us federal reserve: levine

8/14/2019 US Federal Reserve: levine

http://slidepdf.com/reader/full/us-federal-reserve-levine 1/39

Journal of Economic LiteratureVol. XXXV ( June 1997 ) , pp. 688–726

Levine: Financial Development and Economic Growth

Financial Development and EconomicGrowth: Views and Agenda

ROSS LEVINEUniversity of Virginia

I thank, without implicating, Gerard Caprio, Maria Carkovic, David Cole, Robert Cull, Wil-liam Easterly, Mark Gertler, Fabio Schiantarelli, Mary Shirley, Bruce Smith, and KennethSokoloff for criticisms, guidance, and encouragement. This paper was written while I was at

the World Bank. Opinions expressed are those of the author and do not necessarily reflect the

views of the World Bank, its staff, or member countries.

Does finance make a difference . . .? Raymond Goldsmith (1969, p. 408)

I. Introduction: Goals and Boundaries

ECONOMISTS HOLD startlingly dif-ferent opinions regarding the im-

portance of the financial system for eco-nomic growth. Walter Bagehot (1873)

and John Hicks (1969) argue that itplayed a critical role in igniting industri-alization in England by facilitating themobilization of capital for “immense works.” Joseph Schumpeter (1912) con-tends that well-functioning banks spurtechnological innovation by identifyingand funding those entrepreneurs withthe best chances of successfully imple-menting innovative products and pro-duction processes. In contrast, Joan Rob-inson (1952, p. 86) declares that “whereenterprise leads finance follows.” Ac-cording to this view, economic develop-ment creates demands for particulartypes of financial arrangements, and thefinancial system responds automatically to these demands. Moreover, someeconomists just do not believe that thefinance-growth relationship is important.Robert Lucas (1988, p. 6) asserts that

economists “badly over-stress” the role

of financial factors in economic growth, while development economists fre-quently express their skepticism aboutthe role of the financial system by ignor-ing it (Anand Chandavarkar 1992). Forexample, a collection of essays by the

“pioneers of development economics,”including three Nobel Laureates, doesnot mention finance (Gerald Meir andDudley Seers 1984). Furthermore,Nicholas Stern’s (1989) review of devel-opment economics does not discuss thefinancial system, even in a section thatlists omitted topics. In light of these con-flicting views, this paper uses existingtheory to organize an analytical framework of the finance-growth nexus andthen assesses the quantitative impor-tance of the financial system in economicgrowth.

Although conclusions must be statedhesitantly and with ample qualifications,the preponderance of theoretical reason-ing and empirical evidence suggests apositive, first-order relationship betweenfinancial development and economicgrowth. A growing body of work would

push even most skeptics toward the be-688



lief that the development of financialmarkets and institutions is a critical andinextricable part of the growth processand away from the view that the financialsystem is an inconsequential side show,

responding passively to economic growthand industrialization. There is even evi-dence that the level of financial devel-opment is a good predictor of futurerates of economic growth, capital accu-mulation, and technological change.Moreover, cross country, case study, in-dustry- and firm-level analyses documentextensive periods when financial devel-opment—or the lack thereof—crucially affects the speed and pattern of eco-nomic development.

To arrive at these conclusions and tohighlight areas in acute need of addi-tional research, I organize the remainderof this paper as follows. Section II ex-plains what the financial system does andhow it affects—and is affected by—eco-nomic growth. Theory suggests that fi-nancial instruments, markets, and insti-tutions arise to mitigate the effects of

information and transaction costs.1 Fur-thermore, a growing literature showsthat differences in how well financialsystems reduce information and transac-tion costs influence saving rates, invest-ment decisions, technological innova-tion, and long-run growth rates. Also, acomparatively less developed theoreticalliterature demonstrates how changes ineconomic activity can influence financialsystems.

Section II also advocates the func-tional approach to understanding therole of financial systems in economicgrowth. This approach focuses on theties between growth and the quality of the functions provided by the financialsystem. These functions include facilitat-

ing the trading of risk, allocating capital,monitoring managers, mobilizing sav-ings, and easing the trading of goods,services, and financial contracts.2 Thebasic functions remain constant through

time and across countries. There arelarge differences across countries andtime, however, in the quality of financialservices and in the types of financial in-struments, markets, and institutions thatarise to provide these services. While fo-cusing on functions, this approach doesnot diminish the role of institutions. In-deed, the functional approach highlightsthe importance of examining an under-researched topic: the relationship be-tween financial structure—the mix of financial instruments, markets, and insti-tutions—and the provision of financialservices. Thus, this approach discouragesa narrow focus on one financial instru-ment, like money, or a particular institu-tion, like banks. Instead, the functionalapproach prompts a more comprehen-sive—and more difficult—question: whatis the relationship between financial

structure and the functioning of the fi-nancial system?3 Part III then turns to the evidence.

While many gaps remain, broad cross-country comparisons, individual country studies, industry-level analyses, andfirm-level investigations point in the

1 These frictions include the costs of acquiringinformation, enforcing contracts, and exchanging

goods and financial claims.

2 For different ways of categorizing financialfunctions, see Cole and Betty Slade (1991) andRobert C. Merton and Zvi Bodie (1995).

3 The major alternative approach to studying fi-nance and economic growth is based on the semi-nal contributions of John Gurley and EdwardShaw (1955), James Tobin (1965), and RonaldMcKinnon (1973). In their mathematical models,as distinct from their narratives, they focus onmoney. This narrow focus can restrict the analysisof the finance-growth nexus, and lead to a mis-leading distinction between the “real” and finan-cial sectors. In contrast, the functional approachhighlights the value added of the financial sector.The financial system is a “real” sector: it re-searches firms and managers, exerts corporatecontrol, and facilitates risk management, ex-

change, and resource mobilization.

Levine: Financial Development and Economic Growth 689



same direction: the functioning of finan-cial systems is vitally linked to economicgrowth. Specifically, countries withlarger banks and more active stock mar-kets grow faster over subsequent de-

cades even after controlling for many other factors underlying economicgrowth. Industries and firms that rely heavily on external financing grow dis-proportionately faster in countries with well-developed banks and securitiesmarkets than in countries with poorly developed financial systems. Moreover,ample country studies suggest that dif-ferences in financial development have,in some countries over extensive periods,critically influenced economic develop-ment. Yet, these results do not imply that finance is everywhere and always ex-ogenous to economic growth. Economicactivity and technological innovation un-doubtedly affect the structure and qual-ity of financial systems. Innovations intelecommunications and computing haveundeniably affected the financial services industry. Moreover, “third factors,”

such as a country’s legal system and po-litical institutions certainly drive both fi-nancial and economic development atcritical junctures during the growth pro-cess. Nevertheless, the weight of evi-dence suggests that financial systems area fundamental feature of the process of economic development and that a satis-factory understanding of the factors un-derlying economic growth requires agreater understanding of the evolutionand structure of financial systems.

As in any critique, I omit or treat cur-sorily important issues. Here I highlighttwo.4 First, I do not discuss the relation-ship between international finance andgrowth. This paper narrows its concep-tual focus by studying the financial services available to an economy regardless

of the geographic source of those services. In measuring financial develop-ment, however, researchers often do notaccount sufficiently for internationaltrade in financial services. Second, the

paper does not discuss policy. Given thelinks between the functioning of the fi-nancial system and economic growth, de-signing optimal financial sector policiesis critically important. A rigorous discus-sion of these policies, however, wouldrequire a long article or book by itself.5Instead, this paper seeks to pull togethera diverse and active literature into a co-herent view of the financial system ineconomic growth.

II. The Functions of the FinancialSystem

A. Functional Approach: Introduction

The costs of acquiring information andmaking transactions create incentivesfor the emergence of financial marketsand institutions. Put differently, in aKenneth Arrow (1964)-Gerard Debreu

(1959) state-contingent claim framework with no information or transaction costs,there is no need for a financial systemthat expends resources researching proj-ects, scrutinizing managers, or designingarrangements to ease risk managementand facilitate transactions. Thus, any the-ory of the role of the financial system ineconomic growth (implicitly or explic-itly) adds specific frictions to the Arrow-Debreu model. Financial markets andinstitutions may arise to ameliorate theproblems created by information andtransactions frictions. Different typesand combinations of information andtransaction costs motivate distinct finan-cial contracts, markets, and institutions.

4 Also, the theoretical review focuses on purely real economies and essentially ignores work on fi-

nance and growth in monetary economies.

5 The financial policy literature is immense. See,for example, Philip Brock (1992), Alberto Giovan-nini and Martha De Melo (1993), Caprio, Isak Ati- yas, and James Hanson (1994), and Maxwell Fry

(1995).

690 Journal of Economic Literature, Vol. XXXV ( June 1997 )



In arising to ameliorate transactionand information costs, financial systemsserve one primary function: they facili-tate the allocation of resources, acrossspace and time, in an uncertain environ-

ment (Merton and Bodie 1995, p. 12).To organize the vast literature on fi-nance and economic activity, I break thisprimary function into five basic func-tions.

Specifically, financial systems

- facilitate the trading, hedging, diver-sifying, and pooling of risk,

- allocate resources,- monitor managers and exert corpo-

rate control,- mobililize savings, and- facilitate the exchange of goods and

services.

This section explains how particularmarket frictions motivate the emergenceof financial markets and intermediariesthat provide these five functions, and ex-plains how they affect economic growth.I examine two channels through which

each financial function may affect eco-nomic growth: capital accumulation andtechnological innovation. On capital ac-cumulation, one class of growth modelsuses either capital externalities or capitalgoods produced using constant returnsto scale but without the use of nonrepro-ducible factors to generate steady-stateper capita growth (Paul Romer 1986;Lucas 1988; Sergio Rebelo 1991). In thesemodels, the functions performed by thefinancial system affect steady-state growthby influencing the rate of capital forma-tion. The financial system affects capitalaccumulation either by altering the sav-ings rate or by reallocating savings amongdifferent capital producing technologies.On technological innovation, a secondclass of growth models focuses on the invention of new production processes andgoods (Romer 1990; Gene Grossman and

Elhanan Helpman 1991; and Philippe

Aghion and Peter Howitt 1992). In thesemodels, the functions performed by thefinancial system affect steady-stategrowth by altering the rate of technologi-cal innovation. Thus, as sketched in Fig-ure 1, the remainder of this section dis-cusses how specific market frictionsmotivate the emergence of financial con-tracts, markets, and intermediaries andhow these financial arrangements provide five financial functions that affectsaving and allocations decisions in waysthat influence economic growth.

B. Facilitating Risk Amelioration

In the presence of specific informationand transaction costs, financial marketsand institutions may arise to ease thetrading, hedging, and pooling of risk.This subsection considers two types of

risk: liquidity and idiosyncratic risk.

Figure 1. A Theoretical Approach to Financeand Growth

Market frictions- information costs- transaction costs

Financial marketsand intermediaries

Financial functions- mobilize savings- allocate resources- exert corporate control- facilitate risk management- ease trading of goods,

services, contracts

Channels to growth- capital accumulation- technological innovation

Growth




Liquidity is the ease and speed with which agents can convert assets into pur-chasing power at agreed prices. Thus,real estate is typically less liquid thanequities, and equities in the United

States are typically more liquid thanthose traded on the Nigerian Stock Ex-change. Liquidity risk arises due to theuncertainties associated with convertingassets into a medium of exchange. Infor-mational asymmetries and transactioncosts may inhibit liquidity and intensify liquidity risk. These frictions create in-centives for the emergence of financialmarkets and institutions that augment li-quidity. Liquid capital markets, there-fore, are markets where it is relatively inexpensive to trade financial instru-ments and where there is little uncer-tainty about the timing and settlement of those trades.

Before delving into formal models of liquidity and economic activity, some in-tuition and history may help motivatethe discussion. The link between liquid-ity and economic development arises be-

cause some high-return projects requirea long-run commitment of capital, butsavers do not like to relinquish control of their savings for long periods. Thus, if the financial system does not augmentthe liquidity of long-term investments,less investment is likely to occur in thehigh-return projects. Indeed, Sir JohnHicks (1969, pp. 143–45) argues that thecapital market improvements that miti-gated liquidity risk were primary causesof the industrial revolution in England.According to Hicks, the products manu-factured during the first decades of theindustrial revolution had been inventedmuch earlier. Thus, technological innovation did not spark sustained growth.Many of these existing inventions, how-ever, required large injections and long-run commitments of capital. The critical

new ingredient that ignited growth in

eighteenth century England was capital

market liquidity. With liquid capital mar-kets, savers can hold assets—like equity,bonds, or demand deposits—that they can sell quickly and easily if they seekaccess to their savings. Simultaneously,

capital markets transform these liquid fi-nancial instruments into long-term capi-tal investments in illiquid productionprocesses. Because the industrial revolu-tion required large commitments of capi-tal for long periods, the industrial revo-lution may not have occurred withoutthis liquidity transformation. “The indus-trial revolution therefore had to wait forthe financial revolution” (Valerie Ben-civenga, Bruce Smith, and Ross Starr1966, p. 243).6

Economists have recently modeled theemergence of financial markets in re-sponse to liquidity risk and examinedhow these financial markets affect eco-nomic growth. For example, in DouglasDiamond and Philip Dybvig’s (1983)seminal model of liquidity, a fraction of savers receive shocks after choosing be-tween two investments: an illiquid, high-

return project and a liquid, low-returnproject. Those receiving shocks want ac-cess to their savings before the illiquidproject produces. This risk creates in-centives for investing in the liquid, low-return projects. The model assumes thatit is prohibitively costly to verify whetheranother individual has received a shockor not. This information cost assumptionrules out state-contingent insurance con-tracts and creates an incentive for finan-cial markets— markets where individuals

issue and trade securities—to emerge. InLevine (1991), savers receiving shocks

6 The financial revolution included the emer-gence of joint-stock companies with nonredeem-able capital. The Dutch East India Company made capital permanent in 1609, and Cromwellmade the English East India Company capital per-manent in 1650. These financial innovationsformed the basis of liquid equity markets (Larry

Neal 1990).




can sell their equity claims on the profitsof the illiquid production technology toothers. Market participants do not verify whether other agents received shocks ornot; participants simply trade in imper-

sonal stock exchanges. Thus, with liquidstock markets, equity holders can readily sell their shares, while firms have perma-nent access to the capital invested by theinitial shareholders. By facilitating trade,stock markets reduce liquidity risk.7 Asstock market transaction costs fall, moreinvestment occurs in the illiquid, high-return project. If illiquid projects enjoy sufficiently large externalities, thengreater stock market liquidity inducesfaster steady-state growth.

Thus far, information costs—the costsof verifying whether savers have re-ceived a shock—have motivated the ex-istence of stock markets. Trading costscan also highlight the role of liquidity.For example, different production tech-nologies may have a wide array of gesta-tion periods for converting current out-put into future capital, where longer-run

technologies enjoy greater returns. In- vestors, however, may be reluctant to re-linquish control of their savings for very long periods. Thus, long-gestation pro-duction technologies require that owner-ship be transferred throughout the lifeof the production process in secondary securities markets (Bencivenga, B. Smith,and Starr 1995). If exchanging owner-ship claims is costly, then longer-runproduction technologies will be less at-tractive. Thus, liquidity—as measured by secondary market trading costs—affectsproduction decisions. Greater liquidity will induce a shift to longer-gestation,higher- return technologies.

Besides stock markets, financial inter-

mediaries—coalitions of agents that com- bine to provide financial services—may also enhance liquidity and reduce liquid-ity risk. As discussed above, Diamondand Dybvig’s (1983) model assumes it is

prohibitively costly to observe shocks toindividuals, so it is impossible to writeincentive compatible state-contingent in-surance contracts. Under these condi-tions, banks can offer liquid deposits tosavers and undertake a mixture of liquid,low-return investments to satisfy de-mands on deposits and illiquid, high-re-turn investments. By providing demanddeposits and choosing an appropriate mix-ture of liquid and illiquid investments,banks provide complete insurance to sav-ers against liquidity risk while simulta-neously facilitating long-run investmentsin high-return projects. Banks replicatethe equilibrium allocation of capital thatexists with observable shocks. By elimi-nating liquidity risk, banks can increaseinvestment in the high-return, illiquidasset and accelerate growth (Bencivengaand B. Smith 1991). There is a problem,

however, with this description of the roleof banks as reducing liquidity risk. Thebanking equilibrium is not incentivecompatible if agents can trade in liquidequity markets; if equity markets exist,all agents will use equities; none will usebanks (Charles Jacklin 1987). Thus, inthis context, banks will only emerge toprovide liquidity if there are sufficiently large impediments to trading in securi-ties markets (Gary Gorton and GeorgePennacchi 1990).8

7 Frictionless stock markets, however, do noteliminate liquidity risk. That is, stock markets donot replicate the equilibrium that exists when in-surance contracts can be written contingent on ob-

serving whether an agent receives a shock or not.

