NB21-18: Conditional and Unconditional Long-Term Forecasts of Economy-Wide Social Security Variables

This project develops and applies new forecasting methods for long-term economic variables that are integral to Social Security projections, such as trust fund ratios, costs, and income. Examples of questions that can be addressed by the methods: How likely is it that the OASDI Trust Fund ratio will turn negative by 2035? Or, if real wages grow 1% faster than assumed in the intermediate scenario in the Trustees Report, what will be the dynamic effects on trust fund reserves over the next 75 years? The methods use time and frequency domain techniques to model the long-term trends in economic data, ignoring fluctuations that are transitory. Once constructed, the tuned and adjusted models will be used to forecast Social Security aggregates up to 75 years into the future.