How to Estimate Credit Losses Under CECL: Best Practices

CECL fundamentally changes how credit risk is recognized. Unlike the older allowance for doubtful debts model that waited for losses to become probable, CECL requires you to estimate and recognize potential credit losses over the life of a financial asset, rather than waiting for them to be realized. 

Auditors now routinely request CECL documentation as part of their reviews to have early visibility into risk exposure. Let’s break down how you can leverage CECL for better financial health at your company.

What is Current Expected Credit Loss (CECL)

According to the CECL mandate, you now must estimate and record potential credit losses on day 1 upfront, even if there's no immediate indication of non-payment. This is a forward-looking approach based on:

• Historical data: How often do customers from this industry, or with this credit profile, fail to pay in the past?
  
• Current conditions: Are there any indicators that make this customer more likely to default, such as late payments, changes in their financial situation, cash flow issues, or low sales? Supply chain disruptions or demand shocks (due to COVID-19) can also increase default risk.
  
• Economic conditions: Recession, inflation, deflation, or policy changes can directly impact customers' ability to pay.
  
• Future forecasts: What is the outlook for this customer’s ability to pay, considering future economic conditions, industry trends, or any changes in the market? 

For instance, if Company A knows that customers in the same industry have a 20% chance of defaulting on payments over the next year, they would estimate that $2,000 of the $10,000 invoice could be uncollectible and record that loss immediately.

This is a proactive, data-informed estimate—so your provision aligns with actual credit risk earlier.

How is CECL different from the allowance for doubtful debts model

Under the allowance for doubtful debts model, businesses would only recognize credit losses once it became clear that a loss had occurred. In other words, businesses waited until it was almost certain that a customer wouldn’t pay before recording a loss.

For example, let's say Company B is waiting for a customer to pay a $10,000 invoice that is overdue by 90 days. Under the allowance for doubtful debts model, Company A would only record a loss if it was clear (or highly likely) that the customer couldn’t pay, such as if the customer declared bankruptcy or if other clear signs of non-payment appeared.

• Day 90 (Invoice overdue): No loss recognized yet.
  
• Day 120 (Customer declares bankruptcy): Finally, $10,000 loss is recorded, but by this point, the financial statements have been overstating the A/R for 3 months. 

Traditional prudence (or conservatism) says:

• Recognize losses early (even if uncertain)
  
• Delay recognizing gains (until they are certain)

But under the older model (and even under ASC 606), a contradiction appeared:

💡You recognize all revenue upfront if collectibility is probable, but delay recognizing potential credit losses until there's hard evidence.

That creates a prudence flaw:

• You're being optimistic with profits (revenue recognized early)
  
• But cautious with losses (credit losses delayed)

This leads to overstated assets and net income, especially when customers eventually default. 

The older model relies on evidence of risk, making it inherently reactive—you respond after risk becomes apparent.

Shortcomings of the allowance for doubtful debts model

1. Lack of proactive risk management

The allowance for doubtful debts model doesn’t allow businesses to anticipate potential credit losses before they become obvious. This reactive approach can lead to a lack of foresight in financial planning and credit risk management.

Example:

A company may have $1 million in outstanding receivables from 10 customers. Over the past few years, historical data shows that 5% of these customers tend to default on their payments. Under the allowance for doubtful debts model, the company would not recognize a credit loss until one of these customers stops paying or shows signs of financial distress (e.g., overdue by 90 days).

So, the company might continue to show $1 million in A/R, even though there’s a 5% risk that it may only collect $950,000.

• Expected Credit Loss = 5% x $1,000,000 = $50,000.
  
• Under the allowance for doubtful debts model, this $50,000 loss wouldn’t be recognized until one of the customers fails to pay.

2. Challenges in predicting future losses and missed opportunities

The allowance for doubtful debts model does not factor in forward-looking economic conditions, such as market downturns or changes in customer behavior. By waiting for clear signs of loss, businesses are exposed to greater financial risk and may miss opportunities for proactive action, like tightening credit policies or adjusting collection strategies.

Example:

If a recession is forecasted, businesses should anticipate a rise in credit losses but under the allowance for doubtful debts model, they wouldn’t recognize these potential losses until it's too late.

3. Adverse impact on financial ratios and credit ratings

Because credit losses are recognized after the fact, the allowance for doubtful debts model can artificially inflate a company’s financial ratios, such as the current ratio (current assets/current liabilities). These ratios may lead to misleading performance metrics and could negatively impact credit ratings when credit agencies or stakeholders realize that the company has underestimated potential risks.

Example:

Assume Company C has the following financials:

Without accounting for the expected credit losses, the company’s current ratio would look strong. However, if they had accounted for a $50,000 expected loss based on their historical data, the current ratio would appear lower, reflecting more accurate financial health.

How to recognize credit losses under CECL

CECL provides several methods for estimating credit losses, depending on your asset type and data availability. Here's a breakdown:

Method 1: Loss-rate method (most common)

This method applies a simple loss rate based on your past data. You estimate the percentage of A/R that won’t be collected based on your historical loss rates. It's perfect for businesses with a lot of similar customers.

Example:

Your company has a total of $100,000 in A/R and your historical data shows that 3% of accounts typically don't get paid. Using the loss-rate method, you can estimate your bad debt as:

So, based on your data, you would record $3,000 as the allowance for doubtful accounts.

Method 2: Roll-rate method (for accounts with clear trends)

This method uses historical trends (like how many payments were late in the past) to estimate future losses. It works great when you have clear patterns of customer behavior.

Example:

Let’s say Company B sells goods to a group of customers. After looking at your A/R aging report, you notice that 30% of accounts over 90 days overdue eventually default, and you want to apply that to estimate credit losses.

So, based on past data, you estimate $36,500 of your A/R will be uncollectible.

Method 3: Probability of default (PD) method (for complex accounts)

This method calculates how likely an account is to default and then multiplies that by the potential loss. It’s more detailed and works best for larger businesses offering big credit lines.

Example:

If you are dealing with a $200,000 contract and historical data shows the probability of default for this customer is 10%, the uncollectible amount would be:

So, in this case, $10,000 of the contract is likely uncollectible.

Method 4: Discounted cash flow (DCF) method (for complex, long-term contracts)

This method is a bit more complex. You calculate the present value of future cash flows expected from the asset (like payments on long-term contracts) and compare it to the asset's current value. It’s useful for complex financial assets like long-term contracts or loans.

Example:

Let’s say you’re evaluating a $500,000 loan that has a few payments scheduled in the future. You expect the loan to be repaid over the next 5 years, but there’s a chance some payments might be delayed or missed. You calculate the discounted cash flow (the value of future payments in today’s terms) and compare it to the loan's current value. This gives you the amount that could be uncollectible.

So, the uncollectible amount for this loan would be $50,000.

CECL estimation demands cross-functional teamwork

The challenge in estimating CECL lies in how to reliably recognise potential losses when no default has yet occurred.

• It’s a provision, not an actual loss. The customer balance doesn’t reduce naturally; a manual provision is recorded in anticipation for companies to prepare for loss.
  
• This requires judgment, assumptions, and data interpretation, which introduces variability and potential subjectivity.

The A/R team initiates the process and owns much of the underlying data.

• Prepares the aging report, analyzing how long balances remain unpaid.
  
• Reviews customer payment behavior—delays, partial payments, frequency of defaults.
  
• Applies aging logic
  
• Records the provision (allowance for doubtful debts) in the balance sheet.
  
• After estimating the CECL, the A/R team is responsible for populating the 2-3 pages long CECL document, including:
  
  • Historical default data
    
  • Aging analysis
    
  • Estimated threat percentage 

The credit team brings context to the numbers.

• Performs day-to-day credit risk analysis and sets credit limits.
  
• Has deeper insight into:
  
  • The customer’s business model and financial health
    
  • Macroeconomic factors like inflation, deflation, recession
    
  • Industry-specific risks (like liquidity crunch)
    
• This team evaluates economic outlooks and forward-looking risk inputs critical for CECL accuracy.
  

Auditors come in as the final checkpoint, ensuring the CECL estimate is reasonable and compliant.

• They review the CECL document prepared by the A/R team.
  
• Double-check the threat percentage and estimation logic.
  
• If there's variance between their calculation and the client's, they:
  
  • Discuss it with the client (often involving CFOs)
    
  • Evaluate whether the difference is material
    
  • Recommend adjusting the provision if required
    
• For significant variances, they may escalate to partners and validate assumptions with analyst-backed data.

For instance, if the A/R team submits a CECL document estimating a 20% credit loss (threat percentage) based on their internal analysis, the auditor will independently evaluate this figure using external benchmarks, historical data, and industry trends. You can expect to receive an analysis from the A/R team similar to the one below:

If the auditor’s analysis supports only a 10% credit loss, they assess whether the 10% variance is material. If it is, the auditor engages with the client—often including senior stakeholders like the CFO—to understand the assumptions behind the estimate and may request an adjustment to align with more supportable data. If the variance is deemed immaterial, the auditor may accept the client’s estimate but will still document the discrepancy and the rationale for acceptance to ensure compliance with CECL standards.

Together, these teams ensure that credit risk is properly quantified and reflected in financial reporting under the CECL framework.

Zenskar automates A/R tracking to assist accurate credit loss estimation

Current expected credit losses estimation requires continuous, periodic monitoring and reporting of credit risk—it’s not a one-time assessment. While certain inputs, like evaluating a customer’s financial health, still require human judgment, Zenskar strengthens the process by automating A/R management and tracking. This ensures that the foundation of CECL estimation—accurate, up-to-date A/R data—is consistently maintained with minimal manual effort.

Request a demo to see how Zenskar can automate a crucial aspect of your credit loss estimation.As part of the A/R team, you're on the front lines of ensuring that receivables reflect reality—not just expectations. One way to avoid overstating A/R is through the latest Current Expected Credit Losses (CECL) model, introduced by the Financial Accounting Standards Board (FASB) in 2016 and mandated for implementation in 2023.