Developments in Global Portfolio - Harvard Universitysites.fas.harvard.edu/~css318a/handouts/Equities_and...Gloom Boom & Doom report, and is the author of The Great Money Illusion:

InvestingWorldwide VIII

Developmentsin GlobalPortfolio

Management

Investing Worldwide VIII: Developments in Global Portfolio Management

FrontMatter.frm Page 2 Wednesday, November 12, 1997 2:54 PM

An publication

InvestingWorldwide VIII

Developments in Global Portfolio

Management

February 26–28, 1997Bermuda

Sponsored by theAssociation for

Investment Management and Research

and theInternational Society of

Financial Analysts


©1997, Association for Investment Management and Research

All rights reserved. No part of this publication may be reproduced, stored in aretrieval system, or transmitted, in any form or by any means, electronic, mechanical,photocopying, recording, or otherwise, without prior written permission of thecopyright holder.

This publication is designed to provide accurate and authoritative information withregard to the subject matter covered. It is sold with the understanding that thepublisher is not engaged in rendering legal, accounting, or other professionalservices. If legal advice or other expert assistance is required, the services of acompetent professional should be sought.

ISBN 1-879087-93-6Printed in the United States of AmericaSeptember 1997

To obtain the AIMR Publications Catalog, contact:AIMR, P.O. Box 3668, Charlottesville, Virginia 22903, U.S.A.Phone 804-980-3668; Fax 804-980-9755; E-mail [email protected]

orvisit AIMR’s World Wide Web site at www.aimr.org

to view the AIMR publications list.

Editorial Staff

Katrina F. Sherrerd, CFASenior Vice President

Elizabeth A. CollinsEditor

Jaynee M. Dudley Production Manager

Fiona D. RussellAssistant Editor

Lois A. CarrierTypesetting/Layout

Christine P. MartinProduction Coordinator


Contents

ForewordKatrina F. Sherrerd, CFA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii

Biographies of Speakers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii

Political Risk in the World EconomiesMarvin Zonis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Techniques for Today’s Global Asset Allocation StrategiesD. Sykes Wilford . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

Global Equity Analysis: Country, Industry, or Company Selection?Bruno Bertocci . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

Benchmarks for Global PortfoliosJohn C. Stannard, CFA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Implementing Global Tactical Asset Allocation in Developed MarketsR. Charles Tschampion, CFA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

Impact of the European Monetary Union on European Bond Markets and Portfolios

Paul A. Abberley. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

Indexing Emerging MarketsSteven A. Schoenfeld . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

Political Analysis for Investing in the Emerging Asian MarketsRobert Lloyd George . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

Change and the Next Emerging MarketsMarc Faber . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

Fore&Bio.frm Page ii Wednesday, November 12, 1997 2:55 PM

Foreword

The Investing Worldwide seminars are AIMR’s foremost conference for seniormanagers and analysts who are responsible for managing, analyzing, and selectingsecurities for global portfolios. Investing Worldwide VIII: Developments in GlobalPortfolio Management, the most recent in the annual series, was held February 26–28,1997, in Bermuda. The focus of the seminar was on the critical decisions today’smanagers face as they carry out global portfolio management.

More and more managers are thinking globally and acting globally, but theopportunities and risks in global investing constantly shift. In addition to the usualrisks and potential rewards, global investors today must make their decisions inthe face of questions about the European Monetary Union, the avalanche ofprivatizations, an unprecedented entry into the arena of emerging markets fromEastern Europe, and worldwide demographic changes.

This proceedings contains the presentations given at the conference on generaltopics—techniques for global asset allocation, incorporating political analysis in theinvestment process, and selecting appropriate benchmarks. It also includes theworkshop presentations that addressed issues of importance in analyzing specifi-cally the developed markets, the emerging markets, or the pre-emerging markets.

We are grateful to Thomas L. Hansberger, CFA, of Hansberger Global Investors,who served as the moderator of the conference, and to all the authors ofpresentations in this proceedings: Paul A. Abberley, Lombard Odier InternationalPortfolio Management Ltd.; Bruno Bertocci, Stein Roe & Farnham; Marc Faber, MarcFaber Limited; Robert Lloyd George, Lloyd George Management Limited; StevenA. Schoenfeld, Barclay’s Global Investors; John C. Stannard, CFA, Frank RussellCompany; R. Charles Tschampion, CFA, General Motors Investment ManagementCorporation; D. Sykes Wilford, CDC Investment Management Corporation; andMarvin Zonis, University of Chicago and Marvin Zonis + Associates.

Katrina F. Sherrerd, CFASenior Vice PresidentEducational Products

ii ©Association for Investment Management and Research

Fore&Bio.frm Page iii Wednesday, November 12, 1997 2:55 PM

Biographies of Speakers

Paul A. Abberley is director and headof the Fixed Income Department atLombard Odier International PortfolioManagement Limited and a member ofthe Lombard Odier Group StrategyCommittee. Previously, he was a seniorinvestment manager at Saudi Interna-tional Bank, where he was responsiblefor all day-to-day portfolio manage-ment. Mr. Abberley holds a degree fromKeble College, Oxford, in philosophy,politics, and economics.

Bruno Bertocci is a principal at Stein,Roe & Farnham. Previously, he man-aged global equity portfolios andmanaged the London and Hong Kongoffices of Rockefeller & Company. Priorto joining Rockefeller, Mr. Bertocci wasa vice president at T. Rowe PriceAssociates in Baltimore. He holds anA.B. from Oberlin College and anM.B.A. from Harvard University.

Marc Faber is the founder of Marc FaberLimited, an investment advisor, fundmanager, and broker/dealer. He is alsothe investment manager of the Icono-clastic International Fund, a global fundspecializing in unusual investment op-portunities. Mr. Faber publishes amonthly investment newsletter, TheGloom Boom & Doom report, and is theauthor of The Great Money Illusion: TheConfusion of the Confusions. Mr. Faberholds a Ph.D. in economics from theUniversity of Zurich.

Robert Lloyd George is founder, chair,and chief executive officer of LloydGeorge Management Limited. Previ-ously, he served as managing director ofIndosuez Asia Investment Services andspent four years with the Fiduciary TrustCompany of New York researching inter-national securities in the United States

and Europe for the United Nations Pen-sion Fund. Mr. Lloyd George is theauthor of numerous articles and threebooks, A Guide to Asian Stock Markets, TheEast–West Pendulum, and North–South:An Emerging Markets Handbook. He hasbeen honored by the Financial Times Glo-bal Guide to Investing as one of the top10 global investors of the 1990s. Mr.Lloyd George was educated at Eton Col-lege, where he was a King’s Scholar, andat Oxford University.

Steven A. Schoenfeld is a principal inBarclay Global Investors’ emerging mar-kets group. Previously, he worked at theInternational Finance Corporation (IFC),where he developed new investmentvehicles and risk management instru-ments based on the IFC emerging marketindexes. He also worked in the deriva-tives industry, including three years atSIMEX, where he was the first Americanto trade Japanese stock index futures. Heis the author of The Pacific Rim Futures andOptions Markets. Mr. Schoenfeld was aFulbright Scholar in economics at theNational University of Singapore. Heholds a B.A. in history and governmentfrom Clark University and an M.A. ininternational economics and U.S. foreignpolicy from the Johns Hopkins Univer-sity School of Advanced InternationalStudies.

John C. Stannard, CFA, is director ofRussell Data Services for the Frank Rus-sell Company. He has held several previ-ous positions within the Frank RussellCompany, including assignments withthe company’s Tacoma-based U.S. con-sulting group and, in the London office,with the consulting group and withquantitative services and performancemeasurement. Prior to joining the FrankRussell Company, Mr. Stannard worked

Investing Worldwide VIII: Developments in Global Portfolio Management iii

Fore&Bio.frm Page iv Wednesday, November 12, 1997 2:55 PM

for Hallmark Associates, a software firm,and for Barclays Bank as an investmentanalyst and assistant portfolio manager.He holds a B.Sc. in pure mathematicsfrom Royal Holloway College, Univer-sity of London.

R. Charles Tschampion, CFA, is man-aging director of investment strategyand asset allocation for General MotorsInvestment Management Corporation(GMIMCo). He is responsible for devel-opment of long-term strategic asset allo-cation policy and GMIMCo’s tacticalasset allocation process, as well asequity and derivatives trading. Mr.Tschampion serves on the Board of Gov-ernors of AIMR, as a director of theFinancial Analysts Federation, and onthe editorial board for The CFA Digest.He holds a B.S. in industrial engineeringand an M.B.A. in finance from LehighUniversity.

D. Sykes Wilford is a managing directorat CDC Investment Corporation. At thetime of his presentation, he was the chiefinvestment officer of Bankers Trust’s Pri-vate Bank and managing director ofBankers Trust’s Global Investment Man-agement in London. Mr. Wilford isresponsible for global asset allocationstrategies. Previously, he served as man-

aging director of Chase Manhattan BankN.A. and as an economist with the Fed-eral Reserve Bank of New York. He hasauthored and edited books on subjectsranging from economic policy in devel-oping countries to Managing FinancialRisk. Mr. Wilford presently holds visitingprofessorships at City University of Lon-don and at L’Universite de Saint-Louis inBrussels, Belgium. He holds a B.A., M.A.,and Ph.D. in economics.

Marvin Zonis is a professor at the Grad-uate School of Business at the Universityof Chicago and the principal in the inter-national consulting firm of Marvin Zonis+ Associates. Previously, Mr. Zonis wasMiddle East consultant to ABC/CapitalCities Television and International Editorof WBBM-TV, Chicago. He has writtencolumns for the Weekly Tokyo Keizai busi-ness journal of Tokyo and numerousbooks, including The Eastern EuropeanOpportunity: The Complete Business Guideand Sourcebook, Majestic Failure: The Fall ofthe Shah, Khomeini and the Islamic Republicof Iran, and The Political Elite of Iran. Mr.Zonis was educated at Yale University,the Harvard Graduate School of BusinessAdministration, and the Institute for Psy-choanalysis, Chicago. He holds a Ph.D. inpolitical science from the MassachusettsInstitute of Technology.

iv ©Association for Investment Management and Research

Zonis.frm Page 1 Wednesday, November 12, 1997 2:58 PM

Political Risk in the World EconomiesMarvin Zonis

Professor of International Political EconomyUniversity of Chicago

PrincipalMarvin Zonis + Associates

Because the state must provide the crucial economic, financial, legal, and political infrastructure

for the operation of any market economy, political risk considerations need to be part of every investment decision. The model presented here to assess political risk uses 10 variables that have historically been highly

predictive of political instability; the presentation reveals the Top 10 and Bottom 10 of mid-1997.

nternational investors appear to beignoring considerations of politicalrisk in their asset valuations. Thisstrategy could be dangerous. Statesplay an immense role in economiclife, so the fates of governments are

immensely significant to all interna-tional investors. This presentationreviews the role of political risk analysisin investment markets and discusses thekey drivers of a political risk stabilitymodel that I developed two years ago.The political risk rankings produced bythe model of the 10 most stable and 10least stable of 65 countries and somenotes on China and Russia follow.

Political Risk Analysis

Market prices suggest that political riskis not a particularly relevant factor in theminds of international investors. Forexample, the spreads for sovereigndebt—not only of Italy and Spain versusFrance and Germany but also of theemerging markets—have all narrowed.Furthermore, the November/December1996 issue of the Financial Analysts Jour-nal contains an article in which theauthors conclude, “Trading on the basis

of the political-risk measure alone has noability to produce abnormal returns.”1

Political risk has apparently dimin-ished over time. For example, the dan-ger of nationalization has virtuallydisappeared. So, someone who consid-ers political risk to be primarily the dan-ger of nationalization will not want tospend money on insurance againstnationalization or take the risk of nation-alization into account in asset decisions.And in fact, the evidence from the equityand fixed-income markets is that themarkets have discounted this risk.

Another sign that investors aredownplaying political risk is foreigndirect investment. The total global flowof foreign direct investment increasedfrom US$50 billion in 1985 to US$300billion in 1995. Combined with themassive flow of funds into internationalequity markets, as well as into sovereignbonds, foreign direct investment is mov-ing immense amounts of capital aroundthe world—into projects, mergers,acquisitions, and foreign operations.

1Claude B. Erb, Campbell R. Harvey, and Tadas E.Viskanta, “Political Risk, Economic Risk, andFinancial Risk”:29–46.

I

Investing Worldwide VIII: Developments in Global Portfolio Management 1


Evidently, investors have rapidly for-gotten the Mexican peso crisis ofDecember 20, 1994.

Those in the international investmentgame should not forget. Despite the col-lapse of the Soviet Union, despite theabsence of communism, and despite thedisappearance of the risks of national-ization, political risk poses major dan-gers to investments. The reason is thatthe state, especially in emerging mar-kets, plays a determinative role in theeconomy of countries.

The simplistic view of the relation-ship of the state to the economy is thatthe more the state gets out of thebusiness of the economy, the greater theeconomy will flourish. The reality is,however, that a country cannot have asuccessfully functioning market with-out significant state intervention. AsEconomics 101 teaches, markets pro-duce the greatest output economicallybecause markets are about allocatingand distributing goods by price, and ifgoods are allocated and distributed byprice, they are allocated and distributedby the criteria of efficiency. The result isto increase output and minimize inputs,which increases profits and, therefore,increases the total growth of theeconomy. In order for that condition topertain, however, the state must playcertain critical roles: maintaining cur-rency stability, controlling inflation,establishing the legal framework inwhich transactions can occur, prevent-ing monopolies, regulating banks, andregulating securities markets. In short,the state must provide the crucialinfrastructure—economic, financial,legal, and political—for the operation ofany market economy.

In emerging markets, the state playsan even more significant role than that ofproviding infrastructure. For example,emerging markets—especially those inwhich the state previously controlledmuch of the economy—have far moreregulations governing economic activitythan do nonemerging markets. Emerg-

ing markets thus provide massive roomfor bureaucratic interpretation and,therefore, for massive corruption, whichreduces rationality in the operation ofthe economy. States that are privatizingand are seeking to break up monopoliesreplace their ownership of firms withnew regulations that govern the behav-ior of those firms. As a result, privatiza-tion leads to more regulations, more rules,and more government interference in theeconomy than existed before the privati-zation occurred. The same is true fortrade. As free trade increases, govern-ments increase the regulations that theyimpose on imports as a way of guidingthe economy in the absence of the tariffsand quotas that were eliminated in thecourse of supporting free trade.

Political Risk Model

If states play an immense role in eco-nomic life, why is the concern for polit-ical risk virtually disappearing? Onereason is that getting a good handle onpolitical risk has seemed impossible.Various models have been constructed,but the results and the interpretationscan be baffling. Therefore, I set out twoyears ago to develop a model to predictpolitical instability. The model is backedby 35 years of theoretical investigationsinto political stability.

To develop the model, we began with22 theoretical propositions about politi-cal stability, operationalized those prop-ositions, and tested the model by using30 historical instances of politicalinstability—from the student unrest inFrance in 1968 to the Iranian revolutionof 1979 to the Mexican peso crisis of1994. Through a series of regressionanalyses, we identified 10 variables thatseem highly predictive of politicalinstability historically: per capita GDP,rental income, distribution of income,predictability, agriculture, trauma,democracy, competitiveness, quality oflife, and human capital.

2 ©Association for Investment Management and Research


■ GDP per capita. The first factor,weighted slightly more heavily than anyof the others, is per capita GDP. Theproposition is that relatively rich peopleare less discontented than relativelypoor people. Societies that produce ris-ing GDP per capita over time and rela-tively high levels of GDP per capita overtime are likely, everything else consid-ered, to be more politically stable.

■ Rental income. The second mostimportant factor in the market is rentalincome, income that countries enjoy butdo not work for. Rental income encom-passes oil revenues, exports of naturalgas and gold, remittances from foreignworkers, Suez Canal tolls, foreign aidderived from other countries, and so on.The search by the Western powers forcolonies was driven by the desire to cap-ture rental income; the areas with themost natural resources were alwaysconsidered the most desirable coloniesby the imperial states.

The proposition in this model is thatfew countries that enjoy high levels ofrental income in proportion to GDP willever be successful—in the same sensethat it is hard for rich people to raisesuccessful children because if their chil-dren do not have to work for a living,they will never develop a work ethic.For example, Saudi Arabia will never bea successful country because smart,young Saudis do not work for a living;they figure out a way to get in on thedistribution of the rental income. Thecountries that have succeeded since theend of World War II, the countries oftruly great growth (with the exceptionof the oil countries), have all been coun-tries with no natural resources—Japan,Hong Kong, Singapore, Taiwan, SouthKorea. The great growth stories are toldof those countries without significantrental income.

■ Distribution of income. Equality orinequality of income may be an impor-tant issue, but it is not highly relevant toquestions of political stability. Few peo-ple compare themselves with the Rocke-fellers or the Melons of this world. Most

compare themselves in one time periodwith how they were doing in anotherperiod. What is more relevant thanincome equality/inequality is whetheror not the system is distributing eco-nomic benefits throughout society—even to the poorest people—becausepoor distribution of economic benefitsleads to political instability.

Infant mortality is an important wayof measuring income distributionbecause, in developing countries inparticular, the children who die are notthe children of the elite. Their mothers,no matter the country, receive propernutrition, proper medical care, properprenatal care, proper delivery methods,proper hospital care for the infant, andso on. Babies of poor people die. So,infant mortality is a proxy for the abilityof a system to distribute its benefitsthroughout the system—or in a morecynical view, to buy off the lower classes.

■ Predictability. Investors want tocommit capital to systems that are highlypredictable. (The reason investors arenervous about China and Russia, for in-stance, is the lack of predictability inthose markets.) I measure predictabilityby changes in wholesale prices, whichtend to fluctuate less than consumer pric-es. The argument is that stability in therate of inflation is more important thanthe level of inflation. Of course, a highrate of inflation is virtually impossible tomaintain at a stable level over time.

■ Agriculture. In this century, no statecan get rich from agriculture if theagricultural commodity is legal. If theagricultural product is illegal—cocaine,heroin, or marijuana—a state can getrich, as Colombia illustrates, but it isgetting rich from rental income, notagricultural products. The reason nostate can get rich from agriculture is thatno high-value-added agriculture exists.The price of a bottle of wine in compar-ison with the price of grapes is probablythe highest-value-added legal agricul-tural price, and that difference is trivialin comparison with the value-addedprice of virtually any manufactured



product. Agriculture as a high percent-age of GDP is a sign of a country that isnot going to do well.

■ Trauma. From a political point ofview, traumas are wonderful. They elim-inate resistance to change, focus peo-ple’s attention, and unleash energy.Nobody in Germany in 1946 wanted togo through that trauma again. No onewas sitting around in Berlin saying, “Oh,basically everything is fine. If we justkeep going, making some minor adjust-ments, Germany will be terrific.” Statesthat passed through trauma some yearsago are more likely to be successfultoday than not. The great economicpowers that have grown great sinceWorld War II—Germany, Japan, Italy,Hong Kong, Taiwan, Singapore, andKorea—have all suffered massive trau-mas. Countries that have experiencedrecent trauma are still wanderingaround in a depressed haze (in Moscowtoday, people are too stunned to bedoing anything yet), but in time, thestate will move beyond the trauma todevelopment. “Invest when there isblood in the streets” may be understoodas “invest at a time of trauma.”

■ Democracy. Democracy is impor-tant because, in the absence of ideology,the only legitimator of power is democ-racy, popular choice. And the power ofideology is diminishing in most cornersof the earth. Lack of legitimacy isanother factor raising political risk inChina; since Deng Xiaoping’s economicopening, neither political ideology nordemocratic choice legitimates the rule ofhis successors.

■ Competitiveness. Competitiveness isan important variable: Is the economyprepared to participate in the globaleconomy? Competitiveness in the modelis measured by imports plus exports as apercentage of GDP.

■ Quality of life. Another factor help-ing predict political risk is whether thestate is delivering to its people a life thatthey regard as a life of quality. The wayto measure such quality best is the sim-ple variable of life expectancy. Russia is

the only industrial country in the worldin which life expectancy has plummetedin the past 25 years; life expectancy wentfrom close to 70 years for an adult Rus-sian male in 1970 to 57 years for an adultRussian male in 1995. For comparison,note that in Japan, the average lifeexpectancy for an adult male is 80.

■ Human capital. The crucial variablein the performance of states is no longerphysical or material capital. It is humancapital—and specifically, human capitalmobilized in a market economy.

Country Rankings

The model that uses the factors justdescribed generates political risk rank-ings for individual countries. To pro-duce country rankings, we run thosevariables many times in different ways.This section describes the model’s mostrecent ranking of the 10 most stable andthe 10 least stable countries and dis-cusses several interesting scores.

Most StableNot surprisingly, Switzerland is #1,

with a score of 8.23 on a scale from 1 to10. The second most stable country inthe world from the point of view of thepolitical risk model is Japan. The rest areas follows: France #3, the Netherlands#4, the United States #5, Italy #6, Ger-many #7, Australia #8, Finland #9, andSpain #10. The specific rankings may besomewhat surprising—Italy’s place-ment as #6, for example—but mostanalysts would agree that these coun-tries should be in the Top 10 in terms ofstability.

Italy’s placement follows from thedefinition of stability. Consider, forexample, two metaphors for stability.One metaphor involves balance: In NewHampshire, there is a well-known bal-ancing rock, a huge boulder thatappears about to roll off its narrow base,but it has sat on that base for a long time.A different metaphor for stability is anavalanche. An avalanche is an extraor-dinarily stable system—not in the sense



that it can be controlled but in the sensethat its path is quite predictable. Theinternal components of the avalanchemay not be predictable, but for consid-ering its stability, we are less interestedin the internal parts than in its overalldirection and in when and where it isgoing to settle. Italy is stable in the senseof an avalanche’s stability rather thanthe stability of a balancing rock, which,after all, could be knocked off its basewith the right pressure.

Most UnstableThe country that came out at the bot-

tom of the political risk rankings with ascore of 0.38 is Bulgaria—a country thatwe put in the model only in late summer1995. Second from the bottom is Zimba-bwe, third is Nigeria, fourth is Morocco,fifth is Algeria, Turkey (slightly lessunstable than Algeria) is sixth, Pakistanis seventh, Brazil is eighth (which is anamazing output of the model), Iran isninth, and finally, tenth up from the bot-tom is Ukraine.

Russia and ChinaRussia and China—two key coun-

tries in which investors are interested—were not in the Bottom 10 of unstablecountries. These two countries scored3.21 and 3.61, respectively, which is bet-ter than India but not quite as stable asJordan, which came in at 3.62. Becauseof their interest to investment profes-sionals, Russia and China’s rankingsmerit an explanation.

■ Russia. One of the most overratedcountries in terms of foreign investmentat the moment is Russia. Relatively lowlevels of capital are actually moving intoRussia, but the interest is tremendousand has been growing because the Rus-sian stock market has been going uprapidly. Since the election of BorisYeltsin, the market has gone up 55–65percent. Note, however, that this risefollows a period in which the marketessentially lost all of its value, so the riseis from a very low base.

The problem with Russia is exactlythe problem I began this presentationwith: To have a robust market economy,you need a robust state. The problem inRussia is not too much government butrather not enough government. Afterthe collapse of ideology and after thecollapse of the Soviet dictatorial regime,no one in Russia is anxious to see anotherpowerful state constructed. Conse-quently, the Russian government cannotcollect taxes, impose a legal system, cre-ate a criminal investigation and prosecu-tion system, or successfully pursue war.Russia has fought three wars in recenthistory against the Islamic people wholive to its south. It has lost two of thosewars (against Afghanistan and theChechen people) and is about to lose thethird (a war being fought by Islamicforces who oppose the government inTajikistan). Such losses signify a weakstate, and a weak state precludes arobust market economy. In Russia today,six wealthy businessmen control con-glomerates that, in turn, control morethan 50 percent of the Russian economy.In other words, monopoly or oligopolypower is dominating the Russian econ-omy and preventing its emergence intorobust economic growth. If Russia sur-vives, it will survive on the basis of rentalincome. Russia will be able to export rawmaterials and use the wealth those rentsproduce to buy off its population, but itwill not produce a viable market econ-omy in the foreseeable future.

■ China. China is in a very differentposition from Russia. China is not a richcountry, but its vibrant economy hasbeen growing more rapidly than anyother economy in the world. In the1990s, the rapid growth of the Chineseeconomy in the 1980s has begun to paybenefits in the form of rising per capitaincome. Average per capita income inChina has reached US$600, up fromUS$300 in 1980.

Despite this tremendous growth, Iam cautious about the future of China.The reason for my concern, and the



reason China’s low ranking for politicalstability is appropriate, is because Chinahas no governing ideology and nodemocracy. The existence of the dictato-rial regime in China has no justificationother than that it produces economicgrowth, and the reality is that theeconomic growth it is producing is notspread throughout the society. The over-all unemployment rate in China isestimated by the government to be 11–12 percent, but the unemployment ratein the cities is estimated by outsideobservers to be more than 30 percent;that is, only two out of three people inthe labor force are able to find work.Moreover, China has yet to go throughthe political turmoil that will followfrom a state that, in the absence ofideology, has created massive unem-ployment after generating tremendousexpectations. Some very unstable timesmay lie ahead.

Conclusion

Can investors do anything with theseideas? For managing risk, the point isnot to look at the output of this or anyother model and simply avoid countriesbelow some arbitrary number on apolitical risk scale. The point is to use theinsights to understand an investment’strue risk–reward ratio, which investorshave not been doing. The flow of fundsinto equities, the flow of funds into fixedincome, and the decrease in the spreadsindicate that the time has come to restoreconsiderations of risk in investing inemerging markets. A market crisis willrestore those considerations, but wait-ing for a crisis is dangerous. When willit come, and how hard will it hit?Investors hope for a relatively painlesslesson—one on the order of the pesocrisis of 1994—but the lesson could befar more painful. Political risk consider-ations need to be part of every invest-ment decision in emerging markets.


Techniques for Today’s Global Asset Allocation Strategies

D. Sykes WilfordManaging Director and Chief Investment Officer

for Private BankingBankers Trust Company1

Asset allocation is a simple tool, but investors and managers generally do not apply it correctly. The

recommendation here is to go back to basic theory: No one knows the future, so you must use forecasts, not

history; for truly global allocations, forecast excess, not total, returns; make guesses and establish a sense of how good the guesses are and in which direction they tend to go; and let your optimizer do what it is supposed

to do without constraints.

sset allocation is respon-sible for somewhere be-tween 50 percent (fromsomeone who is belit-tling it) and 90 percentof portfolio earnings. If

it is such an important determinant ofreturns, why does everyone want toknow each other’s latest stock pick? Noone comes up to you at a cocktail partyand asks, “What is your asset alloca-tion?” Either asset allocation is very dif-ficult to understand, or it does not giveus much value as it is currently beingimplemented. This presentation revisitsthe theory of asset allocation, focusingon some of the theoretical issues, andlays out certain approaches that willhelp portfolio managers and investorsapply asset allocation correctly.

Theory

Asset allocation theory today haschanged little from the theory devel-oped 40 years ago, but now, for the firsttime, probably, since asset allocation

theory was developed, investors andanalysts can do asset allocation prop-erly. What business school taught aboutasset allocation was largely wrong,which is why we use it so poorly and getso little value from it.

In the basic concepts of modern port-folio theory, the risk of a portfolio can belower than the risks of any of its compo-nents because of diversification. Butwhat does “diversify and get rid of risk”truly mean? Mathematically, investingin both assets A and B is less risky thaninvesting in either one of them sepa-rately only if assets A and B are not oronly slightly correlated. Simply put,reduction in risk depends on a sufficientnumber of weakly correlated assets inthe portfolio.

Investment managers focus on assetallocation because they want the maxi-mum return possible for taking on agiven level of risk; that is, they alwayswant to be on the efficient frontier,shown in Figure 1. Investors do not wantto be below the efficient frontier; inves-tors with portfolios below the curve aretaking risk needlessly and getting no

1Mr. Wilford is now a managing director at CDCInvestment Management Corporation.

