Causes of Long Memory Property in the Highof Asia-Pacific Stock Markets

• 발행기관: 한국자료분석학회
• 저자명: 강상훈(Sang Hoon Kang) Ron Mclver
• 간행물 정보: 『Journal of The Korean Data Analysis Society (JKDAS)』Vol.18 No.6, 2925~2935쪽, 전체 11쪽
• 주제분류: 자연과학 > 통계학
• 파일형태: PDF
• 발행일자: 2016.12.30

영문 초록

This paper has examined the origins of the long memory volatility property using the high frequency data of eight Asian-Pacific stock market returns. We use various long memory models, namely non-parametric (classical and modified rescaled range (R/S) analysis) and semi-parametric (Geweke, Porter-Hudak (GPH) and Local Whittle (LW)) tests, over various time scale intraday returns, such as 10-min, 15-min and 30-min. We provide three important implications in this study. First, we find the persistence in the autocorrelations of 10-min intraday volatility. Second, both non-parametric R/S analyses show absence (presence) of long memory in the intraday returns (volatility). Third, using two semi-parametric long memory tests (GPH and LW), the estimates of long memory parameter (d) are invariant to temporally aggregated intraday volatility returns, implying that a long memory phenomenon is an inherent characteristic of the data generating process, not a result of structural breaks.

목차

1. Introduction

2. Empirical methodology

3. Data

4. Estimation results

5. Conclusions

References

키워드

long memory high frequency data time-scale aggregation Lo’ modified R/S GPH Local Whittle test