The full text on this page is automatically extracted from the file linked above and may contain errors and inconsistencies.
____________Review____________ Vol. 66, No. 3 March 1984 5 A Guide to Foreign Exchange iYlarkets 19 The Money-GNP Link: Assessing Alternative Transaction Measures The Review is published 10 times per year by the Research and Public Information Department o f the Federal Reserve Bank o f St. Louis. Single-copy subscriptions are available to the public fr e e o f charge. Mail requests fo r subscriptions, back issues, or address changes to: Research and Public Information Department, Federal Reserve Bank o f St. Louis, P.O. Box 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 Bank o f St. Louis or the Federal Reserve System. Articles herein may be reprinted provided the source is credited. Please provide the Bank’s Research and Public Inform a tion Department with a copy o f reprinted material. Federal Reserve Bank of St. Louis Review M arch 1984 In This Issue .. . In the first article of this Review, “A Guide to Foreign Exchange Markets,” K. Alec Chrystal guides the reader through the com plexities of foreign exchange markets. Chiystal first describes how currencies are traded, pointing out the key differences between the retail markets and the wholesale or interbank markets for “spot” foreign exchange. He then discusses how the existence of forward currency markets enables importers and exporters to avoid exchange rate risk. The author then considers the newly emergent futures and options markets in foreign exchange and analyzes options as a method of hedging. The im portance of various kinds of arbitrage and speculation in providing an efficient and liquid foreign exchange market is also outlined. Finally, the author discusses the special role of the dollar as the money of the foreign exchange markets. In the second article of this Review, “The Money-GNP Link: Assessing Alternative Transaction M easures,” R. W. Hafer notes that some have questioned the reliability of the link between M l and GNP, given recent velocity developments in 1982 and 1983. He investigates the empirical relationship between econom ic activity and two alternative transaction m easures of money. These two m easures are, respec tively, the narrowly defined m onetaiy aggregate, M l, and one that excludes from M l those checkable deposits that earn explicit interest income, such as NOW accounts. This latter measure is referred to as adjusted M l. Arguing that the introduction of NOW accounts in 1981 represents a m ajor but predictable shift in the relationship between GNP and money, Hafer dem onstrates that the difficulty in explaining GNP movements disappears when the adjusted M l series is used. The author's analysis shows that, when his measure of transaction balances is adjusted for the NOW account effect, the relationship between ad justed M l and GNP displays no deterioration in overall "explanatory pow er” when estim ated through 1983. In contrast, equations estim ated using the current M l m easure experience about a 30 percent reduction in explanatoiy power. This result, Hafer argues, “arises from the public’s willingness to view some portion of interest-bearing checkable deposits as savings-type balances." Based on his em pir ical results, the author denies the claim that the link between transactions money, properly defined, and GNP has been damaged irreparably. 3 A Guide to Foreign Exchange Markets K. Alec Chrystal T M . HE econom ies of the free world are becom ing increasingly interdependent. U.S. exports now amount to almost 10 percent of Gross National Product. For both Britain and Canada, the figure currently exceeds 25 percent. Imports are about the same size. Trade of this magnitude would not be possible without the ability to buy and sell currencies. Currencies must be bought and sold because the acceptable m eans of pay ment in other countries is not the U.S. dollar. As a result, importers, exporters, travel agents, tourists and many others with overseas business must change dol lars into foreign currency and/or the reverse. be described. This will be followed by a discussion of some of the more important activities of market partici pants. Finally, there will be an introduction to the analysis of a new feature of exchange markets — cur rency options. The concern of this paper is with the structure and m echanics of foreign exchange markets, not with the determ inants of exchange rates them selves. The trading of currencies takes place in foreign ex change markets w hose m ajor function is to facilitate international trade and investment. Foreign exchange m arkets, however, are shrouded in mystery. One reason for this is that a considerable amount of foreign exchange market activity does not appear to be related directly to the needs of international trade and invest ment. There is an almost bewildering variety of foreign exchange markets. Spot markets and forward markets abound in a num ber of currencies. In addition, there are diverse prices quoted for these currencies. This section attem pts to bring order to this seeming dis array. The purpose of this paper is to explain how these markets work.1 The basics of foreign exchange will first Virtually every m ajor newspaper, such as the Wall Street Journal or the London Financial Times, prints a daily list of exchange rates. These are expressed either as the num ber of units of a particular currency that exchange for one U.S. dollar or as the num ber of U.S. dollars that exchange for one unit of a particular cur rency. Sometimes both are listed side by side (see table 1). K. Alec Chrystal, professor o f econom ics-elect, University of Sheffield, England, is a visiting scholar at the Federal Reserve Bank of St. Louis. Leslie Bailis Koppel provided research assistance. The author wishes to thank Joseph Hempen, Centerre Bank, St. Louis, for his advice on this paper. 1For further discussion of foreign exchange markets in the United States, see Kubarych (1983). See also Dufey and Giddy (1978) and McKinnon (1979). THE BASICS OF FOREIGN EXCHANGE MARKETS Spot, Forward, Bid, Ask For m ajor currencies, up to four different prices typically will be quoted. One is the "sp ot” price. The others may be “30 days forward,” “90 days forward,” 5 FEDERAL RESERVE BANK OF ST. LOUIS MARCH 1984 Table 1 Foreign Exchange Rate Quotations Foreign Exchange The Dollar Spot and Forward Wednesday, September 7, 1983 The New York foreign exchange selling rates below apply to trading among banks in amounts of $1 million and more, as quoted at 3 p.m. Eastern time by Bankers Trust Co. Retail transactions provide fewer units of foreign currency per dollar. Currency U.S. $ equiv. per U.S. $ Country Wed. Tues. Wed. Tues. Argentina (Peso) .......... . .09652 .09652 10.36 10.36 .8777 Australia (Dollar) .......... . .8772 1.1340 1.1393 Austria (Schilling) ......... . 17.84 .05296 .0560 18.88 Belgium (Franc) Commercial rate ....... . 54.01 53.90 .01851 .01855 .01844 .01846 54.21 54.15 Financial rate ............. . Brazil (Cruzeiro) ............ . .001459 685 671.00 .00149 .6707 Britain (Pound) ............. . 1.4910 1.5000 .6666 1.5004 .6704 6664 30-Day Forward ...... . 1.4915 6697 90-Day Forward ...... . 1.4930 1.5010 6662 .6654 180-Day Forward . 1.4952 1.5028 6688 Canada (Dollar) ............. . .8120 .8123 1.2315 .2310 1.2307 30-Day Forward ...... . .8125 .8128 1.2303 .8134 .8137 1.2293 1.2289 90-Day Forward ...... . 1.2274 180-Day Forward ...... . .8145 .8147 1.2277 .01246 80.21 Chile (Official rate) ........ . .01246 80.21 China (Yuan) .................. . .50499 .50489 1 9802 1 9806 Colombia (P e so )............ . .01228 .01228 81.4 81.40 Denmark (Krone) .......... .10405 9.65 9.6100 .10362 Ecuador (Sucre) .02082 .02082 48.03 Official rate ................ . 48.03 Floating rate ............... . .010917 .010917 91.60 91.60 .17424 .17485 5.7390 5.7190 Finland (Markka) .......... . France (Franc) ............... . .1238 .1238 8 0750 8.0750 30-Day Forward ...... . .1235 .1230 8 0955 8.1300 90-Day Forward ...... . .1224 1223 8.1695 8.1725 180-Day Forward ...... . .1203 .1202 8.3100 8.3150 Greece (D rach m a )......... . .01075 .01078 93. 92.70 Hong Kong (D o lla r)....... . .1297 .13089 7.71 7.6400 India (Rupee) ................ . .0980 0980 10.20 10.20 985 985 Indonesia (R u p ia h )....... . .001015 .001015 1.1775 .8536 8493 Ireland (Punt) ................ . 1.1715 Israel (Shekel) ............... . .0173 57.80 57.80 .0173 .000624 Italy (Lira) ...................... . .0006255 1602. 1598.50 .004067 Japan (Yen) ................... . .004072 245.55 245.85 30-Day Forward ...... . .004083 .004079 244.88 245.15 .004107 243.48 90-Day Forward ...... .004102 243.75 241 10 180-Day Forward ...... .004147 004142 241.39 4.85 Lebanon (Pound) .......... .20618 .20618 4.85 Malaysia (Ringgit) ....... .42462 2.3550 42489 2.3535 Mexico (Peso) 150.00 Floating rate ............... .00665 00666 150.25 Netherlands (Guilder) ... .33288 .3333 3.0040 3.000 1.5327 .6497 6505 1.5397 New Zealand (Dollar) ... 7.48 7.4625 Norway (Krone) ............. . .13368 1340 .07518 13 30 13.30 Pakistan (Rupee) .......... .07518 .0005105 0005105 1958 89 1958.89 Peru (S o l)........................ 11.007 11.007 Philippines (Peso) ....... . .09085 09085 .00804 00807 124 35 123.90 Portugal (Escudo) ......... . Saudi Arabia (Riyal) ... . .28735 .28735 3.48 3.48 .4664 Singapore (D o lla r)......... .46609 2.1455 2.1440 1.1236 South Africa (Rand) ...... . .8870 .8900 1.1273 South Korea (Won) ...... .001285 .001285 778.20 778.20 Spain (Peseta) .............. . .00655 .00658 152 60 151.90 7.9140 7 8950 Sweden (Krona) ............ . .12635 12666 Switzerland (Franc) ...... . 4596 .4591 2.1755 2.1780 30-Day Forward ........ .4619 .4615 216.46 2.1666 .4657 90-Day Forward ...... . .4662 2.1449 2.1470 . .4728 2.1150 2.1172 180-Day Forward .4723 Taiwan (Dollar) ............. . 40.17 40.17 .02489 .02489 Thailand (Baht) ............. . .043459 .043459 23.01 23.01 Uruguay (New Peso) Financial...................... . .02798 .02798 35.73 35.73 Venezuela (Bolivar) Official rate .................. 4.30 .23256 .23256 4.30 .07194 Floating rate .............. .. .07272 13.90 13.75 W. Germany (Mark) .. .3726 .3726 2.6835 2.6835 .3741 30-Day Forward .... .. .3740 2.6731 2 6728 90-Day Forward ....... .3767 .3768 2.6540 2 6538 180-Day Forward .... .. .3808 2.6260 .3808 2.6259 SDR .. 1.04637 1.04903 .955685 UK| I reland t Canada Nethlnd. Belgium Denmark W. Ger. Portugal Spain Italy Norway France Sweden Japan Austria Switz. spread 1.4860-1.4975 1.1665-1.1720 1.2305-1.2320 3.0050-3.0150 54.06-54.20 9.6400-9.6800 2.6850-2.6980 124.20-125.00 152.40-152.70 1604-1608 7.4730-7.4940 8.0775-8.1225 7.9120-7.9265 245.50-246.50 18.89-18.9512 2.1770-2.1875 Close 1.4910-1.4920 1.1710-1.1720 1.2310-1.2315 3.0050-3.0070 54.06-54.08 9.6400-9.6450 2.6865-2.6875 124.40-124.70 152.50-152.60 1605-1606 7.4730-7.4780 8.0825-8.0875 7.9120-7.9170 245.65-245.75 18.89-18.90 2.1800-2.1810 One month 0.02-0.07c dis 0.36-0.30c pm 0.09-0.06c pm 1.12-1.02c pm 7-6c pm 2-21zore dis 1.07-1.02pff pm 115-290C dis 170 220c dis 10 - 10 12lire dis 1.90-2.20ore dis 2.02-2.12c dis 0.90-1.10ore dis 0.69-0.64y pm 7.50-6.70gro pm 1.10-1.05c pm % Three % pa. months pa. -0 .3 6 3.39 0.73 4.26 1.44 -2 .7 9 4.66 -1 9 .5 1 -1 5 .3 3 -7 .6 5 -3 .2 9 -3 .0 7 - 1 .5 1 3.24 4.50 5.91 -0 .5 2 0.17-0.22dis 2.84 0.88-0.78 pm 0.24-0.21 pm 0.73 3.00-2.90 pm 3.92 0.92 14-11 pm -0 .1 0 par-1? dis 3.00-2.95 pm 4.42 330-790dis -1 7 .9 8 675-775dis -1 8 .9 9 -7 .5 3 291r 31 dis 5.90-6.20ds -3 .2 3 9.65 -9 .85ds - 4 .8 1 -1 .1 9 2.25 -2 .45ds 2.11-2.03 pm 3.36 21.00-18.50 pm 4.17 3.10-3.05 pm 5.63 tU K and Ireland are quoted in U.S. currency. Forward premiums and discounts apply to the U.S. dollar and not to the individual currency. Belgian rate is for convertible francs. Financial franc 54.40-54.45. London Financial Times, September 8, 1983 .9536; Special Drawing Rights are based on exchange rates for the U.S., West German, British, French and Japanese currencies. Source: International Monetary Fund. z-Not quoted. 