What can we see on the chart?
Chartoasis Sesame provides a basic and a complex, more detailed chart for calendar effect analysis.
Basic calendar effect chart
Two values are displayed above the names of months (see the example below):
For example: if we have data for three years, and the increase between the first and last days of January during the three years was -1%, 1% and 3%, and the threshold value is 0%, then
the frequency of increase will be 2/3 = 66%, as 1% and 3% are greater, than 0%, and a total of three Januaries have passed.
Average increase will be (-1% + 1% + 3%) / 3 = 1%
Relative frequency of increase is displayed with the bar chart, average returns is displayed with a curve.
What is this chart useful for?
Investments often do not perform equally well in all months of year. As an example, just think about ?Sell in may, and go away? or ?Santa Claus Rally?. These anomalies are called "calendar effects".
One may reasonably ask to what extent will this phenomena influence his/her given investment? Or is there an other period of the year that stands out?
This information is estimated in this monthly statistics.* [e.g.: refer to the following chart].
In the statistics of Google shares (above Figure) one may observe, that December and October months have often yielded decent returns, while during February the shorting of Google shares seemed more deserving.
It must be noted, that relative frequency values around 50% are of the least importance. Months, where the frequency of increase approaches 100% or 0% are more interesting.
Forecasting with a great degree of certainty can be made into money at the stock exchange, in case of either increase, or decrease. This can also be utilized for short-term strategies.
In case of a longer-term investment, the time of opening a position can be optimized with the help of the chart (e.g.: it is advisable to avoid opening a long-term position in the beginning of a month that most often yields decrease).
- Confidence interval may also tell if rising or falling prices are more likely in a particular month (when it is interpreted correctly). It is very important how wide the confidence interval is.
- One needs enough data (like 10 years or so) to use the chart properly.
How can the chart be used?
For statistics, one can include the fact, that for example: an increase near 0% cannot yield returns on the stock market, as with the transaction charges, the result would be loss.
This can be eliminated by setting the statistics threshold to a value greater than 0% (e.g.: 0.5%). In these cases increases below 0.5% are not included in the statistics calculation.
*It must be noted however, that past behavior does not represent any guarantee in terms of future performance!