Contribution of Rural Banks To Regional Economic Development: Evidence From The Philippines
Contribution of Rural Banks To Regional Economic Development: Evidence From The Philippines
Contribution of Rural Banks To Regional Economic Development: Evidence From The Philippines
Université de Limoges, LAPE, 5 rue Félix Eboué, BP 3127, 87 031 Limoges Cedex, France
Abstract:
This paper examines the link between banking and economic development at the regional
level in the Philippines and focuses on the role played by rural banks on economic activity. A
cointegration panel data analysis is applied for the sixteen Philippine regions from 1993 to
2005. When using indicators built at the regional banking industry level, there is no clear-cut
the presence of rural banks on regional economic development for the intermediate and less
developed regions, with a stronger effect for the former, suggesting a threshold effect.
1
This paper was prepared for the ASIA-LINK human resource development project ASIALINK/B7-
301/2005/105-139: Euro-Philippines Network on Banking and Finance, Safety and Soundness of the Financial
System, coordinated by the University of Limoges (http://asialink.philippines.googlepages.com). ASIA-LINK is
a program of the European Commission that seeks to promote regional and multilateral networking among
higher education institutions in Europe and developing economies in Asia.
We would like to gratefully acknowledge the very useful help of P. Rous on panel data econometrics. We would
also like to thank I. Hasan, J. Los Banos, F. Moshirian, P. Wachtel and L. White for their helpful comments on
preliminary versions of this paper. All remaining errors are ours.
During the last thirty years, the extent to which a better-developed financial system
fosters economic development has been the subject of extensive research. The emergence of
the endogenous growth theory shed a new light on the link between financial and economic
development. Levine, 2005, identifies five broad functions provided by the financial sector
that reduce information, enforcement and transaction costs: (i) production of information on
investment projects and capital allocation; (ii) monitoring and effective corporate governance;
(iii) trading, diversification and management of risk; (iv) saving mobilization and (v) easing
the exchange of goods and services. The way these five functions are supplied by the financial
system influences saving rates, investment decisions, technological innovation and hence
economic activity. Since King and Levine, 1993a, 1993b, a large number of empirical studies
have analyzed the finance-growth nexus for developed as well as developing countries (see
Wachtel, 2003, and Demirgüç-Kunt and Levine, 2008, for comprehensive surveys). While
empirical studies used different methodologies to explore the finance-growth nexus, they find
overall consistent results on the sign of the relationship. Countries with better-developed
financial system tend to grow faster. A contentious area of research investigates the causality
of this relationship (King and Levine, 1994; Demetriades and Hussein, 1996; Wachtel and
Rousseau, 1995). Some researchers assert that it is financial development that fuels growth
(King and Levine, 1993a, 1993b; Christopoulos and Tsionas, 2004; Demirgüç-Kunt and
Levine, 2008) while others, following Robinson, 1952, find that improvements in productivity
and economic output would require increased investment and funding (Jung, 1986; Ireland,
1994). Other studies claim that this causality is actually bi-directional (Demetriades and
Hussein, 1996).
The mechanisms through which financial and economic development are linked
remain also an open question. Berger et al., 2004, and Demirgüç-Kunt and Levine, 2008,
mechanism is the reduction of financial constraints for firms that heavily rely on external
finance. Following Rajan and Zingales, 1998, and Demirguc_Kunt and Maksimovic, 1998,
firms or industries that strongly rely on external finance tend to grow faster. The extent to
which institutional characteristics of the banking system could influence access to finance is
also a key issue. More precisely, some authors argue that small, regional and locally-owned
banks could behave very differently from large, national and non locally-owned banks for a
relationship lending segment of the market and a greater commitment to local prosperity
could enable them to better monitor and assess risk of local firms. The presence of these
banks could then have a specific influence on local development by improving financing
opportunities to small and medium size enterprises (Rodriguez-Fuentes, 1998; Collender and
Shaffer, 2003; Carbo Valverde and Fernandez, 2004; Burgess and Pande, 2005; Hakenes et al.,
2009) 3 . This question is of particular interest in countries that have undergone a process of
sharp reduction of the number and the market share of local banks during the last two decades.
In such a context, concerns have been raised that local banks may not be able to compete with
national-wide banks and then to offer specific banking services to local communities in the
future (Avery and Samolyck, 2004). In developing countries where economic development is
hampered by insufficient and inadequate access to financial services in rural areas, this
2
They mention, among others, the size, efficiency and regulation of the banking system or the laws and
regulations that shape the operation of the financial system.
3
Berger et al., 2004, in a cross-country study also highlights such results.
3
This paper aims to extend the existing literature by conducting a market level analysis
in order to assess the impact of the market share of local banks on regional economic
development. Our goal is to analyze the relationship between banking and economic
development in the Philippines by taking into account the weight of the banking industry
We study the case of the sixteen regions of the Philippines. Quite significant
disparities in the level of economic and banking development across the sixteen regions raise
interesting issues on the finance-growth nexus and the specific influence of local banks.
Moreover, as pointed out by Carbo et al., 2007, focusing on a single country enables us to
assume that macroeconomic framework and political governance (monetary and exchange
rate policies, banking regulation, education and health policies, industrial policy …) are
of the dominance of banks in the country as evidenced by the limited presence of equity
markets as source of finance (Gochoco-Bautista, 1999), and the fact that only the largest
corporations are listed in the country’s stock exchange. Hence funding for the majority of
businesses in the country is expected to be sourced primarily from banks and not through
financial markets (Gochoco-Bautista, 1999; Asian Development Bank, 2007). The formal
banking system is composed of three categories of banks: universal and commercial banks,
thrift and private development banks, and regional rural and cooperative banks 4 . Although the
formal banking system is dominated by commercial banks, rural banks in the Philippines were
primarily established to promote and expand the rural economy. They generally cater to small
borrowers including farmers, entrepreneurs, market vendors, business owners, wage earners,
teachers and cooperatives. From the 1960s to the 1980s, rural banks served as conduits of
4
Hereafter, we used the term rural banks for rural and cooperative banks.
