Tuesday, July 20, 2021

Uncertainty and Treasury yields - Forecast disagreement impacts yields



Focusing on the disagreement of yield forecasts from leading economists can serve as a useful factor for improving the accuracy of model-driven Treasury yield forecasts. Disagreement in Blue Chip forecasts has unique and stronger predictive power than the well-known traditional factors like inflation and growth. The positive predictive effect of forecast disagreement is especially strong following recessions. 


This disagreement risk factor was explored in the recent Journal of Finance (February 2021) paper, "Learning from Disagreement in the US Treasury Bond Market". The authors use a Bayesian learning model that updates the parameters driving bond yields within a dynamic term structure model. Including the disagreement among economist forecast improves the accuracy of the model relative to a model that only uses yield information. The authors also find that fundamental factors like output growth and inflation can be viewed as redundant once the model conditions on forecast disagreement. Additionally, the systematic model even without using the disagreement factor is more accurate than the consensus Blue Chip forecast. 




The conclusion is simple. Use the disagreement of forecasts as an added risk factor, but don't use the actual forecasts of economists. Disagreement is a unique factor and helps with the adjustment of model parameters. Disagreement can better explain excess returns when there are transition economic regimes. 

No comments: