Financial Globalization

JAMES TOBINCowles Foundation, Yale Uiversity, New Haven, CT, USA

Summary. Ð The paper studies how global ®nancial markets can be restructured so as to be morestable while also providing some national economic autonomy. An impossibility theorem sets thestage for the analysis of reform alternatives. Then, the relative merits of alternative exchange rateregimes and of di�erent degrees of capital-market-¯ow regulation are considered. The paperconcludes by opting for a combination of : ¯oating exchange rates; slowing down and taxingforeign capital in¯ows; and greater national economic sovereignty. Ó 2000 Elsevier Science Ltd.All rights reserved.

1. INTRODUCTION

Globalization of ®nancial markets a�ectsassets and debtsÐsecurities, bank loans anddeposits, titles to land and physical capital. Letme start by noting that trades in these assetsand debts are much easier to globalize thantrades in commodities and labor. Indeed, theirglobalization has progressed most rapidly.Nothing is involved in ®nancial transactionsbeyond exchanging pieces of paper or makingentries in electronic ledgers. The communica-tions revolution makes these transactions easy,fast and cheap. No physical frontiers have to becrossed by ®nancial assets. The only barriers to®nancial transactions are national regulations.As these have been liberalized in country aftercountry, international ®nancial ¯ows have¯ooded national securities markets and inun-dated banking systems all over the world. These¯ows could be the vehicles by which savings inthe advanced capitalist democracies are chan-neled into productive capital investments in thedeveloping countries of Asia, Africa, and LatinAmerica. Or they could be causes of currencycrises, recessions and depressions, unemploy-ment and deprivation in those countries. Orboth.

The economic rationale for internationaliza-tion of asset markets is that it can assist themovement of productive capital from wealthydeveloped economies to poorer developingcountries. But what matter are the net ¯ows ofcapital, not the gross volume of transactions.The emerging economies of East Asia, as wellas some in Latin America and Eastern Europe,are bene®ciaries of foreign business invest-ments. But much of their capital in¯ows havetaken the form of short-term loans of hard

currencies from banks in ®nancial centers suchas Tokyo, New York and Frankfurt to banks inKorea, Thailand and Indonesia. The periods ofincreasing loans were also periods of increasinggrowth and investment in countries receivingthe loans; however, they were also, necessarily,accompanied by increasing overvaluation ofnational currencies and growing de®cits oncurrent account. Crises came when the lenders,viewing the growing de®cits, became distrustfuland refused to renew the loans.

Are the bene®ts of ®nancial globalizationlikely to outweigh the costs? One hint can beobtained by comparing the magnitudes of grossand net capital transfers. Although developingcountries have increasingly bene®ted fromin¯ows of capital, the investments that havepropelled their growth have been mainly due totheir own internal savings. Capital ¯ows fromthe world economic core to the periphery, only$150 billion a year in the 1990s, have been lessthan 15% of their investment and less than 5%of the savings of the developed capitalisteconomies. These shares are much smaller thancomparable ®gures before 1914, when theywere both close to 50%. By contrast, theworldwide gross volume of foreign exchangetransactions is mind-boggling: US$1.5 trillionper business day and growing rapidly. Nine-tenths of these transactions are reversed withina week, 40% within a day. Clearly, most ofthese foreign exchange transactions are specu-lative. These numbers show that the grossvolume of foreign capital ¯ows dwarfs the netcapital transfers.

It is only the net capital transfers that carrythe economic bene®ts globalization is advertisedto bring. It is the gross, speculative transactionswhich carry with them the destabilizing e�ects

World Development Vol. 28, No. 6, pp. 1101±1104, 2000Ó 2000 Elsevier Science Ltd. All rights reserved

Printed in Great Britain0305-750X/00/$ - see front matter

PII: S0305-750X(00)00008-5www.elsevier.com/locate/worlddev

1101

leading to ®nancial crises and severe realeconomic downturns.

The 1990s have been a decade of distur-bances in international ®nance, beginning inEurope in 1992, followed by Mexico in 1994±95, climaxed by East Asia in 1997±98, Russiaand Brazil in 1998. Is the problem that liber-alization in developing and transition econo-mies is still incomplete? Or has it gone too far?Is there a way of restructuring global ®nancialsystems so that the bene®ts of long-term foreigninvestment will be preserved while speculative¯ows will be discouraged. That is the big debatetoday.

2. TYPES OF REFORMS OF FINANCIALSYSTEMS

Most observers, West and East, public andprivate, in governments and internationalinstitutions, in banks and businesses, in bigcountries and small, now agree that ®nancialglobalization went too far too fast.

(a) An impossibility theorem

The ground rules for reform are posed by a``trilemma,'' which states an impossibilitytheorem the like of which international econo-mists are quite fond: A country can maintainno more than two of the following threeconditions: (i) A ®xed rate of exchange betweenits currency and other currencies. (ii) Unregu-lated convertibility of its currency and foreigncurrencies. Or (iii) a national monetary policycapable of achieving domestic macroeconomicor development objectives.

For example, consider a government andcentral bank that wish to reduce unemploymentby raising aggregate demand for the goods andservices its economy produces. This typicallyrequires cutting the interest rates facingdomestic businesses and households andmaking its products more competitive in worldtrade. But this is not possible if the exchangerate is ®xed while arbitrages across currenciesare unimpeded. The country's central bank willthen be unable to reduce interest rates belowthose available elsewhere in the world, partic-ularly in big centers. For this will trigger capitalout¯ows, precipitate a ®nancial crisis and endup increasing rather than lowering unemploy-ment. Maybe the government can empower itscentral bank by giving up condition (b) andimposing direct controls over movements of

funds across the exchanges. Alternatively, thegovernment could sacri®ce condition (a) and letits currency ¯oat in the market to a lower levelat which activity and employment, especially inexport industries, would be greater, while lowerlocal interest rates would promote investmentand raise international competitiveness. Eitheralternative would be tenable.

In the wake of WWII, it was apparent thatthe economies of Europe and Asia were in noposition to make their currencies whollyconvertible. The articles of the InternationalMonetary Fund adopted at Bretton Woods didnot, and still do not, require that of itsmembers. What they do require is ``currentaccount convertibility,'' namely that foreignersbe free to convert any of a country's currencythey earn in trade. ``Capital account convert-ibility'', which would allow any holder of acurrency, resident or nonresident, to buyforeign currency assets, was put o� to theinde®nite future. Under the Marshall Plan1948±51 the United States encouraged Euro-pean countries to set up a multilateral clearingsystem for their currencies, while restrictingconversions into dollars. Currency exchangerestrictions in Western Europe were not whollyabandoned until the mid-1980s. Today,however, the world ®nancial powers, privateand public, are impatiently pushing developingcountries and transition economies toward fullconvertibility.

Fixed exchange rates, adjustable pegs to hardcurrencies, are the prevailing exchange rateregimes among developing economies,``emerging,'' ``transition'' and others. Typically,the exchange rate regimes they have adoptedare ``managed crawling pegs,'' which do allowfor some exchange rate ¯exibility. Markets areallowed to move the exchange rate within aspeci®ed band. The entire band is itself movedfrom day to day by an announced percentage,usually designed to depreciate the currency tocompensate for a local in¯ation trend in excessof the in¯ation trend in the hard-currency'seconomy. For example, before 1996, the centralparity and the band of Brazil's real rose at amonthly rate of 0.7%. But, if, under marketattack, the price of a dollar in terms of realsshould rise to its upper limit (depreciation ofthe real) then the central bank would have touse its dollar reserves to redeem reals just as if itwere defending a simple ®xed peg. Given thesheer size of transactions in foreign exchange itwould soon run out of foreign exchangereserves with which to defend the real.

WORLD DEVELOPMENT1102

An attack on a currency is like a run on abank. A depositor worried about the ability ofa bank to redeem deposits will want to ask forher cash before the bank runs out, and anydepositor worried about what other depositorswill think and do will act the same way even ifshe thinks the bank is solvent. A country on a®xed exchange rate is like a bank; its holdingsof hard currency reserves are like the bank'scash; and the local currency assets outstandingare like the public's deposits in a bank. Thesame instability and vulnerability apply in bothcases. For a domestic banking system, depositinsurance is an e�ective protection against runs,country's central bank acts as a ``lender of lastresort'' to provide liquidity to banks underattack. The analogous institutions do not existon an international scale, to protect currenciesagainst runs. Hence the incompatibility of ®xedexchange rates, adjustable pegs, or crawlingpegs to hard currency with full capital-accountconvertibility.

The recent epidemic of currency crises makesit unmistakably clear that ®xed but adjustableexchange rates are a bad idea. The only viableregimes in our increasingly globalized ®nancialworld are ¯oating rates, on the one hand, andirretrievably ®xed rates, on the other.

(b) Floating rates

Floating rates have since 1973 worked for theBig Three currencies. They have ¯uctuated, butthere have been no crises. During 1995±97 theyen gradually and unobtrusively fell 50%against the dollar, but this decline never ratedheadlines or evening TV news. (One of thecauses of the yen's substantial decline wasJapan's recession and stagnation, a disaster forJapan itself and for its neighborsÐindeed aprincipal source of their currency crises andeconomic recessions. But this macroeconomicdisaster would have been worse if the yen/dollarrate had been ®xed.)

