elle
(Elle)
#1
Chapter 10. Stochastics
Predictability is not how things will go, but how they can go.
— Raheel Farooq
Nowadays, stochastics is one of the most important mathematical and numerical
disciplines in finance. In the beginning of the modern era of finance, mainly in the 1970s
and 1980s, the major goal of financial research was to come up with closed-form solutions
for, e.g., option prices given a specific financial model. The requirements have drastically
changed in recent years in that not only is the correct valuation of single financial
instruments important to participants in the financial markets, but also the consistent
valuation of whole derivatives books, for example. Similary, to come up with consistent
risk measures across a whole financial institution, like value-at-risk and credit value
adjustments, one needs to take into account the whole book of the institution and all its
counterparties. Such daunting tasks can only be tackled by flexible and efficient numerical
methods. Therefore, stochastics in general and Monte Carlo simulation in particular have
risen to prominence.
This chapter introduces the following topics from a Python perspective:
Random number generation
It all starts with (pseudo)random numbers, which build the basis for all simulation
efforts; although quasirandom numbers, e.g., based on Sobol sequences, have gained
some popularity in finance, pseudorandom numbers still seem to be the benchmark.
Simulation
In finance, two simulation tasks are of particular importance: simulation of random
variables and of stochastic processes.
Valuation
The two main disciplines when it comes to valuation are the valuation of derivatives
with European exercise (at a specific date) and American exercise (over a specific
time interval); there are also instruments with Bermudan exercise, or exercise at a
finite set of specific dates.
Risk measures
Simulation lends itself pretty well to the calculation of risk measures like value-at-
risk, credit value-at-risk, and credit value adjustments.