MACROCAST: A Vintage-Consistent Time Series Foundation Model for Real-Time Macroeconomic Forecasting
arXiv:2606.28670v1 Announce Type: cross Abstract: We introduce MACROCAST, a lightweight Time Series Foundation Model (TSFM) for real-time macroeconomic forecasting. Existing TSFMs suffer from data leakage in two forms: temporal contamination, as the model may have seen the realized values of the series it forecasts, and revision bias, as training on fully revised data diverges from the preliminary, vintage-specific releases available to real-time forecasters. MACROCAST is, to our knowledge, the first TSFM that rules out both forms of leakage entirely: at no stage of training is the model expos
The proliferation of Time Series Foundation Models (TSFMs) has highlighted their limitations in real-time macroeconomic forecasting due to data leakage, making the development of robust, vintage-consistent models a critical next step.
Accurate, real-time macroeconomic forecasting free from data contamination is crucial for policymakers, financial institutions, and businesses to make informed decisions and reduce systemic risk.
The introduction of MACROCAST marks a shift towards more reliable and trustworthy AI models for economic prediction, potentially improving the efficacy of policy responses and investment strategies.
- · Central Banks
- · Financial Institutions
- · Economic Policy Makers
- · Data Scientists in Finance
- · Traditional Econometric Models
- · Forecasters relying on contaminated data
- · Less robust TSFMs
Improved accuracy and reliability of real-time macroeconomic forecasts.
Better capital allocation and risk management within financial systems based on more informed projections.
Enhanced global economic stability through proactive and data-driven policy interventions, potentially mitigating crisis severity.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG