Brokerage Records Retention: What 10 States Actually Require

The brokerage records retention period is set by state real estate license law, not by a single national rule, and in the ten states verified here it runs from three to six years. California requires three years, Texas four, Florida and Nevada and Arizona five, Oregon six. No state in this set requires seven. The seven-year figure repeated across the industry is an IRS convention for financial records, borrowed and misapplied to transaction files.

The seven-year brokerage records retention rule is a tax convention

Search for how long a brokerage must keep its files and the answer comes back fast and confident: seven years. It turns up in brokerage handbooks and in vendor marketing copy about brokerage records retention. It is wrong, or at least it is answering a different question than the one being asked.

Seven years is the retention window associated with federal tax records. It is a reasonable rule of thumb for financial documents that might be examined in an audit of the business. It is not what any state real estate commission requires for transaction files, and applying it as though it were produces two opposite failures at once. Brokerages in short-period states keep files four years longer than required, paying storage and carrying discovery risk for records they were free to destroy. Brokerages in Oregon, where the brokerage records retention period runs six years, may still be under the line but have no idea why.

This is the same failure mode that produces most compliance errors in real estate: a number that is correct in one regulatory category gets carried across a boundary into another. Tax law and license law both impose recordkeeping duties on the same brokerage, over some of the same paper, on different clocks. Getting brokerage records retention right starts with noticing that those are two separate obligations and that satisfying one says nothing about the other.

What state license law actually requires for brokerage records retention

The statutes and administrative rules say the following, quoted to the primary source in each case. These are license-law brokerage records retention periods for transaction files. They are not tax periods, and they are not insurance carrier recommendations.

Primary sources: California Bus. and Prof. Code 10148, RCW 18.85.285, Utah Admin. Code R162-2f, Idaho Code 54-2049, Texas Real Estate Commission, Colorado Division of Real Estate, Florida Administrative Code 61J2-14.012, Nevada Administrative Code 645, A.R.S. 32-2151.01, Oregon Real Estate Agency.

Ten states, four different periods, and ten different descriptions of when counting begins. There is no shortcut that survives contact with this table. A brokerage records retention policy written around a single number is wrong in most of the states it touches.

The brokerage records retention clock test

Most brokerages treat brokerage records retention as one question with one answer. It is three questions, and the second and third are where the failures actually happen. Run every state you operate in through what we call the retention clock test:

1. How long: what period does the license law set?
2. From when: what event starts the clock, and is it the same event for deals that died?
3. What counts: which artifacts are records, and do the ones you never filed count too?

Question one is the only part of brokerage records retention anybody asks about. It is also the only one that is easy. The table above answers it in ten states and the answer is publicly available for the other forty. Questions two and three are the ones that put a brokerage on the wrong side of an audit while its policy document says everything is fine.

Question one, how long the brokerage records retention period runs

The range of brokerage records retention periods across the verified set is three to six years, clustered at three, four and five. That clustering matters because it means the spread between the shortest and longest obligation is a factor of two. A national brokerage applying the California number in Oregon is destroying files three years early, which is half the required window.

There is a tempting simplification here: adopt the longest period you are subject to and apply it everywhere. That is defensible as policy and it is what most multi-state operators settle on. It is worth being precise about what it is, though. Retaining longer than required is a business decision about risk and storage cost, not a legal floor. It also cuts both ways, because records you still hold are records that are discoverable in litigation. The longest-period approach buys regulatory safety and sells some legal exposure to do it.

Question two, when the brokerage records retention clock starts

This is the question that quietly breaks brokerage records retention policies, because the start event is not uniform and is often not the closing.

California runs its three years from the closing date, or from the listing date when the transaction never consummated. Colorado runs four years from consummation, or from expiration of a listing that did not consummate. Nevada runs five years from closing or from the last activity involving the property, whichever applies. Those are three different rules wearing similar language.

Utah and Idaho do something structurally different again, and it is the detail aggregators most often flatten. Both measure in calendar years after the year in which the triggering event happened. Idaho Code section 54-2049 requires records to be kept “for three (3) calendar years after the year in which the event occurred, the transaction closed, all funds were disbursed, or the agreement and any written extension expired.” A deal that closes in February 2026 does not start a three-year clock in February 2026. The clock starts at the end of 2026, which means the file is live until the close of 2029, nearly four years after closing. A system that computes closing date plus three years deletes that file roughly ten months early.

