Ce numéro ou un abonnement.Ajouter au panier Ajouter au panier - Revue économique
|Revue économique 2008/6 (Vol. 59)||22 €|
Versions papier et électronique : le numéro est expédié par poste.
Il est également accessible immédiatement en ligne.
|Abonnement particuliers 2013||84 €|
Tous les numéros en ligne sont immédiatement accessibles.
ATTENTION : cette offre d'abonnement est exclusivement réservée
aux particuliers. Pour un abonnement institutionnel, veuillez
vous adresser à l'éditeur de la revue ou à votre agence d'abonnements.
Vous consultezFinance for GrowthDoes a Balanced Financial Structure Matter?
AuteursLucía Cuadro-Sáez[*][*] Banco de España, International Economics Division c/ Alcala...
suite du même auteur
Alicia García-Herrero[**][**] Banco Bilbao Vizcaya Argentaria, S. A. (bbva) Kong Kong Branch,...
suite du même auteur
Although Bagehot  and Schumpeter  had already introduced the idea, the financial sector started to be considered important for economic growth more recently than capital and labor or technology. Goldsmith  and McKinnon  were the first to argue that the manipulation of the financial sector to achieve development goals was undesirable and that flows of saving and investment should be decentralized in an open capital market operating with market-determined interest rates.
2 Since the early 1990s, a growing body of empirical literature, starting with King and Levine [1993a, b], has showed that financial development leads to growth. Among the different components of the financial system, the banking system has been analyzed most widely finding that a larger share of bank deposits, bank assets or bank credit to the private sector promote economic growth, after controlling for endogeneity. Equity markets have also received attention since they constitute and alternative channel of financing, particularly for large enterprises. Research on other sectors, such as the bond market, is scarce probably due to the data limitations.
3 An interesting and long-standing question is which financial structure –oriented toward the banking sector or the capital markets– performs better. The divergent opinions as to which financial structure is preferred are based on a number of arguments. On the one hand, banks constitute the best means to mobilize capital, identify good projects, monitor managers and manage risk (Levine ). They also maintain the incentives for individual investors to acquire information, since they form long-run relations with firms (Boot et al. ) and information is not made public as in well-developed capital markets (Stiglitz ). On the other hand, deeper capital markets enhance risk management and corporate control (Levine and Zervos ). In addition, they can avoid excessive power concentration in banks’ hands, which could allow them to extract informational rents and protect firms with close bank ties from competition (Hellwig ; Rajan ). Capital markets are also better at fostering innovative but risky projects which would lead to higher growth if successful. Finally, Bencivenga et al.  show that more liquid stock markets reduce the disincentives to invest in long-duration projects because investors can easily sell their stake in the project in they need their saving before the project matures.
4 Most of the existing studies find that neither a bank-based financial structure nor a market-based one is clearly preferred (Rajan and Zingales ; La Porta et al. ; Beck et al. ; Levine et al. ; Beck et al. ; and Levine ). Ergunor , in turn, shows evidence that capital market...
suite This had to hands-off policy recommendation (Levine , is probably the best example), namely that economic authorities should not aim at a specific financial structure but only at developing the financial system in whatever way.
5 The question we pose ourselves in this paper is related but slightly different: While developing only banks or only capital markets does not seem to bring more growth, the question that still remains is whether a balanced mix of the two might be preferable to extreme solutions. There are several reasons why this may be the case. One is that one market could serve as an alternative source of finding –i.e., as “spare wheel” using Greenspan’s parallel– if the other market is under stress. See Greenspan . ...
suite The other is that the two markets seem to influence economic growth through different –but complementary– channels.
6 The definitions of financial structure which have been used until now cannot really address such question because they do not really use measures of “balancedness” of the financial structure but, rather, whether the banking system or the capital market dominates. Furthermore, their concept of financial markets is very restrictive: first they only incorporate domestic sources of financing and, within the domestic capital markets, it only includes the stock exchange.
7 We build upon the existing literature by providing an appropriate measure of how balanced a country’s financial structure is. We, then, use this measure to test whether it contributes to economic growth controlling for other determinants of economic growth. Our a priori, based on the idea that complementarities should exist between bank and capital markets, is that a balanced financial structure should contribute to economic growth. Our results, based on 143 countries, do support that a priori. Our policy conclusion is, thus, quite different from the consensus one, namely that economic authorities should foster the growth of the banking system and the capital markets in a balanced way.
8 The paper is structured as follows. The next section describes our indicator of financial structure and the data used. Section 3 sets out the empirical strategy and the results and Section 4 concludes.
9 In order to account for the differences in financial structure, several measures have been developed in the literature. The most common measure of financial structure is the logarithm of the ratio between the activity or size of stock market relative to the size of banks. Regarding the stock market, Levine  and Ergunor  take the stock market turnover, i.e., the total value of domestic equities traded on domestic exchanges while Levine  also uses the value of all listed shares divided by gdp. As for the size of the banking system, the usual measure is credit to the private sector but also total banks’ assets as in the case of Demirguc-Kunt and Detragiache . Formula 1 depicts such indicator of financial structure.
There are several problems with such indicator, at least when used to measure how balanced a country’s financial structure is. First, the sources or financing included are quite limited: financing by foreign investors and the domestic bond market are excluded. Second, being the natural logarithm of a ratio, the indicator is neither bounded nor linear. Table 1 illustrates this point more carefully. Different sizes of the banking system and the stock market are entered into the formula. The first important problem is that the indicator equals infinite (or minus infinite) when the size of one of the two markets is zero. The second one is that an increase in the stock market size relative to the banking sector has a different impact on the indicator depending on the initial size of the markets’ sector. More specifically, the impact of an increase in the market size for countries with low levels of capital markets will be stronger than the impact of the same increase in market size for countries with bigger capital markets. Such non linearity can certainly affect the empirical analysis.
Table 1 - Traditional indicator of financial structure
10 We construct a new indicator of “balancedness” of the financial structure, which tackles the above caveats. First, it includes more sources of financing, both external and domestic. Second, it is bounded and more linear than the previous one.
11 More specifically, our measure of the financial structure is the absolute value of the distance between the size of banks and markets relative to their joint size. That is:
13 Such indicator decreases the more balanced the financial structure. The minimum value, which is zero, stands for a banking system of equal size of the bond and stock markets together. Table 2 illustrates, for different sizes of the banking system and the capital markets, that this indicator is bounded and more linear.
Table 2 - A new indicator of “balancedness” of the financial structure
14 As an example, we assume a constant size for the banking system (say 45% of gdp) and then consider an increase of 4% in the ratio of capital markets to gdp. Then, for a very small size of the capital markets initially (for instance, 1%), a 4% increase would raise the traditional measure of financial structure by 1.6, whereas for a higher initial level of capital markets (for instance 21%), such 4% increase would raise it by only 0.2 (see Figure 1). In turn, our indicator of “balancedness” of the financial structure would decrease by – 0.2 and – 0.1, respectively since the increase in the markets’ size help get closer to the equilibrium between banks (at 45%) and markets (which move from 1% to 5% in one case and from 21% to 25% in the second case).
15 The second advantage of our indicator is much more comprehensive coverage of the sources of financing. First, it includes the bond market as an additional important component of domestic capital markets. Second, it incorporates financing from abroad, both from foreign banks and foreign capital markets.
16 Going in more detail on the data used for our indicator of financial structure, we measure the size of the banking system in a specific country as the sum of domestic credit from deposit money banks and other financial institutions to the economy as a whole and the country’s borrowing from international banks. The first is drawn from the imf International Financial Statistics (ifs) and the second from the bis International Consolidated Banking Statistics. Figure 2 offers a snapshot of data definitions and sources. The size of the capital markets is measured by the size of the domestic stock market and the bonds (private and public) which are outstanding and have been issued domestically. These data are drawn from the World Bank and the bis, respectively. In addition, we included the bond outstanding abroad from bis statistics. Unfortunately, there is not enough cross-country information to include the financing in foreign stock markets.
17 Apart from constructing a broad indicator of the “balancedness” of the country’s financial structure, which includes domestic and external financing, we also calculate a similar measure for domestic financing only. This will allow us to compare our results with previous ones in the literature although our measure is still more comprehensive as it includes the domestic bond market. For illustrative purposes, Figure 3 depicts the bivariate relation between our measure of “unbalancedness” of the financial structure against the income per capita each of the countries included on our analysis whereas Figure 4 shows the relation between “unbalancedness” and growth.
18 Still, the specific form of these two relationships remains an important question. Thus, both Figures 3 and 4 also include a non-parametric estimation of the corresponding relationship. Although exploring the specific form of the relationship between economic growth and the “unbalancedness” of the financial system goes beyond the scope of this paper, from Figure 4 one can observe that assuming a linear relationship, as we do, simplifies the analysis at a small cost. Of course, there could be potential threshold effects between economic growth and the “unbalancedness” of the financial structure, which would deserve been analyzed but which we leave for future research. In fact, from Figure 4 one can derive that an improvement in the “unbalancedness” of a highly “unbalanced” financial system, that is, higher growth in the more shallow market (banking or capital market) makes the financial system has a much larger impact on growth than the same improvement on a rather balanced financial system.
19 In order to evaluate whether the financial structure affects growth, we need to control for other potential determinants of growth. To that end, we borrow from the endogenous growth literature and test various sets of conditioning information.
20 The narrow set contains measures of initial income, human capital, health and the size of the financial system. For the first, we include the logarithm of initial real per capita gdp. For the second, we take the logarithm of the gross enrolment ratio for secondary education and for the third we use the logarithm of the life expectancy. For the fourth, we have two different definitions for the two different specifications of financial structure: for the broadest one, we include credit granted to the private sector both by the domestic banking system and international banks, as a percentage of gdp. For the narrower measure of financial structure, we only include domestic bank credit to the private sector, as a percentage of gdp.
21 The full conditioning information set contains the previously mentioned variables as well as other macroeconomic variables, such as the logarithm of one plus the rate of inflation, the logarithm of government expenditure as a share of gdp and the logarithm of exports plus imports as a share of gdp. Finally, the institutional characteristics of the country are proxied by an indicator of investment profile The risk rating assigned to the investor profile depends...
suite, in line with previous work by Levine, Loyza and Beck . A short description of all variables is shown in Appendix 1. Tables I and II in Appendix 2 show the statistical properties and the bivariate correlations of the explanatory variables included. Country averages are calculated for the period 1985 to1995, which will be later used for the cross section regression.
22 All in all, we have yearly data for 143 countries, 115 of which are emerging economies. The time frame for which we intend to explain economic growth is ten years, namely from 1991 to 2001.
23 We use two different –but complementary– empirical strategies to assess whether a more balanced financial structure is associated with higher economic growth.
24 We, first, focus on the medium term with a cross section analysis à la Barro (Barro ). We, thus, calculate the average yearly growth rate between 1991 and 2001 and regress it on the initial per capita income (i.e., that of 1990) to minimize endogeneity problems. In the same way, the rest of regressors are taken as the average over the period 1985-1995 except for financial structure variables and investment profile where the average is calculated for a shorter period, 1990-95, due to lack of data.
26 Second, we look into the short term dynamics by estimating the same model with panel data. We include random effects to account for unobserved heterogeneity and we estimate the model using a Maximum Likelihood Estimator.
28 Both in the cross-section and in the panel specification (Tables 3 and 4, respectively below), we find a highly significant and negative coefficient for our financial structure indicator. This means countries where the banking system is of similar size that the capital markets (measured in terms of the stock exchange and the bond market) tend to grow faster, other factors given.
29 This is true when only local sources of financing are included in the definition of financial structure (column 1) or also external sources (columns 2-4). The result is also robust to increasing the number of controls (from the narrow to the full information set).
30 The results found for the control variables are in line with the existing literature. First, a lower initial income is associated with higher economic growth in all model specifications (cross-section and panel). This implies that countries tend to converge in income per capita terms. Second, a higher life expectancy is associated with faster economic growth in all model specifications. Third, better institutions, measured by the investor profile, are positively and significantly associated with economic growth. Fourth, inflation seems to hamper economic growth in the panel specification but it is not significant in the cross-section one.
Table 3 - Cross section analysis for medium-term growth
31 The results are weaker for the size of the government sector, openness to trade and the size of the financial system. However, when significant, the sign is the expected one. It is interesting to note that the size of the financial system, measured as bank credit granted to the private sector, is significant only in the cross section but not when controlling for the quality of the institutions related to finance, namely the investment profile. The results are basically the same when using a broader...
suite Although a more detailed analysis would be warranted, these results bear an important implication for the literature on finance for growth, namely that it is not so much the size that matters but rather the composition of the financial sector and the institutions behind. Finally, our proxy for human capital, namely secondary education, is never significant. This is probably associated with the high correlation between secondary education and life expectancy (over 80%).
32 We, then, move to assessing whether there are differences between higher and lower income countries. We use the World Bank country classification to split the sample into high and upper-middle income countries, and low and lower-middle-income countries. In both cases, a more balanced financial structure is associated with higher economic growth (see Tables 5A and 5B). However, the coefficient is significant for the domestic financial structure in the case of higher income countries and for the total one (i.e., including foreign financing) for lower income ones. This result may be explained by the fact that lower income countries are generally more dependent from foreign financing that higher income ones.
Table 4 - Panel data regression analysis: Baseline exercise
Table 5.A - Panel data regression analysis for High & Upper middle income countries
33 Finally, we perform several robustness exercises. First, we account for the fact that some outliers –i.e. countries which are growing specially fast or slowly– could be driving our results. We, thus, drop the upper and lower 5% of the distribution for economic growth. The results are in line with those of the baseline analysis (see Table 6).
