The economic_curves library

Overview

This library includes Python scripts and notebooks that contain practical algorithms for modeling economic scenarios, many of which are relevant to regulatory requirements in UK and EU countries under the Solvency 2 regime.

See also

This library was based on insurance_python, a library in an external project, the Actuarial Algorithms, developed and maintained by Qnity Consultants and Gregor Fabjan.

See also

• The economic Library for a Hull-White model based on modelx

• Project smithwilson for a primitive Smith-Wilson implementation based on modelx

How to Use the Library

As explained in the Copying a Library section, create you own copy of the economic_curve library. For example, to copy as a folder named curves under the path C:\path\to\your\, type below in an IPython console:

>>> import lifelib

>>> lifelib.create("economic_curves", r"C:\path\to\your\curves")

This library does not use modelx.

This library uses the following packages. Most of the packages are pre-installed in major Python distributions, but if any of them are missing, install them by pip or conda manually.

Additional packages used

• numpy

• pandas

• SciPy (Nelson-Siegel-Svensson only)

• Matplotlib

Algorithms available

Scripts and notebooks for each algorythm are put in a sub-folder dedicated for the algorythm in this library.

Algorithm (Folder)	Source	Description
smith_wilson	Technical-documentation	Interpolation and extrapolation of missing interest rates
stationary_bootstrap_calibration	Whitepaper-2004	Automatic calibration of the stationary bootstrap algorithm
stationary_bootstrap	Politis-Romano-1994	Resampling procedure for weakly dependent stationary observations
bisection_alpha	Technical-documentation	Calibration of the Smith & Wilson’s alpha parameter
correlated_brownian_motion	Wiki Brownian motion	Simple function to generate correlated Brownian motion in multiple dim.
NelsonSiegelSvensson	BIS whitepaper	Nelson-Siegel-Svansson model for approximating the yield curve
black_scholes	Wiki Black&Sholes	Black&Scholes model for pricing option contracts
EIOPA_smith_wilson_test	EIOPA RFR website	Calculation of the risk free rate from the monthly EIOPA publication
Metropolis_Hastings	Original work by OSM	Bayesian maximum likelihood of a Black Sholes stochastic scenario generator

Algorithms planned

New suggestions for algorithms are welcome. If anybody is interested in publishing an algorithm they implemented, or help with the project, contact the original authors, or start a discussion at lifelib Discussions.

Algorithm	Source	Description
Matrix on fraction	TBD	Heuristics for calculating transition matrices on fractions of power
G2++ with piec cons vol	TBD	Calibration of a G2++ model with piecwise constant volatility
Carter-Lee model	TBD	Simple stochastic mortality model
Metropolis-Hastings	TBD	Sampling of probability distributions