general favor itism takes forms that often are not blatantly discriminatory. A frequent device is the use of selective-tender, or single-tender, bids rather than public-tender bids for government con tracts. Bidding often is announced with little notice or information so that foreigners, un familiar with government-procurement policies, are unable to compete. When high-technology items are involved, governments may develop the procurement specifications in consultation with local firms, limiting the ability of foreigners to compete. There are a myriad of similar ex amples of general favoritism. Buy-American policies, like all protectionist measures, result in transfers from consumers via taxes to producers and net losses to society as a whole. Figure 2 illustrates this for the case when r represents the margin established under price favoritism, Po is the foreign-supply price, and P% equals the domestic, pre-trade supply price. Because P2 is less than P0 (1 + r), no for eign goods are purchased by the home govern ment. The shaded area P0P2 a c represents the total cost of the home country’s policy. Of this amount, the lightly shaded area P0 P2 a b is an income transfer to domestic producers, and the darkly shaded area a b c represents a net cost to society associated w ith inefficiencies in production.5 4. D iscrim inatory governm ent-procurem ent policies were discussed at the Toyko Round of GATT. The participants established a detailed set of rules governing procurement policies, w hich the United States adopted beginning in 1979. As w ith most GATT rules, however, many exemptions exist, particularly in the high-technology and defense sectors. 5. The results would be sim ilar under general favoritism, but r would then represent the per-unit dollar value of the restrain t. Also, figure 2 assum es governm ent demand is perfectly price inelastic; consequently, favoritism results in no losses associated w ith consum ption changes. 10 Economic Review □ Winter 1981-82 Fig. 2 Effects of Buy-American Policies Price Q u a n tity Richardson (1972), however, argues that the effectiveness of buy-American policies in pro viding subsidies to domestic producers and in curbing imports depends on the overall ability to substitute foreign and domestic goods and the degree of competition in the domestic market. If, for example, domestic and foreign goods are pro duced in perfectly competitive markets and are perfect substitutes in consumption, no subsidy or import restraint would result from a buyAmerican policy. The shift in government demand away from foreign-produced goods toward domestically produced goods places up ward pressure on the domestic price and down ward pressure on the foreign price, causing an equal, but opposite, shift in private-sector de mand. The government would purchase more of the domestically produced good, while the pri vate sector would purchase more of the foreignproduced good. The total levels of domestic pro duction and imports, however, remain unal tered, and the domestic price would not rise above the world-market price; no subsidy to domestic producers would result. Quality Standards Governments often impose quality standards on goods and services within their jurisdictions. These include health and safety standards, consumer-protection requirements, and environ mental regulations. Governments usually estab lish such standards to improve the efficient functioning of markets. Ideally, quality stan dards compensate for consumers’ lack of perfect information about products and their effects. Often, however, quality standards create market inefficiency, particularly when they discrimi nate against imported goods by intent or design. There are several ways in which quality standards can discriminate against imports. The most obvious case, of course, is when qual ity standards apply only to imported goods or are more rigorous for imported goods than for do mestic goods. Less obvious, but more important, is when quality standards specify design instead of performance criteria. For example, electrical devices usually require insulation, but there are many different insulating materials that can be used successfully. A design criterion that speci fies a particular material to be used in insulation may discriminate against a comparable foreign product that uses a different material but achieves the same degree of insulation. As a general rule, quality standards, written in terms of performance rather than design criteria, have a more neutral impact on imports. Even when written in terms of performance criteria, quality standards can discriminate against imports if their interpretation and enforcement procedures are not clearly explained to foreign producers. This lack of information raises the risks and perceived costs of trade. Just a delay for inspec tion of goods can be very costly, particularly when it is not anticipated. Goods may set in warehouses awaiting inspection, accumulating storage charges and interest costs, while sea sonal goods may miss their peak selling season. Discriminatory quality standards that reduce imports and raise the prices of traded goods have effects similar to tariffs and quotas. If, for example, a quality standard raises the cost and price of imported goods, the standard results in a Federal Reserve Bank of Cleveland transfer of income from consumers to domestic producers and a net cost associated with con sumption and production inefficiency as shown in figure 1, panel B. In this case the quality standard has the same effect as a specific tariff. If, however, the discriminatory quality standard is so stringent that it precludes foreign competi tion in domestic markets, it serves as a prohibi tive tariff or zero quota. Subsidies Although not viewed as trade restraints, sub sidies affect income distribution and economic efficiency in ways strikingly similar to tariffs and quotas. Subsidies that affect international trade are utilized for production of import sub stitutes, exportation of goods, and production of export goods. Subsidies for the production of import substitutes benefit local producers in direct competition with importers. An export subsidy applies only to units exported; a produc tion subsidy on an export good applies to units sold in the domestic market as well. Subsidies can take many forms. A cash bounty paid to a producer is the most direct and obvious subsidy, but many indirect subsidies are pos sible. Governments might provide services or production facilities at less than market values to producers and might excuse producers from certain taxes. Governments also might subsi dize the production or the import of raw mate rials used by an industry. The Export-Import Bank and Domestic International Sales Corpo rations are the primary means in the United States to subsidize exports, but the federal government also influences the international competitiveness of U.S. firms indirectly through programs such as loan guarantees. Some subsidies, such as direct lump-sum pay ments, are easy to identify and measure, while other subsidies may be harder to detect. If, for example, a foreign government grants a lowinterest loan, the subsidy depends on the differ ence between the actual interest rate and the market rate, but it may not be clear how much the firm would pay for a comparable loan in the 11 free market. Similarly, if there is a subsidy to an industry that supplies raw materials, such as coal, to an exporting industry, such as steel, it may be difficult to measure the benefit passed through to the export industry. Foreign govern ments sometimes provide payments to firms to offset burdens placed on them, such as excise taxes or requirements that they continue opera tion of an inefficient plant to avoid unemployment. The income distribution and efficiency impli cations of subsidies are illustrated by the models developed throughout this article. Figure 3 illus trates the comparative statics of subsidizing the production of an import substitute. The subsidy effectively shifts the domestic-supply schedule, (Sd), downward to S'd, and its per-unit cost is equal to the vertical distance between the two schedules (<5). The subsidy results in an income transfer from all taxpayers to the owners and employees of the subsidized firm. The demand curve remains unaltered under the assumption that all three groups are consumers of the good Fig. 3 Effects of a Production Subsidy for an Import Substitute P rice ........— ■ Im p o rts Q u a n tity 12 Economic Review □ Winter 1981-82 with like preferences. Domestic production ex pands by 1 unit, imports fall by 1 unit, and total consumption is unchanged. The total amount of the subsidy, which is borne by taxpayers, is represented by area a c f e. Of this total, the amount indicated by area b c f e is not a net cost to the home nation but an income transfer to domestic producers. Area a b c, however, repre sents a net cost associated with shifting to the less efficient domestic producer, showing the additional resources required for domestic pro duction compared with foreign production. These resources are bid away from production of some other good. Because export subsidies increase trade, pro ponents view them as opposites of tariffs; how ever, export subsidies are similar to tariffs in that both cause inefficiencies and income trans fers. Figure 4 illustrates the domestic market for an export good. Without a subsidy, world and domestic prices are both Pi. Domestic produc Fig. 4 Effects of an Export Subsidy Price tion is 7 units, of which 4 are consumed at home and 3 are exported. When the government offers a subsidy for exports equal to <5 per unit, pro ducers raise the domestic price to equal the sum of the unchanged world price plus the subsidy. The increase in revenue per unit sold induces additional domestic production, which rises to 9 units. Domestic consumption falls to 3 units because of the price increase, and exports rise to 6 units. This analysis assumes that (1) the exporting nation is a relatively small supplier so that its increase in exports does not change world price, and (2) some barrier to imports en ables the new domestic price P2 to remain above world price Pi. The producers’ profits rise by PiP2 b e. Tax payers bear the cost of the subsidy a b c /, as suming a subsidy is paid for all units exported, and domestic consumers pay P1 P2 a c more for the 3 units that they purchase. Producer profits rise by less than the sum of the subsidy and the additional payments from consumers because of the efficiency cost b e f associated with produc ing goods that foreign consumers value, at the margin, less than the marginal cost of producing them. In addition, there is a loss of consumer surplus a c d caused by the reduction of do mestic consumption. Implicit in the foregoing discussion is the assumption that trade barriers or strong con sumer preferences prevent home-country con sumers from importing the traded good. In the absence of such impediments to trade, the do mestic price would remain at the world price Pi, and 4 units of the traded good would be imported. Domestic producers would continue to produce 9 units; under the incentive of an export subsidy, they would sell their entire stock abroad. The total subsidy would equal P 1 P2 b f. Consequently, an export subsidy may raise imports and ex ports, as producers transfer products to the export market and consumers switch to im ported substitutes. The case of a production subsidy for an export good is illustrated in figure 5. Again, the analysis assumes that the exporting nation is a relatively small supplier so that it cannot influence the world price, which also equals the domestic Federal Reserve Bank of Cleveland Fig. 5 Effects of a Production Subsidy for an Export Good P rice price at Pi. Because the subsidy applies to all units, whether sold at home or abroad, domestic producers view it as a reduction in costs rather than an increase in the world price. It can be represented by shifting the supply curve Sa down by <5, the per-unit subsidy, to Sa. Without the subsidy, production is 7 units, of which 4 are consumed at home and 3 are exported. With the subsidy, production rises to 9 units, domestic consumption is unchanged at 4 units, and ex ports rise to 5 units. Again, taxpayers bear the cost of subsidy a d e f. Producers’ profits rise by c d e f, less than the subsidy. The difference reflects the net cost (a c d) associated with transferring produc tion of the eighth and ninth units from more efficient foreign producers to less efficient do mestic producers. There is no loss of consumer surplus, however, because domestic price and consumption are unchanged. 13 Although both an export subsidy and a pro duction subsidy increase production and ex ports, the export subsidy diverts resources more directly to the export market than the produc tion subsidy. Export subsidies, therefore, are more threatening to producers outside the sub sidizing country, which partially may explain why GATT opposes export subsidies more strongly than production subsidies. The foregoing analysis assumes that in creases in exports caused by export and produc tion subsidies are too small to affect the world price of the product. If such is not the case, then the subsidies would lead to a decline in world price, less benefit to domestic producers, and greater loss to foreign producers. Foreign nations would experience a net benefit, as foreign con sumers gain more than foreign producers lose. This can be understood by observing that the reduction in price of the export good transfers wealth from the subsidizing, exporting country to the importing country. If the subsidy is permanent, the importing country must be a beneficiary of the exporting country’s generosity. A subsidy that is not per manent may be disadvantageous for both the importing country and the exporting country. A subsidy that is not known to be temporary in duces shifts of productive resources among industries that later must be reversed when the subsidy is removed. Thus, a temporary subsidy is always costly for the exporting country and may be costly for the importing country, depend ing on how long the subsidy remains in effect. A nti-D um ping and Countervailing Duties Anti-dumping and countervailing duties are tariffs designed to protect domestic industry at the expense of domestic consumers; they differ from traditional tariffs in that they are imposed only when foreign competition is deemed “unfair.” The anti-dumping law imposes duties on foreign goods sold in the United States at less than “fair value” if such sale causes “material injury” to a U.S. industry. Fair value is the price charged in the exporter’s home country. If the goods are not sold in the exporter’s country, 14 Economic Review □ Winter 1981-82 or are sold there at an unreasonably low price, then a fair value is constructed as the sum of materials and labor costs, general expenses, and profit. General expenses are calculated arbitrar ily as 10 percent of materials and labor costs, and profit is 8 percent of materials, labor, and general expenses. Shipping, insurance, and U.S. tariffs also are added to the fair value. Occasionally, it is argued that anti-dumping duties are necessary to prevent predatory pric ing, i.e., situations in which foreign suppliers initially charge low prices until all competition is eliminated and then charge monopoly prices. To eliminate all competition (home-country and worldwide), a foreign firm would require prodi gious financial strength. Anti-dumping duties result in the income dis tribution effects and efficiency costs described in figure 1. These duties involve administrative costs greater than those associated with tariffs because of the costs of determining fair value and material injury.6 Anti-dumping duties are discriminatory in many ways. They permit low foreign prices that reflect, for example, low for eign wages, but do not allow low foreign prices that result from profit margins under 8 percent. A price may fall below the fair-value definition for a product from one country, but not for the same product from another country if it is pro duced there at lower cost. Anti-dumping duties can deny U.S. consumers access to low-priced goods merely because consumers in the export country pay more for the same goods. The anti dumping law also fails to allow for exchangerate fluctuations. A rise in the value of the exporter nation’s currency would increase the fair-value dollar price of their exports, even if the exchange-rate change is temporary. This would imply that foreign exporters, following a temporary exchange-rate change, would have to 6. Material injury is an im precise term . All im ports, w hether sold at fair value or not, are injurious, to some degree, to the domestic firm s against which they compete. U.S. law does not define m aterial injury, except by saying it m eans “harm w hich is not inconsequential, im m aterial, or u nim portant.” Instead, the law directs the International T rade Commission to determ ine w hether there is m aterial injury in each case. raise prices for their goods sold in this country and risk the loss of markets that had been culti vated over time. The countervailing-duty law is based on the premise that it is unfair to U.S. firms to compete with foreign firms that are subsidized by their governments. If subsidized goods are deemed by the International Trade Commission to be inj urious to a U.S. industry, countervailing duties can be imposed to offset the subsidies. By discour aging subsidized imports, countervailing duties reduce the ability of the United States to benefit from the largess of our trade partners. Anti-dumping and countervailing-duty com plaints are difficult paths for U.S. firms to fol low. An anti-dumping petition, for example, must be specific to a product and firm, so dozens of petitions may be necessary in an industry such as steel, with its many foreign producers and categories of products. To obtain relief, injury as well as dumping must be proved. If dumping occurs during a boom, it is hard to demonstrate injury when output, sales, and profits are rising. If dumping occurs during a recession, it is hard to show that the problem is dumping rather than slack domestic demand. And if the recession ends before the case is com pletely adjudicated, there may be political pres sure to drop the complaint. Improved business conditions would have reduced the injury, and recent administrations generally have disliked anti-dumping cases because of adverse reactions of our trading partners. The trigger price mechanism (TPM) is a recent innovation for enforcing U.S. anti dumping and countervailing-duty laws. Imple mented in 1978 and suspended in January 1982, the TPM established reference prices for steel. Im ports below the reference price would “trigger” an anti-dumping or countervailingduty investigation. Thus, the TPM was not a new device for protection but an administrative mechanism for providing better enforcement of existing laws.7 7. The TPM is described fully in Gerald H. Anderson, “The Steel Trigger Price M echanism ,” Economic Commentary, Federal Reserve Bank of Cleveland, forthcoming. Federal Reserve Bank of Cleveland Barriers to Trade in Services Services, an important component of interna tional trade, also are subject to protectionist barriers. Shipping, banking, insurance, data processing, construction, and consulting are examples of services sold internationally. Ser vices, excluding investment income, accounted for 14 percent of U.S. exports in 1980 and 1981. Restrictions on service imports are wide spread. For example, some nations prevent for eign banks from establishing branches or subsid iaries, while others discourage the use of foreign insurance companies. The United States ex cludes foreign ships from transporting goods between U.S. ports. A catalogue of over 2,000 barriers to free trade in services has been com piled by the U.S. government.8 Barriers to trade in services impose the same costs and distortions as barriers to trade in goods. Often they are rationalized by concern about national defense or independence. How ever, in many cases the true motivation is defense of a domestic industry rather than defense of the nation. GATT does not address the issue of barriers to trade in services. The Office of the United States Trade Representative is working to obtain a commitment from other nations to discuss the matter, but it may be years before the start of any negotiations to reduce these barriers. Protectionism and Foreign Investment The term protectionism usually refers to trade in goods and services, but it also pertains to international movements of capital, especially direct foreign investment.9 Advocates of protec tionist measures directed at stemming capital outflows and encouraging capital inflows often base their arguments on the impacts that such capital flows have on labor. It is alleged that capital outflows (1) reduce labor’s share of real 8. 15 earnings relative to capital’s share by lowering the domestic capital-to-labor ratio, (2) increase domestic unemployment, and (3) reduce unionbargaining powers. Capital inflows have an opposite effect. Empirical results on these issues are mixed, depending crucially on what is assumed to happen in the absence of direct foreign invest m ent.10 Labor organizations, for example, have argued that U.S. direct foreign investments reduce domestic investment, employment, and income, and they have sought to limit such investment by raising taxes on foreign profits (see Humpage 1980-81). Implicit in this argu ment is the view that U.S. firms can compete in foreign markets equally as well through domes tic production and export as through direct for eign investment. This view of international competition, however, is not always valid, par ticularly in the long run. Exporting often proves to be the most profit able way for U.S. firms to conduct international business, but in many cases direct foreign invest ment may be a more profitable, or the only prof itable, means to enter and grow in foreign markets. Often trade barriers or high-transportation costs preclude exporting. Other times, the nature of the product may dictate foreign production rather than exporting. Consumergoods sales, for example, may be highly sensitive to design and style and require a degree of tailor ing to local markets or brand identification not afforded through exporting. Similarly, sophisti cated producers’ goods may require servicing, frequent repair, or adjustments. Direct foreign investment also may enable firms to predict market changes more accurately and to secure access to scarce or low-cost resources that cannot be imported. Exporting may be very profitable if a high degree of product differentiation exists. Over time, however, foreign firms may develop competing products and grow rapidly because of their proximity to the market. In such circum- Wall Street Journal, October 5, 1981, p. 1. 9. Direct foreign investm ent generally refers to capital flows through w hich investors attain a significant degree of control over a foreign firm. 10. See Bergsten, Horst, and Moran (1978) and McCulloch (1979) for brief surveys of the empirical work. 16 Economic Review □ Winter 1981-82 stances, local production may be the only method through which a U.S. company can maintain a share of the foreign market. In the absence of foreign investment, U.S. firms may become less profitable and the growing foreign firm may seek a share of the U.S. market through exporting. Stobaugh (1976) suggests that such circumstances eventually may reduce domestic investment. It is similarly unapparent that direct foreign investment increases U.S. imports or decreases U.S. exports (see Bergsten, Horst, and Moran 1978). Direct foreign investment, particularly initial investments, often occurs in marketing and warehouse facilities to promote the firm ’s exports; often investment occurs in assembly operations and encourages trade in parts. But, even when investment occurs in vertically inte grated foreign plants, it may encourage the export of capital goods for expansion and the export of a multitude of complementary goods and services. Furthermore, multinational firms usually are more adept than their domestic counterparts at finding new export markets. Most important, however, even in cases where U.S. firms invest abroad to import goods back into the United States, the possibility exists that if a U.S. firm did not exploit this opportunity, a foreign firm would. Then restraint on invest ment would not reduce imports. For the same reasons that domestic labor organizations have sought to restrict U.S. invest ment abroad, they have sought to encourage for eign investment in the United States. Foreign firms might be induced to produce in the United States by establishing trade barriers and offer ing tax incentives. Many countries impose localcontent requirements that mandate some per centage of the value of a product that must be added within the country where the goods are sold. For example, automobiles assembled in Mexico must have at least 50 percent of their value added locally. U.S. labor organizations have argued that firms selling in the United States should have some production facilities located here. While a shift from imports to domestic production may increase employment and labor bargaining power, the home country is not necessarily better off. A foreign firm whose exports capture a share of a U.S. market, characterized by imperfect competition but fairly homogeneous products, must have a cost advantage over the domestic firm s’. A cost advantage would be necessary to offset the high risks associated with foreign trade. Such a situation is depicted in figure 6, panel A, where St is the foreign firm s’ supply curve and Sa is the domestic firm ’s aggregate supply curve. The vertical distance between the foreign and domestic supply curves represents the per-unit cost advantage enjoyed by the for eign firm. The market supply and demand curves are shown in panel B. At the initial market price of Pi, 10 units of the commodity in question are sold; 3 units are produced domesti cally, and 7 are imported. Assume that because of threatened trade bar riers the foreign firm decides to produce in the United States and consequently loses part of its competitive advantage. This is shown in panel A by an upward shift of curve Sf to Sf. The aggre gate supply curve in panel B similarly shifts upward from S to S', and the price rises from Pi to P2 . At the new price, 9 units of the good are sold. Consumers now pay a higher amount, indi cated by the shaded area P1 P2 a c in panel B, for these 9 units than before the trade restraints. Of this amount, P1 P2 a b merely represents an income transfer from consumers to the firms and does not represent a net loss to the home country, since the foreign firm is now part of the home country. There are, however, two eco nomic costs associated with the shift from importing to domestic production. As with the tariff and quota examples, there is a loss asso ciated with foregone consumption opportuni ties, equal to the darkly shaded area a c d in panel B. There also is a cost associated with a loss of efficiency in production equal to the darkly shaded area a e f g h i. Of this amount, consumers pay a b c, and b e e f g h i comes out of the producers’ profits.11 11. C orresponding areas in panel A are b ' f g ' h ’ for the total efficiency loss, b' c 'd ' for the consum ers’ share, and c 'd 'f 'g ' h 'fo r the share out of profits. Federal Reserve Bank of Cleveland Fig. 6 17 Effects of Forced Local Production Panel A The Firms Panel B The Market Price Price •S' S Q u a n tity It is also possible that the lack of foreign com petition through imports would increase the bargaining power of organized labor. Some or all of the higher profits associated with the price increase (Pi to P2 ) may transfer to workers through higher wage settlements. The increased labor power could result in further reductions in supply, higher transfers from consumers, and larger net losses to the home country. It would appear, however, that forced domes tic production often does not entail costs as great as those associated with tariffs or quotas. This seems especially true in cases such as localcontent requirements where firms transfer only part of the production process abroad. As the section on free trade indicated, the larger the Q u a n tity pre-trade price differentials among nations, the greater the benefits of international trade. A firm, forced to transfer part of its operations abroad, would shift those that least reduce its pre-trade competitive edge. In attempting to maximize its competitive advantage, the multi national firm would minimize the efficiency losses asso ciated w ith forced-dom esticproduction policies. III. Conclusion There are many justifications offered for im posing trade restraints or allowing trade incen tives. Many are not based on purely economic criteria, and most, particularly as applied to spe 18 Economic Review □ Winter 1981-82 cific U.S. industries, are exercises in sophistry. The distribution among political interests of the benefits and costs of trade policies probably explains a great deal about the imposition and continuation of various trade policies. This arti cle has not attempted to analyze the justifica tions offered for trade restraints; it accepts that governments would continue to restrain trade when it is politically advantageous to do so. The primary concern here has been the sub stitution of trade policies favored under the new protectionism for the more traditional trade pol icies. The new-protectionist trade policies arbi trarily transfer real income away from con sumers to producers of protected commodities. While some groups benefit at the expense of others, the country, nevertheless, incurs a net loss associated with lost consumption opportun ities and increased production inefficiencies. A complete ranking of the relative costs of trade policies has not been attempted, but some general observations can be made. Trade re straints such as gentlemen’s agreements, which allow foreigners to capture a portion of the eco nomic rents that the restraints create, seem more costly than equally restrictive tariffs or quotas administered by the home country, whose government would capture those economic rents through tariff revenue or import-licensing fees. Anti-dumping and countervailing duties are merely special-case tariffs, offered special labels so that they seem less onerous to the public. Similarly, excessive quality restrictions are special-case tariffs or quotas, but no revenue accrues to the home country. Buy-American pol icies also function as tariffs of limited scope, but may be completely ineffective in a competitive market. Production subsidies do not appear as costly as tariffs or quotas, but export subsidies may raise domestic prices, creating pressure for additional import barriers. Forced domestic pro duction generally may be less costly than tariffs or quotas, because multinational firms would shift those operations that least reduce their competitive advantage. In short, the new protec tionism is not “new” in terms of efficiency costs and income transfers. A lemon by any other name would taste as sour. R eferen ces Balassa, Bela. “The ‘New Protectionism’ and the International Economy "Journal of World Trade Law, vol. 12, no. 5 (September:October 1978), pp. 409-36. Bergsten, C. Fred. On the Non-Equivalence of Import Quotas and “Voluntary” Export Re straints. Technical Series Reprint T-009. Washington: Brookings Institution, 1975. Bergsten, C. Fred, Thomas 0 . Horst, and Theodore H. Moran. American Multinationals and American Interests. Washington: Brook ings Institution, 1978, chapters 3 and 4. Heller, H. Robert. International Trade: Theory and Empirical Evidence. 2nd edition. New Jersey: Prentice-Hall, 1973. Humpage, Owen F. “U.S. Taxation of ForeignSource Corporate Income: A Survey of Issues,” Economic Review, Federal Reserve Bank of Cleveland, Winter 1980-81, pp. 1-18. Krauss, Melvyn B. The New Protectionism, the Welfare State and International Trade. New York: New York University Press for the International Center for Economic Policy Studies, 1978. Lowinger, Thomas C. “Discrimination in Gov ernment Procurement of Foreign Goods in the U.S. and Western Europe,” Southern Eco nomic Journal, vol. 42, no. 3 Jan u ary 1976), pp. 451-60. Luttrell, Clifton B. “The Voluntary Automobile Import Agreement with Japan—More Protec tionism,” Review, Federal Reserve Bank of St. Louis, vol. 63, no. 9 (November 1981), pp. 25-30. McCulloch, Rachel. “Trade and Direct Invest ment: Recent Policy Trends,” in Rudiger Dornbusch and Jacob A. Frenkel, Eds., Inter national Economic Policy, Theory and Evi dence. Baltimore: Johns Hopkins University Press, 1979. Federal Reserve Bank of Cleveland Morkre, Morris E., and David G. Tarr. The Effects of Restrictions on United States Im ports: Five Case Studies and Theory. Staff Report of the Bureau of Economics to the Fed eral Trade Commission. Washington: U.S. Government Printing Office, June 1980. Richardson, J. David. “The Subsidy Aspects of a ‘Buy American’ Policy in Government Pur chasing,” in Economics of Federal Subsidy Programs, Part II, U.S. Congress, Joint Eco nomic Committee. Washington: U.S. Govern ment Printing Office, June 11,1972, pp. 220-43. 19 Stobaugh, Robert B. U.S. Taxation of United States Manufacturing Abroad: Likely Effects of Taxing Unremitted Profits. New York: Finan cial Executives Research Foundation, 1976, especially chapter 5. Tumlir, Jan. “The New Protectionism, Cartels, and the International Order,” in Ryan C. Amacher, Gottfried Haberler, and Thomas D. Willett, Eds., Challenges to a Liberal Interna tional Economic Order. Washington: Ameri can Enterprise Institute for Public Policy Research, 1979. Operational Policies of Multibank Holding Companies by Gary Whalen Multibank holding company growth has been rapid since the 1970 amendments to the Bank Holding Company Act of 1956.1 Many states have permitted multibank holding companies (MBHCs) for several years, and recent MBHC authorization in Pennsylvania, Illinois, and West Virginia suggests that this growth will con tinue. Legislators and banking regulators alike are concerned with the impact of holding com pany growth on subsidiary banks, unaffiliated bank competitors, and the convenience and needs of the public. Multibank holding company affiliation gener ally has been expected to alter subsidiary bank behavior relative to independent banks, produc ing a variety of impacts. For example, the affilia tion of an independent bank with a larger hold ing company should allow the subsidiary to realize various types of economies (technical and/or pecuniary economies and/or economies of organization), thereby improving its efficiency relative to comparable non-affiliate banks. Reduced costs could benefit consumers through lower prices and/or higher deposit rates. Access to the greater resources and expertise of the holding company may permit subsidiaries to offer more services than would be possible for independent banks, another public benefit. Since a holding company’s sources and uses of funds are typically more diversified than those of independent banks, and because MBHCs can raise capital more easily and more cheaply than independent banks, an affiliate’s performance may improve after acquisition because its man agement may be able to reduce liquid asset hold ings, increase earning assets, and decrease capi tal relative to total assets. Again, the public may benefit from a greater credit flow into the local area. However, since holding company external expansion results in increased statewide con centration and multi-market linkages, and pos sibly a decline in competition, the performance changes described above may result in private rather than social benefits.2 1. These am endm ents subjected single-bank holding com panies to the sam e set of regulations applied to m ultibank holding companies, encouraging holding companies to ac quire additional banks. In 1965 there were 48 m ultibank holding companies controlling 8 percent of deposits in all commercial banks. By the end of 1970, 111 m ultibank hold ing companies controlled 16.2 percent of the commercialbank deposits. At the end of 1979, 330 m ultibank holding companies existed, controlling approxim ately 15 percent of all banks, 27 percent of all bank offices, and 33 percent of all domestic bank deposits. 2. For a discussion of expected performance im pacts and a sum m ary of empirical research on these issues, see Drum (1976) and Board of Governors of the Federal Reserve Sys tem (1978). Gary Whalen is an economist with the Federal Reserve Bank o f Cleveland. 20 Federal Reserve Bank of Cleveland Many empirical investigations of the impact of MBHC affiliation on bank performance have been undertaken over the past decade. Although numerous hypothetical performance benefits have been identified, a very few modest affilia tion impacts have been discovered. Affiliate asset structures have been found to reflect less liquidity and more risk, as expected. However, while affiliation appears to enhance revenues, subsidiary costs generally are higher than those of independents; hence, the profitability of sub sidiaries is not significantly different from that of independent banks. However, there is evidence suggesting that the methodological approach employed in most of these studies has been responsible for the inability of researchers to discover appreciable affiliation-related performance impacts. Typi cally, researchers have assumed that holding company affiliation perse would alter subsidiary bank performance relative to independent banks. That is, in most empirical studies, all holding companies and holding company affiliate banks are assumed to be homogeneous elements of a single group. Several researchers have suggested that this approach is incorrect and seriously biases the results of these empirical studies. These researchers maintain that the operational policies or organizational structure of a particu lar multibank holding company influences the extent to which hypothetical affiliation impacts are actually manifest (see Weiss 1969 and Law rence 1971). More specifically, these researchers hypothesize that the affiliation impact of any MBHC on its bank subsidiaries is contingent on the extent to which subsidiary bank decisions, policies, and operations are centralized in the hands of the parent corporation or lead bank. The contention that a linkage exists between MBHC structure and performance is important, because several studies of MBHC operational policies have revealed that structural centrali zation varies widely among companies.3Further, several researchers have provided a limited 3. See Weiss (1969), Lawrence (1971), Jesser (1973), Stodden (1975), and the Association of Bank Holding Com panies (1978). 21 amount of empirical evidence suggesting that affiliation impacts differ significantly across MBHCs, implying that MBHC structure and performance might be related.4 One researcher concludes that offsetting performance variations attributable to structural differences may be largely responsible for blurring the impact of MBHC affiliation on bank performance (see Fraas 1974). Only one empirical study of the impact of MBHC organizational structure on subsidiary bank performance has been done to date (see Mayne 1976). Although the study is open to crit icism on several grounds, differences in holding company structural centralization were found to be systematically related to differences in several measures of subsidiary bank perfor mance.5 Thus, existing evidence suggests that further research in this area is warranted. Cur rent data on holding company operational poli cies are required, so that additional, perhaps more reliable, empirical evidence concerning the impact of MBHC affiliation on bank perfor- 4. See Fraas (1974), Hoffman (1976), and Mayne (1976). Mayne alone explicitly tested for the im pact of MBHC stru c tu re on performance, for which she found some sup porting evidence. 5. For example, Mayne tested for stru ctu ra l im pacts on subsidiary bank performance in each year over the 1969-72 interval. She assum ed th at holding companies maximize subsidiary rath er th an corporate-level profitability and so do not attem pt to “cap tu re” affiliation-related benefits through the use of excessive m anagem ent fees. She assum ed MBHC stru ctu ral invariance over this period. Evidence contained in the Association of Bank Holding Companies’(ABHC) study suggests th a t this assum ption may not have been correct. F u rth er, she assum ed th a t the stru ctu ral benefits generated by the stru c tu re in place at the outset of the four-year period were completely realized in this time interval. MBHC seniormanagem ent responses in the ABHC study indicate th at stru c tu re benefits accrue over time w ith a considerable lag. In addition, the average asset size of her sam ple holding company banks was quite sm all (less th an $50 million in average assets). T here is evidence th a t affiliation produces net benefits only after subsidiary size exceeds $40 million in total deposits (see Board of Governors, p. 128). Despite all of these problems, Mayne found some evidence supporting the hypothesized relationship. C urrent empirical evidence con firm ing the existence of an organizational structure-performance relationship for MBHCs appears in Whalen (1982). 22 Economic Review □ Winter 1981-82 mance can be obtained. Such evidence continues to be an essential public policy input. Holding company structural data are of interest for other reasons. Although there have been several pre vious studies of MBHC operational policies, most were completed in the early 1970s, at the outset of the period of rapid holding company expansion. The samples generally were small and unrepresentative.6A current study utilizing a broad sample should indicate the validity of the findings contained in previous studies and allow structural trends to be detected. Further, the factors responsible for observed structural variations have not been adequately explored. Recently collected information on MBHC structural centralization is presented in this paper. In Part I centralization-performance rela tionships are discussed, along with the potential causes of structural variation. Following a brief discussion of the 1970 structural benchmark findings of Lawrence, data on current MBHC operational policies are presented. The method used to derive quantitative indexes of policyarea centralization from the survey responses is discussed in Part III. In Part IV structural varia tions are analyzed using a summary index of MBHC centralization, constructed from the in dividual policy-area centralization measures. The summary and conclusions follow. I. The Structure-Perform ance Relationship Previous studies of MBHC operational policies were based on the premise that centralization of certain decisions and operations in holding com panies would enhance subsidiary revenues and/or reduce costs either directly or indirectly.7 Centralization may allow expensive, indivisible pieces of capital equipment to be fully utilized. For example, average computation costs tend to 6. While Lawrence’s sample consisted of 52 MBHCs nationwide, the largest sample size in the other four studies was 16 companies. 7. A detailed discussion of the potential im pacts of cen tralization in MBHCs appears in the Association of Bank Holding Companies’ study. See also Williamson (1980). fall as the size and power of the computer employed are increased. Thus, centralization of data processing ensures that a large computer system would be optimally utilized and so should permit some economies to be realized by the holding company. Centralization of functions such as asset and/or liability management also should generate economies by allowing speciali zation and division of labor to be fully exploited. Efficient use can be made of parent company staff experts if operations such as securities portfolio management are centralized rather than decentralized. Subsidiary capital and materials costs can be reduced if the larger, more diversified holding company raises the bulk of external funds required by subsidiaries and cen tralizes purchasing. Centralization of budget ing, accounting, and auditing, in conjunction with the operation of a centralized incentive sys tem, provides the parent company with the capabilities to monitor, evaluate, and stimulate the performance of subsidiary personnel. Suboptimization with respect to corporate goals can be detected and prevented. Conversely, in de centralized MBHCs, subsidiary banks essentially operate autonomously; hence, there is no reason to expect their performance to differ appreciably from comparable independent banks. Generally, previous researchers have assumed that the greater the extent of parent company centralization, the greater the MBHC affiliation impacts (see Mayne 1976). That is, the net per formance benefits generated by any MBHC are expected to be positively, although not necessar ily linearly, related to its degree of organiza tional centralization. This view reflects the implicit assumption that gross structural bene fits exceed structurally related “coordination costs” as centralization is increased.8 However, 8. C entralization of any activity in a m ulti-plant firm necessitates “coordination costs.” For example, centralized portfolio managem ent requires th a t information be tra n s m itted at some cost between the parent company and its affiliates. T he q u an tity of inform ation required by cen tralized decision-m akers and costs incurred are likely to depend positively (not necessarily monotonically) on the extent to w hich a given operation is centralized and the num ber and complexity of decisions th a t m ust be made. Federal Reserve Bank of Cleveland Lawrence and others exploring the question of structural variation among MBHCs have em phasized that the net performance benefits gen erated by a particular structural alternative, and so observed structure itself, may vary with certain firm-specific characteristics and/or the nature of the particular holding company’s operating environment.9 These writers have suggested that realizable structural benefits and structurally related coordination costs are af fected by factors such as the total holding com pany size, the relative size of the lead bank and other bank subsidiaries, the company’s infor mation processing capability, the degree of the company’s geographic dispersion, competitive pressures, and other factors that influence the feasibility and/or necessity of adopting a partic ular organizational arrangement. These factors therefore may be related to observed MBHC structural centralization. However, since only a limited amount of theoretical and empirical research on this subject has been done to date, the exact nature of these relationships cannot be reliably specified a priori. II. The L aw rence Study The 1971 study by Robert Lawrence is of in terest for two reasons. First, it was done at the outset of the period of rapid MBHC growth and thus serves as a valuable reference point, par ticularly in light of his representative sample of companies. Second, Lawrence converted his survey data into quantitative indexes of organi zational structure and attempted to explain observed variation in MBHC centralization. Based on survey responses, Lawrence sub jectively determined the extent of parent-company control over, or centralization of, subsidiary bank decisions in certain operational areas ex pected to impact affiliate performance. He clas sified each sample company as centralized, 9. These factors influence the decisional and information requirem ents and hence coordination costs necessitated by a given degree of centralization. See the discussion in Law rence (1971), pp. 7-11; Longbrake(1974), pp. 2-7; the Associa tion of Bank Holding Companies (1978), pp.7-10 and 33-34; and Jessup (1980), pp. 492-94. 23 moderately centralized, or decentralized in each policy area, assigning a score of 1 for a rating of centralized, 2 for moderately centralized, and 3 for decentralized. He constructed an aggregate cen tralization index simply by averaging the policyarea scores. Lawrence concluded that MBHC centralization varied greatly by policy area both within and among companies. He found that holding com panies typically centralized capital management, and the management of subsidiary-bank invest ment portfolios (including federal funds trans actions). He concluded that correspondent re lationships (including loan participations) also were closely controlled. Lawrence reported that relatively less parent control was exercised over subsidiary loan-portfolio management and pric ing decisions. Policies in other areas were not characterized because of structural variation. Based on his aggregate structural index, he categorized 23 percent of his sample companies as centralized, 50 percent as moderately central ized, and 27 percent as decentralized. In an attempt to obtain insight as to the reason for observed structural variation, Lawrence cor related his policy area and aggregate central ization measures with MBHC characteristic variables presumed to affect the net perfor mance benefits generated by a particular struc tural alternative. The variables employed were MBHC deposit size, number of subsidiary banks, geographic extent of MBHC operations, relative size of the lead bank, and age of the holding company. He failed to find any significant rela tionships between these variables and his struc tural indexes. III. The Current Study In November 1979 information on the opera tional policies of a sample of MBHCs was col lected through the use of a management survey, based on the one used by Lawrence. Originally, 102 MBHCs were surveyed. The representative sample consisted of 65 responding companies, located in 12 states. Although the identities of the particular respondents are confidential, sum mary data concerning the sample companies are 24 Economic Review □ Winter 1981-82 presented in table l . 10 As in the Lawrence study, questions were asked concerning the extent of parent-company involvement in, and control over, subsidiary bank decisions in operational areas in which MBHC affiliation generally was expected to impact subsidiary bank performance. The ques tions were designed to allow the researcher to determine the degree of MBHC organizational centralization in each policy area and, ulti mately, to construct quantitative indexes of structural centralization amenable to statistical analysis. Unlike the Lawrence survey, however, the respondents were constrained to answer each question with either a “yes” or “no.” This was done to enhance the response rate and to eliminate the need for the researcher to interpret subjectively the heterogeneous responses inevi tably generated by open-ended questions. Questions were asked about holding company involvement in subsidiary bank management, budget policies, capital management, correspon dent relationships, loan participations, federal funds transactions, securities portfolio manage ment, loan portfolio management, liability man agement, pricing, and “miscellaneous areas.” Several questions related to holding company policies in each of these areas, with the number asked varying over policy areas. Essentially, the greater the expected performance impact of cen tralization of decisions in an area, the greater the number of questions asked. For example, numerous questions concerned MBHC involve ment in loan portfolio and capital management. Fewer questions were asked about the parent com pany’s role in su b sid iary correspon dent relationships. In general, quantitative policy-area centrali zation scores were generated by awarding one “centralization point” in a particular area for each response suggesting parent-company in volvement in that area. Thus, the greater the revealed degree of holding company involvement in any area, the higher the centralization score 10. The average deposit size of the responding companies in 1979 w as approxim ately $1.5 billion, while the average num ber of subsidiary banks and bank offices w as 11.3 and 57.4, respectively. T a b le 1 State C h a r a c te r is tic s o f th e S a m p le N um ber R e Total Total su r Resp onse state state vey ed sp o n ses rate banks d ep osits Alabama Colorado Florida M assa chusetts Michigan 5 4 15 5 4 7 5 5 3 2 60.0 40.0 8.5 2.1 29.7 18.9 M issouri New Jersey Ohio Tennessee Texas 8 7 7 5 18 4 6 4 3 10 50.8 85.7 57.1 60.0 55.5 10.4 11.9 15.6 8.0 6.4 23.3 24.1 24.2 24.8 17.8 Virginia Wisconsin 9 14 8 9 88.8 64.2 30.5 11.5 60.5 17.9 100.0% 19.8% 100.0 19.8 46.7 20.6 52.9% 47.7 30.3 assigned. Using this procedure, structural scores were generated for each respondent in each of the 11 policy areas. Since more questions were asked, more centralization points could poten tially be gained in the key policy areas. Responses to selected survey questions in each of the 11 policy areas are detailed in tables 2 through 12.11 Beneath each table are the poten tial maximum, mean, standard deviation, and range of the constructed policy-area centraliza tion index. The pattern of responses and the structural index data indicate structural ten dencies and variations. The responses shown in table 2 and the low mean policy-area score relative to the potential maximum indicate th at MBHCs generally do not attem pt to control subsidiary bank opera tions through board member and/or officer over lap. The standard deviation of the index is large relative to the mean, indicating a great deal of structural variation between companies. 11. Not all questions asked are reported. T his partially accounts for the divergence between the num ber of ques tions in the tables and the maxim um potential centralization score in any policy area. Additionally, some survey re sponses indicating relatively less centralization resulted in a centralization score of 0.5. Federal Reserve Bank of Cleveland T able 2 MBHC G eneral M a n agem en t Q uestions Total affirmative resp o n ses3 MBHC places its board members on subsidiary bank boards 43 (66.2) MBHC places its officers in subsidiary banks 26 (40.0) MBHC publishes organizational m anual detailing subsidiary management responsibilities 20 (30.8) Policy-area centralization scores Potential maximum Mean Standard deviation Range Table 4 C orrespondent R elationships Q uestions MBHC supplies correspondent services to subsidiaries 57 (87.7) Use of correspondent services required 38 (58.5) Use of correspondent services encouraged 18 (27.7) Correspondent relations w ith non affiliates prohibited Correspondent relations w ith non affiliates subject to MBHC approval 5.5 2.1 1.0 0.0 to 5.5 a. N um bers in parentheses represent percent of re sponding companies. Total affirmative resp o n ses3 Policy-area centralization scores Potential maximum Mean Standard deviation Range 7 (10.8) 30 (46.2) 3.0 1.9 0.8 0.0 to 3.0 a. N umbers in parentheses represent percent of re sponding companies. Table 3 Table 5 Capital M anagem ent Q uestions MBHC sets subsidiary dividends Total affirmative resp on ses3 64 (98.5) Liability Management Questions MBHC periodically reviews liability composition of subsidiaries 62 MBHC advises subsidiary with respect to desired time deposit to total deposit ratio 39 (60.0) 56 (86.2) MBHC monitors subsidiary capital requirem ents and stru ctures 65 (100.0) MBHC solely responsible for ensuring capital adequacy 42 (64.6) MBHC raises all external capital 61 (93.8) MBHC advises subsidiaries on negotiable CDs MBHC decides form in which capital injected into subsidiaries 54 (83.1) MBHC approval required before sub sidiaries issue CDs MBHC makes final decision on subsidiary major capital expenditures 62 (95.4) MBHC informed when subsidiaries issue CDs Policy-area centralization scores Potential maximum Mean Standard deviation Range 10.0 8.9 0.9 7.0 to 10.0 a. Numbers in parentheses represent percent of re sponding companies. Total affirmative responses3 Policy-area centralization scores Potential maximum Mean Standard deviation Range (98.4) 5 ( 7.7) 13 (20.0) 6.0 4.1 1.5 0.0 to 6.0 a. Numbers in parentheses represent percent of re sponding companies. 26 Economic Review □ Winter 1981-82 Table 6 Loan Participations Q uestions Table 7 Federal Funds and D iscount W indow Total affirmative re sp o n ses3 Q uestions Subsidiary originating loan requiring participation is required to offer it to coaffiliate Subsidiary originating loan requiring participation custom arily offers it to coaffiliate Subsidiary participation in loans originated by non-affiliates prohibited Subsidiary participation in loan originated by non-affiliates subject to MBHC approval Subsidiaries m ust inform MBHC w hen participating in loans originated by non-affiliates Policy-area centralization scores Potential maximum Mean Standard deviation Range Total affirmative resp o n ses3 (66.2) Subsidiaries use MBHC or lead bank for federal funds transactions 41 (63.1) (30.8) MBHC approval required before sub sidiaries engage in federal funds transactions 14 (21.5) (30.5) MBHC or lead bank centrally affects federal funds transactions for all subsidiaries 32 (49.2) 14 (21.5) Prior consultation w ith MBHC required before subsidiary borrows at discount window 26 (40.0) 18 (27.7) MBHC approval necessary before subsidiary borrows at discount window 9 (13.8) Policy-area centralization scores Potential maximum Mean Standard deviation Range 6.0 3.0 2.9 0.0 to 6.0 43 20 20 3.5 2.5 0.9 0.5 to 3.5 a. N um bers in parentheses represent percent of responding companies. a. N um bers in parentheses represent percent of re sponding companies. The high proportion of affirmative responses to questions about capital management policies suggests that MBHCs typically exert central ized control over subsidiary bank operations in this area (see table 3). Comparison of the mean policy-area score with the potential maximum confirms this finding. Further, the small size of the standard deviation of the structural index relative to the mean suggests little structural variation in this area. MBHCs appear to exercise somewhat less, but still considerable, control over subsidiary bank correspondent relationships, liability manage ment, and loan participation policies (see tables 4, 5, and 6). Centralization appears to be both greater and less variable in the latter two areas. Centralization policies in the area of federal funds transactions follow a similar pattern (see table 7). MBHCs appear to exercise at least a moderate amount of control over subsidiary bank operations in this area. However, the cen tralization index standard deviation is large rel ative to the mean, indicating structural varia tion. Further, responses