economic growth and concentrated ownership in stock markets

Journal of Economic Behavior & OrganizationVol. 59 (2006) 249–286

Economic growth and concentrated ownershipin stock markets

Gonzalo Castaneda∗

Department of Economics, Universidad de las Americas, Puebla. Ex-hacienda de Sta.Catarina Martir, Cholula, Puebla 72820, Mexico

Received 7 May 2001; received in revised form 13 August 2003; accepted 23 April 2004Available online 12 May 2005

Abstract

This paper presents a theoretical framework to analyze the implications on economic growth ofa stock market that grows but is not well-developed in other aspects (concentrated ownership, lowliquidity, poor legal and judiciary systems, and credit constraints). Family firms are modeled asconsisting of risk-averse owners concerned with keeping the control of their firms while deciding togo public. It is suggested that in this type of economies there may be a non-linear relationship betweenstock market size and economic growth and that, in particular, the formation of an equity market mayretard economic growth when institutions are weak.© 2005 Elsevier B.V. All rights reserved.

JEL classification: G32; O40

Keywords: Economic growth; Stock markets; Ownership concentration

1. Introduction

In the theoretical literature on the development of stock markets and economic growth, apositive link is traditionally presented between these two variables. In particular, three causalfactors are commonly stressed as having a positive impact on efficiency or savings: increasedliquidity, diversification, and improvements in corporate control. Yet the theoretical debate isfar from settled since some authors also point out the negative sides of these three elements.

∗ Tel.: +52 222 229 2064; fax: +52 222 229 2110.E-mail address: [email protected].

0167-2681/$ – see front matter © 2005 Elsevier B.V. All rights reserved.doi:10.1016/j.jebo.2004.04.005

250 G. Castaneda / J. of Economic Behavior & Org. 59 (2006) 249–286

The formation of an equity market is endogenized in the model presented in this paperto explain a non-linear relationship between market size and growth, and thus, it is ableto explain why growth in the stock market might retard economic growth under certaininstitutional environment. It is argued that manager/owners might not be risk-taking enoughso that despite the diversification allowed by acquiring shares in other companies, they mayuse low-return technologies that reduce the risk associated with large-scale production.Moreover, a larger capital formation may not offset the drop in economic growth due to thereduction in productivity since managers may become preoccupied with the acquisition offinancial claims on existing projects to the detriment of the firms’ aggregate savings.

In order to derive the previous results, the model incorporates concerns on the controlof ownership and the existence of credit constraints as limits to the financial diversificationof entrepreneurs. These constraints may induce the selection of a low-return technology asan alternative device for risk reduction. In addition, the model considers how capital allo-cation through the stock market,1 particularly in secondary markets, may have the negativeconsequence of diverting firms’ earnings away from internal capital formation.2 Accord-ingly, the model implicitly considers institutional elements that characterize many emergingeconomies, namely, an inadequate legal and judiciary environment as well as social normsthat inhibit the transparency of economic behavior.3 These features tend to produce familyfirms with concentrated ownership, severe failures in financial markets that exacerbate theproblem of credit rationing, and stock markets with liquidity problems that preclude hostiletakeovers and hamper the informational content of prices.4 In the model, no attempt is madeto formalize the causal link between the institutional features and the stylized facts (controlconcerns, credit constraints, and price inefficiency); the latter are only taken exogenouslyto analyze the impact of firms going public on economic growth under the setting describedabove.5

1 This abstracts from the positive benefits of stock markets. See, for instance, the theoretical model inCho (1986)that asserts that capital markets achieve efficient resource allocation since they are not subject to adverse selectionand moral hazard, as bank credit is.

2 Different studies show that stock markets anomalies are more frequent and robust in emerging markets (see,for instance,Durham, 2000; Claessens, 1995), and thus the allocating capabilities of such markets is put intoquestion.

3 SeeLa Porta et al. (1997)for the legal view on financial development, andDe Jong and Semenov (2002)forthe cultural view on stock market development.

4 In recent years, formal theoretical models have shed some light on the workings and performance of family-firms; see, for instance,Caselli and Gennaioli (2003), Burkart et al. (2003), Chami (2001), andBhattacharya andRavikumar (2001). The first two papers emphasize family control due to institutional imperfections despite thatoutside managers might be more talented. The third paper deals with the possibility of reducing agency problemswith altruism and the prospect of succession within the family boundaries. The latter paper analyses the influenceof primary capital markets imperfections on the evolution of family businesses. The owners of these firms arecharacterized by a special business skill that is fixed through time. Leaving aside the debate on the merits ofmanager/owners, the model developed here combines the idea of family control, due to institutional shortcomingsor social norms, with the concern for the owner’s descendants inheriting the firm.

5 Traditionally, one of the main concerns in the literature on ownership structure is the trade-off betweenmanagerial control and liquidity; seeBolton and Von Thadden (1998), Bhide (1993), Maug (1998), Shleifer andVishny (1986). Nonetheless,Burkart et al. (1997)suggest that there is an additional trade-off between monitoringand allowing managerial discretion. Thus, the selection of a relatively low concentration of ownership acts as acommitment device that helps managers to have certain leeway to pursue their initiatives. In the model developed

G. Castaneda / J. of Economic Behavior & Org. 59 (2006) 249–286 251

Pagano (1993)has shown that an entrepreneur with borrowing constraints can improvehis risk-sharing opportunities by going public and using the additional liquidity to buy sharesof other companies.6 In the model presented in this paper, Pagano’s ways of reasoning isextended in three different lines. First, the selection of technology is considered anotherdecision variable. As shown inSaint-Paul (1992), the selection of a flexible (low-return)technology reduces the risk for firms that face an uncertain consumer demand. Second,risk-averse entrepreneurs face a trade-off between their need to diversify and their interestin controlling the firm. In the model presented here, the latter concern is assumed as asocial norm. Third, once other firms decide to become public companies, internal capitalformation has to compete with the stock market to attract savings.

The causal link that explains the reduction in aggregate savings is similar to the argumentdescribed inBencivenga et al. (1996). In their theory, a decrease in transaction costs, perhapsdue to an increase in liquidity, might induce a reallocation of resources away from financingnew projects and into the acquisition of shares in the secondary market. However, themodel presented here analyzes the stock market’s impact on both productivity and savingswhile endogenizing the emergence of an equity market. It is suggested that ownershipconcentration may hinder firms from undertaking risky but highly productive endeavors.Likewise, the model illustrates how portfolio management designed to maintain controlafter entering the stock market may lead to a reduction in savings.

The model developed byPagano and Roell (1998)illustrates a similar concern whenfirms decide to go public. In their paper, the controlling owner cares not only about themarket value of the firm but also about his future private benefits as a manager/owner.For these authors, the firm’s founder uses the ownership structure to limit agency costscaused by distrustful and diffused shareholders. This can be done by accepting a largeexternal shareholder who signals to the market the presence of certain restraint on managers’prerequisites. This strategy allows the firm’s founders to obtain cheaper equity capital.Therefore, when determining the distribution of shares, the owner faces a trade-off between

in this paper these issues are not taken into account. It is plainly assumed that manager/owners are interested inkeeping a high concentration of shares because of social norms and weak institutions, irrespectively of the impactthat this decision may have on the stock market’s turnover rate and on the managers’ decision capabilities.

6 Although equity rationing is also an important economic phenomenon, as pointed out inMyers and Majluf(1984)andGreenwald et al. (1984), outside equity and credit are two complementary sources of financing and,thus, the issuance of equity in the stock market could increase available resources for a manager/owner. Sincethe seminal paper ofJensen and Meckling (1976a), the corporate finance literature has considered that thereis a trade-off between the benefits of managers’ increased diversification, obtained by selling-off a fraction oftheir shares, and the agency costs involved with external financing (debt and equity). Moreover, in a developingeconomy, a reformulated pecking order theory suggests that firms will initially have access to the highly scarcesavings channeled through the banking system and only later to funds raised through the equity market that is justemerging. The model does not intend to imply that any type of firm, in a real economy, can get financing moreeasily through the stock market than through the banking system; it only states that for important and well-knownfamily firms, issuing equity is possible and necessary once their access to bank loans has been drained off. Theempirical literature on this regard is limited.Pagano et al. (1998)show that this strategy is typical for Italian firms.However, according to their paper, portfolio diversification may not be the most important reason for the decisionto go public. Furthermore, in the descriptive analyses of economic historians is commonly mentioned that largefamily firms were able to raise funds for the formation of diversified conglomerates once the transferability ofstocks and their trade in formal markets was legally allowed; see, for instance,Haber (2000)for the case of BrazilandMaurer and Haber (2002)for the case of Mexico.


excessive monitoring by a very large external shareholder and the costs associated withgranting an excessive leeway to manager/owners when going public.7 In contrast, in themodel presented here, only passive investors and large controlling shareholders are included.This is done to emphasize the trade-off between diversification and control and to highlightthat even public firms may reduce their risks by selecting technologies with a lower expectedreturn.8 Moreover, an entrepreneur’s decision to go public is embedded in a growth modelso that different paths of growth can be observed depending on the decision made.9

The rest of the paper is divided as follows: in the second section, the literature on stockmarket development and economic performance is briefly reviewed to provide some contrastwith the model of concentrated ownership developed here. In the third section, a capitalformation model is built from microeconomic foundations and the equilibrium conditionsare specified. The tentative growth paths derived from general equilibrium conditions arepresented in the fourth section. In the fifth section, some stylized facts on cross-countrydata are presented to highlight the possibility of a non-linear relationship between stockmarket size and economic growth. The paper ends with the conclusions and implicationsfor empirical research.

2. Theoretical and empirical background

2.1. Theory

When reviewing the literature on stock market development and economic growth, threechannels are usually mentioned to establish a positive link between them: liquidity, diversi-fication and a market for corporate control.10 With regard to the first factor, it is argued thatthe reduction in transaction costs offered by a secondary market helps diminish liquidityrisks and encourages economic actors to increase their investments in long-term productiveassets. A liquid stock market, by inducing a separation between real and financial decisions,allows firms to have long-term horizons with beneficial consequences for R&D and humancapital formation, which in turn improves ex-ante efficiency (Levine, 1991). Likewise, aliquid stock market favors the acquisition of information since investors can hide the use ofprivate information by trading in a market where prices are not affected by transactions. Themarket improves its allocating capabilities once the information is disseminated through theprice mechanism (Kyle, 1984; Merton, 1987; Holmstrom and Tirole, 1993). With respect

7 Although their model highlights the importance of a large and neutral shareholder, it does not explain howthis neutrality (or the lack there of) is reached. As these authors state, in many countries, it is difficult to findlarge external shareholders performing a neutral monitoring function. In fact, it is common to observe that largenon-controlling shareholders are bought off by the manager/owner through the use of side payments, reciprocalshare deals, favorable supply contracts, and other schemes. Under this scenario, the monitoring exerted by largeshareholders will not be credible, and diffuse shareholders will be reluctant to acquire stocks in the open market.

8 Other papers on the going public decision for firms in general areIbbotson and Ritter (1995), Zingales (1995),Subrahmanyam and Titman (1999), andChemmanur and Fulghieri (1999).

