1 BANK Of ZAMBIA S S U U R R V V E E Y Y O O N N H H O O W W C C O O M M M M E E R R C C I I A A L L B B A A N N K K S S D D E E T T E E R R M M I I N N E E L L E E N N D D I I N N G G I I N N T T E E R R E E S S T T R R A A T T E E S S I I N N Z Z A A M M B B I I A A September 2010 2 1.0 I I n n t t r r o o d d u u c c t t i i o o n n 1.1 As Government has indicated its intention to shift monetary policy away from monetary targeting towards interest rate targeting, the Bank of Zambia (BoZ) has embarked on conducting preliminary research to assess the feasibility of an interest rate targeting framework in Zambia. Gaining a thorough understanding of the interest rate decision-making process undertaken by commercial banks in Zambia would not only assist in the determination of an appropriate policy rate, but would also enable the Bank of Zambia to ascertain the transmission channel through which the policy rate would be most effective. 1.2 Evidence from numerous interest rate targeting central banks indicates that the policy rate should be aimed at influencing developments in the interbank rate, which is then expected to affect borrowing costs along the yield curve. The interbank market is therefore expected to play a crucial role in the implementation of the interest rate targeting framework. 1.3 Investigating why the lending rates are high was also an area of great policy interest. 1.4 The main objective of the survey was therefore to identify the factors, both quantitative and qualitative, that commercial banks consider in making decisions regarding their base lending rates. The specific objectives were twofold: (a) Assess to what extent the interbank market influenced the cost of funds in the interest rate determination process; and, (b) Ascertain which factors have significantly contributed to the high level of lending interest rates currently prevailing in the market. 1.5 The key question posed to commercial banks was: What factors do you take into consideration when determining the base lending rate for Kwacha/Foreign currency loans? A formal model of the calculation method for determining the base rate was also requested, as well as the minutes from Assets and Liabilities Committee (ALCO) meetings in which the interest rate decisions were discussed. All of the 18 registered commercial banks in Zambia were surveyed over the period 1 st – 12 th March, 2010. 1.6 Overall, it was observed that the most common factors considered in the rate setting process were, as expected, the regulatory cash reserve requirements – namely, the statutory reserve ratio (8%), core liquid asset ratio (9%) and the BoZ supervisory fee (0.2% of deposits). Other factors which were considered significant in the determination of base lending rates included: Treasury bill and GRZ bond yield rates; operating costs; cost of funds, i.e. weighted average deposit rates; return on shareholder’s equity and the cost of non- performing loans. The qualitative factors highlighted included, credit risk premiums, the demand and supply for credit and the industry trend in base lending rates. 1.7 The survey results indicated that only half of the banks surveyed considered inflation explicitly in their determination of base lending rates; although some banks indicated that inflation was taken into account when calculating real returns. It was also found that almost all the banks do not consider the interbank rate, or the BoZ overnight facility rate in their calculation of base lending rates. 3 1.8 Furthermore, several of the banks stated that qualitative or “judgemental” factors contributed significantly in the determination of their base lending rates. In particular, they noted that large information asymmetries within the domestic market, as well as the high default culture experienced in Zambia, resulted in large risk premiums being attached to key macroeconomic factors, such as inflation and Treasury bill yield rates. 1.9 These findings have two key implications: the first being for implementation of an interest rate targeting framework in Zambia, and the second being the prevalence of high lending rates. Firstly, as the interbank market is expected to be the transmission channel for the framework, a policy rate that is linked to the interbank rate or overnight rate may not have the desired effects on interest rates in the economy, as it will have no bearing on the banks’ cost of funds. In particular, further analysis indicated that there is a weak correlation between the weighted lending base rates and the interbank rate (0.50) while there is a stronger correlation between the weighted lending base rates and the OMO rates (0.72). This suggests that, as an alternative, a policy rate linked to the OMO rate may be more effective. 1.10 Secondly, with regards to high lending rates, qualitative factors used widely in the rate determination process may dampen the intended effect of a policy decision. For example, a policy rate adjustment intended to lower interest rates in the economy may not be effective if large information asymmetries and high credit default rates remain. 1.11 Overall, it was clear from the survey that there are several issues that need to be addressed before a significant reduction in lending rates in the market is observed; and more importantly, before an effective interest rate targeting framework can be implemented in Zambia. 1.12 The rest of the report is organized as follows. Section 2 outlines the methodology employed. This is followed by a discussion of the survey results. Section 4 concludes, focusing on the way forward. 2 2 . . 0 0 M M e e t t h h o o d d o o l l o o g g y y 2.1 The survey was undertaken using a structured questionnaire over the period 1 st March to 12 th March, 2010. The questionnaire was supplemented with interviews between BoZ staff and representatives from all the 18 registered commercial banks. The questions posed in the questionnaire are listed below: (i) What factors do you take into consideration when determining the base lending rate for Kwacha loans? (ii) What factors do you take into consideration when determining the base lending rate for foreign currency loans? (iii) Kindly rank the importance of these factors, separately for the Kwacha and Foreign Currency lending rates. (iv) Does the importance of these factors change? If yes, under what circumstances? 4 (v) Kindly provide a computer spreadsheet which shows the formula used in computing the base lending rates from 2005 to 2009. It is expected that the spreadsheet contains all the factors mentioned which are used in computing the base lending rates. A soft copy will be preferred. (vi) Kindly provide Assets and Liabilities Committee (ALCO) 1 Minutes and Packs for December 2005, 2006, 2007, 2008 and 2009. (vii) Please assist us with any other relevant information. 3 3 . . 0 0 S S u u r r v v e e y y F F i i n n d d i i n n g g s s a a n n d d A A n n a a l l y y s s i i s s 3.1 This section presents the findings of the survey, based on the information provided by each of the commercial banks. These are discussed in turn below. Determination of Base Lending Rates 3.2 Taking all the commercial banks’ responses into account, we summarised the key factors considered in the base lending rate decision-making process in terms of cost of funds, economic conditions, market conditions and political risks. As can be noted from Table 1, the most common factors considered in the rate setting process are cost of funds: cash reserve requirements – namely, the statutory reserve ratio, core liquid asset ratio and the BoZ supervisory fee; operational costs; and yield rates on Government securities. This is followed by market conditions: credit risk, industry trend, interbank rate, overnight facility and demand and supply of credit. 3.3 The survey results indicated that only half of the banks consider economic conditions, in this case, inflation, explicitly in their determination of base lending rates. Furthermore, while it is understood that the interbank rate represents the cost of short-term liquidity, it is evident from Table 1, that all banks, with the exception of one bank, do not take the interbank rate into account while four banks indicated that they consider the BoZ overnight facility rate in their determination of the base lending rate. 3.4 It was also found that the ranking of factors depended primarily on the bank’s profit motive. For example, while the Treasury bill yield rates are considered by all banks, and by implication one is likely to rank them highly and thus give them a relatively larger weighting in the calculation method, a fall in the yield rates should result in a fall in the base lending rate. However, this is hardly the case. This, therefore, suggests that achieving the required return on equity and covering operational costs are, among other factors, more important factors in the determination of base lending rates. 3.5 From the foregoing, one is bound to ask the following two questions: (i) What are the implications of these findings for the interest rate targeting framework in Zambia? 1 Assets and Liabilities Committee (ALCO) is a senior management committee in a bank or thrift institution, responsible for coordinating the institution's borrowing and lending strategy, and funds acquisition to meet profitability objectives as interest rates change. 5 (ii) What are the implications of these findings for the prevailing high lending rates in the economy? 6 2 Operating costs in this case include the following: management fees, staff costs, transaction costs and communication costs, costs of provisioning, internal cash reserves, projected profit and cost of capital (return on equity). Bank I H G F E D C B A J K L M N O P Q R Cost of Funds √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ Statutory reserve requirement √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ Core liquid asset requirement √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ BoZ supervisory fee √ √ √ √ √ √ √ √ √ √ √ √ Taxation √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ Weighted average deposit rate √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ Operating costs 2 √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ Economic Conditions √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ T-bill/GRZ bond rates √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ Inflation √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ Exchange rate √ √ √ √ √ √ √ √ √ √ √ √ Market Conditions √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ √ Credit risk premium √ √ √ √ √ √ Liquidity premium √ √ Interbank rate √ √ Overnight facility rate √ √ √ √ √ √ √ √ Demand and supply √ √ √ √ √ √ Market expectations √ √ √ √ √ √ √ √ √ √ √ √ √ √ Industry trend √ √ Political Conditions Political risk/Country risk premium √ √ √ √ Table 1: Aggregate results of the factors considered in the determination of Kwacha base lending rates 7 Interbank Rate and Policy Rate 3.6 Since we have observed that the interbank rate is not a significant input in the calculation of banks base lending rates, the introduction of a policy rate which is expected to influence the interbank rate will not have the desired effects on commercial bank interest rates, and ultimately inflation, in the economy. 3.7 Thus, as an alternative, it may be necessary to consider the possibility of using a policy rate that is linked to the Open Market Operations (OMO) rate rather than the interbank or overnight facility rate. This is because other studies conducted in the Bank have shown that there is a strong correlation between the OMO rates and base lending rates in Zambia – with a correlation coefficient of 0.72, compared to a correlation coefficient of 0.50 between the interbank rate and base lending rates. In addition, and as is the case in South Africa, interest rates in the money market are influenced through the use of a repurchase (repo) rate and OMO. This refinancing mechanism has allowed the South African Reserve Bank to effectively administer its inflation targeting regime. Inflation and Lending Rates 3.8 Despite the survey results showing that only half of the banks used inflation in determining the base lending rates, our analysis suggests that inflation is taken into account. This assertion is supported by Graph 1, which depicts a positive relationship between the weighted lending base rate (WLBR) and inflation, over a 10 year period from 2000 to 2009, with a correlation coefficient of 0.75. 3.9 Although Graph 1 shows that overall, the WLBR and inflation moved in the same direction, inflation declined from 16% in 2008 to 9.9% in 2009, but the WLBR rose from 20.8% to 22.6%. This could be due to the fact that changes in the WLBR lag those in inflation, or that there are other factors, such as risk aversion in the recessionary climate and large information asymmetries, that result in the WLBR remaining significantly higher than inflation. Graph 1: Weighted Lending Base Rate (WLBR) and Inflation, 2000 to 2009. 0 10 20 30 40 50 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Weighted Lending Base Rate INFLATION Percent 8 Government Securities yield rates and Lending Rates 3.10 The relationship between the WLBR and Treasury bill yield rates, over the same 10 year period, is shown in Graph 2. As is evident from the graph, there is also a positive relationship between the WLBR and Treasury bill yield rates, with a correlation coefficient of 0.89. This suggests that, as indicated by the banks, Treasury bill yield rates should play a significant role in the determination of the base lending rates. Graph 2: Weighted Lending Base Rate (WLBR) and T-bill yield rate, 2000 to 2009 3.11 Why then have the lending rates not declined in line with the recent fall in Treasury bill yield rates? From our analysis, we can infer that the commercial banks’ base lending rates seem to be sticky downwards in response to declining Treasury bill yield rates. This is especially evident in Graph 2. 3.12 It should also be noted that although inflation and yield rates tend to be relatively unstable, the banks’ base lending rates tend to generally remain stable for long periods of time, suggesting that there could be other factors that dominate the banks determination of base lending rates. However, over a long period of time, a sustained downward adjustment in macroeconomic fundamentals, such as inflation, should eventually result in lower lending rates in the economy. Base Lending Rate used as a Reference 3.13 Although banks set the base lending rates and announce these rates in the market, it is generally expected that the actual lending rates given on loans and advances differ considerably from the base rates. In addition, it appears that the market is divided between prime borrowers, who are able to borrow at the base rate minus some margin, and individual clients, considered more risky, who borrow at the base rate plus some margin. 0 10 20 30 40 50 60 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Weighted Lending Base Rate 91 day Treasury Bill yeild rate Percent 9 3.14 In addition, it was found that most of the banks set their base lending rates qualitatively during the Assets and Liabilities Committee (ALCO) meetings, which primarily assess the borrowing and lending strategy of the bank, among other things, with the view to attaining the profitability objectives of the bank. This suggests that interest rate adjustments seem to be dominated by the banks profit motives rather than the developments in economic fundamentals. Large Information Asymmetries 3.15 The margins charged on loans and advances are, in some cases, excessively high. We observed that this was partly due to the lack of accurate information on borrowers and the “default culture” inherent in the Zambian market. The introduction of the CRB is therefore expected to help in eliminating the information asymmetries, and thus reduce the credit or default risk premium that is included by all banks in the determination of base lending rates. 3.16 Nonetheless, while the CRB was noted as a welcome development by most banks, the survey results indicated that currently the CRB falls short of expectations. Banks were of the view that the CRB’s scope of coverage was too narrow as it was restricted to information provided to it by commercial banks alone. In this regard, it was suggested that the scope of coverage be widened beyond commercial banks, in that information regarding the credit history of clients and employees of other credit-providing institutions be provided to the CRB as a statutory requirement. Excess Liquidity in the Inter-Bank Market 3.17 During the period of the survey, it was found that the excess liquidity in the inter-bank market had resulted in little activity within the market, as well as limited use of the overnight lending facility introduced by the BoZ. The interest rate on the overnight lending facility was found to be punitive (at a 6% margin to the interbank rate), thus giving the impression that a bank accessing the facility is in distress. In this regard, most banks noted that the margin currently applicable on the overnight lending facility should be adjusted downwards and that the interest rate on the facility should not be linked to the inter-bank market. Rather, the interest rate on the central bank’s facility should be independently determined, with the policy rate as a reference. 3.18 Furthermore, it was observed that several smaller banks were unable to access funds within the interbank market, despite the apparent excess liquidity in the market. The reduction of the overnight lending facility rate would therefore make the facility a more viable option for banks that cannot access funds within the interbank market. Operational Costs in the Banking Sector are high 3.19 From the survey, it was found that the operational costs, especially staff costs, for most commercial banks are high and this has a bearing on the determination of base lending rates. In particular, staff loans had, on one occasion, been explicitly included in the calculation of the base lending rate. This, it can be inferred that these loan costs were being passed directly onto clients. The high staff costs may be due to the fact that new banks entering the market have to “poach’’ staff from existing banks, therefore resulting in higher salaries which become sticky downwards. 3.20 In addition, the high operational costs in the banking sector could be an indication of inefficiencies in the banks’ operations, which are then passed on to their clients through high 10 lending rates. In view of the high operational costs, it would be difficult to achieve a significant and sustainable reduction in lending rates regardless of the positive developments in macroeconomic fundamentals, unless competition and innovation are enhanced. 3.21 Further analysis into the nature of the operating costs in the banking sector highlighted specific concerns with regards to efficiency and returns on equity. Table 2 depicts selected operating ratios for each of the commercial banks surveyed, from 2006 to 2009. Efficiency refers to the ability of a bank to generate enough income to cover its non-interest expenses. 3.22 Table 2 also shows that salaries and employment benefits continue to make up a significant portion of operating costs. Although the average salaries to operating costs ratio was between 30% and 50% from 2006 to 2009, salaries for several of the banks reached approximately 60% of operating costs over the last 4 years. 3.23 Given improvements in technology and the relative increase in competition due to the entry of more banks in the market, it is expected that the efficiency in the banking sector should improve over time. Efficiency ratio of 60% or less is considered to be favourable. 3.24 The efficiency ratios presented in Table 2 indicate that operational efficiency within the domestic banking sector has been unfavourable over the period. While it is understood that the global financial crisis had a significant negative impact on the income-generating ability of many banks in 2008 and 2009, several of the banks have had unfavourable efficiency ratios for a number of years. For example, the operational efficiency ratio for Bank H has been above 80% over the past four years, reaching 236% in 2009; and the efficiency ratio for bank L rose from 85% in 2006 to 140% in both 2007 and 2008. 3.25 Further analysis of the relationship between the banking industry efficiency ratio and the lending base rates, as indicated in Graph 3, shows that increased inefficiency partially led to high interest rates. In 2006, based on the efficiency threshold of 60%, the industry was inefficient and correspondingly the base rates were high. However, in 2007 the lending base rate declined despite the efficiency ratio increasing. This can be attributed to the favourable macroeconomic conditions experienced in 2007. In 2008, the lending base rate and industry inefficiency increased and worsened in 2009, as a result of the global financial crisis. [...]... the interbank market influences the cost of funds and thereby the lending interest rates; and (ii) To ascertain which factors have significantly contributed to the high level of lending interest rates in Zambia 17 4.2 From the foregoing, it is clear that there are common factors which all the commercial banks take into account in the determination of base lending rates These include cost of funds, economic... base rate in the market, inter-bank rate, overnight facility and demand and supply of credit However, economic conditions and in this case inflation was found to be directly taken into account in the determination of base lending rates by only half of the surveyed banks The exchange rate is included in the economic conditions 4.3 The factors vary among the banks according to their impact on the cost... inflation, exchange rates and Treasury bills, in determining lending rates That is, other things equal, it is expected that when inflation or Treasury bill rates are low, lending rates should also be adjusted downwards (iii) While qualitative factors are important in the determination of lending rates, in a liberalized economic system like Zambia, macroeconomic factors should be dominant to the determination... 5 that the lending rates declined with an improvement in efficiency from 2006 to 2007 However, in 2008 the lending rates increased despite an improvement in efficiency In 2009, there was an increase in efficiency, and to cover the increasing expenses the lending rates also increased 11 Graph 5: Bank K Efficiency Ratio and Base Rates, 2006-2009 72 26 70 24 Efficiency (%) 66 22 64 Lending Rates (%) 68... line Further, we can conclude from the survey results that the decisions made in setting and adjusting base lending rates are largely decided qualitatively, though some few banks base their decisions on quantitative computations using specified formulas 4.4 The survey results indicated that the interbank rate, which is expected to influence the policy rate, is not a significant input in the determination... respectively, while lending rates were 14% and 18%, respectively Thus, holding other things equal, it can be argued that that the profitability in the Zambian banking sector may be generated primarily through high lending rates charged in the domestic economy, rather than through cost efficiencies Macroeconomic Fundamentals 3.34 The history of consistently high inflation in Zambia has resulted in economic agents... the survey highlighted that there are other qualitative factors such as high default risk and large information asymmetries that contribute to high lending rates, despite the current positive macroeconomic conditions In addition, we found that operational inefficiencies and the need for high returns on the shareholders equity may also have contributed to the high lending rates in the domestic economy... key to base rate determination process were not actually reflected in the formulas provided; and, (iv) Some factors were not disaggregated into various components, especially with respect to operating costs 4.0 Conclusion 4.1 The undertaking of this survey was broadly aimed at identifying the factors that commercial banks consider in making decisions regarding their base lending rates The specific objectives... input in the determination of the lending base rates Our analysis indicated a correlation coefficient of 0.50 between the weighted lending base rate and the interbank rate, versus a correlation coefficient of 0.72 between the weighted lending base rate and the OMO rate In light of these findings, it may be necessary to consider linking the policy rate to an alternative market interest rate, such as... the highest lending rate in the region from 2006 to 2009 Graph 7, Graph 8, Graph 9 and Graph 10 illustrate the return on equity and actual average lending rates for selected countries in the region in 2006, 2007, 2008 and 2009, respectively Graph 7: Return on Equity and Actual Average Lending Rates – 2006 Return on Equity and Lending Rates 70 60 40 30 20 a U ga fr ic A ou th nd a e Return on Equity S . administer its inflation targeting regime. Inflation and Lending Rates 3.8 Despite the survey results showing that only half of the banks used inflation. 3.3 The survey results indicated that only half of the banks consider economic conditions, in this case, inflation, explicitly in their determination of