Financial turbulence and international investment

Robert Z Aliber1

Fifty years ago, an ongoing debate about international monetary reform was initiated by the
publication of Robert Triffin’s Gold and the Dollar Crisis2. Triffin had identified an apparent
inconsistency in international financial arrangements; if the demand for international reserve
assets of various foreign countries were to be satisfied, then the United States would incur
payments deficits year after year, and the U.S Treasury’s gold holdings eventually would be
exhausted. But if the United States adopted measures to avoid balance of payments deficits,
other countries as a group would not be able to satisfy their demand for international reserve
assets. Competition among countries for international reserve assets would be deflationary
and lead to declines in prices.
Three groups of proposals were directed at the Triffin dilemma – two would lead to more
rapid increases in the supplies of international reserve assets, and the claim for the third was
that it would reduce the demand for reserves. One generic approach – the dominant
approach – was to produce “paper gold”; a new international reserve asset that would share
the attribute of gold in that it would be an asset without being the liability of any institution or
government. The motive for the paper gold proposals was the desire to enable the United
States to maintain the U.S. dollar parity of $35 an ounce, then viewed as the centrepiece of
international financial arrangements. The belief was that the annual or periodic increases in
the supply of paper gold would satisfy the increases in the demand for international reserve
assets.
The second approach toward increasing the supply of reserves was that the U.S. dollar price
of gold be increased to $70 or perhaps to $100, with comparable percentage increases in the
price of gold in terms of other most other currencies. (A few countries might use the occasion
of the change in the U.S. dollar price of gold to change their parities in terms of the
U.S. dollar.) The value of the gold owned by central banks immediately would increase in the
same proportion as the increase in the U.S. dollar price of gold. Moreover gold production
would be stimulated. Finally the higher price of gold would lead to a reduction in the private
demand, and central banks would acquire a higher proportion of annual production.
The third approach to resolve the Triffin dilemma was to abandon the system of adjustable
parities for currencies, which would then float, much as the Canadian dollar had from 1950 to
1962. A shock that would have led to a payments deficit if a currency had been pegged
instead would lead to a decline in the price of that country’s currency; similarly a shock that
would have led to payments surplus if the currency had been pegged would have led to an
increase in its price. Since central banks would no longer be committed to maintaining the
value of their currencies, they would no longer acquire international reserve assets.
International monetary arrangements now incorporate each of the three major sets of
proposals. The Special Drawing Rights arrangement embodied a paper gold proposal and
was implemented in 1969 when the SDR equivalent of $3 billion of U.S. dollars was
produced and attached to the International Monetary Fund; a member country could use its
SDR to purchase the currencies of other IMF members. SDR outstanding now total
$308 billion. A floating currency arrangement was adopted, initially in August 1971 when the

1
Professor of International Economics and Finance Emeritus, Booth School of Business, University of Chicago.
2
Full disclosure: Triffin was my thesis advisor at Yale in the late 1950s.

BIS Papers No 58 5
U.S. Treasury formally closed its gold window and sought to achieve the revaluation of the
Japanese yen and the French franc, and then again in February 1973 when the Smithsonian
Agreement faltered. The private market for gold was segmented from the official market in
the spring of 1968, and then the U.S. gold market window was formally closed in August
1971. Market forces led to an increase in the U.S. dollar price of gold to nearly $200 in 1973
and then to nearly $1,000 in January 1980. In the last few months gold has traded above
$1,300; the gold component of central bank reserves is now five or six times larger than the
SDR component. Moreover, the supply of international reserve assets has surged, despite
the earlier argument that the demand for reserve assets would decline once currencies were
no longer pegged.
One dominant feature of the last 40 years is that there have been four waves of financial
crises; each wave has involved the failure of banks and other credit institutions in three, four
or more countries. These financial crises often have occurred at the same time as currency
crisis. The first of these waves of crises was in the early 1980s, when the governments,
government-owned firms, and banks in Mexico, Brazil, and 10 other countries failed. Japan
in the 1990s was the primary country in the second wave. The most recent wave began in
2007; banks in the United States, Britain, Ireland, Spain, and Iceland tumbled into
bankruptcy. Each of these waves of crises was preceded by a wave of credit bubbles when
the indebtedness of a group of borrowers increased by 20–30% a year; most of these credit
bubbles led to rapid increases in the prices of real estate and stocks. The prices of these
assets declined sharply when the credit bubbles were pricked, and financial crises followed.
Most of these waves of credit bubbles followed from an increase in cross-border money flows
to these countries, which led to the appreciation of their currencies and an increase in asset
prices.
These cross-border money flows have been both much larger and much more variable than
when currencies were attached to parities. The rates of return to the investors who undertook
the cross-border money flows have been adversely impacted by the financial crises.
The first of the six sections of this paper is descriptive and summarizes the turbulence in
international financial markets in the last 40 years. The second section is analytical, and
highlights the sources of financial crisis. The third section identifies the impacts of structural
shocks and monetary shocks on currency values. The fourth section highlights the role of
carry trade investors and the impact of their transactions on currency values and asset
prices. The fifth section examines the factors that lead to increases in cross-border money
flows by carry trade investors. The sixth highlights the risks and the returns of cross-border
investments in a world characterized by large movements in currency values. The concluding
section summarizes the main features of the paper.

I. Monetary turbulence and financial crises

The striking development since the early 1970s has been the turbulence in the currency
market and in national financial markets. The proponents of a floating currency arrangement
had suggested that the changes in currency values would be reasonably small and would
reflect changes in the differences in national inflation rates, and that the deviations between
the market prices of currencies and the shadow prices computed from differences in inflation
rates would be smaller than when currencies were pegged. But instead the range of
movement in the currency prices has been much larger and the scope of overshooting and
undershooting much much larger than when currencies were pegged.
There have been four waves of financial crises; each of these waves involved the failure of
banks and other credit institutions in three, four, or more countries at about the same time.
The first of the four waves of crises involved the failure of the governments and government-
owned firms in Mexico, Brazil, Argentina, and 10 or so other developing countries in the early

6 BIS Papers No 58
1980s to pay the interest on their U.S. dollar indebtedness in a timely way; their currencies
depreciated sharply. The domestic banks in these countries failed when many of the
borrowers defaulted on their loans to the domestic banks after the currencies depreciated
sharply, since the borrowers often had liabilities denominated in the U.S. dollar and their
indebtedness surged when their currencies depreciated. The second wave was centred on
the failures of banks and credit institutions in Japan in the early 1990s when property prices
declined; at about the same time the banks in three of the Nordic countries tumbled in
response to sharp declines in real estate prices. The Asian financial crisis was the third wave
and involved the failures of banks in Thailand, Malaysia, Indonesia, and South Korea,
although banks in Russia and Argentina also failed during this wave. Similarly the financial
crisis in Mexico at the end of 1994 and the beginning of 1995 was the bellwether of events
that would impact Thailand and Indonesia thirty months later. The fourth wave of failures of
banks and credit institutions began in 2007 and 2008 and involved the United States, Britain,
Ireland, Spain, and Iceland.
Each of these waves of financial crises followed a period of three, four, or more years when
the indebtedness of a similarly placed group of borrowers in different countries increased at
the rate of 20–30% a year. Most of the waves of indebtedness resulted from cross-border
money flows. Thus bank loans to governments and government-owned firms in Mexico,
Brazil, and the other developing countries increased by 30% a year for nearly a decade, and
the total external indebtedness of these countries increased by 20% a year. Each of the next
three waves led to bubbles in real estate prices Bank loans to buyers of real estate in Japan
increased by 25–35% throughout the 1980s; the increases in the price of real estate led to
comparably large increases in stock prices. The external indebtedness of most of the
countries that were involved in the Asian Financial Crisis and of Mexico had increased
sharply in the early 1990s; the money inflows resulted in part because the overhang of bank
loans that were default was funded into long-term bonds. Moreover some investment banks
had discovered that “emerging market equities were a new asset class”, which led pension
funds and mutual funds to buy these securities. Some of these countries had privatized
government owned firms, and some of the newly privatized firms were acquired by firms
headquartered in the industrial countries. Banks headquartered in the emerging market
countries sourced for money from the banks headquartered in the industrial countries
because the interest rates were below those in the domestic money markets. The United
States, Britain, Ireland, Spain, and Iceland experienced a large money inflows after 2002,
and bank loans for real estate purchases in these countries increased rapidly.
Many of the banking crises have been associated with the abrupt depreciation of currencies;
the principal exception was that most of the banks and many other financial institutions failed
in Japan in the 1990s but there was no crisis in the yen. A second exception is that the
financial crisis in Ireland in 2008 was not associated with a currency crisis, because Ireland
did not have its own currency. The Greek currency crisis led to a significant depreciation of
the euro.
The data on the changes in the prices of currencies belie one of assertions advanced by the
proponents of floating exchange rates, that changes in prices of currencies would be
systematically related to changes in national differences in inflation rates on a week to week
and month to month basis, and in the short run – say intervals of up to four or five years. In
the long run, purchasing power parity concept is validated, but at shorter intervals the
deviations from the values are suggested by the differences in national inflation rates.