8 Goldsmith (1969, p. 396) notes that “Claimsagainst financial institutions are generally easier toliquidate (i.e., to turn into cash without or withonly insignificant delay, formality, and cost) thanare primary debt securities. They have the addi-tional great advantage of being completely divis-ible, whereas primary securities are usually issuedin fixed amounts and often in amounts that makethem very inconvenient for purchase and sale when lenders have small resources and when nu-merous individual purchase and sale transactions

are involved.”




Theory, however, suggests that en-hanced liquidity has an ambiguous affecton saving rates and economic growth.9 Inmost models, greater liquidity (a)increases investment returns and (b)

lowers uncertainty. Higher returns am-biguously affect saving rates due to well-known income and substitution effects.Further, lower uncertainty ambiguously affects savings rates (David Levhari andT. N. Srinivasan 1969). Thus, savingrates may rise or fall as liquidity rises.Indeed, in a model with physical capitalexternalities, saving rates could fallenough, so that growth actually deceler-ates with greater liquidity (Tullio Jap-pelli and Marco Pagano 1994).10

Besides reducing liquidity risk, finan-cial systems may also mitigate the risksassociated with individual projects,firms, industries, regions, countries, etc.Banks, mutual funds, and securities mar-kets all provide vehicles for trading,pooling, and diversifying risk.11 The fi-nancial system’s ability to provide risk diversification services can affect long-run

economic growth by altering resource al-location and the saving rates. The basicintuition is straightforward. While saversgenerally do not like risk, high-returnprojects tend to be riskier than low-re-

turn projects. Thus, financial marketsthat ease risk diversification tend to in-duce a portfolio shift toward projects with higher expected returns (GillesSaint-Paul 1992; Michael Devereux and

Gregor Smith 1994; and MauriceObstfeld 1994). Greater risk sharing andmore efficient capital allocation, how-ever, have theoretically ambiguous ef-fects on saving rates as noted above. Thesavings rate could fall enough so that, when coupled with an externality-basedor linear growth model, overall economicgrowth falls. With externalities, growthcould fall sufficiently so that overall wel-fare falls with greater risk diversifica-tion.

Besides the link between risk diversifi-cation and capital accumulation, risk diversification can also affect technologicalchange. Agents are continuously tryingto make technological advances to gain aprofitable market niche. Besides yieldingprofits to the innovator, successful innovation accelerates technological change.Engaging in innovation is risky, however.

The ability to hold a diversified portfolioof innovative projects reduces risk andpromotes investment in growth-enhanc-ing innovative activities (with sufficiently risk averse agents). Thus, financial sys-tems that ease risk diversification can ac-celerate technological change and eco-nomic growth (Robert King and Levine1993c).

C. Acquiring Information About

Investments and AllocatingResources

It is difficult and costly to evaluatefirms, managers, and market conditionsas discussed by Vincent Carosso (1970).Individual savers may not have the time,capacity, or means to collect and processinformation on a wide array of enter-prises, managers, and economic condi-tions. Savers will be reluctant to invest in

activities about which there is little reli-

9 The analyses described thus far focus on thelinks between liquidity and capital accumulation.Yet, liquidity may also affect the rate of techno-logical change if long-run commitments of re-sources to research and development promotetechnological innovation.

10 Similarly, although greater liquidity unambi-guously raises the real return on savings, more li-quidity may induce a reallocation of investmentout of initiating new capital investments and intopurchasing claims on ongoing projects. This may lower the rate of real investment enough to decel-erate growth (Bencivenga, B. Smith, and Starr1995).

11 Although the recent uses of options and fu-tures contracts to hedge risk have been well publi-cized, the development of these financial con-tracts is by no means recent. Josef Penso de la Vega published a treatise on options contracts, fu-tures contracts, and securities market speculation,

Confusion de Confusiones, in 1688!




able information. Consequently, high in-formation costs may keep capital fromflowing to its highest value use.

Information acquisition costs createincentives for financial intermediaries to

emerge (Diamond 1984; and John Boydand Edward Prescott 1986). Assume, forexample, that there is a fixed cost toacquiring information about a product-ion technology. Without intermediaries,each investor must pay the fixed cost. Inresponse to this information cost struc-ture, however, groups of individuals may form (or join or use) financial intermedi-aries to economize on the costs of ac-quiring and processing informationabout investments. Instead of each individual acquiring evaluation skills andthen conducting evaluations, an interme-diary can do it for all its members.Economizing on information acquisitioncosts facilitates the acquisition of infor-mation about investment opportunitiesand thereby improves resource alloca-tion.

The ability to acquire and process in-

formation may have important growthimplications. Because many firms andentrepreneurs will solicit capital, finan-cial intermediaries, and markets that arebetter at selecting the most promis-ing firms and managers will induce amore efficient allocation of capital andfaster growth (Jeremy Greenwood andBoyan Jovanovic 1990). Bagehot (1873,p. 53) expressed this view over 120 yearsago.

[England’s financial] organization is so usefulbecause it is so easily adjusted. Politicaleconomists say that capital sets towards themost profitable trades, and that it rapidly leaves the less profitable non-paying trades.But in ordinary countries this is a slow pro-cess, . . . In England, however, . . . capitalruns as surely and instantly where it is most wanted, and where there is most to be madeof it, as water runs to find its level.

England’s financial system did a better

job at identifying and funding profitable

ventures than most countries in the mid-1800s, which helped it enjoy compara-tively greater economic success.12

Besides identifying the best produc-tion technologies, financial intermediar-

ies may also boost the rate of technologi-cal innovation by identifying thoseentrepreneurs with the best chances of successfully initiating new goods andproduction processes (King and Levine1993c). As eloquently stated by Schum-peter (1912, p. 74),

The banker, therefore, is not so much pri-marily a middleman, . . . He authorises peo-ple, in the name of society as it were, . . . [toinnovate].

Stock markets may also influence theacquisition and dissemination of infor-mation about firms. As stock markets be-come larger (Sanford Grossman andJoseph Stiglitz 1980) and more liquid(Albert Kyle 1984; and Bengt Holm-strom and Jean Tirole 1993), market par-ticipants may have greater incentives toacquire information about firms. Intui-tively, with larger more liquid markets, it

is easier for an agent who has acquiredinformation to disguise this privateinformation and make money. Thus,large, liquid stock markets can stimulatethe acquisition of information. More-over, this improved information aboutfirms should improve resource alloca-tion substantially with correspondingimplications for economic growth (Mer-ton 1987). However, existing theorieshave not yet assembled the links of thechain from the functioning of stock mar-kets, to information acquisition, and fi-nally to aggregate long-run economicgrowth.

12 Indeed, England’s advanced financial systemalso did a good job at identifying profitable ven-tures in other countries, such as Canada, theUnited States, and Australia during the 19th cen-tury. England was able to “export” financial services (as well as financial capital) to many econo-mies with underdeveloped financial systems

(Lance Davis and Robert Huttenback 1986).




Debate still exists over the importanceof large, liquid, efficient stock markets inenhancing the creation and distributioninformation about firms. Stock marketsaggregate and disseminate information

through published prices. Even agentsthat do not undertake the costly pro-cesses of evaluating firms, managers, andmarket conditions can observe stockprices that reflect the information ob-tained by others. This public goods as-pect of acquiring information can causesociety to devote too few resources to in-formation acquisition. The public goodsfeature of the information thus disclosedmay be sufficiently large, that informa-tion gains from large, liquid stock mar-kets are small. Stiglitz (1985) argues that,because stock markets quickly reveal in-formation through posted prices, there will be few incentives for spending privateresources to acquire information that isalmost immediately publicly available.

D. Monitoring Managers and ExertingCorporate Control

Besides reducing the costs of acquir-ing information ex ante, financial con-tracts, markets, and intermediaries may arise to mitigate the information acquisi-tion and enforcement costs of monitor-ing firm managers and exerting corpo-rate control ex post, i.e., after financingthe activity. For example, firm owners will create financial arrangements thatcompel firm managers to manage thefirm in the best interests of the owners.Also, “outside” creditors—banks, equity,and bond holders—that do not managefirms on a day-to-day basis will create fi-nancial arrangements to compel insideowners and managers to run firms in ac-cordance with the interests of outsidecreditors. The absence of financial ar-rangements that enhance corporate con-trol may impede the mobilization of sav-ings from disparate agents and thereby

keep capital from flowing to profitable

investments (Stiglitz and Andrew Weiss1981, 1983). Because this vast literaturehas been carefully reviewed (Gertler1988; and Andrei Shleifer and Robert Vishny, forthcoming), this subsection (1)

notes a few ways in which financial con-tracts, markets, and institutions improvemonitoring and corporate control, and(2) reviews how these financial arrange-ments for monitoring influence capitalaccumulation, resource allocation, andlong-run growth.

Consider, for example, the simple as-sumption that it is costly for outsiderinvestors in a project to verify projectreturns. This creates important frictionsthat can motivate financial development.Insiders have incentives to misrepresentproject returns to outsiders. Given verifi-cation costs, however, it is socially ineffi-cient for outsiders to monitor in allcircumstances. With “costly state verifi-cation” (and other assumptions includingrisk-neutral borrowers and verificationcosts that are independent of projectquality), the optimal contract between

outsiders and insiders is a debt contract(Robert Townsend 1979; and DouglasGale and Martin Hellwig 1985). Specifi-cally, there is an equilibrium interestrate, r, such that when the project returnis sufficiently high, insiders pay r to out-siders and outsiders do not monitor. When project returns are insufficient,the borrower defaults and the lenderspay the monitoring costs to verify theproject’s return. These verification costsimpede investment decisions and reduceeconomic efficiency. Verification costsimply that outsiders constrain firms fromborrowing to expand investment becausehigher leverage implies greater risk of default and higher verification expendi-tures by lenders. Thus, collateral and fi-nancial contracts that lower monitoringand enforcement costs reduce impedi-ments to efficient investment (Stephen

Williamson 1987b; Ben Bernanke and




Gertler 1989, 1990; Ernst-Ludwig vonThadden 1995).13

Besides particular types of financialcontracts, financial intermediaries canreduce information costs even further. If

borrowers must obtain funds from many outsiders, financial intermediaries caneconomize on monitoring costs. The fi-nancial intermediary mobilizes the sav-ings of many individuals and lends theseresources to project owners. This “dele-gated monitor” arrangement economizeson aggregate monitoring costs because aborrower is monitored only by the inter-mediary, not all individual savers (Dia-mond 1984). Besides reducing duplicatemonitoring, a financial system that facili-tates corporate control “also makes pos-sible the efficient separation of owner-ship from management of the firm. Thisin turn makes feasible efficient speciali-zation in production according to theprinciple of comparative advantage”(Merton and Bodie 1995, p. 14). Thedelegated monitor arrangement, how-ever, creates a potential problem: who

will monitor the monitor (Stefan Krasaand Anne Villamil 1992)? Savers, how-ever, do not have to monitor the inter-mediary if the intermediary holds a diversified portfolio (and agents can easily verify that the intermediary’s portfolio is well diversified). With a well-diversifiedportfolio, the intermediary can alwaysmeet its promise to pay the deposit in-terest rate to depositors, so that deposi-tors never have to monitor the bank.Thus, well-diversified financial inter-mediaries can foster efficient investmentby lowering monitoring costs.14 Further-

more, as financial intermediaries andfirms develop long-run relationships, thiscan further lower information acquisi-tion costs. The reduction in informationasymmetries can in turn ease external

funding constraints and facilitate betterresource allocation (Sharpe 1990).15 Interms of long-run growth, financial ar-rangements that improve corporate con-trol tend to promote faster capital accu-mulation and growth by improving theallocation of capital (Bencivenga and B.Smith 1993).

Besides debt contracts and banks,stock markets may also promote corpo-rate control (Michael Jensen and Wil-liam Meckling 1976). For example, pub-lic trading of shares in stock markets thatefficiently reflect information aboutfirms allows owners to link managerialcompensation to stock prices. Linkingstock performance to manager compen-sation helps align the interests of manag-ers with those of owners (Diamond andRobert Verrecchia 1982; and Jensen andKevin Murphy 1990). Similarly, if take-

overs are easier in well-developed stockmarkets and if managers of under-per-forming firms are fired following a take-over, then better stock markets can pro-mote better corporate control by easingtakeovers of poorly managed firms. Thethreat of a takeover will help align mana-gerial incentives with those of the own-

13 Costly state verification can produce creditrationing. Because higher interest rates are linked with a higher probability of default and monitor-ing costs, intermediaries may keep rates low andration credit using non-price mechanisms (Wil-liamson 1986, 1987a).

14 Diamond (1984) assumes that intermediariesexist and shows that the intermediary arrangement

economizes on monitoring costs. Williamson

(1986) shows how intermediaries arise endogen-

ously. Furthermore, I have only discussed modelsin which state verification proceeds nonstochasti-cally: if borrowers default, lenders verify. Stochas-tic monitoring, however, may further reduce veri-fication costs (Bernanke and Gertler 1989; andBoyd and B. Smith 1994).

15 The long-run relationships between a bankerand client may impose a cost on the client. Be-cause the bank is well informed about the firm,the bank may have bargaining power over thefirm’s profits. If the bank breaks its ties to thefirm, other investors will be reluctant to invest inthe firm. Firms may therefore diversify out of bank financing to reduce their vulnerability

(Raghurman Rajan 1992).




ers (David Scharfstein 1988; and Jeremy Stein 1988). I am not aware of modelsthat directly link the role of stock mar-kets in improving corporate governance with long-run economic growth.

There are disagreements, however,about the importance of stock markets incorporate control. Inside investors prob-ably have better information about thecorporation than outsiders. Thus, if well-informed owners are willing to sell theircompany, less well informed outsidersmay demand a premium to purchase thefirm due to the information asymmetry (Stewart Myers and Nicholas Majluf 1984). Thus, asymmetric informationmay reduce the efficacy of corporatetakeovers as a mechanism for exertingcorporate control. Stiglitz (1985) makesthree additional arguments about take-overs. First, if an acquiring firm expendslots of resources obtaining information,the results of this research will be ob-served by other market participants when the acquiring firm bids for shares.This will induce others to bid for shares,

so that the price rises. The firm that ex-pended resources obtaining informationmust, therefore, pay a higher price thanit would have to pay if “free-riding”firms could not observe its bid. Thus, therapid public dissemination of costly in-formation will reduce incentives for ob-taining information and making effectivetakeover bids. Second, there is a publicgood nature to takeovers that may de-crease the incentives for takeovers. If the takeover succeeds, and the shareprice rises, then those original equity holders who did not sell make a bigprofit without expending resources. Thiscreates an incentive for existing share-holders to not sell if they think the valueof the firm will rise following the take-over. Thus, value-increasing takeoversmay fail because the acquiring firm willhave to pay a high price, which will re-

duce incentives for researching firms in

the hopes of taking them over. Third,current managers often can take strate-gic actions to deter takeovers and main-tain their positions. This argues against animportant role for liquid stock markets in

promoting sound corporate governance.Moreover, liquid equity markets thatfacilitate takeovers may hurt resource al-location (Shleifer and Lawrence Sum-mers 1988; and Randall Morck, Shleifer,and Vishny 1990). A takeover typically involves a change in management. Exist-ing implicit contracts between formermanagers and workers, suppliers, andother stakeholders in the firms do notbind new owners and managers to thesame extent that they bound the originalmanagers. Thus, a takeover allows newowners and managers to break implicitagreements and transfer wealth fromfirm stakeholders to themselves. Whilenew owners may profit, there may be adeterioration in the efficiency of re-source allocation. Overall welfare may fall. To the extent that well-functioningequity markets help takeovers, this may

allow hostile takeovers that lead to a fallin the efficiency of resource allocation.Furthermore, liquid stock markets may reduce incentives for owners to monitormanagers (Amar Bhide 1993). By reduc-ing exit costs, stock market liquidity en-courages more diffuse ownership withfewer incentives and greater impedi-ments to actively overseeing managers(Shleifer and Vishny 1986). Thus, thetheoretical signs on the links in the chainfrom improvements in stock markets tobetter corporate control to faster eco-nomic growth are still ambiguous.16

E. Mobilizing Savings

Mobilization— pooling—involves theagglomeration of capital from disparate

16 Some research also suggests that excessivestock trading can induce “noise” into the marketand hinder efficient resource allocation (Bradford

De Long et al. 1989).




savers for investment. Without access tomultiple investors, many production pro-cesses would be constrained to economi-cally inefficient scales (Erik Sirri andPeter Tufano 1995). Furthermore, mobi-

lization involves the creation of smalldenomination instruments. These instru-ments provide opportunities for house-holds to hold diversified portfolios, invest in efficient scale firms, and toincrease asset liquidity. Without pooling,household’s would have to buy and sellentire firms. By enhancing risk diversifi-cation, liquidity, and the size of feasiblefirms, therefore, mobilization improves re-source allocation (Sirri and Tufano 1995).

