Abstract
The change in collateral price is one of the challenges in modeling mortgage insurance. Current work mostly considers collateral price similar to addressing risky asset modelling, in which geometric Brownian motion is being used to model its underlying processes. This assumption has been heavily criticized due to its lack of fundamental dependencies in its distribution. This work provides empirical investigation towards valuing of loss in mortgage insurance while taking into account the dependency, i.e., memory in its underlying model. Findings suggest that the model with memory and stochastic volatility significantly affects the calculation of loss in mortgage insurance.