Time series analysis
Also known as: time series model, time series forecasting
Time series analysis is the study of data points recorded in time order to identify patterns and forecast future values. It decomposes a series into components such as trend, seasonal variation, cyclical movement, and random noise.
A time series is any variable measured at regular intervals — monthly sales, quarterly revenue, daily production output. What distinguishes time series analysis from ordinary regression is that the observations are ordered and typically correlated with one another, so the sequence itself carries information. The goal is to separate the systematic patterns from the noise, then project the systematic part forward.
The standard decomposition splits a series into four components. Trend is the long-run direction of movement. Seasonal variation is a repeating pattern within a fixed period, such as higher retail sales each fourth quarter. Cyclical variation is a longer, irregular swing tied to the business cycle. Random or irregular variation is what remains. These combine either additively (Y = T + S + C + I), when seasonal swings are roughly constant in size, or multiplicatively (Y = T × S × C × I), when swings grow proportionally with the level of the series.
Common techniques include moving averages to smooth short-term fluctuations and reveal trend, least-squares regression of the variable against a time index to fit a trend line, and seasonal indices that scale a trend forecast up or down for the period being predicted. Because the method assumes past patterns persist, forecasts degrade the further out they run and break down entirely when the underlying process changes.
Time series analysis is tested in the data analytics section of the CMA Part 1 exam and in the quantitative methods portion of the CIMA Certificate in Business Accounting and CGMA Fundamentals of Business Economics. Expect to compute a moving average, apply a seasonal index to a trend forecast, and identify which component explains a described pattern.
Key takeaways
- Time series analysis examines observations recorded in time order to detect patterns and forecast.
- Series are decomposed into trend, seasonal, cyclical, and irregular components.
- Additive models suit constant-size seasonal swings; multiplicative models suit swings that scale with the level.
- Moving averages smooth noise, while regression on a time index fits the underlying trend.
- CMA Part 1 and CIMA business economics exams test moving averages, trend lines, and seasonal indices.