6 Day’s Sept 7 Wall Street Journal, September 8 ,1 9 8 3 MARCH 1984 FEDERAL RESERVE BANK OF ST. LOUIS and “180 days forward.1’ These m aybe expressed either in “European Term s" (such as num ber of $ per £) or in “American Term s" (such as num ber of £ per $). (See the glossary for further explanation.) The spot price is what you m ust pay to buy curren cies for immediate delivery (two working days in the interbank market; over the counter, if you buy bank notes or travelers checks). The forward prices for each currency are what you will have to pay if you sign a contract today to buy that currency on a specific future date (30 days from now, etc.). In this m arket,you pay for the currency w hen the contract matures. Why would anyone buy and sell foreign currency forward? There are some m ajor advantages from hav ing such opportunities available. For example, an ex porter who has receipts of foreign currency due at some future date can sell those funds forward now, thereby avoiding all risks associated with subsequent adverse exchange rate changes. Similarly, an importer who will have to pay for a shipm ent of goods in foreign currency in, say, three m onths can buy the foreign exchange forward and, again, avoid having to bear the exchange rate risk. The exchange rates quoted in the financial press (for example, those in table 1) are not the ones individuals would get at a local bank. Unless otherwise specified, the published prices refer to those quoted by banks to other banks for currency deals in excess of $1 million. Even these prices will vary somewhat depending upon w hether the bank buys or sells. The difference between the buying and selling price is sometim es known as the “bid-ask spread.” The spread partly reflects the banks’ costs and profit margins in transactions; however, m a jo r banks make their profits more from capital gains than from the spread.2 The market for bank notes and travelers checks is quite separate from the interbank foreign exchange market. For smaller currency exchanges, such as an individual going on vacation abroad might make, the spread is greater than in the interbank market. This presumably reflects the larger average costs — includ ing the exchange rate risks that banks face by holding bank notes in denom inations too small to be sold in the interbank market — associated with these sm aller ex changes. As a result, individuals generally pay a higher price for foreign exchange than those quoted in the newspapers. 2Notice the Wall Street Journal quotes only a bank selling price at a particular time. The Financial Times quotes the bid-ask spread and the range over the day. Table 2 Dollar Price of Deutschemarks and Sterling at Various Banks Deutschemark Buy Sell Sterling Buy Sell Retail Local (St. Louis) banks (avg.) .3572-.3844 1.4225-1.5025 W holesale New York banks .3681-.3683 1.4570-1.4580 European banks (high) .3694—.3696 1.4573-1.4583 European banks (low) .3677-3678 1.4610-1.4620 Bankers trust .3681 1.4588 Note: These prices were all quoted on November 28, 1983, be tween 2:00 p.m. and 2:45 p.m. (Central Standard Time). Prices for local banks were acquired by telephoning for their price on a $10,000 transaction. The prices quoted were reference rates and not the final price they would offer on a firm transaction. Figure for Bankers Trust is that given in the Wall Street Journal, November 29, 1983, as priced at 2:00 p.m. (Central Standard Time) on November 28, 1983. Other prices were taken from the Telerate information system at 2:35 p.m. New York prices were the latest available (Morgan and Citibank, respectively). European prices were the last prices quoted before close of trading in Europe by various banks. Deutschemark prices were actually quoted in American terms. The sell prices above have been rounded up. The difference between buy and sell prices for DM in the interbank market actually worked out at $0.00015. An example of the range of spot exchange rates avail able is presented in table 2, which shows prices for deutschemarks and sterling quoted within a one-hour period on November 28,1983. There are two important points to notice. First, all except those in the first line are prices quoted in the interbank, or wholesale, mar ket for transactions in excess of $1 million. The sterling prices have a bid-ask spread of only 0.1 cent (which is only about 0.07 percent of the price, or $7 on $10,000). On DM, the spread per dollars worth works out to be about half that on sterling ($4 on $10,000).3 Second, the prices quoted by local banks for small, or retail, transactions, which serve only as a guide and do not necessarily represent prices on actual deals, in volve a m u ch larger bid -ask spread. T h ese retail spreads vary from bank to bank, but are related to (and larger than) the interbank rates. In some cases, they 3ln practice, the spread will vary during the day, depending upon market conditions. For example, the sterling spread may be as little as 0.01 cents at times and on average is about 0.05 cents. Spreads generally will be larger on less widely traded currencies. 7 FEDERAL RESERVE BANK OF ST. LOUIS may be of the order of 4 cents or less on sterling, though the prices quoted in St. Louis involved average spreads of 8 cents on sterling. The latter represents a spread of about 5 Vi percent (about $550 per $10,000 transaction). The equivalent spread for DM was 7 percent ($700 per $10,000 transaction). The spread on forward transactions will usually be wider than on spot, especially for longer maturities. For interbank trade, the closing spread on one and three m onths forward sterling on Septem ber 8, 1983, was .15 cents, while the spot spread was .10 cents. This is shown in the top line of the Financial Times report in table 1. Of course, like the spot spread, the forward spread varies with time of day and market conditions. At times it m aybe as low as .02 cents. No information is available for the size of spread on the forward prices typically offered on small transactions, since the retail market on forward transactions is very small. HOW DOES “T H E” FOREIGN EXCHANGE MARKET O PERATE? It is generally not possible to go to a specific building and "see” the market where prices of foreign exchange are determined. With few exceptions, the vast bulk of foreign exchange business is done over the telephone between specialist divisions of m ajor banks. Foreign exchange dealers in each bank usually operate from one room; each dealer has severed telephones and is surrounded by video screens and news tapes. Typical ly, each dealer specializes in one or a small num ber of markets (such as sterling/dollar or deutschemark/dollar). Trades are conducted with other dealers who represent banks around the world. These dealers typi cally deal regularly with one another and are thus able to make firm com m itm ents by word of mouth. Only the head or regional offices of the larger banks actively deal in foreign exchange. The largest of these banks are known as "market m akers” since they stand ready to buy or sell any of the m ajor currencies on a more or less continuous basis. Unusually large transac tions, however, will only be accom m odated by market makers on more favorable terms. In such cases, foreign exchange brokers may be used as middlemen to find a taker or takers for the deal. Brokers (of which there are four m ajor firms and a handful of sm aller ones) do not trade on their own account, but specialize in setting up large foreign exchange transactions in return for a comm ission (typically 0.03 cents or less on the sterling spread). In April 1983, 56 percent of spot transactions by value involving banks in the United States were MARCH 1984 channeled through brokers.4 If all interbank transac tions are included, the figure rises to 59 percent. Most small banks and local offices of m ajor banks do not deal directly in the interbank foreign exchange market. Rather they typically will have a credit line with a large bank or their head office. Transactions will thus involve an extra step (see figure 1). The custom er deals with a local bank, which in turn deals with a major bank or head office. The interbank foreign exchange market exists between the m ajor banks either directly or indirectly via a broker. FUTURES AND OPTION MARKETS FOR FOREIGN EXCHANGE Until veiy recently, the interbank market was the only channel through which foreign exchange transac tions took place. The past decade has produced m ajor innovations in foreign exchange trading. On May 16, 1972, the International Money Market (IMM) opened under the auspices of the Chicago M ercantile Ex change. One novel feature of the IMM is that it provides a trading floor on w hich deals are struck by brokers face to face, rather than over telephone lines. The most significant difference between the IMM and the inter bank market, however, is that trading on the IMM is in futures contracts for foreign exchange, the typical busi ness being contracts for delivery on the third W ednes day of March, June, Septem ber or Decem ber. Activity at the IMM has expanded greatly since its opening. For example, during 1972, 144,336 contracts were traded; the figure for 1981 was 6,121,932. There is an important distinction between “forward" transactions and "futures" contracts. The form er are individual agreements between two parties, say, a bank and customer. The latter is a contract traded on an organized market of a standard size and settlement date, which is resalable at the market price up to the close of trading in the contract. These organized mar kets are discussed more fully below. While the m ajor banks conduct foreign exchange deals in large denom inations, the IMM trading is done in contracts of standard size which are fairly small. Examples of the standard contracts at present are £25,000; DM125,000; Canadian $100,000. These are actually smaller today than in the early days of the IMM. Further, unlike prices on the interbank market, price movements in any single day are subject to specific 4See Federal Reserve Bank of New York (1983). MARCH 1984 FEDERAL RESERVE BANK OF ST. LOUIS Figure 1 Structure of Foreign Exchange Markets Customer buys $ with DM \ \ \ \ \ Local bank * Stockbroker \ % \ \ * Foreign exchange broker 4 m m \ Major banks interbank market IMM LIFFE PSE 7 T ~ / / Local bank / / Stockbroker / / * / Customer buys DM with $ NOTE: The International Money Market (IMM) Chicago trades foreign exchange futures and DM futures options. The London International Financial Futures Exchange (LIFFE) trades foreign exchange futures. The Philadelphia Stock Exchange (PSE) trades foreign currency options. 9 FEDERAL RESERVE BANK OF ST. LOUIS limits at the IMM. For example, for sterling futures, prices are not allowed to vaiy m ore than $.0500 away from the previous day’s settlem ent price; this limit is expanded if it is reached in the same direction for two successive days. The limit does not apply on the last day a contract is traded. Unlike the interbank market, parties to a foreign ex change contract at the IMM typically do not know each other. Default risk, however, is m inor because con tracts are guaranteed by the exchange itself. To m ini mize the cost of this guarantee, the exchange insists upon “margin requirem ents” to cover fluctuations in the value of a contract. This m eans that an individual or firm buying a futures contract would, in effect, place a deposit equal to about 4 percent of the value of the contract.5 Perhaps the m ajor limitation of the IMM from the point of view of im porters or exporters is that contracts cover only eight currencies — those of Britain, Canada, West Germany, Switzerland, Japan, Mexico, France and the Netherlands — and they are specified in stan dard sizes for particular dates. Only by chance will these conform exactly to the needs of importers and exporters. Large firms and financial institutions will find the market useful, however, if they have a fairly continuous stream of payments and receipts in the traded foreign currencies. Although contracts have a specified standard date, they offer a fairly flexible method of avoiding exchange rate risk because they are marketable continuously. A m ajor econom ic advantage of the IMM for non bank custom ers is its low transaction cost. Though the brokerage cost of a contract will vary, a “round trip” (that is, one buy and one sell) costs as little as $15. This is only .04 percent of the value of a sterling contract and less for some of the larger contracts. Of course, such costs are high com pared with the interbank market, where the brokerage cost on DM 1 million would be about $6.25 (the equivalent-valued eight futures con tracts would cost $60 in brokerage, taking $7.50 per single deal). They are low, however, com pared with those in the retail market, w here the spread may in volve a cost of up to 2.5 percent or 3 percent per transaction. A market similar to the IMM, the London Interna tional Financial Futures Exchange (LIFFE), opened in Septem ber 1982. On LIFFE, futures are traded in ster 5A bank may also insist upon some minimum deposit to cover a forward contract, though there is no firm rule. Digitized for10 FRASER MARCH 1984 ling, deutschemarks, Swiss francs and yen in identical bundles to those sold on the IMM. In its first year, the foreign exchange business of LIFFE did not take off in a big way. The m ajor provider of exchange rate risk cov erage for business continues to be the bank network. Less than 5 percent of such cover is provided by mar kets such as IMM and LIFFE at present. An entirely new feature of foreign exchange markets that has arisen in the 1980s is the existence of option markets ,B The Philadelphia Exchange was the first to introduce foreign exchange options. These are in five currencies (deutschemark, sterling, Swiss franc, yen and Canadian dollar). Trades are conducted in stan dard bundles half the size of the IMM futures con tracts. The IMM introduced an options market in Ger man marks on January 24, 1984; this market trades options on futures contracts whereas the Philadelphia options are for spot currencies. Futures and options prices for foreign exchange are published daily in the financial press. Table 3 shows prices for February 14, 1984, as displayed in the Wall Street Journal on the following day. Futures prices on the IMM are presented for five currencies (left-hand column). There are five contracts quoted for each cur rency; March, June, September, Decem ber and March 1985. For each contract, opening and last settlement (settle) prices, the range over the day, the change from the previous day, the range over the life of the contract and the num ber of contracts outstanding with the exchange (open interest) are listed. Consider the March and Ju n e DM futures. March futures opened at $.3653 per mark and closed at $.3706 per mark; June opened at $.3698 per mark and closed at $.3746 per mark. Turn now to the Chicago Mercantile Exchange (IMM) futures options (cen ter colum n). These are options on the futures contracts just dis cussed (see inset for explanation of options). Thus, the line labeled "Futu res” lists the settle prices of the March and June futures as above. Let us look at the call options. These are rights to buy DM futures at specified prices — the strike price. For example, take the call option at strike price 35. This means that one can purchase an option to buy DM 125,000 March futures up to the March settlem ent date for $.3500 per mark. This option will cost 2.05 cents per mark, or $2,562.50, plus brokerage fees. The June op tion to buy June futures DM at $ .3500 per mark will cost 2.46 cents per mark, or $3,075.00, plus brokerage fees. 6For a discussion of options in commodities, see Belongia (1983). MARCH 1984 FEDERAL RESERVE BANK OF ST. LOUIS Table 3 Futures and Options Markets Futures Prices Futures Options Tuesday, February 14, 1984 O pen Interest R eflects P revious T rading Day. C hicago M ercantile Exchange .0170 1.6010 .0170 1.5520 .0160 1.5240 .0160 1.4650 .0170 1.4625 21,242, + 7 8 . 1.3930 1.3950 1.3980 1.3990 1.4000 17.694 3,251 157 75 65 .7979 .7983 .7988 .8021 .8023 4,033 740 312 152 50 .4125 .4180 .4354 .4395 25,730 3,908 974 271 .5230 .4470 .5045 .4536 .5020 .4598 .4880 .4665 .4840 .4755 +2 9 6. 24,164 3,165 153 71 5 CANADIAN DOLLAR (IMM)— 100,000 dlrs.; S per Can S Mar June Sept Dec Mar85 Est vol .8010 .8014 .8024 8010 .8020 ...............8169 .8029 .8013 .8023 ...............8168 ............................. 8026 ............... 8147 .8021 .8031 .8021 .8029 ...............8040 .8035 .8035 .8035 .8032 ...............8035 1,087; vol Mon 535; open int 5,287, -1 0 3 . mark Strike Calls— Settle Puts— Settle Mar Jun Mar Jun Price 34 0.01 0.01 35 2.05 2.46 0.01 0.09 1.11 36 1.66 0.06 0.25 37 0.38 1.00 0.33 0.57 0.54 38 0.10 1.00 1.02 39 0.01 0.27 Futures .3706 .3746 Estimated total vol. 2.187. Calls: Mon vol. 180: open int. 2,416. Puts: Mon vol. 73: open int. 1,841. JAPANESE YEN (IMM) 12.5 million yen; $ per yen (.00) Mar June Sept Dec Est vol .4276 .4315 .4354 .4416 9,133; .4297 .4337 .4375 .4420 vol Mon .4276 .4294 + .0011 4396 4312 .4334 + .0011 .4435 .4354 .4374 + .0012 .4450 .4400 .4415 + .0012 .4493 3,306; open int 30,883, +534. SWISS FRANC (IMM)— 125,000 francs; $ per franc Mar June Sept Dec Mar85 Est vol .4495 .4564 .4632 .4705 .4556 .4486 .4549 + .4629 .4557 .4622 + .4692 .4632 .4688 + .4780 .4705 .4747 + ............................. 4830 + 30,610; vol Mon 8,466; open int .0047 .0051 .0052 .0049 .0050 27,558, W. GERMAN MARK (IMM)— 125,000 marks; $ per mark Mar June Sept Dec Mar85 Est vol .3653 .3698 .3743 .3780 .3713 .3650 .3706 + 3754 .3688 .3746 + .3790 .3743 .3780 + .3825 .3780 .3825 + ............................. 3838 + 30,248; vol Mon 9,045; open int .0036 4100 .0037 4002 .0034 4030 .0043 3825 0035 .3699 36,452, +680. .3537 .3568 .3602 .3640 .3699 30,974 4,911 362 204 1 The March call option at strike price $.3900 per mark costs only O.Ol cents per mark or $12.50. These price differences indicate that the market expects the dollar price of the mark to exceed $.3500, but not to rise substantially above $.3900. Notice that w hen you exercise a futures call option you buy the relevant futures contract but only fulfill that futures contract at maturity. In contrast, the Phil adelphia foreign currency options (right column) are options to buy foreign exchange (spot) itself rather than futures. So, w hen a call option is exercised, for eign currency is obtained immediately. The only difference in presentation of the currency option prices as com pared with the futures options is that, in the former, the spot exchange rate is listed for comparison rather than the futures price. Thus, on the Philadelphia exchange, call options on M arch DM 62,500 at strike price $ .3500 per mark cost 1.99 cents per mark or $1,243.75, plus brokerage. Brokerage fees here would be of the same order as on the IMM, about $16 per transaction round trip, per contract. We have seen that there are several different markets for foreign exchange — spot, forward, futures, options Philadelphia Exchange W. GERMAN MARK— 125,000 marks, cents per Lifetim e Open Open High Low Settle Change High Low Interest BRITISH POUND (IMM)—25,000 pounds; $ per pound Mar 1.4150 1.4400 1.4150 1.4370 + June 1.4175 1.4435 1.4175 1.4395 + Sept 1.4285 1.4410 1.4220 1.4410 + Dec 1.4280 1.4435 1.4245 1.4435 + Mar85 1.4280 1.4460 1.4270 1.4470 + Est vol 10,651; vol Mon 1,987; open int Foreign Currency Options O ption & Strike Underlying Price C alls—Last Mar Jun Puts— Last SepMar Jun 12.500 British Pounds-cents per unit. BPound 140 3.40 r 5.70 0.40 1.85 143.00 .145 0.70 2.40 r 3.40 50.000 Canadian Dollars-cents per unit. CDollar . 80 r r 0.68 r 62.500 West German Marks-cents per unit. DMark . . 34 2.67 r r r 36.88 . .35 1.99 2.18 r r 36.88 36 1.04 1.59 r 0.05 0.35 36.88 . .37 0.38 1.00 r 0.37 0.56 36.88 . . .38 0.10 0.62 0.85 r 36.88 . .39 r 0.28 s r 36.88 . . .40 0.01 0.11 s r 6.250.000 Japanese Yen-100ths of a cent per unit. JYen ...42 0.95 1.49 2.04 r 42.75 . . .43 0.30 0.90 r 0.50 0.60 42.75 . 44 0.04 0.45 0.99 62.500 Swiss Francs-cents per unit. SFranc . 44 r r 3.15 r 0.2 45.18 . . .45 0.65 r 0.26 r 45.18 . . .46 0.28 1.09 1.82 r 1.00 45.18 . . .47 0.06 r r 45.18 48 0.02 0.28 r Total call vol. 2,271 Call open int. 37,349 Total put vol. 799 Put open int. 26,173 r— Not traded. s— No option offered. o— Old. Last is premium (purchase price). Sep r Wall Street Journal, February 15, 1984 on spot, options on futures. The channels through which these markets are formed are, however, fairly straightforward (see figure 1). The main channel is the interbank network, though for large interbank transac tion s, foreign exch an g e b rokers may be used as middlemen. FOREIGN EXCHANGE MARKET ACTIVITIES M uch foreign exchange market trading does not appear to be related to the simple basic purpose of allowing businesses to buy or sell foreign currency in order, say, to sell or purchase goods overseas. It is certainly easy to see the usefulness of the large range of foreign exchange transactions available through the interbank and organized markets (spot, forward, fu tures, options) to facilitate trade between nations. It is also clear that there is a useful role for foreign exchange brokers in helping to “m ake” the interbank market. There are several other activities, however, in foreign exchange markets that are less well understood and whose relevance is less obvious to people interested in understanding what these markets accom plish. 11 FEDERAL RESERVE BANK OF ST. LOUIS MARCH 1984 Foreign Exchange Options F ig u re 2 An option is a contract specifying the right to buy or sell — in this case foreign exchange — within a specific period (American option) or at a specific date (European option). A call option confers the right to buy. A put option confers the right to sell. Since each of these options must have a buyer and a seller, there are four possible ways of trading a single option: buy a call, sell a call, buy a put, sell a put. The buyer of an option has the right to undertake the contract specified but may choose not to do so if it turns out to be unprofitable. The seller of the option must fulfill the contract if the buyer desires. Clearly, the buyer must pay the seller some pre mium (the option price) for this privilege. An option that would be profitable to exercise at the current exchange rate is said to be “in the money." The price at w hich it is exercised is the "exercise” or “strike" price. Considera call option on £ 1 000 (although options of this size are not presently available on organized exchanges, it is used to present a simple illustration of the principles involved). Suppose this costs $0.03 per pound or $30 and the exercise price is $1.50 per pound. The option expires in three months. This means that the buyer has paid $30 for the right to buy £1000 with dollars at a price of $1.50 per pound any time in the next three m onths. If the current spot price of sterling is, say, $1.45, the option is "out of the m oney” b ecau se sterling can be bought cheaper on the spot market. However, if the spot price were to rise to, say, $1.55, the option would be in the money. If sold at that time, the option buyer would get a $50 return (1000 x $0.05), which would m o re th a n c o v e r t h e c o s t o f th e o p t io n ($50 —$30 = $20 profit). In contrast, a put option at the same terms would be in the m oney at the cur rent spot price of $1.45, but out of the money at $1.55. Figure 2 presents a diagrammatic illustration of how the profitability of an option depends upon the relationship between the exercise price and the cur rent spot price.1 Figure 2a illustrates the profit avail- 1The pricing of options has been the subject of a large theoretical literature with a major contribution being made by Black and Scholes (1973). The Black-Scholes formula has been modified for foreign exchange options by Garman and Kohlhagen (1983) [see also Giddy (1983)], but the Black-Scholes formula is com plex and beyond the scope of the present paper. Digitized for12 FRASER Profit from Options Buy o Call (b) Buy a Pvt Buy a Straddle One simple relationship which is of interest may be called “ option price parity.” This arises because arbitrage will ensure that the difference between a call option price (per unit) and a put option price (per unit) at the same exercise price will be equal to the present value of the difference between the exercise price and the forward exchange rate at maturity of the options (if the options are marketable, it will also hold for any date to maturity). The relationship may be expressed: when C and P are the call and put option prices at exercise price E. F is the forward exchange