4
subsidized loan funds from the government and international donors and were plagued by
high default rates, insolvent lending programs, and high operating costs to name a few
(Agabin and Daly, 1996). Following the process of financial liberalization that occurs in the
Philippine during the last two decades, the government shifts toward a more market-orientated
approach credit policy for rural areas 5 . Recent government policies have led to strengthen the
place of rural banks by enhancing their role in financing micro-entrepreneurs and poor
an empirical investigation over the period 1993-2005 using an original set of regional banking
data.
The sixteen Philippine regions are ranked in three groups depending on average
developed regions). Rank-order correlation tests provide us some first interesting results.
Whereas a negative and significant correlation between economic development and rural bank
presence is obtained while considering all the sixteen regions together, an opposite result is
Building on the works of Christopoulos and Tsionas, 2004, and Apergis et al., 2007,
which take into account the integration properties of the data, a panel cointegration analysis is
conducted. Our econometric specification enables, on the one hand, to address the
heterogeneity of economic development and banking coverage of the regions and, on the
other hand, to efficiently utilize the limited regional data available presently as annual
banking regional data do not exist prior 1993. This specification also provides some insights
on the causality between economic and banking development. If the estimations do not show
the existence of a strong relationship between regional banking and economic development in
5
For example, removing of interest rate restrictions or easing of new banks and branches opening.
5
the Philippines, the findings highlight a positive effect of the presence of rural banks on
economic development.
regional characteristics of the Philippines. Section 3 presents our research design and results.
Philippines
This paper uses an original dataset made of regional banking data in order to analyze
the banking system at the regional level, underlying the role of rural banks. The
macroeconomic regional data are from the Philippine National Statistics Office and National
Statistical Coordination Board. Bank regional data comes from the Central Bank of the
Philippines (Bangko Sentral ng Pilipinas). The period of our study is from 1993 to 2005. The
dataset could not start prior to 1993 as the organization of the regions in the Philippines was
different 6 .
An originality of this paper is that we used regional level banking data for the three
types of banks (commercial banks, thrift banks and rural banks). The Central Bank aggregates
data per bank branch office to a regional level. For thrift and rural banks which operate
mainly at a regional level, this information is publicly available. However regional data for
commercial banks, which operate at a national level, are not publicly available and are
6
The sample includes the Asian crisis but we did not exclude it as we aim to study a long term relationship
between banking and economic development.
6
2.2 Regional economic development
The Philippines is divided into seventeen geographic regions. For this study however,
we refer to only sixteen regions, having integrated Region 4-A, Calabarzon and Region 4-B,
Mimaropa (Region 4 was divided into two separate jurisdictions only in 2002). The per capita
real gross regional domestic product (PC_RGRDP) is used as a measure of the regional
economic structure and ranking of the regions depending on its variable has remained
relatively constant over the period covered by this study. In view of the heterogeneity of the
stages of economic development, we classify the regions into three groups: less-economically
developed, intermediate developed and developed regions. Table 1 presents the real per capita
gross regional product of the regions. The National Capital Region (NCR) is the most
economically developed region and the Autonomous Region in Muslim Mindanao (ARMM)
has the lowest per capita regional GDP among the regions in the country.
Based on simple statistical analyses of the above data, we identify the less-
economically developed regions to be the following: Ilocos, Cagayan Valley, Bicol, Eastern
(ARMM) and Caraga. These regions are basically agriculture intensive with lower levels of
industrialization. Their regional contribution to the Philippine GDP as of 2005 is below 2.9%
The developed regions, NCR, Cordillera Administrative Region (CAR), and Northern
Mindanao are those with a strong service sector coupled with a vibrant industrial sector
view of the presence of the province of Benguet in the region, which is highly developed and
which greatly improves the ranking of the region despite the significantly poorer economic
performance of the other provinces in the region. Central Visayas (with Cebu province) and
7
Davao (with Davao del Sur province) regions, despite being more highly urbanized than
Northern Mindanao and the CAR, were not classified in this group in view of the lower
were not classified as developed or as less developed and include Central Luzon, South Luzon,
[Insert Table 1]
Table 2 presents descriptive statistics for some banking indicators at the regional level.
Two measures of the regional banking activity are provided: total deposits and total net loans.
To measure banking development (BD), four different measures are used: three measures of
financial depth (the share of total net loans over nominal regional gross domestic product
(Loans), the share of total deposits over regional gross domestic product (Deposits), the
number of banking offices per capita (Banking office density)) and one measure of local
intermediation (total net loans over total deposits (Intermediation)). Recent studies (Berger et
al., 2004; Hasan et al., 2007) suggest using quality-based indicator instead of quantity-based
construct such measures are available only for few commercial banks but not for rural banks
in the Philippines. Finally, two measures of rural banks presence are computed: the share of
net loans granted by rural banks per region over total net loans granted per region (RB Loan
share), and the share of total resources of rural banks per region over total resources for all
banks per region (RB Resource share). To measure the impact of rural banks on the economic
development we will focus on loans variables as the purpose of the rural financial market as
8
defined by policy reforms in the late 1980s is to provide credit access to small borrowers
(Llanto, 2005). We will consider the four following groups of regions: “All regions”,
“Developed regions”, “Intermediate regions” and “Less developed regions”. But, given the
macroeconomic specificities of the NCR region, we will also study the group “All regions
First, considering either the group “All regions” or the group “All regions except
NCR” allows us to show the predominance of the NCR region in terms of banking
development.
