NII simulation is a modelling technique that looks at IRR through an EAR construct. It projects the changes in asset and liability cash flows, expressed in terms of NII, over a specified time horizon for defined interest rates scenarios. Credit unions use income simulations to forecast NII under varying interest rate scenarios to learn about the sources and levels of interest rate risk inherent in their balance sheets. NII simulation analysis allows a credit union to learn which rate scenarios pose the greatest potential threat to its expected earnings stream and to identify which of its assets and/or liabilities are the source of potential earnings volatility under different scenarios.
NII simulations generate insight into the impact of changes in market rates on earnings and guide risk management decisions. Because the levels of future market rates are unknowable, practitioners utilize models to simulate potential outcomes under varying scenarios. Credit unions should simulate a variety of scenarios such as base case, instantaneous parallel rate shock, and alternate interest rate changes to broadly assess different IRR threats that can occur.
NII simulations provide a comprehensive way to stress plausible near-term earnings results. Understanding NII volatility is important because credit unions must always endeavor to do the following: maintain stable earnings, build adequate net worth, and ensure they can smoothly fund normal operations without disruption. Because IRR is inherent to the business model of a credit union, and because certain IRR exposures can materially threaten earnings and net worth, NII simulations provide an important means to comprehend and manage this risk. Simulation results help identify balance sheet-related vulnerabilities, inform management about threats to the earnings stream, and highlight the need for possible risk mitigation strategies. Thus, a credit union’s earnings simulation model provides valuable information: a formal estimate or baseline of future earnings and an evaluation of how earnings would change under different rate scenarios. Together with historical earnings trends, a credit union’s estimate of the IRR sensitivity of its earnings derived from simulation models is an important indication of the potential exposure to changes in rates.
Analyzing the historical behavior of the net interest margin, including the yields on assets, liabilities, and off-balance-sheet positions that make up that margin, can provide useful insights into the relative stability of a credit union’s earnings. It also provides useful empirical data against which simulation model outputs can be compared for reasonableness.
A limitation of NII is that the simulation horizon periods are typically too short to fully measure the income changes resulting from embedded options in longer-dated cash flows (such as those cash flows that occur beyond the horizon of the simulation period). This is one reason why interagency guidance recommends using NEV in conjunction with NII simulations to gain both a short- and long-term perspective on IRR (especially when there are material amounts of longer-term liabilities and assets containing optionality in cash flows that are beyond the simulation horizons).
A key aspect of NII and income simulation modeling involves selecting an appropriate time horizon (or horizons) for assessing IRR exposures. Typically, the forecast horizon for income simulations spans 1–3 years from the run date forward. Simulations allow modelers to produce multiple variations of possible rate moves and identify those scenarios that adversely impact NII (or earnings) compared to the credit union’s base case assumption. Base case represents projected earnings based on the current balance sheet with no change in interest rates.
Simulations can be performed over any period and are often used to analyze multiple horizons identifying short- through longer-term risks. As the simulation time horizon extends, however, the reliability of results diminishes due to uncertainties surrounding how principal and income cash flows are reinvested. Reinvestment assumptions introduce predictions about decisions regarding future sources and uses of funds. Because income simulations are not a point-in-time measure, estimates of future cash flows and holdings must be assumed. As the level of reinvestment decision estimates increases, confidence in the model’s ability to predict diminishes, making simulations with longer horizons less dependable as a risk management tool.
Operational management should recognize that the results of NII might differ substantially between short- and longer-term time horizons. NII is more reliable for short- to medium-time horizons (anywhere from 12 to 36 months) and becomes increasingly uncertain beyond that. It is beneficial to apply 12- to 36-month time horizons to gain a perspective on the short-term versus medium-term risk exposures. The timing of cash flows is significant so a sufficient scenario time horizon is always an important consideration in capturing IRR in income simulations. For example, a credit union may have shifting loan concentrations, rapid share/deposit growth, and/or other strategies the risk of which is materially understated if only viewed over a short time horizon (such as 12 months) that fails to capture relevant longer-term cash flows.
Thus, a credit union modeling a 12-month horizon NII simulation, with a significant amount of adjustable-rate mortgages (ARMs) that reset in 24 months, would not properly capture the IRR aspects of these products (such as repricing, basis, and yield curve risk) in year two.
Some practitioners will address the IRR arising from longer cash flows by lengthening the maturity horizons of their simulations beyond the 1–3 year standard. Medium- to longer-term NII simulations of up to 60 months may be used by some credit unions with material concentrations of assets and liabilities with embedded options. An extended simulation may be able to identify when longer-term mismatches occur (for example, NII can show that a credit union is liability sensitive in years two, three, and four, but asset sensitive in years five, six, and seven), whereas NEV will only aggregate the effect of such mismatches because NEV is a present value methodology that expresses each asset and liability as a single economic amount at a single point in time.
While the confidence in longer-horizon simulations does decline as the horizon extends, their use can provide important insight into the timing of actual cash flow mismatches and help modelers determine when the risk is likely realized in the expected earnings stream.
It is not uncommon for practitioners to run multi-factor scenarios that alter the level and/or timing of market rate changes as well as varying key underlying assumptions. A credit union may vary its simulation rate scenarios based on factors such as pricing strategies, balance sheet compositions, and/or hedging activities. NII simulations may also measure risks presented by non-parallel yield curve shifts (shifting the shape of the yield curve by altering the spread between short rates and long rates). Credit union policy limits are generally only tied to parallel rate shocks results with the additional scenarios being generated only to provide additional information and influence decision making. Credit unions with complex balance sheets are encouraged to produce these alternative NII scenarios, especially ones that involve stressing key drivers of risk (for example, varying prepayment options, non-maturity share assumptions, and key rates) and shifts in the yield curve. When credit unions produce and include such information in their strategic and/or risk management discussions, it reflects favorably on the quality of their process.
Other scenario variations include whether to simulate for changes in the balance sheet. Credit unions can run static or dynamic simulations.
- Static simulations are based on current exposures and assume a constant balance sheet with no new growth. The static simulation approach is focused on gaining an insight into the current portfolio risks. These NII simulations can also include replacement-growth assumptions, where replacement growth is used to offset reductions in the balance sheet during the simulation period.
- Dynamic simulations may assume asset growth, changes in existing business lines, new business, or changes in management or member behaviors. Dynamic simulations can be useful for business planning and budgeting purposes. However, these simulations are highly dependent on key variables and assumptions that are difficult to project with accuracy over an extended period.
When management changes simulation scenarios, it may lose insights on the credit union’s current IRR positions (the risk inherent in the present book of business). Dynamic simulations do provide beneficial information, but also introduce added complexity due to a layering of assumptions. When layering multiple assumptions within a scenario, the modeler is introducing an increasing number of predictions about balance sheet changes and management action. This adds to the risk of growing imprecision and, therefore, potentially misleading results.
It is important for modelers to distinguish between those scenarios designed to highlight risk and those which represent pro forma scenarios used for income budgeting. Projected growth assumptions in dynamic simulations often alter the balance sheet in a manner that reflects reduced IRR exposure. For example, if a liability-sensitive credit union assumes significant growth in one-year ARMs or long-term liabilities that fail to materialize, initial simulation results may have understated the actual exposures to changing interest rates due to reliance upon an overly optimistic budget estimate as opposed to thoughtful scenario analysis designed to capture risk.
Therefore, when performing dynamic simulations for risk management purposes, credit unions should also run static or no-growth simulations to ensure they produce an accurate, comparative description of the credit union’s current IRR exposure. When performing dynamic simulations, credit unions should also run static simulations to provide the ALCO or senior management a complete and comparative description of the credit union’s IRR exposure.
Again, the underlying objective of NII simulations is to generate forward-looking information that demonstrates how changes in market rates impact expected earnings and to challenge management’s expectations and guide its risk management decisions. One of the key aspects of a sound NII analysis process is having the ability to generate a spectrum of varying outcomes that informs the practitioner about which underlying assumptions serve as the key drivers of risk.
As discussed above, NII simulations may not capture certain risks when the simulation horizon periods are too short to fully measure the income changes resulting from embedded options in longer-dated cash flows. As a result, NEV should also be used to broaden the assessment of IRR exposures, particularly in relation to net worth.
Workpapers & Resources
- FFIEC Advisory on Interest Rate Risk Management (January 6, 2010)
Last updated October 11, 2016
