The IRS’s March 2016 IP PIN suspension showed a protection is only as strong as the path to recover it
On March 7, 2016, the IRS suspended its online tool for retrieving identity protection PINs, which TIGTA had found relied on single-factor checks. What held up afterward was stronger identity proofing at both enrollment and recovery.
Looking Back revisits past weeks one at a time, up to the most recent, and asks what it meant for the banks, health systems and public agencies we serve. The series was written in 2026, with the benefit of hindsight, and each article says which week it looks back on.
Two-minute summaryThe dek and the first sentence of each section, quoted from the article
On March 7, 2016, the IRS suspended its online tool for retrieving identity protection PINs, which TIGTA had found relied on single-factor checks. What held up afterward was stronger identity proofing at both enrollment and recovery.
- The week The IP PIN is a six-digit number.
- The credential: a lock tied to the return The IP PIN worked at a single point: acceptance of the return.
- The recovery path: a quiz in front of a spare key TIGTA’s 2015 report laid out the online path step by step. [2] A first-time user entered a name and email address and confirmed a code sent to that email. [2] The user then entered a taxpayer identification number, date of birth, filing status and address from the most recent return. [2] Those had to match IRS records.
- The fallback channels: what closing the door moved Closing the online tool did not remove the need it served.
- The decision record: keeping a breached path open TIGTA’s 2017 report gave the sequence.
- How the case reaches different organizations A federal agency running an online benefit or payment service is a relying party under NIST’s digital identity guidelines. [4] [7] The IRS case shows why the agency should assess the recovery and change-of-details paths at the same level as sign-up.
- Objections to the argument, in order This is true, and the IRS said so in substance when it listed who lost their only alternative. [1] Our argument does not say the suspension was costless.
- Checks a program owner could run Write down every way a person can obtain, reveal, reset or change the delivery of the credential that protects your service.
- With hindsight TIGTA’s March 23, 2017 report put numbers on the season before the suspension.
The week
On March 7, 2016, the Internal Revenue Service deactivated its online Identity Protection PIN application, according to the Treasury Inspector General for Tax Administration (TIGTA). [3] A short statement on the agency’s website said that, as part of its ongoing security review, it had temporarily suspended the Identity Protection PIN tool there. [1] The agency said it was conducting a further review of the application that lets taxpayers retrieve their IP PINs online, and was looking at strengthening the tool’s security features. [1]
The IP PIN is a six-digit number. The statement described it as an additional layer of protection for taxpayers who have been, or could become, victims of tax-related identity theft. [1] Taxpayers who receive one must use it on electronic and paper returns for those returns to be accepted for processing. [1]
The statement gave the scale of the program for that filing season. It said filers received 2.7 million IP PINs by mail, in CP01A letters sent before the 2016 filing season. [1] It said about 5 percent of them, approximately 130,000, had used the online tool to try to retrieve a lost or forgotten IP PIN. [1]
The statement also described what had been done to contain the risk. For taxpayers retrieving a lost IP PIN, it said it had put strengthened processes and filters in place to review those returns. [1] It called these review procedures invisible to taxpayers, and said they had helped detect potential identity theft and stop refund fraud. [1] Through the end of February, the agency said, it had confirmed and stopped 800 fraudulent returns using an IP PIN. [1]
The statement then set out who lost an option. The statement said the online tool was primarily used by people who had lost their IP PINs, and that most holders received their PIN by mail and never used the tool. [1] For certain taxpayers, it said, the online tool had been their only alternative. [1] It named three groups: holders who lost their letter, participants in three pilot locations, and taxpayers invited to use an IP PIN because of non-tax identity theft. [1]
For each group, the statement gave instructions. Holders who lost the CP01A letter would need to call the agency, and if they could verify their identity, the PIN would be mailed. [1] Those who had moved since January 1, 2016 must file a paper return. It said such returns would receive additional scrutiny and take longer to process. [1]
Participants in Florida, Georgia and the District of Columbia who had already retrieved an IP PIN were told to include it. Those who had not would be unable to access the tool at that time but could file as normal. [1] Taxpayers who had filed Form 14039 citing non-tax identity theft and already retrieved a PIN were told to include it as well. [1]
The suspension came after public warnings about the authentication behind the tool. On November 19, 2015, TIGTA had issued a report on the agency’s authentication processes. [2] It found that the methods used for the online Get Transcript and IP PIN applications provided only single-factor authentication, despite government standards requiring multifactor authentication for such high-risk applications. [2] TIGTA said unscrupulous individuals had gained unauthorized access to tax account information as a result. [2]