II. The causes of financial crises

Several features of these cross-border money flows should be noted. The first is that the
increases in money flows to countries have two immediate impacts. One was that their

BIS Papers No 58 7
currencies appreciated and the second was that asset prices in these countries increased in
response to purchases by foreign buyers – who bought the currencies so they could buy
securities. Household wealth increased as asset prices increased, which led to higher levels
of consumption spending and more imports and a larger trade deficit. Increases in asset
prices, household wealth, and imports were an integral part of the adjustment process, which
required that the current account deficit increase by an amount that corresponded with the
autonomous increase in the capital account surplus.
A second feature is that the indebtedness of many of those who had borrowed to buy real
estate was increasing at two to three times the rate of growth of their incomes, which meant
that that the ratio of their indebtedness to their incomes was increasing at a rapid rate – one
which was too high to be sustained. Similarly the external indebtedness of these countries
was increasing more rapidly than their GDPs. The third feature was that the rate of increase
in the indebtedness of these borrowers was two to three times the interest rate on the loans,
which meant that money available to the borrowers from new loans was several times larger
than the interest payments on their outstanding loans. The borrowers were in a “sweet spot”
because all the money they needed to pay the interest on their outstanding loans came from
the lenders in the form of new loans.
This pattern of cash flows could not continue without limit, at some stage the lenders would
reduce the rate at which they would extend more credit to the borrowers, who then would
have to find a new source of money for the scheduled interest payments. When the flow of
money from the lender to the borrower slowed, the borrower’s currency would depreciate.
The implications of changes in cross-border money flows on the price of a currency and on
the prices of assets in a country can be illustrated by reviewing the experience of Iceland
between 2002 and 2008. Iceland had a modest current account surplus in 2002. Then the
foreign demand for Icelandic securities increased sharply, more or less at the same time as
the foreign demand for securities denominated in the U.S. dollar and the British pound
increased. The Icelandic krona appreciated in response to the increase in the foreign
demand for Icelandic securities; Iceland’s capital account surplus and its current account
deficit increased. Moreover the prices of the Icelandic securities increased in response to the
purchases by foreign buyers.
The Icelandic sellers of the securities denominated in the krona then had to decide whether
to use the money from the sale of these securities to buy other Icelandic securities from other
Icelandic investors or to buy consumption goods – they could do both and they did both. To
the extent that they purchased other Icelandic securities, the sellers had the same problem.
In effect the initial purchases of Icelandic securities triggered a series of purchases by those
who sold the securities, who used nearly all of their receipts to buy other Icelandic securities.
The prices of these securities, and the financial wealth of Icelandic households increased.
Their consumption spending increased, which stimulated an economic boom; Iceland’s trade
deficit increased sharply.
This series of transactions in Icelandic securities was an integral part of the adjustment
process whereby the increase in the Icelandic imports and in the country’s current account
deficit matched the increase in the country’s capital account surplus. The intermediate
argument was that Icelandic household wealth increased as the prices of the securities
owned by the borrowers increased which led to an increase in household consumption.
Iceland experienced two bubbles at the same time, one in the currency market and a second
in its asset markets, both for stocks and for bonds. When the foreign demand for Icelandic
krona securities slackened, it was inevitable that the krona would depreciate; at the same
time, it was likely that the prices of Icelandic assets would decline in response to the increase
in domestic interest rates, since some investors would become distress sellers.
The bubble in the U.S. housing market between 2002 and 2007 was similar to the events in
Iceland, although on a much more massive scale. An increase in the foreign demand for U.S.
dollar securities lead to an appreciation of the U.S. dollar (although it mostly dampened a

8 BIS Papers No 58
depreciation that otherwise would have occurred). The U.S current account deficit increased,
and the ratio of the U.S. current account deficit to U.S. GDP increased by 3–4 percentage
points. U.S. real estate prices surged. And then real estate prices started to decline at the
beginning of 2007, much as they did in Britain, Ireland, Iceland, and Spain.
Because the rate of increase of the indebtedness of the borrowers in these countries was so
much greater than their incomes, it was inevitable that at some stage the lenders would
become more cautious about increasing their loans. Similarly, because the rate of increase in
the external indebtedness of these countries was so much higher than the increase in their
GDPs, it was inevitable that lenders would become more cautious. When the flow of money
to these countries slackened, it was inevitable that their currencies would depreciate. The
initial depreciation by itself might induce other lenders to become more cautious. The
combination of the decrease in the pace of money inflows and the depreciation of the
currencies would lead to a decline in asset prices. Economic growth would slow, as
households increased their saving in response to the decline in financial wealth.
Hence the increase in the values of the currencies and the increases in asset prices in the
countries were not sustainable. And the increases in these prices can be considered bubbles
because they were not sustainable.

