The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
____________ Review____________ Vol. 66, No. 4 April 1984 5 International Banking Facilities 12 A Private Central Bank: Some Olde English Lessons 23 Money Growth Variability and GNP The Review is published 10 times per year by the Research and Public Information Department o f the Federal Reserve Rank o f St. Louis. Single-copy subscriptions are available to the public free o f charge. Mail requests fo r subscriptions, back issues, or address changes to: Research and Public Information Department, Federal Reserve Rank o f St. Louis, P.O. Rox 442, St. Louis, Missouri 63166. The views expressed are those o f the individual authors and do not necessarily reflect official positions o f the Federal Reserve Rank o f St. Louis or the Federal Reserve System. Articles herein may be reprinted provided the source is credited. Please provide the Rank’s Research and Public Informa tion Department with a copy o f reprinted material. Federal Reserve Bank of St. Louis Review April 1984 In This Issue . . . In the first article in this Review, “International Banking Facilities,” K. Alec Chrystal gives a brief outline o f the regulatory changes that permitted banking institutions in the United States to establish international banking facilities (IBFs) and conduct international banking business on terms similar to offshore financial centers. Chrystal describes the location and growth o f IBFs and details the extent to which business has been shifted from other accounts and other locations. In addition, he explains the nature o f IBF business as an integral part o f the eurodol lar market. Chrystal concludes that, apart from a change in geographical location, the significance o f IBFs in international banking is small. In the second article, “A Private Central Bank: Some Olde English Lessons," G. J. Santoni investigates the proposition that central bankers respond systematically and predictably when faced with different incentives relating to monetary control. The author focuses on the actions o f the Bank o f England from 1694 to 1930. The Bank o f England provides both an interesting and an apt case study for two reasons. First, in contrast to modern central banks, England's central bank was a privately-owned, for-profit institution from its founding in 1694 until the early 1930s. Further, the bank was structured such that the wealth of the Bank’s owners was inversely related to the rate o f inflation. Second, the government seized the monetary control function from the private owners o f the bank from 1793 to 1821. This period is important because it demonstrates clearly h ow different incentives produce different policy consequences, holding other important institutional factors roughly constant. Santoni concludes that the incentives confronting England’s p rivate central bankers led them to choose relatively low rates o f monetary growth and inflation; in contrast, the government chose significantly faster rates o f monetary growth and inflation when it controlled the Bank o f England. In the third article, "Money Growth Variability and GNP,” Michael T. Belongia explains the sequence o f events through which short-run fluctuations in M l growth make future econom ic conditions more uncertain and, thus, cause reduc tions in the growth o f GNP. Belongia investigates the impact of variable m oney growth on GNP by estimat ing a variant o f the St. Louis equation that incorporates directly a measure o f the variability o f money growth. He finds that the effects o f variable M l growth were negative, but transitory in nature, prior to 1980. Since 1980, however, the variability o f M l growth has tripled, and its negative effects on the path o f GNP have become permanent. In a simulation experiment, the author shows that the growth of nominal GNP w ould have been higher and more stable between 1979 and 1982 if M l growth had been less variable over that period. 3 International Banking Facilities K. Alec Chrystal I n TERNATIONAL Banking Facilities (IBFsl started operation in the United States in early December 1981. Since then, they have grown to the point where they now represent a significant part o f the international banking business worldwide. The purpose o f this arti cle is to examine IBFs and to discuss their significance for international banking. OFFSHORE BANKING A substantial “offshore” international banking sec tor, often called the “eurocurrency” market, grew up in the 1960s and 1970s. Its key characteristic is that bank ing business is transacted in a location outside the country in whose currency the business is denom i nated. Thus, eurodollar transactions are conducted outside the United States, eurosterling transactions are conducted outside Britain, and so on. Much o f this offshore business occurs in major financial centers like London, though some business is literally in islands offshore from the United States, such as the Bahamas or Cayman Islands. Offshore banking business is somewhat different from that conducted onshore. Though, in both cases, banks take deposits and make loans, offshore banks have virtually no checking deposit liabilities. Instead, their deposits are typically made for specific periods of time, yield interest, and are generally in large denom inations. Offshore banking arose as a means to avoid a variety of banking regulations. For example, offshore banks that deal in eurodollars avoid reserve requirements on K. Alec Chrystal, professor of economics-elect, University of Shef field, England, is a visiting scholar at the Federal Reserve Bank of St. Louis. Leslie Bailis Koppel provided research assistance. deposits, FDIC assessments and U.S.-imposed interest rate ceilings. The first two of these regulations increase the margin between deposit and loan rates. Avoiding these costs enables offshore banks to operate on much smaller margins. Interest ceilings, where binding, re duce the ability o f banks subject to such ceilings to compete internationally for deposits. Many "shell” bank branches in offshore centers, such as the Caymans and Bahamas, exist almost solely to avoid U.S. banking regulations. Shell branches are offices that have little more than a name plate and a telephone. They are used simply as addresses for book ing transactions set up by U.S. banks, which thereby avoid domestic m onetaiy regulations. IBFs: ONSHORE OFFSHORE BANKS IBFs do not represent new p h y sica l banking facilities; instead, they are separate sets o f books within existing banking institutions — a U.S.-chartered depository in stitution, a U.S. branch or agency o f a foreign bank, or a U.S. office o f an Edge Act corporation.1 They can only take deposits from and make loans to nonresidents of the United States, other IBFs and their establishing entities. Moreover, IBFs are not subject to the regula tions that apply to domestic banking activity; they avoid reseive requirements, interest rate ceilings and deposit insurance assessment. In effect, they are accorded the advantages o f many offshore banking centers without the need to be physically offshore. 'As a result of a 1919 amendment to the Federal Reserve Act initiated by Sen. Walter Edge, U.S. banks are able to establish branches outside their home state. These branches must be involved only in business abroad or the finance of foreign trade. The 1978 Interna tional Banking Act allowed foreign banks to open Edge Act corpora tions which accept deposits and make loans directly related to inter national transactions. 5 FEDERAL RESERVE BANK OF ST. LOUIS The Establishment o f IBFs Three regulatory or legislative changes have permit ted or encouraged the establishment and growth of IBFs. First, the Federal Reserve Board changed its reg ulations in 1981 to permit the establishment o f IBFs. Second, federal legislation enacted in late 1981 ex empted IBFs from the insurance coverage and assess ments imposed by the FDIC. Third, several states have granted special tax status to the operating profits from IBFs or altered other restrictions to encourage their establishment. In at least one case, Florida, IBFs are entirely exempt from local taxes. Restrictions on IBF Activities While IBFs may transact banking business with U.S. nonresidents on more or less the same terms as banks located offshore, they may not deal with U.S. residents at all, apart from their parent institution or other IBFs. Funds borrowed by a parent from its own IBF are subject to eurocurrency reserve requirements just as funds borrowed from an offshore branch would be. Four other restrictions on IBFs are designed to en sure their separation from domestic money markets. First, the initial maturity o f deposits taken from non bank foreign customers must be at least two working days. Overnight deposits, however, may be offered to overseas banks, other IBFs and the parent bank. This restriction ensures that IBFs do not create a close substitute for checking accounts. Second, the minimum transaction with an IBF by a nonbank customer is $100,000, except to withdraw in terest or close an account. This effectively limits the activity o f IBFs to the “wholesale” money market, in which the customers are likely to be governments, m ajor corporations or other international banks.2 There is no restriction on the size of interbank transac tions. Third, IBFs are not permitted to issue negotiable instruments, such as certificates o f deposit (CDs), be cause such instruments w ould be easily marketable in U.S. money markets, thereby breaking down the in tended separation between IBFs and the domestic money market. Finally, deposits and loans o f IBFs must not be re lated to a nonresident custom er’s activities in the 2Foreign governments are treated like overseas banks for purposes of maturity and transaction size regulations. 6 FRASER Digitized for APRIL 1984 United States.'* This regulation prevents IBFs from competing directly with domestic credit sources for finance related to domestic econom ic activity. Where Are IBFs Located? IBFs are chiefly located in the major financial centers (see table 1). Almost half o f the nearly 500 IBFs are in N ew York; California, Florida and Illinois have the bulk o f the rest. In terms o f value o f liabilities, however, the distribution is even more skewed. O f IBFs reporting monthly to the Federal Reserve (those with assets or liabilities in excess o f $300 million), 77 percent o f total liabilities were in New York, with California (12 percent) and Illinois (7.5 percent) a long wav behind. It is nota ble that Florida, which has 16.5 percent of the IBFs, has only 2 percent o f the liabilities o f reporting banks. While the distribution o f IBFs primarily reflects the preexisting locations o f international banking busi ness, differences in tax treatment between states may have influenced the location o f IBFs marginally. For example, the fact that Florida exempts IBFs from state taxes may well explain w hy it has the largest number of Edge Act corporation IBFs and ranks second to New York in terms o f numbers o f IBFs set up bv U.S.chartered banks. Although Florida has the most advantageous tax laws possible for IBFs, it is not alone in granting them favorable tax status. Nine other states (N ew York, Cali fornia, Illinois, Connecticut, Delaware, Maryland, Georgia, North Carolina and Washington) and the Dis trict o f Columbia have enacted special tax laws that encourage the establishment o f IBFs.4 The reason for the favorable tax treatment for IBFs in states like Florida is not clear. There is no doubt that Florida has tried to encourage its development as an international financial center.5 The benefits from en couragement o f IBFs per se, however, are hard to see. For example, the employment gains are probably triv ial. Since IBFs are merely new accounts in existing institutions, each IBF will involve at m o st the em ploy ment o f a handful of people. In many cases, there may be no extra employment. 3“The Board expects that, with respect to nonbank customers located outside the United States, IBFs will accept only deposits that support the customer's operations outside the United States and will extend credit only to finance the customer's non-U.S. operations.” See “Announcements” (1981), p. 562. 4These provisions vary from case to case. For a summary of the position in New York and California, see Key (1982). 5See "Florida's Baffling Unitary Tax" (1983). APRIL 1984 FEDERAL RESERVE BANK OF ST. LOUIS Table 1 Location of International Banking Facilities Liabilities of Monthly Reporting IBFs, Other than to Parent Entity Percent of total reported Number of banks reporting Total IBFs U.S.chartered banks1 Agencies and branches of foreign banks Edge Act corporations Amount (billions of dollars) TOTAL 477 144 264 69 $173.43 New York California Florida Illinois Texas District of Columbia Pennsylvania Washington Georgia Massachusetts New Jersey Ohio Connecticut Kentucky Michigan N. Carolina Rhode Island Virginia 208 84 79 30 20 38 16 27 6 14 154 57 29 17 0 16 11 27 7 6 133.8 20.1 3.3 13.1 77% 12 2 7.5 90 27 8 11 11 9 7 6 5 4 4 2 2 2 2 1 1 8 7 3 4 3 4 4 2 2 2 2 1 1 3 2 4 1 1 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 3.1 1.8 10 \ > (There are too few reporting banks in other states for a data breakdown to be made available.) NOTE: Figures for numbers of IBFs are as of September 28,1983. Figures for liabilities are as of October 26,1983. Monthly reporting banks are those with assets or liabilities of at least $300 million. SOURCE: Federal Reserve Board Release G.14(518)A and Federal Reserve Board unpublished data. 'One savings and loan association has an IBF that is in the Florida figure. What D o IBFs Do? The assets and liabilities of IBFs on December 30, 1981, December 29, 1982, and October 20, 1983, are recorded in table 2; as o f October 20, 1983, over 98 percent o f their liabilities w ere dollar-denominated. The December 30, 1981, figures largely reflect busi ness switched from other accounts either in the parent bank or an offshore branch. Operations o f the IBFs themselves are reflected more clearly in the later figures. Consider the latest available figures in the third column o f table 2. The most important aspects o f these figures is the proportion o f business with other banks vs. the proportion with nonbank customers. On the asset side, about one-sixth o f total assets are "com m er cial and industrial loans” (Item 5a) and one-ninth are loans to “foreign governments and official institutions” (Item 5c). The remainder, over 70 percent, are claims on either other IBFs, overseas banks or an overseas branch of the parent bank. Claims on overseas banks (Items 3a and 5b) are largest, while claims on other IBFs (Item 2) and overseas offices o f the parent bank (Item 1) are of broadly similar magnitude. The liability structure is even more heavily weighted toward banks. Only about 16 percent o f the liabilities of IBFs (as o f October 26, 1983) w ere due to nonbanks. Of these, one-third was due to "foreign government and official institutions” (Item 10c) and two-thirds were due to “other non-U.S. addressees” (Item lOd). The latter are mainly industrial and commercial firms. The high proportion o f both assets and liabilities of IBFs due to other banking institutions reinforces the conclusion that they are an integral part o f the euro dollar market. A high proportion of interbank business is characteristic o f eurocurrency business in which 7 Table 2 Assets and Liabilities of International Banking Facilities (millions of dollars) December 30, 1981 December 29, 1982 October 26, 1983 $7,188 6,785 403 $20,125 19,150 975 $30,322 29,204 1,118 903 16,577 26,256 8,470 12 26,666 276 29,093 482 438 1,130 1,875 17,081 11,705 7,791 1,164 32,808 30,300 16,960 1,070 36,753 32,237 22,348 958 880 3,839 3,262 49,409 132,569 156,484 48,445 965 129,626 2,943 153,264 3,219 56,597 55,229 1,368 152,694 168,776 3,917 186,806 182,469 4,337 $29,091 28,779 313 $56,372 55,114 1,258 $69,756 68,535 1,221 1,009 10,127 2,834 952 336 17,382 37,045 7,439 13,816 2,756 28,803 42,446 9,115 19,073 2,170 15,686 80,080 103,674 15,258 428 78,439 1,641 101,608 2,066 44,777 44,037 741 136,452 133,552 2,899 173,430 170,143 3,257 $-11,820 -11,193 -627 $-16,242 -15,224 -1,018 $-13,376 -12,325 -1,051 56 122 146 ASSETS 1. Gross Claims on Non-U.S. Offices of Establishing Entity (1) Denominated in U.S. Dollars (2) Denominated in Other Currencies 2. Loans and Balances Due From Other IBFs 3. Gross Due From: A. Banks in Foreign Countries B. Foreign Governments and Official Institutions 4. Securities of Non-U.S. Addressees 5. Loans To Non-U.S. Addressees A. Commercial and Industrial Loans B. Banks in Foreign Countries C. Foreign Governments and Official Institutions D. Other Loans 6. All Other Assets in IBF Accounts 7. Total Assets Other Than Claims on U.S. and Non-U.S. Office of Establishing Entity (1) Denominated in U.S. Dollars (Sum of Items 2 through 6) (2) Denominated in Other Currencies 8. Total Assets Other Than Claims on U.S. Offices of Establishing Entity (Sum of Items 1 and 7) (1) Denominated in U.S. Dollars (2) Denominated in Other Currencies LIABILITIES 9. Gross Liabilities Due To Non-U.S. Offices of Establishing Entity (1) Denominated in U.S. Dollars (2) Denominated in Other Currencies 10. Liabilities Due To: A. Other IBFs B. Banks In Foreign Countries C. Foreign Government and Official Institutions D. Other Non-U.S. Addressees E. All Other Liabilities in IBF Accounts F. Total Liabilities Other Than Due To U.S. and Non-U.S. Offices of Establishing Entity (1) Denominated in U.S. Dollars (Sum of Items 10.A Through 10.E) (2) Denominated in Other Currencies 11. Total Liabilities Other Than Due to U.S. Offices of Establishing Entity (Sum of Items 9 and 10.F) (1) Denominated in U.S. Dollars (2) Denominated in Other Currencies RESIDUAL 12. Net Due From ( + ) / Net Due To ( -) U.S. Offices of Establishing Entity (Item 11 Minus Item 8) (1) Denominated in U.S. Dollars (2) Denominated in Other Currencies Number of Reporters NOTE: Unless otherwise noted, figures include only amounts denominated in U.S. dollars. This report contains data only for those entities whose IBF assets or liabilities are at least $300 million, that is, for those entities that file a monthly report of IBF accounts on form FR 2072. SOURCE: Federal Reserve Board Release G.14 (518). FEDERAL RESERVE BANK OF ST. LOUIS there may be several interbank transactions between ultimate borrowers and ultimate lenders.6 An important role for interbank transactions is to provide "swaps” that reduce either exchange risk or interest rate risk for the parties involved. Suppose, for example, an IBF has a deposit (liability) o f $1 million that will be withdrawn in one month, and it has made a loan (asset) to a customer o f $1 million that will be repaid in two months. There is a risk that when the IBF comes to borrow $1 million to cover the second month o f the loan, interest rates w ill have risen, and it will incur a loss on the entire transaction. If, however, this IBF can find a bank that has the opposite timing prob lem (a deposit o f $1 million for 2 months and a loan of $1 million outstanding for one month), the two banks could arrange a swap. The second bank would loan the IBF $1 million in one month and get it back in two months (with suitable interest). The interest rate in volved will be agreed on at the beginning, so that nei ther bank w ould suffer if interest rates should change in the second month. These swap arrangements enable banks to match the maturity structure o f their assets and liabilities. The existence o f such swaps explains the high levels of both borrowing and lending between IBFs and over seas branches o f their parent bank.7 THE GROWTH OF IBFs Chart 1 shows the growth o f total IBF liabilities since the end o f 1981. Although the most rapid growth oc curred in the first six months o f their operation, IBFs have grown considerably over a period in which inter n ation al banking business in gen eral has been stagnant.8 Within two years, they have come to be a significant part o f the international money market. The liabilities o f IBFs as o f October 1983 (other than to parent banks) represent about 8Vz percent o f gross eurocu rrency liabilities (as m easured bv Morgan Guaranty) or about 7Vi percent o f total international banking liabilities (as measured by the Bank for Inter national Settlements. This includes onshore bank lending). 6See Niehans and Hewson (1976) for an explanation of the intermedi ary function of euromarkets. The interbank market is also discussed in Dufey and Giddy (1978), chapter 5. 7For a discussion of the role of swaps in foreign exchange markets, see Chrystal (1984). 8According to B.I.S. figures, international bank assets grew 8.8 per cent in 1982 in nominal terms. This compares with figures typically in excess of 20 percent throughout the 1970s. The combined assets of overseas branches of U.S. banks declined by 0.6 percent in 1982 [see Press Release (1983)], though this partly reflects the growth of IBFs. APRIL 1984 C h a rt 1 T o ta l L ia b ilitie s o f IBFs Log of levels Log of levels NOTE: Liability levels w ere $44,777 m illion on December 30, 1981, an d $173,430 m illion on O cto b er 26, 1983. Figures exclude liab ilitie s to p arent entity. Where did this growth com e from? Has the creation oflBFsgenerateda large volume o f new business or has business been shifted from elsewhere? The evidence is that IBF business has almost entirely been shifted from elsewhere. Terrell and Mills use regression analysis to test the hypothesis that the creation o f IBFs has led to greater growth o f external bank assets.9 This hypoth esis is decisively rejected. Some evidence concerning the origins o f business shifted to IBFs is available in Key.10 It is convenient to consider separately shifts from existing institutions in the U.S. and shifts from overseas banking centers. Shifts fro m Banks in the United States Up until January 27,1982, about $34 billion o f claims on overseas residents w ere shifted to IBF books from other U.S. banking institutions. The bulk o f this (85 percent) came from U.S. branches o f foreign banks — especially Japanese and Italian. Foreign banks typical ly w ould have had a higher proportion o f assets eligible for shifting to IBFs, w hile Japanese and Italian banks generally had not established shell branches in Carib bean offshore centers. 9See Terrell and Mills (1983). 10See Key (1982). 9 FEDERAL RESERVE BANK OF ST. LOUIS in the same period, shifts of liabilities (due to parties other than overseas branches o f the parent bank) from books o f parent entities w ere much smaller. These amounted to about $6 billion, o f which 90 percent came from branches o f foreign banks . The small shift of liabilities relative to assets was affected by several fac tors: the negotiable nature o f some deposits (CDs); the existence o f penalties for renegotiations before matu rity; the delay in passing N ew York tax relief for IBFs until March 1981; the small proportions o f short-term deposits unrelated to trade with the United States; and the availability o f accounts with similar returns yet fewer restrictions as to maturity and denomination (such as repurchase agreements). If only the domestic books o f U.S.-chartered banks are considered, the shift to IBFs is extremely small. Key reports a shift o f $4.3 billion (through January 27,1982) o f claims on unrelated foreigners and only $0.1 billion o f liabilities to unrelated foreigners. An alternative figure for claims shifted to IBFs is obtainable by looking at the change in commercial and industrial loans to non-U.S. addressees plus loans to foreign banks (F e d e r al Reserve Bulletin, table A18, for large weekly reporting banks with assets o f $750 million or more). This indi cates a decline o f $3.3 billion in the same period. APRIL 1984 shift reflected the redundancy o f shell branches, at least for business with non-U.S. residents, once IBFs were permitted. W hile m uch o f the r a is o n d ’e tr e o f Caribbean branches for business with foreigners has been re moved by the establishment o f IBFs, these branches continue to be important for business with U.S. resi dents. Terrell and Mills report that the proportion o f the liabilities o f Caribbean branches due to U.S. resi dents rose from less than half in mid-1981 to about 70 percent by the end o f 1982. However, the attraction o f offshore deposits to U.S. residents is likely to decrease as interest regulations on domestic U.S. banks are re laxed, thereby narrowing the gap between domestic and offshore deposit rates. Based on the figures o f the Bank for International Settlements, Terrell and Mills estimate that the propor tion o f total international banking assets and liabilities due to U.S. banks’ offshore branches declined by 4 percent in the first year o f IBF operation. Another 3'/2 percent was lost by other overseas banking centers to IBFs. Shifts from Other Offshore Centers THE SIGNIFICANCE OF IBFs FOR INTERNATIONAL BANKING Whereas foreign banks were mainly responsible for shifts to IBFs from banks located in the United States, banks chartered in the United States w ere mainly re sponsible for shifts o f business from offshore centers and other overseas banking locations. Key estimates that U.S.