That report described the steps a first-time user went through. Users entered a name and email address and confirmed a code sent by email. [2] They entered their taxpayer identification number, date of birth, filing status and address from their most recent return, which had to match agency records. [2] For Get Transcript and the IP PIN application, they then answered questions pulled from their credit report and other data sources through a third-party vendor. [2]
TIGTA reported that the agency had not completed an authentication risk assessment for the IP PIN application as required. [2] According to the report, management recognized the application required multifactor authentication but believed requiring it would further burden identity theft victims. [2] The agency did not expect to have multifactor capability in place before the summer of 2016. [2] TIGTA said that, had the assessment been done, it believed management would have concluded that the risk outweighed the burden. [2]
TIGTA also flagged a policy question that was being raised at the time. It noted that it had been suggested that the IP PIN program be expanded to all filers. [2] Before doing so, the report said, the agency must ensure the program provides sufficient assurance that the person requesting the PIN is who he or she claims to be. [2]
The IP PIN was designed as a lock on a taxpayer’s account: without the right six digits, an electronic return would not be accepted. [1] The March 2016 suspension is a record of what happens when the lock is strong and the way to get a spare key is weak. The tool the IRS took down did not issue PINs to new users only. The IRS said it was primarily used by people who had lost their PINs and needed to retrieve them. [1]
Our argument is that the suspension moved the question from the credential to the path used to recover it. A six-digit PIN mailed to a confirmed address carries the assurance of that mailing. A PIN shown on a web page after a quiz carries the assurance of the quiz. TIGTA’s November 2015 report had said, in substance, that the second path did not meet the standard the first one implied. [2]
What held up afterward, in our reading, was one principle. The recovery path must be proofed at least as strongly as enrollment, and both must be stronger than what a fraudster can buy. The IRS restored the tool with multifactor authentication, and later opened the program to everyone who could pass a stronger identity check. [3] [5] NIST later wrote limits on knowledge-based checks into federal guidance. [4] [7]
We take the episode in four parts:
- The credential: what the IP PIN protected, who held one, and what assurance its issuance carried.
- The recovery path: how the online retrieval tool checked identity, and why that check was the weak point.
- The fallback channels: phone, mail, paper returns and manual review, and what the suspension shifted onto them.
- The decision record: how the decision to keep the tool running, then pull it, was assessed and documented.
The parts do not overlap. The first concerns the protection itself, the second the electronic way back to it, the third the human ways back, and the fourth the governance of choosing among them. After them we look at how the case reaches other organizations, take the objections in order, and end with checks a program owner could run.
The credential: a lock tied to the return
What the PIN did
The IP PIN worked at a single point: acceptance of the return. The IRS said taxpayers who receive an IP PIN must use it on electronic and paper returns in order for the returns to be accepted for processing. [1] The control did not depend on spotting a suspicious return. It depended on the thief not having six digits the taxpayer had.
That design is strong when the secret stays secret. The IRS mailed the numbers in CP01A letters, 2.7 million of them before the 2016 filing season. [1] A mailed letter inherits the assurance of the address on file. If the address is right and the mail is not intercepted, the PIN reaches the right person.
Who held one
The holders were not a random sample of taxpayers. TIGTA’s 2015 report said the IRS issues an IP PIN to confirmed victims of identity theft, among others. [2] The IRS statement added participants in three pilot locations and taxpayers invited because of non-tax identity theft. [1] The program protected people whose personal data had already been exposed or misused.
That fact changes the threat model. A knowledge-based quiz asks questions an attacker should not be able to answer. For a person whose identity has already been stolen, the attacker may hold much of the same data. TIGTA said the knowledge-based questions used for the online route could be circumvented by unscrupulous individuals. [2]
Different routes, different assurance
TIGTA found that the IRS’s routes to an IP PIN offered different levels of assurance. [2] A victim filing Form 14039 by mail or fax had to include a photocopy of a valid federal or state government-issued identification. [2] A taxpayer using the online route had to provide basic identifying information and answer knowledge-based questions, with no copy of identification. [2]
The PIN issued at the end was the same six digits either way. In our reading, that is the core of the problem the suspension exposed. A credential’s strength is the strength of the weakest route that can issue or reveal it. The return-acceptance check could not tell which route a PIN had come through.
