Financial Development and Economic Growth: Comparative Analysis across Income Groups of Countries

Analysis across Income Groups of Countries

A. M. Oumer Student No: s1702777

e-mail: s1702777@student.rug.nl or a.m.oumer@student.rug.nl

Thesis Supervisor: Prof. Dr. E. Sterken

Thesis for the Degree of Master of Science in Economics

Department of “Economics”

Faculty of Economics and Business

University of Groningen

The Netherlands

Table of Contents

Abstract:

...

2

1.Introduction

...

2

2.Theoretical Framework

...

4

2.1.Theories of Economic Growth

...

4

2.2.Banks

...

7

2.2.1.Banks as Intermediaries in Imperfect Financial Markets

...

7

2.2.2.Endogenous Formation of Financial Intermediaries

...

10

2.3.Banks and Equity Market

...

15

2.3.1.Financial Structure: the Evolution

...

15

2.3.2.Banks versus Markets: the Demarcation and Implication for Economic Growth

...

16

2.3.3.Financial Markets and Economic Growth: Causal Relation or Simple Correlation

...

19

2.3.4.Legal matters, Financial Market Efficiency and Economic Growth

...

21

3.Empirical Literature: Market, Intermediaries and Growth

...

25

3.1.Summary

...

25

3.2.Discussion

...

26

3.3.Evaluation

...

33

4.Data and Methodology

...

36

4.1.Data

...

36

4.2.Methodology

...

38

5.Results and Discussion

...

43

6.Summary and Conclusions

...

50

References

...

52

Abstract:

This paper analyses the relationship between financial development and economic growth. It investigates whether bank and stock market development have a similar impact on economic growth across countries in different income groups as well as between bank-based and market based economies. The paper also attempt to see whether the impact of the financial development on economic growth tends to decline after some level of development. The relationship between both bank and stock markets development and economic growth is stronger in middle income countries compared to both low income and high income countries. Stock market capitalization tends to have lower impact on economic growth in high income countries; whereas bank concentration has positive impact on economic growth in high income countries as opposed to other income groups. Net interest margin has the strongest link with economic growth throughout all income groups.

Key Words: Economic growth, intermediary development, stock market size, stock market efficiency/liquidity,

1. Introduction

The role of financial sector development in promoting economic growth has received large attention of researchers since the early 1910s. Since Schumpeter’s (1911) argument suggesting productivity and growth-enhancing services increase in a developed financial sector, various research works have emerged. The earlier literature focuses on the question whether the financial sector plays a causal role in economic development or if the financial intermediaries emerge from rapid industrialization (Eschenbach, 2004). However, recently there has been a shift in this interest. Allen and Gale (2000) argue that the current trend is towards market-bases systems. The central issue in the more recent literature has been whether financial structure (stock market or banks) is more appropriate to promote economic growth (Garretsen et al, 2004). These literatures categorize economies into two classes, market-based and bank-based, and analyze them.

on very narrow set of countries or data (Demirguc-Kunt and Levine, 2001) and on formal institutions (Garretsen, et al, 2004). Thus, these literatures extended the analysis to larger number of countries (and data sets) and to informal institutions, respectively.

2. Theoretical Framework

In this section theoretical framework of financial development and economic growth are discussed. Economic growth theories and models are discussed in the first part of this section. The discussion is then extended to banking sector in the second part. And banks and stock markets are discussed in the third part of this section.

2.1.

Theories of Economic Growth

The theory of economic growth and interest of research in economic growth have been evolved over time. The neo-classical growth models explain specific sources of growth such as saving and investment. However, Kaldor (1961, 1963) argue that a satisfactory theory of economic growth should be able to explain a number of stylized facts. Some of these facts include that per capita income grows over time; physical capital per worker increases over time; growth rate of output per worker differs across countries; rate of return to capital is nearly constant; and so on.

Classical economists provided many of the basic ingredients that appear in modern theories of economic growth. These ideas include competition and equilibrium dynamics, diminishing return, accumulation of physical and human capital, per capita income and population growth, technological progress and specialization, discoveries of new goods and methods of production and role of monopoly in innovation (Barro and Sala-i-Martin, 1995). Using these ingredients and modern mathematical methods, modern growth models attempt to explain these facts by extending the earlier models to include human capital, financial wealth and so on. In this section some major growth models are explained. Then theories that relate economic growth with financial development will be discussed.

The neo classical growth model is based on the notion of an aggregate production function of the following form (Barro and Sala-i-Martin, 1995 and Heijdra and Van der Ploeg, 2002):

)],

(

[

)

(

t

f

k

t

y

=

(2.2) where

y

(

t

)

=

Y

(

t

)

/

L

(

t

)

and

k

(

t

)

=

K

(

t

)

/

L

(

t

)

. Moreover, assuming saving is a fraction of output and equals investment which is in turn equal to change in capital over time.

) ( )

(t sY t

S = ; and S(t)₌I(t)_=δK(t)₊K(t)_⇒K_{(t)=I(t)−δK(t) (2.3)}

where

0

<

s

<

1

is a saving rate;

δ

is depreciation rate of capital; and the dot above a variable shows change over time. A condensed Solow-Swan model is given by the following single fundamental differential equation in per capita capital stock (Heijdra and Van der Ploeg, 2002).

