Finance and inequality: the case of India.
Ang, James B.
1. Introduction
Although the relationship between financial development and
economic growth has been extensively studied in the literature (see, for
example, King and Levine 1993; Demetriades and Hussein 1996; Arestis and
Demetriades 1997; Levine, Loayza, and Beck 2000; Bell and Rousseau 2001;
Ang and McKibbin 2007), little is known about how finance impacts income
inequality. The importance of the finance-inequality relationship has
recently been highlighted in an insightful survey article by Claessens
and Perotti (2007). They indicate that while financial development can
help reduce income inequality, financial liberalization captured by
established interests may do the opposite.
The theoretical predictions of the effects of finance on income
inequality are controversial. Rajan and Zingales (2003b) argue that
improvements in the formal financial sector primarily benefit the rich.
Greenwood and Jovanovic (1990) predict a nonlinear relationship between
financial development and income inequality, where it is hypothesized
that income inequality first increases with the degree of sophistication
in the financial systems, then stabilizes and eventually declines.
Others propose that the presence of financial market imperfections
deters the poor from borrowing adequately to invest in human and
physical capital, implying that financial development helps alleviate
income inequality (Banerjee and Newman 1993; Galor and Zeira 1993;
Aghion and BoRon 1997; Mookherjee and Ray 2003, 2006). Given that
theories provide ambiguous predictions regarding the effects of finance
on the distribution of income, it is useful to approach the issue at the
empirical level. This could facilitate our understanding of the
relationship between finance and inequality, and help us to assess the
validity of each theoretical model.
Despite the important role of financial market frictions in the
theories of poverty and income inequality, researchers so far have not
adequately addressed whether financial development, and in particular
financial sector policies, affect income inequality (Demirguc-Kunt and
Levine 2007). In this connection, there are two novel studies that focus
on examining the effect of financial development on income inequality.
Using data for 83 countries over the period 1960-1995, Clarke, Xu, and
Zou (2006) examine the effect of financial development on the level of
the Gini coefficient--a measure of deviations from perfect income
equality. Their results show that financial deepening is associated with
lower income inequality. The finding of a non-linear effect of financial
development is not robust. A more recent study by Beck, Demirguc-Kunt,
and Levine (2007) attempts to assess the impact of financial development
on changes in income distribution and income for the poor. Their main
findings indicate that financial development is associated with a lower
growth rate of the Gini coefficient and a higher growth rate of income
for the poor. While these two studies have established that financial
development helps reduce income inequality, studies examining the direct
impact of financial liberalization on income inequality are particularly
scant (Demirguc-Kunt and Levine 2007). The limited indirect empirical
evidence, based on the survey by Arestis and Caner (2004), seems to
suggest that financial liberalization has ambiguous effects on the poor
and income distribution.
This article attempts to contribute to this rather under-researched
area by considering an important case study--that of India for the
period 1951-2004. Specifically, I analyze the distributional impact of
financial development and financial liberalization on the Gini
coefficient. (1) The article aims to complement the above studies, and
enrich the literature by providing further evidence on how development
of financial systems and implementation of financial sector policies
affect the evolution of income inequality, drawing on the experience of
one of the most rapidly growing developing economies that has undergone
significant financial sector reforms. I focus on India rather than a
larger set of countries given that the effects of financial development
and financial liberalization may be heterogeneous across countries at
different stages of economic development. Moreover, case studies are
particularly useful in disentangling the complexity of the financial
environment and economic history of each country. By analyzing case
studies, the econometric findings of this project can be related to the
prevailing institutional structure (Bell and Rousseau 2001), and
therefore inform academic as well as policy debate.
The main contributions of this study include: (i) empirically
testing the effect of financial development on income inequality by
providing further evidence from a large and fast-growing developing
country. Not only could this enhance our understanding of the
finance-inequality relationship, but also fill the gap in the extant
literature, which is dominated by cross-country analysis; (ii)
contributing to the debate on the effectiveness of financial
liberalization on the Indian economy--although various financial
restructuring programs have been launched since the early 1990s, there
is little empirical evidence informing policy makers of the effects of
these reforms--and (iii) complementing the literature by assessing the
impact of financial liberalization on income inequality. This policy
factor has been somewhat neglected in the analysis of the
finance-inequality nexus. The results show that income inequality
decreases as the financial system deepens and broadens, consistent with
the general findings in the literature. However, liberalization of the
financial systems appears to have a harmful effect on income
distribution, a finding that tends to support the political economy
argument where captured financial sector reforms benefit only small
elites.
The remainder of the article is structured as follows. The next
section describes the financial repression and liberalization experience
of India. Section 3 discusses income inequality and the policies adopted
to alleviate poverty. Section 4 briefly reviews the theoretical
literature on the relationship between finance and inequality. The model
and data are described in section 5. The estimation techniques employed
in this study are explained in section 6. The results are presented and
analyzed in section 7. Finally, section 8 summarizes the main findings
and concludes.
2. Financial Sector Reforms in India
There was little financial repression imposed on the Indian
financial system in the 1950s. However, the Reserve Bank of India
gradually imposed more controls over the financial system by introducing
interest rate controls in the 1960s. The statutory liquidity ratio was
raised from 25% in 1966 to 38% in 1989. The cash reserve rate increased
considerably from 3% to 15% during the same period. These high liquidity
and reserve requirements enabled the Bank to purchase government
securities at low cost. The extent of directed credit programs has also
increased significantly after the nationalization of the 14 largest
private banks in 1969. A number of priority lending rates were set at
levels well below those that would prevail in the free market. This
process culminated in the late 1980s when directed lending accounted for
more than 40% of total lending. Revenue from financial repression was
estimated to be 22.4% of total central government revenue during the
period 1980-1985 (see Giovannini and De Melo 1993).
The major phase of financial liberalization was undertaken in 1991
as part of the broader economic reform in response to the
balance-of-payments crisis of 1990. The objective was to restructure the
entire orientation of India's financial development strategy from
its position as a financially repressed system to one that was more open
in order to provide a greater role for markets in price determination
and resource allocation. Consequently, interest rates were gradually
liberalized, and reserve and liquidity ratios were reduced
significantly. The equity market was formally liberalized in 1992;
although, the first country fund was set up earlier in 1986, allowing
foreign investors to access the domestic equity market directly. There
has also been a change in the capital account regime from a restricted
one to a more open one. The regulatory framework was strengthened
significantly in 1992. In addition, entry restrictions were relaxed in
1993, resulting in the establishment of more private and foreign banks.
Regulations on portfolio and direct investment have since been eased.
The exchange rate was unified in 1993-1994 and most restrictions on
current account transactions were eliminated in 1994.
However, despite this liberalization, the Indian financial system
has continued to operate within the context of repressionist policies
through the provision of subsidized credit to certain priority sectors.
The bank nationalization program in 1969 has enabled the Reserve Bank of
India to effectively implement its credit allocation policy. Although
the government divested part of its equity position in some public banks
in the 1990s, the banking sector has remained predominantly state-owned.
(2) Liberalization of the directed credit programs is only limited to
deregulation of priority lending rates while significant controls on the
volume of directed lending remain in place. Furthermore, the Bank has
tightened supervision and regulation in recent years to ensure that
these priority sector requirements are met. With regard to capital
controls, transactions related to capital outflows have remained heavily
regulated in India. As such, it appears that repressionist measures
coexist with a set of liberalization policies aimed at promoting free
allocation of resources.
3. Macroeconomic Policy and Income Inequality in India
Indian governments have accorded great importance to poverty
eradication and rural employment generation since independence in 1947,
as reflected by the inclusion of poverty alleviation as a major goal in
all the Five-Year Plans. In particular, a number of antipoverty programs
have been implemented since the early 1960s, most notably the National
Rural Employment Program and the Rural Landless Employment Guarantee
Program, to expand employment opportunities in rural areas. Other
strategies include, among others, land reforms, food subsidies, price
controls, concessional loans, and rural housing schemes.
In terms of financial sector policies, the presence of significant
directed credit programs has mandated minimum lending to priority
sectors such as agriculture and small and medium-sized enterprises so
that the poor are not excluded from formal credit. The share of directed
credit in total lending grew by more than 10-fold from 2.5% to 25.9%
during the period 1950-1990. The imposition of interest rate ceilings
from the 1960s to 1980s also ensured that the poor were able to obtain
financing at a reasonably low cost. Furthermore, as part of the effort
to increase bank presence in rural areas, the Reserve Bank of India
imposed a 1:4 branch license policy in 1977, which required banks to
open four branches in rural areas without a bank presence for every new
branch opened in locations that already had a bank presence. Burgess and
Pande (2005) find that this social banking program has substantially
reduced rural poverty. However, this policy ended with the onset of
financial sector reforms in 1991.
The 1950s and 1960s saw a significant reduction in income
inequality, as indicated by a downward trend in the Gini coefficient
(see Figure 1). (3) However, despite this positive development, a
significant number of people continued to live below the poverty line.
Reduction in income inequality, however, slowed in the 1970s and 1980s.
While the reasons for this change are not entirely clear, it was
probably due to lackluster agricultural performance and higher
inflation. Income inequality increased sharply following the 1990
balance-of-payments crisis. Since then, the strategy for alleviating
poverty has been shifted to the acceleration of growth and the creations
of jobs for the poor. Although the Indian economy has achieved
remarkable growth since the reforms in the early 1990s, the reversed
trend in the Gini coefficient suggests that these reforms have been
accompanied by a significant rise in income disparity.
[FIGURE 1 OMITTED]
4. Conceptual Issues on Finance and Inequality
Developing countries are often characterized by the presence of
credit constraints due to market imperfections such as asymmetric
information and moral hazard problems. These credit constraints may
intensify income inequality since the poor may not have equal access to
credit due to the lack of collateral and established relationships with
financial institutions. The relaxation of credit constraints achieved
through improvements in the financial systems enables efficient
allocation of resources and thereby reduces income inequality. In the
models developed by Banerjee and Newman (1993), Galor and Zeira (1993),
Aghion and Bolton (1997), and Mookherjee and Ray (2003, 2006), among
others, only rich agents can borrow enough to invest in human capital
and high-yield investment projects due to credit market imperfections.
Their models imply that borrowing constraints triggered by market
failures could result in greater income inequality, and consequently,
financial development helps alleviate income inequality.
On the other hand, Rajan and Zingales (2003a) argue that in the
presence of weak institutional environments, de jure political
representation is dominated by de facto political influence. This allows
established interests to influence access to finance, implying that
higher financial development induced by captured direct controls is
likely to hurt the poor. Rajan and Zingales (2003b) further argue that
development of financial systems is more likely to benefit the rich and
well connected since they have sufficient wealth for collateral (dubbed
"the tyranny of collateral"). The rich may also be able to
prevent small firms from accessing external finance and reduce the
ability of the poor to improve their economic well-being. Thus, the poor
are often excluded from finance and are therefore unable to invest
sufficiently in human and physical capital.
However, financial development and income inequality can also be
characterized by a hump-shaped relationship. In an influential article,
Greenwood and Jovanovic (1990) present a theoretical model that predicts
a non-monotonic relationship between the two variables. They postulate
that access to finance involves a fixed transaction cost that poor
households cannot afford. Such a market imperfection therefore causes
deterioration in their relative position in the distribution of income.
However, as the economy becomes more developed, the transaction costs of
using financial services decline, allowing the majority to access
finance so that financial deepening narrows the income gap between the
rich and the poor. Hence, it appears that the above theoretical models
offer quite different perspectives about the relationship between
financial development and income inequality.
How financial liberalization impacts income inequality is also
theoretically ambiguous. Arestis and Caner (2004) propose that there are
three main channels through which financial liberalization can influence
poverty and income inequality. The first, known as the economic growth
channel, posits that financial liberalization affects income inequality
through increasing the rate of economic growth based on the financial
liberalization thesis of McKinnon (1973) and Shaw (1973). However, this
depends on the empirical links between financial liberalization and
economic growth as well as economic growth and income distribution,
which are not necessarily unambiguous. The second, the financial crisis
channel, works through changes in macroeconomic volatility triggered by
crises following financial liberalization. The poor are likely to be
more vulnerable to these negative shocks. Finally, the last channel
proposes that improved access to credit and financial services due to
financial liberalization can have a profound effect on income
distribution.
5. Model and Data
Model Specification
My empirical model postulates that income inequality ([Gini.sub.t])
depends on per capita growth rate of real GDP ([GRO.sub.t]), (4)
inflation rate ([INF.sub.t]), trade openness ([TO.sub.t]), and a
variable that captures the effect of finance, as given in Equation 1.
The inflation rate is measured by the growth rate of the GDP deflator
and trade openness is the share of exports plus imports in GDP. For the
finance variable, my focus is on the level of financial development and
financial liberalization; although, I also take into consideration their
growth and non-linear effects. Besides, I also pay attention to stock
market volatility and banking sector efficiency. The model will be
estimated using annual data for India over the period 1951-2004. Except
for the Gini coefficient, all data series are directly obtained or
compiled from the Report on Currency and Finance of the Reserve Bank of
India and National Accounts Statistics of the Central Statistical
Organisation in India.
[Gini.sub.t] = [[beta].sub.0] + [[beta].sub.1] [GRO.sub.t] +
[[beta.sub.2] [INF.sub.t] + [[beta].sub.3] [TO.sub.t] + [[beta].sub.4]
[Finance.sub.t] + [[epsilon].sub.t] (1)
[FIGURE 2 OMITTED]
Measures of Financial Development and Financial Liberalization
Financial development is measured by several commonly used
indicators in the literature: the ratio of claims on the private sector
to GDP, the ratio of broad money M3 minus M1 to GDP, the share of
commercial bank assets in the sum of commercial and central bank assets,
and banking density, which is measured by the number of bank offices per
population. Figure 2 shows the evolution of these variables. It is
evident that all these financial development indicators exhibit an
upward trend over time.
This study uses the financial liberalization measure advanced by
Demetriades and Luintel (1996, 1997). Their approach considers nine
indicators of financial repressionist policies. Six of them are interest
rate controls, including a fixed lending dummy, a minimum lending rate,
a maximum lending rate, a fixed deposit dummy, a minimum deposit rate,
and a maximum deposit rate. These policy controls are translated into
dummy variables that take the value of 1 if a control is present and
that equal 0 otherwise. The remaining three policies are directed credit
programs, the cash reserve ratio, and the statutory liquidity ratio. The
extent of directed credit programs is measured by the share of directed
credit lending in total lending. (5,6) The other two variables are
direct measures expressed in percentages.
Since I want to summarize the financial sector policies to obtain
an overall measure of financial liberalization, the method of principal
component analysis seems to be a natural choice. It is a systematic and
sophisticated way of examining the patterns of relationship among the
variables, with the objective of summarizing the information content of
several observed variables into a handful of representative principal
components. The method involves computing the linear combinations of the
original variables so that the resulting principal components can
capture a large proportion of the variance in the original variables.
This can therefore serve the same purpose as the full set of original
variables, but in a much more succinct manner. Given that the principal
components are uncorrelated to each other and their conciseness, this
approach sufficiently deals with the problems of multicollinearity and
over-parameterization.
To provide a sensitivity check I also consider two alternative
measures of financial liberalization. Firstly, the approach of
Demetriades and Luintel (1996, 1997) is modified to allow for the policy
changes that took place after the liberalization since their work covers
only the period to 1991, prior to the reform programs. The modification
involves taking into consideration privatization in the financial
sector, entry barriers in the banking sector, government regulations on
banking operations, equity market liberalization, and restrictions on
international capital flows. I use dummy variables to represent policy
changes in these dimensions. (7)
In constructing the third summary measure of financial
liberalization, I follow the approach of Abiad and Mody (2005). In
particular, six policy dimensions are considered as the inputs to
construct the measure: (i) credit controls and reserve requirements;
(ii) interest rate restraint; (iii) entry barriers in the banking
sector; (iv) government regulations of operations; (v) privatization in
the financial sector; and (vi) restrictions on international capital
flows. Along each dimension, a score of 0, 1, 2, or 3 is assigned,
indicating fully liberalized, partially liberalized, partially
repressed, and fully repressed, respectively. The aggregation of these
six components is used to obtain an overall measure of financial
repression. (8) Similar to the second approach, this provides a more
broad-based measure of financial sector reforms because it considers
several other dimensions in addition to credit and interest controls.
The inverse of these composite measures can be interpreted as the extent
of financial liberalization (see, for example, Ang and McKibbin 2007;
Ang 2008). All data series are directly obtained or compiled from the
Annual Report and the Report on Currency and Finance of the Reserve Bank
of India.
The resulting three composite financial liberalization indices
displayed in Figures 3a and 3b coincide rather well with the actual
policy changes that took place in India during the sample period, as
discussed earlier. In Figure 3a, the two measures of financial
liberalization show increasing disparity since the early 1970s given
that the second measure captures more dimensions of financial sector
reforms. It therefore necessarily reflects a greater degree of financial
liberalization compared to the first measure that focuses exclusively on
credit and interest controls. The financial liberalization series
depicted in Figure 3b shows a rather different pattern of development,
due largely to the use of a different coding procedure. On the whole, it
is evident that the trend towards financial repression has been reversed
since the early 1990s. The leveling-off observed in the series coincides
with the increase in the extent of directed credit programs in recent
years.
[FIGURE 3 OMITTED]
Correlations
Table 1 presents the correlations between indicators of financial
development and financial liberalization. All four measures of financial
development appear to be positively and strongly correlated with each
other, and the coefficients range from 0.903 to 0.980. Indicators of
financial liberalization show similar correlation patterns, ranging from
a low of 0.708 to a high of 0.998. These two different aspects of
finance appear to be negatively related, suggesting that they are likely
to have different effects on income inequality. (9) However, it may be
difficult to enter both variables in the same regression given the high
correlations between them. All correlations are statistically
significant at the 1% level.
