138
JOURNAL O F E C O N O M I C S A N D F I N A N C E 9 Volume 26 9 N u m b e r 2 9 Summer 2002
Financial Development and Productive Efficiency: A Panel Study of Developed and Developing Countries Farrokh Nourzad*
Abstract This paper uses a stochastic production frontier for panel data to investigate the effect of financial development on productive efficiency. Three panels of a number of countries in different stages of development are used along with eight alternative measures of financial development pertaining to the monetary sector, financial intermediaries, and equity markets. The results indicate that in general the more developed the financial intermediaries sector and equity markets, the higher the productive efficiency. In particular, financial deepening reduces productive inefficiency in both developed and developing countries, although the effect is larger in the former. (JELC23, E44, 047)
Introduction The development of the new endogenous growth theory in the late 1980s (Romer 1986; Lucas 1988) led to a flurry of empirical research into the causes of economic growth, which has continued to date. The early work was concerned with the basic determinants of long-run growth implied by the neoclassical and endogenous growth theories such as investment in real capital, human capital, taxes, and technology (Barro 1991; Barro and Sala-i-Martin 1997; Benhabib and Spiegel 1994; Mankiw, Romer, and Weil 1992). A factor that has received much attention in recent years is financial development. Various measures from the financial sector have been used to examine the relationship between financial development and growth. As Levine (1997, p. 688) summarizes the general finding of the literature, "the preponderance of theoretical reasoning and empirical evidence suggests a positive, first-order relationship between financial development and economic growth." 9 Farrokh Nourzad, Economics Department, Marquette University, Milwaukee, WI, 53201-1881, farrokh.nourzad@ mu.edu. The author wishes to thank three anonymous referees for theirmany helpful comments on an earlierversion of this paper. All remaining errors are the author's responsibility.A slightlydifferentversion of this paper was presented at the 50th annual conference of the InternationalAtlantic Economic Society, Charleston, South Carolina, October 2000. This research is partiallyfunded by a Marquette University College of Business Administration Faculty Research Grant from the Miles Fund and a grant from Marquette University Institutefor InternationalEconomic Affairs.
JOURNAL OF ECONOMICS AND F I N A N C E , Volume 26 * Number 2 9 Summer 2002
139
This paper attempts to investigate the effect of financial development on productivity using the stochastic production frontier model for panel data developed recently by Battese and Coelli (1995). Three panels of developed and developing countries are used along with eight alternative measures of financial development pertaining to the monetary sector, financial intermediaries, and the stock market. The main issue investigated is whether financial variables are statistically significantly correlated with the estimated productive inefficiency effects across space and over time. The results indicate that in general the more developed the financial intermediaries sector and equity markets, the higher the productive efficiency. This effect appears to be larger in industrialized countries relative to developing nations. Section 2 discusses the role of the financial sector in improving efficiency in the real sector. Section 3 outlines the econometric approach used in this paper, while Section 4 introduces the measures of financial development and other data used in this study. Section 5 presents the results. The paper closes with Section 6, which contains a summary of this work along with a few suggestions for further work in this area.
Financial Development and Economic Efficiency As is well known, the two major tenets of classical macroeconomics are monetary neutrality and dichotomy. If one were to rely on these principles, one would conclude that in the long run there is no connection between the financial and real sectors. However, as Levine (1997, p. 689) notes, "[t]he financial sector is a 'real' sector: it researches firms and managers, exerts corporate control, and facilitates risk management, exchange, and resource mobilization." Theoretically, financial development can affect economic growth through two distinct channels. First, as the financial intermediary sector develops, it allocates the economy's savings to firms more efficiently, thus improving productivity a n d growth. Second, better-developed financial intermediaries can increase a nation's saving rate, which can in turn enhance the economy's growth. Nearly 90 years ago Joseph Schumpeter (1912) argued that by allocating the economy's savings to firms, the financial sector plays an important role in the process of growth and development. In the "Schumpeterian view of finance and development," the financial sector's influence on growth and development is transmitted though productivity and technology. Schumpeter (1912, p. 68) defines development as "consist[ing] primarily in employing existing resources in a different way, in doing new things with them, irrespective of whether those resources increase or not." He goes on to state that "credit is primarily necessary to new combinations...[that is] 'financing' as a special act is fundamentally necessary [for development], in practice as in theory" (Sehumpeter 1912, p. 70). In short, development requires more efficient use of available resources, and finance facilitates this process. Recently, Beck et al. (2000) examined the Schumpeterian view of finance and development by considering the effect of financial development on not only economic growth but also various sources of growth such as the saving rate, investment in real capital, and total factor productivity. They found strong support for the "Schumpeterian" view relative to the conventional view that the benefits of financial development are transmitted to economic growth through the saving rate and capital accumulation. This paper also looks at the effect of financial development on a main s o u r c e of growth--labor productivity--but from an entirely different angle. While Beck et al. (2000) employ a conventional regression equation, which contains a measure of financial development as an explanatory variable, this paper uses a stochastic production frontier that does not treat financial development as input. Rather, it treats it as an external factor that can enhance productive efficiency. Financial development can improve production efficiency in a number of ways including a better allocation of the existing stock of capital (Schumpeter 1912). Well-developed financial
140
J O U R N A L O F ECONOMICS A N D F I N A N C E 9 Volume 26 9 Number 2 9 Summer 2002
markets shift capital away from declining industries and toward growing industries (Wurgler 2000). Another way in which financial development can improve productive efficiency is through reduction of transaction costs. As Beck et al. (2000, p. 266) put it, "[b]y providing more accurate information about production technologies and by exerting corporate control, better financial intermediaries can enhance resource allocation and accelerate growth." Financial intermediaries also reduce the cost of acquiring information, thus helping inputs move to their most efficient uses. This is especially true in those instances where the acquisition of information concerning the most efficient production process involves substantial fixed costs. Equity markets can also provide accurate information at low costs by publishing share prices, thus improving production efficiency. These markets also promote efficiency through technology. "For example, different production technologies may have a wide array of gestation periods for converting current output into future capital, where long-term technologies enjoy greater returns. Investors, however, may be reluctant to relinquish control of their savings for very long periods. Thus, longer-gestation production technologies require that ownership be transferred throughout the life of the production process in secondary securities markets" (Levine 1997, p. 693). Not only does a well-functioning financial system provide accurate information at low costs and help mobilize capital, but it can also "spur technological innovation by identifying and funding those entrepreneurs with the best chance of successfully implementing innovative products and production processes" (Schumpeter 1912 as cited in Levine 1997, p. 688). In other words, the financial system can provide for both a larger stock of capital and better use of the existing inputs, thus resulting in improved production efficiency. An issue of concern in the study of the relationship between financial development and economic growth is the direction of causality. Does financial development cause more rapid economic growth, or does a higher rate of economic growth cause the financial sector to develop more fully? Or is there simultaneity between growth and financial development? The possibility of reverse causation was raised first by Joan Robinson (1952, p. 86) who stated that "where enterprise leads finance follows." While a large number of studies find causality flows from financial development to economic growth (Demirguc-Kunt and Maksimovic 1998; Jayaratne and Strahan 1996; Rajan and Zingales 1998; Rousseau and Wachtel 2000), a few single-country, timeseries studies find evidence of bi-directional causality and even reverse causation (Demetriades and Hussein 1996; Jung 1986). Several approaches have been used in the empirical finance-growth literature to control for reverse causation and simultaneity. Beck et al. (2000) and Levine et al. (2000) study the growth effect of financial development using lagged values of explanatory variables and legal origin of countries as internal and external instruments, respectively. Harris (1997) examines the relationship between stock market activity and economic growth using two-stages least squares. Similarly, Rousseau and Wachtel (2000) investigate the effect of stock market liquidity and financial intermediation on economic growth using two-stage least squares.
