simpaq

python-based valuation tools for financial derivatives

simpaq contains valuation tools for financial derivatives. Initialy conceived as an Open-Source pricing tool for convertible bonds, support is for Options, Warrants, and Convertible Bonds. The original idea for this tool came from the fragmented and proprietary nature of quantitative financial tools. In spite of the fact that these models are widely published, bordering on commoditized, it is difficult to find well-documented packages that provide flexible, extensible, and modular valuation tools for complex derivatives.

The module consists of several key components, each extensible from its base form, designed to work together to provide a flexible set of pricing tools that operate seamlessly with one another. That said, the bespoke nature of many financial derivatives calls for bespoke pricing models, so there will always be some need for the user to adapt models to fit the features embedded in certain derivatives.

The "key components" mentioned above consist of the following, each component will maintain a common API across the sub-classes:

Assets - the basic building block of all financial securities, the Asset class has been extended to represent a variety of financial securities, from equity to convertible bonds.
Pricers - Pricers are the methods that can be used to determine relationships between a derivative and its underlying assets. This component includes numerical and analytic solvers for valuing assets in a risk-neutral world.
Features - These are the "embedded features," legal stipulations in the derivative contracts that alter the relationship between a derivative and its underlying asset.

Shown here is an example of the interaction of these components:

from simpaq.assets import Equity, ConvertibleBond
from simpaq.pricers import MCConvertPricer
from simpaq.features import SoftCall

equity = Equity(ticker='XYZ', name='Xylophone Group', price=10, vol=0.25, div=0)
cvt = ConvertibleBond(ticker='XYZ 0 5/30/2019',
                      coupon=0, 
                      maturity='5/30/2019', 
                      underlying=intc, 
                      par=1000, 
                      cr=50)
cvt.add_feature(SoftCall(price=130, price_type='relative', trigger=(20, 30)))
mcpricer = MCConvertPricer(m=10**6, n=500)
print cvt.calc_price(mcpricer, credit_spread=400, vol=35, dt=1/252., save=False, pretty=True)

Using the flag pretty=False, this simple code would return fair value and Greeks for the sample bond created above, formatted for human reading. By leaving pretty to its default value of True, the output can be returned as a tuple with value, greeks, and model parameters.

=======================================================================
    Fair Value Calculations for ABC 0 5/30/2019 Convertible Bond
    -----------------------------------------------
    Fair Value:     131.25      Delta:  0.75
    Gamma:          0.02        Theta:  0.01
    Vega:           0.21        Rho:    0.001
    -----------------------------------------------
    Valuation Date:	2016-10-5
    Versus Stock:	INTC
    Versus Price:	$45.55
=======================================================================

Summary
Components
A. Assets - Classes representing various types of financial securities.
B. Features - Classes representing non-standard features of derivatives.
C. Pricers - Actual implementation of the valuation models.
Examples
WebService
SQLAlchemy Integration
Contact
References

Summary

Components

The valuation models contained in this package operate as linked components that perform various aspects of the analysis. The components work together to separate work into logical groups based on their roles.

Assets represent securities and are responsible for aggregating market data, storing it, and passing data to pricers.
Features belong to a list-attribute of an Asset and serve as modifiers on the Pricer assigned to their parent asset.
Pricers are where the meat of the technical analysis lives, and serve as a means of calculating the fair value of derivatives.

Assets

The Asset-based classes represent the various securities. Securities - both derivatives and primary - are represented by subclasses of the Asset class. Each sub-class inherits its key features from the more generic classes, and contains a from the basic Asset class, Starting with the most basic security type Equity, Bond, and Derivative we combine classes to create derivative securities (Option, ConvertibleBond, etc)

Each asset has a list attribute called Features, which holds a list of Feature objects. These objects are discussed below, but the skinny is that they add functionality to base

Features

Features are modifiers on Assets that operate on both the pricer methods and the asset methods to create altered behavior. For example, adding a SoftCall feature to a ConvertibleBond being priced with a MonteCarlo bond pricer will enforce early conversion on Convertible Bonds when the soft-call provisions have been met. These features handle the complex and unique features that require special model design. A comprehensive list of Features and their respective APIs can be found in the Features page of this documentation.

Pricers

Long story short, pricers are the guts of this operation. Ranging from simple BlackScholes option pricers that can only handle a simple European option, to MonteCarlo simulations capable of pricing complex Options or ConvertibleBonds with path-dependent features. The following pricers are currently supported by this package:

These pricers (especially the MonteCarlo pricers) support a variety of assets, and should be passed into the calc_price() methods of each an asset. A list of assets and their supported pricers can be found in the Assets documentation.

Examples

Web Service

We plan to eventually host a running instance of these pricers on AWS and provide access to these models through our API. See the API Documentation for more details on interacting with the web service. This is a "down-the-road" todo, though, and

Access to the simpaq web service is free after registering for a token, subject to the following constraints:

We will not provide inputs or security terms, just models
Calls to value securities with resource-intensive Pricer models will be throttled to keep us from burning though our budget.
This service is provided under the MIT License, and thus is provided "as is", without warranty of any kind.

SQLAlchemy

Additionally, this package contains a set of SQLAlchemy classes that can be mapped to a relational database and used to are SQLAlchemy classes that correspond to each of the major models. The point of this is to allow Asset classes to be drawn from a database by their ticker, name, etc, without requiring the user to input all features of a security every time they are to be used.

This is a lower priority on the TODO list, so hopefully it can get implemented sometime in the near future.

Contact

If you have problems, questions, ideas or suggestions, please contact us @ exleym@gmail.com for now, or later at a better contact address. In a perfect world, this project will pick up enough steam to get a silly little website of its own, in which case, you can definitely leave messages there.

References

This project would never have even begun without the following brilliant reference material. Some of these models are directly replicated here, while others were a starting point or a learning tool for the development of the models shown here. In all cases, much thanks, and all the credit goes to the mathmaticians, physicists, and all-around smart people that came up with these models. We engineers are keen to enjoy the usefulness of your models without ever contributing to the frontier of knowledge. We hope these implementations can, in some small way, help with the continuation of knowledge in this sphere.

Hull, John. Options, Futures, and Other Derivatives. Pearson, 2016
Longstaff, Francis; Schwartz, Eduardo. "Valuing American Options by Simulation: A Simple Least-Squares Approach. The Review of Financial Studies, Spring 2001, Vol. 14. No. 1, pp. 113-147;
Lvov; Yigitbasioglu; Bachir; "Pricing Convertible Bonds by Simulation"

cliff007 / simpaq