2000년대 이후 전 세계적으로 저금리 기조가 유지됨에 따라 기간프리미엄에 대한 관심이 높아졌음에도 불구하고, 국내에서 기간프리미엄의 정보력에 대한 연구는 매우 제한되어 있었다. 본 연구는 Adrian, Crump, and Moench(2013)의 방법론을 활용해 채권시장의 기간스프레드를 기간프리미엄과 기대단기이자율로 분해한 후, 산업생산지수 변동률 및 주식수익률에 대한 예측 능력을 검증하였다. 이후, 소규모 개방경제로 대외경제에 대한 높은 민감도를 가진 국내 경제의 특성을 반영하여, 미국의 기간스프레드 및 기간프리미엄의 영향력에 대해 분석하였다. 연구의 결과는 다음과 같다. 첫째, 10년물과 1년물 국채수익률의 차로 정의된 기간스프레드는 산업생산지수의 3개월 이상 누적변화율을 유의하게 예측하고 있었으며, 기간스프레드의 예측력은 기간프리미엄에 의한 것으로 나타났다. 둘째, 기간스프레드 및 기간프리미엄은 매우 짧은 미래(1개월 이상)의 KOSPI 수익률을 예측하고 있음을 확인하였다. 기간스프레드를 분해한 예측력 분석에서는 저금리 기간동안 기간프리미엄이 중요한 수익률 예측요인임을 확인하였다. 마지막으로, 미국의 금리정보를 포함한 분석에서는 미국의 기간스프레드가 단독으로 국내 경기에 대한 예측력을 보유하지 않았지만, 국내시장의 기간스프레드를 함께 설명변수로 추가했을 때, 경기예측력이 급격히 증가하는 것을 확인할 수 있었다. 미국의 금리스프레드를 분석한 연구에서 국내 기간프리미엄의 예측력이 주로 미국의 미래 기대단기이자율에 의해 결정되고 있음을 확인하였다.

Prior to the financial crisis, it is known that the term premium reflects uncertainty about future short term interest rates and does not retain significant information about future economic conditions (Hamilton and Kim, 2002; Rudebusch et al., 2007). Rather, the term premium is likely to distort the informativeness of the term spreads regarding the predictability. Consistently, extant literature has shown that the expectation for future short term interest rates, which is extracted from term spreads by excluding term premium, is more informative for predicting future economic conditions. However, as the low interest rate period began after the 2007 financial crisis, some empirical facts have provided new concerns about the forecasting power of term premium. In a situation where expectations for future short-term interest rates are extremely low, the long-term interest rate movement is largely determined by the term premium, thereby raising a question as to whether the forecasting power of the term premiums is still invalid. Many studies, including Dudley (2006), have reported that the term spreads have become less predictable than in the past. This argument results from the fact that the expectation component of the pre-crisis short-term interest rates did not drop significantly, even though the expectation component has been believed to be related to economic forecasting. Under this circumstance, this paper uses the methodology of Adrian et al. (2013, ACM method) to break down the term spreads into risk-neutral expectation component and term premium component, and verifies their predictability on the growth rate of industrial production indices and equity index returns of Korea. In addition, we analyze the impact of the U.S. term spread components on the Korean economy, which is a small open economy with high vulnerability to external shocks. The results of the study can be summarized as follows. First, the term spreads defined as the difference between 10-year spot rate and 1-year spot rate have significant explanatory power for predicting future industrial production index for 3 months or longer horizons. However, the source of predictability is based on the term premium, not the expectation for future short-term interest rates, which is in contrast to the previous studies that had been conducted before the recent low-interest rate period. Second, the KOSPI returns can predict future economic conditions measured by the growth rate of industrial economic index for horizons of 3 months to 9 months, which is inconsistent to the results for the U.S. in that the forecasting horizon of the stock index is much shorter than that of term spreads in the U.S. As a result, the term premiums as well as term spreads predict the future KOSPI returns for very short horizon of 1 month. Third, the unsecured call rate, which can represent the monetary policy stance of a central bank, does not predict future economic condition as well as KOSPI returns. Finally, the U.S. term spreads cannot solely predict the changes in industrial economic index and the stock index returns of Korea. However, when the term spread components of Korea are added together as an explanatory variable, the predictable power of the models with components of both countries sharply increases. In particular, the expectation for short rates alone is not significant, but the inclusion of the U.S. expectation for short rates improves the predictability significantly. These can be interpreted as the result that the future interest rate changes (volatility) of Korea are significantly affected by the expected short-term interest rates of the U.S. This study contributes to the literature in the following aspects. First, this study is the first academic attempt in Korea that use the decomposition method of Adrian et al. (2013). Adrian et al. (2013) is recognized as a standard by the most central banks, including the FRB, and by major financial institutions. Currently, FRB, New York uses the ACM model to disclose the behavior of term premiums on daily basis and the central banks of major countries are also actively analyzing the term premium using this model. Second, recently, the information content of the term premium during the period of low interest rates we use in this study is receiving much attention, as shown by the terminology called Greenspan’s Conundrum. In fact, many recent studies have reported that changes in the U.S. long term interest rates are determined by the term premium, not by expectation for short term rates. Thus, this study is the first study that examines the predictability of the term premium in the low interest rate periods of Korea.