Do Financial Systems Spur Economic Growth in the USA ? An Empirical Investigation

In this paper, we use the autoregressive distributed lag (ARDL) bounds testing approach to examine the dynamic impact of both bank-based financial development and market-based financial development on economic growth in the United States of America during the period 1980 to 2012. In order to adequately capture the depth and width of the USA’s financial system, we used both bank-based and market-based financial development indices as proxies for bank-based and market-based financial systems. These indices were constructed from a number of bankand market-based financial development indicators, using the method of means-removed average. Our empirical results reveal that both bank-based and market-based financial development have a positive impact on economic growth in the USA. These results apply irrespective of whether the regression analysis is conducted in the long-run or in the shortrun. We, therefore, recommend that both pro-bank-based and pro-market-based financial sector development policies should be pursued in the USA – in order to bolster real sector growth and economic development.


Introduction
The debate on the relationship between financial development and economic growth has received considerable attention in recent years from numerous empirical studies in both developed and developing economies.The thrust of this debate has been whether financial sector development has a positive or a negative impact on economic growth.Although conventional wisdom has for a long time been in favour of the former, the empirical findings proving otherwise have been growing steadily in both number and substance.
Three groups of empirical work on the finance-growth nexus have emerged over the years.The first group consists of studies that support Schumpeter's (1911) notion that financial development has a positive impact on economic growth; and that it reflects the close relationship between financial development and economic growth.Studies that fall within this group include: Kargbo and Adamu (2009); Hassan et al. (2011); Levine and Zervos (1996), and Akinlo and Akinlo (2009).The second group is made up of studies that acknowledge the existence of a relationship between financial development and economic growth; but went further to state that this relationship is negative.Thus, according to this group of studies, financial development has a negative impact on economic growth.De Gregorio and Guidotti (1995), Adu et al. (2013), Ujunwa and Salami (2010), and Bernard and Austin (2011) are among the studies that fall within this group.Then, there is the third group that views financial development and economic growth as not related; and concludes that financial development has no impact on economic growth.Studies that anchor this group include those by Ram (1999) and Andersen and Tarp (2003).
Thus, whether financial development has an impact on economic growthpositive or negative or not as postulated by the three groups of studies remains an issue for empirical investigation.Unfortunately, the majority of the previous studies on this topic have concentrated mainly on Asia and Latin America, affording the United States of America (USA) little coverage, even though it is one of the world's biggest economies.
Although the relationship between financial development and economic growth has received considerable attention for many years, most of the earlier studies have focussed primarily on bank-based financial development proxies, giving market-based financial development little attention (see also Christopoulos and Tsionas 2004;Adu et al., 2013).Even where the latter studies have been done, the empirical findings have been largely inconclusive (Levine and Zervos, 1996;Ujunwa and Salami, 2010;Bernard and Austin, 2011).The evidence also suggests that the outcome is country-and time-variant.The outcome was also found to vary depending on the financial development proxy utilised; the model specified; and the estimation method used.
It is against this background that the current study attempts to empirically investigate the long- Third, this study employs the newly developed autoregressive distributed lag (ARDL) bounds testing approach that is appropriate even when the sample size is too smallunlike the commonly used residual-based cointegration test associated with Engle and Granger (1987), or the maximum likelihood test based on Johansen (1988) and Johansen and Juselius (1990) (see Odhiambo, 2008).Finally, this study employs time-series data and econometric techniques, unlike most of the previous studies that over-relied on cross-sectional data, which may not have satisfactorily addressed country-specific issues (Casselli et al., 1996;Ghirmay, 2004;Odhiambo, 2009).The findings of this study will not only contribute to the settlement of the still on-going finance-growth nexus debate, but could also provide policy guidance on financegrowth matters in the USA.
The rest of the paper is organised as follows: Section 2 gives an overview of the financial systems in the USA.Section 3 reviews literature on bank-and market-based financial development and economic growth.Section 4 covers the methodology of the study; while Section 5 presents and discusses the empirical results.Section 6 concludes.

An Overview of the Financial Systems in the USA
By any standard, modern or otherwise, the USA has one of the most highly developed financial systems in the world, which ranks very high in terms of the development and sophistication of its financial intermediaries and marketsas well as the size, depth and access available to its financial services.As a result, the USA was ranked number 1 in 2010 and number 2 in 2011, in terms of financial development, based on the Financial Development Index rankings (World Economic Forum, 2011).Like any other financial system, the USA's financial system consists of two segmentsthe bank-based segment as well as the market-based segmentwhich are both highly developed.However, the securities markets share centre stage with banks in propelling economic growth, hence the USA's financial system is referred to as a "marketbased financial system" (Demirguc-Kunt and Levine, 2001).
The financial system in the USA is unique in many respects.Unlike most financial systems whose apex position is occupied by a single central bank, the USA's financial system's apex is occupied by the Federal Reserve System, often known as the Federal Reserve or just "the Fed".
The Federal Reserve System consists of the regional reserve banks and the Fed that controls and coordinates operations of the regional reserve banks (Federal Reserve Bank of New York, 2012).Among its responsibilities, the Fed is responsible for supervising and regulating banks and other important financial institutions to ensure the safety and soundness of the nation's banking and financial system and to protect the credit rights of consumers.It is also responsible for monetary policy related matters and the maintenance and stability of the financial system (Federal Reserve Bank, 2012).
The uniqueness of the USA financial system can also be explained in terms of the complexity and unevenness of its financial institutions' regulatory structure as well as legal framework governing payment activity.While most countries have only one bank regulator, the USA's banking system is regulated at both federal and state levels (Bank of International Settlement, 2003).Furthermore, unlike most countries where only a few banks dominate the market, in the USA there are over 6000 commercial banks, 1500 savings and loan associations, 400 mutual savings banks, and 10,000 credit unions (Federal Deposit Insurance Corporation, 2012).
From the market-based segment front, it can be observed that there are a number of stock exchanges in the USA.The biggest one is the NYSE Euronext; followed by the NASDAQ OMX; then followed by the Chicago Stock Exchange (CHX) (World Stock Exchanges, 2015).
According to Sylla (1998), the precocity of USA banking development was duplicated in the development of the stock market.Sylla (1998) further points out that even a securities' market crash in early 1792 could not for long arrest the rapid deepening of these markets.This leap in asset liquidity allowed both domestic and foreign investors to overcome their reluctance to hold USA securities.Thus, for the USA, capital market globalisation arrived early in the nation's history, long before the more celebrated capital market globalisations of the late 19 th and late 20 th Centuries (Sylla, 1998).
In order to curb the repeat occurrences of the stock market crushes and the financial crises; as well as to keep pace with the global demand for modernisation, the USA embarked on a chain of financial sector reforms, targeting both the banking institutions and the financial markets.These reforms included, but were not limited to: modernisation of the financial systemboth in terms of clearing and settlement of transactions by banks and stock market trading systems; reducing financial repression; improving the legal, regulatory, judiciary and supervisory environments; rehabilitating the financial infrastructure; and restoring bank soundness.
To date, the landmark reform in the history of the USA's financial system is the Dodd-Frank  200% in 2007(World Bank, 2015).Likewise, the reform processes also gave rise to an increase in stock market capitalisation, total value of stocks traded and turnover ratio, despite a fall in the number of listed companies from 7524 in 2000 to 4171 in 2011; and further down to 4102 in 2012 (World Bank, 2015).Figure 1 tracks the performance and growth of the USA's banking sector (as shown by credit extension to the private sector) and the stock market (based on stock market capitalisation, total value of stocks traded and turnover ratio of stocks traded) during the period 1988-2012.Using the ratio of bank credit to private sector as a measure of bank-based financial development, De Gregorio and Guidotti (1995) examined the empirical relationship between economic growth and financial development in a large cross-country sample.Although on overall they found that bank-based financial development is positively related to economic growth; its impact was found to be negative in a panel data for Latin America.Likewise, Adu et al. (2013), in their study on the long-run growth effects of financial development in Ghana, found the relationship between financial development and economic growth to be negative when broad money stock to GDP ratio was used as a measure of financial development.
From the market-based financial development angle, there is also empirical evidence of the negative impact of market-based financial development on economic growth.Such empirical evidence includes that put forward by Ujunwa and Salami (2010) and Bernard and Austin (2011).When the former examined the impact of stock market development on long-run economic growth in Nigeria, they found evidence of negative association between stock market development and economic growth when stock market liquidity was used as a proxy of stock market development.In the same vein, Bernard and Austin (2011) found a negative association between stock market development and economic growth in Nigeria, using stock market capitalisation and total value traded as proxies of stock market development.
Despite the overwhelming arguments in favour of the positive or the negative impact of financial development on economic growth, there is still the third group of studies that found that financial development does not have any impact on economic growth.Studies, whose results are consistent with this group include Ram (1999) and Andersen and Tarp (2003).Ram (1999) found that financial development does not promote economic growth.Based on his 95country study, the predominant pattern was that of a negligible or weakly negative association between financial development and economic growth.Andersen and Tarp (2003) also found a weak association between financial development and economic growth in their 74-country study.

ARDL Bounds Testing Approach
This study adopts the recently developed autoregressive distributed lag (ARDL) bounds testing approach to examine both the long-run and the short-run relationship between bank-and market-based financial development and economic growth.The choice of this test is based on a number of advantages it has over other conventional cointegration testssuch as the residualbased approach by Engle and Granger (1987); and the full maximum likelihood test proposed by Johansen (1988) and Johansen and Juselius (1990).First, the ARDL-bounds testing approach does not impose the assumption that all the variables under study must be integrated of the same order.Thus, the ARDL approach can be applied to test the existence of a relationship between variables irrespective of whether the regressors are integrated of the same order or not; as long as they are integrated of order not more than one.Second, the ARDL technique normally provides unbiased estimates of the long-run model and valid t-statistics even when some of the regressors are endogenous (Odhiambo, 2008).Third, the ARDL method employs only a single reduced-form equation, unlike the conventional cointegration methods that estimate the long-run relationships within a context of a system of equations (see also Duasa, 2007).Fourth, the ARDL test has superior small sample properties, when compared to the other conventional methods of testing cointegration (Pesaran and Shin, 1999).While other cointegration techniques are sensitive to the size of the sample, the ARDL test is appropriate even when the sample size is small.Because of these reasons, the ARDL approach is, therefore, considered to be very suitable for analysing the underlying relationship.The method has also been increasingly used in empirical research of late.

Specification of the Empirical Model
The ARDL-based empirical model used in this study to test the impact of bank-based and market-based financial development on economic growth is expressed as follows: Where: y = Growth rate of real gross domestic producta proxy for economic growth; as a ratio of GDP, calculated as total shares traded on the stock market exchange divided by GDP (TV); and turnover ratio (TOR) which is equal to the value of total shares traded divided by market capitalisationare computed.The means-removed value of variable X is defined as where ABS(w) refers to the absolute value of w.For mean (X), the average value of X over the 1980-2012 period was used.Second, a simple average of the means-removed CAP, TV and TOR is taken to obtain an overall index of market-based financial development (FDM).The same computations are applied to obtain FDB.
Following the ARDL model specified in equations (1) the ARDL-based error-correction model is specified as follows:

Sources of Data
The annual time series data, covering the period from 1980 to 2012, utilised in this study were obtained from the World Bank Economic Indicators and International Financial Statistics Year Books (IFS, various issues).

Unit Root Tests
Before any analysis is made, the variables are first tested for stationarity, using Dickey-Fuller Generalised Least Square (DF-GLS), Phillips-Perron (PP) and Perron (1997) (PPURoot) unit root tests.The PPURoot test was employed to cater for possible structural breaks within the dataset.The detailed results of stationarity tests for all the variables are presented in Table 2.
Although the ARDL does not require pre-testing of variables, the stationarity test provides guidance as to whether ARDL is appropriate or not, as it is only applicable for the analysis of variables that are integrated of order zero or one.In this case, the variables are found to be integrated of either order 0 [I(0)] or 1 [I(1)], therefore, ARDL bounds testing technique can be performed.

Bounds F-Test for Cointegration
This section examines the long-run relationship between the variables in the specified model using the ARDL bounds-testing approach.The first phase is to get the order of lags on the first differenced variables in equations (1) using either the Schwartz Bayesian Criterion or the Akaike Information Criterion.This is followed by the second phase, which is the application of a bounds F-test to equation (1) to establish the existence of a long-run relationship, if any, between the variables under study.The results of the bounds F-test are reported in Table 3.The results of the ARDL bounds test for cointegration reported in Table 3 show that the calculated F-statistic of 5.486 is higher than the critical values reported by Pesaran et al. (2001) in Table CI(iii) Case III.Based on these results, it can therefore be concluded that the variables in the specified model are cointegrated.

