-
" ~FUNDAÇÃO " GETULIO VARGAS
EPGE Escola de Pós-Graduação em Economia
- -
SE\II'\ \RIOS DF PF~( >l J~ \
" rCU,\()\lJC \
"Global Capital Markets: Integration,
Crisis and Growth"
•
•
Prof. Alan Taylor (Northwestem University/NBER) (co-autoria com
Maurice Obstfeld)
LOCAL Fundação Getulio Vargas
Praia de Botafogo, 190 - 10° andar - Auditório
DATA 12/11/98 (53 feira)
HORÁRIO 16:00h
Coordenação: Prof. Pedro Cavalcanti Gomes Ferreira Email:
[email protected] • (021) 536-9250
-
fH3c1Jd C'it){2-- 5 BIBLIOTECA
'. MARIO HENR'QlJE SIMONSEN FUNJAÇAO G::OU; VARGAS
'( < ;.': ; J ( __ \ \
r _ . - I ) U ')
•
-
•
•
Global Capital Markets
Integration, Crisis, and Growth
This book presents an economic history of intemational capital
mobility since the late nineteenth century. The book begins with a
preamble that introduces the major issues and examines developments
in the eighteeneth century and before, important historical
preconditions that set the stage for a global market in the
nineteenth century. We then discuss the theory and empirical
evidence surrounding the falI and rise of integration in the global
market. A discussion of institutional developments focuses on the
use of capital controls and the pursuit of macroeconomic policy
objectives in the context of changing monetary regimes. A
fundamental macroeconomic policy trilemma has forced policymakers
to trade off beteween conflicting goals, with natural implications
for capital mobility. Understood this way, the present era of
globalization can be seen, in part, as merely the resumption of a
liberal world order that had previously been established in the
years from 1880 to 1914. Marking a reaction against that order, the
Great Depression emerges as the key tuming point in the recent
history of intemational capital markets, and offers important
insights for contemporary policy debates.
MAURICE OBSTFELD is Class of 1958 Professor of Economics at the
University of Califomia at Berkeley and a Research Associate of the
National Bureau of Economic Research.
ALAN M. TAYLOR is Assistant Professor ofEconomics at Northwestem
University and a Faculty Research Fellow of the National Bureau of
Economic Research .
NWOTECA IlARIO HENRIQUE __ qlNDAclO GETOUD 'AlUI
-
Global Capital Markets
Integration, Crisis, and Growth
MAURICE OBSTFELD
ALA'; ;"1. TAYLOR
6C~BRIDGE >, UI'
-
Contents
Foreword
Acknowledgements
Part one: Preamble
Global Capital Markets: Overview and Origins
1.1 Theoretical Benefits
1.1.1 Intemational Risk Sharing
1.1.2 Intertemporal Trade
1.1.3 Discipline
page Xl
X 111
4
5
5
8
9
1.2 Problems of Supranational Capital Markets in Practice 10
1.2.1 Enforcement of Contracts and Informational Problems 10
1.2.2 Loss of Policy Autonomy II
1.2.3 Intemational Aspects of Capital-Market Crises 12
1.3 The Emergence of World Capital Markets 14
1.3.1 Early Modem FinanciaI Development 15
1.3.2 Technological and Institutional Changes 17
1.3.3 The Rise of Global Finance 21
IA Stylized Facts for the Twentieth Century 23
1.5 Trilemma: Capital Mobility, the Exchange Rate, and
Monetary Policy
1.6 Summary
26 28
Vil
-
Vll1 Contents
Part two: Global Capital in Modero HistoricaI Perspective 31
2 Globalization in Capital Markets: The Long-Run Evidence 34
2.1 Introduction 34 .., .., Overview 36 2.3 Quantity Cri teria
41
2.3.1 The Stocks of Foreign Capital 41
2.3.2 The Size ofFlows 52
2.3.3 The Saving-Investment Relationship 59
2.3.4 Caveats: Quantity Cri teria 68
2.4 Price Criteria 71
2.4.1 Covered Interest Parity 72
2.4.2 Real Interest Parity 76
2.4.3 Purchasing Power Parity 79
2.4.4 Caveats: Price Cri teria 85
2.5 Summary 86
3 Globalization in Capital Markets: A Long-Run Narrative 88
3.1 Capital Without Constraints: The Gold Standard, 1870-1931
88
3.1.1 The Classical Gold Standard Era 88
3.1.2 Rebuilding the Gold Standard 91
3.2 Crisis and Compromise: Depression and War, 1931-46 97
3.2.1 Capital Markets and the Great Depression 97
3.2.2 Policy Response: A Consensus on Capital Mobility 103
3.2.3 World War II and its Aftermath 107
3.3 Containment Then Collapse: Bretton Woods, 1946-71 113 3.3.1
Stability Without Integration? 113
3.3.2 Leakage, then Deluge 119
3.4 Crisis and Compromise II: The Floating Era, 1971-99 121
3 A.1 Integration Without Stability? 121
3A.2 The New Global Capital Market 123
-
•
Contents IX
Part three: Lessons for Today 125
4 Open Capital Markets: Worth the Risk? 127
4.1 Introduction 127
4.1.1 Open Capital Markets in Historical Perspective 127 4.1.2
Open Capital Markets Today 127
4.2 Evidence on Benefits versus Costs 127
4.2.1 Convergence 127
4.2.2 Growth 128
4.2.3 Portfolio diversification 128 4.2.4 Consumption smoothing
4.2.5 Output volatility
128
128 4.2.6 Corrleation versus Causation 128
4.3 Implications for Today's Global Economy 128 4.3.1 The Menu
of Policy Choices 128
4.3.2 The Perpetuai Choice: Intervention versus Markets 128
Appendix: An InternatÍona{ Macroeconomic Database 129
References 130
lndex 140
-
•
2
Globalization in Capital Markets: The Long-Run Evidence
2.1 Introduction
In theory and practice, the extent of intemational capital
mobility can have
profound implications for the operation of individual and global
economies.
With respect to theory, the applicability of various classes of
macroeconomic
models rests on many assumptions, and not the least important of
these are
axioms linked to the closure of the model in the capital market.
The predictions
of a theory and its usefulness for policy debates can revolve
critically on this
part of the structure.
The importance of these issues for policy is not surprising
at
-
2.1 /ntroduction 35
hardly describe them as being integrated in a single market, as
the equality of
prices was merely a chance event. Or consider looking at the
size of ftows
between two markets as a gauge of mobility; this is an equally
ftawed criterion,
for suppose we now destroyed the barrier between the two
economies just
mentioned. and reduced transaction costs to zero; we would then
truly have
a single integrated market, but, since on either side of the
barrier prices were
identical in autarky, there would be no incentive for any good
or factor to move
after the barrier disappeared. Thus, convergence of prices and
movements of
goods are not unambiguous indicators of market integration. One
could run
through any number of other putative criteria for market
integration. examining
perhaps the leveIs or correlations of prices or quantities, and
discover essentially
the same kind of weakness: alI such tests may be able to
evaluate market
integration. but only as a joint hypothesis test where some
other maintained
assumptions are needed to make the test meaningful.
Given this impasse, an historical study such as the present
chapter is poten-
tially valuable in two respects. First, we can use a very large
array of data
sources covering different aspects of international capital
mobility over the last
one hundred years or more. Without being wedded to a single
criterion, we can
attempt to make inferences about the path of global capital
mobility with a bat-
tery of tests. using both quantity and price cri teria of
various kinds. As long as
important caveats are kept in mind about each method, especially
the auxiliary
assumptions required for meaningful inference, we can essay a
broad-based
approach to the evidence. Should the different methods all lead
to a similar
conclusion we would be in a stronger position than if we simply
relied on a
single test.
Historical work offers a second benefit in that it provides a
natural set of
benchmarks for our understanding of today's situation. In
addition to the many
competing tests for capital mobility, we also face the problem
that almost every
test is usually a matter of degree. of interpreting a parameter
or a measure of
dispersion or some other variable or coefficient. We face the
typical empirical
conundrums (how big is big:' or how fast is fast?) in placing an
absolute
meamng on these measures. An historical perspective allows a
more nuanced
view, and places all such inferences in a relative context: when
we say that a
parameter for capital mobility is big, this is easier to
interpret if we can say that
by this we mean bigger than a decade or a century ago. The
historical focus
of this chapter will be directed at addressing just such
concerns. I We examine
BUI nOle lha!. agJ..ln. auxiliary assumprions will be necessary.
and lhe caveals will be considered along lhe", ay: for example.
whal if neighboring eco no mies became exogenously more or less
Idenllcal o\er time. but no more or less integrated in terms of
transactlon costs O
-
36 Globalization in Capital Markets: The Long-Run Evidenee
the broadest range of data over the last one-hundred-plus years
to see what has happened to the degree of capital mobility in a
cross section of countries. 2
The empirical work begins by looking at the extent of
international capital
movements over a century or more, employing data on both stocks
and fiows of foreign capital. We then develop more refined quantity
criteria by looking
at the correlations of saving and investment in individual
economies over the long run. In principie, more open economies
should be better able to de-link
saving and investment decisions via externai finance. An
important discussion
of caveats ends this section. The next empirical section focuses
on price-based criteria for capital market
integration, and looks at three parity relations: covered
interest parity (CIP), real interest parity (RIP), and purchasing
power parity (PPP). In principie, ali three
relations should come c10ser to equality the more integrated
markets are. An
examination of long-run price and interest rate series since the
late nineteenth century affords a test ofthese relations. Once
again, an important section details
the caveats with this approach.