34 The second robustness exercise is related to the potential collinearity between education and life expectancy. We test whether results vary when dropping the variable education and we find no significant changes. We keep the variable education in the benchmark since it is a well-known determinant for growth and we prefer to have a model with larger information content (see Table 2).
Table 5.B - Panel data regression analysis for Low & Lower middle income countries
35 The next robustness exercise tackles issues related to autocorrelation and heteroskedasticity across panels, which would prevent us from using random effects. Namely, the disturbances could be heteroscedastic and contemporaneously correlated. To check the robustness of our baseline results, we use panel corrected standard error estimates and we obtain the same results (see Table 4).
Table 6 - Robustness exercise: excluding outliers (below 5% and above 95% growth rate)
Table 7 - Robustness exercise: excluding education as control variable
Table 8 - Robustness exercise: controlling for heteroscedasticity and autocorrelation across panels Panel corrected standard errors estimation
36 In this paper we explore empirically a long-standing question in the literature on finance for growth, namely which financial structure –in terms of the size of the banking system relative to the capital markets– performs better in terms of economic growth.
37 We build upon the existing literature by constructing a new measure of the “balancedness” of the financial structure. Compared to previous indicators, ours has two important advantages; first it is more comprehensive as it includes external financing and the domestic bond market. Second, it is bounded and more linear.
38 Using two different econometric specifications (cross section and panel) for a group of 143 countries for the period 1991 to 2001, we find that a more balanced financial structure –in terms of the size of banks relative to the capital markets– is associated with higher economic growth. This is true not only when the domestic financial structure is considered but also the total one (domestic + external). The results are robust to different sets of control variables and robustness tests.
39 Our findings point to a complementary role of banks and capital markets in fostering economic growth. This might be because one can serve as a “spare tyre” of the other in times of stress, borrowing from Greenspan’s metaphor, or simply because they perform different functions so that one market cannot reach the same clients as the other market.
40 Although our results are still preliminary to draw strong policy conclusions, they go in the direction of a encouraging a more hands-on approach from the part of economic authorities to foster a balanced financial structure. Given the preponderant role of the banking system, particularly in emerging countries, this means fostering the use of capital markets for financing. Several Asian countries, including China, are moving in that direction mainly through the stock market. Latin American countries also are but mostly through the bond market. Within the narrow scope of our paper, no difference is made between the two as long as they grow more than the financing through the banking system while the structure is still unbalanced.
Variables and Sources
Table I - Main summary statistics. Country average data 1991-2001
Table II - Pairwise correlations
Bagehot W. , Lombard Street, London: William Clowes and Sons.
Barro R. , “Economic Growth in a Cross Section of Countries,” The Quarterly Journal of Economics, 106: 407-443.
Beck T., Demirgüç-Kunt A. and Levine R. , “Legal Theories of Financial Development,” Oxford Review of Economic Policy, 17: 483-501.
Beck T., Levine R. and Loayza N. , “Finance and the Sources of Growth,” Journal of Financial Economics, 58: 261-300.
Bencivenga V., Smith B., and Starr R. , “Transaction Costs, Technological Choice, and Endogenous Growth,” Journal of Economic Theory, 67: 153-177.
Boot A., Greenbaum S. and Thakor A. , “Reputation and Discretion in Financial Contracting,” American Economic Review, 83: 1165-1183.
Demirgüç-Kunt A. and Detragiache E. , “Financial Liberalisation and Financial Fragility,” in B. Pleskovic and J. Stiglitz (eds), Annual World Bank Conference on Development Economics 1998, Washington, dc: World Bank, pp. 303-331.
Demirgüç-Kunt A. and Levine R. , “Bank-Based and Market-Based Financial Systems: Cross-Country Comparisons,” The World Bank Policy Research Working Paper Series, No. 2143.
Ergunor O. , “Financial System Structure and Economic Development: Structure Matters,” Federal Reserve Bank of Cleveland, Working Paper 03-05.
Goldsmith R. , Financial Structure and Development, New Haven (ct): Yale University Press.
Greenspan A. , “Do Efficient Financial Markets Mitigate Financial Crises?,” Speech at the 1999 Financial Markets Conference of the Federal Reserve Bank of Atlanta, Sea Island, Georgia, October 19.
Hellwig M. , “Banking, Financial Intermediation and Corporate Finance,” in Giovannini A. and Mayer C. (eds), European Financial Integration, Cambridge: Cambridge University Press, pp. 35-63.
King R., and Levine R. [1993a], “Finance and Growth: Schumpeter Might Be Right,” Quarterly Journal of Economics, 108: 717-737.
King R., and Levine R. [1993b], “Finance, Entrepreneurship and Growth: Theory and Evidence,” Journal of Monetary Economics, 32: 513-542.
La Porta R., Lopez-de-Silanes F., Shleifer A., and Vishny R. , “Investor Protection and Corporate Governance,” Journal of Financial Economics, 58: 3-27.
Levine R. , “Financial Development and Economic Growth: Views and Agenda,” Journal of Economic Literature, 35: 688-726.
Levine R. , “Bank-Based versus Market-Based Financial Systems: Which Is Better?,” Journal of Financial Intermediation, October 2002b, 11: 398-428.
Levine R. and Zervos S. , “Stock Markets, Banks and Economic Growth,” American Economic Review, 88: 537-558.
Levine R., Loayza N. and Beck T. , “Financial Intermediation and Growth: Causality and Causes,” Journal of Monetary Economics, 46: 31-77.
McKinnon R. I., Money and Capital in Economic Development, Washington (dc): Brookings Institution, 1973.
Rajan R. , “Insiders and Outsides: The Choice between Informed and Arm’s-Length Debt,” Journal of Finance, 47: 1367-1400.
Rajan R. and Zingales L. , “Financial Dependence and Growth,” American Economic Review, 88: 559-587.
Schumpeter J. , The Theory of Economic Development, Editorial Oxford University Press, 1961.
Stiglitz J. , “Credit markets and the control of capital,” Journal of Money, Credit and Banking, 17: 133-152.
[ * ] Banco de España, International Economics Division c/Alcala 50, 28014 Madrid, Spain. E-mail: lucia. cuadro@ bde. es.
[ ** ] Banco Bilbao Vizcaya Argentaria, S.A. (bbva) Kong Kong Branch, 43F Two International Finance Centre, 8 Finance Street, Central, Hong Kong. (When the paper was written, Alicia García-Herrero was affiliated with the Bank of International Settlements (bis)). E-mail: alicia. garcia-herrero@ bbva. com. hk.
This paper represents the personal views of the authors and not necessarily those of Banco de España, the bis or bbva. The authors gratefully acknowledge helpful comments from Jean-Paul Pollin, participants at seminar in the Kiel Institute for World Economics and an anonymous referee. All remain errors are the authors’.
 Ergunor , in turn, shows evidence that capital market development promotes economic growth relatively more than a bank-oriented financial system, as long as countries have flexible judiciary systems.
 See Greenspan .
 The risk rating assigned to the investor profile depends on contract viability/expropriation, profits repatriation and payment delays. The index rages from 0 (very high risk) to 12 (very low risk).
 The results are basically the same when using a broader definition of the size of the financial system, which includes the capital markets. Results are available upon request.
Dans cet article, nous étudions empiriquement le problème, bien connu, de l’influence de la finance sur la croissance et plus précisément la question de savoir si la structure du système financier affecte la croissance économique. En nous fondant sur la littérature existante, nous construisons une nouvelle mesure de cette structure, qui est plus large dans la mesure où elle inclut le marché obligataire ainsi que les financements extérieurs. De plus, cette mesure est bornée et plus linéaire que celles actuellement existantes. Nous trouvons qu’une structure financière plus « équilibrée », en terme de taille relative des banques par rapport aux marchés de capitaux, est associée à une croissance économique plus forte. Ce résultat suggère que les banques et les marchés sont plus complémentaires que substituables. Il rejoint l’idée de Greenspan selon laquelle chacune de ces sources de financement peut servir de « roue de secours » à l’autre.
In this paper we explore empirically a long-standing question in the literature on finance for growth, namely whether the financial structure –in terms of the size of the banking system relative to the capital markets– matters for economic growth. We build upon the existing literature by constructing a new measure of the “balancedness” of the financial structure which is broader, as it includes the domestic bond market as well as external sources of financing. It is also bounded and more linear than existing ones. We find that a more balanced financial structure –in terms of the size of banks relative to the capital markets– is associated with higher economic growth. Such finding points to banks and capital markets being more of a complement than a substitute. This is in line with Greenspan’s idea of one market serving as “spare wheel” of the other.
JEL Codes: O16, G15, G21
PLAN DE L'ARTICLE
- A New Measure of Financial Structure and Data Issues
- Empirical Strategy and Results
- Appendix 1
- Appendix 2
POUR CITER CET ARTICLE
Lucía Cuadro-Sáez et Alicia García-Herrero « Finance for Growth », Revue économique 6/2008 (Vol. 59), p. 1075-1096.
URL : www.cairn.info/revue-economique-2008-6-page-1075.htm.
DOI : 10.3917/reco.596.1075.