Transcript Financial Development and Economic Growth in Southern

Financial Development and Economic Growth
in Southern Africa
Presented by
Meshach Aziakpono
Department of Economics
National University of Lesotho
E-mail: [email protected]
At
OECD Conference in Johannesburg:: March 25- 26
SUMMARY OF ISSUES AND MAJOR
FINDING
•
ISSUES:
- The Debate on FI and Econ. Growth
- Alternative views on the link:
(a) Mobilizing res. and ensuring an efficient transfn. Of funds into real
productive capital
(b) Transform maturity of portfolios of savers and investors > providing
sufficient liquidity
© Risk reductn through risk diversification.
(d) sharing, and pooling.
- The increasing International interest in Economic Integration and Monetary
Union
– Led to increase Capital Mobility
- The increasing call for Financial Liberalization (International Financial Inst. and
many Economists)
THE QUESTIONS:
• Whether national fin. Mkts. still matter for growth once
domestic agents have access to foreign fin. Mkts.
• Who gains and losses?
• Why the gains to gainers and losses to losers
WHY THE SACU COUNTRIES:
•
-
SACU and CMA have been cited as examples of successful
Economic Integration in Africa (ADB 2000 and Jenkins and Thomas
1998)
• The Financial Sectors of Member Countries are highly integrated
(Jenkins and Thomas 1998)
•
Therefore if Local Financial Development matters for the Growth of
SACU Countries, it could be concluded that such Financial
Development alongside Regional Financial Integration will continue to
influence future National Growth Rates.
The REST OF THE PRESENTATION
1. Brief Economic Background of the SACU Countries
1.1
Official Integration Arrangements and Implication for the
BLNSS Countries
1.2 Banking in the SACU Countries
1.3 Financial Intermediation
1.4 Economic Performance among the BLNSS Countries
2. Framework for the Analysis
2.1 Variables and data sources
3. Empirical Results
3.1 Effects of FI indicators
3.2 Effects of other variables
4. Conclusion
Brief Economic Background of the SACU
Countries
1.1
Official Integration Arrangements and Implication
for the BLNSS Countries
-The Southern African Customs Union (SACU)
-Free movement of goods
-Free movement of people (large remittances from migrant labour)
-Common external tariff
-RSA compensate the partner countries for loss of fiscal
discretion….
-Share of tariff revenue
Table 1: SACU Revenue Payments, 2001/02
T a ble 1 : S A C U R e ve n ue P ay m e n ts, 2 001 /02
B o tsw a
na
S A C U P a ym en t 2 ,6 22
(R m illion )
%
of
tota l 1 4.5
R even u e P ool
% of tota l G ovt. 1 2.8
R even u e
(exc l.
G ra n ts)
L e so th o
N a m ib ia
S w az ila n d
1 438
2 ,6 41
1 ,5 03
RSA
(r e sid u
a l)
9 ,8 97
7 .9
1 4.6
8 .3
5 4.7
5 1.0
3 0.4
5 4.1
3 .9
S o u rc e : K irk, R an d M . S te rn (2 0 0 3 ) T he N e w S o u th e rn A fric an C u sto m s U n io n A g re e m e n t, W o rld B a n k, P a g e 5 .
CONTD.
• (b) CMA (RMA)
• - Country issue currencies but must be backed and pegged at par with
RSA Rand
• - Countries can establish their central banks
– Swaziland - 1974 (MA); 1979 (CBS and Lilanggeni at par with Rand)
– Lesotho - 1979 (MA); 1980 (Loti) 1982 (CBL)
– Botswana: 1976 (left RMA and introduced Pula linked to Rand in a
basket)
– 1986 - CMA
• Trilateral agreement (RSA, Lesotho and Swaziland)
• Namibia- 1992 (CMA); 1993 (Nam. Dollar) pegged at par with Rand
• Free flow of funds from current and capital account transactions
– Implication - Capital flow to any country where it would earn the highest returns
– Monetary policy follow RSA.
II
Banking in the SACU
• BLNS dominated by RSA banks
• Banks in BLNS follows trend in RSA (CMA and Bank
ownership)
• RSA banks well developed, sophisticated and highly
competitive
• BLNS
–
–
–
–
High mkt concentration
High cost of capital for entrepreneur
High spread (except Botswana)
High excess liquidity - due to low domestic invest and not high
savings
– Private sector credit shift to HH for consumption
– Individuals and firms operates dual accounts = (low Deposit rates,
Banking inefficiencies, high charges for services)
Table B-1:
Banking in the SACU Countries
Country
Number of Commercial Banks
Origin of Banks
Botswana
Five:
Dominated
Lesotho
-
Barclays Bank
-
First National Bank
-
Stanbic Bank
-
Standard Chartered Bank
-
Bank of Baroda
by
Banks
from South Africa
Three:
Dominated
by
-Standard Bank of Lesotho
African Banks.
