A crosscountry analysis of financial growth
The cross-country examination of the finance-growth nexus was initiated by Goldsmith (1969). Using data on 35 countries from 1860 - 1963, he found an evidence of a relationship between economic and financial development over long periods, and that periods of rapid economic growth have often been accompanied by an above-average rate of financial development. He reported a clear relationship between financial development and economic growth albeit statistically weak. King and Levine (1993) were of the pioneers to show the potential for panel datasets such as Beck et al. (1999). They study 80 countries over the period 1960-1989 by controlling for other factors that affect long-run growth. They found a strong predictive component in the relation in addition to consistent contemporaneous nature for the relationship between aggregate measures of financial depth and growth. They argue that current financial depth can predict economic growth over the long run of ten to thirty years. Reporting that the predetermined component of financial development is robustly correlated with future rates of economic growth, physical capital accumulation, and economic efficiency improvements. And conclude that "better financial systems stimulate faster productivity growth and growth in per capita output by funnelling society's resources to promising productivity-enhancing endeavours." (King and Levine, 1993b, p. 540.). King and Levine (1993) sat forth the stances for measuring financial development, which were widely used in later studies. In addition, Atje and Jovanovic (1993) contends a large effect of stock markets on subsequent development yet failed to find a similar effect of bank lending.
Subsequent work investigates which features of the financial system are key for fostering growth. Demirguc¸-Kunt and Levine (2001) is a cross-country comparison of banks, markets, and development. It covers a great deal of empirical work using datasets ranging from micro-level firm data to international comparative studies and employs a wide range of econometric tools.
The debate of a bank-based versus market-based financial system has taken it's share of cross-country studies.Levine (2002) cannot determine an empirical support for either the bank- or market-based but observe a vivid relation between financial development and economic growth. Demirguc-Kunt and Maksimovic (2002) draw on a contemporary survey of some 4,000 firms - small and large - in 54 countries that were asked for their perception of financing, legal and corruption constraints to their growth. The study shows that while financial development can explain the growth of firms neither bank or market based financial systems signal a difference. These results in general suggest that the type of the financial system is of secondary importance in the course of development. The studies add to other empirical work (e.g. Beck and Levine, 2002) that supports whats known as the middle-ground which assigns the importance to the functionality and not type of the financial system. In addition, Levine and Zervos (1998) suggested that equity markets and banks could be exerting complementary services to the economy.#p#分页标题#e#
La Porta et al (1998) opened the venue of legal/financial infrastructure investigation. They assembled a database on the major distinguishing characteristics of legal systems across countries varying in their legal systems as they control and impact financial contracts. In summary the major distinctions lie in the relative protection that is provided to a firm's managers, controlling shareholders and other insiders as against outsider financiers including both creditors and minority shareholders. The results suggest that the relative protection formally granted to those participants correlates with the degree of development of the relevant market. Since finance is based on contracts, legal bodies that produce laws that protect the rights of external investors and enforce those rights effectively will do a correspondingly better job at promoting financial development. Stronger shareholder rights are, on a cross-country findings, correlated with a greater number of listed firms and with higher stock market capitalization; also stronger creditor rights are associated with a higher level of bank credit and bond finance.
La Porta, Lopez-de-Silanes, and Shleifer (2002) measures the effect of public ownership of the banking sector and it's impact on economic growth. Their findings show that state ownership and control of the banking sector in the late-sixties early-seventies specially in developing countries is associated with slower subsequent growth. State ownership is found to has usually insignificant effect on future investment, but a large impact on future productivity growth.
The La Porta et al. (2002) found evidence to the distortion caused by misallocation of resources under state intervention in credit policies. Their results contradict "development" theories of state ownership that emphasize the positive effect that government can have in banking, for example by mitigating negative externalities, encouraging risk-taking investment, financing strategic sectors, etc.
Intrigued by previous cross-country findings, Aghion, Howitt and Mayer-Foulkes (2005) investigate whether financial development increases steady-state growth rates or whether it speeds up convergence to the technological frontier. The authors show that financial development is highly beneficial for converging to the technological frontier. The results imply that countries above some critical level of financial development should converge in growth rates and that in such countries finance has a positive but eventually vanishing effect on steady-state GDP.
The following table provides and over view of a number of cross-country studies.
The advent of time series analysis suffered from sufficiently long time-series of series national accounts data in developing economies. Gupta (1984) was one of the first to tackle this problem in the realm of time series analysis. He utilised data on industrial output to measure the level of economic development. Such data was utilized on quarterly frequency providing a larger frequency than is typically available in national accounts data that was mainly offered in an annual frequency. Data drown from 14 developing countries and the regression analysis revealed a causality link from financial development to economic growth. Gupta (1984) was seen to have inherit limitations, first because of the use of the industrial output which is only a small component of overall output. Second, the fact span of data is considered of much more importance than the number of observations ( Demetrious and Hussein 1996).#p#分页标题#e#
In relation to this first controversy of the direction of the causality relation, Patrick (1966), points the two concepts: demand following and supply leading. In accordance with the first concept, financial institutions and services are created from out of the need to support investment. The financial system follows the economic growth which generates an additional new demand for services. This in turn leads to the financial development as reply to investors demand. The second concept presents the inverse relation and suggest a reverse course of causality. Supply leading, occurs when the creation of financial institutions and the supply of financial services acts as a promote of further economic growth. In this concept, the financial development is not a precondition to initiate a self-supported economy, but it represents a chance to promote the real growth by means of financial instruments.