A


return for it. Investors can move alongthe efficient frontier, trading off the riskin the portfolio for the return that mightresult. Asset allocation supposedlyhelps find efficiency on this frontier. Theproblem is that the efficient frontier is amoving target.

A different, and simpler, way to thinkabout modern portfolio theory is interms of diversifiable risk. For any port-folio, risk can be separated into diversi-fiable and nondiversifiable risk, asshown in Figure 2. If managers addenough assets to their portfolios, theycan get rid of diversifiable risk, which isalso known as bad risk. They cannot getrid of nondiversifiable risk, also knownas good risk. Why should some risk bebad and some be good? Good risk is therisk managers get paid to take. Bad riskis the kind of risk the market will not paya manager to take. A manager can getlucky, of course, and have bad risk turn

out to be good risk because the manageris on the right side of a distribution, butgood risk is the kind of risk the marketwill pay a manager to take ex ante.

Figure 2 also shows locally diversifi-able and globally diversifiable risk. The-oretically, even in a local market,managers can get rid of a lot of risk, butif the market is taken as the wholeworld, they can get rid of more risk.

How many portfolios, however, areglobally diversified? For example, inSpain, Spanish investors say, “Whyshould I invest outside Spain? I do myoptimizations, and they tell me to buySpanish bonds.” In Germany, the Ger-mans say, “Why should I invest outsideGermany? I should buy German bankdeposits.” And they do; about 80 per-cent of their investments are in fixed-income securities or deposits. In theUnited States, people say, “I should buythe S&P 500 Index because it always

Figure 1. The Theoretical Efficient Frontier

Ret

urn

Risk

?

Figure 2. Modern Portfolio Theory: A Simple View

Ris

k

Number of Assets

Nondiversifiable

Locally Diversifiable

Globally Diversifiable


goes up. I shouldn’t buy anything over-seas.” And so it goes around the world.In some markets, investors may ownonly the fixed-income market; in others,they might own bank deposits; in themore adventurous, such as the UnitedStates, they hold some domestic equity.The Spanish have some logic behindtheir position; over the past 30 years,Spain had the highest real rate of returnon investments, and from a risk-adjusted standpoint, the return wasprobably coming from Spain’s fixed-income market. The German and U.S.positions are not so logical.

The reality is that people have a localfocus. Most people tend to ignore diver-sification in general, but even if they dodiversify, it is local. Part of the reason isprobably the difficulty of accessing for-eign markets in the past and the result-ing exorbitant transaction costs forinternational investing. Twenty yearsago, a U.S. investor could hardly con-sider a stock in, say, Italy; the transactioncosts associated with simply gettingone’s hands on such a stock were pro-hibitive. Today, however, global marketsare as accessible as local markets—atleast through indexing. So, the questionremains: If global markets are as acces-sible as local markets, why so little glo-bal diversification? One or two aspectsof the concepts of global diversificationand asset allocation may help explainwhy people tend to stay home ratherthan invest globally.

Use Excess Return SpaceGlobal efficiency should be in excess

return space, not total return space. The-ory discusses risk relative to the no-riskalternative. For years, the risk-free ratemeant the credit-risk-free rate, but thatis not necessarily correct. The risk-freerate is the known rate for the investor’shorizon—the one-year rate for an inves-tor with a one-year horizon. This defini-tion makes sense if one thinks about riskas what an investor has to pay to go lookfor return. Why should investors paymoney for something from which they

will not get excess return? So, the effi-cient frontier should be developed on agraph of risk and return that picturesreturns in excess of the risk-free rate, nottotal returns, which is misleading.

Several years ago, I gave a presenta-tion on this concept to the portfolio man-agers at Bankers Trust, and they said itdoes not matter whether one is talkingabout excess returns above the risk-freerate or total returns: “Anybody can sim-ply make the adjustment.” So, I toldthem to go build an ECU (European cur-rency unit) portfolio, a German markportfolio, a Japanese yen portfolio, anda U.S. dollar portfolio, optimize them,and bring them back. When the exercisewas done—the same optimizations,same risk targets, and same volatilityassumptions—the result was four dif-ferent portfolios, each of which was“globally” efficient. How can four dif-ferent portfolios be the efficient portfo-lio? It is impossible. But investmentmanagers produce them because of thecurrency problem.

Global efficiency should be currencyindependent, not currency dependent.And as soon as analysts move intoexcess-return space, returns are 99 per-cent currency independent. In excess-return space, “globally efficient” is alsono longer home-country dependent. So,an efficient portfolio in U.S. dollar termslooks the same as an efficient portfolioin Spanish peseta terms, with the excep-tion of the currency hedges to bring theportfolio back to the excess return abovethe risk-free return, which is differentfor the Spanish, the German, and theU.S. investor.

Stressing the use of excess returnsmay seem pedantic and unimportant,but it is critical for performing asset allo-cation correctly. Otherwise, not only arethe portfolios wrong, but future perfor-mance has no comparison with the past.

The Past and the Future In local or global portfolios, and even

in excess-return space, portfolios carrycertain risks and returns that change


over time, as the next series of figureswill show. Figure 3 is a graph of risk andreturn in excess-return space for variousasset classes and markets for the five-year period beginning in January 1975.It reveals some interesting informationabout where investors obtained excessreturns for the risk they were taking inthis period. The figure compares theMSCI (Morgan Stanley Capital Interna-tional) country indexes, the MSCI WorldIndex, U.S. stocks (the S&P 500), andU.S. 10-year government bonds; risk, orannualized volatility, was measured asstandard deviation of return. The riski-est market, but it sure provided return,was the United Kingdom. Japan was awinner in providing reasonable risk forhigh return. The MSCI World and theS&P 500 indexes were low risk and lowreturn. For this five-year period, thechoice assets were Japan and maybeGermany. France and the United King-dom were pretty good. The worst allo-cation was U.S. government bonds; theywere low risk but had negative returns.

Figure 4 shows the next five-yearperiod, 1980 through 1984. Supposethat, on the basis of the 1975–79 period,an investor had invested in the United

Kingdom. How would that investmenthave done? The equity was still highrisk, but the return for the period waslow, less than 2 percent. Germany atleast gave a return for the risk takenduring this period. The S&P 500 waslower risk with some return. U.S. gov-ernment bonds simply became morerisky for negative returns. Japan stilllooks like the place to invest: Five yearsbefore, Japan was a great place to invest,and it was a great place to invest thisfive-year period. So, to get ready for thenext five years, in typical recommenda-tion terms, “based on the last five yearsof data, Japan is recommended.”

Look at the next five years. In Figure5, Japan shows up as still the one placean investor can consistently go. Franceis now performing and, in hindsight,should have been bought. And all of asudden, U.S. government bonds and theS&P 500 have started to look good.

Finally, Figure 6 shows the final five-year grouping, 1990 through September1995. The place to be was the S&P 500and U.S. government bonds. The oneasset that was always up in the pastperiods, Japan, gave up everything in

Figure 3. Risk versus Reward, January 1975–December 1979

1815 21 24Risk (annualized volatility)

22

20

18

16

14

12

10

8

6

4

2

0

–2

–4

–6

Ann

ual R

etur

n (%

)

0 39

U.S. Government Bonds

27 30 33 3663 9 12

MSCI World

MSCI Japan

MSCI Germany

S&P 500

MSCI France

MSCI United Kingdom


this period that it had returned in theprevious 15 years.

Today, U.S. investors ask, “Whyshould I invest anywhere but in the S&P500 and U.S. government bonds? Lookwhat a great return they have provided.”Even if this analysis were extended to1997, low risk and high return dominatein the United States. Of course, an inves-tor could have said the same thing for

Japan in January 1990. So, one of the les-sons from this exercise is to go back tobasics: The past does not predict thefuture. If investors simply look at the pastas an indicator of the future—whetherusing five years, three years, or movingaverages—they will choose markets justbefore they go down. A good rule couldbe: All looks bright just before the hurri-cane hits, at least in markets.


Risk (annualized volatility)

16

14

12

10

8

6

4

2

0

–2

–4

Ann

ual R

etur

n (%

)

18150 24


2163 9 12

MSCI World

MSCI Japan

MSCI Germany

S&P 500

MSCI France

MSCI United Kingdom


MSCI France


36

32

28

24

20

16

12

8

4

0

Ann

ual R

etur

n (%

)

18150 30


2163 9 12

MSCI World

MSCI Japan

MSCI Germany

S&P 500

MSCI United Kingdom

24 27


An Approach to Theoretical Correctness

The theorists said that, in a perfectworld, investors should diversify every-where, all the time, instantaneously, intoeverything. The theory assumes notransaction costs and no frictions. Ofcourse, frictions and transaction costs doexist in the world, and investors cannotignore them. What does it mean, then, todiversify a portfolio everywhere, all thetime—in a friction-free world or in aworld with friction?

The meaning starts with a basic prin-ciple: Investors have portfolios becausethey do not know the future. No one isprescient. If investors knew the future,the perfect portfolio would be one bet,because they would know the outcome.

The power of portfolio theory is thatit allows investors and managers to dealwith the world of uncertainty, risk.Because the concept of looking at thepast, which we know perfectly, to builda portfolio is dead wrong, investorsmust begin by setting expectationsabout the future in building a portfolio.Theory demands that investors make aguess about the future. When investorsor managers make such guesses in

building their portfolios, what is criticalis not the guesses but how wrong theycan be—a little wrong or a lot wrong—which depends on the distribution ofexpected returns—that is, the volatilityof what will happen next. Figure 7shows the distribution patterns for beinga little wrong and a lot wrong. Howwrong investors will be has everythingto do with how good they are at forecast-ing, not how volatile the actual marketis. An excellent forecaster will probably

Figure 6. Risk versus Reward, January 1990–September 1995

MSCI France


8

6

4

2

0

–2

–4

–6

Ann

ual R

etur

n (%

)

18150 30


2163 9 12

MSCI World

MSCI Japan

MSCI Germany

S&P 500

MSCI United Kingdom

24 27

Figure 7. How Wrong Will You Be?

Expected Return

Little Wrong

Big Wrong


be a little wrong; a poor forecaster willbe a lot wrong. In portfolio theory, how-ever, that difference does not mattermuch because it is taken into account inthe correlation matrix of risks.

When we in the industry use the lastfive years to get the volatility for the nextfive years, we are forgetting what thetheorists were talking about 40 or 50years ago. We do it because measuringhow wrong we could be is bad forcocktail party discussions and becauseit is difficult to do. So, we make forecastsabout future expected returns, and thenwe turn around and estimate howwrong we could be (future variance)based on something in the past. Nowonder the resulting portfolios appearto be irrational!

Once we have our guess about thefuture and have some idea of howwrong we could be (the standard devia-tion of our guesses), we are also inter-ested in whether we guess everythingwrong in the same direction. In portfoliotheory, it does not matter that one iswrong if one is consistently wrong; port-folio theory allows us to adjust for beingwrong in a consistent manner. Knowingthat any guess will be wrong, what theinvestor needs to do to create an optimalportfolio is have an idea of how wrongthe guesses will be and how closelyassociated those guesses will be in“wrongness” space.

Nothing in the theory says that actualhistorical correlations must be used orthat actual historical standard devia-tions must be used. If we use forecastsof the future from sample data and thenplug in perfect guesses about future cor-relation and volatility, the portfolios areworse than if we had done it theoreti-cally correctly. So, knowing what mayhappen in the future regarding correla-tions and volatilities of underlyingactual prices will not build better port-folios. The idea that we do not knowwhat prices will be but do know thatvolatility in the future will look like vol-atility in the past is wrong.

The correlation data needed for port-

folio allocation should reflect the errorsin our forecasts of the future. For exam-ple, when I get the German mark wrong,I tend to get the French franc wrong thesame way. My models are so interlinkedthat they tend to be wrong the same way.That link is what matters, not whetherthe forecasts are actually right or wrong.

Consider the following forecastingapproaches. The DOLS (dynamic ordi-nary least squares regression) line inFigure 8 is the output of a sophisticatedquantitative forecasting system forglobal asset allocation. It is a Bayesianerror-learning system and is based onvariable autoregressive principles. Themodel is in excess-return space, soanything above the dotted line is greatand anything below that line is losingmoney relative to cash. The model istheoretically correct in excess-returnspace, but it uses a historical variance–covariance matrix for the optimization,which is common practice in theindustry. The returns shown are thekinds that yield a Sharpe ratio over afive-year period of about 1.0. The resultsare real; I did not cheat with statistics.Most managers would be pleased withthe model output.

The DSUR line in Figure 8 is based onan estimation procedure called “dy-namic seemingly unrelated regres-sions.” It reflects the performance of aportfolio created with exactly the sameinput data as DOLS but, instead of somehistorical measure of volatility and cor-relation, the correctly measured errorsof the forecasts. The DSUR cumulativereturn is five times as large as the DOLScumulative return, yet the forecasts arebasically the same. The only differencebetween the two lines is that the DSURone is theoretically correct.

Knowing the future volatility andcorrelation does not help significantly toimprove returns because the theory isbased on the assumption that one doesnot know the future of anything. Wesimply have guesses about the future;we are either good guessers or badguessers, and there is a relationshipbetween the guesses we make. When


people talk about correlation, what theyshould say, theoretically, is the correla-tion of their guesses and the errors intheir guesses.

Asset allocation is a simple tool, butinvestors and managers ended up doingit wrong because obtaining measures ofvolatility and correlation that were notbased on history was difficult. The tech-nology has changed, however, andtoday, analysts can create methodolo-gies so that when they make guessesabout future returns, they can talk aboutthe consistency of the direction of theirerrors, consistency in the volatility asso-ciated with those errors. So, technologyallows correct application of the theory.

When the new technology, such asoff-the-shelf optimizations, is usedincorrectly—not done in excess-returnspace, expected returns are used but notexpected variances and correlations—itreinforces investors’ biases against glo-bal investment as well as global assetallocation. The resulting portfolios donot look appropriate, and managersattempt to sell the product by puttingconstraints on the optimization. “Con-strained optimization” sounds good,but it is the worst kind of error. In thefinancial markets, investors and manag-ers want robust solutions. When weconstrain our optimizations, we are

guaranteeing that we do not have robustsolutions. When we are wrong, we aregoing to make big mistakes.

Conclusion

To do asset allocation correctly, startwith a premise that no one knows thefuture. Forecast excess returns, makeguesses, have a sense of how good aguesser you are and in what directionthose guesses line up, and then do notconstrain the system. Let it do its duty.

Consistency—in excess returns, ex-pected returns, volatility, and correla-tions—is the key. You do not have tohave the fanciest model in the world.You do not have to have the most sophis-ticated system to use portfolio theoryeffectively. Go back to the basics.

In global asset allocation, as soon asyou operate in excess-return space, themodel will come up with portfolios thathave lots of global assets. In a globaloptimization, portfolios look the sameto European and U.S. investors, who canthen simply use the cheapest contract inthe world, the currency forward con-tract, to translate the returns to home-country currency. Going back to thebasics of modern portfolio theory allowsyou to not know the future but still pro-vide your clients with good returns.

Figure 8. DOLS and DSUR Portfolio Performance

6

5

4

3

2

1

0

–1

Cum

ulat

ive

Exc

ess

Ret

urn

(%)

12/876/8712/86 6/88 12/88 6/89 12/89 6/90 12/90 6/91 12/91 6/92 12/92 6/93

DOLS

Cash

DSUR


Bertocci.frm Page 15 Wednesday, November 12, 1997 2:55 PM

Global Equity Analysis: Country, Industry, or Company Selection?

Bruno BertocciPrincipal

Stein Roe & Farnham

Country, industry, and company selection are inseparable in the global arena. One useful analytical approach is to separate the local industries from the

global industries, value local markets after the companies that belong to global industry groups have been excluded, and compare companies only within their appropriate groups—local companies with local

peers, global companies with global competitors.

n the United States, analysts oftenselect individual securities withinthe context of their industry. On aglobal basis, country, industry, andcompany selection decisions arecompletely intertwined. For two

key reasons, country selection outsideof the United States often leads to unin-tended industry concentrations. First,most non-U.S. equity markets are con-centrated in a few industries. Second,global industries—industries that crossborders so the location of the companyis not important—exist, and they aredistributed very unevenly among theworld’s markets. This presentation dis-cusses these issues and outlines anapproach to global investing that recog-nizes these characteristics.

Industry Concentration

The U.S. equity market is the mostdiversified market in the world. Themarket represents many industries and,within each industry, many competi-tors. The largest market capitalization inthe U.S. market is General Electric (GE),and it is barely more than 2 percent ofthe U.S. equity market. A key character-istic of the U.S. market is that investors

can usually identify several companiesin the same industry and compare themwith each other. As a result, most U.S.equity investors and analysts are com-fortable with the idea that when one ismaking a choice among several stocks,one examines companies in the sameindustry and analyzes the differencesbetween them in terms of stock valua-tions and other characteristics.

Non-U.S. markets, in contrast, areoften very concentrated. Most non-U.S.markets contain one or only a few stocksthat take up a significantly higher pro-portion of market capitalization than GEdoes in the United States. A good exam-ple comes from the Netherlands, whereKLM Royal Dutch Airlines makes up ahuge proportion of the equity market. InAsia, Telekom Malaysia dominates theMalaysian market. Fiat makes up a lotof Italy’s equity market.

Not only do a few large companiestend to dominate in non-U.S. markets,but often an industry has only onecompetitor, so identifying stocks in thesame industry to compare with eachother is much more difficult than in theU.S. market. No other stock in theNetherlands can be compared with

I



Royal Dutch; Telekom Malaysia has nocompetitors; no other significant auto-mobile companies exist in Italy that canbe compared with Fiat.

One of the factors that tends toincrease company or industry concentra-tion in non-U.S. equity markets today isprivatization. About 20 percent of theU.K. equity market, for instance, consistsof companies that former U.K. PrimeMinister Margaret Thatcher privatized;these companies did not exist 10 yearsago. Most developing markets have util-ities, banks, and other companies thathave been recently privatized. Theseinstitutions are often enormous compa-nies, and their entry into the privateequity markets changes the structureand dynamics of the markets.

Even if a market is not dominated byone company, non-U.S. markets tend tobe dominated by one or two industries.Figure 1 shows this kind of dominancefor the Finnish, Spanish, and HongKong equity markets. In the Finnishmarket, forest products make up more

than 15 percent and Nokia, a telecom-munications and electronics company,makes up more than 25 percent. Aninvestor who likes Finland is basicallymaking a statement about liking the for-est products industry and liking Nokia.In Hong Kong, a quarter of the marketis bank stocks and about another quar-ter is real estate, so those sectors drivethat market. In Spain, utilities and banksare about half of the market. Every non-U.S. market has a few extremely impor-tant industries and companies.

The industries in which competitiontakes place across borders are extremelyunevenly distributed across countries.Figure 2 contains four examples. Notethat data processing is about 5 percent ofthe U.S. market but is basically nonexist-ent in Singapore, is less than 2 percent inItaly, and so on. Global food and house-hold products companies—Nestlé orProcter & Gamble, for example—haveno market-cap weight in Singapore andvery large weight in the Netherlands

Figure 1. Industry and Company Dominance in Non-U.S. Markets

Source: Based on data from FactSet (January 1997).

Finland Hong Kong

Spain

Other

Other

Other

Banks

BanksNokia

Utilities

RealEstate

ForestProducts



(through Unilever). Japan has no energyresources and thus no energy resourcesindustry. In the United States, on theother hand, about 9 percent of the mar-ket consists of energy companies. Thesecompanies were founded in the UnitedStates, but they compete across the globeand explore for resources in every possi-ble corner of the world. Telecommunica-tions is another example; SingaporeTelecom occupies almost 40 percent ofthe Singapore equity market, but the

industry is a smaller portion (in somecases, a much smaller portion) of theother markets.

History, natural advantages, geogra-phy, climate, and many other factors cre-ate concentrations in some industriesand the absence of certain other indus-tries in national equity markets. Thesefactors drive the development of com-panies and sectors in the markets.Because most global markets are con-centrated in a few industries, a country

Figure 2. Distribution of Four Global Industries

Source: Based on data from MSCI via FactSet (January 1997).

12

9

6

3

0

Mar

ket S

hare

(%)

Food and Household ProductsData Processing

40

30

20

10

0

Mar

ket S

hare

(%)

TelecommunicationsEnergy Sources

United States Japan United Kingdom Germany

Netherlands Italy Singapore



selection will result in an industry con-centration whether the investor wants itor not. The investor could explicitlyomit an industry in a market, of course,but then the investor is making a specificbet against that industry. Either way,because of various industries’ domi-nance in certain markets, going into anon-U.S. market is a bet for or againstthe dominant industry. To avoid makingindustry decisions while selecting coun-tries and to avoid making country deci-sions while selecting industries are quitedifficult.

A Global Approach

The literature does not provide conclu-sive evidence about the relative impor-tance of country and industry effects onstock returns. First, the data are extremelydifficult to decompose. Analysts have ahard time deciphering whether they areanalyzing country returns or industryreturns. Second, most of the data containlarge standard deviations. Multicol-linearity is also a big problem in statisticalanalysis of returns because the factorsthat are being measured are closelyrelated and cannot be separated usingstatistical analysis.

One solution is to view industries aseither global—those that compete inglobal markets—or local—those thatoperate within their local markets. Glo-bal industries are not tied to the fortunesof any one home country but to theglobal economy and to industry dynam-ics. For instance, in the steel industry,steel is priced in dollars and competitiontakes place across borders. To the mostefficient, lowest-cost producer of steel,the location of the company’s factorydoes not matter; the company willexport steel out of its country no matterwhat the state of the country’s economy.Many other industries—gold, semicon-ductors, petroleum, and so on—operatesimilarly in global sectors. Companiesthat compete in global industries areborderless. Their operating characteris-tics are independent of local economic

conditions. Companies that operate inlocal markets are closely tied to theirdomestic markets. They cannot escapethe economic effects of the local econ-omy on business conditions.

The products of global industrygroups are often commodities, anddemand for the products is global. Phar-maceutical companies are a classicexample of a global industry group. Sev-eral pharmaceutical companies arelocated in Switzerland, but little of theirproduct is sold there. They sell all overthe world. Technology companies areanother example; Microsoft first soldWindows 95 in New Zealand. Naturalresources are clearly a commodity: Theprice of oil is the same in London as it isin New York. The automotive industryis a classic global industry; automobilefactories are all over the world. Hondais the third largest auto manufacturer inthe United States and competes head-onwith General Motors Corporation andFord Motor Company.

Local industries are typically locallyregulated. The products are those thatcannot be shipped easily, and thedemand is local. Utilities are perhaps themost classic local business; all utilityregulation is local, and the product can-not go far. Banking is a classic local busi-ness. Most of a bank’s deposit base islocal, the regulations are local, and theinterest rate environment in which thebank operates is local. Construction istypically a local business.

The distinction between global andlocal industries can be used to manageportfolios. First, divide the universe intolocal and global companies using logicalcriteria—the competitive nature of theindustry and export dependence. Thisdivision of the universe permits logicalcomparisons among and within equitymarkets. Figure 3 illustrates theapproach using South Korea and Japan.An important business in Korea is gasdistribution. The government encour-ages businesses and individuals toswitch to the use of natural gas in order



to curtail pollution. In making a decisionabout whether to own, for example,Seoul City Gas or Samchully, two of thelargest gas distribution companies inSouth Korea, the framework is SouthKorea. An investor can compare thesetwo companies directly on the basis ofvaluations or any other factors. Theycompete head-to-head with each other,and there is no need to compare thosecompanies with utilities outside Korea.

On the other hand, if an investor isconsidering Samsung Electronics, one ofthe world’s largest manufacturers ofsemiconductors, the framework must beoutside Korea. For comparisons, inves-tors must go to other companies in thatsame industry—direct competitors ofSamsung, such as NEC Corporation,Texas Instruments, or Intel—irrespectiveof where they are located.

To find the true valuation of a localmarket, the next step in this approach isto recalculate each local market’s keyvaluation ratios excluding the equitiesof the companies in global industry sec-tors. For example, to find the local valu-ation of the Dutch market, one wouldstrip out Unilever, Royal Dutch, andPhilips Electronics N.V., then calculatethe valuation of the market.

Next, in making a choice betweenstocks, compare local companies with

the local market and compare globalstocks with global industry competitors.Comparing local companies can be asmuch an art as a science, because everylocal market has some industry or sectorthat is unique to that market. For exam-ple, no such thing as a plantation sectorexists outside the Malaysian equity mar-ket. Such sectors give a unique flavor toeach market.

Comparing global stocks within glo-bal sectors is the final step. As discussed,it is a natural, logical way of pickingamong global stocks. Look at the indus-try around the world. Do not worryabout where the companies are located;look for the company that is most com-petitive and has the best valuation rela-tive to its worldwide competitors.

Conclusion

Investors should base global countryand industry decisions on the true natureof the respective markets: Separate thelocal industries from the global indus-tries; carry out comparative analysesonly within the appropriate group (localversus local); and value local marketsexcluding the companies that belong toglobal industry groups. We believe thatthis approach leads to relevant and use-ful insights about companies’ relativevaluations.

Figure 3. Comparing Local versus Comparing Global Equities

Japan

Gas Distribution

South Korea

Semiconductors

SeoulCityGas

SamchullySamsung

ElectronicsNEC

GlobalIndustry

GlobalIndustry

LocalIndustry


Stannard.frm Page 20 Wednesday, November 12, 1997 2:57 PM

Benchmarks for Global PortfoliosJohn C. Stannard, CFAFrank Russell CompanyLondon, United Kingdom

The importance of portfolio benchmarks as tools for evaluating the success of an investment strategy should not be overlooked. In particular, benchmarks play the vital role of defining a portfolio’s neutral or low-risk

position. A good benchmark is, first of all, suitable for the strategy being measured; it is also clear, simple, understandable, and unambiguous. In addition to a

discussion of indexes as benchmarks, this presentation describes the factors to be considered in benchmarks

for pension funds with multiple managers, hedged benchmarks, and style benchmarks.

he importance of a portfo-lio benchmark in the mea-surement of an investmentprogram is often under-rated. In defining a newmandate for a manager, the

choice of benchmark is often left until last,at which point both parties are perhapssatisfied to settle for some convenientsolution—for example, using a bench-mark that everybody else uses. A choicemay be made irrespective of the nature ofthe portfolio mandate or the special riskfactors or constraints involved.

Such automatic benchmark choicescan be a mistake. The benchmark isextremely important—not only to theclient and the manager separately butalso to their mutual understanding ofthe contract between them and thefuture success of their relationship. Theportfolio benchmark defines the neutralposition of the investment strategy (interms of market risk), and because mostbenchmarks can be reproduced in a pas-sive portfolio, the benchmark forms thepassive low-cost alternative to activemanagement. Therefore, benchmarksneed to be chosen with care. Moreover,

anyone using a benchmark (the focus inthis presentation is on global bench-marks in particular) needs to under-stand the individual characteristics ofthe various index options and how theirconstruction will affect their effective-ness as benchmarks. This presentationcontains a review of what makes effec-tive benchmarks in theory and a discus-sion of three benchmarks in practice.

Benchmarks in Theory

For their investment benchmarks, fundsand managers should avoid using aseries of targets, each of which might bequite valid on its own but which, likerandom signposts, all point in differentdirections. This type of benchmark struc-ture neither defines an effective strategynor helps in decision making. Rather, aneffective benchmark might be defined as“an independent hurdle rate, forming anobjective test of the effective implemen-tation of an investment strategy.” Inother words, the benchmark sets a hur-dle, and meeting the benchmark perfor-mance means achieving that hurdle andcarrying out a successful strategy.

T



Note that what is being measured bythe benchmark is “an investment strat-egy.” So, the fund or manager must havea strategy that has been logically thoughtthrough and must clearly understandwhat the strategy is. A good benchmarkneeds to be relevant to the strategy andto the markets in which it is applied.

The next need is an “objective test” ofwhether the strategy has been success-ful. The hurdle concept is useful here. Ina hurdle race, a runner can still win therace if he or she knocks down a coupleof hurdles. But the chance of winning issmall if the runner knocks them alldown. Investment managers hope tojump over all the hurdles, realize theywill sometimes miss, but know they can-not always miss and still win the race.