Methodology This paper investigates the productivity effect of financial development using the stochastic production frontier model for panel data due to Battese and Coelli (1995). We begin with a standard Cobb-Douglas production function that expresses the logarithm of output, Q, as a function of logarithms of labor, L, and capital stock, K. s Data availability problems force us to use 5 Moroney and Lovell (1997) point out that,while the use of a more flexible functional form such as the Iranslog may be preferable when using firm-specific or industry-specific dam, at the aggregate level the Cobb-Douglas form performs as well.
JOURNAL OF ECONOMICS AND FINANCE 9 Volume 26 9 Number 2 9 S u t m ~ r 2002
141
three different panels, two that are annual and one that consists of five-year averages. 6 This necessitates that we modify the basic production function in several ways. First, when using the two annual panels, we include a linear time trend, t, to control for disembodied technological change. 7 Second, in order to control for the business cycle effect, which may be present in the annual panels, we include the lagged value of the cyclical component of output, B C , as an additional explanatory variable in the production function) This variable is constructed using the Hodrick-Prescott (1997) filter to isolate the cyclical component of real GDP for every country in the sample. Third, because each of the three panels consists of both developed and developing countries, following Moroney and Lovell (1997) we control for structural differences between the two types of countries using a dummy variable, D, that equals zero for the industrialized countries and one for the developing nations in each panel. We incorporate this into the model both as a differential intercept term and as a series of slope variables expressed as interactions with labor, capital, and time trend. Finally, because we are interested in the effect of financial development on productivity, we subtract the logarithm of labor, L, from both sides of the production function so that the left side becomes the logarithm of output per worker. 9 To summarize, we use the two annual panels to estimate the parameters of the following production frontier, where i is the country index and t is the time index:~~ Qi,- Lit = [~10+(1~,- 1)Li,+ l~Ki,+ [~3t+ 134Dit+ ~5Di~Li,+ [~6DitKit+ ffiDitt + [~BCi, + vit - uit
(1)
As noted earlier, when using the five-year averages panel, the trend and business cycle variables will be excluded from the equation. The production frontier in Equation (1) is an error-components model in that its error term is the difference of two independent random variables: a classical error term, vit, and a non-negative random variable, uit, capturing production inefficiencies in country i in year t. While it is customary to assume vi, is N(0, ~v), the choice of a distribution for uit is arbitrary. We assume uit are independently distributed as truncation at zero of N(p~t, o~). Different stochastic production frontier models have been proposed and estimated, ranging from models that assume production inefficiencies, ui~, are iid, to those that assume they are time invariant, to models in which production inefficiencies vary over time. We employ the specification suggested by Battese and Coelli (1995), which expresses mean inefficiency, P~t, as a function of external factors, which we take to be financial development in country i in period t-l, denoted Fit.m,~ ~'it----~0 + ~lFit-! + ~2Dit+ ~3DitFit-i
(2)
where D is the dummy variable for developing countries defined earlier.
6 The panels are described in the next section. 7 No trend is included in the model when the five-yearaverages panel is used because for each cross section the panel contains only four five-yearobservations. s The business cycle variable is not included in the model when using the five-yearaverages panel because these data are unlikely to pick up short-term fluctuationdue to the smoothingeffect of averaging. 9 Alternatively, one can impose the constant-returns-to-scalerestriction on the parameters of the production function. to Note that as long as labor is subject to diminishing returns (i.e., [3t < 1,) we should find the coefficientof labor in Equation (1), i.e., (1~1- !) to be negative. Once Equation (i) is estimated, we can get back to the labor elasticity of output, ~, by simply adding 1 to the estimated coefficientof L in Equation (1). N The lagged value of financial development is used as an instrument in order to control for reverse causation and simultaneity.
142
JOURNAL O F ECONOMICS A N D F I N A N C E 9 Volume 26 9 N u m b e r 2 9 Summer 2002
The panel model in Equations (1) and (2) is estimated using maximum likelihood yielding point estimates for the parameters of the production function ([3s) as well as the coefficients of the relation between financial development and production inefficiency (6s). The model cannot be estimated using OLS unless u~t = 0 for all i and t, which would be the case only ifu = 60 = 61 = 62= 63 = 0, where u = a 2/0"2v is the variance parameter. This joint hypothesis is tested using a likelihood-ratio test statistic that follows a mixed chi-square distribution.
Measurement and Data As was indicated earlier, we study the effect of financial development on productive efficiency using three separate panels of a number of countries in different stages of development. '2 One is an annual panel of 29 countries for the period from 1966 through 1990 (a total of 725 observations.) The second is also an annual panel covering the years 1970-1990 and containing 18 countries (378 observations.) The third panel consists of five-year averages covering the 1970-1990 period for 28 countries (112 observations). A list of country names is found in the appendix. The larger annual panel consists of variables that represent the degree to which the monetary sector is developed. These include M2; M2 minus currency; and the ratio of currency to demand deposits. Examples of studies using these indicators of financial development include Berthelemy and Varoudakis (1996a, b), Jung (1986), and Demetriades and Hussein (1996). The data for these variables come from the IMF International Financial Statistics (IFS) CD-ROM. Also included in this panel is private sector credit provided by commercial banks. Beck et al. (2000), Demetriades and Hussein (1996), and Levine et al. (2000) among others have used this variable. As noted by Demetriades and Hussein (1996), increases in several of these aggregates represent financial deepening. The second annual panel contains a single financial development variable, namely average share prices traded on all stock markets in each of the countries in the sample. Examples of work using indicators of stock market activity include Levine and Zervos (1998), Atje and Jovanovic (1993), and Harris (1997). The data on this variable are from the IMF IFS CD ROM. The five-year averages panel consists of three alternative measures of the financial intermediary sector. One is the amount of credit extended to the private sector by banks and nonbank financial intermediaries. Another is total liquid liability of the financial system. This captures the size of the financial intermediary sector and thus represents financial depth. The final measure of financial development is the ratio of commercial banks assets to the sum of assets held by commercial banks and the central bank. This represents the degree of financial intermediation by commercial banks, which can exercise better control over entrepreneurs than the central bank, as they are closer to both savers and investors. Beck et al. (2000), De Gregorio and Guidotti (1995), King and Levine (1993a, 1993b), and Levine et al. (2000) are among those who used one or more of these variables. The data used in the present study were taken from Levine et al. (2000). The arguments of the production function are quantified using data from Easterly and Levine (2001). Labor is expressed in terms of the number of workers. Capital stock is based on disaggregated investment on machines, equipment, business structure, etc.
12 For examples of alternative stochastic production frontier models with country-level panel data, see Koop, Osiewalski,and Steel(2000)and Moroneyand Loveil(1997).
JOURNAL O F ECONOMICS AND FINANCE * Volume 26 9 Number 2 * Summer 2002
143
R e s u l t s 13 Table 1 reports the maximum-likelihood estimation results using the two annual panels described above. Columns (1)-(4) contain the results for the panel of 29 countries and four alternative monetary aggregates, M2; M2 minus currency, M2 - C; the ratio of currency to demand deposits, C/DD; and private sector bank credit, CRED. Column (5) reports the results based on the annual panel of 18 countries and stock prices, SP. TABLE 1. MAXIMUMLIKELIHOODESTIMATESOF STOCHASTICPRODUCTIONFRONTIER, ANNUALPANELSOF 29 COUNTRIES 1966-1990 Variable
Intercept L K t D D xL D xK D xt BC.j
Constant F.~ D DF t
y LR
o
(1) M2
(2) M2-C
(3) C/DD
(4) CRED
(5) SP
6.7035 (0.3139)*'" -0.2549 (0.0398)*** 0.2865 (0.0358)"* 0.0070 (0.0012)'** -2.2412 (0.3898) **~ -0.4514 (0.0416)*'* 0.3709 (0.0381) ~176 -0.0051 (0.0027)* 0.4450 (0.1271)'*"
6.7112 (0.3173)'** -0.2539 (0.0402) ~ 0.2856 (0.0362)*'* 0.0070 (0.0012)'*" -2.2428 (0.3975)'** -0.4523 (0.0414)'*" 0.3715 (0.0382)" -0.0052 (0.0028)" 0.4431 (0.1227)**"
6.6844 (0.2998)*** -0.2599 (0.0385)**" 0.2902 (0.0345)*** 0.0074 (0.0012)*** -2.1809 (0.3529) .*~ -0.4405 (0.0419)'** 0.3624 (0.0365)*** -0.0070 (0.0024)*" 0.4292 (0.1284)**"
6.7135 (0.2797)'** -0.2501 (0.0346) "~ 0.4194 (0.1259) ~ 0.0064 (0.0012)'" -2.1114 (0.3317) ~ -0.4578 (0.3771)"" 0.3707 (0.0333)"" -0.0077 (0.0027)*** 0.4194 (0.1259) *.~
5.7389 (0.4102)*** -0.3020 (0.0321)*** 0.3571 (0.0339)*** 0.0026 (0.0012)*" -5.5228 (1.0453) "*" -0.5371 (0.0380)'** 0.5518 (0.0592)'*" -0.0217 (0.0028)'** 0.1782 (0.1001)*
-0.4211 (0.2303)* -0.1423 (0.0790)" 0.7792 (0.2295)**"
-0.4261 (0.2313)" -0.1436 (0.0787)" 0.8053 (0.2300)"*
-0.8593 (0.2363)*'* 0.1072 (0.2789) 1.3212 (0.2190)*"
0.0750 (0.2773) *~ -0.3410 (0.1242) "'~ 0.5190 (0.2901)'*"
0.7836 (0.1103) "~ -0.4620 (0.0820) "~ 0.4134 (0.1818)"
0.1930
(0.0813)'*
0.1894 (0.0831)'"
-0.0428 (0.2823)
0.3089 (0.1258)"
0.1993 (0.1259)*
0.9707 279.9620"** 0.4538
0.9709 279.4873"*" 0.4545
0.9721 296.4098"*" 0.4171
0.9673 285.0301"'" 0.4540
0.9999 198.6686"'* 0.2212
Notes: The panel containing stock prices is for the period, 1970-1990. Asymptotic standard errors are in parentheses. "Significant at the 10 percent level; "'Significant at the 5 percent level; ""Significant at the 1 percent level. "f = variance parameter, LR = likelihood-ration statistic; o = standard error of estimate. 13 All estimations are carried out using the FRONTIER software, version 4.1 by Coelli (1996).
144
J O U R N A L O F ECONOMICS A N D F I N A N C E 9 Volume 26 9 Number 2 9 Summer 2002
The top portion of Table 1 contains the results concerning the deterministic component of the production frontier. They indicate that the estimated coefficients on labor and capital have the expected signs and are highly statistically significant in all five cases. The estimated coefficient on the trend variable is positive across all five equations and is statistically significant at the 5 percent level or better. The four terms involving the developing nations dummy variable are individually and jointly statistically significant at high levels of confidence in all equations. 14 The negative sign of the parameter estimate associated with the intercept dummy indicates that developing nations use production technologies that are inferior to those of the industrialized nations. This result is reinforced by the fact that the estimated coefficient on the product of the dummy variable and time trend representing disembodied technological progress is also negative. The estimated parameter on the dummy• interaction term is negative, suggesting that labor is less productive in developing nations. On the other hand, the positive coefficient on dummy• suggests that physical capital is more productive in developing economies, a result that is consistent with the neoclassical growth model. The estimated coefficient on the business cycle variable is positive and statistically significant at the 10 percent level or better across all equations, suggesting that, as expected, productivity is pro-cyclical. The results in the lower portion of Table 1 indicate that the variance parameter, y, is quite large, ranging from a low of 0.967 (column 4) to a high of 0.999 (column 5). The likelihood ratios, LR, indicate that the hypothesis that there are no productive inefficiencies, i.e., u = ~i0 = 5, = ~2 = ~3 = 0, can be rejected in all cases at the 1 percent level. This coupled with the large value of the variance parameter indicates that much of the variation in the error term is due to inefficiencies. Finally, consider the estimated efficiency effect of financial development in the middle section of Table i. In the case of M2 (column 1), this estimate is negative and statistically significant at the 10 percent level, suggesting that an increase in M2 reduces productive inefficiencies. Demetriades and Hussein (1996) suggest subtracting currency from M2 balances. The rationale is that increased reliance on currency relative to bank deposits is indicative of an underdeveloped financial sector. This is the measure of financial development used in column 2 of Table 1, whose estimated coefficient is negative and statistically significant at the 10 percent level. An alternative approach is to use the ratio of currency to demand deposits, which should increase productive inefficiency. This is what the result in the third column of Table 1 indicates, although the estimate is not statistically significant. The results in the last two columns of Table 1 pertain to two particular financial institutions, namely commercial banks and stock markets. The measure of financial development in column (4) is bank credit to private sector. The estimated coefficient on this variable is negative and statistically significant at the 1 percent level, suggesting that financial deepening reduces productive inefficiency. The result concerning the effect of stock prices is also supportive of the hypothesis that financial development improves the efficiency of the production sector. Turning to the dummy variable for the developing nations, we note that the estimated coefficient is consistently positive and statistically significant at the 5 percent level or lower in all cases. The positive sign of these estimates indicates that there are greater productive inefficiencies in developing countries compared to industrialized nations. The estimated coefficients associated with the dummyxfinancial development interaction terms are also positive and statistically significant at the 10 percent level or better in columns 1, 2, 4, and 5, suggesting that financial development has a smaller effect on productive efficiency in developing countries. This result is consistent with the negative sign of the coefficient in Column 3, except that the point estimate is ,4 The only exceptionsare the estimatedcoefficientson the dummyxtrendvariable in columns I and 2, which are significantat the I0 percentlevel.
JOURNAL O F ECONOMICS AND FINANCE 9 Volume 26 9 Number 2 9 Summer 2002
145
not statistically significant. It is interesting to note that the sum of the estimated coefficients on the measure of financial development and its interaction with the dummy variable is not statistically significant in columns 1, 2, and 4, but is negative and statistically significant in column 5. Thus, it appears that in developing countries the equity market is the primary channel through which financial development influences productive efficiency. ,5 Having examined the efficiency effect o f monetary aggregate, bank credit, and stock prices, we now turn to the results from the five-year average panel o f 28 countries, where we use three separate indicators o f financial intermediary development: total private sector credit by all types o f financial intermediaries, FIC; total liquid liability o f the financial system, LL; and the ratio o f private banks' domestic assets divided by the sum o f private and central banks' domestic assets, AR. The estimates are found in Table 2.
TABLE 2. MAXIMUM LIKELIHOODESTIMATESOF STOCHASTICPRODUCTIONFRONTIER, FIVE-YEAR AVERAGESPANEL OF 28 COUNTRIES 1970-1990
(I)
(2)
(3)
Variable
FIC
LL
AR
Intercept
6.4678 (0.4612)'" -0.2702 (0.0525)" 0.3063 (0.0472)'*"
6.3999 (1.1930)"" -0.3231 (0.0931)"" 0.3424 (0.0766)"
6.1952 (0.5170)" -0.3266 (0.0571)"" 0.3517
-2.3458 (0.9208)" -0.3723
-2.0776 (0.5280)" -0.3912
(0.05285)" -0.7589 (0.9106) -0.3060
(0.0563)"
(0.0894)"
(0.0816)"
D xK
0.3329 (0.0730)"
0.3269 (0.0792)""
0.2195 (0.0747)"
Constant
6.9240 (3.0359)"
-0.0452 (0.2103)
- 1.8180 (1.0298)"
F. I
-0.2080
-0.0230
- 1.0862
(0.0935)"
(0.0063)"
(0.6610)"
6.7531 (2.9589)*" 0.1654 (0.0874)"
0.0386 (0.3093) 0.0515 (0.0150)"
2.8242 (1.0305)" 2.2161 (0.9072)"
0.8902 53.7656"" 0.3359
0.9120 38.8129"" 0.3529
0.9187 53.2810"" 0.3579
L K
D D xL
D D xF. I
y LR
o
Notes: Asymptotic standard errors are in parentheses. "Significant at the 10 percent level; "'Significant at the 5 percent
level; ""Significant at the 1 percent level. u = variance parameter; LR = likelihood-ration statistic: o = standard error of estimate. ~sIn order to investigate whether financialdevelopmentaffects productivitydirectly in addition to influencingproductive efficiency, the model was re-estimated using the two annual panels with financial developmententering both the production frontier Equation (1) and the inefficiency effects Equation (2). The results indicated that financial development generally improves productivityin both developed and developingcountries, the effect being smaller in the latter. The results regarding the efficiencyeffectof financialdevelopmentwere not much differentfrom those reported in Table 1.
146
JOURNAL OF ECONOMICS AND FINANCE e Volume 26 e Number 2 e Surnmer 2002
The results concerning the inputs of labor and capital are quite similar to those in Table 1. This is also true of the intercept and slope dummy variables in the deterministic part of the frontier. The estimated variance parameter, 7, is also large, although not as large as in Table 1, ranging from 0.89 to 0.92 in the present case. The likelihood ratios for testing the null of no productive inefficiencies are significant at the 1 percent level. As far as the estimated efficiency effects of the three measures of financial development are concerned, the results indicate that every one of them exerts a negative effect on productive inefficiency at least at the 10 percent level of significance. The dummy variable identifying the developing nations is positive in all cases, suggesting greater inefficiencies in these countries with the estimate being statistically significant in the first and third columns. The interaction terms involving the dummy variable and alternative measures of financial intermediary development are consistently positive and statistically significant at the 10 percent level or better. The sum of the estimated coefficient on financial development and the dummy interaction term is positive and statistically significant in columns 2 and 3 but not in column 1. These results are in agreement with those of the annual panels (Table 1) that financial development has a larger positive impact on productive efficiency in industrialized nations.
Summary and Suggestions for Further Research This paper examined the effect of eight alternative measures of financial development on productive efficiency using three separate panels of developed and developing countries. The results indicate that seven of the eight measures of financial development exert a negative and statistically significant effect on production inefficiencym implying that the more financially developed the economy, especially in terms of the financial intermediaries sector and equity markets, the more efficient the production of output. This effect appears to be larger in developed nations relative to developing countries. The analysis presented in this paper can be extended in a number of ways. One is to use better measures of the degree of stock market development than share prices used here. For example, following Rousseau and Wachtel (2000), one can use market capitalization or total value traded deflated by share prices. Another possible extension is to consider the role of international financial variables in enhancing productive efficiency. Also valuable may be to control for financial reforms, especially in developing countries. It may also be worthwhile to examine the effect of financial development on human capital. Such a study would complement not only the present work, but also that of King and Levine (1993b).
References Atje, R., and B. Jovanovic. 1993. "Stock Markets and Development." European Economic Review 37: 632-640. Barro, R. J. 1991. "Economic Growth in a Cross Section of Countries." Quarterly Journal o f Economics 106: 407-443. Barro, R. J., and X. Sala-i-Martin. 1997. "Technological Diffusion, Convergence, and Growth." Journal of Economic Growth 2: 1-26. Benhabib, J., and M. M. Spiegel. 1994. "The Role of Human Capital in Economic Development: Evidence from Aggregate Cross-Country Data." Journal o f Monetary Economics 34: 143173.
