Statement
In a no-arbitrage regime-switching affine term structure model fit to US nominal Treasury yields and inflation, essentially all of the upward slope of the nominal yield curve is generated by the inflation risk premium component, not by an upward-sloping expected real-rate path and not by rising expected inflation. The inflation risk premium is itself substantial and time-varying across regimes — large and volatile in high-inflation / high-uncertainty regimes, small and stable in low-inflation regimes.
Evidence summary
The direct evidence is the model-implied component decomposition in ang-bekaert-wei-2008-real-rates-expected-inflation. The decomposition relies on a regime-switching pricing kernel with regime-dependent prices of risk; collapsing the IRP to a constant materially worsens the fit to long- maturity yields, which is consistent with the time-varying-IRP story.
Conditions and scope
- US Treasury yields, pre-TIPS sample.
- Regime-switching three-factor affine ATSM with a regime-dependent inflation risk price.
- Survey inflation forecasts are used as an auxiliary measurement equation to discipline expected inflation.
- The “IRP drives slope” decomposition is conditional on the model class; alternative no-arbitrage models with different price-of-risk specifications could allocate the slope differently.
Counter-evidence
(none ingested yet — to be filled when alternative decompositions enter the wiki)
Linked ideas
(none yet)
Open questions
- How much of the IRP estimate is driven by the survey measurement equation versus the no-arbitrage cross-section restriction?
- Does the IRP-dominated-slope finding survive in the post-2008 low-inflation regime?
- Can the time-varying IRP be tied to identifiable monetary policy shocks or inflation uncertainty proxies, rather than purely a latent regime?