What the credential could not do
The IP PIN protected filing. It did not, by itself, protect the other doors to a taxpayer’s data. TIGTA’s 2015 report grouped the IP PIN application with Get Transcript, which it said provides a tax account or tax return transcript for a specific year. [2] Both sat behind the same authentication framework. [2] A weakness in that framework reached every application behind it.
For the 2016 season, the IRS put locks on accounts of known Get Transcript breach victims to keep their Social Security numbers from being used in the IP PIN application. [3] TIGTA later found those locks did not work as intended for some of them. [3] We return to that finding under the decision record.
Why the credential still mattered
None of the weakness lay in the PIN’s format. Six digits checked at filing remained a sound control for the returns it covered. The IRS said its filters had stopped 800 fraudulent returns using an IP PIN through the end of February 2016. [1] The issue was who could obtain those digits, and how.
For an agency or a bank, the parallel is any one-time code, account number or recovery token that unlocks a valuable action. The token can be well designed. Its value then depends on every path that issues, resends or reveals it.
The recovery path: a quiz in front of a spare key
The steps
TIGTA’s 2015 report laid out the online path step by step. [2] A first-time user entered a name and email address and confirmed a code sent to that email. [2] The user then entered a taxpayer identification number, date of birth, filing status and address from the most recent return. [2] Those had to match IRS records.
For the IP PIN and Get Transcript applications, the user then answered questions drawn from a credit report and other data sources through a third-party vendor. [2] After that, the user could create a username and password. [2] TIGTA counted the result as single-factor authentication. [2]
Why many steps were still one factor
TIGTA addressed a point from the agency’s June 2015 Senate testimony. It noted testimony that taxpayers go through a multistep authentication process, and said multiple steps do not meet the requirements for multifactor authentication. [2] Each step tested something the user knew, or data that could be found about the user.
The report described the levels of assurance then in federal use. Under OMB’s M-04-04 and NIST SP 800-63-2, Level 3 required identity proofing data and account numbers. It also required a second factor, such as a one-time code sent to the individual’s phone or email. [2] TIGTA said NIST standards required multifactor authentication for such high-risk applications. [2]
What a quiz proves
A knowledge-based question proves that the person knows, or can find, an answer. For an ordinary customer, that may be a fair signal. For a population of identity theft victims, it is a weaker one. TIGTA’s description of the questions as circumventable was a description of the people the program served. [2]
NIST’s later guidance put the same point in rules. Its 2017 identity proofing volume said information accessible freely, for a fee in the public domain, or via the black market shall not be used for knowledge-based verification. [4] Its usability considerations advised avoiding knowledge-based verification because it is extremely problematic from a usability perspective. [4]
The burden argument
The IRS had a reason for the design. TIGTA reported that management believed requiring multifactor authentication would further burden identity theft victims trying to obtain an IP PIN. [2] That concern was real. Victims already faced delayed refunds, affidavits and calls.
TIGTA’s response was that a risk assessment would have weighed the burden against the risk of compromised PINs, and in its belief found the risk greater. [2] The disagreement was not about whether victims mattered. It was about whether a weaker path served them better than a slower but stronger one.
The recovery path as an attacker sees it
From an attacker’s side, the retrieval tool offered something valuable: a way to get a victim’s current PIN without intercepting mail. TIGTA later reported that, for some Social Security numbers flagged in the Get Transcript breach, the IRS received a duplicate return after the number was used in the IP PIN application. [3] TIGTA said that indicated a fraudulent return was likely filed by a thief who obtained the PIN from the application. [3]
That is the pattern to look for in any system. Attackers do not need to break the strongest control. They need the cheapest route to a credential that the strongest control will accept.
What a stronger recovery path checks
TIGTA’s 2015 report described the kind of second factor federal standards expected at Level 3: a one-time code sent to a phone number or email address associated with the individual. [2] Its description of the paper route described another kind of evidence: a copy of a government-issued identification. [2] Both tie the request to something other than knowledge of personal data.
A recovery path can combine such checks in several ways. It can send a code to contact details already on file, not to details entered in the same session. It can require evidence that an issuing source validates. It can add a delay before a revealed credential becomes usable, so that the real holder has time to notice. These are design options, not requirements, and each has costs.
The IRS’s later route put recovery behind an account that required stronger verification to create. [5] [8] That approach moves the hard check to one place, account creation, and then relies on the account’s own authentication for every later action. Its strength depends on how the account itself can be recovered, which is the same question one level up.
For that reason, a program owner should follow the chain to its end. If the PIN is recovered through the account, and the account is recovered through an email address, then the email reset is part of the PIN’s protection. The weakest link in that chain sets its strength, and the links may belong to different teams.
The same shape elsewhere
Banks know this shape from password resets, card reissue and changes of contact details. Health systems meet it in patient portal sign-up and proxy access. Agencies meet it wherever a benefit or a payment can be redirected. In each case the account may be well protected while the recovery desk, the reset link or the change-of-address form is not.
We do not suggest the cases are identical. Tax refunds, deposits and medical records carry different stakes and different laws. The shared feature is structural: a recovery path is an issuance path, and should be assessed as one.
The fallback channels: what closing the door moved
The phone and the mailbox
Closing the online tool did not remove the need it served. The IRS told holders who had lost their letter to call. If they could verify their identity, the PIN would be mailed. [1] That shifted load to telephone staff and to postal time, during the filing season.
For holders who had moved since January 1, 2016, the IRS said they must file a paper return. [1] It said those returns would receive additional scrutiny and take longer, because the IRS does not normally accept such returns without an IP PIN. [1] The safer route for the system was a slower route for the taxpayer.
The pilot participants
The pilot in Florida, Georgia and the District of Columbia let residents get an IP PIN online. [2] TIGTA said those locations were chosen for having the highest per capita rate of tax-related identity theft when the initiative was piloted. [3] With the tool suspended, participants who had not yet retrieved a PIN could not get one and could file as normal. [1]
In practical terms, the suspension removed an optional protection from some of the people the IRS had chosen for it because of their risk. The alternative was to keep offering it through a route TIGTA considered unsafe. [2] Neither choice was free.
Manual review as a compensating control
Before the suspension, the IRS relied on manual review to offset the tool’s weakness. Its statement said strengthened processes and filters reviewed returns from retrieved PINs. [1] TIGTA later described the plan in detail: returns filed with a PIN viewed online after January 18, 2016 were to go to tax examiners for manual review. [3]
TIGTA reviewed 32,623 such returns filed between January 19 and May 24, 2016. It found 12,020, or 36.8 percent, were not manually reviewed as required. [3] IRS management told TIGTA some were missed because system programming was not complete at the start of the season and because weekend staffing was insufficient. [3] TIGTA said the IRS paid more than $42.4 million in refunds claimed on those returns. [3]
That finding is the clearest operational result of the episode. A compensating control that depends on a daily list, a staff queue and a weekend roster is a process. TIGTA said internal guidelines required a daily list of Social Security numbers whose PINs were viewed. [3] When the list was incomplete, the control failed without any alarm.
The cost of fallbacks
Every fallback the IRS named had a price. Phone verification needs trained staff. Mailing needs time and a correct address. Paper returns need processing and scrutiny. [1] In-person verification, which the IRS offered later, needs offices and appointments. [5]
For a program owner, the practical step is to size the fallbacks before relying on them. If an online path may be closed in a hurry, the phone, mail and in-person paths will carry its volume. Their capacity is part of the online path’s risk assessment.
Telling people what changed
The IRS statement is a reasonable model of a change notice. It said what was suspended, why in general terms, how many people used the tool, and what each affected group should do next. [1] It told holders who had already retrieved a PIN to keep using it, and told others they could file as normal. [1]
It also avoided a common failure: telling everyone to call. The statement split the population into groups with different instructions, which kept callers to those who needed a call. [1] For an agency or a bank closing a channel at short notice, that segmentation is the difference between a manageable queue and an unmanageable one.
A notice also has to be accurate about timing. The IRS described the suspension as temporary and the review as ongoing, without promising a date. That restraint is worth copying when a fix depends on work not yet scheduled.
Equity in the fallback
Fallbacks do not serve everyone equally. A taxpayer who has moved, works hours that clash with call centers, or lacks documents faces more friction than one who does not. A stronger online check has the same problem: some people cannot pass it.
