Good Morning Wall St.

Insights & Commentary, Education and Resources for Private and Institutional Traders

White Paper - "Trading Seasonality" - Seasonality And Other Instruments

Introduction    What Is Seasonality?    The Yearly Seasonal Cycle
Seasonality Zonestm    Seasonal Heat    Seasonality And Other Instruments
Deploying Seasonality As Part Of An Overall Strategy    Conclusion

Seasonality analysis need not be confined only to stocks. Any instrument that has enough price history can be the subject of seasonal analysis. For instance, the price action of one stock relative to another stock (called pairs trading) can exhibit seasonal patterns. Let's examine the Pepsi (PEP) versus Coca-Cola (KO) pair.


Figure 7 Pepsi/Coca-Cola pair with seasonal patterns


The seasonal cycle for the Pepsi/ Coca-Cola pair in Figure 7 is relative flat until September, when it rises sharply with maximum positive seasonal heat. The pair itself rises when the stock price of Pepsi outperforms the stock price of Coca-Cola. Since the seasonal pattern favors a rise in the paired trade starting in September, traders would expect a "long Pepsi/short Coca-Cola" trade to be setup at this time. Indeed, in 2005 the Pepsi/ Coca-Cola pair did rise from 1.24 on September 7, 2005 to 1.39 on October 19, 2005, when the seasonal cycle flattened out just prior to a weak seasonal zone.

Seasonal analysis can be performed on other indices. Industry group, sector and market indices are prime examples. Many professional money managers believe that the influence of groups, sectors and the stock market itself play a large role in the path taken by individual stocks. We agree, and constantly apply our seasonal microscope to such indices.


Figure 8 Retail sector seasonal patterns

Figure 8 shows the seasonal pattern of the retail sector. The retail sector is a composite of many industry groups related to retail sales, including building material chains, catalog and specialty chains, clothing/shoe/accessory chains , computer/video chains, department stores, discount chains, drug store chains, food chains and other miscellaneous retail chains. These diverse retail industries are reflected in the one retail sector index. As Figure 8 shows, the 11-year seasonal cycle of the retail sector has a high Cycle Rtm statistic of 0.68, indicating the presence of a valid seasonal pattern throughout the past 11 years. There are also moments of strong seasonal heattm, indicating consistency of seasonal zonestm appearing at the same time in prior years. This information is invaluable to professional portfolio managers who often make sector and industry group money management decisions.

In Figure 8, we see a seasonal pattern of weakness for the retail sector in September, followed by strength in October and November. As it turns out, this is similar to the seasonal pattern of the market as a whole, but the task now becomes one of finding retail industry groups and particular stocks that strongly follow the seasonal pattern of the retail sector. In general, most of the retail groups follow this pattern, but the clothing/shoe/accessory chains appear to most suitably mirror the September through November retail sector seasonal setup.


Figure 9 Clothing/shoe/accessory chains with seasonal patterns


In a similar fashion to the seasonal data derived from the retail sector index, Figure 9 shows a weak seasonal curve in September for clothing/shoe/accessory chains, with a large degree of negative seasonal heattm. This turns around in the October and November period, with November showing a large degree of positive seasonal heattm for clothing/shoe/accessory chains. As it turned out, the 2005 price action of the clothing/shoe/accessory chains followed this pattern closely. The task now becomes finding stocks that show comparable seasonal tendencies.


Figure 10 Limited Brands (LTD) with seasonal patterns


Figure 10 shows the seasonal pattern for Limited Brands (LTD), a major retail stock from the clothing/shoe/accessory chains industry group. This is a good candidate because its seasonal curve follows the path taken by its group and sector (and of course the market as a whole,) and also because its seasonal heat patterns of the current seasonal zonestm show great consistency with seasonal zonestm of the past. Moreover, the Cycle Rtm statistic over the past 15 years is a highly correlated 0.69, which means the seasonal pattern is valid over the 15-year look back period. In truth, there are a number of individual retail issues that qualify as portfolio candidates using such seasonal criteria. It is the job of the portfolio manager to sift through the candidates and select those that meet seasonal and other criteria as befits the manager's style.



P.O. Box 280
Acton, MA 01720