Alternative Views Of Exchange-Rate Determination

Page created by Alfred Sharp
 
CONTINUE READING
Alternative Views
Of Exchange-Rate Determination

By Douglas K. Pearce

   The foreign-exchange value of the U.S.                         While much research has been devoted to
dollar has fluctuated widely since fixed ex-                   providing an explanation for the fluctuations in
change rates were abandoned in the early 1970s.                exchange rates, no single theoretical model has
The variation in exchange rates under the                      emerged predominate. In the beginning of
regime of flexible (floating) rates has been a                 the flexible-rate era, exchange-rate movements
matter of concern to policymakers because of                   were usually analyzed in terms of the demands
the fear that uncertainties could have                         for and supplies of currencies in the foreign-
deleterious effects on world trade. Large                      exchange market, with emphasis on the
changes in exchange rates are also thought to                  transactions originating from international
have significant impacts on the level and com-                 trade flows. The large short-run movements in
position of U.S. production by changing the                    exchange rates, however, cast considerable
relative prices of exports and importcompeting                 doubt on the adequacy of this approach and led
goods. Some analysts attribute a substantial                   to "asset models" that view the determination
part of the current U.S. recession to the impact               of the exchange rate as the outcome of the port-
of the recent rise in the exchange value of the                folio behavior of wealth holder^.^ One asset
dollar on the manufacturing sector, which ex-                  model, labeled the "monetary" model, ex-
ports 20 percent of its output. A stronger U.S.                plains exchange-rate fluctuations largely in
dollar, on the other hand, has a beneficial ef-
fect on U.S. inflation in the short run by re-
ducing the domestic prices of imports.'
                                                               found by Richard K Abrarns, "International Trade Flows
                                                               Under Flexible Exchange Rates," Economic Review,
                                                               Federal Reserve Bank of Kansas City, March 1980, pp.
 1 Empirical support for the view that exchange-rate volati-   3-10. For analyses of the economic impacts of the rise of the
lity has a significantly negative effect on trade flows was    dollar, see C. Fred Bergsten, "The Villain is an Overvalued
                                                               Dollar," Challenge, March-April 1982, pp. 25-32, and
                                                               Robert A. Feldman, "Dollar Appreciation, Foreign Trade,
Douglas K. Pearce is an associate professor of economics at    and the U.S. Economy," Quarterly Review, Federal
the University of Missouri-Columbia and a visiting scholar     Reserve Bank of New York, Summer 1982, pp. 1-9.
at the Federal Reserve Bank of Kansas City. The author         2 For a critique of the flow model, see Michael Mussa.,
would like to thank Dan Vrabac for research assistance and     "Empirical Regularities in the Behavior of Exchange Rates
Don Schilling for helpful comments. The views expressed        and Theories of the Foreign Exchange Market," in Policies
here are those of the author and d o not necessarily reflect   for Employment, Prices, and Exchange Rates, Carnegie-
the views of the Federal Reserve Bank of Kansas City or the    Rochester Conference Series on Public Policy, Volume 11,
Federal Reserve System.                                        ed. by Karl Brunner and Allan H. Meltzer, pp. 9-57.

16                                                                               Federal Reserve Bank of Kansas City
terms of changes in the supplies of or demands               rate since the adoption of floating rates. The
for respective money stocks. According to this               second section discusses the influences of such
model, a fall in a country's exchange rate                   variables as inflation, real income, and the in-
reflects excessive growth in its money stock.                terest rate on the exchange rate. The third sec-
Another asset model, the "portfolio-balance"                 tion describes specific models of exchange-rate
model, extends the analysis to consider a wider              determination. The fourth section reports how
range of financial assets. In this framework, in-            well these models explain movements in the
terest rates and exchange rates are determined               U.S.-Canadian exchange rate. The final section
simultaneously as wealthholders adjust their                 summarizes the findings of the article.
financial portfolios. Consequently, imbalances
in government budgets and current accounts af-                    EXCHANGE-RATE POLICIES AND
fect exchange rates by changing the size and                      RECENT DOLLAR MOVEMENTS
distribution of financial-asset stocks.The lack                The choice of exchange-rate policy is an im-
of consensus on which analytical framework is                portant decision for any country. This section
appropriate is an important problem for                      reviews the differences in policies, discusses
policymakers since the predicted effects of                  how policies affect a country's international
domestic economic policy on the exchange rate,               balance of payments and its domestic money
and hence on the trade sector, differ across                 supply, and describes the recent behavior of the
these models.                                                U.S. dollar under a flexible exchange-rate
   This article reviews the factors considered               policy.
important in determining exchange rates and
examines the integration of these factors into               Alternative exchange rate policies
the exchange-rate models. The first section pro-                A country has a choice of three major
vides background on the distinctions between                 exchange-rate policies-flexible, fixed, or
fixed and flexible exchange-rate policies along              managed-which are distinguished. by the ex-
with a brief history of the U.S. dollar exchange             tent to which the government, usually through
                                                             its central bank, intervenes in the foreign-
                                                             exchange market to affect the exchange rate of
3 One version of the flow model is given in Robert A.        its currency.' If a country adopts a flexible
Mundell, "The Monetary Dynamics of International Ad-         (floating) exchange-rate policy, its central bank
justment Under Fixed and Flexible Exchange Rates,"           does not participate in the foreign-exchange
Quarterly Journal of Economics, May 1960, pp. 227-57.
For a discussion of the origins of the monetary model, see   market. Instead, the price of the country's cur-
Jacob A. Frenkel, "A Monetary Approach to the Exchange       rency relative to foreign currencies is deter-
Rate: Doctrinal Aspects and Empirical Evidence," Scan-       mined by supply and demand in the foreign-
dinavian Journal of Economics. May 1976, pp. 200-24.
Several studies employing this framework are collected in    exchange market. The supply comes from
The Economics of Exchange Rates: Selected Studies, ed. by
Jacob A. Frenkel and Harry G. Johnson, Addison-Wesley,
1978. For analyses using the portfolio-balance model, see
William H. Branson, Hanna Halttunen, and Paul Masson,
"Exchange Rates in the Short Run," European Economic           The foreign-exchange market is not in any one location,
Review, December 1977, pp. 303-24, and Joseph Bisignano      as is, say, the New York Stock Exchange. Rather, it is a
and Kevin Hoover, "Some Suggested Improvements to a          worldwide market connected by electronic communica-
Simple Portfolio Balance Model of Exchange Rate Deter-       tions. This market is essentially never closed and has the
mination with Special Reference to the U.S. Dollar/Cana-     largest trading volume of any financial market. See Robert
dian Dollar Rate," Weltwirschaftliches Archiv, Heft 1,       Z. Aliber, The International Money Game. 3rd ed., New
1982, pp. 19-37.                                             York: Basic Books, 1979, pp. 54-55.

