Abstract:We examine the shapes attainable by the forward- and yield-curve in the widely-used Svensson family, including the Nelson-Siegel and Bliss subfamilies. We provide a complete classification of all attainable shapes and partition the parameter space of each family according to these shapes. Building upon these results, we then examine the consistent dynamic evolution of the Svensson family under absence of arbitrage. Our analysis shows that consistent dynamics further restrict the set of attainable shapes, and we demonstrate that certain complex shapes can no longer appear after a deterministic time horizon. Moreover a single shape (either inverse of normal curves) must dominate in the long-run.
What problem does this paper attempt to address?
### What problems does this paper attempt to solve?
This paper aims to explore and classify the achievable shapes of the yield curve and forward curve in the Svensson family (including the Nelson - Siegel and Bliss sub - families), and study the consistent dynamic evolution of these shapes under the no - arbitrage condition. Specifically, the paper mainly solves the following problems:
1. **Classification Problem**:
- Which shapes can be represented by the Svensson family and which cannot?
- By analyzing different regions of the parameter space, the paper gives the parameter conditions corresponding to each shape.
2. **Segmentation Problem**:
- For a given parameter vector \(\beta = (\beta_0, \beta_1, \beta_2, \beta_3)\), what are the specific shapes of the forward curve and the yield curve?
- The paper divides the parameter space into different regions, and each region corresponds to a specific curve shape.
3. **Consistent Dynamics Problem**:
- Under the consistent dynamic evolution, which shapes will appear with a strictly positive probability?
- The paper studies how the dynamic evolution of the Svensson family further restricts the achievable shapes under the no - arbitrage condition, and shows that some complex shapes no longer appear within a certain time range.
#### Main Findings
- **Shape Classification**: The paper proves that the Nelson - Siegel family has at most one local extremum point, the Bliss family has at most two local extremum points, and the Svensson family has at most three local extremum points.
- **Parameter Space Segmentation**: According to different time - scale ratios \(r=\frac{\tau_1}{\tau_2}\), the paper defines three different regimes (scale - regular, weakly scale - inverted, strongly scale - inverted), and describes in detail the achievable shapes in each regime.
- **Dynamic Evolution**: The paper finds that under the consistent dynamic evolution, some complex shapes (such as hdh and dh) will disappear after a certain time, and finally only one shape (inverse or normal curve) will dominate in the long run.
#### Mathematical Formulas
The paper uses several mathematical formulas to describe curves and their derivatives, for example:
- The forward curve formula of the Nelson - Siegel family:
\[
\phi_{NS}(x)=\beta_0+\beta_1 e^{-\frac{x}{\tau}}+\beta_2\frac{x}{\tau}e^{-\frac{x}{\tau}}
\]
- The forward curve formula of the Svensson family:
\[
\phi_S(x)=\beta_0+\beta_1 e^{-\frac{x}{\tau_1}}+\beta_2\frac{x}{\tau_1}e^{-\frac{x}{\tau_1}}+\beta_3\frac{x}{\tau_2}e^{-\frac{x}{\tau_2}}
\]
- The form after parameter re - parameterization:
\[
\phi_S(x)=\beta_0+\beta_3\left(\gamma_{II}e^{-\frac{x}{\tau_1}}+\gamma_I\frac{x}{\tau_1}e^{-\frac{x}{\tau_1}}+\frac{x}{\tau_2}e^{-\frac{x}{\tau_2}}\right)
\]
where \(\gamma = (\gamma_I,\gamma_{II})=\left(\frac{\beta_2}{\beta_3},\frac{\beta_1}{\beta_3}\right)\)