9 As opposed to the model of Bhattacharya and Ravikumar, where the relationship between capital markets andthe evolution of family firms is also analyzed, the paper presented here does not take the degree of developmentof the stock market as given; instead the surge and evolution of such a market is endogenized in the model.10 For longer surveys, seeDemirguc-Kunt and Levine (1996a)andLevine (1997).


to the second factor, it is said that an investor with an enlarged set of alternatives mighthave more incentives to save; furthermore, the firms released from the pressure of having tooffer less risky investments would specialize more and take better advantage of increasingreturns, despite the risk (Saint-Paul, 1992; Obstfeld, 1994). A third factor, monitoring man-agers, becomes relevant when there are ex-post incentives to divert resources. The stockmarket, by increasing the competition for capital and by making hostile takeovers easier,helps put managers’ interests in line with the maximization of stock value (Scharfstein,1988; Stein, 1988; Jensen and Murphy, 1990).

However, different authors have argued that the same three factors might have the oppo-site effect on economic growth. For instance, depending on the parameters that characterizethe risk-aversion of investors, the possibility of reducing risks by diversification mightreduce precautionary savings (Devereux and Smith, 1994). Moreover, financial pressureson managers to come up with attractive results in the short term might make managersmore short-sighted (Bhide, 1993; Shleifer and Vishny, 1986). Likewise, hostile takeoverscan harm growth by threatening the fulfillment of implicit contracts between managersand other stakeholders of the firm (Shleifer and Summers, 1988). Other negative links arealso put forward. It is said that by creating a speculative market with weak connectionsto the real economy, increased volatility distorts the price mechanism and the allocatingcapabilities of the market (Singh, 1993). The fact that noise traders (investors’ sentimentsor fads) influence the behavior of stock prices even in the long run implies that the stockmarket might negatively affect real activity (De Long et al., 1989). Managers guiding theirstrategies with the information content of prices could be misled when the market providesfalse signals about fundamentals. A fad for the stocks of a specific firm or industry mightreduce the price of capital and induce managers to invest in negative present value activitiesif they get private benefits (Morck et al., 1990).

The model presented here also analyses the influence of the stock market on economicgrowth. However, it suggests that the impact of an increase in market size depends on othervariables related to its development, such as ownership concentration, and that, as a result,the relationship between market size and growth may be non-linear. For emerging economiescharacterized by family firms, low liquidity in stock markets and poor legal and judiciarysystems, especially with regard to shareholder protection, an increase in market capitaliza-tion may lower growth if market size is past a certain threshold. Similarly,Berthelemy andVaroudakis (1996)have argued in favor of a ‘threshold’ effect for the relationship betweenfinancial depth and growth. This non-linearity is also in line with the theoretical work pre-sented inGreenwood and Jovanovic (1990), Huybens and Smith (1998), andAcemoglu andZilibotti (1997, 1999)when studying the finance-growth link.

2.2. Empirical literature

Although the empirical literature on financial depth and growth is vast, most studies do notattempt to measure the impact of stock markets on long-term economic performance.11 In

11 This is in part because a stock market development dataset for a 20-year period is available for only about 40countries. For more extensive surveys on empirical work, seeDemirguc-Kunt and Levine (1996b)andBeck andLevine (2004).


the first cross-country study,Atje and Jovanivic (1993)show a positive link between the ratioof traded value to GDP and economic growth.Harris (1997)reassesses Atje and Jovanovic’sestimates by including additional control variables and applying two stages least squares(2SLS) to remove potential endogeneity biases. In their analysis, they divide the sampleinto developed and less-developed countries and the stock market effect is, at best, veryweak.Levine and Zervos (1996)introduce a stock market-development index composedof indicators of size, liquidity, and international integration. Their regression model showsa positive connection between this index and the real economy while controlling for otherfactors affecting growth. Furthermore, they avoid problems of reverse causality by usinginstrumental variables to extract the predetermined component of financial development.A positive link is also statistically found byLevine and Zervos (1998)in a more detailedeconometric analysis. Likewise, by using a different indicator of liquidity (turnover ratio),these authors are able to disentangle the price effect from the liquidity effect.

Other studies use VAR methods.Rousseau and Wachtel (2000)examine the dynamicrelationship between equity markets and growth by estimating panel vector autoregressionswith fixed effects. In their findings, organized equity exchanges granger-causes investmentand output for a sample of emerging markets. Similarly,Arestis et al. (2001)show, throughthe use of quarterly data and the vector error correction model, that both bank and stockmarket development explain growth although the former financial institution exerts a muchlarger effect. Likewise,Caporale et al. (2004)apply trivariate causality tests to analyzewhether financial depth, and in particular stock market development, causes growth. Theirevidence in five out of seven countries favors the finance-causes-growth hypothesis. In astudy of five emerging economies,Spyrou and Kassimatis (2001)find that equity marketsplay a positive role in liberalized economies such as Chile and Mexico or in South Koreawhere government guidance is important, with the difference that for the first two countriesthere is a negative relationship between the banking sector and the real economy.12

Finally, Rousseau and Wachtel (2000)andBeck and Levine (2004)make use of paneleconometric techniques. While the former paper applies GMM-difference, the latter appliesGMM-system. The dynamic nature of a panel data allows GMM to deal with two-waycausality by using proper instruments for the contemporaneous values of the explanatoryvariables. In both cases, the turnover ratio is positive and statistically significant, even whencontrolling for bank development, although in the case of Beck and Levine, the results forone-step estimators are much weaker.

In brief, in a series of first-generation studies, there is not much formal econometricevidence suggesting that an enlarged market capitalization may lead to a slow-down ineconomic growth.13 However,Levine and Zervos (1998)and Levine (2002)show that

12 Because of the use of quarterly data, there is certain skepticism on the usefulness of VAR models to disentanglefrequency factors from long-run elements in establishing a causal nexus; seeBeck and Levine (2004).13 Harris, in his regression analysis using the less-developed countries sub-sample, finds a negative sign in

the coefficient of the interaction term between stock market activity and investment. Since the estimate is notstatistically significant, the author does not pay attention to the fact that this sign reverses the direction of therelationship estimated with the developed countries sample and with the full sample. Rousseau and Wachtel, intheir panel GMM estimates, find a negative sum in the coefficients associated to the lagged values of marketcapitalization when this variable is included with value traded in the VAR system, either with the full sample or


market size per se is not statistically associated with economic growth.14 Furthermore, inthe latter paper, Levine shows that the conglomerate indices of bank and stock market activityare more robust predictors of growth; in other words, the interaction of several variablesassociated with financial development seems to be a more important influence on economicgrowth. Thus, the evidence is not conclusive in so far as only linear regression models areconsidered.15 The removal of this assumption has given birth to a second-generation ofstudies whose results are still tentative but show very distinct results in comparison withthe early findings.

In a paper byKhan and Senhadji (2003), it is shown that a linear model captures the typicalpositive relationship between financial depth and growth when the measure of economicgrowth is long-term GDP. However, for the explanation of growth dynamics of individualcountries (i.e. growth averages in non-overlapping periods) a non-linear relationship seemsto be more adequate. In particular, these authors find statistically significant parameters witha quadratic relationship between stock market capitalization and growth. Thus, the negativesign of the quadratic term indicates that it is plausible that stock market capitalization mayslow growth after a threshold has been surpassed.16

Furthermore,Durham (2002)presents a cross-country study of medium and long-termgrowth that indicates that the typical findings of the first-generation papers are largelydriven by the presence of high-income countries in the sample. Likewise, the non-linearityhypothesis cannot be rejected when running growth regressions with the turnover rate as ameasure of market development and an interaction term between this variable and a seriesof legal base indicators. The same conclusion holds when the author uses a quadratic func-tional form, since the parameter associated with the quadratic turnover term is statistically

with the emerging countries sample. In the former case the sum is small in magnitude but statistically significant,while in the latter case it is not significant.14 In VAR systems in general, the causality link from market size to real activity is also very tenuous.Arestis

and Demetriades (1997), using the Johansen cointegration analysis with quarterly data from 1979 to 1991, showfor Germany that stock market capitalization affects real GDP only through the banking system; likewise, stockmarket volatility has a negative impact on output. For the US case, capitalization is an endogenous variable andreal GDP is weakly exogenous with respect to financial development (market capitalization and the ratio of bankcredit to GDP). Then, Spyrous and Kassimatis find evidence on positive causality running from market size to realeconomic activity for only three of the five countries in their sample; however, for India the equity market doesnot affect growth, and for Taiwan there is even a negative relationship between stock market capitalization andeconomic development.15 The use of cross-country regression models for analyzing long-run growth has been debated byEvans (1997)

precisely because of problems that stem from the heterogeneity of slope coefficients.16 In a more recent study using the level of liquid liabilities of the banking system and the amount of credit

issued to the private sector by other financial institutions,Favara (2003)finds a very weak relationship betweenfinancial development and growth using the one-step GMM-system econometric procedure. Moreover, the authoralso finds evidence of non-linearities in the data, and heterogeneity across countries in the parameters associated tothe financial variables. In particular, the null hypothesis of non-linearity cannot be rejected using a semiparametricmodel where no functional form on the relationship between finance and growth is assumed ex-ante. On theother hand, when estimating the relationship by one-step GMM for different sub-samples, he finds a negativecoefficient, but not statistically significant. Likewise, applying the Pooled Mean Group Estimator (PMG) thatallows for heterogeneous short-run coefficients yet constrains long-run parameters to be identical across countries,his results show negative and statistically significant parameters for the financial development variables in manycircumstances.


significant. This is a result that is very robust to the period used for the growth average, andto the type of sample considered, with or without countries with an inconsequential stockmarket. Hence, irrespective of how the results may be sensitive to the specific non-linearformulation assumed and how the relationship between stock market development andgrowth may be characterized by multiple thresholds, the empirical evidence demonstratesthat further studies are required to understand better the nature of the nonlinearity.

Finally, using a different framework,Morck et al. (2000)show with Canadian data thatlarge family firms are not conducive to growth due to management entrenchment, capitalmarket power, high barriers against outside investment, and low investment in innovation.Thus, along the reasoning behind their econometric model, it can be argued that an ill-functioning stock market might skew wealth distribution even further, thus increasing marketpower in capital markets and rearing barriers to capital mobility. Consequently, the findingthat heir wealth, as opposed to self-made wealth, affects economic growth negatively isconsistent with the main hypothesis presented in this paper,17 although it is not intended tosay that such a finding is an implicit test of the model.

3. Economic decisions and the structure of the economy

3.1. Technology and entrepreneur’s objective

In order to analyze the macroeconomic consequences of the trade-off between risk diver-sification and the concern for control, the model focuses on the choice of technology and thedegree of ownership concentration. To simplify the model, competitive markets are assumedto allocate capital and labor among firms in the economy, so an entrepreneur faces a pro-duction function that exhibits diminishing returns to capital. However, at the macro-level,there are positive externalities from the size of the economy’s capital stock. This positiveexternality may be due to “knowledge spillovers,” as inRomer (1986). Hence, the largerthe economy’s capital stock, the larger the accumulated experience which is spread amongthe firms of the economy.18 Therefore, the production function for the average firm is givenby

yt = Rκθt (Kt)

1−θ(Lt)θ (1)

whereyt is the output,Kt the capital stock of the representative firm,Lt the labor employed,κ the capital stock available in the economy to be split among the firms, andR is theproductivity term; furthermore, 0 <θ < 1 is a technological parameter.

That is, the production function exhibits diminishing returns whenκ remains con-stant. However, an increase in the individual firm’s capital stock increasesκ in the sameamount; hence, there are constant returns to scale at the aggregate level. This specificationassumes that the stock of knowledge grows at the same rate as total investment. Therefore,

17 This finding is also observed in the same study when using cross country macro-data.18 The assumption of spillovers is even more realistic in economies where firms are associated through business

networks, which indeed is a very typical industrial structure in developing economies. Moreover, it is precisely inthese economies where ownership concentration is very large.


expression(1) is derived from a Cobb–Douglas function where technology increases laborproductivity.19

Individual firms take the capital stock for the economy as given in their individualdecisions. Moreover, if labor supply is perfectly inelastic, then firms’ production is givenby a typical ‘AK’ technology, as in expression(2). From the first order conditions in thefirm maximization problem it can be shown that a shareθ of production goes to labor anda share (1− θ) goes to capital.

yt = RKt (2)

On the other hand, the entrepreneur, who retains the control of the firm’s operation andassets, may be reluctant to float shares in the open market.20 Consequently, the controllingshareholder discards firm value as his only objective and tries to maximize his personalutility. The expected utility function of the entrepreneur includes his consumption and thefinal value of productive assets. The latter is incorporated in the objective function to capturethe notion that an entrepreneur may care about the status and privileges conferred by thecontrol of a high-valued firm.21 Likewise, the tight control of the firm can be modeled byintroducing a cost to accepting external shareholders of the firm that would be negativelyrelated to the entrepreneur’s holdings of the firm’s shares.22This cost could be a consequenceof influence activities of outsiders with large holdings, transaction costs due to opportunisticbehavior by unaccountable managers (agency costs), or as a loss of confidentiality due topossible leakage of inside information.23

In the economy there areE infinitely-lived firms, although the controlling owners livefor only one period. When an entrepreneur decides to float shares on the market, he willdiversify risk by acquiring the stock of other firms.24 However, the demand for shares in thestock market also comes from ordinary investors with no voting rights on the firms’ internalaffairs. However, investors can, as a group, interfere with the control privileges of majorityowners. Ordinary investors/workers have a choice between saving with shares in the stockmarket or saving through the internal capital market of their employer. Since capital stock

19 SeeSala-i-Martın (1994).20 In the real world there are other schemes to retain control rights above what is granted according to cash flow

rights; these include shares with superior voting rights, cross-ownership or the use of pyramids to control firmsthrough a chain of companies (seeLa Porta et al., 1999; Castaneda, 1998, Chapter 4).21 If the firm is considered a family business, then the value of productive assets in the utility function also

represents the concern for the value of the inherited capital stock in the family firm.22 This cost is exogenously assumed, although it can be explained as a consequence of asymmetric information,

legal institutions and cultural factors.23 Pagano et al. summarize the cost of going public in the following concerns: adverse selection, administrative

expenses and fees, and loss of confidentiality, while the benefits are derived from overcoming borrowing constraints,greater bargaining power with banks, liquidity and portfolio diversification, monitoring, investor recognition,window of opportunity, and change of control.24 According to the financial services view presented inLevine, 2002, both bank and markets provide comple-

mentary financial services that are helpful to spur growth. The author shows empirically that overall financialdevelopment (measured by the conglomerate indices of bank activity and stock market activity) is positivelyrelated to economic growth. Therefore, it is fair to say that once the equity rationing bottleneck is ameliorated,manager/owners may take advantage of this financing possibility to meet their goals, even if a credit constrainremains.


depreciates in one period, the firms operate in the subsequent period with the real assetssaved internally and controlled by the next generation of entrepreneurs. In other words, inthe model, the decisions of ordinary investors define the saving ratio of the economy, whilethe entrepreneurs’ choices define the depth of the stock market and the productivity of theeconomy. In the next two subsections, the maximization problem and the resulting behaviorof both entrepreneurs and ordinary investors are described in detail.

3.2. Entrepreneurs’ decision problem

Entrepreneurs face four intertwined decisions:

(I) The type of technology to be used, which can be either conventional (c) or innovative(p). The former technology is characterized by a low-return and a low-variance, whilethe latter has a high-return and a high-variance.

(II) The corporate governance of the firms. That is, the decision to keep the firm closedor to list it on the stock exchange. When going public, the entrepreneur must payfees, fill out paperwork and make payments to lawyers, accountants, financiers andthe like. The flotation cost also includes the loss of secrecy caused by the mandatoryrequirement of information disclosure.

(III) The degree of flotation of the firm’s stock. Cultural patterns and inadequate institutionscreate an incentive to maintain a large share of the stock in the hands of the control-ling managers and their families to prevent opportunistic behavior from outsiders. Aconsequence of maintaining extreme control is poor diversification.

(IV) The size of the next generation’s capital stock.25 Since real savings are made onlyby ordinary investors, such savings can be used for productive purpose only whenentrepreneurs have acquired them through internal capital markets.

Thus, in any period, the maximization problem for thejth entrepreneur is as follows:26

Maximize E[U(Cjt ) + V (Pnj

t Kjt+1)]

{kij, T, G, Kjt+1} for j = 1, . . . , E

subject to Pnjt K

jt+1 + C

jt = Y

jt (T, G)

and Pjt K

jt =

E∑i=1

kijt Pi

t + PtF

with F ≤ Kjt

(3)

25 In case of assuming a family business, the productive assets are usually bequeathed to the offspring in chargeof the firm for the next generation.26 The additivity of the utility function is assumed for technical convenience. However, this assumption is not

neutral since it implies that an increase in the price of productive capital exerts only an income effect on its demand.Notice that in this formulation the price of the consumption good is the numeraire.


whereE[·] is an expectation operator, andU(·) andV(·) are additive utility functions forconsumption and the value of the stock of capital, respectively.C

jt is the consumption for

thejth entrepreneur born in periodt; Pnjt the price of new productive assets for thejth firm

bought in the internal market in periodt; Pit the price of the stock for theith firm in the

secondary market;27 Kjt thejth firm capital stock; andE is the total number of entrepreneurs

in the economy. The decision variables are thejth entrepreneur’s holdings in theith firm’sequity (kij

t ), so thatkjjt refers to the degree of ownership concentration;T the binary index

for the technology selection, whereT = c for the use of a conventional technology andT = pfor the use of a productive (innovative) technology;G the binary index for the close/openselection of corporate governance, whereG = 0 indicates a preference to remain completelyclosed andG = 1 means the flotation of shares in the open market; andK

jt+1 is the stock of

productive assets to be used by the next generation in control of thejth firm. The first budgetconstraint in expression(3)signals that thejth entrepreneurs’ income (Y

jt ) is used exclusively

for consumption and for the purchase of productive capital. Because entrepreneurs care forthe status offered by running the firm, they are willing to acquire productive assets withno pecuniary motive.28 The first constraint in(3) highlights the fact that for entrepreneursthere is a trade-off between consumption and capital. Notice that a dynamic path for capitalcan be generated with this formulation. The second constraint indicates that the lack ofcredit forces firms to float their shares in order to get a diversified portfolio and makesindispensable the ex-ante payment of flotation costs (PtF), wherePt is the average relativeprice of firms’ shares.29 In each period, a new generation of entrepreneurs takes charge of thefirm, and technology and ownership decisions are made all over again.30 Each generationof entrepreneurs decide whether the firm will be kept open and what its financial portfoliowill be.

A utility function, such as the one presented in expression(3), is conventionally usedwhen analyzing family firms. In such models, for example, Chami’s considerations of altru-ism and succession are explicitly laid out in the economic agents’ objective functions. In

27 In the model, entrepreneurs derive utility from new productive assets and not from the valuation of equityshares in the secondary market, because, as explained below, they care for the family status coming from theirheirs running the firm and not from their inherited market value of their stockholdings. Consequently, the priceof capital in internal capital markets (where savings for productive uses takes place) is the proper unitary value tocalculate the worthiness of the family firm as an operating business. As in Pagano and Roell, the controlling ownercares about his future private benefits as a manager/owner, which are not necessarily derived from the marketvalue of the firm.28 Therefore, these assets are given for free to the next generation of entrepreneurs. This assumption has more

real life sense if the firm is considered a family business and the stock of capital is inherited to the next generationof entrepreneurs in the family.29 If the flotation cost is interpreted as a registering cost, it is straightforward to see the need of an ex ante payment.

However, for the second interpretation, loss of secrecy, the corresponding pecuniary cost can be made ex-postunless the disclosure of information produces an immediate expenditure (e.g. tax payments.).30 Another approach would have been to specify an infinite-horizon problem where entrepreneurs’ savings were

motivated by future consumption. However, such an approach complicates the algebraic treatment of the model,because of the different sources of income and the introduction of the cost of ownership. Besides, it was preferredto present a model with more realistic and richer assumptions, such as entrepreneurs interested in bequeathing anoperating firm to their heirs, not just financial wealth with the presence of an internal capital market and ordinaryinvestors without means of production.


particular, altruism is one-sided when the child/agent’s utility is embedded in the princi-pal/parent’s utility function; moreover, the impact of the succession is taken into accountby formulating a concern for future utility. In the model presented here, these two elementsare melded into one assumption: the parent’s concern for the future (t + 1) well-being ofthe heirs. Therefore, the first term,U(·), can be interpreted as the parent’s current utilityfrom his own consumption and the second term,V(·), is the utility derived from his concernwith his heirs’ future well-being. Implicitly here, the parameters associated to altruism andpatience are set equal to one; although, this is not a relevant assumption. Because the modelanalyzes the consequences of family firms with a taste for keeping control through time, itfocuses only on future asymmetric altruism, specifically, at the beginning of the next period(t + 1).31

Moreover, the second term of the utility function implies that there is also a concern forstatus and the holding of the firm as an asset of the family’s wealth. This assumption isin contrast with those models where the consumption stream is the only argument in theinter-temporal utility function. For instance, in Bhattacharya and Ravikumar firm ownerscare about the next generation, but they do not mind if the inherited wealth is in the formof cash or productive assets. The perspective taken in this paper is more consistent with thetraditional view that entrepreneurs prefer to keep the firm in the family as long as possible.32

In the first budget constraint presented in expression(3), income is a random variablethat depends on decisions about technology and governance. Because each choice has twopossible combinations, the corresponding income to thejth entrepreneur can take one ofthe following four different values:

(i) The firm stays closed and the selected technology is the conventional one; then

Yjt (Gj = 0, T j = c) = (1 − θ)RcK

jt (4)

where 1− θ is the share of income coming from production that is assigned to entrepreneursas profits. If the productivity termR in expression(2) is assumed a random variable, thenRc defines the random productivity attained from using the conventional technology; thisvariable has meanµc and varianceσ2

c . Hence, entrepreneurs’ mean return on capital is givenby (1− θ)µc.

(ii) The firm stays closed and the technological choice is the innovative one; then

Yjt (Gj = 0, T j = p) = (1 − θ)RpK

jt (5)

where (1− θ)Rp is the return on capital given the use of the innovative technology, whichis a random variable with mean (1− θ)µp and variance (1− θ)2σ2

p, such thatµp >µc and

σ2p > σ2

c .33

31 Diverse studies have shown that there could be a congruence between founder’s utility and profit maximization;see, for instance,Radice (1971), Holl (1975), and other references cited by Chami.32 SeeWard (1987).33 This choice can also be interpreted as a selection between a relatively safe government contract and an innovative

project that is risky but with a good potential return. Alternatively, and in the context of an open economy, somefirms may decide to reduce ownership concentration in order to achieve strategic alliances with foreign firms that


(iii) and (iv) the firm goes public:

Yjt (Gj = 1, T j = h) =

E∑i=1

kijt [(1 − θ)Ri] − C(Kj

t − kjjt ), h = c, p

with a conventional technology :Rj = Rc

with an innovative technology :Rj = Rp

(6)

whereC(Kjt − k

jjt ) is a convex function that represents entrepreneurs’ costs due to inad-

equate institutions. In other words, an ownership structure with reduced inside hold-ings (Kj

t > kjjt ) introduces three types of costs: agency costs due to perks obtained by

owner/managers unaccountable to minority shareholders,34 costs due to influence activitiesof a large presence of external stockholders,35 and a loss of confidentiality.36 These costswill be larger when the legal and social institutions perform poorly to protect property rights.The first term in expression(6) represents the financial earnings of investments made by thejth entrepreneur, which are proportional to their holdings of equity in each company (k

ijt ).

This portfolio is purchased with the proceeds of the sale of his firm’s equity as indicated bythe ex-ante budget constraint in (3). Notice that the return on investments isRh (with h = c,p) regardless of the firm’s controlling owner. This consideration results from assuming thatall the firms are intrinsically identical.

In this formulation there is an important embedded assumption with regard to techno-logical and corporate governance decisions. In each period initial manager/owners have fullcontrol of decision making, irrespectively of the final distribution of cash flow rights. Thissounds plausible for those parameter values that lead to a large concentration of ownership.Nevertheless, the parameters are not restricted in the model since in real life situations,control can also be attained with relatively small shareholdings. This is done with theuse of pyramids or through the issuing of dual shares where the rule one-share one-voteis violated. Obviously, as these financial artifacts are used more extensively, the lower thetrade-off between diversification and control through ownership becomes.37 In other words,the model assumes that no external agent can decideT andG for the firm, so ownershipparticipation is endogenous but control is predefined. For simplicity, external ownershiponly contaminates control through its impact on the cost function inserted in expression(6).

The explanation of capital formation in the model is completed by adding economicagents who save in each period. So far, the entrepreneurs’ decisions generate only a demand

supply technology. Consequently, there is an additional trade-off between the use of an innovative technology andcontrol that does not come from the risk considerations presented here.34 In a context of rational minority shareholders, the potential opportunistic behavior of controlling shareholders

is heavily discounted by the market (seeJensen and Meckling, 1976b). Although this process is not explicitlymodeled here, the cost of this discount net of manager/owner’s perks is introduced through expression(6).35 SeeMilgrom and Roberts (1988).36 SeeYosha (1992).37 Larger flexibility to diversify means lower risks so that the firm is more interested in implementing a high-

return technology. However, when unchecked control remains, despite the degree of ownership concentration, rentextraction is still viable and, hence, productivity suffers due to the lack of commitment of the other stakeholders(seeCastaneda, 1998).


for productive assets. Therefore, in this economy, the supply of real savings is determinedby ordinary investors/workers who assign a fraction of their income for the formation ofcapital. Although this process is done internally in the workers’ firm, it is only a simpli-fying assumption that represents the saving behavior of individuals without access to aspecific technology and lacking an endowment of capital stock. Furthermore, this approachhighlights the importance of internal capital markets for the financing of firms in emergingeconomies where inadequate institutions hamper the functioning of external capital mar-kets. Consequently, the model is capable of analyzing the impact on growth when there isa shift in portfolio away from internal capital markets and into external markets.

3.3. Ordinary investors’ decision problem

Ordinary investors also live for one period, but they do not receive any endowment ofassets. Instead, they have labor income that they allocate between capital investment intheir firm and a financial portfolio in order to diversify. However, when the stock market isabsent all savings are invested within the internal capital market of the firm. Only this type ofsavings is linked to capital formation since it is in this market where the current generation ofentrepreneurs acquires the capital stock needed for future production. As explained above,the acquisition of capital stock in the model is the result of entrepreneurs’ concern for theirheirs in the next period. On the other hand, all earnings obtained by ordinary investors areused for their end-of-the-period consumption; thus their objective function is given by

Maxst ,k

iot

E[U(Cot )] for o = 1, . . . , I (7)

whereE[U(·)] is the operator of expected utility,Cot the consumption at the end of periodt

of theoth ordinary investor working in thejth firm, andI is the number of worker/investorsin the representative firm. Control variables are (i) worker’s investments in those firms (kio

t )that decided to float, and (ii) the fraction of his income (0< st < 1) kept within the firm forthe formation of capital. Yet when there are no shares to buy, the worker is forced to saveall his income (st = 1, kio

t = 0) in the firm’s internal capital market.38

As in the case of entrepreneurs, workers also face a financial constraint in externalcredit markets. Nonetheless, it is assumed that workers can acquire at the beginning oft aportfolio of assets with a line of credit based upon their expected wages. Thus, end-of-periodconsumption comes from their financial earnings, from observed wages, and from capitalgains obtained with the sale of productive assets in the primary (internal) market, once thecredit has been paid for.

Therefore, the representative ordinary investor faces the following constraints when firmsgo public:

38 Notice that the rate st does not determine how income is split between real savings and consumption. In themodel, consumption is made at the end of the period t out of the generated income, financial and salaries. Thus,st indicates the fraction of the ordinary investors’ portfolio held at the beginning of periodt in the internal capitalmarket vis-a-vis equity shares, where only the former type of investment (savings) translates into productive assetsfor the future.


(1 − st)E(wt) =E∑

i=1

Pit k

iot (8)

Cot =

E∑i=1

[(1 − θ)Ri − rPit ] kio

t + wt + stE(wt)(Pnjt − r) with wt = θRjK

jt

I

(9)

wherer = (1 +ρ) andρ > 0 is the interest rate, andwt is the wage earned, ex-post to therandom outcome, by a representative worker of thejth firm. It is assumed that firms’ salariesare the same since firms are identical in all respects except in the realization of the randomvariable. Expression(8) represents the financial portfolio budget constraint, and(9) showsthe income sources for consumption expenditure. The summation term on the right handside of Eq.(9) represents after-credit earnings on his stock holdings, while the last termrepresents the capital gains on savings in the internal market net of borrowing costs. Becausethe model has a static nature, workers have an incentive to save as long as they can get acapital gain in the internal capital market and their end-of-period consumption can beincreased.

On the contrary, when firms have decided to stay completely closed, and thus workersare forced to face a lack of diversification constraint, their only sources of income are wagesand the capital gain obtained in the internal market.

Accordingly, workers’ consumption under financial autarky is given by39

Cot = wt + E(wt)(P

njt − r) (9b)

Before deriving the first order conditions that describe the behavior of workers andentrepreneurs, the next subsection specifies the equilibrium conditions for the capital mar-kets of the economy. These conditions determine the acquisition cost of shares in thesecondary market and the sale value of savings through the internal capital market. In themodel, earnings for stockholders are derived from the random return on the capital stockas a proportion of their holdings. Likewise, prices in the internal capital market produce acapital gain for ordinary investors/workers who save in the firm.

3.4. Capital markets

From the demand side, ordinary investors and entrepreneurs use part of their wages andproceeds of the public issue, respectively, to buy other firms’ stock in the external marketin each period. On the other hand, the supply of shares is positive and identical to the firm’scapital if thejth entrepreneur decides to go public. As can be seen from Eq.(6), a fractionof this capital is repurchased by the initial owner if he decides to retain control in the firm

39 Notice from this formulation that wages are known at the end of periodt and that capital gains are made outof a credit received at the beginning of periodt, when salary was not known, and hence the size of the loan had tobe made in terms of the expected salary.


(i.e.,kjjt is an asset of his portfolio). Hence the equilibrium price in the secondary (external)

market for the equity of thejth firm (Pjt ) is obtained from the following identity:

Kjt =

∑i�=j

kjit + k

jjt +

∑{o not injth firm}

kjot +

I∑oj=1

kj,ojt (10)

where the first summation considers other entrepreneurs’ demand for shares in thejth firm,the second term are the controlling shares of thejth entrepreneur, the third term are thestocks of thejth firm held by the non-jth firm’s workers,40 and the last summation is thedemand for shares in thejth firm by their own workers. The reason to distinguish the cross-shares from the own-shares holdings and the type of investor resides, first, in the differencesin motivations for each type of holding: diversification versus control and, second, in thesource of income to acquire financial assets: equity issuance versus wage income.

With regard to the primary (internal) market, the supply is derived from the savingsaccumulated in the period by the firm’s workers, and the demand comes from the desires ofthe jth entrepreneur to bequeath the business to a new generation of entrepreneurs, that is,to the family offspring. In order to simplify the analysis, neither workers nor entrepreneurshave bargaining power,41 thus the price (Pnj

t ) of capital assets in thejth firm is determinedin a competitive equilibrium:

Kjt+1 =

I∑i=1

stE(wt). (11)

When firms decide not to go public, which in the model is a generalized decision giventhe symmetry of the firms, there is still a primary market. All productive capital depreciatesin one period, and workers provide with their savings the capital to be bequeathed byentrepreneurs to their successors. Notice that for the same level of payroll, the supply ofcapital under these circumstances is larger than in the case of open firms given that ordinaryinvestors do not have any other alternative for investment:

Kjt+1 =

I∑i=1

E(wt). (11b)

According to this formulation, capital prices are known when both agents, workers andentrepreneurs, solve their respective problems at the beginning of periodt. This is so despitethe fact that the inherited capital stock is used productively until periodt + 1.

3.5. Behavioral equations

In order to find an explicit solution for the behavior of investors, a functional form for theutility function needs to be specified. If an exponential utility function is considered, then

40 It is assumed that each company has the same number of workers (I), thusI(E− 1) workers in the economybuy shares of thejth firm.41 Clearly this assumption is unrealistic, yet assuming otherwise does not change the main implications derived

from the model.


the absolute risk aversion is constant according to the Arrow–Pratt measure. Furthermore,if the returns on capital are normally distributed, it is easy to show that the expected utilityfunction is a monotone transformation of the mean-variance utility function:42

E(U(x)) = E(x) −(

b

2

)Var(x) (12)

wherex represents the argument of the utility function, which in the model could be eitherconsumption by entrepreneurs and workers or the value of inherited capital assets, the firstone being a random variable and the second one a known value to be determined by theentrepreneur; b represents a risk aversion parameter for all agents in the economy; andVar(·) is the variance operator. In addition, there is also a cost due to immature institutionsthat represents the cost of accepting interference from outside ownership:

C(Kj − kjjt ) = γ(Kj − k

jjt )

2

2(13)

whereγ > 0 and whose value has to do with the nature of the legal, political and social systemof the economy where the firm operates. That is, a largerγ implies that the entrepreneur ismore concerned for control because of the socio-economic environment. This specificationgives more weight to the cost of outside interference as the difference between firm’s capitalstock and the ownership of the controlling group increases. Notice that by construction thelatter cannot exceed the former.

The symmetry among firm/entrepreneurs and among ordinary investors produces iden-tical behavior for each type of economic agent:k

ijt = kt, k

jjt = k

jt , k

jot = ko

t , kj,ojt = kw

t .Thus, share prices in secondary markets are the same for all firms (Pi

t = Pt). When thisfact is introduced into the ex-ante budget constraint in problem(3), it is straightforward toderive the entrepreneurs’ demand for equity in other firms:

kt = Kjt − k

jt − F

E − 1for i �= j. (14)

While the demand for the entrepreneur’s holdings of his own stock is obtained fromthe first order conditions of problem(3), once(14) has been substituted into the objectivefunction and the budget constraint is defined using income as in(3),43

kjt = K

jt [γ + Bh/(E − 1)] − FBh/(E − 1)

γ + EBh/(E − 1)(15)

where Bh = b(1 − θ)2σ2h, andh = c or p depending on whether the technological selection

of the firm is conventional or innovative.

42 SeeMas-Colell et al. (1995), p. 191 andSargent (1979), pp. 150–151.43 Although firms are identical there still is interest in diversifying given the randomness ofRj, where

cov(Ri,Rj) = 0, for i �= j.


On the other hand, it is assumed that the value function for the inherited capital has thefunctional form

V (Pnjt K

jt+1) = α + β(Pnj

t Kjt+1) − δ(Pnj

t Kjt+1)

2(16)

with parametersα, δ > 0 andβ > 1, such that one can get the proper second order conditionsof the maximization problem.44 Therefore, from problem(3), the optimal choice for theselection of productive assets to be bequeathed by entrepreneurs is given by the followingexpression:

Kjt+1 = β − 1

2δPnjt

(17)

From the first order conditions derived in the problem of the ordinary investors [(7), (8)and (9)] it is easy to get the optimal values for their decision variables, assuming that asecondary market has appeared:

kot = µh(1 − θ) − PtP

nt

bσ2h(1 − θ)2

> 0 (18)

kwt = ko

t − θKjt

(1 − θ)I(18b)

and

st = 1 −{

Ekot − θK

jt

(1 − θ)I

}PtI

θµhKjt

(19)

whereh = c or p depending on the technological selection.As shown inSupplementary data,45 the savings rate is strictly positive and lower than one

when prices in the capital markets are strictly positive. Moreover, when the number of firmsin the economy is relatively small, workers/investors have an incentive to have a positiveholding of their firm’s shares. Intuitively, if the number of public firms is rather small inrelation to the size of the economy, workers’ diversification cannot be exhausted with theirholdings in other companies. Thus, they will invest some funds at the firm where they work.By contrast, when the capital stock is relatively high, income will be derived mostly fromwages, and hence, they will not have a strong incentive to invest in the company’s shares.

The next step in the derivation of the model is to build the alternative growth paths thatthe economy might follow depending on entrepreneurs’ maximization processes. Four alter-natives are possible when combining the technology and the going-public decisions. Theshape and position of these paths are linked to the technological preference of entrepreneursand to the saving rate selected by ordinary investors. In this economy, growth is possiblewith the capital accumulated in each period through internal capital markets. In other words,

44 The value of these parameters depends on cultural factors and especially on the importance assigned to thedesire of having a high-valued firm retain the family status; that is, the importance of keeping the firm as a familybusiness. These parameters can also be interpreted as the entrepreneur’s degree of patience and altruism.45 Supplementary datais available on the JEBO website.


entrepreneurs, when selecting their ownership structure, determine the productivity of theinherited capital stock, while ordinary investors determine the capital investment in theeconomy.

3.6. Equilibrium growth

Once the optimal portfolio for both type of agents is substituted into the market clearingconditions(10) and (11), equilibrium prices are simultaneously derived. Then, when pricesare substituted back into the inherited capital stock(17), it is possible to define the rates ofcapital formation:

* Rate of growth when there is a stock market:

ghst =

(Kt+1

Kt

)− 1 = θµh

1 + Dh(Kjt )

− 1, h = c, p and s= stock market

(20)

where

Dh(Kjt ) =

{(1 − θ)µh −

[F + θK

jt

(1 − θ)

]Bh

IE

}2Fδ

(β − 1);

* Rate of growth with financial autarky:

ght =

(Kt+1

Kt

)− 1 = µhθ − 1, h = c, p. (20b)

Since without secondary markets all savings are channeled through the firm, workers’supply of capital is always sold in the primary market. Hence the price of capital in theinternal market is not a relevant variable and the rate of growth is a constant term, asseen in expression(20b).46 Fig. 1presents four different paths of growth depending on thetechnological selection and on the going-public decision. The shape and position of thesepaths are based upon propositions C–F presented inSupplementary dataavailable on theJEBO website. Briefly, these propositions state that the growth dynamics depicted are morelikely when the risk aversion of invertors is relatively high and the number of ordinaryinvestors is relatively large, as explained below.

The increasing shape of the growth path when firms go public implies continuous changesin workers’ savings rate.47 This result is explained intuitively with the following argument.When the economy expands, there is a larger availability of funds to grant credit lines toworkers so that both the absolute demand for shares in the external market and the supplyof savings through the internal market are increased. Moreover, the needs of diversification

46 When the saving ratest for the workers’ decision problem is exogenously given, the model is slightly simplifiedand a constant rate of growth is also obtained for the stock market case. Nonetheless, this option was not takenhere since the endogeneity ofst allows for a richer explanation of the dynamics without interfering in conveyingthe essence of the model.47 When the equity market is open to any interested investor, a positive growth rate requires a relatively high

saving rate. One way to surpass the needed threshold is by assuming a certain degree of risk aversion and interestin diversification, guaranteeing the required investment in the internal market.


Fig. 1. Possible growth alternative.

and the enlarged demand for other companies’ shares increase their costs and produce achange in the workers’ portfolio composition in favor of saving in the internal market. Inconsequence, the sale value of savings is reduced and the acquisition of productive assetsbecomes more attractive for entrepreneurs. These economic agents decide to enlarge thesize of their inheritance at increasing levels since the prices of new productive assets keepfalling in the internal market. A new round of capital formation stimulates further growthas long as the prices of capital assets remain positive. This scenario is valid for all sizesof capital stock below a certain upper bound and when the number of workers/investors isrelatively large. Otherwise, the very high absolute demand for financial assets combinedwith the diversification needs of relatively few investors would induce a steep hike in shareprices and a sharp fall in the value of the capital stock.

On the other hand, in an economy under financial autarky, a positive rate of growth canbe obtained as long as there is a significant labor’s share of production proceeds (θ) thatare accumulated within the firm.Fig. 1shows that the path of growth for unlisted firms isalways above the corresponding path for an economy with stock markets as long as the listingdoes not induce a change in technology.48 This result is a straightforward consequence ofhaving a unitary saving ratio in the financial autarky case where workers are forced to savein capital assets.49 However, the diagram also shows that the path of growth for public

48 This can be easily proved by assuming strictly positive share prices; consequently, for values ofK that haveeconomic sense, the path growth under autarky is always above the path growth under the presence of a stockmarket.49 Although flotation cost is bounded below and cannot be equal to zero for the hypothesis to hold, as stated in

proposition J inSupplementary dataavailable on the JEBO website, it seems paradoxical that if flotation cost (F)


firms is higher when public firms are enabled to use an innovative technology comparedwith unlisted firms who use conventional technology. The latter scenario may follow froma risk aversion that increases entrepreneurs’ demand for shares in the secondary market.Consequently, the large costs of the financial assets in secondary markets increase workers’supply of internal savings, and this reduces the value for capital assets, in turn encouragingentrepreneurs to acquire more productive assets for future generations. Therefore, capitalformation, although smaller when firms are listed, is not very different from the one observedunder financial autarky. Accordingly, the willingness of public firms to invest in more riskybut more productive technology offsets the impact of its lower saving rate.

In the following section, the model is closed by overlapping the utility functions valuedaccording to the different technology and going-public decisions. The four combinationsoffer partial optimum utilities, and the resulting envelope curve defines a global optimumfunction. This curve specifies for each size of the economy not only what the optimalportfolio and saving ratio under the stock market case are, but also what technology willbe used and whether firms will be better off going public. Each time these decisions aremade, the economy is located in a specific growth path. Therefore, with the help ofFig. 1,it is possible to detect when the economy moves along a particular path and when it shiftsdiscretely from one path to another.

4. Growth implications of entrepreneurs’ decisions

In the model, the capital stock of an economy can lag behind other countries for alengthy period of time when the former is trapped in a low-growth-path. For instance,when the representative firm becomes a public company but invests in projects with a low-return technology. This scenario is described by the bottom path ofFig. 1. In order todetermine the conditions under which this scenario or some other is attained, it is necessaryto compare the entrepreneurs’ expected utility for the four possible paths. Despite howordinary investors’ portfolio decisions determine the savings available for capital formation,the entrepreneurs are, in the end, crucial in defining what path the economy will follow.The interest of entrepreneurs in finding an appropriate balance between diversification andcontrol determines an array of decisions (technology, listing and ownership) that, althoughoptimal for the representative controlling owner, may lead the economy to a low-growthpath.

Consequently, once the optimal demand for financial assets derived in section threeare substituted into the entrepreneurs’ expected utility, the solution of the model can befound by comparing these valued functions for the four possible scenarios: innovative or

goes to zero, then the growth rate with financial autarky is equal to the growth rate when there is a stock market;see expressions(20) and (20b). This result can be explained intuitively as follows: for a very small flotation cost,all the income from the public issuance can be used to purchase shares in the secondary market; this very highdemand from entrepreneurs produces a very steep value for shares so that the price is prohibitive for ordinaryinvestors who decide, instead, to invest exclusively through the internal capital market, making the saving rateequal to one. This situation produces a growth rate identical for the financial autarky case and for the stock marketcase.


Fig. 2. Ecomomy size and risk-reduction decision.

conventional technology (T = p orc, respectively) combined with the going public or stayingclosed selection (G = 1 or 0, respectively). Thus, the envelope curve inFig. 2 (bold line)determines the range ofKt where each of the four alternatives is preferred.

* Expected utility when the firm stays closed and the selected technology is conventionalor productive. From(3) with k

ijt = 0, (4), (5) and (16), (17), it is easy to find that:

E(U|T = h, G = 0) = (1 − θ)µhKt − 12b(1 − θ)2K2

t σ2h + λ (21)

where

λ =[

(β − 1)2

4δ

]+ α, with h = c orp.

* Expected utility when the firm goes public and the technology is conventional or pro-ductive. Substituting the optimal values(14), (15) and (17)into the income formulation(6) and the objective function(3), one can get, after some algebraic manipulations, thefollowing expression:

E(U|T = h, G = 1) = TK2t + HKt + C (22)

where

T = −Bh

2(γ + EBh/(CE − 1))2

[γBh + EB2

h

(E − 1)2+ γ2 + 2γBh

E − 1

],


H = (1 − θ)µh + FB2h

(E − 1)(γ + EBh/(E − 1)),

C = λ − (1 − θ)µhF − Bh(γ + Bh)F2

2(E − 1)(γ + EBh/(E − 1)), with h = c orp

Expressions(21) and (22)are non-linear functions with respect to the size of the economy(capital stock per firm). The shapes and positions of the expected utility curves drawn inFig. 2are derived from propositions G–K presented inSupplementary data. Briefly, thesepropositions state that the relevant utility function varies mainly with the size of the firm(or the economy, given the symmetry assumption). That is, selection of technology and thechoice between being a private or public company is based on the level of the firm’s capitalstock. Likewise, the shape of the curve wrapping up the utility functions is also based onsome additional assumptions linked to the value of the model’s parameters: (i) a ratio ofmean returns for the technology (µp/µc) bounded by both sides, (ii) a flotation cost (F)bounded by both sizes, (iii) a large degree of institutional immaturity (γ), and (iv) a wageshare (θ) above certain threshold. Although a different combination in the value of theseparameters might produce completely different results, the purpose ofFig. 2is to illustrate atypical emerging economy characterized by family firms, ownership concentration, limitedinnovative capacity, labor-intensive industries, a poor regulatory framework for the financialmarkets, and inadequacies in the legal and judiciary systems. The specific influence thateach of these assumptions exerts on the shape of the diagram is explained intuitively below.

As stated in assumption (ii, lower bound), for smaller firms, the flotation cost will besufficiently high as a proportion of their equity issue that it is not worthwhile to go public. Asthe closed firm grows (Kt > K∗

1), there is an initial shift from an innovative technology to aconventional technology. This is so because when firms start small and their incomes do notfluctuate as much, the gains from diversification are limited. Hence, for small-sized firms,entrepreneurs prefer to take advantage of the high mean return offered by the innovativeprojects, as stated in assumption (i, lower bound), despite the risks involved; however, suchbenefit subsides later when firms (and the economy) are bigger.

In the third stage, when firms achieve a certain size (Kt > K∗2), entrepreneurs prefer to

go public in order to reduce their risk even more. They do so by using the funds from a publicoffering to acquire other companies’ stock. Nonetheless, their offering is limited becauseentrepreneurs keep a large percentage of the stock to maintain control of the firm, reducingthe cost of influence activities by outsiders and compensating for the lack of adequateinstitutions in the economy, as stated in assumption (iii). The decision to maintain morecontrol reduces firms’ ability to take advantage of the higher return offered by the moreproductive projects. Moreover, if flotation or secrecy costs are above a certain threshold, asgiven in assumption (ii, lower bound), the net proceeds from issuing stock will be limited,and hence, financial resources will be insufficient to build a fully diversified portfolio.50 This

50 On the other hand, if the flotation cost were very high but still feasible so that it could be paid, the entrepreneurmight be forced to undertake innovative projects to compensate for the lack of diversification and to take advantageof the trade-off between mean and variance. Accordingly, the innovative technology in open firms is dominated ifthe registering costs are not excessive (assumption (ii), upper bound).


Fig. 3. Economic growth and stock exchange atrophy.

situation is aggravated when firms are labor intensive and, thus, the profit share is relativelylow, as given in assumption (iv). This feature encourages public firms to reduce their risks byusing a conventional technology, especially, when they have a limited innovative capacity(e.g. when the ratio of mean returns is not very high, as stated in assumption (i), upperbound).

Since firms are assumed to be identical, they simultaneously decide to go public for aspecific size of the capital stock and for shared beliefs on this decision. Therefore, the stockexchange would become large, but the number of shares floating would be relatively reducedbecause entrepreneurs keep a large portion of their own shares. Thus, public companies donot develop fragmented ownership. Notice that multiple market equilibria are possible sincein the model there are credit market imperfections. This is so because the floating decisionof one firm affects the diversification potential of the market and affects the decision of otherfirms to go public. Therefore, when a sufficient number of firms are interested in floatingtheir stock, then the rational outcome is that all firms will go public. This type of analysis ispresented inPagano (1993). Here, it is simply assumed that, for each size of the company,entrepreneurs beliefs will coincide.51

In Fig. 3, the growth implications of entrepreneurs’ decisions are depicted. This Figure isbuilt by integrating the scenarios described inFigs. 1 and 2and by recalling that risk-averseordinary investors always prefer to have a diversified portfolio.52 Consequently, once firmsissue equity, the ordinary investor acquires stocks on the open market. Thus, the internalmarket equilibrium is defined by(11) and the equilibrium path by(20) with gcs

t . For the

51 Subrahmanyam and Titman describe a similar scenario with two equilibria due to the presence of positiveexternalities related to market size. In their model, increasing the size improves the informational efficiency of thestock market and thus improves the incentives for other firms to go public.52 All the conditions required to integrate the scenario presented inFig. 1are summarized in propositions L–N;

seeSupplementary datafor the formal proofs.


valuesKt < K∗2, entrepreneurs prefer a closed governance structure, and workers/investors

do not diversify.Assuming that the economy starts with a capital stockKo in Fig. 3, the representative firm

grows along the productive technology path with financial autarky (gp), as shown by the boldline, up to the point whereK∗

1 is reached. When crossing this threshold, the entrepreneursdecide to keep the firm closed and choose to switch to conventional technology, droppingdiscretely to the financial autarky lower-path (gc). Later, when firms decide to go public(K∗

2) and new offerings in the stock market flourish, the conventional technology remainsin place. Moreover, at this point, another discrete change ensues and a new lower path istaken (gcs). Hence, this latter scenario leads to the possibility that building a stock marketmight reduce economic growth when entrepreneurs in family firms are very concerned aboutownership concentration. Moreover, if the market capitalization ratio is defined with theshare value (PjKj) to capital’s gross return (RcKj) ratio, then as the economy expands alongpath(20), market prices and the capitalization ratio increase.53 Consequently, after the swiftincrease in market capitalization, when firms went public and growth fell sharply, there is,on average, a continuous increase in market capitalization ratio and economic growth. Inother words, the model exhibits a non-linear relationship between stock market size andgrowth.54

In other words, the model describes how stock market atrophy can emerge, in the sensethat the market evolves and gets larger, but the economy grows stagnant and the numberof shares really floating is relatively small. This follows from risk-averse economic agents,credit rationing and manager/owners worried about losing control. The interest of retainingcorporate control is more relevant in a society where rights are unequally protected andwhere flotation costs are relatively large. These entrepreneurs are willing to sacrifice a high-return technology just to keep the business within family control. As such, they choose a low-risk technology instead of financial diversification as a way to reduce their risks. Moreover,the expansion of an open stock market may also reduce growth because savings are allocatedaway from internal capital markets where financial resources are used productively.55

Likewise, the model highlights the potential benefit of having investors with risk-sharingconstraints.56 When firms decide to issue shares publicly, ordinary investors adopt a diver-

53 Despite equity prices in secondary markets increasing steadily along pathgcs, investors do not arrange theirstrategy accordingly given the short-term horizon assumed in the model.54 In an extrapolation of the model to a real life situation, firms are not identical and, in particular, they are of

different sizes; hence they do not go public all at once. Therefore, instead of observing a discrete drop in economicgrowth, the fall follows a continuous pattern, which is reversed later.55 Caselli and Gennaioli also present a model on dynastic management where family ownership and management

leads to inefficiency and slow growth. In their analysis, untalented family members have to run the firm whenimperfections in financial markets hamper the transfer of control. In their model, it is also the case that small firmscan be more efficient, although not due to the implementation of high-return projects as suggested in this paper, butbecause their owners are more willing to take the outside options offered by the labor market than entrepreneursin large family firms. In other words the larger the firm (and their profits) the less incentive has an untalentedmanager to sell off and the more cumbersome is to get the financing to acquire such firms. Consequently, whenthe economy grows by the dynamics of large firms, they are stuck with inefficient managers and the economy’srate of growth slows down.56 In a static sense, the possibilities of diversification obtained with the public offering are welfare improving;

however, in a dynamic sense the opposite could be the case since less diversification today could mean more


sified portfolio. This removes resources from the financing of new projects and into theacquisition of existing financial claims. The corresponding increase in the price of thecapital stock, whose supply has diminished, reduces entrepreneurs’ incentives to bequeathproductive physical assets due to an income effect.57 However, subsequent increases in thecapital stock ignite a new dynamic process, where an increase in the share of wages investedin the internal market generates a slow recovery of the economy.

5. Some stylized facts

The growth observed during the eighties and nineties in the stock market capitalizationin many developing countries was striking. As described inSingh (1997), the combinedcapitalization for 38 emerging economies in the 1983–1993 period increased from lessthan a hundred billion US dollars to nearly a trillion US dollars. For some leading emergingmarkets this increase was even more astonishing. For instance, relative to GDP, marketcapitalization rose in Mexico over the same period from 2 to 43 percent, and it rose forThailand from 3.8 to 55.8 percent.58 However, the macroeconomic performance in manyof these economies was very bleak, and in many cases the boom ended in a financialcrisis and a severe recession, as witnessed by the Tequila and South East Asian crises.59

India is a concrete case where the negative correlation between market development andreal economic activity is present, as described byNagaraj (1996). Despite the boom intheir stock markets in the eighties, increasing the ability of Indian corporations to raisefunds through the equity market, the output growth of private corporations stagnated,and their investment in fixed assets even declined. The boom largely produced a shiftin financial portfolios, from bank deposits to tradeable securities, without encouragingincreased real savings. Paradoxically, these stylized facts contradict the positive linkbetween stock market development and growth found with the first-generation studies.Irrespective that variables besides those related to financial development may also enterinto the explanation, this paper contends that the appearance of a negative correlation meritsan important reconsideration of the empirical methodology to test the finance-growthlink.

One possible way to solve this paradox, as suggested in the model above, is to considernonlinearity in the growth dynamics. One can interact the effect of an increase in marketcapitalization with a change in ownership concentration or with some other attribute ofthe market that reflects weak institutions and that induces concentration (e.g. lack of trans-

growth and hence better consumption opportunities tomorrow. With a similar line of reasoning,Japelli and Pagano(1994)show that a certain degree of credit constraint in economic households is beneficial in terms of growth andwelfare.57 From constraints in problem(3), income definition(6) and optimal investments(14), (15) and (17), it can

easily be proved that the consumption of entrepreneurs does not change with the price hike. Moreover, accordingto expression(17) the price-elasticity of productive assets is equal to minus one; thus all the increase in prices isabsorbed by a reduction in the acquisition of the capital stock.58 Similar improvements took place with regard to liquidity indicators.59 SeeArestis (2003).


Fig. 4. Real per capita GDP, annual average rate of growth (GWT, 1980–1995). Cross-tabulation between marketcapitalization (MCAP) and market liquidity (MTOR) by quartiles.

parency, reduced protection to minority-shareholders, poor liquidity).60 This test wouldprovide, presumably, evidence consistent with there being a negative correlation betweenstock market size and economic growth when there is a high ownership concentration.

Although the presentation of a formal econometric study is beyond the goal of thispaper, some stylized facts are presented below to support the main claim of the paper: the

60 According toDemirguc-Kunt and Levine (1996b)there are intuitive correlations among different stock marketdevelopment indicators. For instance, (i) large markets tend to be less volatile, more liquid, and less concentratedin a few stocks; (ii) markets that are internationally integrated tend to be less volatile, and (iii) markets supportedby institutions tend to be large and liquid. Nonetheless, it is clear from the empirical evidence that these indicatorsare not perfectly correlated, and hence one specific market could be considered developed when viewed throughthe lenses of one indicator but ranked very low when measured with another indicator.


Fig. 5. Real per capita GDP, annual average rate of growth (GWT, 1980–1995). Cross-tabulation between marketcapitalization (MCAP) and shareholder protection (SHAP) by quartiles.

increasing capitalization of a stock market may be detrimental for growth due to a highconcentration of ownership, which in turn may be related to the poor performance of otherindicators that characterize stock market development such as the turnover rate and thedegree of shareholder protection.

In Figs. 4–7, a sample of countries over the period 1980–1995 has been subdivided inquartiles for each variable, forming sixteen sub-categories. The figures make it possible torecognize that the relationship between stock market development and long-term growthis not always positive.61 Moving from quartile to quartile (Q1–Q4) along any of the axis

61 The definition and sources of the variables used for the elaboration of these figures are presented inSupplementary data. In these figures, quartiles have been constructed arranging observations by ranking onceaverages for the full period have been calculated if information is available although similar results in terms ofa non-linear relationship are also inferred when quartiles are constructed arranging observations by value (notpresented in the paper). This double check helps reduce the possibility that the evidence of multiple thresholds isjust the result of some quartiles with very few observations.


Fig. 6. Real per capita GDP, annual average rate of growth (GWT, 1980–1995). Cross-tabulation between marketcapitalization (MCAP) and ownership diffusion (OWND) by quartiles.

indicates that an increase in one of the indicators of stock market development (size, liq-uidity, shareholders protection, diffusion of ownership) may be associated with a decreasein average growth. It does appear that average growth is positively related with a combinedimprovement of stock market capitalization and some other indicator of market develop-ment, as is seen when moving from the (Q1–Q1) corner to the (Q4–Q4) corner. However,the relationship does not seem to be linear, and the presence of one of more thresholds isplausible. Furthermore, the nature of the relationship between capitalization and growthseems to depend on the level of the other indicators considered in the figures. For instance,when shareholder protection (Fig. 5) is fixed in the first and second quartile, an increasein market capitalization from the third to the fourth quartile is associated with a lowerlong-term growth. In spite of having limited observations in each cell of the matrix andthe findings not being statistically validated, the negative relationship between market sizeand average growth for some fixed level of other indicator is consistent with the paper’s


Fig. 7. Real per capita GDP, annual average rate of growth (GWT, 1980–1995). Cross-tabulation between marketcapitalization (MCAP) and external fragmentation (EXTF) by quartiles.

claim that this association depends on the status of another variable such as ownershipconcentration, minority shareholder protection, and so on.62

In Figs. 8 and 9, the sample was divided in two sub-periods (1980–1986 and 1988–1995),taking as a threshold the Black Monday crash of 1987 that spread into many countries. Thisexercise illustrates that non-linearity between stock market size and growth seems to prevailif shorter periods are taken into consideration and that it is modified through time. Thissuggests that a growth dynamic analysis would be a worthy endeavor to pursue. Furthermore,

62 Similar results not presented in the paper are obtained when the stock market development indicators areaveraged only for the period 1980–1986, while long-run GDP growth is calculated as an average of a laterperiod (1988–1995). That is, this second procedure is more in line with the assumption that the initial level in thedevelopment of equity markets (and not contemporaneous values) affects economic long-run growth. For instance,for a cross-tabulation of market capitalization and the turnover rate, average growth falls when quartiles Q1–Q3in the MTOR axis are fixed and MCAP increases from the first to third quartile.



it highlights that for the pre-crash period an increase of market capitalization from the first tothe fourth quartiles is associated with a smaller average growth for markets with a relativelylow liquidity (first and second quartiles of MTOR). However, this association becomespositive (as stated in the traditional theory of financial development) when referring to thethird and fourth quartiles for the liquidity indicator. In other words, as stated in the model,for emerging economies with inadequate regulatory frameworks that do not encourage sockmarket liquidity or the development of some other indicator for shareholder diffusion, a pureincrease in the size of the stock market might produce a slow down in economic growth.

The possibility of a negative relationship between stock market capitalization and growthfor certain intervals of the other development indicators may be inferred also when con-trolling for other variables. As a first approximation to the problem,Fig. 10presents theaverage residual growth rate, once a regression has been run between per capita GDP



growth and three variables commonly used in growth models (initial real GDP per capita, lifeexpectancy, quality of institutions). To move along the axis, indicates, again, that there mightbe some form of non-linearity between stock market development and growth. In this casethe interacted variables are market capitalization and minority-shareholders protection.63

Notice that when one compares the first and fourth quartiles of MCAP, residual growthdecreases with market size for the first and second quartiles of shareholder protection(SHAP), and when there is above average protection for shareholders, the relationshipbetween residual growth and stock market size is not clear cut, as an approach with multi-ple thresholds would suggest.

63 Although not presented in the paper, similar results are obtained when combining market size with either ofthe following variables: liquidity and ownership concentration.


Fig. 10. Real per capita GDP, annual average rate of growth (RGWT, 1980–1995), GWT based on regressionresiduals. Cross-tabulation between market capitalization (MCAP) and shareholder protection (SHAP) by quar-tiles.

All in all, from the previous evidence, it can be said that there is some empirical foundationto the statement that along certain intervals of stock market size, further increases canhamper growth. Hence the previous model explains why there may be a negative associationbetween stock market capitalization and growth when the market is ill-functioning due to ahigh ownership concentration, perhaps because market liquidity and protection to minority-shareholders are low.

6. Conclusions

Undoubtedly, the presence of a stock market in an economy increases the opportunityfor diversification. However, the family-owned nature of firms in many emerging marketsmight make manager/owners reluctant to diversify too much in a stock market and more


likely to select a lower risk technology. Therefore, conventional modes of production, orsafe government projects, with low returns are preferred means of managing risk. Moreover,in certain scenarios, the appearance of an equity market induces a reduction in capital for-mation by moving resources away from firm’s self-investment and internal capital markets.Consequently, the paper has two related theoretical contributions. First, it offers a differentview on the impact of ownership concentration in the selection of projects. Traditionally,the issue of ownership has been related to agency problems. A large concentration wouldproduce the rent extraction from other stakeholders or provide a solution to free-rider prob-lems by creating incentives to monitor and to undertake projects with a positive net presentvalue. In contrast, here, ownership concentration allows for a general unwillingness to risklosing control of the firm. The unwillingness to be fully diversified promotes the selectionof low-risk, low-productive projects. Second, the paper presents a theory on the existenceof ill-functioning stock markets that may become an obstacle to economic growth. Tradi-tionally the view is that stock market development helps economic growth. However, in themodel, it is suggested that when minority-shareholders are not well-protected and marketsare not very liquid so that share prices do not convey the needed information to improveefficiency in allocation, then the negative side effects of the formation of these marketsmight surpass their positive impact on economic performance. In particular, it is consideredthat these markets do not necessarily encourage the adoption of productive projects but mayinduce a reduction in capital formation.

The empirical implication of the theory is that under certain scenarios there could bea non-linear relationship between market capitalization and economic growth. This linkinvolves ownership concentration and the lack of formal institutions in the economy. Interms of policy-making, these results highlight the fact that certain market institutions, suchas the stock exchange, may not yield their expected outcome (that is, the outcome observedin the developed world). This is especially so if they are not accompanied by the appropriatelegal and political institutions that protect the rights of all stakeholders and if society doesnot have social norms conducive to transparent behavior.64 Although the model’s predictioncontradicts the first generation of empirical studies of the relationship between the devel-opment of the market and growth, the prediction is consistent with the second generationof empirical studies that introduced non-linear functional forms in the finance–growth rela-tionship. The first generation of studies does not consider the possibility that markets couldbe relatively large in terms of capitalization but atrophied in other dimensions. Obviously,econometric tests are required to validate empirically the non-linear implications of themodel. In those tests it is necessary to differentiate between emerging stock markets thatexhibit certain atrophy and those markets characterized by good liquidity, limited ownershipconcentration, strong contract enforcement, and reduced corruption, among other factorscommonly found in well-functioning stock markets. Dynamic panel models estimated byGMM-system are necessary to discover formally possible non-linear relationships in thegrowth dynamics and then to test the possibility of a negative effect of an increase of marketsize for certain atrophied markets.

64 This argument partly explains the drawbacks and financial crises recently observed in Russia, Latin America,East Asia and other emerging markets that attempted to introduce a market system without having the adequateinstitutional framework.


Acknowledgments

This paper is part of a project on business groups, related lending and internal capitalmarkets in Mexico. I am thankful to CONACYT, UDLA and UC-MEXUS-CONACYT forthe financial support, to Felipe Bello for his help in collecting data and elaborating thefigures, and to David Wetzell, Samuel Freije and two anonymous referees who read themanuscript and gave me very helpful comments. Nonetheless, all shortcomings and errorsin the paper are entirely of my responsibility.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, in the online version, atdoi:10.1016/j.jebo.2004.04.005.

References

Acemoglu, D., Zilibotti, F., 1997. Was Prometheus unbound by change? Risk, diversification, and growth. Journalof Political Economy 105, 709–751.

Acemoglu, D., Zilibotti, F., 1999. Information accumulation in development. Journal of Economic Growth 4,5–38.

Arestis, P., 2003. Financial sector reforms in developing countries with special reference to Egypt. The LevyEconomic Institute of Bard College, working paper 383.

Arestis, P., Demetriades, P., 1997. Financial development and economic growth: assessing the evidence. TheEconomic Journal 107, 783–799.

Arestis, P., Demetriades, P.O., Luintel, K.B., 2001. Financial development and economic growth: the role of stockmarkets. Journal of Money, Credit, and Banking 33, 16–41.

