Skip to main content

Looking for Valuant? You are in the right place!

Valuant is now Abrigo, giving you a single source to Manage Risk and Drive Growth

Make yourself at home – we hope you enjoy your new web experience.

Looking for DiCOM? You are in the right place!

DiCOM Software is now part of Abrigo, giving you a single source to Manage Risk and Drive Growth. Make yourself at home – we hope you enjoy your new web experience.

IFSLeaseWorks is now part of Abrigo.

Diversify your portfolio and earn additional interest income. End-to-end lease origination and administration automation make it possible.

Read the press announcement

Looking for TPG Software? You are in the right place!

TPG Software is now part of Abrigo. You can continue to count on the world-class Investment Accounting software and services you’ve come to expect, plus all that Abrigo has to offer.

Make yourself at home – we hope you enjoy being part of our community.

Asset/Liability Asset Liability Modeling Portfolio Risk & CECL CECL Models

Why CECL and ALM prepayment assumptions are not interchangeable

Dave Koch
April 30, 2026
0 min read

This article covers these key topics:

Why the CECL vs. ALM prepayment distinction matters more than most institutions realize

Many financial institutions use the same prepayment assumptions across CECL and asset/liability management (ALM). While this may seem efficient, it introduces hidden risk.

CECL and AML assumptions are often treated as equivalents

Prepayment behavior sits right at the intersection of credit performance and interest rate risk. It’s one of the few areas where accounting, lending, and balance sheet strategy all touch the same underlying loans, which is exactly where confusion tends to begin.

Most institutions are not trying to get this wrong. In fact, what you typically see is a reasonable process playing out. A CECL model is built using historical data, portfolio characteristics, and observed payoff behavior. Prepayment assumptions are developed, documented, and validated in that context. Over time, they become something the institution is comfortable relying on, and from there, it is a short step to reuse them.

If those assumptions are already supported and part of a controlled process, it feels efficient to carry them into ALM. Sometimes they are used directly. Other times they are adjusted. Either way, the underlying logic is the same. On the surface, logic is consistent and disciplined, but it introduces a deeper problem. When assumptions built for one purpose are used for another, the result is distortion that leads to unnecessary risk.

Learn more about asset/liability risks in this webinar, "Reassessing deposit behavior: Strengthening ALM assumptions in a changing rate environment."

WATCH

Two frameworks: CECL vs ALM prepayment assumptions

CECL asks: How much loss will we realize over the life of this asset?

CECL is an accounting framework designed to estimate expected lifetime credit losses. Prepayments determine how long a loan remains exposed to default risk. Once a loan prepays, it is no longer at risk of default. In a CECL framework, prepayments are really about exposure timing, not behavioral response.

ALM asks: How will borrower behavior change as conditions change, and how does that impact earnings and risk?

ALM evaluates how interest rate movements and market conditions impact earnings, value, and liquidity. Prepayments in ALM capture borrower optionality. Borrowers respond to incentives. When rates fall, refinancing accelerates. When rates rise, they slow. That behavior is not linear, and it is not stable. ALM prepayments are dynamic, scenario-driven, and designed to capture that behavior.

Where CECL prepayment assumptions break down in ALM

Many ALM models still rely on some form of industry or vendor-based prepayment assumptions. These are often designed to be broadly applicable, but they are not built around the specific characteristics of an institution’s portfolio. In that context, moving to CECL-based assumptions can feel like a meaningful step forward. Instead of relying on generic inputs, institutions begin using assumptions grounded in their own data and their own borrowers.

That is an improvement, but it is only part of the solution. CECL assumptions are still designed to estimate expected outcomes under stable conditions. When those same assumptions are used in ALM, they may be more institution-specific, but they are still not designed to capture how borrower behavior changes as rates and market conditions shift.

A similar misconception shows up in commercial portfolios. Prepayment penalties reduce activity, but they do not eliminate it. Borrowers still act when the economics make sense. At some level of incentive, behavior accelerates.

The hidden issue: Portfolio mix has changed

Historical CPR reflects a portfolio that may no longer exist. As higher-rate loans refinance and run off, what remains is a different population with different incentives and constraints.

Historical prepayment speed is partly a record of who has already left the pool. Prepayment reflects borrower decisions based on incentive and ability.

Using historical data the right way

The impact of lookback periods

Historical data remains essential, but how it is used matters just as much as the data itself. One of the most overlooked factors in prepayment analysis is the time horizon used to calculate historical speeds. Whether an institution looks back one year, three years, or five years can materially change the result, even if the methodology itself is consistent.

A five-year lookback period often includes multiple rate environments. It may capture both refinance waves and slower periods, which can produce a more stable average. But that stability can be misleading if the current portfolio or rate environment looks very different from earlier years included in that window.

Why “recent” data can still mislead

A three-year window tends to feel more current, but it can still be heavily influenced by prior rate cycles. If a meaningful refinance event occurred during that period, it can continue to shape the average long after conditions have changed.

A one-year lookback may feel the most relevant, but it is also the most sensitive to recent conditions. In a rising rate environment, it can understate prepayment potential. In a declining rate environment, it can overstate it.

Because of this, even when assumptions are refreshed regularly, the output is still anchored to a backward-looking window that may not reflect current borrower incentive or portfolio composition. This creates a subtle but important issue. The model appears dynamic because the number changes over time, but the logic behind it remains tied to past conditions.

What models are missing

What is often missing is a shift in perspective. Instead of asking:

“What has the conditional payment rate (CPR) been over the past X years?”

The more useful question is:

“How has borrower behavior responded to different levels of incentive, and where are we today?”

That shift moves the focus away from selecting the “right” historical window and toward understanding the relationship between incentive and behavior. In many cases, the difference between a one-year and five-year CPR assumption says more about the rate environment than it does about the borrower. 

Why it matters and practical next steps

The historical CPR reflects:

  • The mix of loans that existed at a point in time
  • The rate environment that drove that behavior
  • The timeframe used to measure it.

Change the lookback window, and the number changes. Change the portfolio mix, and the meaning of that number changes again. Even in areas where institutions feel confident, such as commercial portfolios with prepayment penalties, borrower behavior is dynamic. When prepayment is treated as an average, models tend to look more stable than when it is treated as a behavior, and risk becomes clearer.

For many institutions, the gap between CECL and ALM prepayment assumptions is not a lack of data, but how that data is used. Most institutions already have the information needed and simply need a structured way to translate that into forward-looking behavioral insight.

Better visibility into loan production and portfolio runoff

Understanding borrower behavior across different rate environments makes it easier for lending leaders to set realistic production targets, anticipate runoff, and prepare for refinance activity. Instead of relying on portfolio averages, lenders gain visibility into which segments are likely to move, and when.

For the CFO and ALCO, the impact shows up in how clearly risk can be seen and managed. More accurate prepayment behavior leads to more credible interest rate risk measurement, clearer liquidity expectations, and stronger alignment between strategy and risk. It also improves governance by making assumptions easier to explain and defend.

Abrigo’s ALM services support this approach by combining historical performance with forward-looking behavioral modeling and helping busy lenders move from observation to expectation. To gain clarity and find a better approach for your financial institution, start by asking, "Are we modeling what borrowers have done—or what they are likely to do next?"

Find out how dynamic ALM modeling helps you make better strategic decisions.

ALM software
Webinar

Reassessing deposit behavior: Strengthening ALM assumptions in a changing rate environment

Read More
Webinar

How to improve balance sheet strategy with integrated risk and financial management

Read More
Webinar

Scenario planning for liquidity and financial strength

Read More
About the Author

Dave Koch

Director, Advisory Services
Since 1989, Dave has delivered educational programs on Asset/Liability Management and pricing topics to Federal Regulatory Agencies, national and state industry trade groups, Federal Home Loan Banks, and Corporate Credit Unions nationwide.

Full Bio

About Abrigo

Abrigo enables U.S. financial institutions to support their communities through technology that fights financial crime, grows loans and deposits, and optimizes risk. Abrigo's platform centralizes the institution's data, creates a digital user experience, ensures compliance, and delivers efficiency for scale and profitable growth.

Make Big Things Happen.