The IRS’s later design addressed this with several routes at different speeds. In January 2021 it described an online route, a mail-in application with phone verification for taxpayers under an income limit, and in-person verification at a Taxpayer Assistance Center. [5] Each route had its own assurance and its own delay. [5]
The decision record: keeping a breached path open
The dates
TIGTA’s 2017 report gave the sequence. It said the IRS discovered on May 17, 2015 that its single-factor authentication process had been breached by unauthorized individuals. [3] The IP PIN application was taken offline on November 21, 2015 for system maintenance. [3] On January 8, 2016, TIGTA sent an email alert recommending that it not come back online until stronger authentication could be implemented. [3]
The IRS brought the application back online on January 19, 2016, with mitigation strategies. [3] On February 24, 2016, TIGTA again recommended deactivation. [3] The IRS deactivated the application on March 7, 2016. [3]
What the mitigations were
TIGTA listed three mitigations for the 2016 season. [3] Returns using PINs viewed online were to be manually reviewed. Accounts that opted in online were to be marked and run through identity theft filters. Accounts of Get Transcript breach victims were to be locked against use in the application. [3]
TIGTA reported that, before deactivation in the 2016 season, 137,967 individuals revealed their IP PIN through the application and 26,559 opted in. [3] The IRS statement had described approximately 130,000 taxpayers who used the tool to try to retrieve a PIN. [1] The two documents measured different things on different dates, and we report both as their authors stated them.
The missing assessment
TIGTA’s central finding was about documentation. It said the IRS did not complete a sufficient risk assessment for the IP PIN application after discovering the May 2015 breach. [3] It said NIST requires organizations to complete an authentication risk assessment after identifying a system security breach. [3] The assessment should document weaknesses, risks, mitigations considered, and the costs and resources required. [3]
TIGTA recommended that the IRS complete and document such assessments after all future breaches of an online application. The IRS agreed and said it updated its incident response procedures on January 20, 2017. [3] In our reading, that response turned an argument about one tool into a standing rule about decisions.
Why the record matters
A decision to keep a weak path open can be reasonable. The service may be needed, the alternative may harm the same people, and a fix may be close. What makes the decision defensible later is a record made at the time: what was known, what was weighed, who decided and when it would be revisited.
Without that record, the organization is judged on the outcome alone. With it, an inspector, an examiner or a court can see whether the risk was understood. The IRS episode shows the cost of the gap: the inspector general’s reports became the record by default.
Warnings as inputs
TIGTA warned the IRS repeatedly, by its own account. [3] Internal audit, an inspector general or an outside assessor plays that role elsewhere. A warning does not oblige an organization to act as advised. It does oblige the organization to show it considered the warning and decided on stated grounds.
Who should own the decision
The decision to keep a breached path open sits between three owners. The business owner of the service cares about the people it serves. The security owner cares about the attack. The fraud or payments owner cares about losses. Each sees a different part of the risk.
TIGTA’s 2015 report recommended that the IRS establish a function with the authority to make authentication policies consistent across the agency. [2] It found two groups focused on taxpayer authentication, with neither providing cross-functional management and oversight. [2] The IRS agreed to the recommendations. [2]
In our reading, the organizational point is as important as the technical one. A weak recovery path persists when no single owner can weigh the burden on victims against the risk of fraud, and decide. Naming that owner before the next breach is cheaper than finding one during it.
For a bank, the equivalent owner may be a fraud committee or the chief information security officer, depending on the bank’s structure. For an agency, it may be the program office with the chief information officer and privacy official. The structure matters less than the fact that the decision has a name on it.
The second finding
TIGTA’s 2017 report also found a problem unrelated to the online tool. It identified approximately 2 million taxpayers whose identity theft cases were resolved as of December 29, 2015 but whose accounts lacked the marker that generates an IP PIN. [3] IRS officials attributed this to inconsistent guidance among functions where a victim’s address was unknown. [3]
That finding concerns the credential’s issuance, our first part, and we note it here only because it came from the same review. It shows that a protection program can fail at both ends: issuing to too few people through the strong path, and revealing too easily through the weak one.
How the case reaches different organizations
A federal benefits agency
A federal agency running an online benefit or payment service is a relying party under NIST’s digital identity guidelines. [4] [7] The IRS case shows why the agency should assess the recovery and change-of-details paths at the same level as sign-up. Each agency should confirm with counsel and its privacy office which guidance and policies apply to its systems.
A state revenue or motor vehicle agency
State agencies are not bound by NIST guidance in the way federal agencies are, unless state law or policy adopts it. They face the same structural risk wherever a refund, a title or a license can be redirected online. The IRS sequence, from warning to suspension to stronger proofing, is a usable model for their own decision records.