To be a good objective measure, abenchmark must be clear, simple,clearly understood by all who are usingit, and unambiguous. That last point isparticularly important: If two peoplecan arrive at different interpretations ofthe same benchmark, problems arebound to follow.

Benchmark StructuresThe terms “index” and “benchmark”

are often used interchangeably, but theyare not the same, and an understandingof the difference is important. An indexis a mechanism for showing how a mar-ket has changed based on, say, the pricesof the underlying securities. A bench-mark is a tool for evaluating the successof a process. Indexes are often used asbenchmarks, but benchmarks may notalways make valid indexes. Benchmarksare often derived from a standard mar-ket index. Sometimes, managers add avalue-added target rate above the indexreturn—that is, a target of the index plussome percentage. For example, a pen-sion fund might view an investmentmanager as an aggressive, concentrated,high-risk manager, so it may require themanager to beat the index by 1.5 percent.Selecting a value-added target is animportant subjective decision, and bothclient and manager should think realis-

tically about the practicality of such ahurdle rate. For instance, 1.5 percentdoes not sound like much, but in fact, itis not an easy goal for even a quite con-centrated portfolio to meet. A 2 percenttarget is demanding, and few portfolioswill consistently outperform a bench-mark by 3 percent. Also, when settingthe percentage, client and manager needto think about the amount of risk in theportfolio and how the portfolio fits intothe overall strategy.

Another facet of benchmark construc-tion is the weighting of individual com-ponents of the index (stocks, industrygroups, etc.). An investment benchmarkmight be a market-weighted compositeindex (i.e., weighted based on the mar-ket capitalization of each stock). Alterna-tively, it might be weighted by someother factor such as GDP (say, by coun-try) or use some simple equal-weightingformula. In the U.K. market, several pen-sion funds judge themselves against anaverage fund proxy, and indexes havebeen built on the basis of weightings ofthe average fund in the market. Suchindexes can indicate how well a man-ager is doing relative to the aggregate ofall the other funds in the market but donot really take into account special fac-tors, such as the pension fund’s liabilitystructure, unique constraints, or objec-tives. This limitation is becoming morerecognized in the United Kingdom, andas a result, the industry is movingtoward a more U.S.-like model, in whichindividual pension fund benchmarksare based on objectives, liabilities, and soforth.

Finally, a benchmark might be simplysome reasonable representation of thepeer-group return. Comparing againstthe peer group gives an objective test ofhow well a fund is doing in relation toother international bond or equity port-folios.

Sometimes, a fund has two bench-marks. In this case, ensuring that thetwo benchmarks are not conflicting isimportant. For example, in one situationI came across, the pension fund gave its



managers the joint goals of beating theMorgan Stanley Capital InternationalWorld Index when the index is going upand beating cash when the MSCI isgoing down. In principle, that objectivesounds reasonable, but in practice, it ledto compromise and did not motivate theintended behavior: The manager carrieda relatively high cash weight, presum-ably to cushion performance in theevent that the market did go down. Asa result, the fund tended to underper-form the index for long periods and thestrategy was, by definition, unsuccess-ful. For this particular pension fund,leaving the cash objective out wouldhave made more sense. The inherentvolatility of international equities is bet-ter balanced by reasonable diversifica-tion elsewhere in the portfolio (i.e., aspart of the strategic asset allocation deci-sion) than by some unrealistic notionthat a manager can, with perfect fore-sight, suddenly liquidate an equity port-folio immediately ahead of a market fall.

In some cases, a dual benchmarkmakes sense. For example, adding apeer-group benchmark to an “index-plus” hurdle would indicate that beat-ing the competition is the secondaryobjective. That approach is sensible; it iswhat a manger is trying to do anyway.So, primary and secondary benchmarksare workable as long as they are reason-ably homogeneous in terms of the sam-ple and are leading the strategy broadlyin the same direction.

Benchmark structures need to weighsimplicity against perfection. The mar-ket may well be very diverse, and themanager may want to include every sin-gle stock in the benchmark, but in prac-tical terms, such a goal may not beachievable. In most cases, managers andclients can settle for reasonable expo-sure to a market and an index that hassensible construction rules (rules thatdeal with new issues, income, and soon). If no index exists, creating one maybe worthwhile, but that situation willnot often arise. If multiple indexes areavailable, the manager needs to sit back,

judge the strengths and weaknesses ofeach, and see how they fit the portfolio’sgeneral objectives.

Materiality is important. For exam-ple, even if a client fund has a spread ofliabilities across the world, creating anindex that has 1 percent exposure hereand 2 percent exposure there is imprac-tical. At the end of the day, those smallpercentages within the overall bench-mark are irrelevant in terms of the totalreturn, but they will significantly com-plicate the benchmark calculation. Fur-thermore, they are difficult for a fundmanager to match on any kind of prac-tical basis. Less than a 5 percent expo-sure is generally not worth including ina fixed-weight benchmark.

The portfolio benchmark cannotreflect every minute component of amanager’s strategy. Here again, materi-ality (the law of large numbers) comesinto play. A manager with a multifac-eted strategy needs to pick the key com-ponents and reflect those in thebenchmark. Of course, the aim is to beas accurate as possible, but a perfectbenchmark is of little benefit if it isimpossible to interpret and use.

In choosing a benchmark, managersand clients need to make sure it fills theirbasic needs. Instead of simply followingthe herd and using a benchmark becauseeverybody else does, managers need toreview and question the purpose of thebenchmark. The aims are to not compro-mise the strategic objectives and not addthe benchmark as an afterthought.

Index CharacteristicsEven though indexes and bench-

marks are not precisely the same, anindex often forms the vital ingredient ina benchmark. When choosing an indexfor a benchmark, therefore, managersand clients want the index to be as fol-lows:

■ Practical. An index needs to bepractical, in that it provides a basis thatadequately reflects the portfolio man-date and provides a manager withsomething the manager is capable of



beating through good investment man-agement.

■ Investable. A manager must be ableto buy the securities that make up theindex. The question of investability isparticularly important in the emergingmarkets, but surprisingly, many devel-oped markets also have fairly low pro-portions of investable securities. So,investability is a factor in all indexes.

■ Representative. An index should beas representative as possible of theunderlying market. Factors such as mar-ket capitalization and industry groupcoverage are particularly important.

■ Complete. In some markets, includ-ing every security is simply not possible,but the more complete the index, themore useful it is.

■ Widely recognized. Sometimes,broad recognition of an index is betterthan perfection in the index. A perfectindex that has just been developed maybe attractive, but if nobody else is usingit, gaining acceptance of the index, aswell as gathering the underlying infor-mation, may be difficult. Following theherd is all right as long as the herd isgoing in the right direction. But manag-ers and clients must always be cautiousof following the herd. Be prepared tostop and sniff the wind now and againto make sure you and the herd are stillgoing in the right direction.

Index ConstructionSome of the issues to keep in mind

when considering indexes are weight-ing, frequency of reconstitution, howincome and new issues are handled, andease of replication.

■ Weighting. Is the index capitaliza-tion weighted or equal weighted? To saythat a capitalization-weighted index ispreferable may seem obvious today;portfolios perform on a cap-weightedbasis, so a cap-weighted index is themost objective test of how well a portfo-lio has done. Many examples still exist,however, of indexes that are notweighted according to capitalization.One such index is the Dow Jones Indus-

trial Average, and it is used extensivelythroughout the world as an indicator ofmarket performance.

■ Reconstitution. In theory, an indexwould be reconstituted every month oreven every week, but in practical terms,such frequency is impractical. If recon-stitutions are too frequent, the indexitself will be constantly changing, andstocks may move into the index andthen move back out again. Such move-ment causes severe practical problemsfor both passive managers trying totrack the index and active managersusing the benchmark as a performanceproxy. So, frequency of reconstitution isa question of balance. Most indexes arerebalanced about once a year.

■ Income. This issue is relativelyminor compared with weighting andreconstitution, but how income is han-dled is still important. If income issubstantial—as it is for a bond index—and it is reinvested quarterly, themethod for recognizing income andincluding it in the total return can havean impact on the effectiveness of theindex as a measurement tool.

■ New issues. In the normal course ofevents, new stock or debt will be issuedand must be incorporated in the rele-vant index. The index’s rules for treatingnew issues are important. Once again,the best solution is a balance betweenaccurately reflecting changes in theunderlying market, on the one hand,and creating significant turnover andadministrative hassle, on the other.

■ Ease of replication. This factor isbecoming more and more important asindexing increases in the emerging mar-kets. Ease of replication is vital to bench-marks for indexing.

An examination of European equityindexes highlights some of the importantaspects to be considered when choosingindexes for benchmarks. Although thisstudy focuses on Europe, a similarreview would apply to any market andany strategy. Table 1 provides someinformation on the European subindexes



of MSCI, the Financial Times/Standard& Poor’s Actuaries World Index (FT/S&P-AWI), and the Salomon BrothersBroad Market Index (BMI). The tabledemonstrates that indexes can be moredifferent than one might expect. Notethat the BMI Europe is much bigger thanthe MSCI Europe in terms of the numberof stocks in the index. The relative sizesof these two indexes comes as a surpriseto most people. Of course, many of those1,800 securities in the BMI Europe aretiny (in terms of market exposure, 1,000of them may make up less than 5 percentof the index), but they do have someimpact. The FT/S&P-AWI Europe isroughly in between the other two.

Given the differences among indexes,how does a fund choose the appropriateindex for a benchmark? It depends onthe fund’s strategy. For example, if theindex is to be used to benchmark a glo-bal index fund, the fund will want theindex to have as broad a coverage aspossible. If the fund is using a samplingmethodology based on an index, thenthat sampling methodology will itselfcause some error relative to what themarkets are doing. In that case, the mostappropriate index for the fund might bethe BMI. If the fund simply wants someexposure to the global markets in itsbroader asset allocation, then the MSCIEurope is probably fine.

The choice of an index may also relateto what other managers in the region areusing. The MSCI Europe has a longerhistory than the other two indexes andis widely used in the United States,

whereas the FT/S&P-AWI is the morepopular index in Europe and some otherparts of the world. Many index choicesare driven by history. The BMI, which isa very “pure” index in terms of coverageand investability, is not used nearly aswidely as the others.

Available float, the amount of stock“available” to investors, can affect evenmarkets that analysts do not perceive ashaving an issue of cross-ownership (onecompany holding another’s stock), andthe amount of float can have a majorimpact on the country distribution fromone index to another. Figure 1 showsavailable float for Europe and the majorcountries included in European indexes.Most stocks are freely available in theUnited Kingdom; some governmentholding and cross-ownership exists, butnot much. On the other hand, the freefloat of less than 70 percent in Germanyand in France is a surprise to most peo-ple. The reasons are somewhat differentin the two countries. In Germany, thereason is mainly cross-ownership. Deut-sche Telecom, for example, has onlyabout 26 percent free float, and it is oneof the largest stocks in the market. InFrance, the lower available float is morea result of government holdings. So,indexes that adjust for cross-holdings,such as the BMI, tend to have more cap-italization exposure in the United King-dom than in Germany or France, as thedistribution of countries presented inFigure 2 shows. Therefore, a fund thatwants to use a regional or a global indexas a benchmark needs to know how

Table 1. Comparison of Indexes

MSCI EuropeFT/S&P-AWI

Europe BMI Europe

Number of stocks 574 718 1,824Total capitalization

(trillions) US$2.9 US$3.4 US$4.0Inception date 1969 1986 1989Weighting method Capitalization Capitalization Free floatStock selection Sample Sample All stocksMinimum capitalization

(millions) — US$100 Free float > US$100Target capitalization capture 60% 81% 95%



such issues as cross-holdings are han-dled in order to judge the suitability ofthe index for a particular need.

Benchmarks in Practice

As this discussion about applyingbenchmarks unfolds, keep in mind thepoints made so far: First, indexes arenot, by themselves, benchmarks; theyare the building blocks of benchmarks.Second, the benchmark is a hurdle rate,something the fund or manager expectsand hopes to achieve for the long termeven though it will not be achieved fromtime to time. Third, the benchmark must

be practical, unambiguous, and mean-ingful.

This section considers three bench-mark examples—a benchmark for apension fund with multiple managers, ahedged benchmark, and a style index.

Multimanager Pension FundA multimanager, multiasset pension

fund will illustrate the use of bench-marks in a pension fund. This fund isabout as complicated as one might find.It has specialist managers, which is thenorm in the United States and is becom-ing popular in other markets. It includes

Figure 1. Available Float

Eur

ope

9080706050403020100A

vaila

ble

Shar

e C

apit

al (%

)

Uni

ted

Kin

gdom

Net

herl

and

s

Irel

and

Swit

zerl

and

Finl

and

Den

mar

k

Ger

man

y

Fran

ce

Aus

tria

Spai

n

Bel

gium Ital

y

Swed

en

Nor

way

Figure 2. European Country Distribution

MSCI EuropeFT/S&P-AWIBMI

0

United Kingdom

The Netherlands

Switzerland

Germany

France

10 20 30 40 50



balanced managers: Balanced manage-ment is the predominant way of measur-ing money in the United Kingdom andother major markets but is not so com-mon in the United States. The pensionfund also uses index funds. Finally, thefund has a centralized process by whichit sets the asset allocation for the fund.That is, the fund itself has performed anasset/liability study, determined its risktolerance, objectives, and so forth, andhas as a starting point a “strategic bench-mark” that reflects its preferred alloca-tion in terms of the various asset classesavailable for investment.

Suppose we are professionals work-ing for this pension fund and we aretrying to (1) build a framework for eval-uating decision making in the fund, (2)set benchmarks in a consistent wayacross the fund, and (3) evaluate ourstrategic decisions. We start by identify-ing what we want to know to evaluateour decisions. Clearly, we want to knowthe portfolio return. We also want toknow the value added by our managers:Are they achieving the added value thatjustifies the active management fees weare paying? (It is at this point that thenotion of a hurdle rate enters the pic-ture.) Most importantly, we want toknow the impact of each portfolio’s per-formance at the total fund level.

Taking this sort of overall approachavoids bogging us down in details aboutrelatively unimportant portions of thefund—portions representing, say, lessthan 2 percent of assets. Materiality isimportant. Even though small expo-sures (e.g., venture capital ) can be use-ful in a pension fund, the importantpoint is not to spend 50 percent of ourtime evaluating something that is only 2percent of the assets. When we look atour individual portfolio returns,whether they are in balanced or special-ist portfolios, we want to think in termsof the total value added at the total fundlevel.

Next, if we have a rebalancing policy,we want to know if it is working: Are werebalancing too often, not often enough?

Do we have ranges on either side of ourtarget allocations? Are they too narrow?Are they too wide? Do we hold toomuch cash in the fund (our benchmarkprobably holds no cash). Finally, at theend of the day, we want to know howwell we are doing relative to other fundsin the marketplace. We may be unhappywith our relative performance; we maybe happy to be “underperforming”other pension funds because we have adifferent liability structure; we may behappy that we have more in bonds, lessin equities, and so forth. At any rate, weneed to know how our performancecompares with other pension funds.

The result of dealing with these stra-tegic questions is a multilevel frame-work for analysis. Each question fitsbroadly into one level of this frameworkas follows:

■ Total fund level. At the top of thepyramid, our analysis involves howwell we have done in aggregate com-pared with our strategic benchmarksand (if of interest) other pension funds.

■ Broad asset-class level. If we have abalanced manager at this level, we eval-uate how well that manager has done inasset allocation decisions, how well inselecting within the individual sectors,and the aggregate added value. At thislevel, we can also compare the broadcontribution of the various segments ofthe fund (specialist, balanced, indexed,etc.) and measure the impact of rebal-ancing policy.

■ Portfolio level. Finally, at the bottomof the analysis pyramid is the portfoliolevel. The analysis required at this levelinvolves, for example, the specialistmanagers with local or regional assign-ments—for example, a domestic equityportfolio.

At each level, by comparing the port-folio return against our benchmark,(perhaps breaking value down by usingattribution techniques), we can find outwhether value has been added andrelate that result back to the originaldecisions made within the investmentprocess.



Suppose the following were ourresults for a quarter:

Our policy benchmark is just under 2percent, so the total fund return showsthat we have added value to the extentof about 50 basis points. Of that return,before breaking out currency in ourattribution analysis, 0.3 percent camefrom differences in asset allocationbetween the portfolio and the bench-mark and 0.2 percent came from securityselection (basically, buying good stocksand bonds that allowed us to beat theassigned index). These results are goodto know, but what is going on below—that is, within the fund? Is the allocationa result of the balanced manager? Do weperhaps have a tactical asset allocationmanager that we need to measure? Isour rebalancing policy effective? At theselection level, which managers havecontributed to the return? Do certainmanagers have particular strengths andweaknesses?

In order to answer these questions,we need not only a benchmark at thetotal fund level but also a series ofbenchmarks below that level that areconsistent with the total fund. To mea-sure a complete fund in this way, there-fore, we need to build a workable“benchmark structure.” Such a bench-mark structure is provided in Table 2,where we have defined for each of the

broad segments a target asset allocationmix consistent with the total. This pro-cess is complex but not difficult. Oncewe have it in place, we can identify thestrengths and weaknesses in the portfo-lio and where the value is being added.The key to success is to ensure that ourbenchmark structure at the total level isconsistent down through all the levels.Once it is, then we can start to answerthe questions noted.

A practical issue arises if we are look-ing at, say, a European index as the aggre-gate fund benchmark but are measuringcertain countries within Europe sepa-rately. For example, we might pick amore broadly based market index for theUnited Kingdom (such as the FT/S&P-AWI), because it is a relatively large mar-ket in Europe, than we would pick forEurope as a region. If so, we will have amismatch between the specialist man-ager level and the total level that we needto be aware of. Such a factor might notinvalidate our analysis, but it is impor-tant to be aware of the impact such dualbenchmarks may have.

Hedged BenchmarksHedged benchmarks have evolved in

the past 5–10 years as a result of researchconducted on long-term currency move-ments plus the need of some investors tocreate portfolios that are fully hedgedagainst certain base currencies. One con-clusion of the research is that over 10–20years, and taking into account purchas-ing power parity, if currency premiums,

Total fund return 2.4%Policy benchmark 1.9Management effect 0.5

Allocation 0.3Selection 0.2

Average fund return 1.7

Table 2. Total Fund Level

Policy Benchmark Weights

Balanced Specialist Index Property Total

Weight in fund 30% 30% 30% 10% 100%

Equities 55% 40% 70% — 50%Bonds 15 30 — — 14Pacific Basin equities 15 15 15 — 13European equities 15 15 15 — 13Property — — — 100 10

Total 100% 100% 100% 100% 100%



the short-term fluctuations, and so forth,are eliminated, the net return to cur-rency is about zero. Some practitionersargue, therefore, that investors shouldnot take on currency risk; others arguethat if investors do not take on currencyrisk, if they completely hedge that riskaway, they may miss opportunities inthe short term. The swings that occur incurrencies can be dramatic, so those U.S.investors, for example, who are totallyhedged back into dollars will underper-form significantly when all the othercurrencies appreciate. In short, a mas-sive debate has been going on for sometime about currency hedging. At anyrate, those who do decide to hedge stra-tegically will need a hedged benchmark.

Suppose we want to build a hedgedbenchmark. The first step is to focus onthe goal of evaluating the portfolio with-out the currency risk. At first, we mightthink that the approach is simply toeliminate the exchange rate movement.An important part of building a hedgedbenchmark, however, is recognizing thatwhen we hedge, we cannot hedge awaythe complete movement in exchangerates because of interest rate differentialsbetween markets. A second point is thatin any hedging exercise, we need tohedge back to some base currency.Hedged benchmarks have to beuniquely calculated by country; if wehave, say, the JP Morgan World BondIndex hedged back into dollars and theJP Morgan Index hedged back into Ger-man marks, we cannot convert one tothe other without going back to firstprinciples. Additional issues to considerwhen building a hedged benchmark are(1) whether to be totally or partiallyhedged and (2) the impact of forward

premiums on the hedged return. Sup-pose we have a portfolio of hedged yenbonds; that is, we have used either theSalomon Brothers or the JP MorganIndex and hedged a U.S.-dollar-basedportfolio against the yen to eliminatehalf the currency exposure. Table 3shows how this hedged benchmarkworks, illustrates some of the interac-tions that occur within a hedged portfo-lio, and shows how difficult interpretingeven a simple hedged portfolio is.

The first point to make is that thehedge is 50 percent. We have decidedthat we are concerned about the yenvolatility, so we have created a positionin our policy benchmark to hedge halfof that volatility away. Our benchmarkposition is 24 percent, and we sell yenforward at the rate of about 12 percentof the portfolio. The offset weights of theyen forward sale and the matching U.S.dollar forward purchase are the same,but we expect the yen to fluctuate rela-tive to the dollar. So, we will end up withan unrealized gain or loss in yen, whichwill offset some of our currency expo-sure, roughly half of it.

The approach looks fine in theory, butin practice, a “forward premium” existsbecause of differential interest ratesbetween Japan and the United States.Because interest rates in Japan are signif-icantly lower than in the United States,the expectation is that the yen will appre-ciate relative to the U.S. dollar over time.Even if it does not, the forward currencymarket will assume it is going to. Whenwe, as U.S. investors investing in Japan,hedge, we cannot hedge away the antic-ipated interest rate change, which willgive rise to a difference between the yen

Table 3. Hedged Benchmark

Factor Weight Return

Hedged yen bonds 24% –0.9%Yen bonds 24 –3.2a

US$ foreign exchange (purchase) 12 0.0Yen foreign exchange (sale) –12 4.6aYen currency return is –3.4%.



currency return and the return on a yenforward currency position. So, if we holda forward currency contract, if it was apurchase position, we will lose money tothe extent of 4.6 percent—which is sig-nificantly more negative than –3.4 per-cent. The difference is the premium. Ifwe are holding a long position, that pre-mium is a bad thing; if we are holding ashort position, it is a good thing becausewe make more money from it. Most ofthe negative return is from the currencyeffect; only about 0.2 percent is positiveappreciation of the bonds themselves. Inshort, we are hedging away some of thedownside risk in the yen. This hedgingincreases the overall return (because theyen is weak) and has a bigger impactbecause of the premium. That reductionis more than one might expect frommerely taking away half the currencyexposure. Indeed, if we had hedgedcompletely, we would have mademoney from our hedge because of thepremium.

This example may imply that ahedged benchmark is fairly straightfor-ward; with multiple-asset portfolios inmultiple markets, however, it becomesa lot more complicated. Managers beingevaluated against a hedged benchmarkneed to be aware of the benchmark andunderstand how it affects their decisionmaking. As always, the benchmarkneeds to match the strategy.

Style IndexA style index is simply a mechanism

for evaluating a particular kind of strat-egy that may be a subset of a broad mar-ket strategy. Based on their beliefs andtheories, various types of investors pickstocks based on such characteristics asgrowth versus value and/or capitaliza-tion sizes. By defining indexes that repre-sent those different types of stocks, fundsor managers can produce evaluation toolsfor these managers. In general usage, theterm “investment style” describes vari-ous methods of investing in or selectingstocks. In this presentation, however, thefocus is on indexes representing growth

and value investing.Choosing the characteristics for a

style index exemplifies the need to bal-ance intellectual purity—identifying anindex that precisely represents a particu-lar investment management process—against the need to be practical. Supposewe want to differentiate between theperformance of our growth managersand that of our value managers. First,some stocks may well be neither growthnor value, but we have to put themsomewhere; otherwise, we end up withtwo indexes that do not add up to thewhole. Second, many, many differentkinds of growth disciplines exist. Fewerkinds of value discipline exist, but theymay also be different in nature. No sin-gle index will represent every single dis-cipline. Thus, we need to compromiseand be practical in this situation. Weneed an index that is reasonably com-plete and representative but also practi-cal. For differentiating value andgrowth stocks, the Frank Russell Com-pany recommends using the ratio ofprice to book value. P/B is not a perfectmeasure, but it does provide certainadvantages: It broadly reflects earningsgrowth in a company and the extent towhich that earnings growth is rein-vested; it is broadly stable over time; itis readily available; and it applieswidely across markets.

Some other characteristics can beused in individual markets. For exam-ple, dividend yield is a good indicatorof style in the United Kingdom; it iden-tifies divergence in performance of low-yield portfolios relative to high-yieldportfolios. It does not fill the bill in othermarkets, however, and to provide con-sistency for analyzing global portfolios,using a consistent approach makes thebest sense.

The construction methodologyapplied by the Frank Russell Companywhen using P/B as the determinant of anindex begins with sorting all the stocksby descending P/B. Then, as the initialstep, we divide the index in half by mar-ket cap and identify the median P/B



number. At this stage, all the stocksabove the median are growth stocks andall below are value stocks.

This approach is fine in theory, but inpractice, we run into ambiguous situa-tions. For example, British Petroleum isa huge stock, and it is right on the borderbetween value and growth. In addition,although British Petroleum is an oil com-pany, it is also a conglomerate of differ-ent businesses that, individually,probably show growth and value char-acteristics. To arbitrarily define such aconglomerate company as either growthor value in the absence of a clear indica-tion based on P/B would be a leap offaith. So, the Russell methodologyaccepts that stocks near the median maybe part growth and part value. We use asmoothing methodology to define theboundary between growth and valuestocks. Consider the stocks, like BritishPetroleum, with P/Bs slightly above themedian. As the P/B falls, the valueweight gradually increases; as the P/Bapproaches the median, the value andgrowth weights both approach 50 per-cent; then, as the P/B falls below themedian, the value weight continues toincrease. At the third quartile break andbelow (that is, low P/B), the stock isdesignated as being 100 percent value.

This approach also avoids what isoften referred to as the “whale in thebath” syndrome. Take our British Petro-leum example. Suppose the price ofBritish Petroleum sharply deterioratesand it’s P/B consequently suddenlyfalls below the median. Without the

smoothing methodology, this whale of astock would suddenly go from growthto value and, as a result, several (maybemany) “minnow” stocks would beforced the other way. Such a result callsinto question the practicality of theindex structure and makes the indexvery difficult to replicate.

Conclusion

The three most important points aboutchoosing a benchmark are as follows:

First, everything about the selectionof the benchmark must be driven by thestrategy it is to measure.

Second, a benchmark provides abasis on which to make good decisions.Investing is about evaluating risk andreturn; the investment process providesa framework for that evaluation; and abenchmark is an instrument for evaluat-ing risk and return. If managers are tomeasure the impact of their decisionseffectively, they need the frameworkand instruments to be consistent withthose decisions—which brings us backto the first point: Consistency betweenthe decisions and the benchmark isessential if the benchmark is to be effec-tive as a tool in decision making.

Finally, the first point must be bal-anced by the need to chose a benchmarkthat is realistic and practical. To build acomplex benchmark that is intended tobe philosophically perfect may betempting, but a complex benchmarkmay be very difficult to use and may notmeasure the strategy effectively.


Tschampion.frm Page 31 Wednesday, November 12, 1997 2:57 PM

Implementing Global Tactical Asset Allocation in Developed Markets

R. Charles Tschampion, CFAManaging Director, Investment Strategy and

Asset AllocationGeneral Motors Investment Management Corporation

General Motors Investment Management Corporation is using four outside managers and an internal portfolio in

its approach to global tactical asset allocation. The internal GTAA portfolio is actively managed and driven

by quantitative models. The overall model combines traditional and not-so-traditional components to provide small increments of added value to benchmark returns

and to enhance risk control. In addition to providing details of the modeling, the presentation outlines how

GTAA is implemented to achieve the desired benchmark and TAA exposures.

eneral Motors Invest-ment Management Cor-poration (GMIMCo)believes that tacticalasset allocation (TAA)can be used effectivelyto manage a global port-

folio. Based on this belief, we are dedi-cating US$1 billion to global TAA usingfive managers—four external managersand an internal fund. We are also in theprocess of conducting a related “exper-iment” with an additional US$500 mil-lion of pension assets to be managedcollectively as a group portfolio by thefive managers. The challenge of thisexperiment will be to get the managersto work together to produce value notonly beyond the benchmark return,although that is the primary objective,but also beyond the returns from a naivestrategy of taking the collective bets ofthe five managers and merely applyingthem to the group portfolio. In otherwords, we are trying to achieve somesynergy. If this experiment succeeds, we

hope to take the signals, the informa-tion, and the insights from managingthis US$1.5 billion and use them todevelop value-added TAA exposuresfor the entire US$65 billion GM pensionfund. We intend to get together once aquarter with these managers and dis-cuss where they should place the tacticalbets against the benchmark.