6. Estimation Techniques
The dynamic adjustment of the Gini coefficient can be characterized
by a conditional error-correction model (ECM), which can be used to test
for the existence of a long-run relationship using the autoregressive
distributed lag (ARDL) bounds test developed by Pesaran, Shin, and Smith
(2001) and the ECM test of Banerjee, Dolado, and Mestre (1998). The
former involves a standard F-test; whereas, the latter is a simple
t-test. Accordingly, the underlying ECM can be formulated as follows:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (2)
where p is the lag length and [DET.sub.t] is a vector of k
determinants of [Gini.sub.t]. The above equation can be estimated by
ordinary least squares (OLS) because Pesaran and Shin (1998) have shown
that the OLS estimators of the short-run parameters are consistent, and
the ARDL-based estimators of the long-run coefficients are
super-consistent in small sample sizes. Hence, valid inferences on the
long-run parameters can be made using standard normal asymptotic theory.
The main advantage of this approach is that it can be applied to the
model regardless of whether the underlying variables are I(0) or I(1).
The testing procedure involves two stages. In the first stage, the
existence of the long-run relationship between the variables is tested.
Specifically, two separate statistics are employed to test for the
existence of a long-run relationship in Equation 2: (i) an F-test for
the joint significance of coefficients on lagged levels terms of the
conditional ECM ([H.sub.0] : [b.sub.0] = [b.sub.l] = ... = [b.sub.k] =
0), and (ii) a t-test for the significance of the coefficient associated
with [Gini.sub.t-1] ([H.sub.0] : [b.sub.0] = 0). The test for
cointegration is provided by two asymptotic critical value bounds when
the independent variables are either I(0) or I(1). The lower bound
assumes all the independent variables are I(0), and the upper bound
assumes they are I(1). If the test statistics exceed their respective
upper critical values, the null is rejected, and it can be concluded
that a long-run relationship exists. The second stage of the procedure
is to derive the long-run estimates using the underlying ARDL model.
7. Empirical Results
Financial Development and Income Inequality
I begin the empirical analysis by assessing the integration
properties of the underlying variables. Two standard unit root tests
were used to assess the order of integration of the underlying
variables--the Augmented Dickey-Fuller (ADF) test and Phillips-Perron
(PP) test. The results, which are not reported in order to conserve
space but are available upon request, show that all variables appear to
be I(0) or I(1), suggesting that there is no variable integrated at an
order greater than one. This allows legitimate use of the ARDL bounds
and ECM tests since these procedures require all underlying variables to
be integrated at an order less than two. Next, to perform the
cointegration tests, I estimate Equation 2 with only one lag in order to
conserve the degrees of freedom, given the small sample used in this
study (54 annual observations). (10) The results reported in panel I of
Table 2 strongly suggest that a long-run relationship is present for
each of the equations estimated.
However, the relationship between financial development and income
inequality may be driven by reverse causality since lower income
inequality may result in greater political pressure to create a more
market-driven type of financial system in order to ensure efficient
allocation of resources (Beck, Demirguc-Kunt, and Levine 2007).
Moreover, banks may also prefer to open branches in richer areas (Rajan
2006). To address the concern of endogeneity bias, I have attempted to
use financial development as the dependent variable. However, no
evidence of cointegration is found when the measures of financial
development are used as the dependent variables. Hence, the results
suggest that financial development can be interpreted as one of the
long-run forcing variables explaining the evolution in the Gini
coefficient where a reverse causation is absent.
In testing the effect of financial development on the Gini
coefficient, I control for per capital GDP growth, trade openness, and
inflation. These control variables have also been used by Beck,
Demirguc-Kunt, and Levine (2007). Panel II in Table 2 reports the
long-run estimates derived using the underlying ARDL model. The negative
sign and the significance of the coefficient for the growth rate of per
capita real GDP suggest that income growth helps alleviate income
inequality, a result consistent with the general literature on growth
and inequality. In line with the results of Barro (2000), openness to
trade enters positively and significantly in all equations except column
5.
The rate of inflation enters negatively and significantly in all
regressions, suggesting that monetary instability does not seem to hurt
income distribution. In principle, inflation may have an adverse effect
on real agricultural wages and, hence, income inequality. However, it
may also be associated with a decline in unemployment due to lower real
wages and thus benefit the poor. The results are consistent with the
cross-country findings of Cutler and Katz (1991) and Clarke, Xu, and Zou
(2006), but stand in sharp contrast to those of Easterly and Fischer
(2001) and Beck, Demirguc-Kunt, and Levine (2007). However, as
highlighted by Easterly and Fischer (2001), the way inflation affects
the poor may well differ between economies due to the complication of
the tax system, and therefore is an empirical issue.
Financial development appears to be associated with lower income
inequality. The results are compatible with similar studies in this
area, in particular Honohan (2004a), Clarke, Xu, and Zou (2006), and
Beck, Demirguc-Kunt, and Levine (2007). In column 1, an increase in the
ratio of private credit to GDP has a significant and favorable effect on
income inequality (long-run elasticity of -0.066), indicating that
financial deepening has an equalizing effect. In columns 2 and 3, I use
two alternative indicators of financial development, but this does not
alter my main findings.
Column 4 shows the effects of banking density on income inequality.
The use of banking density as an indicator of financial development has
a major advantage--it captures the breadth of financial systems;
whereas, other indicators reflect their depth. This is particularly
important since the theories in finance and income inequality focus on
the importance of broad access to finance (Demirguc-Kunt and Levine
2007). The results indicate that bank branch expansion is associated
with lower income inequality, a finding in line with the results of
Burgess and Pande (2005) for the Indian experience. Thus, the social
banking program launched by the Indian government during the period
1977-1990 appears to have significantly improved the access of the poor
to the formal financial sector.
Studies have suggested that stock market activity may also predict
growth (for example, Levine and Zervos 1998; Arestis, Demetriades, and
Luintel 2001; Beck and Levine 2004). Given that the financial
development measures I have considered so far are primarily bank-based
in nature, I also take into account a market-based measure. The results
in column 5 show that stock market development has no statistically
significant impact on income inequality. However, the results must be
interpreted with caution because the measure of stock market development
is a rather noisy indicator of financial development. The series was
backdated using the share price index for the period before 1976 due to
data unavailability. I am unable to relate the findings to the
literature because the impact of stock market development on the Gini
coefficient has not been studied in the literature.
I also include a proxy for modern sector development in the
estimation due to Clarke, Xu, and Zou (2006). This allows me to examine
how the sectoral structure of the economy affects income inequality. The
influence of the modern sector is measured by the share of the
industrial and service sector's value added in total GDP. When the
economy moves away from subsistence agriculture activities to advanced
service-oriented activities, this may be reflected by credit facilities
becoming more readily and cheaply available. Thus, in general, countries
with more developed financial systems tend to have relatively larger
service and industrial sectors, and therefore this measure may provide
an indirect indicator of financial development. The results reported in
the last column show that development in the non-farm sector has an
equalizing effect, consistent with the findings of Datt and Ravallion
(1998) for India.
Panel III reports some diagnostic statistics. I do not find any
evidence of serial correlation, functional misspecification, and
heteroskedasticity at the 5% level of significance. In sum, the results
suggest that financial development helps in reducing the Gini
coefficient through helping the poor to access finance by reducing
financial market frictions. Thus, quite consistent with the vast
literature showing a positive relationship between financial development
and economic growth, the results reveal that financial development is
effective in reducing income inequality in India.
Financial Liberalization and Income Inequality
Evolution of financial market frictions in the financial systems,
which can have a significant impact on access to finance and thus income
inequality, may be driven by financial sector policies. Therefore, I
want to examine the extent to which financial sector policies matter to
income inequality. This addresses the concern raised by Demirguc-Kunt
and Levine (2007) that there is surprisingly little research that
investigates whether financial sector policies influence the evolution
of the distribution of income. I first look at how each type of
financial sector policy affects income inequality. This includes
examining the three main components of domestic financial sector
reforms--directed credit programs, reserve and liquidity requirements,
and interest rate restraint (see McKinnon 1973; Shaw 1973).
As explained earlier, the extent of directed credit controls is
measured by the share of direct lending in total lending. Reserve and
liquidity requirements are the sum of the cash reserve and statutory
liquidity ratios. The index for interest rate restraint is constructed
using the method of principal component using the six interest control
variables discussed in Measures of Financial Development and Financial
Liberalization. The results reported in Table 3 show that the estimated
elasticity of the Gini coefficient with respect to a steady-state
increase in the extent of directed credit programs is -0.249, and for
reserve and liquidity requirements and interest rate restraint, the
elasticities are -0.229 and -0.035, respectively (see columns 1-3).
Taken together, the results in columns 1-3 appear to suggest that
financial repressionist policies in India were pro-poor--as indeed they
were designed to be--and thus financial liberalization aggravated income
inequality. To confirm this, columns 4-6 examine how income inequality
responds to the overall financial sector reforms in India using
different measures of financial liberalization. The measure used in
column 4 encompasses the three financial sector policies considered
earlier, where they are summarized into just one single variable. I take
the inverse so that this composite variable can be interpreted as
financial liberalization.
