Asset Liability
Management (ALM)

Our ALM engagement begins with a thorough review of our client’s ALM policies, procedures, Asset Liability Committee (ALCO) minutes, and reporting packages. We then have discussions with our client’s senior management team to better understand its goals, objectives and tolerances for interest rate, liquidity, and credit risk. Our next step is reviewing our client’s financial statements for at least the past 10 years to understand the institution’s financial performance. Only after these initial steps do we begin a thorough review of our client’s existing ALM model because we believe we must first have the appropriate context for our analysis. Prior to producing a model for the new reporting period, we obtain inputs from our client’s most recent ALM reporting package and work to successfully replicate the prior results. This step ensures that we have not omitted critical input assumptions in our model review.

We work with our clients to enhance their ALM model input assumptions as we develop the model for the new reporting period. For example, we believe that modeling should specifically address credit risk, and our input assumptions include incidence of expected default (conditional default rate or CDR) as well as the severity of the loss to be incurred on a default. Many valuation firms model the credit risk globally by adjusting the allowance for credit losses. Instead, we believe the credit risk should be “built from the ground up” by considering credit indicators such as the credit score of the borrower and the estimated loan-to-value (LTV) ratio as of the valuation date. We believe that knowing and utilizing these attributes results in a more accurate forecast of voluntary prepayment risk as well. For example, a borrower could have an incentive to refinance, given the interest rate on their loan compared to market interest rates, but they may not have the ability to prepay given their credit score and the estimated LTV ratio as of the valuation date.

We selected the ZM Desk Model because it allows us to model credit risk using CDR and loss severity at the loan or cohort level as appropriate. It also allows us to leverage our existing knowledge and valuation software. For example, we utilize ZM Desk’s direct interfaces with Bloomberg to model “vanilla” investments and with Intex to model investments with complex cash flow waterfalls such as Agency collateralized mortgage obligations (CMOs) and Non-Agency mortgage-backed securities (MBS).

Wilary Winn believes that the ability to measure risk on an integrated basis by explicitly including expected credit losses leads to superior risk measurement and reporting. We further believe that the ability to predict credit losses at the appropriate level of granularity provides for more reliable capital stress testing, concentration risk analyses, budgeting and balance sheet optimization, and optimal loan pricing.