-chartered banks shifted about $25 billion in claims on unrelated foreigners and about $6 billion in liabilities due to unrelated foreigners (through Januaiy 27,1982) to IBFs from overseas branches. The compara ble figures for foreign banks w ere $5Vz billion and $9 billion, respectively. The primary significance o f the experience with IBFs is that it enables us to better understand the forces that led to the growth o f eurocurrency markets. In partic ular, the significant decline in business in Caribbean branches following the creation o f IBFs suggests that the growth of business in this area was almost entirely intended to bypass U.S. monetary regulations. Dereg ulation o f domestic banking in the United States will presumably have further effects, since much o f the remaining business in Caribbean branches o f U.S. banks is with U.S. residents. This difference in the propensity to shift assets to IBFs is probably explained by the differential tax incen tives o f U.S. and foreign banks. U.S. banks pay taxes on w orldw ide income and may benefit from tax advan tages o f IBFs. Foreign banks may increase their tax liability to the United States by establishing an IBF instead o f operating in an offshore center. The regulatory changes that permitted the establish ment o f IBFs w ere intended to ease the burden o f domestic monetary restrictions on U.S. banks in the conduct o f international banking business.11 The ex tent to which this aim has been achieved is probably very limited. This is because IBFs play no role in financ ing either activities o f U.S. residents or the U.S. activities o f nonresidents. The bulk o f business shifted by U.S. banks from their overseas branches has com e from the Bahamas and Cayman Islands (collectively called Caribbean). In the first two months o f operation o f IBFs (11/30/81-1/29/821, liabilities to unrelated foreigners o f branches o f U.S. banks located there fell by $6.8 billion, w hile claims on unrelated foreigners fell by $23.3 billion. Much o f this 10FRASER Digitized for Major U.S. banks that w ere involved in international finance to a significant degree had already found ways around U.S. banking regulations and w ere not re stricted in their ability to com pete internationally. The 11Ibid., p. 566. APRIL 1984 FEDERAL RESERVE BANK OF ST. LOUIS fact that major U.S. banks have shifted business to IBFs from offshore centers means, o f course, that there must be some benefit from having an IBF. This may result from low er transaction costs, some tax advantages or the greater attraction, from a risk perspective, o f de posits located in the United States. However, the big gest gainers among U.S. banks may be medium-sized banks that were big enough to have some international business but not big enough to have an offshore branch.12 Other major beneficiaries from IBFs have been the U.S. branches and agencies o f foreign banks. It is no accident that w ell over half o f all IBFs have been estab lished by these banks. The benefit to them arises from the high proportion o f their existing business that is IBF-eligible, that is, the portion with nonresidents. Not the least o f this w ould be transactions with their parent banks overseas. creased the total volume o f international banking busi ness. Indeed, IBFs have grown at a time when interna tional banking growth has been at its slowest for over two decades. This growth has been largely at the ex pense o f banking offices in other locations. For the U.S. and w orld economies, however, IBFs are not o f great significance. There may be efficiency gains resulting from the relaxation o f U.S. regulations that led to the establishment o f IBFs. But such gains are small. Interest rates in w orld capital markets are unlikely to have been affected. Benefits that accrue to banks lo cated in the United States from their IBF facilities are largely offset by losses in offshore banks, though in many cases the gainers and losers are both branches of the same parent bank. REFERENCES “Announcements." Federal Reserve Bulletin (July 1981), pp. 561-63. CONCLUSIONS The establishment o f IBFs in the United States repre sents a change in the geographical pattern o f interna tional banking. It facilitates the conduct in the United States o f some business that was previously conducted offshore. It also increases the ease with which foreign banks can operate branches in the United States. The creation o f IBFs, however, does not seem to have in 12lt is true that the largest banks have the largest IBFs. However, the cost saving at the margin from IBFs for a bank that had, say, a Caribbean shell operation is much smaller than for a bank that had no offshore booking location. Chrystal, K. Alec. “A Guide to Foreign Exchange Markets,” this Review (March 1984), pp. 5-18. Dufey, G., and Ian H. Giddy. (Prentice-Hall, 1978). The International Money Market “Florida's Baffling Unitary Tax: What Is It, Whom Does It Hurt?” American Banker, December 28, 1983. Key, Sydney J. “International Banking Facilities,” Federal Reserve Bulletin (October 1982), pp. 565-77. Niehans, Jurg, and John Hewson. “The Euro-dollar Market and Monetary Theory,” Journal of Money, Credit and Banking (Febru ary 1976), pp. 1-27. Press Release. 1983. Federal Reserve Board of Governors, August 22, Terrell, Henry S., and Rodney H. Mills. International Banking Facili ties and the Eurodollar Market, Staff Studies No. 126 (Board of Governors of the Federal Reserve System, August 1983). 11 A Private Central Bank: Some Olde English Lessons G. J. Santoni D ISSATISFACTION with persistent and volatile in flation since the mid-1960s has led to numerous calls for a different approach to monetary policy. In some cases, people have suggested that monetary policy de cisions be made more explicitly political, for exam ple, subject to greater control by Congress via congressionally mandated monetary growth targets. In the same vein, Milton Friedman has proposed that m one tary policy be set by the Treasury, thus making the President o f the United States ultimately responsible for its conduct. In contrast, some critics o f the current system have argued for a return to the constraints o f Bretton Woods or the even earlier classical gold standard. Some have suggested that the only lasting solution to the problem entails the private production o f money. Behind these different suggestions is the implication that central bankers will respond systematically and, hence, predictably to the different incentives em bodied in these alternative programs of monetary con trol. If the incentives are changed, so the theory goes, better policy decisions will be made. This paper focuses directly on the theoretical and empirical support for the claim that different incen tives induce policymakers to choose different m one tary growth rates. This paper does not advocate a par ticular set o f incentives or form o f organization for the central bank. Rather, it merely points out that the choices o f m onetary policym akers depend, as all choices do, upon the set o f incentives the individual confronts. G. J. Santoni is a senior economist at the Federal Reserve Bank of St. Louis. Thomas A. Pollmann provided research assistance. 2 A verification o f this proposition is sought by ex amining the behavior o f the Bank o f England over the period from, roughly, 1700 to 1930. This period, which encompasses two significant changes in the incentives facing England’s central bankers, provides support for the view that policymakers, like other individuals, re spond predictably to changes in the cost-reward cir cumstances facing them. WHAT DISTINGUISHES CENTRAL BANKERS? The central banker is the person (or group) holding the enforceable right to control the quantity o f nominal money balances in circulation.1 This right is valuable. Whoever holds it can, among other things, materially influence the rate o f inflation and the flow o f profits from money creation (seigniorage), as well as the pres ent value o f the right itself. WHY THE BANK OF ENGLAND? The Bank o f England presents an interesting case in studying the effect o f different incentives on the be havior o f central bankers. There are two reasons for this. First, the original organization of the Bank differed from its modern counterparts in one fundamental re spect: the Bank o f England was a privately-owned-forprofit central bank from its inception in 1694 until the 'More precisely, they control the issuance of “high-powered money” or base money. Since the long-run link between base money and the transaction balances of the public (money) is so close, the paper treats the right to control base money as synonymous with the right to control money. See Balbach (1981); Johannes and Rasche (1979). FEDERAL RESERVE BANK OF ST. LOUIS early 1930s. Further, the Bank was immersed in a set of institutional arrangements that related the wealth of the Bank’s owners inversely to the rate o f inflation. This paper shows that the costs and benefits o f varying monetary growth rates were different for England's private for-profit central bankers than those typically taken into account by modern central bankersr Conse quently, a different monetary growth rate emerged. A second reason for studying this particular case is that control o f the money supply by the Bank's owners was interrupted from 1793 to 1821, when the govern ment seized the Bank’s monetary control function. For the purposes o f this paper, the interruption is im por tant because it allows a contrast o f monetary growth outcomes produced bv certain identifiable changes in the incentive structure, while other important institu tional factors remained roughly constant. A BRIEF HISTORY OF THE BAMK OF ENGLAND: 1694-1832 Prior to 1694, England's m oney supply consisted mainly o f coins. These coins w ere controlled by the government through regulation o f the mint. The coins were in a continually bad state because the populace persisted in clipping, sweating, filing, washing and boring them. Further, the government resorted to progressive debasement in the form o f fre quent recoinages, a practice that was particularly pro nounced during the reigns o f Henry VIII and Edward VI.3 In some cases, the government expropriated m one tary wealth outright. In 1640, Charles I closed the Lon don Mint and confiscated the funds o f private citizens that had been stored there for safekeeping. Later, in 1672, Charles II expropriated funds deposited with the Treasury by London goldsmiths.4 This irresponsible behavior had important consequences when, in 1692 and 1693, William III floated long-term loans to finance a war with France. Because o f the earlier debasements 2Alchian (1977), pp. 127-50. 3Kemmerer (1944), pp. 34-36. “Charles finally acknowledged one-half the debt, but the promise to pay was never kept. Payments, at the rate of 6 percent, were made only during the period 1677-83. Charles explained the reason for his action as follows: Whereas since the time of our happy Restoration We have been involved in great Forreigne Warrs as well for the Safety of our Government as the vindication of the Rights and Privileges of our Subjects, In the prosecu tion whereof we have been constreyned for some years past, contrary to our Inclinations, to postpone the payment of the moneys due from Us to APRIL 1984 and expropriations, the interest rates dem anded for long-term loans to the Crown contained a substantial premium :’ Establishment o f the Bank W illiam ’s war with France was a costly affair. When additional funds w ere required in 1694, a proposal that twice previously had been put forward by William Paterson was adopted in the Ways and Means Act of that year. The Act provided that those w ho subscribed for and towards the raising and paying into the receipt o f the Exchequer the said sum o f twelve hundred thousand pounds part o f the sum o f fifteen hundred thousand pounds w ere to constitute jointly the Com pany o f the Bank o f England.*’ The loan was a perpetuity, paying interest at the rate of 8 percent. This was considerably below the interest rates that previously had been charged the Crown. The subscribers, however, received additional rights to 1) form a joint stock banking company, 2) deal in bills of exchange, gold and silver, 3) grant advances on secu rity, and 4) issue promissory notes transferable by en dorsement in an amount not exceeding the Bank’s capital.7 These terms apparently were very attractive. The entire loan was subscribed within 12 days. Every sub several Goldsmiths and other upon Tallys s tru ck. . . , And although the present Posture of Our affaires cannot reasonably spare so greate a sum as must be applied to the satisfaction of those debts, Yet considering the great difficulty which very many of our Loving Subjects (who putt their moneys into the hands of those Goldsmiths and others from whom we received it) doe at present Lye under, almost to their utter ruine for want of their said moneys. We have rather chose out of our princely care and compassion towards Our people, to suffer in Our owne Affaires than that our loving subjects should want soe reasonable a Reliefe. Bisschop (1967), p. 48. Incidentally, tallys were pieces of wood upon which government indebtedness was recorded. The govern ment issued tallys when it borrowed from individuals. Goldsmiths regularly accepted these tallys and credited the accounts of the depositors. sHomer (1977), p. 126. Both loans were of 1 million pounds. The first was a life annuity paying interest of 10 percent until 1700 and 7 percent thereafter on a semitontine basis (surviving subscribers split one-half the proceeds due decedent subscribers). The second loan paid interest at the rate of 14 percent. 6Bisschop, p. 74, and Clapham (1958), vol. 1, pp. 16-20. 7Bisschop, pp. 70-71, and Macleod (1897), pp. 773 and 776. The Bank employed three methods of accounting for the transaction accounts (“running cash") of its depositors and these methods de fine how the balances were transferred in the exchange process. The methods were by “Notes payable to Bearer, to be endorsed,” by “Books or Sheets of Paper, wherein their Account to be entered,” or by “Notes to persons to be accomptable." The first method was the forerunner of central bank notes. The third was essentially equivalent to a present-day checking account. The second was much like modern passbook accounts. See Clapham, vol. 1, p. 21. 