Home, choices, barometer, TIMING, backtest, signal, buy, sell, hedge, watch

Day of Month and Day of Week


Timing trades by day of the month will be profitable. Paying attention to day of the week or big moves will be a waste of time.


Day of the month

Daily closing prices for the S&P 500 Index were analyzed during 1950-2011, and during 2000-2011 to see if recent behavior was similar:

1950-2011

2000-2011

The Avg columns show no consistency for day of the month, but the period including 28, 29, 30, 31, 1, and 2 was consistently strong. This period, called Turn of Month, gained 8.09 during 2000-2011 and 6.86 during 2000-2011.

The rallies last longer near some holidays. The best known is the Santa Claus Rally, which occurred during Dec 16 through Jan 6 in the tables but is defined otherwise elsewhere. This period extended the Turn of Month Rally by 2.34 during 1950-2011 and 1.99 during 2000-2011.

There were also significant extensions to the Turn of Month Rally near Thanksgiving, Halloween, and Independence Day. Adding all extensions to the Turn of Month Rally gives a gain of 13.30 during 1950-2011 and 13.45 during 2000-2011. These were almost twice as much as the 6.98 average annual gain in the S&P 500 Index, meaning that the Index declines when not in a favored time period.

The best time to buy is before the 28th, except the 16th in December and the 23rd in November. The best time to sell is before the 3rd, except the 7th in January, the 10th in July, the 6th in November, and the 11th in December.

There were also strong periods that were not Turn of Month Rally extensions. The gain during March 10-18 was 1.27 during 1950-2011 and 2.21 during 2000-2011. The gain during April 5-18 was 1.81 during 1950-2011 and 1.69 during 2000-2011. These rallies might be of interest to traders, but not to investors.

The sell dates are unnecessarily complicated for investors. Most sales for investors will be in response to a trend reversal sell signal or a negative January Barometer signal, which are independent of monthly timing signals. For investors, the best time to sell can be simplified to before the 3rd.

Day of the week

The daily closing prices for the S&P 500 Index during 1950-2011 and during 2000-2011 were analyzed:

day of week

The numbers are small enough to be ignored. The largest magnitude is 0.07, which is less than most changes shown in the charts for day of the month.

Reaction after a big move

Pundits say that a big one-day drop in the stock market is usually followed by an upward move the next day. To see if acting on this theory can be profitable, daily closing prices for the Dow Jones Industrial Average during 1928-2011 were analyzed:

big moves

There is no consistency in the gains for moves <= 2, >= 2, or >= 5. The gains for moves <= 5 are consistently positive, but vary widely.

The gains during 1964-1999 for moves <= 5 boxes averaged 3.68, which is larger than the gain from the Santa Claus Rally, but there were only 6 occurrences in 36 years. The average gains during 1928-1963 were insignificant. The average gains during 2000-2011 were only 1.07, but might also be tradable.

The wide variances in the gains for moves <= 5 boxes raises doubts about their forecasting ability. Furthermore there have been only 18 occurrences since 1963, so there is very little profit opportunity here.