III. Monetary shocks, structural shocks, and changes in currency

values
Large changes in currency values relative to the values based on differences in national
inflation rates may reflect more structural shocks such as sharp increases and declines in oil
prices or more monetary shocks including changes in inflation and interest rates. Large
variations in the prices of currencies can be related to an early literature on currency
movements when currencies are not pegged. When Ragnar Nurkse, in his classic
International Currency Experience, suggested that speculation in the currency market was
destabilizing, he probably was referring to the French experience between 1924 and 1926.
Milton Friedman responded in “The Case for Floating Exchange Rates” that if speculation
had been destabilizing, it would have been unprofitable, and those speculators that had lost
money would leave the market. Nurkse’s statement centred on the empirical properties of
time series of changes in currency values and changes in national price levels. Friedman’s
statement was derived from “first principles” and hence was not a direct refutation of
Nurkse’s observation.
Changes in currency values in the first half of the 1920s were affected by two different
factors. At the beginning of the First World War, most governments suspended the
convertibility of their national currencies into gold; their currencies depreciated relative to the
U.S. dollar in part because their money supplies had increased more rapidly than the U.S.
price level. The view in the immediate post-war period was that currencies would again be
attached to their prewar parities. Initially investors accumulated German marks in anticipation
that the mark would appreciate toward its prewar parity, and then when they reversed their
anticipations and sold their marks, the currency depreciated. When the mark collapsed in
1923, the cliché was that speculative pressure was deflected to the French franc. There were
two “bear raids” on the French franc, one in 1924 and the second in 1926.
Two meanings can be attached to Nurkse’s use of the term “destabilizing speculation” – one
is that speculators caused the amplitude of movements in the currency values attributable to
shocks in the goods market to be larger than they would have been in their absence; in this
sense the speculators are like “tape watchers” or “momentum traders” who followed the
cliché that “the trend is your friend”. The second meaning attached to this term is that the
transactions of speculators would induce changes in domestic income and employment by
their impacts on the trade balance.

BIS Papers No 58 9
The logic is that if there are only two groups of participants – goods market traders and the
speculators – in the currency market, the transactions of one group cause the prices of
currencies to deviate from their long-run average prices, while the transactions of the second
group will limit these deviations. Both the goods market traders and the speculators are
responding to different shocks and different profit opportunities.
Both the goods market traders and the speculators will be impacted by various shocks.
Shocks can be grouped as either structural or monetary; structural shocks include sharp
changes in the prices of oil and other commodities, the loss of an export market, a domestic
crop failure. Monetary shocks include changes in interest rates and changes in the
anticipated inflation rate. If a shock in the form of an increase in the price of imports – say an
oil price shock – might lead to a depreciation of the currency – and if speculators believe the
shock is temporary, they may buy the currency and limit the depreciation. In contrast if
foreign interest rates increase, speculators may move money to the foreign centre which will
cause domestic currency to depreciate. Domestically produced goods will become more
competitive in both the domestic market and in the foreign market, and the increase in the
domestic trade surplus will limit the depreciation of the domestic currency.
Neither Nurkse nor Friedman identified who the speculators were – whether they were
banks, trading firms, brokerage firms, insurance companies, or individual investors.
The debate between Nurkse and Friedman was never joined because they differed in the
source of shocks. Nurkse implicitly suggested that the shocks originated in the money
market, while Friedman believed that the shocks originated in the goods market. An
extension of this distinction is whether the goods market shocks are more frequent than
money market shocks, and the frequency and severity of each type of shock.
Money market shocks and goods market shocks have different impacts on the combination
of changes in the trade balance and the value of the currency. For example, assume that
there is a goods market shock in the form of an increase in the price of oil; the country’s oil
import bill increases and its currency depreciates so that exports will increase to match the
increase in the imports. Speculators may buy the currency and limit the depreciation, which
is the scenario envisioned by Friedman. In contrast, assume a money market shock in the
form of an increase in interest rates in foreign country; investors move money to the foreign
country; the foreign currency appreciates or what is the same thing, the domestic currency
depreciates. The country’s trade surplus increases, which provides goods market traders
with the opportunities to arbitrage. The increase in the trade surplus leads to a higher level of
domestic income and perhaps an increase in upward pressure on the price level. This is the
type of shock envisioned by Nurkse.
The shortcoming of the Nurkse-Friedman debate is that it does not deal with the stylized fact
that large changes in the prices of currencies have been associated with significant changes
in the cross-border movements of money. Again, returning to Iceland, the sharp appreciation
of the currency was associated with a massive flow of money to Iceland; Iceland’s current
account deficit increased as its capital account surplus increased. The money flow to Iceland
might be viewed as consistent with a broad interpretation of Nurkse’s view of destabilizing
speculation, although Nurkse appears to have been concerned that money flows from a
country might put upward pressure on the price level because of the increase in the trade
surplus might lead to excess demand. The Iceland experience is one in which the money
flows to a country led to increases in consumption spending and investment spending as
result of the positive wealth effect.

IV. Carry trade investors and currency movements

One feature of the last 40 years has been large cross-border money movements and the
variations in these flows, which is evident from the changes in the trade balances of

