The Contribution of Financial Development to Income Levels in Sub-Saharan Africa
Draft Copy
VINAY PRASANDJEET NUNDLALL
InternationalBusinessSchool
BrandeisUniversity
Waltham
MA 02452-9110
USA
ABSTRACT
We investigate the contribution of financial development to income levels in a sample of Sub-Saharan African countries over the period 1963-1992. After controlling for capital stock, human capital and quality of institutions, we find evidence that financial development is positively and significantly related to income per capita for countries that have adopted British Common Law as opposed to French Civil Law. The results also explain the cross-country variations in level of financial development.
I.Introduction
We investigate the effect of financial development on income levels in Sub-Saharan Africa (SSA) in this paper using panel data. Applying a hierarchical model, we allow for the slopes of financial development to differ according to the persistent characteristics of countries such as quality of institutions and legal origins. Controlling for human capital development and physical capital stock, we find that financial development can significantly affect GDP per capita. The hierarchical model allows us to further find the determinants of the slope of financial development. We find evidence that financial development positively and significantly affect GDP per capita if the legal origin of the country is British as opposed to French.
Section II summarizes the theory behind the relationship between financial intermediation and economic growth. The discussion condenses the argument that financial systems arise in order to mitigate transaction and information costs. We also discuss some of the most important findings in the literature, particularly those that are more relevant to this study. Section III discusses the data and in Section IV we turn to the methodology. We present our results in Section V and Section VI concludes.
II. Theoretical Framework
Pioneering work by King and Levine (1993), Levine (1997), Levine and Zervos (1998) and Levine, Loayza and Beck (2000), has spawned a growing literature on the positive role of the financial system on growth. The idea however can be traced as far back as Bagehot (1873, in Levine et al. (2000)) who stated that technological innovations, an important factor for growth, rely on external funds to come to fruition. Schumpeter (1912) believed that the ability of good banks to single out potentially successful entrepreneurs and to fund them could spur technological innovation.
In the Arrow-Debreu (1964) state-contingent framework with no information and transaction costs, there is no need for a financial system to facilitate transaction and risk management. However with increasing information and transaction frictions, an uncertain environment is created where financial institutions and markets need to step in to facilitate allocation of resources across time and space. Levine (1997) calls this the primary function of the financial system and breaks it down into five basic functions:
(i)facilitating the trading, hedging, diversifying and pooling of risk,
(ii)allocation of resources,
(iii)monitoring of managers and exerting corporate control,
(iv)mobilizing savings and
(v)facilitating the exchange of goods and services.
The author identifies and examines two channels through which each function affects growth: capital accumulation and technological innovation.
Financial intermediaries can provide insurance to savers through demand deposits and a mixture of liquid and illiquid investments. They simultaneously encourage long-run investment in high return projects. Banks can spur growth by increasing investment in illiquid, high-return assets through the elimination of liquidity risk (Bencivenga and Smith, 1991). On the other hand, the ambiguous effect of liquidity on saving rates and growth is well documented. While higher liquidity increases investment returns and lowers uncertainty, income and substitution effects ambiguously affect saving rates.Jappelli and Pagano (1994) show that with physical capital externalities, saving rates could fall low enough so that growth decelerates with higher liquidity.
Financial institutions, such as banks, also specialize in trading, pooling and diversifying risk. Risk diversifying services induce a portfolio shift in favour of high yield projects, which can affect long run economic growth. Risk diversification can also affect technological change as holding a diversified portfolio of innovative projects will encourage investment in high growth activities. Hence, financial systems may also spur technological change and economic growth through risk diversification.
Diamond (1984) has explained the rationale for the emergence of financial intermediaries in the face of information costs. High information costs prevent capital from flowing to its best destination, as savers, who are the providers of funds, do not want to invest under uncertainty. The fixed cost nature of acquiring information on investment opportunities makes it economically more efficient to have a financial intermediary rather than a number of individuals acting on their own accord in the fund allocation sector.
After the financing activity, financial institutions are very important in enforcing rules, monitoring firms’ managers and exerting corporate control. Since banks mobilize savings from many depositors and then channel funds into projects, there is a ‘delegated monitor’ arrangement. A borrower is in this way monitored only by an intermediary, not by all savers. Further, corporate control allows the separation of ownership from management of the firm, which in turn allows specialization according to the principle of comparative advantage. Lower monitoring costs would thus foster more efficient investment.
Mobilizing savings is a momentous and costly task as firms have to convince savers to entrust their savings unto them. Financial intermediaries can do this better by pooling savings – and then investing thesefunds in multiple firms. For example, banks eschew the bilateral contracts firms would have to enter with each saver if they had to borrow individually. Levine (1997) also argues that the financial system can promote specialization because of the increasing number of transactions and thus reduce transaction costs. He uses the argument that in the 19th century the financial system facilitated technological innovation because exchange was a great improvement over barter.
