New Bounds on Yield Curve Shape Predictability
Mathematical classification of the Hull-White model limits the possible shapes of future yield curves.
The Hull-White model now has a complete classification of all attainable yield curve shapes when using Nelson-Siegel parameterization. By applying Tchebycheff systems and the envelope method, the research partitions the state space to define exactly which curve geometries are possible. This removes the guesswork from how a specific initial curve can evolve over time within this framework.
Fixed-income desks rely on these models to price derivatives and manage hedge ratios. When a model allows for an infinite variety of curve shapes, the risk of "model misspecification" increases. Establishing a finite set of attainable shapes allows traders to identify when a market move is a standard evolution of the curve or a structural break that the model cannot explain.
For those using Bliss or Svensson parameterizations, the work establishes strict upper and lower bounds on local extrema. This means there is a mathematical ceiling on how many "humps" or "dips" a yield curve can realistically exhibit. If a market price implies a curve shape that exceeds these bounds, the model is fundamentally incapable of capturing that reality.
This limitation becomes most acute during periods of extreme volatility or regime shifts. When the initial yield curve is parameterized, the model's ability to project future shapes is constrained by these mathematical boundaries. A desk that assumes a model can fit any arbitrary curve shape may be overestimating its hedging precision.
Long-term projections are further simplified by the model's asymptotic behavior. As time progresses toward infinity, the yield curve only attains three shapes with positive probability: normal, inverse, or humped. This suggests that complex, multi-humped curves are transient phenomena that inevitably collapse into these three primary states.
Quantitative strategies that bet on the persistence of complex curve shapes face a theoretical headwind. The evidence indicates that the mathematical structure of the Hull-White model forces a convergence toward simpler geometries, limiting the long-term viability of trades based on exotic curve curvatures.