Detrending removes a long-run pattern from data so economists can analyze cyclical or short-run behavior more clearly.
Why Economists Detrend Data
Many macro series (GDP, productivity, credit) grow over time. If that trend is not handled, regressions can mistake common trending behavior for real economic relationships.
Common Methods
- Linear or polynomial trend removal.
- First differencing: (\Delta y_t = y_t - y_{t-1}).
- Filter-based methods (for example, HP-style cycle extraction).
Method choice depends on whether the trend is viewed as deterministic or stochastic.
Model Logic
If a series is trend-stationary, subtracting a deterministic trend can recover a stable process. If it has a unit-root-type process, differencing is often more appropriate.
Using the wrong method can remove meaningful long-run information or leave spurious persistence in the residual series.
Practical Context
Detrending is central in business-cycle analysis, output-gap estimation, and policy evaluation. Results can be sensitive to trend assumptions, so empirical work should report robustness to alternative detrending approaches.