Second, the three different groups of regions are characterized by a great heterogeneity
of banking development. As an example, the mean value of the share of total net loans over
nominal regional gross domestic product (Loans) ranges from 0.69 to 0.10 when considering
respectively the group “Developed regions” and the group “Less developed regions”.
Whatever the measure used (intermediation, deposits and banking office density), we still find
heterogeneity through Philippine regions and they show that the wealthiest regions have
The third result is related to the presence of rural banks. Whatever the measure used
(RB Loan share, RB Resource share or RB office density), the presence of rural banks is
higher on average in the less developed regions than in the intermediate developed regions,
itself higher than in the developed regions. 21.22 % of the total average amount of loans are
granted by rural banks in the less developed regions against 13.46 % in the intermediate
[Insert Table 2]
9
To analyze more precisely rural bank presence, Table 3 and Table 4 provide
information respectively on the market share of the different types of banks (commercial
banks, thrift banks and rural banks) at the national level and of the rural bank market share at
The formal banking sector is dominated by commercial banks 7 , which over the 1993-
2005 period represent 56.8% of the total number of bank offices in the Philippines. The thrift
banks represent 17.8% of the total number of bank offices and the remaining 25.37% of the
total banking offices operating in the country are regional rural and cooperative banks.
[Insert Table 3]
Commercial banks remain the major source of funding with an average credit market
share of 89% and 73% when considering respectively the group “All regions” and the group
“All regions except NCR”. However, at the national level, rural bank, on average, account for
37.60% of the total number of banking office and granted 14.44% of the total amount of loans,
when excluding the NCR over the 1993-2005 period. Moreover, since 1998, Figure 1 shows a
decline of the loan market share of commercial banks (from 77% to 65%) and thrift banks
(12.60% to 11%) and, at the same time, an increase of the loan market share of the rural banks
7
In this paper, we do not aim to study the semi-formal and informal financial sectors. For a presentation of the
financial system in the Philippines, see Dauner Gardiol, Helms and Deshpande, 2005. For a detailed study of
rural finance, see Llanto, 2005.
10
Figure 1. Loan market shares of commercial, thrift and rural banks
in the Philippines* (1993-2005)
90,00
80,00
70,00
60,00
50,00
40,00
30,00
20,00
10,00
0,00
1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
* Loan market shares are computed for the group “All regions except NCR”.
Source: Bangko Sentral ng Pilipinas
Table 4 provides information on the evolution of RB market shares between 1993 and 2005,
and shows heterogeneity across regions. We can first notice that whatever the region
considered, RB increase their market share with regard to their resources and the loans they
grant. The analysis of the evolution of RB office density is less straight forward. Their
presence has been strengthened through the period. Indeed they were legally allowed to
increase the number of their branches provided that they develop their microfinance activities
(see Dauner Garniol et al., 2005). But this indicator is altered by the growth of the population
In order to analyze the heterogeneity in the banking structure and in the regional economic
development stressed in tables 2, 3 and 4, rank order correlation tests are conducted.
[Insert Table 4]
11
2.4 Rank order correlation tests
As a preliminary step of our empirical investigation that aims to assess the link
between economic and banking development and the role played by rural banks, we test for
correlation between selected banking and economic development indicators. These tests are
performed using five samples of regions: “All regions”, “All regions except NCR”,
Table 5 presents the results of our correlation analysis for our five sub-samples of regions
using Spearman rank-order tests. The null hypothesis is the absence of rank-order correlation
[Insert Table 5]
Three main results are obtained from the rank order tests. First, a positive and
significant correlation between economic development and financial depth at the regional
level is obtained when financial depth is measured by banking office density and deposits for
four of the five samples. This result is consistent with the existing empirical literature on the
finance growth nexus. The correlation obtained is stronger for the sub-sample “Developed
regions” than for the sub-samples “Intermediate developed regions” and “Less developed
regions”. When the Loans variable is used as an indicator of financial depth, the correlation is
also significant for the economically developed regions but not for the intermediate and less
economically developed regions. When banking development is measured as the ratio of total
net loans to total deposits (Intermediation), we find a positive and significant correlation for
12
Second, rank order tests show different results for the sub-sample “All regions except
NCR”. A positive and significant correlation is obtained between economic and banking
development only when the bank office density is used as a measure of financial depth.
Third, the most interesting result with regard to our issue is related to the role of rural
between economic development and rural banks presence for the samples “All regions”, “All
regions except NCR” and “Developed regions”. On the contrary, a positive and significant
correlation is obtained between the variables PC_RGDRP and the market share of rural banks
which means that the higher is the market share of rural banks, the higher is the regional
economic development.
(its existence, level or sign) is argued to arise primarily from the estimation techniques used to
assess this relationship (times series, panel data, see Apergis et al., 2007). According to
Apergis et al., 2007, the cross-sectional estimation methodology misses (i) to address the
issue of integration and cointegration properties of the data, and (ii) to examine the direction
of causality between economic and financial development. In estimating panel data, Apergis
et al., 2007, point out that using instrumental variables and GMM dynamic panel estimators
alone to account for potential biases induced by simultaneity of regressors, omitted variables
Following this methodology and in order to explore the relationship between banking
development, economic development and the effect of rural banks, we first conduct panel unit
13
root tests on the dataset. We used the Im, Pesaran and Shin t-test 8 . Results are presented in
Table 6.
[Insert Table 6]
Panel unit root tests support the hypothesis of a unit root for most variables in level.
