Credit Risk Determinants in Indonesia

Credit is the main business of the banking industry, therefore, in running the business, the bank is always overshadowed by the credit risk the which can be determined by the ratio of non-performing loans (NPL). The development of technology, finance digital brings the outside could impact on the financial industry both positive and negative. The purpose of this study was to analyze the interest rate, inflation, exchange rates, gross domestic product (GDP), a dummy finance digitalization policies in the long term and the short term of the non-performing loan (NPL) of conventional commercial banks in Indonesia The analytical method used in this research is-EG Error Correction Model (ECM), The Data used in this research is secondary quarterly time series data from the 2008 quarter 1-2017 4. The time series of data are not stationar Often that can cause spurious regression results, the exact models used is-EG Error Correction Model (ECM), This models may explain the behavior of short-term and long-term. The results Showed in the short-term variable interest rates significanly to non-performing loans, while in the long-term variable interest rate, exchange rate, and GDP Significantly, non-performing loans.


INTRODUCTION
Credit is one form of financing provided by banks. Bank in lending is always faced with various risks (Arofu, 2017), One of the risks faced is the high ratio of non-performing loans, the greater the NPL ratio, the level of the bank's health will get worse and have an impact on potential bank failures will be greater and vice versa, non-performing loan is calculated according to the formula, the number of loans classified noncurrent divided by total outstanding loans. (Firman, 2014) Indonesia. The next phase of BI issued Circular No. 16/11 / DKSP on electronic money and BI Circular Letter No. 16/12 / DPAU on the Implementation of the Financial Services Digital (Zahirudin, 2015), Digital Financial Services (LKD) is a merger of the provision of financial services and payments are submitted and managed using mobile technology or web technologies and network agent (Herdiana, 2014), Products from LKD such as mobile payment (m-payment), mobile banking (m-banking), internet banking and electronic money (emoney). Digital Finance to facilitate the public to access the services of financial institutions but on the other also carries the risk if it is not supported by valid data. In terms of credit fraud lot of credit, less precise in lending, and the increase of bad debts. According to Shingjergji (2013); Tanasković & Jandric (2015); Bozdo & Kripa (2015); Ginting (2016) NPL rise much influenced by macroeconomic factors. rate gradually declined from 9% -7.25%, but in the same time frame it is inversely proportional to the state of non-performing loans that show an increase in this period (see figure 1). The condition was repeated in 2016 Q1 -2016 Q4 where the interest rate gradually declined from 7% -4.75% but the level of nonperforming loans actually increased (see figure  1). Inflation in the period of 2008 Quarter 1-2017 Q4 fluctuate, but when viewed as a whole experienced a decrease trend inflation. In 2008 the upward trend of inflation experienced until at its peak in the 2nd quarter which touched 1.5% as a result of the financial crisis that hit the world. after 2008 Q2 economic conditions began gradually improved with the decline in inflation. Since 2014 Quarter 4-2015 Q3 inflation gradually declined from 1.5% -0.3%, but the level of nonperforming loans actually increased during that time (see figure 1.1). The fall in the inflation rate indicates improvements in the economy by increasing purchasing power and further reduce the burden of debtors to repay their loans. Gross Domestic Product (GDP) in the period 2008 Quarter 1-2017 Q4, as a whole experienced a rising trend. In 2008 Q4 GDP gradually began to increase from Rp. 1238754.5, -until 2009 Q2 Rp. 1383350.5, -but in the same time non-performing loans experienced a rise of 3.3% -4% (see figure 1.1). A similar thing happened in 2016 Q1 to 2016 Q4 where GDP showed a trend increase from Rp. 2.26468 million, -to Rp. 2.385..244, -but the level of non-performing loans experienced a rise of 2.8% -3.15% (see figure 1.1). The rise in GDP indicates improvements in the economy by increasing public revenue and increase the ability to repay the loan.