10 BIS Papers No 58
individual countries. During the early 1990s, Mexico’s current account deficit increased to 6%
of its GDP; then the peso depreciated sharply at the beginning of 1995, and Mexico
developed a current account surplus that was 4% of its GDP. Iceland went from a current
account deficit that was more than 20% of its GDP to a current account balance. Similarly
there were large changes in the current account balances of many other countries, although
few were as dramatic as those for Mexico and Iceland.
The shocks that led to these changes in the cross-border money flows originated in the
financial markets; these shocks included changes in interest rates and in anticipated inflation
rates. These shocks have induced changes in cross-border money flows that led to changes
in the values of currencies. (If the shocks that had led to an increase in the current and trade
deficits had originated in the goods market, the currency would have depreciated as the
trade deficit increased.)
These cross-border movements of money are initiated by “carry trade investors”, who
acquire foreign securities with the intent to own them for extended periods. Carry trade
investors are like arbitragers in financial markets, they seek to profit from the difference in
interest rates on comparable securities denominated in different currencies; they realize that
they may incur losses from the depreciation of the foreign currencies – but obviously they
believe that the values in the interest rate differential term is larger than the value in the
currency term. Carry trade investors do not believe that “all the information is in the price”;
instead they believe that the interest rate differential overstates the anticipated or likely
change in the value of the currency during the term to maturity of their investments. The
difference in the two streams of interest income can be considered the revenues for carry
trade investors, and the anticipated change in the price of the currency is the cost.
At times the interest rate term and the currency term are additive. For example, assume that
interest rates in a country increase, perhaps because its central bank has adopted a more
contractionary monetary policy. The carry trade investors move money to the country, and its
currency appreciates. The carry trade investors profit both from the additional interest income
and the gain from the appreciation of the currency. In periods of two or three years, the
additional income from the appreciation of the currency may be larger than from the
difference in interest rates. In the long run, however, the interest rate differential and the
currency term are offsetting, and the currencies of the countries identified with higher interest
rates depreciate.
Carry trade transactions come in 57 varieties. Mrs. Watanabe took the money from one of
her yen deposits in Tokyo to acquire a U.S. dollar annuity from AIG. Citibank used funds
obtained from the sale of dollar deposits in London to fund its U.S. dollar loans to the
Government of Mexico. Nomura acquired dollars in the offshore market to buy the IOUs of
the Landsbanki of Iceland. Individuals in Reykjavik financed the purchase of autos by signing
IOUs denominated in the Japanese yen and the Swiss franc because interest rates were
lower than those on the Icelandic kronor. Similarly individuals in Poland have financed their
purchases of homes by borrowing Swiss francs, and individuals in Australia have borrowed
the yen to finance their home purchases.
Carry trade investors who bought Icelandic krona IOUs in 2002 and 2003 and 2004 profited
from the appreciation of the krona as well as from the excess of interest rates krona
securities over the interest rates on U.S. dollar securities. Similarly Icelandic borrowers who
sold IOUs denominated in the U.S. dollar or the euro profited from the saving in interest
costs.
The efficient market view is that the cross-border money flows surge whenever there is new
“news”; the price of the currency changes immediately until there is no longer an excess
return attached to the cross-border movement of money. However, the appreciation of the
currency of a country that experience an inflow of money is slowed or dampened because of
transactions in the goods market; as the currency appreciates, the opportunities for goods
market arbitrage increase. Hence the anticipated excess return remains.

BIS Papers No 58 11
Carry trade investors can be distinguished from the Friedman’s speculators and from some
of the speculators noted by Nurkse. Friedman’s speculators trade currencies for banks and
other financial firms; they seek to profit from changes in the prices of currencies. These
speculators hold their positions for a relatively short time – a few minutes, a few hours, a few
days. A few of these traders are market makers, many are day traders, and a few are
proprietary traders. The hallmark of these traders as a group is that their anticipated
revenues are from changes in the prices of currencies, while their costs are the difference
between domestic and foreign interest rates. The market makers provide both bid and offer
for transactions of a standard size; while it may seem that they are providing a service, the
information in the order flow is of high value. This group makes its money from the immense
volume of transactions – and they make money regardless of whether their domestic
currency appreciates or depreciates.
Consider the returns to the goods market traders, the carry trade investors, and speculators.
The goods market traders profit from the arbitrage opportunities presented by the divergence
in national price levels created by changes in the values of currencies, the greater the
overshooting and undershooting, the larger their profit opportunities. Their trade transactions
require that they buy and sell currencies as an intermediate transaction prior to the payment
for the purchase of goods, and they may incur a cost for these transactions. Similarly, carry
trade investors must buy foreign currencies before they can buy foreign securities; they incur
a cost. The speculators profit from their market making activities, and from changes in
currency values.
Casual empiricism suggests that the trading revenues of the major international banks have
increased sharply since 1980 and perhaps from the early 1970s. Some of these revenues
are from trading securities and some are from trading currencies and some from trading
commodities. The volume of currency transactions is many times larger than the volume of
trade or the volume of trade and investment. Most of the transactions of the speculators are
with other speculators. Moreover developments in technology and competition have led to
declines in the bid-ask spreads. The increase in the revenues seem larger than the amount
that can be attributed to the bid-ask spreads; the implication is that a large share of these
profits must have come from revaluation gains on their positions.
How can the currency traders and the carry market traders both profit at more or less the
same time? Obviously they can’t in terms of cash flows – the currency traders take money off
the table, minute by minute and hour by hour, and stuff their profits in a sock. Some of that
money may be placed on the table by the goods market traders; the costs of using the
currency market are like transport costs. The carry trade investors are indifferent because
they are continually re-valuing their positions on the basis of current prices. The carry trade
investors earn money for an extended period – until the bubble is pricked, the currencies
depreciate sharply, and firms and banks fail.

