Thorsten Leo Beck (World Bank)

1
Financial DevelopmentEconomic Growth Nexus A
Case Study of Bangladesh

• By
• Thorsten Leo Beck (World Bank)
• and
• M. Habibur Rahman (Bangladesh Bank)

Preliminary Draft Comments and
Suggestions are Welcome
2
Plan of the Presentation

• Two parts
• One Existing literature on Finance-Growth debate
  followed by a sophisticated econometric analysis
  to establish the view that financial development
  is an important factor for economic growth in
  Bangladesh
• Two Analysis of figures and facts identifying
  the causative factors behind financial
  development in Bangladesh

3
Motivation (Finance-Growth Debate)

• Why some countries are developed and some are not
  is a MYSTERIOUS question for development
  economists
• Better infrastructure, institutions, technology,
  more capital could be the possible answer
• But again one can pose another question why they
  are better?

4
Motivation..cont.

• The Role of Financial Intermediation that
  facilitate most of the ingredients for economic
  growth seems to be very important factors in
  newly emerging economies
• The intention, therefore, is to investigate the
  role of Financial Development on capital
  formation and economic growth in light of
  Bangladesh economy

5
From Market Frictions to Economic Growth A
Theoretical Approach to Finance and Growth
6
(No Transcript)
7
LiteratureFinance to Growth

• Goldsmiths (1969) paper on 35 countries is the
  first empirical study that investigates
  finance-growth link
• King-Levine (1993a), Levine (1997 1999),
  Levine-Zervos (1998), Rajan-Zingales (1998),
  Beck-Levine-Loayza (2000)

8
LiteratureFinance to Growth

• Theoretical papers, such as Bencivenga-Smith
  (1991), Diamond (1984), and Williamson (1996
  1998) explain various channels through which
  financial development could contribute positively
  to economic growth

9
LiteratureFinance to Growth

• Studies based on time series technique, such as
  Demetriades-Hussein (1996), Hansson-Jonung
  (1997), Luintel-Khan (1999), and Shan et al.
  (2001) are dominated with the evidence of
  bi-directional causality.

10
LiteratureEconomic growth to financial
development

• Other studies, such as Deveraux-Smith (1994),
  Jappelli-Pagano (1994), Singh (1997),
  Arestis-Demetriades (1997) and Singh-Weisse
  (1998) including Robinson (1952) argue that
  financial development may not always promote
  economic growth.
• They show that depending on the stage of
  development economic growth may promote financial
  development. To the contrary of the previous
  literature they argue that economic development
  generates additional demand for financial
  services and hence establishes a more developed
  financial sector. According to their view
  economic growth leads and financial development
  follows.

11
LiteratureFinance-Growth Joint Evaluation

• Some other papers, however, including Gurley-Shaw
  (1955), Greenwood-Jovanovic (1990), Galetovic
  (1996), Geenwood-Smith (1997), and
  Bencivenga-Smith (1998) observe inextricable link
  between financial development and economic
  growth. They experience both way causality
  between financial development and economic
  growth. They predict joint evolution of the real
  and financial sectors during the growth process.
  They argue that at the initial stage of economic
  development finance follows economy. After a
  certain threshold level when financial
  intermediaries emerge, economy starts to get
  benefit from the financial sectors.

12
Objective

• The main objective of this study is to
  investigate the causal relationship between
  financial development and economic growth in
  Bangladesh, particularly the long-run impact of
  financial development on capital formation and
  per capita income.
• A system of equations based on the hypothesis
  that financial development has long-run impact on
  investment and per capita income is specified and
  estimated using Blanchard-Quahs (1989) technique
  of structural vector autoregressions (SVARs).

13
Objective

• To examine the short-run dynamics among the
  variables in the system, however, the impulse
  response functions (IRFs) and variance
  decomposition (VDCs) are computed based on
  Cholesky factorization where the standard errors
  for VDCs are computed through 1000 Monte Carlo
  simulations.
• To substantiate the causal link among the various
  indicators of financial development, investment
  and income per capita a graphical presentation
  has also been used.

14
An Overview of Financial Development in
Bangladesh

• As financial development lacks any precise
  definitions, following the practice of existing
  literature King-Levine (1993a and 1993b), Levine
  (1997 and 1999), and Levine-Zervos (1998) some
  indicators of financial development may be used
  for effective policy formulation, implementation
  and evaluation.
• Accordingly, three alternative indicators of
  financial development, such as domestic credit to
  the private sector by banks to GDP ratio, total
  deposits to GDP ratio and broad money (M2) to GDP
  ratio for Bangladesh economy have been used.

15
An Overview of Financial Development

• Domestic credit to the private sector as a
  percent of GDP (denoted by cr_y) is one of the
  popular indicators of financial development.
• The second indicator of financial development is
  total deposits (demand plus time) as a percent of
  GDP (denoted by dep_y) which is relatively
  broader measure of financial development as it
  includes all the liquid liabilities of the
  financial system excluding currency.
• A third indicator, broad money as a percent of
  GDP (denoted by m2_y) is basically the liquid
  liabilities of the financial system in Bangladesh
  that includes currency plus demand and
  interest-bearing liabilities of financial
  intermediaries.