A county or city government
Local governments run payment portals, permit systems and benefit intake. Their recovery paths may sit with a vendor. The question to ask is the same one TIGTA asked: whether the route that reveals or resets a credential meets the assurance the service needs, and whether anyone assessed it after a known breach.
A bank or credit union
Banks and credit unions follow their own regulators’ guidance on authentication, not NIST’s federal guidelines as such. The structure of the risk is familiar to them: account takeover through a reset or a change of contact details. The IRS case adds a point about compensating controls. A manual review queue that misses weekends is a control that fails on a schedule.
A health system
Health systems run patient portals, proxy access for family members, and identity checks at registration. HIPAA rules govern access to records, and organizations should check with counsel on what they require. The design lesson is that a portal’s sign-up and recovery checks protect the record as much as its login does.
A broker-dealer or wealth manager
Investment firms face the same pattern when a client asks to change a bank account for withdrawals, add a trusted contact, or reset access. The asset at stake is a transfer of funds rather than a refund. Their own regulators set the rules that apply, and firms should check those with counsel.
The IRS case suggests one specific test for them. A firm should ask whether a person who knows a client’s personal data, and nothing else, could complete a change of payment instructions through any channel. If the answer is yes for one channel, that channel sets the firm’s exposure.
A vendor that runs identity checks
The IRS used a third-party vendor for the knowledge-based questions. [2] A vendor providing identity verification to many clients should expect each client to ask what the check proves, against what data, and how it performs for people whose data has already been exposed. NIST’s later guidelines give vocabulary for those questions. [4] [7]
Objections to the argument, in order
1. The tool served victims and closing it hurt them
This is true, and the IRS said so in substance when it listed who lost their only alternative. [1] Our argument does not say the suspension was costless. It says the cost came from building the online path at a lower assurance than the credential required, which TIGTA had reported months before. [2]
2. Manual review was a reasonable compensating control
Manual review can be reasonable. TIGTA’s finding was that it was not performed for 36.8 percent of the returns TIGTA reviewed for the period. [3] A compensating control is only as reasonable as its execution, and its execution must be measured to be relied on.
3. The exposed data came from elsewhere
TIGTA’s 2015 report pointed to data breaches in the private and public sectors, which it said made personal information available to unscrupulous individuals. [2] Its point was that such data was detailed enough to get through authentication processes. [2] Designing for a world in which personal data circulates is the agency’s task, as NIST later wrote into its rules on knowledge-based verification. [4]
4. Multifactor authentication excludes people
Stronger checks do exclude some people. The IRS’s later design answered this with alternative routes, including mail-in applications with phone verification and in-person verification. [5] The objection argues for several routes of known assurance, not for one weak route open to everyone.
5. This was a tax-specific event
The specifics were tax-specific: refunds, filing seasons and CP01A letters. The structure was not. Any organization that issues a credential through one path and reveals or resets it through another has the same exposure. That is why we think the case belongs in a general reading list on identity programs.
6. NIST guidance came later, so the IRS cannot be judged by it
We agree that the 2017 and 2025 NIST rules came after the event. [4] [6] TIGTA judged the IRS against the standards of 2015, OMB M-04-04 and SP 800-63-2, which already required multifactor authentication at Level 3. [2] We cite the later rules to show where the guidance went, not to apply them backward.
7. A suspension is an overreaction to a few fraud cases
The IRS reported 800 fraudulent returns stopped through February. [1] TIGTA later reported thousands of returns that went unreviewed and refunds paid on them. [3] The scale of exposure was not known on March 7, 2016, which is the argument for deciding on the assessed risk, documented at the time, rather than on the losses seen so far.
Checks a program owner could run
List the paths to the credential
Write down every way a person can obtain, reveal, reset or change the delivery of the credential that protects your service. Include the call center, the mailroom, the branch, the chat channel and any vendor portal.
Rate each path
For each path, record what it checks and what that check proves. Use one vocabulary across paths. NIST’s identity assurance levels give federal agencies such a vocabulary; other organizations can adopt similar terms. [7]
Find the weakest issuing path
Compare each path against the value of what the credential unlocks. The weakest path that can issue or reveal the credential sets its real strength. If that path is weaker than the service needs, the gap is a finding.
