As recently as 2003, 80% of all mortgage origination was agency-backed, with the remaining 20% non-agency. In 2006, non-agency origination, including jumbo-prime, Alt-A and sub-prime loans, accounted for over 55% of total origination. By any measure, there has been a dramatic shift from an agency dominated market to a non-agency majority. This shift has important ramifications for investors, issuers and risk managers involved in MBS & ABS.

1. Participants must now consider adding non-agency holdings to their asset or product mix.
2. Risk management and valuation of non-agencies requires the inclusion of credit components - delinquency, default and loss severity.
3. Prepayment models predicated on data that does not include the most recent shift in the roster of originators of non-agency loans will not adequately capture the prepayment profile of these loans.

How have we responded to this trend?

An entirely revamped suite of non-agency fixed, ARM and Hybrid prepayment models calibrated with loan level data from the largest issuers of non-agency loans from 2000-2005.

The launch of our new Loan Dynamics Model™ which produces required performance metrics such as CPR (prepayment), CDR (default), 60+ delinquency and loss severity. We are hosting a premier of this model at our ASF booth (709) on Monday, January 29th.

Over the last few years there has been a proliferation of sub-prime loan products in the marketplace. AD&Co, a leader in prepayment modeling, has developed models for both fixed and adjustable-rate sub-prime mortgages. The fixed rate models cover 15 and 30 year maturities, whereas the adjustable-rate model can take any first and periodic reset as input, and can adjust for interest-only and loans with prepayment penalties. The new models incorporate AD&Co’s vision to provide a single framework for fixed and adjustable-rate mortgages, across collateral types.

It is important to note that for developing the prepayment models, we did not distinguish between voluntary and involuntary prepayments, but looked at total terminations. A termination is recorded whenever the outstanding balance becomes zero. However, defaults are becoming more and more important in the sub-prime world. The new AD&Co suite of credit models addresses issues faced by market participants while modeling defaults. Click here to read the full article.

The Loan Dynamics Model™ is the culmination of a 2-year company-wide development effort. As a starting point, we examined historical data, including roughly 8 million loans from 144 issuers from 1990 to 2005. Over the course of our analysis, we found that borrower behavior can be described realistically yet parsimoniously using 4 categories of payment status:

Current (0-59 days past due)
Delinquent (60-179 days past due)
Seriously Delinquent (180+ days past due)
Terminated (default or prepay) Click here to read the full article.

The examination of mortgage securities is a multi-faceted challenge. Until now, the analysis of prepayments and the analysis of default risk, for the most part, have been treated as separate pair of problems. Our goal has been to integrate the two.

In studying the problem, we found that these two behaviors represent the end of a more complex process involving the dynamic transition of a borrower from current to delinquent to seriously delinquent. At each stage of the process, the borrower’s sensitivity to economic factors changes. Our model, therefore, simulates the transitions of the borrower among these statuses and the changing economic drivers that occur as a result.

We constructed the model utilizing the active-passive decomposition methods applied to our traditional prepayment models. Thus, it benefits from our experience and knowledge of prepayment modeling in producing a solution which addresses the whole of loan dynamics.

The Loan Dynamics Model™ is really just one step in the overall valuation process for mortgage securities. Valuation of securities involves an assessment of the possible future states of the world, including interest rates, home prices, other economic factors and the interaction among these pieces.

Many investors are interested not just in mortgage loans as investments, but in securities that have mortgage loans as collateral.
These securities, usually, have been structured into classes with varying degrees of credit enhancement. For these securities, it is necessary not only to evaluate the loans, but also how the cash flows of the loans flow through to the securities.

Financial engineers have also created derivatives using these securities [credit default swaps(CDS)] and portfolios of securities and derivatives [collateralized debt obligations (CDOs)]. Each of these adds new levels of complexity to security valuation. We look forward to applying our newest tool to these problems.