Mobilizing the savings of many dispa-rate savers is costly, however. It involves(a) overcoming the transaction costs as-sociated with collecting savings from dif-ferent individuals and (b) overcomingthe informational asymmetries associated with making savers feel comfortable inrelinquishing control of their savings. In-deed, much of Carosso’s (1970) history of Investment Banking in America is a

description of the diverse and elaboratemeans employed by investment banks toraise capital. As early as the mid-1880s,some investment banks used their Euro-pean connections to raise capital abroadfor investment in the United States.Other investment banks established closeconnections with major banks and indus-trialists in the United States to mobilizecapital. And, still others used newspaperadvertisements, pamphlets, and a vastsales force that traveled through every state and territory selling securities toindividual households. Thus, mobilizingresources involved a range of transactioncosts. Moreover, “mobilizers” had toconvince savers of the soundness of theinvestments. Toward this end, interme-diaries are generally concerned about es-tablishing stellar reputations or govern-ment backing, so that savers feel

comfortable about entrusting their sav-

ings to the intermediary (De Long 1991;and Naomi Lamoreaux 1994).

In light of the transaction and infor-mation costs associated with mobilizingsavings from many agents, numerous fi-

nancial arrangements may arise to miti-gate these frictions and facilitate pool-ing.17 Specifically, mobilization may involve multiple bilateral contracts be-tween productive units raising capitaland agents with surplus resources. The joint stock company in which many individuals invest in a new legal entity, thefirm, represents a prime example of mul-tiple bilateral mobilization. To econo-mize on the transaction and informationcosts associated with multiple bilateralcontracts, pooling may also occurthrough intermediaries as discussedabove, where thousands of investors en-trust their wealth to intermediaries thatinvest in hundreds of firms (Sirri andTufano 1995, p. 83).

Financial systems that are more effec-tive at pooling the savings of individualscan profoundly affect economic develop-

ment. Besides the direct effect of bettersavings mobilization on capital accumu-lation, better savings mobilization canimprove resource allocation and boosttechnological innovation (Bagehot 1873,pp. 3–4):

We have entirely lost the idea that any under-taking likely to pay, and seen to be likely, canperish for want of money; yet no idea wasmore familiar to our ancestors, or is morecommon in most countries. A citizen of Long

in Queen Elizabeth’s time . . . would havethought that it was no use inventing railways(if he could have understood what a railway meant), for you would have not been able tocollect the capital with which to make them.At this moment, in colonies and all rudecountries, there is no large sum of transfer-able money; there is not fund from which youcan borrow, and out of which you can makeimmense works.

17 See Sections II.C and II.D for citations on

the emergence of financial intermediaries.




Thus, by effectively mobilizing resourcesfor projects, the financial system may play a crucial role in permitting theadoption of better technologies andthereby encouraging growth. This intu-

ition was clarified 100 years later by McKinnon (1973, p. 13):

The farmer could provide his own savings toincrease slightly the commercial fertilizerthat he is now using, and the return on thismarginal new investment could be calculated.

The important point, however, is the virtualimpossibility of a poor farmer’s financingfrom his current savings the whole of the bal-anced investment needed to adopt the newtechnology. Access to external financial re-sources is likely to be necessary over the one

or two years when the change takes place. Without this access, the constraint of self-finance sharply biases investment strategy toward marginal variations within the tradi-tional technology.

F. Facilitating Exchange

Besides easing savings mobilizationand thereby expanding the of set pro-duction technologies available to aneconomy, financial arrangements that

lower transaction costs can promote spe-cialization, technological innovation, andgrowth. The links between facilitatingtransactions, specialization, innovation,and economic growth were core ele-ments of Adam Smith’s (1776) Wealth of

Nations. Smith (1776, p. 7) argued thatdivision of labor—specialization—isthe principal factor underlying produc-tivity improvements. With greater spe-

cialization, workers are more likely to in- vent better machines or productionprocesses.

I shall only observe, therefore, that the invention of all those machines by which labouris so much facilitated and abridged, seems tohave been originally owing to the division of labour. Men are much more likely to discovereasier and readier methods of attaining any object, when the whole attention of theirminds is directed towards that single object,than when it is dissipated among a great vari-

ety of things. (Smith 1776, p. 3)

The critical issue for our purposes isthat the financial system can promotespecialization. Adam Smith argued thatlower transaction costs would permitgreater specialization because specializa-

tion requires more transactions than anautarkic environment. Smith phrased hisargument about the lowering of transac-tion costs and technological innovationin terms of the advantages of money overbarter (pp. 26–27). Information costs,however, may also motivate the emer-gence of money. Because it is costly toevaluate the attributes of goods, barterexchange is very costly. Thus, an easily recognizable medium of exchange may arise to facilitate exchange (King andCharles Plosser 1986; and Williamsonand Randall Wright 1994).18

The drop in transaction and informa-tion costs is not necessarily a one-timefall when economies move to money,however. For example, in the 1800s, “ it was primarily the development of insti-tutions that facilitated the exchange of technology in the market that enabled

creative individuals to specialize in andbecome more productive at invention”(Lamoreaux and Sokoloff 1996, p. 17).Thus, transaction and information costsmay continue to fall through a variety of mechanisms, so that financial and insti-tutional development continually boostspecialization and innovation via thesame channels illuminated over 200 years ago by Adam Smith.19

18 This focus on money as a medium of ex-change that lowers transaction and informationcosts by overcoming the “double coincidence of wants problem” and by acting as an easily recog-nizable medium of exchange enjoys a long history in monetary theory, from Adam Smith (1776), toStanley Jevons (1875), to Karl Brunner and AllanMeltzer (1971), to more formal models as reviewed by Joseph Ostroy and Starr (1990).

19 Financial systems can also promote the accu-mulation of human capital by lowering the costs of intertemporal trade, i.e., by facilitating borrowingfor the accumulation of skills (Thomas Cooley andB. Smith 1992; and Jose De Gregorio 1996). If

human capital accumulation is not subject to di-




Modern theorists have attempted to il-luminate more precisely the ties be-tween exchange, specialization, and in-novation (Greenwood and B. Smith1997). More specialization requires more

transactions. Because each transaction iscostly, financial arrangements that lowertransaction costs will facilitate greaterspecialization. In this way, markets thatpromote exchange encourage produc-tivity gains. There may also be feedbackfrom these productivity gains to finan-cial market development. If there arefixed costs associated with establishingmarkets, then higher income per capitaimplies that these fixed costs are lessburdensome as a share of per capita in-come. Thus, economic development canspur the development of financial mar-kets.

This approach to linking financial mar-kets with specialization has not yet for-mally completed Adam Smith’s story of innovation. That is, a better market—amarket with lower transactions costs—does not stimulate the invention of new

and better production technologies inGreenwood and B. Smith’s (1997)model. Instead, lower transaction costsexpand the set of “on the shelf” produc-tion processes that are economically at-tractive. Also, the model defines better“market” as a system for supportingmore specialized production processes.This does not explain the emergence of financial instruments or institutions thatlower transaction costs and thereby pro-duce an environment that naturally pro-motes specialized production technolo-gies. This is important because we wantto understand the two links of the chain: what about the economic environmentcreates incentives for financial arrange-ments to arise and to function well or

poorly, and what are the implications foreconomic activity of the emerging finan-cial arrangements?

G. A Parable

Thus far, I have discussed each finan-cial function in isolation. This, however,may encourage an excessively narrow fo-cus on individual functions and impedethe synthesis of these distinct functionsinto a coherent understanding of thefinancial system’s role in economic development. This is not a necessary impli-cation. In fact, by identifying the individ-ual functions performed by the financialsystem, the functional approach can fos-ter a more complete understanding of fi-nance and growth.

Earlier authors often provided illustra-tive stories of the ties between financeand development. For example Schum-peter (1912, pp. 58–74) and McKinnon(1973, pp. 5–18) provide broad descrip-tions—parables—of the roles of the fi-nancial system in economic develop-ment. Just as Smith (1776) used the pin

factory to illustrate the importance of specialization, Schumpeter used the re-lationship between banker and industri-alist to illustrate the importance of thefinancial system in choosing and adopt-ing new technologies, and McKinnonhighlighted its importance in promotingthe use of better agricultural techniques.However, even Schumpeter and McKin-non did not amalgamate all of the finan-cial functions into their stories of financeand development. Consequently, thissubsection synthesizes the individual fi-nancial functions into a simple parableabout how the financial system affectseconomic growth.

Consider Fred, who has just devel-oped a design for a new truck that ex-tracts rocks from a quarry better than ex-isting trucks. His idea for manufacturingtrucks requires an intricate assembly line

with specialized labor and capital.

minishing returns on a social level, financial ar-rangements that ease human capital creation help

accelerate economic growth.




Highly specialized production processes would be difficult without a medium of exchange. He would find it prohibitively costly to pay his workers and suppliersusing barter exchange. Financial instru-

ments and markets that facilitate trans-actions will allow and promote special-ization and thereby permit him toorganize his truck assembly line. More-over, the increased specialization in-duced by easier transactions may fosterlearning-by-doing and innovation by the workers specializing on their individualtasks.

Production requires capital. Even if Fred had the savings, he would not wishto put all of his savings in one risky investment. Also, he wants ready access tosavings for unplanned events; he is re-luctant to tie up his savings in the truckproject, which will not yield profits, if itdoes yield profits, for a long time. Hisdistaste for risk and desire for liquidity create incentives for him to (a) diversify the family’s investments and (b) notcommit too much of his savings to an il-

liquid project, like producing a newtruck. In fact, if Fred must invest dispro-portionately in his illiquid truck project,he may forgo his plan. Without a mecha-nism for managing risk, the project may die. Thus, liquidity, risk pooling, and di-

versification will help him start his innovative project.

Moreover, Fred will require outsidefunding if he has insufficient savings toinitiate his truck project. There areproblems, however, in mobilizing savingsfor Fred’s truck company. First, it is very costly and time consuming to collect sav-ings from individual savers. Fred doesnot have the time, connections, and in-formation to collect savings from every-one in his town and neighboring commu-nities even though his idea is sound.Banks and investment banks, however,can mobilize savings more cheaply than

Fred due to economies of scale, econo-

mies of scope, and experience. Thus,Fred may seek the help of a financial in-termediary to mobilize savings for hisnew truck plant.

Two additional problems (“frictions”)

may keep savings from flowing to Fred’sproject. To fund the truck plant, the fi-nancial intermediaries—and savers in fi-nancial intermediaries—require informa-tion about the truck design, Fred’sability to implement the design, and whether there is a sufficient demand forbetter quarry trucks. This information isdifficult to obtain and analyze. Thus, thefinancial system must be able to acquirereliable information about Fred’s ideabefore funding the truck plant. Further-more, if potential investors feel thatFred may steal the funds, or run theplant poorly, or misrepresent profits,they will not provide funding. To financeFred’s idea, outside creditors must haveconfidence that Fred will run the truckplant well. Thus, for Fred to receivefunding, the financial system must moni-tor managers and exert corporate con-

trol. While this parable does not contain allaspects of the discussion of financialfunctions, it provides one cohesive story of how the five financial functions may interact to promote economic develop-ment.

H. The Theory of Finance andEconomic Growth: Agenda

In describing the conceptual links be-tween the functioning of the financialsystem and economic growth, I high-lighted areas needing additional re-search. Two more areas are worth em-phasizing. First, we do not have asufficiently rigorous understanding of the emergence, development, and eco-nomic implications of different financialstructures. Financial structure—the mixof financial contracts, markets, and insti-

tutions—varies across countries and




changes as countries develop (Boyd andB. Smith 1996). Yet, we do not haveadequate theories of why different finan-cial structures emerge or why financialstructures change. Differences in legal

tradition (Rafael LaPorta et al. 1996) anddifferences in national resource endow-ments that produce different politicaland institutional structures (Stanley Engerman and Sokoloff 1996) might beincorporated into future models of finan-cial development. Furthermore, econo-mists need to develop an analytical basisfor making comparisons of financial struc-tures; we need models that elucidate theconditions, if any, under which differentfinancial structures are better at mitigat-ing information and transaction costs.

A second area needing additionalresearch involves the influence of thelevel and growth rate of the economy on the financial system. Some modelsassume that there is a fixed cost to join-ing financial intermediaries. Economicgrowth then reduces the importance of this fixed cost and more people join.

Thus, economic growth provides themeans for the formation of growth-pro-moting financial intermediaries, whilethe formation of financial intermediariesaccelerates growth by enhancing the al-location of capital. In this way, financialand economic development are jointly determined (Greenwood and Jovanovic1990). Economic development may af-fect the financial system in other waysthat have not yet been formally modeled.For example, the costs and skills re-quired to evaluate production technolo-gies and monitor managers may be very different in a service-oriented economy from that of a manufacturing-basedeconomy or an agricultural-based econ-omy. Building on Hugh Patrick (1966),Greenwood and Jovanovic (1990), andGreenwood and Smith (1997), future re-search may improve our understanding of

the impact of growth on financial systems.

III. Evidence

A. The Questions

Are differences in financial develop-ment and structure importantly associ-ated with differences in economicgrowth rates? To assess the nature of thefinance-growth relationship, I first de-scribe research on the links between thefunctioning of the financial system andeconomic growth, capital accumulation,and technological change. Then, I evalu-ate existing evidence on the ties betweenfinancial structure—the mix of financial

markets and intermediaries—and the

functioning of the financial system. Agrowing body of work demonstrates astrong, positive link between financialdevelopment and economic growth, andthere is even evidence that the level of financial development is a good predic-tor of future economic development.Evidence on the relationship between fi-nancial structure and the functioning of the financial system, however, is moreinconclusive.

B. The Level of Financial Developmentand Growth: Cross-Country Studies

Consider first the relationship be-tween economic growth and aggregatemeasures of how well the financial sys-tem functions. The seminal work in thisarea is by Goldsmith (1969). He uses the value of financial intermediary assets divided by GNP to gauge financial devel-

opment under the assumption that thesize of the financial system is positively correlated with the provision and quality of financial services. Using data on 35countries from 1860 to 1963 (when avail-able) Goldsmith (1969, p. 48) finds:

(1) a rough parallelism can be observed be-tween economic and financial development if periods of several decades are considered;[and](2) there are even indications in the few

countries for which the data are available that




periods of more rapid economic growth havebeen accompanied, though not without ex-ception, by an above-average rate of financialdevelopment.

Goldsmith’s work, however, has sev-

eral weaknesses: (a) the investigation involves limited observations on only 35countries; (b) it does not systematically control for other factors influencing eco-nomic growth (Levine and David Renelt1992); (c) it does not examine whetherfinancial development is associated withproductivity growth and capital accumu-lation; (d) the size of financial intermedi-aries may not accurately measure thefunctioning of the financial system; and

(e) the close association between the sizeof the financial system and economicgrowth does not identify the direction of causality.20

Recently, researchers have taken stepsto address some of these weaknesses.For example, King and Levine (1993a,1993b, 1993c) study 80 countries overthe period 1960–1989, systematically control for other factors affecting long-

run growth, examine the capital accumu-lation and productivity growth channels,construct additional measures of thelevel of financial development, and ana-lyze whether the level of financial development predicts long-run economicgrowth, capital accumulation, and pro-ductivity growth. (Also, see Alan Gelb1989; Gertler and Andrew Rose 1994;Nouriel Roubini and Xavier Sala-i-Martin 1992; Easterly 1993; and the

overview by Pagano 1993.) They use fourmeasures of “the level of financial devel-opment” to more precisely measure the

functioning of the financial system thanGoldsmith’s size measure. Table 1 sum-marizes the values of these measuresrelative to real per capita GDP (RGDP)in 1985. The first measure, DEPTH,

measures the size of financial intermedi-aries and equals liquid liabilities of thefinancial system (currency plus demandand interest-bearing liabilities of banksand nonbank financial intermediaries)divided by GDP. As shown, citizens of the richest countries—the top 25 per-cent on the basis of income per capita—held about two-thirds of a year’s incomein liquid assets in formal financial inter-mediaries, while citizens of the poorestcountries—the bottom 25 percent—heldonly a quarter of a year’s income in liq-uid assets. There is a strong correlationbetween real per capita GDP andDEPTH. The second measure of finan-cial development, BANK, measures thedegree to which the central bank versuscommercial banks are allocating credit.BANK equals the ratio of bank credit divided by bank credit plus central bank

domestic assets. The intuition underly-ing this measure is that banks are morelikely to provide the five financial func-tions than central banks. There are twonotable weaknesses with this measure,however. Banks are not the only finan-cial intermediaries providing valuable fi-nancial functions and banks may simply lend to the government or public enter-prises. BANK is greater than 90 percentin the richest quartile of countries. Incontrast, commercial banks and centralbanks allocate about the same amount of credit in the poorest quartile of coun-tries. The third and fourth measures par-tially address concerns about the alloca-tion of credit. The third measures,PRIVATE, equals the ratio of credit allo-cated to private enterprises to total do-mestic credit (excluding credit to banks).The fourth measure, PRIVY, equals

credit to private enterprises divided by

20 Goldsmith (1969) recognized these weak-nesses, e.g., “there is no possibility, however, of establishing with confidence the direction of thecausal mechanisms, i.e., of deciding whether fi-nancial factors were responsible for the accelera-tion of economic development or whether finan-cial development reflected economic growth whose mainsprings must be sought elsewhere” (p.

48).




GDP. The assumption underlying thesemeasures is that financial systems thatallocate more credit to private firms aremore engaged in researching firms, ex-erting corporate control, providing riskmanagement services, mobilizing sav-ings, and facilitating transactions than fi-nancial systems that simply funnel creditto the government or state owned enter-prises. As depicted in Table 1, there is apositive, statistically significant correla-tion between real per capita GDP andthe extent to which loans are directed tothe private sector.

King and Levine (1993b, 1993c) thenassess the strength of the empirical rela-tionship between each of these four indi-cators of the level of financial develop-ment averaged over the 1960–1989period, F, and three growth indicatorsalso averaged over the 1960–1989 pe-

riod, G. The three growth indicators are

as follows: (1) the average rate of realper capita GDP growth, (2) the averagerate of growth in the capital stock perperson, and (3) total productivity growth, which is a “Solow residual” de-fined as real per capita GDP growth mi-nus (0.3) times the growth rate of thecapital stock per person. In other words,if F( i) represents the value of the ith in-dicator of financial development(DEPTH, BANK, PRIVY, PRIVATE) av-eraged over the period 1960–1989, G(j)represents the value of the jth growth in-dicator (per capita GDP growth, per cap-ita capital stock growth, or productivity growth) averaged over the period 1960–1989, and X represents a matrix of condi-tioning information to control for otherfactors associated with economic growth(e.g., income per capita, education, po-litical stability, indicators of exchange

rate, trade, fiscal, and monetary policy),

TABLE 1FINANCIAL DEVELOPMENT AND REAL PER CAPITA GDP IN 1985

Indictors Very rich Rich Poor Very poor

Correlation with Real per

Capita GDP in1985 (P-value)

DEPTH 0.67 0.51 0.39 0.26 0.51 (0.0001)

BANK 0.91 0.73 0.57 0.52 0.58 (0.0001)

PRIVATE 0.71 0.58 0.47 0.37 0.51 (0.0001)

PRIVY 0.53 0.31 0.20 0.13 0.70 (0.0001)

RGDP85 13053 2376 754 241

Observations 29 29 29 29

Source: King and Levine (1993a)

Very rich: Real GDP per Capita > 4998Rich: Real GDP per Capita > 1161 and < 4998Poor: Real GDP per Capita > 391 and < 1161 Very poor: Real GDP per Capita < 391

DEPTH = Liquid liabilities to GDPBANK = Deposit money bank domestic credit divided by deposit money bank + central bank domestic creditPRIVATE = Claims on the non-financial private sector to domestic creditPRIVY = Gross claims on private sector to GDPRGDP85 = Real per capita GDP in 1985 (in constant 1987 dollars)




then the following 12 regressions are runon a cross-section of 77 countries:

G( j) = α + βF( i) + γ X + ε. (1)

There is a strong positive relationshipbetween each of the four financial devel-opment indicators, F(i), and the threegrowth indicators G(i), long-run real percapita growth rates, capital accumula-tion, and productivity growth. Table 2summarizes the results on the 12 β’s.Not only are all the financial develop-ment coefficients statistically significant,the sizes of the coefficients imply aneconomically important relationship. Ig-noring causality, the coefficient of 0.024on DEPTH implies that a country thatincreased DEPTH from the mean of theslowest growing quartile of countries

(0.2) to the mean of the fastest growing

quartile of countries (0.6) would have in-creased its per capita growth rate by al-most one percent per year. This is large.The difference between the slowestgrowing 25 percent of countries and thefastest growing quartile of countries isabout five percent per annum over this30 year period. Thus, the rise in DEPTHalone eliminates 20 percent of thisgrowth difference.

Finally, to examine whether financesimply follows growth, King and Levine(1993b) study whether the value of fi-nancial depth in 1960 predicts the rateof economic growth, capital accumula-tion, and productivity improvementsover the next 30 years. Table 3 summa-rizes some of the results. In the threeregressions reported in Table 3, the de-

pendent variable is, respectively, real per

TABLE 2GROWTH AND CONTEMPORANEOUS FINANCIAL INDICATORS, 1960–1989

Dependant Variable DEPTH BANK PRIVATE PRIVY

Real Per Capita GDP Growth 0.024***

[0.007]

0.032***

[0.005]

0.034***

[0.002]

0.032***

[0.002]R2 0.5 0.5 0.52 0.52

Real Per Capita Capital Stock Growth 0.022***[0.001]

0.022**[0.012]

0.020**[0.011]

0.025***[0.001]

R2 0.65 0.62 0.62 0.64

Productivity Growth 0.018**[0.026]

0.026**[0.010]

0.027***[0.003]

0.025***[0.006]

R2 0.42 0.43 0.45 0.44

Source: King and Levine (1993b)* significant at the 0.10 level, ** significant at the 0.05 level, *** significant at the 0.01 level.[p-values in brackets]

Observations = 77

DEPTH = Liquid liabilities to GDPBANK = Deposit bank domestic credit divided by deposit money bank + central bank domestic

creditPRIVATE = Claims on the non-financial private sector to total claimsPRIVY = Gross claims on private sector to GDPProductivity Growth = Real Per Capita GDP Growth - (0.3)*Real Per Capita Capital Stock Growth

Other explanatory variables included in each of the 12 regressions: log of initial income, log of initial secondary school enrollment rate, ratio of government consumption expenditures to GDP, inflation rate, and ratio of exportplus imports to GDP.




capita GDP growth, real per capita capi-tal stock growth, and productivity growthaveraged over the period 1960–1989.The financial indicator in each of theseregressions is the value of DEPTH in1960. The regressions indicate thatfinancial depth in 1960 is significantly correlated with each of the growth indi-cators averaged over the period 1960–1989.21 These results, plus those from

more sophisticated time series studies,suggest that the initial level of financialdevelopment is a good predictor of sub-sequent rates of economic growth, physi-cal capital accumulation, and economicefficiency improvements over the next 30 years even after controlling for income, ed-ucation, political stability, and measuresof monetary, trade, and fiscal policy.22

TABLE 3GROWTH AND INITIAL FINANCIAL DEPTH, 1960–1989

Per Capita GDPGrowth, 1960–1989

Per Capita CapitalGrowth, 1960–1989

Per Capita Productivity Growth, 1960–1989

Constant 0.035***[0.001] 0.002[0.682] 0.034***[0.001]Log (Real GDP per

Person in 1960)−0.016***

[0.001]−0.004*

[0.068]−0.015***

[0.001]Log (Secondary school

enrollment in 1960)0.013***[0.001]

0.007***[0.001]

0.011***[0.001]

Governmentconsumption/GDP in 1960

0.07*[0.051]

0.049*[0.064]

0.056*[0.076]

Inflation in 1960 0.037[0.239]

0.02[0.238]

0.029[0.292]

(Imports plus Exports)/GDPin 1960

−0.003[0.604]

−0.001[0.767]

−0.003[0.603]

DEPTH (liquid liabilities)in 1960

0.028***[0.001]

0.019***[0.001]

0.022***[0.001]

R2 0.61 0.63 0.58

Source: King and Levine (1993b)* significant at the 0.10 level, **significant at the 0.05 level, *** significant at the 0.01 level.[p-values in brackets]Observations = 57

21 There is an insufficient number of observa-tions on BANK, PRIVATE, and PRIVY in 1960 toextend the analysis in Table 3 to these variables.Thus, King and Levine (1993b) use pooled, crosssection, time series data. For each country, datapermitting, they use data averaged over the 1960s,1970s, and 1980s; thus, there are potentially threeobservations per country. They then relate the value of growth averaged over the 1960s with the value of, for example, BANK in 1960 and so on forthe other two decades. They restrict the coeffi-cients to be the same across decades. They findthat the initial level of financial development is agood predictor of subsequent rates of economicgrowth, capital accumulation, and economic effi-ciency improvements over the next ten years after

controlling for many other factors associated withlong-run growth.

22 These broad cross-country results hold even when using instrumental variables—primarily indi-cators of the legal treatment of creditors takenfrom LaPorta et al. 1996—to extract the exoge-nous component of financial development (Levine1997). Furthermore, though disagreement exists(Woo Jung 1986 and Philip Arestis and PanicosDemetriades 1995), many time-series investiga-tions find that financial sector developmentGranger-causes economic performance (Paul Wachtel Rousseau 1995). These results are par-ticularly strong when using measures of the value-added provided by the financial system instead of measures of the size of the financial system (KlausNeusser and Maurice Kugler 1996).




The relationship between the initiallevel of financial development andgrowth is large. For example, the esti-mated coefficients suggest that if in 1960Bolivia had increased its financial depth

from 10 percent of GDP to the mean value for developing countries in 1960(23 percent), then Bolivia would havegrown about 0.4 percent faster per an-num, so that by 1990 real per capitaGDP would have been about 13 percentlarger than it was.23 Thus, finance doesnot merely follow economic activity. Thestrong link between the level of financialdevelopment and the rate of long-runeconomic growth does not simply reflectcontemporaneous shocks that affect bothfinancial development and economicperformance. There is a statistically significant and economically large em-pirical relationship between the initiallevel of financial development and fu-ture rates of long-run growth, capitalaccumulation, and productivity improve-ments. Furthermore, insufficient finan-cial development has sometimes created

a “poverty trap” and thus become a severe obstacle to growth even when acountry has established other conditions(macroeconomic stability, openness totrade, educational attainment, etc.) forsustained economic development (Jean-Claude Berthelemy and Aristomene Varoudakis 1996).

Some recent work has extended ourknowledge about the causal relationshipsbetween financial development and eco-nomic growth. For example, Rajan andLuigi Zingales (1996) assume that finan-cial markets in the United States arerelatively frictionless. This benchmarkcountry then defines each industry’s effi-cient demand for external finance (investment minus internal cash flow). They then examine industries across a large

sample of countries and test whether theindustries that are more dependent onexternal finance (in the United States)grow relatively faster in countries thatbegin the sample period with better de-

veloped financial systems. They find thatindustries that rely heavily on externalfunding grow comparatively faster incountries with well-developed interme-diaries (as measured by PRIVY) andstock markets (as measured by stockmarket capitalization) than they do incountries that start with relatively weakfinancial systems. Similarly, using firm-level data from 30 countries, Asli Demir-guç-Kunt and Vojislav Maksimovic(1996b) argue that firms with access tomore developed stock markets grow atfaster rates than they could have grown without this access. Furthermore, whenindividual states of the United States re-laxed intrastate branching restrictions,this boosted bank lending quality and ac-celerated real per capita growth rateseven after controlling for other growthdeterminants (Jith Jayaratne and Philip

Strahan 1996). Thus, using firm- and in-dustrial-level data for a broad cross-section of countries and data on individual states of the United States,recent research presents evidence con-sistent with the view that the level of fi-nancial development materially affectsthe rate and structure of economic development.

Not surprisingly, these empirical stud-ies do not unambiguously resolve the is-sue of causality. Financial developmentmay predict growth simply because fi-nancial systems develop in anticipationof future economic growth. Further-more, differences in political systems, le-gal traditions (LaPorta et al. 1996), or in-stitutions (Engerman and Sokoloff 1996;Douglass North 1981) may be drivingboth financial development and eco-nomic growth rates. Nevertheless, the

body of evidence would tend to push

23 These examples do not consider causal issues

or how to increase financial development.




many skeptics toward the view that thefinance-growth link is a first-order rela-tionship and that difference in financialdevelopment can alter economic growthrates over ample time horizons.

C. Country-Case Studies

Country-case studies provide a richcomplement to cross-country compari-sons. For example, Rondo Cameron etal. (1967) dissect the historical relation-ships between banking development andthe early stages of industrialization forEngland (1750–1844), Scotland (1750–1845), France (1800–1870), Belgium(1800–1875), Germany (1815–1870),Russia (1860–1914), and Japan (1868–1914). These country-case studies do notuse formal statistical analysis. Instead,the researchers carefully examine the le-gal, economic, and financial linkages be-tween banks and industry during the in-dustrialization of these seven countries.Typically, the case studies start by de-scribing the political system, economicconditions, and financial structure at the

start of the period of analysis. Then, they provide a detailed description of the evo-lution of the financial system during aperiod of rapid economic development.Finally, they document critical inter-actions among financial intermediaries,financial markets, government policies,and the financing of industrialization. While providing an informative comple-ment to broad cross country compari-sons, country-case studies rely heavily onsubjective evaluations of banking systemperformance and fail to systematically control for other elements determiningeconomic development. While emphasiz-ing the analytical limitations of country-case studies, Cameron (1967b) con-cludes that especially in Scotland and Ja-pan, but also in Belgium, Germany, En-gland, and Russia, the banking systemplayed a positive, growth-inducing role.

Debate exists, however. Consider the

case of Scotland between 1750 and 1845.Scotland began the period with per cap-ita income of less than one-half of En-gland’s. By 1845, however, per capita in-come was about the same. While

recognizing that the “dominant politicalevent affecting Scotland’s potentialitiesfor economic development was theUnion of 1707, which made Scotland anintegral part of the United Kingdom,”Cameron (1967a, p. 60), argues thatScotland’s superior banking system isone of the few noteworthy features thatcan help explain its comparatively rapidgrowth.24 Other analysts disagree withthe “facts” underlying this conclusion.Some researchers suggest that Englanddid not suffer from a dearth of financialservices because nonfinancial enterprisesprovided financial services in Englandthat Cameron’s (1967a) measures of for-mal financial intermediation omit. Oth-ers argue that Scotland had rich naturalresources, a well-educated work force,access to British colonial markets, andstarted from a much lower level of in-

come per capita than England. Conse-quently, it is not surprising that Scotlandenjoyed a period of rapid convergencetoward England’s income per capitalevel. Finally, still other researchers dis-agree with the premise that Scotland hada well-functioning financial system andemphasize the deficiencies in the Scot-tish system (Sidney Pollard and DieterZiegler 1992). Thus, although AndrewKerr first argued in 1884 that Scotlandenjoyed a better banking system thanEngland from 1750 until 1844, the de-bate about whether Scottish banking ex-plains its faster economic growth overthe period 1750–1845 continues today.