13 FEDERAL RESERVE BANK OF ST. LOUIS scriber became a shareholder of the Bank to the extent o f his subscription, and all or any fraction o f his share could be sold to others.8 The Bank opened for business on July 27, 1694, in Mercers’ Chappell. From its inception, eveiy effort was made by the Governor and Court o f Directors (Board of Directors o f the Bank) to attract depositors and to promote the circulation o f its “running cash notes.’’9 These notes were convertible into legal tender money, gold coins, at a fixed exchange rate upon demand at the Bank."1 Subsequent legislation strengthened the Bank's position. The Bank was granted a m onopoly in joint stock banking in early 1697.” In 1708, the Bank obtained a m onopoly in the issue o f joint stock bank notes.1' Later, in 1742, Act 10 and 11 George II., C. 13 (par. 5), reaffirmed the earlier rights granted to the Bank. Each o f these pieces of legislation was accompa nied by an additional Bank loan to the government. The Bank subscribed an additional 1,001,017 pounds for loan to the government at 8.0 percent in 1697. In 1707, it extended a 1,500,000 pound loan at 4.5 percent to the government and, in 1742, another 1,600,000 pounds at 3.0 percent. No further significant legislative changes regarding the Bank's position occurred until 1826. In that year, the Bank's m onopoly on joint stock banking was lim ited to within a 65-mile radius o f London.1:1Seven years later, in 1833, its m onopoly o f joint stock bank note 8 Bank shares exchanged hands regularly and, in 1747, Gentleman’s Magazine began publishing daily price quotes for Bank of England shares of stock. In 1773 New Jonathan's Coffee House printed the words, “The Stock Exchange" over its door and admittance was permitted only by fee. 9Clapham, vol. 1, pp. 20-23. ,0lt is interesting to note that an official (established by law) gold standard was not enforced in England at this time. The Bank's commitment to redeem its notes at a fixed price in terms of gold was not foisted upon the Bank by the government. Rather, this was a voluntary contract established by the Bank with its customers. En gland’s official gold standard was not established until 1821 (more than 100 years later) and then only as a result of the Bank’s insistence. 11Clapham, vol. 1, pp. 46-50. 12Bisschop, pp. 82-83. Act 7 Anne, C.7, provides “that during the continuance of the said corporation of the Governor and Company of the Bank of England, it shall not be lawful for any body politic or corporate whatsoever, created or to be created (other than said Governor and Company of the Bank of England), or for any other persons whatsoever, united or to be united in covenants, or part nership, exceeding the number of six persons, in that part of Great Britain called England, to borrow, owe, or take up any sum or sums of money on their bills or notes payable at demand, or at a less time then six months from borrowing thereof.” 14 APRIL 1984 issue was also limited to the same area. However, Bank of England notes were made legal tender at this time.14 This legislation provided legal force to the practice that had already been adopted by other banks o f maintain ing their reserves in the form o f Bank o f England notes. The Government Steps In: 1793—1821 The Napoleonic Wars between England and France began in 1793. With the exception o f a m inor truce, the war continued until N apoleon’s abdication on April 6, 1814.15Government demands from the Bank for financ ing rose substantially during the war. O f course, the Bank's contract with its depositors to redeem its notes at a fixed price in terms o f gold got in the w ay o f the government’s interest and, on February 26, 1797, the King and Privy Council ordered the Bank to suspend specie payments, a suspension that was to last for more than 20 years."’ During the suspension, control o f the money supply, which had rested with the Bank’s owners, was largely usurped by the government. Clapham notes that The minutes o f the Court and those o f the Committee o f Treasury are full o f ........ requests for help from Perceval, a n d o f th e B an k's re lu c ta n t bu t in v a r ia b le a c quiescence.17 Th e Bank app a ren tly acq u iesced because o f an “understanding, a gentleman's understanding... to do this business and to do it in the way most convenient to the Treasury."18 Given the transfer o f monetary control to the govern ment, it is, perhaps, not surprising that the Bank's Board o f Directors became unusually lackadaisical in their attention to duties. As a result o f absenteeism, the Bank Court experienced difficulty in maintaining a quorum. Letters w ere sent to a number o f directors that “pointedly asked 'when their attendance could be depended upon’ " for "too much o f the business had 13Bisschop, p. 198. 14Andreades (1924), p. 261; Bisschop, p. 198. 15There was, of course, Napoleon’s “Campaign of 100 Days” be tween his escape from Elba on March 1, 1815, and his defeat at Waterloo on June 18, 1815. Due to its brevity, it is ignored in this analysis. ,6The wording of the order ran as follows: The Bank will “forbear issuing any Cash in Payment until the Sense of Parliament can be taken on that Subject.” Clapham, vol. 1, p. 272. 17Clapham, vol. 2, p. 33; see, as well, Viner (1937), p. 122; Cannan (1919), p. xi. 18Clapham, vol. 2, p. 11. 19Clapham, vol. 2, p. 31. APRIL 1984 FEDERAL RESERVE BANK OF ST. LOUIS been done 'by a Single Director with the assistance of the Head o f the Discount Office’.” 19 Return to Private Control The suspension played an important role in the governm ent’s effort to wrest control o f the money supply from the Bank. In the absence o f suspension, "control” of the stock o f money would have meant little to the government since the requirement to redeem notes at a fixed price in terms of specie eventually w ould have (and did during 1790—96) placed an effec tive constraint on note issue. The public was never keen on the suspension, and the Bank made this the political issue in its fight to regain control of the m oney supply. In October o f 1797, six months after it was ordered to suspend payments, the Bank indicated that it could “with safety resume its accustomed functions (payment o f specie), if the polit ical circumstances o f the country do not render it inexpedient.”211The Bank’s report was virtually ignored by government. In June o f 1810, the "Beport from the Select Committee on the High Price o f Bullion” recom m ended to Parliament that the resumption o f specie payments (at the old par) begin within two years. The issue was not even taken up for discussion until July of the following year.21 A vote on the recommendation was taken in the House o f Commons in 1811. The House voted 180 to 45 against the issue. On its own initiative, the Bank began partial resump tion o f specie payments for notes of 5 pounds or less in Januaiy o f 1817. Early in 1819, however, Parliament required the Bank to discontinue the practice.22 Parlia ment had promised on five different occasions to even tually return to specie payments, but continued to drag its feet on fixing a date. Finally, on July 2,1819, the House o f Commons passed an act permitting the re sumption o f cash payments (bullion a n d coin) after May 1, 1822. At the request o f the Bank, this date even tually was moved forward to May 1, 1821 23 SOME IMPORTANT OBSERVATIONS There are a number o f important points to draw from the previous discussion in analyzing the incen 20Clapham, vol. 1, p. 272. See, as well, Cannan, p. xi. 2'Viner, p. 171. 22Viner, p. 172. 23Viner, pp. 172-73. tives faced by the central bankers. First, given the one exception noted, the right to control the money supply was held privately. This right, in the form of ownei'ship shares in the Bank, was traded in an organized market. Any expected changes in the future profits o f the Bank w ould be reflected by changes in the price o f Bank shares and w ould immediately affect the wealth of Bank owners. Second, the owners of the Bank had loaned consid erable sums to the government in perpetuity at fixed rates o f interest. By 1743, the sum was w ell in excess of nine million pounds. Finally, the Bank’s contract with its customers to redeem its notes at a fixed price in terms o f gold was a voluntary arrangement. An official (established by law) gold standard did not exist in England until 1821. In fact, the gold standard came about largely as a result of the Bank’s continuous prods to an unwilling Parlia ment. The following discusses how this unique incentive structure faced by E ngland’s central bankers in fluenced the monetary growth rate. THE PROFITS FROM THE RIGHT TO CONTROL MONEY Like the right to control the production o f any com modity, the right to control the production of money is valuable. The central bank, at the cost o f a few cents worth o f paper and ink, can produce a $100 bill (or a 100-pound note) that can be exchanged in the market for $100 worth o f resources. The Flow o f Profits The central bank introduces m oney into circulation by exchanging units o f m oney (which it prints) for commodities. These comm odities may be either real or financial assets. Since the bank buys these assets at market prices, the expected flow o f nominal profits generated by the purchase o f the assets is equal to the nominal interest rate times the price o f the assets pur chased. This is equivalent to the nominal interest rate multiplied by the quantity o f money exchanged for the assets. The flow of real revenue is simply the nominal flow divided by the price level (i X M/P). Since w e are interested in relating the bank's real revenue flow to the rate of money production, account must be taken o f the fact that, at higher rates o f money production (higher rates o f inflation), people will want 15 APRIL 1984 FEDERAL RESERVE BANK OF ST. LOUIS Table 1 The Revenue From Money Production TT (2) M/P (3) i = r+ir (4) R = i(M/P) .50 .45 .40 .35 .30 .25 .20 .15 .10 .05 0.00 0 1 2 3 4 5 6 7 8 9 10 .60 .55 .50 .45 .40 .35 .30 .25 .20 .15 .10 0.00 .55 1.00 1.35 1.60 1.75 1.80* 1.75 1.60 1.35 1.00 (1) where: it = the rate of inflation (rate of monetary growth) M/P = the stock of real purchasing power demanded i(M/P) = the profit from monetary production (inflation) population are stationary, that the real interest rate is 10 percent, and that the nominal interest rate is equal to the sum o f the real interest rate and rate o f inflation ( i = r + t t ). The numbers in the first two columns o f table 1 indicate that, as the rate o f inflation rises (falls), the public's demand for real purchasing pow er falls (rises). The third column indicates the nominal rate o f interest at the various rates o f inflation. The fourth column indicates the profit stream at the different rates of inflation. As the rate o f inflation falls from very high rates, the bank’s profit from inflation initially rises because peo ple choose to hold a greater amount of real purchasing power. Reducing the rate o f inflation increases total profits up to a point (1.80 real goods per unit o f time in this example), after which further reductions in the rate o f inflation cause profits to fall. In this example, the profit-maximizing rate o f inflation (monetary growth rate), which is the one the central bankers will choose, is 20 percent.24 ‘ indicates maximum profit to hold less o f their wealth in the form o f real purchas ing pow er (M/P). Other things unchanged, the flow o f real revenue w ould decline at higher rates o f money growth because M/P declines. Even though M is rising, the price level rises faster. However, other things are not unchanged. Faster money growth increases the rate o f inflation and this raises the nominal interest rate (i). Faster m oney prod u ction exerts tw o opposing forces on the bank’s real revenue. One force tends to reduce revenue, w hile the other tends to increase revenue. In general, there is a unique rate o f money growth (and rate o f inflation) that will maximize the flow o f real revenue. THE BANK OF ENGLAND’S UNIQUE CONSTRAINTS The Flow o f Profits Among other things, the above result depends upon the particular set o f operating rules the bank faces. Apart from the particular assum ptions expressed above, the foregoing example does not constrain the bank in any way. If additional rules w ere imposed, the profit function m ay change. As a result, the central banker’s would be confronted with different incentives, causing them to select a different monetary growth rate. The Bank o f England was founded on the condition that the stockholders grant a substantial loan to the A Simple Example Table 1 presents a hypothetical example relating different rates o f inflation (or rates o f m onetary growth), 'ir, to the public’s demand for real purchasing power, M/P. In order to facilitate the calculation o f the rate o f inflation that maximizes the bank’s revenue flow, suppose that the public knows the rate o f infla tion w ith certainty (extreme rational expectations), that changes in the monetaxy growth rate affect only the rate o f inflation and the public’s desire for real purchasing pow er but no other real variables, that the cost o f producing nominal units o f money is zero (so revenue and profits are identical I, that real output and 16FRASER Digitized for 24Since R = i x (M/P) = (r + tt)(M/P), real profits are maximized when d(M/P) tt dir M/P (r + ir) = it Hence, nm(— + 1) = -1, TT where nm = the elasticity of demand for real purchasing power with respect to the rate of inflation. When r = 0, this result reduces to nm = -1 which is the familiar result obtained by others. See Friedman (1953), pp. 251-62; Friedman (1971), pp. 846-56; and Bailey (1956), pp. 93-110. FEDERAL RESERVE BANK OF ST. LOUIS government at a fixed rate o f interest. By 1743, that loan amounted to almost 10 million pounds. At the point when these loans w ere made to the government, the interest rate charged was below the market rate. (Becall the 1G94 loan at 8 percent when the market rate on long-term loans to the Crown was 14 percent.) This subsidized loan rate is a payment made by the Bank to the government for the lease rights to the production of money. The right was never granted to the Bank in perpetuity. Bather, as indicated above, the Bank's char ter came up for review periodically. The cost to the Bank o f its government loan depends upon the market rate o f interest. If the coupon rate is c and the amount loaned is L, the nominal value o f the lease payment per unit o f time is (i —c)L. The real value is the nominal amount divided by the price level, (i —c)L/P. The higher the nominal rate o f interest, i, relative to the coupon rate, c, the larger the cost o f the lease to the Bank. An additional constraint is relevant. The quantity of notes the Bank could issue was restricted by law to an amount less than or equal to the capital invested by stockholders. Since the capital represented the loan to the government, M must be less than or equal to L. Given this constraint, the Bank’s owners will choose M = L because the flow o f real profit is highest in this case, other things the same (see insert). Consequently, the Bank's profit is simply the coupon rate earned on the loan, c, times the loan (which is equal to the quanti ty o f notes issued, M), divided by the price level, P. Table 2 illustrates the effect o f this set o f rules on the profit-maximizing rate o f inflation for the Bank. The first three columns o f table 2 simply reproduce the first three columns o f table 1. Column 4 calculates the real profits o f the Bank under the new set o f rules where the coupon rate, c, is assumed to be 10 percent. Note that profits are maximized at a zero rate o f inflation rather than the 20 percent rate obtained previously.