Fundamental solutions to Kolmogorov equations via reduction to canonical form (Q955495)

In many interesting applications of stochastic ordinary differential equations the knowledge of the fundamental solutions of the backward Kolmogorov equations can provide very useful information concerning the behavior of the system. Such applications include the theory of option pricing in certain models for stochastic volatility, short term interest rate dynamics and zero-coupon bond valuation, credit risk models etc. The present paper derives fundamental solutions to the backward Kolmogorov equations for a wide class of stochastic ODEs that include certain models with affine structure, models with power law diffusion coefficients (as for instance the constant elasticity of variance model) etc. The solution of the equation is obtained through reduction to a canonical form via properly selected transformations. The canonical form of the equation is either a constant coefficient parabolic PDE which is easily solvable or in the form of a heat equation with an inverse square law potential that can be solved using special functions (e.g. Bessel functions). A wealth of interesting examples is given, to illustrate the general theory. The results of the paper are very useful to researchers working in finance or applied probability in general as it provides a large number of analytic solutions for a range of models, that can be used as benchmarks for a variety of applications.