rate and r is the interest rate per period of the contracts. This arises because the simultaneous buying of a call and selling of a put is equivalent to buying currency forward at price E. The forward contract, however, would be paid for at the end of the period, whereas the options are transacted at the beginning. Hence, the forward contract has to be discounted back to the present. FEDERAL RESERVE BANK OF ST. LOUIS able from buying a call option at exercise price A. At spot exchange rate A and anything lower, the option will not be exercised so the loss will equal the price of the option. At a spot exchange rate above a, the option is sufficiently in the m oney to more than cover its cost. Between A and a, the option is in the money but not by enough to cover cost. The profit from selling a call could be illustrated by reversing the + and — signs in figure 2a, or by flipping the profit line about the horizontal axis. Figure 2b illustrates the profit from buying a put option. At spot exchange rates below a, the option with exercise price A will show a profit. Figure 2c illustrates the profit from a sim ul taneous purchase of a put and call at the same Two m ajor classes of activity will be discussed. First, the existence of a large num ber of foreign exchange markets in many locations creates opportunities to profit from "arbitrage.” Second, there is implicitly a market in (foreign exchange) risk bearing. Those who wish to avoid foreign exchange risk (at a price) may do so. Those who accept the risk in expectation of profits are known as “speculators.” Triangular Arbitrage Triangular arbitrage is the process that ensures that all exchange rates are mutually consistent. If, for exam ple, one U.S. dollar exchanges for one Canadian dollar, and one Canadian dollar exchanges for one British pound, th en the U.S. dollar-pound exchange rate should be one pound for one dollar. If it differs, then there is an opportunity for profit making. To see why this is so, suppose that you could purchase two U.S. dollars with one British pound. By first buying C$1 with U.S.$1, then purchasing £ 1 with C$1, and finally buying U.S.$2 with £ 1 , you could double your money im mediately. Clearly this opportunity will not last for long since it involves making large profits with certainty. The process of triangular arbitrage is exactly that of finding and exploiting profitable opportunities in such exchange rate inconsistencies. As a result of triangular arbitrage, su ch in co n sisten cies will be elim inated rapidly. Cross rates, however, will only be roughly con sistent given the bid-ask spread associated with trans action costs. In the past, the possibility of making profits from triangular arbitrage was greater as a result of the prac MARCH 1984 exercise price. This com bination will show a profit at exercise price A if the spot price goes either above b or below a. It is known as a "straddle.” The strad dle is of special interest because it makes clear the role of options as a hedge against risk. The price of a straddle can be regarded as the market valuation of the variability of the exchange rate. That is, the buyer of the straddle will show a profit if the spot price moves from some central value (the exercise price) by more than plus or minus some known percent age. The seller of the straddle accepts that risk for a lump sum. More com plicated “multiple strategies” are also possible.2 2See Giddy (1983). tice of expressing exchange rates in American terms in the United States and in European term s elsewhere. The adoption of standard practice has reduced the likelihood of inconsistencies.7 Also, in recent years, such opportunities for profit making have been greatly reduced by high-speed, com puterized information systems and the increased sophistication of the banks operating in the market. Arbitrage of a slightly different kind results from price differences in different locations. This is “space” arbitrage. For example, if sterling were cheaper in Lon don than in New York, it would be profitable to buy in London and sell in New York. Similarly, if prices in the interbank market differed from those at the IMM, it would be profitable to arbitrage between them. As a result of this activity, prices in different locations will be brought broadly into line. Interest Arbitrage Interest arbitrage is slightly different in nature from triangular or space arbitrage; however, the basic motive of finding and exploiting profitable opportunities still applies. There is no reason why interest rates denom i nated in different currencies should be equal. Interest rates are the cost of borrowing or the return to lending for a specific period of time. The relative price (ex change rate) of money may change over time so that the comparison of, say, a U.S. and a British interest rate requires some allowance for expected exchange rate changes. Thus, it will be not at all unusual to find 7AII except U.K. and Irish exchange rates are expressed in American terms. Futures and options contracts are expressed in European terms. 13 FEDERAL RESERVE BANK OF ST. LOUIS interest rates denom inated in dollars and interest rates denom inated in, say, pounds being somewhat differ ent. However, real returns on assets of similar quality should be the same if the exchange rate risk is covered or hedged in the forward market. Were this not true, it would be possible to borrow in one currency and lend in another at a profit with no exchange risk. Suppose we lend one dollar for a year in the United States at an interest rate of rus. The am ount accum u lated at the end of the year per dollar lent will be 1 + rus (capital plus interest). If, instead of making dollar loans, we converted them into pounds and lent them in the United Kingdom at the rate ruk, the amount of pounds we would have for each original dollar at the end of the year would be S(1 + ruk), w here S is the spot exchange rate (in pounds per dollar) at the beginning of the period. At the outset, it is not known if 1 + rus dollars is going to be worth more than S(l + ruk) pounds in a year’s time because the spot exchange rate in a year’s time is unknown. This uncertainty can be avoided by selling the pounds forward into dollars. Then the rela tive value of the two loans would no longer depend on what subsequently happens to the spot exchange rate. By doing this, we end up with | (l + ruk) dollars per original dollar invested. This is known as the “cov ered,” or hedged, return on pounds. Since the covered return in our example is denom i nated in dollars, it can reasonably be com pared with the U.S. interest rate. If these returns are very different, investors will move funds where the return is highest on a covered basis. This process is interest arbitrage. It is assumed that the assets involved are equally safe and, because the returns are covered, all exchange risk is avoided. Of course, if funds do move in large volume between assets or between financial centers, then in terest rates and the exchange rates (spot and forward) will change in predictable ways. Funds will continue to flow between countries until there is no extra profit to be made from interest arbitrage. This will occu r when the returns on both dollar- and sterling-denominated assets are equal, that is, when (1) MARCH 1984 Speculation Arbitrage in the foreign exchange markets involves little or no risk since transactions can be com pleted rapidly. An alternative source of profit is available from outguessing other market participants as to what fu ture exchange rates will be. This is called speculation. Although any foreign exchange transaction that is not entirely hedged forward has a speculative element, only deliberate speculation for profit is discussed here. Until recently, the main foreign exchange specula tors were the foreign exchange departm ents of banks, with a lesser role being played by portfolio managers of other financial institutions and international corpora tions. The IMM, however, has made it m uch easier for individuals and smellier businesses to speculate. A high proportion of IMM transactions appears to be specula tive in the sense that only about 5 percent of contracts lead to ultimate delivery of foreign exchange. This means that most of the activity involves the buying and selling of a contract at different times and possibly different prices prior to maturity. It is possible, how ever, that buying and selling of contracts before m atu rity would arise out of a strategy to reduce risk. So it is not possible to say that all such activity is speculative. Speculation is important for the efficient working of foreign exchange markets. It is a form of arbitrage that occurs across time rather than across space or b e tween markets at the same time. Just as arbitrage in creases the efficiency of markets by keeping prices consistent, so speculation increases the efficiency of forward, futures and options markets by keeping those markets liquid. Those who wish to avoid foreign ex change risk may thereby do so in a well-developed market. Without speculators, risk avoidance in foreign exchange markets would be more difficult and, in many cases, impossible.9 Risk Reduction Speculation clearly involves a shifting of risk from one party to another. For example, if a bank buys for- (l + rus) = p-(l + ruk). This result is known as covered interest parity. It holds more or less exactly, subject only to a margin due to transaction costs, so long as the appropriate dollar and sterling interest rates are com pared.8 8Since there are many different interest rates, it obviously cannot hold for all of them. Where (1) does hold is if the interest rates chosen are eurocurrency deposit rates of the same duration. In other words, if for 14 FRASER Digitized for rus we take, say, the three-month eurodollar deposit rate in Paris and for ruk we take the three-month eurosterling deposit rate in Paris, then (1) will hold just about exactly. Indeed, if we took the interest rate and exchange rate quotes all from the same bank, it would be remarkable if (1 ) did not hold. Otherwise the bank would be offering to pay you to borrow from it and lend straight back! That is, the price of borrowing would be less than the covered return on lending. A margin between borrowing and lending rates, of course, will make this even less likely so that in reality you would lose. 9This is not to say that all speculative activity is necessarily beneficial. FEDERAL RESERVE BANK OF ST. LOUIS MARCH 1984 Covered Interest Parity: An Example The following interest rate and exchange rate quotations are taken from the London Financial Times of Septem ber 8, 1983 (table 1). Closing Exchange Rate: dollars per pound SP ot 1.4910-1.4920 3-Month Forward .17—.22 discount Eurosterling Eurodollar 9 13/ ie 10*/4 Interest Rates: 3-Month Offer Rate The interest rate on the three-m onth eurodollar deposit is a little higher (.7 percent) than that on an eurosterling deposit. If the exchange rate remains unchanged, it would be better to hold dollars; if the exchange rate falls, the eurosterling deposit would be preferable. Suppose you decide to cover the ex change risk by selling the dollars forward into pounds. Let us com pare the return to holding a sterling deposit with the return to holding a dollar deposit sold forward into sterling (assuming that you start with sterling). Two important points need to be clarified about the above data. First, the interest rates are annual ized so they are not what would actually be earned over a three-m onth period. For example, the threemonth rate equivalent to an annual rate of 10 lA percent is 2.47 percent. Second, the forward exchange rates need some explanation. The dollar is at a discount against ster ling. This m eans the forward dollar buys less ster ling. So we have to add the discount onto the spot price to get the forward price (because the price is the num ber of dollars per pound, not the reverse). Notice also that the discount is measured in frac tions of a cent, not fractions of a dollar! So the ward foreign exchange from a custom er, it increases its exposure to risk while the custom er reduces his. How ever, there is not a fixed am ount of risk that has to be "shared out.” Some strategies may involve a net reduc tion of