Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
PDD WORKING PAPERS Discussion Paper First High-Level Follow-up Dialogue on Financing for Development in Asia and the Pacific Incheon, Republic of Korea
30-31 March 2016
DP/16
March 2016
Monica Das and Sudip Ranjan Basu
FINANCE AND GROWTH: DOES ONE SIZE FIT ALL? EVIDENCES FROM A
NONPARAMETRIC ESTIMATION
For participant only
FINANCE AND GROWTH: DOES ONE SIZE FIT ALL? EVIDENCES FROM ANONPARAMETRIC
ESTIMATION
Monica Das and Sudip Ranjan Basu
Monica Das, Department of Economics, Skidmore
College, 815 N Broadway, Saratoga Springs, NY
12866; [email protected]; 518-580-5096 and
Sudip Ranjan Basu, Economic Affairs Office,
ESCAP
For more information, contact:
Macroeconomic Policy and Financing for Development Division (MPFD) Economic and Social Commission for Asia and the Pacific United Nations Building, Rajadamnern Nok Avenue, Bangkok 10200, Thailand Email: [email protected]
Dr. Aynul Hasan Director Macroeconomic Policy and Financing for Development Division
DP/16
i
Contents
I. Introduction ........................................................................................................................... 2 II. Literature ................................................................................................................................ 4 A. Impact of financial sector development on growth ........................................................ 4 B. Impact of financial structure on growth .......................................................................... 6 III. Empirical Methodology ......................................................................................................... 7 IV. Data and Empirical Model ..................................................................................................... 9 A. Data .................................................................................................................................... 9 B. Dependent and independent variables ............................................................................ 9 C. The Empirical model ........................................................................................................ 10 V. Results ............................................................................................................................... 10 A. Core model results with regional classification ............................................................. 11 B. Core model results with UN regional groupings ............................................................ 12 C. Core model results with size of financial sector ............................................................ 12 D. Core model results with institutional quality................................................................. 13 E. Comparing estimates across bank-based versus market-based system ...................... 13 VI. Conclusions .......................................................................................................................... 14 References ............................................................................................................................... 15 Annex ............................................................................................................................... 17
DP/16
Discussion Paper
Macroeconomic Policy and Financing for Development Division
Finance and Growth: Does One Size Fit All? Evidences from a Nonparametric Estimation
by
Monica Das and Sudip Ranjan Basu
March 2016
Abstract
The objective of the paper is to use nonparametric methodology to examine the
relationship between indicators of financial development and economic growth. The
paper uses the Li-Racine (2004) generalized kernel estimation methodology to
examine the role of financial development in understanding economic performance
across countries and country-groups. We also control for other factors that influence
economic performance such as, trade openness, government policies and institutional
quality. The paper supports the view that countries with larger and well-diversified
financial markets are in a better position to reap benefits from financial sector
deepening and forward-looking public policies. This holds true irrespective of the
initial size of the financial sector in the country as well as the strength of institutional
quality.
Keywords: Bank, Capital market, nonparametric estimation, growth.
Authors’ e-mail addresses: [email protected] and [email protected].
The views expressed in this discussion paper are those of the author(s) and should not necessarily be considered as reflecting the views or carrying the endorsement of the United Nations. Discussion papers describe research in progress by the author(s) and are published to elicit comments and to further debate. This publication has been issued without formal editing.
Financial development and growth: Does one size fit all? Evidences from a nonparametric
2
I. Introduction
The financial sector activity is significantly linked to economic growth. From Armenia
to America, and from Bangladesh to Botswana, the countries have not only seen the
progression in the deepening of the financial sector, but it has also raised challenges in
policymaking that require balancing between the size and depth, and its impact on
economic growth. Increasingly, over the past decades, the volatility and stress in the
financial sectors have created concerns both for the prospects of sustainable growth as
well as in ensuring jobs creation. The transmission channel from finance and economic
growth is not uni-directional, rather there are multiple ways in which these two
macroeconomic factors are related.
This paper uses nonparametric regression methodology to estimate the impact of the
financial system on growth at the cross-country level during the past two decades.
Therefore, the key issue here is to understand the following: Is financial sector
deepening positively linked to economic growth at all levels of development? or does it
depend on the size of the economy and the nature of the financial system? The paper
provides empirical evidence on the issue that “a one-size-fits-all” policies is not
applicable across countries, and during the period.
In the traditional macroeconomic theory, the financial sector development is intertwined
to capital accumulation, one of the input factors in the production function. In particular,
an improvement in the financial market can produce positive outcomes through:
reducing the loss of resources required to allocate capital; increasing the capital saving
ratio; and raising capital productivity. A development in financial sector therefore
contributes to capital accumulation and effectively that subsequently ensure robust
economic growth, and also helps to increase the productivity by selecting the most
profitable investment projects. This linkage between financial sector development and
productivity growth is particularly critical in periods of rapid technological progress and
innovation, which allow countries to benefit and enjoy higher rates of economic growth.
Another discussion that has taken centre stage of the current policy regimes is that of the
expansion of integrated capital market, the evolution of financial structure in different
economies on growth. The countries financial structure is mostly defined either in the
form of a bank-based or market-based financial system. In the bank-dominated financial
system, banks take the dominant role in mobilizing resource, identifying bankable
projects, and managing risk, and incentivizing technological progress and innovation. On
the other hand, with the market-dominated financial system, market operations provides
the necessary lead in identifying investment portfolios, devising risk management
activities, and supporting for high-risk high-return projects, thus reducing the inherent
inefficiencies associated with banks-based financing system.
With the growth of financial services around the world, it is argued that financial
structure does not have added valued to economic growth. Importantly, for finance and
growth to work towards increasing the overall economic prosperity, there is a role of the
property rights and institutional quality to improve the effectiveness of the financial
system in facilitating the economic growth process.
DP/16
3
The contribution of the paper is in the application of the Li and Racine (2007)
nonparametric methodology to estimate the relationship between the size of an
economy and the size of its financial sector in a panel with both time and country
effects. In the estimation of any parametric model, misspecification bias and
endogeneity/omitted variable bias could be present.
The nonparametric estimates in the paper effectively deal with misspecification bias.
Another advantage of the nonparametric methodology is that the paper obtains
estimates of slope coefficients (ΔY/ΔX) for every data point. In other words, paper
estimates the impact of every finance variable on economic growth for every country
in every time period. This allows to provide several group specific aggregates of the
slope coefficients. The standard errors of the group aggregates are obtained via
bootstrapping.
Apart from key role of the size and composition of financial institutions to boost
economic performance, there are growing number of research papers in literature to
document the critical role of efficient domestic institutional conditions as well as
human capital accumulation and geography (Acemoglu et al. 2001; Sachs, 2003;
Easterly and Levine, 2003; Rodrik et al. 2004; and Basu, 2008). Among other group
aggregates, paper provides the nonparametric estimates the economy-finance
relationship for country groups based on property rights (pr) and strength of
institutional quality (ief). These results add to the discussion on the role of
institutional quality in shaping the linkages between financial sector development and
economic growth performance.
The nonparametric estimates find strong support for positive significant impact of
higher level of financial sector development on GDP per capita, by using several
control variables. For majority of the countries examined, the impact of higher level
of financial sector development on the GDP per capita are quite favorable, while the
role of trade policies and institutional quality remain key complementary policy
initiatives. Since the Li-Racine methodology provides weighted estimates (weights
determined by all observations) of the regression function and its slope at every data
point, the paper also examines the nonparametric estimates for various sub-groups by
continents and country characteristics.
The impact of financial sector development, both bank-based and market-based on
GDP per capita is far from uniform across countries or time periods. However, the
favorable relationship between these two or minimal support for a negative relation
between the two variables, is robust to most sub groups and country characteristics.
The paper employs nonparametric methodology that conducts the analysis without
pre-specifying any functional form. Also, it obtains the estimates of the slope
coefficients for every data point. The slope estimates can be aggregated for every sub
groups based on several criteria, such as quality of institutions or size of the financial
sector.
We now sketch a course for the rest of the paper. Section 2 presents a brief survey of the
most relevant papers in the literature. Section 3 provides the Li-Racine estimation
technique for mixed data, utilized in the paper to the estimation of economy-finance
relationship. Section 4 discusses the data set and the empirical model. Main results of the
paper are presented in section 5 and section 6 concludes the paper.
Financial development and growth: Does one size fit all? Evidences from a nonparametric
4
II. Literature
A large body of literature dating as far back as Schumpeter (1911) argues for the positive
relationship between development of a country's financial sector and the rate of growth
of its per capita income (see Levine 1997). The paper by Rajan and Zingales (1998)
argues a reason why financial development leads to economic growth could be that it
reduces the costs of external finance. Their paper demonstrates that industrial sectors that
are relatively more in need of external finance develop disproportionately faster in
countries with more-developed financial markets. Most empirical works seem to agree
with the Schumpeterian arguments.
Similarly, King and Levine (1993) present cross country evidence of a positive
relationship between financial development and economic growth via the mechanism of
physical capital accumulation and growth of economic efficiencies, using data from 80
countries from 1960 to 1989 after controlling for numerous country and policy
characteristics. Furthermore, Levine and Zervos (1998) shows that stock market
liquidity and banking development both positively predict growth, capital accumulation,
and productivity improvements when entered together in regressions, even after
controlling for economic and political factors. Levine et al (2000) estimate a GMM
dynamic panel model with data from 74 countries averaged over 5-year intervals during
1960 – 1995. They find that legal and accounting reforms can boost financial
development and accelerate economic growth.
Some of the papers in the empirical literature question the previous findings. Easterly
and Stiglitz (2000) take a different view that financial development is desirable up to a
point. They find that financial systems that feature debt more prominently than equity
are more vulnerable to growth collapses. Cecchetti and Kharraoubi (2012) conclude as is
the case with many things in economic literature, with financial development countries
can have too much of a good thing. The relationship between financial development and
economic growth is a non-linear inverted-U relationship.
The discussion on this issue can further be grouped into two key areas as follows:
A. Impact of financial sector development on growth
A large number of studies have identified empirically the positive role of financial
activity in economic growth and development. Finance Watch (2014) is a survey study
on relevant literature for Europe, which collectively points to the inverted U-shaped
curve that illustrates the non-linear relation between private sector credit and economic
growth. Higher private credit to GDP ratio is associated with more growth until private
credit reaches 80-100% of GDP, where further financial deepening undermines growth,
as suggested by Arcand et al (2012) and Manganelli (2013). The same pattern emerges
when financial sector’s share of employment surpasses a certain threshold. Whilst
emerging economies benefit from a growing financial sector, it is correlated with weaker
economic growth in advanced economies (Cecchetti and Kharroubi, 2012; Assa, 2012).
Disparate effects on growth are also found for different types of banking activities. Beck
et el (2012) and Bezemer et al (2014) reveal that financial intermediation, especially
lending to non-financial sectors, lead to positive growth yet large European Banks invest
mostly in credit flows to the asset markets (incl. real estate), which brings insignificant
DP/16
5
or negative growth. Explanations for negative growth suggest that the growing capital
markets incur high costs and compete with other sectors in the economies for scarce
resources. The non-linear effect of finance on growth thus implies that current financial
system is too large relative to the real economy, but the trend is hard to revert because as
Benoît Cœuré (2014) suggests, the financial system can extract excessively high
informational rents and pay excessively high wages.
