Smooth Forecast Reconciliation

IMF Working Papers, March 22, 2024

Source details

Research question: How to make forecasts that (1) satisfy constraints, like accounting identities, and (2) are smooth over time?
Motivation: Solving this common forecasting problem manually is resource-intensive, and the existing literature provides little guidance on how to achieve both objectives.
Main contribution: Proposes a new method to smooth mixed-frequency multivariate time series subject to constraints by integrating the minimum-trace reconciliation and Hodrick-Prescott filter.
Closed-form result: With linear constraints, the method has a closed-form solution, convenient for a high-dimensional environment.
Empirical claim: Three examples show that the proposed method can reproduce the smoothness of professional forecasts subject to various constraints and slightly improve forecast performance.

Components integrated:
- Minimum-Trace Reconciliation
- Hodrick-Prescott filter
Problem scope:
- Mixed-frequency multivariate time series
- Constraints include accounting identities (linear constraints)
Solution properties:
- Closed-form solution when constraints are linear
- Suitable for high-dimensional settings

Examples:
- Three examples are used to evaluate the method.
Outcomes reported:
- The method can reproduce the smoothness of professional forecasts subject to various constraints.
- The method slightly improves forecast performance (described qualitatively as "slightly improve forecast performance").

Use cases:
- GDP forecasting
- National accounts
- Any multivariate forecasting task requiring both constraint satisfaction and temporal smoothness
Advantages:
- Reduces manual effort required to enforce accounting identities while preserving smooth time paths
- Computationally convenient in high-dimensional environments due to closed-form solution under linear constraints