CRE concentration and stress testing, explained.

Why CRE loans go bad together, and the concentration and stress tests regulators ask for

The life of a CRE book's risk report, in five steps

1. Loans are made and bookedEach loan gets a line on the call report, the quarterly financial report every bank files with its regulators, plus a property type and a market. Ashgrove has 486 CRE loans worth $1,320.3 million.
2. The month or quarter closesThe loan tape and the capital figures are final. Ashgrove's capital for these tests is $412.6 million at September 30, 2026.
3. Concentration is measured and the book is stress testedRatios against the supervisory markers and the board's limits; every scenario applied to the loans; ratios recomputed after the losses. This is where Larch Street Tower shows up, or does not.
4. The board risk committee reads the reportLimits, results, the loans behind the losses and proposed actions. Ashgrove's committee meets October 22.
5. Limits, lending and capital plans change, or do notThe point of the exercise. Examiners look for the line from the results to a decision.

Step 3 is the subject of this guide. Ashgrove Community Bank, its loans and its people are fictional.

Four surprises about the 300% CRE concentration test

Concentration ratio, NOI, debt service coverage, cap rate

Fourteen terms are defined in the one-page summary.

Which loans count toward the CRE concentration ratio, and which stay out

In shortSupervisory CRE is four call report lines: construction and land, multifamily, nonfarm nonresidential property repaid from third-party rent or sale, and loans that finance real estate without being secured by it. Owner-occupied business property, farmland and 1–4 family mortgages stay out. The classification decides the ratio before any arithmetic starts.

The guidance covers loans whose repayment depends on the property market, and leaves out loans repaid by an operating business that happens to occupy its own building.3 The OCC's CRE booklet names the exact call report lines. For the construction test it is schedule RC-C, Part I, items 1.a.(1) and 1.a.(2) plus Memorandum item 3. For total CRE it is items 1.a, 1.d and 1.e.(2) plus Memorandum item 3.5

Item 1.d is multifamily. Item 1.e.(2) is nonfarm nonresidential property that is not owner-occupied. Memorandum item 3 picks up loans that finance real estate, construction or land development but are not secured by real estate.

What US banks hold, by kind of property

The Federal Reserve's weekly data put bank CRE at $3.14 trillion in September 2026. Banks outside the 25 largest held $2,107.6 billion of the $3,017.2 billion at domestically chartered banks, about 70%.6 That is where concentration lives: in community and regional banks. The OCC's 2012 stress testing bulletin was written for banks of $10 billion or less.7

The CRE concentration guidance and stress testing rules in force in October 2026

In shortTwo documents still set the expectations: the 2006 interagency guidance and the OCC's 2012 community bank stress testing bulletin. Neither has been rescinded. Around them, the denominator changed in 2020, the FDIC re-issued its advice in 2023, model risk guidance was replaced in April 2026, the OCC published a new concentrations booklet in September 2026, and the capital rules are under review.

How to calculate the 100%, 300% and 50% growth tests, with one bank's numbers

In shortAdd the call report lines, divide by capital, and compare the growth of total CRE with its balance 36 months earlier. Then do it again for each property type against the board's sub-limits, convert every percentage into dollars of headroom, and repeat the whole thing with approved loans that have not funded yet.

The guidance treats these as two separate criteria.3 The first, construction and land at 100% or more of capital, stands on its own. The second has two legs that must both be met: total CRE at 300% or more and growth of 50% or more over the prior 36 months. A bank at 340% with 20% growth does not meet it; a bank at 260% with 80% growth does not either. Neither is a reason to relax.

The FDIC's own count of concentrated banks ignores the growth leg and counts every bank above 300% of total CRE or 100% of construction, 1,374 banks or roughly 31% of all banks at the end of 2024.1

Ashgrove's three tests at September 30, 2026

The fictional example from the top of the page. The bank above 300% without 50% growth is the case the guidance still expects to manage carefully: it says a bank that has grown CRE quickly, has notable exposure to one type of CRE, or "is approaching or exceeds" the criteria may be identified for further supervisory analysis.3

Turn every limit into dollars

Boards set limits as percentages, but they decide on lending in dollars. Office at 48.1% sounds comfortable until it is written as $7.7 million, less than half of one large office loan. The arithmetic is simple: headroom equals the limit times capital, less the loans already held. For office, 50% × $412.6M − $198.6M = $7.7M.