Cecchetti and Kharroubi (2012) is one of the first studies to rigorously measure the
impact of financial market on growth using a regression model: five-year average private
credit to GDP ratio and its quadratic form as the main regressors and a five-year average
GDP per worker growth as the dependent variable. Using the estimated coefficients, the
authors approximate the peak of the inverted U-curve at 100% of credit to GDP ratio,
suggesting than a banking system larger than the real economy will have detrimental
effects on growth. Replacing the independent variable with the five-year average
financial intermediation share in total employment, they arrive at the turning-point
employment share at roughly 3.85%.
Gambocorta, Yang and Tssatsaronis (2014) provides additional evidence for the non-
linear linkage between bank credit and GDP growth, and takes one step further to
compare the differential effect on growth from banks and financial market development.
Based on Beck and Levine (2004), their benchmark model uses the turnover ratio (the
ratio of the value of total shares traded to average market capitalization) to measure
financial market development (this ratio is used because it is not affected by asset price
evaluations). The non-linear specification suggests that both bank and market activity
positively affect growth only up to a certain point, setting the peak credit over GDP ratio
at 40% and the critical turnover ratio around 95%. Advanced economies are in the
declining part of both the bank and market expansion curves, yet EMEs can still benefit
from equity market deepening. The authors then use a cross section of downturns and
recoveries from a sample of 24 developed countries over the 1960-2013 period (database
from Bech et al, 2012) and relate financial structure to economic volatility. Findings are
that in the absence of financial crisis, bank-based systems appear more resilient in an
economic downturn, possibly because stronger banks help absorb economic shocks.
However when recession is accompanied by financial crises, countries with bank-based
systems experience three times more severe losses than those with market-oriented
financial structure. Notwithstanding the intuitively non-linear effect of finance on
growth, the implication varies for each country differs because of their stage of
development. For example, Estrada, Noland, Park and Ramayandi (2015) ascertains that
although it is possible that the relationship turns insignificant or even negative beyond
some threshold, developing Asia is well short of that possible turning point.
Beck (2013) summarizes the above arguments from existing literature and provides
guidance for future research. The paper emphasizes the role of theory in linking data and
specific identification strategies in empirical research. It also acknowledges the large
variety of methodologies and data sources applied in the finance for development
literature, ranging from RCTs and micro-experiments to historical macro analyses. Beck
also highlights the importance of close interactions between practitioners and
policymakers, as well as the relevance of developed country experience in the
developing country settings.
Financial development and growth: Does one size fit all? Evidences from a nonparametric
6
B. Impact of financial structure on growth
The financial structure of a country is based on bank-dominated systems or market
dominated system. In the bank-based systems are characterized by a few lenders in the
economy, whereas market-based systems are characterized by a large number of lenders.
Given the cost of information acquisition and the problem of free-riding, banks are more
efficient in the provision of information. However, markets have the capacity to
aggregate views and opinions on the profitability of an investment from different parties,
therefore more efficient in dealing with uncertain investment environment. Markets are
also more conducive for young innovative companies that do not have the collateral
requirements demanded by most banks.
In particular, these two systems also differ in the provision of corporate control. The
plausible long-term relationship with a bank enhances the credibility of the bank’s
commitment to provide staged funding and therefore reduce monitoring costs and
uncertainty in funding sources. On the other hand, the established relationship may
encourage collusion and over-investment in inefficient projects. A decentralized
financial system and stock markets will nevertheless stimulate greater control, facilitate
take-overs and promote operative efficiency.
There is no consensus about the structure of financial system and its correlation with
output growth. Carlin and Mayer (2003) examines the relation between the institutional
structures of 14 advanced OECD countries and their comparative growth and investment
of 27 industries from 1970 to 1995. The analysis includes three proxies for country
structure of financial systems – accounting standards (information disclosure), bank
concentration, and ownership concentration; and three industry characteristics – equity
dependence, bank dependence, and skill dependence. The authors estimate the relation
between the interaction terms and industry growth and investment. Statistical results
support the hypotheses that information disclosure, fragmentation of banking systems,
and concentration of ownership have a strong relation with growth and R&D investment
in equity financed and skill intensive industries, but there is no significant effect on fixed
capital investment. The growth of industries relying on R&D is strongly affected by
financial variables. Thus finance stimulates economic growth by affecting investment in
R&D. In order to study the features of a growth supportive financial structure, Thiel
(2001) researches on the pros and cons of a bank based or market based financial
structure, motivated by the different systems witnessed in the developed world (e.g.
Germany and Japan are bank-based, USA, UK are market-based). The results highlight
the trend of growing reliance on markets and disintermediation through banks,
manifested in the increasing concentration of banks, increasing spread of strategic
alliances and mergers among equity market organizations, and in the growth of new
forms of financial intermediaries such as pension funds, venture capitalists and risk
capital markets.
Another stream of literature, however, plays down the role of financial structure in
predicting economic growth and emphasizes the role of law and finance. Beck and
Levine (2000a), for instance, use a country-industry panel to show that financial
dependent industries do not grow faster in bank- or market-based financial systems.
Levine (2000b) finds strong support that the efficiency of the legal system is positively
related to growth and innovation. Demirgüç-Kunt and Levine (2000) evaluates the
DP/16
7
impact of financial structure on economic growth using the large international dataset
constructed by Beck, Demirgüç-Kunt and Levine (2000) that includes country level data
of 48 countries. Using three indicators for legal system – creditor, anti-director and rule
of law – as instruments for financial development, they find evidence supporting the
financial services and the law and finance views. Financial development and the
component defined by the legal protection of outsider investors explain long-term cross
country growth rates, whilst financial structure does not offer any additional information.
Noticeably, this paper provides a system of indicators of country-level financial
structure, financial development, and legal environment for an array of developed and
developing states, thus providing useful framework for future research on this topic.
III. Empirical Methodology
The basic principle behind the nonparametric estimation technique is to fit a window h
(also known as smoothing parameter) around every observation of the data set and
estimate the relationship of interest between variables in each window. A kernel density
function K(.) is used to give high weights to data points close to the window and low
weights to data points far from the window. Thus the regression relationship is
estimated, piece by piece or window by window as shown in figure (1). One of the
advantages of nonparametric estimation is that it estimates the regression function m(.)
as well as the slope coefficients β(.) at every data point.
Figure 1
If yi is the target variable (GDP per capita) and xi the policy variable (a set of financial
sector variables, policy variables or institutional quality variables), (E(yi|xi) < ) the
relation among them may be expressed in terms of the conditional moment E(yi|xi)
=m(xi). When the actual functional form is unknown, parametric specifications including
complex ones like the translog functions are deemed inadequate. Compared with the
parametric procedures, the nonparametric methodology is more proficient in capturing
non linearities in the underlying system thus dealing with the problem of model
misspecification.
Financial development and growth: Does one size fit all? Evidences from a nonparametric
8
The paper uses the Li-Racine Generalized Kernel Estimation Methodology (by Li and
Racine (2004) and Racine and Li (2004)) to examine the relationship between
financial sector development structue and GDP per capita. Equation (1) represents the
basic regression model.
iii xmy (1)
In equation (1), yi represents the ith
observation on the dependent variable (GDP per
capita) and i indexes country-time observations of N countries and T time intervals.
Also, m(.) is an unknown smooth regression function with argument xi=[ u
i
c
i xx , ],
where c
ix is a NTk vector of continuous variables (a set of financial sector variables,
policy variables or institutional quality variables), u
ix is a NT1 vector of unordered
discrete variables (country effects) and i is a NT1 vector of errors. Following the Li-
Racine methodology, we take a first order Taylor expansion of (1) around xj to obtain
equation (2).
ij
c
j
c
iji xxxxmy (2)
Here, (xj) is the partial derivative of m(xj) with respect to xc. The estimate of (xj)
[m(xj) (xj)]’ is represented by equation (3).
j
j
j x
xmx
ˆ
ˆˆ
i
ic
j
c
ih
ic
j
c
i
c
j
c
i
c
j
c
i
c
j
c
i
hy
xxK
xxxxxx
xxK
1
'
1ˆ
1
ˆ (3)
In equation (3),
r
s
u
s
u
sj
u
si
uq
s s
c
sj
c
si
shxxl
h
xxwhK
11
1ˆ
ˆ,,ˆ
ˆ is the generalized kernel
function. The commonly used product kernel Kh is from Pagan and Ullah (1999),
where w is the standard normal product kernel function with window width hs =
hs(NT) associated with the sth
component of xc. The kernel function l
u is a variation of
Aitchison and Aitken (1976) kernel function which equals one if u
sj
u
si xx and u
s
otherwise.1
It is well known in the nonparametric literature that estimation of the bandwidths (h,
u) is crucial. N implements a number of ‘data-driven’ numerical algorithms to
determine the appropriate bandwidth or smoothing parameters for a given sample.
The paper uses the Least squares cross validation method as discussed in Racine and
Li (2004). Least squares cross validation selects h1, h2, … hq, u
1 , u
2 , … u
r to
minimize the following cross validation function:
n
i
iiii xMxmyCV1
2ˆ (4)
1 See for details of this estimation methodology in Li and Racine (2004) and Racine and Li (2004).
DP/16
9
Here, ii xmˆ = ./. KKy n
ill
n
il is the leave-one-out kernel estimate of m(xi)
and 0M(.)1 is a weight function. The purpose of M(.) is to avoid difficulties caused
by dividing by zero or by the slow convergence rate induced by boundary effects.
IV. Data and Empirical Model
A. Data
Utilizing the Li-Racine nonparametric estimation technique for mixed data, developed
by Li and Racine (2004) and Racine and Li (2004), our paper explores the
relationship between GDP per capita (GDPPC) and various indicators of financial
sector development and structure. The finance indicators are represented by several
macro variables such as, the share of GDP in private credit (pcby), share of GDP in
domestic credit (dcpy), market capitalization (smk), turnover ratio of stocks (str) and
total value of stocks traded (sva). The nonparametric technique of choice allows us to
examine the finance-economy relationship in a data driven specification free manner.
Our paper is based on 156 countries, of which 111 developing countries for the whole
sample. The discussion of results based on smk, str and sva consist of 110 countries
due to lack of data on three market-based variables across the years and countries.
The developing country lists also include 36 Least Developed Countries (LDCs) as
defined by United Nations.2 We obtained data from the UN sources and several
international and research institutions3.
B. Dependent and independent variables
Our main dependent variable is rpcGDP or real GDP per capita (international $, 2005
Constant Prices, Chain series) to identify level of economic performance at the cross-
country level. We use five variables to measure the size and nature of the financial
sector: pcby measures private credit held by banks, dcpy measures domestic credit
held by the private sector, smk is market capitalization of listed companies, str is the
turnover ratio of stocks traded and sva is the value of total stocks. All variables are
measured as a percent of GDP.
To control for non-financial variables, we use two more exogenous variables,
government’s expenditure (GC) and value of merchandise trade (TY), both also
measured as a % of GDP. We also use variables pr and ief to measure the strength of
institutional quality.