That is not a rounding error. It is an entire category of files destroyed inside the required window, produced by a retention rule that looks correct in a policy document.

Question three, what counts as a record

The third brokerage records retention question has moved faster than most policies have. Statutes written around paper files now apply to email threads, e-signature audit trails, shared drives and message apps.

California addressed this directly and the answer has a boundary worth knowing. Section 10148 covers emails and digital documents. A statutory clarification effective January 1, 2015 excludes text messages, instant messages and social media posts from the requirement, with one exception: when those communications are deliberately created to establish a permanent record of the transaction. So the default is that texts are out, and the exception is behavioral rather than technical. Texas points the other way. Its broker responsibility rules reach all records relating to a real estate transaction, and TREC guidance treats text messages as within scope when they are used to conduct business or substantiate a transaction.

Two states, opposite defaults, same messaging app on the same agent phone. A brokerage operating in both cannot resolve this with a single toggle, and the honest answer for most is to capture business communications everywhere and let the stricter rule govern.

Format requirements matter too. Washington requires that electronically stored records be indexed and immediately retrievable at the licensed office, viewable and printable. Oregon permits electronic image storage but conditions it, including allowance for non-erasable write-once systems. For brokerage records retention, having the data is not the same as satisfying the rule. A record you cannot produce on request during an audit is functionally a record you do not have, which is the practical reason retention and end-to-end audit trail design are the same project rather than two adjacent ones.

The dead deal blind spot in brokerage records retention

The single most common brokerage records retention failure almost never appears in the policy document, because nobody thinks of it as retention at all.

Most brokerages archive closed files. Closing triggers the handoff, the commission disbursement and the move to storage. Deals that die do not trigger any of that. The offer gets rejected, the listing expires, the buyer walks, and the paperwork stays wherever it was sitting.

The statutes are explicit that this is not allowed. Nevada requires records kept for at least five years “including, without limitation, offers that were not accepted and transactions that were not completed.” Idaho lists first among its required records “the original or true copy of all accepted, countered or rejected offers.” Utah starts its clock on the year an offer is rejected, which only makes sense if rejected offers are records you were supposed to keep. California measures from the listing date precisely for transactions that never consummated.

Four of the ten states in the table name dead deals specifically. A brokerage whose retention process is triggered by closing has no process at all for the files these rules are most explicitly about. This is the gap an auditor finds quickly, because it is easier to prove a record is missing than to prove a kept record is deficient.

The multi-state brokerage records retention math, worked

Consider brokerage records retention at a firm licensed in California, Nevada and Oregon, closing 400 transactions a year across the three, with roughly 120 additional files a year that go to contract or listing and die.

Under a single seven-year policy borrowed from the tax rule, it retains everything for seven years. That is 3,640 files in storage at steady state, against a legal requirement that tops out at six years in Oregon and sits at three in California. The brokerage is over-retaining roughly 1,400 files, every one of them discoverable.

Under a single three-year policy borrowed from California, storage drops to 1,560 files and the Oregon exposure is severe: every Oregon file is destroyed three years inside a six-year requirement, and every Nevada file two years inside a five-year requirement. Roughly two-thirds of the portfolio is out of compliance while the policy document reads as disciplined and consistent.

Under a per-state policy, California files clear at three years, Nevada at five, Oregon at six, with the dead-deal files entering the same schedule on their own trigger dates rather than never entering it. Storage lands between the two, and every file clears on the rule that actually governs it. The per-state version is the only one of the three that is correct, and it is the only one that cannot be run out of a single number in a policy binder.

Building a brokerage records retention policy that survives an audit

A brokerage records retention policy that holds up has five properties, and most of them are operational rather than legal.

It is per-state, not per-brokerage. The governing rule is the state where the licensed activity happened, which for multi-state operators means retention is a property of the file rather than of the company.

It computes from the correct trigger. Closing date, listing date, last activity, contract termination, or end of the calendar year containing one of those, according to the state. This is arithmetic, and arithmetic is exactly the kind of thing that should never be done by hand on 400 files a year.

It captures dead deals the moment they die. A rejected offer has to enter the retention schedule on the day it is rejected, which means something has to notice that it died.

It defines records broadly and captures them at the source. Email, e-signature certificates, and the communications that substantiate the deal, in a form that is indexed and retrievable rather than merely stored.

It produces on demand. The test of a retention system is not what is in it but how fast a named file and its full history can be put in front of a regulator. If that takes a week of hunting, the brokerage has a storage system rather than a retention system. This is the distinction explored in depth in our guide to what a document custodian does in real estate, and it is the same one that separates working brokerage compliance software from a shared drive with good intentions.