)

(

)

(

)]

(

[

)

(

t

sf

k

t

n

k

t

k



=

−

δ

+

(2.4) Where n is population growth rate. This differential equation states that change in per capita capital over time is the difference between injection ‘

sf

[ t

k

(

)]

’ through saving and leakage ‘

)

(

)

(

+

n

k

t

−

δ

’ through depreciation and population growth. Given that s, n and

δ

are constants, the differential equation depends only on ‘k(t)’. In the steady state k(t) is constant (Barro and Sala-i-Martin, 1995). This imply that

k

(

t

)

=

0

and

sf

[

k

(

t

)]

=

(

δ

+

n

)

k

(

t

)

. This model implies that there is absolute convergence to the steady state in the long run since

k

(

t

)

=

0

. Thus, this model does not provide explanations of the determinants of long run per capita GDP growth. Dissatisfaction with this exogenously driven explanation of long run productivity growth motivated a group of growth theorists led by Romer (1986) to construct growth models in which the key determinants of growth are endogenous to the model. The key property of endogenous-growth models is the absence of diminishing returns to capital. The simplest version of them and the one we discuss here is the AK model. The AK model is given as:

)

(

)

(

t

AK

t

Y

=

(2.5) where A is a positive constant that reflects the level of technology. The idea of absence of diminishing return becomes more plausible if we think of K(t) in a broad sense to include human capital. The average and marginal products of capital are now constant, (and not diminishing), at

A > 0. using

A

t

k

t

k

f

₌

)

(

)]

(

[

we obtain the following steady state growth rate

γ

_k:

)

(

n

sA

k

=

−

δ

+

γ

or

γ

k

=

sA

−

(

g

+

δ

+

n

)

(2.6)

Some growth theorists have introduced the financial sector into growth models. For instance, Chou and Chin (2004) and in their earlier works constructed a growth model with a financial sector. The financial sector in their model consists of financial innovators and financial intermediaries. Financial innovations include novel means of raising funds for expanding firms, as well as new instruments designed to attract more savings from investors. They denote the stock of financial products as

τ

. The existing stock of financial products affects the production of new financial ideas according to:

φ λ

τ

τ

τ



=

F

( L

u

)

(2.7) where

τ



denoted the quantity of financial innovations per unit of time; L denotes labour;

u

_τ is the fraction of labour force employed by financial sector;Fis a productivity parameter; λ∈(0,1) is an elasticity parameter; and

φ

∈(0,1)measures the extent of spillovers from existing financial products. Financial intermediaries, on the other hand, are responsible for intermediating funds between borrowers and lenders. Unlike conventional growth models, Chou and Chin do not assume that all household savings will automatically be costlessly transformed into funds that are utilizable by firms for investment. In particular, some risk-averse savers will continue to hold liquid but unproductive assets until they are offered a sufficient variety of financial products, while the financing needs of some firms will remain unfulfilled. The efficiency by which savings can be transformed into productive investment is determined according to the following.

κ

τ

where Y(t) is the level of aggregate output andC(t)is aggregate consumption level. The remaining parts of the Barro and Sala-i-Martin model type remain the same except the stock financial products and efficiency of intermediation affects the capital accumulation process which is the typical source of economic growth across various growth models. In this model, both the stock financial products and efficiency of intermediation have positive effect on capital accumulation and thus on economic growth.

In the empirical section of this paper the following growth model with financial development indicators is estimated.

y = f[yt-1, X, ex(FD)]; where FD = g(y) (2.10) where y is per capita GDP growth rate; yt-1 is lagged value of per capita GDP growth rate; X is investment indicator; and ex(FD) represents exogenous component of financial development indicators. The expression, FD = g(y) shows endogeneity of financial development indicators and justify the use ex(FD) instead of FD.

2.2.

Banks

2.2.1. Banks as Intermediaries in Imperfect Financial Markets

Quoting McKinnon (1973), they used historical role that banks and insurance companies played to support their argument. They added that this appears to be true in virtually all economies including emerging economies where the development of intermediaries tends to lead the development of financial markets themselves.

Freixas and Rochet (1997) have given a justification for the existence of financial intermediaries. According to them, being financial intermediaries, banks, mutual funds and insurance companies are there to transform financial contracts and securities. In an ideal world of frictionless and complete financial markets, both investors and borrowers would be able to diversify perfectly and obtain optimal risk sharing. But as soon as one introduces even small indivisibilities the perfect diversification is no longer feasible and financial intermediaries are needed. In the process of financial market evolution, Freixas and Rochet have also acknowledged that some of the roles of financial intermediaries are replaced by the recent development of security markets. Yet, they agree with Allen and Gale (1995) that although the existence (development) of more sophisticated financial instruments contributes to financial welfare by allowing agents to hedge their risks, the existence of financial intermediaries smoothes the shocks affecting consumption.

As major intermediaries in the financial sector, banks play important role of resource allocation among different stakeholders and different economic activities. According to Freixas and Rochet (1997), contemporary banking theory classifies banking functions into four main categories: (1) offering access to payment system; (2) transforming assets; (3) managing risks; and (4) processing information and monitoring borrowers. Banks exert fundamental influence on capital allocation, risk sharing and economic growth. Underdeveloped economies with a low level of financial intermediation and small, illiquid, financial markets may be unable to channel savings efficiently. Moreover, the fact that more bank-oriented countries such as Germany and Japan have experienced a higher rate of growth has motivated additional research on the economic role of banks. It is concluded that banks have an important economic function in an economy because of the demand for different monies; for divisible, low risk, short term liabilities; for indivisible, risky, long term capital; and for project monitoring.

monitoring cost (C) reduces the benefit further to b, where b + C < B. Moral hazard can be solved without monitoring if the firms have enough cash assets ‘A’ that they can invest. This is so when the following condition is satisfied:









−

−

−

=

≥

L H H

P

P

B

R

P

I

A

A

ρ

ρ

)

(

(2.11)

where R is loan repayment in case of success and

ρ

is the return (=1 + interest rate) that investors can obtain in the financial market. This model suggests that increased liquidity helps to mitigate the information problem. That is, increased liquidity makes sure that (or increases the chance that) this condition is fulfilled. If this condition is not satisfied, the moral hazard problem may be solved by bank lending. A bank lends Im to the firmwho finance the investment partly by its own funds and partly by borrowing I – Im – A on the financial market. This happens when

)

(

)

,

(

β

ρ

A

A

ρ

A

≤

≤

, where

β

is return on banks loan and:

2.2.2. Endogenous Formation of Financial Intermediaries

A number of theoretical models, dealing with individual behavior and institutional setup in the process of financial intermediation, have been developed over time. Many of such works assume that the financial intermediation process is exogenously given. However, Bencivenga and Smith (1998) explain endogenous formation of financial intermediaries. They discuss three equilibrium conditions: first in an equilibrium with no financial intermediation; second in an equilibrium with financial intermediation; and third in an equilibrium with endogenous financial intermediation.

i)

Equilibrium with no Financial Intermediation

In the absence of financial intermediation, an agent divides his saving between liquid storage investment and illiquid capital investment. The agent then holds these investments directly. Let a

t

q

(“a” for “autarky”) denote the fraction of saving placed in the storage technology at date t. With probability P, the agent consumes in the first period of life. In this event, he liquidates all of his assets, and consumes the proceeds; and any investment in the capital technology is lost. Thus, the consumption of such an agent is given by:

c

1t=

(

t

)

a t

rw

k

q

(2.13)

where r is investment yield; w is real wage rate; and kt is capital-labour ratio at time t. With probability (1 – P), the agent consumes in the second period of life. In this case all assets are held until the second period, and the consumption of such an agent is:

t

c

₂ = [ a t

q

r + (1- a t

q

)Rt]w(kt), (2.14)

where Rt = f ‘(kt+1) denotes a capital rental rate at time t+1. A competitively acting agent maximizes the following expected utility:

E(U) = P ln( a t q r) + (1-P) ln[ a t q r + (1 - a t q )Rt] + ln w(kt) (2.15)

   − − > = otherwise r R P if R r P qa t t t 1 ) 1 ( )] / ( 1 /[ (2.16)

The maximized value of expected utility is given by:

U* = P ln P + (1-P) ln(1-P) + P ln r + ln Rt - P ln(Rt - r) + ln w(kt) (2.17)

This expression is increasing in r, and it is increasing in Rt if and only if Rt (1-P) > r. The equilibrium law of motion for the capital stock in the absence of financial intermediation is:

kt+1=(1-P)(1-qta)w(kt) (2.18)

This low of motion is denoted (LOMa_{). LOM}a _{is monotonically increasing and it} assumed to crosses 45-degree line from above at least once (see figure 2.1 below).

Figure 2.1: Capital-labour ratio (Low of Motion, LOM), under Autarky, bank and endogenous intermediation

Source:

Bencivenga, Valerie R. and Smith, Bruce D. (1998).

When financial intermediaries are active, they accept deposits (in the form of the consumption good), and invest these deposits in the storage technology and the capital investment technology. Following Baumol (1952) and Tobin (1956), they assume that agents incur a fixed cost

γ

in “going to the bank”. At each date each bank chooses the fraction of deposits it places in the storage technology,

b t

q

(“b” for “banks”), the rate of return it will pay on deposits withdrawn at the end of the same period, r1t, and the rate of return it will pay on deposits withdrawn at the end of the next period, r2t, taking as given the return on the capital investment technology, Rt, and the rates of return offered by other banks. Competition among banks for depositors implies that in a Nash equilibrium, the deposit return schedule (r1t, r2t) of each bank, and its portfolio allocation,

b t

q

, must be chosen to maximize the expected utility of a representative depositor:

P ln r1t+(1-P) lnr2t + ln[w(kt) -

γ

] (2.19)

subject to the resource constraints: r b

t

q

= Pr1t and Rt(1-

q

tb) = (1-P)r2t (2.20)

The first constraint requires the yield from investment in the storage technology to equal payments to the fraction P of depositors who withdraw at the end of the same period. The second constraint requires the yield on capital investments to equal payments to the fraction (1-P) of depositors who withdraw at the end of the next period. These constraints incorporate the fact that, if Rt > r, no bank would ever choose to carry low-yielding storage investment between periods. Investment in the storage technology is made only as a form of “reserve-holding” to accommodate the withdrawal demand of agents who turn out to be “impatient”. The solution to the bank’s maximization problem is to set

b t

q

= P, r1t = r, and r2t = Rt. Note that

q

tb<

q

ta necessarily holds, so that the

being invested in productive, but illiquid, capital. The solution to the bank’s problem implies that expected utility of a depositor at date t is given by:

P ln r+(1-P) lnRt + ln[w(kt) -

γ

] (2.21)

Note that this expression incorporates the fixed cost of utilizing the bank’s services,

γ

. The equilibrium law of motion for the capital stock when all saving is intermediated is:

kt+1=(1-

q

tb )[w(kt) -

γ

]

(2.22)

They denote (2.22) as low of motion with banks (LOMb_{). This implies that k}

t+1 is an increasing function of kt. And, if we assume that w’’(kt) < 0 for all k > 0, then kt+1 is a strictly concave function of kt as well.

iii)

Endogenous formation of financial intermediaries

Endogenous formation of financial intermediaries answers such questions as if agents are free to utilize intermediaries or not, when will they choose to save through banks? Given the results of the previous sections, clearly agents will prefer to have their saving intermediated if and only if:

Plnr + (1-P)lnRt+ln[w(kt) -

γ

]

≥

PlnP +(1-P)ln(1-P)+Plnr + lnRt - Pln(Rt-r)+ln w(kt) (2.23)

Rearranging this, we obtain that banks will be in operation if and only if:

P P t t P t t _{P P} k w k w k f r k f ₋ + + _{≥ −}     −     − 1 1 1 _{(1 )} ) ( ) ( ) ( ' ) ( ' γ (2.24)

capital investment technology and storage technology directly. Assuming that f(k) satisfies standard Inada conditions, LOI intersects the horizontal axis when w(kt) = _1−PP_(1−p)1−P

γ

. Bencivenga and Smith have discussed various further scenarios. One

of them is the following and a demonstrative figure is given in figure 2.1 above. Let a H

k

denotes the largest (and possibly unique) positive capital-labour ratio where LOM a _{intersects the 45-degree line; let} I

L

k

and I H

k

denote the capital-labour ratios where LOI intersects the 45-degree line; and let b

H

k

denote the larger capital-labour ratio where LOMb _{intersects the 45-degree line. This leads to the} condition: a H

k

< I L

k

< b H

k

< I H

k

. This condition implies that there are three nontrivial steady state equilibria. One has kt = kt+1 =

k

Ha. In this steady state financial

intermediaries do not operate, and the capital stock is low (as is per capita output). Another steady state has kt = kt+1=

k

Hb . Here all saving is intermediated, a

relatively large fraction of saving is invested in productive but illiquid capital; and, as a result, the steady state capital stock is high. Finally, there is a steady state with kt = kt+1 = kLI . In this steady state, intermediaries are active, but not all

saving is intermediated. It is apparent from graphical and mathematical inspection that the first two steady states are asymptotically stable, while the third steady state is unstable. The local stability of the high and low capital stock steady states implies that costly financial intermediation can easily lead to the existence of development traps. In the low-activity steady state incomes are low, rendering the cost of intermediation prohibitive. A one-time injection of resources, that moves the agent’s income above

(

I

)

L

k

2.3.

Banks and Equity Market

2.3.1. Financial Structure: the Evolution

Financial market structure has been evolved and different financial instruments have been developed over time. Allen and Santomero (2001) explained a dynamic nature of financial intermediaries and financial innovation. They believe that there has been a fundamental shift in the nature of intermediation in the US and also in the UK. Both of these countries are market based economies (Demirguc-Kunt and Levine, 2001).

Allen and Gale (2000) also argue that the current trend is towards market-based systems. Demirguc-Kunt and Levine explain the shift process as follows. In a traditional bank-based economy, where financial markets are not very significant, the main way in which banks deal with risk is through inter-temporal smoothing. They acquire a “buffer” of short-term liquid assets when times are good and run this buffer down when times are bad. As a result of this buffer, households, those hold most of their assets in bank accounts and other fixed income assets are to a large extent shielded from risk and are able to have smooth consumption streams and liquidity insurance. When financial markets develop, they provide competition to banks which makes the intertemporal smoothing they undertake increasingly more difficult. Financial markets allow high returns in good times and there is an incentive for individual investors to withdraw their funds from banks and put them in markets instead.

analysis, Allen and Santomero (1997) have shown that there has been a distinct long term increase in market capitalization relative to GDP.

2.3.2. Banks versus Markets: the Demarcation and Implication for Economic

Growth

Chakraborty and Ray (2006) shed light on the long-standing debate whether bank-based or market-based systems are better for growth. They suggest that such an ‘either-or’ question is, in fact, ill-posed: the growth rate is a function not so much of the financial regime as of the quality of services it delivers. Moreover, we can raise more questions whether the bank and the market are substitutes or complements. They can be seen as substitutes from services point of view and as complements if we see total functioning. For instance, it should be clear that a market-based system does not preclude banks. Yet, it is possible to classify an economy as bank-based or market-based.

The financial sector transforms household savings into capital. Two types of agents participate on the supply side, financial intermediaries (banks) and households themselves. Banks obtain their supply of loanable funds from households. Households have the choice of depositing their savings with banks, or lending directly to firms, or investing it on the international capital markets. Direct lending to firms, which is referred to as direct (or market) finance, is made through the purchase of tradable securities like corporate bonds and equities. The direction and proportion of this flow is useful to establish the demarcation between bank-based and market-based economies before passing on to the discussion whether the bank or the market matters more for economic growth.

Chakraborty and Ray (2006) give a model that can be used as a demarcation between bank-based and market-based economies. They use the following terms and symbols: γ as cost of monitoring; v as residual moral hazard problem under bank monitoring; V as moral hazard problem in the absence of external monitoring; R* is a constant world (gross) rate of return (on perfectly mobile capital markets); and A and 0 < α < 1 are constants in technological efficiency function: At = Akt1- α_{. and k}

t is per capita capital.

But, an entrepreneur with lower level of wealth, bL_{, has to borrow from both financial markets} and banks to make an investment of size qt. The earnings (zt+1), of the entrepreneur who borrows from both in markets and banks, is given by:

t t

t

R

b

A

R

q

z

₊₁

=

*

+

[

α

−

(

1

+

γ

)

*]

, (2.25) The earnings of the entrepreneur who borrows only from the market is given by:

t t

t

R

b

A

R

q

z

₊₁

=

*

+

[

α

−

*]

(2.26) where bt is internal fund. The second identity (2.26) earning line is steeper since the slope is larger by the monitoring cost term,

(

1

+

γ

)

. For an investment level of

q

t, they define the

minimum amount of internal funds required to qualify for indirect and direct finance are

_b

L

(

_q

_t

)

and

_b

H

(

_q

_t

)

, respectively. Defining qI,tand

q

U,tas points of intersection of

(

t

)

L

_q

b

and

_b

H

(

_q

_t

)

with entrepreneur’s wealth bt, they give closed-form solutions for these investment levels as:

t t I

b

R

A

v

R

q

_











+

−

−

−

=

*]

)

1

(

[

)

1

/(

*

,

_π

_α

_γ

and Ut

b

t

R

A

V

R

q

_











−

−

−

=

*]

[

)

1

/(

*

,

_π

_α

(2.27) where

π

represents a probability of success. Since the earning

z

t+1

(

q

t

)

is increasing in the range

]

*,

[

_U,t t

q

q

q

∈

, the maximal earning occurs at

q

t

=

q

U,tand is given by:

t t U t

b

R

A

V

R

V

q

z

_











−

−

−

−

=

+

_/(

₁

₎

₍

_*)

*

)]

1

/(

[

)

(

_, 1

_π

_α

π

(2.28)

where

q

*

represents minimum investment level. Similarly, the maximum earning for

]

,

[

_{U,t I,t}

t

q

q

q

∈

occurs at

q

t

=

q

I,t, and is given by:

t t I t

b

R

A

v

R

v

q

z

_











+

−

−

−

−

=

+

_/(

₁

₎

_[

₍

₁

₎

_*]

*

)]

1

/(

[

)

(

_, 1

_π

_α

_γ

π

(2.29) t I

q

_, represents the case where bank financing is important; whereas

q

U,trepresents the case of

relatively wealthier entrepreneurs, where borrowing on the market is sufficient to finance investment

q

t. It follows that an entrepreneur chooses

q

I,t over

q

U,tif and only if:

This condition of demarcation between market-based and bank-based economies doesn’t depend on borrower characteristics, that is, on bt. It depends on monitoring cost, the level of moral hazard with and without monitoring, and technological efficiency. Intuitively, the financial structure is likely to be bank-based whenever the cost of monitoring (γ) is low and whenever the residual moral hazard problem under bank monitoring (v) is low relative to the moral hazard problem in the absence of external monitoring (V). The financial structure of an economy is bank-based if and only if (2.30) holds. It is otherwise market-based without any dependence on intermediated finance in the long-run. In the short-run, some entrepreneurs with low wealth rely upon intermediated finance even in a market-based system.

Chakraborty and Ray also underlined that a market-based system does not preclude banks. Since the only role banks perform is of monitoring, all it means is that even when banks participate in the loanable funds market, they do not engage in monitoring activities. Based upon their informational content, bank- and market-finance become indistinguishable in that case. The argument is based on the assumption that compared to direct finance, indirect (or bank) finance plays a special role. Banks are endowed with a monitoring technology that allows them to inspect a borrowing firm's cash flows and balance sheet, keep tabs on the owner-manager's activities and ensure the firm conforms to the terms agreed upon in the financial contract. Households do not possess this technology, or even if they do, are too disparate to effectively use it. Hence, banks assume the role of delegated monitors. According to Allen and Gale (2000) market-based financial systems are characterized by dispersed information, and dispersed shareholdings give a large number of people an incentive to gather information on firms and monitor their performances.

those with industries with a significant amount of concentration. Bank finances are more prevalent in countries whose development is dominantly through existing technology rather than entirely new ones, for instance Germany and Japan.

2.3.3. Financial Markets and Economic Growth: Causal Relation or Simple

Correlation

There are different views about the relationship between financial sector development and economic growth. Although it is argued that financial sector development leads to economic growth, some earlier literatures emphasize rather co-evolution of the real and the financial sectors of an economy during the growth process. In the early stages of economic development most investment tends to be self-financed, and financial market activity is severely limited or even non-existent. With increased levels of development come bilateral borrowing and lending. Further development is associated with the appearance of banking and the intermediation of investment. And, ultimately, even more sophisticated financial institutions, such as equity markets, come to be observed (Gurley and Shaw, 1955, 1960, and 1967). Bencivenga and Smith (1998) argue that Gurley and Shaw attributed no particular direction of causation in the co-evolution of the financial system and the level of development. Higher levels of real activity, in their story, do not “cause” financial markets to become more developed. Nor does this financial deepening “cause” the growth of real activity. Rather, in Gurley and Shaw’s account, financial depth contributes to real development and real development contributes to the growth of the financial system. Bencivenga and Smith agree with many other researchers that there are many ways through which financial markets might promote economic growth. Some of these are that financial system of an economy can facilitate exchange, favour the collection of information and the efficient allocation of investment capital, promote specialization, improve the monitoring of managers, or mobilize savings and provide liquidity.

Germany and Russia who entered industrialization when technology was more advanced and investment had a large scale. This time the banking sector provided both capital and entrepreneurship to drive the industrialization process.

According to Patrick (1966) the causal relationship between finance and growth follows two patterns called “demand following” and “supply leading”, attributing to specific stages of the development process. First, economic development establishes a demand for financial services, which is passively satisfied by a growing financial sector. Second, financial intermediation induces economic growth by channeling savings of mostly small savers to large investors. The financial sector channels resources from the traditional to the modern sectors and promotes entrepreneurship in the latter. The supply-leading pattern dominates during the early stages of economic development, and subsequently gradually shifts its leading role to the demand following one. So initially the causality runs from finance to growth, a scenario that should be expected in developing countries. The demand-following pattern should then be expected to establish a causality that runs from growth to finance.

Cameron (1967) argues that financial systems may be both growth-inducing and growth-induced. But he emphasizes the crucial role of the quality of its services and the efficiency with which they are provided. He subsequently summarizes important features of the financial system, in particular of banks: financial intermediation serves as a vehicle for channeling small funds from risk-averse savers to less risk-averse people with entrepreneurial skills. This results in increased availability of funds for the latter. Secondly, financial intermediation provides incentives to investors. Declining costs of borrowing encourage entrepreneurs to make larger investments. An expanding financial sector should reduce the dispersion of interest rates among users, regions, and over periods of seasonal fluctuation. Thirdly, financial institutions create possibilities for a more efficient allocation of the often unproductive stock of initial wealth in the early stages of industrialization. Finally he emphasizes the role of banks in promoting technological progress. Majority of technical innovations are introduced by established firms with access to bank financing.

induce saving rate to fall enough such that overall growth slows with more liquid and internationally integrated financial markets. With more liquidity since it is easier to sell shares, there will be less incentive of monitoring. This in turn impedes effective resource allocation and slows productivity growth. However this negative impact is not supported by their empirical assessment.

For Deidda and Fattouh (2005), what is more important is not whether financial development helps economic growth or not, but it is how it helps economic growth. They argue that a large body of literature on finance and growth offers several explanations as to why financial institutions facilitate economic growth. Financial institutions mobilize savings, diversify risk and produce information about investment opportunities. They also agree that these functions help to improve the productivity of investments. This should result in higher growth rates provided that the returns to accumulating inputs are non-diminishing at an aggregate level.