risk all around. As a general rule, financial institutions (or other firms), operating in a variety of currencies, will tiy to bid-ask spread on the forw ard rate w ould be 1.4927-1.4942. Now let us see if we would do better to invest in a three-m onth eurosterling deposit or a three-month eurodollar deposit where the dollars to be received were sold forward into sterling. The return per £100 invested in eurosterling is £2.369 (annual interest rate of 913/ie), whereas the return on a covered euro dollar deposit is £2.251 = (100 X 1 4910 1.0247) 1.4942 100 . Thus, we could not make a profit out of covered interest arbitrage. Despite the fact that dollar in terest rates are higher, the discount on forward dol lars in the forward market m eans they buy fewer forward pounds. As a result, there is no benefit to the operation. Transaction costs for m ost indi viduals would be even greater than those above as they would face a larger bid-ask spread than that quoted on the interbank market. Consequently, there is no benefit for the typical investor from making a covered or hedged eurocur rency deposit. The return will be at least as high on a deposit in the currency in which you start and wish to end up. That is, if you have dollars and wish to end up with dollars, make a eurodollar deposit. If you have sterling and wish to end up with sterling, make a eurosterling deposit. If you have sterling and wish to end up in dollars, there is likely to be little or no difference between holding a eurosterling de posit sold forward into dollars or buying dollars spot and holding a eurodollar deposit. Of course, if you hold an “uncovered” deposit and exchange rates subsequently change, the result will be veiy different. minimize the risk of losses due to unexpected ex change rate changes. One simple way to do this is to ensure that assets and liabilities denom inated in each operating currency are equal. This is known as “m atch ing." For example, a bank that sells sterling forward to a custom er may simultaneously buy sterling forward. In this event, the bank is exposed to zero exchange rate risk. 15 FEDERAL RESERVE BANK OF ST. LOUIS MARCH 1984 Why Is the D ollar the “Money” of Foreign Exchange Markets? One interesting aspect of the organization of the foreign exchange markets is that the "m oney" used in these markets is generally the U.S. dollar. This is generally true for spot markets and universally true for forw ard m arkets. “C ross-m ark ets’’ betw een many currencies are very thin, and future cross markets are virtually nonexistent. For example, the bulk of foreign exchange trading between £ s and cruzeiro will involve dollar-£ and dollar-cruzeiro transactions instead of direct £-cru zeiro trading. The only exception to this is the transactions involv ing the m ajor Organization for Econom ic Coopera tion and Development (OECD) currencies, especial ly within Europe. Of the $702.5 billion turnover in foreign exchange reported by U.S. banks in April 1983, only $1.5 billion did not involve U.S. dollars. eiy country is in payments balance vis a vis the rest of the world, it will not necessarily be in bilateral balance with each other country. Because som e cur rency has to be used to cover this residual finance, it is natural to choose the currency that has the lowest transaction costs. Chrystal shows there are eco nomic reasons why cross-m arkets between many currencies do not exist.2 It typically will be easier and cheaper to set up a deal in two steps via the dollar than in a single step (cruzeiro-dollar, dollardrachm a rather than cruzeiro-drachma). This is be cause these cross-m arkets, if they existed, would be fairly thin and hence relatively costly for such trans actions. The two markets with the dollar, on the other hand, are well developed. There are two explanations for this special role of the dollar in foreign exchange markets. Both rely upon the fact that transaction costs are likely to be lower if the dollar is used as a medium. Krugman shows that the clearing of foreign exchange markets requires some "interm ediary” currency.1 Even if ev- These analyses refer to the role of the dollar in the interbank market. In the development of the trading places such as the IMM in Chicago and LIFFE in London to date, it is also true that all currency futures are traded against the dollar. 'See Krugman (1980). 2See Chrystal (1982). Banks often use “swaps” to close gaps in the m atu rity structure of their assets and liabilities in a cur rency. This involves the sim ultaneous purchase and sale of a currency for different maturity dates. In April 1983, 33 percent of U.S. banks’ foreign exchange turn over involved swaps as com pared with 63 percent spot c o n tr a c ts an d o n ly 4 p e r c e n t o u trig h t forw ard contracts.10 Suppose a bank has sold DM to a custom er three m onths forward and bought the same am ount of DM from a different custom er six m onths forward. There are two ways in w hich the bank could achieve zero foreign exchange risk exposure. It could either under take two separate offsetting forward transactions, or it could set up a single swap with another bank that has the opposite m ism atch of dollar-DM flows whereby it receives DM in exchange for dollars in three m onths and receives back dollars in exchange for DM in six ,0See Federal Reserve Bank of New York (1983). Digitized for16 FRASER months. Once the swap is set up, the bank’s net profits are protected against subsequent changes in spot ex change rates during the next six months. Within the limits im posed by the nature of the con tracts, a similar effect can be achieved by an appropri ate portfolio of futures contracts on the IMM. Thus, a bank would buy and sell futures contracts so as to match closely its forward com m itm ents to custom ers. In reality, banks will use a com bination of m ethods to reduce foreign exchange risk. Markets that permit banks, firms and individuals to hedge foreign exchange risk are essential in tim es of fluctuating exchange rates. This is especially impor tant for banks if they are to be able to provide efficient foreign exchange services for their custom ers. In the absence of markets that permit foreign exchange risk hedging, the cost and uncertainty of international transactions would be greatly increased, and interna tional specialization and trade would be greatly re duced. FEDERAL RESERVE BANK OF ST. LOUIS CONCLUSION The foreign exchange markets are com plex and, for the outsider, hard to com prehend. The primary func tion of these markets is straightforward. It is to facilitate international transactions related to trade, travel or in v estm en t. Foreign ex ch a n g e m arkets can now accom m odate a large range of current and forward transactions. Given the variability of exchange rates, it is important for banks and firms operating in foreign currencies to Glossary Am erican option — an option that can be exercised any time up to maturity. Am erican term s — an exchange rate expressed as number of currency units per dollar. arbitrage — the simultaneous purchase and sale of currency in separate markets for a profit arising from a price discrepancy between the markets. bid-ask spread — the difference between the buying (bid) and selling (ask) price. covered interest arbitrage — buying a country's currency spot, investing for a period, and selling the proceeds forward in order to make a net profit due to the higher interest rate in that country. This act involves 'hedging” because it guarantees a covered return without risk. The opportunities to profit in this way seldom arise because covered interest differentials are normally close to zero. covered interest pa rity — the gap between interest rates in foreign and domestic currencies will be matched by the forward exchange rate differential, such that the "covered” interest rate differential will be close to zero. eurodollar deposits — bank deposits, generally bearing in terest and made for a specific time period, that are denominated in dollars but are in banks outside the United States. Similarly, eurosterling deposits would be denominated in sterling but outside the United Kingdom. European option — an option that can be exercised only on a specified date. European term s — an exchange rate expressed as number of dollars per currency unit. floa ting exchange rate — an exchange rate that is allowed to adjust freely to the supply of and demand for foreign exchange. MARCH 1984 be able to reduce exchange rate risk whenever possi ble. Some risk reduction is achieved by interbank swaps, but some is also taken up by speculation. Arbi trage and speculation both increase the efficiency of spot and forward foreign exchange markets and have enabled foreign exchange markets to achieve a high level of efficiency. Without the successful operation of these markets, the obstacles to international trade and investment would be substantial and the world would be a poorer place. foreign exchange speculation — the act of taking a net posi tion in a foreign currency with the intention of making a profit from exchange rate changes. forw ard exchange rate — the price of foreign currency for delivery at a future date agreed to by a contract today. futures m arke t— a market in which contracts are traded to buy or sell a standard amount of currency in the future at a particular price. hedging — or covering exchange risk, means that foreign cur rency is sold forward into local currency so that its value is not affected by subsequent exchange rate changes. Say an exporter knows he will be paid E10,000 in two months. He can wait until he gets the money and convert it into dollars at whatever the spot rate turns out to be. This outcome is uncertain as the spot rate may change. Alternatively, he can sell £10,000 two months forward at today’s two-month forward price. Suppose this is $1.5 per £. In two months, he will receive £ 1 0 ,0 0 0 , fulfill his forward contract and receive $15,000. This export contract has been hedged or covered in the forward market. m atching — equating assets and liabilities denominated in each currency so that losses due to foreign exchange rate changes are minimized. options market — a market in which contracts are traded that gives a purchaser the right but no obligation to buy (call) or to sell (put) a currency in the future at a given price. spot exchange rate — the price paid to exchange currencies for immediate delivery (two business days in the interbank market, or over the counter in the retail and travelers check market). swap — the simultaneous purchase and sale of a currency for different maturity dates that closes the gaps in the maturity struc ture of assets and liabilities in a currency. 17 FEDERAL RESERVE BANK OF ST. LOUIS REFERENCES Belongia, Michael T. “ Commodity Options: A New Risk Manage ment Tool for Agricultural Markets,” this Review (June/July 1983), pp. 5-15. Black, Fisher, and Myron Scholes. “The Pricing of Options and Corporate Liabilities,” Journal of Political Economy (May/June 1973), pp. 637-54. MARCH 1984 Federal Reserve Bank of New York. "Summary of Results of U.S. Foreign Exchange Market Turnover Survey Conducted in April 1983” (September 8 , 1983). Garman, Mark B., and Steven W. Kohlhagen. “ Foreign Currency Option Values,” Journal of International Money and Finance (De cember 1983), pp. 231-37. Giddy, Ian H. “ Foreign Exchange Options,” Journal of Futures Mar kets (Summer 1983), pp. 143-66. Chrystal, K. Alec. “ On the Theory of International Money" (paper presented to U.K. International Economics Study Group Confer ence, September 1982, Sussex, England). Forthcoming in J. Black and G. S. Dorrance, eds., Problems of International Finance (Lon don: Macmillan, 1984). Krugman, Paul. “ Vehicle Currencies and the Structure of Interna tional Exchange,” Journal of Money, Credit and Banking (August 1980), pp. 513-26. Dufey, Gunter, and Ian H. Giddy. (Prentice-Hall, 1978). McKinnon, Ronald I. Money in International Exchange: The Con vertible Currency System (Oxford University Press, 1979). 