However the null hypothesis is rejected with the IPS test for bank office density at the 1%
level. In first difference, unit root tests show that all variables are stationary.
As a second step, we conduct panel cointegration tests. To test for the presence of a
long run relationship between banking and economic development, we use the methodology
suggested by Pedroni, 1999, and Pedroni, 2004. This procedure is based on Engle-Granger,
1987, two-step cointegration tests. Pedroni proposed eleven statistics that allow for
heterogeneous intercepts and trend coefficients across cross sections. Two alternatives classes
of statistics are tested: the first one is based on the within dimension of the panel while the
second one is based on the between dimension of the panel. According to Pedroni, 2004, for
very small value of T (time dimension) and a limited number of individuals, the Phillips-
Peron (PP) statistic performs relatively better than the others. Therefore we rely on this
Since the direction of the relationship between economic and banking regional
development is not clear, we perform cointegration tests on the two following models:
8
The IPS test is based on individual ADF regressions and assumes a separate unit roots between the cross-
sections units.
14
PC_RGDRPt = α + βBDt + γRBMSt + εt (1)
where PC_RGRDP t is per capita real gross regional domestic product, BDt is a measure of
banking development and RBMS a measure of rural bank market share. According to
Christopoulos and Tsionas, 2004, the results of the cointegration analysis undertaken on the
two models will give us an insight on the long-run causality between our two variables
In equation (1), cointegration tests are performed using as explanatory variables (i)
alternatively one of the three I(1) measure for banking development 9 , and (ii) alternatively
one of the two I(1) measures for rural banks market share (RB Loan share or RB Resource
share).
In equation (2), cointegration tests are performed using as the explained variable
alternatively one of the three I(1) measures for banking development (FD) and as explanatory
variables (i) the per capita real gross regional domestic product (PC_RGRDP) and (ii)
alternatively one of the two I(1) measures for rural banks market share (RB Loan share or RB
Resource share). Test results for equation (1) and equation (2) are respectively shown in tables
When per capita real gross regional domestic product is used as the dependent variable
(equation (1)), the null hypothesis of no cointegration is rejected for the whole sample and for
the sub-samples “Intermediate developed regions” and “Less developed regions”. Therefore
results show that in the long run regional banking development affects regional economic
9
Loans, Deposits or Local intermediation variables. The bank office density variable is I(0). The statistic
presented is the Phillips-Perron group statistic.
15
development. However, when banking development is used as the dependent variable
(equation (2)), the null hypothesis is only rejected once for all samples when deposits are used
to build the financial depth measure 10 , and in this case the equation might reflect a money
demand based on transactions motive. Overall the results shows that the long run relationship
economic development, and is even unidirectional if we do not consider deposits as the most
appropriate proxy of financial depth in order to assess the impact of the banking system on
11
economic activity . Therefore the analysis of the relationship between economic
development, banking development and the role of rural banks will be focused on equation (1)
for the four sub-samples “All regions”, “All regions except NCR”, “Intermediate developed
The estimation of the long run relationship is performed using alternatively three
different estimators: ordinary least squares (OLS), fully-modified least squares (FMOLS)
initially proposed by Phillips and Hansen, 1990, and the dynamic least squares (DOLS) of
Saikkonen, 1991, and Stock and Watson, 1993. We first use on our panel data set the OLS
estimator. But, as underlined by Kao and Chiang, 2000, this estimator suffers from an non-
negligible bias in finite samples. We then use the FMOLS estimator as suggested by Pedroni
(1996) which performs better than the OLS estimator for small samples as in our case.
Moreover, as shown in Pedroni (2000), the FMOLS methodology addresses the problem of
endogeneity of the regressors. Kao and Chiang, 2000, find from Monte-Carlo simulations that
the DOLS estimator over-performs the FMOLS and OLS estimators in estimating
10
The null hypothesis is also rejected for the economically developed regions when financial depth is proxied
using loans.
11
In order to analyze properly the causality, we would need to distinguish the short and long run causation. To
do so, we need to build the error correction model and then to study the first difference lagged variables which
would provide evidence on the direction of the short run causation while the significance of the error correction
term provide evidence of the long run causation (Canning and Pedroni, 2008; Narayan et al., 2008).
Unfortunately, we do not have enough time observations for such an analysis.
16
cointegrated panel regressions, therefore, we also present results using the DOLS
methodology.
Table 7 displays the long run relationship between economic, banking development and
the presence of rural banks for the four sub-samples of regions for which the Pedroni test is
conclusive.
[Insert Table 7]
First, we can not clearly identify a consistent impact of our banking development
variables on regional economic development. Depending on the proxies, the samples and the
methodologies, the coefficient of the banking development variable can be either positive or
negative, and either significant or not. While analyzing the impact of banking development on
economic development, OLS estimations show opposite results depending on the proxy
retained. Financial depth when measured as the ratio of total net loans on the nominal regional
gross domestic product (Loans variable) has a negative impact on economic development for
the samples “All regions” and “Less developed regions” and not significant for the two other
samples. The literature has often pointed out that variables such as loans have an ambiguous
status. They are good measure of the size of the financial sector and could also well predict
banking crisis. Indeed we might explained our result by a strong decrease in the level of loans
granted by commercial banks following the Asian crisis, whereas the economic activity
recovered more rapidly (Podpiera and Singh, 2007).This negative link between financial
depth and economic development when data set includes 1997-1998 Asian crisis is in line
with the finding of Rousseau and Wachtel, 2005. Unfortunately, given the availability of the
17
data, it was not possible to work on a period excluding the Asian crisis. The role of
commercial banks is of main importance for the country because of their strong presence
especially in the wealthy regions. This negative link holds for three out of the four samples if
we use the Local intermediation variable instead of the Loans variable. However, when we
use the ratio of total deposits on the nominal regional gross domestic product (Deposits
development for three out of the four samples. Using the FMOLS estimation procedure leads
mainly to the same overall explanation of the results even if the results don’t tally for each
sample and if the significance of the coefficients is stronger for the Local intermediation
variable and than for the Loans variable. However if we consider the coefficients obtained
with the DOLS methodology, the link between banking development and economic
development collapses. This finding of the sensitivity of our results to the econometric
methodology used has been highlighted by a number of studies such as Favara, 2003, and
Dufrénot et al., 2007, and is often explained by the difficulty to specify correctly the origin of
non-stationary variables. An answer could be the use of common factor models such as the
Components) proposed by Bai and Ng, 2004. Unfortunately, the sample does not able us to
use this technique as it requires a large time and individual dimension panels.
