BAS Case Study - March 2023

Board of Directors

Page2 June 8, 2017

IRR Measurement, Assumptions, and Modeling Principles

Inspection of the Bank's March 2016, June 2016, September 2016, and December 2016 IRR Reports revealed the methodology employed to measure, monitor, and report on a quarterly basis that the Bank's sensitivity to IRR appears reasonable. The integrity of the information used to generate the December 2016 report was then challenged. Through comparison of the general ledger, call report, and other supporting documentation of the prepared report, we noted the amounts used to generate the NII and EVE reports appeared reasonable. In our review of December 31, 2016, Interest Rate Risk Monitor Report prepared by The Baker Group, we noted it appears that all major instruments are included in the model, such as Spread/Price Assumptions, Prepayment Assumptions, Interest Rates, Deposit Sensitivities, Decay Rates, etc. We performed a review of the model using the December 2016, Interest Rate Risk Monitor Reports. The Baker Group utilizes their IRRM modeling software to model the Bank's assets and liabilities. Management uses the model to measure IRR in several different manners. The primary method is to stress-test the balance sheet as of the reporting date against instantaneous and sustained rate shocks. The intervals used are +400 basis points for the rising rate environment and -100 basis points for the falling rate, assuming various intervals and minimal asset growth. These ranges are subject to change based on the current level of market rates and appear to be in compliance with the interagency guidance. These generally reflect' the changing rate environment reflected in the policy. The majority of the data sources used to update the model are based on electronic reports from the Bank's core processing system. Current loan portfolio data is provided by the Bank's internal operating systems. Repricing information and cash flow replacement assumptions are also built into the model. It was noted that the loan portfolio is composed of fixed and variable rate loans, which present some modeling challenges with respect to ensuring the data in the model takes into consideration rate changes, rate in-dexes, caps, floors, and reset dates. We noted, through review of the data download provided to The Baker Group, the files appear to capture the loan data correctly. To ensure the quality of the data used in the model, we reconciled the December 31, 2016, general ledger and call report to the model reports on a select basis. We also reconciled the totals from the various downloaded files to both the general ledger and the model. There were a few differences for some of the accounts, since the general ledger is not going to tie to the model exactly due to certain groupings of assets and liabilities. These differences were immaterial. In the end, the total assets and liabilities in the model were in agreement with the general ledger. The inputs regarding pricing appear reasonable and conservative, since variable rate loans are tied to a key driver rate established for each account and adjusted for the assigned spread. The majority of the pricing correlations in the model are very straightforward.

Made with FlippingBook Online newsletter creator