Intra-Decade Seasonality — ^DJI

SeasonAlpha's Decade Cycle explores a single question: do stock-market years sharing the same last digit of the year behave similarly on average? To find out, the tool sorts every complete year of a time series into ten cohorts — from x0 (1990, 2000, 2010 …) to x9 (1989, 1999, 2009 …) — and averages the normalized yearly path for each group. This reveals which last digits have historically tended to perform strongly and which weakly. The default is the Dow Jones (^DJI) with over 130 years of history; any other ticker can be loaded from the sidebar.

Each line in the main chart is the average yearly path of a last-digit cohort, normalized to 0 percent at the start of the year. If a line rises over the course of the year, the average return of those years was positive. The gold curve highlights the current year, letting you compare the present level directly with the typical path of its decade cohort. In addition, the bar chart shows the average return per last digit, and the table shows each cohort's win rate (share of positive years), median, volatility and the underlying years. A high win rate combined with a solid average return points to a historically more reliable pattern — high volatility points to large swings between individual years.

Methodologically, the tool uses normalized returns: the first trading day of each year is set to 0 percent, and the following days accumulate on top of it as logarithmic returns. This makes years with different price levels directly comparable — what matters is the percentage path, not the absolute price. Each yearly curve is interpolated to 252 trading days so that the cohorts can be averaged cleanly; years with fewer than 200 trading days are excluded. Grouping by last digit produces ten cohorts whose sample size n is limited: even for the Dow Jones, which spans over 130 years, only about 13 years are available per digit.

The Decade Cycle is a descriptive tool for putting historical patterns into context — not investment advice and not a price forecast. Because the sample per last digit is small, individual extreme years (such as 2008 or 1931) can heavily skew a cohort's average, which is why win rate, median and the spread of individual years should always be viewed together. A historically strong decade year is no guarantee that the next year with the same last digit will behave the same way. Use the cycle as additional context alongside fundamentals and your own research, not as your sole basis for decisions. Past patterns do not guarantee future results.

Frequently Asked Questions

What is the decade cycle in the stock market? The decade cycle groups all available years of a time series by the last digit of the year (0 to 9) and averages their price paths. This creates ten cohorts — for example all years ending in 5 (1985, 1995, 2005, 2015 …). The tool descriptively shows how the years of a given last digit performed on historical average. It is a statistical analysis of the past, not a forecast.

How do I read the decade path and the win rate? Each line in the main chart is the averaged yearly path of a last-digit cohort, normalized to 0 percent at the start of the year. If the line rises, the average return of those years was positive over the course of the year. The win rate states in what percentage of a cohort's years the year closed positive. The sample size n indicates how many years the respective average is based on — for long series like the Dow Jones, usually only about 13 years per digit.

Can the decade cycle predict future prices? No. The decade cycle is a descriptive tool for putting historical patterns into context and is not investment advice. Each last digit covers only a few years, which is why individual extreme years can heavily skew the average. A historically strong x5 year is no guarantee of a positive future x5 year. Past patterns are no guarantee of future results — use the cycle as context, not as your sole basis for decisions.