Estimated ARDL Model
Having found that y, FDB, FDM, INV, SAV and TOP are cointegrated, the ARDL technique is employed in the estimation of the model.The first stage in this process is the determination of the optimal lag-length for the specified model using the Akaike information criterion (AIC) or the Bayesian information criterion (BIC).The optimal lag-length selected based on BIC is ARDL(1,0,0,1,0,1).The BIC-based model was chosen because it was more parsimonious than the AIC-based model.The long-run and short-run results of the selected model are reported in  The regression results reported in Table 4 show that the coefficient of bank-based financial development is positive and statistically significant, as expected.This implies that in the USA, bank-based financial development has a positive impact on economic growth.Thus, an increase in the level of bank-based financial development in the USA leads to an increase in economic growth.These results apply both in the long run and in the short run.Other results show that the coefficient of investment is positive and statistically significant as expected, suggesting that investment impacts positively on economic growth, but only in the short run.In the long run, the coefficient of investment is statistically insignificant.The coefficient of savings ratio is positive and statistically significant as expected, both in the long run and in the short run.Thus in the USA, an increase in the savings ratio leads to an increase in economic growth.The other results, however, show that while the coefficient of trade openness is statistically insignificant in the short run, it is negative and statistically significant in the long run.These findings, though contrary to the expectations of the current study, are consistent with the results obtained in some of the previous studies (see Odedokun, 1996;Güryay et al., 2007).The coefficient of ECM (-1) is also found to be negative and statistically significant, as expected.
The regression for the underlying ARDL model fits well, as indicated by an R-squared of 88.8%.On a battery of tests performed for serial correlation, functional form, normality and heteroscedasticity, the results displayed in Table 5 show that the model passed all tests except normality.However, an inspection of the Cumulative Sum of Recursive Residuals (CUSUM) and the Cumulative Sum of Squares of Recursive Residuals (CUSUMSQ) graphs in Figures 2   and 3 shows that there is stability and that there is no systematic change identified in the coefficients at 5% significance level over the study period.Thus, the CUSUM and CUSUMSQ graphs show that the parameters in this model are stable over the sample period.
and short-run relative impact of bank and market-based financial development on economic growth in the USA.This study differs fundamentally from the majority of previous studies on the subject in several ways.First, it splits financial development into bank-based and marketbased components and examines the relative impact of each component on economic growth.Second, unlike the majority of previous studies that used one or a few indicators of bank-based financial development, which might not sufficiently capture the breadth and depth of a financial sector, this study constructs bank-and market-based financial development indices from a number of bank-and market-based financial development indicators to ensure a holistic picture of the relationship between financial development and economic growth in the study country is captured.
Wall Street Reform and Consumer Protection Act (Pub.L. 111-203, H.R. 4173) (DFWSRCP Act) of 2010 that was passed in response to the recession of the late 2000s (United States Government Printing Office, 2012).It brought the most significant changes to financial regulation in the USA since the post-Great Depression regulatory reform.This legislative piece made changes in the American financial regulatory environment that affect all federal financial regulatory agencies and almost every part of the nation's financial services industry, including the stock market.Over time, these reforms have given rise to a developed and well-regulated financial system in the USA.In the banking sector, this growth is evidenced by an increases in the number of Automated Teller Machines (ATMs) from 352,000 ATMs in 2002 to 396,000 in 2005, to 425,010 in 2008; and slightly down to 403,000 in 2009 (United States Department of State, 2012).Credit extension to the private sector has also grown from an average of 120% of GDP between 1975 and 1981 to 150% in the late 1980s.Historically, between 1975 and 2012, private sector lending reached an all-time low, of 115.2% in 1981; and an all-time high of over

Figure 1 :
Figure 1: Trends in Banking Sector and Stock Market Growth in the USA (1988-2012)

FDB=
An index of bank-based financial development; which is a means-removed average of M2, M3 and credit provided to the private sector by financial intermediariesa proxy for bank-based financial development (see also Demirguc-Kunt and Levine, 1996); FDM = An index of market-based financial development; which is a means-removed average of stock market capitalisation, stock market traded value and stock market turnovera proxy for market-based financial development (see also Demirguc-Kunt and Levine, 1996); INV = Gross fixed capital formation as a percentage of GDP; SAV = Gross savings as a percentage of GDP; TOP = Trade openness, which is the sum of the share of total imports in GDP and the share of total exports in GDP; conglomerate index of market-based financial development (FDM), the meansremoved values of the three indicators of market-based financial development are averaged, in a two-step procedure.First, the means-removed valuesof stock market capitalisation ratio, calculated as the value of listed shares divided by GDP (CAP); the total value of stocks traded The long-run positive impact is supported by the coefficient of bank-based financial development in Panel 1, that is positive and statistically significant, while the short-run economic impact of bank-based financial development on economic growth is shown by the coefficient of bank-based financial development in Panel 2, that is also positive and statistically significant.The results displayed in the same table further show that the coefficient of market-based financial development (FDM) is positive and statistically significant, as also expected.This indicates that, market-based financial development has a positive impact on economic growth in the USA.These results apply irrespective of whether the model is estimated in the long run or in the short run.The long-run positive impact is evidenced by the coefficient of marketbased financial development in Panel 1, that is positive and statistically significant, while the short-run economic impact of market-based financial development is supported by the coefficient of market-based financial development in Panel 2, that is also positive and statistically significant.

Figure 2 :
Figure 2: Plot of Cumulative Sum of Recursive Residuals

Table 1 : Studies Showing the Nature of Impact of Bank-and Market-Based Financial Development on Economic Growth Author(s) Region/Country Methodology Nature of Impact Panel 1: Bank-Based Financial Development and Economic Growth
Table 1 summarises the empirical studies on the impact of financial development on economic growth.Panel 1 indicates studies on bank-based financial development and economic growth while Panel 2 shows studies on market-based financial development and economic growth.