Patrick (1966) suggested that the reality is an interaction between the two phenomena. In accordance with his findings, supply leading must induce the initial modernizing growth by transferring resources from traditional sectors to modern sectors; however, demand following must assume a gradual importance when higher growth creates more needs for financial services and modern financial institutions. Jung (1986) conducts Granger causality tests for 56 countries from 1950 to 1981. While the results provide more support for the supply-leading hypothesis, they yield inconclusive results for reverse temporal causality patterns.
Neusser and Kugler (1998) report that financial sector GDP Granger-caused manufacturing sector GDP in a sample of thirteen OECD countries which is in line with the supply-leading view that finance plays an important role in economic development.
On the other hand, Demetriades and Hussein (1996) use the stationarity tests and cointegration analysis, while most other prior studies had used least squares estimation. The study tested the existence of a stable long-run relationship between real GDP per capita and the proxies adopted for financial development which are bank deposit liabilities to nominal GDP and the ratio of bank claims on the private sector to nominal GDP in a sample of 16 developing countries with at least 27 annual observations. With results showing substantial variation across countries, Demetriades and Hussein find that the relationship between financial development and economic growth tends to be bi-directional. Their findings also suggest that there is little or no evidence that causality is unidirectional from finance to GDP growth whilst the opposite seems to be the case for some of the sample countries.
In perusing the reasons behind the variation of causality direction across tested sample , Arestis and Demetriades (1996) provide several accounts. Firstly, it is due to the variation in financial systems and institutional structures that the relation differ in it's magnitude and direction. Secondly, financial sector policies play a determining role in the extent to which financial development fosters economic growth. Thirdly, even if we assume identical financial systems and financial sector policies there still a room for the extent of the effectiveness of those institutions when they design and implement the policies. In effect, it shows the finance growth relationship is driven by both factors affecting each other. Arestis and Demetriades also suggests that in the case of developed economies credit-based indicators are more likely to exhibit a stable long-run relationship with output than deposit-based ones. Therefore, the finance-growth nexus is largely determined by the nature and operation of the financial institutions and policies pursued in each country.#p#分页标题#e#
Luintel and Khan (1999) finds bi-directional long-run causality link between financial development and economic growth in a sample of 10 countries with at least 36 years of data using a multivariate vector autoregression. Attributing their findings to i. analysis of a higher dimensional system, ii. A new method of identifying the long-run economic relationships, and iii. a new approach to long-run causality testing. Thus supporting the findings of Demetriades and Hussein (1996).
Arestis, Demetriades and Luintel (2000) find additional support for the view that finance stimulates growth but raise some cautions on the size of the relationship. Using quarterly data and apply time series methods to five developed economies and substantially augmenting time series. Their findings suggest that banking sector and stock market development contribute to subsequent growth, with significant weight given to the first. The fact that the sample size was small casts a hint of limitation on the study ( Levine 2004).
Further to Arestis to the investigation of stock market development vs banking sector development and the link to growth, Arestis et al. (2001) rigorously compare and contrast the
country-specific financial structure, that is, an assortment of financial markets, instruments and intermediaries in operation, and conclude that financial structure is important for economic growth. Their findings include that banks are more powerful in promoting economic growth and emphasise the superiority of bank-based systems with clear implications for developing economies. They argue that the role of stock markets has been overemphasized by cross-country studies and further to it, showing that one in two of the five developed economies examined, stock markets tend to have negative effects on economic growth.
The issue of the direction of the causality link was revisited by Ang and McKibbin (2007) who conduct multivariate cointegration and several causality tests in the small open economy of Malaysia. The findings suggest that output growth causes financial development in the long run in contrast to the prevailing view of a finance-to-growth direction of causality. Although the country has more features of a bank-based financial system, the findings did not specify that this form of system has a significant contribution to growth in the long-run.
Time-series methodology was also applied to investigate economic views on financial repression policies vs liberalization. Demetriades and Luintel (1997) find that financial repression in India had large negative effects on financial development, over and above the retarding influence of low real rates of interest.
In a consecutive study testing for South Korea, Demetriades and Luintel (2001) provide time-series evidence but this time obtaining a contrary findings in which an index of financial repression is found to have a large positive impact on financial development. Using a monopoly-bank model they show that mild repression of lending rates increases the amount of financial intermediation. The authors attribute these variations in results as reflecting institutional differences and differences in the severity of repression. The presence of a sound institutional system and mild financial repression according to their findings, may bare positive effects under certain conditions. Arestis et al. (2002) collected data on a number of financial restraints from six developing economies finding that the effects of financial policies vary considerably across the sample of countries. In their conclusion suggested that financial liberalisation is a much more complex and should not be handled with sweeping generalizations.#p#分页标题#e#
The main findings of the time series studies are provided in Table 2.
There has also been a movement away from applying time-series methods to a variety of
countries and toward specific country studies. This allows the focus on country specific measures of financial development and expand the time-series dimension of the analyses in some cases. Focusing in a single country is generally believed to give more confidence that the finance-growth correlation is not driven by the difficulty to control for country characteristics, such as social capital, property right protection, law and regulation. This bypasses some of the shortcomings of cross-country studies which inherently suffer from high correlation between financial, institutional, legal and regional factors, which makes it difficult to identify and hardly serve a broad-brush picture for the effect of finance on growth.