Seasonality and the election cycle
Many markets behave differently at different times of the year. Grains follow the planting and harvest cycle, energy follows driving and heating seasons, and currencies and indices are moved by fiscal year-ends, tax deadlines and fund flows. Seasonality measures those tendencies.
How a seasonal chart is built
Take the last 10 to 30 years of prices. For each year, convert the path into a percentage path starting from the first trading day. Average all the years together and you get the typical shape of the year. Projected forward from today, it shows what the market usually does over the next weeks.
Reading it
- Direction: is the seasonal line rising or falling over your trade horizon?
- Turning points: seasonal tops and bottoms often line up with real ones
- Consistency: a tendency that held in most of the years is more useful than one driven by a single big year
The presidential election cycle
US stock indices have historically behaved differently across the four years of a presidential term. Popularised by Yale Hirsch's Stock Trader's Almanac, the pattern is often summarised like this: pre-election years have tended to be the strongest, midterm years the weakest with a low often forming in the second half, and post-election years mixed. Looking at the same calendar window only in years of the same cycle type gives a sharper seasonal read than averaging every year together.
How seasonality fits the method
Seasonality is a supporting factor, never a trade on its own. In the two-step method it's the third check after positioning and valuation. When all three line up, the bias is strong. Our UTC Seasonality indicator draws the seasonal path on your chart, marks turning points and has an election-cycle mode.
The Underdog Trade Club is an independent community and is not affiliated with Bernd Skorupinski, Larry Williams or their official brands.
Questions
Does seasonality work in forex?
Yes, especially around fiscal year-ends and in currencies tied to commodities, but it should confirm a bias, not create one.
How many years of data should I use?
Between 10 and 30 years. Fewer years react more to recent behaviour, more years give a steadier average.