Trading Day of the Week (TDW) on the PSE Index

The concept of Trading Day of the Week was introduced by Larry Williams in his trading literature. The idea was applied to commodity trading, particularly treasury bonds and S&P 500 futures, and was later extended to practically any trading instrument. The idea behind TDW is that markets exhibit strong tendencies depending on the day of the week beyond the level that random variations can explain. This is my attempt to replicate the same research on the Philippine Stock Exchange Index.

Some basic information and background on the research:

• The data used on this study was daily data on the PSE Index from 3 January 2000 updated to 27 June 2008. There are 2,092 data points.
• The performance data of each days are primarily as follows:
  • Profit – is measured as the distance in points from the opening and close of day. A gain is if the close is higher than the open, while a loss is the opposite.
  • Win Rate – is the percentage of days in the category that registered a positive close.
  • Avg Win – is the total gains of profitable days over the number of profitable days.
  • Avg Loss – is the total losses of losing days over the number of losing days.
  • Win/Loss – is the average win figure over the average loss figure.

All Days Results

The data shows a tendency for losses on the first 3 days of the week and gains for the last two days. Winning percentages are below 45% for Mondays and Tuesdays, while at 55% for Thursdays.

Average Wins are about even for all days except Tuesdays, where average gains were 89% of the size of average losses. Computing for the expectancy of all days results in:

Thursdays are the most favorable for gains while Tuesdays are the worst favorable.

Months Of The Year

Before I got acquainted with Williams’ methodology, I was already conducting some studies on seasonality. Last November 2007, I posted on Finance Manila a study of 20-day holding period results for all days of the year and found some interesting bias for gains in October. The problem with this study is that it tracks only winning percentages, and not the magnitude of the gains.

Another take on the same problem is to apply the TDW analysis for months in the year. Clearly if a bias in winning percentages exists for some months, the magnitude of the returns would also support this, and the resulting expectancy would say the same.

Before I continue, readers of the original Seasonality Problem should note the difference between that study and this one. The original Seasonality Problem examined the likelihood of a winning trade after 20-days having bought in a certain month. This study examines gains made within the month itself.

The data shows that the Months of January, September, and December have a tendency for gains while the months from February, March, and August have a tendency for declines. Winning percentages are highest for January, September, and December while lowest for February, June, July, and October.

The months of October and November have the best win/loss ratios with aveage wins at least 134% of average losses, while the months of February, March and August are the worst with average wins less than 81% of losses.

The expectancy of all months are a varied spread:

The months of January, September, and December have the best expectancy for gains, while the months of February, March , and August have the worst expectancy.

Insight and Application

The results of the study indicate that there are strong behavioural tendencies for days and months in trading which are inconsistent with random variation. A significance test of both studies shows the odds of random chance producing those results at 1 in 2,116 for daily returns, and 1 in 727 for monthly returns.

The spread of returns may indicate a cyclical or systematic trending component in trading activity–i.e. liquidation in the first half of the week and accumulation in the other half, or liquidation from February to August and accumulation from October to January.

These seasonal components can be combined with standard technical and fundamental inputs to spot ideal times for position entries and exits. Traders can also use seasonal input to help them adjust their position sizes, scaling down during less favorable times, and stepping up volumes during more favorable periods.