The results reported in column 4 show that the measure of financial
liberalization is significantly and positively associated with income
inequality, with a long-run elasticity of 0.074. To provide some
sensitivity checks, I also consider two other broader indicators of
financial liberalization (see details in Measures of Financial
Development and Financial Liberalization). Financial liberalization
continues to enter positively and significantly even when I use
different measures in columns 5 and 6. The estimates are found to be
0.077 and 0.019, respectively. Thus, the results unanimously show that
financial liberalization appears to have a harmful effect on the
distribution of income in India, confirming the earlier findings in
columns 1-3. In all cases, I continue to find evidence of cointegration
(panel I). There is also little evidence of econometric problems (panel
III).
Similar to the effect of financial development, changes in the
distribution of income may also affect the political economy in shaping
financial sector policies so that variations in the Gini coefficient may
influence the composite indices of financial liberalization and bias the
results. I address this concern by also treating the measures of
financial liberalization as the dependent variables but no evidence of
cointegration is found (results not reported). This provides some
support for entering financial liberalization "exogenously" in
the specification.
The Joint Effects of Financial Development and Financial
Liberalization
The evidence presented thus far suggests that while financial
development has an equalizing effect on income distribution, financial
liberalization tends to work in the opposite direction. The results are
obtained by entering these two finance variables separately in the
regressions. Nevertheless, given that financial development and
financial liberalization are two distinct aspects of finance emphasized
in this article, it would be interesting to see if the results remain
robust to the inclusion of these two variables in the same
specification. (11) However, this may present some estimation
difficulties since these two variables are strongly correlated (see
Table 1). To mitigate the problems of multicollinearity, I have chosen
the third financial liberalization measure, which is based on the
approach of Abiad and Mody (2005), given that it is least correlated
with all other measures of financial depth. It would also be interesting
to see how financial breadth (proxied by banking density) influences
income inequality. Its inclusion with financial depth variables,
however, can cause severe collinearity problems given their high
correlations. I have therefore considered the average value of banking
density and financial depth as an overall measure of financial
development.
As can be seen from Table 4, this set of results is econometrically
less robust compared to the case where the finance variables are entered
individually in the regressions. This is probably due to the strong
correlations between the finance variables, although attention has been
paid to mitigate any problems associated with multicollinearity. (12) In
only half of the cases, financial development is found to be
significantly associated with lower income inequality. While financial
liberalization has the sign consistent with the previous findings, it is
only significant in one out of six cases. The evidence of cointegration
is also considerably weaker. Nevertheless, a general picture emerges
from these analyses--financial development and liberalization appear to
have different effects on income inequality--a result consistent with
the findings reported in Tables 2 and 3 earlier.
Although in principle financial sector reforms may reduce market
frictions and thereby alleviate income inequality, the results seem to
suggest that liberalization of the domestic and international financial
system has led to an increase in income inequality in India, providing
some support to the arguments of Claessens and Perotti (2007). How could
these results be interpreted within the specific context of India?
In India, directed credit has been extended to the agricultural
sector and small and medium enterprises over the last few decades. These
programs have significantly benefited farmers and small traders,
allowing the poor direct access to financial services. Therefore,
reducing the extent of these programs as part of the financial sector
reforms is likely to hurt the poor. Similarly, the deregulation of
interest rates may increase the costs of borrowing to the poor since
this involves higher transaction costs relative to the size of the loan.
The resulting higher borrowing costs, along with the reduction of direct
lending, can have an undesirable effect on income inequality since these
policies deter the poor from adequately accessing finance. (13)
Financial liberalization in India did not necessarily lead to a
relaxation of credit constraints to poor individuals that result in
lower inequality. Before the liberalization, banks were required to open
a certain number of branches in rural areas, and this policy was an
important factor behind the savings rate increases of the 1970s and
1980s in India. However, this requirement was relaxed in 1991 following
the launch of financial reforms. Thus, foreign and private banks would
necessarily have a bias in providing consumer credit to richer areas,
and access to finance by the poor would fall as banks withdrew branches
from rural areas.
As Aghion, Bacchetta, and Banerjee (2004) have shown, unrestricted
financial liberalization may induce instability. While financial
repression may not be desirable, the evidence presented in this article
does provide some support to the argument that some form of financial
restraint may help in alleviating income inequality in developing
countries. However, as noted by Demetriades and Luintel (2001) and
Honohan and Stiglitz (2001), financial restraints are more likely to
work well in environments with strong regulatory capacity, pinpointing
the importance of strengthening the institutional framework.
For instance, Beck, Levine, and Levkov (2008) find that bank branch
deregulation reduces the Gini coefficient of income inequality in the
United States, a result that contradicts the findingsfor India. This
highlights the fact that the effect of financial deregulation on income
inequality may depend on the quality of institutions. As Rajan and
Zingales (2003b) propose, the process of financial liberalization is
likely to be harmful for countries with a weak institutional
environment. Although the legal system in India was originally based on
the British model that emphasizes protection of property rights, India
ended up with a much less effective institutional framework since the
legal system was modified in a way that benefited the small number of
Europeans that settled in and ran the economy (Mishkin 2006).
In sum, financial sector reforms may lead to well-connected elites
capturing most of the gains from new opportunities (Rajan and Zingales
2003a, b; Claessens and Perotti 2007). Therefore, the presence of these
established interests may deepen rather than broaden access to finance,
resulting in higher income inequality. For example, Das and Mohapatra
(2003) have shown that stock market liberalization in emerging markets
has mainly benefited high income individuals at the expense of others.
Furthermore, the presence of weak institutional environments in many
liberalizing markets has allowed insiders to expropriate the interests
of minority shareholders (Claessens et al. 2002; Claessens 2006). Unlike
central and eastern Europe, where the extent of captured reform and rent
seeking is limited (Roland 2002), in India financial liberalization has
primarily benefited the well-connected rich, leaving the poor to fall
further behind due to unequal access to finance.
Recent opinion polls conducted by the National Election Study (NES)
in 2004 complement the findings. This election analysis shows that more
than two-thirds of the respondents believe that the benefits of the
reforms have primarily been reaped by a small number of affluent
individuals (Suri 2004). In this connection, Varshney (2007) also argues
that elite-oriented reforms have continued whereas more pro-poor reforms
have been delayed in India. Similarly, Jha (2004) suggests that reforms
allow the rich to move financial resources more easily in their favor,
and this leads to greater income disparity. There is further evidence
suggesting that finance in India is subject to political capture. For
instance, using data for the period 1992-1999, Cole (2009) finds that
agriculture lending by government-owned banks increases significantly in
election years and the surge is most prominent in the highly contested
districts. Given that these banks often prefer lending to firms with
connections to business groups or politicians (Gormley and Gopalan
2008), the results suggest that these loans are subject to substantial
capture in the post-reform period. Using individual tax return data,
Banerjee and Piketty (2005) show that the top 0.1% of income tax payers
have gained disproportionately in the 1990s, suggesting that the
ultra-rich have benefited significantly from the reforms.
An important component of financial liberalization is the easing of
priority sector loans. Although the actual share of directed loans in
total lending remains high in recent years, bank compliance with these
targets reduced sharply after financial liberalization following a
change in the priority sector definition to include many other
activities. As a result, most banks have avoided lending to small
farmers and small entrepreneurs who are deemed less creditworthy,
thereby depriving the relatively poor of institutional lending for
investing in their businesses. This has negatively affected income
distribution since a majority of the Indian population is dependent on
the agriculture sector. Alongside this development, Pal and Ghosh (2007)
note that there has been anecdotal evidence suggesting that a number of
large Indian industrial incumbents are in fact responsible for a
significant share of the non-performing loans in recent years. However,
strong connections with political elites prevent any legal actions
against these default firms.
The Effects of Financial Growth, Non-Linearity, Efficiency, and
Volatility on Income Inequality
Next, I analyze the results further by examining the growth and
non-linear effects of financial development and financial
liberalization. Since the earlier results reveal that the relationship
between income inequality and finance is not sensitive to the measures
of financial development and financial liberalization, I use only the
ratio of private credit to GDP and the approach of Demetriades and
Luintel (1996, 1997), respectively, in performing these additional tests
for brevity. The results in the first two columns in Table 5 show that
while the growth rate of financial development is found to be
statistically significant at the conventional levels, the growth rate of
financial liberalization is not.
In addition, I do not find any evidence of a non-linear effect of
finance, a finding consistent with Beck, Demirguc-Kunt, and Levine
(2007). When the linear and squared terms for financial development are
added to the base model, these two terms become jointly insignificant,
suggesting that a threshold effect of financial development is not
present in the relationship (column 3). Similarly, I do not find any
significant non-linear effect of financial liberalization. The
coefficient of squared financial liberalization is insignificant;
although, its linear term is weakly significant at the 10% level (column
4).
In order to shed additional light, I test the effects of banking
efficiency and volatility in the share market on the degree of economic
opportunity. Demirguc-Kunt and Levine (2007) propose that financial
innovation can affect the level and evolution of financial market
frictions, and therefore has an impact on income inequality. On the
other hand, although in principle macroeconomic volatility may affect
economic growth (see Ramey and Ramey 1995), empirical research has not
yet established whether financial volatility has an impact on income
distribution.
The interest rate spread is measured by the difference between the
average lending rates and average deposit rates. (14) 1 use the rolling
standard deviation of growth rate of the ratio of share price index to
GDP deflator with a five year rolling window as the measure of share
price volatility. The results in column 5 show that efficiency in the
banking system is effective in reducing income gaps. However, volatility
in the stock market, measured by variations in relative share prices,
has no statistically significant impact on income inequality (see column
6).
8. Summary and Conclusions
Since theories provide different predictions about the impact of
finance on income inequality, more empirical analysis is necessary to
shed light on their relationship. In this article, I examined the
determinants of the Gini coefficient in an autoregressive distributed
lag framework, paying particular attention to testing the effects of
financial development and financial sector reforms. Employing the ECM
cointegration test and the ARDL bounds technique, the empirical evidence
showed a significant steady-state relationship between the Gini
coefficient and its determinants. After documenting these basic
cointegration results, the long-run estimates were derived using the
underlying ARDL model.
The evidence suggests that underdevelopment of financial systems
hurts the poor more than the rich, resulting in higher income
inequality. Both the level and growth effects of financial development
are found to be significant. The results therefore highlight the
importance of developing financial systems in order to alleviate income
inequality. However, both domestic and international financial sector
reforms do not seem to reduce unequal access to finance, but rather tend
to aggravate income inequality in India. My results are not sensitive to
the measures of financial development and financial liberalization. In
addition, increased banking density and banking efficiency are found to
have a favorable effect on income inequality in India. Finally, there is
no evidence to support the presence of a non-linear effect in the
finance-inequality relationship, providing no support to the Greenwood
and Jovanovic (1990) hypothesis.
Although the finding that financial liberalization is leaving the
poor behind seems plausible in the context of India, the results do not
necessarily suggest that repressing the financial system is an effective
device for reducing inequality. One of the main objectives for priority
sector lending is to expand financial inclusion. While the evidence
suggests that this policy may have succeeded in ameliorating income
inequality to some extent, directed lending is generally associated with
misallocation of resources, which is costly for the financial system.
Moreover, in recent years, the definition of priority sectors has been
modified to include more sectors. To the extent that the poor tend to
have higher default rates, banks have increasingly preferred to provide
lending to established entrepreneurs who are deemed more creditworthy.
For a large developing country with a prevalence of widespread
poverty and a high concentration of poor in the rural sector, any
structural adjustment program that requires efficient allocation of
financial resources and price corrections is likely to harm a vast
majority of the poor who have been supported by subsidies of various
kinds in the past. Hence, the immediate undesirable effects on income
distribution associated with financial reforms are inevitable. However,
it remains to be seen how financial reforms may impact the poor in the
longer term. Probably some of these adverse consequences are only
transitory, and it will take a longer period of adjustment for the
benefits of reforms to trickle down to the poor. This suggests that the
relationship between financial liberalization and income distribution
may indeed be a non-linear one; although, currently available data do
not support such a conjecture. Thus, the results are not necessarily
contrary to the Washington consensus.
Achieving greater depth and breadth in the financial system is only
one element of an effective strategy to reach a more equitable income
distribution in India. It is also important to strengthen the regulatory
framework and improve financial supervision. Reforms that work in favor
of politically connected interests may undermine financial access. Thus,
future policy design should focus on expanding financial inclusion with
a view to broadening and improving access to finance for the poor.
References
Abiad, Abdul, and Ashoka Mody. 2005. Financial reform: What shakes
it? What shapes it? American Economic Review 95:66-88.
Aghion, Philippe, Philippe Bacchetta, and Abhijit V. Banerjee.
2004. Financial development and the instability of open economies.
Journal of Monetary Economics 51:1077-106.
Aghion, Philippe, and Patrick Bolton. 1997. A theory of
trickle-down growth and development. Review of Economic Studies
64:151-72.
Andrianova, Svetlana, Panicos Demetriades, and Anja Shortland.
2008. Government ownership of banks, institutions, and financial
development. Journal of Development Economics 85:218-52.
Ang, James B. 2008. What are the mechanisms linking financial
development and economic growth in Malaysia? Economic Modelling
38:38-53.
Ang, James B., and Warwick J. McKibbin. 2007. Financial
liberalization, financial sector development and growth: Evidence from
Malaysia. Journal of Development Economics 84:215-33.
Arestis, Philip, and Asena Caner. 2004. Financial liberalization
and poverty: Channels of influence. Levy Economics Institute, Economics
Working Paper No. 411.
Arestis, Philip, and Panicos O. Demetriades. 1997. Financial
development and economic growth: Assessing the evidence. Economic
Journal 107:783-99.
Arestis, Philip, Panicos O. Demetriades, and Kul B. Luintel. 2001.
Financial development and economic growth: The role of stock markets.
Journal of Money, Credit, and Banking 33:16-41.
Asian Development Bank. 2007. Key indicators. Asian Development
Bank, Manila: Philippines.
Banerjee, Abhijit, and Thomas Piketty. 2005. Top Indian incomes,
1922-2000. The Worm Bank Economic Review 19:1-20.
Banerjee, Abhijit V., and Andrew F. Newman. 1993. Occupational
choice and the process of development. Journal of Political Economy
101:274-98.
Banerjee, Anindya, Juan Dolado, and Ricardo Mestre. 1998.
Error-correction mechanism tests for cointegration in a single-equation
framework. Journal of Time Series Analysis 19:267-83.
Barro, Robert J. 2000. Inequality and frowth in a panel of
countries. Journal of Economic Growth 5:5-32.
Beck, Thorsten, Asli Demirguc-Kunt, and Ross Levine. 2007. Finance,
inequality and the poor. Journal of Economic Growth 12:27-49.
Beck, Thorsten, and Ross Levine. 2004. Stock markets, banks, and
growth: Panel evidence. Journal of Banking and Finance 28:423-42.
Beck, Thorsten, Ross Levine, and Alexey Levkov. 2008. Big bad
banks: The impact of U.S. branch deregulation on income distribution.
Brown University mimeo.
Bell, Clive, and Peter L. Rousseau. 2001. Post-independence India:
A case of finance-led industrialization? Journal of Development
Economies 65:153-75.
Burgess, Robin, and Rohini Pande. 2005. Dorural banks matter?
Evidence from the Indian social banking experiment. American Economic
Review 95:780-95.
Claessens, Stijn. 2006. Access to financial services: A review of
the issues and public policy objectives. Worm Bank Research Observer
21:91-122.
Claessens, Stijn, Simeon Djankov, Joseph Fan, and Larry Lang. 2002.
Disentangling the incentive and entrenchment effects of large
shareholdings. Journal of Finance 57:2741-71.
Claessens, Stijn, and Enrico Perotti. 2007. Finance and inequality:
Channels and evidence. Journal of Comparative Economics 35:748-73.
Clarke, George, Lixin Colin Xu, and Heng-fu Zou. 2006. Finance and
income inequality: What do the data tell us? Southern Economic Journal
72:578-96.
Cole, Shawn. 2009. Fixing market failures or fixing elections?
Elections, banks and agricultural lending in India. American Economic
Journal: Applied Economies 1(1):219-50.
Cutler, David M., and Lawrence Katz. 1991. Macroeconomic
performance and the disadvantaged. Brookings Papers on Economic Activity
2:1-61.
Das, Mitali, and Sanket Mohapatra. 2003. Income inequality: The
aftermath of stock market liberalization in emerging markets. Journal of
Empirical Finance 10:217-48.
Datt, Gaurav, and Martin Ravallion. 1998. Why have some Indian
states done better than others at reducing rural poverty? Eeonomica
65:17-38.
Deininger, Klaus, and Lyn Squire. 1996. A new data set measuring
income inequality. Worm Bank Economic Review 10:565-91.
Demetriades, Panicos O., and Khaled A. Hussein. 1996. Does
financial development cause economic growth? Time- series evidence from
sixteen countries. Journal of Development Economies 51:387-411.
Demetriades, Panicos O., and Kul B. Luintel. 1996. Financial
development, economic growth and banker sector controls: Evidence from
India. Economic Journal 106:359-74.
Demetriades, Panicos O., and Kul B. Luintel. 1997. The direct costs
of financial repression: Evidence from India. Review of Economics and
Statistics 79:311-20.
Demetriades, Panicos O., and Kul B. Luintel. 2001. Financial
restraints in the South Korean miracle. Journal of Development Economics
64:459-79.
Demirguc-Kunt, Asli, and Ross Levine. 2007. Finance and
opportunity: Financial system and intergenerational persistence of
relative income. Brown University mimeo.
Dollar, David, and Aart Kraay. 2002. Growth is good for the poor.
Journal of Economic Growth 7:195-225. Easterly, William, and Stanley
Fischer. 2001. Inflation and the poor. Journal of Money, Credit and
Banking 33:160-78.
Galor, Oded, and Joseph Zeira. 1993. Income distribution and
macroeconomics. Review of Economic Studies 60:35-52.
Giovannini, Alberto, and Martha De Melo. 1993. Government revenue
from financial repression. American Economic Review 83:953-63.
Gormley, Todd, and Radhakrishnan Gopalan. 2008. Stock market
liberalization and the decision to go public. Washington University
mimeo.
Greenwood, Jeremy, and Boyan Jovanovic. 1990. Financial
development, growth, and the distribution of income. Journal of
Political Economy 98:1076-107.
Honohan, Patrick. 2004a. Financial development, growth and poverty:
How close are the links? World Bank Policy Research Working Paper No.
3203.
Honohan, Patrick. 2004b. Financial sector policy and the poor:
Selected findings and issues. Washington, DC: The World Bank.
Honohan, Patrick, and Joseph E. Stiglitz. 2001. Robust financial
restraint. In Financial liberalization: How far, how fast?, edited by G.
Caprio, P. Honohan, and J. E. Stiglitz. Cambridge, UK: Cambridge
University Press, pp. 31-62.
Jha, Raghbendra. 2004. Reducing poverty and inequality in India:
Has liberalization helped? In Inequality growth and poverty in an era of
liberalization and globalization, edited by G. A. Cornia. Oxford: Oxford
University Press, pp. 297-326.
King, Robert G., and Ross Levine. 1993. Finance and growth:
Schumpeter might be right. Quarterly Journal of Economics 108:717-37.
Levine, Ross, Norman Loayza, and Thorsten Beck. 2000. Financial
intermediation and growth: Causality and causes. Journal of Monetary
Economics 46:31-77.
Levine, Ross, and Sara Zervos. 1998. Stock markets, banks, and
economic growth. American Economic Review 88:537-58.
McKinnon, Ronald I. 1973. Money and capital in economic
development. Washington, DC: Brookings Institution.
Mishkin, Frederic S. 2006. The next great globalization: How
disadvantaged nations can harness their financial systems to get rich.
Princeton and Oxford: Princeton University Press.
Mookherjee, Dilip, and Debraj Ray. 2003. Persistent inequality.
Review of Economic Studies 70:369-94.
Mookherjee, Dilip, and Debraj Ray. 2006. Occupational diversity and
endogenous inequality. Boston University, The Institute for Economic
Development Working Paper No. 142.
Pal, Parthapratim, and Jayati Ghosh. 2007. Inequality in India: A
survey of recent trends. United Nations, Department of Economics and
Social Affairs Working Paper No. 45.
Pesaran, M. Hashem, and Yongcheol Shin. 1998. An autoregressive
distributed-lag modelling approach to cointegration analysis. In
Econometrics and economic theory in the twentieth century: The
Ragnar-Frisch centennial symposium, edited by S. Strom. Cambridge, UK:
Cambridge University Press, pp. 371-413.
Pesaran, M. Hashem, Yongcheol Shin, and Richard J. Smith. 2001.
Bounds resting approaches to the analysis of level relationships.
Journal of Applied Econometrics 16:289-326.
Pritchett, Lant, Asep Suryahadi, Sudarno Sumarto, and Yusuf
Suharso. 2000. The evolution of poverty during the crisis in Indonesia,
1996-1999. The World Bank Policy Research Working Paper No. 2435.
Rajan, Raghuram G. 2006. Separate and unequal. Finance and
Development 43:56-7.
Rajan, Raghuram G., and Luigi Zingales. 2003a. The great reversals:
The politics of financial development in the twentieth century. Journal
of Financial Economics 69:5-50.
Rajan, Raghuram G., and Luigi Zingales. 2003b. Saving capitalism
from the capitalists: Unleashing the powerof financial markets to create
wealth and spread opportunity. New York: Crown Business.
Ramey, G., and V. A. Ramey. 1995. Cross-country evidence on the
link between volatility and growth. American Economic Review 85:1138-51.
Roland, Gerard. 2002. The political economy of transition. Journal
of Economic Perspectives 16:29-50.
Rousseau, Peter L. 1998. The permanent effects of innovation on
financial depth: Theory and U.S. historical evidence: From unobservable
components models. Journal of Monetary Economics 42:387-425.
Shaw, Edward S. 1973. Financial deepening in economic development.
New York: Oxford University Press.
Suri, K. C. 2004. Democracy, economic reforms and election results
in India. Economic and Political Weekly 39:5404-11.
United Nations Development Program. 2007. Human development report.
New York: United Nations.
Varshney, Ashutosh. 2007. India's democratic challenge.
Foreign Affairs 86:93-106.
James B. Ang, Department of Economics, Monash University, 900
Dandenong Road, Caulfield East, Vic 3145, Australia; E-mail
james.ang@buseco.monash.edu.au.
Constructive suggestions from two anonymous referees of this
journal are much appreciated.
Received April 2008; accepted September 2008.
(1) Although financial crises may also have significant effects on
income inequality, as demonstrated by Pritchett et al. (2000) in the
case of Indonesia during the Asian financial crisis of 1997-1998, this
issue is not addressed because the economy of India was largely
unaffected by the 1997-1998 crisis.
(2) In a study that examines the influence of institutional quality
on government ownership of banks, Andrianova, Demetriades, and Shortland
(2008) report that India has the highest ratio of state-owned or
state-controlled bank assets to total commercial bank assets in a sample
of 83 countries.
(3) The Gini coefficient is the ratio of the area between the
Lorenz curve (which plots share of population against income share
received) to the area below the diagonal. The value ranges from 0 to 1,
where 0 means perfect income equality and 1 implies perfect income
inequality. I use the Gini data from Deininger and Squire (1996) and
Dollar and Kraay (2002). These data are updated with more recent data
points available from ADB (2007) and UNDP (2007).
(4) I use the growth rate of per capita real GDP as a control
variable rather than the level of per capita real GDP since the
influence of finance on income inequality operate through growth
(Demirguc-Kunt and Levine 2007). Moreover, the use of a level income
variable in the specification may lead to some eonometric problems
because per capita real GDP is highly correlated with the measures of
financial development.
(5) Ideally, one should use a de jure (rather than a de facto)
measure reflecting the strength of directed credit policies designed to
repress the financial system. Unfortunately, such information is not
available on a consistent and reliable basis for India.
(6) Demetriades and Luintel (1996, 1997) set the measure to 0 when
the directed credit programs are not implemented, and to 1, 2, and 3
when the programs cover up to 20%, 21-40%, and over 40%, respectively,
of total bank lending. However, I use the share of directed credit in
total credit in order to allow for more variation in the series,
particularly in the 1950s and 1960s, when the ratio was always below
20%.
(7) For the first and second measures of financial liberalization,
I extract six principal components, which are able to account for 97%
and 95% of the total variation in the policy variables, respectively.
These components are then summarized into just one composite measure
using eigenvalues as the weights. I have also tried using just one and
all principal components. However, the results remain insensitive to the
number of principal components extracted due to their high correlation
structure.
(8) I have also explored using the principal component analysis.
Correlation analysis shows that this simple arithmetic mean is
significantly and highly correlated with the first principal component
and weighted average of all six principal components, suggesting that
the results would not vary significantly with the use of any of these
measures.
(9) This may not necessarily imply that one has to pay the costs
associated with financial liberalization in order to deepen the
financial system. A more in-depth analysis is required to draw such a
policy conclusion. For instance, regressing financial development on
financial liberalization and other conditioning variables such as per
capita real GDP and real interest rate may shed some light on this
issue. This additional analysis is not performed here because my focus
is on the finance-inequality relationship.
(10) In estimating the Gini coefficient equation, my focus has been
on the long run, that is, how the variables are related in the steady
state. Results on the short-run dynamics are not reported. My
specification with only one lag shows that the short-run effects of both
financial development and financial liberalization are highly
significant with signs consistent with their long-run counterparts.
Allowing for two lags in the estimation does not vary the results in any
significant way; although, the coefficients of two-period lagged changes
on financial development and financial liberalization become
statistically insignificant.
(11) I am grateful to a referee for recommending this approach.
(12) I have also considered using the lagged values of financial
liberalization. However, this does not improve the results considerably.
(13) Although many farmers in developing countries obtain credit
through microfinance, this issue is not formally addressed in this
article due to the lack of reliable time series data. Moreover, the
microfinance finance penetration rate, defined as borrowing clients as a
percentage of population, stands at only 1.1% for India. This ratio is
relatively lower than many other developing countries such as
Bangladesh, Indonesia, and Thailand, where micro finance claims 13.1%,
6.7%, and 6.5% of the population as clients, respectively (see Honohan
2004b).