South
-Nedbank
-Lesotho Bank (1999) Ltd
Namibia
Four:
Namibia banking sector
-First National Bank of Namibia (Ltd) is dominated by South
(which merged with SWABOU Bank African Banks. The only
South Africa
in 2003)
fully
Namibian-owned
-Standard Bank of Namibia (Ltd)
bank, SWABOU Bank
-Commercial Bank of Namibia (Ltd)
merged
-Bank of Windhoek (Ltd)
National Bank in 2003.
with
First
- 42 Registered Banks as at December
2001
-3 Mutual Banks as at 31 December
2001
-15 Local branches of foreign banks as
at 31 December 2001
-61
Foreign
banks
representative
offices
with
as
local
at
31
December 2001.
Swaziland
Three:
Dominated
Nedbank
banks from South Africa.
Standard Bank
by
foreign
Table 2: Financial Intermediation
Table: 2
Sample: 1980 – 2000: Balanced panel.
Country
FI1
FI2
FI3
FI4
IDR
Mean Std.D. Mean Std.D. Mean Std.D. Mean Std.D. Mean Std.D.
Botswana
Lesotho
Namibia
South Africa
Swaziland
46.63 7.45 79.54 15.08
17.83 3.85 34.28 4.11
33.80 21.09 35.97 6.84
57.76 9.50 49.72 6.26
20.80 4.12 28.13 3.59
72.30
66.77
86.47
89.05
79.99
10.92
11.29
7.48
2.16
2.46
117.00
99.52
100.53
117.15
100.54
15.17
11.47
14.84
12.31
10.06
3.26 1.73
7.19 2.59
8.31 1.33
4.46 1.04
6.80 0.91
Note: FI1 is the ratio of credit extended to the private sector by commercial banks to GDP; FI2 is the ratio
of liquid liabilities of commercial banks to GDP, where liquid liabilities equals demand deposit plus time
and savings deposits; FI3 is the ratio of commercial bank assets to the sum of commercial banks and central
banks’ assets; and FI4 is a composite index computed from the combination of the other three indexes.
Source: Aziakpono (2003)
Table 3:SACU Countries Basic Data,
Table 3:
SACU Countries Basic Data: 2001
Count Area Pop
GDP
GD Av.
HD
ry
(‘000s (mn)
($bn)
P
Growt I
q.km)
per h rate ran
capi GDP
k
ta $ per
capita
19902001
Botsw
ana
Lesoth
o
Namib
ia
RSA
600
1.7
5.2
306 2.5
(22.3) (3.3)
(3.6)
6
30
1.8
0.8
386 2.1
(1.1) (3.5)
(0.6)
824
1.9
3.1
173 2.2
(30.6) (3.7)
(2.2)
0
1221( 44.4(8 133.3
262 0.2
45.4) 7.2)
(92.8) 0
Swazil 17
1.1
1.3
117 1.9
and
(0.6) (2.2)
(0.9)
5
SACU 2692 50.9
143.7
249
(100) (100) (100)
1.4
Source: UNDP Human Development Report 2003
Life
Expe
ct. at
birth
(year
s)
Imp
ort
as
%
of
GD
P
Exp
ort
as
%
of
GD
P
125 39.7
Av.
Infl
atio
n
rate
199
0200
1
10.0
35
51
137 35.1
8.8
86
34
124 44.3
9.5
66
54
111 47.7
8.3
25
28
133 34.4
9.3
81
69
Table 4:
Country
Botswana
% growth
Per Capita GNP (current US $ and % compound annual growth rate)
1965 1970
70
140
14.9
1980
1985
1990
1996
430
25.16
1020
18.85
1120
1.88
2490
17.32
3210b
5.21
3070
3066
550
7.67
660
3.08
570
386
Lesotho
% growth
60
100 250
10.8 20.11
440
11.97
380
-2.88
Namibia
% growth
NA
NA
NA
NA
NA
1900 2250
9.08 2.85
South Africa 530
% growth
770 1590
7.75 15.60
2490
9.38
2100 2860
-3.35 7.87
Swaziland 180
% growth
230 570
5.02 19.90
910
9.81
760
1110
-3.53 7.87
3520
3.52
1210
1.45
1998
2001c
1975
1940
1730
3310
2620
1390
1175
Source: World Bank, World Development Indicators 1998 in Allen and Ndikumana (1998) and African
Development Bank, African Development Report 2000
GNP is in current US$, World Bank’s Atlas method.
b
For Botswana, GNP for 1996 is missing; the value reported here is for 1995
c GDP per capita, obtained from UNDP, Human Development report, 2003.
2.
Framework for the Analysis
2.1 Variables and data sources
Two Indicators of FI
- FIC – Ratio of Private Credit to GDP
- FIL - Ratio of Liquid liabilities to GDP
-Dependent Variable - log of real GDP and Growth of real GDP
- Control Variables
- Inflation,
- Size of government = Govt. Exp. /GDP
- Openness to trade = (export + import)/ GDP
- Exchange rate.
- All the data came from IMF International Financial Statistics 2001
and earlier issues.
Econometrics Techniques
-
The Panel Econometrics Analysis applied the Zellner seemingly
unrelated regressions estimation (SURE) method.
-
The main features of SURE as a method for pooling time-series
and cross-sectional data are:
- The assumption of contemporaneous correlation in the
disturbances, and
-That each cross-sectional unit has a different coefficient vector
- The objective of the econometric technique is to overcome two
major weaknesses of most cross-country approaches:
- These approaches give all countries, either small or large, an
equal weighting since they are assumed to be
homogeneous; and
- The coefficients represent only an average relationship, which
may or may not apply to individual countries in the sample
(Bloch and Tang 2003: 250).
RESULTS
• Effects of FI indicators
• Overall: weak positive effects of FI
– Allan and Ndikumana (1998) = reflect pervasive inefficiencies in
credit allocation mechanism- poor legal and banking supervision
– Country -specific
• South Africa -highest gains
– Efficiency in deposit utilization
– Positive externalities
Results contd.
• Botswana- weak effects
– “Demand following finance”
– Reflects general shift of resources from productive
invest to consumption
• Lesotho: some positive effects but not strong
– Most credit not used to fin dom. Private invest
– Restrictive govt policies
• Credit ceiling by CBL to control MS (1986-1996)
• MLAR (1981-2000) - risk free invest opp for banks
– Weak institutional, structural and legal environment
• confusion on property right
• weak and slow legal system to enforce contract and debt
repayment
• non-loan repayment culture
Results cont.
• Swaziland
Swaziland’s case looks very gloomy
It appears the role of FI in promoting Growth is becoming
less important
• Problems similar to Lesotho
• Attributable to negative externalities, especially since
political and economic stability return to the RSA.
•
•
•
Table:5 Empirical Results
Table 5: Financial Intermediation and Economic Performance- SURE Estimation Results
Country /
Variable
Botswana
FIL
FIC
Lesotho
FIL
FIC
South Africa
FIL
FIC
Swaziland
FIL
FIC
M odel 1 (GRY)
M odel 2 (LRY)
-0.042 (-0.49)
-1.71 (-7.32)a
0.02 (9.05)a
0.019 (3.48)a
0.017 (0.84)
0.297 (1.34)
-0.011 (-6.89)a
0.0004 (0.19)
0.51 (4.8)a
0.018 (1.79)a
0.01 (14.69)a
0.002 (6.13)a
-0.41 (-.41)
-0.57 (-1.57)
-0.005 (-0.97)
-0.001 (-0.14)
Note: a-significant at 1% level of significance, b-significance at 5% level of significance and c- significance at
10% level of significance.
Effects of the controlled variables
Ø Openness has a significant negative on growth in most
of the SACU, except for South Africa where it was
positive and significant
Ø Government Expenditure had a mixed signs.
o For Botswana it was positive and Significant
o For the remaining countries, the negative coefficient
dominates and was significant in most cases
Ø Inflation was negative in almost all the countries but
was insignificant
Ø The coefficients of Exchange rate were positive and
Significant in most of the equations.
4. The Untold Story
• Are there any other possible effects of the
integration arrangement? (positive or negative)
– yes
– The credibility effects on the smaller countries
Table A-1: Covariance tests for regression-coefficient homogeneity across-sectional countries
Equation /
Degrees of Freedom
Hypothesis
F-Stat Numerator Denominator
F (table-value)
Decision
Eq1a: GRY (FIL)
H1
8.05
H2
2.87
Eq1b: GRY (FIC)
H1
8.45
H2
2.94
Eq1c: LRY (FIL)
H1
592.4
H2
9.16
Eq1d: LRY (FIC)
H1
381.66
H2
12.05
15
27
216
216
2.13
1.29
Reject
Reject
15
27
216
216
2.13
1.79
Reject
Reject
15
27
216
216
2.13
1.79
Reject
Reject
15
27
216
216
2.13
1.79
Reject
Reject
Note: H1 is the hypothesis of overall homogeneity and H2 is the hypothesis of slope homogeneity. The tests were based on
the estimated model using least squares method. Eq1 uses growth in real GDP as the dependent variable, and FIL and FIC
(1) were included among the explanatory variables in the option (a) and (b) respectively; while in the
same vein Eq2 uses the log of real GDP as the dependent variable.
Table B-2 Seemingly Unrelated Regression Estimation Results for Growth in Real GDP
Equation (1) Botswana
Constant
-11.52 (-1.47)
FIL (-4)
-0.042 (-0.49)
OPN (-2)
-0.33 (-2.98)a
GEY(-4)
2.72 (4.96)a
INF(-4)
-1.75 (-1.5)
ER(-1)
3.77 (1.1)
SER
17.611
Adj.R2
0.491
Lesotho
24.2 (4.1)a
0.017 (0.84)
-0.44 (-5.13)a
0.175 (0.796)
0.427 (1.03)
0.77 (0.89)
6.37
0.494
South Africa
11.19 (1.8)c
0.51(4.8)a
0.15 (2.01)b
-0.17 (-0.81)
-0.18 (-2.87)
0.59 (1.26)
2.416
.448
Swaziland
24.56 (1.95)b
-0.41(-1.07)
-0.22 (-1.44)
1.06 (1.31)
-0.26 (-0.48)
-2.42 (-1.52)
13.585
0.364
Note: a-significant at 1% level of significance, b-significance at 5% level of significance and c- significance at
10% level of significance.
Table B-3: Seemingly Unrelated Regression Estimation Results for Growth in Real GDP
Swaziland
South Africa
Lesotho
Equation (2) Botswana
27.14 (3.14)a
15.44 (2.42)b
17.79 (1.59)
0.35 (0.17)
Constant
-0.57 (-1.57)
0.018 (1.79)c
0.297 (1.34)
-1.71(-7.32)a
FIC (-4)
-0.43(-2.24)b
0.185 (2.45)b
-0.26 (-2.91)a
-0.41 (-3.97)a
OPN (-2)
1.57 (1.66)c
-0.38 (-1.75)c
0.64 (2.48)a
4.81 (8.61)a
GEY(-4)
-0.396(-0.63)
-0.19 (-2.96)a
0.93 (2.0)b
-1.01 (-0.71)
INF(-4)
-2.04 (-1.48)
0.137 (0.27)
-1.79 (-2.16)b
-1.54 (-0.86)
ER(-1)
13.267
2.44
6.799
13.552
SER
0.402
0.546
0.501
0.495
Adj.R2
Note: a-significant at 1%level of significance, b-significance at 5% level of significance and c- significance at
10% level of significance.
Table B-4: Seemingly Unrelated Regression Estimation Results for log of Real GDP
Equation (3)
Botswana
Lesotho
South Africa
Swaziland
Constant
FIL (-4)
OPN (-2)
GEY(-4)
INF(-4)
ER(-1)
SER
Adj.R2
7.68 (17.29)a
0.02 (9.05)a
-0.007 (-3.78)a
0.016 (1.37)
-0.025 (-0.76)
0.36 (6.012)
0.2438
0.7806
9.71 (78.05)a
-0.011 (-6.89)a
-0.012 (-9.64)a
-0.017 (-4.9)a
-0.0009 (-0.14)
0.098 (12.16)a
0.065
0.9536
11.74 (96.77)a
0.01 (14.69)a
0.003 (1.79)c
0.017 (4.49)a
-0.001 (-1.0)
0.069 (12.5)a
0.045
0.9335
8.65 (36.97)a
-0.005(-0.97)
-0.012(-3.9)a
-0.065(-4.26)a
-0.004(-0.72)
0.21 (13.86)a
0.1137
0.9333
Note: a-significant at 1% level of significance, b-significance at 5% level of significance and c- significance at 10% level
of significance.
Table B-5: Seemingly Unrelated Regression Estimation Results for log of Real GDP
Equation (4) Botswana
Lesotho
South Africa
Swaziland
Constant
7.96 (12.67)a 9.32 (54.42)a 12.182 (125)a 8.01 (84.02)a
FIC (-4)
0.019 (3.48)a 0.0004 (0.19) 0.002 (6.13)a
-0.001(-0.14)
OPN (-2)
-0.007(-2.83)a -0.008 (-9.5)a 0.003 (2.99)a
-0.01 (-2.99)a
GEY(-4)
0.029 (1.97)b -0.18 (-8.16)a 0.002 (0.56)
-0.05 (-2.51)a
INF(-4)
-0.04 (-1.03) 0.005 (1.52) -0.001(-1.62)
-0.007(-0.61)
ER(-1)
0.41 (11.35)a 0.109 (14.66)a 0.05 (9.02)a
0.25 (25.16)a
SER
0.246
0.0688
0.0296
0.098
Adj.R2
0.784
0.945
0.945
0.949
Note: a-significant at 1% level of significance, b-significance at 5% level of significance and c- significance
at 10% level of significance.