As noted earlier, there is some skepticism in the theoretical literature over the role of finance over economic growth. Bencivenga and Smith (1991) note that higher returns from more efficient allocation of funds could depress savings rate and hence hamper growth. Lucas (1988) further counters by saying that economists have badly over-stressed the contribution of the financial system. Robinson (1952) has also expressed doubts over its influence on the economy, hypothesizing that banks respond passively to economic growth.
In empirical work, Patrick (1966) and Goldsmith (1969) are the earliest to uncover a positive correlation between financial development and growth. However, the question of causality is not properly addressed. Patrick theorizes that causality can be either supply-leading or demand-following. McKinnon (1973) and Shaw (1973) specifically address the supply-leading hypothesis and recommend governments to liberalize their financial sector in order to spur growth. More recent studies like Jung (1986) delve into the time series aspect of the problem. Using bivariate causality tests to detect temporal patterns in causality, Jung does not find any clear evidence of the direction of causality. Xu (2000) finds a negative relationship between bank-based financial development and growth in 14 middle and low income countries (mostly African), but finds significant positive long run effects of financial development on growth in 27 other countries. Wachtel and Rousseau (2000) show that banks and stock market development both explain growth. Arestis, Demetriades and Luintel (2000) use quarterly data from five OECD countries and find that banks and stock markets both cause growth, but that the effect of bank lending is larger on growth.
More recent studies by Beck and Levine (2002, 2004) and Beck, Demirguc-Kunt and Levine (2002) stress the importance of legal factors and institutional preconditions in influencing financial development. They find that countries that have inherited British Common Law (countries that were colonized by Great Britain) tend to have better performing financial system. Bordo and Rousseau (2006) also find evidence of a relationship between a country’s legal origin and its financial development. La Porta, Lopez-de-Silanes, Shleifer and Vishny (1999) (henceforth LLSV) using dummy variables to distinguish countries by their legal tradition, manage to explain the wide variation in cross-country financial development.
Existing institutions also seem to have a lasting, persistent effect on the functioning of the financial system. A new literature suggests that the existence of institutions that protect property rights is a precondition to sound development. Acemoglu, Johnson and Robinson (2001) (henceforth AJR) use settler mortality rates from previously colonized countries to proxy for the quality of institutions. AJR argue and in fact find evidence that the rate of mortality in the colonies affected the development path that these economies took because wherever mortality rate was high, Europeans were likely to set up extractive industries and neglect development of institutions. Further, former colonies where Europeans settled, as opposed to where they exploited natural resources, end up having good institutions which in turn determine financial development. Using the AJR settler mortality rate, Beck, Demirguc-Kunt and Levine (2002) show that the type of settlement established by colonizers has a lasting impact on financial development even today.
The model in this paper uses panel data to test the significance of financial development in the economy of a sample of Sub-Saharan African countries. The approach is similar to Jamison, Lau and Wang (2004) who adapt multi-level modeling technique to an aggregate production function. The model is an improvement on previous panel data approach because it relaxes the assumption of homogeneity across countries in the slope of policy parameters. Therefore, the study is also consistent with previous heterogeneity in panel data detected by Boskin and Lau (2000), Lee, Pesaran and Smith (1997) and Dougherty and Jorgensen (1996).
We allow for heterogeneity in the slope of financial development as the rate at which finance affects GDP may be different for each country based upon the preconditions established before modern economic growth.
III. The Data
The study covers the period 1963-1992 for Sub-Saharan Africa (SSA). While there are about 50 states (continental and islands) that qualify to represent the region, paucity of data reduces the model to only 24 countries. Data attrition is a problem regularly encountered when working with SSA. The study utilizes data on PPP-adjusted income and physical capital per capita from the Penn World Table (Version 5.6) (Heston and Summers 1996). Data on educational attainment (of the total population aged 15 and over) comes from the Barro-Lee (1996) data set. This measures the average years of school and is given for 5 year intervals.
Consistent with the current literature, we use the ratio of banks’ credit to the private sector to GDP from International Financial Statistics database to measure financial development. Earlier studies have also used the ratio of M2 to GDP as a proxy for financial development. However, this is perhaps a better measure for the extent to which the economy is monetized, rather than a measure of banking activity. Banks’ credit to the private sector is the preferred measure for financial development because it is assumed that banks allocating credit to private firms are more engaged in researching firms, exerting corporate control, providing risk management services, mobilizing savings, and facilitating transactions ( as opposed to say, banks that channel credit to the state, or state owned enterprises.) Table A1 in the Appendix shows a wide variation in the average value of credit ratio for the countries in the sample. The Republic of South Africa has the highest ratio at nearly 0.698, while Uganda has the lowest at 0.026 (which is in fact an extremely low figure).