Economic Review        February 1983                                                                                17
holders of domestic currency that need foreign                  domestic currency, or shifts in the private de-
currency to buy foreign goods and services (im-                 mand for the currency, cause fluctuations in the
ports) or assets denominated in foreign curren-                 central bank's holdings of foreign exchange
cies. The demand comes from foreigners that                     rather than fluctuations in the exchange rate.
want to buy domestic goods and services (ex-                        If a country adopts a managed exchange-rate
ports) or assets denominated in the domestic                    policy, its central bank participates in the
currency. Under this policy, the exchange rate                  foreign-exchange market when it decides a
moves to keep the amount of currency de-                        movement in its exchange rate is undesirable.
manded just equal to the amount supplied.' An                   There is no formal commitment to defend a
increase in the demand for (supply of) domestic                 specific exchange rate. Under a managed
currency, arising, say, from an increase in de-                 exchange-rate policy, the effect of a shift in the
mand for domestic (foreign) goods by for-                       supply of domestic currency, or the demand for
eigners (domestic residents), causes an im-                     it, is uncertain. If the central bank wants the ex-
mediate appreciation (depreciation) in the ex-                  change rate change that would result from the
change rate. The exchange rate, then, reflects                  shift, it takes no action and the exchange rate is
the activities of private economic agents or                    allowed to move to its new equilibrium value. If
foreign central banks but not the direct actions                the central bank does not want the change, it
of the domestic central bank.6                                  enters the market to keep the rate constant. If
   If a country adopts a fixed exchange-rate                    the central bank merely wants to smooth the
policy, its government or central bank is active                movement in the exchange rate, as is often the
in the foreign-exchange market, buying or sell-                 case, it buys or sells just enough currency for
ing the country's currency when its exchange                    the exchange rate to adjust slowly to its new
rate starts to deviate from the fixed or pegged                 equilibrium value.
value.' If there is an excess demand for the
country's currency at the fixed rate, the central               Exchange rate policy, the balance of
bank must satisfy the excess demand by buying                   payments, and the money supply
foreign exchange-that is, by supplying its own                     A country's transactions with the rest of the
currency-to keep the exchange rate from                         world are reported for specific periods as its
rising. If there is an excess supply of the coun-               balance of payments statistics. Private transac-
try's currency, the central bank must purchase                  tions are classified either as current or capital
its own currency to prevent the exchange rate                   transactions. Included in the current account
from falling. This is done by supplying foreign                 are purchases or sales of goods and services and
exchange. Hence, shifts in the private supply of                transactions involving interest payments.
                                                                Transactions involving the exchange of finan-
                                                                cial claims appear in the capital a c c ~ u n tThe
                                                                                                               .~
5 The exchange rate discussed in this paper is the spot rate,
the price of foreign exchange for immediate delivery. The
forward exchange rate is the price of foreign currency that
                                                                8 The current account is essentially the sum of the trade
will be delivered at a specific date in the future.             balance (the value of exports minus imports) and net in-
  Domestic monetary policies that affect interest rates, in-    terest income (interest earned from foreign assets less in-
flation, or real incomes may, of course, lead to exchange-      terest paid to foreigners). For a description of alternative
rate changes.                                                   methods of reporting the international balance of
  In practice, there is usually a narrow band in which the      payments, see Leland B. Yeager, International Monetary
exchange rate can fluctuate without the central bank in-        Relations: Theory, Hisrory, and Policy, 2nd ed., New
tervening.                                                      York: Harper & Row, 1976, chap. 3.

18                                                                                Federal Reserve Bank of Kansas City
net private capital flow is the value of domestic                1 traces the bilateral exchange rates between the
financial assets purchased by foreigners minus                   dollar and the West German mark, the Japa-
foreign assets purchased by domestic private                     nese yen, the French franc, the Canadian
residents. If the current account plus the net                   dollar, and the English pound. The chart il-
private capital flow balance out, the country                    lustrates an important point: the dollar may
has a zero balance of payments. If the sum is                    simultaneously appreciate relative to one cur-
positive, the country has a balance of payments                  rency and depreciate relative to another.
surplus. If the sum is negative, the country has                 Generally, however, it fell against most curren-
a balance of payments deficit.                                   cies immediately after the mid-1971 collapse of
   Under a cleanly floating exchange-rate                        the fixed exchange-rate system and has ap-
policy, the balance of payments is always zero.                  preciated across the board since mid-1980.11Be-
This is because any surplus (deficit) implies an                 tween these two periods the dollar fell substan-
excess demand for (supply of) the domestic cur-                  tially relative to the "hard" currencies of West
rency in the foreign-exchange market that an                     Germany and Japan, despite considerable inter-
appreciation (depreciation) of the exchange rate                 vention by the central banks of these
would eliminate. There is no direct relation be-                 countries." Over the same period, the dollar
tween the foreign-exchange market and the                        rose relative to the Canadian dollar, stayed
domestic money supply. Under a fixed                             roughly constant relative to the French franc,
exchange-rate policy, a balance of payments                      rose and then fell back relative to the British
surplus (deficit) raises (lowers) the domestic                   pound. To give an overview of the exchange
money supply unless the central bank takes off-                  rate of the dollar, Chart 2 shows a weighted
setting action^.^ Hence, the choice of exchange-                 average of the dollar's value relative to 10 ma-
rate policy has important implications for the
control of the domestic money supply.
                                                                 United States generally did not intervene in the foreign-ex-
U.S. dollar under flexible exchange rates                        change market, leaving defense of the pegged rates to the
                                                                 countries involved, even though the United States typically
  The foreign-exchange value of the dollar has                   ran balance of payments deficits. As a result of foreign cen-
varied considerably since the effective end of                   tral banks exchanging much of their dollar reserves for
the Bretton Woods regime in mid-1971.1° Chart                    gold-the U.S. gold stock fell about 50 percent from 1950
                                                                 t o 1970-President Nixon eliminated the right of central
                                                                 banks to convert U.S. dollars to gold in August 1971. This
                                                                 led to the Smithsonian Agreement of December 1971 in
                                                                 which exchange rates were realigned. This arrangement did
                                                                 not last long, however, and the United States formally
9 Assume, for example, the country runs a $10 billion            adopted a flexible-rate policy in March 1973. For a detailed
balance of payments surplus. To keep the exchange rate           account of the Bretton Woods agreement, see Kenneth W.
from appreciating, the central bank has to supply the $10        Dam, The Rules of the Game, Chicago: University of
billion excess demand for the home currency so that the          Chicago Press, 1982, chap. 4. For a review of the fined-rate
domestic monetary base (currency plus bank reserves) will        period, see Richard K Abrams, "Federal Reserve Interven-
rise $10 billion. All else constant, this would lead to an in-   tion Policy," Economic Review, Federal Reserve Bank of
crease in the domestic money supply. To offset or                Kansas City, March 1979, pp. 15-23.
"sterilize" the balance of payments surplus, the central         11 Canada adopted a floating exchange rate in June 1970
bank would have to sell $10 billion of domestic government       and the Canadian dollar immediately appreciated against
securities from its portfolio.                                   the U.S. dollar.
10 The 1944 Bretton Woods conference of the Allies set up        l 2 For a discussion of this intervention, see Victor Argy,
a system of fixed-exchange rates among most currencies.          Exchange Rate Management in Theory and Practice,
Under this system, the U.S. dollar was fixed in terms of         Princeton Studies in International Finance, No. 50, Oc-
gold and other currencies were pegged to the dollar. The         tober 1982.

Economic Review          February 1983
Chart 1
 EXCHANGE RATES- U.S. DOLLAR AND MAJOR CURRENCIES
Currency Per Dollar                                                Currency Per Dollar
                                                                                    4.0

                                                                                     3 .O

2.0                                                                                  2.0

 '-L L :
 .4

 .3
      1970 '71
                 British Pound

                   '72   '73     '74   '75 '76   '77   '78   '79    '80 '81    '82
                                                                                      .6

                                                                                      .5

                                                                                      .4
                                                                                       .3

                                                                   Federal Reserve Bank of Kansas City
Chart 2                                  Relative inflation
TRADE-WEIGHTED EXCHANGE RATE                             Because international trade in goods and ser-
        OF U.S. DOLLAR                                vices underlies many of the transactions in the
               (March 1973 = 100)                     foreign-exchange market, changes in domestic
                                                      prices relative to foreign prices are thought to
                                                      affect the exchange rate. If domestic inflation
                                                      exceeds that of a country's trading partners, the
                                                      demand for domestic goods falls, the demand
                                                      for foreign goods rises, and the exchange rate
                                                      of the home currency falls." However, the ex-
                                                      tent and speed of the adjustment of exchange
                                                      rates to different inflation rates are unresolved
                                                      issues. According to the theory of purchasing
                                                      power parity, the exchange rate moves quickly
                                                      to keep the effective prices of goods equal
                                                      across countries. In its strict form, this theory
                                                      asserts that the exchange rate always equals the
                                                      ratio of the foreign price level to the domestic
                                                      price level. For example, if a particular good
                                                      costs $3.00 in the United States and 15 francs in
                                                      France, the exchange rate must be 5 francs to
                                                      the dollar. The theory predicts that domestic in-
                                                      flation higher than world inflation results in an
                                                      immediate depreciation of the domestic ex-
                                                      change rate. Empirical evidence suggests,
                                                      however, that the relationship between infla-
jor currencies, with the weights based on the         tion rates and exchanges rates is much looser
dollar-volume of trade with each country. This        than this theory maintains."
composite measure indicates the dollar                Relative real growth
depreciated generally over the 1970s but has re-
bounded in the last two years.                          Another factor affecting trade flows-and
                                                      thus supplies of and demands for the home cur-
            FACTORS AFFECTING
              EXCHANGE RATES
                                                      13 This assumes the sum of the absolute values of the price
   Before discussing specific theories of             elasticities for domestic exports and domestic imports ex-
exchange-rate determination, it is useful to          ceeds one.
review factors generally thought to influence ex-     14 The literature on purchasing power parity is substantial.
change rates. The factors include a country's         For a historical review, see Jacob A. Frenkel, "Purchasing
                                                      Power Parity: Doctrinal Perspective and Evidence From
inflation rate, real economic growth rate, in-        the 1920s," Journal of International Economics, May
terest rates relative to the rest of the world, and   1976, pp. 169-91. A thorough review of the issues involved
private speculation. Theories of the exchange         in this theory is given by Lawrence H. Officer, "The
                                                      Purchasing-Power-Parity Theory of Exchange Rates: A
rate differ because of the assumptions they           Review Article," Staff Papers, International Monetary
make about the importance of these factors.           Fund, March 1976, pp. 1-61.

Economic Review     February 1983                                                                              21
rency in the foreign-exchange market-is the                     change in the exchange rate. If, say, new one-
growth rate of domestic real income relative to                 year U.S. and West German Treasury notes are
the rest of the world. With all else held con-                  perfect substitutes and pay 10 percent and 5
stant, high domestic real growth is thought to                  percent, respectively, the expected appreciation
weaken a currency's exchange rate because in-                   of the mark over the year must be 5 percent.
creases in domestic real income raise the de-                      This suggests, however, that the interest-rate
mand for imports and hence the demand for                       differential will widen if investors come to
foreign currency relative to the available sup-                 believe for some reason that the mark will ap-
ply.I5 This line of reasoning assumes, however,                 preciate more than 5 percent. In that case, a
that higher domestic growth affects only the                    larger interest differential could occur without
current account. If investors at home and                       encouraging a capital flow from West Germany
abroad view the higher income growth as an                      to the United States. The interest differential
indication of higher retunis on capital, there                  could be just enough to compensate for a
could be a net capital inflow that would more                   higher expected appreciation of the mark.
than offset the current-account deficit. In that                Thus, the source of the interest-rate differential
case, the home currency would appreciate                        determines whether a widening of the differen-
rather than depreciate.                                         tial causes an exchange-rate appreciation.
Relative interest rates
                                                                Private speculation
   A rise in interest rates that makes domestic
assets more attractive to investors (at home and                   Speculation-the purchase (sale) of foreign
abroad) can cause a capital inflow leading to an                exchange for the sole purpose of profiting from
appreciation in the exchange rate. This                         an expected fall (rise) in the domestic exchange
result-that an increase in interest rates creates               rate-is often said to account for much of the
a comparative advantage in the return on                        volatility of exchange rates. Volatility, then, is
domestic over foreign assets and tends to in-                   seen as stemming from the actions of specula-
crease the exchange rate-depends crucially on                   tors rather than from changes in the factors
the reason for the widening interest differential.              determining the equilibrium exchange rate. One
  Consider a case where iinvestors see foreign                  such view assumes that a fall (rise) in the ex-
and domestic assets as perfect substitutes. Their               change rate leads speculators to think a further
portfolios will be in equilitlrium only when the                decline (increase) is imminent and prompts
expected returns on alternative assets are equal.               sales (purchases) of the domestic currency in
The expected return on (a foreign asset, as                     the foreign-exchange market that drive its price
viewed by a domestic resident, is the foreign in-               down (up) further. According to this view,
terest rate plus the expecte:d change in the ex-                speculation is a destabilizing force that
change rate. Perfect substitutability, then, im-                magnifies fluctuations in flexible exchange
plies that in equilibrium the: interest differential            rates and makes fixed rates preferable.
between two countries just equals the expected                     Some analysts, however, see speculation as a
                                                                stablizing force. Since, to make profits, specu-
                                                                lators must buy when the exchange rate is below
                                                                its equilibrium level and sell when it is above its
' 5 This result presumes that the rise in domestic income did
not originate from an increase in net exports caused, for ex-
                                                                equilibrium level, the action of profitable
ample, by an exogenous shift in the demand for domestic         speculators (the only ones that can survive over
goods.                                                          time) push the exchange rate toward its

22                                                                             Federal Reserve Bank of Kansas City
equilibrium rather than away from it.I6 In any                    plies in the foreign-exchange market. The ex-
case, to argue for government intervention to                     change rate is in equilibrium when supply just
counteract speculation is to argue that govern-                   equals demand-when any current-account
ment officials are better judges of the                           imbalance is just matched by a net capital flow
equilibrium exchange rate than private                            in the opposite direction. The current account is
speculators . I 7                                                 assumed to be determined by relative prices and
                                                                  relative real incomes. Increases in domestic
           ALTERNATIVE EXCHANGE                                   prices relative to foreign prices are predicted to
               RATE MODELS                                        have a negative effect on the current account
   This section describes three models that have                  and hence, all else constant, to cause a
been proposed to explain movements in ex-                         depreciation. Goods prices, however, are
change rates. The first focuses on the demand                     presumed to move sluggishly so that purchasing
and supply flows in the foreign exchange                          power parity is not imposed. This allows
market and is referred to here as the "tradi-                     exchange-rate changes originating from other
tional flow" model. The other two models are                      sources to change the relative prices of domestic
asset models. In their analytical framework, the                  and foreign goods. An increase in domestic real
exchange rate is determined by the equilibrium                    income is thought, all else being equal, to cause
conditions in asset markets. One of these is the                  the exchange rate to fall. This is because an in-
"monetary" model, which looks solely at the                       crease in income tends to increase imports,
supply of and demand for money in each coun-                      reducing the current account, with no offsetting
try. The other is the "portfolio-balance"                         effect on capital flows.
model, which extends the analysis explicitly to
                                                                     The model posits that foreign and domestic
include other assets.                                             assets are imperfect substitutes in a portfolio.
Traditional flow model                                            An increase in the domestic interest rate, with
                                                                  no change in the foreign interest rate, is
   The                        used in many                        predicted to cause a net capital inflop that
textbooks         the flow demands and s u p                      results in an appreciation of the exchange rate.
                                                                  Thus, according to this model, a country that
16 For arguments that speculation is likely to be destabi-
                                                                  wants to strengthen the exchange value of its
lizing, see Paul Einzig, The Case Against Floating Ex-            currency must adopt policies to lower prices,
changes, London: MacMillan, 1970, chap. 9. For                    raise interest rates, and reduce real growth.
arguments that speculation is likely to be stabilizing, see
Milton Friedman, "The Case for Flexible Exchange                     The main theoretical criticism of the tradi-
Rates," in Essays in Positive Economics, Chicago: Univer-         tional flow model is directed at its implications
sity of Chicago Press, 1953.                                      for the asset market. The model predicts that an
17 Political, as well as economic, instability also affects a     exchange rate could be in equilibrium when a
country's exchange rate although its impact is difficult to
quantify. Political decisions that result in trade restrictions   country is running a current-account deficit if
and capital controls create artificial barriers that interfere    the domestic interest rate is high enough to
with the economic forces bearing on exchange rates. More          maintain an offsetting net capital inflow. This
dramatic actions, such as the nationalization of banks in
Mexico or the election of the Socialist party in France,          implies that at a constant interest differential,
make investments appear riskier and often lead to domestic        there is a steady, potentially infinite, accumula-
capital outflows. The political stability of the United States    tion of domestic assets by foreigners. No ac-
makes it the natural recipient of such capital flows. Conse-
quently, the dollar usually appreciates when international        count is given of how the portfolios of for-
disruptions occur.                                                eigners are brought into equilibrium.