18 The International Money Market Kubarych, Roger M. Foreign Exchange Markets in the United States. (Federal Reserve Bank of New York, 1983). The Money-GNP Link: Assessing Alternative Transaction Measures R. II. Hafer F, I J MPIRICAL research strongly suggests that the growth of M l — a m easure of transaction balances — is more closely related to GNP growth than are the broad er monetary m easures.1 Yet, at its October 1982 m eet ing, the Federal Open Market Com m ittee (FOMC), which establishes m onetary policy for the Federal Re serve System, decided to attach relatively less im por tance to observed movements in M l in formulating policy. Instead, it placed increased significance on the behavior of broader, nontransaction-oriented m ea sures, such as M2 and M3. This decision cam e about for two reasons: First, some members of the FOMC believed that the behavior of M l had been and would continue to be distorted by the shifting of funds among new types of monetary instrum ents that resulted from financial deregulation. Second, velocity developments in 1982, which con- R. W. Hafer is a senior economist at the Federal Reserve Bank of St. Louis. Jane Mack provided research assistance. NOTE: The empirical work presented here is based on the unrevised M1 data. 'Transaction balances refer to those balances that are available for immediate spending, such as demand deposits. Empirical evidence comparing narrow (that is, transactions-oriented) and broad mone tary definitions is presented in Carlson and Hein (1980), Hafer (1981) and Batten and Thornton (1983). An alternative view, advocating the use of broader measures of money or debt, is expressed in Friedman (1981, 1982) and Morris (1982). The use of broad monetary aggre gates or debt measures in the conduct of policy is examined critically by Lawler (1981) and Davidson and Hafer (1983). tinued into 1983, raised doubts about the stability of the relationship betw een M l and nom inal incom e (GNP).2 Much of the uncertainty about the usefulness of Ml as a transactions measure arises because it includes currency and dem and deposits — traditionally re garded as “m oney" — plus interest-bearing checkable deposits, su ch as negotiable order of withdrawal (NOW) acco u n ts, autom atic transfer system (ATS) accounts, and credit union share drafts.3 Some have argued, however, that these latter deposits, “while 2For a general discussion, see “ Monetary Policy Report to Congress" (1983), especially pages 132-35. See also Solomon (1983). 3The concept of money as that asset used expressly for transaction purposes has a long history in monetary economics. Lauchlin Currie (1935), for example, makes clear the distinction between the concept of money, defined as currency plus demand deposits, and broader measures that incorporate savings-type deposits: There is, however, an important distinction between means of payment and what may be regarded by individuals as equivalent to means of payment. Time deposits, in this respect, do not differ essentially from holdings of government securities, call loans, or, indeed, any property possessing good marketability which by sale can be converted into means of payment. It is no more correct to say that one can spend a time deposit than a government security. Both must first be exchanged for cash or deposits subject to check before they can be spent. This distinction between money and “ near money” also is noted by Martin Bronfenbrenner (1945): “ No monetary commodity can have any use other than cash balance uses,” where “cash balance uses” refers to those items "held expressly (consciously) for the purpose of future direct exchange for other goods." Recent attempts to deter mine empirically the transaction uses of current monetary measures are represented by Barnett (1980) and Spindt (1983). 19 FEDERAL RESERVE BANK OF ST. LOUIS serving the transaction needs of holders, have many of the characteristics of savings accounts.”4 Thus, the nature of M l as a m easure of transaction balances has com e under question. In this article, we investigate the relationship be tween two alternative m easures of transaction bal ances and GNP. One m easure is the current M l aggre gate. Because of the difficulty in reliably determining what proportion of other checkable deposits is held as savings instead of transaction balances, they are ex cluded from our alternative measure, called adjusted M l. Thus, adjusted M l is simply M l m inus other checkable deposits, that is, M l less its interest-bearing com ponents.5 By exam ining the evidence obtained from using these polar definitions of transaction bal an ces, som e light m ay b e sh ed on the question w hether recent movements in M l, especially those in 1982 and early 1983, accurately reflect the actual m one tary stimulus to the economy. A MODEL OF THE DEMAND FOR TRANSACTION BALANCES Useful theoretical models have been developed to analyze the effect of the interest payment prohibition on demand deposits. These models provide a founda tion from which to analyze the im pact of the introduc tion of interest-bearing checkable deposits. From these models, we can predict some of the effects of the repeal of interest prohibition on transaction deposits which, in essence, occurred w hen NOW accounts becam e avaifabfe nationwide.6 In a general m odel developed by Santomero, the household is assum ed to allocate its wealth among various assets in order to maximize the return from its consum ption activities.7 The household’s initial en dowment of wealth may be held as currency, demand deposits, savings deposits or com m odity inventories. The savings deposit pays a positive, explicit interest rate, r*. Demand deposits yield some implicit interest, rd, 0 i'1 ? rsK Because savings cannot be traded MARCH 1984 directly for commodities, the model also posits trans action costs for currency and dem and deposits that are strictly lower than those for savings.9 Thus, savings are viewed as being a tem poraiy store of funds. More over, the theoretical model predicts that “the savings asset will only be used as a temporary store of working balances for intra period use if the interest rate differ ential [rs — rd] is sufficient to com pensate the house hold for the extra cost of going to the bank. If this condition is not satisfied, the savings asset will not be used and demand deposits will become the temporary store o f funds."'0 Thus, as the rate paid on demand deposits (implicit or explicit) approaches the rate paid on savings deposits, households will increase their average holdings of demand deposits relative to sav ings deposits.11 In this event, funds stored in savings deposits will be converted into dem and deposits, which will now possess the dual characteristics of a tran sactio n s m edium and a “tem porary store of funds.”12 9Let a DG and aCG represent the transactions cost of obtaining com modities (G) by means of using demand deposits (D) and currency (C), respectively. If aSD and asc represent the cost of converting savings deposits into demand deposits or currency, respectively, then the transactions cost of using savings deposits to acquire com modities ( a S c . ) i s either <x s g = a s D ^ <* d g o r s g = o c s c "*■ (* c g ■The household’s cost of transferring funds from savings to demand de posits (ignoring currency) and the relative return from holding sav ings deposits are crucially related. As Santomero notes, “the return from the short-term interest bearing asset [rs] must be sufficient to compensate the household for the additional cost of withdrawing funds from S [savings] and not D [demand deposits].” See Santo mero (1974). 10lbid., p. 97, italics added. "S e e also the analyses of Barro and Santomero (1972) and M. Klein (1974). 12Formally, the solutions for average demand deposit holdings (D) and average savings deposit holdings (S) are given as q _ / Y(gs ~ “ dc) \f o 1 \/-r 2(r® - r°) _ / V jj a DC~ _ \ J 2(r« - r°) / Y (1 ~ h) g De 2(r* - r°) /Y (a s — « dc) r^) where Y = rate of consumption of lump sum income payment X across intervals T(Y = X/T), 4“ Monetary Policy Report to Congress," p. 134. 5"Adjusted M1” is not identical to the pre-1980 M1 definition. Unlike the previous measure, adjusted M 1 includes travelers checks and excludes deposits due to foreign commercial banks and official in stitutions. For a comparison between old and current M1, see Hafer (1980). 6lt should be noted that the analysis concerns household behavior only: businesses currently are not allowed to hold NOW accounts. 7Santomero (1974). 8These are returns on the marginal dollar held in each deposit group. 20 h = proportion of transactions using currency, (1 - h) = proportion of transactions using demand deposits, a DC = cost of converting demand deposits into currency, and r9 = return on commodity inventories (i^fO). Holding transactions costs constant, as the rate on demand de posits (O approaches that on savings deposits (r*), the first term in the demand deposit equation becomes indefinitely large as does the expression under the radical sign in the savings equation. JThe consequence, clearly, is for average demand deposit holdings (D) to increase relative to average savings deposit holdings (S). MARCH 1984 FEDERAL RESERVE BANK OF ST. LOUIS The crucial element in this analysis is the difference between the rates on dem and deposits and savings. If the demand deposit rate is both "com petitive,” as sug gested by Klein, and Barro and Santomero, and close to the rate paid on savings accounts, removing the in terest prohibition on dem and deposits (assuming that r11 cannot exceed rs) should not appreciably affect the household's allocation of funds. Evidence by Startz, however, indicates that the implicit rate paid on de mand deposits (essentially through rem ission of ser vice charges) is only about one-half of the alternative savings rate.13 Consequently, allowing explicit interest payments on checkable deposits that approach the rate paid on savings deposits, according to the model, would attract funds from savings deposits that pay a similar rate of return and are relatively less liquid. THE IMPACT OF INTEREST PAYMENTS ON CHECKABLE DEPOSITS: SOME EVIDENCE NOW accounts were made available to households on a nationwide basis beginning in January 1981. Be fore then, they were available only in the New England states.14 Frodin and Startz examined the effects of the early NOW experience on money dem and estim ates for the New England states relative to the rest of the United States.15 Their results indicate that, after 1975, the in troduction of NOW accounts increased personal trans action balances by about 37 percent; in terms of total money demand, the result was an increase of about 9 percent. In another recent study, Badecki and Wenninger examine m oney dem and functions for the consum er and n on fin an cial b u sin ess se cto rs to determ ine, among other things, the effect of NOW accounts on the two groups during 1981 and 1982. Based on a series of post-sam ple forecasts, they conclude that “the in crease in NOW accounts during that year [1981] could not have represented just a substitution of demand 13Startz (1979) estimates two series on the implicit interest on de mand deposits. In 1975, the rate was calculated to be 2.47 percent and 2.80 percent. These implicit returns paid on demand deposits are compared with the passbook savings rate at commercial banks of 4.87 percent and the passbook rate at savings and loans of 5.24 percent. 14NOW accounts were offered first in June 1972 by the Consumer Savings Bank of Worcester, Massachusetts. Initially, NOWs were limited to mutual savings banks. In January 1974, New England commercial banks were authorized to offer NOW accounts. See Klein (1978). 15Frodin and Startz (1982). Table 1 Growth of M1, Adjusted M1 and Other Checkable Deposits: 1/1982 Through 11/1983______________________________ Quarter M1 1/1982 11 .0 % II 3.3 Adjusted M1 3.2% - 0 .5 O ther Checkable Deposits 54.1% 21.3 III 6.3 2.4 23.4 IV 13.7 7.8 38.1 1/1983 14.9 5.4 55.5 II 12.7 6.7 34.3 deposits for NOW account deposits, leaving the de mand for total money balances unchanged.”16 More over, their evidence indicates that the rapid growth of M l during 1982 was due to a continuing flow of funds away from non-M l sources into NOW accounts as new accounts were opened.17 Specifically, they claim that about $8 billion of the new NOW accounts originated outside M l. The results of other studies by Johannes, and Joh an nes and Basche, on forecasting the M l money multi plier imply that there was a portfolio shift between time deposits and transaction accounts during the early part of 1981.18 They found that a level shift adjust ment was necessary for five of the seven ratios used in calculating the multiplier. Their results are roughly consistent with the Board of Governors’ staff projec tions that, during eariy 1981,20 percent to 25 percent of the funds shifting to NOW accounts were from nondemand-deposit sources. During 1982, the growth of M l far exceeded that of adjusted M l. The figures in table 1 indicate that M l averaged about an 8.5 percent growth rate in that year. Adjusted M l, on the other hand, grew an average rate of only 3.2 percent. In early 1983, this divergence was 16Radecki and Wenninger (1983), p. 5, italics in original. It should be noted that the results of Radecki and Wenninger are based on data that has been questioned. Consequently, some caution is advised in interpreting their findings. 17Data presented by Radecki and Wenninger suggest that the number of new NOW accounts opened between November 1981 and November 1982 totaled 3.32 million, an increase of 22 percent. 18Johannes (1981), Johannes and Rasche (1981). An opposite con clusion is reached by Tatom (1982). 21 FEDERAL RESERVE BANK OF ST. LOUIS MARCH 1984 even greater: M l increased at an average annual rate of 13.8 percent and adjusted M l at a 5.9 percent average rate. MONEY AND ECONOMIC ACTIVITY: WHICH M l? A recent study by Judd and McElhattan helps ex plain these divergent growth rates. In their study, Judd and M cElhattan argue that the M l series overstated the “effective” money growth rates during 1982-83. This overstatement arises from an interest-rate-induced in crease in the quantity of m oney balances dem anded by the public. That is, the sharp drop in market rates during late 1982 precipitated an increase in the quanti ty of m oney balances dem anded to w hich “the Federal Reserve responded by allowing money to grow faster than originally targeted."19 The nationw ide in trod u ction of NOW accou n ts attracted funds from both dem and deposits and nonM1 sources. During 1981, the growth of dem and de posits fell dramatically as households shifted some of these funds into NOW accounts. For example, adjusted M l decreased at rates of 21.4 percent, 4.7 percent and 2.3 percent, respectively, during the first three quarters of 1981. This drop signified that the public was less willing to hold transaction balances that did not pay explicit interest at eveiy level of real incom e and inter est rates.20 O ther things unchanged , ad ju sted -M l velocity should have shown a marked upward level shift during this period. The data in table 1 indicate that this increase in money growth exists largely in the interest-bearing com ponent of M l, not in the adjusted M l series. The Judd-M cElhattan analysis, com bined with the data in table 1, suggests that demand deposits and currency have reacted differently to changes in market interest rates than did the interest-bearing com ponent of M l. Indeed, other checkable deposits appear to be more in terest-e la stic than the n on -in terest-bearin g bal ances. Moreover, Judd and M cElhattan find that an M l series "adjusted” for the increased quantity of money demanded due to the sharp interest rate decline in late 1982 explains econom ic activity behavior better than M l during the 1982-83 period. Thus, the implication is that the increased quantity of m oney demanded was not used to fund transactions but, rather, was held as a store of funds. The discussion thus far indicates that the increase in M l in 1981 is partially attributable to the shifting of funds from time deposits to transaction balances. In 1982, the divergent behavior of M l and adjusted M l also suggests that the growth in the interest-bearing com ponents of M l was, in part, for non-transaction purposes. This result is "predicted” by the theoretical model discussed above. The interesting policy ques tion that emerges from these results is: Does M l have the same influence on econom ic activity as it did be fore these new interest-bearing deposits were made available? Moreover, do transaction balances that do not ca n y explicit interest payments display the same relationship to total spending before and after the change in the financial environment? The rem ainder of this article attem pts to answ er these questions. ' 9Judd and McElhattan (1983), p. 46. Digitized for22 FRASER Chart 1 plots the levels of adjusted M l and M l veloci ties for the period 1/1960 to 11/1983. There is no discern ible difference between the two series before the mid1970s, because other checkable deposits were a minor part of the public’s money holdings. The introduction of ATS accounts, New England NOWs and credit union share draft accounts produced the divergent behavior of the two series since the mid-1970s. The biggest de- 20This assertion is borne out by estimates of a conventional money demand equation for adjusted M1. For example, using the period 1/1960 through 11/1983, the adjusted-M1 equation yields the result (absolute value of t-statistics in parentheses): In (M/P), = - 0.247 - 0.011 D1 + 0.047 In y, (2.64) (2.63) (3.24) - 0.032 In r, + 0.970 InfM /P )^, (6.81) (63.35) R2 = 0.987 SE = 0.0096 Dh = -0 .4 6 where Pis the GNP deflator (1972 = 100),yisrealG N P($1972),ris the three-month Treasury bill rate and D1 is a (0,1) dummy term that equals 1.0 for the period 11/1974 to 11/1983, zero elsewhere. These results indicate an abnormally slow adjustment speed (3 percent per quarter) and long-term income and interest elasticities that are quite large relative to standard results. Accounting for a level shift in the function in 1981, however, restores the underlying economic relationship between real money balances and its determinants. Introducing another intercept shift term (D2), defined as 1.0 for the period 1/1981 to 11/1983 and zero elsewhere, the results are In (M/P), = - 0.373 - 0.023 D1 - 0.042 D2 (4.56) (5.46) (5.67) + 0.081 In y, - 0.028 In r, + 0.825 ln(M/P)t _, (5.97) (6.93) (29.01) R2 = 0.991 SE = 0.0081 Dh = 0.80 These results are similar to numerous other studies in terms of the estimated speed of adjustment (18 percent per quarter) and the income and interest elasticities. The significance of the D2 coeffi cient supports the contention of a downward level shift in the func tion. MARCH 1984 FEDERAL RESERVE BANK OF ST. LOUIS C h a rt 1 Velocity of M l and Adjusted Ml 9.0 9.0 I 8.5 8.5 ■ 8.0 8.0 7.5 7.5 I IS 7.0t 6.5 AdjustFed Mt 6.0 / ■ 5.5 7.0 6.5 I SJ 5.0 4.5 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 S h a d e d a re a s re p re s e n t p e r io d s o f bus in e s s recessions. viation in the respective velocity m easures occurs in 1981 when NOW accounts were made available nation wide. For example, in 1/1981, M l velocity increased at a 13.8 percent rate, while adjusted M l velocity increased at an unprecedented 40.0 percent rate. For the year as a whole, M l velocity increased at an average rate of 5.3 percent, within two standard deviations of its 3.2 per cent average growth since 1960. Adjusted M l velocity, in contrast, grew at an average rate of 17.4 percent. Again in 1982, the growth of velocity m easured by adjusted M l diverged sharply from that of M l. For example, during 1982, adjusted-M l velocity declined at an average 0.72 percent rate; M l velocity declined, on average, at a 5.62 percent rate. Several researchers have attem pted to explain this sharp drop in M l velocity. Tatom, for example, argues that some of the drop in M l velocity growth during the last recessio n can be accounted for by the cyclical response of velocity to the recession.21 As noted earlier, Judd and McElhattan argue that the M l m easure overstates the growth of transaction balances in 1982 that influences econom ic activity. Using their adjusted-M l series, they find that “[slimulations of velocity, real GNP and inflation . . . were more accurate than those using m easured M l.’’22 Has the behavior of the interest-bearing com ponent of M l during the past one and one-half years led to a substantial change in its empirical relationship with GNP growth? Once we have captured the expected velocity shift in adjusted M l due to financial innova tions in 1981, has there been any deterioration in its relationship with GNP growth? To determine which measure of money, M l or ad justed M l, better explains GNP growth, both were used 2'Tatom (1983). 22Judd and McElhattan, p. 46. 23 FEDERAL RESERVE BANK OF ST. LOUIS in estimating a variant of the reduced-form St. Louis GNP equation.23 First, in-sam ple estim ates using M l and adjusted M l are presented for the period 1/1960 to IV/1979 and are used as a basis for com parison.24 The sample period then is updated through II/1983, and the equation is re-estimated. MARCH 1984 Table 2 GNP Equation Estimates: 1/1960—IV/1979 Coefficient Because the constant term in the reduced-form GNP equation represents the average growth rate of velocity, a failure to capture the intercept shift caused by reac tion to the introduction of nationwide NOW accounts would lead to biased coefficient estim ates 25 Conse quently, a (0,1) dummy variable is used to capture the short-lived aberration in adjusted-M l velocity growth during 1981. This term (D1981) equals 1.0 for 1/1981, 11/1981 and III/1981, and zero elsewhere. In-Sample Estimates: 1960—1979 To gauge the presum ed deterioration in the moneyGNP link, the two alternative money m easures are used initially to explain econom ic activity during a previous, relatively untroubled period. The results of estimating the reduced-form GNP equation using both monetary definitions for the period I/1960-IV/1979 are presented in table 2.26 Not surprisingly, the em pirical estim ates are quite similar. In terms of overall fit, the coefficient of deter mination (R2) of the M l equation is slightly greater than that for adjusted M l, albeit by less than 3 percent. This slight improvement also is reflected in the relative stan dard errors of the equation (SE). Moreover, as the Dur- 23The model estimated here is presented in Tatom (1981). The basic model is expressed in the form M N GNP = a 0 + Pi 2 mi M,_i + p2 ^ ej E t-j i= 0 j=0 Q + p3 2 pek Pf_ 1 _ k + St + k=0 where M is the growth of money, E is the growth of high-employment federal expenditures, Pe is the change in the relative price of energy and S is a variable entered to capture the effect of major strikes on GNP. 24This specific sample period is used because monetary policy proce dures changed after this date, monetary policy in 1980 was influ enced by the Special Credit Controls program, NOW accounts were legalized nationwide in 1981 and, finally, financial deregulation accelerated after this period. 25See Maddala (1977), pp. 155-57. 26The monetary and fiscal actions measures are estimated using a fourth-degree Almon polynomial with both endpoints constrained. The relative energy price variable is estimated using a third-degree polynomial without endpoint constraints. 24 Constant M1 (2.46) (2.40) (5.57) (3.14) (1.69) (0.42) (6.26) (2.15) (2.55) (5.45) (3.12) (1.97) - 0 .0 1 2 (0 . 