GoKawiilThe authors show that, contrary to this belief, at long horizons nothing is lost by ignoring cointegration when the forecasts are evaluated using standard multivariate forecast accuracy measures. In fact, simple univariate Box-Jenkins forecasts are just as accurate. The authors' results highlight a potentially important deficiency of standard forecast accuracy measures — they fail to value the maintenance of cointegrating relationships among variables — and the authors suggest alternatives that explicitly do so.
Cointegration and Long-Horizon Forecasting (2025)
Why This Matters
This article challenges the conventional wisdom that cointegration is essential for long-horizon forecasting accuracy, revealing that simple univariate models can perform just as well as more complex multivariate approaches. It underscores the need for improved forecast accuracy measures that account for cointegrating relationships, which are often overlooked. This insight has significant implications for the development of more effective forecasting models in the tech industry and beyond.
Key Takeaways
- Simple univariate models can match complex multivariate forecasts at long horizons.
- Standard accuracy measures may undervalue the importance of cointegration.
- New evaluation methods are needed to better capture the value of cointegrating relationships.
Get alerts for these topics