24 It is also worth noting that Scottish banking was comparatively stable over this period, suffer-ing fewer and less severe panics than its southernneighbor. For more on Scottish banking, see Syd-ney Checkland (1975) and Tyler Cowen and Ran-

dall Kroszner (1989).




The relationship between financialand economic development has beencarefully analyzed for many other coun-tries. For example, Stephen Haber(1991, 1996) compares industrial and

capital market development in Brazil,Mexico, and the United States between1830 and 1930. He finds that capitalmarket development affected industrialcomposition and national economic per-formance. Specifically, Haber shows that when Brazil overthrew the monarchy in1889 and formed the First Republic, italso dramatically liberalized restrictionson Brazilian financial markets. The liber-alization gave more firms easier access toexternal finance. Industrial concentra-tion fell and industrial productionboomed. While Mexico also liberalizedfinancial sector policies, the liberaliza-tion was much more mild under the Diazdictatorship (1877–1911), which “reliedon the financial and political support of asmall in-group of powerful financialcapitalists” (p. 561). As a result, the de-cline in concentration and the increase

in economic growth was much weaker inMexico than it was in Brazil. Haber(1996, p. 40) concludes that “differencesin capital market development had a sig-nificant impact on the rate of growth of industry. . . . [and that a] lack of accessto institutional sources of capital becauseof poorly developed capital markets wasa non-negligible obstacle to industrialdevelopment in the nineteenth cen-tury.”25

Finally, but perhaps most influen-

cially, McKinnon’s (1973) seminal bookMoney and Capital in Economic Devel-opment studies the relationship betweenthe financial system and economic devel-opment in Argentina, Brazil, Chile, Ger-

many, Korea, Indonesia, and Taiwan inthe post World War II period. McKin-non interprets the mass of evidenceemerging from these country-case stud-ies as strongly suggesting that betterfunctioning financial systems supportfaster economic growth. Disagreementexists over many of these individualcases, and it is extremely difficult to iso-late the importance of any single factorin the process of economic growth.26

Thus, any statements about causality are—and will remain—largely impres-sionistic and specific to particular coun-tries and specific periods. Nonetheless,the body of country-studies suggeststhat, while the financial system respondsto demands from the nonfinancial sector, well-functioning financial systems have,in some cases during some time periods,greatly spurred economic growth.

D. Financial Functions and Growth: Liquidity and Risk

I now turn to evidence on the ties be-tween measures of the individual finan-cial functions and economic growth.First, consider liquidity. Deposit-takingbanks can provide liquidity by issuingliquid demand deposits and making illiq-uid, long-term investments. Isolating thisliquidity function from the other finan-cial functions performed by banks, how-ever, has proven prohibitively difficult.In contrast, economists have studied ex-tensively the effects of the liquidity of anindividual security on its price. Substan-tial evidence suggests a positive correla-tion between the liquidity of an asset

25 Interestingly, these political and legal im-pediments to financial development are appar-ently difficult to change. In Mexico, the largestthree banks control the same fraction of commer-cial banking activity today, about two-thirds, asthey did 100 years ago. Also, Mexico has the low-est ranking of the legal protection of minority shareholder rights of any country in La Porta etal.’s (1996) detailed comparison of 49 countries, which may facilitate the concentration of eco-

nomic decision making.

26 For more on Mexico see Robert Bennett(1963). For more on Asia, see Cole and Yung Park(1983), Park (1993), and Patrick and Park (1994).

Fry (1995) provides additional citations.




TABLE 4STOCK MARKET LIQUIDITY MEASURES: SELECTED COUNTRIES, ANNUAL A VERAGES 1976–1993

TurnoverRatio

ValueTraded Ratio

Per CapitaGDP Growth

Low-incomeBangladesh 0.015 0.000 1.89%Cote d’lvoire 0.028 0.001 −2.50%Egypt 0.060 0.030 3.56%India 0.537 0.036 2.43%Nigeria 0.006 0.000 −0.11%Pakistan 0.105 0.008 3.13%Zimbabwe 0.059 0.010 −0.97%

Lower-middle -incomeColombia 0.087 0.004 1.95%Costa Rica 0.013 0.001 0.89%Indonesia 0.193 0.010 4.18%

Jordan 0.154 0.085 3.01%Philippines 0.250 0.026 0.21%Thailand 0.739 0.144 5.90%Turkey 0.207 0.026 2.32%

Upper-middle-incomeArgentina 0.266 0.013 0.22%Brazil 0.355 0.041 0.65%Chile 0.060 0.021 3.61%Korea 0.832 0.186 9.67%Malaysia 0.230 0.243 4.27%Mauritius 0.059 0.003 1.76%Mexico 0.498 0.044 0.85%Portugal 0.108 0.014 2.85%

High-incomeAustralia 0.256 0.124 1.57%Germany 0.704 0.156 0.95%Great Britian 0.349 0.253 1.75%Hong Kong 0.372 0.471 6.20%Israel 0.669 0.144 1.72%Italy 0.253 0.028 2.68%Japan 0.469 0.406 3.42%Netherlands 0.490 0.123 1.43%Norway 0.318 0.059 2.48%Spain 0.216 0.045 1.75%Switzerland 0.467 0.442 1.16%

United States 0.493 0.299 1.67%

Sources: International Finance Corporation, and Morgan Stanley Capital InternationalTurnover Ratio = Value of Domestic Equities Traded on Domestic Exchanges Divided by Market Capitalization Value Traded Ratio = Value of Domestic Equities Traded on Domestic Exchanges Divided by GDP Income

classifications from the World Bank’s 1995 World Development ReportLow-income economies = average GNP per capita of $380 in 1993Lower-middle-income economies = average GNP per capita of $1,590 in 1993Upper-middle-income economies = average GNP per capita of $4,370 in 1993High-income economies = average GNP per capita of $23,090 in 1993




and its price (e.g., Yakov Amihud andHaim Mendelson 1989; and Gregory Kadlec and John McConnell 1994). Putdifferently, agents must be compensated with a lower price for purchasing an as-

set that is difficult to sell. These secu-rity-level studies of the relationship be-tween the liquidity of individualsecurities and their prices, however, donot link liquidity with national long rungrowth rates.

To evaluate the relationship betweenstock market liquidity and nationalgrowth rates, capital accumulation rates,and rates of technological change,Levine and Sara Zervos (1996) build onRaymond Atje and Jovanovic’s (1993)study and focus on two measures of liq-uidity for a broad cross-section of 49countries over the period 1976–1993.The first liquidity indicator, the valuetraded ratio, equals the total value of shares traded on a country’s stock ex-changes divided by GDP. The valuetraded ratio measures trading relative tothe size of the economy. While not a di-

rect measure of trading costs or the un-certainty associated with trading and set-tling equity transactions, theoreticalmodels of liquidity and growth directly motivate the value traded ratio (Ben-civenga, B. Smith, and Starr 1995). Asshown in Table 4, the value traded ratio varies considerably across countries. Forexample, the United States had an aver-age annual value traded ratio of 0.3 dur-ing the 1976–1993 period, while forMexico and India it was about 0.04. Thesecond indicator, the turnover ratio,equals the total value of shares traded ona country’s stock exchanges divided by stock market capitalization (the value of listed shares on the country’s ex-changes). The turnover ratio measurestrading relative to the size of the market.It also exhibits substantial cross-country variability. Very active markets such as

Japan and the United States have turn-

over ratios of almost 0.5, while for lessliquid markets, such as Bangladesh,Chile, and Egypt they are 0.06 or less.27

The turnover ratio may differ from the value traded ratio because a small, liquid

market will have high turnover ratio buta small value traded ratio. For example,India’s average turnover ratio of 0.5 overthe 1976–1993 is greater than theUnited States’, but India’s value tradedratio is about one-tenth the size of theUnited States’. These measures seek tomeasure liquidity on a macroeconomicscale: the objective is to measure the de-gree to which agents can cheaply,quickly, and confidently trade ownershipclaims of a large percentage of the econ-omy’s productive technologies.28

The researchers then assess thestrength of the empirical relationshipbetween each liquidity measure and thethree growth indicators: economic growth,capital accumulation, and productivity growth. They conduct a cross-country analysis with one observation per coun-try. Namely, six basic regressions are

run: economic growth, capital accumula-tion, and productivity growth averagedover the 1976–1993 period are regressedfirst on the value traded ratio in 1976and then on the turnover ratio in 1976 while controlling for various factors asso-ciated with economic growth (initial in-come per capita, education, political sta-bility, indicators of exchange rate, trade,fiscal, and monetary policy) to see whether stock market liquidity predictssubsequent economic growth. Impor-tantly, the level of banking sector devel-opment (bank credit to private enter-prises divided by GDP) measured in

27 Note, Germany’s very large turnover ratio(0.7) reflects the explosion in stock market trans-actions during unification.

28 Levine and Zervos (1996) also construct andexamine two measures of stock trading relative tostock price movements: (1) the value traded ratiodivided by stock return volatility, and (2) the turn-

over ratio divided by stock return volatility.




1976 is included in the regressions to as-sess the independent link between stockmarket liquidity and growth after con-trolling for other aspects of financial development. The results are summarizedin Table 5. The initial level of stock mar-ket liquidity—measured either by theturnover ratio or the value traded ratio—is a statistically significant predictor of economic growth, capital accumulation,and productivity growth over the next 18 years. The sizes of the coefficients alsosuggest an economically meaningful rela-tionship. For example, the results imply that if Mexico had had the sample aver-age value traded ratio in 1976 (0.044) in-stead of its realized 1976 value (0.004),per capita GDP would have grown at a0.4 percent faster rate (0.04*0.098). Ac-

cumulating over the 18 year period, this

implies each Mexican would have en- joyed an almost 8 percent higher incomein 1994. The results are consistent withthe views that the liquidity services provided by stock markets are indepen-dently important for long-run growthand that stock markets provide differentfinancial services from those provided by financial intermediaries (or else they would not both enter the growth regres-sions significantly).29

Besides the difficulty of assigning acausal role to stock market liquidity,there are important limitations to mea-suring it accurately (Sanford Grossmanand Merton Miller 1988; and Stephen

TABLE 5GROWTH AND INITIAL STOCK MARKET LIQUIDITY, 1976–1993

Dependant Variable Value Traded

RatioTurnover

Ratio

Real Per Capita GDP Growth 0.098***[0.003]

0.027***[0.006]

Adjusted R2 0.33 0.34Real Per Capita Capital Stock Growth 0.093***

[0.005]0.022***[0.023]

Adjusted R2 0.38 0.35

Productivity Growth 0.075***[0.001]

0.020**[0.030]

Adjusted R2 0.21 0.21

Source: Levine and Zervos (1996)

* significant at the 0.10 level, ** significant at the 0.05 level, *** significant at the 0.01 level.[p-values in brackets]Observations = 42

Value Traded Ratio = Value of domestic equity transactions on domestic stock exchanges divided by GDPTurnover Ratio = Value of domestic equity transactions on domestic stock exchanges divided by domestic marketcapitalization.

Other explanatory variables included in each of the six regressions:log of initial income, log of initial secondary school enrollment, initial ratio of government expenditures to GDP,initial inflation rate, initial black market exchange rate premium, initial ratio of commercial bank lending to privateenterprises divided by GDP.

29 Stock market size, as measured by marketcapitalization divided by GDP, is not robustly cor-related with growth, capital accumulation, and

productivity improvements.




Wells 1994). Theory suggests that econo-mies will benefit from the ability totrade ownership of an economy’s pro-ductive technologies easily. Stock mar-kets, however, are only one mechanism

for providing liquidity. Banks and bondmarkets may also provide liquidity. Thus,measures of stock market liquidity mightomit important financial arrangementsfor providing liquidity. Moreover, theliquidity indicators measure stock trans-actions on a country’s national stock ex-changes. The physical location of thestock market, however, should not neces-sarily matter. That is, Californian saversand firms would probably not havegreater access to liquidity if the NewYork Stock Exchange were to move toLos Angeles. Thus, measures of the trad-ing of equities on a country’s exchangesmay not gauge fully the degree of stockmarket liquidity available to the econ-omy. This measurement problem willincrease over time if economies be-come more financially integrated andfirms list and issue shares on foreign ex-

changes.Besides liquidity risk, the financial sys-tem also provides mechanisms for hedg-ing and trading the idiosyncratic risk as-sociated with individual projects, firms,industries, sectors, and countries. Whilea vast literature examines the pricing of risk, there exists very little empirical evi-dence that directly links risk diversifica-tion services with long-run economicgrowth. Moreover, the only study of therelationship between economic growthand the ability of investors to diversify risk internationally through equity mar-kets yields inconclusive results (Levineand Zervos 1996).

One common weakness in empirical work on liquidity, idiosyncratic risk, andeconomic growth is that it focuses on eq-uity markets. Bond markets and financialintermediaries may also provide mecha-

nisms for diversifying risk. Indeed, tech-

nological, regulatory, and tax differencesacross countries may imply that differentfinancial structures arise to provide liq-uidity and risk diversification vehicles.For example, in one economy the costs

of establishing an intermediary may behigh while the costs of conducting equity transactions are low. The reverse may hold in a second economy. The firsteconomy may provide liquidity and riskdiversification services primarily throughequity markets, while the second does itthrough financial intermediaries. Thefirst economy has an active stock ex-change, so that existing empirical studies would classify it as providing substantialliquidity and risk diversification services.In contrast, existing studies would class-ify the second economy as financially un-derdeveloped. Thus, measuring the per-formance of one part of the financialsystem may generate a misleading indi-cator of the functioning of the whole fi-nancial system.

E. Financial Functions and Growth:Information

Theory strongly suggests that financialintermediaries play an important role inresearching productive technologies be-fore investment and monitoring manag-ers and projects after funneling capitalto those projects. Although it is very dif-ficult to measure whether a country’s fi-nancial system is comparatively adept atreducing information acquisition costsfirm level studies provide insights intothe role played by financial intermediar-ies in easing information asymmetries(Schiantarelli 1995). Theory suggeststhat as the costs to outsiders of acquiringinformation about a firm rise, a firm’s investment decisions become more tightly constrained by retained earnings andcurrent cash flow. Thus, studies test whether the investment decisions of firms with particular traits that proxy for

the costs to outsiders of acquiring infor-




mation are more sensitive to cash flowthan firms without those traits. The sam-ple selection criterion varies across stud-ies.

The empirical evidence suggests that

the investment decisions of firms withmore severe asymmetric informationproblems are more sensitive to cash flowthan those where it is less expensive foroutsiders to monitor. This conclusionholds when firms are classified accordingto whether they have received bondratings (Toni Whited 1992; CharlesCalomiris, Charles Himmelberg, and Wachtel 1995), whether they are issuinglarge or small dividends (Steven Fazzari,Glen Hubbard, and Bruce Peterson1988; Hubbard, Anil Kashyap, and Whited 1995), whether they are large orsmall (James Tybout 1983; Gertler andSimon Gilchrist 1994), whether they place a relatively high or low shadow value on internal funds based on theirresponse to taxes (Calomiris and Hub-bard 1995), and whether regulationsrestrict bank credit allocation (Fidel

Jaramillo, Schiantarelli, Weiss 1996;John Harris, Schiantarelli, and MirandaSiregar 1994). In sum, when outsidersfind it expensive to evaluate and fundparticular firms, those firms find itrelatively difficult to raise capital for investment and rely disproportionately oninternal sources of finance. Thus, finan-cial innovations or policies that lowerinformation asymmetries ease firm fi-nancing constraints on more efficientfirms.

More relevant for this section, a largebody of work shows that when firms haveclose ties to financial intermediaries, thisreduces information costs and eases firmfinancing constraints. Specifically, firms with close ties to banks tend to be lessconstrained in their investment decisionsthan those with less intimate, less ma-ture banking relationships as shown for

Japan (Takeo Hoshi, Kashyap, and

Scharfstein 1990), Italy (Schiantarelliand Alessandro Sembenelli 1996), andthe United States (Petersen and Rajan1994). Furthermore, borrowers withlonger banking relationships pay lower

interest rates and are less likely topledge collateral than those with lessmature banking relationships (Petersenand Rajan 1994; and Allen Berger andGregory Udell 1995). Finally, stock priceevidence also indicates that banks pro-duce valuable, private information aboutborrowers. When banks sign loan agree-ments with borrowers, borrower-firmstock prices respond positively (Christo-pher James 1987; Scott Lummar andMcConnell 1989; and James and Peggy Weir 1990). The value of the informationobtained by banks about firms can alsobe exemplified by Continental Illinois’troubles in the mid-1980s. Myron Slovin,Marie Sushka, and John Polonchek(1993) show that the banks’ impendinginsolvency negatively affected the stockprices of its client firms and that theFDIC’s rescue efforts positively affected

the stock prices of those same clients.These findings are consistent with the view that the durability of bank-bor-rower relationship is valuable. The evi-dence directly indicates an importantrole for financial intermediaries in re-ducing informational asymmetries be-tween firm insiders and outside inves-tors. Indirectly, the evidence suggeststhat countries with financial institutionsthat are effective at relieving informa-tion barriers will promote faster eco-nomic growth through more investmentthan countries with financial systemsthat are less effective at obtaining andprocessing information.

F. Patterns of Financial Development

I now turn to the question: Does fi-nancial structure change as countries develop and does it differ across countries?

Again, Goldsmith pioneered the cross-




country work in this area. He traced therelationship between the mix of financialintermediaries and economic develop-ment for 35 countries over the period1860–1963. The World Bank (1989) andDemirguç-Kunt and Levine (1996b) re-cently extended Goldsmith’s work by examining the association between the

mix of financial intermediaries, markets,

and economic development for approxi-mately 50 countries over the period1970–1993. This work finds that finan-cial structure differs considerably acrosscountries and changes as countries develop economically.

Four basic findings emerge fromthese studies, which are illustrated in

Figure 2. As countries get richer over

P e r c e n t a g e o f G D P

FinancialDepth

180

160

140

120

100

80

60

40

20

0

Figure 2. Financial Structure in Low-, Middle-, and High-Income Economies, 1990

Sources: IMF (International Financial Statistics), IFC (Emerging Markets Data Base), and individualcountry reports by central banks, banking commissions, and stock exchanges.

Notes: (1) The data are for 12 low-income economies (Bangladesh, Egypt, Ghana, Guyana, India,Indonesia, Kenya, Nigeria, Pakistan, Zaire, Zambia, and Zimbabwe), 22 middle-income economies

(Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, the Dominican Republic, El Salvador, Greece,Guatemala, Jamaica, the Republic of Korea, Malaysia, Mexico, Paraguay, The Philippines, Taiwan, Thailand,Tunisia, Turkey, Uruguay, and Venezuela), and 14 high-income economies (Australia, Canada, Denmark,Finland, Germany, Italy, Japan, The Netherlands, Singapore, Spain, Sweden, the United Kingdom, and theUnited States) data permitting. In 1990, low-income economies had an average GDP per capita of $490;middle-income economies, $2,740; and high-income economies, $20,457.(2) Non-bank financial institutions include insurance companies, pension funds, mutual funds, brokeragehouses, and investment banks.(3) Financial depth is measured by currency held outside financial institutions plus demand deposits andinterest-bearing liabilities of banks and nonbank financial intermediaries.(4) For stock market trading as a percentage of GDP, Taiwan is omitted because its trading/GDP ratio in1990 was almost ten times larger than the next highest trading/GDP ratio (Singapore). With Taiwanincluded, the middle-income stock trading ratio becomes 37.3 percent.

CentralBank Assets

CommercialBank Assets

NonbankAssets

Stock MarketCapitalization

Stock MarketTrading

Low Income Middle Income High Income




time or as one shifts from poor to richercountries,

(1) financial intermediaries get largeras measured by the total assets orliabilities of financial intermediar-

ies relative to GDP;(2) banks grow relative to the central

bank in allocating credit;(3) non-banks—such as insurance

companies, investment banks, fi-nance companies, and private pen-sion funds—grow in importance;and

(4) stock markets become larger, asmeasured by market capitalization

relative to GDP, and more liquid,as measured by trading relative toGDP, market capitalization, andstock price variability.

While these “patterns” pose a chal-lenge to financial theorists, they must betreated cautiously because the data suf-fer from numerous problems. For exam-ple, it is difficult to distinguish privatefrom public banks and development

banks from commercial banks in many countries. Similarly, the definition of abank and of a non-bank are not alwaysconsistent across countries. Further-more, there is nothing causal about theserelationships. These patterns alone donot suggest that poor countries can ac-celerate their growth rates by changingthe structure of their financial systems.Finally, many differences exist acrosscountries at similar stages of economicdevelopment (World Bank 1989). For ex-ample, the assets of deposit banks com-posed 56 percent of financial system as-sets in France, while the comparablenumber in the United Kingdom was 35percent. The assets of contractual sav-ings institutions composed 26 percentof total financial system assets in theUnited Kingdom, while in France thefigure was only 7 percent in 1985. Thus,

while there is a general trend involving

financial structure and the level of GDPper capita, there are important excep-tions and differences within incomegroups. While one must be hesitant indrawing conclusions about patterns of fi-

nancial development, an even greaterdegree of hesitancy is called for in link-ing financial structure to economicgrowth.

G. Financial Structure and EconomicGrowth

There exists considerable debate, withsparse evidence and insufficient theory,about the relationship between financialstructure and economic growth. Afterbriefly outlining the major examplesused in discussions of financial structure,I describe the major analytical limita-tions impeding research on financialstructure and economic growth. Theclassic controversy involves the compari-son between Germany and the UnitedKingdom. Starting early in this century,economists argued that differences inthe financial structure of the two coun-

tries help explain Germany’s more rapideconomic growth rate during the latterhalf of the 19th century and the firstdecade of the 20th century (AlexanderGerschenkron 1962). The premise is asfollows. Germany’s bank-based financialsystem, where banks have close ties toindustry, reduces the costs of acquiringinformation about firms. This makes iteasier for the financial system to identify good investments, exert corporate con-trol, and mobilize savings for promisinginvestments than in England’s more se-curities market oriented financial sys-tem, where the ties between banks andindustry are less intimate. Indeed, quitea bit of evidence suggests that Germanbankers were more closely tied to indus-try than British bankers. Unlike En-gland, nearly all German bankers startedas merchants. The evolution from entre-

preneur to banker may explain the com-




paratively close bonds between bankersand industrialists. For example, Germanbankers frequently “mapped out a firm’spaths of growth, conceived farsightedplans, decided on major technological in-

novations, and arranged for mergers andcapital increases” (Gerschenkron 1968,p. 137). Private German bankers also or-ganized and promoted an impressive ar-ray of major manufacturing companiesduring the mid-19th century (RichardTilly 1967, p. 179). Besides this en-trepreneurial role, some evidence sug-gests that German bankers tended to bemore committed to the long-term fund-ing of their clients than English bankers.Short-term credits could be transformedinto longer-term securities more easily in Germany (Tilly 1967, pp. 178–81).Thus, various pieces of evidence suggesta closer relationship between banker andindustrialist in Germany. While bank-industry relationships may have beencloser in Germany, this does not imply that the German financial system wasbetter at risk management, providing

liquidity, or facilitating exchange.Furthermore, economists disagree over whether the growth differential betweenthe two was really very large. AlthoughGerman manufacturing production grewnoticeably faster than Britain’s in the sixdecades before World War I, Germany’soverall per capita GNP growth rate was1.55 while the U.K.’s was 1.35 over theperiod 1850 to 1913 (Goldsmith 1969,pp. 406–07). Thus, aggregate growth dif-ferences are not very large, the signifi-cant differences that do exist are indus-try specific, and other factors besidesdifferences in financial structure may ex-plain industry specific growth differen-tials over this period.

The debate concerning bank-based versus market-based systems eventually expanded to include comparisons withthe United States. German banks are

larger as a share of GDP than U.S. banks

and German bankers tend to be more in-tricately involved in the management of industry than U.S. bankers (RandallPozdena and Volbert Alexander 1992;Franklin Allen and Gale 1995; and

Demirguç-Kunt and Levine 1996a). Fur-thermore, historical evidence suggeststhat German universal banks were moreefficient (lower cost of capital) than U.S.banks over the 1870–1914 period andsuffered less systemic problems than theU.S. banking system (Calomiris 1995). Incontrast, the U.S. financial system istypically characterized as having a com-paratively larger, more active securitiesmarkets with more equities held by households. These observations suggestthat the German bank-based system may reduce information asymmetries andthereby allow banks to allocate capitalmore efficiently and to exert corporatecontrol more effectively. In contrast, theUnited States’ securities market-basedfinancial system may offer advantages interms of boosting risk sharing opportuni-ties (Allen and Gale 1995). While this

functional approach highlights the relevant issues, substantially more researchis needed before drawing conclusionsabout the dominance of one financialstructure over another.30

Many of the arguments involvingbank-based versus securities market-based financial systems have been usedto compare Japan and the United States.For example, research suggests thatJapanese bankers are more closely tiedto industrial clients than U.S. bankers.This closer connection may mitigate in-formation asymmetries (Hoshi, Kasyap,and Sharfstein 1990), which may fosterbetter investment and faster growth.Thus, the structure of the Japanese fi-nancial system is sometimes viewed assuperior to the financial structure of the

30 Park (1993) compares the structure and func-tioning of the financial systems of Korea and Tai-

wan in relation to their industrial composition.




United States and an important factor inJapan’s faster growth rate over the lastfour decades. Interestingly, however, therecent banking problems and slowergrowth in Japan have led some to argue

that the absence of a credible takeoverthreat through efficient stock marketshas impeded proper corporate gover-nance and competitiveness. These con-flicting analyses highlight the need forbetter empirical measures of financialstructure and the functions provided by financial systems.

There are severe analytical problems with linking financial structure to eco-nomic performance. First, existing re-search on financial structure does notquantify the structure of financial sys-tems or how well different financialsystems function overall. For example,German bankers may have been moreclosely connected to industrialists thantheir British counterparts, but less capa-ble at providing liquidity and facilitatingtransactions. Similarly, while JapaneseKeiretsu may lower information acquisi-

tion costs between banks and firms, thisdoes not necessarily imply that the Japa-nese financial system provides greaterrisk sharing mechanisms or more accu-rately spot promising new lines of busi-ness. Furthermore, while Japan is some-times viewed as a bank-based system, ithas one of the best developed stock mar-kets in the world (Demirguç-Kunt andLevine 1996a). Thus, the lack of quanti-tative measures of financial structureand the functioning of financial systemsmake it difficult to compare financialstructures.

Second, given the array of factors in-fluencing growth in Germany, Japan, theUnited Kingdom, and the United States,it is analytically difficult—and perhapsreckless—to attribute differences ingrowth rates to differences in the finan-cial sector. Moreover, over the post

World War II period, the devastated

Axis powers may simply have been con- verging to the income levels of theUnited States, such that observedgrowth rate differentials have little to do with financial structure. Thus, before

linking financial structure with economicgrowth, researchers need to control forother factors influencing long-rungrowth.

A third factor that complicates theanalysis of financial structure and eco-nomic growth is more fundamental. Thecurrent debate focuses on bank-basedsystems versus market-based systems.Some aggregate and firm level evidence,however, suggest that this dichotomy isinappropriate. The data indicate thatboth stock market liquidity—as meas-ured by stock trading relative to GDPand market capitalization—and the levelof banking development—as measuredby bank credits to private firms dividedby GDP predict economic growth oversubsequent decades (Levine and Zervos1996). Thus, it is not banks or stock mar-kets; bank and stock market develop-

ment indicators both predict economicgrowth. Perhaps, the debate should notfocus on bank-based versus market-based systems because these two compo-nents of the financial system enter thegrowth regression significantly and pre-dict future economic growth. It may bethat stock markets provide a differentbundle of financial functions from thoseprovided by financial intermediaries. Forexample, stock markets may primarily offer vehicles for trading risk and boost-ing liquidity. In contrast, banks may fo-cus on ameliorating information acquisi-tion costs and enhancing corporategovernance of major corporations. Thisis merely a conjecture, however. Thereare important overlaps between the services provided by banks and stock mar-kets. As noted above, well-functioningstock markets may ameliorate informa-

tion acquisition costs, and banks may




provide instruments for diversifying riskand enhancing liquidity. Thus, to under-stand the relationship between financialstructure and economic growth, we needtheories of the simultaneous emergence

of stock markets and banks and we needempirical proxies of the functions per-formed by the different components of financial systems.

A fourth factor limiting our under-standing of the links between financialstructure and economic growth is thatresearchers have focused on a few indus-trialized countries due to data limita-tions. The United States, Germany, Ja-pan, and the United Kingdom havebasically the same standard of living. Av-eraged over a sufficiently long time pe-riod, they must also have very similargrowth rates. Thus, comparisons of fi-nancial structure and economic develop-ment using only these countries will tendto suggest that financial structure is un-related to the level and growth rate of economic development. Future studies will need to incorporate a more diverse

selection of countries to have even achance of identifying patterns betweenfinancial structure and economic devel-opment.

Finally, there are important interac-tions between stock markets and banksduring economic development that havenot been the focus of bank-based versusmarket-based comparisons. As noted,greater stock market liquidity is associ-ated with faster rates of capital forma-tion. Nonetheless, new equity sales donot finance much of this new investment(Colin Mayer 1988), though importantdifferences exist across countries (AjitSingh and Javed Hamid 1992). Most newcorporate investment is financed by retained earnings and debt. This raisesa quandary: stock market liquidity ispositively associated with investment,but equity sales do not finance much of

this investment. This quandary is con-

firmed by firm-level studies. In relatively poor countries, enhanced stock marketliquidity actually tends to boost corpo-rate debt-equity ratios; stock market li-quidity does not induce a substitution

out of debt and into equity finance(Demirguç-Kunt and Maksimovic 1996a).However, for industrialized countries,debt-equity ratios fall as stock market li-quidity rises; stock market liquidity in-duces a substitution out of debt finance.The evidence suggests complex interac-tions between the functioning of stockmarkets and corporate decisions to bor-row from banks that depend on the over-all level of economic development. Thus, we need considerably more researchinto the links among stock markets,banks, and corporate financing deci-sions to understand the relationship be-tween financial structure and economicgrowth.

IV . Conclusions

Since Goldsmith (1969) documented

the relationship between financial andeconomic development 30 years ago, theprofession has made important progress.Rigorous theoretical work carefully illu-minates many of the channels through which the emergence of financial mar-kets and institutions affect—and are af-fected by—economic development. Agrowing body of empirical analyses, in-cluding firm-level studies, industry-levelstudies, individual country-studies, andbroad cross country comparisons, dem-onstrate a strong positive link betweenthe functioning of the financial systemand long-run economic growth. Theory and evidence make it difficult to con-clude that the financial system merely—and automatically—responds to industri-alization and economic activity, or thatfinancial development is an inconse-quential addendum to the process of

economic growth. I believe that we will




not have a sufficient understanding of long-run economic growth until we un-derstand the evolution and functioningof financial systems. This conclusionabout financial development and long-

run growth has an important corollary:although financial panics and recessionsare critical issues, the finance-growthlink goes beyond the relationship be-tween finance and shorter-term fluctua-tions.

Undoubtedly, the financial system isshaped by nonfinancial developments.Changes in telecommunications, com-puters, nonfinancial sector policies, insti-tutions, and economic growth itself in-fluence the quality of financial servicesand the structure of the financial system.Technological improvements lowertransaction costs and affect financial ar-rangements (Merton 1992). Monetary and fiscal policies affect the taxation of financial intermediaries and the provi-sion of financial services (Bencivengaand B. Smith 1992; Roubini and Sala-i-Martin 1995). Legal systems affect finan-

cial systems (LaPorta et al. 1996), andpolitical changes and national institu-tions critically influence financial devel-opment (Haber 1991, 1996). Further-more, economic growth alters the willingness of savers and investors to pay the costs associated with participating inthe financial system (Greenwood andJovanovic 1990). While economists havemade important advances, we needmuch more research on financial devel-opment. Why does financial structurechange as countries grow? Why do coun-tries at similar stages of economic devel-opment have different looking financialsystems? Are there long-run economicgrowth advantages to adopting legal andpolicy changes that create one type of fi-nancial structure vis-à-vis another?Much more information about the deter-minants and implications of financial

structure will move us closer to a com-

prehensive view of financial develop-ment and economic growth.

REFERENCES

AGHION, PHILIPPE AND HOWITT, PETER. “AModel of Growth Through Creative Destruc-

tion,” Econometrica, Mar. 1992, 60(2), pp. 323–51.ALLEN, FRANKLIN AND GALE, DOUGLAS. “A

Welfare Comparison of Intermediaries andFinancial Markets in Germany and the US,’’Europ. Econ. Rev., Feb. 1995, 39(2), pp. 179–209.

AMIHUD, YAKOV AND MENDELSON, HAIM . “TheEffects of Beta, Bid-Ask Spread, Residual Risk,and Size on Stock Returns,” J. Finance, June1989, 44(2), pp. 479–86.

ARESTIS, PHILIP AND DEMETRIADES, PANICOS.“ ‘Finance and Growth: Schumpeter Might BeRight’ A Comment on King and Levine.” U. of

East London mimeo, 1995.ARROW, KENNETH J. “The Role of Securities inthe Optimal Allocation of Risk Bearing,” Rev.Econ. Stud., Apr. 1964, (2), pp. 91–96.

ATJE , RAYMOND AND JOVANOVIC, BOYAN. “StockMarkets and Development,” Europ. Econ. Rev.,Apr. 1993, 37 (2/3), pp. 632–40.

BAGEHOT, WALTER. Lombard street. Homewood,IL: Richard D. Irwin, [1873] 1962 Edition.

BENCIVENGA, VALERIE R. AND SMITH, BRUCE D.“Financial Intermediation and EndogenousGrowth,” Rev. Econ. Stud., Apr. 1991, 58(2),pp. 195–209.

———. “Deficits, Inflation, and the Banking Sys-

tem in Developing Countries: The Optimal De-gree of Financial Repression,” Oxford Econ.Pap., Oct. 1992, 44(4), pp. 767–90.

———. “Some Consequences of Credit Rationingin an Endogenous Growth Model,” J. Econ. Dy- nam. Control, Jan.-Mar. 1993, 17 (1/2), pp. 97–122.

BENCIVENGA, VALERIE R.; SMITH, BRUCE D.AND STARR, ROSS M. “Transactions Costs,Technological Choice, and EndogenousGrowth,” J. Econ. Theory, Oct. 1995, 67 (1), pp.153–77.

BENNETT, ROBERT L. “Financial Innovation andStructural Change in the Early Stages of Indus-

trialization: Mexico, 1945–59,” J. Finance, Dec.1963, 18(4), pp. 666–83.BERGER, ALLEN N. AND UDELL, GREGORY F.

“Relationship Lending and Lines of Credit inSmall Firm Finance,” J. Bus., July 1995, 68(3),pp. 351–81.

BERNANKE, BEN AND GERTLER, MARK . “Agency Costs, Net Worth, and Business Fluctuations,” Amer. Econ. Rev., Mar. 1989, 79(1), pp. 14–.

———. “Financial Fragility and Economic Perfor-mance,” Quart. J. Econ., Feb. 1990, 105(1), pp.87–114.

BERTHELEMY , JEAN-CLAUDE AND VAROUDAKIS,ARISTOMENE. “Economic Growth, Conver-

gence Clubs, and the Role of Financial Devel-




opment,” Oxford Econ. Pap., Apr. 1996, 48(2),pp. 300–28.

BHIDE, AMAR . “The Hidden Costs of Stock Mar-ket Liquidity,” J. Finan. Econ., Aug. 1993,34(1), pp. –51.

BOYD, JOHN H. AND PRESCOTT, EDWARD C. “Fi-nancial Intermediary-Coalitions,” J. Econ. The-

ory , Apr. 1986, 38(2), pp. 211–32.BOYD, JOHN H. AND SMITH, BRUCE D. “HowGood Are Standard Debt Contracts? Stochastic Versus Nonstochastic Monitoring in a Costly State Verification Environment,” J. Bus., Oct.1994, 67 (4), pp. 539–61.

———. “The Co-Evolution of the Real and Finan-cial Sectors in the Growth Process,” WorldBank Econ. Rev., May 1996, 10(2), pp. 371–96.

BROCK, PHILIP L., ed. If Texas were Chile: A primer on banking reform. San Francisco, CA:ICS Press, 1992.

BRUNNER, KARL AND MELTZER, ALLEN H. “TheUses of Money: Money in the Theory of an Ex-

change Economy,” Amer. Econ. Rev., Dec.1971, 61(5), pp. 784–805.CALOMIRIS, CHARLES W. “The Costs of Rejecting

Universal Banking: American Finance in theGerman Mirror, 1870–1914,” in Coordinationand information: Historical perspectives on theorganization of enterprise. Eds.: NAOMI R.LAMOREAUX AND DANIEL M. G. RAFF . Chi-cago: U. of Chicago Press, 1995, pp. 257–321.

CALOMIRIS, CHARLES W.; HIMMELBERG,CHARLES P. AND WACHTEL, PAUL . “Commer-cial Paper, Corporate Finance, and the Busi-ness Cycle: A Microeconomic Perspective,”Carnegie-Rochester Confer. Ser. Pub. Policy,

June 1995, 42, pp. 203–50.CALOMIRIS, CHARLES W. AND HUBBARD, R.GLENN. “Internal Finance and Investment:Evidence from the Undistributed Profits Tax of 1936–37,” J. Bus., Oct. 1995, 68(4), pp. 443–82.

CAMERON, RONDO. “Scotland, 1750–1845,” 1967ain RONDO CAMERON ET AL. 1967, pp. 60–99.

———. “Conclusion,” 1967b in RONDO CAMERONET AL. 1967, pp. 290–321.

CAMERON, RONDO ET AL. Banking in the early stages of undustrialization: A study in comparative economic history. New York: Oxford U.Press, 1967.

CAPRIO, GERARD, JR.; ATIYAS, IZAK AND HAN-

SON , JAMES A. Financial reform: Theory and experience. New York: Cambridge U. Press, 1994.

CAROSSO, VINCENT. Investment banking in America. Cambridge, MA: Harvard U. Press,1970.

CHANDAVARKAR, ANAND. “Of Finance and Devel-opment: Neglected and Unsettled Questions,”World Devel., Jan. 1992, 20(1), pp. 133–42.

CHECKLAND, SYDNEY G. Scottish banking: A history, 1695–1973. Glasgow: Collins, 1975.

COLE , DAVID C. AND PARK, YUNG CHUL . Finan-cial development of Korea 1945–1978. Cam-bridge, MA: Harvard U. Press, 1983.

COLE , DAVID C. AND SLADE, BETTY F. “Reform

of Financial Systems,” in Reforming economic

systems in developing countries. Eds.: D WIGHTPERKINS AND MICHAEL ROEMER. Cambridge,MA: Harvard U. Press, 1991, pp. 3–40.

COOLEY, THOMAS F. AND SMITH, BRUCE D. “Fi-nancial Markets, Specialization, and Learningby Doing.” Cornell U. mimeo, 1992.

COWEN, TYLER AND KROSZNER, RANDALL. “Scot-

tish Banking Before 1845: A Model for Laissez-Faire?” J. Money, Credit, Banking, May 1989, 21(2), p. 221–.

DAVIS, LANCE E. AND HUTTENBACK, ROBERT A.Mammon and the pursuit of empire: The politi-cal economy of British imperlialism, 1869–1912.Cambridge: Cambridge U. Press, 1986.

DEBREU, GERARD. Theory of value. New York: Wiley, 1959.

DE GREGORIO, JOSE . “Borrowing Constraints,Human Capital Accumulation, and Growth,” J.Monet. Econ., Feb. 1996, 37 (1), pp. 49–71.

DE LONG, J. BRADFORD. “Did J. P. Morgan’sMen Add Value? An Economist’s Perspective

on Financial Capitalism,” in Inside the businessenterprise: Historical perspectives on the use of information. Ed. PETER TEMIN. Chicago: U. of Chicago Press, 1991, pp. 205–36.

DE LONG , J. BRADFORD ET AL. “The Size andIncidence of the Losses from Noise Trading,” J.Finance, July 1989, 44(3), pp. 681–96.

DEMIRGÜÇS-KUNT , ASLI AND LEVINE, ROSS .“Stock Market Development and Financial In-termediaries: Stylized Facts,” World BankEcon. Rev., May 1996a, 10(2), 291–322.

———. “Stock Markets, Corporate Finance, andEconomic Growth: An Overview,” World BankEcon. Rev., May 1996b, 10(2), 223–40.

DEMIRGÜÇS-KUNT , ASLI AND MAKSIMOVIC, VO-JISLAV. “Stock Market Development and Fi-nancing Choices of Firms,” World Bank Econ.Rev., May 1996a 10(2), pp. 341–70.

———. “Financial Constraints, Uses of Funds,and Firm Growth: An International Compari-son.” World Bank mimeo, 1996b.

DEVEREUX, MICHAEL B. AND SMITH, GREGOR W. “International Risk Sharing and EconomicGrowth,” Int. Econ. Rev., Aug. 1994, 35(4), pp.535–50.

DIAMOND, DOUGLAS W. “Financial Intermedia-tion and Delegated Monitoring,” Rev. Econ.Stud., July 1984, 51(3), pp. 393–414.

DIAMOND, DOUGLAS W. AND DYBVIG, PHILIP H.“Bank Runs, Deposit Insurance, and Liquidity,” J. Polit. Econ., June 1983, 91(3), pp. 401–19.

DIAMOND, DOUGLAS W. AND VERRECCHIA,ROBERT E. “Optimal Managerial Contracts andEquilibrium Security Prices,” J. Finance, May 1982, 37 , pp. 275–87.

EASTERLY, WILLIAM. “How Much Do DistortionsAffect Growth?” J. Monet. Econ., Nov. 1993,32(4), p. 187–212,

ENGERMAN, STANLEY L. AND SOKOLOFF, KEN-NETH L. “Factor Endowments, Institutions, andDifferential Paths of Growth Among New World Economies: A View from Economic His-

torians of the United States,” in How Latin




America fell behind. Ed.: STEPHEN HABER.Stanford, CA: Stanford U. Press, 1996, pp. 260–304.

FAZZARI, STEVEN M.; HUBBARD, R. GLENN ANDPETERSEN, BRUCE C. “Financing Constraints,and Investment,”Brookings Pap. Econ. Act.,Aug. 1988, 1, pp. 141–206.

FRY, MAXWELL J. Money, interest, and banking ineconomic development. Baltimore: John Hop-kins U. Press, 1995.

GALE , DOUGLAS AND HELLWIG, MARTIN. “In-centive-Compatible Debt Contracts: The One-Period Problem,” Rev. Econ. Stud., Oct. 1985,52(4), pp. 647–63.

GELB , ALAN H. “Financial Policies, Growth, andEfficiency.” World Bank PPR Working PaperNo. 202, June 1989.

GERSCHENKRON , ALEXANDER. Economic back- wardness in historical perspective—A book of essays. Cambridge: Harvard U. Press, 1962.

———. Continuity in history and other essays.

Cambridge, MA: Harvard U. Press, 1968.GERTLER, MARK . “Financial Structure and Aggre-gate Economic Activity: An Overview,” J.Money, Credit, Banking, Aug. 1988, 20(3, Pt.2), pp. 559–88.

GERTLER, MARK AND GILCHRIST, SIMON.“Monetary Policy, Business Cycles, and the Be-havior of Small Manufacturing Firms,” Quart. J.Econ., May 1994, 109(2), pp. 309–40.

GERTLER, MARK AND ROSE , ANDREW. “Finance,Public Policy and Growth,” in GERARD CAPRIO,JR., IZAK ATILJAS, AND JAMES A. HANSON.1994, pp. 13–45.

GIOVANNINI, ALBERTO AND DE MELO , MARTHA.

“Government Revenue from Financial Repres-sion,” Amer. Econ. Rev.., Sept. 1993, 83(4), pp.953–63.

GOLDSMITH, RAYMOND, W. Financial structureand development. New Haven, CT: Yale U.Press, 1969.

GORTON, GARY AND PENNACCHI, GEORGE. “Fi-nancial Intermediaries and Liquidity Creation,” J. Finance, Mar. 1990, 45(1), pp. 49–71.

GREENWOOD, JEREMY AND JOVANOVIC, BOYAN.“Financial Development, Growth, and the Dis-tribution of Income,” J. Polit. Econ., Oct. 1990,98(5,Pt.1), pp. 1076–1107.

GREENWOOD, JEREMY AND SMITH, BRUCE. “Fi-

nancial Markets in Development, and the De- velopment of Financial Markets,” J. Econ. Dy- namics and Control, Jan. 1997, 21(1), pp.145–81.

GROSSMAN, GENE M. AND HELPMAN, ELHANAN.“Quality Ladders in the Theory of Growth,”Rev. Econ. Stud., Jan. 1991, 58(1), pp. 43–61.

GROSSMAN, SANFORD J. AND MILLER, MERTONH. “Liquidity and Market Structure,” J. Fi- nance, July 1988, 43(3), pp. 617–33.

GROSSMAN, SANFORD J. AND STIGLITZ, JOSEPHE. “On the Impossibility of Informationally Ef-ficient Markets,” Amer. Econ. Rev., June 1980,70(3), pp. 393–408.

GUESLIN, ANDRE. “Banks and State in France

from the 1880s to the 1930s: The ImpossibleAdvance of the Banks,” in Finance and finan-ciers in European history 1880–1960. Ed.YOUSSEF CASSIS. New York: Cambridge U.Press, 1992, pp. 63–92.

GURLEY, JOHN G. AND SHAW , EDWARD S. “Fi-nancial Aspects of Economic Development,”

Amer. Econ. Rev., Sept. 1955, 45(4), pp. 515–38.HABER, STEPHEN H. “Industrial Concentration

and the Capital Markets: A Comparative Study of Brazil, Mexico, and the United States, 1830–1930,” J. Econ. Hist., Sept. 1991, 51(3), 559–80.

———. “Capital Immobilities and Industrial De- velopment: A Comparative Study of Brazil,Mexico, and the United States, 1840–1930.”Stanford U. mimeo, 1996.

HARRIS, JOHN R.; SCHIANTARELLI, FABIO ANDSIREGAR, MIRANDA G. “The Effect of FinancialLiberalization on the Capital Structure and In- vestment Decisions of Indonesian Manufactur-

ing Establishments,” World Bank Econ. Rev.,Jan. 1994, 8(1), pp. 17–47.HICKS, JOHN . A theory of economic history. Ox-

ford: Clarendon Press, 1969.HOLMSTROM, BENGT AND TIROLE, JEAN . “Mar-

ket Liquidity and Performance Monitoring,” J.Polit. Econ., Aug. 1993, 101(4), pp. 678–709.

HOSHI, TAKEO; KASHYAP, ANIL K. ANDSHARFSTEIN, DAVID S. “Bank Monitoring andInvestment: Evidence from the ChangingStructure of Japanese Corporate Banking Rela-tionships,” in Asymmetric information, corpo-rate finance, and investment. Ed.: R.GLENNHUBBARD. Chicago: U. of Chicago Press, 1990,

pp. 105–26.HUBBARD, R. GLENN; KASHYAP, ANIL K. AND WHITED, TONI M. “International Finance andFirm Investment,” J. Money Credit Banking,Aug. 1995, 27 (3), pp. 683–701.

INTERNATIONAL FINANCE CORPORATION. Emerg- ing markets data base. Washington, DC: Vari-ous years.

JACKLIN, CHARLES J. “Demand Deposits, TradingRestrictions, and Risk Sharing,” in Contractualarrangements for intertemporal trade. Eds.:EDWARD D. PRESCOTT AND NEIL WALLACE.Minneapolis: U. of Minnesota Press, 1987, pp.26–47.

JAFFEE, D WIGHT M. AND RUSSELL, THOMAS.“Imperfect Information, Uncertainty, andCredit Rationing.” Quart. J. Econ., Nov. 1976,90(4), pp. 651–66.

JAMES, CHRISTOPHER. “Some Evidence on theUniqueness of Bank Loans,” J. Finan. Econ.,Dec. 1987, 19(2), pp. 217–36.

JAMES, CHRISTOPHER AND WIER, PEGGY. “Bor-rowing Relationships, Intermediation, and theCost of Issuing Public Securities,” J. Finan.Econ., Nov.-Dec. 1990, 28(1/2), pp. 149–71.

JAPPELLI, TULLIO AND PAGANO, MARCO. “Saving,Growth, and Liquidity Constraints,” Quart. J.Econ., Feb. 1994, 109(1), pp. 83–109.