-5 25This result is completely general as long as the demand for real purchasing power is inversely related to the rate of inflation. In this case, dR d(M/P) „ Since the derivative of the profit function with respect to the rate of inflation is negative, it does not pay the Bank to generate an inflation by continuously expanding M and, of course, the constraint that M=sL will eventually become binding. Given that profit maximization requires M = L from expression 1, a deflation would not benefit the Bank because it would require M to fall below L. As a result, the Bank will choose a zero rate of inflation. In addition, the Bank’s owners tended to be net monetary credi tors as a class and this further reduced their incentive to inflate. APRIL 1984 Table 2 An Example of the Bank’s Profit Function (1) IT (2) M/P (3) i = r+w (4) R'=c(M/P) .50 .45 .40 .35 .30 .25 .20 .15 .10 .05 0.00 0 1 2 3 4 5 6 7 8 g 10 .60 .55 .50 .45 .40 .35 .30 .25 .20 .15 .10 0.00 .10 .20 .30 .40 .50 .60 .70 .80 .90 1.00* 'indicates maximum profit The Role o f Specie Payments: A Contract f o r Price Level Stability It is in the interest o f Bank owners to inform the public o f their intention to maintain a relatively low inflation rate. Demand for the Bank's product does not rise until the anticipated rate o f inflation declines. This eventually will result from the Bank’s policy o f main taining a relatively low monetary growth rate. The ow n ers o f the Bank, however, chose to hurry the adjust ment o f expectations by "marketing” their bank notes in a particular way. In marketing its notes, the Bank guaranteed its cus tomers a low rate o f price inflation. This guarantee took the form o f a contract to redeem Bank notes at a fixed price (a fixed weight o f gold). The contract can be thought o f as insurance against the overissue o f Bank notes, because it pledged the original investment o f the Bank’s stockholders as surety for meeting the con tract.26 If bank notes w ere issued in such quantity as to There were about 1,300 original subscribers to the Bank stock. Many of them were London businessmen who were “linked sneeringly with the rather ill-famed money-lending scriveners.” Others were Gentlemen and Esquires, “people who . .. live idly as gentlemen'.” See Clapham, vol. 1, pp. 273-89. 26“That double event, (1) a low identification cost to everyone about the intermediate commodity and (2) specialist-experts who provide quality assurance and information more cheaply than novices can provide for themselves, explains the use of a low identification cost commodity as a general intermediary medium of exchange-money. It permits purchase of information from lower cost sources, a cost reduction that exceeds the added cost of using an intermediary good for indirect exchange.” Alchian, pp. 117-18. 17 FEDERAL RESERVE BANK OF ST. LOUIS APRIL 1984 The Profit Function and Sharing Arrangement BANK PROFITS The Bank’s flow o f real profit is i(M/P), as before, minus the opportunity cost o f the government loan, (i-c)IV P . (1) R' = .M i (1 — c ) .L - P = .M 1— P •L . L l— + c— P P Given the constraint that M < L, it is clear from expression 1 that the Bank’s owners will choose M = L. The Bank’s profits are highest in this case, other things the same. As a result, the profit function reduces to ,,, (2) L M R = c— = c— SHARING THE GAIN FROM MONEY PRODUCTION The capitalized value o f the nominal profit stream Since the above expression takes account o f the opportunity cost o f issuing the government loan, expression 3 is the net increase in stockholder wealth that derives from the right to issue money. The increase is proportional to the loan IL) where the proportion is determined by the ratio c/i. The higher is c relative to i, the greater is the stockhold ers' share and the low er is the government’s share. For the initial loan o f 1,500,000 pounds, the stock holders' share was 857,142 pounds (1,500,000 X .08/.14) in terms o f the prices that existed in 1694. The government’s share was 642,857 pounds. Long-term interest rates fell dramatically after the formation o f the Bank.1 In part, this was due to the eventual effect o f the Bank’s choice o f a zero rate o f inflation on price expectations. O f course, the de cline in the nominal interest rate had the effect o f raising the Bank’s share o f the net wealth derived from issuing money (c/i in expression 3 rises as i falls). As noted above, the government renegotiated c downward each time the Bank’s charter was re newed in an effort to maintain the relative shares. ’ See Homer, pp. 131 and 161; Martin (1865), pp. 24-25; and Rogers (1887), p. xiv. cause their market price in terms o f gold to fall below the price promised by the Bank, people w ould arbi trage the difference by trading gold for notes in the market at the low price and exchanging the notes for gold at the Bank for the higher price. In the process, wealth would be transferred awav from stockholders to those engaging in the arbitrage. The guarantee was believable because customers knew that stockholders would lose wealth if the Bank overissued its notes relative to the supply of goods in general and gold in particular.27 270f course, the guarantee is not perfect. New gold discoveries or improvements in mining technology would cause the price of gold and Bank notes to fall in terms of, say, a standard commodity basket. However, the guarantee, while imperfect, was operational. It provided a relatively low-cost method of metering the Bank’s rate of note production and policing the guarantee. Digitized for18 FRASER DIFFERENT CONSTRAINTS FOR THE GOVERNMENT When, for all practical purposes, the government took control o f the money supply in 1793, the con straints facing the decisionmakers changed substan tially. Recall that the government did not expropriate ownership rights in the Bank outright. Had they done so, it w ould have been a clear land, possibly, politically unsavory) transfer o f wealth from stockholders to the government. The government, however, did the next best thing from its point o f view. It expropriated the wealth o f the stockholders by a m ore circuitous route. When the government took over, the Bank held a loan which, while an asset to the Bank, was a liability to the government. In terms of the example used here, an increase in the rate o f inflation increases the nominal APRIL 1984 FEDERAL RESERVE BANK OF ST. LOUIS interest rate, i, and increases the Bank's opportunity cost o f the governm ent loan, (i —c)L/P. In effect, accelerating the rate o f inflation raised the lease pay ment the Bank made to the government. Further, during its period o f control, the government continuously violated the constraint that the quantity o f notes in circulation not exceed the capital o f the Bank. The government did not wish to be bound by the same rule that it believed appropriate in regulating the behavior o f the Bank's stockholders. As a result of the different constraints faced by Bank ow ners vs. the governm ent, w e should expect to observe relatively low rates o f monetary growth and inflation during periods when the money supply is controlled bv the private owners o f the Bank o f En gland and more rapid rates o f m onetaiy growth and inflation during the periods o f government control. In addition, the dem and for real purchasing pow er should be low er during the period of government con trol and the price o f Bank stock should decline. EV ID E N C E The Behavior o f English Prices One o f the more interesting pieces o f evidence con cerning the effect o f different incentives is England’s history o f inflation during the period o f private m one tary control. England’s m oney supply was under private control for almost 200 years, and the rate of inflation during this period was statistically indistin guishable from zero. From the establishment of the Bank in 1694 until the beginning o f the Napoleonic War's in 1793 when the government usurped control o f the money supply, the annual average rate o f inflation in England was .01 percent. In 1821, after the Napoleonic Wars, the govern ment returned control o f the m oney supply to the Bank and, at the Bank’s insistence, established an official gold standard. Private control continued until 1913. During this period, 1822-1913, England's annual aver age rate o f inflation was .42 percent which, again, is statistically indistinguishable from zero. In contrast, by 1931, the English government had taken complete con trol o f the m oney supply. Since that date, the annual average rate o f inflation has been significantly positive at 6.47 percent.28 T able 3 The Real Price of Bank Stock, 1780-18321____________________________ Estimate2 Ps/P = 1.375 ,232r + 005RW - ,141D0 (7.40)* (8.22)* (7.27)* (2.21)* Rho = .70 (7.02)* R2 = .72 DW = 2.02 where: Ps/P = the real price of Bank stock. r = the yield on 3 percent consols. RW = real weekly earnings. D0 = a dummy variable for the period subsequent to government's seizure control. D0 = 1 for the years 1793-1832 and zero otherwise. 'The price of Bank stock is the annual average calculated from the daily price quotes reported in Gentleman's Magazine for the years 1780-1832. 2Corrected for first-order autocorrelation, t-values in parentheses. 3The E. W. Gilboy series for average weekly earnings is spliced to the A. L. Bowley and G. H. Wood series by a factor which is the ratio of the average levels of the two indexes over the years 1791-93 and then divided by the price level. See R. B. Mitchell, Abstract for British Historical Statistics (Cambridge University Press, 1962), pp. 347-48. ‘ Significantly different from zero at the 5 percent level. The Napoleonic Wars: Additional Evidence The above data regarding the history o f English infla tion are consistent with one o f the implications o f the theory. Other aspects o f the theory can be examined by considering data from the period immediately before, during and after- the Napoleonic Wars. The war is im portant because the transfer o f control o f the money supply during the war was accomplished through a “gentleman’s understanding” rather than an outright government expropriation o f Bank ownership. As a result, ownership shares in the Bank continued to be exchanged by private individuals, and changes in the value o f these shares along with changes in the de mand for real purchasing pow er provide further evi dence in regard to the theory. The Market Price o f Bank Stock 28This period includes the Depression and World War 11. If these years are excluded and the inflation rate is calculated over the period 1946-82, the mean rate of inflation is 6.87 percent (t-score = 9.10). The t-scores for the periods 1694-1793, 1822-1913 and 1931-82 are .145, .489 and 6.62, respectively. Table 3 presents an estimate o f the effect o f the government takeover and suspension o f specie pay ments on the real price o f Bank stock. The estimate 19 FEDERAL RESERVE BANK OF ST. LOUIS controls for the effect o f business cycles (proxied by average annual real weekly earnings o f men in full-time employment) and the interest rate, and includes a dummy variable for the period subsequent to the sus pension. The coefficients o f the business cycle proxy and the interest rate are significant and have the ex pected sign. The coefficient o f D„ is negative and sig nificant. Its magnitude implies that the government takeover caused the real price o f Bank stock to fall ''permanently" by about 11 percent.21’ In short, the government chose a rate o f inflation that was inconsis tent with maximizing the real flow o f Bank profits (from all sources) and this was reflected in the price o f Bank stock. The price o f the stock did not return to its original level w hen monetary control was returned to the stockholders in 1819. “’ This was tested by including a second dummy variable that assumes the value of 1 for the period 1793-1818 and zero otherwise, along with D„ that assumes a value o f 1 for the period 1793-1832 and zero otherwise. The coefficient o f the second dum my is insignificant, indicating that the variable is re dundant. That is, singling out the 1793-1818 period adds nothing to the explanatory pow er o f the equation. In addition to this evidence, monthly data for the price o f Bank and India Company stock are available for the period 1780-1801.31 There appeal's to be a break in the ratio o f the price o f Bank stock to India Company stock in 1793. Before then, the mean o f the ratio was .93 with a standard deviation o f .11. After 1793, the mean fell to .83 with a standard deviation o f .03. The decline in the ratio is statistically significant.32 The price of Bank stock apparently declined relative to India Com pany stock by about 11 percent, virtually identical to the estimated decline produced by the regression in table 3. Disgruntled Stockholders Understandably, the stockholders were restive dur ing the suspension. In 1801, Alexander Allardvce, 29The estimates are adjusted for first-order autocorrelation. They were checked for second-order autocorrelation with the result that Rho 2 was insignificant. 