--- title: "Reading Bank Loss Reserves: Provision, Allowance and Adequacy" description: "How the allowance for credit losses moves with the provision and charge-offs, and how to judge reserve adequacy from call report and UBPR data" canonical_url: https://www.bankregreports.com/blog/what-is-cecl/ source: BankRegReports published: "2026-06-17" updated: "2026-10-28" --- In Q1 2020, JPMorgan Chase posted an $8.3 billion provision for credit losses, of which $6.8 billion was a reserve build, up from roughly $1.4 billion the prior quarter. The bank's loan portfolio had not yet deteriorated. The charge came from CECL forcing management to book their forward estimate of lifetime losses on the entire portfolio the moment the economic outlook shifted. That is the mechanism in a single data point: reserves move on expectations, not evidence. [CECL](/glossary/cecl/), the **current expected credit losses** standard, is the most significant change to bank loss accounting since the savings-and-loan crisis. Every analyst reading a bank's [call report](https://www.bankregreports.com/blog/what-is-a-bank-call-report/) needs to understand how the allowance for credit losses works under it, because the numbers mean something different now than they did before 2020. ![Allowance for Credit Losses to Loans - Community Banks, Q3 2021 to Q1 2026](figures/what_is_cecl-1.svg) The 2020 reserve build was a pandemic response, and CECL set its speed. It was not the CECL day-one charge: JPMorgan adopted the standard on January 1, 2020, and booked that one-time adjustment to retained earnings, not to the provision. The first-quarter build was the first time a sharp change in the economic forecast ran straight through a large bank's reserve under the new standard. ---
Loan-loss reserve coverage by bank asset size
Allowance coverage of loans by bank asset size (FDIC).
## What CECL Actually Requires CECL stands for **Current Expected Credit Losses**. FASB introduced it in Accounting Standards Update (ASU) 2016-13, with the model codified at ASC 326. The requirement is direct: when a bank originates a loan, it estimates the credit losses it expects to incur over the **entire contractual life** of that loan and records that estimate immediately as a reserve. The estimate must incorporate historical loss experience, current conditions, and "reasonable and supportable" forecasts of future economic conditions. Not a probability-weighted worst case. An expected value, over the full term, booked on day one. The reserve that results is the **allowance for credit losses (ACL)**, which sits as a contra-asset against gross loans on Schedule RC of the FFIEC 031/041/051. The provision for credit losses (RIADJJ33, Schedule RI item 4, on the call report) is the income-statement expense that funds it each period. [FDIC guidance on CECL](https://www.fdic.gov/accounting/current-expected-credit-losses-cecl) has a clean summary of the mechanics if you want the regulator's framing. --- ## What CECL Replaced Was Genuinely Broken The prior model, the **incurred-loss standard**, required a bank to reserve for a loss only once it was probable the loss had already been incurred. In practice, that meant waiting for observable evidence of credit deterioration before booking any reserve. A performing loan with no delinquency signs carried almost no allowance, even if it was underwritten loosely at the top of a cycle. The problem showed up clearly in 2007 and 2008. Reserve levels entering the crisis were thin. As charge-offs mounted, banks had to build allowances rapidly and simultaneously, which compressed earnings and capital precisely when they could least afford it. The pro-cyclical dynamic amplified the stress. FASB spent years on CECL partly in direct response to that failure. The conceptual difference is worth stating plainly: | | Incurred-loss (prior) | CECL (current) | |---|---|---| | When reserves are built | After losses appear | At origination, over the loan's life | | Forecast incorporation | None | Required: current conditions + projections | | Cycle behavior | Reserves lag deterioration | Reserves build during good times | | P&L sensitivity | Low in good times, spikes in stress | Higher baseline volatility, less cliff risk | The trade-off CECL makes is more provision expense volatility in exchange for fewer surprise reserve builds in downturns. Whether individual banks execute that well depends entirely on model quality and the reasonableness of their economic assumptions, which are not disclosed in any detail in public filings. --- ## The Accounting Mechanics: Provision, Allowance, Charge-offs Three line items do all the work: **Provision for credit losses (RIADJJ33)**: the periodic income-statement charge a bank records to build or adjust the ACL. This is what hits earnings. A bank growing its portfolio aggressively in a deteriorating economy will run elevated provision expense even if no loans have gone bad yet. **Allowance for credit losses (RCON3123)**: the cumulative balance-sheet reserve against expected losses. It sits as an offset to gross loans. The allowance does not reduce the contractual principal owed; it reduces the carrying value of the loan book on the bank's books. **Net charge-offs**: when a loan is deemed uncollectible, the bank writes it off against the allowance. Recoveries on previously charged-off loans flow back in. The net of these two is the net charge-off figure. The identity is: beginning ACL + provision − net charge-offs = ending ACL. A bank can hold its ACL flat by setting provision roughly equal to net charge-offs, or it can build reserves by provisioning above charge-offs, or it can release reserves by doing the opposite, which boosts earnings but may signal that management thinks credit conditions are improving. Reserve releases are worth watching. They are accretive to earnings but reduce the cushion. Several large banks released substantial reserves in 2021 as the pandemic credit losses they had anticipated did not materialize at the scale modeled. --- ## When Banks Had to Adopt CECL Adoption was phased: - **Larger SEC-filing institutions** (non-smaller reporting companies) adopted for fiscal years beginning after December 15, 2019, effectively January 1, 2020 for calendar-year banks. - **Smaller reporting companies, private banks, and other entities** received an extended deadline and adopted for fiscal years beginning after December 15, 2022, effectively January 1, 2023. The stagger matters for peer analysis. Through 2022, you had a mix of CECL and incurred-loss reporters in the same peer group. A community bank's ACL/loans ratio of 1.1% in mid-2022 was not directly comparable to a CECL-reporting peer at 1.5%: the methodologies were different, not the credit culture. The [Federal Reserve's CECL FAQ](https://www.federalreserve.gov/supervisionreg/topics/faq-new-accounting-standards-on-financial-instruments-credit-losses.htm) covers the capital transition elections that accompanied adoption. Today, adoption is complete across the industry. All call report filers are on CECL. --- ## How to Read Reserve Adequacy Four measures do most of the work in an asset quality review: **ACL to total loans**: the headline coverage ratio. Calculated from RCON3123 ÷ RCON2122. The industry median has historically run in the 1.0–1.5% range in calm periods and peaked near 1.6% after the 2008 crisis, while the dollar-weighted industry ratio reached about 3% in 2010. What matters more than the level is the trend relative to peers and relative to the bank's own loan mix: a CRE-heavy portfolio should carry more reserve than a plain-vanilla residential mortgage book. **ACL to nonperforming loans**: how well the allowance covers the loan book already showing distress. A ratio above 100% means reserves exceed the balance of nonperforming loans. A ratio below 100% is not automatically a problem, but it warrants understanding why: strong collateral coverage is a common legitimate explanation. **Provision to average loans (annualized)**: what the bank is spending to maintain or build its reserve, relative to the size of the portfolio. Watch this alongside loan growth. If loans are growing at 15% and the provision rate is flat, the allowance is not keeping pace with the risk being added. **Net charge-off rate**: the losses actually realized, which the allowance is supposed to anticipate. A bank consistently provisioning well below its realized charge-off rate is drawing down its cushion, which will eventually require a catch-up provision hit. All of this connects directly to asset-quality frameworks like the [Texas Ratio](/glossary/texas-ratio/), which adds the allowance to tangible common equity in its denominator. A bank that has systematically under-reserved can look healthier on that ratio than it actually is. --- ## Where the CECL Numbers Live For CECL analysis, the allowance-to-loans ratio is RCON3123 ÷ RCON2122, and the provision rate is RIADJJ33 annualized over average RCON2122. (RIAD4230 is the older loans-and-leases-only provision, now reported in the Schedule RI-B allowance rollforward.) **One note on prefixes before you pull anything.** The same call report item carries a different prefix depending on which form the bank files. FFIEC 041 and 051 filers report `RCON`, for domestic offices. FFIEC 031 filers report `RCFD` on a consolidated basis, and `RCFN` for foreign offices, so total deposits for an 031 filer is RCON2200 plus RCFN2200 rather than RCON2200 alone. Schedule RC-R Part I (regulatory capital and the capital ratios) moved off `RCON` with the Basel III revisions of 2014 and 2015 and now uses `RCOA` on the 041 and 051 and `RCFA` on the 031; Part II (risk-weighted assets) still reports under `RCON` and `RCFD`. Query a single prefix and you will silently drop a whole population rather than get an error, and for an RCON-only query that population is the largest banks in the country. ![Provision and Net Charge-Off Rates - Community Banks, Q4 2021 to Q1 2026](figures/what_is_cecl-2.svg) The divergence between provision and actual charge-offs in 2020, before this chart's window, was CECL working as designed, or at least as intended. Whether the models that generated those estimates were actually good is a separate question, and one the current data cannot fully answer. --- ## Frequently Asked Questions **What does CECL stand for?** CECL stands for Current Expected Credit Losses. It is the FASB accounting model, introduced in ASU 2016-13, under which banks estimate and reserve for the credit losses they expect over the entire life of a loan at the time it is originated. **How is CECL different from the old incurred-loss model?** Under the prior incurred-loss model, a bank reserved for a loss only once it was probable the loss had already occurred. CECL is forward-looking: banks estimate expected losses over the full life of a loan up front, incorporating current conditions and reasonable forecasts, so reserves are built ahead of the credit cycle rather than after losses appear. **When did CECL take effect?** Larger SEC-filing banks adopted CECL in 2020 (fiscal years beginning after December 15, 2019). Smaller reporting companies and private entities received an extended timeline and adopted it in 2023 (fiscal years beginning after December 15, 2022). **What is the allowance for credit losses (ACL)?** The allowance for credit losses is the balance-sheet reserve a bank holds against expected loan losses under CECL. It is built through the provision for credit losses (an income-statement expense) and reduced by net charge-offs when loans become uncollectible. **What is the difference between the provision and the allowance?** The provision for credit losses is the periodic expense a bank records to build or adjust its reserve. The allowance for credit losses is the cumulative reserve balance on the balance sheet. Provision builds the allowance; charge-offs draw it down. **Where can I find a bank's allowance for credit losses?** The allowance, provision, charge-offs, and coverage ratios are reported in the call report (Schedule RC and Schedule RI) and standardized in the [UBPR](https://www.bankregreports.com/blog/how-to-read-ubpr-report/). [BankRegReports](https://www.bankregreports.com) displays them for every U.S. bank with historical trends and peer comparisons. --- The one thing worth sitting with: CECL shifted loss recognition forward, but it did not make reserve estimates more accurate. It made them more visible, more forward-looking, and more sensitive to economic assumptions that are disclosed only in broad strokes. The next credit cycle will test whether the models banks built actually anticipated what happened, and the data to answer that question will live in the call report. --- BankRegReports publishes call report, UBPR, FDIC, Federal Reserve, NCUA, and SEC EDGAR data for every U.S. bank and credit union, with peer benchmarks and quarterly history. [Look up a bank](https://www.bankregreports.com/banks/) ## About this document This is the markdown twin of the page named in `canonical_url` above. Source filings are public record; the cleaning, normalization, cross-source linkage, and derived metrics are BankRegReports' own. Attribute with a link to the canonical URL. Markdown twins are published for editorial content only. Institution data pages do not have them. For data at scale use the API, which is documented at https://www.bankregreports.com/api/ and specified at https://www.bankregreports.com/openapi.json. It returns typed JSON with explicit units and periods, and one call replaces hundreds of page fetches. ## Sitemap - [Glossary](https://www.bankregreports.com/glossary/): every metric, with formula and thresholds - [Blog](https://www.bankregreports.com/blog/): analysis and explainers - [API documentation](https://www.bankregreports.com/api/): the programmatic way to get this data - [XML sitemap](https://www.bankregreports.com/sitemap.xml): the complete crawlable index