Economics Letters North-Holland
353
26 (1988) 353-356
SETTLEMENT
SYSTEM AND THE DAY-OF-THE-WEEK
EFFECT
Tom K. LEE National University of Singapore, Singapore 0511 Received
16 December
In this paper, stock returns.
1987
I empirically
show that a settlement
system
matters
importantly
in determining
the day-of-the-week
effect of
1. Introduction Since the birth of the Stock Exchange of Singapore (SES) on June 4, 1973, the settlement system of the exchange has undergone various changes. On January 1, 1975, a settlement system was introduced. All contracts for ready delivery were to be completed within fourteen days which included Sundays and holidays. The important aspect of this form of settlement system was that it made no difference when a buyer purchased a stock in terms of the opportunity cost of funds. Irrespective of which day of the week he purchased the stock, there were fourteen days grace for him to meet payments. One should expect stock returns in different days of the week to be the same under this settlement system. After the sharp downturn of stock prices of the SES in mid-1981, a temporary settlement system was introduced on September 4, 1981. All contracts were to be completed on an immediate delivery basis. As the stock market returned to normal trading, the ruling on immediate delivery was relaxed on June 1, 1982. A new settlement system was introduced. All contracts for ready delivery done between Monday to Friday were to be completed on any trading day not later than 4 p.m. on Tuesday of the following week. The important aspect of this form of settlement system was that it made a difference when a buyer purchased a stock in terms of the opportunity cost of funds. For example, if contracts were transacted on Monday (Tuesday, Wednesday, Thursday) at 4 p.m., there would be eight (seven, six, five, respectively) calendar days grace before payments of stock purchase had to be made, but if contracts were transacted on Friday at 4 p.m., contract notice would not be issued until Monday of the next week, and then there would be nine calendar days grace before payments of stock purchase had to made. Hence, one should expect the rate of return on stock purchases to be lowest on Mondays, to increase progressively over the days of the week and to be the highest on Fridays under this system of settlement. From December 2 to December 4,1985, trading of shares was suspended for three days due to the sudden collapse of a large conglomerate firm listed in the exchange. From December 5, 1986, trading resumed on an immediate delivery basis. On January 6, 1986, normal trading was resumed and the previous settlement system was restored. Contracts for ready delivery transacted on any day of the week were due by 4 p.m. Tuesday of the following week. 0165-1765/88/$3.50
0 1988, Elsevier Science Publishers
B.V. (North-Holland)
T.K. Lee / Settlement
354
system, day-of-the-week
effeci
In light of the above account of the history of the settlement system of the SES, it seems ideal to study the effect of a settlement system on the day-of-the-week effect of stock returns in Singapore. In particular, I define two time periods: the first time period is from January 13, 1978 to June 30, 1981, and the second time period is from June 1, 1982 to November 18, 1985. Each time period consists of 833 daily observations of stock returns with returns on the days immediately after holidays excluded. If the settlement system is important in determining the day-of-the-week effect, one should expect to find no day-of-the-week effect in the first period but a day-of-the-week effect in the second period. The rest of the paper is organized as follows. Section 2 describes the data and their sources. Section 3 reports the statistical analysis and results and draws the conclusion of the paper.
2. Data On January 1, 1984, a new series of SES Share Indices were introduced. They are the SES All-Shares, SES All-Commercial and Industrial, SES All-Finance, SES All-Hotel, SES All-Property, SES All-Plantation and SES All-Mining Indices. These new series of stock price indices are based on 100 percent coverage of the common stock listed under different industrial sectors of SES and are value-weighted. They are computed using the daily closing prices of the respective stocks traded in the exchange. The new series are adjusted for the changes in capitalization such as right issues, bonus issues and new share issues. They are backdated to January of 1975. The stock price index data used in this study are obtained from three sources:
(i> Data from January
2, 1975 to October 31, 1983 are obtained from Saw and Tan (1983). Data for the last two months, November and December, of 1983 are obtained from the SES library. from the Singapore Srock Exchange (iii) Data from January 2, 1984 to the present are obtained Journal, a monthly official publication of SES. (ii)
3. Statistical analysis and conclusions To study the day-of-the-week
effect, I specify the following
regression
model:
n
R,=a+
c
b,D,,+
u,,
(1)
j=2
where = one day stock return on data t; = 1 if 1 is on the jth day of the week, = 0 otherwise: and ut is an error term. The interpretation of the coefficients of the regression equation is as follows: a stands for the average rate of return on Mondays. b, stands for the difference between returns on Mondays and the jth day of the week. Hence a statistically significant b coefficient indicates the presence of a day-of-the-week effect. R* %
T.K. Lee / Settlement
system, day-of-the-week
Table 1 Estimates of variation in trading day to trading day daily (%) rates of return 13, 1978 to June 30, 1981 (833 observations, with t-statistics in parentheses).
Tuesday
Wednesday
Thursday
All share
Industrial and commercial
Finance
Hotel
- 0.081 ( - 0.942)
- 0.705 ( - 0.749)
-0.196 (- 1.396)
-0.182 (- 1.139)
0.027 (0.317)
0.088 (0.944)
-0.148 (- 1.059)
- 0.001 ( - 0.013)
- 0.040 (-0.463)
355
effect
on the SES stock indices over the period
Property
January
Plantation
Mining
0.004 (0.029)
- 0.059 ( - 0.544)
- 0.021 (-0.190)
0.024 (0.150)
0.038 (0.291)
- 0.029 ( - 0.271)
0.160 (1.489)
- 0.231 (- 1.655)
0.008 (0.051)
- 0.077 (- 0.600)
- 0.003 ( - 0.030)
0.103 (0.958)
Friday
0.065 (0.762)
0.101 (1.077)
- 0.0632 (-0.449)
0.103 (0.645)
0.016 (0.125)
0.147 (1.358)
0.069 (0.635)
Intercept
0.155 (2.555)
0.127 (1.904)
0.275 (2.764)
0.250 (2.205)
0.173 (1.878)
0.103 0.343)
0.076 (0.990)
Tables 1 and 2 report the statistical results of the one-way analysis of variance as specified in eq. (1) for the seven SES stock indices for the first time period from January 13, 1978 to June 30, 1981 and for the second time period from June 1, 1982 to November 18, 1985, respectively. Notice that there is no day-of-the-week effect of stock returns in the first period but there is a statistically significant day-of-the-week effect of stock returns in the second period. In general the rate of return appears to increase progressively over the days of a week. Since these two time periods correspond to a significant shift of the settlement system, I conclude that a settlement system matters importantly in determining the day-of-the-week effect of stock returns.
Table 2 Estimates of variation in trading day to trading day daily (W) rates of return on the SES stock indices with t-statistics in parentheses). 1982 to November 18, 1985 (833 observations, All share
Industrial and commericial
Finance
Hotel
Property
over the period June 1,
Plantation
Mining
Tuesday
0.086 (1.216)
0.089 (0.935)
0.149 (1.626)
0.067 (0.916)
0.020 (0.587)
0.101 (1.207)
- 0.034 ( - 0.292)
Wednesday
0.130 (1.844)
0.136 (1.436)
0.190 (2.092)
0.123 (1.703)
0.036 (1.071)
0.164 (1.981)
0.132 (1.134)
Thursday
0.206 (2.952)
0.255 (2.721)
0.210 (2.326)
0.155 (2.160)
0.082 (2.426)
0.202 (2.464)
0.257 (2.222)
Friday
0.172 (2.445)
0.211 (2.225)
0.205 (2.251)
0.166 (2.284)
0.057 (1.678)
0.178 (2.145)
0.147 (1.254)
- 0.166 (- 3.197)
- 0.056 ( - 2.304)
- 0.118 (- 2.000)
Intercept
- 0.127 (-2.516)
- 0.144 (- 2.122)
-0.138 (-2.119)
-0.152 (-1.817)
356
T.K. Lee / Settlement
system, day-of-the-week
effect
References French, K.R., 1980, Stock returns and the weekend effect, Journal of Financial Economies 8, 55-69. Gibbons, M.R. and P. Hess, 1981, Day of the week effect and asset returns, Journal of Business 54, no. 4, 579-596. Saw, S.H. and K.C. Tan, 1983, The S.E.S. all-share price indices, 1975-1983, (The Stock Exchange of Singapore Singapore). The Singapore Stock Exchange Journal, various monthly issues.
Limited,