Atje, R., Jovanivic, B., 1993. Stock markets and development. European Economic Review 37, 632–640.Beck, T., Demirguc-Kunt, A., Levine, R., 2005. The database on financial development and structure. The World

Bank (data are available at:http://www.worldbank.org/research/projects/finstructure/database.htm).Beck, T., Levine, R., 2004. Stock markets, banks, and growth: correlation and causality. Journal of Banking and

Finance 28, 423–442.Bencivenga, V., Smith, B., Starr, R., 1996. Equity markets, transaction costs, and capital accumulation: an illus-

tration. World Bank Economic Review 10, 241–266.Berthelemy, J.C., Varoudakis, A., 1996. Financial Development Policy and Growth. Development Center OECD,

Paris.Bhattacharya, U., Ravikumar, B., 2001. Capital markets and the evolution of family businesses. Journal of Business

74, 187–220.Bhide, A., 1993. The hidden cost of stock market liquidity. Journal of Financial Economics 34, 1–51.Bolton, P., Von Thadden, E.L., 1998. Blocks, liquidity, and corporate control. Journal of Finance 53, 1–25.Burkart, M., Gromb, D., Panunzi, F., 1997. Large shareholders, monitoring and the value of the firm. Quarterly

Journal of Economics 112, 693–728.Burkart, M., Panunzi, F., Shleifer, A., 2003. Family firms. Journal of Finance 58, 2173–2207.Caporale, G.M., Howells, P., Soliman, A., 2004. Stock market development and economic growth: the causal

linkage. Journal of Development Economics 29, 33–50.Caselli, F., Gennaioli, N., 2003. Dynastic management. National Bureau of Economic Research, working paper

9442.Castaneda, G., 1998. La empresa mexicana y su gobierno corporativo, antecedentes y desafıos para el siglo XXI.

Universidad de las Americas-Puebla y Alter Ego Editores, Puebla, Mexico.

http://dx.doi.org/10.1016/j.jebo.2004.04.005

http://www.worldbank.org/research/projects/finstructure/database.htm


Chami, R., 2001. What is different about family businesses? International Monetary Fund, working paperWP/01/70.

Chemmanur, T., Fulghieri, P., 1999. A theory of the going-public decision. The Review of Financial Studies 12,249–279.

Cho, Y.J., 1986. Inefficiencies from financial liberalization in the absence of well-functioning equity markets.Journal of Money, Credit and Banking 18, 191–200.

Claessens, S., 1995. The emergence of equity investment in developing countries: overview. The World BankEconomic Review 9, 1–18.

De Jong, E., Semenov, R., 2002. Cross-country differences in stock market development: a cultural view. Universityof Nijmegen, Department of Economics working paper, The Netherlands.

De Long, J.B., Shleifer, A., Summers, L., Waldmann, R., 1989. The size and incidence of the losses from noisetrading. Journal of Finance 44, 681–696.

Demirguc-Kunt, A., Levine, R., 1996a. Stock markets, corporate finance and economic growth: an overview.World Bank Economic Review 10, 223–240.

Demirguc-Kunt, A., Levine, R., 1996b. Stock market development and financial intermediaries: stylized facts.World Bank Economic Review 10, 291–322.

Devereux, M., Smith, G., 1994. International risk sharing and economic growth. International Economic Review35, 535–550.

Durham, J.B., 2000. Which anomalies are robust in emerging and developed markets? Emerging Markets Quarterly4, 50–67.

Durham, J.B., 2002. The effects of stock market development on growth and private investment in lower-incomecountries. Emerging Markets Review 3, 211–232.

Easterly, W., Sewadeh, M., 2001. Global development network growth database. The World Bank (data are availableat: http://www.worldbank.org/research/growth/GDNdata.htm).

Evans, P., 1997. Consistent estimation of growth regressions. Ohio State University, working paper.Favara, G., 2003. An empirical reassessment of the relationship between finance and growth. International Mon-

etary Fund, working paper WP/03/123.Greenwald, B., Stiglitz, J.E., Weiss, A., 1984. Informational imperfections in the capital market and macroeconomic

fluctuations. American Economic Review 74, 194–199.Greenwood, J., Jovanovic, B., 1990. Financial development, growth, and the distribution of income. Journal of

Political Economy 98, 1076–1077.Haber, S., 2000. The political economy of financial market regulation and industrial productivity growth in Brazil

1866–1934. In: Haber, S. (Ed.), Political Institutions and Economic Growth in Latin America. Hoover Insti-tution Press, Stanford, CA, pp. 69–120.

Harris, R., 1997. Stock markets and development: a re-assessment. European Economic Review 41, 139–146.Heston, A., Summers R., Bettina A., 2002. Penn World Table Version 6.1. Center for International Comparisons at

the University of Pennsylvania (CICUP) (data are available at:http://pwt.econ.upenn.edu/aboutpwtto.html).Holl, P., 1975. Effect of control type on the performance of the firm in the UK. Journal of Industrial Economics

23, 257–271.Holmstrom, B., Tirole, J., 1993. Market liquidity and performance monitoring. Journal of Political Economy 101,

678–709.Huybens, E., Smith, B., 1998. Financial market frictions, monetary policy, and capital accumulation in a small

open economy. Journal of Economic Theory 81, 353–400.Ibbotson, R., Ritter, J., 1995. Initial public offerings. In: Jarrow, R., Maksimovic, V., Ziemba, W. (Eds.), Handbooks

of Operation Research and Management Science: Finance, vol. 9. Amsterdam, North-Holland, pp. 993–1016.Japelli, T., Pagano, M., 1994. Savings, growth, and liquidity constraints. Quarterly Journal of Economics 109,

83–109.Jensen, M., Meckling, W., 1976a. Theory of the firm: managerial behavior, agency costs, and ownership structure.

Journal of Financial Economics 3, 305–360.Jensen, M., Meckling, W., 1976b. Agency costs of free cash flow, corporate finance and takeovers. American

Economic Review 76, 323–339.Jensen, M., Murphy, K., 1990. Performance pay and top-management incentives. Journal of Political Economy

98, 225–264.

http://www.worldbank.org/research/growth/gdndata.htm

http://pwt.econ.upenn.edu/aboutpwtto.html


Khan, M.S., Senhadji, A.S., 2003. Financial development and economic growth: a review and new evidence.Journal of African Economies 12, 89–110.

Kyle, A., 1984. Market structure, information, futures markets and price formation. In: Storey, G., Schmitz, A.,Sarris, A. (Eds.), International Agricultural Trade: Advanced Readings in Price Formation, Market Structure,and Price Instability. Boulder Co., Westview, pp. 45–64.

La Porta, R., Lopez-de-Silanes, F., Shleifer, A., 1999. Corporate ownership around the world. Journal of Finance54, 471–517.

La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R., 1997. Legal determinants of external finance.Journal of Finance 52, 1131–1150 (data are available at La Porta’s web page:http://mba.tuck.dartmouth.edu/pages/faculty/rafael.laporta/publications.html).

Levine, R., 1991. Stock markets, growth and tax policy. Journal of Finance 46, 1445–1465.Levine, R., 1997. Financial development and economic growth: views and agenda. Journal of Economic Literature

35, 688–726.Levine, R., 2002. Bank-based or market-based financial systems: which is better? Journal of Financial Intermedi-

ation 11, 1–30.Levine, R., Zervos, S., 1996. Stock market development and long-run growth. World Economic Review 10,

323–340.Levine, R., Zervos, S., 1998. Stock markets, banks, and economic growth. American Economic Review 88,

537–557.Mas-Colell, A., Whinston, M., Green, J., 1995. Microeconomic Theory. Oxford University Press, Oxford, U.K.Maug, E., 1998. Large shareholders as monitors: is there a trade-off between liquidity and control? Journal of

Finance 53, 65–98.Maurer, N., Haber, S., 2002. Institutional change and economic growth: banks, financial markets, and Mexican

industrialization 1878–1913. In: Bortz, J., Haber, S. (Eds.), The Mexican Economy 1870–1930 Essays on theEconomic History of Institutions, Revolution and Growth. Stanford University Press, Sanford, CA, pp. 23–49.

Merton, R., 1987. A simple model of capital market equilibrium with incomplete information. Journal of Finance42, 483–510.

Milgrom, P., Roberts, J., 1988. An economic approach to influence activities in organizations. American Journalof Sociology 94, s154–s179.

Morck, R., Shleifer, A., Vishny, R., 1990. The stock market and investment: is the market a sideshow? BrookingsPapers on Economic Activity 2, 157–215.

Morck, R., Stangeland, D., Yeung, B., 2000. Inherited wealth, corporate control and economic growth: the Canadiandisease? In: Morck, R. (Ed.), Concentrated Corporate Ownership. University of Chicago Press, Chicago, pp.319–369.

Myers, S., Majluf, N., 1984. Corporate financing and investment decisions when firms have information thatinvestors do not have. Journal of Financial Economics 13, 187–221.

Nagaraj, R. 1996. India’s capital market growth: trends, explanations and evidence. Indira Gandhi Institute ofDevelopment Research working paper, New Delhi.

Obstfeld, M., 1994. Risk-taking, global diversification and growth. American Economic Review 84, 10–29.Pagano, M., 1993. The flotation of companies on the stock market, a coordination failure model. European

Economic Review 37, 1101–1125.Pagano, M., Panetta, F., Zingales, L., 1998. Why do companies go public? An empirical analysis. Journal of

Finance 53, 27–64.Pagano, M., Roell, A., 1998. The choice of stock ownership structure: agency costs, monitoring, and the decision

to go public. Quarterly Journal of Economics 113, 187–225.Radice, H., 1971. Control type, profitability and growth in large firms: an empirical study. Economic Journal 81,

547–562.Romer, P., 1986. Increasing returns and long-run growth. Journal of Political Economy 94, 1002–1037.Rousseau, P., Wachtel, P., 2000. Equity markets and growth: cross-country evidence on timing and outcomes.

Journal of Business and Finance 24, 1933–1957.Saint-Paul, G., 1992. Technological choice, financial markets and economic development. European Economic

Review 36, 763–781.Sala-i-Martın, X., 1994. Apuntes sobre crecimiento economico. Antoni Bosch, Barcelona, Spain.

http://mba.tuck.dartmouth.edu/pages/faculty/rafael.laporta/publications.html


Sargent, T., 1979. Macroeconomic Theory. Academic Press, Orlando, FL.Scharfstein, D., 1988. The disciplinary role of takeovers. Review of Economic Studies 55, 185–199.Shleifer, A., Summers, L., 1988. Breach of trust in hostile takeovers. In: Auerbach, A.J. (Ed.), Corporate Takeovers:

Causes and Consequences. University of Chicago Press, Chicago, pp. 33–56.Shleifer, A., Vishny, R., 1986. Large shareholders and corporate control. Journal of Political Economy 94, 461–488.Singh, A., 1993. The stock market and economic development: should developing countries encourage stock-

markets. UNCTAD Review, 1–25.Singh, A., 1997. Financial liberalization, stock markets, and economic development. The Economic Journal 107,

771–782.Spyrou, S.I., Kassimatis, K., 2001. Stock and credit market expansion and economic development in emerging

markets: further evidence utilizing cointegration analysis. Applied Economics 33, 1057–1064.Stein, J., 1988. Takeovers threats and managerial myopia. Journal of Political Economy 96, 61–80.Ward, J., 1987. Keeping the Family Business Healthy. Jossey-Bass Publishers, London, UK.Subrahmanyam, A., Titman, S., 1999. The going-public decision and the development of financial markets. The

Journal of Finance 54, 1045–1082.Yosha, O., 1992. Essays on the structure of financial markets, Ph.D. thesis. Harvard University.Zingales, L., 1995. Insider ownership and the decision to go public. The Review of Economic Studies 62, 425–448.

economic growth and concentrated ownership in stock markets

Documents