JARAMILLO, FIDEL; SCHIANTARELLI, FABIO AND




WEISS, ANDREW. “Capital Market Imperfec-tions, Financial Constraints and Investment:Econometric Evidence from Panel Data forEcuador,” J. Devel. Econ., Dec. 1996, 51(2), pp.367–86.

JAYARATNE, JITH AND STRAHAN, PHILIP E. “TheFinance-Growth Nexus: Evidence from Bank

Branch Deregulation,” Quart. J. Econ., Aug.1996, 111(3), pp. 639–70.JENSEN, MICHAEL AND MECKLING, WILLIAM R.

“Theory of the Firm: Managerial Behavior,Agency Costs, and Ownership Structure,” J. Fi- nan. Econ., Oct. 1976, 3(4), pp. 305–60.

JENSEN, MICHAEL C. AND MURPHY, KEVIN J.“Performance Pay and Top-Management Incen-tives, J. Polit. Econ., Apr. 1990, 98(2), pp. 225–64.

JEVONS, W. STANLEY. Money and the mechanismof exchange. London: Appleton, 1875.

JUNG , WOO S. “Financial Development and Eco-nomic Growth: International Evidence,” Econ.

Devel. Cult. Change, Jan. 1986, 34(2), pp. 333–46.KADLEC, GREGORY B. AND MCCONNELL, JOHN

J. “The Effect of Market Segmentation and Illi-quidity on Asset Prices: Evidence from Ex-change Listings,” J. Finance, June 1994, 49(2),pp. 611–36.

KERR , ANDREW W. History of banking in Scot-land. Glasgow, 1884; 3rd ed. London: A&CBlack, 1918.

KHAN , AUBHIK. “Financial Development and Eco-nomic Growth.” U. of Virginia mimeo, 1994.

KING , ROBERT G. AND LEVINE, ROSS . “FinancialIntermediation and Economic Development,”

in Financial intermediation in the constructionof Europe. Eds.: COLIN MAYER AND XAVIER VIVES. London: Centre for Economic Policy Research, 1993a, pp. 156–89.

———. “Finance and Growth: Schumpeter MightBe Right,” Quart. J. Econ., Aug. 1993b, 108(3),pp. 717–37.

———. “Finance, Entrepreneurship, and Growth:Theory and Evidence,” J. Monet. Econ., Dec.1993c, 32(3), pp. 513–42.

KING , ROBERT G. AND PLOSSER, CHARLES I.“Money as the Mechanism of Exchange,” J.Monet. Econ., Jan. 1986, 17 (1), pp. 93–115.

KRASA, STEFAN AND VILLAMIL, ANNE P. “Moni-

toring the Monitor: An Incentive Structure fora Financial Intermediary,” J. Econ. Theory,June 1992, 57 (1), pp. 197–221.

KYLE , ALBERT S. “Market Structure, Information,Futures Markets, and Price Formation,” inInternational agricultural trade: Advancedreadings in price formation, market structure,and price instability. Eds.: GARY G. STOREY,ANDREW SCHMITZ, AND ALEXANDER H.SARRIS. Boulder, CO: Westview, 1984, pp. 45–64.

LAMOREAUX, NAOMI R. Insider lending: Banks, personal connections, and economic development in industrial New England, New York:

Cambridge U. Press, 1994.

LAMOREAUX, NAOMI AND SOKOLOFF, KENNETHL. “Long-Term Change in the Organization of Inventive Activity,” Proceedings of the National Academy of Science, Nov. 1996, 93, pp. 12686–92.

LAPORTA, RAFAEL ET AL. “Law and Finance.” Na-tional Bureau of Economic Research Working

Paper No. 5661. July 1996.LEVHARI, DAVID AND SRINIVASAN, T. N. “Opti-mal Savings Under Uncertainty,” Rev. Econ.Stud., Apr. 1969, 36(1), pp. 153–63.

LEVINE, ROSS . “Stock Markets, Growth, and TaxPolicy,” J. Finance, Sept. 1991, 46(4), pp. 1445–65.

———. “Law, Finance, Economic Growth.” U. of Virginia mimeo, 1997.

LEVINE, ROSS AND RENELT, DAVID. “A Sensitiv-ity Analysis of Cross-Country Growth Regres-sions,” Amer. Econ. Rev., Sept. 1992, 82(4), pp.942–63.

LEVINE, ROSS AND ZERVOS, SARA . “Stock Mar-

kets, Banks, and Economic Growth.” WorldBank Policy Research Working Paper, No.1690, Dec. 1996.

LUCAS, ROBERT E., JR. “On the Mechanics of Economic Development,” J. Monet. Econ., July 1988, 22(1), pp. 3–42.

LUMMER, SCOTT L. AND MCCONNELL, JOHN J.“Further Evidence on the Bank Lending Pro-cess and the Capital Market Response to BankLoan Agreements,” J. Finan. Econ., Nov. 1989, 25(1), pp. 99–122.

MAYER, COLIN. “New Issues in Corporate Fi-nance,” Europ. Econ. Rev., June 1988, 32(5),pp. 1167–83.

MCKINNON, RONALD I. Money and capital in economic development. Washington, DC: Brook-ings Institution, 1973.

MEIER, GERALD M. AND SEERS, DUDLEY. Pio- neers in development. New York: Oxford U.Press, 1984.

MERTON, ROBERT C. “A Simple Model of CapitalMarket Equilibrium with Incomplete In- for-mation,” J. Finance, July 1987, 42(3), pp. 483–510.

———. “Financial Innovation and Economic Per-formance,” J. Applied Corporate Finance, Win-ter 1992, 4(4), pp. 12–22.

MERTON, ROBERT C. AND BODIE, Z VI. “A Con-

ceptual Framework for Analyzing the FinancialEnvironment,” in The global financial system: A functional perspective. Eds.: D WIGHT B.CRANE ET AL. Boston, MA: Harvard BusinessSchool Press, 1995, pp. 3–.

MORCK, RANDALL; SHLEIFER, ANDREI AND VISHNY, ROBERT W. “Do Managerial Objec-tives Drive Bad Acquisitions?” J. Finance, Mar.1990, 45(1), pp. –48

MYERS, STEWART C. AND MAJLUF, NICHOLAS S.“Corporate Financing and Investment Deci-sions When Firms Have Information That In- vestors Do Not Have,” J. Finan. Econ., June1984, 13(2), pp. 187–221.

NEAL , LARRY. The rise of financial capitalism: In-




ternational capital markets in the age of reason.New York: Cambridge U. Press, 1990.

NEUSSER, KLAUS AND KUGLER, MAURICE.“Manufacturing Growth and Financial Develop-ment: Evidence from OECD Countries.” U. of Berne mimeo, 1996.

NORTH, DOUGLASS C. Structure and change in

economic history. New York: Norton, 1981.OBSTFELD, MAURICE. “Risk-Taking, Global Di- versification, and Growth,” Amer. Econ. Rev.,Dec. 1994, 84(5), pp. 10–29.

OSTROY, JOSEPH M. AND STARR, ROSS M. “TheTransactions Role of Money,” in, Handbook of monetary economics. Vol. 1. Eds.: BENJAMINFRIEDMAN AND FRANK H. HAHN . Amsterdam:North-Holland, 1990, pp. 3–62.

PAGANO, MARCO. “Financial Markets andGrowth: An Overview,” Europ. Econ. Rev., Apr.1993, 37 (2–3), pp. 613–22.

PARK , YUNG CHUL . “The Role of Finance in Eco-nomic Development in South Korea and Tai-

wan,” in Finance and development: Issues andexperience. Ed.: ALBERTO GIOVANNINI.Cambridge: Cambridge U. Press, 1993, pp.121–50.

PATRICK, HUGH T. “Financial Development andEconomic Growth in Underdeveloped Coun-tries,” Econ. Devel. Cult. Change, Jan. 1966,14(2), pp. 174–89.

PATRICK, HUGH T. AND PARK , YUNG CHUL . Fi- nancial development in Japan, Korea, and Tai- wan. New York: Oxford U. Press, 1994.

PENSO DE LA VEGA , JOSEF. Confusion de confusiones, 1688. Boston: Baker Library, HarvardGraduate School of Business, 1957.

PETERSEN, MITCHELL A. AND RAJAN, RAGHU-RAM G. “The Benefits of Lending Relation-ships: Evidence from Small Business Data,” J.Finance, Mar. 1994, 49(1), pp. 3–37.

POLLARD, SIDNEY AND ZIEGLER, DIETER. “Bank-ing and Industrialization: Rondo CameronTwenty Years On,” in Finance and financiers inEuropean history 1880–1960. Ed.: YOUSEFCASSIS. New York: Cambridge U. Press, 1992,pp. 17–38.

POZDENA, RANDALL J. AND VOLBERT, ALEXAN-DER . “Bank Structure in West Germany,” inBanking structures in major countries. Ed.:GEORGE G. KAUFMAN. Norwell, MA and Dor-

drecht: Kluwer, 1992, pp. 555–90.RAJAN, RAGHURAM G. “Insiders and Outsiders:The Choice Between Informed and ArmslengthDebt,” J. Finance, Sept. 1992, 47 (4), pp. 1367–1400.

RAJAN, RAGHURAM G. AND ZINGALES, LUIGI.“Financial Dependence and Growth.” U. of Chicago mimeo, May 1996.

REBELO, SERGIO T. “Long-Run Policy Analysisand Long-Run Growth,” J. Polit. Econ., June1991, 99(3), pp. 500–21.

ROBINSON, JOAN . “The Generalization of theGeneral Theory,” in The rate of interest, andother essays. London: Macmillan, 1952, pp. 67–

142.

ROMER, PAUL M. “Increasing Returns and Long-Run Growth,” J. Polit. Econ., Oct. 1986, 94(5),pp. 1002–37.

———. “Endogenous Technological Change,” J.Polit. Econ., Oct. 1990, 98(5, Pt. 2), pp. S71–102.

ROUBINI, NOURIEL AND SALA-I-MARTIN,

XAVIER. “Financial Repression and EconomicGrowth,” J. Devel. Econ., July 1992, 39(1), pp.5–30.

———. “A Growth Model of Inflation, Tax Eva-sion, and Financial Repression,” J. Monet.Econ., Apr. 1995, 35(2), pp. 275–301.

SAINT-PAUL , GILLES. “Technological Choice, Fi-nancial Markets and Economic Development,”Europ. Econ. Rev., May 1992, 36(4), pp. 763–81.

SCHARFSTEIN, DAVID. “The Disciplinary Role of Takeovers,” Rev. Econ. Stud., Apr. 1988, 55(2),pp. 185–99.

SCHIANTARELLI, FABIO. “Financial Constraints

and Investment: A Critical Rev. MethodologicalIssues and International Evidence,” in Is banklending important for the transmission of monetary policy? Eds.: JOE PEEK AND ERIC S.ROSENGREN. Boston, MA: Federal ReserveBank of Boston Conference Series No. 39, June1995, pp. 177–214.

SCHIANTARELLI, FABIO AND SEMBENELLI, ALES-SANDRO. “Form of Ownership and FinancialConstraints: Panel Data Evidence FromLeverage and Investment Equations.” WorldBank Policy Research Working Paper No. 1629,1996.

SCHUMPETER, JOSEPH A. Theorie der Wirt-

schaftlichen Entwicklung [The theory of economic development].Leipzig: Dunker & Hum-blot, 1912; translated by REDVERS OPIE .Cambridge, MA: Harvard U. Press, 1934.

SHARPE, STEVEN A. “Asymmetric Information,Bank Lending, and Implicit Contracts: A Styl-ized Model of Customer Relationships,” J. Fi- nance, Sept. 1990, 45(4), pp. 1069–87.

SHLEIFER, ANDREI AND SUMMERS, LAWRENCEH. “Breach of Trust in Hostile Takeovers,” inCorporate takeovers: Causes and consequences.Ed.: ALAN J. AUERBACH. Chicago: U. of Chi-cago Press, 1988, pp. 33–56.

SHLEIFER, ANDREI AND VISHY, ROBERT W.

“Large Shareholders and Corporate Control,” J.Polit. Econ., June 1986, 94(3), pp. 461–88.

———. “A Survey of Corporate Governance,” J.Finance, forthcoming.

SINGH, AJIT AND HAMID, JAVED. “Corporate Fi-nancial Structures in Developing Countries.” Washington, DC: International Finance Corpo-ration Technical Paper 1, 1992.

SIRRI, ERIK R. AND TUFANO, PETER. “The Eco-nomics of Pooling,” in The global financial system: A functional perspective. Eds.: D WIGHT B.CRANE ET AL. Boston, MA: Harvard BusinessSchool Press, 1995, pp. 81–128.

SLOVIN, MYRON B.; SUSHKA, MARIE E. AND

POLONCHEK, JOHN A. “The Value of Bank Du-




rability: Borrowers as Bank Stakeholders,” J. Fi- nance, Mar. 1993, 48(1), pp. 247–66.

SMITH, ADAM . An inquiry into the nature andcauses of the wealth of nations. London: W. Sta-han & T. Cadell, 1776.

STEIN, JEREMY C. “Takeover Threats and Mana-gerial Myopia,” J. Polit. Econ., Feb. 1988, 96(1),

pp. 61–80.STERN, NICHOLAS. “The Economics of Develop-ment: A Survey,” Econ. J., Sept. 1989, 99(397),pp. 597–685.

STIGLITZ, JOSEPH E. “Credit Markets and theControl of Capital,” J. Money, Credit, Banking,May 1985, 17 (2), pp. 133–52.

STIGLITZ, JOSEPH E. AND WEISS, ANDREW.“Credit Rationing in Markets with ImperfectInformation,” Amer. Econ. Rev., June 1981,71(3), pp. 393–410.

———. “Incentive Effects of Terminations: Appli-cations to Credit and Labor Markets,” Amer.Econ. Rev., Dec. 1983, 73(5), pp. 912–27.

VON THADDEN, ERNST-LUDWIG. “Long-TermContracts, Short-Term Investment and Moni-toring,” Rev. Econ. Stud., Oct. 1995, 62(4), pp.557–75.

TILLY, RICHARD. “Germany, 1815–1870,” inBanking in the early stages of industrialization: A study in comparative economic history. Eds.:RONDO CAMERON ET AL. New York: Oxford U.Press, 1967, pp. 151–82.

TOBIN, JAMES. “Money and Economic Growth,”Econometrica, Oct. 1965, 33(4), pp. 671–84.

TOWNSEND, ROBERT M. “Optimal Contracts andCompetitive Markets with Costly State Verifica-tion,” J. Econ. Theory, Oct. 1979, 21(2), pp.

265–93.

TYBOUT, JAMES R. “Credit Rationing and Invest-ment Behavior in a Developing Country,” Rev.Econ. Statist., Nov. 1983, 65(4), pp. 598–607.

WACHTEL, PAUL AND ROUSSEAU, PETER. “Finan-cial Intermediation and Economic Growth: AHistorical Comparison of the U.S., U.K., andCanada,” in Anglo-American financial systems:

Institutions and markets in twentieth-century North America and the United Kingdom. Eds.:MICHAEL D. BORDO AND RICHARD SYLLA.Homewood, IL: Business One Irwin, 1995, pp.329–82.

WELLS, STEPHEN. “Turnover Statistics and TradeReporting,” London Stock Exchange Quarterly with Quality of Markets Rev., Spring 1994, pp.23–26.

WHITED, TONI M. “Debt, Liquidity Constraints,and Corporate Investment: Evidence fromPanel Data,” J. Finance, Sept. 1992, 47 (4), pp.1425–60.

WILLIAMSON, STEPHEN D. “Costly Monitoring,

Financial Intermediation, and EquilibriumCredit Rationing,” J. Monet. Econ., Sept. 1986,18(2), pp. 159–79.

———. “Costly Monitoring, Loan Contracts, andEquilibrium Credit Rationing,”Quart. J. Econ.,Feb. 1987a, 102(1), pp. 135–45.

———. “Financial Intermediation, Business Fail-ures, and Real Business Cycles,” J. Polit. Econ.,Dec. 1987b, 95(6), pp. 1196–1216.

WILLIAMSON, STEVE AND WRIGHT, RANDALL.“Barter and Monetary Exchange Under PrivateInformation,” Amer. Econ. Rev., Mar. 1994,84(1), pp. 104–23.

WORLD BANK . World development report Wash-

ington, DC: Oxford U. Press, 1989.


us federal reserve: levine

Documents