Economic Review          February 1983                                                                           23
Monetary model                                           mand that, with a fixed nominal money supply,
   There are several variations of the monetary          results in a reduction in domestic prices and,
model, but they all share the premise that               through purchasing power parity, pulls the ex-
movements in the exchange rate between two               change rate up. An increase in the domestic in-
currencies can be explained by changes in the            terest rate, which is assumed to reflect higher
demand for or supply of money in the two                 expected inflation, lowers money demand,
countries. In contrast to the traditional model,         raises prices, and lowers the exchange rate.
in which the exchange rate is determined by              Changes in foreign variables have symmetric ef-
trade and capital flows, this model asserts that         fects. The domestic exchange rate is increased
the equilibrium exchange rate depends on the             by a rise in the foreign money supply, by a
stock-equilibrium conditio~lsin each country's           reduction in foreign real income, and by an in-
money market. The model is derived from                  crease in the foreign interest rate.
several assumptions. First, purchasing power                Like the traditional flow model, the
parity holds continuously so that the exchange           monetary model predicts that changes in
rate always equals the ratio of price levels in the      domestic real income and interest rates affect
two countries. Second, domestic and foreign              the exchange rate. The effects are in the op-
bonds are perfect substitutes so that any dif-           posite direction, however, since the monetary
ference in interest rates equals the expected rate       model asserts that rapid economic growth and
of change in the exchange rate. These two                low interest rates should cause the exchange
assumptions imply that interest differentials            rate to appreciate rather than depreciate.
just equal differences in expected inflation                Criticism of the monetary model centers on
rates. Third, the demand for money in each               its assumptions. First, several investigators
country is a stable functiorl of the domestic in-        have reported evidence that purchasing power
terest rate and real income. Fourth, if out of           parity does not hold in the short run.19 In par-
equilibrium, the money market adjusts rapidly,           ticular, it is argued that prices are "sticky" in
with domestic prices moving quickly to                   the short run and do not have the required flex-
eliminate any excess supply of & demand for              ibility to keep the money market in equilibrium.
money.                                                   Second, if domestic and foreign bonds are not
   These assumptions yield an equation for the           perfect substitutes, as the monetary model
equilibrium exchange rate in terms of dif-               assumes, the model must take into account
ferences between the two countries' money sup-           changes in the composition of portfolios with
plies, interest rates, and real i n c ~ m e s . 'The
                                                 ~       respect to these two assets. This consideration
partial effects of these variables are predicted to      leads to the portfolio-balance model.
be as follows. An increase in the domestic               Portfolio-balance model
money supply reduces the ,exchangerate as the
initial excess money supply drives domestic                The portfolio-balance model views the ex-
prices up and hence, through purchasing power            change rate and interest rates as determined
parity, the exchange rate down. An increase in
domestic real income causes excess money de-
                                                         19 The strength of commodity arbitrage in keeping in-
                                                         dividual prices in line was found to be weak by Irving B.
                                                         Kavis and Robert E. Lipsey, "Price Behavior in the Light
18 This assumes that the demand for money functions in   of Balance of Payments Theories," Journal of Interna-
each country have identical parameters.                  tional Economics, May 1978, pp; 193-246.

                                                                          Federal Reserve Bank of Kansas City
simultaneously by the portfolio equilibrium                      original interest rates and exchange rate, port-
conditions for wealthholders in each country.                    folios would no longer be in their desired pro-
Residents of each country are assumed to                         portions, since domestic wealthholders would
allocate their net financial wealth among three                  want to redistribute their wealth increase
assets: the domestic monetary base, domestic                     toward domestic bonds and foreign bonds.
government bonds, and net foreign bonds                          With the foreign interest rate fixed, actions of
denominated in foreign currency.z0The desired                    domestic investors to realign their portfolios
proportions of these assets are assumed to de-                   would result in a drop in the domestic interest
pend on their respective yields, with domestic                   rate and a depreciation of the exchange rate."
and foreign bonds considered imperfect                           An increase in net holdings of foreign bonds
substitutes. An increase in the domestic                         resulting from a current-account surplus would
(foreign) interest rate causes investors to in-                  also increase domestic wealth and disturb port-
crease the desired proportion of their wealth in                 folio equilibrium. In that case, domestic
domestic (foreign) -bonds and to lower the                       wealthholders would want to hold some of the
desired proportions in the monetary base and                     wealth increment in the form of domestic
foreign (domestic) bonds. The outstanding                        assets. This would lead to a fall in the domestic
stocks of these assets are fixed at any point in                 interest rate and an appreciation of the ex-
time so that the exchange rate and the two in-                   change rate.
terest rates equal the values at which wealth-                      Unlike the first two cases, an increase in
holders are just willing to hold existing assets,                domestic government bonds has an uncertain
assuming asset markets clear continuously.                       effect on the exchange rate. On the one hand,
   Stocks of financial assets change over time,                  the increase in wealth would increase domestic
causing interest rates and the exchange rate to                  demand for foreign assets resulting in an ex-
change. Bond-financed government deficits                        change-rate depreciation. On the other hand,
(surpluses) increase (decrease) the private                      the increase in domestic government debt
holdings of government bonds. Money-                             would raise the domestic interest rate, making
financed deficits (surpluses) increase (decrease)                foreign bonds less attractive. If this substitution
the monetary base. Current-account surpluses                     effect were larger than the wealth effect, the net
(deficits) increase (decrease) net domestic                      result would be an appreciation of the exchange
holdings of foreign debt.                                        rate.
   An increase in the domestic monetary base
would increase domestic wealth and raise the
proportion of wealth held in this asset. At the
                                                                 z1 An increase in the monetary base causes an excess supply
                                                                 of this asset at the original exchange rate and interest rates
                                                                 under the usual assumption that the partial derivative of the
20 The model is concerned with the allocation of the net         demand function for each asset with respect to wealth is less
wealth of all private domestic wealthholders. Since demand       than one. The excess supply is matched by excess demands
deposits are liabilities of domestic banks, the monetary base    for domestic and foreign bonds. The excess demand for
rather than the money supply appears in the model. The in-       domestic bonds raises their price-lowers the domestic in-
clusion of domestic government debt in the hands of              terest rate-which increases the proportion of wealth held
domestic residents assumes that this too is an "outside"         in domestic bonds. The excess demand for foreign bonds
asset-that residents do not take account of the present          (denominated in foreign currency) increases the demand for
value of the implied tax liability associated with the govern-   foreign currency resulting in a depreciation that increases
ment debt. Note that an appreciation of the exchange rate        the proportion of wealth held in foreign bonds.
lowers the domestic currency value of foreign assets and         Equilibrium is restored when these proportions reach their
hence lowers domestic wealth.                                    higher desired levels.

