Premium Market Analysis, Trader Education, Software and Trading Strategies

Uncategorized

Trading Moon Phases and Related Folklore

I cannot discount the possibility that Astrology may provide clues for occasional market timing but as far as contributing to a systematic methodology I cannot see the implied causality that could facilitate that.

Markets are driven by financial facts and expectations rather than by the motion of celestial bodies. Except, if one can prove that financial developments in macro and micro scale are caused by the motion of celestial bodies. I am not going to deal with such metaphysical issues here but instead I will attempt to analyze the results of the study by a major bank in a practical manner that everyone can understand. In the study only the results were shown but not the code that generated them. Thus, I have no way of checking the validity of the results. What we can can do though, is to take the conclusions and run a few examples to check them out. The hypothesis to test in that specific regard is the conclusion of the study and specifically the claim that buying on a new moon and selling on a full moon is a trading/investing strategy that significantly outperforms buy and hold.

As I said I am a practical person. For the purpose of testing the hypothesis mentioned (MPS from now on) I did not attempt to develop a program to calculate the moon phases and then use it for back-testing. I had a feeling that I will never use it again and it would be a waste of time in that regard. Instead I used SPY daily data and I inputted in Excel the dates for a new and a full moon for various years using this lunar calendar. Click on the images to enlarge. 

The first year I studied was 2003. This was a bull market year during which SPY rose from $88.85 to $112.22. In the screenshot below we can see the new moon and full moon trading days, the buy and sell price, the P/L of each trade and the net result. If the exact date falls on a Saturday or Sunday, or on a holiday, the trading day is moved to the nearest previous trading day. Buy and sell orders are entered at the open of the day.
Results for year 2003. MPS outperforms Buy and Hold by 18.5%

It may be seen from the results that the moon phase strategy (MPS) outperformed buy and hold by 18.15% not counting commissions of course. The next year I studied was 2007, a year that included a strong rally that brought the S&P 500 to new highs but also included a correction after that and high volatility. In the following screenshot we can see that the MPS underperformed the positive buy and hold result by 258% and produced a negative net return. This is the first study that started raising questions about the performance of the MPS.

Results for year 2007. MPS underperformed by 258%

The next year studied was 2008, the year of the financial crisis and the huge plunge in stocks. The MPS outperformed buy and hold by 49.1% but both had negative net result. Can we attribute the better negative performance of MPS to some causal powers due to moon phases? Certainly, not. It is simply the fact that the market went straight down and MPS was more or less 50% of the time invested in it. Hence, the performance was increased by 50%. This realization shows that when one performs back-testing studies should not only look at the final results but also have a knowledge of the structure of the inputs. It is rational to attribute the improved performance to the fact that the MPS was almost half of the time invested in a market that went straight down than to some magical causal power affecting financial behavior.

Results for year 2008. MPS performed better by a factor of 49%

The last year I ran the study for was 2009, a very good year for the market. The MPS underperformed buy and hold by 33.22% and this is probably due to the opposite effect of year 2008. During a strong bull market, it does not pay to invest half of the time and this is the reason for the lower performance of MPS. In my opinion, if there was some causal relationship between moon phases and financial market behavior in favor of buying on a new moon day and selling on a full moon day for better performance, year 2009 would be a good one to demonstrate it. Instead, we noticed lower performance.  We can thus draw the conclusion that the MPS performance is better in plunging markets due to the fact that it stays almost half of the time not invested, like in year 2008, and it underperforms buy and hold in sharply rising markets due to the same fact of not being fully invested in the markets, like it did in 2007 and 2009, and  may outperform or underperformed in volatile markets, like in 2003 where it outperformed, depending on the market behavior during the moon phases.

Results for 2009. MPS underperformed buy and hold by 33%
I was not able with these examples to produce numbers that would contribute to a spectacular performance of the MPS as in the original study mentioned. That particular study took into consideration about 26 years and I am not sure the dividends of S&P 500 were accounted for in the performance figures. The dividends of 500 stocks are a sizable percentage of investment performance over the years due to the yield but also due to their reinvestment. Thus, one cannot compare just plain index performance to performance that is invested about 50% of the time without considering the impact of dividends. Regardless, what is important is that MPS did not turn a negative SPY performance to positive during year 2008, neither outperformed the low but positive performance of 2007, rather  failed with negative performance, so I can safely attribute all this talk about moon phases and trading to folklore rather than to anything that has any sound basis.