V. The sources of financial instability

The much greater variability in cross-border money flows since the early 1970s can be
attributed to a larger number of shocks. Some of these shocks might be structural, including
sharp changes in the price of oil, dramatic increases or decreases in the rates of return on a
particular group of assets, and changes in financial regulations. Some of these shocks might
be monetary, including significant changes in anticipated inflation rates, or in interest rates,
or in currency values.
One of the principal arguments advanced by the proponents of a floating currency
arrangement is that in the absence of parities, national central banks would be able to follow
“independent” national monetary policies; they would no longer be constrained by the need
to minimize their payments imbalance at their established parities. When currencies were

12 BIS Papers No 58
pegged, national inflation rates were closely linked because countries could not finance large
trade deficits. Because currencies are no longer attached to parities, national inflation rates
are more likely to differ – and the larger possible difference in these rates means that the
scope for changes in these differences is much larger. When interest rates change relative to
the inflation rate, carry trade investors may recognize an exceptional profit opportunity.
The necessary condition for a significant increase in cross-border money flows is a shock
that leads to an increase in the return on securities available in a particular country, or a
shock that leads to a relaxation of restrictions that previously had restricted investor
purchases of certain securities, or a change in controls on cross-border money movements.
One of the two sufficient conditions for an increase in cross-border flows involves the
willingness of carry trade investors to take on the risks associated with the cross-border
movements of money, and the other is a pool of money that these investors can access. The
large payments imbalances since the mid-1960s have led to a surge in international reserve
assets which is an enormous pool of accessible money. (Central banks are more likely than
others to hold funds in the offshore deposits.)
That there have been four waves of credit and asset price bubbles in 40 years suggests that
there may be a connection between several of these waves – more precisely, between the
implosion of one wave of bubbles, and the formation on another wave. That three, four, or
more countries have been involved in each of the several waves of credit bubbles suggests a
common cause. The shock that preceded the first wave of credit bubbles in the 1970s was a
surge in the world inflation rate that led to significant increases in commodity prices and in
the anticipated rates of growth of GDP in the countries that produce primary commodities.
The shock that preceded the bubble in Japanese real estate and stocks was the decline in
interest rates on US dollar securities, which lead to an increase in money flows toward Tokyo
and a tendency toward the appreciation of the yen. The Japanese authorities relaxed
restrictions on bank loans for real estate. The shock that led to the surge in money flows to
the Nordic countries was the relaxation of restriction that limited the ability of banks
headquartered in these countries to source for money in the offshore market. Several shocks
contributed to the increase in money flows to the emerging market countries in the early
1990s, including the appreciation of the yen, and the liberalization of restrictions that had
limited the ability of banks headquartered in these countries to source for money in foreign
markets. The sharp depreciation of the Thai baht and the currencies of other emerging
market countries in mid-1997 contributed significantly to the bubble in U.S. stocks. The shock
that led to a rapid increase in the supply of credit available for real estate in the United
States, Britain and other countries was the surge in China’s trade surplus.
One feature of these shocks is that the adjustment process induced by the flow of carry trade
money to these countries leads to increases in their rates of growth of GDP as a result of
increases in consumption spending and investment spending in response to higher levels of
household wealth. It is as if there is a feedback loop; an initial shock leads to increases in
cross-border money flows, and then the increases in wealth induced by these flows lead to
the more rapid growth of GDP – which induces carry trade investors to move more money
abroad.
Although the cross-border money movement is induced by the increase in the rates of return,
the primary impact of this movement is to finance higher levels of consumption spending –
the story is that the increase in wealth induced by the money inflow leads to a decline in
domestic saving as consumption spending increases.
Because currencies are not pegged, changes in national monetary policies lead to changes
in anticipated values for currencies and induce changes in cross-border money flows; a
move to more contractionary monetary policies may attract money because of the higher
inflation-adjusted rate of return and the downward revision in the anticipated inflation rate. In
this way the appreciation of the currency in response to the adoption of a more
contractionary policy may be like a self-fulfilling prophecy.

BIS Papers No 58 13
Once a shock leads to an increase in the anticipated returns on securities denominated in a
group of currencies or a reduction in the restrictions on the cross-border movement of
money, the conditions are appropriate for the formation of a bubble. The money is there, and
the initial movement of money across national border is likely to enhance the anticipated
returns on the money market arbitrage. No one foresees the inevitable crunch because the
rate of increase in indebtedness is not sustainable.

VI. Managing wealth in turbulent times

Keynes wrote several articles in the mid-1920s that questioned whether British foreign
investment “paid” – whether British GDP was higher because London-based firms increased
their investments abroad and because households bought foreign securities. His argument
centred on the distinctions between private rates of return and social rates of return. One
was that when the British owned the New York subway and the subway went bankrupt, the
equity investors lost all their money and the American bondholders became the owners. (If
the Americans had owned the equity and the British had owned the bonds, the conclusion
would have differed.) His second point was that the private rate of return to the owners of the
foreign investments was higher than the social rate of return to Britain as a country, and for
two reasons – one is that the U.S. government rather than the British government would
collect the taxes on the investment. His second was that the capital stock available to British
workers was smaller because of the capital outflow, which involved a comparison between
the decline in domestically produced GDP and the return to Britain on its foreign investment.
His third argument was that anticipated rates of return by the first to invest abroad were
lowered by others who followed them and increased their investments.
The textbook answer to the question whether foreign investment pays is an extension of the
argument about the gains from trade; both the capital-exporting country and the capital-
importing country gain, and a lot of country-specific factors determine the shares of the gain
to each country. This textbook answer is in a “real economy” – one without money.
There are three primary characteristics of cross-border money flows when the currencies
have been floating. One is that the risk of these flows has increased because of the much
larger range of movement in the prices of currencies. A second is that the economic booms
in the money-receiving countries are associated primarily with increases in consumption
spending. The third is that there is a likelihood of financial failure when the bubble implodes.
The implicit assumption was that the increases in the flows of money to a country would be
associated with an increase in the investment in the country. One of the unique features of
cross-border money flows since the early 1970s is that they have been associated primarily
with increases in household consumption or an increase in the fiscal deficits in the countries
that have experienced increases in money inflows. For example, the bank loans to Mexico,
Brazil, Argentina, and other developing countries in the 1970s primarily financed increases in
the fiscal deficits of the governments in these countries and increases in the capital
expenditures of government-owned firms. The surge in money flows to the United States
after 2002 contributed to an increase in the supply of credit available for real estate; a
significant part of this credit financed mortgage equity withdrawals. One feature of this period
was that the household savings rate declined sharply and was not significantly different from
zero; in effect the increase in the supply of foreign saving available to Americans induced a
set of market developments that led to a decline in the household saving rate. The surge in
money flows to Iceland enabled the domestic banks to finance large loans to firms that
wanted to invest abroad; increases in household consumption accounted for 90% of the
increase in the current account deficit.
The implication is that cross-border money flows have not been primarily associated with
increases in investment and in the rate of economic growth in the countries that have

14 BIS Papers No 58
received the money inflows. Still it may be that some countries have been able to achieve
somewhat higher rates of economic growth, since they are no longer obliged to maintain
parities for their currencies. Nevertheless there appears to have been a disproportionate
increase in the risk relative to the increase in return from cross-border investments. Hence
there has been a significant reduction in the “all-in” return available to the carry trade
investors from the sum of the additional interest income and the currency losses and gains
relative to the risk of revaluation losses and credit defaults.
These statements about increases in return and increases in risk follow from first principles.
Obviously investors who get the timing right – who buy low and sell high and repatriate their
money before the bubble implodes – will have a much higher rate of return. A few investors
can pursue this strategy, however if many were to produce this strategy, the currency would
depreciate and the bubbles would implode.
Consider the market in junk bonds otherwise known as high-yield bonds as a metaphor.
Promises were made about a large supply of “free lunches” on junk bonds or high-yield
bonds by Michael Milken in the 1980s, who convinced investors that there was market
inefficiency because the rating agencies did not rank these bonds, and hence there was an
excess return on these bonds. The excess return persisted until the market in these bonds
collapsed, which occurred soon after the savings and loan associations and insurance
companies that were managed by Milken’s buddies were no longer the “buyers of last resort”
for these bonds. Subsequent studies have shown that the additional interest income was not
sufficiently large relative to the credit losses that investors incurred.
The same point is made by considering the appropriate premium for selling flood insurance
in New Orleans. How should the underwriters set the appropriate premium – high enough to
cope with the losses due to the floods that occurs every 10th year? But then the premiums
will not be large enough to reimburse the losses due to the exceptional flood that occurs
every 50th year. If the premiums are set to cover the losses from the more frequent, less
severe floods, they may be too low to cope with the more severe floods. The most severe
flood may lie in the future.
The dominant implication of the increase in risk relative to return on cross-border carry trade
investments is that market participants should devote more attention to determining the
currency composition of their assets and liabilities that minimizes their exposure to loss from
changes in currency values. The managers of the international reserves of central banks are
in a position much like multinationals and other firms involved in international trade; they first
need to determine the currency composition of their reserves that minimizes their exposure
to gain and loss from changes in the price of currencies, and they then need to determine
whether the anticipated interest income from maintaining a different composition is
worthwhile in terms of the exposure to loss from changes in currency values.

VII. Summary and conclusion

During the 1960s the dominant concern of those involved in international finance was that
the shortage of international reserve assets would lead to deflationary pressures on national
economies. The waves of credit bubbles since the 1970s reflect in part the surge in the
supply of international reserve assets. There have been four waves of financial crisis since
the early 1980s; each of these waves has led to the failures of large number of banks in
three, four, or more countries at the same time as many borrowers defaulted on their loans.
The first wave occurred in the early 1980s and involved the governments of and government-
owned firms in Mexico, Brazil, and 10 other developing countries. The inability or
unwillingness of these borrowers to repay in a timely way led to a depreciation of their
currencies, and domestic firms with liabilities denominated in the U.S. dollar failed because
of the large revaluation losses on these loans, which contributed to the failure of the

BIS Papers No 58 15
domestic banks. The second wave of financial crises occurred in the early 1990s, when
banks and other financial institutions failed in Japan and three of the Nordic countries –
Finland, Norway, and Sweden. The Asian Financial Crisis that began in mid-1997 was the
third wave; the financial turmoil in Mexico at the end of 1994 was a prelude to the collapse in
values in Asia. The fourth wave of crises that began in 2008 resulted from the sharp decline
in real estate prices in the United States, Britain, Ireland, Iceland, and other countries; the
decline in the value of mortgages and mortgage-related securities led to large losses by
mortgage bankers, investment banks, and commercial banks, and other lenders.
Each of these waves of crises followed from increases in the indebtedness of a group of
borrowers at rates of 20–30% a year for three, four, or more years. Each of these four waves
of credit bubbles involved the cross-border movement of money; the principal exception was
that the rapid increases in real estate prices in Japan followed from the rapid growth in the
domestic supplies of money and credit. In contrast, the rapid growth in the credit in the
Nordic countries in the late 1980s resulted from money inflows as domestic banks sourced
money in the offshore market.
One central aspect of the period since the early 1970s has been that the range of
movements in the prices of currencies has been much larger than the range that would have
been forecast based on contemporary or lagged differences in national inflation rates. These
very large changes in the value of currencies have resulted from changes in cross-border
money flows; increases in the money flows to countries have led to extensive appreciations
of their currencies.
The increases in cross-border money flows to countries have two immediate effects – their
currencies appreciate and the asset prices in these countries increase in response to
purchases by those who had moved money to these countries. The increases in asset prices
were an integral part of the adjustment process; asset prices and household wealth
increased until the increase in consumption spending and in imports led to an increase in the
current account deficit that matched the autonomous increases in money inflows. The
counterpart of the increase in the money flows to the country was that its savings rate
declines as household consumption spending increased.
The increase in the indebtedness of the borrowers in these several waves was several times
higher than the increase their incomes and GDPs; similarly the increase in the indebtedness
was several times higher than the interest rate on the indebtedness. As long as the
indebtedness of the borrowers was higher than the interest payments, the borrowers were in
the sweet spot, since all the money needed to pay the interest on the indebtedness came
from the lenders in the form of new loans. But it was inevitable that the lenders would reduce
the rate of growth of new loans, which automatically would lead to a depreciation of the
currencies of these countries.
The minimum requirement to generate a bubble is that the rate of the flow of money to a
country is too high to be sustained; when the rate slackens, it is inevitable that the currency
depreciates and interest rates increase, in part in response to the decline in the supply of
credit.
Three factors have contributed to the four waves of bubbles. One is that since the early
1970s, there has been a large pool of “idle money” parked with the international banks,
available to be tapped by those who have concluded that they enhance their own returns by
taking on credit risk or currency risk. This pool has been inflated by the large payments
imbalances since the late 1960s. The second is that there have been a series of shocks at
national borders, which either have increased the anticipated returns available on securities
in certain countries or increased the scope for cross-border investment by reducing
restrictions at the border. The third is that the early stages of cross-border money flows
enhance the returns in countries that receive the money, so that the flows are self-justifying –
at least for a while.

16 BIS Papers No 58
Bibliography
Aliber, Robert Z., The New International Money Game, Palgrave, Basingstoke, sixth edition,
2005.
Federal Reserve Bank of Boston, The International Adjustment Mechanism, conference
series No. 2, Proceedings of a Conference held in October 1969.
Friedman, Milton, and Robert V. Roosa, The Balance of Payments: Free Versus Fixed
Exchange Rates, American Enterprise Institute for Public Policy Research, Washington,
1967.
Friedman, Milton, “The Case for Floating Exchange Rates”, reprinted in Essays in Positive
Economics, University of Chicago Press.
International Monetary Fund, International Financial Statistics, Washington.
Kemp, Murray C. “Foreign Investment and the National Advantage – the Lending Country”,
Economic Record, XXVIII, March 1961. reprinted in The Pure Theory of International Trade,
Prentice Hall, Englewood Cliffs, 1964.
Kindleberger, Charles P. and Robert Z. Aliber, Manias, Panics, and Crashes, Palgrave, fifth
edition 2005.
Nurkse, Ragnar, International Currency Experience: Lessons of the Inter-War Period,
League of Nations, Princeton University Press, 1944.
Sohmen, Egon, Flexible Exchange Rates, revised edition, Chicago, 1969.
Triffin, Robert, Gold and the Dollar Crisis, Yale University Press, New Haven, 1960
Sohmen, Egon, Flexible Exchange Rates, revised edition, Chicago, 1969.
Triffin, Robert, Gold and the Dollar Crisis, Yale University Press, New Haven, 1960.

BIS Papers No 58 17