The Figure A.2 provides graphs of the nonparametric density estimates of various
financial quality indicators (dcpy, dcpy, smk, str and sva) and real per capita GDP. We
look at how the density functions shift between 1993 and 2011. These graphs plot the
values of the financial indicators along the x-axis and the frequency of countries
(proportion of countries in that range) along the y-axis. The results are mixed while
2 See Annex Table A1 for a complete list of countries.
3 See Annex Table A2 for data sources of the variables used in the paper.
Financial development and growth: Does one size fit all? Evidences from a nonparametric
10
comparing pdfs in 1993 and 2011; for some indicators (dcpy, dcpy, smk) we see the
density function shifts to the right, thus suggesting a greater fraction of countries in
our dataset have better quality of financial institutions. However, for others, (str and
sva) the density function does not move much or even shifts upwards. The density
function shifts to the right for the variable measuring real per capita GDP. Fewer
fraction of countries in our data are in the low income range between 1993 and 2011.
The descriptive statistics of the variables are reported in Table A.3.
C. The Empirical model
The objective is to examine the impact of the size and structure of financial sector
(measured by pcby, dcpy, smk, str and sva) on GDP per capita (rpcGDP). The main
model is a semi-log function converted to a nonparametric model represented by
equation (5). Here, m(.) is an unknown smooth function of the covariates, i are
unobserved country characteristics that are constant over time. This flexible
estimation strategy helps to avoid any functional form misspecification bias and
enables to explore the shape of the underlying relationship without superimposing any
a priori functional form restriction.
ln(rpcGDPit) = m(i, Fit, GCit, TYit) (5)
Fit = pcbyit,, dcpyit,, smkit, strit or svait
V. Results
This section discusses results for the core empirical model as represented by equation
(5), which has three main independent variables for a sample of 156 countries over
the period of 1993-2011. The three independent variables are the financial variable
(pcby/ dcpy/ smk/ str/ sva), government expenditures (GC) and value of trade (TY).
The results reported by dividing the data into various sub-groups of countries.
The dependent variable (rpcGDP) is measured in logs. All independent variables are
measures as percentage of GDP. All estimated coefficients are interpreted as a model
like in a semi log model. If the value of any x-variable changes by 1%, the real per
capita GDP (rpcGDP) changes by the coefficient multiplied by 100%.
As noted earlier, the nonparametric estimation technique gives an estimate of the
value of the regression function (the conditional moment) and its slope at every
country-time period combination. To help with the analysis and interpretation of
results, paper provides the slope estimates at the 25th
, 50th
and 75th
percentiles (labeled
quartiles 1, 2 and 3 or Q1, Q2 and Q3) and their standard errors obtained via
bootstrapping. The table also indicates which estimates are significant at the 90%,
95% or 99% confidence level.
In section A, the results reported for three group of countries, namely: regional
groupings as East Asia and Pacific, Europe and Central Asia, Latin America & the
Caribbean, Middle East and North Africa, South Asia, and Sub-Saharan Africa, as
well as for two income groups such as high income OECD countries and high income
DP/16
11
non OECD countries. The results are also shown for regional groupings based on the
United Nations Regional Commissions country classifications, namely: ESCAP,
ECA, ECE, ECLAC and ESCWA. We also discuss robustness checks of the results
for the model which then include two additional types of classification of results
based on the size of the financial sector and institutional quality.
In table 1, the subgroups are: (i) East Asia and Pacific, (ii) Europe and Central Asia,
(iii) Latin America & the Caribbean, (iv) Middle East and North Africa, (v) South
Asia, (vi) Sub-Saharan Africa, (vii) high income OECD countries and (viii) high
income non OECD countries. Classification is table 2 is based on United Nations
regional classifications (unrgcr).
In table 3, we divide the countries by size of the financial sector. Tables 4 and 5
divide the available countries on the basis of property rights (pr) and strength of
institutions (ief).
A. Core model results with regional classification
Table 1 displays the nonparametric estimates of the responsiveness of real GDP per
capita (rpcGDP) to changes in the financial variables (pcby/ dcpy/ smk/ str/ sva) for
various country groups. For countries in the East Asia and Pacific region, in 75% of
the observations, we find a positive relationship between economic growth
performance and the financial sector variables.
First, the results are shown for the East Asia and Pacific region, in columns 2 – 4, which
is based on the estimates in row titled pcby. In this region, as indicated by the 3rd
Quartile being equal to .017, in 75% of cases considered (in this case 34 countries across
7 time periods), during certain time periods, countries with similar sized governments
(GC) and degree of openness (TY) show a positive relationship between pcby and
economic growth. Specifically, in these cases, if share of GDP in private credit held by
banks or pcby increases by 1%, the per capital GDP increases by 1.7%. This relationship
is also statistically significant. In other words, the ceteris paribus relationship between
pcby and economic growth is positive significant for 75% of cases considered. Similarly,
from row titled GC, we can infer that the ceteris paribus relationship between GC and
economic growth is positive significant for 75% of cases, however the relationship is
significant negative for 50% of cases considered. And also from the row titled TY, the
ceteris paribus relationship between TY and economic growth is positive significant for
75% of cases. Unlike parametric results which provide a single global estimate for the
entire dataset, nonparametric methodology provides more heterogeneous results,
highlighting the differences across countries and country groups.
Second, the results for dcpy are similar as seen from columns 5 - 7. The ceterus paribus
relationship is significant positive for 75% of cases considered for the relationships
between economic growth and dcpy, economic growth and GC as well as economic
growth and TY.
Our general observation from the set of two parts under table 1 is a positive
relationship between the finance variables, especially the bank-based system and
growth. Almost all 3rd
quartile estimates are positive and significant. This observation
Financial development and growth: Does one size fit all? Evidences from a nonparametric
12
holds true for market-based system as show in the second panel of table 1 with
financial variables such as smk/ str/ sva. However, across regions, we find that the
size and significance of the results vary as well in descending order: Sub-Saharan
Africa, Middle—East and North Africa, East Asia and the Pacific, South Asia, East
Asia and the Pacific and Latin America and Caribbean in the case of pcby variable,
while the order for dcpy is the following: Sub-Saharan Africa, Europe and Central
Asia, South Asia, Middle—East and North Africa, East Asia and the Pacific and Latin
America and Caribbean.
For example, consider Table 1 (g) presenting the 3rd
Quartile estimates for high
income OECD countries. The table at the end of this set of tables presents a
comparison of the nonparametric estimates among various financial variables. We
find the magnitude of the estimates is the highest for pcby. A 1% increase in pcby will
increase real per capita GDP by 3.7%. At the same time, the magnitude is the lowest
for str; a 1% increase in str will increase per capita GDP by only 0.2%. All the
financial variables have a positive impact on economic growth. However, a better
banking system may have a stronger impact on growth compared with a larger stock
market.
B. Core model results with UN regional groupings
Table 2 presents medians, 1st quartiles and 3
rd quartiles of nonparametric estimates
categorized by the United Nations regional commissions classifications (unrgcr). The
theme continues of a ceteris paribus positive significant relationship between the
finance variables (pcby/ dcpy/ smk/ str/ sva) and economic growth for 75% of the
sample for all sub groups considered. Irrespective of the nature of classification, in
75% of the sample, the size of the financial sector has a positive impact on growth, for
countries with similar sized governments and degree of openness.
C. Core model results with size of financial sector
With the set of results under table 3, we try to evaluate the claims of Easterly and Stiglitz
(2000) and Cecchetti and Kharraoubi (2012) who find the positive relationship between
growth of the financial sector and GDP only under certain conditions. We particularly
want to see if a bloated financial system slows down the economy as claimed by
Cecchetti and Kharraoubi (2012). To do so, we divided the data into three subgroups
based on size of the financial sector.
For every financial variable (pcby/ dcpy/ smk/ str/ sva), we roughly sort the data in
increasing order of size of the variable and then divide it by three quartiles. The purpose
is to see if the positive finance-economy relationship is different for countries with the
small sized, medium sized and large sized financial sectors, or if some kind of U-shaped
relationship exits between size of the financial sector and economic growth. We find the
3rd
quartile is significant positive for all financial variables for all sub-groups in this
category. No matter what the size of the financial sector, we find the positive and
significant relationship does exist between the finance variables and growth
DP/16
13
D. Core model results with institutional quality
As mentioned earlier, the nonparametric methodology gives us a slope estimate for every
country in every year. Table 4 presents medians, 1st quartiles and 3
rd quartiles of
nonparametric estimates categorized by property rights or pr (weak to ideal), and table 5
provides the same based on institutional quality or ief (weak to ideal), which is a much
broader concept than the property rights. Again, the 3rd
quartile is positive significant for
all finance variables for all sub groups. The positive significant finance-economy
relationship holds irrespective of the strength or property rights or institutional quality.
In other words, even with different levels of institutions or property rights, a bigger
financial sector may have a transitory higher positive impact on growth.
E. Comparing estimates across bank-based versus market-based system
To compare the 3rd
quartile estimates of the relationship between the various financial
variables (pcby/ dcpy/ smk/ str/ sva) and growth, we add a comparison table 6 for each
of the groups presenting the 1st, 2
nd and 3
rd quartile estimates obtained from the
nonparametric methodology. In general we find the magnitude of the estimates for
pcby and dcpy are higher than the magnitude of the estimates for smk or str.
Here, we examine the impact of a growing financial system on growth or evaluate the
views of Easterly and Stiglitz (2000) as well as Cecchetti and Kharraoubi (2012)
about an inverted U relationship between financial development and growth. Table 7
presents a comparison table of the 3rd
quartile results across all financial variables and
the size of the financial sector. First we look at countries with small sized financial
sectors, i.e. pcby <15, dcpy < 19, smk < 30, str < 5 and sva < 1. The data are sorted
into three quartiles by magnitude of the various financial variables to determine the
cut off points. In this category of results, the banking sector has a significant positive
impact on growth. Increasing pcby or dcpy by 1% will increase per capita GDP by
3.7% or 3.8%. At the same time increasing smk or str by 1% will increase per capita
GDP by 1.4% or 1.3%. Total value of stocks traded also has a significant impact on
growth. Increasing sva by 1% will increase per capita GDP by 7.3%. The impact of
these financial variables on growth deteriorates as the size of the financial sector
expands. To understand the declining impact of a bloated financial sector in growth,
we look at countries with large sized financial sectors, i.e., pcby <15, dcpy < 19, smk
< 30, str < 5 and sva < 1. Here, all financial variables have a weaker impact on
growth. Increasing pcby or dcpy or smk or str or sva by 1% will increase per capita
GDP by only 0.8% or 0.9% or 0.3% or 0.2% or 0.45%, a much smaller magnitude
compared with the results for the sub group with smaller financial sector. A very large
financial sector may be too much of a good thing.
Now we examine the impact of property rights and institutions on the finance-growth
relationship. Table 8 compares the nonparametric 3rd
quartile estimates of the finance-
growth relationship for various sub groups of countries based on an index of property
rights (pr). For countries with low pr (0 < pr < 39), the impact of increasing pcby or
dcpy by 1% on real per capita GDP is 3.3% or 3.1%. The impact is much lower in
magnitude for stock market variables such as smk or str. However, if sva increases by
1%, real per capita GDP will increase by 5.1%. The magnitude of the impact is much
lower for the countries grouped together on the basis of high values of pr (70 < pr <
100). For this group, if pcby or dcpy increases by 1% then real per capita GDP rises
Financial development and growth: Does one size fit all? Evidences from a nonparametric
14
by 1.1% or 0.9%. The magnitudes of the impact on economic growth is much smaller
for str or sva. However, smk does have a strong impact of growth in this subgroup. If
smk increases by 1%, per capita GDP increases by 6%.
Table 9 conducts a similar comparison across financial variables by the value of the
index of institutional quality (ief). For the groups of countries with low values of ief
(i.e., 0 < ief < 39), pcby, dcpy and sva have a large impact on economic growth. In
this subgroup if pcby, dcpy or sva increase by 1%, per capita GDP will increase by
4.7%, 3.1% or 5.3%. However, the magnitude of this impact is much lower for the
subgroup of countries with high values of ief ((i.e., 70 < ief < 100). If pcby, dcpy or
sva increase by 1%, per capita GDP will increase by 0.7%, 0.8% or 1%.
Given the general relationship between well-defined property rights, good
institutional quality and income, one could claim, countries with high values of pr and
ief would on average have high incomes. It appears, financial variables have a low
impact on growth in these countries. Although the estimates may be statistically
significant, their economic significance is considerably low.
VI. Conclusions
The impact of finance variables on economic performance has enormous policy
implications for international institutions such as the United Nations to achieve the
newly adopted Sustainable Development Goals (SDGs). In this paper, we reassess the
relationship between two types of finance variables, bank-based (pcby/ dcpy) and
market-based (smk/ str/ sva) and GDP per capita by utilizing the Li-Racine
methodology.
We examine here a dataset of 156 countries over 1993-2011 time periods. There is
strong evidence of a statistically significant, positive impact of financial sectors on
GDP per capita. It’s worth noting that this positive significant trend is fairly uniform
for all subgroups by country characteristics, size of the financial sector, strength of
property rights and institutional quality. Flow of credit as well as functioning financial
markets are essential to support higher level of economic performance across
countries.
Our paper supports the view that countries with larger and better financial markets are
in a better position to reap benefits from trade integration and public policies. This
holds true irrespective of the initial size of the financial sector in the country as well
as the strength of property rights or institutional quality. The financial sector
deepening and its diversification are catalyst of economic growth by moving capital to
the highest value user without substantial risk of loss through moral hazard, adverse
selection, or transactions costs. Although some of the empirical literature supports this
argument, these studies establish correlation rather than causation.
In the long run, it is important that the financial structure is complete, stable,
transparent, diversified and able to offer financing through both banks and capital
markets, while being sufficiently adaptive to new developments. Policies aimed at
financial structure reform also depend on legal issues in creditor protection,
accounting standards in the particular country, and the stage of a country’s
development, as previously discussed.
DP/16
15
References
Acemoglu D., Johnson S., and Robinson J.A. (2001). The Colonial Origins of
Comparative Development: An Empirical Investigation. American Economic
Review, 91, December 2001. pp. 1369-1401.
Anderson, T.W. (1984). An Introduction to Multivariate Statistical Analysis, 2nd
ed.
New York: JohnWiley and Sons.
Basu, S.R., L.R. Klein, and A.L. Nagar (2005). Quality of life: comparing India and
China. Paper presented at Project LINK Meeting. UN Office, Geneva, 1
November.
Beck, T. and Levine R. (2000a). External dependence and industry growth - does
financial structure matter?. Mimeo, University of Minnesota, February 2000.
_________ (2000b). New Firm formation and industry growth: Does having a market
or bank-based system matter?. World Bank Working Paper No. 2383, May
2000.
_________ (2004). Stock markets, banks, and growth: Panel evidence. Journal of
Banking and Finance, 28(3), 423–442.
Beck, T., Levine, R., and Loayza, N. (2000). Finance and the sources of growth.
Journal of Financial Economics, 58(1-2), 261–300.
Beck, T., Demigruc- Kunt, A. & Levine, R. (1999). A New Database on Financial.
Development and Structure, World Bank Policy Research Working Paper No
2146
Beck, T., Büyükkarabacak, B., Rioja, F. K., Valev, N. T., 2012. Who gets the credit?
and does it matter? Household vs. firm lending across countries. The B.E.
Journal of Macroeconomics 12 (1), 1–46.
Benoît Cœuré (2014)
Bezemer, D., Grydaki, M., Zhang, L., 2014. Is Financial Development Bad for
Growth? Research Report 14016-GEM, University of Groningen, Research
Institute SOM
Carlin W. and C. Mayer. 2003. Finance, Investment and Growth. Journal of Financial
Economics. 69 (1). pp. 191–226.
Cecchetti, S.G., and E. Kharraoubi (2012). Reassessing the impact of finance on growth.
Working Papers, No. 381. Basel, Switzerland: Bank for International
Settlements.
Demirgüç-Kunt, Asli and Ross Levine (2001). Bank-based and Market-based Financial
Systems: Cross-Country Comparisons. In: Demirgüç-Kunt, Asli and Ross
Levine (Eds.):: Financial Structure and Economic Growth: A Cross-Country
Comparison of Banks, Markets, and Development, Cambridge, MA: MIT Press.
Easterly, W. and Levine, R. (2003). Tropics, Germs, and Crops: How Endowments
Influence Economic Development. Journal of Monetary Economics, 50(1), 3-39.
Easterly, I., and Stiglitz (2000). Shaken and stirred: explaining growth volatility.
Discussion Paper. Washington, D.C.: World Bank.
Financial development and growth: Does one size fit all? Evidences from a nonparametric
16
Estrada, Noland, Park and Ramayandi (2015). Financing Asia’s Growth. ADB
Economics Working Paper. Manila: Asian Development Bank.
Gambocorta, Yang and Tssatsaronis (2014). Financial structure and growth. BIS
Quarterly Review, March 2014. Page 21-35
King, Robert G., and Ross Levine. 1993. Finance and Growth: Schumpeter Might Be
Right. Quarterly Journal of Economics 108 (August) pp. 717–737.
Levine, R. (1997). Financial development and economic growth: views and agenda.
Journal of Economic Literature, vol. 35, No. 2 (June), pp. 688-726.
Levine, R., N. Loayza, and T. Beck (2000). Financial intermediation and growth:
causality and causes. Journal of Monetary Economics.
Levine, R., and S. Zervos (1998). Stock markets, banks and Economic growth.
American Economic Review, vol. 88, pp 537-558.
Li, Q., and J. Racine (2004). Cross-validated local linear nonparametric regression.
Statistica Sinica, vol. 14, No. 2, pp. 485-512.
N , Nonparametric Software by J. Racine. Available from
www.economics.mcmaster.ca/racine/.
Nagar, A.L., and S.R. Basu (2002). Weighting socio-economic variables of human
development: a latent variable approach. In Handbook of Applied
Econometrics and Statistical Inference, A. Ullah and others, eds. New York:
Marcel Dekker.
Pagan, A., and A. Ullah (1999). Nonparametric Econometrics. New York:
Cambridge University Press.
Racine, J., and Q. Li (2004). Nonparametric estimation of regression functions with
both categorical and continuous data. Journal of Econometrics, vol. 119, No.
1, pp. 99-130.
Rajan, R., and L. Zingales (1998). Financial dependence and growth. American
Economic Review, vol. 88, pp. 559-586.
Rodrik, D., Subramanian, A. and Trebbi, F. (2004). Institutions Rule: The Primacy of
Institutions Over Geography and Integration in Economic Development. Journal
of Economic Growth, 9(2), 131-165.
Sachs, Jeffrey D., 2003. Institutions Don’t Rule: Direct Effects of Geography on Per
Capita Income. NBER Working Paper 9490 (Cambridge, Massachusetts:
National Bureau of Economic Research).
Schumpeter, J.A. (1911). A Theory of Economic Development. Cambridge, MA:
Harvard University Press.
Silverman, B.W. (1986). Density Estimation for Statistics and Data Analysis. New
York: Chapman Hall.
DP/16
17
ANNEX TABLES
Table A.1. List of countries
Country
UN
Classification
UN Regional
Commissions Regional Classification Income Classification
Albania Transition ECE Europe and Central Asia Lower-middle-income economies
Algeria Developing ECA Middle East and North Africa Upper-middle-income economies
Antigua and
Barbuda SIDS ECLAC Latin America & the Caribbean Upper-middle-income economies
Argentina Developing ECLAC Latin America & the Caribbean Upper-middle-income economies
Armenia Transition ECE Europe and Central Asia Lower-middle-income economies
Australia Developed ESCAP High-income OECD members High-income OECD members
Austria Developed ECE High-income OECD members High-income OECD members
Azerbaijan Transition ECE Europe and Central Asia Upper-middle-income economies
Bahamas, The SIDS ECLAC High-income nonOECD members High-income nonOECD members
Bahrain SIDS ESCWA High-income nonOECD members High-income nonOECD members
Bangladesh LDCs ESCAP South Asia Low-income economies
Barbados SIDS ECLAC High-income nonOECD members High-income nonOECD members
Belarus Transition ECE Europe and Central Asia Upper-middle-income economies
Belgium Developed ECE High-income OECD members High-income OECD members
Belize SIDS ECLAC Latin America & the Caribbean Lower-middle-income economies
Benin LDCs ECA Sub-Saharan Africa Low-income economies
Bhutan LDCs ESCAP South Asia Lower-middle-income economies
Bolivia Developing ECLAC Latin America & the Caribbean Lower-middle-income economies
Botswana Developing ECA Sub-Saharan Africa Upper-middle-income economies
Brazil Developing ECLAC Latin America & the Caribbean Upper-middle-income economies
Brunei Darussalam Developing ESCAP High-income nonOECD members High-income nonOECD members
Bulgaria Developed ECE Europe and Central Asia Upper-middle-income economies
Burkina Faso LDCs ECA Sub-Saharan Africa Low-income economies
Burundi LDCs ECA Sub-Saharan Africa Low-income economies
Cambodia LDCs ESCAP East Asia and Pacific Low-income economies
Cameroon Developing ECA Sub-Saharan Africa Lower-middle-income economies
Canada Developed ECLAC High-income OECD members High-income OECD members
Central African
Republic LDCs ECA Sub-Saharan Africa Low-income economies
Chad LDCs ECA Sub-Saharan Africa Low-income economies
Chile Developing ECLAC Latin America & the Caribbean Upper-middle-income economies
China Developing ESCAP East Asia and Pacific Upper-middle-income economies
Colombia Developing ECLAC Latin America & the Caribbean Upper-middle-income economies
Comoros LDCs/SIDs ECA Sub-Saharan Africa Low-income economies
Congo, Dem. Rep. LDCs ECA Sub-Saharan Africa Low-income economies
