검색트렌드를 활용한 통화정책 환경 불확실성지수 구축

Abstract

글로벌 금융위기 이후 최근의 코로나19 팬데믹에 이르기까지 불확실성이 일상화되면서 통화정책결정에 미치는 불확실성의 영향이 증대되고 있다. 본 연구는 이에 주목하여 한국은행의 주된정책 수행상황 보고서인 󰡔통화신용정책보고서󰡕를 기초로 통화당국이 인식하는 정책환경불확실성의 지수화를 시도하였다. 󰡔통화신용정책보고서󰡕에서 불확실성 관련 어휘를 선정하고이들 어휘에 대한 온라인 상에서의 검색빈도를 ｢구글 트렌드｣에 의해 측정하는 방식으로 작성된불확실성지수는 기존 불확실성지표와의 교차상관계수 비교, 산업생산과의 인과관계 검정 등을통해 그 유용성을 확인할 수 있었다. 거시경제변수와의 관계에서는 충격반응함수 추정 결과불확실성 충격에 산업생산 증가율이 급락 후 급반등하면서 원래 수준을 회복하는 움직임을 보인반면 콜금리는 불확실성 충격이 해소된 이후에도 상당 폭 하락하여 불확실성이 저금리기조 장기화의주된 요인으로 작용하고 있음을 보여주었다. As COVID-19 spread worldwide last year, uncertainties have increased significantly with the duration of infectious diseases, the development of vaccines and treatments, and changes in consumption patterns. Since the global financial crisis, uncertainty has already been a major consideration when central banks in many countries make policy decisions. This study tries to construct an index of the uncertainty of the monetary policy environment from the policy authority’s point of view and examines the usefulness of the index as a policy decision factor. Prior to this study, various ways have been carried out to find indicators to measure economic uncertainty. Before the global financial crisis, volatility of financial market indicators was used as an uncertainty measure in many cases. After the crisis, with high interests in uncertainty, more creative ways of developing the index that can measure the uncertainty felt by economic agents, such as big data, were proposed. Referring to Castelnuovo and Tran(2017), among the recent studies dealing with big data, this study constructs an uncertainty index for the Korean economy using Google Trends. This study attempts to differentiate from the existing domestic uncertainty indicators in the sense that the uncertainty is measured from the policy authority’s point of view. With reference to Evans et al.(2015), which uses FOMC minutes to measure the uncertainty evaluated by the members of the FOMC, we select search words from the Monetary Policy Report, a policy implementation report issued by the Bank of Korea and use Google Trends(GT) with those search words to construct the uncertainty index. Concretely, 65 search words were selected from 41 Monetary Policy Reports issued from the first half of 2000 to 2020. Excluding six words, of which the GT values were near 0, 59 words were used to construct the Total Index, the main uncertainty index. And the General, Endogenous, Domestic Exogenous, and Foreign Exogenous Indices were constructed as sub-indices of uncertainty. The General Index measures monetary policy uncertainty, and the Endogenous Index measures uncertainties with models and parameters. The Exogenous Index reflects exogenous shocks to the economy. According to the Granger causality test, GT indices, including the Total Index and the sub-indices unilaterally Granger cause industrial production. In the case of the call rate, the Total, General, and Endogenous uncertainty indices unilaterally Granger cause the call rate, but the Domestic and Foreign Exogenous Indices’ causalities with the call rate are unclear. This is because the call rate as the policy rate responds mainly to the endogenous variables. The GT index for the uncertainty of the policy environment perceived by the central bank constructed in this study reflects recent academic trends. After the global financial crisis, the focus of research has moved from the analysis of the relationship between uncertainty and the economy to the development of uncertainty indicators. The objective of the uncertainty index has also shifted from the wide range of macroeconomic uncertainty to the specific one, such as the monetary policy environment. The basis of measurement of uncertainty is also moving from financial markets to online web and SNS. The precedence supports the usefulness of the GT uncertainty index by the cross-correlation and the exogeneity by the Granger causality test. The Total Index developed in this study precedes the corporate bond spread, the financial stress index for Korea by the Asian Development Bank, and the economic and business sentiment indicators surveyed by the Bank of Korea. Also, the fact that the Total Index unilaterally Granger causes industrial production unlike other existing uncertainty indicators except the corporate bond spread means that the index satisfies the real option effect and unpredictability requirements. The GT uncertainty index confirms that uncertainty is one of the main factors in determining the policy rate by the Bank of Korea. The forecast error variance decomposition shows that a quarter of the variance of the call rate is due to uncertainty. The impulse response function also indicates that uncertainty is the cause of the prolonged low policy rate in recent years.