Intra-day seasonalities on stock returns: evidence from the Turkish Stock Market

Abstract

One of the interesting findings among the seasonalities in stock markets is that the return, volume and volatility of the stock prices and bid-ask spreads all follow a U-shaped pattern over the trading day. This study examines the intra-daily seasonalities of the stock returns in the emerging Turkish Stock Market which is an order-driven market using electronic trading without market makers, in the period from 1996 to 1999, by using 15-min (and also 5- and 1-min) interval data. Results show that stock returns follow a U-shaped or more precisely a W-shaped pattern over the trading day at the Istanbul Stock Exchange (ISE) since there are two separate trading sessions in a day. Opening (Overnight) and closing returns are significantly large and positive. Volatility is higher at the openings and follows an L-shaped pattern during the both sessions. Interestingly, the daily average close-to-close returns are generated only during the opening and closing intervals and the average intra-day return is negative when the returns at the opening and/or closing intervals (even the first and the last minutes of the day) are excluded from the analyses. Relatively higher mean return and standard deviation at the openings of the trading sessions seem to be significantly generated by the accumulated overnight information and the closed-market effect (halt of trade). Large day-end returns are strongly affected by the activities of fund managers and speculators for the window-dressing around the close. Finally, intra-day seasonalities exist significantly also in the Turkish Stock Market as consistent with those of the international stock markets.

Introduction

Price behavior in financial markets has generated much research interest. The existence of seasonalities that contradicts the Efficient Markets Hypothesis in stock markets has increased the importance of intra-day data which offer further insight in understanding the price behavior of stocks. One of the interesting findings among the seasonalities in stock market prices is that the return, volume and volatility of the stock prices and the spread between the bid and ask quotes, all broadly follow a U-shaped pattern.² All these variables are at the highest point at the opening, fall rapidly to lower levels during mid-day and then rise again towards the close.

Harris (1986) shows that there is a general increase in New York Stock Exchange (NYSE) stock prices on the last trade of the day for the period December 1, 1981 through January 31, 1983. Harris found that the weekend effect spills over into the first 45 min of trading on Monday, with prices falling during this period. On all other days, prices rise sharply during the first 45 min. Also, returns are high near the very end of the day, particularly on the last trade of the day. Furthermore, the day-end price changes are greatest when the final transaction is within the last 5 min of trading. He indicated that stocks traded in the last 5 min went up 0.1241% (1.75 cents), while the increase on the last trade for all stocks was only 0.0507% (0.595 cents). This is a large increase; assuming 250 trading days per year, a 0.1241% per day increase amounts to 31% without compounding.

Wood et al. (1985), Brooks and Chiou, 1995, McInish and Wood, 1990a with a 1980–84 NYSE dataset and Jain and Joh (1988) with hourly price changes using the S&P's index for the 1979–1983 period have also obtained similar results that report very high returns for the last minute of the day. Furthermore, some researchers³ indicate to the relationship between intra-day and day-of-the week effects.

It is widely documented that the return volatility varies systematically over the trading day and that this pattern is highly correlated with the intraday variation in trading volume and bid-ask spreads. Jain and Joh (1988) document a positive relation between volume and contemporaneous and lagged returns that is especially strong for positive returns. They study hourly aggregate NYSE volume and report that the volume is particularly high at the beginning and towards the close of trading. Foster and Viswanathan (1993) and Brock and Kleidon (1992) also identify U-shaped patterns in volume and volatility for NYSE stocks. The highest volume coincides with the highest variance, which is supportive of the concentrated trading model of Admati and Pfleiderer (1988). They find a significant positive relation between the volume and volatility of the stocks. Wood et al. (1985), Harris (1986), McInish and Wood (1990a), Foster and Viswanathan (1993), Ho and Cheung (1991), Chang et al. (1995), Madhavan et al. (1997) find similar results that the variance of returns is much higher at both the open and the close than for the rest of the day.

Moreover, bid-ask spread,⁴ risk premium,⁵ the number of shares traded,⁶ number of trades,⁷ quote revisions⁸ and trade size⁹ all have also been shown to have U-shaped patterns over the day.

Additional to the evidence from US markets, Cheung (1995), Lam and Tong (1999) for the Hong Kong Stock Market, Miller (1989), Yadav and Pope (1992) for the UK market and Choe and Shin (1993) for Korean, Copeland and Jones (2000) for Korean and Hong Kong Markets provide evidence for the intra-day effects. However, existence of intraday patterns and more specifically U-shaped patterns has been demonstrated for other markets such as: by Peterson (1990) and Aggarwal and Gruca (1993) for equity options market, Muller et al. (1990) and Baillie and Bollerslev (1990) for the foreign exchange market, Cornett et al. (1995) for the foreign exchange futures market, and Jordan et al. (1987) for the commodities market.

The main motivation here in this study is the search for evidence to support the microstructural hypotheses regarding intradaily trading patterns which are given in Section 2. In this context, the Turkish market is particularly interesting because of the shut down at lunchtime with two trading sessions in a day similar to the most of the Asian markets. For instance, if the Brock and Kleidon hypothesis is correct, this market should have four rather than two peak in activity during the day. Existing literature concentrates on markets where floor trading predominates. The ISE is a continuous order-driven market which uses computerized trading system but without a specialist or market maker.

However, finding an evidence for the existence of intra-day effect in an emerging market would imply that there may be significant opportunities for the international investors to make profit if they have discretion to time their trades. This study will also refer to this issue in order to find whether large profits can be obtained by using a simple trading rule such as buying and selling stocks at a particular time during the trading day based on the intra-day seasonality in stock returns observed.