The major point of the modeling is to determine whether the cash flow from the mutual fund fees can cover the repayment of the notes (the principal and the interest). In the static model, the actual shortfalls or excess cash flows are put to the output pages. In the dynamic model, the outputs include both indicative and quantitative measures of the excess/loss. According to the Structured Asset Securitization Group, when the dynamic model is used only the indicative output values are considered important.

The VRM MBS calculator treats conversions using a constant annualized conversion rate that is applied as a monthly mortality rate. This treatment of conversions ignores the impact of ageing, seasonality and interest rate volatility. However, since the VRM coupon resets each month, we expect ageing/seasonality/ interest rate volatility to impact the price marginally; therefore, the extra spread to be paid to investors because of the early principal payment, especially due to interest rate volatility, is marginal as well.

A model of mortgage prepayment rates based on the neural net approach is proposed. The model for insured, closed, five-year term mortgages has been developed. The neural net prepayment model behaves consistently across the training and testing sets and outperforms a simpler predictor, the linear regression model.

The proposed enhancement computes the constant liquidation rate for which the remaining principal balance at the month before the first maturing principal tranche is the same as remaining principal balance (RPB) using variable liquidation rate (Standard Vector). The MBS PE then also displays a warning message that a different constant equivalent liquidation rate was computed.