Congo, Rep. Developing ECA Sub-Saharan Africa Lower-middle-income economies
Costa Rica Developing ECLAC Latin America & the Caribbean Upper-middle-income economies
Cote d'Ivoire Developing ECA Sub-Saharan Africa Lower-middle-income economies
Croatia Developed ECE High-income nonOECD members High-income nonOECD members
Cyprus Developed ECE High-income nonOECD members High-income nonOECD members
Czech Republic Developed ECE High-income OECD members High-income OECD members
Financial development and growth: Does one size fit all? Evidences from a nonparametric
18
Country
UN
Classification
UN Regional
Commissions Regional Classification Income Classification
Denmark Developed ECE High-income OECD members High-income OECD members
Djibouti LDCs ECA Middle East and North Africa Lower-middle-income economies
Dominica SIDS ECLAC Latin America & the Caribbean Upper-middle-income economies
Dominican Republic SIDS ECLAC Latin America & the Caribbean Upper-middle-income economies
Ecuador Developing ECLAC Latin America & the Caribbean Upper-middle-income economies
Egypt, Arab Rep. Developing ECA Middle East and North Africa Lower-middle-income economies
El Salvador Developing ECLAC Latin America & the Caribbean Lower-middle-income economies
Equatorial Guinea LDCs ECA High-income nonOECD members High-income nonOECD members
Eritrea LDCs ECA Sub-Saharan Africa Low-income economies
Estonia Developed ECE High-income OECD members High-income OECD members
Ethiopia LDCs ECA Sub-Saharan Africa Low-income economies
Fiji SIDS ESCAP East Asia and Pacific Lower-middle-income economies
Finland Developed ECE High-income OECD members High-income OECD members
France Developed ECE High-income OECD members High-income OECD members
Gabon Developing ECA Sub-Saharan Africa Upper-middle-income economies
Gambia, The LDCs ECA Sub-Saharan Africa Low-income economies
Georgia Transition ECE Europe and Central Asia Lower-middle-income economies
Germany Developed ECE High-income OECD members High-income OECD members
Ghana Developing ECA Sub-Saharan Africa Lower-middle-income economies
Greece Developed ECE High-income OECD members High-income OECD members
Grenada SIDS ECLAC Latin America & the Caribbean Upper-middle-income economies
Guatemala Developing ECLAC Latin America & the Caribbean Lower-middle-income economies
Guinea LDCs ECA Sub-Saharan Africa Low-income economies
Guyana Developing ECLAC Latin America & the Caribbean Lower-middle-income economies
Honduras Developing ECLAC Latin America & the Caribbean Lower-middle-income economies
Hong Kong SAR,
China Developing ESCAP High-income nonOECD members High-income nonOECD members
Hungary Developed ECE High-income OECD members High-income OECD members
Iceland Developed ECE High-income OECD members High-income OECD members
India Developing ESCAP South Asia Lower-middle-income economies
Indonesia Developing ESCAP East Asia and Pacific Lower-middle-income economies
Iran, Islamic Rep. Developing ESCAP Middle East and North Africa Upper-middle-income economies
Israel Developing ECE High-income OECD members High-income OECD members
Italy Developed ECE High-income OECD members High-income OECD members
Japan Developed ESCAP High-income OECD members High-income OECD members
Jordan Developing ESCWA Middle East and North Africa Upper-middle-income economies
Kazakhstan Transition ECE Europe and Central Asia Upper-middle-income economies
Kenya Developing ECA Sub-Saharan Africa Low-income economies
Korea, Rep. Developing ESCAP High-income OECD members High-income OECD members
Kuwait Developing ESCWA High-income nonOECD members High-income nonOECD members
Kyrgyz Republic Transition ECE Europe and Central Asia Low-income economies
Lao PDR LDCs ESCAP East Asia and Pacific Lower-middle-income economies
Latvia Developed ECE Europe and Central Asia Upper-middle-income economies
Lebanon Developing ESCWA Middle East and North Africa Upper-middle-income economies
Lesotho LDCs ECA Sub-Saharan Africa Lower-middle-income economies
Libya Developing ECA Middle East and North Africa Upper-middle-income economies
DP/16
19
Country
UN
Classification
UN Regional
Commissions Regional Classification Income Classification
Lithuania Developed ECE Europe and Central Asia Upper-middle-income economies
Macao SAR, China Developing ESCAP High-income nonOECD members High-income nonOECD members
Macedonia, FYR Transition ECE Europe and Central Asia Upper-middle-income economies
Madagascar LDCs ECA Sub-Saharan Africa Low-income economies
Malawi LDCs ECA Sub-Saharan Africa Low-income economies
Malaysia Developing ESCAP East Asia and Pacific Upper-middle-income economies
Mali LDCs ECA Sub-Saharan Africa Low-income economies
Malta Developed ECE High-income nonOECD members High-income nonOECD members
Mauritius SIDS ECA Sub-Saharan Africa Upper-middle-income economies
Mexico Developing ECLAC Latin America & the Caribbean Upper-middle-income economies
Moldova Transition ECE Europe and Central Asia Lower-middle-income economies
Mongolia Developing ESCAP East Asia and Pacific Lower-middle-income economies
Morocco Developing ECA Middle East and North Africa Lower-middle-income economies
Mozambique LDCs ECA Sub-Saharan Africa Low-income economies
Namibia Developing ECA Sub-Saharan Africa Upper-middle-income economies
Nepal LDCs ESCAP South Asia Low-income economies
Netherlands Developed ECE High-income OECD members High-income OECD members
New Zealand Developed ESCAP High-income OECD members High-income OECD members
Nicaragua Developing ECLAC Latin America & the Caribbean Lower-middle-income economies
Niger LDCs ECA Sub-Saharan Africa Low-income economies
Nigeria Developing ECA Sub-Saharan Africa Lower-middle-income economies
Norway Developed ECE High-income OECD members High-income OECD members
Oman Developing ESCWA High-income nonOECD members High-income nonOECD members
Pakistan Developing ESCAP South Asia Lower-middle-income economies
Panama Developing ECLAC Latin America & the Caribbean Upper-middle-income economies
Papua New Guinea SIDS ESCAP East Asia and Pacific Lower-middle-income economies
Paraguay Developing ECLAC Latin America & the Caribbean Lower-middle-income economies
Peru Developing ECLAC Latin America & the Caribbean Upper-middle-income economies
Philippines Developing ESCAP East Asia and Pacific Lower-middle-income economies
Poland Developed ECE High-income OECD members High-income OECD members
Portugal Developed ECE High-income OECD members High-income OECD members
Russian Federation Transition ESCAP Europe and Central Asia Upper-middle-income economies
Rwanda LDCs ECA Sub-Saharan Africa Low-income economies
Saudi Arabia Developing ESCWA High-income nonOECD members High-income nonOECD members
Senegal LDCs ECA Sub-Saharan Africa Lower-middle-income economies
Sierra Leone LDCs ECA Sub-Saharan Africa Low-income economies
Singapore Developing ESCAP High-income nonOECD members High-income nonOECD members
Slovak Republic Developed ECE High-income OECD members High-income OECD members
Slovenia Developed ECE High-income OECD members High-income OECD members
Solomon Islands LDCs/SIDs ESCAP East Asia and Pacific Lower-middle-income economies
South Africa Developing ECA Sub-Saharan Africa Upper-middle-income economies
Spain Developed ECE High-income OECD members High-income OECD members
Sri Lanka Developing ESCAP South Asia Lower-middle-income economies
St. Kitts and Nevis SIDS ECLAC High-income OECD members High-income OECD members
St. Lucia SIDS ECLAC Latin America & the Caribbean Upper-middle-income economies
Financial development and growth: Does one size fit all? Evidences from a nonparametric
20
Country
UN
Classification
UN Regional
Commissions Regional Classification Income Classification
St. Vincent and the
Grenadines SIDS ECLAC Latin America & the Caribbean Upper-middle-income economies
Sudan LDCs ECA Sub-Saharan Africa Lower-middle-income economies
Swaziland Developing ECA Sub-Saharan Africa Lower-middle-income economies
Sweden Developed ECE High-income OECD members High-income OECD members
Switzerland Developed ECE High-income OECD members High-income OECD members
Syrian Arab
Republic Developing ESCWA Middle East and North Africa Lower-middle-income economies
Tajikistan Transition ECE Europe and Central Asia Low-income economies
Tanzania LDCs ECA Sub-Saharan Africa Low-income economies
Thailand Developing ESCAP East Asia and Pacific Upper-middle-income economies
Togo LDCs ECA Sub-Saharan Africa Low-income economies
Tonga SIDS ESCAP East Asia and Pacific Lower-middle-income economies
Trinidad and Tobago SIDS ECLAC High-income nonOECD members High-income nonOECD members
Tunisia Developing ECA Middle East and North Africa Upper-middle-income economies
Turkey Developing ECE Europe and Central Asia Upper-middle-income economies
Uganda LDCs ECA Sub-Saharan Africa Low-income economies
Ukraine Transition ECE Europe and Central Asia Lower-middle-income economies
United Arab
Emirates Developing ESCWA High-income nonOECD members High-income nonOECD members
United Kingdom Developed ECE High-income OECD members High-income OECD members
United States Developed ECLAC High-income OECD members High-income OECD members
Uruguay Developing ECLAC Latin America & the Caribbean Upper-middle-income economies
Vanuatu LDCs/SIDs ESCAP East Asia and Pacific Lower-middle-income economies
Venezuela, RB Developing ECLAC Latin America & the Caribbean Upper-middle-income economies
Vietnam Developing ESCAP East Asia and Pacific Lower-middle-income economies
Yemen, Rep. LDCs ESCWA Middle East and North Africa Lower-middle-income economies
Zambia LDCs ECA Sub-Saharan Africa Lower-middle-income economies
Zimbabwe Developing ECA Sub-Saharan Africa Low-income economies
Source: United Nations, and classification is based on the World Bank
DP/16
21
Table A.2. Description and sources of variables
Variable/code Description Source
rpcGDP GDP per capita (constant 2005 US$) UN
pcby Private credit by deposit money banks to GDP
(%) IMF/World Bank
dcpy Domestic credit to private sector (% of GDP) IMF/World Bank
smk Market capitalization of listed companies (% of
GDP)
IMF/World Bank
str Turnover ratio of stocks traded (% of GDP) IMF/World Bank
sva Total stock value traded IMF/World Bank
GC General government final consumption
expenditure (% of GDP) World Bank
TY Merchandise trade (% of GDP) UNCTAD
pr Property rights Heritage Foundation
ief Institutions classification Heritage Foundation
Note: All variables are converted in logs, denoted by ‘ln’ in the text, tables and figures.
Table A.3. Descriptive statistics
ln(rpcGDP) pcby dcpy smk str sva GC TY
Mean 8.02 42.32 46.72 46.15 47.06 26.13 15.81 65.88
Std Dev 1.59 40.65 44.48 58.92 72.15 53.48 6.01 41.26
Min 4.79 0.12 0.83 0.03 0.02 .002 2.92 8.62
Max 11.11 272.7 282.7 539.33 605.02 673 55.87 358.4
N 1001 1001 1092 677 700 700 1001 1001
Financial development and growth: Does one size fit all? Evidences from a nonparametric
22
Figure A.2. List of Graphs
0
.005
.01
.015
.02
0 100 200 300x
kdensity pcby_1993 kdensity pcby_2011
(1993 vs 2011)
pcby density estimates
0
.005
.01
.015
.02
0 100 200 300x
kdensity dcpy_1993 kdensity dcpy_2011
(1993 vs 2011)
dcpy density estimates0
.005
.01
.015
.02
.025
0 100 200 300 400x
kdensity smk1993 kdensity smk2011
(1993 vs 2011)
smk density estimates
0
.005
.01
.015
0 200 400 600x
kdensity str_1993 kdensity str_2011
(1993 vs 2011)
str density estimates
0
.02
.04
.06
.08
0 200 400 600 800x
kdensity sva_1993 kdensity sva_2011
(1993 vs 2011)
sva density estimates
0
.00002
.00004
.00006
0 20000 40000 60000x
kdensity rpcGDP_1993 kdensity rpcGDP_2011
(1993 vs 2011)
rpcGDP density estimates
DP/16
23
Table 1. Nonparametric slope coefficients by region
(a) Estimates for East Asia and Pacific
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd Quartile 1
st Quartile Median 3
rd Quartile
pcby -0.0007
(0.0006)
0.005
(0.003)
.017***
(0.002)
dcpy
.002***
(.0004)
.007**
(.002)
.014***
(.001)
GC -0.042**
(0.019)
-.013*
(.007)
0.029**
(0.014)
-.01
(.009)
-.0005
(.003)
.08***
(.02)
TY 0.0006
(0.0008)
0.003***
(0.0006)
0.006***
(0.001)
.001***
(.0005)
.003***
(.0002)
.008***
(.001)
Dependent
variable
ln(rpcGDP)
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
smk .000
(.000)
.002***
(.000)
.009***
(.002)
str
-.003**
(.001)
.001
(.001)
.004***
(.001)
sva
.000
(.001)
.005**
(.002)
.027*
(.016)
GC -.048***
(.016)
-.001
(.009)
.033
(.024)
-.059***
(.009)
-.014
(.012)
.057***
(.016)
-.035**
(.017)
.008
(.010)
.063***
(.011)
TY -.005***
(.001)
.002
(.001)
.007***
(.001)
-.004***
(.001)
.003***
(.000)
.007***
(.001)
-.002*
(.001)
.003***
(.001)
.008***
(.001)
(b) Estimates for Europe and Central Asia
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd Quartile 1
st Quartile Median 3
rd Quartile
pcby 0.004***
(0.001)
0.012***
(0.002)
0.021***
(0.003)
dcpy
.006***
(.001)
.015***
(.001)
.024***
(.001)
GC -0.080***
(0.014)
-0.019**
(0.008)
0.020**
(0.009)
-.059***
(.005)
-.024***
(.002)
.001
(.003)
TY -0.004
(0.003)
0.006***
(0.0009)
0.012***
(0.001)
-.008***
(.002)
.005***
(.0007)
.008***
(.0009)
Dependent
variable
ln(rpcGDP)
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
smk .006***
(.001)
.012***
(.002)
.032***
(.008)
str
-.007
(.004)
-.000*
(.000)
.005**
(.002)
sva
.014***
(.003)
.042***
(.004)
.063***
(.004)
GC .080***
(.025)
.007
(.015)
.041***
(.015)
-.11***
(.02)
-.024***
(.009)
.016
(.011)
-.077***
(.021)
-.022**
(.010)
.023**
(.010)
TY -.004
(.006)
.004***
(.001)
.011***
(.001)
-.002
(.006)
.009***
(.001)
.018***
(.005)
-.003
(.002)
.005***
(.001)
.014***
(.001)
Financial development and growth: Does one size fit all? Evidences from a nonparametric
24
(c) Estimates for Latin America & the Caribbean
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd Quartile 1
st Quartile Median 3
rd Quartile
pcby -0.001
(.001)
0.004***
(.0008)
0.01***
(.001)
dcpy
.005***
(.0006)
.0088***
(.0003)
.013***
(.0005)
GC -0.025***
(.003)
0.001
(.002)
0.018***
(.005)
-.025***
(.004)
.006
(.003)
.025***
(.005)
TY -0.002***
(.0006)
0.001**
(.0009)
0.007***
(.0009)
-.002***
(.0007)
.003***
(.0004)
.007***
(.0008)
Dependent
variable
ln(rpcGDP)
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
smk -.002**
(.001)
.002***
(.000)
.005***
(.001)
str
.000
(.000)
.003***
(.000)
.008***
(.002)
sva
.000
(.003)
.007***
(.001)
.025***
(.004)
GC -.006
(.007)
.021***
(.005)
.075***
(.009)
-.003
(.004)
.045***
(.009)
.102***
(.016)
-.000
(.004)
.030***
(.007)
.075***
(.011)
TY .001**
(.000)
.004***
(.000)
.009***
(.001)
-.000
(.001)
.007***
(.001)
.016***
(.001)
.003*
(.001)
.007***
(.000)
.013***
(.001)
(d) Estimates for Middle East and North Africa
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd Quartile 1
st Quartile Median 3
rd Quartile
pcby .006***
(.001)
.011***
(.002)
.027***
(.006)
dcpy
.005***
(.001)
.007***
(.001)
.017***
(.002)
GC -.035***
(.006)
-.013*
(.008)
.009
(.020)
-.027***
(.008)
-.011*
(.006)
.014**
(.006)
TY -.002**
(.001)
.002
(.002)
.013***
(.0009)
.0004
(.001)
.004*
(.002)
.012***
(.001)
Dependent
variable
ln(rpcGDP)
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
smk -.003*
(.001)
.001*
(.000)
.005***
(.001)
str
-.016***
(.004)
-.001
(.002)
.006**
(.002)
sva
.000
(.003)
.012***
(.004)
.031***
(.004)
GC -.087
(.058)
-.020*
(.012)
.074*
(.041)
-.062***
(.017)
-.012
(.023)
.14**
(.05)
-.1***
(.029)
-.009
(.029)
.12***
(.036)
TY -.003
(.003)
-.000
(.001)
.008**
(.003)
-.004***
(.001)
.000
(.002)
.016**
(.007)
-.01***
(.004)
-.001
(.001)
.003
(.003)
DP/16
25
(e) Estimates for South Asia
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd Quartile 1
st Quartile Median 3
rd Quartile
pcby .011***
(.003)
.016***
(.001)
.019***
(.0009)
dcpy
.006**
(.002)
.016***
(.001)
.018***
(.0009)
GC -.017
(.012)
.024**
(.01)
.063***
(.008)
.048**
(.019)
.062***
(.012)
.11***
(.009)
TY .001
(.002)
.006***
(.001)
.017***
(.005)
.002
(.002)
.003***
(.0004)
.015***
(.002)
Dependent
variable
ln(rpcGDP)
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
smk -.008**
(.003)
-.001
(.002)
.01**
(.005)
str
.000
(.000)
.001**
(.000)
.003
(.003)
sva
-.001
(.0010
.005
(.005)
.038***
(.013)
GC .000
(.033)
.059***
(.007)
0.11***
(.031)
-.039**
(.017)
.039
(.034)
0.15***
(.03)
.040
(.037)
.084***
(.011)
.127***
(.020)
TY .002
(.004)
.017***
(.004)
0.03***
(.008)
-.000
(.003)
.013***
(.004)
.023***
(.003)
.009***
(.002)
.023***
(.008)
.041***
(.002)
(f) Estimates for Sub-Saharan Africa
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd Quartile 1
st Quartile Median 3
rd Quartile
pcby .008***
(.001)
.022***
(.002)
.038***
(.001)
dcpy
.016***
(.001)
.026***
(.001)
.039***
(.001)
GC -.03***
(.003)
-.007**
(.003)
.014***
(.002)
-.017***
(.002)
-.003
(.001)
.019***
(.002)
TY -.001*
(.0008)
.003***
(.0004)
.007***
(.0007)
-.0007
(.001)
.003***
(.0003)
.009***
(.0006)
Dependent
variable
ln(rpcGDP)
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
smk .000
(.000)
.002*
(.001)
.013***
(.003)
str
-.000
(.001)
.008***
(.001)
.015***
(.001)
sva
.037***
(.004)
.057***
(.005)
.083***
(.003)
GC -.029***
(.006)
-.007
(.005)
.014*
(.008)
-.023***
(.007)
-.009*
(.005)
.019***
(.003)
-.035***
(.007)
.001
(.005)
.022***
(.003)
TY -.003***
(.001)
.001
(.001)
.008***
(.001)
-.005***
(.001)
.003**
(.001)
.009***
(.001)
-.008***
(.002)
.000
(.001)
.008***
(.002)
Financial development and growth: Does one size fit all? Evidences from a nonparametric
26
(g) Estimates for high income OECD countries
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd Quartile 1
st Quartile Median 3
rd Quartile
pcby -.037***
(.01)
-.004
(.005)
.037***
(.011)
dcpy
.001***
(.000)
.005***
(.000)
.008***
(.000)
GC -.004**
(.001)
.001*
(.001)
.008***
(.002)
-.025***
(.003)
.022***
(.008)
.053***
(.004)
TY -.002***
(.000)
.000**
(.000)
.003**
(.001)
-.006***
(.001)
.002***
(.000)
.006***
(.000)
Dependent
variable
ln(rpcGDP)
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
smk -.000
(.000)
.002***
(.000)
.005***
(.000)
str
-.000*
(.000)
.000***
(.000)
.002***
(.000)
sva
.000*
(.000)
.003***
(.000)
.007***
(.000)
GC -.017***
(.005)
.015***
(.004)
.059***
(.007)
-.041***
(.008)
.005
(.003)
.038***
(.007)
-.017**
(.008)
.020***
(.005)
.085***
(.010)
TY -.004***
(.001)
.004***
(.000)
.008***
(.000)
-.005***
(.001)
.003***
(.000)
.007***
(.000)
-.014***
(.001)
-.001
(.001)
.006***
(.001)
(h) Estimates for high income non OECD countries
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd Quartile 1
st Quartile Median 3
rd Quartile
pcby -.000
(.001)
.003***
(.000)
.009***
(.002)
dcpy
.001
(.001)
.005***
(.000)
.010***
(.001)
GC -.051***
(.007)
-.011**
(.004)
-.000
(.002)
-.040***
(.004)
-.013
(.008)
.012**
(.005)
TY -.003**
(.001)
.000
(.000)
.004***
(.001)
-.001
(.001)
.003***
(.000)
.006***
(.001)
Dependent variable
ln(rpcGDP)
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
smk .000
(.000)
.002***
(.000)
.003***
(.000)
str
-.002***
(.000)
.000**
(.000)
.003**
(.001)
sva
.000
(.000)
.006***
(.001)
.020***
(.002)
GC -.029***
(.008)
-.008
(.006)
.017
(.014)
-.037***
(.009)
-.004
(.004)
.037***
(.011)
-.032**
(.014)
.003
(.008)
.048***
(.013)
TY -.002***
(.000)
.000
(.001)
.006***
(.002)
-.004**
(.002)
.001*
(.001)
.008***
(.002)
-.007***
(.001)
.000
(.001)
.005***
(.001)
All standard error are in parentheses and are obtained via bootstrapping
* indicates significance at 10% level
** indicates significance at 5% level
*** indicates significance at 1% level
DP/16
27
Table 2. Nonparametric slope coefficients by UN regional commissions
(a) Estimates for unrgcr = ECA
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd Quartile 1
st Quartile Median 3
rd Quartile
pcby .006***
(.0009)
.018***
(.002)
.037***
(.002)
dcpy
.01***
(.002)
.02***
(.001)
.037***
(.001)
GC -.031***
(.003)
-.006**
(.003)
.014***
(.002)
-.016***
(.001)
-.002*
(.001)
.02***
(.001)
TY -.0008
(.0008)
.004***
(.0004)
.010***
(.001)
-.0001
(.0008)
.003***
(.0003)
.01***
(.0006)
Dependent
variable
ln(rpcGDP)
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
smk .000
(.000)
.002*
(.001)
.011***
(.002)
str
-.002
(.002)
.006***
(.001)
.014***
(.001)
sva
.027***
(.004)
.051***
(.004)
.080***
(.003)
GC -.029***
(.005)
.001
(.006)
.034**
(.015)
-.023***
(.006)
.001
(.006)
.033***
(.008)
-.032***
(.008)
.006
(.006)
.037***
(.010)
TY -.003*
(.001)
.001
(.001)
.009***
(.001)
-.005***
(.001)
.003**
(.001)
.010***
(.001)
-.006***
(.002)
.000
(.000)
.009***
(.001)
(b) Estimates for unrgcr = ECE
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd Quartile 1
st Quartile Median 3
rd Quartile
pcby .002***
(.0002)
.005***
(.0004)
.01***
(.001)
dcpy
.002***
(.0002)
.006***
(.0003)
.013***
(.0009)
GC -.058***
(.003)
-.021***
(.006)
.02***
(.003)
-.027***
(.002)
-.002
(.004)
.045***
(.003)
TY -.001*
(.0007)
.003***
(.0004)
.009***
(.001)
-.004***
(.001)
.004***
(.0005)
.007***
(.0003)
Dependent
variable
ln(rpcGDP)
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
smk .000
(.000)
.003***
(.000)
.010***
(.000)
str
-.001***
(.000)
.000
(.000)
.002***
(.000)
sva
.000
(.000)
.005***
(.001)
.032***
(.005)
GC -.018***
(.006)
.017***
(.005)
.062***
(.006)
-.047***
(.007)
.001
(.004)
.037***
(.004)
-.036***
(.005)
.013**
(.005)
.077***
(.012)
TY -.001
(.001)
.004***
(.000)
.008***
(.000)
-.002*
(.001)
.005***
(.000)
.010***
(.000)
-.010***
(.001)
.001**
(.001)
.008***
(.000)
Financial development and growth: Does one size fit all? Evidences from a nonparametric
28
(c) Estimates for unrgcr = ECLAC
Dependent
variable
lp(rpcGDP)
1st Quartile Median 3
rd Quartile 1
st Quartile Median 3
rd Quartile
pcby .002***
(.0002)
.005***
(.0004)
.01***
(.001)
dcpy
.004***
(.0003)
.008***
(.0005)
.013***
(.0007)
GC -.058***
(.003)
-.021***
(.006)
.02***
(.003)
-.031***
(.002)
-.002
(.003)
.023***
(.001)
TY -.001*
(.0007)
.003***
(.0004)
.009***
(.001)
-.002***
(.0008)
.002***
(.0004)
.007***
(.0009)
Dependent
variable
ln(rpcGDP)
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
smk -.001**
(.000)
.002***
(.000)
.005***
(.000)
str
.000
(.000)
.003***
(.000)
.007***
(.001)
sva
.000
(.001)
.007***
(.001)
.020***
(.003)
GC -.023**
(.012)
.017***
(.003)
.071***
(.010)
-.009
(.007)
.028**
(.012)
.073***
(.012)
-.004
(.004)
.026***
(.005)
.067***
(.007)
TY .000
(.001)
.004***
(.000)
.009***
(.000)
-.003**
(.001)
.005***
(.001)
.014***
(.0010
-.000
(.001)
.006***
(.000)
.013***
(.001)
(d) Estimates for unrgcr = ESCAP
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd Quartile 1
st Quartile Median 3
rd Quartile
pcby .0004
(.0009)
.012***
(.001)
.019***
(.0009)
dcpy
.002***
(.0004)
.007***
(.001)
.017***
(.0009)
GC -.056***
(.009)
-.014**
(.006)
.019***
(.007)
-.035***
(.002)
.003
(.002)
.053***
(.011)
TY -.004**
(.001)
.002***
(.0004)
.006***
(.0007)
-.001
(.002)
.003***
(.0002)
.007***
(.001)
Dependent
variable
ln(rpcGDP)
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
smk -.000
(.001)
.001***
(.000)
.006***
(.001)
str
-.000
(.000)
.001***
(.000)
.004***
(.001)
sva
.000
(.000)
.006***
(.001)
.020***
(.003)
GC -.044***
(.009)
.010
(.008)
.066***
(.012)
-.060***
(.007)
-.003
(.010)
.062***
(.009)
-.025**
(.011)
.029***
(.007)
.084***
(.010)
TY -.004***
(.001)
.002***
(.001)
.009***
(.001)
-.005***
(.000)
.001**
(.000)
.007***
(.002)
-.003***
(.001)
.002**
(.000)
.010***
(.002)
DP/16
29
(e) Estimates for unrgcr = ESCWA
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd Quartile 1
st Quartile Median 3
rd Quartile
pcby .001
(.002)
.012***
(.002)
.02***
(.004)
dcpy
.004**
(.002)
.01***
(.001)
.017***
(.001)
GC -.03***
(.011)
-.009*
(.005)
.003
(.003)
-.03*
(.016)
-.011
(.009)
.02***
(.007)
TY -.002***
(.0006)
.0006
(.0006)
.001
(.001)
.0008
(.0007)
.004***
(.0003)
.006
(.006)
Dependent
variable
ln(rpcGDP)
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
smk -.000
(.000)
.001***
(.000)
.002***
(.000)
str
-.002**
(.001)
-.000
(.000)
.001***
(.000)
sva
-.000
(.001)
.004***
(.001)
.010**
(.004)
GC -.045***
(.018)
-.015***
(.005)
.003
(.004)
-.040***
(.007)
-.017***
(.004)
.002
(.004)
-.089***
(.018)
-.015
(.010)
.005
(.012)
TY -.003
(.002)
-.000
(.001)
.007**
(.003)
-.001
(.001)
.000
(.001)
.010***
(.003)
-.012***
(.003)
-.003*
(.001)
.002
(.002)
All standard error are in parentheses and are obtained via bootstrapping
* indicates significance at 10% level
** indicates significance at 5% level
*** indicates significance at 1% level
Financial development and growth: Does one size fit all? Evidences from a nonparametric
30
Table 3. Nonparametric slope coefficients by size of the financial sector
(roughly arranging by quartiles)
(a) Estimates for Countries with a Small Sized Financial Sector
pcby <15 dcpy<19
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd Quartile 1
st Quartile Median 3
rd Quartile
pcby .006***
(.001)
.02***
(.002)
.037***
(.001)
dcpy
.014***
(.001)
.024***
(.0008)
.038***
(.001)
GC -.034***
(.005)
-.003
(.002)
.018***
(.002)
-.024***
(.002)
-.005***
(.001)
.018***
(.002)
TY -.002***
(.0008)
.003***
(.0005)
.01***
(.001)
-.002***
(.0009)
.003***
(.0005)
.008***
(.0006)
smk <30 str < 5 sva < 1
Dependent
variable
ln(rpcGDP)
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
smk -.000
(.000)
.004***
(.000)
.014***
(.001)
str
-.003*
(.002)
.004***
(.001)
.013***
(.001)
sva
.012***
(.003)
.042***
(.004)
.073***
(.004)
GC -.02***
(.006)
.007***
(.002)
.055***
(.007)
-.023***
(.006)
.003
(.003)
.037***
(.010)
-.034***
(.007)
.005
(.003)
.041***
(.007)
TY -.000
(.000)
.004***
(.000)
.008***
(.000)
-.004***
(.001)
.002***
(.000)
.009***
(.001)
-.004***
(.001)
.003***
(.001)
.010***
(.001)
(b) Estimates for Countries with a Medium Sized Financial Sector
15 < pcby <50 19<dcpy<50
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd Quartile 1
st Quartile Median 3
rd Quartile
pcby .002***
(.0007)
.009***
(.0008)
.017***
(.0009)
dcpy
.003***
(.0004)
.009***
(.0007)
.015***
(.0006)
GC -.04***
(.003)
-.008***
(.003)
.016***
(.002)
-.026***
(.003)
-.001
(.002)
.03***
(.003)
TY -.0009**
(.0004)
.003***
(.0005)
.009***
(.0009)
.0002
(.0007)
.004***
(.0005)
.01***
(.0006)
DP/16
31
30<smk <60 5 < str < 20 1 < sva < 4.5
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd
Quartile
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
smk -.000
(.000)
.002***
(.000)
.004***
(.000)
str
-.003*
(.001)
.001**
(.000)
.009***
(.001)
sva
.005***
(.002)
.026***
(.002)
.045***
(.002)
GC -.044***
(.008)
.006
(.006)
.045***
(.01)
-.049***
(.006)
-.007
(.0050
.049***
(.012)
-.045***
(.006)
.003
(.006)
.085***
(.010)
TY -.005***
(.001)
.002**
(.001)
.011***
(.001)
-.005***
(.001)
.003***
(.001)
.012***
(.002)
-.004***
(.001)
.004***
(.001)
.011***
(.000)
(c) Estimates for Countries with a Large Sized Financial Sector
pcby > 50
dcpy>50
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd Quartile 1
st Quartile Median 3
rd Quartile
pcby .001***
(.0003)
.004***
(.0005)
.008***
(.0005)
dcpy
.002***
(.0002)
.005***
(.0003)
.009***
(.0004)
GC -.05***
(.003)
-.019***
(.004)
.012***
(.003)
-.03***
(.001)
.006*
(.003)
.048***
(.003)
TY -.003***
(.0006)
.002***
(.0002)
.006***
(.0003)
-.003**
(.001)
.002***
(.0002)
.005***
(.0005)
smk > 60 str > 20 sva > 4.5
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd
Quartile
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
smk -.000
(.000)
.001***
(.000)
.003***
(.000)
str
-.000***
(.000)
.000***
(.000)
.002***
(.000)
sva
.005**
(.002)
.026***
(.002)
.045***
(.002)
GC -.026***
(.008)
.009***
(.003)
.059***
(.008)
-.045***
(.006)
.004
(.003)
.050***
(.007)
-.045***
(.006)
.003
(.007)
.085***
(.011)
TY -.004***
(.001)
.002***
(.000)
.009***
(.001)
-.003***
(.000)
.004***
(.000)
.010***
(.000)
-.004***
(.001)
.004***
(.001)
.011***
(.000)
All standard error are in parentheses and are obtained via bootstrapping
* indicates significance at 10% level
** indicates significance at 5% level
*** indicates significance at 1% level
Financial development and growth: Does one size fit all? Evidences from a nonparametric
32
Table 4. Non parametric slope estimates by property rights (weak to ideal)
(a) Estimates for countries with low pr (0 – 39)
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd Quartile 1
st Quartile Median 3
rd Quartile
pcby .005***
(.000)
.016***
(.001)
.033***
(.002)
dcpy
.007***
(.001)
.018***
(.001)
.031***
(.001)
GC -.050***
(.006)
-.014***
(.006)
.012***
(.004)
-.033***
(.002)
-.011***
(.001)
.013***
(.004)
TY -.001*
(.000)
.004***
(.000)
.009***
(.001)
-.001*
(.001)
.003***
(.000)
.009***
(.000)
Dependent
variable
ln(rpcGDP)
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
smk -.000
(.000)
.005***
(.001)
.016***
(.002)
str
-.001
(.001)
.002***
(.000)
.010***
(.001)
sva
.004*
(.002)
.022***
(.003)
.051***
(.003)
GC -.045***
(.008)
.001
(.007)
.074***
(.008)
-.063***
(.008)
-.010*
(.005)
.090***
(.028)
-.057***
(.008)
-.007
(.007)
.063***
(.012)
TY -.001
(.001)
.004***
(.000)
.010***
(.000)
-.004***
(.001)
.004***
(.000)
.011***
(.001)
-.002*
(.001)
.006***
(.000)