Based on these data we can see that on a quarterly basis in the year 2008-2017 interest rates, inflation experienced a decline trand, although exchange rate depreciated but still looks stable and GDP shows trand increased. This shows that the economy can be said to be improving. There is a gap phenomenon here, where non-performing laon occur trand increase (see figure 1.1). It is counterproductive when viewed macroeconomic indicates improving conditions, improving economic conditions indicate the economy running smoothly, which means that people and companies can pay back the loan to the bank with the better so as to reduce the ratio of non-performing laon. Namnun according to Figure 1.1 NPL even increased when the economy goes toward the better.
This study aimed to analyze the relationship between macroeconomic variables and their influence on the bank's non-performing laon common in Indonesia.

RESEARCH METHODS
This study uses a quantitative approach, according Sugiyono (2010) emphasizes the quantitative approach to the dates in the form of numbers or qualitative data diangkakan. Quantitative methods will be acquired significance significant group differences or relationships between the variables studied. The data used in this research is secondary data time series with the first quarter of the period 2008 to 2017 4th quarter were sourced from various agencies related publications, such as Bank Indonesia, the Financial Services Authority, the Central Statistics Agency (BPS).
The method of data analysis is a research method used in analyzing the data. The steps are performed in the data analysis in this study is the correlation analysis method Error (Error Correction Model / ECM). Analysis tools using a computer program Econometric Views 9.0 (Eviews). According toBasuki (2016) method Error Correcttion Model (ECM) as a means of econometric calculations and is also used analisisi descriptive method aims to identify long-term and -term due to the presence of cointegration among variables. The model will be used as a research model that is encapsulated in the ECM-EG as follows : Long-term models are: Based on table 1 note that all significant variables on the level 1st different-trend andintercept, the probability value of each variable NPLs 0.0000, 0.0143 annual interest rate, inflation of 0.0000, the exchange rate as big as GDP amounted to 0.0143 and 0.0020, which means α is smaller than 5%. Source: Processed Eviews 9.0 * Significant at the level α = 5% The results of long-term estimates in Table 2. model can be written as follows: NPLt = 5.649081 -.3375Irt + 0,16INFt + 0,000370KURSt -0,0000024PDBt-0,22Dt + μt The results of short-term estimates in Table 3. model can be written as follows:

* Significant at the level α = 5%
Model ECM-EG is said to be valid if the error correction coefficient signs (ECT) is negative and statistically significant (Widarjono 2009). Based on estimates in Table  1.3 in mind that the ECT significant at α = 5% with a coefficient value of-0.544573 which means that the model ECM-EG is a legitimate use and able to correct imbalances in the short term into long-term balance of 54.4%.
Influence the variable interest rate on short-term estimates have a coefficient of -0.254185meaning that any reduction in interest rates by 1% would cause an increase in NPLs0.254185%, Furthermore, interest rates significant negative effect in the short term to NPL it is shown on the probability value <critical limit (0.0239 <0.05). or H0 received, is in line with research (Saba, Kouser, & Azeem, 2012); Bekhet & Eletter (2014), The estimation results of short-term interest rates significantly affect non-performing loans, the negative effect can be interpreted in case of interest rate cuts, the public and business people tend more to borrow money in the bank because the interest rate on the pay is not low, thus lending more and potentially increase credit risk.
Variable interest rates on long-term estimates have a coefficient of -0.337305 meaning that any reduction in interest rates of 1% would cause a rise in NPLs amounting to 0,337 305%, Furthermore,Interest Rates significant negative effect in the long term to NPL it is shown on the probability value <critical limit (0.0001 <0.05).or H0 be accepted. In the long term the interest rate cuts to spur credit growth and further drive the economy, but the decline in interest rates continuously increasing number of credit could trigger uncontrollable and lead to greater risk of bad loans.
Variables influence inflation in the short term estimates have a coefficient of 0.148429 means that any increase in inflation by 1% would cause an increase in NPLs0.148429%, Furthermore, inflation probability value> critical limit (0.0903> 0.05) so that it can be concluded in the short-term inflation variable has no effect on the NPL. or H0be accepted. Variable inflation in the long-term estimates have a coefficient of 0.160525 means that any increase in inflation by 1% would cause an increase in NPLs0.160525%, Furthermore, inflation probability value> critical limit (0.335> 0.05) so that it can be concluded in the long-term inflation variable has no effect on the NPL. or H0be accepted. The results of this study support (Firman, 2014).
The estimation results indicate a longterm short-term inflation does not affect the NPL ratio, despite inflation variable has a high fluctuation but it is only temporary as it approached the religious holiday, the moment of beginning and end of the year and the inflation rate is mostly caused by food. The rise of high inflation followed by a variety of market control policies by the government so that the inflation rate back controlled by the government and does not affect bad credit.
The effect of variable rate on short-term estimates have a coefficient value of 0.0000944 means that any increase in inflation by 1% would cause an increase in NPLs by 0.0000944%, Furthermore, rate probability value> critical limit (0.3395> 0.05) so that it can be concluded in the short-term variable rate has no effect on the NPL. or H0be accepted. The variable rate on a long-term estimate has a coefficient of 0.000370 means that any increase in inflation by 1% would cause an increase in NPLs0.00037%, Furthermore, rate probability value <critical limit (0.0000 <0.05) so that it can be concluded variable rate long-term significant effect on the NPL. or H1be accepted. The results of this study supportShingjergji (2013); Tanasković & Jandric (2015).
The exchange rate in the short term does not burden the public and the business community in carrying out activities using foreign exchange transactions. But, in the long term depreciation of the rupiah in the long term can shake the world of business because the majority of the raw materials are imported so the more burdensome the business world in paying off debt obligations and may increase credit risk. GDP variable effect on short-term estimates have a coefficient of-0.000000608 meaning that any increase GDP by 1% would cause a decrease in NPLs in Indonesia amounted to 0.000000608%. Furthermore, GDP probability value> critical limit (0.4308> 0.05)It can be concluded GDP variable in the short term has no effect on NPLs in Indonesia. or H0be accepted. Variable GDP in the longterm estimates have a coefficient of-0.0000024 meaning that any increase GDP by 1% would cause a decrease in NPLs in Indonesia amounted to 0.0000024%. Furthermore, GDP probability value <critical limit (0.0000 <0.05)It can be concluded variables of GDP in the long term significant negative effect on NPLs in Indonesia. or H1 received, is in line with research Ekanayake et al. If there descend revenue in the short term has little impact on financing problems. So in the short term revenue is not significant in influencing credit risk. But, in the long termdecrease in public income may increase credit risk, this is in accordance with the theory of the business cycle in which each occurred penuruanan business cycle will have an impact on the economy and may increase credit risk.
Financial Digitization policy variables influence on short-term estimates have a coefficient of-0.206543 negative sign, Furthermore, Dummy probability value> critical limit (0.5189> 0.05) It can be concluded financial digitization policy variable in the short term has no effect on NPLs in Indonesia or H0be accepted. Financial Digitization policy variables on the estimation of the long-term has a coefficient of-0.222347 are negative with probability> critical limit (0.555> 0.05). It can be concluded financial digitization policy variable in the long run no effect on NPLs in Indonesia. or H0 be accepted.
Digitization finance in the short term and the long term due to less influence is not yet even technology infrastructure, communication and information are still concentrated in large cities, the low abiding by the public related to the use of mobile technology, improvement of governance data base. (Bank Indonesia, 2015) CONCLUSION Based on the results of research and discussion, it can be seen partially in long-term interest rates and exchange rates positive and significant impact on non-performing loans, GDP is negative and significant effect on nonperforming loans, while inflation and the policy of digitizing finance has no effect on non-performing loans , In the short-term interest rates and a significant negative effect on non-performing loans, while the exchange rate, inflation, GDP, and the policy has no effect on non-performing loans.
Increasing economic growth and push the economy higher quality of government should be to encourage the business community in order to improve the quality of the output through serngkaian policies such as control of competitive interest rates for the business world, inflation control, and control of the exchange rate and continue to encourage the expansion of exports to countries potential such as Africa, Latin America and Eastern European countries. In an effort to improve access to government financial institutions need to improve the digital infrastructure into the countryside so that the wider community can obtain financial services through digitization.