본 논문은 가계신용 증가율과 장기금리에 대하여 한국은행을 통하여 제공된 지수를 이용하여 상호간의 정보전달 메커니즘을 규명하는데 있다. 표본자료는 1995년 3분기부터 2011년 3분기까지 17년간 분기별 자료를 이용하여 동태적 분석방법인 VAR 모형을 이용하여 그랜즈 인과관계 분석(Granger Causality test)과 충격반응분석(Impulse Response Function) 및 분산분해(Variance Decomposition)를 실시하였다. 주요 분석결과는 다음과 같다. 첫째, 그랜즈 인과관계 분석결과 장기금리는 가계신용금액에 선행하여 예측력이 있어 금리변동에 따라 가계신용금액이 변동함을 알 수 있었다. 그러나 가계신용금액은 장기금리 중에서 회사채(AA-, 3년)에 한해서만 선행하여 예측력을 지니고 있는 것으로 나타났다. 둘째, 충격반응함수의 분석결과 장기금리가 가계신용 증가율에 영향을 미치다가 일정시점이후에 충격이 사라져, 금리가 상승할 경우 가계신용금액이 감소하며, 금리가 하락할 경우에는 가계신용금액이 증가하는 것으로 추론할 수 있다. 마지막으로 분산분해 분석 결과 가계신용금액증가율은 장기금리에 의하여 일정부분 영향을 받는 것으로 나타났으나, 장기금리인 국민주택 1종금리(5년), 국고채권금리(3년, 5년), 회사채(AA-, 3년)금리는 자기 자신의 오차에 의해 거의 영향을 받는 것으로 나타났다. 본 연구는 기존 단기금리와 단기대출간의 상호연관성에 관한 연구를 확대하여 가계신용금액과 장기금리와의 상호연관성을 분석하였다는 점에서 가계신용금액과 장기금리와의 정보이전 메커니즘을 파악하였을 뿐만 아니라 은행들의 대출정책 수립에 기여할 수 있을 것으로 판단된다. 즉 금리상승은 대출원리금 상환을 지연 및 연체시킬 뿐만 아니라, 신규 가계대출을 감소시켜 은행의 건전성 유지에 영향을 미칠 수 있기 때문에 금리변동에 따라 은행의 예금 및 대출정책을 탄력적으로 조정하는 정책이 필요하다고 사료된다. 따라서 금리가 상승할 경우에는 은행대출 감소에 따른 은행수익 감소를 다른 수익을 통하여 상쇄시켜야 하며, 금리가 하락할 경우에는 가계대출 증가에 따른 가계대출자들의 신용상태를 재평가할 수 있는 시스템이 필요할 것이다
This dissertation is to verify the efficiency (correlation) of household credit and long-term interest rates by revealing mechanism of their interrelationships from related indices provided by Bank of Korea. Both long-term interest rates and household credit would not have any effect on each other if they had perfect efficiency. However, if there is inefficiency between two markets, they have an effect on each other, and the size of household credit may differ according to changes in long-term interest rates. Sample data are based on the amount of household credit for 65 samples in 17 years from 1995 3Q to 2011 3Q, and its increase rates are calculated in quarterly basis. Granger causality test, impulse response function and variance decomposition are computed with VaR (dynamic analysis method) using yield rates of housing bond 1 (5 years), Government bond (3 years), Government bond (5 years) and corporate bond (AA-, 3 years) which periods same as that of household credit.
As a result of analysis on basic statistics subject to household credit increase rates and long-term rates, average increased rates of household credit amounts is 2.99% in quarterly basis during the analyzed periods. Average yield rate of housing bond 1 (5 years) and Government bond (3 years) are 7.04% and 6.75% respectively and that of both Government bond (5 years) and corporate bond (AA-, 3 years) is shown as 7.01%. Standard deviation which shows volatility is figured out as 2.91% for housing bond 1 (5 years), 3.35% for Government bond (3 years) and 3.22% for both Government bond (5 years) and corporate bond (AA-, 3 years).
To verify stability of time series model for household credit increase rates and long-term interest rates, a unit roottest was taken. As a result, the null hypothesis is accepted as “there exists a unit root in case of standard variables.” However, it is dismissed in case of defference variables. Thus, analyses have used a VaR model which does not consider errors. Main analyses results are as follows;First, Granger causality test results showed that the long-term interest rate leads household credit amount. In other words, the analysis demonstrated that the F statistical amount values were 3.38, 2.87, 2741 and 6.96 each, which were dismissed from the significant level. Because long-term interest rate led the household credit amount, it is possible to know that the household credit amount changed according to the change in interest rate when it comes to forecasting ability. However, household credit amount did not have forecasting ability, leading some form of long-term interest rate. In other words, it is possible to know that the household credit amount changed according to the change in interest rate since there is forecasting ability by leading the household credit amount when it comes to national housing type 1’s interest rate(five years), treasury bond’s interest rate (three years, five years) and corporate bond’s (AA-, three years)interest rate, which are long-term interest rates. However, it was shown that the household credit amount does not have forecasting ability, leading certain long-term interest rate.
Second, the results of the impulse response function analysis demonstrated that the household credit growth rate exerted instant effect on the long-term interest rate while the shock disappeared at the time difference 3 and time difference 5. Moreover, household credit amount continued to exert negative (-) effect. Thus, household credit decreased when long-term interest rate increased. It is possible to deduce that the household credit increased when the long-term interest rate decreased. In other words, national housing type 1’s interest rate(five years), treasury bond’s interest rate (three years, five years) and corporate bond’s (AA-, three years)interest rate, which are long-term interest rates continued to exert negative (-) effect on the household credit amount. Thus, household credit amount decreased when the interest rate increased. When interest rate decreased, household credit amount increased.
Lastly, analysis on the variance decomposition demonstrated that the household credit amount growth rate is partially affected by long-term interest rate. However, national housing type 1’s interest rate (five years), treasury bond’s interest rate (three years, five years) and corporate bond’s (AA-, three years)interest rate, which are long-term interest rates appeared to be affected only by their own margin of error. In conclusion, national housing type 1’s (five years)interest rate, treasury bond’s (three years and five years)interest rate, corporate bond’s (AA-, three years)interest rate, which are long-term interest rates, have forecasting ability, leading the household credit amount. Moreover, it is possible to know that the household credit amount changes according to the change in interest rate. Accordingly, given that the long-term interest rate is an important factor that determines household loan and credit based sales, these research results are expected to contribute to the development of all types of loan policies and credit policies by the banking and non-banking industries. This research can be considered meaningful in the sense that the lead/lag effect with the household credit and long-term interest rate were analyzed.