11 ) 1.082 (6.15) 0.076 (1.89) 0.019 (0.56) -0 .0 3 6 (0.97) 0.074 (1.81) 0.021 (0.59) -0 .0 2 9 (0.79) -0.03 1 (0.93) 0.014 (0.35) 0.043 (0.41) -0.02 1 (0.64) 0.022 (0.56) 0.066 (0.63) 0.004 0.001 -0 .0 1 4 -0 .0 2 7 -0 .0 2 3 pe -5 pe -6 £pe_i -0 .0 1 5 -0 .0 0 3 - 0.011 -0 .0 2 3 - 0 .0 2 0 0.013 0.095 0.035 S -0.64 1 (3.51) mo m _, m_2 m_3 m_4 2 m( 6o e_i e ^2 e- 3 e 4 Sei pe 0 p e -, pe z pe 3 pe -4 R2 SE DW 2.466 0.290 0.383 0.300 0.114 -0 .0 4 7 1.039 Adjusted M1 0.502 2.772 2.09 (0.56) (0.20) (0 .6 6 ) (1.74) (1.18) (0.80) (3.20) (0.67) 2.228 0.295 0.375 0.294 0.130 (0.15) (0.07) (0.80) (2.03) (1.33) 0 .0 1 2 (0 .6 8 ) 0.091 (3.04) 0.044 (0.85) -0 .6 4 8 (3.50) 0.487 2.813 2.04 NOTE: Absolute values of t-statistics appear in parentheses. R2 is the coefficient of determination adjusted for degrees of freedom; SE is the regression standard error; and DW is the Durbin-Watson test statistic. bin-Watson (DW) test statistics indicate, neither equa tion is ham pered by first-order serial correlation. The results for the individual variables also show little difference. In each instance, the pattern of the estim ated lags is sim ilar in magnitude and signifi cance. For example, the hypothesis that the cum ula tive effect on GNP growth of a change in money growth is unity cannot be rejected for M l (t = 0.39) or adjusted M l (t = 0.60). Similarly, we cannot reject the hypoth esis that fiscal actions and changes in relative energy prices have no lasting, significant effects on the growth of GNP. Thus, in term s of overall fit and individual coefficient estimates, there appears to be little differ ence between M l or adjusted M l in explaining GNP growth during the period 1960-79. FEDERAL RESERVE BANK OF ST. LOUIS MARCH 1984 Table 3 GNP Equation Estimates: 1/ 1960- 11/1983_____________________________________ (3) Coefficient Constant (1) M1 (2) Adjusted M1 2.743 (2.33) 3.130 (2.88) 0.262 (2.56) 0.125 0.211 0.230 0.178 0.083 0.827 D1981 m0 m , m ...2 m_3 m _4 XrTii 0.353 0.268 0.074 -0 .0 8 7 (5.22) (2.96) (1.04) (0.82) 0.870 (4.81) A djusted M1 w ith Intercept Shift 2.017 (2.20) 13.112 (4.69) (1.60) (3.96) (3.52) (3.41) (1.09) (5.86) 0.387 (4.34) 0.024 (0.53) 0.014 (0.34) 0.034 (0.86) 0.393 0.250 0.106 0.029 1.164 (6.41) (4.28) (2.16) (0.42) (8.03) 0.044 0.012 -0 .0 1 7 -0.011 0.016 0.044 (0.99) (0.31) (0.40) (0.27) (0.35) (0.36) (0.40) (0.05) (0.44) (0.83) (0.24) (1.04) (1.91) (1.17) -0 .0 0 6 (0.20) 0.008 (0.48) -0 .0 0 2 (0.12) -0 .0 0 8 (0.29) -0.00 1 (0.07) -0 .0 0 5 (0.30) pe 5 pe-e Spe_i 0.011 0.001 -0 .0 0 8 -0.01 1 -0 .0 0 4 0.017 0.057 0.062 -0 .0 1 6 -0 .0 1 6 0.021 0.112 0.101 -0 .0 0 9 -0 .0 0 5 0.017 0.067 0.056 S -0 .7 1 8 (3.34) -0 .5 8 8 (2.78) 6o 6-1 e -2 6 -3 e -4 2e, pe0 p e -i pe -2 pe-3 p e _4 R2 SE DW 0.360 3.369 1.71 0.011 0.026 0.039 0.113 0.371 3.339 1.78 (0.26) (0.70) (0.90) (0.94) (1.26) (0.90) (1.27) (3.66) (1.87) 0.016 -0 .0 0 5 -0 .0 0 5 0.008 0.049 (0.46) (0.13) (0.15) (0.21) (0.46) (0.79) (0.34) (1.15) (2.31) (1.14) - 0.654 (3.46) 0.500 2.980 2.03 NOTE: Absolute values of t-statistics appear in parentheses. R2 is the coefficient of determination adjusted for degrees of freedom; SE is the regression standard error; and DW is the Durbin-Watson test statistic. In-Sample Estimates: 1960—1983 The GNP estim ates using the post-1979 data indicate a substantial deterioration in the equation's explana tory power. As reported in colum ns 1 and 2 of table 3, there is almost a 30 percent reduction in explanatory power regardless of the M l measure used.27 Moreover, 27A similar deterioration is documented, although not explained, in Batten and Thornton (1983). the summed effect of money growth has declined sub stantially. For example, using the 1960-79 sample, a 1 percentage-point change in M l growth has a cum ula tive 1.039 percentage-point change in GNP growth. When the 1960-83 sample is used, however, the esti mate of this cumulative effect drops to a 0.870 percentage-point change. A similar result occurs when the sample period is updated and adjusted M l is used as th e m o n e ta ry v a ria b le (1.082 p e r c e n t to 0.827 percent).28 The problem with the adjusted M l results shown in colum n 2 of table 3, as noted earlier, is that the adjusted M l results are not reliable u nless the 1981 NOW account effect has been taken into account. Thus, the GNP equation using adjusted M l was re-estimated for the 1960-83 period incorporating the intercept shift term. These results, presented in colum n 3, show that the intercept shift term (D1981) is positive and statisti cally significant; thus, the hypothesis that the constant term was subject to a significant displacem ent during 1981 is not rejected by the data. The im portance of capturing this effect is evidenced by the dramatic change in the equation’s explanatory power and in the coefficient estim ates of the money variable.29 When compared with the 1960-79 estimation re sults, the adjusted M l equation with the intercept adjustm ent shows no deterioration in overall fit; the R2 in creases from 0.487 to 0.500, com pared with the approximately 30 percent decline found using M l.30 Not only is the overall fit of the equation actually im proved, but the drop in the sum med coefficient esti mates on adjusted M l that appears when comparing 28lt should be noted that neither sum estimate is statistically different from unity. 29Another procedure also was used to account for the rapid adjustedM1 velocity growth in 1981. Because GNP does not respond im mediately to changes in money growth, a rapid increase (decrease) in money growth during a quarter will appear as a sharp decline (increase) in velocity. Thus, to abstract from the declines in adjusted M1 growth during the NOW account introduction, a (0,1) dummy term is used to form an interaction variable with the adjusted M1 growth. This variable takes on the value of zero in all quarters except 1/1981, 11/1981 and 111/1981, when it equals actual adjusted M1 growth. As expected, the outcome using this approach is quite similar to the shift-adjusted model in table 3. Again the R2 (0.50) is increased by about 40 percent relative to the M1 equation. The deterioration in the coefficient on the summed effect of money growth (2 mi) found using M 1 disappears; the estimated cumulative effect is 1.139. This result provides further evidence on the relative usefulness of adjusted M1 in explaining GNP growth. 30For completeness, we also estimated the M1 equation with the D1981 variable; the estimated coefficient was not statistically sig nificant. 25 MARCH 1984 FEDERAL RESERVE BANK OF ST. LOUIS the results in tables 2 and 3 vanishes as well: a 1 percentage-point change in adjusted M l growth is now estim ated to have a cumulative im pact of 1.164 percentage-point change in GNP growth, slightly high er than the 1.082 percentage-point change reported for the 1960-79 sample. Thus, w hen the velocity change during 1981 caused by the NOW account introduction is taken into account, the adjusted M l m easure ex plains the growth of GNP better than does M l.31 CONCLUSION It has been argued that M l, as it is currently defined, may give a distorted view of actual policy actions on the economy. This problem arises from the public’s will ingness to view som e portion of interest-bearing checkable deposits as savings-type balances. Unfortu nately, there currently is no reliable procedure by which we can disentangle the transaction from the non-transaction shares of these deposits. This is espe cially true in terms of anticipating what those propor tions will be in the future. To investigate the validity of the alleged problem with M l, an alternative M l m easure was derived that excluded all interest-bearing checkable deposits. This adjusted M l m easure — defined simply as M l less other checkable deposits — was used in a reducedform GNP equation, and the results were compared with estim ates obtained using M l. Estim ates derived from the 1960—83 sample period indicate that, once the distorting effects of the NOW account introduction in 1981 are accounted for, the adjusted M l series explains GNP growth better than M l. Although the results suggest that recent criticism of the Ml-GNP link is not unwarranted, they strongly deny the associated claim that the link between trans actions money and GNP has been damaged irrepa rably. Instead, the evidence suggests that a more fruit ful approach would be to sharpen the distinction between transaction deposits and those held for both transactions and savings. 31The results suggest that the “other checkable deposit” (OCD) component of M1 may be dominated by the growth of adjusted M1 in explaining the growth of GNP. To test this, OCDs were added to the adjusted M 1 equation as a separate set of independent variables. The equation was then re-estimated for the 11/1964-11/1983 sample period; the sample period is shorter due to the limited data availabil ity for OCDs. Based on a standard F-test, adding OCDs does not significantly increase the explanatory power of the equation (F l 2 = 1.62). 26 REFERENCES Barnett, William A. “The Optimal Level of Monetary Aggregation,” Journal of Money, Credit and Banking (November 1980), pp. 687-710. Barro, Robert J., and Anthony M. Santomero. “ Household Money Holdings and the Demand Deposit Rate," Journal o f Money, Credit and Banking (May 1972), pp. 397—413. Batten, Dallas S., and Daniel L. Thornton. “ M1 or M2: Which is the Better Monetary Target?” this Review (June/July 1983), pp. 3 6 42. Bronfenbrenner, M. “ Some Fundamentals in Liquidity Theory," Quarterly Journal of Economics (May 1945), pp. 405-26. Carlson, Keith M., and Scott E. Hein. “ Monetary Aggregates as Monetary Indicators,” this Review (November 1980), pp. 12-21. Currie, Lauchlin. The Supply and Control o f Money in the United States (Harvard University Press, 1935). Davidson, Lawrence S., and R. W. Hafer. “ Some Evidence on Selecting an Intermediate Target for Monetary Policy,” Southern Economic Journal (October 1983), pp. 406-21. Friedman, Benjamin M. “The Roles of Money and Credit in Macroeconomic Analysis," in J. Tobin, ed., Macroeconomics, Prices, and Quantities: Essays in Memory of Arthur M. Okun (Brookings, 1983), pp. 161-89. --------------- “ Time to Reexamine the Monetary Targets Frame work,” New England Economic Review (March/April 1982), pp. 15-23. Frodin, Joanna H., and Richard Startz. “ The NOW Account Experi ment and the Demand for Money,” Journal of Banking and Fi nance (June 1982), pp. 179-93. Hafer, R. W. “The New Monetary Aggregates,” this Review (Febru ary 1980), pp. 25-32. __________ “ Much Ado About M2," this Review (October 1981), pp. 13-18. Johannes, James M. “Testing the Shift Adjustment in the Federal Reserve’s New Shift Adjusted M1B,” Economics Letters (8,1981), pp. 367-72. Johannes, James M., and Robert H. Rasche. "Forecasting Multi pliers for the New-New’ Monetary Aggregates,” Shadow Open Market Committee Policy Statement and Position Papers, Work ing Paper No. PPS-81-8 (Center for Research in Government Policy & Business, University of Rochester, September 1981), pp. 39-49. Judd, John P., and Rose McElhattan. “The Behavior of Money and the Economy in 1982-83,” Federal Reserve Bank of San Francis co Economic Review (Summer 1983), pp. 46-51. Klein, Benjamin. “ Competitive Interest Payments on Bank Deposits and the Long-Run Demand for Money,” American Economic Re view (December 1974), pp. 931-49. Klein, Michael A. “ Deposit Interest Prohibition, Transactions Costs, and Payments Patterns: A Theoretical Analysis,” Metroeconomica (November/December 1974), pp. 144-52. __________ “ The Implicit Deposit Rate Concept: Issues and Ap plications,” Federal Reserve Bank of Richmond Economic Review (September/October 1978), pp. 3-12. Lawler, Patrick J. “The Large Monetary Aggregates as Intermediate Policy Targets," Voice of the Federal Reserve Bank of Dallas (November 1981), pp. 1-13. Maddala, G. S. Econometrics (McGraw-Hill, 1977). “ Monetary Policy Report to Congress.” (March 1983). Federal Reserve Bulletin FEDERAL RESERVE BANK OF ST. LOUIS MARCH 1984 Morris, Frank E. "Do the Monetary Aggregates Have a Future as Targets of Federal Reserve Policy?" New England Economic Re view (March/April 1982), pp. 5-14. Spindt, Paul A. “ Money is What Money Does: A Revealed Produc tion Approach to Monetary Aggregation," (Board of Governors of the Federal Reserve System, 1983, processed). Radecki, Lawrence J., and John Wenninger. “ Shifts in Money De mand: Consumers versus Business,” Federal Reserve Bank of New York Quarterly Review (Summer 1983), pp. 1-11. Startz, Richard. “ Implicit Interest on Demand Deposits,” Journal of Monetary Economics (October 1979), pp. 515-34. Santomero, Anthony M. “ A Model of the Demand for Money by Households," Journal of Finance (March 1974), pp. 89-102. Solomon, Anthony M. “ Interpreting Monetary Policy in a Period of Change," Remarks before the Second Annual Conference on Cur rent Financial Issues, Financial Executives Institute, Washington, D.C. (December 7, 1983). Tatom, John A. “ Energy Prices and Short-Run Economic Perfor mance,” this Review (January 1981), pp. 3-17. __________ "Recent Financial Innovations: Have They Distorted the Meaning of M1?” this Review (April 1982), pp. 23-35. __________ "Was the 1982 Velocity Decline Unusual?” this Review (August/September 1983), pp. 5-15. 27