The conclusion conjectures some reasons why international
capital mobility might have varied from time to time in the
international economy over the last
century or so. Important constraints on policy makers included a
fundamental tradeoff between monetary policy choice, policies as
regards capital mobility,
and the desire for activist domestic monetary policy. We have
termed this the
maeroeeanomie palie)' trilemma 3 Consideratio!'! ofthe trilemma
illustrates the
tensions facing policymakers during the interwar period, a major
turning point in the evolution of markets in the twentieth century,
and helps us understand
the changing commitments of governments to monetary regimes,
their attempts
at sterilization, and their confiict over adherence to the
previously sacrosanct "rules of the game" under the orthodoxy of
the gold standard. In this political
economy context, the empirical evidence appears consistent with
the stylized facts of twentieth-century social, political, and
economic history.
2.2 Overview
The broad trends and cycles in the world capital market that we
will document refiect changing responses to the fundamental
trilemma. Before 1914, each of
~ Given the limitations of the data. we will frequently be
restricted to looking at between a dozen and twenty countries for
which long-run macroeconomic statistics are available. and this
sample will be dominated by today's developed countries. including
most of the OECD countries. However, we a1so have long data series
for some developing countries such as Argentina. Brazil. and
Mexico: and in some criteria. such as our opening look at the
evolution of the stock of foreign investments, we can examine a
much broader sample.
3. See section 1.5.
-
2.2 Overview 37
the world's major economies pegged its currency's price in
terrns of gold, and
thus. implicitly, maintained a fixed rate of exchange against
every other major
country's currency. Financiai interests ruled the world of the
classical gold
standard and financiai orthodoxy saw no alternative mode of
sound finance. 4
Latin American interludes of floating exchange rates were viewed
from the
main financiai centers with "fascinated disgust," to use the
words of Bacha and
Dlaz Alejandro (1982). Thus the gold standard system met the
trilemma by
opting for fixed exchange rates and capital mobility, sometimes
at the expense
of domestic macroeconomic health. Between 1891 and 1897, for
example, the
United States Treasury put the country through a harsh deflation
in the face
of persistent speculation on the dollar's departure from gold.
These policies
were hotly debated; the Populist movement agitated forcefully
against gold, but
lost. The balance of political power began to change only with
the First World
War, which brought a sea-change in the social contract
underlying the industrial
democracies. Organized labor emerged as a political power, a
counterweight
to the interests of capital 5
Although Britain's return to gold in 1925 led the way to a
restored interna-
tional gold standard and a limited resurgence of international
finance, the system
helped propagate a worldwide depression after the 1929 New York
stock mar-
ket crash. Following (and in some cases anticipating) Britain's
example, many
countnes abandoned the gold standard in the early 1930s and
depreciated their
currencies: many also resorted to trade and capital controls in
order to manage
independently their exchange rates and domestic policies. Those
countries in
the "gold bloc," which stubbornly clung to gold through the
mid-1930s, showed
the steepest output and price-Ievel declines. But eventually in
the 1930s, ali
countries jettisoned rigid exchange-rate targets and or open
capital markets in
favor of domes ti c macroeconomic goals. 6
Thcse decisions reflected the shift in political power
solidified by the First
World \Var. They also signaled the beginnings of a new consensus
on the role of
economic policy that would endure through the inflationary
1970s. As an im-
mediate consequence. howevcr. the Great Depression discredited
gold-standard
orthodoxy and brought Keynesian ideas about macroeconomic
management to
the fore. It also made financiai markets and financiai
practitioners unpopular.
Their supposcd ex cesses and attachment to gold became
identified in the public
mind as causes of the economic calamity. In the United States.
the New Deal
brought a Jacksonian hostility toward eastern (read: New York)
high finance
back to Washington. Financiai products and markets were banned
or more
.1 See Bordo and Schwartz (198.1): Eichengreen (1996). ~. See
Temm ( 1989) 6 See Temm ( 1989): Eichengreen (1992)
-
38 Globalization in Capital Markets: The Long-Run Evidence
closely regulated, and the Federal Reserve was brought under
heavier Treasury influence. Similar reactions occurred in other
countries.
Changed attitudes toward financiai activities and economic
management un-derlay the new postwar economic order negotiated at
Bretton Woods, New
Hampshire, in July 1944. Forty-four allied countries set up a
system based on fixed, but adjustable, exchange parities, in the
belief that floating exchange
rates would exhibit instability and damage international trade.
At the center
of the system was the International Monetary Fund (IMF). The
IMF's prime
function was as a source of hard-currency loans to governments
that might oth-erwise have to put their economies into recession to
maintain a fixed exchange
rate. Countries experiencing pennanent balance-of-payments
problems had the option of realigning their currencies, subject to
IMF approval. 7
Importantly, the IMF's founders viewed its lending capability as
primarily a substitute for, not a complement to, private capital
inflows. Interwar experience had given the latter a reputation as
unreliable at best and, at worst, a dangerous
source of disturbances. Encompassing controls over private
capital movement,
perfected in wartime, were expected to continue. The IMF's
Articles of Agree-
ment explicitly empowered countries to impose new capital
controls. Article VIII of the IMF agreement did demand that
countries' currencies eventually be
made convertible - in effect. freely saleable to the issuing
central bank, at the of-
ncial exchange parity, for dollars or gold. But this privilege
was to be extended
only to ;;:nresidents (not a country's own citizens), and only
if the country's
currency had been earned either through merchandise sales or as
à return on
past lending. Convertibility on capital account, as opposed to
current-account convertibility, was not viewed as mandatory or
desirable.
Unfortunately, a wide extent even of current-account
convertibility took many
years to achieve. In the interim, countries resorted to
bilateral trade deals
that required balanced or nearly balanced trade between every
pair of trading
partners. IfFrance had an export surplus with Britain, and
Britain a surplus with
Germany, Britain could not use its excess marks to obtain
dollars with which
to pay France. Germany had very few dollars and guarded them
jealously for
criticai imports from the Americas. Instead, each country would
try to divert
import demand toward countries with high demand for its goods,
and to direct its exports toward countries whose goods were favored
domestically.
Convertibility gridlock in Europe and its dependencies was ended
through
a regional multilateral clearing scheme, the European Payments
Union (EPU).
The clearing scheme was set up in 1950 and some countries
reached de facto
7. On lhe Brelton Woods syslem. see Bordo and Eichengreen
(1993).
-
2.2 Overv/ew 39
convertibility by mid-decade. But it was not until December 27,
1958 that
Europe officially embraced convertibility and ended the EPU.
Although most European countries still chose to retain extensive
capital con-
trols (Germany being the main exception), the retum to
convertibility, important
as it was in promoting multilateral trade growth, also increased
the opportuni-
ties for disguised capital movements. These might take the form,
for example,
of misinvoicing and accelerated or delayed merchandise payments.
Buoyant
growth encouraged some countries in further
financialliberalization, although
the U.s.. worried about its gold losses, raised progressively
higher barriers
to capital outftow over the 1960s. Eventually, the Bretton Woods
system 's
very successes hastened its collapse by resurrecting the
"inconsistent trinity"
or tr/lemma.
Key countries in the system, notably the U.S. (fearful of slower
growth) and
Germany (fearful of higher inftation), proved unwilling to
accept the domestic
policy implications of maintaining fixed rates. Even the limited
capital mobility
of the early 1970s proved sufficient to allow furious
speculative attacks on the
major currencies, and after vain attempts to restore fixed
dollar exchange rates,
the industrial countries retreated to ftoating rates early in
1973. Although
viewed at the time as a temporary emergency measure, the
ftoating-dollar-rate
regime is still with us a quarter-century on.
Floating exchange rates have allowed the explosion in
intemational financiai
markets experienced over that same quarter-century. Freed from
one element
of the trilemma - fixed exchange rates - countries have been
able to open their
capital markets while still retaining the ftexibility to deploy
monetary policy in
pursuit of national objectives.8
There are several valid reasons for countries to fix their
exchange rates -
for example. to keep a better lid on inflation or to counter
exchange-rate insta-
bility due to financial-market shocks. However, few countries
that have tried
have succeeded for long; eventually, exchange-rate stability
tends to come into
conflict with other policy objectives. the capital markets catch
on to the gov-
ernment's predicament. and a crisis adds enough economic pain to
make the
authorities give in. In recent years only a very few major
countries have ob-
served the discipline of fixed exchange rates for at least five
years, and most
of those were rather special cases.9 One puzzling case,
Thailand, has dropped
off the list - with a resounding crash. Even Hong Kong, which
operates as a
8. Aksma. Grilli. and ~ilesi-Ferreni (1994) and Grilli and
Milesi-Ferreni (1995) report on panel studles of the incidence of
capital controls (for 20 industrial countries over the years 1950
to 1989. and for 61 industrial and developing countries over the
years 1966 to 1989). They find that more tlexible exchange rate
regimes and greater central bank independence lower the probability
of capnal controls.
9 See Obstfeld and Rogoff (1995).
-
,
40 Globalization in Capital Markets: The Long-Run Evidenee
currency board supposedly subordinated to maintaining the Hong
Kong-U.S.
dollar peg, suffered repeated speculative attacks in 1997.
Another currency-
board country, Argentina, has now held to its 1: 1 dollar
exchange rate since
April 1991, and so joins the exclusive five-year club. To
accomplish this feat,
the country has relied on IMF credit and has suffered
unemployment higher
than many countries could tolerate. The European Union members
that have
maintained mutually fixed rates have been aided by market
confidence in their
own planned solution to the trilemma, a full currency merger due
to be con-
summated in J anuary 1999. The trend toward greater financiai
openness has
been accompanied - inevitably, we would argue - by a declining
reliance on
pegged exchange rates in favor of greater exchange rate
flexibility.
In short, the limitations that open capital markets place on
exchange rates
and monetary policy are summed up by the idea of the
"inconsistent trinity" or,
as we term it, the maeroeeonomie poliey trilemma: a country
cannot simultane-ously maintain fixed exchange rates and an open
capital market while pursuing
a monetary policy oriented toward domestic goals. Governments
may choose
only two of the above. If, monetary policy is geared toward
domestic consid-
erations, capital mobility or the exchange-rate target must go.
If instead, fixed
exchange rates and integration into the global capital market
are the primary
desiderata, monetary policy must be entirely subjugated to those
ends.
The details of this argument form the core of this book, based
on empirical
evidence and the historical r~cord, but we can already pinpoint
lhe ke:' turning
points (see Table 2.1). The Great Depression stands as the
watershed here,
in that it was caused by an ill-advised subordination of
monetary policy to an
exchange-rate constraint (the gold standard), which led to a
chaotic time of
troubles in which countries experimented, typically
noncooperatively, with al-
ternative modes of addressing the fundamental trilemma. Interwar
experience,
in turn, discredited the gold standard and led to a new and
fairly universal policy
consensus. The new consensus shaped the more cooperative postwar
interna-
tional economic order fashioned by Harry Dexter White and John
Maynard
Keynes. but also implanted within that order the seeds of its
own eventual de-
struction a quarter-century bter. The global financiai nexus
that evolved since
is based on a solution to the basic open-economy trilemma quite
different than
that envisioned by Keynes or White - one that allows
considerable freedom
for capital movements, gives the major currency areas freedom to
pursue in-
ternai goals, but largely leaves their mutual exchange rates as
the equilibrating
residual.
-
2.3 QlIantiry Criteria 41
Table 2.1. The Trilemma and Major Phases of Capital Mobility
Resolutlon of tnlemma -Countries choose to sacrifice:
Activist Capital Fixed
Era Eolicies mobilitv exchange rate Notes
Gold standard Most R are Rare Broad consensus. (crises)
lnterv.·ar Rare Several Most Capital controls esp. in (when off
goldl C. Europe. La!. America.
Bretton Woods None(O) Most R are Broad consensus. (crisies)
Float Rare Gening Gening Some consensus; except to be rare to be
common currency boards. others.
2.3 Quantity Criteria
This section employs data on the stocks and ftows of capital
between coun-
tries. that is. quantity data. to examine how the extent of
capital mobility has
changcd o\'cr the last hundred or so years. We begin by looking
at long-term
capital mobility. and we first discuss the size of foreign
investment stocks and
flows. Ceteris pariblls. a greater degree of capital mobility
should lead to larger
tlows ando with cumulation over time, larger stocks of foreign
investment. We
then relate the size of ftows to saving and investment pattems,
to see to what
extent the externaI ftows mattered in terms of the overalI
composition of saving
and in\cstment. \Ve next consider the saving-investment
correlation. an oft-
employed test that asks whether saving and investment activities
lean toward
being dc-Ilnked. as in a theoretically open economy, or tend in
the direction of
equallty. as in a closed economy.
2.3.1 The Stocks of Foreign Capital
In this section we examine the extant data on foreign capital
stocks to get some
sense of the evolution of the global market. AIthough the
concept is simple, the
measurement is not. Perhaps the simplest measure of the activity
in the global
capital market is obtained by looking at the total stock of
overseas investment at
a point In time. Suppose that the total asset stock in country
or region i, owned
by country or region j, at time tis Aij/' lncluded in here is
the domestically-
, '.
•
-
42 Globali:.atioll in Capital Markets: The Long-Rull
Evidence
owned capital stock Aiit. Of interest are two concepts: what
share of lhe total
assets of country j are held overseas? and what share ofthe
Iiabilities of country i are held overseas? Essentially, we are
interested in the measures
ForeignAssetsShareit = LAjir!LAjit; (2.1 ) j#1 j
Foreign Liabilities Shareit LAijr!LAijt. (2.2) j#i j
Note that here we are concerned with net wealth and asset
measures, since
we want to identify the extent to which the net wealth of a
country is held in its
own versus others' portfolios. There is, then, a complication to
our measures,
since, over the long-run timescales we are dealing with, there
has been a vast
multiplication in the ratio oftotal assets to net weath and
total assets to GDP. This
is because financiai development, and the increasing
sophistication of national
capital markets, has allowed the capital stock of each economy
to be packaged
and repackaged in various asset bundles, which may be held by a
chain of
assets and Iiabilities in various financiai intennediaries
between the physical
asset itself and the ultimate net wealth owner. At the
international levei, we
also need to keep the net wealth question in perspective, but
the problem is
somewhat simpler in the sense that ali net foreign claims are
true net c1aims on
a natIonal economy: should ali creditors show up demanding
paymenl, lhen,
even after a country Iiquidates its own foreign holdings, it
will still need to hand
over an amount of its own net wealth equal to the net c1aim. In
that sense, net
foreign Iiabilities represent a claim on an economy's net
wealth. Thus, in ali
that follows we must be careful to keep this distinction in
mind. 10
A relatively easy hurdle to surmount concerns nonnalization of
the data;
foreign investment stocks are commonly measured at a point in
time in current
nominal terms, in most cases U.S. dollars. Obviously, both the
growth of the
national and international economies might be associated with an
increase in
such a nominal quantity, as would any long run inflation. These
trends would
have nothing to do with market integration per se. To overcome
this problem,
we elected to normalize foreign capital at each point in time by
some measure
of the size of the world economy, dividing through by a nominal
size indexo
The ideal denominator. given that the numerator is the stock of
foreign-
owned capital, would probably be the total stock of capital.
However, construct-
ing long-run time-series for national capital stocks is fraught
with difficulty. 11
10. For cross-country evidence on the evolution of financia!
assets as a fraction of output see Gold-smith (1985).
II Only a few countries have reliable data from which to
estimate capita! stocks. Most of these
-
2.3 Quantity Criteria 43
Given these problems we chose a simpler and more readily
available measure
of the size of an economy, namely the levei of output Y measured
in current prices in a common currency unit. 12 Over short
horizons, unless the capital-
output ratio were to move dramatically, the ratio of foreign
capital to output
should be adequate as a proxy measure of the penetration of
foreign capital in
any economy. Over the long run, difficulties might arise if the
capital-output
ratio has changed significantly over time - but we have little
firm evidence to
suggest that it has. 13 Thus. our analysis focuses on
capital-to-GDP ratios of
the form
Foreign Assets-to-GDP Rati0it
Foreign Liabilities-to-GDP Rati0it
L Ajir/Yit ; Ui
L Aijr/Yi!. Ui
(2.3)
(2.4)
A still irksome empirical problem, however, arises for the
numerator. It is
in fact very difficult to discover the extent of foreign capital
in an economy
using both contemporary and historical data. For example, the
IMF has always
reported balance-of-payments ftow transactions in its
lnternationai Financiai Statistics. It is straightforward for most
of the recent postwar period to discover
the annual flows of equity, debt, or other forms of capital
account transactions
fram these accounts. Conversely, it was only in 1997 that the
IMF began reporting the corresponding stock data, namely, the
international investment position of each country. This data is
also more sparse, beginning in 1980 for less than a dozen
countries. and expanding to about 30 countries by the mid-
estlmates are accurate only at benchmark censuses. and in
between census dates they rely on comblnations of interpolation and
estimation based on investment fiow data and depreciation
assumpttons. Most ofthese esttmares are calculared in real
(constanr price) rather than nominal (current pnce) rerrns. which
makes them incommensurate with the nominally measured foreign
caflital data. At the end of the day. we would be unlikely to find
more than a handful of countries for whlch thls technique would be
feasible for the entire twentierh century. and certainly nothing
lik.: global coverage would be posslble even for recent years.
12 For the GDP data we rely on Maddison's (1995) constant price
1990 U.S. dollar estimares of output for the period from 1820.
These figures are then "reflared" using a U.S. price deflator to
obtaln estimates of nominal USo dollar "World" GDP at each
benchmark date. This approach is crude. stnce. in particular. it
relies on a PPP assumption. Ideally we would want historical series
on nominal GDP and exchange rates. to estimate a common (U.Sdollar)
GDP figure at various hlstoncal dates. This is possible for a small
sample of countries. notably the main creditor nattons In the
disrant pasr; it is also possible for the last few decades for
almost alI countries. \Ve follow this route when applying our
method to a smaller sample group of countries.
13 But for exactly the reasons just mentioned. since we have no
capital stock data for many countries. it is hard to forrn a sample
of capital-output ratios to see how these differ across time and
space. What IS typlcally the case. and the working assumptton for
most studies. is that the capital-output ranges from 3 to 4 for
most countries. both developed and developing. [Literature on
capltal-output ratios.]
-
44 Globalization in Capital Markets: The Long-Run Evidence
1990s. The paucity of data is understandable, since the
collection burden for
this data is much more significant: knowing the size of a bond
issue in a single
year reveals the tlow transaction size; knowing the implications
for future stocks
requires, for example, tracking each debt and equity item, and
its tluctuating
market value over time, and maintaining an aggregate of these
data. The stock
data is not simply a temporal aggregate of fiows: the stock
value depends on
past tlows, capital gains and losses, and any retirements
ofprincipal or buybacks
of equity, and a host of other factors. Not surprisingly this
kind of data is hard
to collect and rarely seen. Just as the IMF has had difficulty
assembling this
data, so toa have economic historians. Looking back over the
nineteenth and
twentieth centuries an exhaustive search across many different
sources yields
only a handful of benchmark years in which estimates have been
made, an
effort that draws on the work of dozens of scholars in official
institutions and
numerous other individual efforts. 14 It is based on these
efforts that we can put
together a fragmentary, but still potentially illuminating,
historical description
in Table 2.2 and Figure 2. I. Displayed here are nominal foreign
investment
and output data for major countries and regions, grouped
according to assets
and Iiabilities. Many cells are empty because data is
unavailable, but where
possible summary data have been derived to illustrate the ratio
of foreign capital
to output, and the share of various countries in foreign
investment activity.
What do the data show? On the asset side it is immediately
apparent that
for alI of the nineteenth century, and until the intcrwar
period, the British werc
rightly terrned the'bankers to the worId"; at its peak, the
British share of total
glob::d foreign investment was almost 80 perccnt. This is far
above the current
U.S. share of global foreign assets, a mere 24 percent in 1995,
and still higher
than the maximum U.S. share of 50 percent circa 1960. The only
rivais to the
British in the early nineteenth century were the Dutch, who
according to these
figures held perhaps 30 percent of global assets in 1825. This
comes as no
surprise given what we know of Amsterdam's early preeminence as
the first
global financiai center before London 's rise to dominance in
the eighteenth and
nineteenth centuries. IS By the late nineteenth century both
Paris and Berlin had
also emerged as major financiaI centers, and, as their own
economies grew and
industrialized. French and German holdings offoreign capital
rose significantly,
each ecIipsing the Dutch position. In this era the United States
was a debtor
rather than a creditor nation, and was only starting to emerge
as a major lender
and foreign asset holder after 1900. European borrowing from the
United States
in WorId War One then suddenly made the United States a big
creditor. This
14. See. for example. Paish (19xx), Staley (19xx), Woodruff
(19xx), and Twomey (19xx). 15. Indeed, the Amsterdam market was an
important source of externai finance for Britain during
the Industrial Revolution. See chapter I.
-
2.3 Quantity Criteria 45
Table 2.2. Foreign Capital Stocks
1825 1855 1870 1900 1914 1930 1938 1945
Assers
Umted Kingdom 0.5a 0.7a 4.9a 12.la 19.5a 18.23 22.ge 14.23
Franee O.la 2.5a 5.2a 8.6a 3.5a 3.ge
Germany 4.8a 6.7a l.la 0.7e
Netherlands 0.3a 0.2a O.3a l.la l.2a 2.3a 4.8e 3.7a
United States O.Oa O.Oa O.Oa O.5a 2.5a 14.7a 11.5e 15.3a Canada
O.la 0.2a 1.3 a l.ge
Japan l.2e
Other Europe 4.6c
Other 6.0c 2.0a Ali 0.9a 0.9a 7.7a 23.8a 38.7a 4 1.1 a 52.8e
35.13
World GDP Illb 128b 221b 491b 491b 722b Sample GDP 43f 76f 149f
182f 273f Sample slze 7f 7f 7f 7f 7f
Assets/Sample GDP 0.55 0.51 0.28 0.26 0.12 Assets/World GDP 0.07
0.19 0.18 0.08 0.11 0.05
UKJAII 0.56 0.78 0.64 0.51 0.50 0.44 0.43 0.40 CS/AII 0.00 000
0.00 0.02 0.06 0.36 0.22 0.43 Lwbilules
Europe 5.4a 12.0a 10.3a North Amenca 2.6a I !.la 13.7a Oceanla
l.6a 2.3a 4.5a LatIn Amen~a 2.9g 8.9g l1.3g ASla lexd. Japan) 2.4g
6.8g 10.6g Afn.:a 3.0g 4.lg 4.0g Developlng Countries 6.0g 13.0g
25.9g Ali 17.9a 45.5a 55.0a
World GDP Illb 128b 221b 491b 491b 722b Sampk GDP
Sampk slze
LiabIlIlles/Sample GDP
LiabIlltIes/World GDP 0.14 0.21 0.11
Developlng Counuies/AII 0.34 0.29 0.47
--.-':"--:~.-, ... --
-
46 Globalization in Capital Markets: The Long-Run Evidence
Table 2.2. (Continued)
1960 1971 1980 1985 1990 1995
AsseIS
United Kingdom 26.4a 551d 857d l.757d 2 .. ~89d
Franee 736d 1.105d
Germany 1.2a 257d 342d 1.I00d 1.672d
N etherlands 27.6a 178d 418d
United States 63.6a 775d 1.296d 2.178d 3,353d , Canada 129d 227d
302d
Japan 160d 437d 1.858d 2.725d
Other Europe
Other 5.9a
All 124.7a 1.963d 4.025d 10.321d 14.25ld
World GDP 1.942b 4.733b 11.118e 12.455e 21.141e 25.llOe
SampleGDP 67lf 5.922d 8.873d 17.584d 21.479d
Sample size 7f 10d 19d 28d 29d
AssetslSample GDP 0.18 0.29 0.37 0.47 0.54
Assets!World GDP 0.06 0.18 0.32 0.49 0.57
UKlAIl 0.21 0.28 0.21 0.17 0.17
US/AIl 0.51 0.39 0.32 0.21 0.24
Liabilitles
Europe 7.6a
N orth Ameriea 12.5a
Oeeania 2.2a
Latin Ameriea 9.2a 5~G '", 250g 505g 768g
Asia lexel. Japan) 2.7a 29g 129g 524g 960g Afriea 2.23 199 124g
306g 353g Developing Countries 14.la 107g 506g 1.338g 2.086g
All 39.9a 1.569d 3.685d 10.311d 1".735d
World GDP 1.942b 4.733b 11.118e 12.455e 21.141e 25.11Oe
Sample GDP 5.922d 8.873d 17.584d 21.479d
Sample size 10d 19d 28d 29d
Liabilities/Sample GDP 0.26 0.42 0.59 0.69
Liabilities/World GDP 0.02 0.14 0.30 0.49 0.59
Developing Countries/All 0.35 0.32 013 0.14
Notes and Sources:
VOlts for foreign investment and GDP are billions of eurrent
U.S. dollars.
a = from Woodruff(1967. 150-159). b = from t\laddlson (1995):
sample of 199 eountries: 1990 US dollars converted to eurrent
dollars using US GDP deflator: some interpolation.
e = from Lewis (1945. 292-97). d = from lFS (9/97). Maximum 40
eountry sample 1980-1997. Sample size varies. e = from World Bank
(1994). f = excludes "Other Europe" and "Other": GDP data from
appendix. g = from Twomey (1998: unEublished worksheets).
". .,-,
-
2.3 Quantity Criteria
1.0
0.9
0.8 ~
0.7 ~
0.6 -
0.5 ~
0.4 ~
0.3 -
0.2 -
0.1
/ /
'" / , / , /
--- AsselS/Sample GDP • AsselS/World GDP
- - - - UK share of ali asseIS - - - - - - US share of ali
assets
0.0 -------~~=--=--:=-----------------
1820
1.0
0.9 ~
0.8
0./
06
0.5
OA
0:1
02
0.1
00
IS20
1840 1860
18.\0 1860
1880 1900 1920 1940 1960 1980 2000
---Liabilities/Sample GDP • Liabilities/World GDP
- - - - LDC share of allliabilities
1880
"'-/ "-
/ "-...... / '-
...... /
1900 1920 1940 1960
Fig. 2.1. Foreign capital stocks
\....-
1980 2000
47
:;
, ." -,.,'
-
48 Globa/ization in Capital Markets: The Long-Run Evidence
carne at a time when she was ready, if not altogether willing,
to assume the
mantle of "banker to the world," following Britain 's abdication
of this position
under the burden of war and recovery in the 1910s and 1920s.16
But the dislocations of the interwar years were to postpone the
United States' nse as a foreign creditor, and New York's pivotal
role as a financiai center. After
1945, however, the United States decisively surpassed Britain as
the major
international asset holder, a position that has never been
challenged. 17
How big were nineteenth century holdings of foreign assets? In
1870 we estimate that foreign assets were just 7 percent of World
GDP; but this figure
rose quickly, to just under 20 percent in the years 1900-14 at
the zenith of the classical gold standard. During the interwar
penod, the collapse was swift,
and foreign assets were only 8 percent of world output by 1930,
11 percent in 1938, and just 5 percent in 1945. Since this low
point, the ratio has climbed,
to 6 percent in 1960, 18 percent in 1980, and then climbing
drarnatically to
57 percent in 1995. Thus, the 1900-14 ratio of foreign
investment to output in the world economy was not equaled again
until 1980, but has now been
approximately doubled. 18
An alternative measure recognizes the incompleteness of the data
sources:
for many countries we have no information on foreign investments
at ali, so
a zero has been placed in the numerator, although that country
's output has
been included in the denominator as part of the World GDP
estimate. This is an unfortunate aspect of our estirr.:!t;:::;)
procedure, and makcs ~he above
ratio a likely an underestimate, or lower bound, for the true
ratio of foreign
assets to output. One way to correct this is to only include in
the denominator
the countries for which we actually have data on foreign
investment in the numerator. 19 This procedure yields an estimate
we term the ratio of foreign
16. This Anglo-American transfer of hegemonic power is discussed
by Kindleberger (1986). 17. or course. lhis is lhe gross foreign
inveslmenl position. nOllhe nel position. The Uniled Slales
is also now lhe world's number one debtor nalion. in bOI h gross
and net lerms. She holds more liabililies lhan any olher country.
ando since the early 19805, has been. on neto a debtor country.
18. However. even lhen we cannOl necessarily infer lhallhere has
been an increase in foreign asseI ownership relalive 10 10lal
asseIS. since lhe asseI 10 GDP ralio has risen across lhe twentielh
century wilh financial developmenl (see Goldsmilh 1985). This is
nOl a problem if we view ali foreign asseIs as "outside" - bUl
allhough lhat mighl be reasonable for lhe nineleenth century. il
mighl nOl be 50 reasonable now. In lhe pasl. mOSl assel-liabilily
posilions were one way aI lhe national levei (examp!e: Brilain
circa 1900). bUlloday lhe nel ftows are much less lhan lhe gross
ftows (example: mOSl OECD countries loday). Hence. nOl ali foreign
asseIs may be "outside" and 50 the multiplicalion of asseIs can be
an issue lhal introduces biases even aI lhis levei - as when. say,
a domeslic asseI is held overseas by an instilution lhal is in lum
held by domestic investors. Given lhe macroeconomic aggregale dala
we are dealing with here. however. lhe resolulion of lhis kind of
dala problem seems impractical.
19. Thal sample of countries is much less lhan lhe entire world.
as we have nOled. Unlil 1960. il inc1udes only lhe seven major
credilor countries nOled in Table 2.2; after 1980. we rely on lhe
IMF sample from which we can identify 10. rising to 30. countries
wilh foreign investment and GDP dala.
-
2.3 Quantity Criteria 49
assets to sample GDP. This is likely an overestimate, or upper
bound, for the
true ratio, largely because in historical data, if not in
contemporary sources,
attention in the collection of foreign investment data has
usually focused on
the principal players, that is, the countries which have
significant foreign asset
holdings. 20 Given ali these concerns, does the ratio to sample
GDP evolve in a
very different way? No. It is, as expected, higher at most
points, except 1985.
The two ratios are very dose after 1980. But before 1945 they
are quite far apart:
from 1870 to 1914, the sample of seven countries has a foreign
asset to GDP
ratio of over 50 percent, far above the world figure of7 to 20
percent. Clearly,
these seven major creditors were exceptionally internationally
diversified in
the late nineteenth century in a way that no group of countries
is today. By
this reckoning, in countries like today's United States, we
still have yet to see
a return to the extremely high degree of international portfolio
diversification
seen in, sal'. Britain in the 1900-14 period, a historical
finding that sheds light
on the ongoing international diversification puzzle. 21
Is the picture similar for liabilities as well as assets?
Essentially, yes. The
data is much more fragmentary here, with none in the nineteenth
century, when
the inforrnation for the key creditor nations was simpler to
collect than data
for a multitude of debtors, perhaps. Even so, we have some
estimates running
from 1900 to the present at a few key dates. The ratio of
liabilities to World
GDP follows a path very much like that of the asset ratio, which
is reassuring:
thel' are each approximations built from different data sources
at certain time
points. though, in principie, they should be equal. Again, the
ratio reaches a
local maximum in 1914 of 21 percent, collapsing in the interwar
period to 11
percent in 1938, andjust 2 percent in 1960. By 1980 the ratio
had risen to 14
percent. and by 1985 it had exceeded the 1914 levei and stood at
30 percent.
By 1995, the ratio was 59 percent.
Finally, what about the distinction between net and gross
stocks? A cursory
glancc at the data reveals that this problem is very serious in
recent decades,
but relatively unimportant in the pre-1914 era of globalization.
The reason is
simple: in the late nineteenth century the principal ftows were
long-terrn in-
vestment capital, and virtually unidirectional at that. The key
creditor nations,
20 Th~t 15. we are probably restricted in these samples to
eountries with individually high ratios of forelgn assets to GOP.
For example. in the rest ofEurope eirea 1914. we would be unlikely
to find countries with portfolios as diversified intemationally as
the British. French. Germans. and Outch. If we included those other
eountries it would probably bring our estimated ratio down.
Howc'er. In the 19805 and 1990s IMF data the problem is much less
severe sinee we observe m~ny more countries. and both large and
small asset holders. Sample selection might not be as biased in
thls data set: for example. one of the biggest foreign investors in
1990 is France. but French data IS unavailable for most of lhe
1980s. This is interesting because it is only at the 1985 benchmark
that this ratio to sample GOP is below the ratio to "World
GOP."
21 On the intemational dlversification puzzle see K. K. Lewis
(1996)
-
50 Globali:ation in Capital Markets: The Long-Run Evidence
principally Britain, but also France and Gennany, engaged in the
financing of
other countries' capital accumulation, and in doing so,
developed enonnous
one-way positions in their portfolios. For example, circa 1914
the scale of Ar-
gentine assets in Britain 's portfolio was very large, but the
converse holding of
British assets by Argentine 's was trivial by comparsion. Thus,
the nineteenth
century was an era of one-way asset shifts, leading to great
portfolio diversifi-
cation by the principal creditor/outftow nations like Britain,
but relatively little
diversification by the debtor/inftow nations. To a first
approximation, the gross
asset and liability positions were very c10se to net in that
distant era. The 1980s
and 1990s are obviously very different: for example, the United
States became
in this period the world's largest net debtor nation. But whilst
accounting for
the biggest national stock of gross foreign liabilities, the
United States also held the largest stock of gross foreign
assets.
Thus, our discussion of the stock data, and our inferences
conceming the
recovery of foreign asset and liabilities in the world economy
after 1980 needs
considerable modification to take into account this problem.
And, indeed, it is
a significant problem for ali of the countries concemed: the
rank of countries
by foreign assets in the IMF sample, is very highly correlated
with the rank by
foreign liabilites. Countries such as Britain, Japan, Canada,
Germany, and the
Netherlands are ali big holders of both foreign assets and
liabilities. Strikingly,
when we net out the data. the result is that, since 1980, there
has been virtually
/lO change in the net foreign asset position (or liability)
position in the wcdd
economy, as inàicatcd by Figure 2.2. Thus, for ali the
suggestion that we have
retumed to the pre-1914 type of global capital market, here is
one major qualita-
tive difference between then and now. Today's foreign asset
distribution is much
more about asset swapping by countries, than about the
accumulation of large
one-way positions. It is therefore more about hedging, portfolio
diversifcation,
and risk sharing than it is about long-term finance and the
mediation of saving
supply and investment demand between countries. In the latter
sense, we have
never come c10se to recapturing the heady times of the pre-1914
era, when a
creditor like Britain could persist for years in satisfying half
of its accumulation
of assets with foreign capital, or a debtor like Argentina could
simlarly go on
for years generating liabilities of which one half were taken up
by foreigners.
Instead. still to a very great extent today, a country's net
wealth will depend, for
accumulation, on the pravision of financing from domestic rather
than foreign
sources. and issue we will shortly take up again in the
discussion of long-run
trends in capital ftows. An interesting, and c10sely related,
insight also follows fram looking at the
share of less-developed countries (LDCs) in global liabilities.
This is now at
an ali time low. In 1900. LDCs in Asia. Latin America, and
Africa accounted
-
2.3 Quantity Criteria
08~-------------------------------------------------------
~ Assets/Sample GDP 07; ---o- LiabilitieslSample GDP
I _ Assets/World GDP ___ LiabilitiesIWorld GDP _ Net
Assets/World GDP
06; ___ Net Liabilities/World GDP
05 .
o' .
o:. '
02 j
OI • .... .. -- • • • • • • • • • ~ I
00-------------------------------------------------------
IQ!l5 1900 109\
Fig. 2.2. Foreign capital stocks
51
for 34 percent of global Iiabilities. The global capital market
of the nineteenth
century centered on Europe, especially London, extended
relatively more credit
to LDCs than does today's global capital market. Is this
surprising? There are
various interpretations for this observation. One is that
capital markets are
biased now, or were biased in the past; for example, did
Britain, as an imperial
power. favor LDCs within her orbit with finance? or, today, does
the global
capital market fail in the sense that there are insufficient
capital flows to LDCs,
and an excess of flows among developed countries (DCs)? These
are hard
c1aims to prove, as market failure could be a cause, as could a
host of other
factors including institutions and policies affecting the
marginal product of
capital in different 10cations.22 Of course, this resultjust
follows from the fact
that many of the top asset holders also figure in the top
Iiability holders, and
most of them are developed OECD countries. A rival explanation
for the recent
fali in the LDC share of liabilities, and the rise of DC
Iiabilities, could be just
a move toward greater - c10ser to optima!? - global portfolio
diversification,
22. See Lucas ( 1990).
-
52 Globalization in Capital Markets: The Long-Run Evidence
and we might see this as an efficient rebalancing given the
large weight that DC
capital has in the global capital stock.
Figuring whether toa much or toa little diversification existed
at any point
must remain conjecturai, and conclusions would hinge on a
calibrated and
estimated portfolio model applied historically. This is
certainly an object for
future research. However, unless the global economy has
dramatically changed
in terms of the risk-return profile of assets and their global
distribution, we have
no prior reason to expect the efficient degree of
diversification to have changed.
For the present we can just say that, unless a massive such
change did occur in
the 1914-45 period, and unless ir was then promptly reversed in
the 1945-90 period, we cannot explain the time path of foreign
capital stocks seen in Table
2.2 and Figures 2.1 and 2.2 except as a result of a dramatic
decline in capital
mobility in the interwar period, and a very slow recovery of
capital mobility
thereafter.
2.3.2 The Size of Flows
In contrast to the previous discussion of stocks, this section
now attempts an
analogous historical survey of global foreign investment ftows
since the late
nineteenth century. 23 The stock data suggested a amrked
diminution of foreign
investment activity in the middle of the twentieth century, with
recovery to the
1900--14 leveis only seen as recently as the 1980s and 1980s.
Can the fiow data
detect a similar historical evolution?
Some basic definitions and notation will now prove useful. To
simplify, we
may define gross domestic product Q as the sum of goods
produced, which, with imports M, may be allocated to private
consumption C, public consumption G, investment 1, or export X, so
that Q + M = C + 1 + G + X. Rearranging, GDP is given by
GDP == Q = C + I + G + NX, (2.5)
\vhere N X = X - M is net exports. If the country's net credit
(debt) position vis-a-vis the rest of the world is B (-B), and
these claims (debts) earn (pay)
interest at a world interest rate r, then gross national product
is GDP plus (minus) this net factor income from (to) the rest of
the world,
GNP == Y = Q + rB = C + I + G + NX + rB. (2.6)
It is then simple to show that the net balance on the current
account is
C A == N X + r B = (Y - C - G) - 1 = S - 1, (2.7) 23. The
next!Wo sections draw heavily on A. M. Taylor (1998).
-
2.3 Quantity Criteria 53
where 5 == Y - C - G is gross national saving. Finally, the
dynamic structure of the current account and the credit position is
given by the equality of the
current account surplus (C A) and the capital account deficit (-
K A), so that
t::,.BI = BI - BI_I == CAI = -K AI' (2.8)
This section, and some that follow, will focus on the patterns
of saving (5),
investment (I), and the current account (C A) as defined above.
The basic
identity (2.7), C A = 5 - I, is central to the analysis. In
terrns of historical data collection, it proves essential to
utilize the identity to measure saving
residually, as 5 = 1 + C A. because no national accounts before
the 1940s supply independent saving estimates; rather, we have
access only to investrnent
and current-account data.
A sense of the changing patterns of international financiai
ftows can be
gleaned by examining their trends and cycles. However, a
norrnalization is
again needed. Measurement traditionally focuses on the size of
the current
account balance C A, equal to net foreign investment, as a
fraction of national
income Y. Thus (C A / n/I becomes the variable of interest, for
country i in
period t. a convention we follow here. Table 2.3 and Figure 2.3
present some
basic trends in foreign capital ftows. We measures the extent of
capital ftows
with the mean absolute value I'lICA/Yi.I' The average size of
capital ftows in
this sample was often as high as 4 to 5 percent of national
income before World
\Var I. At its first peak it reached 5.1 percent in the overseas
investment boom
of the late 1880s. This fell to around 3 percent in the
depression of the 1890s.
The figure approached 4 percent again in 1905-14, and wartime
lending pushed
the figure over 6 percent in 1915-19. Flows diminished in size
in the 1920s,
however. and international capital ftows were less than 2
percent of national
in come in the late 1930s. Again. wartime loans raised the
figure in the 1940s,
but in the 19505 and 1960s, the sizc of international capital
ftows in this sample
declincd to an ali time low. around 1.3 percent of national
income. Only in the
late 19705 and 1980s have ftows increased. though not to leveis
comparable to
those of a century ag0 24
Indl\idual country data supply some detail to fill in this
general picture. Some
countries were clearly very dependent on foreign capital inftows
before 1914,
including the well-known cases of the settler economies,
Argentina, Australia,
and Canada. ~lany of these countries had typical capital inftows
in excess of
5 percent of GDP, and in some years in excess of 10 percent. The
Argentine
figure beforc 1890 is inaccurate and surely an overstatement, as
it derives from
24 The open c"eles in Figure 2.3 (and later figures) denote gaps
in data coverage due to the two world wars. The cireles' positlons
are determined by the incomplete sample of countJies for whlch data
are available.
-
54 Globalization in Capital Markets: The Long-Run Evidence
Table 2.3. Si::.e of Capital Flows: Average Absolute Value of
CA/Y, By
Country, Selected Periods, Annual Data
ARe AOS i:AN IJ1'lK t:11'l t:RA IJEO IrA 1870-1889 .187 .097
.072 .018 .062 .029 .019 .012 1890-1913 .062 .063 .076 .027 .059
.023 .014 .019 1914-1918 .027 .076 .035 .054 .142 .031 .117
1919-1926 .049 .088 .023 .012 .039 .117 .022 .043 1927-1931 .037
.128 .036 .007 .029 .037 .018 .015 1932-1939 .016 .037 .016 .008
.029 .025 .004 .007 1940-1946 .048 .071 .065 .024 .069 .018 .034
1947-1959 .031 .034 .023 .014 .014 .015 .020 .014 1960-1973 .010
.023 .012 .019 .017 .006 .010 .021 1974-1989 .019 .036 .017 .032
.024 .008 .021 .013 1989-1996 .020 .045 .040 .018 .051 .007 .027
.016
lPN NLD NOR ESP SWE GBR USA AlI 1870-1889 .005 .060 .016 .010
.031 .045 .015 .040 1890-1913 .022 .053 .041 .014 .023 .045 .008
.037 1914-1918 .066 .043 .033 .063 .029 .035 .058 1919-1926 .021
.069 .027 .020 .029 .017 .039 1927-1931 .006 .004 .019 .018 .016
.020 .008 .027 1932-1939 .011 .018 .013 .012 .015 .011 .006 .015
1940-1946 .010 .049 .013 .019 .073 .010 .039 1947-1959 .013 .038
.031 .023 .011 .012 .006 .020 1960-1973 .010 .013 .024 .012 .007
.008 .005 .013 1974-1989 .018 .026 .052 .020 .015 .015 .014 .022
1989-1996 .021 .030 .029 .032 .020 .026 .012 .026
the rather poor quality data from this era. Even so it reveals
the extent to which
foreign finance was willing to fuel an investment boom before
the Baring crash
in 1890. AIso. note that. unlike the settler economies, the U.S.
economy had
matured by the tum of the century, and was on the verge of
becoming a capital
exporter. with saving and investment almost in balance.
The major capital exporter was obviously Britain. with between 5
percent and
10 percent of GDP devoted to overseas investment in a typical
year before 1914.
This coincided with the years of so-called "Edwardian failure"
at home, and
the increasingly promising ventures for capital within and
beyond the empire. 25
This extraordinary net ftow of capital as a share of output has
never been matched
since by any overseas investing country. AlI countries shared in
the collapse of
capital ftows in the interwar period, and few have recovered the
pre-1914 leveI
of ftows as a share of output. 26
25. See Edelstein (1982) and HaIl (1968) for more on this
phenomenon ofBritish capital outlfow. 26. That is. excepting brief
upsurges during and after the wanime periods when credits.
especiaIly
-
2.3 QlIarztity Criteria
.07 -
06 -
05 -
.Oo! .
.02 -
.0 I -
00
1860 1880 1900
o " I I I' I ,
I
1920
9, I
19o!O 1960 1980 2000
Fig. 2.3. Slze af capital ftaws: average absolute value of CNY,
15 cauntries. quinquennia. annual data
55
Gi\en that the size of flows is still smaller then a century
ago, we would
have to take this data as indicative of an incomplete recovery
of global capital
markets relative to their levei of integration in 1914. There
still cou1d be other
explanations for this path of capital flows over time, but. as
in the caveats for
the stock data, we would have to posit some large shock that
made countries
more alike. reducing the incipient flows after 1914. This is
potentially plau-
sibk \\ithin the group of most developed OECD countries where
productivity
convergence has taken place: but it stillleaves out the
potential flows betweerz core to periphery that. given the still
large development gap between rich and
poor countries today.
In order to confront that questiono Figure 2.4 examines the same
kind of flow
data on the time series of C A / Y for the postwar period, but
expanding the samplc to encompass developed and developing
countries. 27 We here divide
the world into two samples. and look at the size of flows in
each as a share of
from the l:nited States to Europe. inftated the size of
intemational transactions. For clarity . .... anlme qumquennia
(1914-18 and 19o!~6) are shown as open circles in the chart: note
that in these penods the averages are based on incomplete
samp1es.
2~ We dra .... on data from the lr"vrFs InrernarlOIlQ1 Financiai
Srarisrics here and in Figure 2.6. Note that because of errors and
omissions it seems tha! planet earth is usually running a current
account Imbalance.
"
" ....... , .. ~
-
56 Globalization in Capital Markets: The Long-Run Evidence
.w'I--------------------------------------------{)--
Industrialized Countries CNGDP
I - Developing Countries CNGDP .03 1
I
.02 ~
-.02 ~ I I
-.03 J i
-.04 -'---------------------------' 1940 1950 1960 1970 1980
1990 2000
Fig. 2.4. Size of capital ftows, postwar: CA/Y, developed and
developing country samples. annual data
each region 's output. It is apparent that there has been a
surge in capital flows to
devclopir.g countries in the 1930s and 1990s, far exceeding any
previous flows
in the preceding fifty years. At peak times this flow has
amounted to about 3
to -+ percent of dcvloping country GDP. and a very much smaller
fraction of developed country GDP. However,judged next to the size
offlows see in the late
nineteenth century, one is struck by two features of this
postwar data: first. how
small even the large flows in the 1990s are as a fraction of the
receiving region 's
output. as compared to similar receiving region in the 1890s and
19OOs; second,
one is struck by the fact that this surge in inflows was not
witnessed sooner
in the postwar period. taking about thirty to forty years to
overcome whatever
impediments to capital movement there were between core and
periphery. Thus,
even expanding to a global sample, we argue that, most likely,
the U-shape in
the long-run flow data reflect the considerable shifts in the
transactions costs
for capital arising from policy environments which became more
inimical to
capital movements after 1914, and especially so after 1929. This
phase of
relative capital immobility has perhaps only just disappeared in
the last decade
or so.
In order to further contrast the situation now and a hundred
years ago it is
worth spending a few moments examining how important capital
flows are now,
and were then. relative not just to GDP, but relative to total
capital formation.
-
2.3 Quantity Cri te ria
.60
.50 -
40 ~
.30 ~
:0 .
10 ~
00
- 10
(a) Ratio of capital inflows to investment for periphery
economies
1 78
ARG FIN AUS CAN NOR SWE
01870-1889
1111890-1913
DNK ESP
(b) Ratio of capital outflows to saving for core economies
40 -----------------------------
.30 ~
10 -
00
- 10 -
01870-1889
1111890-1913
-20------------------------------GBR NLD FRA DEU USA lPN
57
Fig.2.5. CapItal flows in relation to saving and investment,
1870--1913: 15 countries, quinquennia, annual data
Some natural questions to ask are: how important are the inftows
as a fraction of
total capital forrnation in recipient countries? and how
important were outftows
as a fraction of total saving for source countries?
Let us look first at the data for the late nineteenth century.
Figure 2.5 displays
the ratio of average capital inftows to average investment for
periphery countries
in our l5-country sample, and the ratio of capital outftows to
saving for core
countries. looking at subperiods 1870-89 and 1890-1913. Once
again, the
Argenune figure pre-1890 must be taken with a pinch of salt, but
in several cases,
especially the settler economies, we see the remarkable
importance of capital
inftows for capital formation. In several economies foreign
capital supplied up to half of investment demando This squares with
well-known data on the
stocks of foreign capital in some of the settler economies: by
1914 about 50
percent of the Argentine capital stock was in the hands
offoreigners: for Canada
, . - '
-
58 Globa/i;:ation in Capital Markets: The Long-Run Evidence
.15 -,--------------------,----------
.10
.05
-.05
-.10 ,
--o-- Industrialized Countries CNS _ Developing Countries
-CM
-.15 -'-I----------------------~ 1940 1950 1960 1970 1980 1990
2000
Fig. 2.6. Capital flows in relation to saving and investment,
postwar: developed and developing country samples, annual data
and Australia the figure was in the range 20 percent to 30
percent. 28 Clearly
these large ftows cumulated over time into a vcry strong foreign
(read, mostly
British) interest in the total capital stock of many nations
before 1914. On the
sending side. the British dominance is readily apparent in the
figure: about one
third of total British saving was devoted to overseas investment
in the 1870-
1914 period; moreover, it is acknowledged that in some periods,
for example
1900-13, this fraction crept as high as one half. 29 In
contrast, few other capital
exporters in the core could register anything like so high a
fraction of foreign
investment relative to total savings, with France, Germany, and
the Netherlands
each registering less than 10 percent of domestic savings as
destined for foreign
countries in the 1890-1913 era.
Next. we again make a comparison to the contemporary era with
this type
of measure. Using our long run database we would find postwar
ftows much
lower relative to saving and investment as compared to the
pre-1914 era, just as
we did relative to output. But clearly we cannot satisfactorily
use our existing
long-run database for this question, or we open ourselves to the
criticism that in
focusing on our long-run fifteen-country sample we are missing
the heart of the
action in today's global capital market. Instead we should tum
our attention to
28. See A. M. Taylor (1992) for more discussion of these
comparative data and sources. 29 On British foreign investment in
this era see Edelstein (1982).
-
2.3 Qlwfltiry Criteria 59
the importance of capital ftows in relation to saving and
investment in today's
core and periphery. We can use the postwar data to look at how
important
developed countries outftows were as a share of their total
saving, and how
important developing country inftows were as a share of their
total investment.
These measures would then accord with the pre-1914 data in
Figure 2.5.
For the postwar period Figure 2.6 supplies the details. As with
the discussion
of ftows as a share of output in Figure 2.4, one is forced to
note the relatively
small impact offtows until the 1980s and 1990s: inftows have
never amounted to
more than about lOto 15 percent of developing country
investment, and outftows
neve r more than about 5 percent of developed country saving.
This contrasts
with nineteenth century experience, where a country like Britain
exported as
much as half her annual savings. and where countries such as
Australia, Canada,
and Argentina, imported up to half their savings supply. And
even now, the
ftows for the core group are still small compared to the total
size of the core
capital market as measured by aggregate saving. lndeed, in many
years, both
periphery and core countries appear to have sizeable inftows,
due to the usual
data problems. But corrections to the data could probably not
affect the two key
qualitative messages of this chart. 30 First, ftows have grown
large in the last
ten years, for the first time in the postwar era, but they have
not yet surpassed
their importance in the pre-1914 era relative to receiving and
sending region
capital markets. Second. and as in the pre-1914 era, because the
capital market
in the core is so much bigger than the periphery, these ftows
weigh as a much
larger share of periphery investment than they do as a share of
core saving.
ar course, mere ftow data, as a quantity criterion, serve only
as weak evi-dence of changing market integration. However, these
basic descriptive tables
and f1gures do illustrate the record of capital ftows, and offer
prima facie ev-
idence that the globalization of the capital market has been
subject to major
dislocations. most notably in the inter-war period, with a
dramatic contraction
of tlows seen in the Depression of the 1930s. Moreover, this low
levei in the
volume of ftows persisted long into the postwar era, and
possibly even to to-
day. We now turn to more formal tests to see whether this
description, and the
conventional historography of world markets that points to the
Depression as
an era of dlsintegration, has broader support.
2.3.3 The Saving-Investment Relationslzip
Feldstein and Horioka (1980) proposed cross-country
saving-investment corre-
lations as a measure of international capital mobility. They
reasoned that, in a
10 The concluslon holds éVen if we were to shift the focus to
the key capital importers and exporters. we suspcct
-
60 Globalization in Capital Markets: The Long-Run Evidence
world of perfectly mobile capital, domestic savings would seek
out the highest
retums in the world capital market independent of local
investment demand,
and by the same token the world capital market would cater to
domestic invest-
ment needs independent of domestic savings supply. Thus, they
expected to
find low correlations of domestic saving and investment rates
among developed
countries given the conventional wisdom that intemational
capital markets were
well integrated by the 1960s and 1970s. In a surprising and
provocative result,
they discovered a high and significant investment rate-saving
rate correlation
with coefficients typically close to unity for a cross section
of OECD countries
with five-year period averaging. It appeared that changes in
domestic saving
passed through almost fully into domestic investment, suggesting
imperfect
intemational capital mobility.31
A substantialliterature has evolved following Feldstein and
Horioka (FH) to
assess whether incrementai savings were retained in the home
country or else
entered the global capital market seeking out the highest retum.
But as is well
known, the same literature has criticized the FH methodology on
both theoret-
ical and empirical grounds. This section develops and extends
the historical
application of saving-investment analysis, and seeks to extend
its theoretical
and empirical scope in several ways.32 Methodologically, we
apply not only
the traditional FH tests, but offer an altemative time-series
approach based on
a more explicit mode!. A major criticism of the FH method is
that one might
expect saving and investment te be ~~ghly corre!ated once
time-~""ragirg is
performed in CI oss-section, simply because ali countries must
abide by a long-
run version of current account balance in order to satisfy their
intertemporal
budget constraint. This idea leads to a very different modelling
approach, and
to a hypothesis that saving and investment may have trends or
unit roots, but
that the current account will be stationary - that is,
investment and saving will
be cointegrated - and an error-correction model (ECM) emerges as
a natural theoretical framework.
The standard technique employed for estimating long-run capital
mobility is
the generic FH regression of average invcstment rates on a
constant and average
31 Their finding of a large and significant coefticient has been
replicated many times for various cross-section and time-series
samples using post-World War Two and historical data. so much so as
to be now considered a robust result - a stylized fac!. as it were
(Dombusch 1991; Feldstein and Bacchetta 1991; Frankel 1991;
Obstfeld 1991; Tesar 1991; Sinn 1992).
32. Our statistical tests have enhanced power compared to other
historical studies. since we have increased the sample size: we use
annual data for the full period 1850 to the present, and we
increase the cross-section size from the usual nine or ten up to
fifteen, by adding various countires missing from earlier studies.
Some applications of the Feldstein-Horioka approach in economic
history, the natural antecedents of this chapter, include Bayoumi
(1989), Zevin (1992), and Eichengreen (1990)
-
2.3 Quantiry Criteria 61
saving rates. The regression is perfonned on a cross section of
countries indexed
by i. where we use a "short" averaging penod T (say, five or ten
years):
-- FH FH--(l/n,=a +b (S/ni+Ui, (2.9)
-- IT --- IT where (l/ni = T LI (f/nit, (s/ni = T LI (S/nit. The
cross-section slope coefficient bFH has the conventional
interpretation of a long-run savings-
retention coefficient for the sample of countries.33 The results
are shown in
Figure 2.7. The figure displays the coefficient bFH for both
5-year and ten-year
averaging patterns with two-standard-error intervals above and
below indicated
by vertical bars. Before FH-type coefficients can be interpreted
we require a
benchmark for what constitutes a "high" or a "Iow" bFH . Put
another way,
bFH comes equipped with no intrinsic absolute yardstick: we need
to find
a period we consider one of undisputed capital mobility and then
compare
other bFH observations to this benchmark.34 Alternatively, we
might search for
movements in bFH as indicators of whether saving-investment
interdependence
was relatively high or low for a given penod. There is certainly
variance in
the coefficient. The estimated coefficient is significant in
most periods, and
usually positive. It occasionally exceeds unity. The results
reveal considerable
tluctuation in the magnitude of bFH over the long run.
What do the results show? First. international capital market
integration, as
measured by bFH for this admittedly small sample, has shown no
marked trend
over the last hundred or so years: we can say that, in this
narrow sense, saving-
investment association has been no weaker in the last decade for
these countries
than it was circa 1900. Second, the coefficient bFH has not
evolved unifonnly
or monotonically over time: saving-investment association was
relatively low
during the 1880s, when financiaI markets were engaged in a
frenzy of foreign
investment. The crash of the early 1890s brought this phase to a
halt (higher
!JFH). Gradually. closer to \Vorld War One, saving-investment
association again
decreased (falling bFH ) in the last great foreign investment
boom during the
age 01' high imperialismo The Great Depression ushered in a time
of increased corre lations ior several decades (rising bFH). The
saving-investment association
began to decline in the late 1970s and 1980s. albeit to a
limited degree. 35 We
:n In 3 s~n\~. bFH is measuring the extent to which the sample
of countries deviates. on average over the sample penod. from the
closed-economy property whereby saving rates equal investment r3tes
by identitl'. Smce period averaging is employed the procedure
anempts to abstract from buslness cl'cle fluctuations that
simultaneously affect both saving and investment. The averaging
:Dsa Implies that this procedure has nothing to sal' abou!
l'ear-to-year (short-run) capital mobility: the ability of
countries to temporanly run current-account surpluses or deficits
in response to shocks a! hlgh frequencies (meaning approximately
annual frequencies. or higher) .
. '4 See Obstfeld (1986: 1994) ,5 For comparslOns on the size of
the recent postwar coefficient see Feldstein and Bacchetta
1991:
Obstfeld 1994: and A. M. Taylor 1994a.
-
62
I·
Globali-:.ation in Capital Markets: The Long-Run Evidence
2.0 ~----------------------
1.5
1.0 , .0 I'. ce . ... 0, ~.OC.O
• U-' , Cit' ~~ cJP cJk)~ 00 ° °
° 0.5 l
°
i 1 I
1 I
00 +-' -----------------------+
, 0 - , -) T
I
-1.0~' -----------------------'-
1860 1880 1900 1920 1940 1960 1980 2000
2.0~i ~~~1
1.5~
O ~~i~ 1.0 -05 +
, ~~~
-
2.3 Quantity Criteria 63
may note the remarkable relationship between the results of the
FH test and
the stylized facts concerning institutional change, monetary
experiments, and
policy regimes. Still, we might ask, what does the FH
coefficient mean and how do we
measure it accurately? The criterion. despite the attention it
has commanded
and ItS widespread use, is handicapped by two distinct sets of
problems: First,
do the regressions of investment on saving measure true and
unbiased "savings
retention"'l Second. even if estimated accurately. what does the
coefficient say
about international capital mobility?36
An alternative to the Feldstein-Horioka approach of estimating
the savings-
retention coefficient is to employ time-series analysis for an
individual country.
However. for technical reasons. there is no reason to expect
cross-section and
time-series coefficients to bear any relation to each other
since the time-series
and cross-section properties of the estimators differ
dramatically.37 The time-
series estimation ofthe coefficient also raises additional
concems, as it embodies
an assumption of stationarity in the series (1/ Y)it and (S/
Y)it. Such an as-sumption may indeed be valido especially for short
time frames. However,over
longer horizons, several studies have cast doubt on the
assumption that saving
and investment rates are truly stationary.38
Ali the same. the above ideas would be entirely familiar to
anyone applying
the standard small-open-economy macroeconomic model, or an
open-economy
growth model, to the problem at hand: for a poor (capital
scarce) country.
accumulation-Ied growth would begin at low income leveis using
borrowed
funds for investment; at the same time, permanent consumption
leveis would
be attained by borrowing against future output. Saving would
start low and
end high. to repay these debts; investment would start high and
end low, as
36 Th~s~ concems hinge of the samc typ~s of caveats already
mentioned: what auxiliary assump-tlons we thlnk are relevant for
the structural modeling of saving and investment in the long run.
and how we think the structure may have changes so as to possibly
account for the parameters secn. Sce Obstfeld (1986: 1994)
3 7 :-'lor~o\~r. the natural interpretation ofthe two
coeffjcients differs substantially - typically bCS is consld~red an
Indlcator of long-run Intemational capital mobility in a sample
group of countries.
w hereas b T 5 is seen as a measure of short-run. year-to-year
intemational capital mobility in a slnglé country (Obstfcld 1986:
1994).
38 Thls problem will come as no surprise to economic historians:
the notion of saving and in-vestment shlfts as important features
of growth and structural transformation in the long run is embodled
in many national histories. includlng the famous "grand traverse"
described by the Lnlted States economy in the nineteenth century
(Abramovitz and David 1973: David 1977). A similar upward shift In
the investment rate has been documented in nineteenth century
Britain (Crafts 1985. chapter 4. for examplel. The stationarity of
some settler-economy saving and Investment rate series could also
be called into question based upon an examlnation of long-run
trends in the Australian and Canadian data since 1870 (McLean 1994
illustrates the pattems). W A Lewis (1954.155) considered such
shifts in saving and investment the essential problem of economic
developmenL
-
64 Globalization in Capital Markets: The Long-Run Evidence
steady-state leveIs were reached. The long-run intertemporal
budget constraint
would be met, and the current account would tend toward a
stable, steady-state
leveI, implying stationarity. We used routine stationarity tests
to check on the stationarity of the series
(5/nit, (Ilnit, and (CA/nit in our long-run sample of annual
data for 15 countries. The current-account ratio is stationary in
ali cases, encouraging
the view that it may be subject to equilibrium tendencies as
theory suggests. However, the saving and investment series were not
stationary. 39 Thus, the new
strand of cointegration research into the properties of saving
and investrnent cor-relations may have a particular resonance in
long-run historical applications.40
An important implication is that there exists an
error-correction representation
of the relationship. The simplest such representation, commonly
adopted in
the Iiterature, is the first-order error-correction model (ECM),
which may be
written
= a ECM + bECM .6.(51 nt + c
ECM((5 I nt-I - (I I nt-I) +
dECM (51 nt-I + Ut· (2.10) This equation states that
instantaneous changes in investrnent may be driven by
a pass-thorugh from instantaneous changes in saving, as in a
c10sed economy, or through an error-correction term, which tends to
drive investment and saving
back towards some long-run equilibrium relationship in
levels.41
In the ECM equation, each of the terms has implications not only
for the
dynamics. but for our interpretation of the system as decribing
capital mobility. The coefficient bECM has the natural
interpretation: even given a long-run
equilibrium relationship between saving and investment, bECM
measures to
what extent a temporary annual shock to domestic saving .6. (5 I
nt passes through (ful1y? partially? not at ali?) into domestic
investment .6.(1 I nt.
39. Standard Dickey-Fuller test were used. On these econometric
maners see A. M. Taylor (1997). 40. The cOlntegration approach 10
the saving-investment relationship has been much discussed in
th~ Inerature. and the details need not be repeated at length
here (see Miller 1988; Leachman 1991; Vikoren 1991; Jansen and
Schulze 1996; A. M. Taylor 1997).
41 This approach follows Jansen and Schulze (1996). A. M. Taylor
(1997) applied (2.10) as a con-venient estimating equation for
implementing a well-specified test of the saving-investment
correlation. It is immediately obvious that (2.10) embodies a
long-run equilibrium rela-tionship between saving and investment
prescribed by theory. In the steady-state equilib-rium 1':,.( I /
Y)/ = /':;(5/ Y)/ = O. The implied equilibrium relationship in
leveis is given by (JECM + ~CM«5/Y) - (l/n) + d ECM (5/Y) = O.
Parameter restrictions may be used 10 test various natural
hypotheses conceming the nature of this long-run equilibrium.
If
dECM = O then the relevant error-correction terrn is just the
current account (C A/ n/; here. «(l/n/. (5/ nrl have a
cointegrating veClOr that is (1. -I); and the long-run equilibrium
cur-rent account is equal 10 a constant. C A' = (5/ Y)' - (l/ n' =
_aECM / ~CM; moreover. if a ECM = d ECM = O, then this constant is
zero.
-
2.3 QI/antit)" Criteria 65
.70 -
• Pass Through (b) O Adjustment Speed (c)
.60 .
I .50 " f 40 " ! í .30 " f .20 "
.10 "
00
1880--1913 191+-1945 1946-1971 1972-1992
Fig. 2.8. The Saving"Investment Error"Correction Model: Pooled
Coefficients with plus or minus 2 s.e. bars
It is thus a "short-run" adjustment parameter, with a meaning
not unlike that of a conventional FH coefficient. The coefficient
cECM also has a natural mterpretation: it measures the speed of
convergence of the system toward equilibrium Converted into a
"half-life" measure, this provides an indication of the
sustainability of saving-investment inequality for each country in
the sample period under study. It is thus a kind of "long-run"
measur