Then add what is approved but not yet funded

The call report carries outstanding balances, and the FDIC's count states that "Unfunded commitments are not included".1 A board still needs the second column, because construction draws raise the balance without any new approval. In Ashgrove's case $41.2 million of the $96.5 million pipeline is construction, which would take that ratio to 99.4%. In 2025 the OCC wrote the same point into a formal agreement with Patriot Bank, asking for "procedures for monitoring concentration reports monthly based upon total committed amounts relative to Board-approved limits".19

The same bank on two denominators

Illustrative arithmetic. Which way the gap runs depends on what else the bank counts in tier 2 capital. The examiner's number uses the 2020 denominator.4 Banks that use the community bank leverage ratio, eased from 9% to 8% from July 1, 2026, do not compute risk-based capital at all, so for them the newer denominator is the only one available.12

Owner-occupied or not: the 50% rent test that moves the CRE ratio

In shortA nonfarm nonresidential loan counts as CRE when 50% or more of its repayment comes from rent paid by unrelated tenants, or from selling or refinancing the property. The decision is made once, when the loan is made or bought, and need not be revisited. Every dollar classified the wrong way moves the ratio by that dollar divided by capital.

The call report instructions put item 1.e.(2) loans in the CRE ratio when "50 percent or more of the source of repayment comes from third party, nonaffiliated, rental income" or from the sale, refinancing or permanent financing of the property, and they list hotels and motels on that line.20

A machine shop that owns its plant through a family property company is owner-occupied. An investor who leases the same plant to an unrelated manufacturer is not. A hotel counts as CRE even though its income comes from running the business, because its value moves with the property market.

The instructions also say the owner-occupied decision is made "upon acquisition (origination or purchase) of the loan" and "need not be reviewed thereafter".20 That keeps the call report stable, and it lets the label drift from the risk. A family that sells its operating company and keeps the building, now leased to the buyer, still reports the loan as owner-occupied. The building's loan is now repaid by a tenant. A concentration report that tracks such loans as their own segment shows the board risk the call report hides.

The pressure on the label is one-sided. A loan officer has no reason to prefer the non-owner-occupied label, which pulls a loan into the ratio and its limits. On a bank with $400 million of capital, $40 million classified the wrong way moves the total CRE ratio by 10 points, more than many stress scenarios move it. A sound practice is a written reason on file for every 1.e loan: who owns the property, who occupies it, what share of repayment came from third-party rent at origination, and what has changed since.

Four kinds of CRE stress test, and what a loss-rate spreadsheet can and cannot say

In shortThe OCC names four kinds: transaction (one loan), portfolio, enterprise and reverse. A portfolio test can be built from the top, by applying loss rates to groups of loans, or from the bottom, by stressing each loan and adding up the results. The OCC's own worked example is a top-down spreadsheet, and it accepts either method as long as it fits the bank's risk.

The OCC's own example, a top-down spreadsheet

Appendix B of the bulletin multiplies each category's balance by a two-year loss rate, carries the losses through two years of earnings and subtracts the result from tier 1 capital. On its $600 of loans (the example states no unit), losses come to $61.8, two-year net income turns to a loss of $27.2, tier 1 capital falls from 88 to 60.8 and the leverage ratio from 11% to 8.2%.7

For many banks with a seasoned, low-leverage book, that is a complete and defensible annual test. The 2006 guidance itself says that well-margined, seasoned multifamily loans normally need far lighter testing than construction.3

Where the loss-rate method runs out

The OCC put the scenario factors in a table of its own. In appendix C, four factors apply to every property type: lower collateral values, higher utility and energy costs, higher interest rates and higher capitalization rates. Falling rents and rising vacancy apply to most income property, and sale prices and absorption to construction.7 A test that moves all of those through one haircut on value has skipped the analysis the table asks for.

Stress testing one office loan: income, the rate at refinance, value and loss

In shortA loan-level stress test moves the property's income, the interest rate it will pay if it refinances inside the horizon, and its value, then asks whether coverage falls below 1.00x or the loan rises above the value. Where it fails, the loss is the shortfall of the stressed value after selling and workout costs. Larch Street Tower, the loan from the top of the page, shows every step.

The facts, from three documents

The four steps of the severe scenario

Two shocks break this loan, and they come from different documents. The income shock comes from the operating statement and lands on a building already less full than at origination, whose biggest tenant can leave three months before the loan matures. The rate shock comes from the note: a loan written at 4.15% meets an 8.75% refinance rate.

Even with no fall in income, the refinance alone at today's 6.75% would take coverage to about 1.14x, below the bank's 1.25x policy and below the 1.2 to 1.5 minimum the Federal Reserve says lenders generally require.21 A test that reads both documents also says when the trouble arrives: the lease ends in March 2027 and the loan matures in June, eight months after the October board meeting.

How cap rates move value

A wider cap rate lowers value even when income holds, and the two shocks compound. That is why the OCC's CRE booklet says that "capitalization rates, interest rates, and DSCRs should be stress-tested" when a loan is underwritten.5

Research gives a sense of how far office values can move: a study of New York offices published in the American Economic Review in February 2026 estimates a 46% decline in long-run value, and $556.8 billion of value destruction across US offices.24 A severe scenario that takes office values down 35% sits inside that range. A 10% haircut applied to every property type alike would not be a severe office scenario.

Adding up 486 stressed loans, and the concentration ratios after the losses

In shortRun every loan through every scenario and add up the losses. Take the losses out of capital, then divide the book by what is left: that gives the ratios after stress, a number a report of losses alone leaves out. Because the result is a sum of loans, it can be cut by property type, market or maturity, and it names the loans behind every dollar.

The board's three scenarios at Ashgrove

Fictional example. The scenario library is the board's, approved April 23, 2026; losses are taken against capital before any earnings, a conservative choice the board made. Loans maturing or repricing in the horizon, $866.8M or two thirds of the book, refinance at 6.75% plus the scenario's shock.

The ratio after stress is the board's second number. A bank with $123.8 million of room under its total CRE limit today has none left after a severe nine quarters, and is $37.9 million over. Office crosses its 50% limit as soon as losses anywhere in the book exceed $15.4 million, because that is the point at which capital falls to $397.2 million, twice the office book.

The OCC's 2025 consent order with EH National Bank put the link in writing: concentration limits must be "consistent with the Board's risk tolerance based on current market conditions, the declining capital base, and results of the most recent stress test results".25

From 486 loans to the one the board will ask about

Loans fail for different reasons, and the reason points to the action. In the severe case at Ashgrove, 50 loans fail on coverage only, their stressed income below their payments; 23 fail on value only, with income that still covers the payments but a stressed value below the balance; and 6 fail on both, Larch Street Tower among them. A coverage failure suggests a conversation about rates, reserves or paydowns. A value failure suggests a new appraisal and a look at what a sale would recover.

What the 2007–2011 failures show about the concentration screens, and the mistakes that repeat

In shortThe screens predicted failure well in the last real CRE downturn: banks over both criteria failed at about 46 times the rate of banks over neither, and construction concentration drove most of the insurance fund's losses. The banks that failed had been building concentration for years. The common errors today come from classification, timing and follow-through; the arithmetic is the easy part.

OCC and Federal Reserve staff tracked 7,379 banks through the downturn. The study also estimated that banks over the construction criterion accounted for about 80% of the losses to the FDIC's insurance fund from 2007 to 2011.26 That is the reason to watch the construction ratio monthly: construction books move fastest, through draws on loans already approved.

