Central Bank Reform and Bank Lending:
Does the Credit Channel Work?
Laksmi K. Pande N, s1941429
Master Thesis General Economics, Rijksuniversiteit Groningen Supervisor: Prof. Jakob de Haan
August 2010
Abstract
I. Introduction
One important aspect of monetary policy institutions is central bank independence (CBI). Some studies have tried to give definitions on CBI. Bade and Parkin (1988) explained that a central bank will be independent if it is free from political interventions. Meanwhile, Grilli et al. (1992) argued that a central bank becomes independent when it has both political and economic independence in formulating monetary policy. Moreover, Cukierman et al. (1992) also considered financial independence as another important aspect of CBI. Financial independence implies that a central bank is not allowed to finance government’s budget deficits.
Many studies have explored the effect of CBI on inflation. It is widely believed that CBI has a negative relationship with inflation. De Grauwe (2009) found that an independent central bank can create price stability by reducing expected inflation. Moreover, Alpanda and Honig (2009) asserted that price stability can be attained since an independent central bank eliminates the political-monetary cycles in the election time. However, CBI is not free of costs. Some studies show that disinflation leads to output losses that are higher in countries with an independent central bank than in countries with a less independent central bank (Debelle & Fischer, 1994; Walsh, 1994; Jordan, 1997). There is still no general conclusion about the effect of CBI on economic growth.
Theoretically, the trade-off between inflation and output growth is explained by a Phillips curve relationship (Walsh, 2003). Nevertheless, the relationship between CBI and the output growth is still debatable. By observing industrialized countries, De Long and Summers (1992) found that CBI has a positive relationship with economic growth. After including developing countries, Cukierman et al. (1992) explained that CBI tends to increase economic growth. Moreover, Klomp and De Haan (2009) found that financial stability will increase due to central bank independence. On the other hand, Grilli et al. (1991); Alessina and Summers (1993); and Akhmad (1998) did not find any relationships between CBI and economic growth. These unclear impacts of CBI on economic growth might be caused by the ineffectiveness of monetary policy transmissions.
rates as the instruments. When a central bank conducts a loose monetary policy, the required reserves and market interest rates will decrease. Consequently, high excess reserves will allow banks to offer more loans. Meanwhile, on the demand side, firms and households will ask for more loans for investment and consumption which induce economic growth. However, the credit channel does not work if there are other alternative sources of funds and investments, especially when financial markets are well-developed (Bernanke & Blinder, 1988). Since CBI is associated with lower inflation, the real interest rate will be higher through Mundell-Tobin effect (Eijffinger and De Haan, 1996). Based on McKinnon (1973), a high real interest rate will attract people to save their money in banking sectors. Hence, this leads to more advance financial market development. Moreover, financial independence of central bank forces the government to issue more bonds in financing fiscal deficits. If there is a close substitute between zero risk government bonds and bank credit, it may shift the portfolio of the banking sector from private credit to government bonds. Thus, the credit channel will not be effective.
Based on the explanation above, this paper aims to examine the relationship between CBI and bank credit to the private sector, which is conditional on the magnitude of government budget balance and the extent of financial deepening. In general, there are two questions that will be addressed in this paper. First, whether CBI has a significant effect on bank credit. Second, whether the magnitude of budget balance and the extent of financial deepening significantly influence the effect of central bank reforms on the credit channel. By applying a panel data estimation method for 44 countries observation from 1980 to 2008, this paper finds that CBI has a negative effect on bank credit to the private sector. This fact is more likely to happen in developing countries than in developed countries. Moreover, central bank reform tends to reduce bank credit to the private sector, which is conditional on the budget balance and financial deepening. Since there is no previous study trying to explore this issue, it becomes a main contribution of this paper.
estimation; section VI explains estimation results, and section VII gives general conclusions.
II. Literature Review
Central bank independence is an important institutional aspect of monetary policy. Empirical evidences support that CBI will benefit for price stability, which is the main goal of monetary policy. However, the effect of an independent central bank on the other economic performances especially on output growth is still debatable. It depends on the effectiveness of monetary transmissions to the real sector. One of the monetary policy transmissions is the credit channel. This part will explain the definitions of CBI and the effects of CBI on macroeconomic variables, such as inflation and growth. Afterward, this part will also explain the credit channel of monetary policy. Finally, the relationship between CBI and the credit channel will be explained briefly in the end this section.
Central Bank Independence
Many studies have tried to give definitions on CBI. In general, there are three aspects of CBI: political independence, economic independence, and financial independence. Political independence means that a central bank is independent from political interventions in obtaining policy objectives (Bade & Parkin, 1988). Moreover, they suggested that to be independent, a central bank and a government should not have institutional relationships which can be used by a government to intervene central bank’s decisions.
However, political independence is not enough in ensuring the independence of a central bank. Therefore, an independent central bank should have both political and economic independence (Grilli et al., 1991). Economic independence means that the central bank has freedom in applying its monetary policy instrument without any restrictions from governments.
that a monetary policy is dominant over a fiscal policy. Although the CBI eliminates the monetary financing of fiscal deficits, the magnitude of government deficits will not be affected (Sikken & De Haan, 1998). Consequently, the government should manage its deficits and find other sources to finance the fiscal deficits, for instance by issuing government bonds.
Many countries have considered reforming their central bank laws to be independent from political interventions since 1990s. One of the main reasons is that an independent central bank could maintain price stability. An independent central bank can lower inflation as it can reduce the expected inflation of economic agents (De Grauwe, 2009). Moreover, De Grauwe (2009) also described that an independent central bank has more credibility in implementing monetary policy. Hence, the equilibrium of inflation will be lower. Meanwhile, when the central bank is not independent, the economic agents will set their expected inflation based on politicians’ preferences. The politicians with unemployment aversion will push the central bank to deviate from its inflation target. Hence, the equilibrium of inflation will be higher than the target of inflation. It is the so-called time inconsistency problem of monetary policy.
In addition, CBI can lower inflation by eliminating political monetary cycles during elections (Alpanda & Honig, 2009). When elections come, political parties will try to attract their voters by providing a lower unemployment and a higher output. Alpanda and Honig (2009) explained that if the central bank is not independent, the political parties will force the central bank to loose its monetary policy to induce the economy. Moreover, politicians will use monetary policy to finance government budget deficits due to the increasing of government spending or the tax cut campaign. Consequently, monetary financing of fiscal deficits generates higher inflation. Therefore, Alpanda and Honig (2009) suggested that the CBI is especially needed by developing countries, in which political-monetary cycles usually exits.
This is called the sacrifice ratio of disinflation. Some studies found that countries with an independent central bank tend to have higher sacrifice ratio than countries with a less independent central bank (Debelle & Fischer, 1994; Walsh, 1994; Jordan, 1997). In other words, we can interpret that an independent central bank could maintain price stability, but it creates output losses.
In industrialized countries, some studies argued that CBI has no relationship with economic growth. For instance, Grilli et al. (1991) did not find any significant relationships between CBI and economic growth. This result is supported by Alessina and Summers (1993) that compared the pattern of CBI and average output growth. They also found that among countries with high degree of CBI, some countries had low outputs, while others had high outputs. On the other hand, De Long and Summers (1992) argued that the CBI has a positive effect on long term output growth because the price system works effectively after central bank reform.
By including developing countries into observations, the effect of CBI on economic growth is still unclear. On the one hand, Cukierman et al. (1992) found that CBI has positive effects on economic growth. This result is supported by Klomp and De Haan (2009) which found that an independent central bank will decrease financial instability. Afterwards, financial stability leads to better economic conditions. On the other hand, by examining the divergence between legal and actual independence of central bank, Akhand (1998) explained that there is no robust correlation between CBI and output growth.
The Credit Channel of Monetary Policy
the people who have excessive funds to the people who do not have sufficient funds in the economy.
There are two important parts of credit channel through which monetary policy affects the real economy: bank lending channel and firm’s balance sheet (Mishkin, 1996). The bank lending channel emphasizes an effect of monetary policy on the supply of credit through bank’s balance sheet. Monetary policy can affect the supply of credit by setting reserve requirements of the banking sector. The higher the reserve requirements are, the lower the funds will be available for credit. Therefore, the supply of credit to the private sector will decrease.
A firm’s balance sheet channel is viewed from a demand side of credit (Mishkin 1996). Since the demand for credit depends on the net worth of the borrowers, monetary policy can affect the demand for credit through affecting the net worth of the borrowers. The net worth of the borrower is affected by the asset price and interest rate. The higher the asset price, the higher the value of collateral. Moreover, the lower the interest rate, the lower the cost of funds. Hence, the demand for credit will increase.
In conclusion, credit channel of monetary policy is important in enhancing economic growth. When central bank conducts a loose monetary policy, the required reserves and market interest rate will decrease. High excess reserves will allow banks to offer more loans. Meanwhile, firms and households will ask for more loans for investments and consumptions which induce the economic growth.
Central Bank Independence and Credit Channel
The effect of CBI on the credit channel depends on the extent of financial deepening and the amount of fiscal deficits. An independent central bank induces financial deepening through a high interest rate policy. Eijffinger dan De Haan (1996) explained that CBI is associated with lower inflation rates. Hence, through the Mundell-Tobing effect, this may lead higher real interest rates. Based on McKinnon (1973), a high real interest rate will attract people to save their money in banking sectors. By observing three Asian countries (South Korea, Thailand, and Indonesia), Agrawal (2001) found that a higher interest rate induces a more financial deepening because foreign investors invest their money in a domestic market. It implies that financial deepening provides other sources of funds besides bank credit. Consequently, the credit channel of monetary policy to real sectors will not be effective.
In addition, the effect of CBI on the credit channel is also determined by the amount of government fiscal deficits. Sikken & De Haan (1998) found that fiscal deficits are not affected by central bank reforms. Since an independent central bank is not allowed to finance fiscal deficits, government should try to find the other sources of financing. One of the alternatives for financing government deficits is by issuing government bonds in the domestic market. This will affect the credit transmission of monetary policy through banking sectors if the government bonds and bank credit are closely substituted.
III. General Pictures of Central Bank Independence and Credit Channel.
There are two proxies used to express the credit channel. The first proxy is the ratio between banking institutions’ claims on the government and the private sector (Govt/pvt). The second proxy is the growth of banking institutions’ claims on the private sector (pvtcrgr). The data is provided by International Financial Statistics (IFS) for 44 countries from 1980 to 2008.
According to the cross countries data, the ratio between banking institutions’ claims on the government and the private sector (Govt/pvt) reached the highest value in Mexico in 1987. Meanwhile, Australia counted the lowest value in 2005. The highest and the lowest values of the ratio between government credit and private credit were 249.74 percent and -0.44 percent, respectively. In the beginning of 1990s, the government of Mexico limited private credit by nationalizing banking systems (De Cordoba & Kehoe, 2009). Moreover, in that period, credit with a low interest rate was only provided for some large firms. Low credit for government in Australia is caused by a surplus on its government budget. Therefore, issuing new government bonds was not desired (Battellino & Chambers, 2006). On average, the ratio of banking institutions’ claims on the government and the private sector was 21.32 percent and the standard deviation was 25.48 percent (see Table 1).
The highest and the lowest growth of private credit were reached by Argentina and Indonesia, respectively. Argentina reached the highest growth of private credit in 1990 with the value 736.50 percent. This achievement was caused by the government’s stabilization plans (Saxton, 2003). On the other hand, Indonesia experienced the lowest growth of private credit in 1999 with the value 55.71 percent. It was caused by a deep banking crisis in 1998 (Wie, 2003). In general, the average and the standard deviation of private credit growth in 44 countries from 1980 to 2008 were 21.51 percent and 40.94 percent, respectively (see Table 1).
their central bank laws in the 1990s. Chile was the first country that reformed its central bank law in year 1989, while Mauritius reformed its central bank law in 2004 (see Appendix Table A.2.). Chile mandated its central bank as an independent institution because they experienced high inflation (Carstens & Jacome, 2005).
The other variables used in this paper are budget balance, financial deepening, inflation and income per capita. The government budget balance is measured by the ratio between budget balance and GDP (BalL). The positive value of government budget indicates a surplus in government budget, while the negative value indicates a deficit in government budget. The highest surplus was 22.66 percent of GDP, while the highest deficit of government budget was 30.89 percent of GDP. The highest surplus of government budget was experienced by Botswana in 1985, while the highest government deficit was experienced by Elsavador in 1992. Moreover, the average government budget balance per GDP was -2.06 percent and the standard deviation was 4.48 percent (see Table 1).
Financial deepening is expressed by broad money (M2) per GDP (DeepL). The country that has the highest value of financial deepening was Switzerland in 1999. On the other hand, the country that has the lowest value of financial deepening is Uganda in 1989. On average, the level of financial deepening for 44 countries from 1980 to 2008 was 48.20 percent of GDP and the standard deviation was 29.31 percent of GDP (see Table 1).
The percentage change of consumer price index is used to represent inflation in each country (InfL). Inflation reached 3079.81 percent as the highest rate of inflation in our observations. This condition happened in Argentina in 1989. On the other hand, the highest deflation rate is 2.64 percent experienced by Bahrain in 1985. Therefore, the average and the standard deviation of inflation for data observations were 19.40 percent and 112.91 percent, respectively (see Table 1).
the data of 11 developed countries and 33 developing countries, the average income per capita was US$ 6,796 and the standard deviation was US$ 10,235 (see Table 1).
Table 1: Statistical Descriptive
Pvtcrgr Govt/pvt BalL DeepL InfL GDPCapL
Max 7.3650 2.4974 0.2266 1.5833 30.7981 65175.9488
Min -0.5571 -0.0044 -0.3089 0.0554 -0.0264 85.5415
Mean 0.2151 0.2132 -0.0206 0.4820 0.1940 6796.2530
Std.Dev 0.4094 0.2548 0.0448 0.2931 1.1291 10,235.6752
Note: Pvtcrgr and Govt/pvt stand for private credit growth and ratio between government and private credit. BalL is the lag of government budget balance per GDP. DeepL is the lag of broad money per GDP. InfL is the lag of inflation and GDPCapL is the lag of income per capita.
In order to compare the variables used for estimation before and after central bank reform, the mean analysis is employed. Before central bank reform, the average ratio between government credit to private credit (Govt/pvt) was 19.81 percent, while after the central bank reforms this ratio became 24.98 percent. It means that after central bank reform, the ratio of government credit to private credit increased 5.17 percentage points on average (see Table 2).
After central bank reform, the growth of private credit from banking sectors decreased. On average, private credit growth before central bank reforms and after central bank reform was 23.05 percent and 17.78 percent, respectively. In other words, private credit growth decreased by 5.27 percentage points after central bank became independent (see Table 2).
The average government budget deficit declined after central bank reform. The average government budget balance per GDP increased from -2.31 percent to -1.42 percent (see Table 2). The increase of government budget balance might give an illustration about the absence of monetary financing of fiscal deficits after the independence of central bank (Cukierman et al., 1992; Grilli et al., 1991).
inflation as the main target of monetary policies. It is showed by the significant decrease of inflation from 23.36 percent to 9.95 percent after central bank reform (see Table 2). Meanwhile, the income per capita after central bank reform decreased from US$6958 to US$6415 (see Table 2). It reflects that many less developed countries reformed their central bank laws later in our sample periods. However, the t-test does not show that the difference of income per capita is significant.
Table 2: The mean analyses
Variable Before After t-test p-value
Govt/pvt 0.1981 0.2498 0.0010 Pvtcrgr 0.2305 0.1778 0.0370 BalL -0.0231 -0.0142 0.0013 DeepL 0.4839 0.4772 0.7104 InfL 0.2336 0.0995 0.0536 GDPCapL 6958.092 6415.831 0.3892
Note: Pvtcrgr and Govt/pvt stand for private credit growth and ratio between government and private credit. BalL is the lag of government budget balance per GDP. DeepL is the lag of broad money per GDP. InfL is the lag of inflation and GDPCapL is the lag of income per capita. The p-value of the t-test shows whether there are no significant differences between the period before and after a reform.
Data descriptions might not give clear explanations about the relation between CBI and the credit channel. Many variables are not taken into consideration by the mean analysis. Hence, an empirical estimation can provide better evidence about the relationships since other factors in the economy could also be controlled for robust estimation (Wooldridge, 2005). Therefore, this paper will try to deepen our understanding of the relationship between CBI reform and the credit channel by using the panel data estimation method.
IV. Methodology
method (Wooldridge, 2005). The Hausman test will be applied to determine whether the fixed effects or the random effects model is preferred.
By using the fixed effects panel method, the model assumes that the coefficients of CBI for all countries (γ) are the same across countries (see equation 1). However, the heterogeneity of each country can be captured by individual specific effect, which is represented by μi in equation 1. The individual specific effect captures the components
that are unobserved by the econometrician (Wooldridge, 2005).
Similar to Klomp and De Haan (2010), this paper applies the time dummy of central bank reform developed by Acemoglu et al. (2008) and expanded by Daunfeldt et al. (2010). Based on the literature review in previous section, the other factors that possibly influence bank credit to the private sector are budget balance, financial deepening, and macroeconomic variables such as income per capita and inflation. In order to avoid endogeneity problems, this paper uses a lag of one period for all independent variables. The basic estimation model used in this paper is as follows:
In equation (1), the dependent variable (CRi,t)reflects a bank credit indicator of
country i at time t. The credit indicator is proxied by two variables: the ratio between banks’ claims on the government and the private sector, and the growth of bank credit to the private sector. The variable of interest is CBIi,t-1, which represent the time dummy of
central bank reform. The periods before central bank reform are signed zero, while the periods after central bank reform are given one. If the growth of private credit is used as a dependent variable, the parameter of CBI (γ) is expected to be negative. It means that central bank reform reduces the incentive of banking sectors to provide credit for private sectors. This hypothesis is related to a close substitute between bank credit and government bonds.
The other variables which possibly determine the bank credit to the private sector are budget balance and financial deepening, which are respectively represented by Bali,t-1
and Deepi,t-1. The parameter of government budget balance (Ø) and the parameter of
the government tends to issue more bonds. The government bonds can be alternative investment for banking sectors. A more developed financial market leads to higher substitutions between government bonds and bank credit to the private sector. As a result, the growth of bank credit to the private sector will decrease.
Macroeconomic indicators, i.e. income per capita and inflation, are included in the estimation. Those variables are represented by Zi,t-1 in equation 1. Income per capita
is expected to have a positive effect on bank credit (Walsh, 2003). Inflation could have a negative or positive effect on the credit channel. Since inflation is similar to taxes on money, it will decrease the value of collateral and the supply of credit (Walsh, 2003). Meanwhile, high inflation can increase the value of collateral because of illusionary households (Piazzesi & Scheneider, 2007). In equation (1), μi reflect the country specific
effects and εi,t denotesanerror term.
The effect of CBI on bank credit could be different across countries. The differences are caused by different characteristics of countries in the sample. The evidences of heterogeneity can be captured by separating observations into developed and developing countries. Moreover, separating observations are also used to test whether the estimation results are robust. If the effect of CBI on private credit is statistically different between developed and developing countries, it indicates that the heterogeneity problems exist in the observations.
In order to analyze the conditional effect of one variable on the dependent variable, variable interactions can be applied (Wooldridge, 2005). In our model, the effect of CBI on private credit depends on budget balance and financial deepening. Hence, for estimation, we use two dummy interactions, which are the multiplication of central bank reform with budget balance and financial deepening. The model of estimation which uses the dummy interaction between CBI and budget balance can be seen in the following model:
interaction’s parameter (δ) and budget balance (Bali,t-1). The total marginal effects of CBI
on private credit are computed by the following equation:
The effect of CBI on private credit also depends on the extent of financial deepening. By interacting CBI and financial deepening, the equation can be presented as follows:
Based on equation (3), the effect of CBI on private credit depends on its parameter (γ) and the dummy interaction’s parameter (φ) and financial deepening (Deepi,t-1). The total marginal effect of CBI on private credit becomes:
From equation (3.a), the CBI is expected to have a negative effect on bank credit after considering the extent of financial deepening.
V. Data Description
The data for estimations are available across countries and years. Since the availability of data varies among countries, this paper applies an unbalanced panel data method. Moreover, this paper observes 44 countries, which consists of 11 developed countries and 33 developing countries from 1980 to 2008 (see Appendix Table A2).
whereas bank credit to government is approximated by banks’ claims on the government. The data are provided by International Financial Statistics (IFS).
In this paper, the time dummy variable is used to represent the period of central bank reform. The data is obtained from Acemoglu et al. (2008) covering 55 countries during period 1980 to 2005. Afterwards, Daunfeldt et al. (2009) extended Acemoglu’s data set by adding the number of countries into 133 countries. However, since the availability of other variables used in the estimation varies, this paper can only use time dummy reform of 44 countries.
This study also considers other factors which affect the provisions of bank credit to the government and the private sector, such as government budget balance, financial deepening, inflation, and income per capita. The data is provided by IFS from 1980 to 2008. The government budget balance is measured by the share of budget balance to GDP. Financial deepening is proxied by the ratio between broad money (M2) and GDP. Inflation is represented by the percentage change of consumer price index for each country. The detail descriptions about data availability can be seen in table A1 appendix.
Table 3: Correlation matrix of exogenous variables
BalL InfL DeepL CBIL GDPCapL
BalL 1.0000
InfL -0.0122 1.0000
DeepL 0.0991* -0.1182* 1.0000
CBIL 0.0900* -0.0541 -0.0104 1.0000 GDPCapL 0.2385* -0.0592* 0.4010* -0.0241 1.0000
Note: Number of observation is 1272. * is significant at 5 percent level. BalLis the lag of government budget balance per GDP. DeepL is the lag of broad money per GDP. InfL is the lag of inflation and GDPCapL is the lag of income per capita.
VI. Results
In order to analyze the effect of central bank reform on bank credit to the private sector, this paper follows some estimation procedures. To estimate the equation (1), (2), and (3), there are two dependent variables used: the ratio between banks’ claims on the government and the private sector, and the growth of banks’ claims on the private sector. Both dependent variables are used to capture the behavior of banking sectors prior to and after central bank reform. The estimation results which use the ratio between banks’ claims on the government and the private sector as the dependent variable can be seen in the table 4. Meanwhile, the estimation results which use the growth of banks’ claims on the private sector as the dependent variable can be seen in table 5.
Ratio between government and private credit as the dependent variable
The column (1) of table 4 is the estimation result of equation 1 (without dummy interactions). Based on the Hausman test, the fixed effects method1 is preferred. The estimation result shows that the lag of central bank reform dummy variable (CBIL) positively significant affects the ratio between banks’ claims on the government and the private sector. It implies that after central bank reform, banking sectors tend to change their asset portfolio by giving more credit to the government than to the private sector. In addition, lag of budget balance (BalL) also significantly affect the ratio between banks’ claims on the government and the private sector. The sign of budget balance coefficients
1 Fixed effect does not rule correlation between the individual specific effect (
i
is negative. It indicates that the higher budget deficits, the larger credit of banking sectors to the government. However, the financial deepening (DeepL) is not statistically significant. Moreover, two control variables such as income per capita and inflation are also included. Income per capita (GDPCapL) is statistically significant with a negative sign. It means that banking sectors in countries with higher income per capita tend to reduce the credit for government. Meanwhile, inflation (InfL) does not statistically significant determine the behavior of banking sectors.
Developed countries and developing countries are estimated separately to capture the heterogeneity problems. For the developed countries, the results show that the ratio between banks’ claims on the government and the private sector after central bank reform is lower than that before central bank reform (column 2 of Table 4). It implies that central bank reform in developed countries increase the incentive of banking sectors to provide credit for the private sector. In this case, the credit channel still works in developed countries after central bank reform. Meanwhile, the sign of central bank reform coefficient (CBIL) in developing countries (column 3 of Table 4) is positive. This result indicates that after central bank reform, banking sectors tend to give more credit for government than for private sector. It confirms the hypothesis that the credit channel of banking sectors is damped after central bank reform. Moreover, the financial deepening variable has a positive effect on the ratio between banks’ claims on government and private sectors in both developed and developing countries.
This paper considers that the effect of CBI on bank credit to the private sector depends on the other factors, such as the magnitude of budget balance and the extent of financial deepening. Therefore, in column (4) and (5) of table 4, two dummy interactions are considered. The first dummy variable is the interaction between central bank reform and the magnitude of government budget balance (CBIL*BalL). The second dummy variable is the interaction between central bank reform and the extent of financial deepening (CBIL*DeepL).
(column 4 of Table 4). The total marginal effect of central bank reform on the ratio between banks’ claims on the government and the private sector is represented by the following calculation:
There are two ways can be used to analyze the marginal effect of CBI on the ratio of bank credit: the average value analyses and the marginal effect graph. Based on the equation (4.a), since the average budget balance per GDP equals to -2.06 percent for 44 countries from 1980 to 2008 (see Table 1), the total marginal effect of central bank reform is 5.65 percent. This result reconfirms that central bank reform tends to increase the credit of banking sectors to the government relatively to the private sector.
Based on the marginal effect graph in figure (1), central bank reform has significant negative and positive effects on the ratio between government credit and private credit. The negative effect of central bank reform on the ratio credit is experienced by highly government budget deficit countries. Moreover, when the government deficits decrease (becomes surplus), the independence of central bank can increase the ratio between government credit and private credit. The threshold of government budget balance exists when the marginal effect equals to zero (see equation 4.a). Therefore, based on the observations, the CBI has a positive effect on the credit ratio when the government budget deficit is higher than 8.4 percent. Since the average value of budget deficit in the sample is higher than 8.4 percent, it implies that CBI will increase the ratio between banks’ claims on the government and the private sector.
In addition, the effects of other variables are also considered. The government budget balance and income per capita have significantly negative effects on the ratio between bank’s claims on the government and the private sector. On the other hands, financial deepening and inflation have no significant effect on the ratio between government and private credit (see column 4 of Table 4).
banks’ claims on the government and the private sector depends on the extent of financial deepening (see column 5 of Table 4). By using a fixed-effect estimation method, the marginal effect of central bank reform on the ratio between banks’ claims on the government and the private sector is as follows:
Similar to the previous explanations, the effect of CBI on the ratio between government credit and private credit can be analyzed through the average analysis and the marginal effect graph. Since the average value of financial deepening is 48.20 percent (see Table 1), the marginal effect of central bank reform on the ratio between banks’ claims on the government and the private sector is 5.13 percent. This result means that after central bank reform, the ratio between banks’ claim on the government and the private sector increases.
The graph of marginal effect shows that CBI has a significantly positive and negative effect on the ratio between government credit and private credit (see Figure 2). At a five percent significant level, the upper and lower bounds of marginal effect estimation lines lie in positive and negative areas. It means that by considering financial deepening of each country, the central bank reform will increase the ratio between government credit and private credit. At a certain level of financial deepening, the central bank reform will decrease the ratio credit. The threshold value of financial deepening is obtained when the CBI’s marginal effect equals to zero (see equation 4.b). With the threshold value of financial deepening equals to 64.3 percent, it can be interpreted that when the level of financial deepening is higher than 64.3 percent, central bank reform will decrease the ratio between government credit and private credit. Meanwhile, if the level of financial deepening is lower than the threshold, central bank reform will increase the ratio between government credit and private credit.
threshold, this paper concludes that central bank reform increases the ratio between banks’ claims on the government and on the private sector.
The other variables also affect the ratio between government and private credit. At a five percent significant level, government budget balance and income per capita have a significantly negative effect on the ratio between banks’ claims on the government and the private sector. Whereas, the financial deepening has a significantly positive effect on the ratio between banks’ claims on the government and the private sector (see column 5 of Table 4). However, inflation has no significant effect on bank credit.
Table 4: Results for the ratio between government and private credit
Variables 1 2 3 4 5 Constanta 0.199** (9.00) 0.107** (4.48) 0.197** (5.13) 0.190** (6.25) 0.129** (5.43) CBIL 0.065** (4.60) -0.090** (-6.50) 0.112** (6.45) 0.082** (5.52) 0.206** (8.46) BalL -0.418** (-2.67) -0.389** (-3.19) -0.412** (-2.16) -0.705** (-3.99) -0.383** (-2.49) DeepL 0.074 (1.55) 0.215** (6.22) 0.155** (2.50) 0.077 (1.63) 0.214** (4.22) Control Variables GDPCapL -7.07e-06** (-5.47) -4.77e-06** (-8.54) -3.00e-05** (-6.05) -7.18e-06** (-5.58) -6.26e-06** (-4.92) InflationL -0.006 (-1.22) -0.143** (-2.52) -0.007 (-1.31) -0.005 (-1.19) -0.003 (-0.55) Interaction CBIL*BalL 0.976** (3.43) CBIL*DeepL -0.321** (-7.04) R2 5.66 36.87 2.24 6.86 8.53 #Observation 1269 287 982 1269 1269 Hausman Test 19.58 (FE) 3.62 (RE) 50.17 (FE) 23.19 (FE) 12.37 (FE)
Note: t-values in parenthesis, ** significant at a 5 percent level, * significant at a 10 percent level
Column (1) describes the result of estimation includes all country observations without using variable interactions.
Column (2) describes the result of estimation which only includes the developed countries (11 countries) without using variable interactions. Column (3) shows the result of estimation which only includes the developing countries (33 countries) without using variable interactions. Column (4) shows the result of estimation for all country observations with considering the interaction between central bank reforms and government budget balance.
Figure 1.
The Marginal Effect of CBIL on Govt/Pvt for Different Values of BalL
Figure 2.
The Marginal Effect of CBIL on Govt/Pvt for Different Values of DeepL
Note:
• Vertical axis describes a marginal effect of Central Bank Independence on the ratio between government credit and private credit.
• Horizontal axis shows government budget balance.
• The middle line describes the marginal lines which significant at five percent
• The upper and lower lines show the bounds of marginal effect.
Note:
• Vertical axis describes a marginal effect of Central Bank Independence on the ratio between government credit and private credit.
• Horizontal axis shows financial deepening.
• The middle line describes the marginal lines which significant at five percent
• The upper and lower lines show the bounds of marginal effect.
Growth of private credit as the dependent variable
The second estimation uses the growth of banks’ claims on the private sector as the dependent variable (see table 5). The estimation method applied in table 5 is the same as the estimation method used in table 4. Based on the estimation results (in column 1 of Table 5), the central bank reform have a significant negative effect on the growth of private credit. It means that after central bank reform, the credit of banks to the private sector decrease.
With respect to dummy interaction, the interaction between central bank reform and budget balance is not statistically significant (see column 4 of Table 5). Moreover, other variables i.e. the magnitude of budget balance, the extent of financial deepening, and income per capita also do not have any impacts on the growth of private credit. Meanwhile, inflation has a significantly positive effect on private credit.
The marginal effect of central bank reform on the growth of private credit can be analyzed by using average value analyzes and marginal effect graph. The total marginal effect of central bank reform on the growth of bank credit is measured by the following equation:
Since the dummy interaction between central bank reform and budget balance is not significant, the effect of central bank reform is not influenced by the magnitude of government budget balance (see equation 5.a). Hence, the total marginal effect of central bank reform on private credit growth is -9.3 percent.
From the marginal effect graph in figure 3, the central bank reform at various budget balance does not have a significant effect on the growth of banks’ claims on the private sector. The significance of central bank reforms is obtained by analyzing the upper and lower bounds of marginal effect. The upper bound lies on a positive area, while the lower bound lies on a negative area. This implies that the level of budget balance does not significantly influence the effect of central bank reform on private credit growth.
The second dummy interaction (between central bank reform and financial deepening) is statistically significant at a five percent significant level (see column 5 of Table 5). It means that the degree of financial deepening determines the effect of central bank reform on private credit growth. In addition, the financial deepening itself and the income per capita significantly affect private credit growth. Meanwhile, government budget balance and inflation have no effect on private credit growth.
Based on equation (5.b), the effect of central bank reform on private credit growth is influenced by the level of financial deepening. Since the average value of financial deepening in the observations is 48.20 percent, the total marginal effect of central bank reform on private credit growth is -8.62 percent. It means that when a central bank becomes independent, private credit growth will decrease.
The effect of central bank reform on the growth of private credit at the various level of financial deepening is proven by analyzing the marginal effect graph. The central bank reform has positive and negative effects on private credit growth since the upper and lower bounds of the marginal effect lie in the same area (see Figure 4). By taking the
Table 5: Results for the growth of private credit
Variables 1 2 3 4 5 Constanta 0.258** (6.68) 0.128** (2.55) 0.264** (6.23) 0.256** (6.60) 0.309** (7.31) CBIL -0.097** (-3.93) 0.035 (0.79) -0.090** (-3.07) -0.093** (-3.54) -0.201** (-4.65) BalL 0.340 (1.25) 0.916** (2.59) 0.120 (0.38) 0.278 (0.90) 0.314 (1.15) DeepL -0.078 (-0.95) -0.083 (1.35) 0.046 (0.44) -0.078 (-0.94) -0.182** (-2.03) Control Variables GDPCapL -3.65e-07
(-0.16) 1.23e-06 (0.87) -2.00e-05** (-2.47) -3.90e-07 (-0.17) -9.63e-07 (-0.43)
InflationL 0.165** (18.74) 0.783** (2.50) 0.162** (17.30) 0.165** (18.74) 0.162** (18.43) Interaction CBIL*BalL 0.212 (0.42) CBIL*DeepL 0.238** (2.94) R2 30.73 10.26 29.96 30.77 30.96 #Observation 1267 287 980 1267 1267 Hausman Test 105.16
(FE) 4.35 (RE) 123.75 (FE) 101.25 (FE) 118.05 (FE)
Note: t-values in parenthesis, ** significant at a 5 percent level, * significant at a 10 percent level
Column (1) describes the result of estimation includes all country observations without using variable interactions.
Column (2) describes the result of estimation which only includes the developed countries (11 countries) without using variable interactions. Column (3) shows the result of estimation which only includes the developing countries (33 countries) without using variable interactions. Column (4) shows the result of estimation for all country observations with considering the interaction between central bank reform and government budget balance.
marginal effect of equation 5.b equals to zero, the threshold of financial deepening is 84.5 percent. Therefore, central bank reform will decrease private credit growth when the financial deepening is lower than the threshold. However, when the financial deepening is higher than the threshold, central bank reforms will increase private credit growth. Based on the data from 1980 to 2008 in 44 countries, there are six countries that experienced a high level of financial deepening, i.e. Egypt, Jordan, Korea, Malaysia, Singapore, and Switzerland. Since most of countries in the observations have the level of financial deepening lower than the threshold, this study concludes that the central bank reform decreases private credit growth.
Figure 3.
The Marginal Effect of CBIL on PVTCrGr for Different Values of BalL
Figure 4.
The Marginal Effect of CBIL on PVTCrGr for Different Values of DeepL
Note:
• Vertical axis describes the marginal effect of Central Bank Independence on private credit growth.
• Horizontal axis shows government budget balance.
• The middle line describes the marginal lines which is not significant at five percent of significant level.
• The upper and lower lines show the deviation of marginal effect.
Note:
• Vertical axis describes the marginal effect of Central Bank Independence on private credit growth.
• Horizontal axis shows financial deepening.
• The middle line describes the marginal lines which significant at five percent of significant level.
• The upper and lower lines show the deviation of marginal effect.
In conclusion, central bank reform gives negative effects on the credit channel. The ratio between banks’ claims on the government and the private sector increases after central bank reform. It has been proven by the estimation results in table 4 and by the marginal effect in figure (1) and figure (2). Moreover, the growth of bank credit to the private sector also decreases after central bank reform (Table 5, Figure 3, and Figure 4). When the estimations are separated for developed and developing countries, the negative
effect of central bank reform on bank credit to the private sector is more likely to happen in developing countries than in developed countries. In addition, the countries with a high level of financial deepening get more benefits from central bank reform since private credit growth increases significantly.
VII. Summary and Conclusions
Monetary policies could affect the real economy through its transmission mechanisms. One of the monetary transmissions is the credit channel. Theoretically, based on Bernanke and Blinder (1988), the credit channel does not work if there are other alternative sources of funds and investment, especially when the financial market is well-developed. Related to central bank reform, an independent central bank is not allowed to give credit to government. Moreover, based on empirical evidences, CBI does not reduce the magnitude of government budget deficits (Siken & De Haan, 1998). Consequently, the government should find other alternatives to finance fiscal deficits. One of sources to finance fiscal deficits is by issuing government bonds. If the government bonds and bank credit are a close substitute, the credit channel will not work.
By considering conditional factors such as government budget balance and financial deepening, this paper finds that central bank reform has a negative impact on bank credit to the private sector. After central bank becomes independent, bank credit growth to the private sector significantly decreases. Moreover, the ratio between bank credit to government and to private sectors significantly increases after central bank reform. The negative effect of central bank reform on bank credit to the private sector is likely to happen in developing countries than in developed countries. In addition, central bank reform will lead the credit channel works if particular countries have a high level of financial deepening.
illusionary households which explain that the value of collateral increases because of higher inflation (Piazzesi & Scheneider, 2007). Hence, the higher the collateral value, the higher the bank credit to private sectors (Walsh, 2003).
References
Acemoglu, Daron, Simon Johnson, Pablo Querubin, and James A. Robinson. 2008.
“When does Policy Reform Work? The Case of Central Bank Independence”. National Bureau of Economic Research Working Paper 14033.
Agenor, Pierre-Richard, Joshua Aizenman, and Alex Hoffmaister. 2000. “The Credit
Crunch in East Asia: What can Bank Excess Liquidity Assets Tell US?” National Bureau of Economic Research Working Paper 7951.
Agrawal, Pradeep. 2001. “Interest Rates, Exchange Rates and Financial Deepening in
Selected Asian Economies.” ASEAN Economic Bulletin 18(1): 83-93.
Akhand, Hafiz A. 1998. “Central Bank Independence and Growth: A Sensitivity
Analysis.” The Canadian Journal of Economics 31(2): 303-317.
Alessina, Alberto, and Lawrence H. Summers. 1993. “Central Bank Independence and
Macroeconomic Performance: Some Comparative Evidence.” Journal of Money,
Credit, and Banking 25(2): 151-162.
Alpanda, Sami, and Adam Honig. 2009. “The Impact of Central Bank Independence on
Political Monetary Cycles in Advanced and Developing Nations.” Journal of
Money, Credit, and Banking 41(7): 1365-1388.
Bade, Robert, and Michael Parkin. 1988. “Central Bank Laws and Monetary Policy.”
http://economics.uwo.ca/faculty/parkin/CentralBankLaws.pdf.
Battellino, Ric, and Mark Chambers. 2006. “An Overview of Australian Bond
Market”. Bank for International Settlements Papers 26: 45-55.
Bernanke, Ben S, and Alan S. Blinder. 1988. “Credit, Money, and Aggregate
Demand.” The American Economic Review 78(2): 435-439.
Carstens, Agustin, and Luis I. Jacome. H. 2005. “Latin America Central Bank
Reforms: Progress and Challenges.” International Monetary Fund Staff Papers 5(114): 1-43.
Cukierman, Alex, Steven Webb, and Bilin Neyapti. 1992. “Measuring the
Independence of Central Banks and Its Effect on Policy Outcomes.” World Bank
Daunfeldt, Sven-Olov, David Granlund, Mats Landstrom, and Niklas Rudholm.
2010. “Central Bank Independence Reforms and Price Stability: The Role of Democracy and Political Stability.”
http://www.pubchoicesoc.org/papers_2010/Daunfeldt_Gran_Land_Rud.pdf
De Cordoba, Gonzalo F, and Timothy J. Kehoe. 2009. “The Current Financial Crisis:
What Should We Learn from the Great Depressions of the Twentieth Century?” Federal Reserve Bank of Minneapolis. Research Departement Staff Report 421.
De Grauwe, Paul. 2009. Economics of Monetary Union. 8th ed. New York: Oxford
University Press.
De Long, J.Bradford, and Lawrence H. Summers. 1992. “Macroeconomic Policy and
Long-Run Growth.” Economic Review Q(IV): 5-29.
Debelle, Guy and Stanley Fischer. 1994. “How Independent Should a Central Bank
be?” Conference Series: Proceedings, Federal Reserve Bank of Boston: 195-225.
Eijffinger, Sylvester, and Jakob de Haan. 1996. “The Political Economy of
Central-Bank Independence.” Special Papers in International Economics 19. International Finance Section, Princeton University.
Grilli, Vittorio, Donato Masciandaro, and Guido Tabellini. 1991. “Political and
Monetary Institutions and Public Finance Policies in the Industrial Countries.”
Economic Policy 6(13): 342-392.
Jordan, Thomas J. 1997. “Disinflation Costs, Accelerating Inflation Gains, and Central
Bank Independence.” Review of World Economics 133 (1): 1-19.
Klomp, Jeroen, and Jakob de Haan. 2009. “Central Bank Independence and Financial
Instability.” Journal of Financial Instability 5(4): 321-338.
Klomp, Jeroen, and Jakob de Haan. 2010. “Do Central Bank Law Reforms Affect The
Term in Office of Central Bank Governors?” Economics Letters 106: 219-222.
McKinnon, Ronald I. 1973. Money and Capital in Economic Development. Washington
DC: Brookings Institution.
Mishkin, Frederic S. 1996. “The Channels of Monetary Transmission: Lessons for
Monetary Policy.” National Bureau of Economic Research Working Paper 5464.
Piazzesi, Monika, and Martin Scheneider. 2007. “Inflation Illusion, Credit, and Asset
Saxton, Jim. 2003. “Argentina’s Economic Crisis: Causes and Cures.”
http://www.hacer.org/pdf/Schuler.pdf
Sikken, Bernd J., and Jakob de Haan. 1998. “Budget deficits, monetization, and
central bank independence in developing countries.” Oxford Economic Papers 50(3): 493-511.
Walsh, Carl E. 1994. “Central Bank Independence and the Short-Run Output-Inflation
Trade-Off in EC.” Central of German and European Studies Working Paper.
Walsh, Carl E. 2003. Monetary Theory and Policy. 2nd ed. Cambridge, MA: MIT Press. Wie, Thee K. 2003. “The Indonesian Economic Crisis and the Long Road to Recovery”.
Australian Economic Historical Review 43(2): 183-196.
Wooldridge, Jeffrey M. 2005. Introductory Econometrics: A Modern Approach. 3rd ed.
Appendix
Table A1: Data Availability
Variable Num. Countries Periods Sources
Dependent:
Ratio Between Banks’ Claims on the government and the private sector (Govtcr/pvtcr)
44 1980 - 2008
(not complete for every year)
International Financial Statistics (IFS), IMF Credit Growth:
‐ Private credit growth (Pvtcrgr) ‐ Government credit growth
(govcrgr)
44 1980 - 2008
(not complete for every year)
International Financial Statistics (IFS), IMF
Independent:
Lag of Central Bank Institutional Reforms (DCBIL)
44 1980 – 2008 Acemoglu et.al (2008)
Daunfeldt et.al (2010) Lag of Government Budget Balance
(BalL)
44 1980 – 2008
(not complete for every year)
International Financial Statistics (IFS), IMF
Lag of Financial Deepening (DeepL) 44 1980 - 2008
(not complete for every year)
International Financial Statistics (IFS), IMF
Lag of Income per Capita (GDPCapL) 44 1980- 2008
(not complete for every year)
International Financial Statistics (IFS), IMF
Lag of Inflation (InfL) 44 1980 - 2008
(not complete for every year)
Table A2: The list of countries and the central bank reform
Developed Countries Developing Countries
Countries Law Reform Countries Law Reform Countries Law Reform
Australia 1996 Argentina 1992 Malaysia 1994
Bahamas 2000 Barbados - Mauritius 2004
Bahrain - Botswana 1996 Mexico 1994
Canada - Burundi - Nepal 2002
Denmark - Chile 1989 New Guinea 2000
Iceland 2001 Colombia 1992 Pakistan 1997
Korea 1997 Costa Rica 1995 Paraguay 1994
Singapore - Dominican Rep 2002 Philippines 1993
Sweden 1999 Ecuador 1992 South Africa 1996
Switzerland 1999 Egypt 1993 Sri Lanka 2002
United States - El Savador 1991 Swaziland -
Guatemala 2002 Thailand - Hounduras 1996 Turkey - India - Uganda 1993 Indonesia 1999 Uruguay 1995 Jordan - Venezuela 1992 Kenya 1995
