Academy of Economic Studies, Bucharest
Doctoral School of Finance and Banking
Financial Development and Economic Growth
in Romania
Dissertation Paper
Supervisor:
Professor Moisa Altar
MSc Student:
Mihai Tarau
CONTENTS
1. Introduction
2. Economic Growth-Financial Development Nexus
.
3. Theoretical Model of Analysis
4. Data Used and Econometric Estimations .
5. Conclusions
1. Introduction
Reasons why countries grow at different rates
The pozitive link between financial intermediation
and economic growth
Financial depth and monetization of economy, as
exogenous variables for output growth
Simplified mathematical model: ARDL(1,1) & ECM
Econometric estimations: OLS, UVAR & VECM
(The case of Romania output growth from
January 1992 to September 2001)
2. Economic growthfinancial development nexus
Adams (1819) asserted that banks harm the
“morality, tranquility, and even wealth” of nations
Hamilton (1781) argued that “banks were the
happiest engines that ever were invented” for
spurring economic growth
Robert Lucas (1988) dismisses finance as a major
determinant of economic growth
Merton Miller (1998): “financial markets
contribute to economic growth is a proposition
almost too obvious for serious discussion.”
Banks-Markets
Distiction and rivalries
Advantages and Disadvantages
It’s not banks or markets, it is banks and markets;
these different components of the financial system
ameliorate different information and transaction
costs.
Five Key Components (Prerequisites)
of a Performant Financial System:
• Sound public finances and public debt management
• Stable monetary arrangements
• Variety and influence of banks
• Credible Central Bank in stabilizing domestic finances
and managing international financial relations
• Well-functioning securities market
3. Theoretical Model of Analysis
ARDL (1,1):
(1)
(2)
(3)
Long-run relationship (y stable) requires |b|<1.
(4)
t is distributed independently from εt
ECM from ADL:
(1-b) = speed of adjustment
The long-run (steady-state) relationship implied by
the dynamic system in equations (1)-(4) is given by:
(7)
or
(8)
The main assumption is that there exist a single long-run
relationship between the endogenous and forcing variables. The
pre-requisites for consistent and efficient esti-mation are that the
shocks in the dynamic specification be serially uncorrelated and
that the forcing variables be strictly exogenous.
4. Data used and econometric estimations
The time series used are:
qind = monthly value of industrial production in ROL;
rtcrqind = (Δqindt/qindt-1) = industrial production monthly value
rate of growth (chain), as a proxy for GDP monthly rate of
growth;
fd = monthly credit to non-governments in ROL;
fdinq = (fd/q) = financial depth (credit to non-governments
weight in monthly industrial production);
rtcrfd = (Δfdinqt/fdinqt-1) = financial depth monthly rate of
growth (chain);
m2 = monthly value in ROL of monetary aggregate M2;
m2inq = (m2/q) = monetary aggregate M2 weight in monthly
industrial production;
rtcrm2inq = (Δm2inqt/m2inqt-1) = monthly growth rate of M2
weight in industrial production (chain).
Figure 1 :
1.2E+08
Time Series Used
5
2.5E+08
1.0E+08
2.0E+08
4
8.0E+07
1.5E+08
6.0E+07
3
1.0E+08
4.0E+07
2
5.0E+07
2.0E+07
0.0E+00
92 93 94 95 96 97 98 99 00 01
1
92 93 94 95 96 97 98 99 00 01
FD
0.0E+00
92 93 94 95 96 97 98 99 00 01
M2
FDINQ
5.0
600 00 00 0
0.6
4.5
500 00 00 0
0.4
4.0
400 00 00 0
0.2
3.5
300 00 00 0
3.0
0.0
200 00 00 0
2.5
-0.2
100 00 00 0
2.0
1.5
92 93 94 95 96 97 98 99 00 01
0
92 93 94 95 96 97 98 99 00 01
QIND
M2INQ
0.6
0.6
0.4
0.4
0.2
0.2
0.0
0.0
-0.2
-0.2
-0.4
92 93 94 95 96 97 98 99 00 01
-0.4
92 93 94 95 96 97 98 99 00 01
RTCRQIND
RTCRFD
-0.4
92 93 94 95 96 97 98 99 00 01
RTCRM2INQ
Granger Tests Conclusions:
Industrial production is the endogenous variable in any
relationship (at least under an one year lag) involving credit to nongovernments and monetary aggregate M2, as it do not Granger
cause any of the last two indicators.
On the other hand, Granger tests reveal an extremely stable
causality link between industrial production (as predicted variable)
and credit to non-governments or monetary aggregate M2 (as
predicting variables) in any lag of at most one year.
Credit to non-governments is, in his turn, Granger caused by the
monetary aggregate M2, all over one year period (thus having
another time stable causality relationship).
Financial depth rate of growth has a strong capacity of prediction
over the industrial production rate of growth, for periods from one
month to six months; the same relationship occur between monthly
growth rate of M2 weight in industrial production and the industrial
production rate of growth, but only for one quarter or one semester
lags.
Selected Regression of rtcrqind
Dependent Variable: RTCRQIND
Method: Least Squares
Date: 06/26/02 Time: 16:34
Sample(adjusted): 1994:01 2001:09
Included observations: 92
Excluded observations: 1 after adjusting endpoints
Variable
Coefficient
Std. Error
t-Statistic
Prob.
C
RTCRQIND(-24)
RTCRFD(-1)
RTCRFD(-3)
RTCRM2INQ(-3)
RTCRM2INQ(-24)
DUMMAR97
DUMJAN98
DUMJAN2000
0.050523
-0.357926
0.135798
-0.514537
0.638021
-0.415388
0.488365
0.455259
-0.331202
0.011701
0.159371
0.061531
0.146024
0.150882
0.153725
0.065471
0.065044
0.073907
4.317840
-2.245867
2.206980
-3.523658
4.228618
-2.702148
7.459235
6.999200
-4.481318
0.0000
0.0274
0.0301
0.0007
0.0001
0.0084
0.0000
0.0000
0.0000
R-squared
Adjusted R-squared
S.E. of regression
Sum squared resid
Log likelihood
Durbin-Watson stat
0.688115
0.658054
0.063748
0.337297
127.4523
1.856950
Mean dependent var
S.D. dependent var
Akaike info criterion
Schwarz criterion
F-statistic
Prob(F-statistic)
0.040367
0.109016
-2.575050
-2.328353
22.89045
0.000000
It can be also remarqed that the growth rates of financial
depth and M2 weight in industrial production (which explain
in the selected regression over 65% of growth rate from
industrial production) proved themselfes earlier to be relevant
regarding Granger causality tests.
15
Regression
10
5
Stability
Test:
0
-5
-10
-15
00:04
00:07
00:10
CUSUM
01:01
01:04
5% Significance
01:07
Figure 2 : Impulse Responses Functions of
rtcrqind, rtcrfd and rtcrm2inq
Respons e to One S.D. Innovations ± 2 S.E.
Re s p o n s e o f RTCRQIND to RT CRQIND
Re s p o n s e o f RTCRQIND to RTCRF D
Re s p o n s e o f RT CRQIND to RTCRM 2 INQ
0. 12
0. 12
0. 12
0. 08
0. 08
0. 08
0. 04
0. 04
0. 04
0. 00
0. 00
0. 00
-0. 04
-0. 04
1
2
3
4
5
6
7
8
9
10
11
12
-0. 04
1
Re s p o n s e o f RTCRFD to RT CRQIND
2
3
4
5
6
7
8
9
10
11
12
1
Re s p o n s e o f RT CRFD to RTCRF D
0. 08
0. 08
0. 04
0. 04
0. 04
0. 00
0. 00
0. 00
-0. 04
-0. 04
-0. 04
-0. 08
-0. 08
-0. 08
-0. 12
1
2
3
4
5
6
7
8
9
10
11
12
2
3
4
5
6
7
8
9
10
11
12
Re s p o n s e o f RTCRM 2 INQ to RTCRF D
1
0. 08
0. 08
0. 04
0. 04
0. 04
0. 00
0. 00
0. 00
-0. 04
-0. 04
-0. 04
-0. 08
-0. 08
-0. 08
-0. 12
1
2
3
4
5
6
7
8
9
10
11
12
5
6
7
8
9
10
11
12
2
3
4
5
6
7
8
9
10
11
12
Re s p o n s e o f RTCRM 2 INQ to RTCRM 2 INQ
0. 08
-0. 12
4
-0. 12
1
Re s p o n s e o f RT CRM 2 INQ to RT CRQIND
3
Re s p o n s e o f RTCRF D to RTCRM 2 INQ
0. 08
-0. 12
2
-0. 12
1
2
3
4
5
6
7
8
9
10
11
12
1
2
3
4
5
6
7
8
9
10
11
12
Figure 3 : Variance Decomposition for
rtcrqind, rtcrfd and rtcrm2inq
Varianc e Decompos ition
Pe rc e n t RT CRQIND v a ri a n c e d u e to RT CRQIND Pe rc e n t RTCRQIND v a ri a n c e d u e to RTCRF D Pe rc e n t RTCRQIND v a ri a n c e d u e to RTCRM 2 INQ
100
100
100
80
80
80
60
60
60
40
40
40
20
20
20
0
0
1
2
3
4
5
6
7
8
9
10
11
12
Pe rc e n t RTCRF D v a ri a n c e d u e to RT CRQIND
0
1
2
3
4
5
6
7
8
9
10
11
12
Pe rc e n t RTCRFD v a ri a n c e d u e to RTCRF D
1
100
100
80
80
80
60
60
60
40
40
40
20
20
20
0
1
2
3
4
5
6
7
8
9
10
11
12
3
4
5
6
7
8
9
10
11
12
Pe rc e n t RT CRFD v a ri a n c e d u e to RTCRM 2 INQ
100
0
2
0
1
2
3
4
5
6
7
8
9
10
11
12
1
2
3
4
5
6
7
8
9
10
11
12
Pe rc e n t RTCRM 2 INQ v a ri a n c e d u e to RT CRQINDPe rc e n t RT CRM 2 INQ v a ri a n c e d u e to RTCRFPe
D rc e n t RT CRM 2 INQ v a ri a n c e d u e to RTCRM 2 INQ
100
100
100
80
80
80
60
60
60
40
40
40
20
20
20
0
0
1
2
3
4
5
6
7
8
9
10
11
12
0
1
2
3
4
5
6
7
8
9
10
11
12
1
2
3
4
5
6
7
8
9
10
11
12
After proceeding a Johansen test for three I(1)
integrated variables (qind, fdinq and m2inq), the
cointegrated equation obtained support the
initializing of a VEC model.
Figure 4 : Impulse Responses Functions of
qind, fdinq and m2inq
Res pons e to One S.D. Innov ations
Re s p o n s e o f QIND to QIND
Re s p o n s e o f QIND to F DINQ
Re s p o n s e o f QIND to M 2 INQ
2500000
2500000
2500000
2000000
2000000
2000000
1500000
1500000
1500000
1000000
1000000
1000000
500000
500000
500000
0
0
0
- 500000
- 500000
10
20
30
40
50
60
70
80
90
100
- 500000
10
Re s p o n s e o f FDINQ to QIND
20
30
40
50
60
70
80
90
100
10
Re s p o n s e o f F DINQ to F DINQ
0. 3
0. 3
0. 2
0. 2
0. 2
0. 1
0. 1
0. 1
0. 0
0. 0
0. 0
- 0. 1
- 0. 1
- 0. 1
- 0. 2
10
20
30
40
50
60
70
80
90
100
20
30
40
50
60
70
80
90
100
10
Re s p o n s e o f M 2 INQ to F DINQ
0. 2
0. 2
0. 1
0. 1
0. 1
0. 0
0. 0
0. 0
- 0. 1
- 0. 1
- 0. 1
- 0. 2
- 0. 2
- 0. 2
- 0. 3
10
20
30
40
50
60
70
80
90
100
50
60
70
80
90
100
20
30
40
50
60
70
80
90
100
Re s p o n s e o f M 2 INQ to M 2 INQ
0. 2
- 0. 3
40
- 0. 2
10
Re s p o n s e o f M 2 INQ to QIND
30
Re s p o n s e o f FDINQ to M 2 INQ
0. 3
- 0. 2
20
- 0. 3
10
20
30
40
50
60
70
80
90
100
10
20
30
40
50
60
70
80
90
100
Figure 5 : Variance Decomposition for
qind, fdinq and m2inq
Variance Decompos ition
Pe rc e n t QIND v a ri a n c e d u e to QIND
Pe rc e n t QIND v a ri a n c e d u e to F DINQ
Pe rc e n t QIND v a ri a n c e d u e to M 2 INQ
100
100
100
80
80
80
60
60
60
40
40
40
20
20
20
0
0
10
20
30
40
50
60
70
80
90
100
0
10
Pe rc e n t FDINQ v a ri a n c e d u e to QIND
20
30
40
50
60
70
80
90
100
10
Pe rc e n t FDINQ v a ri a n c e d u e to F DINQ
100
100
80
80
80
60
60
60
40
40
40
20
20
20
0
10
20
30
40
50
60
70
80
90
100
20
30
40
50
60
70
80
90
100
Pe rc e n t M 2 INQ v a ri a n c e d u e to F DINQ
10
100
100
80
80
80
60
60
60
40
40
40
20
20
20
0
10
20
30
40
50
60
70
80
90
100
50
60
70
80
90
100
20
30
40
50
60
70
80
90
100
Pe rc e n t M 2 INQ v a ri a n c e d u e to M 2 INQ
100
0
40
0
10
Pe rc e n t M 2 INQ v a ri a n c e d u e to QIND
30
Pe rc e n t F DINQ v a ri a n c e d u e to M 2 INQ
100
0
20
0
10
20
30
40
50
60
70
80
90
100
10
20
30
40
50
60
70
80
90
100
5. Conclusions
 Long-run stability of the dynamic relationship between output
growth and financial deepening (observed like weighted nongovernmental credit and monetary aggregate M2 by the industrial
production value)
 The limited effectiveness of the banking system is revealed by
the extremely low level of banking sector credit in the economy, as
banks are unable or unwilling to lend (the last possibility is best
mirrored by the crowding-out effect that had a peak in 1999).
 Although the monetary aggregate M2 weight in industrial
production seems to Granger cause the other variables and to be
exogenous (indicating policy effectiveness), the responses of
economic growth (and even financial depth) to shocks in its level
are counterintuitive.
 The model may be augmented in case of finding better “proxy”
variables for economic growth and financial intermediation, or we
add other relevant variables omited (e.g. fiscal variables, bank
supervision, asset prices a.s.o.).