Prepayment Risk Model

Definition

A Prepayment Risk Model is a specialized quantification framework that aims to estimate the range of future realizations of Prepayment in a variety of loan or other fixed income portfolios and thereby help manage Prepayment Risk.

Conceptual Approaches

There are various approaches to modeling prepayments, depending on the context and use of Model Outputs.

Market Instrument Approach

Prepayment rates are determined in such a way that valuations of prepayment sensitive securities equal the prices observed in the market for such securities. This approach is useful in pricing derivatives. The estimated parameters are Risk Neutral and therefore not directly usable in Risk Measurement.

Option Theoretic Approach

Borrowers are assumed to use optimal exercise strategy of their Prepayment Option. This approach treats borrowers (option holders) as rational agents exercising their prepayment option optimally (factoring in any constraints and external conditions). Such a rationality assumption is (in general) the more appropriate the more sophisticated the economic agent, but even professional entities may not exercise options optimally on pure fianancial value grounds but may take into account other less tangible factors such as Reputation Risk. Under those caveats, option theoretic approaches can be a useful means to generate prepayment estimates in the absence of alternative data sources.

Empirical or Statistical Approach

Models that aim to estimate prepayments behavior from (macro)economic and potentially other social data on the basis of past (historical) realisations and under the assumption that these behaviors are stable. Within a statistical approach there are further distinct families of models that can be used together or as alternatives

• Portfolio (Pool) based Models
• Prepayment Scoring Models

Model Elements

In general, the following elements might appear in prepayment risk models. The precise manner is model dependent.

• Valuation of contractual terms (cash flows)
• Macroeconomic variables (housing prices, interest rates, economic conditions), both current, historical and forward expectations
• Borrower characteristics (socioeconomic indicators, geographical)
• Seasonality