JOURNAL OF ECONOMICSAND FINANCE 9 Volume 26 9 Number 2 9 Summer 2002
147
Battese, G. E., and T. J. Coelli. 19959"A Model for Technical Inefficiency Effects in a Stochastic Production Frontier for Panel Data." Empirical Economics 20: 325-32. Beck, T., Levine, R., and N. Loayza. 2000. "Finance and the Sources of Growth." Journal of Financial Economics 58: 261-300. Berthelemy, J. C., and A. Varoudakis. 1996a. "Models of Financial Development and Growth: A Survey of Recent Literature." In Financial Development and Economic Growth: Theory and Experiences from Developing Countries, edited by N. Hermes and R. Lensink. Studies in Development Economics, vol. 6. London and New York: Routledge, 7-34. 9 1996b. "Financial Development, Policy and Economic Growth." In Financial Development and Economic Growth: Theory and Experiences from Developing Countries, edited by N. Hermes and R. Lensink. Studies in Development Economics, vol. 6. London and New York: Routledge, 7-34. Coelli, Tim J. 1996. "A Guide to FRONTIER Version 4.1: A Computer Program for Stochastic Production and Cost Function Estimation9 Centre for Efficiency and Productivity Analysis, University of New England, Armidale, NSW, Australia, working paper 96/07. De Gregorio, J., and P. E. Guidotti. 1995. "Financial Development and Economic Growth." Worm Development 23: 433-448. Demetriades, P. O., and K. A. Hussein. 1996. "Does Financial Development Cause Economic Growth? Time-Series Evidence from Sixteen Countries." Journal of Development Economics 51. Demirguc-Kunt, A., and V. Maksimovic. 1998. "Law, Finance, and Firm Growth." Journal of Finance 53: 2107-21379 Easterly, W., and R. Levine. 2001. "It's Not Factor Accumulation: Stylized Facts and Growth Models." World Bank Economic Review 15: 177-219. Harris, R. D. F. 1997. "Stock Markets and Development: A Re-assessment." European Economic Review 41: 139-146. Hodrick, R. J., and E. C. Prescott. 1997. "Postwar U.S. Business Cycles: An Empirical Investigation." Journal of Money, Credit, and Banking 29: 1-16.
Jayaratne, J., and P. E. Strahan 1996. '~I'he Finance-Growth Nexus: Evidence from Bank Branch Deregulation." Quarterly Journal of Economics 111: 639-6709 Jung, W. S. 1986. "Financial Development and Economic Growth: International Evidence." Economic Development and Cultural Change 34: 333-346. King, R. G., and R. Levine. 1993a. "Finance and Growth: Schumpeter Might Be Right." Quarterly Journal of Economics 108:717-737 9 1993b. "Finance, Entrepreneurship, and Growth: Theory and Evidence." Journal of Monetary Economics 32:513-542. Koop, G., Osiewalski, J, and M. F. J. Steel. 2000. "Modeling the Sources of Output Growth in a Panel of Countries." Journal of Business and Economic Statistics 18: 284-299. Levine, R. 1997. "Financial Development and Economic Growth: Views and Agenda." Journal of Economic Literature 35: 688-726.
148
JOURNALOF ECONOMICSAND FINANCE 9 Volume 26 9 Number 2 9 Summer 2002
Lcvine, R., Loayza, N., and T. Beck. 2000. "Financial Intcrmediation and Growth: Causality and Causes." Journal of Monetary Economics 46: 31-77. Levine, R., and S. Zervos. 1998. "Stock Markets, Banks, and Economic Growth." American Economic Review 88: 537-558. Lucas, R. J. 1988. "On the Mechanics of Economic Development." Journal of Monetary Economics 22: 3-42. Mankiw, N. G., Romer, D., and D. Weft. 1992. "A Contribution to the Empirics of Economic Growth." Quarterly Journal of Economics 107: 407-437. Moroney, J. R., and C. A. K. Lovell. 1997. "The Relative Efficiencies of Market and Planned Economies." Southern Economic Journal 63: 1084-1093. Rousseau, P. L., and P. Wachtel. 2000. "Equity Markets and Growth: Cross-Country Evidence on Timing and Outcomes, 1980-1995." Journal of Banking and Finance 24: 1933-1957. Rajah, R. G., and L. Zingales. 1998. ''Financial Dependence and Growth." American Economic Review 88: 559-586. Robinson, J. 1952. The Rate of Interest and Other Essays. London: McMillan. Romer, P. M. 1986. "Increasing Returns and Long-Run Growth." Journal of Political Economy 94: 1002-1037. Schumpeter, Joseph. 1912. Theorie der Wirtschafilichen Entwicklug (The Theory of Economic Development), 1934 translated ed. Cambridge, Massachusetts: Harvard University Press. Wurgler, J. 2000. "Financial Markets and the Allocation of Capital." Journal of Financial Economics 58: 187-214.