rippy 0.0.6

Reinsurance Pricing in Python

rippy

ReInsurance Pricing in PYthon!

Getting Started

Install rippy from pypi:

pip install rippy

Introduction

Rippy is a simple, fast and lightweight simulation-based reinsurance modeling package. It has two main components:

1. Frequency-Severity, otherwise known as compound distribution simulation

   Rippy contains the FrequencySeverityModel class to set up and simulate from compound distributions with a variety of frequency and severity distributions

   from rippy import FrequencySeverityModel, Distributions
   claims = FrequencySeverityModel(Distributions.Poisson(5),Distributions.Pareto(0.5,100000)).generate()
   

2. Reinsurance contract calculation

   Rippy can then calculate the recoveries of a number of types of reinsurance contract for the simulated claims.

   from rippy import XoL
   xol_layer = XoL(name="Layer 1", limit=1000000, excess=100000, premium=1000)
   xol_layer.apply(claims)
   xol_layer.print_summary()
   

Rippy is based on the scientific python stack of numpy and scipy for fast performance. It can optionally run on a GPU for extremely fast performance. It is designed for interoperability with numpy and ndarrays, so for example the simulated claims can be operated on with numpy ufuncs:

import numpy as np
capped_claims = np.minimum(claims,2000000)

Under the hood/bonnet, rippy represents the simulations of the Frequency-Severity model in sparse storage, keeping two lists of simulation indices and values:

sim_index	values
0	13231.12
0	432.7
2	78935.12
3	3213.9
3	43843.1
...	...

Frequency-Severity simulations can be aggregated (summed within a sim_index), which results in a standard np.ndarray

aggregate_claims = claims.aggregate()

np.array([13663.82,0,78935,12,47957.0,....])

Configuring the simulation settings

The global number of simulations can be changed from the config class (the default is 100,000 simulations)

from rippy import config
config.n_sims = 1000000

The global random seed can also be configured from the config class

config.set_random_seed(123456)

Rippy uses the default_rng class of the numpy.random module. This can also be configured using the config.rng property.

Using a GPU

GPU support requires a CUDA compatible GPU. Internally rippy uses the cupy library. Install the dependencies by running

pip install rippy[gpu]

To enable GPU mode, set the RIPPY_USE_GPU environment variable to 1.

export RIPPY_USE_GPU=1

on Linux or

set RIPPY_USE_GPU=1

on Windows. Set it to anythin else to revert to using a CPU

Project Status

Rippy is currently a proof of concept. There are a limited number of supported distributions and reinsurance contracts. We are working on:

• Adding more distributions and loss generation types
• Adding support for Catastrophe loss generation and reinsurance contracts
• Adding support for more reinsurance contract types (Surplus, Stop Loss etc)
• Grouping reinsurance contracts into programs and structures
• Stratified sampling and Quasi-Monte Carlo methods
• Reporting dashboards

Issues

Please log issues in github

Contributing

You are welcome to contribute pull requests