Floating rates would work for most curren-cies. They would forestall extreme crises. Ofcourse, exchange rates would go up and down,people would speculate on them, and often the¯uctuations would be unpleasant for the econ-omies a�ected. But the trauma of discreteexchange-rate regime change, default of solemno�cial promises, and the bandwagon momen-tum these events generate, would be avoided.Foreign lenders would be more careful if theyunderstood that exchange rates were not guar-anteed. Events that triggered the Asian crises

would have likewise pushed down thosecurrencies had they been ¯oating, but surelynot nearly as far as they plunged in the panickyfree-falls following the collapse of ®xed rates.Fixed rates are, after all, a hangover from thepre-globalized Bretton Woods system.

(c) Fixed exchange rates

At the other extreme is the alternative of®xing the national currency irretrievably to thedollar or some other hard currency standard.The trouble with this course is that, withglobalized ®nancial markets and no lender oflast resort, it surrenders monetary sovereignty.This is what the 11 European countries aredoing. They will no longer have their individualmonetary policies, or even discretionary ®scalpolicies. It remains to be seen whether politicaland economic advantages, comparable to thoseof the 200-year old monetary union of the USstates, can be quickly manufactured in Europe.

An individual country, developing or transi-tional, can tie itself tightly and permanently toa hard currency. Examples are Hong Kong andArgentina, which are e�ectively dollarized. Therule is that local currency outstanding must becovered 100% by the central bank's hardcurrency reserves. Indeed, the dollar may partlyor wholly replace local currency as unit ofaccount and means of payment. This is theessence of a ``currency board''Ðone wellenough endowed with reserves of the hardcurrency to convince the world of convertibil-ity, and convincingly determined to protectthose reserves. The idea is to sacri®ce everyother possible objective of monetary and ®scalpolicies to the defense of the exchange rate. Forexample, if it takes double- or triple-digitinterest rates to attract and hold enoughreserves, so be it, regardless of macroeconomicconsequences. Similarly, for unemploymentrates. In terms of the trilemma, the countrymeets condition (i) ®xed exchange rate, and (ii)convertibility. But it sacri®ces (iii) monetarysovereignty, thus forfeits all possibility ofcontrolling its own macroeconomic (anddevelopmental) fate.

In contrast, consider China. Like HongKong and Taiwan, China was evidentlyimmune from the ``contagious'' currency crisesthat began in East Asia in 1997. But the reasonfor the stability of the renminbi is quite di�er-ent. It is not currency-board austerity or anyother capitalist virtue. China allows no ``capitalaccount convertibility,'' only ``current account

FINANCIAL GLOBALIZATION 1103

convertibility,'' as had European countries inthe early days of Bretton Woods. In terms ofthe trilemma, violating condition (ii) (convert-ibility) enables China to maintain the other twoconditions, (i) ®xed exchange rate, and (iii)monetary sovereignty.

(d) Regulation

Some reforms are the responsibility of theborrowing countries. They need to develop theinstitutions that make ®nancial markets workin the developed world, banking regulation andsupervision, transparency requirements like theUS Securities and Exchange Commission,bankruptcy procedures. Transparency andbetter governance of domestic banking systems,will not, however, in and of itself su�ce toavoid ®nancial crises.

Since the country's international reserves areat stake, those governments should limit thehard currency exposures of banks and busi-nesses. They should feel free to slow downin¯ows of liquid capital, by devices such asextra reserve requirements on new foreigndeposits in their banks, used successfully inChile. They should stress import of capital inthe form of direct investment and equity. Asargued above, they should let their exchangerates ¯oat. In terms of the trilema, it would

make convertibility more expensive, thusreducing the pro®tability of (ii), therebyenabling the combination of either ®xedexchange rates or ¯oating ones with greaterautonomy of monetary policy possible.

I have proposed a system-wide internationalmeasure to slow down ¯ows of ``hot money,''without interfering signi®cantly with currencytransactions related to trade and productiveinvestment. This is a simple small tax onforeign exchange transactions, levied at anagreed common rate by all countries wheresuch transactions originate in signi®cantamount. The tax, perhaps only 0.1 or 0.2%means nothing for a round trip of a year ormore from one currency to another and back.But for one-week round trips it would beequivalent to a di�erence between interest ratesin the two markets of 10 or 20% per year, apalpable protection of monetary sovereignty.Alas, the lords of ®nance throughout the worldwill have none of the ``Tobin Tax.'' How wouldyou like to have a tax named after you?

The economic triad I personally favor wouldconsist of a combination of: ¯oating exchangerates; slowing down foreign capital in¯ows andmaking them more expensive; and greaternational economic sovereignty. While thiscombination would not entirely eliminatecrises, it would make them less frequent, lesscostly and of reduced virulence.

WORLD DEVELOPMENT1104