Where brokerage records retention breaks in practice

Brokerage records retention failures are consistent across brokerage size. Records live in the transaction management system, but the email thread that documents the disclosure timing lives in an inbox and the texts that document the extension live on a phone. Retention is assigned to whoever runs operations, on top of their actual job, with no trigger for dead files. When a state audit arrives, the brokerage discovers its records are complete in the sense that nothing was deliberately destroyed and incomplete in the sense that nothing can be assembled.

The underlying problem is that brokerage records retention is a coordination task sitting across four or five systems that do not talk to each other. That is why it fails at exactly the moment it matters, and why it cannot be fixed by writing a better policy document. It gets fixed by making the systems agree.

ReBillion approaches this as an AI control plane that orchestrates your stack rather than another archive to feed. It reads every contract, tracks every deadline, and coordinates your TMS, CRM, signature and comms, so the brokerage records retention trigger fires from the transaction record itself and applies the rule for the state the deal was written in. Because it is trained on your contracts and customized to your workflow, it knows that a rejected Idaho offer starts a clock at the end of the calendar year and a rejected California offer starts one at the listing date. Pricing is one predictable monthly fee that starts at $199 and is set by your volume, so compliance coverage does not get billed per file or per seat. You can see how ReBillion is priced before talking to anyone.

Retention is not the only place where a dead or stalled deal creates an obligation that nobody owns. The same pattern shows up in holdover seller situations, where the transaction is functionally over but the file is not. And the question of who inside the brokerage is even permitted to handle these records is governed separately, by the unlicensed assistant rules in each state. Operators building state-specific processes can start from our guides to transaction coordination in Nevada and Oregon, the two longest-clock states in the table above.

Frequently asked questions

How long does a brokerage need to keep transaction records?

It depends on the state. Across the ten states verified against primary sources here, the brokerage records retention period runs from three years in California, Washington, Utah and Idaho, to four in Texas and Colorado, five in Florida, Nevada and Arizona, and six in Oregon. There is no single national period, and a policy built on one number will be wrong in most states it touches.

Is there a seven year brokerage records retention requirement?

Not under state real estate license law in any of the states verified here. Seven years is a federal tax records convention that has been widely repeated as though it governed transaction files. It does not. Brokerages may choose to retain for seven years as policy, but they should know they are applying a tax rule of thumb rather than meeting a license law requirement.

When does the brokerage records retention clock start?

It varies by state and it is frequently not the closing date. California runs from closing or from the listing date on deals that did not close. Colorado runs from consummation or listing expiration. Nevada runs from closing or last activity. Utah and Idaho run in calendar years after the year the triggering event occurred, which pushes the real deadline up to a year later than a simple closing-plus-three calculation suggests.

Do brokerages need to keep offers that were rejected?

Yes, in many states, and this is the most common gap. Nevada explicitly includes offers that were not accepted and transactions that were not completed. Idaho requires the original or a true copy of all accepted, countered or rejected offers. Utah starts its clock on the year an offer is rejected. A retention process triggered only by closing misses every one of these files.

Do text messages count as brokerage records?

It depends on the state and the purpose. California excludes text messages, instant messages and social media posts from its requirement unless they were deliberately created to establish a permanent record of the transaction. Texas reaches all records relating to a transaction, and treats texts as in scope when used to conduct business or substantiate the deal. Multi-state brokerages generally capture business communications everywhere and apply the stricter rule.

What happens if a brokerage destroys records too early?

Early destruction breaks the brokerage records retention rule and is a license law violation exposing the brokerage to discipline from the state commission, and it removes the evidence the brokerage would need to defend itself in a civil dispute over the same transaction. Florida addresses the second risk directly by extending retention to two years after the conclusion of any civil action or appeal, and in no case less than five years.

Written by Vikas Malpani

Vikas Malpani is the CEO and Co-Founder of ReBillion and a CAR-Certified Transaction Coordinator. A serial real estate technology entrepreneur with 15+ years across technology and real estate operations, he was named to MIT Technology Review's TR35 list of young innovators. At ReBillion he leads the AI systems that deliver compliant, accurate transaction coordination for brokerages and agents across the 10 states ReBillion serves. Connect with Vikas on LinkedIn: https://www.linkedin.com/in/vikasmalpani/

View Profile