Responses of Retail Interest Rate to Policy Rate in Indonesia
Aditya Suselo1 and Dony Abdul Chalid2[*]
1Departement of Management, Faculty of Economics and Business, Universitas Indonesia Kampusbaru UI, Depok 16424, West Java, Indonesia
2Departement of Management, Faculty of Economics and Business, Universitas Indonesia Kampusbaru UI, Depok 16424, West Java, Indonesia
ABSTRACT
Transmission of monetary policy have several channels, and one of them is through interest rate channel. Using banks data in Indonesia for period 2006-2016, this study attempt to measure the change of bank’s interest rate toward the change in central bank’s policy rate, by seeing the momentum of interest rate pass through using error – correction model.The results show the adjustments of retail bank rates as between loans and deposit, between type, and across maturities are different. The results provide some evidences about the inefficiency of monetary policy through interest rate channel of in Indonesia.
Keywords : Bank, Interest rate pass through, central bank policy rate, commercial bank interest rate, Indonesia
1. Introduction
The effectiveness of monetary policy is influenced by the monetary policy transmission mechanism that influences the magnitude of the influence of monetary policy on real sector conditions. The magnitude of the interest rate pass-through¸ is defined as a change in the interest rate of the bank when there is a change in the Central Bank's benchmark interest rate (Amarasekara, 2005; Gigineishvili, 2011), which is often referred to as monetary policy transmission power can be known (De Bondt, 2002). The interest rate pass through is said to be perfect if one unit changes at the central bank's benchmark interest rate, resulting in a single unit of change in the interest rate of banking (Amarasekara, 2005). Thus, the more perfect the rate of interest rate pass through that occurs, the process of monetary policy transmission mechanism through the interest rate channel will also be more effective (Mojon, 2000). Conditions such as economic cycles, imperfect information, internal banking conditions and the presence of volatility risks that prevent banks from responding fully to changes in benchmark interest rates resulting in asymmetric interest rate pass through or imperfect interest rate pass through (Kuan M. , Binh NTT, and Hui WS, 2008). This imperfect interest rate pass throughs varies for every economic condition (Gigineishvili, 2011). Cottareli and Kourelis (1994) define imperfect interest rate pass through as interest rate stickiness, where changes in money market and banking interest rates change smaller and tend to have delays in both the short term and long term.
Empirical studies that discuss the issue of asymmetry of momentum of interest rate pass through have been done quite a lot (De Bondt, 2002; Epinosa Vega and Rebucci, 2003; Tieman, 2004). This study wanted to see how far the interest rates of banks can move following policy rates, as measured by the momentum of interest rate pass through in Indonesia. In its development, based on UU No. 3 of 2004, Pasal 7, it is explained that Bank Indonesia which functions as Central Bank has monetary authority, in order to maintain the stability of Indonesian rupiah currency. One of the efforts of Bank Indonesia to maintain the stability of currency values is to establish the Inflation Targeting Framework (ITF) in 2005.
The more perfect momentum of the interest rate passthrough between the reference rate and the interest rate of the bank, the more effective the monetary policy applied by the Central Bank (Mojon, 2000). This study will try to measure the momentum component of the interest rate pass through to measure the effectiveness of the BI rate as an operational instrument of monetary policy through the interest rate channel, which is shown by the magnitude of the pass-through coefficient between the interest rate and the Banking interest rate.
2. Literature Review
Previous studies by several researchers (Hannan & Berger, 1991, Schnolnick, 1996, Dueker, 2000; Lim, 2001; Sarno & Thornton, 2003; Hofmann & Mizen; 2004; GambacortaLannotti, 2007; Payne & Waters, 2008 ) found that the interest rates of Commercial Banks between one Bank and another were asymmetric. In addition, they also found that in more concentrated markets the level of price stiffness is also greater, and changes in deposit rates tend to be more rigid when the magnitude of the change is increased rather than declining.
Referring to the framework that has been made by (Hannan & Berger, 1991; Schnolnick, 1996) to analyze the asymmetric rate adjustment of deposit and lending rates at commercial banks in Singapore and Malaysia, using cointegration method. It was found that Banks that would adjust their interest rates by decreasing, would tend to be faster than adjustments by increasing interest rates.
Then there is research on Banking in Australia conducted by Lim (2001). The research used multivariate asymmetric error correction model. The results of this study indicate that the Bank tends to make easy adjustment of interest rates when monetary easing compared to the time of monetary tightening. Then, it was found that the rate of adjustment of interest rates in the short term is still asymmetry, but has become symmetry in the long term.
A study conducted by Cecchetti (1999) suggests that the problem of asymmetry in monetary policy transmission in the European Banking system is due to the differences in the financial structure of each country in Europe. Then, still on the continent of Europe, research conducted by Favero et al. (1999) suggests that credit channels are a critical channel in identifying the asymmetry of monetary policy transmission in Europe.
Several studies have shown that the rate of asymmetry of Banking interest rates with short-term periods is lower than long-term, due to the effect of sticky lending rates (Borio & Fritz, 1995; DonnayDegryse, 2001; Gambacorta, 2008). The notion of such sticky lending rates is the time required by the Bank to make adjustments to changes in the benchmark interest rates made by the Central Bank. According to (Stiglitz & Weiss, 1981; Calem et al., 2006) the explanation of the delay can be due to the agency cost and customer switching cost.
Burgstaller (2003) looks at the dynamic response of lending rates to changes in policy rates and money market rates in Austria. Through the Structural Vector Autoregression (SVAR) method, Burgstaller shows that the power and rate of transmission of interest rates depends on whether the interest rate increasing or declining. Implementation of the European Monetary Union (EMU) in 1999 also had a significant impact on reducing asymmetric effects and accelerating transmission.
Hovarth, et al (2004) also tested the interest rate pass-through mechanism in Hungary. Through the Error Correction Model (ECM) method it was found that the adjustment of Banking interest rate incomplete pass-through and stiffness. Meanwhile, through the Threshold Auto Regressive (TAR) method it is found that the rate of adjustment of the Bank's interest rate depends on the size of the official interest rate change and the long-term balance. The yield shock and volatility also affect the speed of adjustment.
Added some other studies related to interest rate pass-through, including Sander and Kleimeier (2002) which obtained results that with imperfect competition and cost adjustment frameworks, there has been an increasing rigidity in deposit rates and reduced stiffness in loan rates in European countries.
Based on research by ApergisCooray (2015), this study will examine the amount of Interest Pass Through in Indonesian Banking industry. To see these magnitudes, ApergisCooray (2015) connects the benchmark interest rate with deposit rates and loan rates . The relationship between the benchmark interest rate with the deposit rate and loan rate is measured through the amount of Interest Rate Pass Through, where as the magnitude of the Interest Rate Pass Through indicates that the relationship is going on perfectly and quickly. (Amarasekara, 2005) it is explained that a perfect Pass Through occurs when a single unit change in the central bank's benchmark interest rate, will result in a single unit of change in the interest rate of the Banking. Empirically, these changes can be observed through changes in the prices of assets (Credit) and liabilities (Deposit) money offered by the Banking.
Based on a study conducted by Zulkhibri (2012), it was found that the short-term and long-term interest rate pass through in Malaysia is imperfect, with the amount of momentum different for each financial institution and its products. Then, Rocha (2012) also found that the interaction between loan rates, deposit rates and interbank rates in Banking Portugal is asymmetric, and has diverse values between loan and deposit sectors with different tenors.
3. Research Method
3.1. Data
This research uses banking data in Indonesia in 2006 - 2016 period. The sources used for data at Banking level are from Eikon, Datastream, BI publication, annual financial report of Bank and Report of Central Bureau of Statistics "BI Rate and Interest Rate of Rupiah Credit by Group of Banks ".
3.2. Research Model
In line with research conducted by ApergisCooray (2015), this study will use regression model to know the amount of Interest Rate Pass Through from Banking industry, measured by short term and long term, by looking at deposit rates and loan rates. The research model is based on and adapted by the presence of cointegration among the variables to be tested, so it is used Error Correction Model (Liu et.al, 2008; Zulkhibri, 2012; Rocha, 2012). This research will use the main model as follows:
?yt = β0?xt + δ(yt-1 – α0 – α1Xt-1) + i?xt-i + i?yt-i
Where yt is the interest rate of the Bank which is the time series of the dependent variable:
• Working capital loanrate (LNRATEMK)
• Investment loan rate (LNRATEINV)
• Consumer loan rate (LNRATEC)
• Deposit rate with 1 month tenor (DRATE1M)
• Deposit rate with 3 month tenor (DRATE3M)
• Deposit rate with 6 month tenor (DRATE6M)
• Deposit rate with 12 month tenor (DRATE12M)
• Deposit rate with 24 month tenor (DRATE24M)
While Xt is the central bank's benchmark interest rate (BIRATE) which is the time series of independent variables.
Non-perfect long-term pass through will also have different coefficient values from one statistically, which is smaller. The faster and perfect rate of interest rate pass-through between the benchmark interest rate and the Bank's interest rate, it indicates that the monetary policy applied by the Central Bank can be well implemented, and it also shows that the lower the friction in the monetary policy transmission mechanism through the interest rate channels (Mojon, 2000).
4. Results
4.1.Descriptive Statistics
% / BI RATE / DRATE1M / DRATE3M / DRATE6M / DRATE12M / DRATE24M / LNRATEC / LNRATEINV / LNRATEMKMean / 7,54 / 7,61 / 7,99 / 8,23 / 8,51 / 8,83 / 14,97 / 12,75 / 13,13
Median / 7,5 / 7,18 / 7,49 / 7,87 / 8,23 / 8,99 / 14,48 / 12,32 / 12,76
Maximum / 12,75 / 12,01 / 12,32 / 12,2 / 12,38 / 12,93 / 17,88 / 15,94 / 16,35
Minimum / 4,75 / 5,35 / 5,61 / 5,87 / 5,81 / 5,36 / 13,03 / 11,14 / 11,35
Standar Deviation / 1,76 / 1,58 / 1,68 / 1,6 / 1,81 / 1,98 / 1,53 / 1,32 / 1,35
Observation / 133 / 133 / 133 / 133 / 133 / 133 / 133 / 133 / 133
Table 1.0Descriptive Statistics Variable
Source: Processed Researcher (2017)
From the loan side, the descriptive statistics in Table 1.0 confirm that there is a significant difference between the BI Rate and loan rates, which indicates that one of the ways banks earn income is through loan channeling. From the mean value of each type of loan, investment loan has the lowest mean value, followed by working capital loan and the highest consumption loan. Intuitively, investment loan is supposed to have the lowest margin. This is explained by the purpose of financing from investment loan, which is to finance the fixed assets which will generally be a guarantee of the loan transaction. From the standard deviation value of each type of loan, it is seen that investment loan has the lowest standard deviation followed by working capital loan and consumption loan. The high standard deviation on consumption shows that the change in the consumption loan rates is relatively dynamic, since banks make periodic adjustments to consumer loans, because users of consumption loans are bulk, so banks have a high risk because they can not pay much attention to user profiles consumption loans appropriately. Then, for working capital loans and investment loans have a lower standard deviation. Indicates that the volatility or changes in lending rates are relatively more rigid. This can be due to the fact that the use of investment and working capital generally has a relatively long period of time, and the bank can clarify loan users more accurately. The three types of loans also appear to have a standard deviation below the BI rate as the reference rate.
4.2. Results
The stationary test results indicate that all time series variables used are not stationary at stationary level but stationary at the first derivative level I (1), which is tested at a 95% confidence level. Therefore, the Johansen cointegration test can be performed on all time series variables. The selection of lag uses two tests: Akaike's Information Criterion (AIC) and Schwarz Information Criterion (SIC). Both methods have the advantage of other methods because these two methods are suitable for time series data. AIC and SIC can explain the suitability of the model with existing data and the value occurring in the future (Gujarati, 2003: 536). The best model is the model that has the smallest AIC and SIC value (GklezakouMylonakis, 2010: 318). The maximum time lag selection is limited to 6 months, which refers to studies and observations made by Mojon (2000) on the ECB, which explains that studies with monthly data intervals should use the maximum time lag 6. Based on the optimal lag time test results, it can be seen that each variable has an optimal lag that is different from one another. The determination of a large lag was decided by the suitability of AIC and SIC tests. When there is a difference between the decision and the AIC and the SIC, the researcher adjusts to the smallest selection of AIC and SIC.
The Johansen Cointegration Test can be seen that there is no cointegration between (DRATE1M, BIRATE); (DRATE3M, BIRATE); (DRATE12M, BIRATE) and (LNRATEC, BIRATE). Thus, the error-correction model estimation will be performed using BIRATE as an independent variable, and (DRATE6M, DRATE24M, LNRATEINV and LNRATEMK) bivariate (1 dependent 1independent), which can be seen in table 1.1 below.
The error correction model estimation is performed by performing optimum lag specifications, as well as the use of assumptions in accordance with (Wojcik, 2011). In accordance with the model specification that can provide the most appropriate estimation results, the estimated long-term and short term interest rate pass through can be done as follows:
This table summarizes the estimated value of coefficientsα1, obtained from the VECM estimation results between independent and dependent variables.
Dependent / Independent / Assumtion / α1 / Pass Through RateDRATE6M / BIRATE / #4 / 0,481299* / Imperfect
DRATE24M / BIRATE / #4 / 1,804644* / Perfect
LNRATEINV / BIRATE / #4 / 0,232985* / Imperfect
LNRATEMK / BIRATE / #2 / 0,690603* / Imperfect
*significant at α = 5%
Table 1.1Estimated Long Term Interest Rate Pass Through Indicator
Source: Processed Researcher (2017)
The assumptions used in ECM estimates are based on the characteristics of the given interest rate. Deposit rate with a tenor of 6 months and 24 months is a deposit rate that has a long period of time so it can bring up a pattern. Similarly for investment loan rate are generally given for long-term purposes such as buying fixed assets, so that long-term nature will bring a pattern. Therefore, the deposit rate with the tenor of 6 months and 24 months and the investment loan rate will use assumption # 4. In contrast to the working capital loanrates, which are generally provided for immediate purposes, such as salaries of employees, to buy raw materials, and the magnitude of these requirements is very fluctuating making it difficult to find a pattern. Thus, the working capital loan rate will use assumption # 2.
Based on the result of research, the estimation of error correction model on long term is significant at 95% confidence level. In general, the momentum of interest rate pass through deposits and loan, is in line with the Bank Indonesia reference rate. The interest rate response also depends on the internal condition of the banking system itself. Magnitude pass through coefficients smaller than 1 indicate that there is friction within the monetary policy transmission mechanism through long-term interest rates. It can be interpreted that after reaching the long-term equilibrium, a change of 1 unit of BI Rate will have a change effect of less than 1 unit of Banking interest rate except for deposit rate with 24 month tenor, which can be seen the pass through degrees are declared perfect because the magnitude α1 is greater than 1 (1.8046> 1), which has the sense that there is a strong reaction of the Banking in the form of deposit products with a longer tenor.
Like an empirical study conducted by some former researchers, the Bank's interest rate will reach a long-term equilibrium, which means that the Bank's interest rate will be more sticky to the benchmark interest rate when the time period of the Bank's interest rate is longer. In this study, the deposit rate with 24 months tenor and investment loan rate were classified as long-term, but pass through rate was found to be perfect only at the deposit rate with 24 monthstenor and was not found in the investment loan rate.
The pass through rate is found to be perfect at the deposit rate with 24 months tenor because the deposit can be one of the Bank's funding sources, due to the long period of time. Thus, in this case banks will adjust interest rates with reference rates, so that deposit products with a 24-month tenor becomes attractive. On this long-term deposit, what the Bank intends to obtain is the certainty that the Bank's internal condition has adequate liquidity conditions, so that the margin to be achieved through deposit products with 24-months tenoris also not too significant. In general, a perfect pass through is due to the different internal conditions of the banking system. For banks with stable and large internal conditions, then the bank has the power to provide a relatively low interest rate, because the bank has a reputation that leads to public trust that the bank will not go bankrupt. Conversely, for banks with relatively inadequate internal conditions, the bank will provide interest rates to customers with larger quantities.