Monetary Transmission in Brazil: Has the Credit Channel Changed?
Author: Ms. Mercedes Garcia-Escribano1
  • 1 0000000404811396https://isni.org/isni/0000000404811396International Monetary Fund

Contributor Notes

Author’s E-Mail Addresses: mgarciaescribano@imf.org
Publication Date:
17 Dec 2013
eISBN:
9781484312001
Language:
English
Keywords:
WP; lending; bank; private bank; credit concession; bank characteristic; lending capacity
Download PDF (842.4 KB)

This paper investigates the transmission of monetary policy by private banks in Brazil during the recent easing cycle. The analysis presented uses a panel dataset with information on lending by private banks in Brazil and concludes that monetary transmission through lending volumes was not impaired. Instead, the observed diminished lending appears to be related to supply and demand factors, as well as to the rapid expansion of public banks’ lending.

Abstract

This paper investigates the transmission of monetary policy by private banks in Brazil during the recent easing cycle. The analysis presented uses a panel dataset with information on lending by private banks in Brazil and concludes that monetary transmission through lending volumes was not impaired. Instead, the observed diminished lending appears to be related to supply and demand factors, as well as to the rapid expansion of public banks’ lending.

I. Introduction

During the past two years, private banks’ credit seems to have been less responsive to monetary policy changes than in previous loosening cycles. In August 2011, the Brazilian central bank started an easing cycle. Since then and until the start of the tightening cycle in April 2013, the policy rate was cut by 525 bps to 7.25 percent, a decade’s low. Despite the substantial monetary stimulus in place, credit growth by private banks continued to decline, partly reflecting that new lending operations were recovering only very gradually.

Several factors could help explain the apparently diminished monetary transmission. The surge in NPLs that private banks experienced since late-2011 could explain the more cautious and limited supply of credit. On the demand side, weak consumer and business confidence, coupled with elevated household indebtedness and relatively high debt service ratios could have also been holding back demand for credit by corporates and individuals. Another factor that could have played a role in explaining the dynamics of private banks’ credit is the expansion of credit by public banks, which in recent years have been growing at sizable rates and may be competing with private credit.

This paper explores why monetary transmission to lending volumes has been lower than in the past. The paper examines if the diminished monetary transmission has been due to a weakening of the credit channel or if, shifts in underlying supply and/or demand factors could help explain the delay in transmission. The paper also analyzes if during the recent monetary easing cycle, there were differences in the monetary transmission across banks depending on their ownership (private domestic or foreign banks). The impact of the expansion of public bank lending on lending by private banks is also analyzed; in particular, we test if competition between private and public banks has changed.

II. Stylized facts

Has monetary transmission via the lending channel changed? Figure 1 shows the negative co-movements of cumulative changes in monetary policy and new credit growth. The negative correlation between Selic changes and credit growth seems to have weakened in the recent monetary easing cycle, particularly for bank lending to corporates.

Figure 1.
Figure 1.

Brazil. Credit Growth and Monetary Policy

Citation: IMF Working Papers 2013, 251; 10.5089/9781484312001.001.A001

Sources: Central Bank of Brazil; and Fund staff calculations.

The delayed monetary transmission to credit does not appear to have affected lending rates as the pass-through of the policy rate to loan rates has been complete. In contrast with the previous easing cycle, the interest rate pass-though during the last easing cycle was initially delayed, but eventually, private banks fully passed the policy cuts to lending rates for firms and individuals, with even a tightening of the spreads for some loan segments.

Figure 2.
Figure 2.

Brazil: By Borrower, Changes in Selic and Lending Growth

Citation: IMF Working Papers 2013, 251; 10.5089/9781484312001.001.A001

Sources: Central Bank of Brazil; and Fund staff calculations.
Table 1.

Brazil: Monetary Transmission Through Lending Rates

article image
Sources: Central Bank of Brazil; and Fund staff calculations.

Other factors affecting credit supply or demand may help understand credit dynamics. Elevated household indebtedness and the surge in NPLs may have prompted banks to be more cautious and selective in their lending practices. Uncertain global and domestic economic conditions—reflected in weakened consumer and business confidence—as well as elevated household indebtedness may have lowered demand for credit by consumers and corporates. The indicator of credit requests by Serasa shows that demand for credit by consumers was relatively weak during 2012 and picked up in early 2013. In contrast, corporates’ demand for credit that had also weakened since end-2011 continued to be soft.

Increased lending by public banks may have also played a role. Public banks’ credit expanded at an annual rate of nearly 25 percent during 2011–12, resulting in an increase in their share of total credit to 48 percent (compared to 42 percent by end-2010). Heightened competition from public banks during this period may also help explain the lower sensitivity of private banks’ credit growth to Selic changes.

Figure 3.
Figure 3.

Brazil: Serasa Indicator of Credit Requests

Citation: IMF Working Papers 2013, 251; 10.5089/9781484312001.001.A001

Sources: Serasa Experian Brazil; and Fund staff calculations.
Figure 4.
Figure 4.

Brazil: Credit Growth by Bank Ownership

Citation: IMF Working Papers 2013, 251; 10.5089/9781484312001.001.A001

Sources: Central Bank of Brazil; and Fund staff calculations.

III. Methodology and Data Description

A panel dataset with quarterly data comprising 37 private banks (19 domestic and 18 foreign) is used to explain the credit channel of the monetary transmission mechanism. During the sample period that extends from the first quarter of 2005 to the fourth quarter of 2012, monetary conditions were eased and tightened in several occasions. The observed volume of new loans by private banks reflects the equilibrium between credit supply and credit demand. The empirical specification presented in this paper explains the growth in new lending as a function of supply and demand factors, including changes in the Selic rate, bank characteristics, proxies for corporate and individual demand for credit, and new lending extended by public banks.

The empirical specification is as follows:

Δlnxijt = β0+β1ΔSelict+β2Δln⁡publicit+β2kit+β3Aijt+β4Zjt+υij+εijt

where xijt stands for new loans to sector i, either a corporate or an individual, by private bank j at time t; Selic is the policy rate of the Brazil Central Bank; publicit represents new loans extended by public banks to sector i; kit is a variable representing the demand for credit by sector i (proxied by confidence index, expectations index, or the Serasa credit request indicator); Aijt includes factors that could affect private banks’ perception for credit risk (some factors may be bank and sector specific, such as the non-performing loan ratio, while other factors are common to all banks, such as VIX, EMBI, the Bovespa stockprice index or the level of household debt service); Zjt are factors that limit bank’s lending capacity (some could be bank specific like bank capitalization or liquidity, or common to all banks, like reserve requirements); υij is the bank specific fixed effect when lending to sector i; and ∊ijt is the error term.

A dummy variable δt that takes the value of 1 during 2012Q1-2012Q4 will allow identifying if the growth of private banks’ new credit was lower during the last monetary easing cycle than during the rest of the sample period. The interaction of δt with a few of the other explanatory variables—such as the change in the Selic—will allow examining if the sensitivity of credit growth to any of these variables changed recently.

Tables 2.a and 2.b present the definition of the variables and the descriptive statistics. Figure 5 shows the time evolution of these variables.

Table 2.a.

Descriptive statistics

article image
Source: Fund staff calculations.
Table 2.b.

Definition of Variables

article image
Source: The Brazil Central Bank provided a dataset containing time series information on new credit concessions, non-performing loans, loans, total regulatory capital, risk-weighted assets, total assets, liquid assets, short-term liabilities, and ROA for each of the private banks included in the sample. The dataset also included information on new credit concessions by public banks. The rest of the sources are Haver and Serasa Experian Brazil.
Figure 5.
Figure 5.
Figure 5.

Description of the Variables

Citation: IMF Working Papers 2013, 251; 10.5089/9781484312001.001.A001

Sources: Fund staff calculations.

IV. Results

This section summarizes the main results. Table 3 explores the presence of changes in monetary transmission during the recent easing cycle. Table 4 tests the role of the Selic as well as demand and supply factors in explaining changes in lending growth during the sample period. Table 5 focuses on understanding the lower lending growth during the recent cycle, in particular, if the sensitivity of lending growth to either of the explanatory variables has changed recently. Table 6 assesses if public bank lending has any impact on private banks’ credit.

Table 3.

Monetary Transmission and Changes during the 2012 Easing Cycle (Dependent Variable: dlnx)

article image
*p<0.1 **p<0.05 ***p<0.01Source: Fund staff estimates.
Table 4.

Impact of Selected Factors on Lending Growth (Dependent Variable: dlnx)

article image
Source: Fund staff estimates.
article image
Source: Fund staff estimates.
article image
*p<0.1 **p<0.05 ***p<0.01Source: Fund staff estimates.
Table 5.

Testing for Changes in the Sensitivity of Lending Growth to Factors

article image
*p<0.1 **p<0.05 ***p<0.01Source: Fund staff estimates.
Table 6.

Impact of Public Banks’ Lending

article image
article image
*p<0.1 **p<0.05 ***p<0.01Sources: Fund staff estimates.

Table 3 confirms that lending growth was weaker since end-2011. Changes on the Selic have a negative and statistically significant impact on credit growth. Lending growth during the past cycle has been significantly lower than during the rest of the sample period, as shown by the negative coefficient on the dummy variable δt in Column I. The coefficient for the interaction between Selic changes and δt in Column II is positive and statistically significant indicating that monetary transmission has been weaker during the recent cycle. However, the coefficient sign reverts and is statistically significant when the dummy δt is included in the regression (Column III), implying that when controlling for other factors contributing to the lower credit growth in the past cycle, the reductions in the Selic did have a positive impact on loan disbursements; thus, it is not possible to conclude that the lower lending growth recently observed was due to limited or impaired monetary transmission mechanism. Similar results are obtained when the sample includes only lending to individuals (Columns V–VI) or to corporates (Columns VII–VIII). Column IV explores if lending from foreign banks differs from domestic private banks during the past cycle; the coefficient on the interaction term is not statistically significant and thus, suggests this is not the case.

Demand and supply factors are main determinants of lending growth. Table 4 shows the estimates for a range of demand and supply factors using the total sample. Three different of variables—confidence index, expectations index and Serasa credit requests—are tested as proxies for demand for credit. The three of them have a positive and significant impact on credit growth. Perception of changes in credit risk matters for credit supply. Heightened market risk (proxied with either increases in the EMBI or VIX, or declines in the Bovespa stockprice index) has a negative impact on credit growth. Similarly, a deterioration in the credit portfolio (captured in an increase in the NPL or decrease in the ROA) leads to lower supply of credit by banks. Table 4 also displays the impact of bank’s balance sheet variables related to the bank’s lending capacity on credit growth. Increases in reserve requirements should limit bank’s funding, but the coefficient sign is instead positive indicating the presence of endogeneity (i.e. the central bank increases reserve requirements in response to excesses in credit growth). Bank capitalization and liquidity have the expected sign and are statistically significant.

The lending channel was not weakened during the last monetary easing cycle. Column I in Table 4.c displays the regression results when including the full set of explanatory variables. Column II adds the dummy variable δt in the set of regressors; the coefficient is negative and statistically significant, confirming that the behavior of credit growth during the past year was different than during the rest of the sample period. Column IV displays the results when interacting the change in the Selic with the dummy δt. The negative sign on the Selic indicates not only that the lending channel was not impaired, but that after controlling for the rest of the factors that weakened lending growth, the lending channel was more effective than during the rest of the sample. The results are similar when the sample is limited to lending to individuals (Column VII) or lending to corporates (Column X) in the sense that the interaction term is not statistically significant, suggesting that the monetary transmission mechanism during the recent cycle was working similarly as during the rest of the sample period.

Recently, banks have become more cautious in their lending practices. Banks became more sensitive in their lending to changes in monetary policy during the past easing cycle (as shown by the negative and significant coefficient on the interaction of the dummy δt and the Selic change in Table 5), and also to macroeconomic or global environments, captured by the EMBI (Column III). The coefficient on the interaction of the EMBI and δt is negative and statistically significant; further, when the interaction term is added in the regression, the coefficient on the dummy δt reverts its sign, indicating that the weaker monetary transmission during the past year is explained by the heightened sensitivity of banks to the Selic changes and to the EMBI. Other coefficients on the interaction terms with δt are not statistically significant (as the one for non-performing loans), but the signs are as expected, providing further evidence of the increased caution of banks in their lending practices.

Public banks compete with private banks in some credit segments, but the competition did not intensify recently. Coefficient results on the variable lending growth by public banks, dlnpublic, displayed in Columns I–IV in Table 6 indicate that when lending to individuals, public banks’ lending moves in tandem with private banks’ lending, suggesting differences regarding lending products or type of borrower. However, there seems to be competition between private and public banks when lending to corporates (Columns V–VIII). The interaction of dlnpublic with δt is negative but not statistically significant, thus, providing no strong evidence of increased competition in the recent period in the sense of public banks targeting more similar customers or offering more similar products to those by private banks than during the rest of the sample. Columns IX–XIII display similar results when the sample is limited by lending product.

V. Conclusions

The transmission of monetary policy has worked efficiently during the last monetary easing cycle. This paper has provided evidence that though private bank lending has been weaker since end-2011 despite the substantial monetary policy cuts, the monetary transmission neither through lending rates or volumes was impaired, but instead, the sensitivity of lending to Selic changes seems to have increased.

The observed diminished lending appears to be related to supply and demand factors. Shifts in the demand for credit have led to a weaker lending growth. In addition, loan supply was impacted by tighter banks’ profitability and level of capitalization. Banks have also exhibited an increased caution in their lending practices given their recent experience with the surge in non-performing loans and heightened market risk.

Public banks lending seems to have also contributed to the weaker lending by private banks. While public banks’ lending have traditionally moved in tandem with private banks’ lending when extending loans to individuals, there seems to be competition/substitution effect in other lending areas. Though this paper does not find evidence that the competition has increased, to the extent that public banks’ credit has expanded rapidly, public banks may have contributed to the diminished impact of monetary policy on private banks’ lending.

VI. References

  • Arena, M., Reinhart, C. and F. Vazquez, 2007, “The Lending Channel in Emerging Economies: Are Foreign Banks Different?” IMF Working Paper 07/48 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

  • Coelho, C., De Mello, J. M. P., and M. Garcia, 2010, “Identifying the Bank Lending Channel in Brazil Through Data Frequency” Departamento de Economia PUC-Rio Working Paper No. 574.

    • Search Google Scholar
    • Export Citation

  • Castro, P, and J. M. P. De Mello, 2010, “Há assimetria no repasse dos juros bancários de variações na taxa Selic?” Departamento de Economia PUC-Rio Working Paper No. 580.

    • Search Google Scholar
    • Export Citation
1

The author would like to thank Martin Kaufman and the Brazil Central Bank for their feedback and extensive discussions.

Same Series

Other IMF Content

Other Publishers

Cover IMF Working Papers
Monetary Transmission in Brazil: Has the Credit Channel Changed?
Author: Ms. Mercedes Garcia-Escribano
Volume/Issue:
Volume 2013: Issue 251
Publisher:
International Monetary Fund
ISBN:
9781484312001
ISSN:
1018-5941
Pages:
20
DOI:
https://doi.org/10.5089/9781484312001.001

Brazil. Credit Growth and Monetary Policy

Brazil: By Borrower, Changes in Selic and Lending Growth

Brazil: Serasa Indicator of Credit Requests

Brazil: Credit Growth by Bank Ownership

Description of the Variables

Description of the Variables