Re-assess after every breach
After any breach that touches the data your checks rely on, repeat the assessment and document it. TIGTA’s recommendation to the IRS sets out a usable list: weaknesses, risks, mitigations considered, and costs and resources required. [3]
Measure compensating controls
If a manual review or a filter offsets a weak path, count what it reviewed against what it should have reviewed, daily. The IRS gap TIGTA found arose from incomplete lists and weekend staffing. [3] Both can be measured.
Size the fallbacks
Estimate the call, mail and in-person volume if the online path closes tomorrow. Confirm staff, scripts and identity checks exist for that volume. A fallback that cannot absorb the load will push decision makers to keep a weak path open.
Keep the decision record
When you keep a weak path open, write down why, who decided, which warnings were considered and when the decision expires. Give the record to the risk owner and to audit. Revisit it on the stated date.
Watch the outcomes of each path
Track what happens after each recovery: disputes, duplicate filings or claims, complaints and fraud reports, by path. TIGTA used duplicate returns as a signal that a PIN had been obtained by a thief. [3] An equivalent signal may exist in other programs, and it can show which path is being used against you.
Report those outcomes to the owner of the decision record, on a fixed schedule. A recovery path that looks safe in design and produces disputes in operation needs a new assessment.
Retire knowledge-based questions as proof
Plan to remove knowledge-based questions as the basis for identity verification. NIST’s 2025 guidelines say knowledge-based verification shall not be used for identity verification, while allowing it as part of a fraud management program. [7] Non-federal organizations should decide with counsel how far to follow that line.
With hindsight
The IRS brought the IP PIN application back online on July 19, 2016, with improved security including multifactor authentication, according to TIGTA. [3] The IRS’s own statement page carries an update dated January 10, 2017, saying the IP PIN was now supported by Secure Access. [1]
TIGTA’s March 23, 2017 report put numbers on the season before the suspension. It found 12,020 of 32,623 returns filed with a PIN viewed online were not manually reviewed as required. [3] It found 2,347 Social Security numbers from the Get Transcript breach had been used in the application, and 299 of them later had duplicate returns. [3] The IRS agreed to document authentication risk assessments after future breaches. [3]
NIST revised its digital identity guidelines in June 2017. The identity proofing volume, SP 800-63A, set rules for knowledge-based verification at IAL2. [4] It barred the use of information accessible freely, for a fee in the public domain, or via the black market. [4] It required at least four questions, barred questions whose answers do not change, and advised avoiding knowledge-based verification for usability reasons. [4]
On January 12, 2021, the IRS expanded the IP PIN Opt-In Program to all taxpayers who can verify their identities. [5] It said taxpayers must pass a rigorous identity verification process, using Secure Access if they did not already have an IRS account. [5] It also described a mail-in route with phone verification for taxpayers with adjusted gross income of $72,000 or less, and in-person verification. [5]
That release answered the question TIGTA had raised in November 2015. TIGTA had said the IRS must ensure sufficient assurance before expanding the program to all taxpayers. [2] The IRS expanded it after moving the online route onto stronger verification. [5] In our reading, the order mattered: assurance first, then scale.
NIST published SP 800-63-4 in July 2025, superseding SP 800-63-3. [6] Its identity proofing volume says knowledge-based verification or knowledge-based authentication shall not be used for identity verification. [7] It allows credential service providers to use knowledge-based verification as part of a fraud management program. [7] The direction of travel since 2016 is clear in the text: from limits on quizzes to no quizzes as proof of identity.
As checked on November 16, 2026, the IRS page on getting an IP PIN said anyone with a Social Security number or ITIN who can verify their identity is eligible. [8] It said the fastest way is through the IRS app or an online account, and that people without an account must register to validate their identity. [8] It listed alternatives, including Form 15227 for filers with adjusted gross income below $84,000, or $168,000 for joint filers. [8]
The same page announced that the online IP PIN system would be offline for maintenance effective Monday, November 16, 2026, until mid-January 2027. [8] The IRS had described an annual offline period between November and mid-January in its 2021 release. [5] The seasonal pattern is a reminder that fallback routes still carry load every year.
What held up from March 2016 was the idea behind the suspension: a credential is only as strong as the route that reveals it. TIGTA’s reports, the IRS’s later design and NIST’s guidance all moved in that direction. [3] [5] [7] What did not hold up was the hope that manual review could carry a weak route through a filing season. TIGTA’s count showed how much of it went undone. [3]
Spotted an error? Tell us and we will check it and correct the article where needed.