Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Examining the effect of credit on monetary policy with Markov regime switching: Evidence from Turkey

Examining the effect of credit on monetary policy with Markov regime switching: Evidence from Turkey Economics and Business Review, Vol. 8 (22), No. 4, 2022: 68-87 DOI: 10.18559/ebr.2022.4.4 Examining the ee ff ct of credit on monetary policy with Markov regime switching: Evidence from Turkey Ali İlhan Abstract: iTh s paper analyses the ee ff ct of credit on monetary policy responses for die ff rent regimes in Turkey. To do so, the Taylor rule augmented with the credit gap is estimated by using a Markov regime switching model from January 2006 to December 2019. The empirical findings identify two regimes: the low- and high-interest rate re - gimes. The prevalence of the former indicates policy authorities’ growth priorities. Furthermore, die ff ring responses across the regimes ree fl ct that the Central Bank of the Republic of Turkey has an asymmetric policy stance. In the low-interest rate regime, the monetary policy only signic fi antly responds to inflation. In the high-interest rate regime, both inflation and credit have signic fi ant positive impacts on interest rate set - ting. This indicates that credit conditions ae ff cted the tightening of the monetary policy stance in Turkey despite the use of macroprudential tools ae ft r the global financial crisis. Keywords: credit, financial stability, monetary policy, macroprudential policy, Markov regime switching, Turkey. JEL codes: C24, E44, E52, E58. Introduction Since the global financial crisis (GFC) showed that price stability is not suf - ci fi ent to ensure financial stability, the search for alternative policy tools and frameworks has accelerated. While there were debates on the role of monetary policy in financial stability, international financial institutions focused on es - tablishing a macroprudential policy framework that directly targets systemic risk in the financial system (FSB, IMF & BIS, 2011a, 2011b). Article received 26 May 2022, accepted 15 November 2022. Department of Economics, Faculty of Economics and Administrative Sciences, Tekirdağ Namık Kemal University, 59030, Tekirdağ, Turkey, ailhan@nku.edu.tr, https://orcid.org/0000- 0001-6201-5353. Before the GFC, many economists were already pointed out that price stability might not guarantee financial stability and monetary policy should consider financial stability (Borio & Lowe, 2002; Borio, English & Filardo, 2003; White, 2006). A. İlhan, Examining the ee ff ct of credit on monetary policy with Markov regime switching 69 Despite the consensus reached for a proactive response to n fi ancial instabili - ties, views die ff r regarding which policy or policy mix best does this. Suggestions for the best policy framework for achieving price and financial stability mainly fall into two groups. The first group proposes a policy framework in which monetary policy focuses on price stability while macroprudential policy pur- sues financial stability. Ee ff ctive macroprudential policies, which should be the first line of defence against financial instabilities, allow monetary policy to focus on price stability (Svensson, 2012, 2017; Ekholm, 2013; Ozkan & Unsal, 2014; Laeven, 2016). In contrast, the second group claims that macroprudential policies are usually inadequate, so monetary policy should lean against macro- n fi ancial imbalances. Central banks should consider financial stability in their interest rate setting, while monetary policy and macroprudential policy should complement each other to achieve their goals (Woodford, 2012; Angeloni & Faia, 2013; Borio, 2014; Rungcharoenkitkul, Borio, & Disyatat, 2019; Adrian, 2020). Reasonable credit growth is vital for macro-financial stability, especially in emerging market economies (EMEs). Jordà, Schularick, and Taylor (2011) showed that the best indicator of financial instability is credit growth. Excessive credit growth increases the risk of a price bubble when it flows into asset mar - kets. As the share of credit in financing consumption and investment increases, it leads to an unsustainable debt burden and damages economic activity (Gross & Zahner, 2021). Furthermore, the rapid credit growth increases depreciation pressure in the foreign exchange market, whose stability is crucial for EMEs (Aizenman & Binici, 2016; İlhan, Akdeniz, & Özdemir, 2022). Agénor and Pereira da Silva (2019) proposed a framework leaning against the credit cycles for EMEs called integrated inflation targeting (IIT). IIT suggests that the credit growth gap should be included in central banks’ reaction function while mon- etary policy and macroprudential instruments should be calibrated together. One of the EMEs that adopted the lean against the wind strategy ae ft r the GFC is Turkey. The domestic and external demand die ff rentiation associated with accelerating capital ino fl ws created dilemmas for the country’s existing policy framework. Accordingly, in late 2010, a new policy framework, namely the new policy mix, was implemented under the Central Bank of the Republic of Turkey (CBRT) leadership to simultaneously ensure price and n fi ancial sta - bility. In this new policy mix, n fi ancial stability was adopted as a supplementary goal of monetary policy while many macroprudential tools were introduced. Monetary and macroprudential policies were implemented in coordination to complement each other. The intermediate targets were controlling credit growth and slowing short-term capital ino fl ws while the intermediate variables were credit and the exchange rates (Başçı & Kara, 2011; Kara, 2013). i Th s paper analyses the ee ff ct of credit on monetary policy responses with a Markov regime switching (MS) model in Turkey. More specic fi ally, the cred - See Smets (2014) for a detailed review of policy frameworks for financial stability. 70 Economics and Business Review, Vol. 8 (22), No. 4, 2022 it-gap augmented Taylor rule is estimated using the Markov regime switching intercept autoregressive heteroscedasticity (MSIAH) model between January 2006 and December 2019. Unlike previous studies, the ee ff ct of credit on mon - etary policy is examined for die ff rent regimes in Turkey. For example, TVP- VAR models allow quantifying the gradual evolution of the interaction between series throughout the sample period, and parameters can change without the subsamples (Çatik & Akdeniz, 2019). However, the MSIAH model den fi es dif - ferent regimes and captures the ee ff cts of policy stance by providing regime dependent coec ffi ients (Baharumshah, Soon & Wohar, 2017). Thus, the MSIAH model can estimate the impact of credit on monetary policy depending on the CBRT’s policy stance. i Th s paper contributes to the literature as follows. First, it analyses two dif - ferent interest rate regimes with differing coefficients and provides asymmetric monetary policy behaviour in Turkey. Second, the prevalence of the low-inter- est rate regime throughout the sample period ree fl cts the growth priorities of the policy authorities. Finally, in the low interest rate environment, monetary policy focused only on inflation whereas also the credit gap ae ff cted monetary policy decisions in the high-interest rate regime. The signic fi ance of the ee ff ct in the high-interest rate regime indicates that credit conditions contributed to a tightening of monetary policy stance in Turkey. e p Th aper proceeds as follows. Section 1 explains Turkey’s monetary and macroprudential policy responses to credit developments in full-fledged in - flation targeting. Section 2 reviews the empirical literature. Section 3 describes the data and methodology. The following section provides the empirical results and discussion. The final section concludes the paper. 1. Policy responses to credit developments in Turkey e T Th urkish banking sector, which operated as a public debt financier in the 1990s due to high bond yields, returned to its traditional intermediary function ae ft r the 2001 crisis and credit volume grew. The improvement in credit quality due to a decline in the ratio of non-performing loans encouraged the percep- tion that rapid credit growth indicated the normalisation of the banking sector (Kenc, Turhan, & Yildirim, 2011). In May 2006, however, a currency shock in- terrupted credit growth when the policy rate was increased by 425 basis points in two months (CBRT, 2022a) due to a sharp depreciation of the Turkish Lira (TL) and increased inflation expectations (CBRT, 2008). The monetary policy responses to restore price and exchange rate stability slowed credit growth (see Figure 1). Despite the modest recovery ae ft r the first months of 2007, credit growth sue ff red from the GFC. The policy rate was reduced by 650 basis points between December 2008 and April 2009 to limit the GFC’s harmful ee ff cts on economic activity and financial stability (CBRT, 2009). A. İlhan, Examining the ee ff ct of credit on monetary policy with Markov regime switching 71 e q Th uantitative easing and unconventional monetary policies of advanced economies to alleviate the adverse ee ff cts of the GFC accelerated capital ino fl ws to EMEs, thereby creating new macro-financial risks. Exposed to these risks, as are many EMEs, Turkey was forced to switch its policy framework. In the new policy mix, the CBRT announced that monetary policy considers n fi an - cial stability while emphasising the use of non-rate instruments under vari- ous scenarios (CBRT, 2010). Accordingly, the CBRT employed reserve require- ments and unique tools for n fi ancial stability, such as the asymmetric interest rate corridor and the reserve option mechanism. Moreover, in the second half of 2011, the Banking Regulation and Supervision Agency (BRSA) introduced various macroprudential instruments to control credit growth (CBRT, 2014b). Due to these measures and the impact on capital flows of the sovereign debt crisis in the euro area, credit growth declined from the third quarter of 2011 (see Figure 1). 20 70 Lehman currency the taper Brothers second shock talk macroprudential bankruptcy macroprudential loosening 60 tightening policy rate 5 30 hike first currency macroprudential –5 shock credit-to-GDP gap beginning of tightening 0 the new credit growth policy mix –10 –10 Note: The left axis stands for the credit-to-GDP gap, while the right denotes credit growth. All values are in percent for both axes. The credit-to-GDP is the ratio of credit from all sectors to private non-financial sectors to GDP. The credit-to-GDP gap is the die ff rence between the actual value of the credit-to-GDP and its trend. Credit growth is the annual growth of total credit volume in the banking sector’s balance sheets. Figure 1. Credit-to-GDP gap and credit growth in the Turkish economy: 2006Q1–2019Q4 Source: Author’s construction based on CBRT (2022b) and BRSA (2022). From the second half of 2012, an increasing global risk appetite accelerated capital ino fl ws to Turkey while falling interest rates to ease appreciation pres - sure in TL encouraged credit growth (CBRT, 2013). However, once this exceeded a reasonable level (Kara, Küçük, Tiryaki, & Yüksel, 2013), the authorities un- der the leadership of the BRSA further tightened the macroprudential policy stance by introducing new tools and strengthening existing tools (CBRT, 2014b). 2006-Q1 2007-Q1 2008-Q1 2009-Q1 2010-Q1 2011-Q1 2012-Q1 2013-Q1 2014-Q1 2015-Q1 2016-Q1 2017-Q1 2018-Q1 2019-Q1 72 Economics and Business Review, Vol. 8 (22), No. 4, 2022 Besides these macroprudential instruments, a tightening monetary policy stance associated with global financial conditions also helped restrain credit growth. Following the taper talk in May 2013 (Bernanke, 2013), its implemen- tation in December 2013 accelerated capital outo fl ws from EMEs. The CBRT hiked interest rates sharply in January 2014 due to worsening TL depreciation (CBRT, 2015). Ae ft r a  series of shocks in the second half of 2016 deepened the slow - down in credit growth, Turkey’s macroprudential policy stance was eased in September 2016 (CBRT, 2016; İlhan, Özdemir, & Eryigit, 2021). While mon- etary policy tightened in 2017 to limit the depreciation in TL associated with global shocks (CBRT, 2018), credit growth was supported by increased credit guarantee fund incentives (IMF, 2018). Although the monetary policy stance was not loosened in the first half of 2018, the pace of credit growth increased. However, the currency shock in August 2018 and the subsequent rise in in- flation resulted in a sharp interest rate hike and falling credit growth (CBRT, 2019). The policy rate, which had remained constant in the first half of 2019, declined by 1200 basis points until the end of the year, resulting in a recovery in credit growth (CBRT, 2020). 2. Literature review Many studies have estimated the exchange rate-augmented Taylor rule in an- alysing the CBRT’s actions to achieve financial stability (Hasanov & Omay, 2008; Civcir & Akçağlayan, 2010; Caporale, Helmi, Çatık, Ali & Akdeniz, 2018; Yağcıbaşı & Yıldırım, 2019; Soybilgen & Eroğlu, 2019; Özdemir, 2020; Tetik & Yıldırım, 2021). However, few have examined the role of credit in Turkey’s monetary policy. Çamlıca (2016) investigated the CBRT’s responses to financial stress using a composite index of systemic stress, which is an indicator of financial risk in the credit, money, equity, forex and bond markets, for 2005:01 – 2015:10. Compared to the pre-GFC period, the CBRT responded more to n fi ancial stress, and adopt - ed a lean against the wind strategy ae ft r mid-2010. Erdem, Bulut, and Kocak (2017) analysed the exchange rate gap- and credit gap-augmented Taylor rule in a time-varying manner using a cointegration test and the Kalman filter for 2006:01 – 2016:02. Although the nominal domestic credit gap and exchange rate gap ae ff cted the interest rate settings, they did not change the priorities of the CBRT in the new policy mix. Chadwick (2018) explored the impact of monetary and macroprudential policies on consumer credit growth with panel VAR for the period from 2005:12 to 2017:12. A contractionary monetary policy has a restrictive impact on credit growth. Moreover, this impact is greater when combined with macroprudential policy. Kurowski, Rogowicz, and Smaga (2020) examined the Taylor rule ex- A. İlhan, Examining the ee ff ct of credit on monetary policy with Markov regime switching 73 tended with the credit-to-GDP gap using the TVP-VAR model from 2002Q1 to 2018Q3. The interest rate settings were adjusted following the Taylor rule and credit conditions were considered in monetary policy decisions ae ft r 2010 in Turkey. The ee ff ct of interest rate settings on credit and inflation increased with the strengthening of monetary transmission mechanisms ae ft r the GFC. 3. Data and methodology iTh s study used an MS model to investigate the ee ff ct of credit on Turkey’s monetary policy between January 2006 and December 2019. e Th starting date corresponds to the adoption of full-fledged inflation targeting while the end - ing date excludes the impact of COVID-19. Taylor (1993) suggested a simple interest rate rule, known as the Taylor rule, for policy authorities focusing on price and output stability. However, increas- ing financial stability concerns have led others to expand this rule by including n fi ancial variables (Käfer, 2014). Following this literature, an augmented Taylor rule is employed to analyse the impact of credit on monetary policy in Turkey. e cr Th edit-augmented Taylor rule to be estimated is as follows: i = α + βi + δ(π – π *) + θ(y – y *) + ω(cr – cr*) (1) t t–1 t t t t t t Here, i stands for the policy rate. The CBRT used die ff rent short-term inter - est rates as policy rates throughout the sample period. Moreover, from late 2010, the multi-policy rate framework was employed in line with an unconventional approach (Binici, Kara & Özlü, 2016). Alp, Kara, Keleş, Gürkaynak and Orak (2010) showed that the strongest predictor of Turkey’s monetary policy stance is the one-week Turkish Lira Libor (TRLIBOR) rate. Similarly, Gürkaynak, Kantur, Taş and Yıldırım-Karaman (2015) used the one-week TRLIBOR to measure the ee ff ctive policy rate. Therefore, the one-week TRLIBOR rate is used in this study to represent the CBRT’s policy rate. π is actual ina fl tion, obtained from the annual percentage change of the con - sumer price index, while π * is the inflation target. The inflation gap ( π – π *) t t t is calculated as the die ff rence between actual inflation and its target. y is the industrial production index, which proxies for output while cr represents the total credit volume in the banking sector. For both variables, the actual values (y , cr ) are transformed into the logarithmic form before applying the Hodrick- t t -Prescott filter for the trend values ( y *, cr *). The trend values are then sub - t t tracted from their actual values to obtain gaps. The series representing season - The Hodrick-Prescott smoothing parameter was set as λ = 14400, while suggest power for λ was set as 2. 74 Economics and Business Review, Vol. 8 (22), No. 4, 2022 al ee ff cts are adjusted with the Census X-13. To eliminate level die ff rences, all series are standardised. e Th series are taken from various databases. e Th one-week TRLIBOR rate is retrieved from the Banks Association of Turkey. The consumer price index and production index of total industry are obtained from the Federal Reserve Economic Data. The banking sector’s total credit volume is retrieved from the CBRT. Table 1 shows the time series properties and descriptive statistics of the variables. Table 1. Descriptive statistics and unit root test results Descriptive statistics Standard Variables Observations Mean Maximum Minimum deviation i 168 12.534 26.318 5.089 5.199 π 168 9.547 25.240 3.986 3.593 π * 168 5.178 7.500 4.000 0.860 y 168 87.524 119.990 56.995 18.714 cr 168 1,010,904 2,587,738 136,063 755,481 Unit root test results Variables LM DT DT 1t 2t i –6.2557*** 2009:08 2018:03 (π – π *) –7.2967*** 2008:11 2018:06 t t (y – y *) –5.9626** 2008:08 2011:01 t t (cr – cr *) –5.3337* 2009:01 2010:10 t t Note: Descriptive statistics are the level forms of the series. ***, **, and * show stationary at 1%, 5%, and 10% signic fi ance level, respectively. Source: Author’s estimation. Inflation and the interest rate reached their maximum levels in September 2018, right ae ft r the currency shock in August. While the minimum value of the interest rate corresponded to the beginning of the taper tantrum, inflation fell to its lowest value in the first months of 2011, when global liquidity was abundant. Actual inflation remained well above the target levels due to recent jumps in inflation. Output was lowest around the GFC whereas it peaked in the last month of 2017. e t Th ime-series properties of the variables were analysed using Lee and Strazicich’s (2003) unit root test, allowing for two structural breaks. The trend break model indicates that all series are stationary in their levels. Furthermore, A. İlhan, Examining the ee ff ct of credit on monetary policy with Markov regime switching 75 the first break dates of all variables correspond to the GFC while the second break dates for the inflation gap and the interest rate fall in the first and sec - ond quarters of 2018, respectively, before TL’s sharp depreciation. The second break of the credit gap corresponds to the beginning of the new policy mix. iTh s study used an MS model to estimate the augmented Taylor rule. MS models identify die ff rent regimes and allow the dynamic behaviour of the vari - ables to be examined depending on each regime. In these models, past regimes can reoccur throughout the sample while the number of regimes generally var- ies from two to at most four (Baharumshah et al., 2017, pp. 249–250). In MS models, the switch in regime is not den fi ed as the outcome of a de - terministic event but by an unobserved random variable (s ), called the regime or state. Since s takes only discrete values in determining the regime, a Markov chain is used for the regime switching process (Hamilton, 1994, pp. 677–678). The N-state Markov chain with transition probabilities is den fi ed as follows: P{s = j|s = i, s = k, …} = P{s = j|s = i} = p (2) t t–1 t–2 t t–1 ij Here, p is the transition probability, which is the probability that state j is ij preceded by state i, and the sum of the transition probabilities is equal to 1. e t Th ransition probabilities can be collected in the following N ∙ N transition matrix (Hamilton, 1994, pp. 678–679): p p  p   11 21 N1   p p   p 12 22 N 2   P =      (3)         p p   p   1N 2N NN   e m Th aximum likelihood (ML), which is the expectation-maximisation (EM) algorithm-based method, is used for the estimation of the MS model. e v Th alue of the ML function increases with each iteration of the EM algorithm (Hamilton, 1994, p. 689). These processes continue until the parameters con - verge (Çatık & Önder, 2011, p. 128). Many MS models allow for switches in the intercept, mean, and variance of the residuals throughout the regimes governed by an unobserved state vari- able. This study used the MSIAH model as it takes into account the entire pa - rameter shift and changes in the variance of the residuals throughout the state (Çatık & Önder, 2011, p. 127). The model is shown in linear form in equation (1), but it can be re-written for the two-regime MSIAH model: e Th EM algorithm, which was developed by Dempster, Laird and Rubin (1977), was em - ployed by Hamilton (1990) for the ML estimation (Krolzig, 1997). 76 Economics and Business Review, Vol. 8 (22), No. 4, 2022 i = β(s )(i ) + δ(s )(π – π *) + θ(s )(y – y *) + ω(s )(cr – cr *) + ε (4) t t t – 1 t t t t t t t t t t e e Th stimated coeci ffi ents are strongly dependent on the state variable ( s ). e Th integer variable s can take the values 1 or 2 to indicate, respectively, that the low-interest or high-interest rate regime prevails. 4. Empirical results and discussion e e Th mpirical findings indicate the existence of both the low- and high-inter - est rate regimes, called Regime 1 and Regime 2, respectively. That is, the CBRT adopted two die ff rent monetary policy stances. Regime 1 ree fl cts a loose stance whereas Regime 2 ree fl cts a tight stance. Table 2 shows the transition matrix and the properties of the two regimes. Table 2. Transition matrix and regime properties Transition matrix Regime 1 Regime 2 Regime 1 0.881 0.118 Regime 2 0.156 0.843 Regime properties Observation Probability Duration Regime 1 97.7 0.568 8.43 Regime 2 70.3 0.431 6.40 Source: Author’s estimation. As the transition matrix in Table 2 shows, the switching probability from the high- to the low-interest rate regime is higher (15.6%) than that of switching from the low- to the high-interest rate regime (11.8%). Furthermore, the regime properties indicate that the low regime was more likely to be implemented and to last for longer. e Th se n fi dings ree fl ct the policy authorities’ monetary policy preferences, which favoured growth. Figure 2 shows changes in regime prob- abilities over the study period. e r Th egime probabilities in Figures 2(b) and 2(c) clearly show that the low- interest rate regime dominated before a surge in capital flows threatened finan - cial stability. The first switch from a low- to high-interest rate regime ae ft r the GFC took place when the banking sector’s annual credit growth was close to e r Th egime classic fi ations are reported in Table A1 in the Appendix. A. İlhan, Examining the ee ff ct of credit on monetary policy with Markov regime switching 77 a) Interest rate b) Low-interest rate regime c) High-interest rate regime Figure 2. Regime probabilities Source: Author’s estimation. 40% nominal. During this period, macroprudential policy practices started to restrain credit growth (Kara, 2016). In the next high-interest rate regime, the taper tantrum began (Bernanke, 2013), and the CBRT hiked the policy rate, citing TL depreciation (CBRT, 2014a). In addition, macroprudential policies were further tightened in response to accelerating credit growth (CBRT, 2014b). Excluding short-lived switches, the subsequent prevalence of the high-inter- est rate regime corresponds to the jumps in the nominal exchange rate in the spring of 2018. While interest rate declined from the second half of 2019, the sample ends with a high-interest rate regime. Prior to estimating the model, its nonlinearity was evaluated with diagnostic tests. As seen in Table 3, the likelihood (LR) linearity test rejects the null hy- 78 Economics and Business Review, Vol. 8 (22), No. 4, 2022 pothesis of linearity and supports the nonlinear model. e l Th og-likelihood and Akaike information criterion (AIC) values of the linear and nonlinear models also verify the two-regime model (Çatık & Önder, 2011; Kumah, 2011). Table 3 presents the empirical results for the estimation. Table 3. Estimation results Regime 1 (Standard error: 0.051) Variables Coefficient Standard error t-value α –0.006 0.007 –0.927 i 0.965*** 0.009 99.072 t–1 (π – π *) 0.031*** 0.010 2.981 t t (y – y *) 0.013 0.009 1.504 t t (cr – cr *) 0.012 0.008 1.461 t t Regime 2 (Standard error: 0.297) Variables Coefficient Standard error t-value α –0.067 0.043 –1.551 i 0.759*** 0.068 11.096 t–1 (π – π *) 0.209*** 0.093 3.188 t t (y – y *) –0.003 0.044 –0.077 t t (cr – cr *) 0.156*** 0.050 3.098 t t LR Linearity Test: 160.059 Chi (6) = (0.000)*** Chi (8) = (0.000)*** Log-likelihood AIC Nonlinear model 98.540 –1.006 Linear model 18.510 –0.148 Note: ***, **, and * show signic fi ance at 1%, 5%, and 10%, respectively. Source: Author’s estimation. As Table 3 shows, the coeci ffi ents die ff r in sign, size, and signic fi ance be - tween the regimes, which supports the nonlinear interest rate setting and the two-regime model. Furthermore, the differentiation of responses is consistent with previous studies reporting an asymmetric monetary policy behaviour for Turkey (Hasanov & Omay, 2008; Caporale et al., 2018; Öge Güney, 2018; Bulut, 2019; Özdemir, 2020). In the low-interest rate regime, inflation positively and signic fi antly impacts the policy rate, whereas the output and credit gap ee ff cts are insignic fi ant. That is, monetary policy focused on the traditional goal, namely price stability, in a low interest rate environment. The coeci ffi ents of the high-interest rate re - A. İlhan, Examining the ee ff ct of credit on monetary policy with Markov regime switching 79 gime indicate that the monetary policy took an additional responsibility. The ee ff ct of the inflation gap is signic fi ant and positive while the coeci ffi ent is larg - er. However, as in the low-interest rate regime, the output gap is insignic fi ant. On the contrary, the credit gap signic fi antly and positively impacts interest rate setting. This indicates that the rising credit gap had an impact on the tighten - ing of monetary policy in Turkey. e Th se empirical findings conr fi m those of Erdem and others (2017), Chadwick (2018) and Kurowski and others (2020), who found that credit conditions af- fected monetary policy decisions in Turkey. Furthermore, the findings consist - ent with those of Çamlıca (2016), who reported that the CBRT adopted a lean against the wind strategy ae ft r mid-2010, to some extent. A robustness check is performed by augmenting the Taylor rule with an- other credit variable. Since many macroprudential tools, especially those im- plemented by the BRSA, mainly aim to control consumer credit growth (CBRT, 2014b), it is worth examining whether consumer credit ae ff cted monetary policy. To this end, the Taylor rule augmented with the consumer credit gap (ccr – ccr *) is estimated. t t e Th empirical findings indicate that the regime probabilities and properties are largely similar to those of the credit-augmented Taylor rule estimation (see Table A3 in the Appendix). In the low-interest rate regime, monetary policy aligns with the standard Taylor rule. Both inflation and the output gap posi - tively and signic fi antly impact the policy rate, whereas the consumer credit gap is insignic fi ant. In contrast, in the high-interest rate regime, inflation and the consumer credit gap signic fi antly and positively ae ff ct the policy rate, where - as the output gap has no signic fi ant ee ff ct. Thus, consumer credits ae ff cted the policy rate despite the availability of many macroprudential tools. In conclu- sion, the robustness check n fi dings verify that credit developments impacted the tightening of the monetary policy stance in Turkey between January 2006 and December 2019. Conclusions i Th s study investigates the ee ff ct of the credit gap on monetary policy responses in Turkey employing the MSIAH model between January 2006 and December 2019. The empirical findings show that in the low interest rate environment, When the credit variable is removed from the equation and the standard Taylor rule is estimated, results are similar to credit-augmented Taylor rule findings. In the high-interest rate regime, only inflation has a signic fi ant and positive impact on interest rate setting, while the out - put gap is insignic fi ant (see Table A2 in the Appendix). Consumer credit is retrieved from the BRSA and similar processes are performed for the consumer credit gap (see Section 3). 80 Economics and Business Review, Vol. 8 (22), No. 4, 2022 monetary policy focuses only on inflation, whereas the credit gap also inu fl - ences monetary policy decisions in the high-interest rate regime. This indi - cates that credit conditions contributed to a tightening of the monetary policy stance in Turkey. e r Th ole of monetary policy on financial stability substantially depends on the performance of macroprudential policy. When macroprudential policy falls short of ensuring financial stability, monetary policy might support macropru - dential policy (Gerlach, 2012). Many studies have reported that macropruden- tial policies have a limiting ee ff ct on credit growth in Turkey (Binici, Erol, Kara, Özlü & Ünalmış, 2013; Bulut, 2015; Bumin & Taşkın, 2016; Yüceyılmaz, Altın & Tunay, 2017; Alper, Binici, Demiralp, Kara & Özlü, 2018; Chadwick, 2018; İlhan et al., 2021). However, empirical findings show that macroprudential policy was not the only way to control credit growth. Similar to what Kurowski and others (2020) pointed out, the determinants of the monetary policy stance in the high-interest rate regime indicate that the policy framework was partially consistent with the IIT strategy. e e Th e ff ct of credit developments on interest rate settings also indicates that monetary policy was complementary to macroprudential policy in Turkey. However, this led to adverse side ee ff cts on price stability. In the new policy mix, monetary policy, which was also concerned with financial stability, devi - ated from its primary goal (Gürkaynak et al., 2015). On the other hand, macro- prudential policy stance loosened prematurely with the increasing impact of growth priorities. Furthermore, macroprudential policy has not been compre- hensive enough, and measures have not directly covered controlling commer- cial credit growth (Kara, 2016; IMF, 2018). Therefore, implementing a  more ee ff ctive macroprudential policy would ease the burden on monetary policy and increase its room for manoeuvre. Price stability-focused and clearer mon- etary policy framework would reduce uncertainty and help achieve the ulti- mate goal of the CBRT. A. İlhan, Examining the ee ff ct of credit on monetary policy with Markov regime switching 81 Appendix Table A1. Regime classification Regime 1 Regime 2 2006:01 – 2006:05 2006:06 – 2006:07 2006:08 – 2008:05 2008:06 – 2008:06 2008:07 – 2008:11 2008:12 – 2009:03 2009:04 – 2011:07 2011:08 – 2012:08 2012:09 – 2013:02 2013:03 – 2015:01 2015:02 – 2015:08 2015:09 – 2016:01 2016:02 – 2016:12 2017:01 – 2017:03 2017:04 – 2018:03 2018:04 – 2018:11 2018:12 – 2019:02 2019:03 – 2019:12 Source: Author’s estimation. Table A2. Estimation results of the standard Taylor rule Regime 1 (n: 99.2, Prob.: 0.577, Duration: 8.61, Standard error: 0.052) Variables Coefficient Standard error t-value α –0.011 0.007 –1.573 i 0.965*** 0.010 95.682 t–1 (π – π *) 0.033*** 0.010 3.258 t t (y – y *) 0.023*** 0.006 3.620 t t Regime 2 (n: 68.8, Prob.: 0.422, Duration: 6.29, Standard error: 0.323) Variables Coefficient Standard error t-value α –0.032 0.046 –0.696 i 0.738*** 0.077 9.564 t–1 (π – π *) 0.252*** 0.072 3.477 t t (y – y *) 0.038 0.050 0.766 t t LR Linearity Test: 156.871 Chi (5) = (0.000)*** Chi (7) = (0.000)*** Log-likelihood AIC Nonlinear model 92.380 –0.956 Linear model 13.944 –0.106 Note: ***, **, and * show signic fi ance at 1%, 5%, and 10%, respectively. Source: Author’s estimation. 82 Economics and Business Review, Vol. 8 (22), No. 4, 2022 Table A3. Estimation results of the consumer credit-augmented Taylor rule Regime 1 (n: 98.8, Prob.: 0.574, Duration: 8.67, Standard error: 0.052) Variables Coefficient Standard error t-value α –0.008 0.007 –1.229 i 0.965*** 0.009 98.427 t–1 (π – π *) 0.032*** 0.010 3.135 t t (y – y *) 0.017* 0.009 1.926 t t (ccr – ccr *) 0.007 0.008 0.970 t t Regime 2 (n: 69.2, Prob.: 0.425, Duration: 6.41 Standard error: 0.307) Variables Coefficient Standard error t-value α –0.074 0.046 –1.596 i 0.741*** 0.070 10.513 t–1 (π – π *) 0.265*** 0.068 3.708 t t (y – y *) –0.022 0.051 –0.436 t t (ccr – ccr *) 0.153** 0.068 2.557 t t LR Linearity Test: 161.223 Chi (6) = (0.000)*** Chi (8) = (0.000)*** Log–likelihood AIC Nonlinear model 96.199 –0.978 Linear model 15.587 –0.114 Note: ***, **, and * show signic fi ance at 1%, 5%, and 10%, respectively. Source: Author’s estimation. References Adrian, T. (2020). “Low for long” and risk-taking. (IMF Departmental Paper No. 15). Retrieved from https://www.elibrary.imf.org/view/journals/087/2020/015/arti- cle-A001-en.xml Agénor, P. R., & Pereira da Silva, L. A. (2019). Integrated inflation targeting: Another per - spective from developing world. Retrieved from https://www.bis.org/publ/othp30.htm Aizenman, J., & Binici, M. (2016). Exchange market pressure in OECD and emerging economies: Domestic vs. external factors and capital flows in the old and new nor - mal. Journal of International Money and Finance, 66, 65–87. http://doi.org/10.1016/j. jimonn.2015.12.008 fi Alp, H., Kara, H., Keleş, G., Gürkaynak, R., & Orak, M. (2010). Türkiye’de piyasa gösterge- lerinden para politikası beklentilerinin ölçülmesi. İktisat İşletme ve Finans, 25(295), 21–45. https://doi.org/10.3848/iif.2010.295.2849 A. İlhan, Examining the ee ff ct of credit on monetary policy with Markov regime switching 83 Alper, K., Binici, M., Demiralp, S., Kara, H., & Özlü, P. (2018). Reserve requirements, liquidity risk, and bank lending behavior. Journal of Money, Credit and Banking, 50(4), 817–827. https://doi.org/10.1111/jmcb.12475 Angeloni, I., & Faia, E. (2013). Capital regulation and monetary policy with fragile banks. Journal of Monetary Economics, 60, 311–324. http://dx.doi.org/10.1016/j. jmoneco.2013.01.003 Baharumshah, A. Z., Soon, S. V., & Wohar, M. E. (2017). Markov-switching analysis of exchange rate pass-through: Perspective from Asian countries. International Review of Economics and Finance, 51, 245–257. http://doi.org/10.1016/j.iref.2017.05.009 Başçı, E., & Kara, H. (2011). Financial stability and monetary policy. (CBT Working Papers No. 08). Retrieved from https://www.tcmb.gov.tr/wps/wcm/connect/EN/ TCMB+EN/Main+Menu/Publications/Research/Working+Paperss/2011/11-08 Bernanke, B. S. (2013). e Th economic outlook . Retrieved from https://www.federalre- serve.gov/newsevents/testimony/bernanke20130522a.htm Binici, M., Erol, H., Kara, H., Özlü, P., & Ünalmış, D. (2013). Interest rate corridor: A new macroprudential tool?. CBT Research Notes in Economics, 20. Retrieved from https:// www.tcmb.gov.tr/wps/wcm/connect/EN/TCMB+EN/Main+Menu/Publications/ Research/Research+Notes+in+Economicss/ Binici, M., Kara, H., & Özlü, P. (2016). Unconventional interest rate corridor and the monetary transmission evidence from Turkey. (CBRT Working Paper No. 08). Retrieved from https://www.tcmb.gov.tr/wps/wcm/connect/EN/TCMB+EN/ Main+Menu/Publications/Research/Working+Paperss/ Borio, C. (2014). Monetary policy and financial stability: What role in prevention and recovery. Capitalism and Society, 9(2), 1. Borio, C., English, W., & Filardo, A. (2003). A tale of two perspectives old or new chal- lenges for monetary policy?. (BIS Working Papers No. 127). Borio, C., & Lowe, P. (2002). Asset prices, financial and monetary stability: Exploring the nexus. (BIS Working Papers No. 114). BRSA. (2022). Banking sector data—quarterly data. Retrieved March 15, 2022 from https://www.bddk.org.tr/BultenAylik/en/Home/Gelismis Bulut, U. (2015). The interest rate corridor as a macroprudential tool to mitigate rapid credit growth in credits: Evidence from Turkey. e Th oretical and Applied Economics , 4(605), 133–144. Bulut, U. (2019). Does the Central Bank of the Republic of Turkey respond asymmetri- cally to inflation and output?. Margin—e Th Journal of Applied Economic Research , 13(4), 381–400. https://doi.org/10.1177/0973801019868374 Bumin M., & Taşkın, F. D. (2016). e i Th mpact of macro-prudential measures on retail loans: Evidence from Turkey. International Journal of Contemporary Economics and Administrative Sciences, 5(1–2), 48–46. Çamlıca, F. (2016). Responsiveness of monetary policy to financial stress in Turkey. Central Bank Review, 16, 143–150. http://doi.org/10.1016/j.cbrev.2016.11.003 Caporale, G. M., Helmi, M. H., Çatık, A. N., Ali, F. M. & Akdeniz, C. (2018). Monetary policy rules in emerging countries: Is there an augmented nonlinear Taylor rule?. Economics Modelling, 72, 306–319. https://doi.org/10.1016/j.econmod.2018.02.006 84 Economics and Business Review, Vol. 8 (22), No. 4, 2022 Çatik, A. N., & Akdeniz, C. (2019). e e Th volution of the monetary transmission chan - nels in Turkey: Evidence from a TVP-VAR model. Applied Economics Letters, 26(13), 1072–1079. https://doi.org/10.1080/13504851.2018.1537468 Çatık, A. N., & Önder, A. Ö. (2011). Inflationary ee ff cts of oil prices in Turkey: A re - gime-switching approach. Emerging Markets Finance and Trade, 47(5), 125–140. https://doi.org/10.2753/REE1540-496X470506 CBRT. (2008). Monetary and exchange rate policy for 2009. Retrieved from https://www. tcmb.gov.tr/wps/wcm/connect/EN/TCMB+EN/Main+Menu/Publications/Moneta ry+and+Exchange+Rate+Policy+Texts/ CBRT. (2009). Monetary and exchange rate policy for 2010. Retrieved from https://www. tcmb.gov.tr/wps/wcm/connect/EN/TCMB+EN/Main+Menu/Publications/Moneta ry+and+Exchange+Rate+Policy+Texts/ CBRT. (2010). Financial stability report No. 11. Retrieved from https://www.tcmb. gov.tr/wps/wcm/connect/EN/TCMB+EN/Main+Menu/Publications/Reports/ Financial+Stability+Report/ CBRT. (2013). Annual report 2012. Retrieved from https://www.tcmb.gov.tr/wps/wcm/ connect/EN/TCMB+EN/Main+Menu/Publications/Reports/Annual+Reports/ CBRT. (2014a). Press release on interest rates, No. 06. Retrieved from https://www. tcmb.gov.tr/wps/wcm/connect/EN/TCMB+EN/Main+Menu/Announcements/ Press+Releases CBRT. (2014b). Financial Stability Report No. 19. Retrieved from https://www.tcmb. gov.tr/wps/wcm/connect/EN/TCMB+EN/Main+Menu/Publications/Reports/ Financial+Stability+Report/ CBRT. (2015). Annual report 2014. Retrieved from https://www.tcmb.gov.tr/wps/wcm/ connect/EN/TCMB+EN/Main+Menu/Publications/Reports/Annual+Reports/ CBRT. (2016). Financial stability report No. 23. Retrieved from https://www.tcmb. gov.tr/wps/wcm/connect/EN/TCMB+EN/Main+Menu/Publications/Reports/ Financial+Stability+Report/ CBRT. (2018). Annual report 2017. Retrieved from https://www.tcmb.gov.tr/wps/wcm/ connect/EN/TCMB+EN/Main+Menu/Publications/Reports/Annual+Reports/ CBRT. (2019). Annual report 2018. Retrieved from https://www.tcmb.gov.tr/wps/wcm/ connect/EN/TCMB+EN/Main+Menu/Publications/Reports/Annual+Reports/ CBRT. (2020). Annual report 2019. Retrieved from https://www.tcmb.gov.tr/wps/wcm/ connect/EN/TCMB+EN/Main+Menu/Publications/Reports/Annual+Reports/ CBRT. (2022a). Central Bank interest rates. Retrieved from https://www.tcmb.gov.tr/wps/ wcm/connect/EN/TCMB+EN/Main+Menu/Core+Functions/Monetary+Policy/ Central+Bank+Interest+Rates/CBRT+INTEREST+RATES CBRT. (2022b). Banking sector total credit volume—quarterly data. Retrieved March 15, 2022 from https://evds2.tcmb.gov.tr/index.php?/evds/serieMarket Chadwick, M. G. (2018). Ee ff ctiveness of monetary and macroprudential shocks on consumer credit growth and volatility in Turkey. Central Bank Review, 18(2), 69–83. https://doi.org/10.1016/j.cbrev.2018.03.001 Civcir, I., & Akçağlayan, A. (2010). Inflation targeting and the exchange rate: Does it matter in Turkey?. Journal of Policy Modeling, 32, 339–354. https://doi.org/10.1016/j. jpolmod.2010.03.002 A. İlhan, Examining the ee ff ct of credit on monetary policy with Markov regime switching 85 Dempster, A. P., Laird, N. M., & Rubin, D. B. (1977). Maximum likelihood from in- complete data via the EM algorithm. Journal of the Royal Statistical Society. Series B (Methodological), 39(1), 1–38. Ekholm, K. (2013, March 13). Monetary policy, business cycle stabilisation and macro- prudential policy. Speech at the Confederation of Swedish Enterprise, Stockholm. Retrieved from https://www.bis.org/review/r130319d.pdf Erdem, E., Bulut, U., & Kocak, E. (2017). Have financial stability concerns changed the priority of the Central Bank of the Republic of Turkey?. Studies in Business and Economics, 12(2), 3–45. https://doi.org/10.1515/sbe-2017-0019 FSB, IMF, & BIS. (2011a, February). Macroprudential policy tools and frameworks— Update to G20 Finance Ministers and Central Bank Governors. Retrieved from https:// www.imf.org/external/np/g20/pdf/021411.pdf FSB, IMF, & BIS (2011b, October), Macroprudential policy tools and frameworks— Progress report to G20. Retrieved from https://www.imf.org/external/np/g20/ pdf/102711.pdf Gerlach, S. (2012). Monetary policy ae ft r the crisis. (Speech at the 44th Annual Money, Macro and Finance Conference, Trinity College, Dublin, 8 September). Retrieved from https://www.bis.org/review/r120910c.pdf Gross, J., & Zahner, J. (2021). What is on the ECB’s mind? Monetary policy before and ae ft r the global financial crisis. Journal of Macroeconomics, 68. https://doi. org/10.1016/j.jmacro.2021.103292 Gürkaynak, R. S., Kantur, Z., Taş, M. A., & Yıldırım-Karaman, S. (2015). Monetary pol- icy in Turkey aer ft Central Bank independence. İktisat İşletme ve Finans, 30(356), 9–38. https://doi.org/10.3848/iif.2015.356.4592 Hamilton, J. D. (1990). Analysis of time series subject to changes in regime. Journal of Econometrics, 45(1–2), 39–70. https://doi.org/10.1016/0304-4076(90)90093-9 Hamilton, J. D. (1994). Time series analysis. Princeton: Princeton University Press. Hasanov, M., & Omay, T. (2008). Monetary policy rules in practice: Re-examining the case of Turkey. Physica A, 387, 4309-4318. https://doi.org/10.1016/j.physa.2008.02.075 İlhan, A., Akdeniz, C., & Özdemir, M. (2022). Analyzing exchange market pres- sure dynamics with Markov regime switching: The case of Turkey. Organizations and Markets in Emerging Economies, 13(1), 238–259. https://doi.org/10.15388/ omee.2022.13.78 İlhan, A., Özdemir, M., & Eryigit, K. Y. (2021). The role of macroprudential policies on controlling credit growth: The case of Turkey. Journal of Yasar University, 16(61), 393–413. https://doi.org/10.19168/jyasar.821930 IMF. (2018). 2018 Article IV Consultation with Turkey. IMF Country Report, No. 18/110. Retrieved from https://www.imf.org/en/Publications/CR/Issues/2018/04/30/ Turkey-2018-Article-IV-Consultation-Press-Release-Staff-Report-and-Statement- by-the-45822 Jordà, Ò., Schularick, M., & Taylor, A. M. (2011). Financial crises, credit booms, and ex- ternal imbalances: 140 years of lessons. IMF Economic Review, 59, 340–378. https:// doi.org/10.1057/imfer.2011.8 Käfer, B. (2014). The Taylor rule and financial stability—A literature review with applica - tion for the eurozone. Review of Economics, 65(2), 159–192. https://doi.org/10.1515/ roe-2014-0204 86 Economics and Business Review, Vol. 8 (22), No. 4, 2022 Kara, A. H. (2013). Monetary policy ae ft r the global crisis. Atlantic Economic Journal, 41, 51–74. https://doi.org/10.1007/s11293-012-9358-7 Kara, A. H. (2016). Turkey’s experience with macroprudential policy. BIS Papers, 86, 123–140. Kara, H., Küçük, H., Tiryaki, S.T. & Yüksel, C. (2013). In search of a reasonable cred- it growth rate for Turkey. Research Notes in Economics, 3. Retrieved from https:// www.tcmb.gov.tr/wps/wcm/connect/EN/TCMB+EN/Main+Menu/Publications/ Research/Research+Notes+in+Economicss/ Kenc, T., Turhan, M. I., & Yildirim, O. (2011). e e Th xperience with macro-prudential poli - cies of the Central Bank of the Republic of Turkey in response to the global financial crisis. (World Bank Policy Research Working Paper No. 5834). Krolzig, H. M. (1997). Markov-switching vector autoregressions: Modelling, statistical in- ference, and application to business cycle analysis. Lecture Notes in Economic and Mathematical Systems, 454. Berlin: Springer. Kumah, F. Y. (2011). A Markov-switching approach to measuring exchange market pressure. International Journal of Finance and Economics, 16, 114–130. https://doi. org/10.1002/ijfe.415 Kurowski, L., Rogowicz, K., & Smaga, P. (2020). What would the economy with credit- enhanced interest rates look like? The case of Turkey. Turkish Studies, 21(4), 620– 644. https://doi.org/10.1080/14683849.2019.1659735 Laeven, L. (2016). Monetary policy in the clutches of financial stability?. In J. Ulbrich, C. C. Hedrich & M. Balling (Eds.), e SS Th M at 1, SUERF Conference Proceedings (pp. 17–32). Vienna: SUERF. Retrieved from http://hdl.handle.net/10419/163519 Lee, J., & Strazicich, M. C. (2003). Minimum Lagrange Multiplier Unit Root Test with two structural breaks. e Th Review of Economics and Statistics , 85(4), 1082–1089. https://doi.org/10.1162/003465303772815961 Öge Güney, P. (2018). Asymmetries in monetary policy reaction function and the role of uncertainties: the case of Turkey. Economic Research—Ekonomska Istraživanja, 31(1), 1367–1381. https://doi.org/10.1080/1331677X.2018.1481445 Özdemir, M. (2020). The role of exchange rate in inflation targeting: The case of Turkey. Applied Economics, 52(29), 3138–3152. https://doi.org/10.1080/00036846.2019.17 Ozkan, F. G., & Unsal, D. F. (2014). On the use of monetary and macroprudential policies for small open economies. (IMF Working Paper No. 112). https://doi. org/10.5089/9781498375429.001 Rungcharoenkitkul, P., Borio, C., & Disyatat, P. (2019). Monetary policy and hysteresis and the financial cycle . (BIS Working Papers No. 817). Smets, F. (2014). Financial stability and monetary policy: How closely ınterlinked?. International Journal of Central Banking, 10(2), 263–300. Soybilgen, B., & Eroğlu, B. A. (2019). Time-varying Taylor rule estimation for Turkey with e fl xible least square method. Boğaziçi Journal Review of Social, Economic and Administrative Studies, 33(2), 122–139. https://doi.org/10.21773/boun.33.2.3 Svensson, L. E. O. (2012). Comment on Michael Woodford, “Inflation targeting and fi - nancial stability” . Sveriges Riksbank Economic Review, 1, 33‒39. Retrieved from http:// archive.riksbank.se/en/Web-archive/Published/Published-from-the-Riksbank/ Economic-Review/Economic-Review/2012/index.html A. İlhan, Examining the ee ff ct of credit on monetary policy with Markov regime switching 87 Svensson, L. E. O. (2017). Cost-benet a fi nalysis of leaning against the wind. Journal of Monetary Economics, 90, 193–213. http://dx.doi.org/10.1016/j.jmoneco.2017.07.004 Taylor, J. B. (1993). Discretion versus policy rules in practice. Carnegie-Rochester Conference Series on Public Policy, 39, 195–214. https://doi.org/10.1016/0167- 2231(93)90009-L Tetik, M., & Yıldırım, M. O. (2021). Distortionary ee ff cts of economic crises on policy coordination in Turkey: Threshold GMM approach. Economics and Business Review, 7(3), 83–102. https://doi.org/10.18559/ebr.2021.3.6 White, W. (2006). Is price stability enough?. (BIS Working Papers No. 205). Woodford, M. (2012). Inflation targeting and financial stability. Sveriges Riksbank Economic Review, 1, 7–32. Retrieved from http://archive.riksbank.se/en/Web- archive/Published/Published-from-the-Riksbank/Economic-Review/Economic- Review/2012/index.html Yağcıbaşı, Ö. F., & Yıldırım, M. O. (2019). Estimating Taylor rules with Markov switch- ing regimes for Turkey. Romanian Journal of Economic Forecasting, 22(3), 81–95. Yüceyılmaz, H. F., Altın, O., & Tunay, K. B. (2017). Makro İhtiyati Tedbirler Çerçevesinde Tüketici Kredilerine İlişkin Düzenlemelerin Etkinliği: Türkiye Örneği. Ege Akademik Bakış, 17(4), 449–459. https://dergipark.org.tr/tr/pub/eab/issue/39982/475204 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Economics and Business Review de Gruyter

Examining the effect of credit on monetary policy with Markov regime switching: Evidence from Turkey

İlhan, Ali
Economics and Business Review , Volume 8 (4): 20 – Dec 1, 2022
Download PDF
20 pages

Loading next page...
 
/lp/de-gruyter/examining-the-effect-of-credit-on-monetary-policy-with-markov-regime-ZabvBwirNt

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher
de Gruyter
Copyright
© 2022 Ali İlhan, published by Sciendo
eISSN
2450-0097
DOI
10.18559/ebr.2022.4.4
Publisher site
See Article on Publisher Site

Abstract

Economics and Business Review, Vol. 8 (22), No. 4, 2022: 68-87 DOI: 10.18559/ebr.2022.4.4 Examining the ee ff ct of credit on monetary policy with Markov regime switching: Evidence from Turkey Ali İlhan Abstract: iTh s paper analyses the ee ff ct of credit on monetary policy responses for die ff rent regimes in Turkey. To do so, the Taylor rule augmented with the credit gap is estimated by using a Markov regime switching model from January 2006 to December 2019. The empirical findings identify two regimes: the low- and high-interest rate re - gimes. The prevalence of the former indicates policy authorities’ growth priorities. Furthermore, die ff ring responses across the regimes ree fl ct that the Central Bank of the Republic of Turkey has an asymmetric policy stance. In the low-interest rate regime, the monetary policy only signic fi antly responds to inflation. In the high-interest rate regime, both inflation and credit have signic fi ant positive impacts on interest rate set - ting. This indicates that credit conditions ae ff cted the tightening of the monetary policy stance in Turkey despite the use of macroprudential tools ae ft r the global financial crisis. Keywords: credit, financial stability, monetary policy, macroprudential policy, Markov regime switching, Turkey. JEL codes: C24, E44, E52, E58. Introduction Since the global financial crisis (GFC) showed that price stability is not suf - ci fi ent to ensure financial stability, the search for alternative policy tools and frameworks has accelerated. While there were debates on the role of monetary policy in financial stability, international financial institutions focused on es - tablishing a macroprudential policy framework that directly targets systemic risk in the financial system (FSB, IMF & BIS, 2011a, 2011b). Article received 26 May 2022, accepted 15 November 2022. Department of Economics, Faculty of Economics and Administrative Sciences, Tekirdağ Namık Kemal University, 59030, Tekirdağ, Turkey, ailhan@nku.edu.tr, https://orcid.org/0000- 0001-6201-5353. Before the GFC, many economists were already pointed out that price stability might not guarantee financial stability and monetary policy should consider financial stability (Borio & Lowe, 2002; Borio, English & Filardo, 2003; White, 2006). A. İlhan, Examining the ee ff ct of credit on monetary policy with Markov regime switching 69 Despite the consensus reached for a proactive response to n fi ancial instabili - ties, views die ff r regarding which policy or policy mix best does this. Suggestions for the best policy framework for achieving price and financial stability mainly fall into two groups. The first group proposes a policy framework in which monetary policy focuses on price stability while macroprudential policy pur- sues financial stability. Ee ff ctive macroprudential policies, which should be the first line of defence against financial instabilities, allow monetary policy to focus on price stability (Svensson, 2012, 2017; Ekholm, 2013; Ozkan & Unsal, 2014; Laeven, 2016). In contrast, the second group claims that macroprudential policies are usually inadequate, so monetary policy should lean against macro- n fi ancial imbalances. Central banks should consider financial stability in their interest rate setting, while monetary policy and macroprudential policy should complement each other to achieve their goals (Woodford, 2012; Angeloni & Faia, 2013; Borio, 2014; Rungcharoenkitkul, Borio, & Disyatat, 2019; Adrian, 2020). Reasonable credit growth is vital for macro-financial stability, especially in emerging market economies (EMEs). Jordà, Schularick, and Taylor (2011) showed that the best indicator of financial instability is credit growth. Excessive credit growth increases the risk of a price bubble when it flows into asset mar - kets. As the share of credit in financing consumption and investment increases, it leads to an unsustainable debt burden and damages economic activity (Gross & Zahner, 2021). Furthermore, the rapid credit growth increases depreciation pressure in the foreign exchange market, whose stability is crucial for EMEs (Aizenman & Binici, 2016; İlhan, Akdeniz, & Özdemir, 2022). Agénor and Pereira da Silva (2019) proposed a framework leaning against the credit cycles for EMEs called integrated inflation targeting (IIT). IIT suggests that the credit growth gap should be included in central banks’ reaction function while mon- etary policy and macroprudential instruments should be calibrated together. One of the EMEs that adopted the lean against the wind strategy ae ft r the GFC is Turkey. The domestic and external demand die ff rentiation associated with accelerating capital ino fl ws created dilemmas for the country’s existing policy framework. Accordingly, in late 2010, a new policy framework, namely the new policy mix, was implemented under the Central Bank of the Republic of Turkey (CBRT) leadership to simultaneously ensure price and n fi ancial sta - bility. In this new policy mix, n fi ancial stability was adopted as a supplementary goal of monetary policy while many macroprudential tools were introduced. Monetary and macroprudential policies were implemented in coordination to complement each other. The intermediate targets were controlling credit growth and slowing short-term capital ino fl ws while the intermediate variables were credit and the exchange rates (Başçı & Kara, 2011; Kara, 2013). i Th s paper analyses the ee ff ct of credit on monetary policy responses with a Markov regime switching (MS) model in Turkey. More specic fi ally, the cred - See Smets (2014) for a detailed review of policy frameworks for financial stability. 70 Economics and Business Review, Vol. 8 (22), No. 4, 2022 it-gap augmented Taylor rule is estimated using the Markov regime switching intercept autoregressive heteroscedasticity (MSIAH) model between January 2006 and December 2019. Unlike previous studies, the ee ff ct of credit on mon - etary policy is examined for die ff rent regimes in Turkey. For example, TVP- VAR models allow quantifying the gradual evolution of the interaction between series throughout the sample period, and parameters can change without the subsamples (Çatik & Akdeniz, 2019). However, the MSIAH model den fi es dif - ferent regimes and captures the ee ff cts of policy stance by providing regime dependent coec ffi ients (Baharumshah, Soon & Wohar, 2017). Thus, the MSIAH model can estimate the impact of credit on monetary policy depending on the CBRT’s policy stance. i Th s paper contributes to the literature as follows. First, it analyses two dif - ferent interest rate regimes with differing coefficients and provides asymmetric monetary policy behaviour in Turkey. Second, the prevalence of the low-inter- est rate regime throughout the sample period ree fl cts the growth priorities of the policy authorities. Finally, in the low interest rate environment, monetary policy focused only on inflation whereas also the credit gap ae ff cted monetary policy decisions in the high-interest rate regime. The signic fi ance of the ee ff ct in the high-interest rate regime indicates that credit conditions contributed to a tightening of monetary policy stance in Turkey. e p Th aper proceeds as follows. Section 1 explains Turkey’s monetary and macroprudential policy responses to credit developments in full-fledged in - flation targeting. Section 2 reviews the empirical literature. Section 3 describes the data and methodology. The following section provides the empirical results and discussion. The final section concludes the paper. 1. Policy responses to credit developments in Turkey e T Th urkish banking sector, which operated as a public debt financier in the 1990s due to high bond yields, returned to its traditional intermediary function ae ft r the 2001 crisis and credit volume grew. The improvement in credit quality due to a decline in the ratio of non-performing loans encouraged the percep- tion that rapid credit growth indicated the normalisation of the banking sector (Kenc, Turhan, & Yildirim, 2011). In May 2006, however, a currency shock in- terrupted credit growth when the policy rate was increased by 425 basis points in two months (CBRT, 2022a) due to a sharp depreciation of the Turkish Lira (TL) and increased inflation expectations (CBRT, 2008). The monetary policy responses to restore price and exchange rate stability slowed credit growth (see Figure 1). Despite the modest recovery ae ft r the first months of 2007, credit growth sue ff red from the GFC. The policy rate was reduced by 650 basis points between December 2008 and April 2009 to limit the GFC’s harmful ee ff cts on economic activity and financial stability (CBRT, 2009). A. İlhan, Examining the ee ff ct of credit on monetary policy with Markov regime switching 71 e q Th uantitative easing and unconventional monetary policies of advanced economies to alleviate the adverse ee ff cts of the GFC accelerated capital ino fl ws to EMEs, thereby creating new macro-financial risks. Exposed to these risks, as are many EMEs, Turkey was forced to switch its policy framework. In the new policy mix, the CBRT announced that monetary policy considers n fi an - cial stability while emphasising the use of non-rate instruments under vari- ous scenarios (CBRT, 2010). Accordingly, the CBRT employed reserve require- ments and unique tools for n fi ancial stability, such as the asymmetric interest rate corridor and the reserve option mechanism. Moreover, in the second half of 2011, the Banking Regulation and Supervision Agency (BRSA) introduced various macroprudential instruments to control credit growth (CBRT, 2014b). Due to these measures and the impact on capital flows of the sovereign debt crisis in the euro area, credit growth declined from the third quarter of 2011 (see Figure 1). 20 70 Lehman currency the taper Brothers second shock talk macroprudential bankruptcy macroprudential loosening 60 tightening policy rate 5 30 hike first currency macroprudential –5 shock credit-to-GDP gap beginning of tightening 0 the new credit growth policy mix –10 –10 Note: The left axis stands for the credit-to-GDP gap, while the right denotes credit growth. All values are in percent for both axes. The credit-to-GDP is the ratio of credit from all sectors to private non-financial sectors to GDP. The credit-to-GDP gap is the die ff rence between the actual value of the credit-to-GDP and its trend. Credit growth is the annual growth of total credit volume in the banking sector’s balance sheets. Figure 1. Credit-to-GDP gap and credit growth in the Turkish economy: 2006Q1–2019Q4 Source: Author’s construction based on CBRT (2022b) and BRSA (2022). From the second half of 2012, an increasing global risk appetite accelerated capital ino fl ws to Turkey while falling interest rates to ease appreciation pres - sure in TL encouraged credit growth (CBRT, 2013). However, once this exceeded a reasonable level (Kara, Küçük, Tiryaki, & Yüksel, 2013), the authorities un- der the leadership of the BRSA further tightened the macroprudential policy stance by introducing new tools and strengthening existing tools (CBRT, 2014b). 2006-Q1 2007-Q1 2008-Q1 2009-Q1 2010-Q1 2011-Q1 2012-Q1 2013-Q1 2014-Q1 2015-Q1 2016-Q1 2017-Q1 2018-Q1 2019-Q1 72 Economics and Business Review, Vol. 8 (22), No. 4, 2022 Besides these macroprudential instruments, a tightening monetary policy stance associated with global financial conditions also helped restrain credit growth. Following the taper talk in May 2013 (Bernanke, 2013), its implemen- tation in December 2013 accelerated capital outo fl ws from EMEs. The CBRT hiked interest rates sharply in January 2014 due to worsening TL depreciation (CBRT, 2015). Ae ft r a  series of shocks in the second half of 2016 deepened the slow - down in credit growth, Turkey’s macroprudential policy stance was eased in September 2016 (CBRT, 2016; İlhan, Özdemir, & Eryigit, 2021). While mon- etary policy tightened in 2017 to limit the depreciation in TL associated with global shocks (CBRT, 2018), credit growth was supported by increased credit guarantee fund incentives (IMF, 2018). Although the monetary policy stance was not loosened in the first half of 2018, the pace of credit growth increased. However, the currency shock in August 2018 and the subsequent rise in in- flation resulted in a sharp interest rate hike and falling credit growth (CBRT, 2019). The policy rate, which had remained constant in the first half of 2019, declined by 1200 basis points until the end of the year, resulting in a recovery in credit growth (CBRT, 2020). 2. Literature review Many studies have estimated the exchange rate-augmented Taylor rule in an- alysing the CBRT’s actions to achieve financial stability (Hasanov & Omay, 2008; Civcir & Akçağlayan, 2010; Caporale, Helmi, Çatık, Ali & Akdeniz, 2018; Yağcıbaşı & Yıldırım, 2019; Soybilgen & Eroğlu, 2019; Özdemir, 2020; Tetik & Yıldırım, 2021). However, few have examined the role of credit in Turkey’s monetary policy. Çamlıca (2016) investigated the CBRT’s responses to financial stress using a composite index of systemic stress, which is an indicator of financial risk in the credit, money, equity, forex and bond markets, for 2005:01 – 2015:10. Compared to the pre-GFC period, the CBRT responded more to n fi ancial stress, and adopt - ed a lean against the wind strategy ae ft r mid-2010. Erdem, Bulut, and Kocak (2017) analysed the exchange rate gap- and credit gap-augmented Taylor rule in a time-varying manner using a cointegration test and the Kalman filter for 2006:01 – 2016:02. Although the nominal domestic credit gap and exchange rate gap ae ff cted the interest rate settings, they did not change the priorities of the CBRT in the new policy mix. Chadwick (2018) explored the impact of monetary and macroprudential policies on consumer credit growth with panel VAR for the period from 2005:12 to 2017:12. A contractionary monetary policy has a restrictive impact on credit growth. Moreover, this impact is greater when combined with macroprudential policy. Kurowski, Rogowicz, and Smaga (2020) examined the Taylor rule ex- A. İlhan, Examining the ee ff ct of credit on monetary policy with Markov regime switching 73 tended with the credit-to-GDP gap using the TVP-VAR model from 2002Q1 to 2018Q3. The interest rate settings were adjusted following the Taylor rule and credit conditions were considered in monetary policy decisions ae ft r 2010 in Turkey. The ee ff ct of interest rate settings on credit and inflation increased with the strengthening of monetary transmission mechanisms ae ft r the GFC. 3. Data and methodology iTh s study used an MS model to investigate the ee ff ct of credit on Turkey’s monetary policy between January 2006 and December 2019. e Th starting date corresponds to the adoption of full-fledged inflation targeting while the end - ing date excludes the impact of COVID-19. Taylor (1993) suggested a simple interest rate rule, known as the Taylor rule, for policy authorities focusing on price and output stability. However, increas- ing financial stability concerns have led others to expand this rule by including n fi ancial variables (Käfer, 2014). Following this literature, an augmented Taylor rule is employed to analyse the impact of credit on monetary policy in Turkey. e cr Th edit-augmented Taylor rule to be estimated is as follows: i = α + βi + δ(π – π *) + θ(y – y *) + ω(cr – cr*) (1) t t–1 t t t t t t Here, i stands for the policy rate. The CBRT used die ff rent short-term inter - est rates as policy rates throughout the sample period. Moreover, from late 2010, the multi-policy rate framework was employed in line with an unconventional approach (Binici, Kara & Özlü, 2016). Alp, Kara, Keleş, Gürkaynak and Orak (2010) showed that the strongest predictor of Turkey’s monetary policy stance is the one-week Turkish Lira Libor (TRLIBOR) rate. Similarly, Gürkaynak, Kantur, Taş and Yıldırım-Karaman (2015) used the one-week TRLIBOR to measure the ee ff ctive policy rate. Therefore, the one-week TRLIBOR rate is used in this study to represent the CBRT’s policy rate. π is actual ina fl tion, obtained from the annual percentage change of the con - sumer price index, while π * is the inflation target. The inflation gap ( π – π *) t t t is calculated as the die ff rence between actual inflation and its target. y is the industrial production index, which proxies for output while cr represents the total credit volume in the banking sector. For both variables, the actual values (y , cr ) are transformed into the logarithmic form before applying the Hodrick- t t -Prescott filter for the trend values ( y *, cr *). The trend values are then sub - t t tracted from their actual values to obtain gaps. The series representing season - The Hodrick-Prescott smoothing parameter was set as λ = 14400, while suggest power for λ was set as 2. 74 Economics and Business Review, Vol. 8 (22), No. 4, 2022 al ee ff cts are adjusted with the Census X-13. To eliminate level die ff rences, all series are standardised. e Th series are taken from various databases. e Th one-week TRLIBOR rate is retrieved from the Banks Association of Turkey. The consumer price index and production index of total industry are obtained from the Federal Reserve Economic Data. The banking sector’s total credit volume is retrieved from the CBRT. Table 1 shows the time series properties and descriptive statistics of the variables. Table 1. Descriptive statistics and unit root test results Descriptive statistics Standard Variables Observations Mean Maximum Minimum deviation i 168 12.534 26.318 5.089 5.199 π 168 9.547 25.240 3.986 3.593 π * 168 5.178 7.500 4.000 0.860 y 168 87.524 119.990 56.995 18.714 cr 168 1,010,904 2,587,738 136,063 755,481 Unit root test results Variables LM DT DT 1t 2t i –6.2557*** 2009:08 2018:03 (π – π *) –7.2967*** 2008:11 2018:06 t t (y – y *) –5.9626** 2008:08 2011:01 t t (cr – cr *) –5.3337* 2009:01 2010:10 t t Note: Descriptive statistics are the level forms of the series. ***, **, and * show stationary at 1%, 5%, and 10% signic fi ance level, respectively. Source: Author’s estimation. Inflation and the interest rate reached their maximum levels in September 2018, right ae ft r the currency shock in August. While the minimum value of the interest rate corresponded to the beginning of the taper tantrum, inflation fell to its lowest value in the first months of 2011, when global liquidity was abundant. Actual inflation remained well above the target levels due to recent jumps in inflation. Output was lowest around the GFC whereas it peaked in the last month of 2017. e t Th ime-series properties of the variables were analysed using Lee and Strazicich’s (2003) unit root test, allowing for two structural breaks. The trend break model indicates that all series are stationary in their levels. Furthermore, A. İlhan, Examining the ee ff ct of credit on monetary policy with Markov regime switching 75 the first break dates of all variables correspond to the GFC while the second break dates for the inflation gap and the interest rate fall in the first and sec - ond quarters of 2018, respectively, before TL’s sharp depreciation. The second break of the credit gap corresponds to the beginning of the new policy mix. iTh s study used an MS model to estimate the augmented Taylor rule. MS models identify die ff rent regimes and allow the dynamic behaviour of the vari - ables to be examined depending on each regime. In these models, past regimes can reoccur throughout the sample while the number of regimes generally var- ies from two to at most four (Baharumshah et al., 2017, pp. 249–250). In MS models, the switch in regime is not den fi ed as the outcome of a de - terministic event but by an unobserved random variable (s ), called the regime or state. Since s takes only discrete values in determining the regime, a Markov chain is used for the regime switching process (Hamilton, 1994, pp. 677–678). The N-state Markov chain with transition probabilities is den fi ed as follows: P{s = j|s = i, s = k, …} = P{s = j|s = i} = p (2) t t–1 t–2 t t–1 ij Here, p is the transition probability, which is the probability that state j is ij preceded by state i, and the sum of the transition probabilities is equal to 1. e t Th ransition probabilities can be collected in the following N ∙ N transition matrix (Hamilton, 1994, pp. 678–679): p p  p   11 21 N1   p p   p 12 22 N 2   P =      (3)         p p   p   1N 2N NN   e m Th aximum likelihood (ML), which is the expectation-maximisation (EM) algorithm-based method, is used for the estimation of the MS model. e v Th alue of the ML function increases with each iteration of the EM algorithm (Hamilton, 1994, p. 689). These processes continue until the parameters con - verge (Çatık & Önder, 2011, p. 128). Many MS models allow for switches in the intercept, mean, and variance of the residuals throughout the regimes governed by an unobserved state vari- able. This study used the MSIAH model as it takes into account the entire pa - rameter shift and changes in the variance of the residuals throughout the state (Çatık & Önder, 2011, p. 127). The model is shown in linear form in equation (1), but it can be re-written for the two-regime MSIAH model: e Th EM algorithm, which was developed by Dempster, Laird and Rubin (1977), was em - ployed by Hamilton (1990) for the ML estimation (Krolzig, 1997). 76 Economics and Business Review, Vol. 8 (22), No. 4, 2022 i = β(s )(i ) + δ(s )(π – π *) + θ(s )(y – y *) + ω(s )(cr – cr *) + ε (4) t t t – 1 t t t t t t t t t t e e Th stimated coeci ffi ents are strongly dependent on the state variable ( s ). e Th integer variable s can take the values 1 or 2 to indicate, respectively, that the low-interest or high-interest rate regime prevails. 4. Empirical results and discussion e e Th mpirical findings indicate the existence of both the low- and high-inter - est rate regimes, called Regime 1 and Regime 2, respectively. That is, the CBRT adopted two die ff rent monetary policy stances. Regime 1 ree fl cts a loose stance whereas Regime 2 ree fl cts a tight stance. Table 2 shows the transition matrix and the properties of the two regimes. Table 2. Transition matrix and regime properties Transition matrix Regime 1 Regime 2 Regime 1 0.881 0.118 Regime 2 0.156 0.843 Regime properties Observation Probability Duration Regime 1 97.7 0.568 8.43 Regime 2 70.3 0.431 6.40 Source: Author’s estimation. As the transition matrix in Table 2 shows, the switching probability from the high- to the low-interest rate regime is higher (15.6%) than that of switching from the low- to the high-interest rate regime (11.8%). Furthermore, the regime properties indicate that the low regime was more likely to be implemented and to last for longer. e Th se n fi dings ree fl ct the policy authorities’ monetary policy preferences, which favoured growth. Figure 2 shows changes in regime prob- abilities over the study period. e r Th egime probabilities in Figures 2(b) and 2(c) clearly show that the low- interest rate regime dominated before a surge in capital flows threatened finan - cial stability. The first switch from a low- to high-interest rate regime ae ft r the GFC took place when the banking sector’s annual credit growth was close to e r Th egime classic fi ations are reported in Table A1 in the Appendix. A. İlhan, Examining the ee ff ct of credit on monetary policy with Markov regime switching 77 a) Interest rate b) Low-interest rate regime c) High-interest rate regime Figure 2. Regime probabilities Source: Author’s estimation. 40% nominal. During this period, macroprudential policy practices started to restrain credit growth (Kara, 2016). In the next high-interest rate regime, the taper tantrum began (Bernanke, 2013), and the CBRT hiked the policy rate, citing TL depreciation (CBRT, 2014a). In addition, macroprudential policies were further tightened in response to accelerating credit growth (CBRT, 2014b). Excluding short-lived switches, the subsequent prevalence of the high-inter- est rate regime corresponds to the jumps in the nominal exchange rate in the spring of 2018. While interest rate declined from the second half of 2019, the sample ends with a high-interest rate regime. Prior to estimating the model, its nonlinearity was evaluated with diagnostic tests. As seen in Table 3, the likelihood (LR) linearity test rejects the null hy- 78 Economics and Business Review, Vol. 8 (22), No. 4, 2022 pothesis of linearity and supports the nonlinear model. e l Th og-likelihood and Akaike information criterion (AIC) values of the linear and nonlinear models also verify the two-regime model (Çatık & Önder, 2011; Kumah, 2011). Table 3 presents the empirical results for the estimation. Table 3. Estimation results Regime 1 (Standard error: 0.051) Variables Coefficient Standard error t-value α –0.006 0.007 –0.927 i 0.965*** 0.009 99.072 t–1 (π – π *) 0.031*** 0.010 2.981 t t (y – y *) 0.013 0.009 1.504 t t (cr – cr *) 0.012 0.008 1.461 t t Regime 2 (Standard error: 0.297) Variables Coefficient Standard error t-value α –0.067 0.043 –1.551 i 0.759*** 0.068 11.096 t–1 (π – π *) 0.209*** 0.093 3.188 t t (y – y *) –0.003 0.044 –0.077 t t (cr – cr *) 0.156*** 0.050 3.098 t t LR Linearity Test: 160.059 Chi (6) = (0.000)*** Chi (8) = (0.000)*** Log-likelihood AIC Nonlinear model 98.540 –1.006 Linear model 18.510 –0.148 Note: ***, **, and * show signic fi ance at 1%, 5%, and 10%, respectively. Source: Author’s estimation. As Table 3 shows, the coeci ffi ents die ff r in sign, size, and signic fi ance be - tween the regimes, which supports the nonlinear interest rate setting and the two-regime model. Furthermore, the differentiation of responses is consistent with previous studies reporting an asymmetric monetary policy behaviour for Turkey (Hasanov & Omay, 2008; Caporale et al., 2018; Öge Güney, 2018; Bulut, 2019; Özdemir, 2020). In the low-interest rate regime, inflation positively and signic fi antly impacts the policy rate, whereas the output and credit gap ee ff cts are insignic fi ant. That is, monetary policy focused on the traditional goal, namely price stability, in a low interest rate environment. The coeci ffi ents of the high-interest rate re - A. İlhan, Examining the ee ff ct of credit on monetary policy with Markov regime switching 79 gime indicate that the monetary policy took an additional responsibility. The ee ff ct of the inflation gap is signic fi ant and positive while the coeci ffi ent is larg - er. However, as in the low-interest rate regime, the output gap is insignic fi ant. On the contrary, the credit gap signic fi antly and positively impacts interest rate setting. This indicates that the rising credit gap had an impact on the tighten - ing of monetary policy in Turkey. e Th se empirical findings conr fi m those of Erdem and others (2017), Chadwick (2018) and Kurowski and others (2020), who found that credit conditions af- fected monetary policy decisions in Turkey. Furthermore, the findings consist - ent with those of Çamlıca (2016), who reported that the CBRT adopted a lean against the wind strategy ae ft r mid-2010, to some extent. A robustness check is performed by augmenting the Taylor rule with an- other credit variable. Since many macroprudential tools, especially those im- plemented by the BRSA, mainly aim to control consumer credit growth (CBRT, 2014b), it is worth examining whether consumer credit ae ff cted monetary policy. To this end, the Taylor rule augmented with the consumer credit gap (ccr – ccr *) is estimated. t t e Th empirical findings indicate that the regime probabilities and properties are largely similar to those of the credit-augmented Taylor rule estimation (see Table A3 in the Appendix). In the low-interest rate regime, monetary policy aligns with the standard Taylor rule. Both inflation and the output gap posi - tively and signic fi antly impact the policy rate, whereas the consumer credit gap is insignic fi ant. In contrast, in the high-interest rate regime, inflation and the consumer credit gap signic fi antly and positively ae ff ct the policy rate, where - as the output gap has no signic fi ant ee ff ct. Thus, consumer credits ae ff cted the policy rate despite the availability of many macroprudential tools. In conclu- sion, the robustness check n fi dings verify that credit developments impacted the tightening of the monetary policy stance in Turkey between January 2006 and December 2019. Conclusions i Th s study investigates the ee ff ct of the credit gap on monetary policy responses in Turkey employing the MSIAH model between January 2006 and December 2019. The empirical findings show that in the low interest rate environment, When the credit variable is removed from the equation and the standard Taylor rule is estimated, results are similar to credit-augmented Taylor rule findings. In the high-interest rate regime, only inflation has a signic fi ant and positive impact on interest rate setting, while the out - put gap is insignic fi ant (see Table A2 in the Appendix). Consumer credit is retrieved from the BRSA and similar processes are performed for the consumer credit gap (see Section 3). 80 Economics and Business Review, Vol. 8 (22), No. 4, 2022 monetary policy focuses only on inflation, whereas the credit gap also inu fl - ences monetary policy decisions in the high-interest rate regime. This indi - cates that credit conditions contributed to a tightening of the monetary policy stance in Turkey. e r Th ole of monetary policy on financial stability substantially depends on the performance of macroprudential policy. When macroprudential policy falls short of ensuring financial stability, monetary policy might support macropru - dential policy (Gerlach, 2012). Many studies have reported that macropruden- tial policies have a limiting ee ff ct on credit growth in Turkey (Binici, Erol, Kara, Özlü & Ünalmış, 2013; Bulut, 2015; Bumin & Taşkın, 2016; Yüceyılmaz, Altın & Tunay, 2017; Alper, Binici, Demiralp, Kara & Özlü, 2018; Chadwick, 2018; İlhan et al., 2021). However, empirical findings show that macroprudential policy was not the only way to control credit growth. Similar to what Kurowski and others (2020) pointed out, the determinants of the monetary policy stance in the high-interest rate regime indicate that the policy framework was partially consistent with the IIT strategy. e e Th e ff ct of credit developments on interest rate settings also indicates that monetary policy was complementary to macroprudential policy in Turkey. However, this led to adverse side ee ff cts on price stability. In the new policy mix, monetary policy, which was also concerned with financial stability, devi - ated from its primary goal (Gürkaynak et al., 2015). On the other hand, macro- prudential policy stance loosened prematurely with the increasing impact of growth priorities. Furthermore, macroprudential policy has not been compre- hensive enough, and measures have not directly covered controlling commer- cial credit growth (Kara, 2016; IMF, 2018). Therefore, implementing a  more ee ff ctive macroprudential policy would ease the burden on monetary policy and increase its room for manoeuvre. Price stability-focused and clearer mon- etary policy framework would reduce uncertainty and help achieve the ulti- mate goal of the CBRT. A. İlhan, Examining the ee ff ct of credit on monetary policy with Markov regime switching 81 Appendix Table A1. Regime classification Regime 1 Regime 2 2006:01 – 2006:05 2006:06 – 2006:07 2006:08 – 2008:05 2008:06 – 2008:06 2008:07 – 2008:11 2008:12 – 2009:03 2009:04 – 2011:07 2011:08 – 2012:08 2012:09 – 2013:02 2013:03 – 2015:01 2015:02 – 2015:08 2015:09 – 2016:01 2016:02 – 2016:12 2017:01 – 2017:03 2017:04 – 2018:03 2018:04 – 2018:11 2018:12 – 2019:02 2019:03 – 2019:12 Source: Author’s estimation. Table A2. Estimation results of the standard Taylor rule Regime 1 (n: 99.2, Prob.: 0.577, Duration: 8.61, Standard error: 0.052) Variables Coefficient Standard error t-value α –0.011 0.007 –1.573 i 0.965*** 0.010 95.682 t–1 (π – π *) 0.033*** 0.010 3.258 t t (y – y *) 0.023*** 0.006 3.620 t t Regime 2 (n: 68.8, Prob.: 0.422, Duration: 6.29, Standard error: 0.323) Variables Coefficient Standard error t-value α –0.032 0.046 –0.696 i 0.738*** 0.077 9.564 t–1 (π – π *) 0.252*** 0.072 3.477 t t (y – y *) 0.038 0.050 0.766 t t LR Linearity Test: 156.871 Chi (5) = (0.000)*** Chi (7) = (0.000)*** Log-likelihood AIC Nonlinear model 92.380 –0.956 Linear model 13.944 –0.106 Note: ***, **, and * show signic fi ance at 1%, 5%, and 10%, respectively. Source: Author’s estimation. 82 Economics and Business Review, Vol. 8 (22), No. 4, 2022 Table A3. Estimation results of the consumer credit-augmented Taylor rule Regime 1 (n: 98.8, Prob.: 0.574, Duration: 8.67, Standard error: 0.052) Variables Coefficient Standard error t-value α –0.008 0.007 –1.229 i 0.965*** 0.009 98.427 t–1 (π – π *) 0.032*** 0.010 3.135 t t (y – y *) 0.017* 0.009 1.926 t t (ccr – ccr *) 0.007 0.008 0.970 t t Regime 2 (n: 69.2, Prob.: 0.425, Duration: 6.41 Standard error: 0.307) Variables Coefficient Standard error t-value α –0.074 0.046 –1.596 i 0.741*** 0.070 10.513 t–1 (π – π *) 0.265*** 0.068 3.708 t t (y – y *) –0.022 0.051 –0.436 t t (ccr – ccr *) 0.153** 0.068 2.557 t t LR Linearity Test: 161.223 Chi (6) = (0.000)*** Chi (8) = (0.000)*** Log–likelihood AIC Nonlinear model 96.199 –0.978 Linear model 15.587 –0.114 Note: ***, **, and * show signic fi ance at 1%, 5%, and 10%, respectively. Source: Author’s estimation. References Adrian, T. (2020). “Low for long” and risk-taking. (IMF Departmental Paper No. 15). Retrieved from https://www.elibrary.imf.org/view/journals/087/2020/015/arti- cle-A001-en.xml Agénor, P. R., & Pereira da Silva, L. A. (2019). Integrated inflation targeting: Another per - spective from developing world. Retrieved from https://www.bis.org/publ/othp30.htm Aizenman, J., & Binici, M. (2016). Exchange market pressure in OECD and emerging economies: Domestic vs. external factors and capital flows in the old and new nor - mal. Journal of International Money and Finance, 66, 65–87. http://doi.org/10.1016/j. jimonn.2015.12.008 fi Alp, H., Kara, H., Keleş, G., Gürkaynak, R., & Orak, M. (2010). Türkiye’de piyasa gösterge- lerinden para politikası beklentilerinin ölçülmesi. İktisat İşletme ve Finans, 25(295), 21–45. https://doi.org/10.3848/iif.2010.295.2849 A. İlhan, Examining the ee ff ct of credit on monetary policy with Markov regime switching 83 Alper, K., Binici, M., Demiralp, S., Kara, H., & Özlü, P. (2018). Reserve requirements, liquidity risk, and bank lending behavior. Journal of Money, Credit and Banking, 50(4), 817–827. https://doi.org/10.1111/jmcb.12475 Angeloni, I., & Faia, E. (2013). Capital regulation and monetary policy with fragile banks. Journal of Monetary Economics, 60, 311–324. http://dx.doi.org/10.1016/j. jmoneco.2013.01.003 Baharumshah, A. Z., Soon, S. V., & Wohar, M. E. (2017). Markov-switching analysis of exchange rate pass-through: Perspective from Asian countries. International Review of Economics and Finance, 51, 245–257. http://doi.org/10.1016/j.iref.2017.05.009 Başçı, E., & Kara, H. (2011). Financial stability and monetary policy. (CBT Working Papers No. 08). Retrieved from https://www.tcmb.gov.tr/wps/wcm/connect/EN/ TCMB+EN/Main+Menu/Publications/Research/Working+Paperss/2011/11-08 Bernanke, B. S. (2013). e Th economic outlook . Retrieved from https://www.federalre- serve.gov/newsevents/testimony/bernanke20130522a.htm Binici, M., Erol, H., Kara, H., Özlü, P., & Ünalmış, D. (2013). Interest rate corridor: A new macroprudential tool?. CBT Research Notes in Economics, 20. Retrieved from https:// www.tcmb.gov.tr/wps/wcm/connect/EN/TCMB+EN/Main+Menu/Publications/ Research/Research+Notes+in+Economicss/ Binici, M., Kara, H., & Özlü, P. (2016). Unconventional interest rate corridor and the monetary transmission evidence from Turkey. (CBRT Working Paper No. 08). Retrieved from https://www.tcmb.gov.tr/wps/wcm/connect/EN/TCMB+EN/ Main+Menu/Publications/Research/Working+Paperss/ Borio, C. (2014). Monetary policy and financial stability: What role in prevention and recovery. Capitalism and Society, 9(2), 1. Borio, C., English, W., & Filardo, A. (2003). A tale of two perspectives old or new chal- lenges for monetary policy?. (BIS Working Papers No. 127). Borio, C., & Lowe, P. (2002). Asset prices, financial and monetary stability: Exploring the nexus. (BIS Working Papers No. 114). BRSA. (2022). Banking sector data—quarterly data. Retrieved March 15, 2022 from https://www.bddk.org.tr/BultenAylik/en/Home/Gelismis Bulut, U. (2015). The interest rate corridor as a macroprudential tool to mitigate rapid credit growth in credits: Evidence from Turkey. e Th oretical and Applied Economics , 4(605), 133–144. Bulut, U. (2019). Does the Central Bank of the Republic of Turkey respond asymmetri- cally to inflation and output?. Margin—e Th Journal of Applied Economic Research , 13(4), 381–400. https://doi.org/10.1177/0973801019868374 Bumin M., & Taşkın, F. D. (2016). e i Th mpact of macro-prudential measures on retail loans: Evidence from Turkey. International Journal of Contemporary Economics and Administrative Sciences, 5(1–2), 48–46. Çamlıca, F. (2016). Responsiveness of monetary policy to financial stress in Turkey. Central Bank Review, 16, 143–150. http://doi.org/10.1016/j.cbrev.2016.11.003 Caporale, G. M., Helmi, M. H., Çatık, A. N., Ali, F. M. & Akdeniz, C. (2018). Monetary policy rules in emerging countries: Is there an augmented nonlinear Taylor rule?. Economics Modelling, 72, 306–319. https://doi.org/10.1016/j.econmod.2018.02.006 84 Economics and Business Review, Vol. 8 (22), No. 4, 2022 Çatik, A. N., & Akdeniz, C. (2019). e e Th volution of the monetary transmission chan - nels in Turkey: Evidence from a TVP-VAR model. Applied Economics Letters, 26(13), 1072–1079. https://doi.org/10.1080/13504851.2018.1537468 Çatık, A. N., & Önder, A. Ö. (2011). Inflationary ee ff cts of oil prices in Turkey: A re - gime-switching approach. Emerging Markets Finance and Trade, 47(5), 125–140. https://doi.org/10.2753/REE1540-496X470506 CBRT. (2008). Monetary and exchange rate policy for 2009. Retrieved from https://www. tcmb.gov.tr/wps/wcm/connect/EN/TCMB+EN/Main+Menu/Publications/Moneta ry+and+Exchange+Rate+Policy+Texts/ CBRT. (2009). Monetary and exchange rate policy for 2010. Retrieved from https://www. tcmb.gov.tr/wps/wcm/connect/EN/TCMB+EN/Main+Menu/Publications/Moneta ry+and+Exchange+Rate+Policy+Texts/ CBRT. (2010). Financial stability report No. 11. Retrieved from https://www.tcmb. gov.tr/wps/wcm/connect/EN/TCMB+EN/Main+Menu/Publications/Reports/ Financial+Stability+Report/ CBRT. (2013). Annual report 2012. Retrieved from https://www.tcmb.gov.tr/wps/wcm/ connect/EN/TCMB+EN/Main+Menu/Publications/Reports/Annual+Reports/ CBRT. (2014a). Press release on interest rates, No. 06. Retrieved from https://www. tcmb.gov.tr/wps/wcm/connect/EN/TCMB+EN/Main+Menu/Announcements/ Press+Releases CBRT. (2014b). Financial Stability Report No. 19. Retrieved from https://www.tcmb. gov.tr/wps/wcm/connect/EN/TCMB+EN/Main+Menu/Publications/Reports/ Financial+Stability+Report/ CBRT. (2015). Annual report 2014. Retrieved from https://www.tcmb.gov.tr/wps/wcm/ connect/EN/TCMB+EN/Main+Menu/Publications/Reports/Annual+Reports/ CBRT. (2016). Financial stability report No. 23. Retrieved from https://www.tcmb. gov.tr/wps/wcm/connect/EN/TCMB+EN/Main+Menu/Publications/Reports/ Financial+Stability+Report/ CBRT. (2018). Annual report 2017. Retrieved from https://www.tcmb.gov.tr/wps/wcm/ connect/EN/TCMB+EN/Main+Menu/Publications/Reports/Annual+Reports/ CBRT. (2019). Annual report 2018. Retrieved from https://www.tcmb.gov.tr/wps/wcm/ connect/EN/TCMB+EN/Main+Menu/Publications/Reports/Annual+Reports/ CBRT. (2020). Annual report 2019. Retrieved from https://www.tcmb.gov.tr/wps/wcm/ connect/EN/TCMB+EN/Main+Menu/Publications/Reports/Annual+Reports/ CBRT. (2022a). Central Bank interest rates. Retrieved from https://www.tcmb.gov.tr/wps/ wcm/connect/EN/TCMB+EN/Main+Menu/Core+Functions/Monetary+Policy/ Central+Bank+Interest+Rates/CBRT+INTEREST+RATES CBRT. (2022b). Banking sector total credit volume—quarterly data. Retrieved March 15, 2022 from https://evds2.tcmb.gov.tr/index.php?/evds/serieMarket Chadwick, M. G. (2018). Ee ff ctiveness of monetary and macroprudential shocks on consumer credit growth and volatility in Turkey. Central Bank Review, 18(2), 69–83. https://doi.org/10.1016/j.cbrev.2018.03.001 Civcir, I., & Akçağlayan, A. (2010). Inflation targeting and the exchange rate: Does it matter in Turkey?. Journal of Policy Modeling, 32, 339–354. https://doi.org/10.1016/j. jpolmod.2010.03.002 A. İlhan, Examining the ee ff ct of credit on monetary policy with Markov regime switching 85 Dempster, A. P., Laird, N. M., & Rubin, D. B. (1977). Maximum likelihood from in- complete data via the EM algorithm. Journal of the Royal Statistical Society. Series B (Methodological), 39(1), 1–38. Ekholm, K. (2013, March 13). Monetary policy, business cycle stabilisation and macro- prudential policy. Speech at the Confederation of Swedish Enterprise, Stockholm. Retrieved from https://www.bis.org/review/r130319d.pdf Erdem, E., Bulut, U., & Kocak, E. (2017). Have financial stability concerns changed the priority of the Central Bank of the Republic of Turkey?. Studies in Business and Economics, 12(2), 3–45. https://doi.org/10.1515/sbe-2017-0019 FSB, IMF, & BIS. (2011a, February). Macroprudential policy tools and frameworks— Update to G20 Finance Ministers and Central Bank Governors. Retrieved from https:// www.imf.org/external/np/g20/pdf/021411.pdf FSB, IMF, & BIS (2011b, October), Macroprudential policy tools and frameworks— Progress report to G20. Retrieved from https://www.imf.org/external/np/g20/ pdf/102711.pdf Gerlach, S. (2012). Monetary policy ae ft r the crisis. (Speech at the 44th Annual Money, Macro and Finance Conference, Trinity College, Dublin, 8 September). Retrieved from https://www.bis.org/review/r120910c.pdf Gross, J., & Zahner, J. (2021). What is on the ECB’s mind? Monetary policy before and ae ft r the global financial crisis. Journal of Macroeconomics, 68. https://doi. org/10.1016/j.jmacro.2021.103292 Gürkaynak, R. S., Kantur, Z., Taş, M. A., & Yıldırım-Karaman, S. (2015). Monetary pol- icy in Turkey aer ft Central Bank independence. İktisat İşletme ve Finans, 30(356), 9–38. https://doi.org/10.3848/iif.2015.356.4592 Hamilton, J. D. (1990). Analysis of time series subject to changes in regime. Journal of Econometrics, 45(1–2), 39–70. https://doi.org/10.1016/0304-4076(90)90093-9 Hamilton, J. D. (1994). Time series analysis. Princeton: Princeton University Press. Hasanov, M., & Omay, T. (2008). Monetary policy rules in practice: Re-examining the case of Turkey. Physica A, 387, 4309-4318. https://doi.org/10.1016/j.physa.2008.02.075 İlhan, A., Akdeniz, C., & Özdemir, M. (2022). Analyzing exchange market pres- sure dynamics with Markov regime switching: The case of Turkey. Organizations and Markets in Emerging Economies, 13(1), 238–259. https://doi.org/10.15388/ omee.2022.13.78 İlhan, A., Özdemir, M., & Eryigit, K. Y. (2021). The role of macroprudential policies on controlling credit growth: The case of Turkey. Journal of Yasar University, 16(61), 393–413. https://doi.org/10.19168/jyasar.821930 IMF. (2018). 2018 Article IV Consultation with Turkey. IMF Country Report, No. 18/110. Retrieved from https://www.imf.org/en/Publications/CR/Issues/2018/04/30/ Turkey-2018-Article-IV-Consultation-Press-Release-Staff-Report-and-Statement- by-the-45822 Jordà, Ò., Schularick, M., & Taylor, A. M. (2011). Financial crises, credit booms, and ex- ternal imbalances: 140 years of lessons. IMF Economic Review, 59, 340–378. https:// doi.org/10.1057/imfer.2011.8 Käfer, B. (2014). The Taylor rule and financial stability—A literature review with applica - tion for the eurozone. Review of Economics, 65(2), 159–192. https://doi.org/10.1515/ roe-2014-0204 86 Economics and Business Review, Vol. 8 (22), No. 4, 2022 Kara, A. H. (2013). Monetary policy ae ft r the global crisis. Atlantic Economic Journal, 41, 51–74. https://doi.org/10.1007/s11293-012-9358-7 Kara, A. H. (2016). Turkey’s experience with macroprudential policy. BIS Papers, 86, 123–140. Kara, H., Küçük, H., Tiryaki, S.T. & Yüksel, C. (2013). In search of a reasonable cred- it growth rate for Turkey. Research Notes in Economics, 3. Retrieved from https:// www.tcmb.gov.tr/wps/wcm/connect/EN/TCMB+EN/Main+Menu/Publications/ Research/Research+Notes+in+Economicss/ Kenc, T., Turhan, M. I., & Yildirim, O. (2011). e e Th xperience with macro-prudential poli - cies of the Central Bank of the Republic of Turkey in response to the global financial crisis. (World Bank Policy Research Working Paper No. 5834). Krolzig, H. M. (1997). Markov-switching vector autoregressions: Modelling, statistical in- ference, and application to business cycle analysis. Lecture Notes in Economic and Mathematical Systems, 454. Berlin: Springer. Kumah, F. Y. (2011). A Markov-switching approach to measuring exchange market pressure. International Journal of Finance and Economics, 16, 114–130. https://doi. org/10.1002/ijfe.415 Kurowski, L., Rogowicz, K., & Smaga, P. (2020). What would the economy with credit- enhanced interest rates look like? The case of Turkey. Turkish Studies, 21(4), 620– 644. https://doi.org/10.1080/14683849.2019.1659735 Laeven, L. (2016). Monetary policy in the clutches of financial stability?. In J. Ulbrich, C. C. Hedrich & M. Balling (Eds.), e SS Th M at 1, SUERF Conference Proceedings (pp. 17–32). Vienna: SUERF. Retrieved from http://hdl.handle.net/10419/163519 Lee, J., & Strazicich, M. C. (2003). Minimum Lagrange Multiplier Unit Root Test with two structural breaks. e Th Review of Economics and Statistics , 85(4), 1082–1089. https://doi.org/10.1162/003465303772815961 Öge Güney, P. (2018). Asymmetries in monetary policy reaction function and the role of uncertainties: the case of Turkey. Economic Research—Ekonomska Istraživanja, 31(1), 1367–1381. https://doi.org/10.1080/1331677X.2018.1481445 Özdemir, M. (2020). The role of exchange rate in inflation targeting: The case of Turkey. Applied Economics, 52(29), 3138–3152. https://doi.org/10.1080/00036846.2019.17 Ozkan, F. G., & Unsal, D. F. (2014). On the use of monetary and macroprudential policies for small open economies. (IMF Working Paper No. 112). https://doi. org/10.5089/9781498375429.001 Rungcharoenkitkul, P., Borio, C., & Disyatat, P. (2019). Monetary policy and hysteresis and the financial cycle . (BIS Working Papers No. 817). Smets, F. (2014). Financial stability and monetary policy: How closely ınterlinked?. International Journal of Central Banking, 10(2), 263–300. Soybilgen, B., & Eroğlu, B. A. (2019). Time-varying Taylor rule estimation for Turkey with e fl xible least square method. Boğaziçi Journal Review of Social, Economic and Administrative Studies, 33(2), 122–139. https://doi.org/10.21773/boun.33.2.3 Svensson, L. E. O. (2012). Comment on Michael Woodford, “Inflation targeting and fi - nancial stability” . Sveriges Riksbank Economic Review, 1, 33‒39. Retrieved from http:// archive.riksbank.se/en/Web-archive/Published/Published-from-the-Riksbank/ Economic-Review/Economic-Review/2012/index.html A. İlhan, Examining the ee ff ct of credit on monetary policy with Markov regime switching 87 Svensson, L. E. O. (2017). Cost-benet a fi nalysis of leaning against the wind. Journal of Monetary Economics, 90, 193–213. http://dx.doi.org/10.1016/j.jmoneco.2017.07.004 Taylor, J. B. (1993). Discretion versus policy rules in practice. Carnegie-Rochester Conference Series on Public Policy, 39, 195–214. https://doi.org/10.1016/0167- 2231(93)90009-L Tetik, M., & Yıldırım, M. O. (2021). Distortionary ee ff cts of economic crises on policy coordination in Turkey: Threshold GMM approach. Economics and Business Review, 7(3), 83–102. https://doi.org/10.18559/ebr.2021.3.6 White, W. (2006). Is price stability enough?. (BIS Working Papers No. 205). Woodford, M. (2012). Inflation targeting and financial stability. Sveriges Riksbank Economic Review, 1, 7–32. Retrieved from http://archive.riksbank.se/en/Web- archive/Published/Published-from-the-Riksbank/Economic-Review/Economic- Review/2012/index.html Yağcıbaşı, Ö. F., & Yıldırım, M. O. (2019). Estimating Taylor rules with Markov switch- ing regimes for Turkey. Romanian Journal of Economic Forecasting, 22(3), 81–95. Yüceyılmaz, H. F., Altın, O., & Tunay, K. B. (2017). Makro İhtiyati Tedbirler Çerçevesinde Tüketici Kredilerine İlişkin Düzenlemelerin Etkinliği: Türkiye Örneği. Ege Akademik Bakış, 17(4), 449–459. https://dergipark.org.tr/tr/pub/eab/issue/39982/475204

Journal

Economics and Business Review – de Gruyter

Published: Dec 1, 2022

Keywords: credit; financial stability; monetary policy; macroprudential policy; Markov regime switching; Turkey; C24; E44; E52; E58

There are no references for this article.