The GAO's comparison shows how slowly the danger built. The banks that later failed were already well above the others in 2001, and their concentration rose faster for seven years.2 The FDIC's inspector general later found that from 2007 to 2013 poorly managed concentration risk contributed to the failure of 492 institutions and $74.6 billion of losses to the Deposit Insurance Fund.27 The same review is blunt about stress tests: some examiners "did not realize what 'good' concentration stress tests or concentration strategic contingency plans were" until they reviewed one at a bank.27

Ten mistakes that make a concentration report wrong

Bank CRE in 2026: a larger book, office distress outside the banks, and a wall of maturities

In shortBank CRE loans reached $3.14 trillion in September 2026. The median bank of $1 billion to $10 billion holds CRE above 300% of tier 1 capital plus the allowance. Office distress is visible in commercial mortgage securities, while bank loss rates stay low; the risk supervisors name in 2026 is refinancing loans written when rates were low.

The FDIC's 2026 Risk Review put the industry's median CRE ratio at 200% at year-end 2025 and the median for banks of $1 billion to $10 billion at 311%, so more than half of the banks in that band sit above 300% on the ratio alone.23

The Atlanta Fed reported in 2024 that concentration ratios had risen about 15 percentage points since the start of 2021, for US banks overall and in its own district.30 The Chicago Fed's look at community banks found that of 2,668 with CRE at 25% or more of loans in mid-2023, 42.7% were at or above 300% of total capital.31

Loss rates at banks remain low. The median net charge-off ratio on CRE was 0.07% in 2025, and of $11.6 billion of modified CRE loans, 82% were performing.23 At June 30, 2026, noncurrent rates across insured institutions were 0.99% for construction and development, 1.23% for nonfarm nonresidential and 1.10% for multifamily.32

The worst office losses sit at the largest banks, whose non-owner-occupied CRE past-due and nonaccrual rate fell to 3.08% from a 4.99% peak in the third quarter of 2024, against a pre-pandemic average of 0.59%.33 Federal Reserve research finds small banks' lower delinquency mostly reflects "their low holdings of large-sized office loans".34

Office: the market's weak spot

Loans coming due at higher rates are the 2026 worry

Commercial banks are the largest holder of commercial and multifamily mortgage debt, with 38% of $5.1 trillion in the second quarter of 2026.39 The Mortgage Bankers Association counted $875 billion of commercial and multifamily mortgages, 17% of $5.0 trillion, scheduled to mature in 2026, $396 billion of it at depositories, and $652 billion more in 2027.29

The OCC's Spring 2026 report says CRE headwinds "have not resulted in significant credit defaults, although refinance risk merits continued attention".22 The Federal Reserve's May 2026 Financial Stability Report adds that "Lenders' willingness to extend or modify maturing loans has helped to limit this risk to date. However, this option may be increasingly limited going forward".40 New York Fed research on large banks puts CRE loans expiring within three years at about 37.0% of marked-to-market capital for less-capitalized banks, against 27.8% for better-capitalized ones.41

Credit standards have stopped tightening. In the Federal Reserve's survey of senior loan officers, the net share of banks tightening standards for construction and land loans fell from 73.8% in the second quarter of 2023 to −3.7% in July 2026: slightly more banks eased than tightened.42 Looser standards and growing books are the conditions the agencies warned about in December 2015, when they observed "substantial growth in many markets, increased competitive pressures, rising concentrations, and an easing of underwriting standards".9

What the board approves, what the report should show, and what examiners grade

In shortThe board owns the limits, the CRE strategy and the scenarios; management measures, tests and proposes. A good report shows each limit with its headroom in dollars, the results of every approved scenario, the ratios after losses, the loans behind the losses, the assumptions and the data quality, and proposed actions. Examiners grade the quality of the test and what it changed.

The 2006 guidance gives the board "ultimate responsibility for the level of risk assumed by the institution". It should approve the CRE lending strategy, "Ensure that management implements procedures and controls", review information that quantifies CRE concentration risk, and "Periodically review and approve CRE risk exposure limits and appropriate sublimits".3 Management should have "a contingency plan to reduce or mitigate concentrations in the event of adverse CRE market conditions", and should stratify the book by "property type, geographic market, tenant concentrations, tenant industries, developer concentrations, and risk rating".3

The inspector general counted 11 of 41 reports that used the page reserved for matters requiring board attention.27 That finding is the case for a board report that does not wait for the examiner. A board that sees its own numbers, often enough to act and in a form that names loans and markets, does not depend on the report of examination to learn that concentration has become a problem.27

What two 2025 OCC orders spell out

The six pages a board report needs

The layout is ours, drawn from the guidance, the OCC's examination procedures and the 2025 orders.3,5,25,19 The OCC's procedures ask examiners to "Evaluate the quality and results of portfolio stress testing", and to assess whether management and board reports address growth, asset quality, concentrations and trends.5

Automating CRE stress tests: what SR 26-2 treats as a model, and what stays with people

In shortSoftware can read rent rolls, statements and appraisals, classify the tape, compute the ratios, re-stress every loan and draft the report. Under the April 2026 model risk guidance, arithmetic on board-approved shocks is outside the definition of a model, a valuation or default engine probably inside it, and generative AI outside the guidance's scope altogether. Scenarios, limits, owner-occupied judgments and signatures stay with named people.

SR 26-2 defines a model as "a complex quantitative method, system, or approach that applies statistical, economic, or financial theories to process input data into quantitative estimates". It excludes "simple arithmetic calculations, such as those found within spreadsheets" and "deterministic rule-based processes and software where there are no statistical, economic, or financial theories underpinning their design or use".11,16

The OCC's bulletin says generative AI and AI that acts on its own "are not within the scope of this guidance", and repeats that the guidance does not "require community banks to perform annual model validation".16 The Federal Reserve expects it to be most relevant to organizations over $30 billion.11

How examiners will judge an automated CRE stress test

The board's scenarios, limits and owner-occupied calls stay human

• The board chooses the scenarios and the limits. Software applies them; approving them stays with the board.
• A borrower who has not sent this year's operating statement still has not sent it. Software can mark the stale input and draft the request.
• Whether a loan is owner-occupied is a credit judgment made at origination, and a person should decide any change.20
• The allowance for credit losses, interest rate risk and liquidity tests remain separate work.
• The responsibility stays with the bank that signs the report.

Who builds the CRE stress test today, where the six weeks go, and three ways to do it

In shortA portfolio manager or credit risk analyst builds it, credit analysts chase property financials, the chief risk officer owns the scenarios, the chief credit officer signs, and the board risk committee reads it. Most of the time goes to collecting rent rolls and operating statements, which the loan system does not hold. That is why many banks run the full test once a year.

Who builds the CRE stress test, from portfolio manager to board

Five hand-offs from loan tape to the board risk committee

Thirty working days, two fifths spent collecting rent rolls

Thirty working days is six calendar weeks for one person, often longer between other duties. Because of that, the full test happens once a year, close to the OCC's floor of "at least an annual basis".7 Between tests the board sees a ratio table, and a director's question about a market move waits for the next cycle.

Loss-rate spreadsheet, outside consultant or stress testing software

From a 40-basis-point market move to a signed board report, on the scenario board

In shortThis is the example from the top of the page, inside the product. Renee Castellano, the CRE portfolio manager, answers the board's likely question by moving one slider, saves the answer into the write-up, runs the month's three board scenarios on every loan, opens the loan that drives the loss, and sends a drafted report. Sunita Varma, the chief risk officer, reviews it; Martin Hale, the chief credit officer, signs.

1

The book on one board, with rents already 8% lower

Renee opens her desk view: in-place rents down 8% as leases roll. The top line labels the bank's $412.6 million "total risk-based capital", the 2006 guidance's term; since March 31, 2020 supervisors measure capital for these ratios as tier 1 capital plus the allowance for credit losses.

Every loan has been re-stressed from its own rent roll. Losses to December 2028 are $13.8 million, capital after losses $398.8 million, and 14 loans sit below 1.00x coverage, Larch Street Tower first. No limit is crossed, but office is at 49.8% of capital against its 50% limit, and total CRE at 331.1% of 350%.

The run log carries the morning's market note: downtown Columbus office cap rates are 40 basis points higher than in September.

2

Cap rates 150 basis points wider: office crosses its limit

She clicks +150 on the cap-rate slider. The spread is added to each property's own cap rate from its appraisal, and each value is recomputed from its stressed income. Losses rise to $24.5 million and capital after losses falls to $388.1 million. The office bar crosses its red line at 51.2%. Thirty-seven loans are now below 1.00x, 23 more than the desk view, and the newly broken ones move to the top of the list with their occupancy, coupon, refinance rate, income against debt service and value.

3

The question goes into the write-up before anyone asks it

Renee saves the move as "Cap rates +150 bp · rents −8%", with every loan's result and every assumption. One switch puts it in the stress test write-up as a fourth column beside the board's three approved scenarios, with the loans that break named in a paragraph. The board's three approved scenarios stay as they were.

4

The month's run: three scenarios on every loan in about two minutes

She presses Run October. The run reads the September 30 loan tape and capital, the rent rolls, statements and appraisals, and the bank's market data; tests every limit; and runs 1,458 loan-scenarios, 486 loans under three scenarios. In the sample data it finishes in 2 minutes 8 seconds. Two steps carry amber marks: 11 operating statements older than 15 months were used as they are and marked, and the severe case fails 79 loans.

The limits table shows the result: total CRE at 360.4% after severe losses, construction at 100.7%, office over its limit under both the adverse and the severe case.

5

Why Larch Street Tower fails, line by line

Renee opens the loan that carries the largest single loss. The walk shows income from the June 2026 operating statement cut 22%, debt service at the June 2027 refinance at 8.75% on 25-year amortization, the March 2025 appraisal cut 35%, and the loss: stressed value less 10% selling costs, plus 5% workout costs, $5,280,000. The property panel shows occupancy down from 84% at origination to 71%, and the largest tenant's lease ending in March 2027. Each line opens to the rent roll, statement or appraisal it came from.

6

A six-page draft, four proposed actions, then two signatures

The report drafter has written the board report: summary, concentration with and without the pipeline, stress results with Renee's scenario as a fourth column, the loans behind the loss, the assumptions and the actions. It proposes four: a new appraisal for Larch Street Tower, an early refinance conversation with its borrower, funding the construction pipeline in step with payoffs, and collecting the 11 stale statements.

Renee can remove any of them or add her own; in the sample story she adds a fifth, no new office loans until office is under 45% of capital, about $12.9 million less office. She sends it; Sunita Varma reviews; Martin Hale signs. The report goes to the committee for October 22, every step on the activity record.

The solution's design targets are concentration and stress results every month instead of once a year, and every one of the 486 loans stressed property by property with every assumption shown. They are targets, to be measured on your own book.

The rest of the desk: every loan as a square, the three tests, scenarios, markets, inputs and settings

In shortOne run is the story. These screens show the rest: every loan by property type with the ones that broke ringed, the supervisory tests and sub-limits, the scenario results and the library behind them, losses by market, the data and how current it is, every loan under stress, the signed report and the board's settings.

Select any screen to open it larger. "Total risk-based capital" on the screens is the 2006 guidance's label for the sample bank's $412.6M; a bank loads tier 1 capital plus the allowance for credit losses on loans, the measure since March 31, 2020 (section 04). The bank, people, properties, borrowers and market data source are fictional.

Six agents read, calculate and draft. The board's people decide and sign.

In shortEach agent has one job and limits it does not cross. None of them changes a board limit or a scenario, estimates a missing income figure, changes a loan's owner-occupied status or sends a report. The portfolio manager prepares and proposes, the chief risk officer reviews, the chief credit officer signs, and the board approves limits, scenarios and actions.

Six agents, from loan tape reader to board report drafter

People decide; every number traces to its loan

The board risk committee approves limits, sub-limits, scenarios and actions; a change in Settings is a proposal until it does. Losses are taken against capital before earnings, as the board chose. Every tape loaded, run, saved scenario, action, review and signature is on the activity record with who and when, agent steps and human decisions kept apart.

Renee Castellano's 486-loan book and the people who sign for it

Month-end loan tape and property files in, a signed concentration and stress report out

In shortIt reads the data a bank already has: the loan tape, call report capital, rent rolls, operating statements, appraisals, its market data and its credit policy. It leaves behind every ratio against every limit, stress results by loan, property type and market, the loans that fail under each scenario, and a signed board and examiner report with every assumption shown.

Loan tape, call report and market data a monthly run reads

For each, check whether it offers an API, a file feed or only screens, and who owns it.

Your loan system, credit files and board limits stay yours

Loan by loan, the board's scenarios, people sign

What a once-a-year CRE stress test costs you in staff weeks, late answers and missed loans

In shortStart from your own numbers. No regulator or study publishes how long a CRE stress test takes, what it costs, or how often concentration reports are wrong. What is published frames the stakes: what a risk analyst's hour costs, how concentrated banks fared the last time, and how little recent loss history tells you.

Counting your stress test hours and stale operating statements

A limit crossed in March, found in the December test

What monthly, loan-level CRE stress testing could return for your bank

In shortFor most banks the return is more tests for less staff time per test, and answers that arrive while there is time to act. The calculator starts from a published wage and asks you for the rest. Every default marked illustrative is an assumption to replace with your own.

Fewer staff weeks, twelve tests a year, same-day board answers

Your staff hours back from monthly loan-level stress tests

A bank that wants monthly tests clears every line but the last; one that wants quarterly tests breaks even near the $75,000 line. A bank with a small, seasoned book that the OCC's loss-rate spreadsheet already serves, and a board that asks no "what if" questions between tests, probably is not, and should say so before buying anything.7

The loan tape, property financials and board limits a loan-level CRE stress test needs from your bank

In shortPick the answer closest to your bank today. Eight questions score 0 to 2; the ninth is for reflection. The meter adds up your score as you go, and each question says why it matters and what a strong answer sounds like.

Explore, rerun last year's test, or move to monthly runs

Checking a CRE stress testing tool's data, concentration ratios and stress method before you sign

In shortAsk every vendor the same questions, us included. The answers show where property income comes from, how the rate shock is timed, whether the denominator is current, who owns the limits and scenarios, and whether you can take any loss apart.

We found no published independent comparison of CRE stress testing tools, as of October 2026. Rerunning last year's test on your own book is the fairest comparison, and it tests your current process at the same time.

How to start: rerun last year's stress test loan by loan in ten weeks

In shortFix the denominator and the tags first, then rerun your last stress test on the same book, date and scenarios, and compare loan by loan. Decide with four numbers. None of these steps commits you to a vendor.

1. Recompute last quarter's ratios on tier 1 capital plus the allowance on loans, with committed amounts beside funded.
2. Tag the tape: property type and market for every CRE loan; a reason on file for every owner-occupied decision.
3. Gather last year's inputs: the tape for the date of your last test, the property files as they stood then, capital, and the scenarios the board approved.
4. Rerun loan by loan and compare totals, and which loans each method flags.
5. Decide with four numbers: loans flagged that the spreadsheet missed, share of balance on current income, staff hours per run, and days from month-end to a signed report.

Where loan-level testing earns its keep first

Positions are illustrative. Place your own segments with two numbers from your tape: the share of balance maturing or repricing in the horizon, and how far your severe scenario moves income and value.

Loans the loss-rate spreadsheet missed, and days to a signed report

The 300% CRE guidance, owner-occupied loans and loan-level stress tests, explained one question at a time

In shortThe questions credit and risk teams and buyers search for most, each answered on its own.