Second, a very interesting finding is the positive and always significant except once
impact of the presence of rural banks on economic development whatever the proxies, the
samples and the methodologies used. We show for all samples studied that rural bank
presence affects positively the economic activity even for the sample “All regions” for which
we obtained a negative relationship from Spearman rank-order tests 12 . The results are robust
to the econometric estimators used: OLS, FMOLS and DOLS estimation procedures give
12
We remind the reader that the “Developed regions” sub-sample is not included because we do not find a
cointegration relationship between the variables.
18
mostly the same results. However in the case of the sample “All regions”, using the FMOLS
estimators we find an abnormal high value for the coefficient of the rural bank variable
whatever the proxy used for banking development. This result could be explained by
specificities of the National Capital Region (where the presence of rural banks is negligible)
as we do no longer find such coefficient for the sample “All regions except NCR”. Results
also show that the impact of rural banks on economic development for the intermediate
developed regions is usually stronger than for the less developed regions and than the average
effect of rural banks on economic development. It might suggest that a threshold exists.
Threshold effects are also found in the literature of cross-countries growth finance nexus,
built on country ranking using ex-ante economic or financial development criteria (see
Dermirgüç-Kunt and Levine, 2008). In particular, Rioja and Valev, 2004, show in their study
where the Philippines are classified as a low-income country, the existence of a threshold
development is enhanced not only by capital accumulation but also through productivity
growth. Thus this positive impact of financial development is all the more important that the
country has a high level of per capita GDP. This work show that in a low-income country
words, a minimum level of economic development should be required for the influence of
To check for the robustness of the results, we also estimate the long run equilibrium
using the total resources market share of rural banks as a proxy of rural bank presence. The
results found are mainly the same while a little less significant (See Annex II).
19
Section 4. Conclusion
This paper aims to contribute to the finance/growth literature by analyzing the specific
effect of local banks on regional economic performance. More precisely it focuses on the
influence of rural banks, which are mainly dedicated to foster expansion of rural areas, on
economic development in the sixteen regions of the Philippines. The regional market analysis
undertaken in this paper relies on regional balance sheet data for the three types of Philippine
banks even those from nationwide banks which are usually confidential information 13 .
When examining the relationship between banking and economic development using
indicators built at the regional banking industry level, there is no clear-cut evidence of a
banking-led economic development. But, if we focus on the specific effect of rural banks
presence, a positive impact is found on economic development for the intermediate and less
developed regions, with a stronger impact for intermediate regions. This result might indicate
the existence of threshold effect that is a minimum level of yield per capita required for rural
The Philippines experience shows that the presence of rural banks which have an expertise in
economic activity especially in the rural areas of developing countries. The results suggest a
segment of the market favoring therefore the financing of projects that commercial banks
would not have done. Further research using individual bank data could consist in analyzing
the threshold effect and in explaining the comparative advantage by assessing the
13
This information was obtained here thanks to the courtesy of the Central Bank of the Philippines.
20
References
Agabin, M. and Daly, J.L. (1996) An Alternative Approach to Rural Financial Intermediation:
Apergis, N., Filippidis, I. and Economidou, C. (2007) Financial Deepening and Economic
Growth Linkages: A Panel Data Analysis, Review of World Economics 143(1), 179-198.
Avery R. and Samolyk, K. (2004) Bank consolidation and small business lending: the role of
Bai, J. and Ng, S. (2004) A PANIC Attack on unit roots and cointegration, Econometrica 72,
1127-1177.
Berger, A.N., Hasan, I. and Klapper, L.F. (2004) Further Evidence on the Link between
Burgess, R. and Pande, R. (2005) Do Rural Banks Matter? Evidence from the Indian Social
Canning, D. and Pedroni, P. (2008) Infrastructure, long-run economic growth and causality
Carbo Valverde, S. and Fernandez, F.R. (2004) The Finance-Growth Nexus: A Regional
Carbo Valverde, S., Lopez del Paso R. and Fernandez, F.R. (2007) Financial Innovations in
evidence from panel unit root and cointegration tests, Journal of Development Economics 73,
55–74.
21
Collender, R. N. and Shaffer, S. (2003) Local bank office ownership, deposit control, market
structure and economic growth, Journal of Banking and Finance 27, 27-57.
Dauner Gardiol, I., Helms, B. and Deshpande, R. (2005) Philippines. Country-Level Savings
387-411.
Demirguc_Kunt A. and Maksimovic V. (1998) Law, Finance, and Firm Growth, Journal of
Demirgüç-Kunt, A. and Levine, R. (2008) Finance, Financial Sector Policies, and Long-Run
DeYoung, R., William C.H. and Udell, G. (2004) The Past, Present, and Probable Future for
Dufrénot, G., Mignon V. and Péguin-Feissolle, A. (2007) Testing the Finance-Growth Link:
is there a Difference between Developed and Developing Countries?, CEPII Working Paper,
2007-24.