Economic Review         February 1983
The portfolio-balance model incorporates                            EMPIRICAL EVIDENCE ON
elements of both the traditional flow model and                       ALTERNATIVE THEORIES:
the monetary model. By including foreign                         THE U.S.-CANADIAN EXCHANGE RATE
currency-denominated assets, it allows current-                     This section examines how well the different
account imbalances to affect exchange rates as                  t h e o r i e s explain movements i n t h e
the flow model predicts. By including the                       U.S.-Canadian exchange rate. This particular
monetary base of each country, it allows dif-                   case was chosen for several reasons. First,
ferences in monetary growth to affect exchange                  Canada and the United States are each other's
rates, as the monetary model predicts. The                      largest trading partner, so fluctuations in this
channels of influence differ, however. Because                  exchange rate can have substantial effects on
the portfolio model focuses only on distur-                     trade flows. Second, Canada adopted a flex-
bances to asset portfolios, it ignores the                      ible-rate policy in June 1970, before most other
underlying determinants of trade as well as the                 countries, so a longer time period is available
role of purchasing power parity."                               for empirical tests. Third, data on bilateral
   There have been reservations about the                       asset holdings, required for estimation of the
portfolio-balance model. Because comparative                    portfolio-balance model, are more extensive for
returns on domestic and foreign assets are im-                  these countries than for most other countries.
portant in the model, expected exchange-rate
movements must be considered. Different                         Graphical overview
assumptions about how a.gents form their ex-                      Before examining statistical estimates of the
pectations can lead to very different predictions               models, a graphic overview may help show
from the model. This issue is particularly im-                  broad relationships between the exchange rate
portant if one country is a net debtor rather                   and its possible determinants. Chart 3 com-
than creditor in foreign currency-denominated                   pares movements in the ratio of Canadian to
bonds. Under some assumptions about the for-                    U.S. price levels, measured by the wholesale
mation of exchange-rate expectations, the                       price indexes, with the exchange rate measured
model may be unstable.23The model may also                      as Canadian dollars per U.S. dollar. According
be difficult to use in empirical work because of                to the theory of purchasing power parity, these
the scarcity of data on domestic holdings of                    two series should move together quite closely.
foreign financial assets.                                       As the chart shows, the price ratio does roughly
                                                                correspond with the exchange rate. While the
                                                                relationship is not very close in terms of
22 Changes in the composition of wealth that leave its level    quarterly movements, the fall in the exchange
initially unchanged also have exchange rate effects accord-
ing to this theory. An open market purchase (sale) of           value of the Canadian dollar since 1976 has
domestic government bonds by the central bank reduces the       coincided with higher Canadian inflation, as
domestic interest rate and cause!; a depreciation (apprecia-    the theory of purchasing power parity predicts.
tion) of the exchange rate. 1ntc:rvention in the exchange
market by the central bank in the form of purchases or sales
of foreign bonds has the same qualitative effects on the
domestic interest rate and the exchange rate as open market
operations.
23 If one country is a large net debtor in foreign currency     tional and speculation is stabilizing. See Dale W. Hender-
denominated financial claims and agents have static expec-      son and Kenneth Rogoff, "Negative Net Foreign Asset
tations-if they assume that exchange rates will not             Positions and Stability in a World Portfolio Balance
change-the portfolio-balance model is generally unstable.       Model," Journal of International Economics, August
This instability disappears, however, if expectations are ra-    1982, pp. 85-104.

26                                                                                Federal Reserve Bank of Kansas City
Chart 3                                                           Chart 4
   CANADIAN-U.S. EXCHANGE RATE                                    CANADIAN-U.S. EXCHANGE RATE AND
    AND RATIO OF PRICE LEVELS                                        INTEREST RATE DIFFERENTIAL
                                           Canadian Dollars       Interest Rate                  Canadian Dollars
Ratio of Prices                            Per U.S. Dollar        Differential                   Per U.S. Dollar
1.35                                                       1.25

                                                                           Interest Rate   ( ~ i ~ h ~
       -

               Ratio of Price

           I   I   I   I   I   I   I   I   I   I   I   I

   Chart 4 compares the interest-rate differen-                   the decline in the differential from mid-1976 to
tial, measured by the Canadian commercial-                        mid-1978 occurred as the value of the Canadian
paper rate less the U.S. commercial-paper rate,                   dollar declined steeply.
with the exchange rate. The traditional flow                         Chart 5 compares the differences in real
model predicts that the two series should be in-                  growth rates, measured as the annualized rate
versely related, because it asserts that a wider                  of change in real GNP in Canada less the U.S.
differential will cause a net capital flow into                   counterpart, with the exchange rate. The tradi-
Canada, raising the value of the Canadian                         tional flow model asserts that higher economic
dollar. The monetary model, on the other                          growth in Canada should cause the Canadian
hand, predicts a positive relationship, because                   dollar to fall, due to its adverse effects on the
it assumes a wider differential reflects higher                   current account. In contrast, the monetary
expected inflation in Canada and, thus, a                         model asserts that faster real income growth
depreciation in the Canadian dollar. Graphic                      should strengthen the Canadian dollar, since it
evidence suggests that larger interest differen-                  raises the demand for Canadian money. Al-
tials are associated with a rising Canadian                       though the graphic evidence indicate no dis-
dollar, although the relationship appears weak.                   cernible short-run relationship between relative
The bulge in the interest differential in 1975-76                 growth rates and the exchange rate, the Cana-
coincided with a stronger Canadian dollar and                     dian economy grew faster than the U.S.

Economic Review            February 1983                                                                        27
. Chart 5                                 used in past work to capture the essence of the
  CANADIAN-U.S. EXCHANGE RATE                       alternative theories.14 The purchasing power
 AND REAL GROWTH DIFFERENTIAL                       parity equation is also included because strict
Growth Rate                     Canadian Dollars    purchasing power parity is an assumption of
Differential                    Per U.S. Dollar     the monetary model and is assumed by the
 6.0 1                                     11.25    traditional flow model to hold partially. Above
                                                    each coefficient is the sign expected from each
                                                    theory. Table 2 presents estimates of the dif-
                                                    ferent models, based on quarterly data over the
                                                    flexible-rate period from 1971:Q1 to 1982:Ql.
                                                       The estimation results point to the conclusion
                                                    that none of the theories is fully supported by
                                                    the Canadian-U. S. experience.15 The estimate
         leal ~ r o w t t
                                                    of the purchasing power parity model implies
         Differential                               that a Canadian inflation rate one percentage
     1   (Left Scale)       v                       point above the U.S. inflation rate is associated
                                                    with a depreciation of the Canadian dollar of
                                                    only 0.5 percentage points, half the impact
                                                    predicted by the theory.26

                                                    24 Other representations of the models have been pro-
                                                    posed. For a version of the monetary model that relaxes the
                                                    assumption of strict purchasing power parity, see Jeffrey
                                                    A. Frankel, "On the Mark: A Theory of Floating Exchange
economy in the first half of the 1970s and the      Rates Based on Real Interest Differentials," American
Canadian dollar was strong. In the second half      Economic Review, September 1979, pp. 610-22. Another
of the 1970s, U.S. growth was generally higher      monetary model that allows for central bank intervention is
                                                    given in Lance Girton and Don Roper, "A Monetary
and the Canadian dollar fell. Thus, the             Model of Exchange Market Pressure ~p'plied to the
evidence does not support the traditional flow      Postwar Canadian Experience," American Economic
model but is weakly consistent with the             Review, September 1977, pp. 537-48. An alternative version
                                                    of the portfolio model does not assume that the foreign in-
monetary model. These graphic implications          terest rate is exogenous and therefore includes the cor-
could be misleading, however, since the             responding foreign asset holdings.
theories predict the effect of one variable         25 Similar negative results, based on forecasting perfor-
                                                    mance, were found for other exchange rates by Richard A.
holding all others constant and the graphs allow    Meese and Kenneth S. Rogoff, "Empirical Exchange Rate
other variables to change. The next section,        Models of the Seventies: Are Any Fit to Survive?" Interna-
therefore, presents statistical evaluations of      tional Finance Discussion Papers, No. 184, Board of
                                                    Governors of the Federal Reserve System, 1981.
each exchange-rate model.
                                                    26 The effect of the inflation differential on the exchange
                                                    rate was also not estimated precisely. Only at the 6 percent
Estimates of exchange rate models                   significance level can one reject both the hypothesis that the
   Table 1 presents the single-equation represen-   change in the exchange rate is unrelated to the inflation dif-
                                                    ferential and the hypothesis that the exchange rate moves
tations of the traditional flow, monetary, and      on a one-to-one basis with the inflation differential (pur-
portfolio-balance models, which have been           chasing power parity holds).

                                                                       Federal Reserve Bank of Kansas City
statistically significant effect on the exchange
             Table 1
                                                               rate, while the interest-rate differential did not
ALTERNATIVE MODELS OF EXCHANGE
                                                               have the positive impact specified by the
      RATE DETERMINATION
                                                               theory.=' The portfolio-balance model fares
A. Purchasing Power Parity                                     little better. Canadian asset stocks had no
                            (+I                                discernible effect on the exchange rate,
     In et = ag       +     a1 ln(Pc/Pu)t        +   ft        although the U.S. interest rate did have the
B. Traditional Flow Model                    -                 predicted result. A higher U.S. interest rate was
                                                               associated with a lower exchange rate for the
                          (+I
     In q = bo        + bl      l n ( ? / ~ ~+) ~              Canadian dollar ."
                (+I                        (-I                    There could be several reasons for the lack of
          b2 In(Pc/Pu)t + b3(rc- ru)t + e t                    success with the models. The models assume
C. Monetary Model                                              stable asset-demand functions, a premise that
                                                               may not have held in the 1 9 7 0 ~The . ~ ~models
     In q = co        + (c1+I ln(MC/MU)t+                      do not incorporate the behavior of speculators,
                (-1                        (+I                 and the frequent economic and political shocks
            c2 l n ( f / ~ ~+ )c3(rC-rU)t
                                ~        + et                  over the period may have made speculation an
D. Portfolio Balance Model                                     important factor. The models, in assuming
                      (+I            (*I             (-I       freely floating exchange rates, do not allow for
     et   = Q    +    dl M B +
                             ~ d2 E$             +   dj    +   intervention by central banksa30
           (+)
               d4ry   +     et
                                             3

Definitions:                                                   27 While the lack of support for purchasing power parity
                                                               may be viewed as a priori evidence that the monetary model
   e = spot exchange rate defined as number of                 must be invalid, it has been argued that because published
   . Canadian dollars per U.S. dollar                          price indexes are inadequate for evaluating purchasing
 P1 = price level in country i                                 power parity, monetary models should be tested directly.
  y! = real income in country i                                This argument is made, for examule, by John F. 0. Bilson,
  r1 = nominal interest rate in country i                      " ~ a t i o n a lExpectations and the-~xchangeRate," in The
                                                               Economics o f Exchanne Rates: Selected Studies. Estimates
M B =~ monetary base in Canada                                 of the monetary model developed by Frankel ("On the
 BC = Canadian government debt held by Canadian                Mark...,"), which relaxes strict purchasing power parity,
       residents                                               also yield results that do not support the monetary model.
 FC = net U.S. dollar claims held by Canadian residents        28 Estimates of the portfolio-balance model that include
   c = Canada                                                  U.S. asset stocks instead of the U.S. interest rate produce
   u = United States                                           essentially similar results.
   E = error term
                                                               29 The demand for money appears to have been unstable
                                                               over the 1970s and, assuming at least part of money de-
   The estimate of the traditional flow model in-              mand is for wealth portfolio needs, this implies instability
dicates that only the prediction that higher                   in other asset demand functions. For a survey of the money     ,
                                                               demand evidence, see John P. Judd and John L. Scadding,
Canadian interest rates were associated with an                "The Search for a Stable Money Demand Function," Jour-
appreciation of the Canadian dollar is consis-                 nal of Economic Literature, September 1982, pp. 993-1023.
tent with the evidence. Even this effect is not                30 Allowing for a reaction function by the Bank of Canada
strongly supported. The estimate of the mone-                  along the lines suggested by Branson, Halttunen, and
                                                               Masson, "Exchange Rates in the Short Run," does not,
tary model shows similarly negative results. The               however, significantly change the estimation results for the
difference in money-supply growth had no                       portfolio-balance model.

Economic Review              February 1983
, ~.
         l       $'     n   tPi*d
                                            .: Table 2
                                YESTIMATEDEXCHANGE RATE MODELS
Estimation Period: 19713-1982:I
           Model                                                      (rC - rU)
                            Constant In (PC/PU) In (MC/MU) In (yC/yU) -                              R2 -
                                                                                                     -  SEE -  9
                                                                                                            DW -
A. Purchasing Power                      .511                                                                .015
    Parity*                            (1.906)
B. Traditional Flow            .l$3      .347 7:              -.I37     - .004                       .I36 .015 1.41       .99
      .,-    "             :( (,321)   (1.243) . "1"       (- 1.155) (- 1,821)        '                  '
C. Monetary                 - ,285                    0.54    - .154    - .004                       .093 .016 1.24 .99
                              (.500)                 (351) (- 1.245) ( - 1.704)
Estimation Period: 1971:I-1981 :IV
           Model            Constant MBC             - BC   -  Fc      -  ru                            SEE -
                                                                                                     R2 -
                                                                                                     -         $
                                                                                                            DW -
D. Portfolio Balance            -,   .917        .001 .'.. .003         - .001               .004    321' .019 1.84 .91
     ~   -
         -   p              ,   ..,.(15,.881)   (.086) ..:5-(1.451)
                                                       ?,              ( - .972)' .        (3.222)
*~stim'atedin fist-differen2iform.
Note: See Table 1 for definitions of variables.
       Numbers in parentheses are t-statistics.
         R2 = multiple correlation coefficient
       SEE = standard error of estimate
        D y = Durbin-Watson statistic
      ,., P  = estimated autocorrelation coefficient
                             ,
                           r".

     SUMMARY AND CONCLUSIONS                                      government-budget or current-account im-
                                                                  balances.
   The recent rise in the foreign-exchange value                     Each of these models was estimated to see if
of the U.S. dollar has important implications                     it accounted for changes in the Canadian-U.S.
for many sectors of the U.S. economy as well as                   exchange rate over the flexible-rate period. The
for other countries. This article has reviewed                    empirical results suggest that none of the
the major factors thought to influence the ex-                    models can be considered an adequate guide for
change rate, especially relative inflation,                       economic policy. There was little evidence of a
relative growth in real income, and interest-rate                 systematic, short-run relationship between the
differentials. Three models of exchange-rate                      exchange rate and differences in money growth,
determination were discussed.                                     differences in economic growth, or changes in
   The first model focuses directly on the flow                   asset portfolios. Two regularities did emerge
demands and supplies in the foreign-exchange                      from the empirical work, however: first,
market arising from international trade in                        although the short-run relationship is im-
goods and assets. The second attributes                           precise, inflation higher than in other countries
exchange-rate changes to differences in the                       is linked to exchange-rate depreciation, and se-
growth rates of money supplies. The third                         cond, domestic interest rates higher than in
asserts that exchange-rate movements reflect                      other countries a r e associated with an
asset portfolio readjustments caused by                           exchange-rate appreciation.

30                                                                                        Federal Reserve Bank of   Kansas City
The Economic Review (USPS 417-850), published, edited, and owned by the Federal Reserve Bank of
Kansas City, is distributed monthly except for the July-August and September-October issues. Sub-
scriptions and additional copies are available to the public without charge by writing the Research
Division, Federal Reserve Bank of Kansas City, 925 Grand Avenue, Kansas City, Missouri 64198. If
any material is reproduced from this publication, please credit the source. Controlled circulation
postage paid at Kansas City, Missouri. The purpose and status of this publication has not changed
during the preceding 12 months. The circulation has been reported as follows:

                                               Average No. Copies Each    Actual No. Copies of Single
        Extent and Nature of Circulation        Issue During Preceding     Issue Published Nearest to
                                                      12 Months                    Filing Date

Total Number of Copies (Not Press Run)                34,213
Free Distribution by Mail, Carrier or Other
 Means, Samples, Complimentary, and
 Other Free Copies                                    32,213
Total Distribution                                    32,213
Copies Not Distributed (Office Use, Left
 Over, Unaccounted, Spoiled After
 Printing)                                             2,m
TOTAL (Should Equal Total Shown Above)                34,213

All other portions of USPS Form 3526, Statement of Ownership, Management, and Circulation, do
not apply to this publication.
You can also read