We are only in the start-up phase ofthis experiment. In October 1996, wehired four outside managers to manageUS$200 million each in a global TAA(GTAA) mode. In February 1997, wefunded an internal portfolio with thesame amount of money. We do not know,of course, exactly how many managerswould be the best number; we chose fivemanagers to achieve the benefits ofdiversification without the redundan-cies of too many managers. Having 2firms provides no real diversificationpower; having 10 is apt to be redundant.Table 1 shows the correlations of excessreturns from March 1989 through Janu-ary 1996 between the internal manager

G



and three of the four outside managers(one did not have a sufficient trackrecord to examine correlations). The cor-relations are based on combining a back-test of manager results and real returnsfrom the start of the program. These lowcorrelations indicate that we have beenreasonably successful in choosing man-agers that do not perform like oneanother. In one pair, in fact—Manager 1and Manager 2—the correlation isslightly negative.

The balance of this presentation pro-vides details of the process GMIMCo isusing to manage the internal GTAAportfolio.

The Internal GTAA Portfolio

The GMIMCo approach for its internallymanaged GTAA portfolio is active anddriven by quantitative models. Webelieve we have put in place anapproach that will not be taking anyextreme departures from the benchmarkbut, instead, will take many little posi-tions that will have a good informationratio over time and thus add reasonablysignificant value. The models we areusing rely heavily on the theoretical con-cept that an equilibrium risk premiumexists in stock and bond markets in thelong term. That is, the models assumethat the markets are basically efficient inthe processing of risk and return overthe long run but may deviate from equi-librium in the short run and, therefore,provide opportunities to add value ifone can identify those deviations. Themodels also assume that the equilibrium

levels are not constant over time, thatthey change. So, the process is dynamic;the models try to incorporate new infor-mation and new risks in terms of estab-lishing that equilibrium level. We alsobelieve, in line with the capital assetpricing model (CAPM), that higher-riskinvestments will have higher expectedreturns and higher actual returns overtime.

In the GTAA process, the long-termequilibrium views on the market andour views on short-term disequilibriumare combined, based on our confidencein the views, into a single set of expectedreturns. This set is put into an optimizertogether with an asset covariance matrixand certain constraints to produce finalmodel allocations. Figure 1 illustratesthe pieces that are combined to producethe integrated model.

ViewsThe process of generating the GTAA

positions begins with establishing viewson the various global markets. We arecurrently using four models to generateshort-term views—that is, whether dis-equilibrium exists in a market. The mod-els look at earnings yield gaps, earningsrevisions, foreign currency attractive-ness, and relative real bond yields. Theearnings yield gap model takes currentstock earnings yield less the long bondyield in each of the G–7 countries (Can-ada, Italy, France, Germany, Japan, theUnited Kingdom, and the United States)and compares that figure for each coun-try with the country’s history of that

Table 1. Correlations between GTAA Managers’ Monthly Returns,March 1989–January 1996

Manager 1 Manager 2 Manager 3 GMIMCo

Manager 1 1.00 –0.05 0.38 0.16Manager 2 1.00 0.07 0.28Manager 3 1.00 0.43GMIMCo 1.00

Note: Composite of simulated and live returns.



relationship. This simple model worksin all seven countries. We do not use themodel findings in cross-sectional analy-ses to determine which is the bettercountry to invest in. Instead, we take thecapitalization-weighted score of all ofthe country models, and if it favorsequity over bonds, then we tilt the over-all 60/40 mix more in favor of equities,and vice versa. So, the earnings yieldgap model is used to make the broadglobal stock and bond decision.

For intra-equity decisions—to calcu-late which countries we should over-weight or underweight—we use a cross-sectional model, the model of earningsestimates revisions. This model calcu-lates the ratio for each country of thefour-month cumulative number ofearnings estimates that were raisedversus the four-month cumulative num-ber of estimates that were lowered. Thismodel also has worked in all the G–7countries, but we have data for only sixand a half years, so our confidence in theresults is not as high as for the othermodels, which have more years of data.

The model related to relative real

bond yields first calculates the spread ofthe current real bond yield in each coun-try with respect to the median real yieldin existence in that country over the pasttwo years. These spreads are then exam-ined cross-sectionally to determine theweighting of the various countries’bond markets.

So, these three models deal with,respectively, the overall stock and bondmix, the country mix for the equities,and the country mix for the bonds.

The fourth model is a foreign cur-rency attractiveness model, which pro-vides a cost–benefit analysis of hedgingeach currency. We view the costs andbenefits in terms of the premiums anddiscounts an investor pays in hedging.For example, a U.S.-based investorwould capture a great premium rightnow by hedging the yen back into theU.S. dollar because of the interest ratedifferential. We compare that benefitwith the expected change in the spotrate, which we try to capture by compar-ing the real cash yields in the differentcountries.

We are currently using only these

Figure 1. GMIMCo GTAA Process Schematic

EquilibriumExpected Returns

Earnings YieldGap Model

Views

AssetCovariance

MatrixConstraints

OptimizationSingle Set of

ExpectedReturns

FinalModel

AllocationsEarnings Revisions

Model

Foreign CurrencyAttractiveness

Model

Relative RealBond Yield Model

ManagerViews

(Enhancement)

Bond Slope andReal Cash Model(Enhancement)

Black–Litterman Framework Optimization



four models, but as Figure 1 indicates,we are considering additional models ifthey provide the prospect for addingvalue. The two shown are possibilities:One involves putting in the views of theexternal managers who are managingagainst an identical benchmark; theother is a model to enhance the countryallocation for bonds based on relativeyield-curve slopes and real cash returns.

As with picking the members of agroup of outside managers or the secu-rities in a portfolio, we want pieces, ourshort-term view models, that are notgenerally going to behave like oneanother over time. If we are successfulin obtaining such diversification, whenthe models do give us the same signal,it is reinforcing, and when they give usdifferent signals, it is risk reducing.

In selecting models, we focus on fourbasic characteristics. First, we try tomake sure a model is simple and trans-parent—that is, that it has a direct link tocapital market and economic theory.These models are not regression models;they are basically models that deal witha factor, either in a time series or cross-sectionally, and allow the analyst tomake judgments about that kind of out-put.

Second, we look for models that havesubstantial predictive power. We didextensive out-of-sample tests of each ofthe view models and found that they do,individually, have predictive power.

Third, we look for models that haveconsistency over time and across bor-ders. Again, we found that the viewmodels generally apply to all the majordeveloped markets. The only exceptionis the bond market model, whichprovides useful signals in only six of theG–7 countries. The model does not workfor Italy, so we do not include Italianbonds in our selection set.

Finally, as noted previously, we wanteach model to add input to the overallprocess without creating much redun-dancy. Having additive models with lowcorrelations with each other reduces ourtracking error risk while maintaining

whatever alpha-generating potential wehave.

Integrated Optimization Our approach uses a combination of

classic optimization and a model devel-oped by Robert Litterman and FisherBlack of Goldman, Sachs & Company togenerate the final allocations. Figure 2depicts the flow in a traditional opti-mizer in Panel A and the flow in theBlack–Litterman model, a sort of reverse-engineered optimizer, in Panel B.

In a traditional optimizer, theexpected returns and the covariancematrix, which includes the volatilities ofindividual returns and the correlationsof those returns with each other, areplugged into the optimizer, which thenproduces optimal weights as output. TheBlack–Litterman model begins withdetermining what the market weightsare—that is, what the weights are in thebenchmarks; then, using the same cova-riance matrix used in the traditional opti-mizer, it produces the expected returnsas output. The model makes the assump-tions that the markets are in equilibrium,that the CAPM works, that people havepriced the market capitalization in themarkets correctly, and therefore, that thereturns the model produces must be theones the market is expecting.

The views from the four models, theasset covariance matrix, and the con-straints are then used as inputs to theoptimization to modify the equilibriumreturns generated by the Black–Litterman model, and this set ofmodified expected returns is the inputinto a traditional optimizer to generatetactical allocations. Naturally, the equi-librium expected returns with no viewsyield benchmark allocations; the viewsprovide the appropriate modificationsto expected returns to modify thoseallocations. For example, if the viewsindicate that Asset A will outperformAsset B by more than the equilibriumspread, the first part of the processwould raise the expected return of Aabove equilibrium and lower theexpected return of B below equilibrium.



The traditional optimization processwould thus shift the weighting towardAsset A and away from Asset B.

The incorporation of equilibriumreturns into our GTAA process providesa level of risk control not found in tradi-tional optimization. In a traditional opti-mization process, if our earningsrevision model, for example, said weshould overweight Japan and theUnited States and underweight Ger-many, we would raise our expectedreturns for Japan and the United Statesin the model and lower our expectedreturn for Germany and then run thoserevisions through the optimizer to findout what the weights would be. Becausethere is no tie back to global equilibrium,that process can come up with some out-rageous solutions; the optimizer may allof a sudden say it wants to be 80 percentin the United States, 20 percent in Japan,and totally out of Germany. With ourprocess, we never get to those extremesbecause the returns are modifiedtogether with the risks and covariances.Thus, running the single set of expectedreturns generated by this approach will

wind up tilting our allocations, not byextreme weightings against the bench-mark, but by 1–2 percent.

Also incorporated are some tradi-tional constraints. For example, we limitthe extent of our global stock/bond betsto no more than 15 percentage pointseither way; so, for example, we can goonly as high as 75 percent and as low as45 percent in equities against the 60/40benchmark. We also limit our departurefrom the 50 percent hedged position inforeign currencies of the benchmark to

20 percent; so, we can become as muchas 70 percent hedged or as little as 30percent hedged. In addition, we haveindividual country constraints: Cur-rently, we are prohibiting our managersfrom shorting markets and have set alimit of 10 percentage points around themarket weighting for the larger devel-oped markets and a limit of 5 percent-age points for the smaller markets. Wehave similar market constraints on thebond markets, and the individual cur-rencies cannot be more than 5 percent-age points from their benchmarkweights. The constraints provide risk

Figure 2. Traditional and Black–Litterman Optimization Flows

aCap-weighted benchmark equivalent to optimal weights under the CAPM.

ExpectedReturns

Panel A. Traditional Optimization Flow

Panel B. Black–Litterman Expected Returns Implied by Equilibrium

CovarianceMatrix

OptimizerOptimalWeights

BenchmarkWeightsa

CovarianceMatrix

OptimizerExpectedReturns

±

±

±



controls to our overall approach.

Integrated Model Backtest

We performed a backtest of the inte-grated model, totally out of sample, forDecember 1989 through September 1996and produced the results shown inFigure 3. Transaction costs would, ofcourse, erode these alphas, so we sepa-rately made estimates of what thetransaction costs of this process wouldbe. This estimation assumed the use offutures and country baskets to keepcosts low and assumed an overall cost of15 basis points (bps) annually merely tomaintain the overall structure. We esti-mated that our transaction costs wouldbe 25–30 bps a year on the overall fund,so that estimate should be kept in mindin viewing Figure 3. The 12-monthrolling line shows some periods, such asmid-1993, when results are substantiallynegative, but with a three-year timehorizon, the integrated model addedsignificant value in all periods.

This backtest was for only six and ahalf years, the longest period for whichwe had data for all the models. Some ofthe models that we backtested for longerperiods gave us confidence that we

would continue to find added value ifwe could backtest the integrated modelfor a longer period.

The average alpha reflected in Figure3 (keeping in mind the short period andthe absence of transaction costs) isapproximately 310 bps a year, with an expost tracking error in attaining thatalpha of about 180 bps, which is aninformation ratio of about 1.7. Taking asubstantial, 50–60 percent, haircut onthat alpha, we are targeting as a perfor-mance benchmark 150 bps in excessreturn with the expectation of 200 bps expost in tracking error. The resultinginformation ratio is 0.75.

Implementation

The first step in implementing theGTAA is to achieve the benchmarkexposures, then we pursue the TAAexposures we want. The matrix shownin Exhibit 1 notes the vehicles we use forimplementation by stocks and bonds onone axis and by U.S. domestic and inter-national arenas on the other axis.

We define risk in our internal GTAAprocess the same way we define it forthe outside managers: Risk is underper-

Figure 3. Integrated Model Backtest Results: Rolling 12-Month and 36-Month Alpha, December 1989–September 1996

Note: Returns are gross of transaction costs.

7

6

5

4

3

2

1

0

–1

–2

Ann

ualiz

ed A

lpha

(%)

12–Month Rolling

36–MonthRolling

12/89 12/9612/90 12/91 12/92 12/93 12/94 12/95



forming the benchmark. In the GTAAexperiment, each of the five GTAA man-agers is managing against the samebenchmark, which is 60 percent of theMorgan Stanley Capital International(MSCI) All Country Index for equitiesand 40 percent of the Salomon BrothersGlobal Bond Index for fixed-incomeinvestments. For non-U.S. assets, thisbenchmark is also 50 percent hedgedback into the U.S. dollar.

Thus, in implementing our strategy,we have to determine where to take therisk. We do so by setting a risk budget.In this case, our risk budget is the exante tracking error of 150 bps. We coulduse the budget 100 percent for takingthe models’ views and generating betsor positions that will potentially addvalue, or we could use part of the bud-get for tracking error associated withstrategies that lower transaction costsbut have basis risk against the bench-mark. We try to make the trade-offbetween tracking error and transactioncosts. For example, for our domestic orU.S. exposures, to achieve both the ini-tial benchmark exposure and the TAAexposure, we use S&P 500 Indexfutures because they cost little andhave low tracking error to the U.S.stock market. Even though the MSCI

U.S. equity component is not the S&P500, it is so closely correlated that weexperience little tracking error againstthat component.

Similar to U.S. stocks, a deep bondfutures market exists for U.S. domesticbonds. So, a hedge combining 5-, 10-,and 30-year bond futures can be used; ithas low transaction costs and a reason-ably low tracking error. The cost of abasket or a bond index fund to reducethe tracking error would be too high forthose vehicles to be used in the U.S.market for benchmark exposure. In thecase of U.S. domestic equities andbonds, futures also provide an opportu-nity to add value from cash manage-ment: Because these futures are pricedoff of LIBOR, if we can invest cash inthem to generate an incremental returnabove LIBOR, that strategy can addsome value.

On the international side, we usecommingled country baskets to achieveour benchmark positions. Futures arenot useful in non-U.S. markets becausemost of them track stock market indexesin the various countries, such as theCAC in France or the DAX in Germany,which do not match the individualcountry components of the MSCI well.Also, in some cases, the use of futures

Exhibit 1. Implementation

U.S. Domestic International

S&P 500 index futures (to gain benchmark exposure and for TAA)

Low transaction costs

Commingled country equity baskets (to gain benchmark exposure andplug holes)

Stocks Low tracking error High transaction costs Low tracking error

Commodity Futures Trading Commission (CFTC)-approved index futures (for TAA)

Low transaction costsHigh tracking error

5-, 10-, and 30-year bond futures (to gain benchmark exposureand for TAA)

Commingled country bond baskets (to gain benchmark exposure andplug holes)

Bonds Low transaction costs Medium transaction costsMedium tracking error Low tracking error

CFTC-approved bond futures (for TAA)Low transaction costsMedium tracking error



internationally is not possible because,in addition to the tactical bets we wantto make in the G–7 countries, we alsohave to gain the benchmark exposure toall of the other developed markets—Bel-gium, Spain, and so on— many of whichdo not have futures contracts we canuse. We use the State Street index fundsas country baskets because they havebeen designed to track the benchmarkswith low tracking error. The funds haverelatively high transaction costs, partic-ularly going in, but because we hold thebenchmark positions for the long term,when the cost of rolling over futures isconsidered, the funds quickly matchfutures in costs. We also use commin-gled country baskets to plug the “holes”between international equity and bond

positions dictated by our active modelsand the benchmark positions.

With respect to the tactical bets inter-nationally, we do use futures becausethe transaction costs of moving moneyin and out of the country baskets woulddefeat the GTAA alpha. The low cost ofthe futures makes up for the medium-to-high benchmark tracking error theyhave.

Conclusion

We are very excited about this approachto tactical asset allocation and believethat it will add value over time. We alsohope to have a successful outcome to thegroup portfolio experiment that we canreport in the future.


Abberley.frm Page 39 Wednesday, November 12, 1997 2:54 PM

Impact of the European Monetary Union on European Bond Markets and Portfolios

Paul A. AbberleyDirector and Head of Fixed Income

Lombard Odier International Portfolio Management Ltd.

EMU should broaden and deepen the European capital markets, improve liquidity, fill out the yield curves in European bonds, and encourage development of a

corporate bond market and the introduction of high-yielding government bonds and sophisticated

instruments. The major effects of EMU implementation will relate to changes in the relationship between European bonds and U.S. Treasuries, determining where the yield curve should lie, the relationship

between EMU and non-EMU European countries, and sovereign spreads. Portfolio managers are likely to see

a reduction in top-down investment management, a major increase in bottom-up opportunities, and the

emergence of a new analytical breed—the European fixed-income analyst.

he hotels in Europe mightbe a metaphor for the bondmarkets of Europe. Travel-ers to Europe who havestayed in any of the nicerhotels there know that

these hotels are elegant, some of themhave interesting histories, and they tendto have very nice dining facilities. But iftravelers ask someone to direct them tothe gymnasium or the pool for someexercise, they will get a look of blankamazement. The hotels in Europe sim-ply do not have such facilities. The Euro-pean bond markets also can be veryinteresting; some have long and fasci-nating histories; many have plenty ofcharacter. The one thing they lack is car-diovascular fitness. Compared withU.S. bonds, European bonds lack length,depth, and breadth. For that reason, asa fixed-income portfolio manager, I amencouraged by the prospect of Euro-

pean Monetary Union. Assuming thatEMU happens, the EMU process willgive the European bond markets theirown personal fitness trainer. The mar-kets should shape up in a healthy fash-ion, which will be very positive forfixed-income portfolio management.

EMU and Market Structure

The first impact of the EMU should beimprovement in the structure of theEuropean capital markets. The Euro-pean bond markets account for about 42percent of the global bond market, withthe U.S. market accounting for about 30percent of the remainder and the Japa-nese market, about 24 percent. At themoment, the European bond marketsare actually 15 individual markets, butclearly, as more and more countries jointhe EMU, that market could dominatethe U.S. and Japanese markets in size.

T



The increasing size of the marketshould have a positive impact on liquid-ity. Figure 1 presents a general view ofliquidity in 1996 by comparing the U.S.bond market with selected Europeanmarkets. In 1996, the liquidity in U.S.Treasury bonds was about double that inthe German or French bond markets. Inthe other markets shown—Belgium, Ire-

land, and the ECU (European currencyunit) basket—liquidity is greatly infe-rior. The hope is that a larger bond mar-ket for the euro under EMU will greatlyimprove liquidity throughout Europe.

The distribution of government debtacross maturities varies quite a bitbetween the United States and Europe.Figure 2 shows the distributions in 1996

Figure 1. Approximate Relative Liquidity in European and U.S. Bond Markets, 1996

Figure 2. Distribution of Outstanding Government Debt: United States and Europe, 1996

Ireland

220

200

180

160

140

120

100

80

60

40

20

0

US$

Mill

ions

BelgiumFranceGermanyUnited States ECU

10+

40

35

30

25

20

15

10

5

0

Perc

ent

7–105–73–51–3Number of Years

United States Europe



by the percentage of the debt in eachmaturity grouping. Note that theUnited States has considerably morebonds with durations of 10 years andlonger than does Europe. Moreover,Figure 2 does not do full justice to thedifferences, because the long end of themarket in Europe does not necessarilyfollow the rest of the yield curve. Onlycertain types of investors use longbonds, so holes exist in the yield curve.To some degree, 30-year bonds inEurope are commodities.

With EMU, the yield curves inEurope should fill out and the variouscountry markets should introducemore-sophisticated instruments—strips, for example. The key reasons are,first, that the individual countries willlose their captive savings pools. Theywill no longer be able to keep domesticsavings and direct them only to theirown government bonds, so they willhave to compete for capital. The prizesfor the winners will be substantial andwill come in the form of reducedborrowing costs. Second, everyone willwant to produce the benchmark yieldcurve, and the probability is that themore sophisticated the range of issues,the greater the likelihood of achievingthat benchmark status. These sorts offactors should help duration manage-ment and term structure management atthe portfolio level. Counteracting thesepositive factors is an initially negativeone: Europe will lose a range of bondmarkets immediately because many ofthem will be subsumed into the euro,which will reduce the flexibility possiblein portfolio management.

Compared with the United States, theEuropean bond market not only lackslength and depth, it also lacks breadth.Lacking the range of sectors that existsin the United States, the market is domi-nated by government bonds. EMU mayprovide scope for broadening the mar-ket. For example, the potential exists fordevelopment of a corporate bond mar-ket in Europe, although progress islikely to be slow. One of the constraints

on the development of a corporate bondmarket is the close relationship inEurope between corporations and theirbanks. Borrowing tends to be in the formof loans, partly because of these rela-tionships and partly because thespreads borrowers can obtain on loanstend to be much narrower than spreadsavailable in the corporate bond market.

Initial progress may come more in theform of securitization, including mort-gage- and asset-backed securities, thanin direct corporate bond issuance. Manyasset-backed bonds have been issued inEurope since 1995, and the trend islikely to continue when the euro is cre-ated. But EMU should also increase thedemand for corporate bonds. One of theproblems historically for corporatebonds in Europe has been the lack ofbuyers; many European bond investorshave been too conservative to buy cor-porates. That attitude also might changebecause of the development of morefunded pension schemes in Europe.Therefore, professional bond manage-ment should increase and pensionfunds should be interested in taking therisk to achieve the high yields affordedby corporate bonds.

Initially, municipal bonds may bemore likely to develop than corporatebonds. A major trend toward politicaldecentralization is occurring; variousregions want a greater say in their ownfuture and less reliance on the oldnation-state structure. The process isleading to disparity, however, betweenthe responsibilities some of the regionsachieve and the money they have tocarry out the responsibilities. They oftenget the increased responsibility they askfor, but because of fiscal constraints,they do not get the money and resourcesthey need. This outcome leads to agrowing need for debt financing, andthe bond market is the one place theseregions can turn to. Another factor isthat the role of traditional local financeproviders is changing. The environmentis much more competitive than in thepast, so the old European approach of



directing captive savings through tothese entities is changing.

If the municipal sector within the euromarket does emerge as expected, thequestion becomes how to analyze thosetypes of bonds. Given a range of munic-ipal bonds, how should the analystdecide where they ought to trade rela-tive to government bonds? Which oneswill offer relative value? Moody’s Inves-tors Service suggests a series of factors toexamine:• the structure of government in the

particular country,• the relationship between the munic-

ipalities and the central government(what sort of authority and powersdoes the municipality have?),

• the legal and institutional frame-work in which that relationshipoperates,

• economic and demographic factors(is the area wealthy or poor, andwhat sort of demographic structuredoes it have?),

• the political climate (is it a regionwith a stable political background,or is it a more contentious environ-ment?),

• budgetary dynamics of the region(what sort of responsibilities forspending does the locality have?),

• borrowing powers (does the localityhave the power to borrow, and if so,in what form?), and

• financial policies and management(does the locality have a track recordthat reveals whether the area hasbeen competent in its manage-ment?).

This type of analysis involves funda-mental examination of the issues—aprocess that is not typically carried outfor today’s European bond portfolio,which usually is managed through atop-down process.

The general economic reform in Eu-rope should produce a host of new high-yielding government bond markets—the markets of Eastern Europe. The dis-tinction being drawn in Europe betweenemerging bond markets and mature

bond markets is far too stark at present.Why should the Czech Republic be con-sidered an emerging market and beanalyzed completely separately from“mainstream” Portugal? I believe theanalysis of the emerging bond marketsof Eastern Europe will merge with theanalysis of mainstream European bondmarkets. In time, more markets could beadded—from the Czech Republic toHungary, to Romania, Croatia, the Slo-vak Republic, and Slovenia, to Bulgaria,Estonia, Latvia, and Lithuania, then toRussia, Ukraine, Belarus, Moldavia, andKazakhstan. The big unknown is wheth-er the new countries will want bondmarkets. Their macroeconomic funda-mentals imply that sovereign credit willbe available, but whether they will actu-ally issue bonds with a decent maturitythat could compete with the bonds ofmainstream Europe is another question.

EMU and Market Dynamics

EMU is likely to change the way mar-kets operate considerably. Key changesrelate to the relationship between Euro-pean bonds and U.S. Treasuries, deter-mining where the yield curve should lie,the relationship between EMU and non-EMU European countries, and sover-eign spreads.

Relationship with U.S. TreasuriesNothing is more frustrating than ask-

ing someone for a view on German bondyields, being told, “I think they will gohigher,” and when asking why, beingtold, “Well, I am bearish on Treasuries.”The response is all too common, and ofcourse, it is based on common sense:The U.S. Treasury market does domi-nate much of what happens in globalcapital markets. One reason may be that,depending on the time horizon exam-ined, correlations are increasingbetween various bond markets, particu-larly among the markets of the Organi-zation for Economic Cooperation andDevelopment (OECD). (Some research-ers question the idea that correlations



are increasing, but most practitionersbelieve that the markets are movingtogether.)

Macroeconomic sense also lies be-hind the influence of U.S. Treasuries.Nowadays, analysts are using a singlemacroeconomic model, one that every-one believes to be correct (and let’s hopethey are right!), so there is little argu-ment about what constitutes good fiscalpolicy, good currency policy, and soforth. Therefore, a convergence in realeconomic performance has been occur-ring. The range of inflation rates amongcountries is much narrower than everbefore. So, perhaps the primary key toknowing whether euro bond yields arelikely to go up or down is knowingwhether U.S. T-bond yields are likely togo up or down. Other factors, however,may also enter the picture.

The correlation between euro bondyields and Treasuries is liable to be lessin a few years than the correlation is nowbetween, say, German bonds and Trea-suries. The example of Japan is relevant

here. The Japanese bond market seemsto move on its own fundamentals; itapparently does not slavishly followU.S. T-bonds. One reason may be thescale of the Japanese market, and theeuro capital market will probably be suf-ficiently large that the direction of itsinterest rates also will be dominated byits own domestic determinants.

Determining Where the Yield Curve Should Lie

If the euro bond market does begin tomove separately, then global bond port-folio managers need to address wherethe yield curve will be in relation to theU.S. Treasury market. Figure 3 shows acomparison of the U.S. and Europeanyield-curve structures (with the ECUmarket as a proxy for the euro) prevail-ing in April 1996. The euro curve isclearly below the U.S. curve, but thisfigure does not reveal where equilib-rium lies across the cycle because it mustbe taken in the cyclical context, in that

Figure 3. Prevailing U.S. and ECU Yield-Curve Structures, April 1996

10

7.0

6.5

6.0

5.5

5.0

4.5

4.0

3.5

3.0

Yie

ld (%

)

7532 3016 months3 months

Maturity (years)

ECU Yield

U.S. Yield



the European economy was weak at thetime and, therefore, European bondyields would be expected to be belowU.S. yields.

Several analytical approaches sug-gest that the euro yield curve could welllie beneath the U.S. yield curve. Oneapproach is to recognize that nominalbond yields are simply the sum of threeparts—the real yield, the inflation differ-ential, and the risk premium—and com-pare those three parts between the twomarkets. This task is not easy, however.• Real yield. Some people argue that

the real yield in the United Statesand in Europe should be the samebecause the capital market is globalin nature and thus efficient. Otherscontest the efficiency of the marketand the idea that real yields shouldbe the same. Moreover, what is thereal yield?

