By narrowly defining asset and liability management (“ALM”) as the process of monitoring, managing and anticipating the interest rate and liquidity risks that arise under various interest rate scenarios on a static balance sheet, Wilary Winn believes that the ALM processes employed by many community-based financial institutions failed in the past and will do so again in the future.
During the most recent Financial Crisis, covering the 2007 through 2009 time period, the ALM models utilized by most community-based financial institutions failed to identify the most significant risk facing them – increased credit risk. At best, credit losses caused depressed earnings and, at worst, caused multiple bank and credit union failures. In our experience, most community-based financial institutions incorporate credit risk into their ALM process by projecting a loss provision as a single line item. The loss provision estimate is typically based on the amount recorded for the previous year. Thus, in early 2009, community financial institutions were not modeling significant increases in loss provision due to credit conditions even though the environment had changed dramatically for the worst. Outsourcing of ALM to external vendors, did not improve modeling results. Most ALM vendors focus on providing investment advice. Thus, they typically rely on the financial institutions to provide guidance on the loss provision estimate, thereby providing no real additional insight from an ALM perspective as to the potential devastating impact that credit risk could have to the institution. Financial institutions and their ALM vendors were thus modeling the interest rate risk of cash flows which were never to occur. In short, we question how much could a move in interest rates really matter to an organization with significant levels of non-performing loans. Their issue was not an upward move in interest rates over the medium term – it was survival.
In contrast to the previously described method typically used by others to incorporate credit risk, Wilary Winn believes a financial institution’s risk measurement and management processes can be significantly improved by incorporating credit risk on a bottom up approach – that is developing potential credit losses at the loan cohort or even individual loan level. Wilary Winn includes an explicit default and loss severity assumption for each loan based upon individual loan attributes. The attributes from which default assumptions are derived include refreshed credit scores and updated property values for residential real estate loans. Loans with lower credit scores and higher loan-to-values are modeled with higher conditional default rate assumptions as well as higher loss severities. In addition to considering the credit risk on loans, Wilary Winn’s ALM processes also take credit risk into account for securities such as private label CMOs by including explicit default and loss severity assumptions, as these securities are likely to receive less than the full amount of the outstanding balance. Incorporating these losses obviously improves the measurement of interest rate risk because we are not including interest income that will not be received into our forecasts. More importantly, incorporating estimated credit losses into the forecasts results in better measurements of overall net income and capital.
We believe that Wilary Winn’s bottom up approach to forecasting credit losses as part of the ALM process meets the main objectives of Enterprise Risk Management (“ERM”). Enterprise Risk Management is an emerging best practice and an area of increased emphasis from the regulators. Simply stated, ERM means measuring and managing risks on an integrated basis across the institution. As an additional benefit, the Wilary Winn approach to measuring credit risk is in full accordance with the provisions of the proposed Current Expected Credit Loss (“CECL”) model and ensures our clients will be able to meet its requirements if and when it is adopted.
While we acknowledge that we have largely emerged from the recent credit crisis, we believe latent credit risk remains as we detail below.
We also believe traditional ALM processes are masking another significant risk – they are overstating net interest income in the rising rate scenarios. Net interest income simulation under various interest rate scenarios is a component of essentially every financial institution’s ALM process. Financial institutions differ with respect to the timeframe in which net interest income is measured (1 year, 2 year, 5 year, etc.) and on the amount of permitted variability (typically expressed as a percentage) from the base case scenario for various rate shocks. However, all institutions perform net interest income simulation in some form and net interest income simulation is typically based on a constant balance sheet assumption – where all asset and liability balances remain unchanged during the projection period for the net interest income simulation.
Wilary Winn believes that the constant balance sheet assumption will prove to be incorrect if and when interest rates rise significantly and that most financial institutions will be worse off in comparison to their ALM model projection for net income in rising interest rate scenarios. Our rational for this is primarily liability-based but is also partially asset-based.
With respect to the liabilities on the balance sheets of community-based financial institutions, the Financial Crisis caused a significant increase in non-maturity deposits as customers made additional deposits in a flight to quality. These “surge” balances are at increased risk should depositors seek a higher level of return elsewhere now that the crisis has abated. We calculate surge balances for clients by analyzing non-maturity deposit balances and growth trends prior to the Financial Crisis and comparing those balances projected forward based on a linear function to the actual balances which now exist.
We note that at the same time, time-based deposits have significantly decreased since the Financial Crisis as the low overall rates of today offer little incentive for customers to lock in to a low overall yield for an extended period of time.
Wilary Winn believes that market interest rates will eventually increase and when they do, we believe three things will happen to the liabilities of community-based financial institutions:
- Surge balances will flow out of non-maturity deposits
- Surge balances will flow primarily into time-based deposits
- Existing time deposit customers will incur early withdrawal penalties and re-invest at higher rates
Based on these three events, actual interest expense in the rising rate scenarios will be higher than the amounts most community-based financial institutions are projecting in their current ALM processes using constant balance sheet assumptions. The reason is the flow back to time based deposits. These deposits are more expensive in comparison to non-maturity deposits and the level of difference becomes greater in rising rates scenarios as time deposits move more in line with the overall level of interest rates than non-maturity deposits.
Although Wilary Winn calculates net interest income for clients under a constant balance sheet assumption, we also calculate scenarios that simulate surge balances flowing into time based deposits and include early withdrawals on certificates. The result is higher interest expense. With this analysis, we believe our clients are better prepared for rising market rates, including an unforeseen rise resulting from an adverse reaction the bond market could have to Fed policy changes.
As we indicated earlier, we believe that major credit risk lies largely behind us for the moment. That said, Wilary Winn believes a latent credit risk remains. We believe that the default rate on adjustable rate loans will increase in rising rate scenarios due to payment shock. As a result, Wilary Winn runs scenarios for its ALM clients which measure the potential negative impact that rising rates will cause on variable rate loans. We believe this risk could be particularly acute on HELOCs nearing the end of their interest-only payment terms. Many of these loans were made near the top of the housing market. Borrowers will face increased payments as they must begin amortizing their loans with limited chances to refinance given the loans’ loan-to-value ratios. The risk will increase in a rising rate scenario as the amortizing payment will be based on higher market rates. As previously stated, most internal or outsourced solutions ignore these potential effects as the ALM models are limited to a single line loss provision assumption.
Wilary Winn believes community-based financial institutions need to strengthen their ALM processes by understanding traditional ALM modeling’s recent failures and current assumption limitations. As the operating environment continues to change, community-based financial institutions that are able to better measure risk and develop strategies in advance will be far better off in comparison to their peers. Please let us know if we can be of help.