ETF 괴리율 결정요인은 유형별로 상이한가?: 국내 ETF 시장의 실증 분석 (Determinants of ETF Differentials by Type)

서지정보

· 발행기관 : 한국금융학회
· 수록지 정보 : 금융연구 / 32권 / 4호 / 151 ~ 178페이지
· 저자명 : 김세완, 김영민, 김경록

초록

본 연구는 우리나라 ETF 시장에서 유형별로 괴리율에 미치는 요인을 GMM 모형을 이용하여실증분석하였다. 유형으로는 기초자산, 투자지역, 투자전략 등이 고려되었다. 실증분석 대상은2017년 9월 기준 3년 이상 설정된 국내 ETF 중 순자산 규모가 가장 큰 19개 종목이다. 괴리율발생은 비효율적 시장에서의 마찰적 현상이라는 인식 하에, 괴리율 결정요인으로 시장에서의거래에 마찰(friction)을 일으키는 변수를 최대한 수용하고 기존의 연구결과를 참고하여 대체투자수단 여부 및 기관투자자 거래비중 등을 괴리율의 설명변수로 사용하였다. 유형 전체를 포함한분석에서는 전기 괴리율, 거래대금, 호가스프레드율, 대체투자수단, 기관투자자 거래비중, 가격변동성 등이 유의한 영향을 주었다. 유형별로 나누어 추정한 결과에서는 국내주식형의 괴리율에가격변동성을 제외한 모든 변수가 유의한 영향을 준 반면 국내채권형 괴리율에는 전기 괴리율, 거래대금, 호가스프레드율만이 영향을 미치는 것으로 나타났다. 해외주식형 괴리율에는 호가스프레드율을 제외한 모든 변수들이 유의한 영향을 주었다. 섹터주식형 괴리율에는 전기 괴리율, 기관투자자 거래비중, 가격변동성이 유의한 영향을 주었다. 본 연구는 전기 괴리율이 익일에도해소되지 않는 시계열의 관성효과(momentum effect)를 발견하고, 유형별로도 ETF 괴리율의결정요인이 상이함을 밝혔다. 최근 ETF의 종류가 다양해지면서 시장의 활성화를 위하여 괴리율을줄이는 것이 중요한 과제다. 본 논문의 연구결과는 기관투자자의 거래 활성화 및 LP 역할의확대 등이 필요함을 시사한다.

영어초록

Nowadays ETF (Exchange Traded Fund) is one of the most important asset management financial products in Korea. In fact, ETF of trading volume account for 17% of the total stock market as of July of 2017 from 1.1% in 2002. The market size also reaches at 31.9 trillion won from 0.3 trillion won during the same period. In particular, the number of ETF has grown up to 315 from 4. Due to the fast growth of the ETF market, institutional investors utilize the ETF products: for example, GEPS (Government Employees Pension Service) invest more than 200 billion won in EMP (ETF Managed Portfolio).
Therefore, the differentials of ETF, the difference between price and net asset value of ETF, is very important topic to be discussed. There are not many previous studies about differentials of ETF yet.
This study aims to investigate the determinants of ETF differentials by type based on underlying assets, investment region, and investment strategy. For the study, we employ the largest assets of 19 ETFs which have been listed more than three years as of August of 2017. The independent variables include previous differentials, trading volume, spread, dummy of alternative investment vehicles (whether the underlying index has futures and options or not), institutional investors’ trading ratio and price volatility of ETF. In particular, we choose two new independent variables: alternative investment vehicles as dummy and institutional investors’ trading ratio.
Alternative investment vehicles such as futures and options might contribute to lessen the differentials through arbitrage. And institutional investors have better information and their trading volume is bigger than individuals, so their trading ratio could lessen the differentials. Moreover, GMM (Generalized Method of Moments) has been used considering the endogeneity among independent variables.
The hypotheses to be addressed are as follows. Hypothesis 1 : Bond ETFs do not have alternative investment vehicles such as futures and options and institutional investors’ trading volume for Bond ETFs is relatively lower than Stock ETFs. Therefore, determinants of differentials might be different by underlying assets.
Hypothesis 2 : Investors have relatively less information about overseas stocks compared to domestic. Also, there is time difference between home country and overseas. Therefore, the affecting factors for the differentials might be different by investment region.
Hypothesis 3 : Sector ETFs has relatively less alternative investment vehicles and trading volume. Therefore, the differentials of determinants might be different by investment strategy.
Before the estimation by type, at first, we estimate the total. We find that all of independent variables affect the differentials: previous differentials, trading volume, spread, dummy for alternative investment vehicle, institutional investors’ trading ratio, and price volatility. Based on this result, we estimate by type and the results are as follows. First of all, the differential of domestic stock ETFs is affected by all variables except price volatility of ETF while that of domestic bond ETFs is affected only by previous differential, trading volume and spread. Second, the differential of overseas equity ETFs is affected by all variables except spread.
Third, the differential of sector ETFs is affected by previous differentials, institutional investors’ trading ratio, and price volatility. In summary, we find that the differentials of ETFs is differently affected by type.
The estimation results reveal the policy implications. First of all, since trading volume and institutional investors’ trading ratio affect negatively to the differentials, we need to develop various portfolio strategy for especially institutional investors. Moreover, since Kim et al. (2014) shows that individuals herd institutions, more institutional investors’ trading is needed to activate the market.
Second, the role of Liquidity Provider (LP) should be expanded. We find there are momentum effect in the differentials of ETFs. However, the momentum effect with high institutional investors’ trading ratio is relatively less. Therefore, LP should be more actively involved to lessen differentials.
Third, providing more information about ETFs might be activated the market. For example, ETFdb.com in US provide various information on ETF such as liquidity, performance, volatility. Fund evaluation company, asset evaluation company in Korea could play this role.

참고자료

· 없음