30During the later part of the suspension, various moves to resume specie payments were afoot. As early as 1810, the Bullion Report advocated a return to specie payments in 1812. In addition, the Bank had begun a partial resumption in 1817, and in 1819 Parlia ment finally committed itself to a specific date for resumption. For purposes of the following test, I terminate the period in 1818, the year prior to Parliament's decision to return control to the Bank. 31See Sinclair (1803), pp. 22-48. 32The t-score = 10.78. 20 APRIL 1984 Table 4 Growth of Bank Money and Inflation, 1780-1832 Time period Money growth1 t-score Inflation2 t-score 1780-1792 1793-18133 1814-1821 1822-1832 1.96 4.04* -2.95 .47 .41 2.20 1.05 .16 .72 4.85* -6.99 -.87 .60 2.67 2.06 .41 1B. R. Mitchell (1962), pp. 442-43. 2lbid., pp. 469-70. The Schumpeter index for consumer goods is spliced to the Gayer, Rostow, Schwartz index of domestic and imported commodities by a factor which is the ratio of the average levels of the two indexes over the years 1821-23. 3The period of the Peace of Amiens (the year 1802) is excluded from this period. 'Significantly different from zero at the 95 percent confidence level. spokesman for the critics, m oved that a complete accounting o f the Bank’s financial condition be pre sented to the stockholders so that the Court might “ declare a dividend o f the w hole profits, the Charges of Management only excepted, as the Law directs.”33 Beal dividend payments, inclusive o f bonuses, did not increase during the suspension and the real value o f Bank stock declined. These two factors along with interest-free loans made by the Bank to the govern ment must have appeared to critics as a thinly veiled expropriation o f wealth.34 They no doubt recognized the spirit o f Charles II lurking in the government. Prices As was the case for the more extended period dis cussed above, the rate o f inflation is indistinguishable from zero (see table 4) in the years im mediately preced ing the government takeover and those following re sumption o f specie payments in 1821. In contrast, the price level rose significantly (at an average annual rate 33Clapham, vol. 2, p. 40. 34ln 1799, when the market rate of interest stood at 5.07 percent on long-term securities, the Bank made a "loan” to the government of 3 million pounds interest-free for six years. The present value of this gift was, roughly, 770,000 pounds. In addition, when the loan came due in March of 1806, the government asked that the loan be renewed until a point in time six months after a “Definitive Peace.” The government offered to pay 3 percent interest. At the time, the long-term interest rate was considerably higher and 3 percent con sols were selling at about a 40 percent discount. 35The years 1814-21 are treated separately in table 4. During this period, various steps were being taken to return to specie payment (see footnote 30). FEDERAL RESERVE BANK OF ST. LOUIS Table 5 The Demand for Real Purchasing Power, 1780-1832_____________________ Estimate1 M/P = -1.535 + 2.002t + .573RW - 18.678D0 (.08) (5.88)* (5.00)* (1.88)* Rho = .56 (3.92)* R2 = .70 DW = 1.97 where: M/P = the stock of real purchasing power, t = time in years. RW = real weekly earnings. D0 = a dummy variable for the period subsequent to government’s seizure of monetary control. D0 = 1 for the years 1793-1832 and zero otherwise.2 APRIL 1984 Table 5 presents a regression that estimates the effect o f the government’s seizure o f control over the money supply on the demand for real purchasing pow er by controlling for the effect o f business cycles, population and the alternative cost o f holding money. Annual population data are not available back to 1780, so time is used as a rough proxy to control for popula tion growth. The interest rate on 3 percent consols was included as a measure o f the alternative cost o f holding wealth in the form o f money. In addition, a dummy variable is included to test for a shift in the relationship in the period subsequent to the government takeover. The coefficients o f the proxies for the business cycle and population are significant and have the expected signs. The interest rate proved insignificant and was excluded from the estimate. The coefficient o f D„ is negative and significant. Its magnitude implies that the demand for real purchasing pow er fell by about 12 percent when the government seized control o f the Bank 37 'Corrected for first-order autocorrelation, t-values in parentheses. 2Since the hypothesis excludes positive values for D0, a one-tailed t-test is employed. "Significantly different from zero at the 5 percent level. o f 4.85 percent) during the years 1793-1813.35 This was a result o f a significant increase in the m onetary growth rate. Note that the rate o f inflation closely cor responds to the rate o f growth in the money supply during this period.30 The Real Value o f Bank Money The theory implies that the demand for real pur chasing pow er will fall if the guarantee regarding the low rate of inflation is broken.30 36The issue of private vs. government control of the money stock might seem to be a red herring since the money supply and price level always rise during wars. This, however, was not the case in two previous instances. During the Seven Years' War (1755-63), the government did not tamper with the Bank's control over the money supply, and the mean rates of growth in money and prices were 2.94 and 1.18 percent, respectively. Neither of these magnitudes differs significantly from zero. Similarly, during the War of Jenkin's Ear (1739-43), the mean rates of growth in money and prices were -1.62 and .66 percent, respectively. Again, neither of these differs significantly from zero. 37Exactly when the public became aware of a break in the trust is problematical. They certainly were aware of it by 1797 when the government ordered the Bank to suspend specie payments. The data, however, suggest an earlier date. Prices began rising rapidly in 1790 and, shortly afterwards, the public began arbitraging the difference between the price of gold in terms of notes at the Bank and its price in the foreign market. The Bank’s bullion account began to decline in 1791, then fell substantially in 1792. The following C O N C L U SIO N The above analysis suggests that decisions regarding the control o f m oney depend more on the incentives individuals face in making choices than on the partic u lar in d ivid u a ls w h o m ake the ch oice. Various methods o f organizing monetary control produce dis tinct policy outcomes insofar as they confront policy makers with different incentives. Since it is unclear, for example, that the incentives confronted by the policy maker w ould be much different if monetary control were placed in the hands o f Congress or the Treasury instead o f the Board o f Governors, it is unclear that the adoption o f either o f these alternatives would cause a noticeable change in policy. Additional research along these lines may prove helpful in suggesting a system of incentives that will induce the present-day equivalents o f the Court o f Directors to assign the desired weights (whatever they happen to be) to present and future consequences in reaching decisions regarding m one tary control. assumes the public became aware of the break in 1793 when the war with France began. 38As was the case with the price of Bank stock, the demand for real purchasing power did not return to its original level when the govern ment returned monetary control to the Bank and the guarantee was reinstated. This was tested by the same procedure as that employed in the case of the price of Bank stock. The results were the same. The coefficient of the second dummy was insignificant, indicating that it is redundant. 21 APRIL 1984 FEDERAL RESERVE BANK OF ST. LOUIS Alchian, Armen A. REFERENCES ________ “Government Revenue from Inflation,” Journal of Politi cal Economy (July/August 1971), pp. 846-56. "Some Economics of Property Rights,” Eco Homer, Sidney. A History of Interest Rates (Rutgers University Press, 1977). nomic Forces at Work (Liberty Press, 1977). Andreades, A. 1924). History of the Bank of England (P. S. King and Son, Bailey, Martin L. “The Welfare Cost of Inflationary Finance," The Journal of Political Economy (April 1956), pp. 93-110. Balbach, Anatol B. "How Controllable is Money Growth?” this Re view (April 1981), pp. 3-12. Bisschop, W. R. The Rise of the London Money Market: 1640-1826 (Burt Franklin, 1967). Cannan, Edwin. Son, 1919). The Paper Pound of 1797-1821 (P. S. King and Johannes, James M., and Robert H. Rasche. “Predicting the Money Multiplier,” Journal of Monetary Economics (July 1979), pp. 30125. Kemmerer, Edwin W. 1944). Gold and the Gold Standard (McGraw-Hill, Macleod, M. A. The Theory of Credit, 2nd ed. (Longmans, Green and Co., 1897), vol. 2, part 2. Martin, Frederick. Co., 1865). Stories of Banks and Bankers (Macmillan and Rogers, Thorold. The First Nine Years of the Bank of England (Clarendon Press, 1887). Clapham, Sir John. The Bank of England a History, vols. 1 and 2 (Cambridge University Press, 1958). Sinclair, Sir John. The History of Public Revenue of the British Empire (A. M. Kelly: New York, 1803). Friedman, Milton. Viner, Jacob. Studies in the Theory of International Trade (Harper and Brothers, 1937). “Discussion of the Inflationary Gap," Essays In Positive Economics (University of Chicago Press, 1953). 2 Money Growth Variability and GNP Michael T. Belongia D i MjiCENTLY, a number o f economists have argued that sharp fluctuations in the short-run growth rate of M l since 1979 have reduced GNP growth, raised in terest rates and generated expectations of higher fu ture inflation. Milton Friedman, for one, has concluded that variable money growth — bv producing these conditions — was responsible for the shorter and more abrupt cycles in real income experienced over that period.1 Based on slightly different analyses, Bomhoff, and Mascaro and Meltzer also have concluded that variable money growth has tended to lower the level of output.2 Finally, a recent conference sponsored by The Cato Institute was devoted entirely to the adverse effects o f variable m oney growth and methods by which money growth could be made more stable. ’ E con om ic th eory im p lies that variable m oney growth could low er the level o f GNP by reducing its short-run growth rate, if this variability were associated with certain changes in money demand and velocity. This article reviews the theoretical case for such a link and provides empirical evidence on the existence of this relationship. The results support the notion that variable money growth — by increasing money d e mand and reducing velocity — has had significant negative effects on both the level and the growth rate of nominal GNP in recent years. Michael T. Belongia is an economist at the Federal Reserve Bank of St. Louis. John G. Schulte provided research assistance. ’ Friedman (1983). 2Bomhoff (1983); Mascaro and Meltzer (1983). 3See The Search for Stable Money (1983). THEORETICAL RELATIONSHIPS The most common approach to constructing a link between variable money growth and GNP is based on intermediate relationships involving money demand. Although the theory behind these relationships sug gests that more variable m oney growth will increase uncertainty about future econom ic conditions and in crease the demand for money, the empirical evidence on this hypothesis has been mixed.4 The discussion that follows, however, proceeds with a standard model of m oney dem and and shows how more variable money growth — by increasing uncertainty — can be linked to a decline in the level o f income and, possibly, the long-run growth rate o f GNP. Since the expected effects o f variable m oney grow th on inflation are assumed to be small, the conclusions that follow apply to real GNP as well."’ The Basic Tobin Model A money demand m odel derived by Tobin suggests that there is an explicit relationship between uncer- 40ne statement of uncertainty's effect on money demand and interest rates is found in Friedman and Schwartz (1982), p. 39: Another variable that is likely to be important empirically is the degree of economic stability expected to prevail in the future. Wealth holders are likely to attach considerably more value to liquidity when they expect economic conditions to be unstable than when they expect them to be highly stable. This variable is likely to be difficult to express quantitatively even though the direction of the change may be clear from qualitative information. For example, the outbreak of war clearly produces expecta tions of instability, which is one reason war is often accompanied by a notable increase in real balances — that is, a notable decline in velocity. 5For one argument to support this assumption, see Friedman. 23 FEDERAL RESERVE BANK OF ST. LOUIS APRIL 1984 Figure 1 The Links Between Variable Money Growth and GNP taintv about future values o f interest rates and money demand.1’ In its most basic form, the model assumes that an individual can hold both money and govern ment bonds in his portfolio. Moreover, if the yield on money is zero, both the expected return o f the portfolio and its variance depend on only the bond yield and the proportion o f the total portfolio held in bonds. Therefore, in this simple world, an individual who seeks to maximize utility bv holding some combination o f cash balances and bonds in his portfolio faces a tradeoff between return and risk. That is to say, he can hold more bonds and increase the return on his port folio only at the cost o f increased risk: if the interest rate rises, the value of his bonds will fall. He can reduce risk, however, only by holding more cash balances, which reduces earnings. This m odel implies that risk and money demand are negatively related.' If more variable money growth in creases uncertainty about future values o f interest rates, greater money growth variability will result in an increase in money demand. This inverse relationship has been supported empirically in several studies.K What remains to be seen, however, is whether this type of shift in money demand can be linked to a decrease in the level o f GNP. 6Tobin (1958). 7Some economists disagree with this conclusion. For discussions of the theoretical indeterminacy of a sign relating uncertainty to money demand and supporting evidence, see Blejer (1979), Levi and Makin (1979), Smirlock (1982), Fieleke (1982), and Berson (1983). 8Klein (1977), Slovin and Sushka (1983), and Mascaro and Meltzer. Digitized for 24 FRASER Money Demand, Velocity and GNP The sequence of events depicted in figure 1 illus trates the first-round effects o f greater money growth variability on uncertainty, m oney demand, velocity and GNP.!1 Reading from the figure’s left side, more variable m oney grow th is h yp oth esized to cause greater uncertainty about future econom ic conditions. Increased uncertainty increases the precautionary de mand for money. A higher level o f money demand implies lower velocity (V). From the equation o f ex change, MV = Y, low er velocity clearly implies a lower level for GNP (Y). Because GNP will shift to a low er level with some lag, this level shift will be observed as a tem poraiy decline in the growth rate o f GNP. After the adjustment process is complete, the growth o f GNP should return to its long-run equilibrium path unless further changes in uncertainty and risk premia (or other exogenous shocks) set off another round o f shifts in the levels o f money demand and velocity. Theoretical Indeterminacy: Several Paths f o r GNP Are Possible Whether increased uncertainty about future money growth has any effect on GNP, however, is an empirical issue. Moreover, if increased uncertainty does have some effect on these variables, the nature o f its effect could cause GNP to follow one o f several different paths. For example, if the effect o f greater uncertainty is a once-and-for-all shift in m oney demand, the level of 9This figure is adopted from a similar figure in Bomhoff, p. 98. FEDERAL RESERVE BANK OF ST. LOUIS APRIL 1984 Table 1 Estimates of a Reduced-Form GNP Equation Adding a Measure of Money Growth Variability__________________ n . p Y, = aQ + 2 b,Mt , + 2 Cj(EP - P), , i =0 j =0 + q 2 gk VARMt_k + dQSt + et affected permanently. A third possibility is that greater uncertainty will alter investment decisions in a man ner that also changes the econom y’s long-run capitallabor ratio: in this case, both the level and growth rate of GNP would be permanently lower. Finally, money growth variability may have no observable effect on uncertainty, money demand and velocity; in this event, neither the level nor the growth rate o f GNP w ould be affected. Hypotheses concerning the impact o f in creased money growth variability and these alternative paths for GNP are tested in the next section. k =0 Sample: 11/1962--IV/1983 Sample: 11/1962--111/1979 aQ 5.536 (6.06)' 3.002 (1.77) bD bi b2 2b 0.373 0.164 0.499 1.036 (3.70) (1.48) (4.63) (0.26)2 0.448 0.244 0.347 1.039 (2.62) (1.18) (2.03) (0.23) Co 0.000 0.064 -0.081 -0.011 -0.022 0.035 0.099 0.084 (0.00) (1.45) (1.85) (0.28) (0.60) (0.97) (3.25) (2.38) - 0.047 0.114 -0.085 -0.012 -0.007 0.035 0.059 0.057 (1.26) (2.12) (1.55) (0.23) (0.13) (0.68) (1.54) (1.28) 93 94 95 2g 0.529 -0.341 -1.675 1.577 0.826 -2.311 -1.395 (1.22) (0.58) (3.03) (2.55) (1.21) (5.06) (4.75)3 1.596 -1.794 -1.034 2.541 0.590 -1.894 0.005 (2.05) (1.87) (1.16) (2.76) (0.58) (2.38) (0.00) S -0.621 (3.36) -0.634 (3.05) R2 0.62 0.58 DW 2.33 2.14 SE 2.692 2.646 c, c2 c3 c4 C5 c6 2c 9o 9i 92 'Absolute values of t-statistics in parentheses. 2 2This t-statistic applies to the null hypothesis 2 b, = 1. i =0 3The F-statistic for the null hypothesis gD = g, = ... = g5 = 0 is 10.09, which is greater than the critical value of F6 69 — 2.22. GNP will be permanently lower, but its growth rate eventually will return to its former path. If the shift in money demand is transitory, however, there w ill be a short-run decline in the growth rate o f GNP, but nei ther the level nor the growth rate o f income will be SOME TESTS OF THE HYPOTHESES RELATING MONETARY VARIABILITY TO INCOME The effects o f variable money growth on GNP can be tested by adding a measure o f m oney growth variability to a basic reduced-form monetarist m odel o f nominal GNP growth. The general reduced-form GNP equation to be estimated is shown at the top o f table 1. This equation expresses nominal GNP growth (Y) as a func tion o f the growth rate o f M l (M), the relative price of energy (EP — P), the variability o f m oney growth (VARM) and S, a variable that denotes periods o f major strikes: the strike variable is defined as the change in the quarterly average o f days lost due to strikes, de flated by the size o f the civilian labor force.10 The measure o f money growth variability chosen is the square root o f a four-quarter moving average of squared errors o f m oney growth forecasts over the I/1950-IV/1983 sample period.11 The errors then were used to construct a measure o f error variability meant to represent changes in the risk or uncertainty faced by econom ic agents as the pattern o f m oney growth changes. Intuitively, one might conclude that risk has 10The model chosen is discussed in Tatom (1981). The initial speci fication of the equation in table 1 also includes high-employment government expenditures as a right-hand-side variable. Pre-test statistics, however, indicated no significant marginal contribution to the model’s explanatory power from this variable. This pre-test result is consistent with earlier studies that have found no long-run effect of government spending on GNP growth. See, for example, Andersen and Jordan (1968); Carlson (1978); and Hafer (1982). For these reasons, the variable was omitted from the equation esti mated in this paper. 1'See Berson on the construction of a similar measure. The transfor mation is defined as: [(UM?_, + UMf_2 + UMf-3 + UM?_4) - 4]1/2, where UM represents unanticipated money growth, i.e., the re siduals from an autoregressive model of money growth. Errors were generated by fitting a sixth-order autoregressive model to the growth rate of M1. 25 APRIL 1984 FEDERAL RESERVE BANK OF ST. LOUIS Chart 1 Estimates of M onetary Uncertainty Under Actual and Simulated Ml G row th increased if forecasting errors begin to fall over an increasingly w ider range. After all, the probability of making an incorrect econom ic decision increases with the probability o f making a large forecasting error. This measure o f money growth variability, represented by the solid red line in chart 1, shows that forecast errors for M l growth have been considerably more variable since 1979. Pre-Test Estimation and Lag Length Selection The unknowns to be determined in this equation prior to estimation are the lag lengths for money growth, relative energy prices and m oney growth variability (i.e., the n, p and q shown in table 1). These values were chosen following procedures discussed recently by Batten and Thornton.1- Pre-testing indi 12Batten and Thornton (1983a, b) summarize an approach to the selection of lag length and polynomial degree based on the work o' 26 cated the use o f contemporaneous and two lagged quarterly values of the growth rate o f M l, contem pora neous and six lags for the relative price o f energy, and contemporaneous and five lags for the measure of money growth variability.13 The choice o f five lags for the measure of money growth variability reflects the lagged responses of money demand, velocity and GNP suggested bv theory and depicted in figure 1. That is to say, increased variability in m oney growth is expected to affect GNP only after some lag; econom ic agents require sufficient time both to discover the w ider band o f errors on money growth forecasts and to adjust their behavior accordingly. To test whether increased uncertainty Geweke and Meese (1981); Mallows (1973); Schwartz (1978); Akaike (1969); and Pagano and Hartley (1981). ,3The Pagano-Hartley t-ratios, final prediction errors and Mallows’ test statistic all suggested these lag lengths. These lag lengths were fitted and chosen using ordinary distributed lag models without polynomial smoothing. FEDERAL RESERVE BANK OF ST. LOUIS APRIL 1984 Table 2 Implications of Alternative Pairs of Test Results for Estimated Coefficients on Monetary Variability Practical Implication Individual Coefficients Sum of Coefficients 1) Neither the level nor the growth rate of GNP is affected each equals zero equals zero 2) The level of GNP is temporarily or permanently lower but its long-run growth rate is unaffected some initial coefficients are significantly negative equals zero 3) Both the level and long-run growth rate of GNP are permanently lower some initial coefficients are significantly negative is significantly negative has an effect on the level o f income, the relevant null hypothesis is g() = g, = ... = g5 = 0, as shown in table 2. Failure to reject this hypothesis w ould imply that m oney growth variability had no effect on GNP. If one or more individual coefficients indicate a sta tistically significant negative relationship between GNP growth and money variability, the second issue o f in terest is whether this effect on the level and the growth rate is transitoiy or permanent. In other words, it is important to know whether greater money growth variability causes a tem poraiy or permanent reduction in the level and growth rate of GNP. This result can be 5 determined by testing the null hypothesis that X k= 0 gk = 0. If this sum is not significantly different from zero but some individual coefficients are significantly negative, the results w ould im ply a transitoiy decline in the growth rate o f GNP and either a temporary or permanent reduction in its level. If this hypothesis also is rejected, however, it can be determined that both the level and growth rate o f GNP are permanently lower. Implications o f possible test results are summarized in table 2. TESTING THE IMPACT OF VARIABLE MONEY GROWTH The results o f estimating the augmented GNP equa tion over the II/1962-IV/1983 sample period are given in the first column o f table 1. The results reject each o f the null hypotheses discussed above: some initial indi vidual coefficients for money growth variability are sig nificantly negative and their sum is significantly nega tive. Within the context o f the specified equation, these results indicate that greater short-term variations in the rate o f money growth tend to increase uncertainty and money demand; as a result, permanent reductions in both the level and the growth rate o f nominal in come are produced. It also is important to note that the sum o f the 2 coefficients on monev growth ( X b,l is not signifii= 0 cantly different from one after the addition o f a direct measure o f money growth variability. This shows that the one-to-one long-run correlation betw een the growth rates o f money and nominal GNP remains, even after the effect o f variable money growth is directly taken into account.14 The significance tests on the other variables in cluded in the regression indicate that the strike vari able has negative effects on incom e growth. Also, changes in the relative price o f energy have exhibited som e significant positive long-run effects on GNP growth. This latter result is not surprising; the impacts o f short-run changes in relative energy prices are usually measured as changes in inflation. Thus, the relative energy price effect shows up in nominal GNP (via the price change); and this explains the positive sum coefficient for relative energy prices in this model. 14These results hold for a variety of variability measures, including a moving standard deviation of money growth, squared money growth rates and a multi-state Kalman filter estimate of the variance of errors associated with one-quarter-ahead forecasts of money growth. Unlike the criticisms of Allen with regard to uncertainty results for money demand, these results for a GNP equation appear to be robust with respect to the measurement of money growth variability. See Allen (1982). 27 APRIL 1984 FEDERAL RESERVE BANK OF ST. LOUIS Robustness As a check o f the m odel’s robustness, the equation in table 1 was re-estimated over a shorter II/1962-III/1979 sam ple period. This p eriod was chosen for tw o reasons. First, the Federal Reserve changed its operat ing procedures in October 1979. Second, as shown in chart 1, there was a sharp increase in money growth variability after IV/1979. The results o f re-estimating the GNP equation over the shorter sample period with new values for VARM are given in the second column of table 1.1S The results for the shorter estimation period still indicate that variable money growth temporarily low ers the growth rate o f GNP. The long-run impacts on the level and growth rate o f GNP, however, are no longer significantly different from zero. Apparently, the considerably lower variability o f money growth that existed prior to 1980 did not produce anv long-run impact on the growth o f GNP. Or, viewed differently, even though variable money growth has a significantly negative effect on GNP in both periods, permanent reductions in its level and growth rate are found only after 1980, w hen the variability o f m oney growth tripled. The effects of money growth and relative energy prices also follow lag patterns similar to those for the longer sample period. However, the long-run effect of relative energy prices is no longer significantly positive. T h e o n ly o th e r a p p a re n t c h a n g e fro m th e full p e rio d estim a tio n to this re s tric te d o n e is a d e c lin e in th e e s tim a te d g r o w th rate o f v e lo c ity (the m o d e l’s co n sta n t term ) to 3.0 fro m 5.5. H o w ever, sin ce th e g r o w th rate o f velocity in this model is reallv a() + 5 2 gk, the implied k= 0 velocity growth for the full-sample model is actually 4.14, which is not significantly different from 3.0.lfa In all other respects, the results for both models are qualita tively similar and w ould seem to indicate that the addition o f a money variability measure is robust with respect to choice o f sample period. ' 5To reflect the less volatile pattern of money growth that prevailed prior to 1980, the autoregressive model of money growth used to generate values for the money variability measure was reestimated. An AR(1) model was found to whiten the residuals for a model of money growth estimated over the pre-1980 sample. 16The F-statistic for H0: ac + 5 £ k= 0 critical value for F, 60 = 4.00. 28 gk = 3.0 is 1.83, less than the SOME IMPLICATIONS OF REDUCED MONETARY VARIABILITY FOR MONETARY POLICY The estimates reported in table 1 support the hy pothesized negative relationship betw een variable money growth and GNP discussed elsewhere.17 H ow ever, the statistical measure o f m oney growth variabil ity is not expressed in units that have a clear econom ic meaning. Therefore, the results in table 1 may be dif ficult to interpret directly, especially for policy pur poses. It may be useful to illustrate more intuitively w hy some economists are concerned about the poten tial negative effects o f m oney volatility. This is done below by using the equation in table 1 to repeat an experiment recently suggested by Friedman.18 Friedman asked what the path o f GNP w ould have been in recent years if the money stock had grown at the following rates over these intervals: 7.1 percent from III/1979 to III/1980; 6.1 percent from III/1980 to III/1981; and 5.1 percent from III/1981 to III/1982.1UThe 6.1 percent three-vear average growth rate described above is equal to its actual average over the same period. The plots of both actual M l growth and Fried man’s smoothed m oney path are shown in the upper panel o f chart 2. While maintaining the same average growth rates of money over four quarters, the Friedman scenario sig nificantly reduces the large quarter-to-quarter varia tions in M l growth that actually occurred over this period. This result is shown clearly by the sharp de cline in money growth variability that is generated bv these data; this new measure o f monetary uncertainty is represented by the dashed line in chart 1. Over the III/1979—III/1982 period, the more stable path o f M l growth w ould have produced — in terms o f Fried man's analysis — a longer but less severe recession in 1980 and, beginning around mid-1981, an expansion typical o f the postwar period (lasting about three years). The projected path o f GNP under stable M l growth is contrasted in the low er panel of chart 2 with the projected path o f GNP under actual money growth. The solid black line in the low er panel of chart 2 is the path of GNP produced by a simulation o f the model reported in the second column o f table 1 based on the 17For example, Friedman and Schwartz (1963b); Friedman; The Search for Stable Money. 18Friedman. ,9The experiment stops at this point because money growth acceler ated sharply and varied substantially over subsequent quarters. FEDERAL RESERVE BANK OF ST. LOUIS APRIL 1984 Chart 2 GNP Growth and Alternative M oney Growth Paths Actual and Smoothed Money Series Actual and Smoothed Paths of GNP Growth Percent Percent 24 24 Simulat on with actual money * /\ / \y / / 20 16 / 1 1 Simulation with smoi jthed m o n e y / 1 12 1 \\\ \ / r I \ 8 \ / 1 / \ / 1/ \ A » / /I 1 i f # \ Y/ 1i 1 / 1 / M N1 1' / \ V '' /# g / \ 16 1 \\ 111 y \\ \ J \ I V II A ' I — ■v v 7 \ ' 11 1 \ / \l ' 1 / \i » \\ i M / // / 20 4 M - 1 11 1 11 // U / # // * Actual / Y ' A\ '» ' / Iku ' / / \v /,/ ^ \1Vv \ \' l f \» L \ / // \I 12 I / > S A\. / #/ #/ . \ \ X X X X '' '' t \ / ## # # 1980 8 »\ \\ ■4 1979 ' -4 1981 1982 29 APRIL 1984 FEDERAL RESERVE BANK OF ST. LOUIS smoothed m oney growth figures listed above.20 The results are quite similar to Friedman's conjecture; moreover, they depict clearly what some economists claim are the prospective benefits o f more stable money growth. The simulated path o f GNP growth — under reduced quarter-to-quarter variation in M l growth — shows higher average growth and much narrower variation than does actual GNP growth over this period. For example, actual GNP growth ranged between —2 and 20 percent; under more stable money growth, however, the simulated rates of growth in GNP vary between 7 and 12.5 percent. Moreover, while simulated GNP growth using actual money growth rates fell to zero in III/1982 and was 5 percent or below in three o f the 12 quarters shown, the simulated path of GNP growth under less variable money growth fell below 7.5 percent on only one occasion. In summary, the contrasting results shown in chart 2 suggest that more stable money growth could promote a higher average level o f GNP growth and reduce the range in which GNP growth fluctuates. CONCLUSIONS A number o f recent studies have argued that variabil ity in the quarter-to-quarter growth rate of money has increased m oney demand and, therefore, decreased the growth rate o f GNP in the short run. This article investigates the link between variable money growth and GNP bv adding a measure o f m oney growth variability to a specific m odel o f GNP. Th e results suggest that increased quarter-toquarter variation in the growth rate o f M l has some transitory negative effects both on the level and growth rate of nominal GNP; moreover, in more recent year's, when the variation in money growth has increased nearly threefold, there is some evidence that the effects on the level and growth rate o f GNP have been permanent reductions. If the effect of money variability on inflation is small, as is generally thought, these results imply a permanently low er level and, perhaps, smaller growth rate o f real GNP. A simulation experiment based on these results illustrates the potential benefits o f more stable money growth. Within the context o f the m odel used, growth in nominal GNP w ould have been higher, on average, and more stable since 1979 if the quarter-to-quarter growth in M l had been substantially less variable than it actually has been since then. REFERENCES Akaike, H. “Statistical Predictor Identification,” Annals of the Insti tute of Statistical Mathematics, vol. 21 (1969), pp. 203-17. Allen, Stuart D. “Klein’s Price Variability Terms in the U.S. Demand for Money,” Journal of Money, Credit and Banking (November 1982), pp. 525-30. Andersen, Leonall C., and Jerry L. Jordan. “Monetary and Fiscal Actions: A Test of their Relative Importance in Economic Stabiliza tion,” this Review (November 1968), pp. 11-24. Batten, Dallas S., and Daniel L. Thornton. “Polynomial Distributed Lags and the Estimation of the St. Louis Equation,” this Review (April 1983a), pp. 13-25. ----------- “Lag-Length Selection Criteria: Empirical Results from the St. Louis Equation,” Federal Reserve Bank of St. Louis Work ing Paper No. 83-008 (1983b). Berson, David W. "Money Growth Volatility, Uncertainty, and High Interest Rates," Federal Reserve Bank of Kansas City Economic Review (November 1983), pp. 23-38. Blejer, M. I. "The Demand for Money and the Variability of the Rate of Inflation: Some Empirical Results,” International Economic Re view (June 1979), pp. 545-49. Bomhoff, Eduard J. Holland, 1983). Monetary Uncertainty (Amsterdam: North- Carlson, Keith M. “Does the St. Louis Equation Now Believe in Fiscal Policy?” this Review (February 1978), pp. 13-19. Evans, Paul. “Price-Level Instability and Output in the U.S.,” Eco nomic Inquiry (April 1983), pp. 172-87. Fey, R. A., and N. C. Jain. "Identification and Testing of Optimal Lag Structures and Causality in Economic Forecasting,” in O. D. Anderson and M. R. Perryman, eds., Applied Time Series Analysis (Amsterdam: North Holland, 1982), pp. 65-73. Fieleke, Norman S. “Fluctuations in Economic Activity and the Money Supply: An Overview,” New England Economic Review (May/June 1982), pp. 5-14. Friedman, Milton. “What Could Reasonably Have Been Expected From Monetarism: The United States,” paper presented at the Mont Pelerin Society Meeting, Vancouver, British Columbia, Au gust 29, 1983. Friedman, Milton, and Anna J. Schwartz. A Monetary History of the United States, 1867-1960 (Princeton University Press, 1963a). 20The dashed line was derived by using actual money growth rates and the errors from an autoregressive model fit to actual money data. These data provided the basis for projected GNP growth from 111/1979-111/1982 under actual monetary conditions. Actual M1 data then were replaced with Friedman’s figures for the 111/1979-111/1982 interval. The autoregressive model for money growth then was re-estimated over 1/1960-111/1982 to generate a new error series and a new measure of money growth variability. The coefficient estimates reported in the second column of table 1 were used to re-simulate GNP growth over 1979-82 in an environment of more stable money growth. These simulated results are shown by the solid black line in chart 2. 30 ________ “Money and Business Cycles,” supplement to The Re view of Economics and Statistics (February 1963b), pp. 32-64. ________ Monetary Trends in the United States and the United Kingdom: Their Relation to Income, Prices and Interest Rates, 1867-1975 (University of Chicago Press, 1982). Geweke, J., and R. Meese. “Estimating Regression Models of Finite But Unknown Order," International Economic Review (February 1981), pp. 55-70. Hafer, R. W. “The Role of Fiscal Policy in the St. Louis Equation,” this Review (January 1982), pp. 17-22. FEDERAL RESERVE BANK OF ST. LOUIS Klein, Benjamin. “The Demand for Quality-adjusted Cash Balances: Price Uncertainty in the U.S. Demand for Money Function,” Journal of Political Economy (August 1977), pp. 691-715. Levi, Maurice D., and John H. Makin. "Fisher, Phillips, Friedman and the Measured Impact of Inflation on Interest,” Journal of Fi nance (March 1979), pp. 35-52. Mallows, C. L. “Some Comments on Cp," Technometrics (Novem ber 1973), pp. 661-75. Mascaro, Angelo, and Allan H. Meltzer. “Long- and Short-Term Interest Rates in a Risky World,” Journal of Monetary Economics (November 1983), pp. 485-518. Pagano, Marcello, and Michael J. Hartley. “On Fitting Distributed Lag Models Subject to Polynomial Restrictions,” Journal of Econo metrics (June 1981), pp. 171-98. APRIL 1984 Schwartz, G. “Estimating the Dimension of a Model," Annals of Statistics, vol. 6 (March 1978), pp. 461-64. Slovin, Myron B., and Marie Elizabeth Sushka. "Money, Interest Rates and Risk," Journal of Monetary Economics (September 1983), pp. 475-82. Smirlock, Michael. “Inflation Uncertainty and the Demand for Money,” Economic Inquiry (July 1982), pp. 355-63. Tatom, John A. “Energy Prices and Short-Run Economic Perfor mance," this Review (January 1981), pp. 3-17. “The Search for Stable Money.” James A. Dorn, ed., The Cato Journal (Spring 1983). Tobin, James. “Liquidity Preference as Behavior Toward Risk," The Review of Economic Studies (February 1958), pp. 65-86. 31