The data on settler mortality rate comes from Acemoglu, Johnson and Robinson (2001). They construct the data from archival records on mortality rates among soldiers, sailors and bishops during the 17th, 18th and 19th centuries. Further studies by North and Thomas (1973), North (1981), Hall and Jones (1999), Rodrik (1999) and Johnson, McMillan and Woodruff (1999) all find that good institutions (e.g. secure property rights and good governance)matter for economic development. ‘Settler’ countries like Republic of South Africa and Mauritius have relatively lower settler mortality rates (log values of 2.74 and 3.42 respectively), while ‘extractive’ countries like Cote d’Ivoire and Nigeria have relatively higher rates (log values of 6.50 and 7.60 respectively)
Geographical variables and the measure of economic openness are from the HarvardCenter for International Development. For each country, we use the percentage of the country’s land area that lieswithin the tropics. We also use access to the coast, measured as the fraction of the country land area situated within 100 km of the sea coast and navigable waterways. Recent findings show that tropical and isolated countries have lower GDP per capita (Sachs and Warner 1997, Bloom and Sachs 1998). Geography matters because deadly diseases (malaria, dengue fever) are more prevalent in the tropics. Easterly and Levine (1997) provide an alternative measure in independence, which is the ratio of number of years a country has been independent since 1776. A country that has been independent for longer will have developed better institutions. Openness also seems to influence growth as evidenced from Sachs and Warner (1997) and Frankel and Romer (1999), as it allows countries to explore their comparative advantage potentials while also acquiring new technology. The data comes from Sachs and Warner (1997) and measures the ratio of number of years the country had an open economy over the number of years in the sample. Many countries are totally closed over the period (value of 0), while one country, Mauritius, is totally open (value of 1).
Legal origin data comes from LLSV (1999). Hayek (1960) mentions that checks and balances in Anglo-American courts play an important part in judicial independence. British government adopted Common Law to limit Crown intervention in the 17th century as a response to conflict between Parliament (mainly composed of land owners) and the English Monarchy. Common Law in effect dismantled the feudal system in England and established itself as a counterbalance that promoted private property rights. So under British Common Law, there is more protection for the individual through private property rights and economic freedom. On the other hand, the French under Napoléon reinforced government intervention though the Civil Code (also known as the Code Napoléon) in 1802. Prior to the Civil Code the Crown would sell judgeships to rich families resulting in judges essentially protecting the interests of the elites. Napoléon codified the French civil law giving more power to the State, relegating judges to minor bureaucratic roles. LLSV attribute higher economic performance under Common Law due to protection of property rights (and economic freedom), and lower economic performance under Civil Law due to its stiflingof individual economic freedom through state intervention, high taxation and generally less efficiency.
IV. Methodology
The analysis is on SSA countries which limits data availability in many cases to 5-year intervals (for example, data on education), or in some cases to a smaller period (bank credit). The model estimates a variant of a Cobb-Douglas specification. We explicitly include time as a left hand side variable, following Jamison, Lau and Wang (2004), which captures cumulative progress – that the authors call technical progress1. Following their model, we allow for cross-country variation in technical progress by using a multi-level modeling technique. A critical source of variation in rates of growth seems to come from the fact that each country may progress at different rates depending on its persistent characteristics2. We also allow for the slope of financial development to vary with the country’s legal origin and the quality of its institutions.
The multi-level modeling technique adopted is the Hierarchical Linear Model (HLM) developed by Bryk and Raudenbusch (1992). The application is available in SAS 9.1. It is a maximum likelihood procedure allowing us to model country-specific intercepts – this is similar to a GLS estimated random effects model when a common production function with varying intercepts is imposed across countries. However, more generalized HLM procedures that allow estimation of country-time, where time measures technical progress and country-financial development interactions are also employed. These permit us to assess determinants of cross-country variations in technical progress and in financial development. We thus address parameter heterogeneity for technical progress and financial development.
Usually, studies use either panel or cross section data with common coefficients across countries. However, following Jamison, Lau and Wang (2004), the model takes into consideration the possibility that important sources of cross-country variations in income growth result from, among other things, persistent differences in the characteristics of countries. As in Hall and Jones (1999), the paper tries to explain the fixed effects. The aggregate product function is modeled by specifying equation (1) below, supplemented with equations (3), (4) and (5). All the equations are modeled simultaneously and explain the country specific intercepts, rates of technical progress and elasticity of financial development in equation (1).