In the model above we have assumed that prepayments are at 100% of PSA. Expressing prepayments in terms of the PSA model is an industry standard but what are the primary drivers for prepayments?
The principal drivers for prepayments are housing turnover and refinancing. There are other causes of prepayments too such as defaults, curtailments or partial prepayments and full payoffs.
We will discuss each of these briefly below:
When a home owner sells his house he uses the proceeds to pay off the outstanding balance on the existing mortgage. There are a number of social and economic reasons why a homeowner may choose to sell his existing home such as family relocation, trading up or down based on changes in the level of income, interest rates and home prices, divorce, death, children leaving home etc.
The factors impacting the decision to sell include:
- The mortgage rate on his existing loan in comparison to those prevalent in the market. If the latter is higher relative to the former the home owner will most likely not prepay his mortgage. This is known as the lock in effect. The effect is strengthened if a new buyer can assume the existing loan at the existing rate provided of course that the differential between outstanding loan and house price is not substantial, i.e. a high LTV. The effect is diminished due to amortization of the loan over time and house price appreciation which both serve to reduce the Loan-to-Value (LTV) ratio. A lower LTV enhances the likelihood of prepayment regardless of the whether the current rates exceed the existing rates or not.
- The cumulative house price appreciation since the origination of the loan. The higher the appreciation is the greater the likelihood of prepayment as homeowners will sell their houses to realize price gains. The impact of house price appreciation is intensified the greater the proportion of refinanced to purchased loans in the pool as the refinanced borrowers have purchased their homes a while back and are more likely to have experienced appreciation in the prices of their homes, hence making them more likely to prepay.
- The age of the loan or loan seasoning. New home buyers will be less likely to sell in the near future given the transaction and moving costs they experienced in the recent purchase. However the age of the loan could be deceptive because the pool of mortgages may consist of both purchased loans, which will exhibit the characteristic mentioned above, and refinanced loans. In the case of the latter, though the loan is newly acquired the homeowner has been living in the existing house for a number of years and therefore is more likely to consider moving. In general the prepayment rate increase gradually as the loan seasons and then levels off. Under the 100% PSA prepayment model this “seasoning ramp” indicates a 0% prepayment rate in month 0 which increases by 0.2% each month till it finally levels off at 6% after 30 months.
- The monthly seasonal pattern of home sales. This is dictated by the borrowers’ mobility primarily due to factors such as the school year and weather. Home sales tend to peak when school is out and slack during winter months.
- The points paid effect. This refers to the amount of down payment that the borrower has made at the origination of the loan. If the borrower has made a high down payment he will have extra points which will entitle him to a lower contract rate which in turn means a greater resistance to prepayment.
- The loan type. For example, a balloon mortgage tends to prepay faster than conventional mortgages. This is due to the fact that borrowers who select a balloon mortgage expect to move again soon as compared to those who select the conventional form.
Modeling of the housing turnover rate involves the projection of existing home sales numbers considering the affordability (as measured by the ratio of median income to median monthly payment on a median home) against the desirability (as measured by the prospective inflation in home prices) of such a move to the homeowner; the change in interest rates and the appreciation in house prices; the impact of seasonality (usually by the application of monthly seasonal multipliers to base prepayment rates); the proportion of refinanced to purchased loans in the pool; and loan characteristics such as the loan age, points paid effect and loan type.