This study examines the effects of the initial margin requirements on the stock market by using daily indexes of the U.S. and some Pacific-Basin countries. Do margin requirements curb speculative activity and reduce stock market volatility? This question has recently generated considerable controversy among academics and regulators following the market crash, 1987. The recent experiences of Japan, Korea, and Taiwan with margin requirements provide significant additional information about the contemporary impact of margin policy.
Chapter II deals with the related questions about the short-term relations of stock market response to margin change: (1) whether or not stock price levels change when margins change and (2) whether or not the volatility of stock return change when margins change. To investigate the relations between margin requirements and market behavior, we use the switching regression and modified Levene statistic. Bootstrap simulation is used to resolve the small-sample property. We find no consistent evidence that margin requirements affect market in the short-term. Stock price levels in Taiwan, however, seem to respond mildly to margin changes. We also show that there exists weak evidence of a liquidity effect in the relation between margin requirements and market volatility.
The third chapter provides another evidence about the relation between margin requirements and stock market volatility. To find the rational measure of volatility, we consider three different proxies of volatility-rolling measure, daily volatility, and conditional measure. The multiple regression and vector auto-regressive (VAR) model are analyzed to investigate the long-term relation between margin requirements and volatility. For the short-term relation between margins and volatility, only margin decreases in Japan are found to have a significant relation. We show that the effect of margins on volatility is dominated by liquidity effect rather than speculative effect. We find no convincing evidence that margins affect volatility in the long-term and that volatility Granger-causes margins.
The last chapter in this study, in addition, deals with an interesting issue in security pricing problem - robustness of goodness-of-fit criteria for factor identification of Arbitrage Pricing Theory (APT). A recent issue in empirical tests of APT has been factor identification. Several financial researchers have cautioned against using the formal likelihood ratio (LR) statistic in financial application, because several violations of the assumptions of the standard factor model typically occur with actual return data. Using the simulated returns generated under the violations of the assumptions of the factor model, we compare goodness-of-fit criteria based on the maximum likelihood analysis by changing the number of factors, the number of securities, and the number of observations. For a variety of simulation designs, schwartz's baysean criterion is robust to all the violations of the assumptions in the factor model. The cross-validation performs well for the small number of securities, but as the number of securities increases, its robustness is limited to only a large sample. It is found that by actual returns in the Korean stock market, schwartz's baysean criterion identifies three factors.
본 연구는 미국 및 환태평양 국가들의 시장지수를 사용하여 신용거래제도가 증권시장에 대하여 미치는 영향을 고찰한다. 신용거래보증금의 변경이 증권시장에 있어서 투기적 활동을 통제하며 분산성(Volatility)을 축소시킬 수 있는가? 이러한 의문은 1987년 10월에 전세계적으로 발생한 월요일의 대폭락(Black Monday)이후에 학계와 실무계에 상당한 관심을 불러 일으켰다. 따라서 환태평양국가들 특히 일본, 대만, 그리고 한국의 신용거래와 관련된 실증적 연구는 신용정책이 미치는 영향에 대하여 중요한 의미를 제공한다.
제 2장에서는 신용제도가 증권시장에 미치는 단기적인 효과에 대하여 주로 다루고 있다. 신용거래제도와 시장행태간의 관계를 파악하기 위해, 우리는 스위칭(Switching) 회귀분석과 수정레베느(Modified Levene)통계량을 사용하며 소표본문제를 해결하기 위해 부트스트랩(Bootstrap) 시뮬레이션이 행해진다. 우리는 신용거래제도가 시장에 단기적으로 영향을 미친다라는 단정적인 증거를 발견할 수가 없으며, 정부가 규제할려는 투기적 효과보다는 차라리 역규제효과인 유동성효과가 나타나고 있다는 것을 발견한다.
제 3장에서는 관심의 대상을 분산성에 국한시켜서 신용거래와 분산성간의 관계를 장.단기로 나누어서 자세히 분석한다. 먼저, 분산성의 합리적인 척도를 발견하기 위해, 롤링(Rolling) 표준편차, 일별(Daily) 분산성, 및 조건부(Conditional) 척도가 고려되어지며, 장기적관계를 파악하기 위해 다중회귀분석 및 벡터자기회귀(VAR)모델이 사용된다. 신용제도와 분산성간의 단기적인 관계에 대해, 오로지 일본의 신용거래보증금이 감소했을 경우만이 통계적으로 유의한 관계를 가진다는 것이 발견된다. 또한 신용거래보증금의 변경이 장기적으로 분산성에 영향을 미친다는 단정적인 증거를 발견할 수가 없다.