16
An Overview of Financial Development
Period lr cr_y dep_y m2_y i_y y_pcap
1976-1980 11.09 6.59 14.86 19.03 10.44 160.0
1981-1985 13.68 13.67 20.23 24.54 10.51 192.0
1986-1990 14.71 19.08 24.75 28.67 13.87 242.0
1991-1995 13.90 16.58 23.07 26.68 17.93 283.0
1996-2000 13.83 23.17 26.7 31.01 21.51 353.0
2001-2005 12.33 28.83 35.08 40.02 22.63 395.0
17
An Overview of Financial Development

• It has been observed from the Table that the
  average credit, deposit and broad money to GDP
  ratios increased substantially respectively from
  6.6 percent, 14.9 percent and 19.0 percent in
  1976-1980 to respectively 28.8 percent 35.01
  percent and 40.0 percent in 2001-2005.
• Investment as a percent of GDP and per capita
  income (in current USD) also display a similar
  pattern and move broadly together reflecting a
  close association among financial development,
  investment and per capita income during the
  period

18
An Overview of Financial Development
19
An Overview of Financial Development
20
An Overview of Financial Development
21
An Overview of Financial Development
22
An Overview of Financial Development

• The scatter-plots of the three indicators of
  financial development vis-à-vis investment as
  well as per capita income strongly supports the
  co-movement of financial development and economic
  activity.
• Besides, almost a linear relationship is also
  observed in a scatter-plots between
  investment-GDP ratio and per capita income.

23
Methodology

• Structural macroeconometric models, such as the
  Klein interwar model, the Brooking model, the BEA
  model, the St. Louis model and the Taylor model
  that are based on hundreds of equations are
  replaced by the vector autoregressions (VARs).
  The problem of identification and endogeneity are
  associated with these structural macroeconometric
  models which can easily be overcome by the VARs
  approach (Simss 1980)

24
Methodology

• Because it does not impose any a priori
  restrictions and is based on reduced form
  equations, it is difficult to reconcile VARs with
  economic theory and to provide any meaningful
  interpretations of the estimated parameters
• In order to overcome the above difficulties with
  the standard unrestricted VARs some studies, such
  as Bernanke (1986), Blanchard-Watson (1986) and
  Sims (1986) come up with a structural VARs
  (SVARs) model that allows contemporaneous
  structural restrictions

25
Methodology

• As the objective of this paper is to investigate
  long-run relationship between financial
  development and economic growth in Bangladesh, a
  Blanchard-Quah (1989) type of long-run structural
  model is estimated
• To examine the short-run dynamics among the
  variables in the system, however, the impulse
  response functions (IRFs) and variance
  decomposition (VDCs) are computed based on
  Cholesky factorization

26
Methodology

27
Methodology

• The restrictions stated in previous slide have
  some interesting implications regarding financial
  development-economic growth relationship
• it asserts financial development has long-run
  effect on investment and per capita income
• Income per capita, on the other hand, has no
  long-run effect on financial development.

28
Preliminary data analysis
Variables (in natural log) without trend without trend without trend with trend with trend with trend Decision
Variables (in natural log) DF PP KPSS DF PP KPSS Decision
Rate Lending rate (lr) f Lending rate at 1st difference (dlr) f I(1) I(0) I(1) I(0) I(0) I(0) I(1) I(0) I(1) I(0) I(1) I(0) I(1) I(0)
Financial development Domestic credit to the private sector as a percent of GDP (cr_y) Total deposit as a percent of GDP (dep_y) Broad money as a percent of GDP (m2_y) I(0) I(0) I(1) I(0) I(0) I(1) I(1) I(1) I(1) I(0) I(0) I(0) I(0) I(0) I(0) I(0) I(0) I(0) I(0) I(0) I(0)
Investment Per capita gross fixed capital formation as a percent of GDP (i_y) I(1) I(0) I(1) I(1) I(0) I(0) I(0)
Income Per capita GDP at current USD (y_pcap) I(1) I(1) I(1) I(1) I(0) I(0) I(0)
Notes f without log, I(1) unit-root and I(0) stationary. Lag length for ADF tests are decided based on Akaikes information criterion (AIC). Maximum Bandwidth for PP and KPSS test are decided based on Newey-West (1994). All the tests are performed on the basis of 5 significance level. Notes f without log, I(1) unit-root and I(0) stationary. Lag length for ADF tests are decided based on Akaikes information criterion (AIC). Maximum Bandwidth for PP and KPSS test are decided based on Newey-West (1994). All the tests are performed on the basis of 5 significance level. Notes f without log, I(1) unit-root and I(0) stationary. Lag length for ADF tests are decided based on Akaikes information criterion (AIC). Maximum Bandwidth for PP and KPSS test are decided based on Newey-West (1994). All the tests are performed on the basis of 5 significance level. Notes f without log, I(1) unit-root and I(0) stationary. Lag length for ADF tests are decided based on Akaikes information criterion (AIC). Maximum Bandwidth for PP and KPSS test are decided based on Newey-West (1994). All the tests are performed on the basis of 5 significance level. Notes f without log, I(1) unit-root and I(0) stationary. Lag length for ADF tests are decided based on Akaikes information criterion (AIC). Maximum Bandwidth for PP and KPSS test are decided based on Newey-West (1994). All the tests are performed on the basis of 5 significance level. Notes f without log, I(1) unit-root and I(0) stationary. Lag length for ADF tests are decided based on Akaikes information criterion (AIC). Maximum Bandwidth for PP and KPSS test are decided based on Newey-West (1994). All the tests are performed on the basis of 5 significance level. Notes f without log, I(1) unit-root and I(0) stationary. Lag length for ADF tests are decided based on Akaikes information criterion (AIC). Maximum Bandwidth for PP and KPSS test are decided based on Newey-West (1994). All the tests are performed on the basis of 5 significance level. Notes f without log, I(1) unit-root and I(0) stationary. Lag length for ADF tests are decided based on Akaikes information criterion (AIC). Maximum Bandwidth for PP and KPSS test are decided based on Newey-West (1994). All the tests are performed on the basis of 5 significance level.
29
Empirical Results

30
Empirical Results
31
Empirical Results
Variance Decompositions of Financial Development Variance Decompositions of Financial Development Variance Decompositions of Financial Development Variance Decompositions of Financial Development Variance Decompositions of Financial Development
Time Horizon (Year) Explained by shocks in Explained by shocks in Explained by shocks in Explained by shocks in
Time Horizon (Year) Lending Rate Financial Development Investment Income per Capita
4 8 12 16 20 27.61 (-16.84) 52.89 (-17.81) 63.72 (-18.80) 70.36 (-19.54) 70.72 (-19.77) 40.57 (-19.77) 21.60 (-16.48) 18.84 (-17.16) 12.72 (-17.10) 14.63 (-17.72) 31.30 (-17.74) 20.17 (-16.43) 12.40 (-16.36) 11.15 (-16.54) 8.64 (-16.90) 0.52 (-3.75) 5.33 -(6.28) 5.04 (-6.15) 5.77 (-6.26) 6.01 (-5.92)
32
Empirical Results
2. Variance Decompositions of Investment 2. Variance Decompositions of Investment 2. Variance Decompositions of Investment 2. Variance Decompositions of Investment 2. Variance Decompositions of Investment
4 8 12 16 20 31.86 (-16.95) 43.29 (-17.84) 50.09 (-18.92) 62.32 (-19.59) 57.43 (-19.89) 2.02 (-14.64) 5.95 (-16.17) 14.15 (-16.75) 15.56 (-17.01) 28.07 (-17.84) 61.57 (-19.07) 45.68 (-18.21) 31.19 (-17.60) 17.73 (-17.41) 10.56 (-17.38) 4.54 (-5.01) 5.08 (-5.91) 4.57 (-5.50) 4.40 (-5.71) 3.94 (-5.73)
33
Empirical Results
3. Variance Decompositions of Income per Capita 3. Variance Decompositions of Income per Capita 3. Variance Decompositions of Income per Capita 3. Variance Decompositions of Income per Capita 3. Variance Decompositions of Income per Capita
4 8 12 16 20 13.92 (-16.54) 41.35 (-18.80) 66.24 (-19.64) 70.69 (-19.67) 65.60 (-20.14) 38.56 (-17.96) 28.39 (-17.04) 11.94 (-17.24) 14.13 (-17.32) 20.35 (-17.57) 33.56 (-16.22 19.56 (-16.14 15.14 (-16.46 8.94 (-16.86 8.34 (-17.00 13.96 (-6.90) 10.70 (-6.53) 6.68 (-5.94) 6.24 (-5.99) 5.71 (-5.92)
34
Summary and Conclusion

• The graphical presentation as well as estimated
  coefficients of the long-run response matrix
  indicates that various indicators of financial
  development and investment have long-run impact
  on per capita income
• The estimated results also support the argument
  that in the long-run financial development
  stimulates investment activities. The estimated
  coefficients of the long-run response matrix,
  however, do not provide any statistical evidence
  that the lending rate has any impact on financial
  development, investment or on per capita income

35
Summary and Conclusion

• Regarding the short-run dynamics among the
  variables in the system, the results from IRFs
  indicate that both the financial development and
  investment have short-run impact on per capita
  income at the immediate year of initial shocks
• The results from VDCs, on other hand, imply that
  all the variables in the system, such as lending
  rate, indicator of financial development and
  investment contain very useful information in
  predicting the future path of per capita income

36

• Thank You