2.3.4. Legal matters, Financial Market Efficiency and Economic Growth

Legal factors and financial market efficiency can affect investment activities and economic growth by affecting the probability of the repayment rate of bank loans (i.e. by reducing default rate). In a theoretical analysis of loans, risks and growth, the role of government and public banking, Pelozo (2006) shows how financial market modernization and the relationship between projects risk and interest rate affect economic growth. The model follows from Waller and Lewarne (1994) and Pagano (1993). The model is believed to be able to explain the banking sector in Paraguay, where businesses are relatively small. From this model, it is also easy to infer about the role of legal matters in affecting returns on investment projects and so economic growth. Loans have an effect on economic growth through the default probability and other risks faced by intermediaries. What is important in the model is the distinction between different borrowers; credit suppliers are concerned about the interest rate they charge and the riskiness of borrowers.

Investment projects are indexed by their gross return; R1 < R2. Ri = (1 + øi). Where øi is the rate of return of project i. R1 represents a group of projects with low expected returns and low risk (here assumed risk-free). On the other hand, type 2 projects are risky. Labeling the loan to the projects L1 and L2, respectively, the total loans in the economy LT is:

Banks can distinguish if a borrower invests in a type 1 or type 2 project. The payoff of project 1 is R1 with certainty. On the other hand, if borrowers invest in type 2 projects, their payoff is a random variable, i.e., project 2 pays R2 with probability δ and 0 with probability 1 − δ. The loans are fixed-rate for the two types of projects; type 1 projects have an interest rate of r1, and type 2 of r2. The availability of funds depends on the amount of deposits that lenders receive from the public, D and their own equity, E.

LT = ( 1 - π)D + E (2.32)

where (1 − π) is the loans-to-deposits ratio; and the constant π, represents the legal reserve requirements. The loans-to-deposits ratio depends on the probability of repayment, δ, and the probabilities of other risks, α, faced by banks, i.e., π = f(δ, α), as δ increases π will decrease and as α increases π will increase. Lenders maximize their expected profits:

d

Dr

r

L

r

L

Max

.

Π

=

₁₁

+

δ

_{2 2}

−

(2.33) Subject to: D = ω ( L 1 + L 2), (2.34) where rd is the exogenous interest rate paid to depositors and 1< ω = 1/(1 - π). Each bank is considered to be small in the market and unlimited liability of bank shareholders is assumed. From the First Order Conditions we can get the interest rate charged for every type of loan:

r1 = (rdω) ε1 , and r2 = (rdω)(ε2/δ) (2.35) where εi is the elasticity of loan i supplied with respect to changes in the interest rate i. Thus the interest rate charged for each type of loan depends on the bank's costs, the loans-to-deposits ratio, the return paid by deposits, and the repayment probability. When this probability is larger the bank will charge less for type 2 loans. Assuming that elasticities in both markets are the same:

r1 = δr2 (2.36) It is easy to see that an increase in the exogenous δ (probability of loan repayment), ceteris paribus, will make the difference between r1 and r2 become smaller, that is the spread between these two rates will narrow. To see the role that can be played to improve the returns, through improvement of financial sector efficiency (in terms of information as assumed that there is knowledge spillover, legal framework and so on), the default probability or credit risk can be written as the inverse of the repayment probability in the following way:

where Λ is total credit risk, as perceived by banks; Λ1 is systemic risk; and Λ2 is micro risk. We will not do anything about the micro risk because it can be eliminated by diversification. but the systemic risk behavior can be affected by government investment in financial sector. Systemic risk and government investment in the financial sector are inversely related. When the government invests in the financial sector it reduces Λ1. This means systemic risks can easily be identified and eliminated. In addition, this investment increases spillover benefits to all banks in the financial sector. Also when firms invest in upgrading their information systems or their human capital, they can get better assessment of the risk; their loans portfolio reduce their costs, and increase the spill over effect, which in turn reduces Λ1 further. The government's role here could be very important, i.e., it could modernize the legal framework for the financial institutions and eliminate barriers to different operations. In addition, it could impose better control for all businesses, improve accounting systems, reduce market failures using public banks and promote a capital market (stock market, derivatives market, etc.). In general, these and any similar type of investment could promote a more efficient financial system.

Initial investment in the modernization of the financial sector, will decrease the default probability of the overall economy. Then this decrease in the overall probability of default implies a decreased total risk and increased repayment probability. Thus, banks can decrease the interest rate charged for the risky projects. Assuming a downward sloping investment demand, the lower interest rate will imply higher investment. Also investors will become less risk averse, and thus they may be willing to undertake the higher return projects. The latter will improve the economy's resource allocation and increase the long-run growth rate. Using a simple macro model, Pelozo (2006) explains how improvement in financial sector performance positively influences economic growth. Given a simple production endogenous growth model:

Yt = A Kt (2.38) where A is the social marginal productivity of capital. Assume that population is constant and that the economy produces a single good that can be invested or consumed; and there exists a depreciation rate, δ, associated with the investment. Then gross investment equals:

In a closed economy, in equilibrium we require that gross investment equals gross savings. We assume that all savings, S are deposited into banks, and Investments, I are equal to total loans, i.e., S = D and thus I = LT. Using these in (2.32) with E = 0, we have:

LT = (1 − π)D, equivalently I = (1−π)S (2.40) This equation implies that a proportion π of savings is “lost” in the intermediation process. The growth rate of this economy is:

(

)

t t t t t t t _K K K Y Y Y 1 ( 1 ) 1 − = − = + + + θ (2.41)

The steady-state growth for this model is:

θ = A(I/Y) − δ = A(1 − π)s − δ (2.42) where s = S/Y is the savings rate. We can easily see from (2.42) that a decrease in the default probability will influence the long-run growth rate of the economy, by increasing investments, I, and increased loans-to-deposits ratio, (1 − π). In the same way a decrease in other risks, for example the liquidity risk by the activity of the public sector bank, will increase (1 − π) and the loans provided by financial intermediaries, which in turn will increase the equilibrium growth rate.

3. Empirical Literature: Market, Intermediaries and Growth

3.1.

Summary

A number of empirical works on the relationship between financial markets development and different economic growth indicators are produced. Per capita GDP growth is often used as a proxy of economic growth. This section is arranged as follows. First the table below summarizes the main empirical literatures that are discussed in this section. In second part of this section, the literatures are discussed in detail; and in the third part critiques of the empirical literatures are given.

1) Levine and Zervos (1998)

Main tasks or assessment

investigating whether measures of stock market liquidity, size volatility and integration with world capital markets are robustly correlated with current and future rates of economic growth, capital accumulation, productivity investments, and saving rates.

Focus area both banks and equity markets with more emphasis on equity markets development Methodology Least squares as well as Instrumental Variable (IV) estimation and sensitivity tests Data coverage 49 countries for years 1976 through 1993

Major result both the initial level of banking development and the initial level of stock market liquidity have statistically significant relationships with future values of output growth, capital stock growth, and productivity growth

2) Levine, Loayza and Beck (1999)

Main tasks or assessment

evaluating whether the level of financial intermediary development exerts a causal influence on economic growth; and whether cross-country differences in particular legal and accounting system explain cross-country differences in the level of financial intermediary development

Focus area only banks (or financial intermediary) development

Methodology (a) pure cross-sectional estimation and (b) panel estimation; and Instrumental Variable (IV) approach is used in both estimations

Data coverage 71 countries over the period of 1960 to 1995.

Major result strong connection between exogenous component of financial intermediary development and long run economic growth.

3) Beck and Levine (2001)

Main tasks or assessment

analyzing the impact of overall financial development on growth and about separate effects of stock markets on growth and banks on economic growth.

Focus area both banks and equity markets

Methodology panel estimation with Instrumental Variable (IV) approach Data coverage 40 countries over the period 1976 to 1998

Major result both stock market development, measured in terms of stock market turnover has both statistically and economically large positive impact on economic growth, but bank credit to private sector is not always significant

4) Demirguc-Kunt and Levine (2001)

Main tasks or

assessment assessing interrelation among non-bank financial intermediaries, financial market structure including classification into market-based and bank-based, equity markets across countries, financial markets size, efficiency, legal environment

Methodology used simple graphs, correlations, and regressions Data coverage data on a cross-section of up to 150 countries

stock markets also tend to be larger, more active and more efficient; overall financial system becomes larger, more active and more efficient; and domestic stock markets tend to become more active relative to domestic banks in richer countries. In higher income countries, financial system tends to be more market-based.

5) Fisman and Love (2003)

Main tasks or assessment

identifying the distinction between the role of financial development on industry growth in the short-run and the long-run.

Methodology Least Squares estimation with specific country and time effects Data coverage data from a panel of 37 industries and 44 countries.

Major result an improvement in financial development will result in an increased responsiveness to global growth shocks; in the short-run, financial development would facilitate the reallocation of resources to any industry with high growth potential; and actual growth is more highly correlated with the measure of growth opportunities in economies with higher financial development.

6) Beck, Demirguc-Kunt, Laeven and Levine (2006)

Main tasks or assessment

assessing whether financial development boosts the growth of small firms more than large firms and hence provides information on conflicting theoretical predictions about the distributional effects of financial development.

Methodology both Least Square estimation and Instrumental Variable (IV) approach are used Data coverage 44 countries and 36 industries in the manufacturing sector

Major result small-firm industries grow disproportionately faster in economies with well-developed financial systems

7) Garretsen, Lensink and Sterken (2004)

Main tasks or assessment

investigating whether informal institutions together with formal institutions are relevant in explaining cross-country differences in financial development and its effect on economic growth.

methodology both panel estimation and cross-section estimation methods and Instrumental Variable (IV) approach.

Data coverage data on societal variables from 43 countries

Major result societal norms are significant determinant of stock market capitalization but not of bank credit.

3.2.

Discussion

and generation/estimation and use of additional explanatory variables such as international capital markets integration and measure of volatility in stock market size.

They use the following stock market development indicators. (1) Stock market size (or Capitalization) equals the value of listed domestic shares on domestic exchanges divided by GDP. (2) Stock markets turnover equals the value of the trades of domestic shares on domestic exchanges divided by the value of listed domestic shares. (3) Value traded equals the value of the trades of domestic shares on domestic exchanges divided by GDP. While capitalization indicates market size turnover and value traded are liquidity indicators. As an indicator of banking development they use bank credit measured as the value of loans made by commercial banks and other deposit banks to the private sector divided by GDP. They argue that bank credit improves upon traditional financial depth measures of banking development (measured by broad money to GDP ratio) by isolating credit issued by banks, as opposed to credit issued by the central bank or other intermediaries, and by identifying credit to the private sector, as opposed to credit issued to governments.

They estimated three additional variables. Two of them are measures of international integration: (i) computed using international capital asset pricing model called CAPM integration and (ii) computed using international arbitrage pricing theory called APT integration. The third one is volatility of stock returns as twelve-months rolling standard deviation estimate that is based on market returns. They use least-squares regressions to study the ties between the growth indicators and measures of banking development, stock market liquidity, market size, and stock return volatility. They apply an Instrumental Variables (IV) approach to examine the links between the growth indicators, banking development, and measures of capital market integration. They use the IV approach because the international integration measures are estimated regressors and thus very likely exhibit endogeneity.

Moreover, the level of banking development, as measured by bank loans to private enterprises divided by GDP, also enters these regressions significantly. Banking development and stock market liquidity are both good predictors of economic growth, capital accumulation, and productivity growth.

The other stock market indicators do not have a robust link with long run growth. Volatility is insignificantly correlated with growth in most specifications. Similarly, market size and international integration are not robustly linked with growth, capital accumulation, and productivity improvements. Finally, they find that none of the financial indicators is robustly related to private saving rates. Moreover, they do not find any support for theories that more liquid or more internationally integrated capital markets negatively affect saving and growth rates or that greater liquidity retards productivity growth. Their evidence does not support the belief that stock return volatility hinders investment and resource allocation.

Levine, et al (1999) examines the relationship between financial intermediaries’ development and economic growth. They use data of 71 countries over the period of 1960 to 1995. They totally focus on banking sector. They address the issue of causality and also provide suggestive evidence concerning the determinants of financial development. Specifically, they evaluate whether the level of financial intermediary development exerts a causal influence on economic growth; and whether cross-country differences, in particular legal and accounting system characteristics, explain cross-country differences in the level of financial intermediary development. They used three major indicators of financial intermediary development. These are (a) Liquid liabilities, which equal currency plus demand and interest-bearing liabilities of banks and nonblank financial intermediaries divided by GDP; (b) the ratio of commercial bank assets to the sum of commercial banks and central bank assets; and (c) private credit which equals the value of credits by financial intermediaries to the private sector divided by GDP.

They used two major approaches: one purely cross-sectional estimator and panel data estimators. For the pure cross-sectional analysis they averaged the data over the period of 1960 to 1995 and have only one observation for each country. They estimate the following regression model:

i i i

i

FINANCE

CONDITIONI

NG

SET

country i; [CONDITIONING SET]i is a vector conditioning information that controls for other factors associated with economic growth; and

ε

_i is the error term. In the panel procedure a five-year average is used. They employed a dynamic panel GMM model of the following form:

t i i t i t i t i

y

X

y

_,

=

α

_,−1

+

β

'

_,

+

η

+

ε

_, (3.2) where

y

_i,tis logarithm of real per capita GDP of country i in period t;

X

_i,tis a set of other explanatory variables of country i in period t (other than lagged per capita GDP);

η

_i is an unobserved country specific effect;

ε

it is the error term; and i and t represent country and time,

respectively. To eliminate country specific unobserved effects, they use first-order-differences and obtain the following.

)

(

)

(

'

)

(

_{, 1 , 2 , , 1 , , 1} 1 , ,t

−

it−

=

it−

−

it−

+

it

−

it−

+

it

−

it− i

y

y

y

X

X

y

α

β

ε

ε

(3.3)

They used instrumental variable (IV) estimation method to account for the likely endogeneity of explanatory variables and correlation between the new error term and the dependent variable that arises from construction in taking the first-difference. The estimation results of both models show the same thing. The cross-sectional model result shows that each of the three financial intermediary development indicators (liquid liabilities, ratio of commercial bank asset to total banks assets and private credit) is significantly correlated with economic growth at 5% significance level. Their dynamic panel estimate also suggests that financial intermediary development exerts a large and positive causal impact on economic growth.

Methodologically they employ the same approach as Levine, et al (1999) we discussed above and come up with similar results. Their result shows that both stock market development, measured in terms of stock market turnover ratio, and bank development, measured by bank credit to private sector, have both statistically and economically large positive impact on economic growth. The impact of stock markets turnover is found to be stronger than that of bank credit. For instance, bank credit does not always enter the regression significantly.

Demirguc-Kunt and Levine (2001) discuss various aspects of financial market and economic growth. They use data on a cross-section of up to 150 countries to illustrate how financial systems differ as one compares poorer with richer countries. Without going to a detailed analysis of specific issue their discussion covers relatively larger number of issues compared to other works on finance and economic growth. One of the extensions they produced is their classification of many countries’ financial structure into either market-based or bank-based. Their discussion covers non-bank financial intermediaries, financial market structure (classification into market-based and bank-market-based, included in the appendix A), equity markets across countries, financial markets size, efficiency, legal environment including legal origin, legal codes, enforcement and corruption, regulatory environment such as accounting system, bank regulations, deposit insurance and taxes, and inflation.

They used simple graphs, correlations, and regressions to illustrate the relationship between financial structure and economic development. As opposed to rigorous test of specific hypothesis, their main goal was to compile and compare different indicators of financial structure, identifying certain interesting patterns and highlight (suggestive) correlations. Some of these patterns and correlations they come up with are the following. Banks and other financial intermediaries tend to be larger, more active, and more efficient in higher income countries than they are in lower income countries. Stock markets also tend to be larger, more active and more efficient in higher income countries. The overall financial system becomes larger, more active and more efficient in higher income countries. Domestic stock markets tend to become more active relative to domestic banks in richer countries. Their conglomerate index of financial structure shows that in higher income countries, financial system tends to be more market-based.

positive link between corruption and financial underdevelopment. Countries with lower level of corruption tend to have more market-based financial systems. Countries with strong accounting standards tend to have market-based financial systems. Countries with explicit deposit insurance systems are less likely to have market-based financial systems. There is not a strong link between financial structure and tax distortions favoring either dividends or capital gains relative to interest income. High inflation economies are much more likely to have underdeveloped financial systems, but inflation is not strongly linked with whether a country’s financial system is bank-based or market-bank-based.

Fisman and Love (2003) also provide an empirical analysis on the relationship between financial development and growth. They emphasize the distinction between the role of financial development on industry growth in the short and the long run. In their empirical approach they assess two major aspects of finance and growth. The first one is the relationship between industry growth and growth opportunities; the second one is relationship between industry share and needs for finance. Because of well-developed financial market institutions, they used United States Growth (USGrowth) as proxy for worldwide shocks to growth opportunities. In the analysis of the first aspect, they test whether financial development facilitates efficient responses to these shocks at time t using the following model.

Growthict = αi + αc + FDc *USGrowthit + εic (3.4) where α’s represent industry i and country c specific shocks; FD is financial development; εic is error term; and i, c and t, respectively, stands for industry, country and time. Growthict is calculated as: Growthict = USTradeict *USGrowthit. In the second aspect, they used a measure of external finance dependence constructed by Rajan and Zingales (1998), which they call USNeedsi and assume that: