Central bank coordinated policies and bank market power: an insight from the African context

Figures & data

Table 1. Descriptive statistics

Variables	Mean	Std. Dev.	Min	Max
Market Power	0.474	1.712	−4.383	9.665
Monetary Policy	0.0854	0.0568	0.0233	1
Macro-Prudential Policy	0.012	0.143	−1	1
Deposit Funds	0.752	0.218	0.001	0.97
Credit Risk	0.0099	0.0074	0.0008	0.453
Concentration	0.5532	0.2769	0.28	1
Foreign Bank Entry	0.593	0.491	0	1
Money Supply	0.433	0.3879	0.261	1.13
Inflation	0.0774	0.063	−0.0241	0.4436
Central Bank Independence	0.5806	0.1649	0.2456	1
Regulatory Quality	−0.483	0.545	−1.66	0.853
Business Cycle	0.0738	0.0087	0.0553	0.0923
Instruments Used
Banking activity restrictions	5.292	5.571	4	9.
Financial freedom	46.54	13.108	0	70

Note: Market Power is measured with the Lerner Index; Monetary Policy is the central bank policy rate of a country; Macro-Prudential Policy Action Policy, an aggregate index of 17 indicators of macro-prudential action obtained from the iMaPP database constructed by Alam et al. (2019); Deposit Funds (ratio of total deposit to total asset); credit risk (ratio of nonperforming to gross loan); Credit Risk is the ratio of nonperforming to gross loan; Bank Concentration is the industry asset concentration of banks, measured as the ratio of asset of the three largest commercial natural logarithm of total bank assets; Foreign Bank Entry is a dummy variable (1 = year of foreign Bank entry, 0 otherwise); Money Supply (broad money (M2+) to GDP ratio); Inflation (consumer price index); Central Bank Independence is a measure of de jure central bank independence, which is based on a weighted aggregation of 16 legal indicators in four categories regarding the tenure of the bank’s governor, policy formation, objectives and limitations on lending to the government, using the criteria and weights in the Cukierman, Webb and Neyapti (CWN)); Regulatory Quality is an index that captures the ability of government and central bank to formulate policies or regulations; Business cycle is measured as real GDP growth rate. Activity restrictions measures the limits imposed on commercial banks to participate in the financial market. This measure varies from 4 to 16 with higher scores indicating more restrictions. The financial freedom variable provides an overall measure of the openness of the financial sector and the extent to which financial institutions are free to operate their business. It ranges from 0 percent (lowest freedom) to 100 percent (highest freedom).

Table 2. Pairwise Correlation Matrix

Variables	Mean VIF = 3.11	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)
(1) Macro-Prudential Policy	−3.37	1.000
(2) Monetary Policy	3.91	0.541*	1.000
(3) Central Bank Independence	3.35	−0.180*	0.112*	1.000
(4) Business Cycle	4.34	−0.425*	−0.551*	0.108*	1.000
(5) Deposit Funds	5.65	0.612*	0.320*	−0.653*	−0.377*	1.000
(6) Credit Risk	1.65	0.127*	0.201*	−0.268*	−0.297*	0.371*	1.000
(7) Concentration	3.14	−0.313*	−0.225*	−0.543*	−0.006	0.172*	0.208*	1.000
(8) Foreign Bank Entry	1.09	0.113*	0.131*	0.077*	−0.108*	0.043	−0.017*	−0.139*	1.000
(9) Money Supply	2.98	0.297*	−0.204*	−0.281*	−0.248*	0.403*	−0.037*	−0.250*	0.067	1.000
(10) Inflation	1.50	0.066	0.190*	−0.141*	−0.259*	0.155*	0.400*	−0.036*	−0.048	0.012	1.000
(11) Institutions	3.21	0.063	0.001	0.317*	0.539*	−0.354*	−0.197	−0.378*	0.136*	−0.178*	−0.309*	1.000

Note: Table 5.2 shows the pairwise correlation matrix. Monetary Policy is the central bank policy rate of a country; Macro-Prudential Policy Action Policy, an aggregate index of 17 indicators of macro-prudential action obtained from the iMaPP database constructed by Alam et al. (2019); Deposit Funds (ratio of total deposit to total asset); credit risk (ratio of nonperforming to gross loan); Credit Risk is the ratio of nonperforming to gross loan; Bank Concentration is the industry asset concentration of banks, measured as the ratio of asset of the three largest commercial natural logarithm of total bank assets; Foreign Bank Entry is a dummy variable (1 = year of foreign Bank entry, 0 otherwise); Money Supply (broad money (M2+) to GDP ratio); Inflation (consumer price index); Central Bank Independence is a measure of de jure central bank independence, which is based on a weighted aggregation of 16 legal indicators in four categories regarding the tenure of the bank’s governor, policy formation, objectives and limitations on lending to the government, using the criteria and weights in the Cukierman, Webb and Neyapti (CWN)); Regulatory Quality is an index that captures the ability of government and central bank to formulate policies or regulations. Note that * (p < 0.5) is the level of significance of correlation coefficient; Business cycle is measured as real GDP growth rate.

Table 3. Impact of Individual Central Bank Regulatory Policies on Market Power

Variables	Model 1	Model 2	Model 3
Monetary Policy	0.144***		0.0944***
	(0.0059)		(0.0047)
Macro-prudential Policy		0.0391***	0.0153**
		(0.0082)	(0.0075)
Deposit Funds	0.0031***	0.0012***	0.0002**
	(9.37e-05)	(7.44e-05)	(8.44e-05)
Credit Risk	−0.00176	−0.0045*	−0.0083***
	(0.0035)	(0.0025)	(0.0021)
Concentration	0.0006***	0.0005***	0.001***
	(0.0001)	(8.53e-05)	(9.79e-05)
Foreign Bank Entry	−0.361***	−0.0532	−0.0139
	(0.0782)	(0.0573)	(0.0499)
Money Supply	−0.0009***	−0.0024***	−0.0010***
	(0.0003)	(0.0002)	(0.0002)
Inflation Rate	0.0052	0.0325***	0.0075**
	(0.0049)	(0.0035)	(0.0035)
Central Bank Independence	0.281***	0.183***	0.182***
	(0.00475)	(0.0229)	(0.0233)
Regulatory Quality	−0.820***	−0.395***	−0.404***
	(0.0583)	(0.0510)	(0.0469)
Business Cycle	0.469***	0.0900***	0.309***
	(0.0406)	(0.0345)	(0.0326)
Trend	0.0668***	−0.0253***	−0.0202**
	(0.0120)	(0.009)	(0.0079)
Time Effect	Yes	Yes	Yes
Country Effect	Yes	Yes	Yes
Constant	−4.686***	−0.577*	−3.265***
	(0.366)	(0.310)	(0.298)
Instruments
Banking Activity Restrictions	Yes	Yes	Yes
Financial Freedom	Yes	Yes	Yes
Observations	676	676	676
R-squared	0.7411	0.5848	0.9107
Wald Chi^2	761.09***	514.56***	627.96***
Sargan’s Test	13.77	12.65	0.195
P-value	0.131	0.179	0.207
Hansen’s Test	12.34	9.679	8.55
P-value	0.862	0.260	0.201

Note: Table 3 shows the impact of individual central bank regulatory policies (monetary policy (model 1) and macro-prudential policy (model 2) on banks’ market power. The dependent variable is market power. Description of variables is found below Tables 1 and 2. They are not reported here because of space. The regressions use instruments for the central bank regulatory policy variables and the instruments are (1) banking activity restrictions, an index of regulatory restrictions on bank activity; and (2) financial freedom, which provides an overall measure of the openness of the banking sector. The Sargan and Hansen Tests are reported for overidentifying restrictions of instrument exogeneity—confirming that the instruments are valid. The parameters are estimated with the robust standard errors in parentheses: *** p < 0.01, ** p < 0.05, * p < 0.1

Table 4. Impact of Coordinated Central Bank Regulatory Policies on Market Power

Variables	Coordination impact	Countries in weak CBI framework	Countries in stringent CBI framework
VARIABLES	Model 4	Model 5	Model 6
Monetary Policy	0.115***	0.0198***	0.0593***
	(0.0143)	(0.0015)	(0.0093)
Macro-prudential Policy	0.0554***	−0.0045	−0.0588***
	(0.0159)	(0.0170)	(0.0089)
Monetary Policy*Macro-prudential Policy	0.0042**	−0.004**	0.0035***
	(0.0019)	(0.0017)	(0.00126)
Deposit Funds	0.0013***	0.0039***	0.001***
	(7.66e-05)	(6.97e-05)	(6.41e-05)
Credit Risk	−0.0044*	−0.0335***	−0.0052***
	(0.0024)	(0.0025)	(0.0013)
Concentration	0.0004***	0.0027***	0.0008***
	(8.65e-05)	(7.24e-05)	(6.70e-05)
Foreign Bank Entry	0.0627	−0.308***	−0.0977***
	(0.0560)	(0.0513)	(0.0309)
Money Supply	−0.0021***	−0.0018***	−0.0022***
	(0.0002)	(0.0002)	(0.0002)
Inflation Rate	−0.0002	0.0155***	0.0044*
	(0.0039)	(0.00325)	(0.00266)
Central Bank Independence	0.182***
	(0.0234)
Regulatory Quality	−0.468***	−0.416***	−0.176***
	(0.0541)	(0.0481)	(0.0229)
Business Cycle	0.171***	−0.122**	0.175***
	(0.0371)	(0.0515)	(0.0277)
trend	−0.0138	−0.0621***	0.0227***
	(0.009)	(0.0094)	(0.0049)
Time Effect	Yes	Yes	Yes
Country Effect	Yes	Yes	Yes
Constant	−1.937***	1.496***	−0.215***
	(0.323)	(0.171)	(0.0771)
Instruments
Banking Activity Restrictions	Yes	Yes	Yes
Financial Freedom	Yes	Yes	Yes
Observations	676	321	353
R-squared	0.6599	0.9637	0.7308
Wald Chi^2	886.56***	5339.06***	616.21***
Marginal Effect	0.1151***	0.0197***	0.0593***
Diagnostics
Sargan’s Test	51.36	10.07	2.502
P-value	0.491	0.347	0.286
Hansen’s Test	18.72	7.426	8.24
P-value	0.149	0.206	0.221

Note: Table 5 shows the coordination impact (interaction effect) of central bank regulatory policies on market power. The dependent variable is market power. CBI: de jure central bank independence. Weak CBI framework are countries strictly below the average de jure central bank independence index, and Stringent CBI framework are those equal or above the average de jure central bank independence index. The regressions use instruments for the central bank regulatory policy variables and the instruments are (1) banking activity restrictions, an index of regulatory restrictions on bank activity; and (2) financial freedom, which provides an overall measure of the openness of the banking sector. The Sargan and Hansen Tests are reported for overidentifying restrictions of instrument exogeneity—confirming that the instruments are valid. The parameters are estimated with the robust standard errors in parentheses: *** p < 0.01, ** p < 0.05, * p < 0.1

Figure 1. Marginal Plots: Coordinated Impact of Monetary and Macro-prudential Policies on Market Power.

Note: Figure 1 shows market power (Lerner index) on the y-axis. The x-axis shows monetary policy variable. The schedules show the level at which macro-prudential influence the positive effect of monetary policy on market power (Figure 1).

Figure 2. Marginal Plots of the Coordination of Monetary and Macro-prudential in Different CBI Framework.

Note: Market power is on the y-axis while the individual policies are on the x-axis. Steep line denotes marginal impacts of the coordination in the Stringent CBI framework while the flat line denotes impacts of the coordination in the Weak CBI framework.

Table

Application of the Use of the 3SLS: Equation 1(1) $MarketPowe{r}_{jt}=\sum _{l=1}^2{\alpha }_{l}CentralBankRegulatoryPolicies{ }_{jt}+\sum _{k=1}^N{\beta }_k{X}_{jt}+{\gamma }_j+{\mu }_t+{\epsilon }_{jt}$(1) In estimating the equation for the 3SLS, the study forms a number of equations which satisfy both the ORDER condition (the necessary condition) and the RANK condition (the necessary and sufficient condition of identification. According to Greene (2003: 392), the ORDER condition of identification for an equation is that the number of exogenous variables excluded from the equation be greater than or equal to the number of endogenous variables included in the same equation. According to Greene (2003: 392), the RANK condition imposes a restriction on a sub-matrix of the reduced-form coefficient matrix in order to ensure that there is exactly one solution for the structural parameters given the reduced-form parameters. To proceed, the study first arranges the structural parameters in a table and examines the sub-matrices one by one. The study forms a sub-matrix of the structural coefficients in the other equations on variables that are excluded from the equation and the study checks if all the elements of any column or row are all zero. Such a result indicates that the equation is not identified. On the contrary, if there are no columns or rows consisting of only zeros, it is concluded that the equation is identified. Therefore, the specified equations in the study (equations 5(5) $MarketPowe{r}_{jt}={\lambda }_{1}MonetaryPolic{y}_{jt}+{\lambda }_{2}MacroprudentialPolic{y}_{jt}+{\delta }_1{\left(MonetaryPolicy\ast MacroprudentialPolicy\right)}_{jt}+\sum _{k=1}^N{\beta }_k{C}_{jt}+{\sigma }_j+{\theta }_t+{\mu }_{jt}$(5) and equations 6) are identified. They all satisfy the ORDER condition (the necessary condition) of identification. Each of the equations employed is over-identified. They also satisfy the RANK condition (the necessary and sufficient condition) of identification. Since all the equations are over-identified, the 3SLS is suitable. The study reports on the 3SLS, because the 3SLS estimator is a systems estimator, it is presumptively superior in terms of efficiency to the 2SLS estimator, at least asymptotically (see Greene, 2003: 413; Iyoha, 2004: 118–9).

ereturn list scalars: e(rank) = 12 e(df_m) = 11 e(chi2) = 386.8009261261271 e(p) = 3.86917547959e-76 e(sigma_u) = 0 e(sigma_e) = .5880967529576227 e(sigma) = .5880967529576227 e(rho) = 0 e(rmse) = .6006624530249924 e(N) = 192 e(Tbar) = 28.6786389680177 e(Tcon) = 0 e(N_g) = 6 e(g_min) = 21 e(g_avg) = 32 e(g_max) = 47 e(that_min) = 0 e(that_5) = 0 e(that_50) = 0 e(that_95) = 0 e(r2_w) = .6911679329382696 e(r2_b) = .8986483998314698 e(r2_o) = .6824281829773407 e(thta_max) = 0

Table

Algeria	Comoros	Gambia, The	Mali	Senegal	Zambia
Angola	Congo, Dem. Rep.	Ghana	Mauritania	Seychelles	Zimbabwe
Benin	Congo, Rep.	Guinea	Mauritius	Sierra Leone
Botswana	Côte d’Ivoire	Guinea-Bissau	Morocco	Somalia
Burkina Faso	Djibouti	Kenya	Mozambique	South Africa
Burundi	Dominican Republic	Lesotho	Namibia	South Sudan
Cameroon	Egypt, Arab Rep.	Liberia	Niger	Sudan
Cape Verde	Equatorial Guinea	Libya	Nigeria	Tanzania
Central African Republic	Ethiopia	Madagascar	Rwanda	Togo
Chad	Gabon	Malawi	Sao Tome and Principe	Uganda

Alam, Z., Alter, A. E., Gelos, J., Kang, G., Narita, H., Nier, M., & Wang, N. (2019). Digging deeper – Evidence on the effects of macro-prudential. Monetary and capital markets department. IMF Working Paper

 Google Scholar