Engle, R.F. and Granger, C.W.J. (1997) Cointegration and error-correction: representation,
Favara, G. (2003) An empirical reassessment on the relationship between finance and growth,
Gochoco-Bautista, M. (1999) The Past Performance of the Philippine Banking Sector and
Challenges in the Postcrisis Period, Rising to the Challenge in Asia: A Study of Financial
Hao, C. (2006) Development of financial intermediation and economic growth: the Chinese
22
Hakenes H., Schmidt, R. and Xie, R. (2009) Regional banks and economic development:
Hasan, I., Koetter, M. and Wedow, M. (2009) Regional growth and finance in Europe: is there
a quality effect of bank efficiency?, Journal of Banking and Finance 33, 1446-1453.
Kao, C. and Chiang, M. H. (2000) On the estimation and inference for cointegration in panel
data, Advances in Econometrics, 15, edited by B. Baltagi and C. Kao, Elsevier Science, 179-
22.
King, R. G. and Levine, R. (1993a) Finance and growth: Schumpeter might be right,
King, R. G. and Levine, R. (1993b) Finance, entrepreneurship and growth: theory and
Levine, R. (2005) Finance and Growth: Theory and Evidence, in Aghion P. and S. Durlauf
Llanto, G. (2005) Rural Finance in the Philippines: Issues and policy challenges, Agricultural
Narayan, P.K., Nielsen, I. and Smyth, R. (2008) Panel data, cointegration, causality and
Wagner’s law: Empirical evidence from Chinese provinces, China Economic Review 19, 297-
307.
23
Pedroni, P. (2004) Panel Cointegration: Asymptotic And Finite Sample Properties Of Pooled
Time Series Tests With An Application To The Ppp Hypothesis, Econometric Theory,
Pedroni, P. (2000) Fully Modified OLS for Heterogeneous Cointegrated Panels, (Eds) Non-
Stationary Panels, Panels Cointegration and Dynamic Panels, vol 15, 93-130, Elsevier.
Pedroni, P. (1999) Critical values for Cointegration Tests in Heterogeneous Panels with
Pedroni, P. (1996) Fully Modified OLS for Heterogeneous Cointegrated Panels and the Case
Phillips, P.C.B. and Hansen, B.E. (1990) Statistical inference in individual variables
Podpiera, R. and Singh, R. (2007) Credit growth and Bank Balance-Sheets in the Philippines,
Rajan, R.G. and Zingales, L. (1998) Financial Dependence and Growth, American Economic
Rioja, F. and Valev, N. (2004) Finance and the source of growth at various stages of
Wachtel, P. (2003) How much do we really know about growth and finance?, Federal Reserve
24
Wachtel, P. and Rousseau, P. (1995) Financial Intermediation and Economic Growth: A
Historical Comparison of the United States, United Kingdom and Canada. In M.D. Bordo and
R. Sylla, Eds. Anglo-American Financial Systems. New York: Irwin Professional Publishing.
Annex I
Annex II
25
Table 1. Per Capita Real Gross Regional Domestic Product: Summary statistics
and ranking indicators (1993-2005)
1993 1993 Rank 2005 2005 Rank
Developed regions
NCR 879 1 1452 1
Northern Mindanao 516 2 619 2
CAR 373 3 585 3
Intermediate developed regions
Socksargen 293 8 481 4
Central Visayas 320 6 432 5
South Luzon 368 4 418 6
Western Visayas 287 9 416 7
Central Luzon 313 7 357 8
Davao 352 5 310 9
Less developed regions
Zamboanga Peninsula 258 10 280 10
Eastern Visayas 192 12 258 11
Ilocos 181 14 257 12
Cagayan Valley 187 13 240 13
Caraga* 209 11 223 14
Bicol 178 15 210 15
ARMM 111 16 125 16
Mean 344 468
Mean excluding NCR 312 410
Median 290 333
Median excluding NCR 287 310
Source: National Statistical Coordination Board; *Caraga figure corresponds to 1997. Real gross
regional domestic product is expressed in thousands of pesos at 1990 prices.
26
Table 2. Descriptive statistics (Average value of the variables over the 1993-2005 period)
Banking RB office
Total net RB Resource
Total deposits* Intermediation PC_RGRDP Deposits Loans office RB Loan share density
loans* share
density
DEVELOPED REGIONS
NCR 1 299 034 1 274 290 0,99 1,13 1,19 1,19 0,24 0,11% 0,18% 2,14%
Northern Mindanao 22 226 15 989 0,82 0,52 0,16 0,13 0,08 7,21% 9,55% 34,99%
CAR 16 754 4 470 0,26 0,51 0,24 0,06 0,07 5,36% 17,28% 36,28%
Mean 446 005 431 583 0,69 0,72 0,53 0,46 0,13 0,23% 0,37% 27,47%
Standard deviation 738 750 729 828 0,38 0,36 0,57 0,63 0,10 0,04 0,09 0,16
INTERMEDIATE DEVELOPED REGIONS
Socksargen 10 579 4 778 0,54 0,35 0,10 0,05 0,04 9,56% 14,68% 40,50%
Central Visayas 90 404 43 462 0,56 0,38 0,41 0,23 0,08 3,58% 4,68% 26,52%
South Luzon 120 274 38 084 0,37 0,39 0,26 0,09 0,10 13,05% 24,92% 38,93%
Western Visayas 50 817 19 782 0,47 0,34 0,24 0,11 0,06 5,39% 10,30% 32,42%
Central Luzon 84 780 33 918 0,46 0,33 0,33 0,14 0,09 10,95% 17,27% 36,95%
Davao 38 855 23 996 0,67 0,35 0,23 0,15 0,06 5,88% 7,33% 30,47%
Mean 65 952 27 337 0,51 0,36 0,26 0,13 0,07 8,85% 13,46% 34,30%
Standard deviation 39 831 14 113 0,10 0,02 0,10 0,06 0,02 0,04 0,07 0,05
LESS DEVELOPED REGIONS
Zamboanga Peninsula 17 158 5 508 0,36 0,26 0,22 0,08 0,03 4,13% 11,15% 23,53%
Eastern Visayas 13 694 4 449 0,36 0,22 0,17 0,06 0,03 6,44% 12,95% 39,47%
Ilocos 37 820 11 922 0,38 0,23 0,38 0,14 0,08 12,84% 26,96% 49,74%
Cagayan Valley 15 529 8 779 0,64 0,23 0,24 0,15 0,07 16,40% 20,73% 57,75%
Caraga 8 954 4 395 0,51 0,22 0,19 0,09 0,04 18,98% 26,47% 53,29%
Bicol 19 403 9 183 0,51 0,19 0,22 0,11 0,04 12,72% 20,79% 42,53%
ARMM 4 568 1 315 0,31 0,13 0,21 0,06 0,02 2,75% 8,50% 23,57%
Mean 16 732 6 507 0,44 0,21 0,23 0,10 0,05 12% 21,22% 41,01%
Standard deviation 10 576 3 612 0,12 0,04 0,07 0,04 0,02 0,06 0,07 0,13
ALL REGIONS EXCEPT NCR
Mean 115 678 94 020 0,51 0,36 0,30 0,18 0,07 8,45% 14,60% 37,56%
Standard deviation 317 366 315 014 0,194 0,23 0,25 0,27 0,05 0,05 0,08 0,10
ALL REGIONS
Mean 36 788 15 335 0,48 0,31 0,24 0,11 0,06 1,55% 2,34% 35,31%
Standard deviation 34 985 13 634 0,15 0,11 0,08 0,047 0,02 0,05 0,07 0,13
Loans: total net loans/nominal regional gross domestic product, Deposits: total deposits/regional gross domestic product, Banking office density: number of banking offices per capita,
Intermediation: total net loans/total deposits, RB Loan share: net loans of rural banks per region/total net loans per region, RB Resource share: total resources of rural banks per
region/total resources per region.* In millions of pesos. Source: Bangko Sentral ng Pilipinas; National Statistical Coordination Board.
27
Table 3. Market share per type of banks in the Philippines 1993-2005
Commercial banks Thrift banks Rural Banks
56,81 % 17,81 % 25,37 %
Banking office density
45,45 % 16,96 % 37,60 %
90,67 % 7,77 % 1,55 %
Total resources
79,07 % 12,05 % 8,88 %
89,13 % 8,52 % 2,34 %
Total net loans
73,12 % 12,44 % 14,44 %
Numbers in italics are the market share computed for the group “All regions except NCR”, Source: Bangko
Sentral ng Pilipinas
28
Table 4. Rural banks market share per region
1993 2005
RB resource RB loan RB office RB resource RB loan RB office
1
Rank 2
Rank 3
Rank Rank Rank Rank
share share density share share density
DEVELOPED REGIONS
NCR 0,05% 16 0,07% 16 1,21% 16 0,21% 16 0,53% 16 2,61% 16
Northern Mindanao 4,92% 11 5,96% 14 31,38% 13 10,37% 7 18,29% 11 43,31% 4
CAR 4,01% 12 14,08% 4 40% 7 7,94% 10 34,94% 5 38,18% 10
INTERMEDIATE DEVELOPED REGIONS
Socksargen 9,18% 5 12,72% 6 41,54% 6 10,01% 8 20,34% 10 39,18% 8
Central Visayas 3,92% 13 4,88% 15 31,46% 12 5,20% 15 8,82% 15 26,59% 14
South Luzon 11,25% 2 18,01% 2 38,55% 8 13,50% 5 37,74% 3 39,82% 7
Western Visayas 5,23% 9 8,06% 10 37,06% 9 6,43% 13 16,85% 12 33,18% 12
Central Luzon 8,44% 6 12,54% 7 32,77% 11 14,30% 4 26,59% 7 41,32% 5
Davao 5,22% 10 6,44% 13 29,86% 14 8,97% 9 15,37% 14 33,60% 11
LESS DEVELOPED REGIONS
Zamboanga 2,87% 14 6,98% 11 24,39% 15 7,08% 12 22,87% 9 29,91% 13
Peninsula
Eastern Visayas 5,67% 8 12,45% 8 42,71% 5 7,40% 11 16,66% 13 39,10% 9
Ilocos 12,90% 1 26,82% 1 51,36% 2 13,21% 6 38,16% 2 50,92% 3
Cagayan Valley 10,58% 3 13,15% 5 53,03% 1 19,25% 2 30,90% 6 63,39% 1
4
Caraga 10,22% 4 16,51% 3 47,94% 3 23,20% 1 45,10% 1 61,40% 2
Bicol 6,61% 7 11,24% 9 45,51% 4 17,26% 3 35,39% 4 39,91% 6
ARMM 2,62% 15 6,75% 12 36,96% 10 6,11% 14 25,56% 8 19,23% 15
1 2 3
Share of total resources of rural banks over total resources of all types of banks Share of net loans granted by rural banks over total net loans granted ; Number bank offices
4
for of rural banks over total number of bank offices; CARAGA figure corresponds to 1996. Source: Bangko Sentral ng Pilipinas.