in this area constitute evidence that MBHC centralization generally has increased since Lawrence’s study. None of Lawrence’s sample companies reported that advance approval was required before a sub sidiary bank could engage in federal funds trans actions; 14 MBHCs indicated that such approval was required (see Lawrence 1971). Security portfolio management appears to be relatively centralized (see table 8). Virtually all companies report substantial involvement in this area. This is reflected by the proximity between the index mean and potential max imum. Structural variation does not appear to be considerable. Federal Reserve Bank of Cleveland T able 8 S e c u r itie s P ortfolio M an a g em en t Q uestions MBHC offers advice on securities investm ents to affiliates MBHC suggests proportion of funds allocated to investm ents vs. loans MBHC specifies maximum or mimmum am ount of state and local issues to be held MBHC handles buy and sell orders for all subsidiaries’ securities transactions MBHC handles securities portfolio m anagement entirely for subsidiaries Total affirmative resp on ses3 62 (95.4) 55 48 45 38 (84.6) (73.8) (69.2) 24 (36.9) MBHC prohibits subsidiaries from obtaining portfolio advice from non-affiliates 49 (75.4) Questions Total affirmative responses3 MBHC periodically reviews schedule of subsidiary bank-loan interest rates 50 MBHC suggests pricing loans at floating vs. fixed rates of interest or vice versa 53 (81.5) MBHC offers advice on service changes, deposit interest rates, fees, etc. 55 (84.6) MBHC makes final decision on prices charged at subsidiary banks 10 (27.8) (76.9) Prices set uniformly throughout MBHC system 6 ( 9.2) Policy-area centralization scores Potential maximum Mean Standard deviation Range 7.5 3.7 1.7 0.0 to 7.5 a. Numbers in parentheses represent percent of responding companies. 11.0 8.2 2.9 0.0 to 11.0 a. Numbers in parentheses represent percent of responding companies. MBHCs appear to exert a moderate amount of control over subsidiary bank pricing decisions (see table 9). In particular, holding company policies in this area appear to be much more cen tralized in 1979 than they were in 1970. Lawrence rated companies as centralized in this area (16 of 52 companies) if they only reviewed subsidiary bank interest rates. In contrast, 50 of 65 com panies in 1979 reviewed subsidiary interest rates, and 10 reported that the holding company made the final decision on the prices to be charged. MBHC control over subsidiary budget policies appears to be highly centralized (see table 10). Pricing Policies (58.5) MBHC manages security holdings of subsidiaries collectively, as single portfolio Policy-area centralization scores Potential maximum Mean Standard deviation Range Table 9 27 Little structural variation in this area is evident. The responses in table 11 suggest that MBHCs exert at least a moderate amount of centralized control over subsidiary bank loan portfolio management. The evidence indicates that MBHC policies in this area can no longer be cate gorized as relatively decentralized as they were by Lawrence. Finally, centralization of other operations and services appears to vary (see table 12). Data pro cessing, auditing, and purchasing seem to be centralized. Most companies also appear to operate a centralized incentive system. Trust activities, purchasing, and accounting seem to be centralized less often. Centralization seems to have increased in some of these areas over the decade as well. For example, only 38 percent of Lawrence’s sample indicated that trust opera tions were centralized in 1970; 58 percent of the 28 Economic Review □ Winter 1981-82 sample reported centralization in this area in 1979 (see Lawrence 1971). IV. Structural Variation Structural variation was explored by regress ing an aggregate measure of MBHC centraliza tion on variables expected to influence the net benefits attributable to a particular organiza tional alternative. The final form of the esti mated regression appears in table 13.12 The summary centralization index, CT, was formed by simply summing the 11 policy-area central ization scores.13 The mean and standard de viation of this measure, as well as those of the independent variables employed, appear in the table. Comparison of the standard deviation of this structural measure relative to its mean suggests that aggregate centralization, as well as policy-area centralization, varies widely among holding companies. The /-statistics in table 13 indicate that the coefficients of all of the independent variables used in the regression are significant. Similarly, the significant F-statistic suggests that the explanatory power of the estimated equation is considerable. The regression results reveal a quadratic sizecentralization relationship, centralization rising with MBHC size until holding company total deposits reach approximately $2.2 billion. Beyond that point, size and centralization are inversely related. Evidently, structurally related “coordination costs” rise relative to realizable 12. Additional explanatory variables (e.g., a geographic dispersion variable, a grow th rate variable, and holding company age) were employed in prelim inary runs. All were insignificant and so were dropped from the estim ated equation. 13. An equally w eighted sum m ary index also was con structed by sum m ing the 11 policy indexes after each had been divided by the policy area potential m aximum . T hus, all policy area indexes were scaled to vary between 0 and 1. The equally w eighted sum m ary index was highly correlated w ith C T (the correlation coefficient w as approximately 0.9), and regression resu lts were virtually identical w hen this m easure was employed as the dependent variable. T hus, these results were not reported. T a b le 1 0 B u d g e t P o lic ie s Q uestions Total affirmative re sp o n ses3 MBHC subsidiaries m ust subm it budgets 63 (96.9) MBHC approval of subsidiary budgets required 56 (86.2) MBHC monitors budget variances 63 (96.9) MBHC evaluates subsidiary CEOs relative to budgeted figures 59 (90.8) Policy-area centralization scores Potential maximum Mean Standard deviation Range 5.0 4.5 1.0 0.0 to 5.0 a. N umbers in parentheses represent percent of re sponding companies. structural benefits as MBHC size increases, ceteris paribus. A negative relationship was discovered be tween the subsidiary size configuration vari able, H, and MBHC centralization. The more equal the sizes of its subsidiary banks, the lower the value of H. Thus, subsidiary size equality and the degree of holding company central ization are positively related. Centralization thus may be less costly and/or more necessary when all bank subsidiaries are relatively homo geneous in terms of size. The positive coefficient of the LBSIZE vari able suggests that the larger the lead bank rela tive to the other holding company banks, the greater the degree of MBHC centralization. Pre sumably, the large lead banks have extensive managerial and financial resources and can pro vide strong support to smaller affiliates, all of which could lead to some form of centralization. The coefficient on the information-processing dummy MISDUM was also positive, indicating that holding company centralization and information-processing sophistication are directly related. Operation of a management-information system evidently lowers structurally related “coordination costs,” ceteris paribus. Federal Reserve Bank of Cleveland Table 11 Table 12 Services Loan Portfolio M anagem ent Questions M iscellaneous Centralized Total affirmative responses8 Questions MBHC publishes loan policies binding on all affiliates 30 (46.2) Total affirmative responses8 MBHC has centralized incentive program 52 (80.0) 20 (30.8) MBHC has centralized systems and procedures group 41 (63.1) MBHC suggests a loan-to-deposit ratio for each bank 55 (84.6) MBHC has central purchasing department 40 (61.5) MBHC advises affiliates on maturity structure of loan portfolios 42 MBHC has central printing department 27 (41.5) MBHC advises affiliates on “mix” of loan portfolios 59 (90.8) MBHC centralized data processing available 61 (93.8) Use of MBHC data-processing facilities required 52 (80.2) MBHC has central audit department 62 (95.4) MBHC has centralized trust operations 38 (59.5) Demand deposit/time deposit ac counting centralized 35 (53.8) General ledger accounting centralized 50 (76.9) MBHC has management information system 35 Subsidiary banks must publish loan guidelines subject to holding company approval MBHC monitors amount of loans made at floating vs. fixed rates of interest MBHC advises subsidiaries on non-price loan terms Subsidiaries inform MBHCs when lines of credit extended MBHC requires approval of subsidiary credit-line extensions Policy-area centralization scores Potential maximum Mean Standard deviation Range (64.6) 45 (69.2) 51 (78.5) 24 (36.9) 9 (13.8) 11.5 7.4 2.4 0.5 to 11.5 a. N um bers in parentheses represent percent of re sponding companies. Finally, the lim ited-area and statew ide branching dummy variables exhibit positive coefficients, indicating that branching freedom and MBHC centralization are directly related. This finding can be interpreted as evidence of a positive association between competitive pres sures and holding company centralization, as suming the intensity of competition and branch ing freedom are directly related. V. Sum m ary and Conclusion The survey findings indicate that MBHC organizational structures remain diverse. Oper 29 Policy-area centralization scores Potential maximum Mean Standard deviation Range (53.8) 34.5 22.1 6.7 4.0 to 33.0 a. N um bers in parentheses represent percent of re sponding companies. ational policies appear to vary within and among companies by policy area. In general, MBHCs typically exercise relatively centralized control over subsidiary bank budgets, capital manage ment, and securities and loan portfolio manage ment policies. Somewhat less, but still con siderable control is exercised in the areas of correspondent relationships, loan participations, federal funds transactions, liability management, 30 Economic Review □ Winter 1981-82 Table 13 Variable R egression of Sum mary Centralization Index on P otential D eterm inan ts3 SIZE SIZESQ H -l.ixio-6 LBSIZE MISDUM BR D U M l BRDUM 2 Constant 65.01 0.0049 -34.5655 1.6309 8.0987 6.0864 10.3847 /-Statistic 1.87* 2.66** 3.36** 3.42** 2.89** 1.88* 2.68** Mean 1.49X103 4.39X106 0.411 3.11 0.539 0.565 0.226 10.61X106 0.212 5.83 0.502 0.500 0.422 Standard deviation 'fe X 00 Coefficient F R2 7.29 0.42 a. T he mean and the standard deviation of the sum m ary centralization index, CT, are 68.7 and 13.05, respectively. * Significant at 10 percent level—two-tail test. ** Significant at 5 percent level—two-tail test. V ariables U sed in Estim ated Equations: CT: SIZE : SIZE SQ : Sum of 11 policy-area centralization indexes. Holding company total deposit size, year-end, 1978 (in millions). The square of SIZE. H: Subsidiary size variation m easure. The sum of the squared shares of holding company total deposits held by each subsidiary bank. LB SIZE : Total deposits held by the lead holding company bank relative to total deposits in all other subsidiary banks. M ISD U M : Equal to 1 if holding company reported it had a management-information system; equal to 0 otherwise. B R D U M l: Equal to 1 if holding company operated in limited branching state; equal to 0 otherwise. BRDUM2: Equal to 1 if holding company operated in a statew ide branching state; equal to 0 otherwise. and pricing. It should be noted that in the empir ical studies in which structural differentials have been ignored, MBHC affiliation is consist ently found to impact subsidiary bank asset allocation and leverage. This is suggestive evi dence that centralization in the areas of capital management and securities and loan portfolio management is responsible for these differences in performance. Further, the data reveal that MBHC centralization, on average, has increased over the last decade. This trend also suggests that centralization and performance may be pos itively related. The evidenced levels of centrali zation indicate that MBHCs may be effective substitutes for branch systems. The regression results provide some insight on the possible causes of observed structural variations between different holding companies. Counter to the earlier findings of Lawrence, it seems that MBHC structural policies can be estimated or predicted from available data. The evidence indicates that MBHC total size and centralization become inversely related after a critical size threshold is reached. Assuming that parent company centralization has a sys tematic effect on subsidiary bank performance, the implication of this finding is that increases in statewide concentration and multi-market linkages resulting from external expansion of the largest holding companies may not be a cause for concern, since these companies typi cally exercise less centralized control over the operations of their bank affiliates. The structural information presented in this study clearly implies that it is incorrect to view all holding companies and their bank subsidiaries as homogeneous elements of a single group, the standard practice in previous affiliation impact studies of holding companies. If MBHC struc ture does influence performance, reliable empir ical evidence on the affiliation impacts of hold ing companies can only be obtained from studies in which differences in holding company cen tralization are explicitly taken into account. Federal Reserve Bank of Cleveland R eferen ces Association of Bank Holding Companies. Bank Holding Company Centralization Policies. Washington, D.C.: Golembe Associates, Inc., February 1978. Board of Governors of the Federal Reserve Sys tem .“The Bank Holding Company Movement to 1978: A Compendium.” Staff Study. Wash ington, D.C.: Board of Governors, September 1978. Drum, Dale S. “MBHCs: Evidence after Two Decades of Regulation,” Business Conditions, Federal Reserve Bank of Chicago, December 1976, pp. 3-15. Fraas, A rthur G. “The Performance of Individ ual Bank Holding Companies.” Staff Study No. 84. Washington, D.C.: Board of Gover nors of the Federal Reserve System, 1974. Hoffman, Stuart G. “The Impact of Holding Company Affiliation on Bank Performance: A Case Study of Two Florida Multibank Hold ing Companies.” Working Paper Series. Fed eral Reserve Bank of Atlanta, January 1976. Jesser, Edward A., Jr., and Kenneth H. Fisher. “Guidelines for Bank Holding Company Management,” Bankers Magazine, vol. 156, no. 3 (Summer 1973), pp. 13-20. 31 Lawrence, Robert J. “Operating Policies of Bank Holding Companies: Part I.” Staff Study No. 59. Washington, D.C.: Board of Governors of the Federal Reserve System, April 1971. Longbrake, William A. “The Differential Effects of Single-Plant, Multi-Plant, and Multi-Firm Organizational Forms on Cost Efficiency in Commercial Banking.” Working Paper No. 74-7. Washington, D.C.: Federal Deposit In surance Corporation, 1974. Mayne, Lucille S. “Management Policies of Bank Holding Companies and Bank Performance,” Journal of Bank Research, vol. 7, no. 1 (Spring 1976), pp. 37-48. Stodden, John R. “Survey of the Operating Poli cies and Performance of Texas Multibank Holding Companies: Preliminary Findings,” in Conference on Bank Structure and Competi tion, Federal Reserve Bank of Chicago, 1975. Weiss, Steven J. “Bank Holding Companies and Public Policy,” New England Economic Re view, Federal Reserve Bank of Boston, Janu ary/February 1969, pp. 3-29. Whalen, Gary. “Multibank Holding Company Organizational Structure and Performance.” Working Paper 8201. Federal Reserve Bank of Cleveland, March 1982. Williamson, Dan. “Dozens of Portfolios—One Investment Policy,” ABA Banking Journal, vol. 72, no. 2 (February 1980), pp. 56-66. Gary Whalen, “Multibank Holding Company O rg an ization al S tru ctu re and P er fo r m an ce,” Working Paper 8201, March 1982, 30 pp. Bibliography. Multibank holding company growth has been rapid since 1970. Recent passage of legislation au thorizing multibank holding companies (MBHCs) in Pennsylvania, Illinois, and West Virginia and similar proposals in other states suggest that this growth will continue. MBHC affiliation generally has been expected to alter the perfor mance of subsidiary banks relative to indepen dent banks, thereby impacting consumer welfare and bank competition and soundness. The net social benefits produced by holding company expansion are debatable. Many empirical investi gations of the impact of MBHC affiliation on bank performance have been undertaken over the past decade. Contrary to a priori expecta tions, very few modest affiliation impacts have been discovered. Affiliate asset structures have been found to reflect less liquidity and more risk. While affiliation appears to enhance revenues, subsidiary costs generally are higher than those of independents; hence, their profitability is not sig n ifican tly different from th a t of independent banks. There is evidence suggesting that the metho dological approach employed in most of these studies has been responsible for the failure of researchers to discover appreciable affiliationrelated performance impacts. Generally, re searchers have assumed that holding company affiliation per se would alter subsidiary perfor mance relative to independent banks. Several w riters have countered that MBHC operational policies—and more specifically, the extent to which subsidiary bank decisions and operations are centralized in the hands of the parent corpo ration or lead bank—influence the degree to which hypothetical affiliation impacts are actu ally manifest. These researchers have supplied a considerable amount of empirical evidence suggesting that MBHC structural centralization varies widely among companies. The implica tion is that offsetting performance variations attributable to structural differences may be responsible for blurring the impact of MBHC affiliation on bank performance. Researchers have suggested that MBHCs may attempt to maximize corporate rather than subsidiary-level performance; hence, the parent may attem pt to “capture,” either totally or par tially, affiliation benefits realized by bank sub sidiaries through the use of intra-company reve nue transfers. If this were the case, beneficial affiliation impacts, particularly lower costs re sulting from scale economies, would not be detectable at the subsidiary bank level. This study represents an attem pt to deter mine empirically w hether differences in MBHC organizational centralization are systematically related to differences in bank affiliate perfor mance and, through this channel, consolidated holding company performance. The sample is cross-sectional, consisting of 62 MBHCs whose management responded to a survey of opera tional policies in November 1979. Quantitative indexes of the organizational centralization of the sample companies, derived from the survey data, were related to summary measures of hold ing company profitability, and a positive, signif icant relationship was found. Given that MBHC organizational structure varies considerably, the analysis implies that it is inappropriate to treat all holding companies and their subsid iaries as members of a single, homogeneous group. Public policy governing future intra- and interstate expansion by MBHCs should be guided by empirical evidence obtained from studies in which differences in MBHC organizational structure are explicitly taken into account. Copies of this working paper are available from the Federal Reserve Bank of Cleveland, Research Department, P.O. Box 6387, Cleveland, OH 44101. 32