.013***
(.001)
(b) Estimates for countries with medium pr (40 – 69)
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd Quartile 1
st Quartile Median 3
rd Quartile
pcby .001*
(.000)
.008***
(.000)
.019***
(.001)
dcpy
.003***
(.000)
.009***
(.000)
.018***
(.000)
GC -.032***
(.004)
.001
(.001)
.021***
(.004)
-.014***
(.003)
.009***
(.002)
.030***
(.002)
TY -.001**
(.000)
.004***
(.000)
.010***
(.001)
-.000
(.001)
.004***
(.000)
.009***
(.000)
Dependent
variable
ln(rpcGDP)
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
smk -.031***
(.005)
.000
(.004)
.044***
(.006)
str
-.002***
(.000)
.001**
(.000)
.007***
(.001)
sva
.003***
(.001)
.019***
(.003)
.048***
(.003)
GC -.002***
(.000)
.002***
(.000)
.007***
(.000)
-.031***
(.004)
.002
(.004)
.051***
(.008)
-.023***
(.006)
.009***
(.003)
.055**
(.005)
TY -.000
(.000)
.002***
(.000)
.007***
(.001)
-.001*
(.001)
.004***
(.000)
.014***
(.001)
-.002***
(.000)
.004***
(.001)
.011***
(.001)
DP/16
33
(c) Estimates for countries with high pr (70 – 100)
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd Quartile 1
st Quartile Median 3
rd Quartile
pcby .002***
(.000)
.005***
(.000)
.011***
(.000)
dcpy
.001***
(.000)
.005***
(.000)
.009***
(.001)
GC -.048***
(.005)
-.017***
(.005)
.014***
(.003)
-.026***
(.004)
.018***
(.004)
.048***
(.003)
TY -.002***
(.000)
.002***
(.000)
.006***
(.000)
-.001
(.000)
.003***
(.000)
.006***
(.000)
Dependent
variable
ln(rpcGDP)
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
smk -.017***
(.006)
.014***
(.002)
.060***
(.007)
str
-.000
(.000)
.001***
(.000)
.003***
(.000)
sva
.000**
(.000)
.004***
(.000)
.010***
(.001)
GC -.002**
(.001)
.004***
(.000)
.009***
(.000)
-.040***
(.006)
.006**
(.003)
.044***
(.006)
-.005
(.007)
.025***
(.003)
.085***
(.008)
TY .000
(.000)
.001***
(.000)
.004***
(.000)
-.004***
(.000)
.003***
(.000)
.008***
(.000)
-.014***
(.001)
-.000
(.001)
.006***
(.001)
Financial development and growth: Does one size fit all? Evidences from a nonparametric
34
Table 5. Non parametric slope estimates by overall institutions classification
(weak to ideal)
(a) Estimates for countries with low ief (0 – 39)
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd Quartile 1
st Quartile Median 3
rd Quartile
pcby .017***
(.005)
.034***
(.004)
.047***
(.007)
dcpy
.002
(.003)
.019***
(.004)
.031***
(.010)
GC -.023***
(.004)
-.006
(.007)
.033
(.035)
-.023***
(.007)
-.012***
(.004)
.016
(.013)
TY -.006**
(.002)
.000
(.001)
.003**
(.001)
-.007**
(.003)
.000
(.001)
.005**
(.002)
Dependent
variable
ln(rpcGDP)
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
smk .000
(.000)
.001
(.001)
.004
(.003)
str
.006*
(.004)
.012***
(.003)
.016***
(.001)
sva
.001
(.013)
.027**
(.015)
.053***
(.015)
GC -.042*
(.023)
-.014
(.010)
-.000
(.006)
-.074***
(.0110
-.048***
(.018)
-.022***
(.008)
-.056***
(.022)
-.034
(.033)
-.001
(.17)
TY -.013*
(.005)
-.000
(.006)
.003
(.003)
-.014***
(.003)
-.003
(.005)
.000
(.004)
-.009
(.006)
-.001
(.003)
.003
(.003)
(b) Estimates for countries with medium ief (40 – 69)
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd Quartile 1
st Quartile Median 3
rd Quartile
pcby .002***
(.000)
.009***
(.000)
.020***
(.000)
dcpy
.004***
(.000)
.012***
(.000)
.022***
(.000)
GC -.044***
(.003)
-.007***
(.001)
.018***
(.001)
-.024***
(.002)
.000
(.002)
.031***
(.001)
TY -.000**
(.000)
.003***
(.000)
.009***
(.000)
-.000
(.000)
.004***
(.000)
.009***
(.000)
Dependent
variable
ln(rpcGDP)
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
smk -.000
(.000)
.002***
(.000)
.009***
(.000)
str
-.001***
(.000)
.001***
(.000)
.006***
(.000)
sva
.002***
(.000)
.012***
(.001)
.045***
(.002)
GC -.032***
(.005)
.009***
(.003)
.060***
(.005)
-.040***
(.004)
.002
(.003)
.054***
(.006)
-.032***
(.004)
.014***
(.003)
.066***
(.005)
TY -.002***
(.000)
.003***
(.000)
.009***
(.000)
-.002***
(.000)
.004***
(.000)
.010***
(.000)
-.005***
(.000)
.003***
(.000)
.010***
(.000)
DP/16
35
(c) Estimates for countries with high ief (70 – 100)
Dependent
variable
ln(rpcGDP)
1st Quartile Median 3
rd Quartile 1
st Quartile Median 3
rd Quartile
pcby .001***
(.000)
.004***
(.000)
.007***
(.001)
dcpy
.001***
(.000)
.005***
(.000)
.008***
(.000)
GC -.064***
(.006)
-.020***
(.007)
.010**
(.005)
-.031***
(.004)
.000
(.005)
.035***
(.007)
TY -.006**
(.002)
.002***
(.000)
.007***
(.001)
-.001
(.001)
.003***
(.000)
.005***
(.000)
Dependent
variable
ln(rpcGDP)
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
1st
Quartile
Median 3rd
Quartile
smk -.000
(.000)
.001**
(.000)
.003***
(.000)
str
-.000
(.000)
.001***
(.000)
.004***
(.000)
sva
.000
(.000)
.002***
(.000)
.010***
(.002)
GC -.017**
(.007)
.011***
(.004)
.045***
(.010)
-.046***
(.011)
.005
(.004)
.043***
(.006)
-.018***
(.006)
.022***
(.005)
.072***
(.016)
TY -.001
(.000)
.003***
(.000)
.008***
(.001)
-.006***
(.001)
.002***
(.001)
.007***
(.001)
-.012***
(.003)
.000
(.001)
.007***
(.001)
Financial development and growth: Does one size fit all? Evidences from a nonparametric
36
Table 6. Nonparametric slope coefficients by financial sector structure
(a) Estimates by region
Dependent variable
ln(rpcGDP)
pcby dcpy smk str sva
3rd
Quartile
East Asia and Pacific .017***
(0.002)
.014***
(.001)
.009***
(.002)
004***
(.001)
.027*
(.016)
Europe and Central Asia 0.021***
(0.003)
.024***
(.001)
.032***
(.008)
.005**
(.002)
.063***
(.004)
Latin America & the
Caribbean
0.01***
(.001)
.013***
(.0005)
.005***
(.001)
.008***
(.002)
.025***
(.004)
Middle East and North
Africa
.027***
(.006)
.017***
(.002)
.005***
(.001)
.006**
(.002)
.031***
(.004)
South Asia .019***
(.0009)
.018***
(.0009)
.01**
(.005)
.003
(.003)
.038***
(.013)
Sub-Saharan Africa .038***
(.001)
.039***
(.001)
.013***
(.003)
.015***
(.001)
.083***
(.003)
(b) Estimates by regions income level
Dependent variable
ln(rpcGDP)
pcby dcpy smk str sva
3rd
Quartile
high income OECD
countries
.037***
(.011)
.008***
(.000)
.005***
(.000)
.002***
(.000)
.007***
(.000)
high income non OECD
countries
.009***
(.002)
.010***
(.001)
.003***
(.000)
.003**
(.001)
.020***
(.002)
(c) Estimates by UN regional commissions
Dependent variable
ln(rpcGDP)
pcby dcpy smk str sva
3rd
Quartile
ECA .037***
(.002)
.037***
(.001)
.011***
(.002)
.014***
(.001)
.080***
(.003)
ECE .01***
(.001)
.013***
(.0009)
.010***
(.000)
.002***
(.000)
.032***
(.005)
ECLAC .01***
(.001)
.013***
(.0007)
.005***
(.000)
.007***
(.001)
.020***
(.003)
ESCAP .019***
(.0009)
.017***
(.0009)
.006***
(.001)
.004***
(.001)
.020***
(.003)
ESCWA .02***
(.004)
.017***
(.001)
.002***
(.000)
.001***
(.000)
.010**
(.004)
DP/16
37
Table 7. Nonparametric slope coefficients by size of the financial sector
(roughly arranging by quartiles)
(d) Comparing Estimates for Countries by size of Financial Sector
Dependent variable
ln(rpcGDP) pcby <15 dcpy<19
smk <30 str < 5 sva < 1
3rd
Quartile .037***
(.001)
.038***
(.001)
.014***
(.001)
.013***
(.001)
.073***
(.004)
Dependent variable
ln(rpcGDP) 15 < pcby
<50
19<dcpy<50 30<smk <60 5 < str < 20 1 < sva < 4.5
3rd
Quartile .017***
(.0009)
.015***
(.0006)
.004***
(.000)
.009***
(.001)
.045***
(.002)
Dependent variable
ln(rpcGDP) pcby > 50
dcpy>50 smk > 60 str > 20 sva > 4.5
3rd
Quartile .008***
(.0005)
.009***
(.0004)
.003***
(.000)
.002***
(.000)
.045***
(.002)
Table 8. Non parametric slope estimates by property rights (weak to ideal)
(d) Comparing 3rd
Quartile Estimates for countries by pr
Dependent variable
ln(rpcGDP)
low pr (0 – 39) medium pr (40 – 69) high pr (70 – 100)
pcby .033***
(.002)
.019***
(.001)
.011***
(.000)
dcpy .031***
(.001)
.018***
(.000)
.009***
(.001)
smk .016***
(.002)
.044***
(.006)
.060***
(.007)
str .010***
(.001)
.007***
(.001)
.003***
(.000)
sva
.051***
(.003)
.048***
(.003)
.010***
(.001)
Table 9. Non parametric slope estimates by overall institutions classification
(weak to ideal)
(d) Comparing 3rd
Quartile Estimates for countries by ief
Dependent variable
ln(rpcGDP)
low ief (0 – 39) medium ief (40 – 69) high ief (70 – 100)
pcby .047***
(.007)
.020***
(.000)
.007***
(.001)
dcpy .031***
(.010)
.022***
(.000)
.008***
(.000)
smk .004
(.003)
.009***
(.000)
.003***
(.000)
str .016***
(.001)
.006***
(.000)
.004***
(.000)
sva
.053***
(.015)
.045***
(.002)
.010***
(.002)
All standard errors in parentheses are obtained via bootstrapping
* indicates significance at 10% level.
** indicates significance at 5% level.
*** indicates significance at 1% level.
MPDD Working Papers WP/16/..
About Economic and Social Commission for Asia and the Pacific (ESCAP) ESCAP is the regional development arm of the United Nations and serves as the main economic and social development centre for the United Nations in Asia and the Pacific. Its mandate is to foster cooperation between its 53 members and 9 associate members. ESCAP provides the strategic link between global and country-level programmes and issues. It supports Government of countries in the region in consolidation regional positions and advocates regional approaches to meeting the region’s unique socio-economic challenges in a globalizing world. The ESCAP office is located in Bangkok, Thailand. WWW.UNESCAP.ORG TWITTER.COM/UNESCAP FACEBOOK.COM/UNESCAP YOUTUBE.COM/UNESCAP
the Australian experience of reform?