Every deal package leads with the same pair: a rent roll and a T-12. They are usually read separately, the rent roll for the tenancy and the T-12 for the numbers, and that is the mistake. Each document is a check on the other. The rent roll is a set of claims about contracts; the T-12 is a record of what the accounting actually saw. Read together, they either reconcile, which tells you the seller's story is coherent, or they do not, which tells you exactly where to dig.
The short version
Establish what each document is claiming and as of when. Build gross potential rent from the rent roll, then walk it down to the T-12's collected rent through four lines: vacancy, concessions, bad debt, and non-revenue units. Check delinquency against bad debt, effective rent against face rent, and the T-12's other income against a source document, because the rent roll will not show it. Date both documents, cite every number to its line, and treat any gap the walk cannot explain as a finding, not a rounding error.
What each document is actually claiming
A rent roll is a census taken on one day: unit by unit, who holds the lease, at what rent, from when to when, with what deposit and what status. Its claims are contractual. Nothing in a rent roll tells you whether the tenant in suite 210 has paid in six months.
A T-12 is the operating history, month by month for twelve months: scheduled rent, the deductions that separate scheduled from collected, the other income the property throws off, and every expense of running it. Its claims are accounting. Nothing in a T-12 tells you that the anchor tenant's lease expires in nine months.
That is why neither is sufficient. The rent roll describes the future the leases imply; the T-12 describes the past the property delivered. Underwriting is the act of deciding how much of each to believe, and the reconciliation below is how you earn that decision.
The five reconciliations
1. Gross potential rent to rental income. Multiply the rent roll's total in-place monthly rent by twelve and set it against the T-12's gross scheduled rent line. These will not be equal, because the rent roll is one day and the T-12 is a year of changing rents, but they should be explainably close. If the annualized rent roll runs well above the T-12's schedule, rents were raised recently or the rent roll is dated after a lease-up push; either way, the trailing numbers understate or overstate the future, and you need to know which.
2. The vacancy walk. Take gross potential rent and subtract, using the T-12's own lines: vacancy loss, concessions, bad debt, and non-revenue units. What remains should land on collected rental income. When the walk does not close, something is being presented on a different basis, and the usual suspects are concessions netted into the rent line, or vacancy calculated against a different unit count than the rent roll shows.
3. Delinquency to bad debt. A good rent roll carries a delinquency or balance column. Set it against the T-12's bad debt expense. A rent roll showing meaningful balances alongside a T-12 showing near-zero bad debt means someone has not written off what they will not collect, and the trailing income is better than reality.
4. Face rent to effective rent. Where the rent roll shows concessions, free months, or a separate effective-rent column, compare the spread against the T-12's concession line and its trend across the twelve months. Concessions that appear only in recent months are a leasing market telling you something the occupancy number is not.
5. Other income to a source. Parking, laundry, utility reimbursements, fees: the T-12 shows them, the rent roll almost never does. Each recurring other-income line above trivial size needs its own source, whether that is a parking agreement, a RUBS program summary, or a fee schedule. Other income that exists only as a T-12 line is the easiest number in the package to inflate.
The traps in both documents
The conveniently dated rent roll. A rent roll is a snapshot, and the seller picks the day. A roll dated right after a lease-up push, or right before known move-outs, is technically accurate and materially misleading. Ask for the same report ninety days earlier and compare.
Month-to-month at face rent. Holdover and month-to-month tenancies listed at their old contract rent overstate the durability of the rent roll. Count them separately; they are occupancy you cannot underwrite as term.
Units that are occupied but not paying. Model units, employee units, and down units sit in the occupied column and produce no revenue. If the rent roll does not flag them, the physical occupancy number is quietly better than the economic one.
The adjusted T-12. Owner add-backs, one-time credits, insurance recoveries, and management-fee arrangements with related parties all flow through small line items. A T-12 that has been cleaned for presentation will reconcile suspiciously well; ask for the general ledger export behind any line that looks smoothed.
Twelve months hides a trend. A single annual column can average away six good months and six deteriorating ones. Read the T-12 in its monthly columns, and be suspicious of any package that offers a trailing three annualized in place of a true trailing twelve; annualizing a quarter is how seasonality and recent softness get laundered.
A one-page bridge that holds up
The working product of this exercise is a single page: gross potential rent at the top, the four deduction lines in the middle, collected rent at the bottom, with the T-12's figure beside each line and a variance column. Date it twice, once for the rent roll's as-of date and once for the T-12's period. Under it, three or four sentences on what the variances mean. Every number on the page cites the document and line it came from, because six weeks later the question that comes back is where a number came from, and the page already answers it.
Build the page so a stranger can audit it. Each line names the figure, the document it came from, and the page or tab, with the variance and a short note where the variance has an explanation. The header carries both dates and the preparer's name. A bridge with no source column is a set of assertions, and the first person to challenge one will send you back to the documents anyway.
Keep a second version showing your underwritten numbers beside the seller's actuals, since those are two different claims. The seller's column describes what happened; your column describes what you believe happens next, and the notes between them are the underwriting. Investment committees read the notes.
Keep the documents next to the bridge
The bridge is a summary, and summaries drift. Keep the rent roll, the T-12, and the bridge in the same folder, and when a number is challenged, go back to the source line and settle it there. The habit of citing every figure to its document is what separates a reconciliation you can defend in an investment committee from one you have to redo.
Store the pair with the bridge and the leases in one deal folder, and keep the original files with their original names alongside anything you reformatted. Analysts re-cut rent rolls into spreadsheets constantly, and the re-cut version is where transcription errors enter and then propagate into the model. When a number is questioned, the answer should come from the seller's own file, not from your copy of it.
Date the folder to the delivery, not to the analysis. Sellers issue updated rent rolls and refreshed T-12s through diligence, and a deal file that overwrites instead of versioning loses the ability to answer the most useful question of all, which is what changed between the package that priced the deal and the package that closed it.
The expense side is half the document
Everything above walks the income side, and most rent roll reviews stop there, which is how expense surprises survive diligence. The T-12's expense lines deserve the same reconciliation discipline, starting with the two that reprice at sale. Property taxes in the trailing period reflect the current owner's assessment, and in many jurisdictions a transaction triggers reassessment; underwriting the seller's tax line forward is one of the most common and most expensive T-12 mistakes, so build the go-forward tax number from the local assessor's method, not the trailing statement. Insurance has the same character in a harder market: the trailing premium tells you what coverage cost the seller under policies you are not inheriting, and a current quote outranks twelve months of history.
Then come the lines a manager can steer. A management fee paid to the seller's own affiliate can sit above or below market, and either direction distorts. Repairs and maintenance versus capital expenditure is the same boundary dispute the income side never sees: recurring costs pushed below the line flatter net operating income, while genuine capital items expensed in a bad year understate it. Read the R&M detail month by month; a smooth line at a round number is an allocation, and an allocation is a choice someone made for a reason.
Loss to lease is not vacancy
There is a third gap the two documents create together, distinct from both physical and economic vacancy: the spread between in-place rents and market rents, usually called loss to lease. A rent roll can be fully occupied at rents signed two years ago and thirty dollars under today's market, which is either the upside story the broker is selling or a sign the units cannot actually command the sheet the seller quotes, and the trailing twelve cannot tell you which. Test it at the unit level: take the most recent five leases signed and compare their rates to the older leases on identical unit types. Recent signings above the in-place average confirm the market story with the property's own paper; recent signings below it mean the loss to lease is not upside, it is the market's opinion. This is also where concession timing from the income walk earns its second use, because new leases at headline rents with two months free are the same information wearing a different disguise.
Separate the two things loss to lease can mean. In a rising market it is genuine upside, realized unit by unit as leases roll, and its speed is set by the expiration schedule, not by the size of the spread. In a soft market it is the gap between an asking rent nobody is paying and the rents actually signed, which is not upside at all. The expiration schedule tells you which one you are looking at, since upside you cannot capture for three years is worth materially less than upside arriving next quarter.
Underwrite the capture cost too. Rolling a unit to market usually costs a turn: downtime, make-ready, a concession, sometimes a commission. A model that lifts rents to market without those costs will beat reality in every period, and the difference compounds across a lease-up plan.
Tie the roll to the leases themselves
The rent roll is a report generated from someone's records, and the leases are the records. Close the loop on a sample: pick five to eight units across the mix, pull the actual lease files, and verify the roll's claims against the signatures, including rate, term, deposits, concessions, and any amendments or transfer addenda the roll's software never learned about. In commercial properties the same tie-out runs through estoppels, where the tenant certifies the facts directly and disagreements between estoppel and rent roll outrank both documents. A clean sample earns the roll trust for the rest of diligence; a sample with variances converts the whole review from confirmation to investigation, and finding that out on five units is cheap compared to finding it out on fifty after close.
Choose the sample deliberately instead of randomly. The largest tenants by rent, anything with an unusual term or a below-market rate, recent signings, and any unit the roll flags as month-to-month or in some non-standard status. Those are where variances live, and a clean result on the hard cases says more than a clean result on eight identical units.
Record each tie-out as a line with the unit, the fields checked, and the outcome, then keep it with the bridge. If a variance turns up, expand the sample in the same category before expanding it everywhere; errors in rent rolls cluster by cause, usually a lease type the software handles badly or a period when the property changed managers.
What the lender does with the same two documents
The rent roll and T-12 you are reconciling will shortly be read by a second audience with veto power: the lender's underwriter, who runs the same walk with less generosity. Lenders typically underwrite to their own adjusted net operating income, applying a vacancy factor at least equal to a floor even when the property runs fuller, marking taxes to the post-sale assessment, marking insurance to quote, imposing a management fee whether or not one is paid, and deducting replacement reserves per unit or per square foot regardless of what the seller spent. The result is an underwritten NOI reliably below the broker's, and debt sizing follows it through the debt service coverage and debt yield tests, not the offering memorandum. Running the lender's version of the reconciliation yourself, before application, tells you what the debt will actually be and converts the loan process from a negotiation about hope into a review of arithmetic. It also gives the equity story its honest denominator: a deal that only works at the broker's NOI does not work.
Ask early which of the lender's adjustments are policy and which are negotiable, because the answer changes the deal. Vacancy floors and replacement reserves tend to be fixed by program; the management fee, the treatment of other income, and the tax assumption sometimes move with evidence. A current tax bill, a signed management agreement, and a documented other-income history are the three exhibits most likely to recover dollars in underwriting.
Run the debt service coverage and debt yield tests yourself on the lender's version, then on your own, and note the gap. That gap is the honest measure of how much of your equity story depends on being right where the lender is conservative, and it belongs in the memo next to the bridge.
A worked walk
The reconciliation is easier to run than to describe. Take a property whose rent roll annualizes to gross potential rent of $1,200,000. The T-12 shows vacancy loss of $84,000, which is 7 percent and should roughly match the unit math on the roll; concessions of $36,000; bad debt of $18,000; and non-revenue units of $12,000. The four deductions total $150,000, so the walk lands on $1,050,000 of expected collections. The T-12's collected rental income line shows $1,046,500. The residual is $3,500, and the point of the exercise is that the residual now has a size: small enough to be a utility reimbursement netted into the wrong line or a mid-month move-out proration, large enough to spend ten minutes confirming which. A walk that lands within a rounding story is a seller whose numbers hang together. A residual of $40,000 on the same walk is not a rounding story, and finding it took one page of arithmetic that most buyers skip.
The walk also has a follow-through. Chase the residual to a general ledger line before accepting any explanation of it, because "that's probably utilities" is a hypothesis and the GL is an answer. Then keep the completed walk as page one of the diligence file, dated to its two source documents, and re-run it once more before closing against the updated rent roll the seller delivers at proration time. Properties change during a sixty-day escrow, and the second walk is how you find out whether they changed in the direction the first one promised. Ten minutes twice beats a surprise once.
For larger properties, run the walk by unit type as well as in total, because a clean total can hide offsetting stories: one-bedrooms outperforming while the townhomes carry the vacancy, or commercial suites propping a struggling residential side. The by-type walk uses the same five lines and the rent roll already carries the segmentation; it is one more column of arithmetic and it is where the business plan either finds its target or loses its premise.
Frequently asked questions
What is a rent roll?
A point-in-time census of a property's leases: every unit, its tenant, dates, current rent, deposits, and status, as of a stated date. It is a claim about contracts, not a record of collections.
What is a T-12?
The trailing twelve months of operating history, month by month: scheduled rent, vacancy and concession losses, other income, and operating expenses. It records what the accounting saw, not what the leases promise.
What is the difference between the two?
The rent roll is a snapshot of contractual terms; the T-12 is a year of actuals. One answers what the rent should be, the other what was billed, collected, and spent. Each checks the other, which is why they are read together.
Why don't they match?
Some gap is structural: vacancy, concessions, bad debt, and mid-year rent changes all separate contractual rent from collected rent. The question is whether the T-12's own lines explain the gap. An unexplained residual usually points to a flatteringly dated rent roll, misclassified units, or income presented on different bases.
What is economic vacancy?
The share of gross potential rent not collected from all causes: physical vacancy, concessions, bad debt, and non-revenue units. A property can be 95 percent physically occupied and meaningfully worse economically, which is exactly what a rent roll alone will not show.