• Inflation differential. To compare five-year euro bonds with five-year Trea-suries, one needs an opinion onwhat the inflation differential islikely to be over that period.

• Risk premium. Where the risk pre-mium is concerned, the Europeanoutlook is far less certain than thatof the United States.

My very subjective opinion is that thesethree parts taken together lead to a con-clusion that euro yields will be belowU.S. Treasury yields over the full cycle.

Another approach is to look at for-eign exchange—whether the euro as acurrency is likely to be weak or strong.If a portfolio manager makes a bet thatthe euro will be weak in relation to theU.S. dollar over time, then that managershould be rewarded with a higher inter-

est rate, and vice versa. To address thatissue from an economic viewpoint,Table 1 presents some comparative datafor Europe, the United States, and Japan.Assuming all the relevant countries areparticipating in the euro, the Europeaneconomy, with 39 percent of total GDP,will be the greatest of the three. Thesavings rate in Europe will be compara-ble with Japan’s and much higher thanin the United States. The debt-to-GDPratio (which reflects indebtedness) inEurope will probably average some-where between those of the UnitedStates and Japan. In terms of the currentaccount, another indication of how acurrency might move, Europe will havea surplus.

The current-account situation inEurope is generally healthy. As Figure 4shows, in recent years the Japanese sur-plus has been declining, the U.S. current-account deficit is seemingly boggeddown, and in Europe, the surplus hasbeen gradually climbing. Our view isthat the healthy current-account situa-tion for Europe will support the euro.Keep in mind again, however, that thisfactor is cyclical. With a relatively weakEuropean economy, Figure 4 is probablyflattering to the current-account situa-tion in Europe.

Another aspect of judging where theyield curve should lie is the potential ofthe euro to become a reserve currency.Logically, the euro ought to have a roleas a reserve currency. If one looks atworld trade, identifies the country oforigin or the region of origin for all thetrade, extracts any trade betweenEurope, the United States, and Japan,and counts everything that is left, the

Table 1. Comparative Macroeconomics if EMU Implemented

Europe United States Japan

Share of GDP 39% 36% 25%Savings rate 13 5 13Debt/GDP 74 64 93Current account/GDP 1 –2 2

Note: Numbers have been rounded.Source: OECD.



three regions stack up as follows assources of trade for the rest of the world:Japan 21 percent, the United States 35percent, and the European Union (EU)44 percent. The implication is that theeuro should have a role as a reservecurrency.

An issue related to the euro’s poten-tial as a reserve currency is the questionof foreign exchange reserves. Accordingto Goldman, Sachs & Company data,total EU foreign exchange reserves for1996 were about US$238 billion. Themajority, about US$193 billion, was innon-European currencies; U.S. dollarswould be the largest component, butsome would also be in yen. When theEuropean Central Bank is created, it willneed about US$50 billion of that total.The rest would be surplus. On Day 1 ofEMU, the central banks of Europeshould thus have substantial surplusreserves, most of which will be in non-European currencies, and reserves ofthat size should no longer be neededbecause much of the foreign trade willturn into regional trade within the single

currency zone. The result could be anoverhang of non-European currencies,which could find their way into the mar-ket. This circumstance is likely to makethe euro a stronger rather than weakercurrency over time.

These factors taken together supportthe notion that the euro yield curvecould well lie beneath the U.S. yieldcurve over time.

Relationships of EMU with Non-EMU European Countries

The next impact of EMU to consider iswhat will happen to those Europeancountries that do not join. What will bethe relationship between euro govern-ment bond markets and, say, the U.K.bond market if the United Kingdom doesnot participate? The euro bond marketmay move independently of the U.S.Treasury market, but within Europe, cor-relations among bond markets shouldincrease. Figure 5 shows the pattern ofthree relationships beginning in Decem-ber 1987 and ending in December 1996.The Germany–United Kingdom and

Figure 4. Current-Account Trends for the European Union, Japan, and the United States, 1988 through Estimates for 1996 and 1997

1995

150

100

50

0

–50

–100

–150

Cur

rent

Acc

ount

(US$

bill

ions

)

1994199319921991 1996E199019891988 1997E

Japan United StatesEuropean Union



Germany–France bond market correla-tions have been relatively high and stablein recent years. The relationship betweenGerman and Italian bonds has been morevolatile, but in 1996, that correlation rose.

When EMU is created, an exchangerate mechanism is likely to be intro-duced for most non-euro currencies,and eventually, a system will be estab-lished of virtually fixed exchange ratesbetween the euro and the other Euro-pean currencies. For example, if Italydoes not join on Day 1, Italy is almostcertain to have a fixed exchange ratewith the euro, which ought to lead to ahigh correlation between the bond mar-kets of Italy and the EMU countries.

How that situation will affect portfo-lio management depends on one’s per-spective. On the one hand, that highcorrelation is bad, because low correla-tions provide more opportunities forstrategic positioning. When correlationsare low, a manager can look for majormoves in relative performance betweenmarkets. On the other hand, high andstable correlations create greater oppor-tunities for tactical positioning. If one

bond market moves out of line withanother and a manager believes thatsooner or later the currencies will revertback to their means, the manager canput an anomaly switch on to capitalizeon the expected reversion.

Sovereign SpreadsOne obvious question is: How stable

are the sovereign spreads likely to be?The answer, based on the past, is: Prettystable. Managers are not likely to obtainspectacular returns by selecting “theright country.” Figure 6 uses the exam-ple of Sweden versus the United King-dom in 1996 for U.S. dollar five-yearbond issues. For most of the time, thespread moved within a 10–15 basispoint range. If a portfolio manager iswilling to put on a bond position, a bondswitch, and take it off for a quarter of apoint, this spread relationship hasenough room for such a strategy. But therelationship is certainly not the sort thatwill provide substantial returns.

The example given in Figure 7 ofPortugal and Ireland in the Germanmark bond market shows even lessopportunity. Portugal and Ireland are

Figure 5. Trends in European Bond Correlations, December 1987– December 1996

Source: Data from Merrill Lynch.

1.0

0.8

0.6

0.4

0.2

0

–0.2

–0.4

Cor

rela

tion

88 89 90 91 92 94 95 9693

Germany–Italy

Germany–FranceGermany–United Kingdom

87



two seemingly very different countries,but the spread relationship between thetwo has been highly stable. One lessonfrom this analysis is that sovereignspreads will need to be analyzed afterEMU but the scope to add materialvalue is going to be very small. Never-theless, country selection has to be done,and the question is how to decidebetween the various countries.

Typically, managers tend to view cur-rency risk as very different from creditrisk, but in sovereign bond analysis, theyshould not be separated. The reason isseigniorage: If a sovereign state wishes torepay its debts, it can simply printmoney and pay off the debt. If its cur-rency is convertible, it can even follow

that course for overseas debt; it can printnotes, exchange them on the foreignexchange markets, and then redeem thehard-currency debt. In that case, all thecountry is doing is transferring creditrisk to currency risk. Because it willremove the ability to follow thatapproach, the European Central Bankwill have a major impact on currencyand credit risk after EMU. Seignioragewill no longer exist for EMU countries.Individual countries will lose control ofmonetary policy, so the option of simplyprinting money to pay back sovereigndebt will no longer exist.

Whether the end of seigniorage mat-ters in the real world is a question ofsome controversy. Moody’s and Stan-dard & Poor’s apparently have different

Figure 6. Stability in Sovereign Spreads (Five-Year Bond Issues in U.S. Dollars): Sweden minus the United Kingdom, 1996

Figure 7. Stability in Sovereign Spreads (Five-Year Bond Issues in U.S. Dollars): Portugal minus Ireland, 1996

Aug

30

20

10

0

Spre

ad (b

asis

poi

nts)

JulJunMayApr SepMarFebJan1996

Oct Nov Dec Jan1997

Aug

30

20

10

0

Spre

ad (b

asis

poi

nts)

JulJunMayApr SepMarFebJan1996

Oct Nov Dec Jan1997



views, with one thinking it is significantand the other thinking it is not. If it issignificant, then the day a country signsup for the EMU, its credit rating is likelyto decline, at least at the margin. Theirony is that a country that chooses notto join, and thus retains seigniorage,could end up with a higher credit ratingthan a country that gives up its mone-tary policy independence. Portfoliomanagers will have to decide, as EMUapproaches, how the markets will reactto the situation. Bond investors maysimply assume that the issue is aca-demic, but I suspect that it will have animpact on spreads.

Sovereign credit analysis after EMUwill involve a number of new issues inaddition to the usual factors. In general,the tougher the EMU environment is,the wider the spreads are likely to bebecause of pressure on the weaker mem-bers. Some questions to consider are asfollows:• To what extent will fiscal sover-

eignty be retained by individualcountries?

• A stability pact is to be put in place,but will it be a serious pact? If it istough and countries have little fiscalflexibility, the spreads of the weakercurrencies and the more indebtedcountries will suffer.

• How deflationary will EMU be?Everyone at the moment is usingEMU as a deflationary project. If itcontinues to be so used, again, itputs pressure on the weaker curren-cies and their spreads are likely towiden.

• Will a bailout of the weaker debtorsoccur? The Maastricht Treaty sug-gests that no bailouts of sovereignstates will take place, but manyinvestors assume that if the ques-tion actually arises, support will begiven. The question of bailoutscould influence the general level ofspreads.

• The final issue concerns the possi-bility of countries leaving the EMU.Even if EMU goes ahead, what hap-

pens if countries begin to leave?Spreads could have much greatervolatility. The example of Quebectrading in Canadian dollars couldbe used to see the potential impactof countries leaving the EMU.

EMU and Portfolio Management

An improved structure of the Europeanbond markets and the changing marketdynamics that EMU will spur haveimplications for the management ofEuropean bond portfolios and the Euro-pean portions of global bond portfolios.

One way to look at the expectedimpact of EMU is in terms of efficientfrontiers. Figures 8, 9, and 10 reproduceresearch carried out by Nomura Securi-ties International on the returns ofhedged and unhedged bond portfolios.The figures highlight the types of issuesthat will have to be addressed. They aredrawn from the viewpoint of a German-mark-based investor who is looking at aEuropean bond portfolio. Nomura usedmodeling techniques to examine effi-cient frontiers before EMU and projectedefficient frontiers after EMU. Figure 8shows the pre-EMU efficient frontiersfor unhedged and hedged Europeanbonds. The two lines are somewhat dif-ferent from what one would see from aU.S.-dollar perspective. In Nomura’sanalysis, the unhedged efficient frontieris the more attractive, and part of thereason is the expectation of currencygain over time from a German mark per-spective. Currency’s role is quite power-ful in this example.

For the post-EMU environment, theNomura researchers made two assump-tions. One is that correlations betweenEuropean countries—outside and insidethe EMU—will be higher than beforeEMU. Nomura researchers also assumedthat the euro market itself will be lessvolatile after EMU than its pre-EMUcomponent parts have been. If the Maas-tricht criteria have been achieved, theeconomic results ought to encourage lessvolatility in the bond markets. Under



Figure 8. Pre-EMU Efficient Frontiers for Hedged and Unhedged Bond Portfolios

Source: Research by Nomura Securities International.

Figure 9. Pre- and Post-EMU Efficient Frontiers: Unhedged


Figure 10. Pre- and Post-EMU Efficient Frontiers: Hedged


7.0

6.5

6.0

5.5

5.0

4.5

4.0

3.5

3.0

Ret

urn

(%)

2.01.51.00.5Risk (%)

HedgedUnhedged

7.0

6.5

6.0

5.5

5.0

4.5

4.0

3.5

3.0

Ret

urn

(%)

2.01.51.00.5Risk (%)

Post-EMU

Pre-EMU

7.0

6.5

6.0

5.5

5.0

4.5

4.0

3.5

3.0

Ret

urn

(%)

2.01.51.00.5Risk (%)

Post-EMU

Pre-EMU



those assumptions, as Figure 9 shows,the efficient frontier for an unhedgedportfolio looks less attractive than it didbefore EMU. The frontier has moveddown almost to the level of the pre-EMUhedged portfolio in Figure 8. The reasonfor this deterioration is the lack of cur-rency gain. The absence of the foreignexchange dimension under EMU will bea negative for an unhedged portfolio.

Figure 10 draws a different picture.The profile of the currency-hedged effi-cient frontier has moved up after EMUfrom its pre-EMU position. The reasonfor the improvement is simply the lowervolatility assumed to follow EMU imple-mentation. Before EMU, the unhedgedbond portfolio dominated the hedgedportfolio, but after EMU, because of theabsence of currency gains and reducedvolatility, the hedged portfolio is pre-dicted to dominate, as Figure 11 shows.

In general, although the structure ofthe bond markets is expected toimprove and to provide greater flexibil-ity for duration and term structure man-agement, EMU’s reduction in thenumber of individual capital markets islikely to reduce top-down investmentmanagement opportunities. At thesame time, the industry can expect amajor increase in bottom-up opportuni-ties from a more vibrant European cap-ital market with more sectors than

simply government bonds. These devel-opments are likely to have profoundimpacts on the management of Euro-pean bond portfolios. One outcome willbe the emergence of the European fixed-income analyst—a rare animal at themoment. Most fixed-income operationsin Europe are a top-down exercise witha strong generalist component; the staffis composed of portfolio managers whodo a mixture of long-term strategy andportfolio management. The operationmay have a trading desk, but althoughit does some bottom-up investing, it islargely devoted to execution.

Moreover, if the importance ofbottom-up investing increases, then aspecialist structure, with employees whoare doing purely the bottom-up creditanalysis, for example, may becomeappropriate. Such a change could alsohave an impact on organizational struc-ture, which might then include specialistresearch teams.

These changes could lead to style dif-ferentiation. Presently, the styles inwhich European bond portfolios aremanaged are broadly similar. Somefirms are better than others, but the tech-niques they use are similar. If bottom-upcapability is added to top-down capabil-ity, it is only a matter of time beforesomeone thinks of firing the economist

Figure 11. Post-EMU Efficient Frontiers: Hedged and Unhedged


7.0

6.5

6.0

5.5

5.0

4.5

4.0

3.5

3.0

Ret

urn

(%)

2.01.51.00.5Risk (%)

HedgedUnhedged



and simply running money with thebottom-up team. So, one result might be“nondirectional” European bond man-agement, and the range of managementstyles available in the United Statesmight soon be available in Europe.

Finally, the change in market dynam-ics will have an impact on global bondportfolios. If a full range of bond typesbecomes available in Europe, the optionof bottom-up management in Europe aswell as in the United States will surelytransform global bond management.Simply running global bond portfoliosfrom an asset allocation viewpoint willno longer be sufficient. Bond analysis inEurope as well as in the United Stateswill require bottom-up analysis.

The worldwide industry contains agreat reservoir of professionals withanalytical ability in the U.S. fixed-incomearea, but those analysts may not have theknowledge of the local borrowers inEurope. In Europe, fund managers tendto know the regions well but may lackthe skills to do the new type of analyticalwork. Therefore, changes in mandatesfor the management of global bond port-folios will, in turn, present organiza-tional challenges. In theory, firms willwant to have bottom-up bond teams forboth continents, but having investmentpersonnel spread across centers is moredifficult than having them all under thesame roof. The investment process itselfcould come under some scrutiny in theyears ahead as EMU develops.

Conclusion

EMU will affect European capital mar-kets, market dynamics within those mar-

kets, and bond portfolio management—both within Europe and around theworld. The first impact of EMU shouldbe improvement in the structure of theEuropean capital markets. The capitalmarkets should both broaden anddeepen. The second impact should be onthe dynamics of the market. UnderEMU, the way the markets operate—thesorts of factors they respond to—is likelyto change considerably. On the negativeside, portfolio diversification opportuni-ties are likely to decrease, so the scope fortop-down portfolio management maybe reduced. The bottom-up opportuni-ties, however, are likely to improve con-siderably, which leads to a third impact:The investment management processitself is likely to change. Indeed, theimpact of EMU on the process may be sogreat that the organizational structure ofinvestment management firms inEurope could be affected. Overall, EMUis likely to be a positive development forthe European bond market.

I have made the assumption in thispresentation that EMU will start, but Idid not assume it will necessarily work.If the whole project ends in tears, no oneknows what will happen. The MaastrichtTreaty has no mechanism for exit; noallowance is made for the whole thingnot working particularly well. Whatwould happen to spreads if countriesstarted to leave or the project looked as ifit could not continue and what wouldhappen to the value of the euro in thecurrency markets if some of the membersbegan to talk about leaving are part of thegreat unknown.


Schoenfeld.frm Page 52 Wednesday, November 12, 1997 2:56 PM

Indexing Emerging MarketsSteven A. Schoenfeld

Principal, Emerging Markets Strategy GroupBarclays Global Investors

Investors define “emerging markets” in practical ways. A criterion that is increasingly important is investability; without it, investors cannot obtain the high returns these markets promise. Today, several families of emerging market indexes allow investment managers to create

innovative index-based—but far from passive—strategies for emerging market investing. Indexing can

provide diversification, minimal transaction costs, reduced management fees, and consistent long-term

returns in line with the asset class.

egardless of how fasci-nating, potentially re-warding, and dynamicemerging markets ap-pear to be, investors needto decide how to actually

invest in these markets. In ever-changingmarkets such as those in developingeconomies, this decision can be a chal-lenge. This presentation reviews the def-inition of emerging markets—the officialdefinitions and the definitions used inpractice—and analyzes the characteris-tics of the primary emerging marketbenchmarks. In addition, the presenta-tion addresses the challenges for tradi-tional active managers in emergingmarkets and the advantages of index-based strategies for emerging markets. Itconcludes with an overview of index/quantitative portfolio management tech-niques, with a focus on implementationissues and tactics for using index-basedapproaches to efficiently access this dy-namic asset class.

Defining the Asset Class

Emerging markets can be defined in avariety of ways. Most practitioners start

with the definition established by theInternational Finance Corporation (IFC),the private-sector arm of the WorldBank, which coined the term “emergingmarket” in 1984: An emerging market isa stock market in a low- or middle-income developing country (as definedby the World Bank). The per capita GNPfigure changes this definition each year,but as of 1996, a “developing country”was one with a per capita GNP of lessthan US$9,386. By that criterion, morethan 170 countries currently qualify asdeveloping economies; of these coun-tries, more than 60 have stock marketsthat function well enough that investorscan recognize them as real, albeit less-developed, stock markets.

Figure 1 shows a variety of developedcountries (on the right) and developingcountries (on the left) by 1995 marketcap to GNP and GNP per capita. (Notethat both axes in Figure 1 are log scales.)Countries with GNP per capita of aboutUS$9,400 are officially not developingcountries by the World Bank standard,so Taiwan, for example, is an outlier. Inpractice, the definition of emergingstock markets is broader than the officialeconomic definition, but even the pure

R



economic definition does not conformexactly with investment practice.

Another approach is to look at coun-tries in terms of the ratio of their “invest-able” market capitalization to GNP.Surprisingly, several emerging marketshave high market capitalizations as apercentage of GNP; as Figure 1 shows,the percentage for Malaysia, SouthAfrica, and Chile is higher, in fact, thanthat of most developed markets. But themarket cap in the emerging markets isnot all investable by foreigners. Whenmarkets are screened for investability,these countries generally move down tobelow 100 percent of GNP. As a result,investors have developed a more prac-tical definition of an emerging marketthan the more formal economists’ defi-nition: An emerging market is “a stockmarket that does not function as well asthose in developed markets.”

Investors approach a definition ofthe emerging market asset class alongthree paths. The first path is simply tosingle out the countries that are not“developed”—that is, for a U.S. inves-tor, countries not in the Morgan StanleyCapital International (MSCI) Europe/Australia/Far East (EAFE) Index andfor a U.K. investor, countries not in theFinancial Times/Standard & Poor’sActuaries World Index (FT/S&P-AWI).This straightforward path has someproblems, however, some of which are

created by the index vendors them-selves. For example, although Malaysiais firmly in the emerging market cate-gory by most definitions, it is in theEAFE Index as well as the MSCI Emerg-ing Market Free (EMF) Index and theIFC Investable (IFCI) Composite Index.South Africa, Brazil, and five otheremerging markets are in the FT/S&P-AWI. Israel is a developed market byWorld Bank standards, but it was put inthe EMF in 1993, partly on the basis ofa survey of practitioners, and added toIFCI coverage in 1997.

The second path to a practical defini-tion of an emerging market singles outthe markets that are simply more difficultto access than developed marketsbecause of statutory restrictions on for-eign investments or because of insuffi-cient liquidity and/or substandardtrading and settlement mechanisms.Restrictions on portfolio investment canbe onerous even in some of the wealthiestand most advanced emerging markets,particularly in Asia—markets such asTaiwan and South Korea. Liquidity andtrading mechanisms can create prob-lems; cumbersome scrip-based physicalsettlement procedures in India and Ven-ezuela are an example, as are the lessliquid and premium-priced “alienboard”-quoted stocks in Thailand andIndonesia. Such markets have aspectsthat managers should make clients

Figure 1. Market Capitalization as a Percentage of GNP: Developed and Emerging Markets, 1995

JordanPhilippinesIndia

South Africa

Thailand

Zimbabwe IndonesiaEgypt

Pakistan

Hungary

Peru

Poland

VenezuelaTurkey

CzechRepublic South KoreaMexico

Colombia Brazil

Chile

Morocco

GNP per Capita (log scale, US$)

1,000

100

10

1

Mar

ket C

apit

aliz

atio

n/G

NP

(log

sca

le, %

)

10,0001,000100 100,000

ItalyGermany

France

JapanNetherlands

Canada

United Kingdom

United States

Hong Kong

Greece Portugal

Switzerland

Taiwan

Malaysia

Sri Lanka

ChinaNigeria

Russia



aware of, aspects that might justify a plansponsor “red flagging” them for exclu-sion from a portfolio even if they are partof the generally accepted emerging mar-ket universe.

A consensus has evolved within theinstitutional investment communityaround a third path to achieving a prac-tical definition of emerging markets—that of focusing on “investability.” Thisapproach starts with the broad assetclass that can be accessed to somedegree—if not 60 markets, then at least45–50 markets—and segments it by cri-teria: openness to foreign investors,liquidity, and transactional efficiency.The result is a narrower universe thanthe original 45–50 but still a diverse uni-verse of 20–30 emerging markets repre-senting all developing regions.

Emerging Market Benchmarks

The emerging market universe is definedby three indexes that have become wellknown to institutional and internationalinvestors—the IFC indexes, the MSCIindexes, and the ING Barings indexes.Each index portrays the markets in a con-sistent way, so an investor or managercan compare apples with apples, orangeswith oranges, while having a practicaldefinition of the asset class, namely, theinvestable universe. And like emergingmarkets themselves, the indexes evolveand expand regularly to include newmarkets.

The IFC indexes were the first familyof emerging market indexes and havethe longest time series. It is a rapidlygrowing family, which now includesfour distinct index series, the broadest ofwhich covers 44 markets. The IFC GlobalIndex (IFCG) is the original index series;it is not screened for investability forforeigners. The IFCI is rigorouslyscreened for investability. The IFC Trad-able Index (IFCT) is even more narrowlydefined on the basis of liquidity foractive trading and derivative use. Thenewest family member, the IFC FrontierIndex (IFCF), is broader than the IFCG

and covers the newest emerging mar-kets, many of which are not yet appro-priate for institutional investors. TheIFCG indexes were launched in 1981,with some history extending back to1975. The IFCI is derived from the IFCGbut screened at the market and securitylevel for investablility and liquidity. TheIFCI indexes are currently the dominantbenchmarks for index funds in emergingmarkets. As of the end of 1996, approxi-mately 75 percent of indexed emergingmarket assets were tracking the IFCI.

The MSCI emerging market indexeswere launched in 1987. The broadestindex, known simply as the EM Index,covers 26 markets; the EMF Index coversonly investable emerging markets andstocks. The EMF is similar in construc-tion to the EAFE Free Index and MSCI’sother investable indexes for the devel-oped markets, such as the MSCI WorldFree Index. MSCI recently enhanced theEMF methodology to account for foreignboard share prices as well as major free-float considerations. Its primarystrengths are that it can be linked to theMSCI developed market indexes—forexample, the MSCI All Country WorldIndex covers 47 developed and emerg-ing stock markets using a singlemethodology—to provide seamless cov-erage of developed and emergingmarkets.

The ING Barings Emerging Marketindexes (EMI) were launched in 1992 toaddress the liquidity and tradablity con-cerns of some investors. They have beenadopted as benchmarks for some indi-vidual funds but have not gained broadacceptance as benchmarks among pen-sion plan sponsors and consultantsbecause they are considered too narrow;they generally cover only about 40 per-cent of total market capitalization.

An investable emerging market uni-verse as defined by the three indexes con-tains about half the countries of a full index.Table 1 shows the 26 markets and theirweights in the IFCI Composite and in theEMF at the end of 1996. The indexes coverthe same number of countries but are not



in harmony on which countries areincluded. Israel is currently in the EMF butnot the IFCI; Zimbabwe, which is in theIFCI, is not yet in the EMF. Furthermore,Taiwan was added to the EMF only in 1996.The IFCI plans to expand before the end of1997 to 31 countries. The five planned addi-tions are Egypt, Israel, Morocco, Russia,and Slovakia. The IFCG universe alsoincludes Nigeria; the IFCF adds a dozenmore countries—Bangladesh, Botswana,Bulgaria, Côte d’Ivoire, Ecuador, Ghana,Jamaica, Lithuania, Mauritius, Slovenia,Trinidad and Tobago, and Tunisia.

Roles of Benchmarks in Emerging Market Investing

The purpose of benchmarks inemerging market investing is no differ-ent from their purpose in developedmarket investing. In many ways, how-ever, benchmarks are even more impor-tant in emerging markets because theasset class itself is both amorphous anddynamic. Therefore, the first role of abenchmark in emerging markets is sim-ply to define the asset class.

The second role is to serve as a tool fordetermining strategic allocations to

Table 1. Comparison of Country Coverage and Weights: IFCI Composite and EMF as of December 31, 1996

Region/Market IFCI Composite EMF

Total number of stocks 1,225 1,021

Latin America 33.2% 30.9%Argentina 3.9 3.4Brazil 11.4 12.3Chile 5.2 3.6Colombia 1.0 0.7Mexico 9.8 8.1Peru 1.0 1.0Venezuela 0.9 1.1

East Asia 13.7% 17.2%China 0.7 0.6Philippines 3.3 3.4South Korea 2.4 4.3Taiwan 7.3 8.9

South Asia 32.8% 33.0%India 1.9 5.4Indonesia 4.5 5.5Malaysia 22.9 16.6Pakistan 0.5 0.6Sri Lanka 0.1 0.1Thailand 2.9 4.8

Europe/Mideast/Africa 20.4% 19.7%Czech Republic 0.5 1.3Greece 1.5 1.2Hungary 0.5 0.4Israel — 2.1Jordan 0.1 0.1Poland 0.8 0.5Portugal 1.9 2.0South Africa 12.6 10.9Turkey 2.4 1.4Zimbabwe 0.1 —

Sources: IFC; MSCI.



emerging markets. Beyond simply defin-ing the asset class, benchmark indexesallow investors to determine the optimalstrategic allocation to emerging markets.A primary benefit of emerging marketsfor plan sponsors is diversification, andeven the most active of active managersneeds tools for analyzing such benefits.If a manager decides not to follow theweights the indexes are using, theindexes are nevertheless useful (and pri-mary) tools in the allocation process.Thus, benchmarks and indexes do notnecessarily have to enter the debateabout indexing versus active manage-ment. Figure 2 shows the diversificationeffects, using historical returns for aneight-year period, of adding the IFCIComposite to an EAFE portfolio. Suchuse of index returns in an efficient fron-tier diagram is an essential part of deter-mining overall policy allocation foremerging and developed markets.

The third role of benchmarks is tomeasure manager performance. Certaininvestment managers dealing in emerg-ing markets have in the past stated: “Youcannot benchmark me because there isno good benchmark to use.” But bench-marks are necessary to measure managerperformance and can now be usedappropriately in connection with emerg-ing market investing.

Finally, benchmarks and indexes in

emerging markets, as in developed mar-kets, are increasingly becoming aninvestment vehicle in their own right.Indeed, the most important new charac-teristic of the two main emerging marketbenchmarks is that they are becomingindexes that one can actually invest in.They are realistic benchmarks. Five yearsago, a fund manager could say, “That’snice that the IFC or MSCI has an index foremerging markets, but you cannot actu-ally get those returns.” Since 1993, inves-tors have been able to get emergingmarket benchmark returns, and the effi-ciency of doing so is rapidly improving.

Index Construction IssuesOnce investors determine “how

much” to allocate to emerging markets,they can then determine “how to” getexposure to the markets. But investorsshould first have a good understandingof how the benchmarks are constructed.

The key issues related to benchmarksin emerging markets concern countryand security coverage, constructionmethodology and investability, marketweights, availability, and objectivity.Although several other emerging mar-ket benchmarks exist, in this compari-son, I focus on the two benchmarks thathave gained wide acceptance fromemerging market investors: the IFCI andthe MSCI EMF.

Figure 2. Efficient Frontier: Adding the IFCI Composite to the EAFE Index, December 1988–February 1997

Sources: Data from MSCI, the IFC, and Barclays Global Investors.

20

20

18

16

14

12

10

8

6

4

Ann

ualiz

ed R

etur

n (%

)

13Annualized Risk (%)

100% EAFE

60% EAFE/40% IFCI

100% IFCI

191814 15 16 17



■ Coverage. Index coverage relates tothe number of stocks and markets in abenchmark. Both index providers havegreatly expanded their coverage in thepast several years, and by late 1997, thecoverage of each is expected to expandeven more, with the IFCI growing to 31markets. Despite its lower market capi-talization, the result of more rigorousscreening for foreign investability, theIFCI currently includes 1,225 securities tothe EMF’s 1,021.

■ Construction methodology and inve-stability. Investors need to assesswhether an emerging market indexoffers practical investability—that is,whether international institutions caneffectively realize the returns of theindex. The key issues relate to liquidityof the markets and securities, legalrestrictions on foreign investors, andcross-holding/free-float adjustments.These issues are complex in mostemerging markets because informationis difficult to access. The problems haveno simple solutions, but the advantage

goes to the index that defines investabil-ity consistently across and withinmarkets. The index should examinewhat is truly investable by foreignersand remove the percentage of marketcapitalization that is not available. Theindex should also adjust for cross-holdings and government holdings thatare not available to portfolio investors.

The index providers use differentmethodologies for handling these issues.The IFC deals with availability to foreigninvestors, cross-holdings, and govern-ment holdings in excruciating detail. If,for example, Company A owns 30 per-cent of Company B and Company B owns15 percent of Company C, the IFCindexes (assuming a three-stock indexand that the stocks are all the same size)will have a market capitalization thatlooks like the bottom row of Figure 3. TheMSCI EMF, in contrast, uses an all-in orall-out methodology. The EMF makes ajudgment about whether a stock is openor closed and, in the case of the threestocks in Figure 3, would be likely either

Figure 3. IFC Adjustments for Cross-Ownership

Company A owns 30% of Company B

Company B owns 15% of Company C

Cross-Holdings =

A

IFCI MarketCapitalization =

B C+ +

A B C+ +



to include B and C at full market cap orto include A at full market cap. The morelikely scenario would be for the EMF totake out B and C completely. The problemis that B and/or C might be very liquidstocks in the index. In South Africa, forexample, which is probably the mostextensively cross-held market in theworld, Anglo-American Corporation ofSouth Africa Ltd. and DeBeers Consoli-dated Mines Ltd. are the two most liquidstocks, but MSCI removes Anglo-Ameri-can. The IFC would not drop one or theother out but would put in a slice of each.So, using the IFC indexes as benchmarksprovides exposure to the markets with-out taking out liquid shares.

■ Market weights. Another key issue isthe weight of individual markets in anoverall asset-class index. The IFCI andEMF contain a large number of emergingcountries, but when the weights areaggregated, significant concentrationsappear in both indexes. In both, the topfive markets in a 26-market universecomprise more than 50 percent of theindex market capitalization. And atmany times in both benchmarks’ histo-ries, single markets have been more than20 percent of the weight. The most recentexample is Malaysia, but South Africawas heavily weighted in 1995, as wasMexico in 1994. The effect is not asextreme as the weight of Japan in theEAFE Index (Japan is currently about 30percent but was 60 percent in 1989), butsuch concentration can be an issue formany investors. The solution that Bar-clays Global Investors (BGI) has devel-oped to this problem is to provide ourclients with alternatively weightedbenchmarks. These custom indexes canreduce market concentrations (for exam-ple, by constraining the weight of largermarkets) or tailor the weighting to otherclient preferences. We may, for example,use a l iquidity-tiered weightingapproach, which would group emergingmarkets into two or three tiers based onsize, correlations, and trading costs andthen equal weight allocations withineach tier.

■ Transparency and broad availability.

The transparency of an index is maxi-mized when the calculators rigorouslyfollow a set of well-defined and well-understood rules. The IFCI has had aclear advantage in this area because theEMF’s rules for security inclusion aregenerally more subjective and the IFCI’smethodology for treating markets andsecurities is more consistent and predict-able. For example, the IFC publishes spe-cific “investable weight factor” ratios foreach stock in the index. The ratioincludes the degree to which the stock isopen to foreigners, the degree of cross-holding, and the degree of governmentownership. MSCI has made significantstrides in this area in the past two years,and although the edge in transparencystill goes to the IFCI, the gap is narrowerthan it was in the mid-1990s. Both indexcompilers must rely on numerous excep-tions to their rules, however, to accountfor the diversity of market microstruc-ture in emerging markets.

Information on the IFCI and EMF isbroadly available from screen-basedinformation services, in print, and onthe Internet.

■ Objectivity. This issue involves thereasons behind an index’s inclusion ofmarkets and securities. Both the EMFand the IFCI can be considered objective.The IFC, as a member of the World BankGroup, is committed to the developmentof emerging stock markets, but it retainsstrict neutrality with regard to marketinclusion and stock selection for theindexes. The IFCI’s objectivity isenhanced by its Index Advisory Panel,which provides guidance from marketpractitioners about index policy. TheEMF is constructed and maintained byCapital International Perspective inGeneva, which maintains a solid “firewall” between its operations and Mor-gan Stanley’s investment banking andbrokerage activities.

Challenges for Active Managers

Delivering the potential returns fromemerging markets poses many chal-



lenges to active managers. In the pastfive years, the growth of funds undermanagement that are targeted to emerg-ing markets has been phenomenal. Thefunds are both institutional and retail.Many of the investments are comingfrom dedicated emerging market funds,managers with specific mandates toexpose portfolios to the emergingmarkets, but there are also a lot of“dabblers”—EAFE-mandated manag-ers who have permission from their plansponsors to go, say, 5–10 percent intoemerging markets or managers with anAsia or Pacific Rim mandate, whichinherently entails a 20–30 percentexposure to emerging markets. Suchmandates may be the most dangeroustypes from a manager’s as well as a plansponsor’s perspective. If emergingmarkets are not specifically in themanager’s mandate, the amount ofemerging market exposure in theportfolio is unclear to both parties. Insuch circumstances, the manager mightbe taking inappropriate risks, and theplan sponsor could receive nastyperformance surprises.

At the same time as emerging marketfunds under management are increas-ing, the number of undiscovered mar-kets is shrinking. Getting sufficientexposure to Sri Lanka, Bangladesh, orSlovakia in an emerging market portfo-

lio to make a difference at the portfoliolevel is difficult. Even if a manager couldbuy all the securities in such a market,allocating into such markets the 3–5percent of the portfolio needed todeliver benefits to overall portfolio per-formance would be very difficult.

The availability of information anddata, however—earnings estimates andso on—is rapidly improving. The qual-ity of the data is not as high as for thedeveloped markets, but so much moreinformation is available in the emergingmarkets than previously that, althoughthe emerging markets are significantlyless efficient than the developed mar-kets, their efficiency is growing. Never-theless, transaction costs—the actualcosts of moving in and out of these mar-kets—are not diminishing much; costsare about three times as high as in devel-oped markets and five times the cost inthe U.S. domestic markets.

A key question for practitioners is:Can active management still beat theindexes in the emerging markets?Despite most people’s intuition thatmeeting this challenge should be rela-tively easy, the answer is far from conclu-sive. Beating the indexes may depend onthe universe and the time frame. Table 2shows the average performance of threepools of managers from Micropal’s man-ager universe—global emerging market

Table 2. Performance of Average Micropal Active Emerging Market Manager versus IFCI as of September 1996

Management Area

Active Manager

IFCI Total Return

Number of Funds Returns

One-year returnsGlobal 268 7.23% 8.66%Asia 303 3.43 7.08Latin America 137 4.84 10.53

Three-year returnsGlobal 100 19.87 19.53Asia 191 25.38 32.20Latin America 55 17.57 15.16

Note: Cumulative returns.Source: Micropal Emerging Markets Fund Monitor (Micropal); IFC.



managers, managers with Asian emerg-ing market mandates, and managerswith Latin American mandates. The one-year and three-year cumulative data areinconclusive. The average active man-ager is doing better or worse dependingon how you slice the asset class. Returnsfor the five-year period ending Decem-ber 1996, shown in Figure 4, reflect smallperformance differences between theaverage global manager performanceand the indexes, although the indexesare ahead.

Table 3 compares the performancesof the InterSec median, not average,emerging market manager, the IFCI,and the EMF for one year and threeyears. Again, one-year data are not con-clusive, although the median activemanager beat both indexes. For threeyears, median performance is very closeto the indexes.

Another universe, compiled by FrankRussell Company, is illustrated in Figure5. This performance universe shows thatthe annualized median three-year rate ofreturn is below the two major bench-marks. These data also highlight thelarge variance between managers in dif-ferent quartiles and thus highlight theactive risk plan sponsors might bear inaddition to the risks inherent in theemerging market asset class. Investors

also need to remember that most activemanagers’ performance is reported beforefees; fees for active management in theemerging markets start above 100 basispoints (bps), and although they movedown the scale relatively rapidly forlarge institutions, they remain muchhigher than in developed markets. Thus,when fees are deducted, the medianmanager, and certainly the overall aver-age universe of active managers, is doingworse than the market.

Clients want consistent performancefrom active managers, and judging con-sistency requires examination of the per-formance of specific investment firmsrather than the average or the median.Table 4 reports the one-year and three-year cumulative performance of some ofthe biggest names in active emergingmarket management. Some of them arehandily beating the benchmark overthree years and some are not. For the fiveyears ending December 1996, the cumu-lative performance numbers (rounded)are 119 percent for Manager S, 64 percentfor Manager T, 101 percent for ManagerC, 49 percent for Manager F, and 47 per-cent for Manager E. The IFCI Compositeduring that five-year period returnedabout 64 percent, and the sector averagewas 59 percent.

Figure 4. Performance of Average Micropal Active Emerging Market Manager versus IFCI Composite and EMF for Five-Year Period Ending December 31, 1996

Notes: The manager pool is open- and closed-end global emerging market funds, before some fees.Sources: Micropal; IFC; MSCI.

5-Year Period

70

60

50

40

30

20

10

0

Cum

ulat

ive

Ret

urn

(%) 59.4%

63.5% 64.9%

Micropal Global Average IFCI Composite EMF



Clearly, certain active managers canconsistently beat the benchmarks. Thequestions for investors are: Can you pickthat manager? And more importantly,will that manager produce those returnsin the future, particularly as it gainsassets? Manager C, for example, whichhas produced extraordinary returns for along period, has more than US$10 billionin dedicated emerging market assetsunder management. Will this firm beable to continue to produce that perfor-mance? Should a plan sponsor be willingto bet on that outcome? These issues are

at the core of why indexing makes somesense for emerging markets.

Table 5 shows that, in general, beat-ing the benchmark, at least the interna-tional benchmarks, is not easy evenwithin countries. One reason relates tothe structure of these markets. The secu-rity concentrations in emerging marketsare very high. The top 10 stocks by mar-ket capitalization are often 50 percent,and sometimes even more, of marketcapitalization.

In addition, the country factor domi-nates returns in emerging markets.

Table 3. Performance of Median InterSec Active Emerging Market Manager versus IFCI and EMF for Periods Ending June 1996

Manager or Index One-Year Return Three-Year Return

InterSec median emerging marketmanager 11.8% 12.8%

IFCI Composite 10.8 12.3EMF 8.3 13.8

Notes: The pool is active global emerging market managers for U.S. tax-exempt investors, before fees. Annualized returns.Sources: InterSec Research Corp.; IFC; MSCI.

Figure 5. Performance of Median Manager of Frank Russell Universe versus the IFCI Composite and EMF for Three Years Ending December 31, 1996

Notes: Annualized rate of return.Source: Frank Russell Company.

6

4

2

0

–2

–4

–6

–8

Rat

e of

Ret

urn

(%)

5th Percentile

25th Percentile

75th Percentile

95th Percentile

Median

EMF

IFCI Composite



Table 4. Performance of Major Institutional Fund Managers versus IFCI Composite as of September 1996

Manager/Universe or Index One Year Three Years

Number of funds 268 100

M 6.59% 29.33%S 11.00 43.37T 8.31 13.51GE 16.08 20.63C 8.74 28.97J 4.59 —F 6.16 16.12E –4.08 15.05GO 3.94 —P –3.20 6.37Universe average 7.23 19.87IFCI Composite 8.66 19.53

Note: Cumulative returns.Sources: Micropal; IFC.

Table 5. Single-Country Emerging Market Funds versus IFCI Country Returns

One Year (7/95–6/96) Three Years (7/93–6/96)

Emerging Market Funds

Emerging Market Funds

Country Number Return IFCI Return Number Return IFCI Return

Argentina 7 30.45% 38.72% 2 40.52% 65.05%Brazil 47 29.35 26.17 18 88.19 106.30Indonesia 26 5.80 12.53 21 21.67 58.40Malaysia 18 1.08 9.30 13 66.09 67.92Mexico 12 16.67 19.81 7 –15.35 –12.74South Korea 76 –9.29 –15.71 36 25.20 8.54Taiwan 48 1.65 22.05 26 43.05 79.89Thailand 47 –8.60 –14.01 32 62.17 55.25

Sources: Micropal, August 1996; IFC.

Figure 6. Influence of Factors in Returns of Developed and Emerging Markets, December 1994

Source: BARRA data.

Emerging Markets

Industry Factors 16%(global)

Stock-SpecificFactors 38%

(nonsystemic, local)

Country Factors 46%(local)

Developed Markets

Industry Factors 22%(global)

Stock-SpecificFactors 48%

(nonsystemic, local)

Country Factors 30%(local)



Using the BARRA model, Figure 6 com-pares the return factors significant to thereturns in emerging markets with thosethat are significant in developed mar-kets as of December 1994. The factors arealmost a mirror image. In emerging mar-kets, the country factor dominates; indeveloped markets, the country factordrops drastically in importance as stock-specific factors and industry factorsexplain more of the returns. Because thecountry factor dominates in emergingmarkets, investors would do well to pur-sue broad exposure within the countriesor, depending on the investor’s objec-tives, implement top-down asset alloca-tion country by country.

Advantages of Index Strategies for Institutional Investors

Application of index-based strategies inemerging markets can provide investorswith broad diversification, minimaltransaction costs, reduced managementfees, and consistent long-term returns inline with the asset class—all of whichlead to a very compelling bottom line:efficient, low-cost core exposure toemerging markets.

DiversificationThe prime advantage of an index-

based strategy for emerging markets isbroad diversification across and withinmarkets. Indexing is an efficient way togain deep exposure within 20 or moreemerging markets. The portfolios ofmany active managers look very similarto each other and often comprise the larg-est blue chip stocks in each market: thecement company, the telecommunica-tions company, some other utilities, andone or two companies in whateverindustry that country has a comparativeadvantage. Index portfolios go muchdeeper into the markets. They may stillbe concentrated in whatever industrygroups dominate that country’s market,but the diversity of securities is broad.The number of stocks BGI holds in ourindex portfolios, for example, is signifi-cantly larger than the average holdingsof active emerging market managers. Wedo not always buy every stock in the

index, but whenever possible and cost-effective, we buy most of the stocks, notsimply the top five names or the bluechip stocks.

Minimal Costs and FeesFew things are certain in emerging

markets, but high transaction costs are.Broadly defined, transaction costs aremuch higher than in developed markets;they average three times those of devel-oped international markets. In theuncertain return environment character-istic of emerging markets, a reduction inexplicit trading costs through pursuit ofefficient trading and low turnover will,by definition, lower transaction costs.Add to this approach the ability to cross-trade between portfolios and the lowermanagement fees incurred throughindexing, and the result is to put theindex investor significantly ahead of thegame in relation to the active investor.

Table 6 shows that the great advan-tage of indexing lies not so much inreduced custody charges or commis-sions and fees but in reduced marketimpact and management fees. The rea-son for the difference is that information-driven active stock selection requires thatwhen a manager has an idea, the man-ager must often move quickly into thatsecurity; the manager is trading on infor-mation and, therefore, demandingliquidity of the market and demandingliquidity of the broker. Of course, whenthat manager comes in with that bigblock trade, the bid–offer spread widenssubstantially. In indexing, portfolio man-agers actually often provide liquidity in

Table 6. Active versus Passive Manage-ment Expenses

Expense Active Passive

Custody fees 0.27% 0.24%Management fees 1.00 0.35Commissions/fees 1.35 0.72Market impact 2.29 0.94

Total 4.91 2.35Note: Calculations assume 50 percent turnover a year for active management and 15 percent turnover a year for passive management.Source: IFC survey of brokers and fund managers, June 1996.



the market, and they can use a number oftrading techniques to attenuate marketimpact. Furthermore, crossing trades,either at the unit level or the securitylevel, has no market impact.

Consistent Returns in Line with the Asset Class

The delivery of consistent perfor-mance is one of indexing’s greatestattributes. Suppose a plan sponsor hasperformed an asset allocation study andis determined to put 4–6 percent of aplan’s assets into emerging markets.Using index-based strategies will assurethat the portfolio reflects performance ofthe emerging market asset class for thatpercentage, and an index portfolioenables a plan sponsor to be fullyinvested. Plan sponsors, who are gener-ally putting only a small percentage offund assets into emerging markets,might not want a slice of that relativelysmall percentage to be in cash and shouldnot pay 80–100 bps in management feeson cash when that cash is not going tohave much effect at the plan level. Inaddition, large index fund managers areoften able to partially cross new clientsinto their emerging market positions,which can greatly reduce the cost of get-ting emerging market exposure.

Integrating Index-Based Strategies in Emerging Markets

Indexing and active management arenot mutually exclusive; indexing can becomplementary to active management.An index or index-based approach canprovide the efficient, low-cost core expo-sure as the entry point or anchor for anemerging market portfolio. Around thiscore, the manager or investor can• tilt by region,• use active managers that are expected

to add value through stock selection,whether broadly defined or, perhaps,confined to small-cap stocks,

• add private equity in emergingmarkets,

• add managers that specialize in

“frontier” markets,• tilt by management style, and/or• add a hedge fund vehicle that fol-

lows a long/short or market-neutralstrategy.

Furthermore, in emerging markets,indexing is anything but passive. Index-based investment strategies are asdiverse as those available for developedmarkets.

Alternative Weights and Quantitative Approaches

Market-capitalization weighting, theapproach of all the major commercialindex providers, is only the startingpoint for emerging market investmentstrategies, even a strategy based on anindex. For example, most of BGI’s clientsare using alternative benchmarks. Wework closely with them to structure solu-tions to their particular needs. We havedeveloped equal-weighted strategies;strategies using benchmarks with two-or three-tier weighting, GDP weighting,value-traded or liquidity weighting, orcustomized weighting; single-countryor single-region approaches; and mostrecently, active/passive strategies. Indexportfolios are relatively easy to mold toa client’s existing emerging market expo-sure or global portfolio characteristics.Clients can even specify that they do notwant market-cap weighting but do wantbroad exposure. For example, if they arewilling to take 3 percent tracking riskagainst the benchmark while still match-ing market exposure, we can tailor asolution.

Indexing in emerging markets issomewhat controversial and is counter-intuitive to many in the investmentcommunity. Indexing is in its infancy,but it has come a long way; Table 7,which starts one year after the introduc-tion of the IFCI Index, shows that as ofmid-1996, about US$10 billion wasunder management by index/quantita-tive techniques, which is less than 5 per-cent of assets invested by foreigners intoemerging markets. I believe it is reason-able to predict that emerging market



assets being indexed or using index-type structures will grow to 10 percentby the end of this decade.

Management of Emerging Market Index Portfolios

Putting money to work in emergingmarkets, whether actively managed orindex based, is a significant challenge,but index managers face some of thegreatest challenges. They must matchthe benchmark every day, and anyerrors are more transparent than thosein active management. Implementationmust deal with foreign ownership andother restrictions, liquidity, taxes, settle-ment, and high transaction costs in anenvironment of tracking a cost-free, fric-tionless index.

■ Investment problems. The first andtrickiest problem is the issue of foreignownership restrictions, which createsdiverse problems in different countries.South Korea, which has a very opaqueOTC market for foreign shares, presentsparticular difficulties. Gray market pre-miums on Korean stocks can reach 60percent or more; we do not blindly buy astock if it has that kind of premium, evenif it is part of the index. We look at thetrade-offs between the guaranteed track-ing error if we do not buy, buying some-thing at a premium, and the trackingerror of buying a stock substitute. Indo-nesia and Thailand have “alien” boards,on which the pricing is different from thelocal boards. The index providers for themost part use the local prices and do notquote the alien board prices because thealien board prices are often ratcheted upby locals who know foreigners are goingto be forced to use it, so the pricing is not

consistent. MSCI and the FT/S&P-AWIhave added a foreign premium factorinto their indexes, but such adjustmentsdo not fully alleviate tracking problemsin Indonesia and Thailand.

Taiwan and South Korea also havelimits on the percentage of local equitiesthat can be held by foreigners. So, aswith an elevator, only so many peoplecan be in and if an investor is not in onthe allocation, that investor will have topay a higher price. Indonesia, Thailand,Malaysia, China, and India also haverestrictions on foreign ownership.

Chile has specific restrictions onrepatriation of capital gains that limit amanager’s ability to adjust client assetallocations. Therefore, we primarily relyon American Depositary Receipts to getexposure in Chile. If we need to reduceChilean exposure, ADRs allow us to doit quickly.

Liquidity can be a major issue forforeign investors. Many stocks inemerging markets trade essentially “byappointment,” and a manager must becareful with whom the appointment ismade because they may not show up or,worse, they may jack the price up.

Both dividend withholding tax andcapital gains tax are issues. India, forexample, has high taxes, and the mecha-nisms for the tax treatment of foreigninvestors are different from those usedfor local investors. Some investmentvehicles are tax efficient and some arenot, and some of the tax-efficient vehiclesare not appropriate for ERISA investors.Thus, managers have to work their waythrough these issues to develop optimalvehicles for their clients.

Settlement failures are another pitfall

Table 7. Indexed Emerging Market Equity Assets under Management, June 1996(US$ millions)

DateGlobal Passive

StrategiesSingle-Country and Regional Strategies

Active/Passive and Tilt Strategies

March 1994 490 180 230October 1994 890 490 400September 1995 2,300 640 850June 1996 6,200 1,600 1,900

Source: IFC survey of investment managers, September 1996.



in emerging market investment imple-mentation, and BGI follows a number ofprocedures to avoid them. The simpleissues of settlement mismatches and tim-ing present numerous difficulties. Inemerging markets, the settlement periodranges from one day before to five daysafter trade date, and in South Africa thatt + 5 day can only be on a Tuesday. Whenmanagers are moving assets around,they have to make sure they will have theright cash amounts for the exchangeswhen they settle.

The simple costs of custody move-ment and market impact create worseproblems in some markets than in others,but in almost every case, these costs aremuch higher in emerging markets thanin developed markets. Therefore, anyefforts that managers can make to reducethe costs is value added for the client.

Figure 7 provides some recent esti-mates of trading costs in the largeremerging markets. The average tradingcost in emerging markets is about threetimes that of EAFE countries and aboutfive times that of the United States. It isthat hurdle that all managers have to dealwith, and index-based approaches aremore effective at minimizing those costs.

■ Solutions and tactics. How one actu-ally gets exposure in emerging marketsin an index-based strategy is not passive.At BGI, we look at three dimensions ofinvestment management performance—returns, risks, and costs. The foremostobjective is getting the returns of the mar-kets and the asset class, but managingcosts is also important. Portfolios shouldalso be structured in ways that reduce therisk inherent in the indexes by usingappropriate country allocations and

Figure 7. One-Way Trading Costs in Emerging Markets: Commissions, Bid–Offer Spreads (Halved), Fees, and Taxes

aADRs: Chile, 0.55 percent; Venezuela, 0.70 percent.Sources: IFC EMDB survey of brokers and fund managers, June 1996; BGI survey of international broker trading desks, January 1997.

1.00 0.5 1.5 2.0 2.5 3.0Cost (%)

Chilea

Peru

Argentina

Venezuelaa

Indonesia

Hungary

Brazil

Philippines

Portugal

Greece

Mexico

Turkey

Thailand

South Korea

Malaysia

Poland

South Africa

Average

EAFE

United States

2.50

2.40

1.70

1.70

1.65

1.59

1.57

1.50

1.47

1.45

1.40

1.40

1.20

1.15

1.11

1.00

0.85

1.51

0.65

0.30



delivering accurate tracking of thebenchmarks.

Figure 8 shows the variety of instru-ments available to get returns in emerg-ing markets. Our preference is always tobuy the local market, but if thatapproach is not the most efficient andeffective way, we will consider ADRsand Global Depositary Receipts (GDRs).The reason may not be simply tradingcosts; investment restrictions, as inChile, may also be a concern.

The use of equity swaps is not attrac-tive currently, but we are constantlyassessing the development of that mar-ket. Liquidity in emerging market swapsis growing, and when we, and moreimportantly our clients, are comfortablewith it, we hope to use equity swaps. Theswap markets are gradually becomingmore liquid, and most importantly, morefairly priced. The development of listedfutures, options, and other derivatives inemerging markets will also improve thepricing transparency of swaps. Convert-ible securities in emerging markets canalso be an effective way to get exposureto stocks. In South Korea, for example,opportunities usually arise that help toboth get around the foreign premiumand get exposure to the market.

We use country funds opportunisti-cally for some very difficult markets, andsome managers have used vehicles thatI call “insects.” This generic term refersto listed and tradable index fundsknown as WEBS—or World EquityBenchmark Shares—which track indi-vidual MSCI country indexes and tradeon the American Stock Exchange. Theseinstruments offer an efficient way forsmall managers to get exposure to suchemerging markets as Malaysia and Mex-ico. BGI is not using emerging marketstock index futures and options at themoment, but these products are rapidlydeveloping, and we intend to be in theforefront of their use when their liquidityand regulatory status improve.

A number of emerging marketsalready have listed equity derivativemarkets. Some futures and optionstraded on exchanges in developed mar-kets are based on emerging marketindexes, such as those of Mexico, Taiwan,Israel, Hungary, the Czech Republic,Poland, Slovakia, and China. One cantrade futures on the Mexican stock indexon the Chicago Mercantile Exchange.The Singapore International MonetaryExchange has launched a Taiwanese

Figure 8. Advanced Investment Management Techniques in Emerging Markets

Internal CrossingTrading Tactics

Cost Management

Securities Lending

Emerging MarketReturns

Convertibles

Equity Swaps

ADR or GDR Purchase(Foreign Market)

Stock Purchase(Local Market)

Stock IndexFutures/Options

Country Fundsand Insects

Risk Management

Alternative WeightingStrategies

Return in Line withAsset Class



stock index future that is reasonably liq-uid and has already developed stronginstitutional use. This contract is espe-cially useful in light of Taiwan’s currentrestrictions on foreign short sales in thecash market and the country’s day t – 1settlement environment.

