Magazine article Mortgage Banking

Recalibrating Loan Risk Assessment

Magazine article Mortgage Banking

Recalibrating Loan Risk Assessment

Article excerpt

Slowing home-price appreciation, the subprime credit shakeout and rising delinquencies have led lenders to revisit technology for underwriting and pricing credit, collateral, compliance, operational, prepayment and other risks. There is always discussion about automated underwriting systems (AUSes) and automated property valuations (APVs) using automated valuation models (AVMs), but what other technology for risk-evaluation and risk-pricing activities needs examination? This column examines the need for lenders to recalibrate their lending risk-assessment technologies.

Industry trends and risk assessment

Weaker home prices and rising delinquencies have led to revisions in loan underwriting guidelines. Zero-down-payment and negative-amortization loan programs are falling out of favor, as are reduced-document loan programs, because lenders can no longer rely on rising home prices as security for the loan if a borrower defaults.

The double-whammy for lenders is that lending volume will continue to decline as underwriting guidelines continue to tighten. The Mortgage Bankers Association (MBA) data show that mortgage lending volume declined at a 10 percent compound annual growth rate (CAGR) from 2003-2006, and will decline a further 4 percent CAGR from 2006-2009. Five-year trend data from the 2006 MBA Cost Study show that declining profit margins are directly correlated with declining lending volume. Net loan production financial income declined from $1,102 per loan in 2003 to an estimated $240 per loan in 2006.

So with lending volume and profits for the average lender expected to decline for the rest of this decade, what should individual lenders do to outperform the averages? The different types of risk that lenders need to re-evaluate and recalibrate include credit, loan product, pricing, collateral, hedging, interest-rate and prepayment risks.

Credit and collateral risk-assessment tools

Subprime lenders have relied excessively on credit scores for underwriting and pricing credit risk. Credit scores are great tools and inputs into a comprehensive underwriting process that includes an AUS, but most were developed to predict consumer loan credit losses 18 to 24 months after origination--not throughout the three- to five-year average mortgage term. Further, credit scores do not account for loss severity, loan product risk, loan-to-value (LTV) ratio or declining home prices.

In addition, there are new scores on the market, and the vendors themselves are recalibrating their models. Many lenders are evaluating Stamford, Connecticut-based VantageScore Solutions LLC's VantageScore[sm] to see how it performs versus the Classic FICO[R] score, especially for loan applicants with young (recent), thin (little) or subprime credit. In June, Minneapolis-based Fair Isaac Corporation announced Classic FICO 08, scheduled for release this September.

Lenders should rethink and refine their use of credit scoring in loan underwriting and pricing. By doing so, they can outsmart the competition and improve profit margins.

Automated underwriting systems. AUSes need recalibration as well, in particular to account for updated loss-performance data by loan product type. For example, payment-option adjustable-rate mortgages (option ARMs) not underwritten at the fully indexed accrual rate and with negative amortization are performing poorly.

AVM and AVM-assisted valuations. Excitement over usage of AVMs as the primary collateral-assessment tool has slowed as home-price growth has declined and the appetites of most mortgage investors for AVM-only loans have been lacking. Nevertheless, AVMs remain essential tools to spot-check appraisals in loan origination and revalue loan-portfolio risks.

Financial, product and pricing risk-assessment tools

The secondary marketing department remains the epicenter of most aspects of loan pricing, as well as market, interestrate, price, product and fallout risk management. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.