(14) Honohan (2004a) proposes that the interest rate spread is a
better measure that reflects the efficiency of the financial system,
compared to other rudimentary measures of financial development commonly
used in the literature. This measure has also been used by Rousseau
(1998) as a measure of technical progress in the banking sector.
Table 1 Correlations between Financial Development and
Financial Liberalization
PCY M2Y
PCY 1.000
M2Y 0.980 (0.000) *** 1.000
COM 0.925 (0.000) *** 0.909 (0.000) ***
BANK 0.971 (0.000) *** 0.962 (0.000) ***
FL-DL -0.904 (0.000) *** -0.851 (0.000) ***
FL- [DL.sub.E] -0.890 (0.000) *** -0.828 (0.000) ***
FL-AB -0.511 (0.000) *** -0.402 (0.003) ***
COM BANK
PCY
M2Y
COM 1.000
BANK 0.903 (0.000) *** 1.000
FL-DL -0.831 (0.000) *** -0.946 (0.000) ***
FL- [DL.sub.E] -0.817 (0.000) *** -0.932 (0.000) ***
FL-AB -0.382 (0.004) *** -0.567 (0.000) ***
FL-DL FL-[DL.sub.E] FL-AB
PCY
M2Y
COM
BANK
FL-DL 1.000
FL- [DL.sub.E] 0.998 (0.000) *** 1.000
FL-AB 0.708 (0.000) *** 0.743 (0.000) *** 1.000
PCY = ln(private credit/GDP); M2Y = ln[(M3 - M1)/GDP]; COM =
ln[commercial bank assets/(central bank assets + commercial
bank assets)]; BANK = ln(number of bank offices per
population); FL-DL = ln(financial liberalization index -
Demetriades and Luintel approach); FL-[DL.sub.E] = ln(financial
liberalization index extended Demetriades and Luintel
approach); and FL-AB = ln(financial liberalization index *** -
Abiad and Mody approach). Figures in parentheses are p-values.
indicates 1% level of significance.
Table 2 Income Inequality and Financial Development
(Dep. = Ln Gini Coefficient)
(1) (2)
I. Cointegration tests
ARDL bounds test
(Pesaran et
a1. 2001) 4.817 ** 4.549 **
ECM t-test (Banerjee
et al. 1998 -4.522 ** -4.341 **
II. ARDL estimate
Intercept 3.659 (0.000) *** 3.743 (0.000) ***
Per capita real
GDP growth -0.331 (0.094) * -0.529 (0.082) *
Ln Trade openness 0.142 (0.000) *** 0.153 (0.000) ***
Inflation rate -0.351 (0.008) *** -0.495 (0.009) ***
Ln (Private
credit / GDP) -0.066 (0.000) ***
Ln [(M3 - M1)/GDP] -0.037 (0.021) **
Ln [Comm. bank
assets/ (central
bank assets +
comm. bank
assets)]
Ln Banking density
Ln (Stock market
capitalization
/ GDP)
Ln (Modern sector
/ GDP)
III.Diagnostic checks
LM test for serial
correlation 1.076 (0.301) 2.821 (0.093) *
Ramsey's RESET test 0.327 (0.568) 0.079 (0.778)
Heteroskedasticity
test 0.991 (0.319) 0.345 (0.557)
(3) (4)
I. Cointegration tests
ARDL bounds test
(Pesaran et
a1. 2001) 4.768 ** 4.218 **
ECM t-test (Banerjee
et al. 1998 -4.371 ** -4.087 **
II. ARDL estimate
Intercept 3.729 (0.000) *** 3.388 (0.000) ***
Per capita real
GDP growth -0.535 (0.063) * -0.479
Ln Trade openness 0.143 (0.000) *** 0.135 (0.000) ***
Inflation rate -0.545 (0.002) *** -0.470 (0.011) **
Ln (Private
credit / GDP)
Ln [(M3 - M1)/GDP]
Ln [Comm. bank
assets/ (central
bank assets +
comm. bank
assets)] -0.123 (0.009) ***
Ln Banking density -0.037 (0.000) ***
Ln (Stock market
capitalization
/ GDP)
Ln (Modern sector
/ GDP)
III.Diagnostic checks
LM test for serial
correlation 1.843 (0.175) 1.772 (0.183)
Ramsey's RESET test 0.185 (0.667) 0.001 (0.974)
Heteroskedasticity
test 0.686 (0.407) 0.386 (0.534)
(5) (6)
I. Cointegration tests
ARDL bounds test
(Pesaran et
a1. 2001) 3.626 * 4.659 **
ECM t-test (Banerjee
et al. 1998 -3.954 * -4.491 **
II. ARDL estimate
Intercept 3.709 (0.000) *** 3.673 (0.000) ***
Per capita real
GDP growth -0.771 (0.028) ** -0.412 (0.044) **
Ln Trade openness 0.061 (0.125) 0.149 (0.000) ***
Inflation rate -0.675 (0.001) *** -0.375 (0.007) ***
Ln (Private
credit / GDP)
Ln [(M3 - M1)/GDP]
Ln [Comm. bank
assets/ (central
bank assets +
comm. bank
assets)]
Ln Banking density
Ln (Stock market
capitalization
/ GDP) 0.019 (0.195)
Ln (Modern sector
/ GDP) -0.267 (0.000) ***
III.Diagnostic checks
LM test for serial
correlation 3.043 (0.081) * 1.034 (0.309)
Ramsey's RESET test 0.001 (0.991) 0.252 (0.615)
Heteroskedasticity
test 0.994 (0.319) 0.411 (0.521)
The optimal lag structure for the resulting ARDL model was
chosen using SBC. The test statistics of the bounds tests are
compared against the critical values reported in Pesaran et
al. (2001). The estimation allows for an unrestricted
intercept and no trend. The 10%, 5%, and 1% critical value
bounds for the F-test are (2.45, 3.52), (2.86, 4.01), and
(3.74, 5.06), respectively. The 10%, 5%, and 1% critical value
bounds for the t-test are (-2.57, -3.66), (-2.86, -3.99), and
(-3.43, -4.60), respectively. Numbers in parentheses indicate
p-values. *, **, and *** indicate 10%, 5%, and 1% levels of
significance, respectively.
Table 3. Income Inequality and Financial Sector Policies
(Dep. = Ln Gini Coefficient)
(l) (2)
I. Cointegration tests
ARDL bounds test
(Pesaran et al.
2001) 4.024 ** 3.908 *
ECM t-test (Banerjee
et al. 1998 -4.033 ** -4.193 **
II. ARDL estimate
Intercept 3.797 (0.000) *** 3.803 (0.000) ***
Per capita real
GDP growth -0.527 (0.081) * -0.512 (0.098) *
Ln Trade openness 0.123 (0.000) *** 0.106 (0.000) ***
Inflation rate -0.502 (0.008) *** -0.465 (0.020) **
Directed credit
programs -0.249 (0.000) ***
Reserve and
liquidity
requirements -0.229 (0.028) **
Interest rate
restraint
Ln Financial
liberalization
index
(Demetriades-
Luintel)
Ln Financial
liberalization
index (Extended
Demetriades-
Luintel)
Ln Financial
liberalization
index (Abiad-Mody)
III.Diagnostic checks
LM test for serial
correlation 2.125 (0.145) 1.077 (0.299)
Ramsey's RESET test 0.041 (0.841) 0.256 (0.613)
Heteroskedasticity
test 0.756 (0.384) 0.506 (0.477)
(3) (4)
I. Cointegration tests
ARDL bounds test
(Pesaran et al.
2001) 3.857 * 4.661 **
ECM t-test (Banerjee
et al. 1998 -3.746 * -4.467 **
II. ARDL estimate
Intercept 3.709 (0.000) *** 3.759 (0.000) ***
Per capita real
GDP growth -0.276 -0.294
Ln Trade openness 0.083 (0.000) *** 0.109 (0.000) ***
Inflation rate -0.353 (0.009) *** -0.338 (0.015) **
Directed credit
programs
Reserve and
liquidity
requirements
Interest rate
restraint -0.035 (0.000) ***
Ln Financial
liberalization 0.074 (0.000) ***
index
(Demetriades-
Luintel)
Ln Financial
liberalization
index (Extended
Demetriades-
Luintel)
Ln Financial
liberalization
index (Abiad-Mody)
III.Diagnostic checks
LM test for serial
correlation 1.578 (0.209) 1.538 (0.215)
Ramsey's RESET test 0.264 (0.607) 0.051 (0.821)
Heteroskedasticity
test 1.875 (0.171) 1.848 (0.174)
(5) (6)
I. Cointegration tests
ARDL bounds test
(Pesaran et al.
2001) 4.660 ** 3.715 *
ECM t-test (Banerjee
et al. 1998 -4.467 ** -4.011 **
II. ARDL estimate
Intercept 3.751 (0.000) *** 3.621 (0.000) ***
Per capita real
GDP growth -0.305 (0.134) -0.574 (0.072) *
Ln Trade openness 0.105 (0.000) *** 0.081 (0.001) ***
Inflation rate -0.348 (0.012) ** -0.556 (0.004) ***
Directed credit
programs
Reserve and
liquidity
requirements
Interest rate
restraint
Ln Financial
liberalization
index
(Demetriades-
Luintel)
Ln Financial
liberalization
index (Extended
Demetriades-
Luintel) 0.077 (0.000) ***
Ln Financial
liberalization
index (Abiad-Mody) 0.019 (0.041) **
III.Diagnostic checks
LM test for serial
correlation 2.248 (0.134) 2.092 (0.148)
Ramsey's RESET test 0.289 (0.591) 0.036 (0.851)
Heteroskedasticity
test 2.871 (0.091) * 1.371 (0.242)
Table 4. Income Inequality, Financial Development
and Financial Liberalization
(1) (2)
I. Cointegration tests
ARDL bounds test
(Pesaran et
al. 2001 3.373 * 2.937
ECM t-test (Banerjee
et al. 1998) -4.262 ** -3.984 *
II. ARDL estimate
Intercept 3.656 (0.000) *** 3.682 (0.000) ***
Per capita real
GDP growth -0.329 (0.099) * -0.503 (0.093) *
Ln Trade openness 0.141 (0.000) *** 0.129 (0.004) ***
Inflation rate -0.352 (0.009) *** -0.479 (0.011) **
Ln (Private credit
/ GDP) -0.065 (0.004) ***
Ln [(M3 -M1) / GDP] -0.025 (0.233)
Ln [Comet. bank
assets / (central
bank assets +
comm. bank
assets)]
Ln [1/2 (private
credit/GDP + BD)]
Ln [1/2 (M3 - M1/
GDP + BD)]
Ln [1/2 (ComBa/ComBA
+ CenBA + BD)]
Ln Financial
liberalization
index (Abiad-Mody) 0.001 (0.932) 0.009 (0.419)
III. Diagnostic checks
LM test for serial
correlation 0.077 (0.781) 0.147 (0.701)
Ramsey's RESET test 6.409 (0.011) ** 5.871 (0.015) **
Heteroskedasticity
test 1.534 (0.215) 3.876 (0.049) **
(3) (4)
I. Cointegration tests
ARDL bounds test
(Pesaran et
al. 2001 3.274 3.429 *
ECM t-test (Banerjee
et al. 1998) -3.981 * -4.399 **
II. ARDL estimate
Intercept 3.601 (0.000) *** 3.929 (0.000) ***
Per capita real
GDP growth -0.298 (0.136) -0.281 (0.171)
Ln Trade openness 0.089 (0.004) *** 0.143 (0.000) ***
Inflation rate -0.524 (0.000) *** -0.362 (0.008) ***
Ln (Private credit
/ GDP)
Ln [(M3 -M1) / GDP]
Ln [Comet. bank
assets / (central
bank assets +
comm. bank
assets)] -0.041 (0.433)
Ln [1/2 (private
credit/GDP + BD)] -0.025 (0.011) **
Ln [1/2 (M3 - M1/
GDP + BD)]
Ln [1/2 (ComBa/ComBA
+ CenBA + BD)]
Ln Financial
liberalization
index (Abiad-Mody) 0.019 (0.040) ** 0.005 (0.554)
III. Diagnostic checks
LM test for serial
correlation 1.088 (0.297) 0.201 (0.609)
Ramsey's RESET test 4.076 (0.044) ** 6.756 (0.009) ***
Heteroskedasticity
test 0.751 (0.386) 1.385 (0.239)
(5) (6)
I. Cointegration tests
ARDL bounds test
(Pesaran et
al. 2001 2.817 2.925
ECM t-test (Banerjee
et al. 1998) -3.833 -3.780
II. ARDL estimate
Intercept 3.909 (0.000) *** 4.029 (0.000) ***
Per capita real
GDP growth -0.491 (0.099) * -0.307 (0.144)
Ln Trade openness 0.131 (0.004) *** 0.128 (0.000) ***
Inflation rate -0.473 (0.011) ** -0.398 (0.004) ***
Ln (Private credit
/ GDP)
Ln [(M3 -M1) / GDP]
Ln [Comet. bank
assets / (central
bank assets +
comm. bank
assets)]
Ln [1/2 (private
credit/GDP + BD)]
Ln [1/2 (M3 - M1/
GDP + BD)] -0.028 (0.215)
Ln [1/2 (ComBa/ComBA
+ CenBA + BD)] -0.053 (0.041) **
Ln Financial
liberalization
index (Abiad-Mody) 0.007 (0.551) 0.003 (0.801)
III. Diagnostic checks
LM test for serial
correlation 0.026 (0.872) 1.138 (0.286)
Ramsey's RESET test 6.407 (0.011) ** 6.709 (0.011) **
Heteroskedasticity
test 2.972 (0.085) * 1.971 (0.161)
The dependent variable is Ln Gini coefficient. "BD" is banking
density, "Com BA" is commercial bank assets, and "CenBA" refers
to central bank assets. The regressions involve six variables
and therefore a different set of critical values applies. The
relevant 10%, 5%, and I% critical value bounds for the F-test
are (2.26, 3.35), (2.62, 3.79), and (3.41, 4.68), respectively.
The 10%, 5%, and 1% critical value bounds for the t-test are
(-2.57, -3.86), (-2.86, -4.19), and (-3.43, -4.79),
respectively.
Table 5 Income Inequality, Financial Growth, Nonlinearity,
Efficiency, and Volatility
(1) (2)
I. Cointegration tests
ARDL bounds test
(Pesaran et
al. 2001) 3.432 3.762 *
ECM t-test (Banerjee
et al. 1998 -3.869 * -3.903 *
II. ARDL estimate
Intercept 3.815 (0.000) *** 3.701 (0.000) ***
Per capita real
GDP growth -1.281 (0.007) *** -0.651 (0.089) *
Ln Trade openness 0.116 (0.000) *** 0.089 (0.006) ***
Inflation rate -1.051 (0.002) *** -0.562 (0.023) **
Growth rate of
private credit
/ GDP -0.485 (0.034) **
Growth rate of
financial
liberalization
index 0.188 (0.329)
Ln (Private credit
/ GDP)
Squared Ln (Private
credit / GDP)
Ln Financial
liberalization
index
Squared Ln
(Financial
liberalization
index)
Interest spread
Ln Share price
volatility
III. Diagnostic checks
LM test for serial
correlation 2.901 (0.089) * 4.252 (0.039) **
Ramsey's RESET test 0.002 (0.961) 0.239 (0.625)
Heteroskedasticity
test 1.452 (0.228) 0.789 (0.374)
(3) (4)
I. Cointegration tests
ARDL bounds test
(Pesaran et
al. 2001) 3.975 ** 3.459 *
ECM t-test (Banerjee
et al. 1998 -4.438 ** -4.154 *
II. ARDL estimate
Intercept 3.781 (0.000) *** 3.753 (0.000) ***
Per capita real
GDP growth -0.344 (0.096) * -0.291 (0.170)
Ln Trade openness 0.121 (0.000) *** 0.109 (0.000) ***
Inflation rate -0.319 (0.037) ** -0.266 (0.103)
Growth rate of
private credit
/ GDP
Growth rate of
financial
liberalization
index
Ln (Private credit
/ GDP) 0.118 (0.488)
Squared Ln (Private
credit / GDP) 0.047 (0.277)
Ln Financial
liberalization
index 0.086 (0.096) *
Squared Ln
(Financial
liberalization
index) 0.015 (0.753)
Interest spread
Ln Share price
volatility
III. Diagnostic checks
LM test for serial
correlation 0.789 (0.374) 1.189 (0.275)
Ramsey's RESET test 0.016 (0.898) 0.040 (0.841)
Heteroskedasticity
test 1.287 (0.257) 2.335 (0.127)
(5) (6)
I. Cointegration tests
ARDL bounds test
(Pesaran et
al. 2001) 4.698 ** 3.969 *
ECM t-test (Banerjee
et al. 1998 -4.272 ** -4.112 **
II. ARDL estimate
Intercept 3.765 (0.000) *** 3.640 (0.000) ***
Per capita real
GDP growth -0.601 (0.052) * -0.793 (0.041) **
Ln Trade openness 0.112 (0.000) *** 0.084 (0.018) **
Inflation rate -0.529 (0.006) *** -0.633 (0.016) **
Growth rate of
private credit
/ GDP
Growth rate of
financial
liberalization
index
Ln (Private credit
/ GDP)
Squared Ln (Private
credit / GDP)
Ln Financial
liberalization
index
Squared Ln
(Financial
liberalization
index)
Interest spread -0.694 (0.000)***
Ln Share price
volatility 0.020 (0.337)
III. Diagnostic checks
LM test for serial
correlation 2.778 (0.096)* 4.775 (0.029) **
Ramsey's RESET test 0.549 (0.459) 0.067 (0.795)
Heteroskedasticity
test 1.578 (0.209) 1.175 (0.278)
The dependent variable is Ln Gini coefficient. *, **, and
*** indicate 10%, 5%, and 1% levels of significance,
respectively.