Some stock index futures and optionsare trading within emerging markets.This development is bringing down thecost of swaps because swap intermedi-aries can price swaps and then lay off therisk in the domestic markets. Brazil andSouth Africa have very liquid stockindex futures, but they are not yetapproved for U.S. investors’ use. Malay-sia’s futures and options exchange hasgotten off to a slow start, but it willprobably ultimately succeed. Israel hasa liquid stock options market. Hungaryalready has a somewhat liquid stockindex futures market. Korea hasdecided to replicate some of the dys-functionality of its equity market in thefutures market and has a futures marketthat is highly impractical for foreigninvestors.

As far as managing the high costs inemerging markets, internal crossing isone of the most powerful ways to pro-vide value to investors. When differentclients have different country alloca-tions, instead of trading the positions inthe markets, whenever possible wecross internally. BGI has the industry’slargest internal crossing network, andwe have added much value for our cli-ents through this capability.

Crossing creates an internal market,which allows a manager to virtuallyeliminate trading costs. Table 8 shows

what BGI has been able to provide forour developed market investors in theirEAFE portfolios in 1996 and what weachieved during the same period in ouremerging market strategies. The tablegives the percentage of total trades thatwere crossed. The reason the net buynumbers are lower in both markets isbecause U.S. institutional investors(ERISA-qualified plans) are still verymuch net investors into developed andemerging markets. Lining up a buy tocross in is much easier for us if we havean investor who is exiting. The data onactive versus passive managementgiven in Table 6 gave an idea of thepower of crossing: Considering tradingcosts, crossing about one-third of netinflows allows us to eliminate aboutone-third of trading and market impactcosts. Slowly, over time, we will be ableto bring down trading costs and raisenet returns close to the index numbers.This advantage in trading and manag-ing emerging market portfolios willbecome even more sophisticated andcompelling in the future.

Finally, as the bottom diagram in Fig-ure 8 shows, where we cannot cross, weput the trading expertise of a large andsophisticated trading desk behind us andthen, over time, add securities-lendingrevenue. Securities lending is just begin-ning in emerging markets. As spreadshave come down in developed markets,this market has been growing. Securitieslending will be a valuable, securemethod to add returns to investmentportfolios as the futures and optionsmarkets in emerging markets mature.

Table 8. 1996 Crossing Activity for BGI Non-U.S. Developed Market and Emerging Market Index Funds

ActivityNon-U.S. Developed Market Index Funds Emerging Markets

Net buys 45.40 36.30Net sells 99.00 81.47

Note: Total percentage crossed includes unit-level crosses and security crosses.



Clearly, as this discussion has stressed,indexing in emerging markets is any-thing but passive. A manager must takedistinct approaches to different markets.In our S&P 500 Index portfolios and, forthe most part, in our EAFE portfolios, weaim for full replication of the indexesbecause that approach produces the leastamount of tracking error. In emergingmarkets, however, we pursue a variety ofapproaches and implement whichever ismost practical—optimized techniques,sampling techniques, or full replication,as appropriate for each market. Eventu-ally, the techniques and the technology ofdeveloped markets will prevail in emerg-ing markets. This trend will bring bene-fits for all investors, active or indexbased.

The growth of indexing in emergingmarkets is helping increase investors’focus on the three dimensions of invest-ment management performance. Per-formance is more than absolute return;managers and their clients should lookat the risks in their exposures and con-centrations in stocks and, of course, thecosts of initial and ongoing investment.Emerging markets are somewhat likerapidly growing children or adoles-cents: One certainly cannot always pre-dict their behavior (returns), but one canand should control some of the risks andmuch of the costs.

The Transformation of Emerging Markets

I will close with several broad themes Iforesee evolving in the next severalyears. Emerging markets will remain acompelling asset class. Developingcountries’ equities are not a fad, andthese markets are now too big to ignore.The emerging markets were less than 2percent of world market capitalization

10 years ago; now, they are about 11 per-cent of world market cap and 20 percentof world GDP, and their populationscomprise more than 80 percent of worldpopulation. Moreover, the trend is forthe emerging market universe to con-tinue expanding—to include marketsthat a few years ago would cause laugh-ter if someone suggested investing inthem. For example, even as recently as1994, it would have been difficult toenvision the phenomenal rise of the Rus-sian equity market in 1996 and 1997.

The rigorous use of benchmarks foremerging markets should greatlyimprove the transparency of all man-aged portfolios and improve investors’understanding of the risks and rewardsin emerging markets. The growth ofindexing will continue—and likelyaccelerate—and it should have animportant impact on lowering tradingcosts and management fees. Techniquesthat build on indexing and apply activeideas in a systematic, quantitative wayare already being used in the emergingmarkets, and this trend is likely toaccelerate.

In many ways, one could say the tra-ditional approach to investing in emerg-ing markets is over—the game of goingout into unexplored markets, setting upcamp, talking to the friendly local bro-kers, and getting invested before every-one else gets in. As emerging marketsbecome increasingly integrated into glo-bal equity markets, the investment man-agement techniques prevalent indeveloped markets will proliferate—and eventually prevail. The science andart of structured investing is just begin-ning in emerging markets, and it willdramatically transform them into more-efficient capital markets.


George.frm Page 70 Wednesday, November 12, 1997 2:56 PM

Political Analysis for Investing in the Emerging Asian Markets

Robert Lloyd GeorgeChairman and Chief Executive Officer

Lloyd George Management Limited

Asia continues to be the most rapidly growing region in the world, but political risk analysis continues to be

critical for stock pickers in the emerging Asian markets. That analysis should take a historical

perspective and should encompass culture, money flows, savings rates, and inflation. History shows the sharp effect of some single critical events on Asian

markets, so one vital task of investors is to distinguish political events that will have long-term consequences

from events of little importance.

But nothing of [the conflicts of Brit-ish party politics] counted muchagainst the great movements in his-tory. None of our struggles matteredmuch, wars or revolutions or whatyou will, as compared with the sheerbiological and geographical facts.Whatever happened, in two hundredyears, perhaps sooner, the balance ofthe world would have changed.—David Lloyd George as quoted

by C.P. Snow in Variety of Men.

n the 20th century, the balance ofthe world has changed severaltimes, and it has already changedagain from the post-World War IIbalance. Consider the followingfacts: 61 percent of the world’s

population is in Asia, but only about 26percent of the world economy is in Asia.Japan by itself represents 17 percent ofthe world economy; India and Chinatogether are about 4 percent, and the restof Asia represents only about 5 percent.Few analysts doubt that the 26 percentwill rise in the next generation to 40 or50 percent. Real GDP growth and, moreimportantly, the growth of per capita

income and the growth of the middleclass that underlies the political stabilityand the economic potential of thesecountries is very encouraging. Com-pared with Organization for EconomicCooperation and Development coun-tries, Asia is the place to be in the future.But investing in Asia requires some spe-cial skills.

John Templeton has said that socialawareness and political awareness arekey attributes of an investor. This prin-ciple can be taken a step farther bybecoming aware of the historical experi-ence of investors in each Asian country;the proper analysis of risk and rewarddepends on a knowledge of past events.The national response to foreign tradersand foreign investors in China duringthe past 50–100 years, for example, pro-vides a good guide to how China willrespond in the future.

This presentation reviews the criticalelements of political analysis andsummarizes the current political envi-ronment for several interesting marketsfrom an investment manager’s per-spective.

I


Elements of Political Analysis

Political analysis is a necessary part ofthe investment manager’s task in anymarket, but it is especially important inemerging markets, where political sys-tems may be immature and volatile.

Critical EventsIn political analysis, an important

aspect is to distinguish the times whenpolitics matters to investors and thetimes when it is irrelevant. Analystsneed to ask themselves whether the bigpicture is changing or whether a changeof government is irrelevant because thesame economic policy will continue.

About 10 years ago, I received a tele-phone call at 1:00 a.m. Hong Kong timefrom George Soros. He said, “Robert,have you noticed the Politburo changesin Hanoi?” I tried to wake up fast, but Ihad to confess that I had not. His nextwords were, “Start buying shares inBangkok.” As the great chess player andgeopolitical strategist that he is, he wastwo or three moves and at least 12months ahead of the crowd in identify-ing the shift in policy from militaryadventurism to economic reconstruc-tion in Vietnam that would lead to awithdrawal from Cambodia and a sig-nificant improvement in the way U.S.investors perceived the political risk ofinvesting in Thailand. Within the next 12months, Thailand did indeed becomeestablished as a stock market for inter-national investors.

CultureThe historic hand-over of Hong Kong

to China on June 30, 1997, is the mostwell-anticipated, well-advertised eventof the century and should not providemany unanticipated surprises. Re-sponses to the hand-over, however, maynot be entirely rational. What investorsshould keep in mind about 1997 is that itis not 1949. The world has changed;China has changed. Communism is nolonger in the ascendant but is rapidlydeclining as an ideology. Privatization,

not nationalization, is now the success-ful policy of governments worldwide.Human nature and the fears and re-sponses of people, however, changeslowly. The psychology of individuals inthe Hong Kong business community isstill greatly colored by the experiencesof a generation ago in Shanghai. Knowl-edge of such history and psychology isimportant to investors, particularly interms of stock selection. For example,the family who controlled Hong KongLand through Jardine Matheson had alltheir assets expropriated in 1950 inShanghai, despite the promises that hadbeen made beforehand. I believe thatevent is still seared in their corporatememory, and therefore, they expectsome Chinese takeovers of British-owned assets.

In the Asian context, investors andmanagers need to understand howmuch political analysis shades into cul-tural, social, and economic analysis.Japan is a perfect example. The bull mar-ket of the 1980s in Japan went on for farlonger than most foreign managersexpected. P/E multiples rose far higherthan expected because of cultural factorsinvolving land values, equity cross-holdings, and political factors. The bearmarket of the 1990s is also confoundingforeign forecasters, who every yearexpect a recovery, because of these samestubborn political and social factors.Japan seems curiously helpless andunable to make decisions in the face ofthe continuing banking and real estateproblems. The rigidity of the labor mar-ket and the habits of lifetime employ-ment have made rationalization amonglarge Japanese corporations much morepainful and drawn out than in the West.South Korea, which shares many of theseConfucian characteristics with Japan,shares many of the same problems.

What the West so admired on theupside of Asian society—social disci-pline, manufacturing excellence, long-term corporate planning rather thanshort-term earnings targets, social andpolitical stability under a virtual one-



party state—has come back to haunteveryone on the downside. I do notbelieve investors should be bearish onAsia, but the Asian societies are havingproblems adjusting. Few Asian societieshave the flexibility that the United Stateshas demonstrated in the past decade.

Unparalleled speed of economicgrowth and social change in South Koreamay be the reason that market is nowexperiencing problems. The growthresulted in an economy overburdenedwith debt and with wage increases thatoutstripped productivity. Political anal-ysis of Korea is difficult, primarilybecause the country is so unpredictable.All investors and managers were con-founded by the weakness of the Koreanmarket in 1996; I do not know of a singleforeign fund manager who got Korearight that year. At the beginning of theyear, we were anticipating 30 percentearnings growth; the actual figure forcorporate profits at the end of the yearwas –30 percent. Such a complete about-face is highly unusual in Asia, and nowinvestors are responding with deep dis-illusionment and distrust to the forecaststhat Korean companies will recover. Fora contrarian, of course, the opportunityis interesting. The current situationappears to be very high risk because of

uncertainty in North Korea, but if NorthKorea collapses peacefully, the resultcould be a strong bull market in Seoul.

Capital FlowsIn southeast Asia, one of the rules I

have followed in the past 15 years is tofollow the smart money—in general, thatis, where the overseas Chinese busi-nesses are themselves investing, whetherin real estate or in manufacturing plants.Figure 1 shows the capital flows amongthe United States, Japan, and the rest ofAsia; note in particular that the moneyearned in Asia from exports to the UnitedStates—the trade surplus with theUnited States—is not recycled into U.S.Treasury bonds as the Japanese surplusis. It is reinvested in Asia. This money iscoming from the overseas Chinese, whoare the major investors in Asia. The tre-mendous rise in foreign direct invest-ment, US$100 billion out of the US$300billion net total global worldwide FDI, isaccounted for by Hong Kong, Taiwan,and overseas Chinese investors. Indeed,80 percent of all the money that has goneinto China in the past decade has comefrom overseas Chinese sources.

Tracking the FDI capital flows is impor-tant for identifying portfolio investment

Figure 1. Trade and Capital Flows among the United States, Japan, and the Rest of Asia, 1995(billions)

Source: Data from the World Bank.

Japan

Asia United States

Net Trade Payments:US$75

Capital Recycling:US$20

Capital Recycling:US$50

Capital Recycling: US$12

Net TradePayments

Surplus: US$70

Net TradePayments

Surplus: US$60



opportunities. Thailand in the mid-1980salso experienced a boom based on foreigninvestment, which was coming fromJapan as well as Taiwan and Hong Kong.In Thailand, political analysis seemed fora long time to be relatively simple. Theking represented the ultimate guaranteeof stability; a military coup d’état occurredevery year or two, and prime ministersand governments changed constantly, butthe growth rate of the country continuedat almost 10 percent every year for adecade. The economic policy continued tobe friendly and fair to foreign investors.History again supplies some backgroundnecessary for understanding these mar-kets. Thailand has dealt so much morewisely and successfully with foreigninvestors than either of its neighbors,Burma and Vietnam, which were colo-nized in the 19th century and, therefore,have more-difficult relationships todaywith foreign investors. Thailand hasalways been a successfully run economyand presented a stable political situation.The Thais have an excellent civil serviceclosely modeled on the Japanese Ministryof International Trade and Industry, andthey have always been willing to recog-nize the need for foreign capital and tech-nology.

Today, however, Bangkok has becomea victim of its own success. The infra-structure has not kept up with the rate ofgrowth, and the economic boom is caus-ing tremendous bottlenecks. So, thesmart Chinese money today is targetingother places—the Philippines and, to aslight extent, Myanmar (Burma). A lot ofit is going to Sydney, London, and Van-couver, which may be revealing.

Note that, although the cultural ten-dency in Thailand toward a laissez-fairemarket policy is the opposite of the rigidcentral planning that has been commonin Korea, the end result has been thesame. The political crisis that hasembroiled Thailand in the past two yearscame from the same Asian cultural traitsat work in Korea and Japan—the seeking

of social and political consensus—andthe same social habits of awarding favorsand giving gifts. The results have beenwidespread corruption and an inabilityto make key decisions. The political situ-ation has put enormous pressure on thecurrency and the stock market. As inJapan, what appeared in the past to be thestrong elements of the system—forexample, the central bank’s rigid adher-ence to maintaining the baht against thecurrency basket—have become leadingfactors in Thailand’s economic crisistoday. Because of the very strength ofThailand’s foreign reserves and its pasteconomic success, Thailand’s current-account deficit has widened, the U.S.dollar loans taken out by Thai companieshave increased rapidly, and instead of aonce-and-for-all Mexican-style devalua-tion, Thailand is likely to experience anongoing “crisis,” with high interest ratesand economic stagnation.

Savings RateA major driver of growth in Asia has

been the countries’ high savings rates, asFigure 2 shows. For example, althoughSingapore’s Lee Kuan Yew has beenwidely criticized by Western academics,nobody can deny the extraordinary suc-cess that Singapore’s economy hasenjoyed in the past 30 years under one-party rule. Political and social stabilityhave been the foundation of this success,and in addition to the investment inhuman capital, fiscal conservatism hasbeen important to this success. Indeed,throughout Asia, the importance of sav-ings rates in promoting economic devel-opment is obvious.

At almost 50 percent, Singapore leadsin savings rate as a percentage of GDP,and the reason is its Central ProvidentFund, which effectively creates a man-datory savings rate of more than 40 per-cent. To this farsighted policy ofmandated savings, Singapore owes itsoutstanding road network, splendid air-port, and fine housing developments.



Few political analysts would haveexpected this result from the early man-ifesto from Lee Kuan Yew’s People’sAction Party.

Behind the savings figures shown inFigure 2 may be an important trait of theoverseas Chinese communities in allthese countries. There are no welfarestates in Asia, and therefore, the Chinesefamily, on average, saves 30 percent ofits income to provide for old age, sick-ness, unemployment, and education ofchildren. These tremendous savingssurpluses and the financial strengthunderlying these countries are a con-tinuing important foundation for thefuture development of the capital mar-kets in Asia. In the years ahead, if theChinese domestic savings rate contin-ues at the level of the other Asian coun-tries, that trend is a positive indicator ofthe future growth of China. The Philip-pines, which Figure 2 shows laggingbehind for so long in economic growthrate, is now moving up to a 6–8 percentrate, and its savings rate also is risingtoward the typical Asian average.

InflationA general point about political stabil-

ity in Asia is the link of stability to low

inflation. The importance of inflation forpolitical stability is clear. In the case ofthe Philippines, when the inflation ratewas 50 percent, Ferdinand Marcos wasoverthrown. In China in 1989, inflationrose to nearly 30 percent, and a politicalcrisis took place that finished at Tianan-men Square. Similarly striking havebeen the results of the close link betweeninflation and stability in many otherAsian countries. In Singapore andJapan, however, very low (even nega-tive) inflation has been linked to verystable political situations. Table 1 showsinflation rates for the Asian markets inJanuary 1997.

The tremendous success of Malaysiais perhaps even more surprising than thesuccess of Singapore. As in Singapore,fiscal conservatism played a role in keep-ing inflation low in Malaysia. Malaysia’ssuccess stems from the implementation,beginning in 1981, of the New EconomicPolicy, the brainchild of MahathirMohammad. In the course of a genera-tion, he succeeded in transferring 30 per-cent of the wealth of the country fromforeign and Chinese hands into those ofthe Bumiputra (Malay community), and

Figure 2. Savings Rates in Asia and Nine Asian Countries Relative to Real GDP Growth

Note: Rates are 1991–96 averages.Source: Data from Global Information Services.

50

45

40

35

30

25

20

15

10

5

0

Savi

ngs

Rat

es (%

GD

P) Hong Kong

Philippines

Taiwan

AsiaMalaysia

Singapore

ChinaThailand

Indonesia

Korea

Real GDP Growth (%)

2 4 6 8 10 12



he achieved this success without the riseof political and social tensions that led tothe struggle between Muslim Malaysand ethnic Chinese in the 1950s and1960s, the bloody riots, and the 22-month suspension of parliament in 1969.Malaysia has enjoyed a steady 8 percentGDP growth in the past 20 years.

Effects of Political Factors on Stock Markets

The Asian markets provide somespecific examples of political factorsaffecting stock markets. The examplescovered here involve Indonesia, Thai-land, Taiwan, South Korea, and HongKong. Russia is another interestingplace to invest now, so it is discussedbriefly in this section. Finally, twocountries that are not attractive—Vietnam and Myanmar—are noted.

IndonesiaThe Jakarta Stock Exchange (JKSE)

became truly established for foreigninvestors in 1989. At that time, most ofthe economic ministers and advisorygroup in Jakarta had been educated atthe University of Chicago, rather likethe case in Chile. These advisors per-suaded President Suharto that the stockmarket could become a useful tool ofnational economic policy by attractingforeign capital and also by encouragingmany of the wealthiest Chinese familiesin Indonesia to list their companies. Theresult was astonishing. The marketgrew from about a US$1 billion capital-ization in 1988 to nearly US$20 billion in

two to three years. Much of the hugewealth that many analysts knew was inIndonesia started to be reflected in itsnational capital market.

As with all these emerging markets,the Indonesian stock market has had itsproblems. For example, the marketgrew too fast and corrected quite dra-matically in 1991. The JKSE’s perfor-mance was and still is very dependenton energy prices. Oil is, indeed, a moreimportant factor in Indonesia than inother Asian countries; 40–50 percent ofIndonesia’s export earnings come fromoil and gas.

In Indonesia, analysts should weighthe political factor carefully. Manyinvestment analysts today continue tobe positive about the economic pros-pects and earnings prospects of compa-nies listed on the JKSE, but Suharto’s ageand health have become factors in thestock market’s performance. As in someother Asian countries, the influence of asingle personality on the stability of thecountry is immense in Indonesia. Presi-dential elections will be held in 1998,and Suharto may well decide not tostand again, which will affect the stockmarket profoundly. No mechanism forsuccession exists, and no successor ismarked out. A political crisis could,therefore, arise even before the end of1997. I am not predicting a crisis such asthe abortive Communist coup and mili-tary takeover by Suharto in the mid-1960s, because Indonesia has a growingmiddle class, but the transition afterSuharto will be difficult, and politicalfactors will weigh on the market.

Table 1. Interest Rates in Asia

Country Nominal Rate Inflation Rate Real Rate

China 7.5% 5.0% 2.5%Hong Kong 5.6 5.2 0.4India 10.1 6.1 3.9Indonesia 13.5 7.0 6.5Korea 12.0 5.1 6.9Malaysia 7.4 3.6 3.8Philippines 11.5 4.4 7.1Singapore 3.2 1.5 1.7Taiwan 5.3 3.8 1.5Thailand 9.7 4.6 5.1



ThailandThe stock market in Thailand reflects

the political situation in the country. Thecoup d’état in February 1991 was inter-esting because it led to a strong rise inthe market, which went up nearly 40percent in the next four months. Thisreaction is in contrast to what happenedin Manila in 1989, when a failed coupwas followed by the market going down50 percent because of political instabil-ity. With the successful coup in Thailand,political stability was not in question;the coup, in fact, removed a corruptadministration. Since then, the politicalsituation in Thailand has changed.Today, the failure of successive adminis-trations has led to a drop in politicalconfidence, as well as the economicproblems of a current-account deficit.

The situation in Thailand is seriousand will take some time to resolve.Although value is beginning to emergein stocks, particularly in the bankingstocks and in the blue chips, such as SiamCement, investors face the same ques-tion as they did with the Japanese banks:How long will it take the Thai banks toresolve what is a growing bad-debt, non-performing-loan situation? And as withTokyo, it may take several years beforeThailand emerges from this crisis andlending and economic growth resumethe pattern of the last 10 years. Whetherinvestors should buy Thai equities atwhat is apparently five or six times earn-ings today is questionable.

TaiwanTaiwan also had a tremendous bull

market in the 1980s followed by a dra-matic correction in the early 1990s. Themarket was beginning to recover, asdemocracy in Taiwan became moreestablished in 1994 and 1995, when itwas hit by the pressure from China onPresident Lee Tung-hui, and in March1996, during the missile tests that wereconducted off the island during the elec-tion, the Taiwan market fell dramati-cally. The tests had the opposite effect ofwhat China expected; Lee was reelected

with a large majority, and since then, themarket has doubled. The reason is thatthe economy in Taiwan has been in rea-sonably good shape, with a lot of liquid-ity in the market. Earnings are pickingup even though corporations in Taiwanare seen as a play on China.

Investors should never ignore theChina–Taiwan relationship in assessingpolitical risk in Taiwan. The recentannouncement of planned direct ship-ping links between Taiwan and main-land China is of great significance, eventhough the announcement did not havemuch immediate impact because it willtake some time for these links to develop.The shipping links suggest great poten-tial for Taiwan to increase its trade withthe mainland; at the same time, they mayhave a negative impact on Hong Kong.So, if the reversion of Hong Kong toChina goes well, gradual economic (andperhaps political) integration of Taiwanwith China, with Taiwan preservingsome of its autonomy, may be in store.

South KoreaThe Korean stock market has given

investors a ride similar to those in otherAsian developing economies: a tremen-dous rise in the 1980s followed by acorrection. In the early 1990s, the marketwas strongly affected by the surge ininflation and also by political factors.The arrest of the two ex-presidents, RohTae Woo and Chun Doo Hwan, in 1996was a big shock to the market. A similareffect on the market occurred in the lat-ter half of January 1997; the Hanbo scan-dal deeply affected confidence in the bigKorean companies and in Korea’s polit-ical leadership.

Hong KongHong Kong provides some of the

most interesting and detailed examplesof stock selection being influenced bypolitical factors. A major shift hasoccurred in the Hong Kong market sincethe early 1980s. At that time, investorswere concentrated on the old “Hongs”—Hong Kong Bank, Jardine Matheson,



Swire Pacific, Hong Kong Land, ChinaLight and Power, and so on—British-controlled trading companies, banks,and utilities. Then, during the 1980s, ashift occurred toward the new HongKong/Chinese elite. The most promi-nent company, Cheng Kong, is con-trolled by Li Ka Shing, who in turn,controls Hutchinson, which togetherconstitute a large percentage (almost 20percent) of the Hong Kong stock markettoday. New World, Sun Hung Kai Prop-erties, Henderson Land—all of thesegroups are controlled by the major Chi-nese billionaires in Hong Kong.

Since about 1992, a third group hastaken over in long-term importance andsignificance: “red chip” companiesunder the control of mainland China.Initially, these stocks were mainly ofmanufacturing companies with opera-tions in China, but they are increasinglytrading and property groups with par-ent companies controlled by Beijing.The classic red chip is CITIC (ChinaInternational Trust and Investment Cor-poration) Pacific, and it exemplifies theapplication of political analysis to spe-cific stocks. CITIC was establishedwhen Deng Xiaoping came into powerin 1979 by Rong Yi-ren, a close friend ofDeng, and the company was in theunique position of reporting directly tothe State Council. This unique politicalconnection, what the Chinese callguang-xi, was a major reason behind themeteoric rise in the share price of CITICafter its Hong Kong subsidiary waslisted in 1986. The chair’s son, LarryYung, is now on the board of the HongKong Jockey Club, which is more impor-tant than it sounds, and is now perhapsthe leading new mainland China tycoonin Hong Kong.

A political connection, however, canbe a two-edged sword. Since the deathof Deng Xiaoping, a shift is alreadyoccurring in who a company should beconnected with and how to value thatconnection. Recently, the parent com-pany of CITIC, the Chinese government,sold down its stake from 35 percent to

18 percent to allow Yung and his col-leagues to acquire more shares. Thissomewhat questionable transaction hadthe effect of decreasing investor confi-dence in this stock; it fell nearly 20 per-cent. The way mainland China respondsto shareholders is still fairly unsophisti-cated, which affects the ratings investorsshould give the red chip companies.

In addition, China obviously wantsto control certain strategic industries inHong Kong, and CITIC is a reflection ofthat stance. It has stakes in CathayPacific, the airline, and in Hong KongTelecommunications, the telephonecompany; and it recently announced amajor stake in China Light and Power,the electric utility. Another two indus-tries China is likely to focus on are bank-ing and the media. In analyzing theirportfolio compositions in the next yearor two, investors should think aboutthese political factors: What does Chinawant to control? Will China develop itscontrol of the economy in Hong Kong bytaking stakes through the stock market?

The long-term effect of China’sactions on Hong Kong stock perfor-mance is hard to discern. For example,at the time, Tiananmen Square affectedHong Kong investors deeply, with mar-ket values falling some 40 percent. Butwith time, the effect of TiananmenSquare is hardly apparent in the long-term Hang Seng Index performance.Although several months passed beforepeople believed that business would goon as usual, it did, and this normalcywas followed by a long, long bull mar-ket that has pushed the Hang SengIndex from 2,000 to 13,000 in six to sevenyears. So, sentiment in Hong Kong haschanged remarkably since TiananmenSquare.

The key to individual stocks in HongKong is property. The consensus inHong Kong today is very bullishbecause people think that property canonly go up so the stock market can onlygo up. However, although an enormousamount of money is coming into HongKong from China, the current level of



the property market will be difficult tosustain. The first measures the new gov-ernment under Tung Chee-hwa takesmay have to be to make property moreaffordable to the middle-class people inHong Kong. The tycoons live extremelywell, but the ordinary people in HongKong are having increasing difficultyfinding affordable property to buy.Thus, the property market, the keyinfluence on the Hong Kong stock mar-ket in the past 10 years, is subject topolitical considerations.

The future of the Hong Kong stockmarket will increasingly lie with themainland companies. Some of the keyChina holdings, all those businesses inShanghai that I noted for their politicalconnections in China, also have favor-able balance sheets, earnings, and creditratings. The underlying businesses areshowing strong growth, but the P/Es insome cases are reasonable. ShanghaiIndustrial, for example, which waslisted in early 1996, has done extremelywell and is in a strong position. Goodeconomic and business reasons justifybuying these stocks, and within the nextfive years, 50 percent of the Hong Kongmarket will probably be accounted forby China listings.

RussiaOne of the potentially most lucrative

areas to invest in, I believe, is Russia.Lloyd George Management hesitated along time to invest in Russia because ofthe obvious political and market risks.We have just opened a fund, however, toinvest in energy companies in Siberia.Companies that will supply the rapidlygrowing areas of Asia in the future areunquestionably undervalued assets atthe moment. Moreover, a qualitativeimprovement in political risk hasoccurred in Russia as free markets havedeveloped in the last two years. The freemarket is clearly in Russia to stay, socompanies exploiting assets such as oil

will increasingly rise toward the valua-tions typical of such companies else-where in the world.

VietnamSome fund management groups have

bravely gone into Vietnam, but invest-ments there are not easy to find.

MyanmarMyanmar, the former Burma, has a

stock exchange, but I see no political oreconomic basis for investing there. It hasone of the least desirable politicalregimes in the region; no investor withany ethical principles would want to beassociated with it, and doing businessthere would inevitably involve such anassociation.

Conclusion

Asia continues to be the number onegrowth region in the world, and theunleashing of capitalist energy in Chinaallied with the powerful resources of theoverseas Chinese is a secular phenome-non that will continue to dominateworld trade for years to come. The long-term trend for investing in Asian devel-oping markets is up. That is why buyingopportunities occur whenever a majorbreak in the trend occurs, as in HongKong in 1989 and, more recently, in Thai-land and Korea. It is also why I recom-mend a contrarian approach to stocks.

The experience of investing in Asialeads to one key conclusion: Be welldiversified. Events in these countries arehighly unpredictable, as demonstratedby Tiananmen Square and crises in thePhilippines and Thailand. For the longterm, investors should strive for bal-anced portfolios in which Hong Kongwould be slightly outweighed by theother three major markets—Singapore,Malaysia, and Thailand—as a groupand which would have increasing por-tions of Taiwan and South Korea.


Faber.frm Page 79 Wednesday, November 12, 1997 2:55 PM

Change and the Next Emerging Markets Marc Faber

Managing DirectorMarc Faber Limited

Signs are that investors now inhabit a New Era of expanding economic opportunities, accelerated

information transmission and economic development, and renascence of the capitalistic system. In a New Era, shifts occur in the centers of prosperity, consumption

patterns change, and growth industries do not necessarily produce profitable stocks. While seeking

the next emerging markets, investors need to remember that growth has drawbacks as well as benefits and that the history of foreign investments has not been rosy. Moreover, today’s slowdown in the growth of Asian

developing economies may reflect a structural change that will necessitate greater investment selectivity than

in the past decade.

he expression “punctu-ated equilibrium” is usedin the natural sciences todescribe how an estab-lished equilibrium is sud-denly disturbed by a shock

or a major event. Consider, for instance,the dinosaurs: They dominated theearth for approximately 500 millionyears, and then something happened(we do not know exactly what), and in ashort period of time, they disappeared.Or consider an example from militarywarfare: Napoleon’s armies could movehardly any faster than Julius Caesar’sLegion approximately 2,000 years ear-lier; more likely, they were even slowerbecause Roman roads had been supe-rior. Then came the invention of the rail-road, and within 60 years, an entirearmy could be shifted at about 60 milesper hour. That punctuation changed theentire art of war almost overnight.

In the last few years, I believe, theinvestment management world has been

part of an environment of punctuatedequilibrium. In politics, economics, andthe social sphere, major changes haveoccurred in the world. Because of thesechanges, many people are referring tothe present age as a New Era, and I agree.

The changes are occurring in a varietyof areas and for a variety of reasons.First, many formerly closed countrieshave opened because of the breakdownof the socialist/communist ideology orthe abolition of policies of national self-reliance. As a result, the economic uni-verse has expanded by about as much asit expanded at the end of the 15th cen-tury when the explorers opened up theAmericas and the East Indies or by aboutas much as the world was enlarged bythe entry of the U.S. economy into theglobal economy in the 19th century. Ifthey have know-how or capital or spe-cial technologies, companies suddenlyhave many more business opportunitiesto capitalize on than previously.

Following the failure of communism

T



and socialism, the accepted doctrine isthat the market economy and the capi-talistic system are superior to anythingelse. Investment managers, then, wouldseem to be living in, in Voltaire’s words,the best of all possible worlds. Theacceptance of free market doctrines andcapitalistic systems has caused an erup-tion of new financial markets all over theworld. Financial markets are growingrapidly, and whereas only 10–15 yearsago capital was concentrated in nationalmarkets, capital can now flow from onecorner of the world to another at thetouch of a computer key.

The second area of major changetoday relates to the new technologies incommunications, computing, and otherspheres that have brought about anacceleration of information transmis-sion. A hundred or so years ago, sendinga letter from Asia to Europe took threemonths. Today, because of instant com-munication, a company that introducesan innovative new product may findthat product can be copied, improved,and made obsolete within a year, some-times within weeks. Technology transfercan occur almost overnight. As a result,new countries are opening up, industri-alizing, and emerging into the globaleconomy at an unprecedented pace.Therefore, the pace of economic devel-opment has accelerated dramatically.

Of course, these changes have somedrawbacks. On the ideological side isthe issue of whether the market and thecapitalistic system can fulfill every-body’s expectations in the future, espe-cially in the emerging economies. Otherissues relate to the drawbacks that comewith growth—current problems thatemerging economies, not in Asia alonebut also in other regions of the world,are facing.

Growth Issues

When new eras are beginning, thegrowth of new industries, companies,and products is frequently underesti-mated and the supply is overestimated.

For example, Ferdinand Foch, a profes-sor at the Ecole Supérieure de Guerre inParis before World War I, opined that“airplanes are an interesting toy but of nomilitary value.” Tom Watson, Jr., thenchair of IBM Corporation, said in 1943, “Ithink there is a world market for maybefive computers.” As late as 1977, KenOlson, founder of Digital Equipment,said “There is no reason for any individ-ual to have a computer in their home.”

On the other hand, whether relatingto the weight of a white rat, or Spanishtrade in the 16th century, or the horse-drawn vehicles industry, or railroads inthe last century, growth always deceler-ates after a rapid growth accelerationperiod.

For investors, the issue of most impor-tance is obviously timing—investingduring the acceleration phase. Expecta-tions are exceeded in the accelerationphase and tend to be disappointed dur-ing the growth deceleration phase. Thispattern is also relevant to assessing polit-ical risk in the emerging economies, par-ticularly some of the countries that areonly now opening up. As long as peo-ple’s standards of living are improvingand their expectations are beingexceeded, social stability will prevail; ifone day expectations are disappointed,the result may be tensions in the socialsphere.

Uncertain Stock PerformanceThe performance of stocks does not

necessarily mirror the growth curve of acountry’s economy or of an industry.Early in the growth cycle, stocks tend toperform well. Later, even though theeconomy or industry may continue togrow for a number of years, the stocks,because of competition and sometimesregulation, may no longer perform well.The U.S. railroad industry is a goodexample. In recent years, citing how rail-roads improved productivity in theworld in the 19th century has becomefashionable. This claim is absolutelytrue, but the point is that railroad stockspeaked in the United States in 1854 to



1862 and performed rather badly there-after; by 1893, 85 percent of all railroadsin the United States were either bank-rupt, in reorganization, or had to be refi-nanced. Even in the 20th century,although industrial stocks were risingbetween 1906 and 1921, railroad stocksperformed poorly. Because of regulation,they could not increase rates sufficiently.By 1932, at the bottom of the market,railroad stocks were no higher than theyhad been about 100 years earlier.

Frequently, the industry that is grow-ing the most, such as railroads in the lastcentury, is not the best place for inves-tors to put their money. For example, theprimary beneficiaries of U.S. railroadswere other industries, not the railroads.Railroads benefited the farmers in theMidwest and allowed the industrializa-tion of the Great Lakes area because theyenabled products to be shipped from theMidwest to the coastal areas for ship-ment abroad. Then came the automo-bile, and the railroad industry suffered.When people compare software andmodern communications to railroads,they should think not only about thecompanies involved in the industry,which do not always make the money,but about the others that can use the newtechnology.

The automobile industry is anotherexample of a growth industry that wasnot necessarily a profitable long-terminvestment. Around 1910, the automo-bile industry could be termed a growthindustry. Car sales rose from 100,000units in 1910 to about 4.4 million unitsin 1923. More than 100 car manufactur-ers were operating in the United Statesin 1910. The problem was that many ofthe companies were not particularlyprofitable. Many fell by the wayside—leaving the United States with the threecar manufacturers of today.

An investor who recognized thegrowth potential of the automobileindustry but realized that picking outthe survivors would be a problem mighthave decided to invest in suppliers to theautomobile industry. The obvious choice

in those days was a rubber plant or rub-ber plantation, but this choice also wasnot good. Rubber prices peaked in 1908,and in spite of global car sales goingfrom 100,000 units in 1910 to more than45 million units today, the price of rubbernever made a new high. Why? Supplywas increasing continuously, new pro-duction methods for rubber allowedmore efficiency, and prices never fullyrecovered. Those who bought rubberplantations in 1908 or 1912 had lost a lotof money by the time the Great Depres-sion was in full swing, when the pricehad dropped by more than 90 percent.

Shifting Centers of ProsperityIn the New Era, in which technology

can be transferred almost instantly,developing countries can join the indus-trialized process very rapidly. The resultis tremendous shifts in the centers ofprosperity. New cities are rising; oldones are decaying. New markets areemerging; some developed markets aresuffering. New regions—Asia, parts ofLatin America, and now also parts ofEastern Europe—are displacing olderones. Such changes are not new, but theyare taking place much faster than previ-ously. Of the 10 largest cities in theUnited States in 1850, seven—Baltimore,New Orleans, Cincinnati, St. Louis, Pitts-burgh, Buffalo, and Cleveland—are nolonger on the list of the 10 largest citiesin the United States. Los Angeles was noton the list until about 1930; it is now thesecond largest city in the United States.

Asia will grow and Eastern Europewill grow, and some benefits will result,but each instance of a major punctuationin the equilibrium destroys someparticipants as wealth shifts to new cen-ters. In Asia, the older, established cen-ters of prosperity are growing at a muchslower rate than the explosive rate of thenew regions. For instance, previously,there were no links between the islandssouth of the Philippines—Mindanoa,Sabah, Kalimantan, and the CelebesIslands—with Manado, which was thespice center in the 16th century. Now,



because of political developments,direct shipping and airline links exist.When such links are built, trade canimmediately explode. Other growthareas are northeast of Myanmar(Burma), north of Laos, and northwestof Vietnam and the Guiyang Province ofChina. Because the borders have openedup, trade is growing tremendously. Thearea abounds with rapidly growing(and extremely sleazy) Wild West cities.Unfortunately, investing in that regionis not possible unless you move thereand establish yourself as a big-timesmuggler or nightclub operator.

Another wildly growing area is theYangtze River basin. When China beganto open up in 1978, growth was concen-trated in the Pearl River delta near HongKong. It later expanded to GuangdongProvince, and growth has now movedup into the Shanghai–Tianjin corridorand the whole Yangtze River basin. Thatregion is now growing above thenational average, whereas the south isencountering a growth slowdown.Northern Asia offers another area of tre-mendous growth potential, with Japanand South Korea offering technologies,Asian Russia offering its resources, andChina providing land and labor.

Changing Patterns of Consumption

Rising standards of living and the open-ing of countries to trade change con-sumption patterns. Figure 1 illustratesthis phenomenon by showing the trendsin spending for Hong Kong from 1975through 1993. Note the decline in thepercentage of consumption spent onfood and the rise in percentage of con-sumption spent on consumer services.Spending on food is being replaced bydiscretionary spending on goods suchas soft drinks, cigarettes, cosmetics, andcars, and on services such as financialservices, credit cards, insurance, bank-ing, and so forth.

The breakdown in communism hasmeant that this pattern is being repeatedin the growth areas of Asia and EasternEurope. As the many previously closedcountries open up and grow, however,the noncommunist countries that werebeneficiaries of communism (becausethey did not have any competitionbetween 1950 and 1978) are strugglingto maintain their previous growth rates.When countries open up that were pre-viously planned economies, consumerspending proliferates because discre-tionary consumption is practically non-

Figure 1. The Switch into Consumer Services in Hong Kong, 1975–93

Source: Based on Hong Kong government estimates of GDP.

30

28

26

24

22

20

18

16

14

12

10

Per

cent

Food as a Percentage ofTotal Consumption

75 9376 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92

Consumer Services as a Percentage ofTotal Consumption



existent under planning. People havenarrowly limited choices; a suit mightbe either blue or green, shoes might bebrown or black, but there are no brands.The market economy brings choices,and people want to distinguish them-selves from other people. This environ-ment benefits companies that have abrand; they can go into a market andimmediately grab 50–70 percent of themarket for, say, soft drinks, in the case ofCoca-Cola, or cigarettes, in the case ofPhilip Morris. Because there were nobrands in the market under planning,brands are imported or develop quickly.

As they become richer and their stan-dard of living rises, people also workfewer hours. In many of these countries,the number of hours worked per monthhas declined significantly since 1990,and people have more free time forentertainment. This development bene-fits the sports and consumer industries.Soccer, for instance, is likely to becomethe sport of the masses in China; golf isthe sport of the rich in Asia. Show busi-ness, movies, music, leisure activities,and shopping have all grown. Whenpeople have free time, they also have apropensity to be sick more often, whichbenefits the health care industry. Theconstruction of hospitals leads, in turn,to a boom in ancillary services.

Higher standards of living andincreased leisure time also encouragethe travel industry. In England, thedeparture rate—that is, the percentageof the population that leaves the countryat least once a year—is 55 percent. InJapan, the departure rate is about 12percent and growing rapidly; it shouldgrow to 40 percent in the next 10–20years. In China, the departure rate isonly 0.25 percent. The Chinese maynever have a 55 percent departure rate,which would mean 550 million touristsleaving China a year, but a 10 percentdeparture rate is not unreasonable. Thisdevelopment will create a boom in thetravel industry worldwide.

Another issue related to changingconsumption that is important for the

developing world is that, as it rapidlygrows, its demand for resources rapidlygrows. The developing countries havebecome much more important percent-ages than in the past of the global marketfor commodities. The developing coun-tries are consuming more coffee, morecocoa, and more metal—more of allcommodities. The per capita consump-tion of cocoa and coffee is extremely lowin Asia but growing rapidly. The Japa-nese have adopted coffee drinking, ashave the Koreans, the Taiwanese, andthe Hong Kong Chinese. When Chineseper capita consumption of coffee reachesthe same levels as in the Western world,the Chinese will consume roughly fivetimes the global coffee production. Inthe future, when China is a big buyer ofcommodities, commodity prices will goup, and when China overstocks and liq-uidates, commodity prices will tumble,which will add an element of volatilityto the commodity markets.

Investment Implications

In the life cycle of emerging markets,depicted in Figure 2, the economy of adeveloping country rises from adepressed stage to a boom stage andthen to a phase of euphoria, followed bydisappointment; then, the cycle repeatsbut without the extreme highs and lows.The question for the emerging marketsis: Having underperformed the U.S.market since January 1994, where arethe markets in this life cycle? This ques-tion cannot be answered for the emerg-ing stock markets in general becausedifferent emerging economies dance todifferent tunes. Some are poised for thenext bull market, others are not. In theyears to come, therefore, selectivity—not only among emerging markets orcountries but also within those marketsin terms of industries and companies—will become increasingly important.Despite the rising popularity of index-ing around the world, the stage follow-ing disappointment may be the timewhen indexing will fail.



Changing EconomicsAsian economies are in the midst of

various changes, but considerabledebate is going on about whether thechanges are structural or cyclical.Because many of the Asian stock mar-kets are lower today than in 1990,despite strong economic growth in suchcountries as Indonesia, South Korea,and Thailand, I believe that some struc-tural changes have taken place.

Some observers believe that a majorimprovement will occur in the growthof Asian economies in the next two yearsas a result of a pickup in exports. I dis-agree. First, the Western world, which isthe consumer of Asian products, is in asecular decline in the rate of growth ofconsumption because of demographics.The group that consumes the most,young people, especially the 25-yearolds, is declining. The year 1997 marksthe biggest decline in that group ever,

which has led some economists to arguethat this year will bring a recession,which is quite possible. Second, the con-sumer in the Western world is heavilyindebted and thus cannot consume orincrease consumption by as much as inthe 1980s. Finally, real wages are declin-ing globally. The addition of 3.5 billionpeople who are prepared to work forUS$50 to US$60 a month to the globallabor force is depressing the wages, aswell as abilities to consume, of unskilledlabor around the world.

Another reason for the slowdown inexports from Asia is that Asian exportshave grown enormously and now, at 25percent of global exports, are alreadysubstantial. That growth is obviouslydecelerating. Exports from Asia havebeen slowing since 1989. In the 1985–90period, exports grew at about 30 percenta year, but that growth rate has slowedto 5–12 percent.

Figure 2. The Life Cycle of Emerging Markets

PhaseOne

PhaseZero

PhaseTwo

PhaseThree

PhaseFour

PhaseFive

PhaseSix



This slowdown has brought Asiancountries from large current-accountsurpluses and excess liquidity in the late1980s to large current-account deficitsand capital deficiency. In other words,Asian developing countries, like mostother emerging economies, are depen-dent on foreign capital to finance eco-nomic growth.

In the case of Asia, the slowdown inexports has come at an inopportunetime. In the first growth phase in Asia,the economies were growing rapidlybecause of private initiatives but theinfrastructure was neglected. Now, asTable 1 shows, the region must spendabout US$1,500 billion on infrastructureprojects. Unless the infrastructure isbrought up to standards, Asia cannotgrow and bottlenecks will develop.China needs to invest US$300 billion intransport alone. (The Chinese railroadsystem today has fewer miles in opera-tion than the United States had in 1870,but China today has 1.2 billion peoplewhereas the United States in 1870 had apopulation of 60 million.) The magni-tude of the investment needed in thearea is obviously draining liquidity.

In summary, I think emerging econo-mies, particularly in Asia, will experi-ence slower economic growth than inthe past. It may be 4–6 percent ratherthan the 6–10 percent it was in the last10 years or so. The drain on liquidity

will make it difficult for financial mar-kets to rise to the heights of the late1980s. These markets may echo Ger-many’s experience after World War II,when the stock markets had a hugeboom but the high was reached in 1962and, even though Germany was grow-ing very rapidly in the 1960s and 1970s,was not exceeded until 1986.

Moreover, economic growth in Asiawill not necessarily translate into risingstock prices across the board. Investorswill have to be selective; one year, onemarket will perform well, and anotheryear, another market that may not haveperformed well previously will performwell. Markets such as South Korea,Thailand, Sri Lanka, Pakistan, and Indiawere all hard hit in the last few years;maybe they will bottom out and thenprovide good opportunities in the nextfew years.

The Impact of ChinaTo understand China, one must

understand how the economic land-scape of Asia looked before the Commu-nist Party took over China in 1949. Thecenters of economic activity wereShanghai and Manchuria. Table 2shows the distribution of foreign directinvestments in China before the com-munist takeover. Almost 50 percent offoreign investment in China (includingHong Kong) was in Shanghai. Even the

Table 1. East Asia: Infrastructure Investment, 1995–2004(US$ billions except percentages)

Country PowerTelecom-

municationsTrans-

portationWater and Sanitation Total

Total as Percent of

GDP

China 200 141 302 101 744 7.4%Indonesia 82 23 62 25 192 6.8South Korea 101 32 132 4 269 5.6Malaysia 17 6 22 4 49 4.8Philippines 19 7 18 4 48 6.8Thailand 49 29 57 10 145 7.2Othera 25 18 14 4 61 7.5

Total East Asia 493 256 607 153 1,508 6.8%aCambodia, Fiji, Kiribati, Laos, Maldives, Mongolia, Myanmar, Solomon Islands, Tonga, Vanuatu, Vietnam, and Western Samoa.Source: Data from “Infrastructure Development in East Asia and Pacific,” World Bank (1995).



British invested more than 75 percent oftheir US$963 million in Shanghai andonly 9 percent in Hong Kong. Before thecommunist takeover, Hong Kong wasan insignificant port village. Hong Kongbecame the global center that it isthrough a historical accident. WhenChina closed down under the commu-nists and the embargo of China duringthe Korean War, Shanghai’s entrepre-neurs moved to Hong Kong and desti-tute workers came to Hong Kong asrefugees. The refugees had to work atany salary, which gave the lift to the firstindustrial production in Hong Kong—textiles and, later, toys, plastics, andelectronics.

Indeed, when the communists tookover China in 1949, all of Asia changed.Taiwan, for example, would be an islandwithout any economic importance butfor the closing of China and, later, theclosing of Vietnam, Burma, Laos, andCambodia. Now, as these countriesbegin to open up, Asia is once againchanging.

When China began to open up in1978, growth centered in the special eco-nomic zones that were put around HongKong in the Pearl River delta, notablyShenzhen. In recent years, growth hasshifted to the Shanghai–Tianjin corridor.Jiangsu Province and Zhejiang Province,surrounding Shanghai, already havehigher GDPs per capita today thanGuangdong Province, even thoughJiangsu and Zhejiang began to open uponly in 1988. Shanghai’s growth nowexceeds that of the Guangdong region.The central government has a lot at stake

in Pudong, a new city across the riverfrom Shanghai, and will ensure its healthand growth. The government will giveprivileges to people who settle in Pud-ong, so it will eventually compete withHong Kong.

Asian Real EstateOne of the growth industries in Asia

but also one of the problem areas at themoment is real estate. Massive construc-tion everywhere has led to buildingoversupply in many countries and cit-ies. In particular, China has a tremen-dous oversupply in the commercialproperty market. These properties willbe prizes in Shanghai and Beijing withinthe next two years. Eventually, propertyprices in Shanghai will exceed those inHong Kong. Hong Kong property pricesare selling at a huge premium, aboutfive times higher than in Shanghai andother Chinese cities, but with HongKong becoming a Chinese city after July1, 1997, arbitrage is likely. Therefore,investors should be short Hong Kongproperties and long properties in Shang-hai and other centers.

Property investments in Asia—con-sidering that the property market iscyclical—will have generally goodreturns if bought at reasonable prices,because urbanization in Asia, althoughprogressing rapidly, is low. South Koreaillustrates the progress of urbanizationin Asia: 74 percent of the population wasliving in cities in 1990, up from only 21percent in 1950. In China, about 25percent live in cities, but I expect thispercentage to increase to 50 in the next

Table 2. Geographical Distribution of Foreign Direct Investments in China in 1931(US$ millions except percentages)

AreaGreat

Britain Japan RussiaUnited States Total

Percent of Total

Shanghai 737.4 215.0 — 97.5 1,049.9 46.4%Manchuria — 550.2 261.8 — 812.0 36.0Rest of China

(including HongKong) 226.0 108.9 11.4 52.7 399.0 17.6Total 963.4 874.1 273.2 150.2 2,260.9 100.0%

Source: Data from C.F. Remer, Foreign Investments in China (New York: Macmillan, 1933).



15 years. Therefore, growth in the sub-urbs of cities will be dramatic. Duringthe next 15–20 years, a billion peoplewill move from the countryside in Asiato the cities.

Population GrowthThe growth in population continues

to explode in Asia, as Figure 3 illus-trates. In China, estimates are that thepopulation began climbing in the 1700s,dipped only once, in the late 1800s toabout 415 million, and has been climb-ing ever since then. It reached about 835million in 1975 and could exceed 1.2billion by the year 2000. On the Indiansubcontinent, the population began tosoar in the early 1800s; by 1975, it hadgrown from about 190 million to about775 million and is on the way to 1.3billion by the year 2000.

The population growth of Asia is amixed blessing. It may be good for someproperty markets, but the explosion of

populations in India, China, and Indo-nesia is leading to rising wealth dispar-ities within Asian countries and amongAsian countries. GDP per capita varieswidely among the countries of Asia. Sin-gapore with 2.5 million people has alarger GDP than Bangladesh with 110million people. Shanghai’s GDP is nowslightly higher than Singapore’s GDP. Inthe long run, such disparities can createtensions between countries and lead toinstability in the financial markets.

Conclusion

Several geographical areas will offergreat opportunities in Asia in the next10 years. Regions that will open upinclude Asian Russia, the cities in north-eastern China and the countries in theShanghai region—Vietnam, Laos, Cam-bodia, Myanmar—and even NorthKorea. In general, politically stablecountries with low wealth disparity;

Figure 3. Asia’s Population Explosion

Note: China includes Turkestan and Tibet, China proper, and Inner Mongolia and Manchuria. The Indian subcontinent includes Pakistan, India, Bangladesh, Sri Lanka, and Nepal.Source: Data from the Atlas of World Population History.

BC

400

800

700

600

500

400

300

200

100500

The Indian Subcontinent

Popu

lati

on (m

illio

ns)

AD

1

400

800

1100

1300

1500

1600

1700

1800

1900

2000

BC

400

800

700

600

500

400

300

200

100500

China

Popu

lati

on (m

illio

ns)

AD

1

400

800

1100

1300

1500

1600

1700

1800

1900

2000



selected cities in large countries that candevelop special skills; commodity-based economies, such as Australia andNew Zealand; and Singapore andMalaysia should perform relativelywell. This performance does not mean,however, that their stock markets willdo well.

As for industries, real estate insecondary locations and on the outskirtsof large cities should do well, as shouldagricultural commodities (especiallysugar and cocoa) and precious metals(gold and silver). Other opportunitiesshould lie in consumer goods, travel,entertainment, services, and insurance.

In the next 10 years, investors shouldavoid the stock markets of Hong Kong,the Philippines, southern China,Malaysia, and any large country with alarge population and great wealth dis-parity. Industries that are likely to bedisappointing are manufacturers ofcommodity-type products, such assemiconductors, personal computers,and so on, telephone companies, infra-structure-related projects, and realestate in downtown areas—especiallyHong Kong. At one point, we at MarcFaber Limited thought that fixed-income securities would outperformequities in general in the emergingmarkets, but equities have had a strongrally since the Mexican peso crisis in1994–95. In fact, some money may flowout of fixed-income securities intoequity in some markets.

Promising emerging markets outsideAsia include Eastern Europe and thestates of the former Soviet Union, theBaltic countries—particularly Romania.In a few years, someone driving fromMoscow to Frankfurt will hardly see adifference because standards of livingwill have risen so substantially through-out Eastern Europe. At the moment,

however, Russian equities are not agood buy because of their strong rise. InLatin America, the countries of Bolivia,Ecuador, and eventually, Cuba shouldopen up in the next 10 years. Africa(particularly for its resources) and Cen-tral Asia are the last frontiers for theemerging market investor.

The emergence of so many countriesoffers investors tremendous opportuni-ties, especially through the multina-tional corporations, which know how tocapitalize on this change. The world hasalways gone in phases, however—phases in which foreigners were thrownout and then phases where they wereinvited back in. Following the end of thecolonial system after World War II, mostemerging economies were hostile to for-eigners; they instituted policies of self-reliance and restricted foreign invest-ments. In recent years, these emergingeconomies have begged foreigners tocome back and invest. If anything goeswrong, however, and the expectations ofthe populations in these countries arenot met—in other words, if the public isdisappointed and social problemsarise—the foreigners are likely to takethe blame and be expelled again.

Finally, the history of the U.S. stockmarket, the stock market that has lastedfor more than 150 years, is unusual.Most other markets in the world havebeen totally destroyed at one time oranother. The Russian stock market,which was a huge market before WorldWar I, and all the East European marketsceased to exist after World War II.Nationalizations—as in Egypt in 1954—have wiped out investors’ assets numer-ous times. If a slowdown occurs in thegrowth of emerging markets, rememberthat the history of foreign investmentshas not been particularly kind.


Documents

Developments in Global Portfolio - Harvard Universitysites.fas.harvard.edu/~css318a/handouts/Equities_and...Gloom Boom & Doom report, and is the author of The Great Money Illusion: