‘Turn-of the-month’ return effects for small cap Hong Kong stocks

Abstract

Intra-month returns for a liquid pool of small-cap stocks in Hong Kong are analysed over the period January 2000 to June 2005. A strong and persistent ‘turn-of-the-month’ effect is apparent where returns, measured between the close of trading on the penultimate day of business in a calendar month and the subsequent close five business days later, are recorded at levels averaging almost 1.5%. This compares to returns over the preceding five-day period–capturing the ‘end of month’ effect–of almost −0.3% on average. Further analysis also points to differential ‘day-of-the-week’ effects with Tuesday and Friday returns in particular at noticeably lower levels during the final week of a month as compared to earlier weeks. The Tuesday effect may well be an analogue of the US Monday effect given Wang et al .'s evidence of notably lower Monday returns toward month-ends, and the suggestion of a ‘delayed Monday effect’ in Asian time zones due to the opening of US Monday trading hours after Hong Kong trading hours.

Notes

¹ See Ariel (1988, p. 111 for details). Significantly, Ariel reported larger mean returns for the first part of the calendar month for the whole 1963–1981 time frame, as well as for the 1963–1966, 1967–1971 and 1977–1981 subperiods. Returns for the nine-day period preceding the month-end run to the last but one business day of the month in Ariel.

² For further evidence of ‘turn-of-the-month’ effects, see Lakonishok and Smidt's analysis of US returns between 1897 and 1986.

³ See Wong (1995, p. 288) for a summary of such results. Note that returns for the ‘first half’ of a month in Wong (1985) run from the last trading day of a month to the eighth day thereafter with returns over a preceding nine-day period capturing what Wong describes as the ‘last half’ of the month.

⁴ Thanks are also due to Lilian Lam and S.L. Wong for answering various questions relating to the data.

⁵ Details of the selection criteria for the Hang Seng Index are contained in a downloadable brochure under ‘Criteria’ for the HSI at http://www.hsi.com.hk/.

⁶ As is apparent from the Hang Seng Composite Index Series brochure (p. 1) at ‘Criteria’ (at http://www.hsi.com.hk/).

⁷ See Hang Seng Composite Index Series (op. cit.) brochure (p. 1) at ‘Criteria’ (at http://www.hsi.com.hk/).

⁸ For the 65 observations, mean cumulative returns on the HSI and HSSCI, for the period from the close of trading on the penultimate day of a calendar month to the close of the ninth business day of the ensuing month, were as reported in Table 2 at levels of 0.743% and 1.212% respectively. For the 10-day period prior to this, the mean cumulative returns were, respectively,–0.712% and–0.688%.

⁹ Note that as the five-day return is a close-to-close return, consideration of the whole trading day (with regard to NONTRDAYS) means that 6 days are considered: m, t − 2; m, t − 1; m, t + 1; m + 1, t + 2; m + 1, t + 3; and m + 1, t + 4. Values evident for NONTRDAYS were 0, 1 and 2 across the 65 return measures for RetHSSCI_{t −2 to t +4}.

¹⁰ Ho (1990, pp. 62–63) reports a Hong Kong April return subordinate in size to the January return for the shorter January 1975 to November 1987 period.

¹¹ Rosenberg (2004), however, has recently presented evidence using US data pointing to ‘… an end-of-month effect distinct from … Mondays …’(p. 1). Rosenberg relates the phenomenon to business cycles.