29
Table 5. Correlation Analysis: Spearman rank-order with PC_RGRDP as referent variable
All regions All regions Developed Intermediate Less developed
except NCR regions developed regions regions
Banking Development (BD)
Financial depth
- Loans 0.233*** 0.066 0.841*** -0.824 0.043
- Deposits 0.244*** 0.072 0.884*** 0.247** 0.246**
- Banking office density 0.652*** 0.576*** 0.948*** 0.314*** 0.358***
Local intermediation
- Intermediation 0.181*** 0.011 0.485*** -0.397*** 0.110
Rural banks market share (RBMS)
- RB Loan share -0.261*** -0.103 -0.489*** 0.297*** 0.284***
- RB Resource share -0.380*** -0.134* -0.653*** 0.080 0.313***
Boldface values denote a significant presence of a rank-order correlation. (***), (**) and (*) signify rejection of the null
hypothesis of absence of rank-order correlation at the 1%, 5% and 10% levels respectively.
Table 6. Im, Pesaran and Shin (IPS) panel unit root tests
Variable in level Variable in first difference
PC_RGDRP 2.77 -3.86***
Financial depth
- Loans 0.25 -2.75***
- Deposits -1.48 -3.11***
- Banking office density -2.83***
Local intermediation
30
Table 7. Long run relationship between economic development, banking development (BD) and the
role of rural banks (RB)1 using OLS, DOLS and FMOLS estimators
All regions
BD -0.20*** -0.02 0.01 -0.06 0.12 -0.18* -0.06* -0.04** -0.01
31
Table A1. Pedroni panel cointegration test
Dependent variable: PC_RGDRP
Rural bank market share
RB Loan share RB Resource share
All regions (N1 = 204 ; N2 = 16)
Financial depth (Loans) -3.80*** -2.10**
Financial depth (Deposits) -5.98*** -3.22***
Local intermediation -3.53*** -2.96***
All regions except NCR (N1 = 195 ; N2 = 15)
Financial depth (Loans) -4.24*** -2.23**
Financial depth (Deposits) -6.41*** -3.45***
Local intermediation -3.86*** -3.19***
Economically developed regions (N1 = 39 ; N2 = 3)
Financial depth (Loans) 0.61 1.24
Financial depth (Deposits) 0.83 1.08
Local intermediation 1.33 1.30
Intermediate developed regions (N1 = 78 ; N2 = 6)
Financial depth (Loans) -4.25*** -0.41
Financial depth (Deposits) -7.48*** -1.76
Local intermediation -3.45*** -0.34
Less developed regions (N1 = 91 ; N2 = 7)
Financial depth (Loans) -2.61*** -3.77***
Financial depth (Deposits) -2.66*** -3.95***
Local intermediation -3.43*** -5.12***
32
Table A2. Pedroni panel cointegration test
Dependent variable: Loans
Rural bank market share
RB Loan share RB Resource share
All regions (N1 = 204 ; N2 = 16)
Output per capita (PC_RGDRP) -0.34 -0.48
All regions except NCR (N1 = 195 ; N2 = 15)
Output per capita (PC_RGDRP) -0.47 -0.57
Economically developed regions (N1 = 39 ; N2 = 3)
Output per capita (PC_RGDRP) -1.47* -1.95**
Intermediate developed regions (N1 = 78 ; N2 = 6)
Output per capita (PC_RGDRP) 0.37 0.66
Less developed regions (N1 = 91 ; N2 = 7)
Output per capita (PC_RGDRP) 0.28 0.19
Dependent variable: Deposits
Rural bank market share
RB Loan share RB Resource share
All regions (N1 = 204 ; N2 = 16)
Output per capita (PC_RGDRP) -7.67*** -5.81***
All regions except NCR (N1 = 195 ; N2 = 15)
Output per capita (PC_RGDRP) -8.00*** -6.17***
Economically developed regions (N1 = 39 ; N2 = 3)
Output per capita (PC_RGDRP) -3.80*** -4.19***
Intermediate developed regions (N1 = 78 ; N2 = 6)
Output per capita (PC_RGDRP) -3.13*** -3.76***
Less developed regions (N1 = 91 ; N2 = 7)
Output per capita (PC_RGDRP) -4.07*** -1.63*
Dependent variable: Intermediation
Rural bank market share
RB Loan share RB Resource share
All regions (N1 = 204 ; N2 = 16)
Output per capita (PC_RGDRP) 0.84 0.62
All regions except NCR (N1 = 195 ; N2 = 15)
Output per capita (PC_RGDRP) -0.88 -0.70
Economically developed regions (N1 = 39 ; N2 = 3)
Output per capita (PC_RGDRP) -0.58 -0.84
Intermediate developed regions (N1 = 78 ; N2 = 6)
Output per capita (PC_RGDRP) 1.68 1.28
Less developed regions (N1 = 91 ; N2 = 7)
Output per capita (PC_RGDRP) 0.30 0.50
(***), (**) and (*) signify rejection of the null hypothesis of absence of long run relationship at the 1%, 5% and 10%
levels respectively. N1 and N2 are respectively the number of observations and the number of cross-section units.
33
Table A3. Long run relationship between economic development, banking development (BD) and the role
of rural banks (RB)1 using OLS, DOLS and FMOLS estimators
BD: Financial depth: Loans Financial depth: Deposits Local intermediation
OLS FMOLS DOLS OLS FMOLS DOLS OLS FMOLS DOLS
All regions
BD -0.24*** -0.14*** -0.28*** 0.02 0.24*** 0.06 -0.09*** -0.07*** -0.13**
34