doi: 10.15659/3.sektor-sosyal-ekonomi.19.07.1146
Önerilen Atıf /Suggested Citation
Ordu Akkaya, B.M. 2019, US Economic Policy Uncertainty And Loan Growth: Evidence From Turkey, Üçüncü Sektör Sosyal Ekonomi Dergisi, 54(3), 1049-1063
Research Article
US Economic Policy Uncertainty And Loan Growth: Evidence From Turkey
ABD Ekonomik Politika Belirsizliği ve Kredi Büyümesi: Türkiye’den Kanıt
Beyza Mina ORDU AKKAYA
Ankara Sosyal Bilimler Üniversitesi İşletme Bölümü.
beyza.akkaya@asbu.edu.tr https://orcid.org/0000-0003-4353-3977
Makale Gönderme Tarihi
15.05.2019 Revizyon Tarihi 9.07.2019 Kabul Tarihi 11.07.2019 Abstract
In this paper we aim to examine the transmission channel of US economic policy uncertainty on Turkish economy. We proxy economic uncertainty via the recent measure calculated by Baker et al. (2016), which calculates uncertainty through counting selected word frequencies in a number of newspapers. Previous literature show that economic uncertainty has significant and negative impact on economic growth through investments and consumption. However, there are very few studies which check the transmission channel of uncertainty across economies. Hence, we investigate whether economic policy uncertainty (EPU) spills over to bank loan growth in Turkey for the periods between 1985 and 2018 using Hafner and Herwartz (2006) methodology. Using monthly data, we find that, indeed EPU has an impact on bank loans and therefore economic uncertainty mainly diffuses across countries through financial institutions.
Keywords: Economic Policy Uncertainty, Bank Loan, Volatility Spillover, Financial Institutions. Jel Classification: G10, G15, G21
Öz
Bu araştırmanın amacı Amerika Birleşik Devletleri’ndeki ekonomik belirsizliğin Türkiye ekonomisine hangi vasıtalar ile yayıldığını incelemektir. Ekonomik politik belirsizliği, Baker vd. (2016) tarafından geliştirilmş ve seçilen bazı kelimelerin önde gelen gazetelerde ne sıklıkla geçtiği üzerinden hesaplanan bir endekstir. Literatür, ekonomik belirsizliklerin, ertelenen yatırım ve harcama kararları nedeniyle ekonomi büyümesi üzerinde negatif etkileri olduğunu göstermiştir. Fakat belirsizliğin ekonomiler arasında nasil geçişkenlik gösterdiği nispeten daha az çalışılmıştır. Bu nedenle, ekonomik politika belirsizliğinin 1985 ve 2018 arasında Türkiye’deki banka kredileri büyümesi üzerindeki oynaklık yayılımı Hafner ve Herwartz (2006) yöntemi kullanılarak incelenmektedir. Aylık veri kullanarak bulunan sonuçlar ekonomik politika belirsizliğinin banka kredileri üzerinde etkisi olduğunu göstermektedir. Yani,ekonomik belirsizlik diğer ülkelere esas olarak finansal kurumlar aracılığıyla yayılmaktadır.
1050
Anahtar Kelimeler: Ekonomik Politika Belirsizliği, Banka Kredisi, Oynaklık Yayılımı, Finansal Kurumlar.
Jel Sınıflandırması: G10, G15, G21 INTRODUCTION
Economic policy uncertainty (EPU hereafter) has always been a critical phenomenon for all market participants; since their decisions and outcomes of those decisions are highly dependent upon the monetary policy. However, the importance got even larger after the wake of financial crisis in 2008. US has experienced the second biggest crisis of its history and recovery process was highly supported by monetary policy decisions. However, the big uncertainty again after the crisis periods led to slower recovery process (Antonakakis et al., 2013).
The importance of uncertainty lies in its impact over the economy through the success of policies and actions taken by government officials. Although some actions would have certain implications, uncertainty can mask these effects and officials might not end up with what they had been aiming for. Moreover, during uncertainty, investors generally delay their investment decisions since investment decisions are generally irreversible (Bernanke, 1983). Some other scholars find, uncertainty also has negative impact on economic growth and inflation (Bloom et al., 2018). Furthermore, the spillover is found to be increasing even in higher across emerging markets during the quantitative easing period of United States of America (US) (Glick and Leduc, 2012).
Since there is a wide literature which measures uncertainty, Baker et al. (2016) computes economic policy uncertainty commencing from 1985 for US. This index is calculated via the selected word frequencies scanned in a number of newspapers. The robustness of the index is also double-checked with human readings and plus the trend of the index conforms well to the economic conditions in the last three decades. Therefore this index is one of the mostly cited and employed index in the current literature. The relationship of this index with the economic growth and business cycles are studied for US (Bloom et al., 2018). However, the transmission mechanism of economic policy uncertainty to emerging markets has not been studied extensively. Given the importance of financial institutions on integrating economic policies, we aim to investigate whether US policy uncertainty transmit to Turkish market through banks and the availability of credit to households and corporations. Although Bordo et al. (2016) find confirmatory results for US, as of our knowledge other economies has not been studied, yet. Our major findings propose that the volatility in EPU leads to an increase in volatility for the bank loan growth in Turkey. This underlines the importance of financial institutions on the transmission of economic conditions across boundaries. Given that the seminal paper by Kaufman (1994) argue that the contagion occurs faster and more severely amongst financial institutions, financial markets are the first to be get affected from uncertainty. Next, since banks are the major lending institutions, the criticality in lending availability leads to shaky business environment.
In the next section, we summarize the related literature. Next, we explain the methodology and data, respectively. Lastly, we discuss results of the paper and conclude.
LITERATURE REVIEW:
Our paper is mainly related with literature which investigates the impact of uncertainty on the economy. Since growth of economy is largely shaped through investment decisions, researchers mainly focus on how uncertainty affects investment. For instance, Bernanke (1983) indicates that firms postpone their investment and employment decisions during times when uncertainty is high. Especially the irreversibility of investment decisions lead to hesitancy in management team. Bloom (2014) shows that uncertainty increases significantly during recessions and state the underlying theory for this reasoning is as follows: During good times, firms buy and sell actively which increases the available information. However, when things are not going that well, firms feel doubtful about the future and hence prefers to wait and see. This decreases the information
1051
availability and uncertainty increases significantly. Hence, there is an intertwining nature between uncertainty and economic conditions and literature suggests that uncertainty leads to a decrease in output and trade (Bloom, 2014). During the financial crisis in 2008 many market participants and policy makers indicated that uncertainty further exemplifies the shrinkage in US economy (Federal Reserve Open Market Committee, 2008). On the empirical side, papers by Stein and Stone (2012) and Bloom et al. (2012) find that uncertainty specifically is responsible for one third of the 9 percent drop in US GDP during 2008 crisis.
Pastor and Veronesi (2012) particularly check the government policy uncertainty, which are including but not limited to tax rates, completion regulation and subsidies. Authors find that US stock returns on average fall after a policy change, since any policy change that increases stock prices are already incorporated in prices. In a similar vein, Yonce (2015) show that firms intentionally decrease their investment before major elections. Therefore, the political uncertainty might have significant effects especially on the discretionary spending of households and investment decisions of firms (Gilchrist et al., 2010).
Moreover, policy uncertainty can have impact on welfare of an economy, as well as inflation and capital flows (Pastor and Veronesi, 2012). For instance, Hermes and Lensink (2001) indicate that uncertainty on issues which are critical for foreign direct investors lead to decrease in capital inflow to the country. Some of these issues are tax payments, government controls and inflation. From another end, uncertainty on social security policies lead to decrease in consumer spending and adjustment in portfolio selection (Gomes et al., 2012).
This strand, next narrows down to the impact of uncertainty on stock returns. The term uncertainty could, of course change on the definition. For instance, Connolly et al. (2005) proxy uncertainty though implied volatility from equity index options and show stock returns are lower when uncertainty is high. Zhang (2006) approach the question from information uncertainty and find uncertainty leads to higher expected returns following good news vice versa following bad news. Chau et al. (2014) investigate political uncertainty during Arab Spring in Middle East and North Africa region (MENA) and show that market volatility significantly increased following the turmoil.
So, the term uncertainty is dependent upon the definition, though economy and financial markets somehow respond to this uncertainty. Especially the 2008 crisis pushed researchers to find the hints of crisis, retrospectively. Because, if the crisis could have been caught about a year before – such global economy would not had experienced such a vast downturn. Stemming from this, Baker et al. (2016) calculate the economic policy uncertainty index for US based on newspaper coverages. Authors also double checked the accuracy of the index with some major turmoil periods in history, as well as human readings of over 12 thousand articles. They measured the aggregated economic policy uncertainty index (EPU) back to 1900s and categorically divided into a few such as health care policy uncertainty, sovereign debt uncertainty etc. Since the data is available in a very long history, researchers benefited vastly from EPU and recently started to employ in their studies.
One of the first studies employing EPU is by Colombo (2013) and study spillover from US to Euro area via the EPU. He shows that Euro area economic indicators are negatively affected from an unexpected increase in US EPU. Similarly Bernal et al. (2016) study the spillover and cover 10 Eurozone countries – Austria, Belgium, Finland, France, Germany, Greece, Italy, The Netherlands, Portugal and Spain Results show that especially the EPU spills over firstly to bond markets of France and Germany and then transmits to other individual countries. Another study is by Antonakakis et al. (2013), who study the dynamic conditional correlation of VIX, policy uncertainty and US stock market returns. Results propose that correlation between stock returns and EPU is consistently negative between 1985 and 2013. In a parallel study, Ko and Lee (2015) particularly investigate the co-movement between EPU and 11 countries’ stock returns via wavelet analysis. Results propose that co-movement has been significantly more negative in early 2000’s and late 2000’s.
1052
Kang and Ratti (2013) numerically show the significance of EPU on stock returns; authors indicate that EPU accounts for 19% variation in the US stock returns. On the other hand, Bordo et al. (2016) approach the impact of uncertainty from another critical perspective which is the aggregate bank credit growth. Authors show that loan growth across US banks significantly slows down following an increase in economic policy uncertainty.
Arouri et al. (2016) study interaction of stock markets and policy uncertainty for a reasonably long period of time. Findings show that the interaction is negative and is even stronger during high volatility episodes. Karnizova and Li (2014), on the other hand, look how likely EPU predicts recessions. Their findings show that alongside with financial variables such as term spread, change in S&P500 and corporate spread, EPU adds further on prediction power.
Another critical strand of literature is investigating the emerging market perspective. Since emerging countries might be highly affected from changes in monetary policy in US, could have significant spillovers to other countries. For instance, Carrière-Swallow and Céspedes (2013) check and compare whether there is a difference in the impact of EPU on 20 developed and 20 emerging countries. Their findings imply that emerging markets experience a much higher decline in investment compared to developed economies and moreover recovery of emerging markets are, again, longer. Therefore, the spillover could be more traumatic for emerging markets. Arouri and Roubaud (2016) find similar results for China and India. On the contrary, Mensi et al. (2014) conclude that EPU has no effect on BRICS market, of which are Brazil, Russia, India, China and South Africa.
The findings on Turkish market, though, is not as extensive as US market. Akkus (2017) investigates the impact and his findings propose that economic growth of Turkey got negatively affected from the EPU. Another critical paper is by Sahin (2018), who interrelates different types of uncertainties with income or transactions velocity and examine which type of uncertainty affects velocity more. Although this paper measusres uncertainty through XXXBearing Akkus’ (2017) paper in mind, we approach question from a financial market perspective. Through this way, we aim to fill the gap partially via analyzing how EPU affects Turkish banking sector and financial markets.
METHODOLOGY
We utilize Hafner and Herwartz (2006) (HH, hereafter) volatility spillover test to examine the force of EPU on Turkish financial markets. Earlier tests by Cheung and Ng (1996) and Hong (2001) base their methodologies on cross-correlation functions; which are estimated from the standardized residuals of univariate GARCH’s. On the other hand, one major deficiency of these methods, is having oversizing problems especially when the sample series’ volatility are leptokurtic. Another critical deficiency of previous methods is being sensitive to lead and lag order selection (Nazlioglu et al., 2013) and this gives rise to question marks on stability and robustness of the method.
On the other hand, HH method surmounts these deficiencies, since it is based on Lagrange Multiplier principle (Nazlioglu et al., 2015). Furthermore, since our sample size is quite large, HH works even much better under these circumstances.
Another critical advantage of volatility spillover tests is the ability to analyze the relationship between selected variables more easily.
The null hypothesis of HH test is that there is no volatility spillover between selected two return series:
𝜀𝑖𝑡 = 𝜑𝑖𝑡√𝜎𝑖𝑡2(1 + 𝑧′𝑗𝜋) (1)
1053
𝑧𝑗𝑡 = (𝜀2𝑗𝑡−1, 𝜎2𝑗𝑡−1)′
(2)
In the above equations, 𝜑𝑖𝑡 and 𝜎𝑖𝑡2 are standardized residuals and conditional volatility of the
series 𝑖; 𝜀2
𝑗𝑡−1 and 𝜎2𝑗𝑡−1 are disturbance term and conditional standard deviation, respectively,
both squared of series 𝑗. Standardized residuals are regressed on derivatives and the null
hypothesis for the process is dependent upon the 𝜋 in the equation (Nazlioglu et al., 2015). If 𝜋 = 0, this means there is no volatility spillover between variables 𝑖 and 𝑗. The alternate hypothesis for, though, is 𝜋 ≠ 0 which rejects there is no volatility spillover from 𝑗 to 𝑖..
HH method checks the null test via LM test statistics as follows:
𝜆𝐿𝑀 = 1 4𝑇(∑(𝜎𝑖𝑡 2 𝑇 𝑡=1 − 1)𝑧′𝑗𝑡)𝑉(𝜃𝑖)−1(∑(𝜎𝑖𝑡2 𝑇 𝑡=1 − 1)𝑧𝑗𝑡) (3) Where, 𝑉(𝜃𝑖) = ĸ 4𝑇(∑ 𝑧𝑗𝑡𝑧 ′ 𝑗𝑡 𝑇 𝑡=1 − ∑ 𝑧𝑗𝑡𝑥′𝑖𝑡 𝑇 𝑡=1 (∑ 𝑥𝑖𝑡𝑥 ′ 𝑖𝑡 𝑇 𝑡=1 )−1∑ 𝑥𝑖𝑡𝑧′𝑗𝑡 𝑇 𝑡=1 ), (4) ĸ =1 𝑇∑(𝜎𝑖𝑡 2 𝑇 𝑡=1 −1)2 (5)
Since we have two variables and two misspecifications are extant in 𝑧𝑗𝑡, the test statistics is
compared with a chi-square distribution of two degrees of freedom.
DATA
Our sample period is between January 1986 and June 2018 and the dataset is open to public and available in monthly frequency. Given the long period we include in our analysis, we believe that the interaction between economic policy uncertainty and Turkish financial markets could be analyzed more thoroughly. This sample period does not only include the 2008 crisis, but also several significant events such as Gulf Wars 1 and 2, Black Monday and 9/11. Therefore, fiscal and monetary result ends of these events and how they affected Turkish system remains to be an interesting question.
We employ economic policy uncertainty index calculated via Baker et al. (2016). This index is built upon the newspaper coverages and counting the frequencies of selected wordings. Authors include 10 major newspapers of US which are USA Today, Miami Herald, Chicago Tribune, Washington Post, Los Angeles Times, Boston Globe, San Francisco Chronicle, Dallas Morning News, New York Times, and Wall Street Journal. To make sure that one article is about economic
policy, authors urge one article to include trio of selected words1.Next, authors calculate the EPU
index in 11 different categorical basis; which are Taxes, Government Spending & Other, Fiscal Policy, Monetary Policy, Healthcare, National Security, Financial Regulation, Regulation, Sovereign Debt & Currency Crises, Entitlement Programs and Trade Policy. So one can analyze
1054
not only the impact of aggregated economic policy uncertainty, but also the categorically how particularly sovereign debt or monetary policy.
We employ loan growth in Turkey and obtain the data from the Central bank of Turkey database. Even though economic policy uncertainty might have significant impact on a number of financial indicators, specific channel on the impact of Turkish economy is a more interesting question. As Mercan (2013) finds out there is a close and bi-directional relationship between credit and economic growth in Turkey. Therefore, following Bordo et al. (2016) we proxy the bank lending via domestic bank loans provided to individuals and corporations. If bank lending in Turkey gets stricter following a policy uncertainty in US, this means the transmission channel is through financial institutions, which is the heart of financial system.
Figure 1 shows the historical index levels and aggregated credit growth below. As one can see that EPU (economic policy uncertainty index) is higher during recession periods which are depicted in gray shaded areas. These recession periods are set by National Bureau of Economic Research and is publicly available. Furthermore, following the 2008 crisis, we see quite a volatile trend in EPU up until 2014 indicating the QE period. The variable KRED shows the bank loan growth in Turkey and in line with the growth in Turkish economy we see a consistent upward trend in the last three decades.
When we look to the categorical uncertainty indices, every of them show different historical levels depending upon the certain changes in different categories. For instance, NS which stands for national security shows the extreme levels during Gulf War and 9/11 events. Therefore, every category might or might not have an impact on Turkish loan growth.
Figure 1: Historical index levels and credit growth
Notes: EP stands for Entitlement Programs, EPU stands for aggregated Economic Policy
Uncertainty index, FP stands for Fiscal Policy, FR stands for Financial Regulation, GS stands for Government Spending, HC stands for healthcare, KRED stands for the aggregate loan growth in Turkey, MP stands for Monetary Policy, NS stands for National Security, R stands for Regulation, SD stands for Sovereign Debt, T stands for Taxes, TP stands for Trade Policy. Gray bars refer to recession periods determined by National Bureau of Economics Research.
We calculate monthly returns of each series via conventional log-return formula: 𝑅𝑖𝑡 = ln(𝑖𝑡) −
ln(𝑖𝑡−1). The descriptive statistics for our variables in returns are as presented below in Table 1.
0 100 200 300 400 500 600 1990 1995 2000 2005 2010 2015 EP 0 50 100 150 200 250 300 1990 1995 2000 2005 2010 2015 EPU 0 100 200 300 400 1990 1995 2000 2005 2010 2015 FP 0 200 400 600 800 1,000 1990 1995 2000 2005 2010 2015 FR 0 200 400 600 800 1990 1995 2000 2005 2010 2015 GS 0 100 200 300 400 500 600 1990 1995 2000 2005 2010 2015 HC 0 500,000,000 1,000,000,000 1,500,000,000 2,000,000,000 2,500,000,000 1990 1995 2000 2005 2010 2015 KRED 0 100 200 300 400 500 1990 1995 2000 2005 2010 2015 MP 0 200 400 600 800 1990 1995 2000 2005 2010 2015 NS 0 400 800 1,200 1,600 1990 1995 2000 2005 2010 2015 SD 0 100 200 300 400 500 1990 1995 2000 2005 2010 2015 T 0 200 400 600 800 1,000 1,200 1990 1995 2000 2005 2010 2015 TP
1055 Table 1: Descriptive Statistics
EP EPU FP FR GS HC KRED Mean -0.0007 -0.0009 0.0000 -0.0011 -0.0012 0.0010 0.0321 Median -0.0368 0.0048 -0.0091 -0.0390 0.0157 0.0088 0.0267 Maximum 1.6725 1.0885 1.3209 2.5397 2.2672 1.3722 0.2309 Minimum -2.3986 -0.8128 -1.2525 -2.4706 -1.9646 -2.6499 -0.0773 Std. Dev. 0.5836 0.2543 0.3472 0.7983 0.5615 0.4901 0.0358 Skewness 0.0065 0.4014 0.0935 0.1713 0.1591 -0.3882 1.5434 Kurtosis 3.4115 4.5217 3.4396 3.5334 3.7737 4.8680 8.5643 Observations 385 385 385 385 385 385 385 MP NS R SD T TP Mean -0.0045 0.0014 -0.0014 -0.0096 -0.0004 0.0049 Median -0.0048 -0.0141 -0.0180 0.0000 -0.0102 -0.0278 Maximum 1.7674 2.8526 1.1883 3.3579 1.1724 1.8986 Minimum -1.5213 -1.3091 -1.1744 -3.3602 -0.8676 -2.0922 Std. Dev. 0.5193 0.4234 0.3613 0.9413 0.3487 0.6836 Skewness 0.2403 0.8123 0.2337 0.0511 0.1273 -0.0397 Kurtosis 3.4719 8.0837 3.6406 3.9739 2.9738 2.9448 Observations 385 385 385 385 385 385
Notes: EP stands for Entitlement Programs, EPU stands for aggregated Economic Policy
Uncertainty index, FP stands for Fiscal Policy, FR stands for Financial Regulation, GS stands for Government Spending, HC stands for healthcare, KRED stands for the aggregate loan growth in Turkey, MP stands for Monetary Policy, NS stands for National Security, R stands for Regulation, SD stands for Sovereign Debt, T stands for Taxes, TP stands for Trade Policy.
RESULTS
We mainly investigate the transmission channel of US economic policy uncertainty to Turkish markets through bank lending. The criticality of bank lending lies in the significant role of banks on transmitting liquidity across markets. Moreover, contagion literature argues that banking originated crises occur much faster and wider compared to real sector originated crises (Kaufman, 1994). Therefore, financial institutions hold a crucial role on transmission of economic conditions across markets.
We employ HH volatility spillover method for the sample period between 1985 and 2018. We also check which categorical policy uncertainty has or does not have impact on bank lending. Therefore, we can comment on which category specifically affects bank lending and Turkish economy on the whole.
First of all, we check the stationarity of our time series, through employing Augmented Dickey Fuller (Dickey and Fuller, 1979) and Phillips-Perron (1988) unit root tests. The null hypothesis for these tests assert that each of the series has unit root and is non-stationary. If we reject (fail to reject) the test statistic, this would imply that the respective time series is stationary (not stationary). Results are presented in Table 2, below. As one can note that results imply that all series are stationary and has no unit root. Therefore, we could easily employ HH methodology.
1056 Table 2: Unit Root Test Results
ADF PP
t-stat p-value t-stat p-value
EP -10.84 0.00 -50.84 0.00 EPU -14.99 0.00 -33.27 0.00 FP -11.58 0.00 -33.68 0.00 FR -14.98 0.00 -59.84 0.00 GS -18.35 0.00 -49.27 0.00 HC -16.02 0.00 -45.51 0.00 KRED -2.59 0.28 -19.48 0.00 MP -11.57 0.00 -79.64 0.00 NS -25.84 0.00 -39.54 0.00 R -16.11 0.00 -46.57 0.00 SD -16.76 0.00 -44.93 0.00 T -11.51 0.00 -35.54 0.00 TP -14.81 0.00 -63.39 0.00
Notes: Table presents Augmented Dickey Fuller (ADF) and Phillips-Perron (PP) unit root test
results. EP stands for Entitlement Programs, EPU stands for aggregated Economic Policy Uncertainty index, FP stands for Fiscal Policy, FR stands for Financial Regulation, GS stands for Government Spending, HC stands for healthcare, KRED stands for the aggregate loan growth in Turkey, MP stands for Monetary Policy, NS stands for National Security, R stands for Regulation, SD stands for Sovereign Debt, T stands for Taxes, TP stands for Trade Policy.
Table 3: Volatility Spillover between EPU and KRED
TO KRED EPU FR O M KRED - 0.3863 EPU 0.0046 -
Notes: The values provided are p-values and hence if the figure is smaller than 0.10 it implies
there is a significant volatility spillover from the respective variable to the other.
Next, we run HH volatility spillover test, first between aggregated economic policy uncertainty index and credit growth. Findings propose that there is a significant volatility spillover from EPU to KRED, but not vice versa. Therefore, we can conclude that the policy uncertainty in US has an impact on the loan growth in Turkey. This means that the transmission mechanism from economic policy uncertainty happens mainly through financial institutions. A potential economic downturn leads to liquidity problems and this has a direct impact on emerging markets (Chen et al., 2014). Our findings support this finding and indicate that loan growth becomes volatile following the volatility in EPU.
1057
Table 4: Volatility Spillover between Categorical EPU and KRED
TO TO KRED FR KRED EP FR O M KRED - 0.4219 FR O M KRED - 0.0159 FR 0.0660 - EP 0.6461 - TO TO KRED MP KRED R FR O M KRED - 0.1036 FR O M KRED - 0.2022 MP 0.6078 - R 0.0170 - TO TO KRED T KRED TP FR O M KRED - 0.1578 FR O M KRED - 0.3921 T 0.0729 - TP 0.1929 - TO TO KRED GS KRED SD FR O M KRED - 0.1844 FR O M KRED - 0.1858 GS 0.2512 - SD 0.8278 - TO TO KRED HC KRED FP FR O M KRED - 0.0002 FR O M KRED - 0.2982 HC 0.8689 - FP 0.5204 - TO KRED NS FR O M KRED - 0.5073 NS 0.2929 -
Notes: Table presents Augmented Dickey Fuller (ADF) and Phillips-Perron (PP) unit root test
results. EP stands for Entitlement Programs, EPU stands for aggregated Economic Policy Uncertainty index, FP stands for Fiscal Policy, FR stands for Financial Regulation, GS stands for Government Spending, HC stands for healthcare, KRED stands for the aggregate loan growth in Turkey, MP stands for Monetary Policy, NS stands for National Security, R stands for Regulation, SD stands for Sovereign Debt, T stands for Taxes, TP stands for Trade Policy.
Next, we further analyze which categorical policy uncertainty is more critical on the economic policy uncertainty.
Findings imply that most of the categorical policy uncertainties have no effect on Turkish credit market growth. The only exceptions are financial regulation and taxes, which spillover to Turkish loan growth significantly. This means that the uncertainty especially on taxes and regulation affect credit availability, which could indicate that financial institutions are critically influenced by restrictive regulations or tightening tax decisions. Given that most of the banks have multinational arms in Turkey, any regulatory changes could have direct effect on their profitability. Hence, they might be highly hesitant on giving out loans following an uncertainty in regulations. Quite surprisingly results also show that the loan growth spills over to healthcare index. This odd result is attributed to healthcare index proxying for another variable, which we are not aware of.
1058 CONCLUSION
The impact of uncertainty on economies has been evidenced in the literature. Uncertainty, especially, affects the investment decisions of corporations and consumption behavior of consumers. Investment, being an irreversible one, raises worries of investment-makers during uncertainty. Therefore, economy gets a big hit with uncertainty, if it continues for a considerably long time. Recently, Baker et al. (2016) come up with a measure of policy uncertainty measure via the selected word counts in a number of newspapers. Hence, the index is easy to measure and is not affected by statistical problems of other measures. Although Baker et al. (2006) influenced extensive amount of study, how the transmission of EPU towards other economies has not been studied, significantly. Following Bordo et al. (2016), we investigate the volatility spillover between EPU and credit growth in Turkey. The spillover of economic uncertainty to emerging markets is an interesting question. Plus, it allows us to understand the contagion mechanism further.
We investigate the spillover of economic policy uncertainty and credit growth effect via Hafner and Herwartz (2006) methodology. The importance of credit growth lies in its direct impact on the economy. As credit channels become more easy to attain, households tend to consume more and businesses tend to invest further. However, if credit channel is getting tighter, efficient allocation of capital gets disrupted and economic conditions might hurt for a short and sometimes long-term.
The findings propose that, indeed, as we expected the spillover is uni-directional and from the EPU to loan growth. Hence, economic policy uncertainty in the US spills over to Turkish economy through credit market growth. Apparently, financial institutions located in Turkey tighten the credit belt and anticipate the market to assess the underlying reason behind increasing uncertainty. Therefore, our study shows that the transmission channel of policy uncertainty is through credit growth via financial institutions.
Moreover, the categorical EPU show that regulations and tax decisions are the most significant impact-makers of all other categories. As a future line of research, one can extend the analysis to other countries and check whether government or private banks have a more dominant role over transmission.
1059
REFERENCES
Akkus, Ömer., (2017), “Ekonomik Politika Belirsizliği ve Politik İstikrarsızlığın Büyüme Üzerindeki Etkisi” Anadolu Üniversitesi Sosyal Bilimler Dergisi, 17(3), s:27-42.
Antonakakis, N., Chatziantoniou, I., Filis, G. (2013), “Dynamic co-movements of stock market returns, implied volatility and policy uncertainty” Economics Letters, 120(1), s: 87-92. Arouri, M., Estay, C., Rault, C., Roubaud, D. (2016), “Economic policy uncertainty and stock markets: Long-run evidence from the US”. Finance Research Letters, 18, s: 136-141.
Arouri, Mohamed; Roubaud, David. (2016), “On the determinants of stock market dynamics in emerging countries: the role of economic policy uncertainty in China and India”. Economics
Bulletin, 36(2), s: 760-770.
Baker, S. R., Bloom, N., Davis, S. J. (2016), “Measuring economic policy uncertainty” The
Quarterly Journal of Economics, 131(4), s: 1593-1636.
Bernal, O., Gnabo, J. Y., Guilmin, G. (2016), “Economic policy uncertainty and risk spillovers in the Eurozone.” Journal of International Money and Finance, 65,s: 24-45.
Bernanke, Ben S., (1983), “Irreversibility, uncertainty, and cyclical investment” The Quarterly
Journal of Economics, 98(1), s: 85-106.
Bloom, Nicholas, (2014), “Fluctuations in uncertainty” Journal of Economic
Perspectives, 28(2), s: 153-76.
Bloom, N., Floetotto, M., Jaimovich, N., Saporta‐Eksten, I., Terry, S. J., (2018), “Really uncertain business cycles”. Econometrica, 86(3), s: 1031-1065.
Bordo, M. D., Duca, J. V., Koch, C., (2016), “Economic policy uncertainty and the credit channel: Aggregate and bank level US evidence over several decades” Journal of Financial
Stability, 26, s: 90-106.
Carrière-Swallow, Yan; Céspedes, Luis Felipe, (2013), “The impact of uncertainty shocks in emerging economies” Journal of International Economics, 90(2), s: 316-325.
Chau, F., Deesomsak, R., Wang, J., (2014), “Political uncertainty and stock market volatility in the Middle East and North African (MENA) countries” Journal of International Financial
Markets, Institutions and Money, 28, s: 1-19.
Chen, M. J., Griffoli, M. T. M., Sahay, M. R., (2014), “Spillovers from United States monetary policy on emerging markets: different this time?” International Monetary Fund, (No. 14-240).
Cheung, Yin-Wong; Ng, Lilian K.(1996), “A causality-in-variance test and its application to financial market prices” Journal of Econometrics, 72(1-2), s: 33-48.
Colombo, Valentina, (2013), “Economic policy uncertainty in the US: Does it matter for the Euro area?” Economics Letters, 121(1), s: 39-42.
Connolly, R., Stivers, C., Sun, L., (2005), “Stock market uncertainty and the stock-bond return relation” Journal of Financial and Quantitative Analysis, 40(1), s: 161-194.
Dickey, David A.; Fuller, Wayne A., (1979), “Distribution of the estimators for autoregressive time series with a unit root” Journal of the American Statistical Association, 74(366a), s: 427-431.
Gilchrist, S., Sim, J. W., Zakrajšek, E., (2014), “Uncertainty, financial frictions, and investment dynamics” (No. w20038). National Bureau of Economic Research.
Glick, Reuven; Leduc, Sylvain, (2012), “Central bank announcements of asset purchases and the impact on global financial and commodity markets”. Journal of International Money and
1060
Gomes, F. J., Kotlikoff, L. J., Viceira, L. M., (2012), “The excess burden of government indecision” Tax Policy and the Economy, 26(1), s: 125-164.
Hafner, Christian M.; Herwartz, Helmut, (2006), “A Lagrange multiplier test for causality in variance” Economics Letters, 93(1), 137-141.
Hong, Yongmiao, (2001), “A test for volatility spillover with application to exchange
rates” Journal of Econometrics, 103(1-2), s: 183-224.
Hermes, Niels; Lensink, Robert, (2001), “Capital flight and the uncertainty of government policies” Economics letters, 71(3), 377-381.
Kang, Wensheng; Ratti, Ronald A., (2013), “Oil shocks, policy uncertainty and stock market return” Journal of International Financial Markets, Institutions and Money, 26, s: 305-318. Karnizova, Lilia; Li, Jiaxiong Chris, (2014), “Economic policy uncertainty, financial markets and probability of US recessions” Economics Letters, 125(2), s: 261-265.
Kaufman, George G, (1994), “Bank contagion: A review of the theory and evidence” Journal
of Financial Services Research, 8(2), s: 123-150.
Ko, Jun-Hyung; Lee, Chang-Min, (2015), “International economic policy uncertainty and stock prices: Wavelet approach” Economics Letters, 134, s: 118-122.
Mensi, W., Hammoudeh, S., Reboredo, J. C., Nguyen, D. K., (2014), “Do global factors impact BRICS stock markets? A quantile regression approach” Emerging Markets Review, 19, s: 1-17.
Mercan, M. (2013). “Kredi hacmindeki değişimlerin ekonomik büyümeye etkisi: Türkiye ekonomisi için sınır testi yaklaşımı”. Bankacılar Dergisi, 84, s: 54-71.
Nazlioglu, S., Erdem, C., Soytas, U., (2013), “Volatility spillover between oil and agricultural commodity markets” Energy Economics, 36, s: 658-665.
Nazlioglu, S., Soytas, U., Gupta, R., (2015), “Oil prices and financial stress: A volatility spillover analysis” Energy Policy, 82, s: 278-288.
Pastor, Lubos; Veronesi Pietro, (2012), “Uncertainty about government policy and stock prices” The Journal of Finance, 67(4), s: 1219-1264.
Phillips, Peter CB; Perron, Pierre, (1988), “Testing for a unit root in time series regression”. Biometrika, 75(2), s: 335-346.
Sahin, Afsin. (2018), “Staying vigilant of uncertainty to velocity of money: an application for oil‐producing countries”. OPEC Energy Review, 42(2), s: 170-195.
Stein, Luke CD; Stone, Elizabeth “The Effect of Uncertainty on Investment, Hiring, and R&D: Causal Evidence from Equity Options”
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1649108 (05.3.2018)
Yonce, Adam, (2015), “US Corporate Investment Over the Political Cycle”, The Quarterly
Journal of Finance, 5(01).
Zhang, X. Frank., (2006), “Information uncertainty and stock returns”, The Journal of
1061 Araştırma Makalesi
US Economic Policy Uncertainty And Loan Growth: Evidence From Turkey
ABD Ekonomik Politika Belirsizliği ve Kredi Büyümesi: Türkiye’den Kanıt
Beyza Mina ORDU AKKAYA
Ankara Sosyal Bilimler Üniversitesi İşletme Bölümü.
beyza.akkaya@asbu.edu.tr https://orcid.org/0000-0003-4353-3977
Genişletilmiş Özet
Ekonomik politika belirsizliği (EPB) piyasa oyuncularının yatırım, harcama ve diğer iktisadi kararlarının hepsini etkilediği için, oldukça önem arz etmektedir. Ancak, EPB’nin önemi, özellikle 2008 krizi ile daha da öne çıkan bir konu olmuştur. Zira, belirsizlik, ekonominin toparlanmasını nispeten uzatmaktadır (Antonakakis vd., 2013).
Belirsizliğin ekonomi ile olan ilintisi ise, politika yapıcıların seçtiği politikaların başarısı ile kendin göstermektedir. Zira politikaların bilinen etkisi, belirsizlik altında çok farklı şekilde kendini gösterebilir ve bu nedenle de politikaların sonucu beklenenden sapabilir. Yatırımcılar ise yatırım kararlarının geri dönülemez olması nedeniyle, belirsizlik ortamlarında yatırımlarını erteleme yoluna gitmektedirler. (Bernanke, 1983). Bir diğer önemli nokta da belirsizliğin, ekonomilerin büyümesi ve enflasyon üzerinde negatif etkisi olmasıdır (Glick and Leduc, 2012). Bu çalışma ana olarak, ABD’deki ekonomik politika belirsizliğin Türkiye’deki kredi büyümeleri üzerinden etkisini incelemektedir. Dolayısıyla da çalışma, belirsizliğin ekonomi üzerindeki etkisini inceleyen yazın kolu ile yakından ilgilidir. Örneğin, Bernanke (1983) belirsizlik altında yatırım kararlarının ertelendiğini bulurken, Bloom (2014), belirsizliğin resesyon dönemlerinde daha da arttığını göstermiştir. Dolayısıyla da belirsizlik ve ekonomik gidişat arasında yakın ve iç içe geçmiş bir ilişki olduğu bulunmuştur. Örneğin, Stein ve Stone (2012) ve Bloom vd. (2012) tarafından yapılmış ampirik çalışmada “sadece” belirsizliğin ABD ekonomisinde yaşanan %9’luk küçülmenin üçte birinden sorumlu olduğunu göstermiştir. Benzer şekilde, Hermes ve Lensink (2001) vergi ödemeleri, denetim ve enflasyon konusundaki belirsizliğin yabancı yatırımcıyı tedirgin eden ve doğrudan yatırımı ciddi anlamda düşüren bir etkisi olduğunu da bulmuştur. Belirsizliği hisse senedi opsiyonlarının zımni oynaklığı kullanarak ölçen çalışmalar ise, bahsi geçen belirsizliğin yüksek olduğu dönemlerde hisse senedi getirilerinin az olduğunu göstermiştir. Dolayısıyla, belirsizlik olgusunun tanımı duruma göre değişmekle birlikte, ekonomi üzerindeki etkisi çeşitli şekillerde gösterilmiştir. Ancak belirsizlik çalışmalarındaki ilgi çeken konu ise, belirsizliğin nasıl ölçüleceği konusundadır. Baker vd. (2016), EPB endeksini 1985 yılından başlayarak Amerika Birleşik Devletleri (ABD) için hesaplamışlardır. Bahsi geçen endeks, belirli gazetelerde ekonomi ve belirsizliğe dair seçili kelimelerin kaçar kere geçtiğini gösteren bir endekstir. Endeksin doğru ve düzgün bir şekilde belirsizliği ölçüp ölçmediği de çeşitli istatistiki yöntemlerle de kontrol edilmiş ve bu nedenle kısa zamanda oldukça fazla atıf alarak yazında yeni bir yayın kolu yaratmıştır. Yazarların en önemli savı ise 2008 krizinin daha önceden fark edilebilmesi, krizin dünya ekonomisi üzerindeki etkisini azaltabileceği yönündedir.
EPB endeksine ilişkin ilk çalışmalardan biri Colombo (2013) tarafından yapılmış ve ABD’den Avro bölgesine olan belirsizlik yayılımı incelenmiştir. Sonuçlar EPB’deki beklenmedik bir artışın Avro bölgesindeki ekonomik göstergelerde negatif bir değişime yol açtığını göstermiştir. Benzer şekilde, Bernal vd. (2016) Avro bölgesindeki 10 ülkeyi incelemiş ve sonuçlar, yayılmanın önce
1062
ABD’den Fransa ve Almanya’nın devlet tahvili piyasalarına doğru olduğunu, daha sonra bu ülkelerin kalan Avro bölgesi ülkelerini etkilediğini göstermiştir. Antonakakis vd. (2013) ise EPB ve ABD hisse senedi getirileri arasında dinamik koşullu korelasyonu incelemiş ve sürekli negatif seyreden bir ilişki olduğunu göstermiştir. Daha sonra, Kang ve Ratti (2013) EPB’nin hisse senedi getirilerinin %19’unu açıkladığını göstermişlerdir. Bordo vd. (2016) ise belirsizliğin ekonomi üzerindeki etkisine başka bir açıdan yaklaşmış ve bankaların sağladığı kredi büyümesi üzerinden incelemiştir. Sonuçlar, ekonomik belirsizliğin arttığı dönemlerde, ABD’de yer alan banka kredilerinin büyümesinde ciddi bir yavaşlama olduğunu göstermiştir.
EPB’nin ilgilendiği bir diğer önemli yazın kolu ise EPB’nin gelişmekte olan piyasalar üzerindeki etkisidir. Zira, gelişmekte olan piyasaların ABD’deki politikalardan etkilenebilmesi, bu ülkeler arasındaki yayılma etkisinin EPB üzerinden olabileceği fikrini doğurmuştur. Örneğin Carrière-Swallow and Céspedes (2013) EPB’nin 20 gelişmiş ve 20 gelişmekte olan piyasa üzerindeki etkilerini ayrı olarak incelemiş ve gelişmekte olan piyasalardaki yatırımın çok daha fazla negatif etkilendiğini göstermişlerdir. Ayrıca, yazarlar gelişmekte olan piyasaların negatif etkilendikten sonra toparlanma süreçlerinin de fazlasıyla uzun sürebildiğini göstermiştir. Bu nedenle de EPB gelişmekte olan piyasalar için daha travmatik etkiler doğurabilmektedir.
Türkiye piyasası için yapılan çalışmalar ise oldukça sınırlıdır. Akkus (2017) EPB’nin Türkiye ekonomisi üzerindeki etkisini incelemiş ve olumsuz etkilendiğini bulmuştur. Akkus’un (2017) makalesi göz önünde bulundurularak, soruna finansal piyasalar açısından bakılması öngörülmektedir. Bordo vd.’nin (2016) çalışmasını da göz önünde bulundurarak EPB’nin Türk bankacılık ve hisse senedi piyasaları üzerindeki etkisi incelenecektir.
Bu çalışmada yöntem olarak Hafner ve Herwartz’ın (2006) (HH) oynaklık yayılma analizi kullanılarak EPB’nin Türkiye finansal piyasaları üzerindeki etkisi incelenmiştir. Erken oynaklık yayılma analizlerinden olan Cheung ve Ng (1996) ve Hong’un (2001) çalışmaları tek değişkenli GARCH’ların standardize edilmiş artıkları üzerinden hesaplanan çapraz korelasyon fonksiyonları üzerinde temellenmiştir. Fakat bu testlerin en büyük eksiği, örneklem oynaklığının basık olduğu durumlarda testin sağlıklı sonuçlar vermemesidir. Aynı şekilde, bu metotların sunduğu bulgular öncü ve artçı seçimlerine fazlasıyla hassastır (Nazlioglu vd., 2013). Bu da, yöntemlerin sonuçlarının sorgulanabilirliğini doğurmaktadır. Ancak HH yöntemi bahsi geçen problemleri ortadan kaldırmakta ve örneklemin büyük olduğu durumlarda daha da sağlıklı sonuçlar vermektedir (Nazlioglu vd., 2015).
Örneklem dönemimiz Ocak 1986 ve Haziran 2018 arasındadır ve veri frekansı aylık olarak belirlenmiştir. Dönemin uzun olması nedeniyle EPB’nin Türkiye piyasaları ile olan ilişkisi daha da net incelenebilecektir. Baker vd.’nin (2016) yarattığı endeks ABD’deki en önde gelen 10 gazetesinde (örneğin, Washington Post, Wall Street Journal, New york Times vb.) seçilmiş bazı kelimelerin yoğunluğuna göre hesaplamaktadır. Ayrıca EPB 11 farklı kategoriye ayrılmış ve kategorik olarak artış ve azalışların incelenmesi fırsatı doğmuştur. Bu kategoriler, Vergi, Hükümet Harcamaları, Mali Politika, Para Politikası, Sağlık, Ulusal Güvenlik, Finansal Regülasyon, Regülasyon, Kamu Borcu ve Döviz Krizi, Yetkilendirme ve Ticaret Politikası olarak ayrılmıştır. Çalışmada her bir kategorinin ayrıca kredi büyüklüğü ile ilişkisi de incelenmiştir. Türkiye’de yer alan bankaların sunduğu kredilerin tutarı ise Merkez Bankası verisetinden elde edilmiştir. Mercan’ın (2013) bulduğu üzere kredi büyümesi ile ekonomik büyüme arasında çift yönlü yakın bir ilişki mevcuttur ve bu nedenle de ekonomi üzerinde direkt bir etkisi beklenebilir.
Verilerin ham ve konvansiyonel getiri yöntemi (𝑅𝑖𝑡 = ln(𝑖𝑡) − ln(𝑖𝑡−1))ile hesaplanmış ve getiri
yöntemine ilişkin betimleyici istatistikler de Tablo 1’de sunulmuştur. Görülebileceği üzere örneklem dönemimiz içinde ortalama kredi büyümesi %3 olmuş ve bu da Türkiye’nin son 30-40 yıldaki ekonomik büyümesinin göstergelerinden biridir.
Bu çalışmadaki esas amaç, ABD ekonomik politika belirsizliğinin, Türkiye’de yerleşik bankaların kredi sağlama davranışlarına olan etkisini incelemektir. Bankaların kredi sağlamasının en önemli özelliklerinden biri piyasadaki likidite ihtiyaçlarına karşılık vermeleri ve piyasalar arasındaki likidite taşımını sağlamalarıdır. Aynı zamanda bulaşıcılık literatürü, bankacılık sektöründen
1063
doğan krizlerin, reel sektörden doğan krizlerden çok daha kalıcı ve hızlı yayıldığını göstermiştir (Kaufman, 1994).
Çalışmanın ilk aşamasında HH’nin oynaklık yayılma analizinden evvel, birim kök testi bütün veriler için yapılmıştır. Bütün serilerin birim kökü olmadığı ve durağan oldukları tespit edilmiştir. Daha sonra Tablo 3’te HH oynaklık yayılma analizinin esas sonuç tablosu görülmektedir. Tablodaki veriler, HH test sonuçlarının p-değerini göstermektedir. Bu nedenle, değerin 0.10’dan küçük olduğu durumda istatistiki olarak anlamlı bir sonuç olduğu kanısına varılmaktadır. Görülebileceği üzere, EPB’den kredi büyümesine anlamlı bir oynaklık yayılımı varken, beklenebileceği üzere kredi büyümesinden EPB’ye benzeri bir yayılım mevcut değildir. Bu da bize ABD’deki ekonomik politika belirsizliğinin Türkiye’de yerleşik bankaların kredi büyümesini direkt olarak etkilediğini gösteren önemli bir bulgudur. Dolayısıyla, ekonomik belirsizliklerin Türkiye ekonomisine olan etkisinin bankalar vasıtası ile de geçtiğini söylemek yanlış olmayacaktır.
Tablo 4’te ise kategorik olarak ayrılmış EPB’lerin her biri ile kredi büyümesi arasındaki ilişki incelenmiştir. Sonuçlardan görülebileceği üzere, finansal regülasyon ve vergi ile ilişkili belirsizlikler Türkiye’deki banka kredisi büyümelerinde oynaklık olarak kendini göstermektedir. Bu sonucun altında yatan sebep ise ABD’de artan regülasyon veya vergi kanunundaki değişikliklerin kimi şekillerde diğer gelişmekte olan piyasalardaki yatırımcı davranışını etkileyebileceği şeklindedir. Beklenmedik olan bulgu ise, Türkiye’deki kredi büyümesinin EPB’nin sağlık hizmetleri kategorisinde oynaklık yayılımına sebep olmasıdır. Muhtemelen sağlık hizmetleri bilinmeyen başka bir değişkeni yansıtmaktadır.
Sonuç olarak, ekonomilerdeki belirsizliğin pek çok sektör ve diğer piyasalar üzerindeki etkileri yazında bolca tartışılmıştır. Belirsizlik, özellikle şirketlerin yatırım ve kişilerin tüketim davranışlarını çok ciddi ölçüde etkilemektedir. Ancak belirsizliğin nasıl ölçüleceği konusu literatürdeki esas sorulardan biri olmuştur. Baker vd. (2016)’nin önde gelen günlük gazeteler vasıtası ile hazırladığı EPB endeksi halihazırda piyasada kullanılan en iyi göstergelerden biridir. Önceki çalışmalarda ABD’deki kredi büyümelerinin EPB’deki bir artışla ciddi anlamda etkilendiği gösterilmiştir (Bordo vd., 2016). Bu nedenle de Bordo vd. (2016), bulgularında EPB’nin ABD piyasalarına olan etkisinin esas olarak banka kredileri vasıtasıyla olduğunu incelemişlerdir. Benzer şekilde yayılımın diğer ülkelere de krediler vasıtasıyla olup olmadığı ise ilgi çeken bir konudur. Bu nedenle EPB’nin Türkiye’de yerleşik bankaların kredi büyümesi üzerindeki etkileri HH yöntemi ile incelenmiştir. Sonuçlar, beklenildiği üzere EPB’nin tek yönlü olarak bankaların kredi sağlama davranışı üzerinde etki ettiğini göstermiştir. Dolayısıyla, bankaların likidite sağlama ve piyasalar arasındaki köprü vazifesi göze alındığında kredi büyümesinin ne kadar hassas olduğu göze çarpmaktadır.
