On broker fraud in the ACA marketplace

The marketplace necessarily invests a lot of trust in brokers and agents. But HealthCare.gov needs new fraud controls.

I’ve got *what* in my pocket?

Julie Appleby of Kaiser Health News reported last week that health insurance brokers are switching a number of ACA marketplace enrollees in HealthCare.gov states from one plan to another “without their express permission” and often without their knowledge.

Unauthorized enrollment or plan-switching is emerging as a serious challenge for the ACA, also known as Obamacare. Brokers say the ease with which rogue agents can get into policyholder accounts in the 32 states served by the federal marketplace plays a major role in the problem, according to an investigation by KFF Health News.

Indeed, armed with only a person’s name, date of birth, and state, a licensed agent can access a policyholder’s coverage through the federal exchange or its direct enrollment platforms. It’s harder to do through state ACA markets, because they often require additional information.

The story is well-sourced and illustrated, with individuals recounting that they suddenly were unable to use the plans they thought they were enrolled in. In one case, a victim found himself on the hook for months of tax credits after disenrolling because he’d obtained employer-sponsored insurance and then being re-enrolled in another plan by a broker unknown to him. Appleby also cites brokers who claim that hundreds of their clients were poached and re-enrolled in plans other than the ones they’d chosen. Appleby links to key CMS documents, online ads, and insurer advisories.

I spoke to brokers and web brokers to delve further into how the fraud works, how the harms are redressed, and how it might be prevented. A few takeaways below.

---

The fraud is egregious, but hard to pin down. As Appleby noted, CMS guidance states explicitly that a broker must obtain consent before switching an enrollee’s plan or making any other changes to her application. Ronnell Nolan, president and CEO of Health Agents for America (HAFA), said that a whistleblower has revealed training materials from a call center perpetrating this fraud that explicitly instruct agents not to ask for consent and not to tell prospective clients that the agent will switch them into a new plan. The premise is that clicking on one of their misleading ads — which present the federal tax credit used to pay the health plan premium as money that can be used for other purposes — constitutes consent. Some of the ads in question, posted in the Facebook ad library (to which Appleby linked), bear this out. One invites the user to “Pre-qualify yourself below to speak to one of our agents.”

Shutting down the bad actors — not to mention prosecuting them — is tricky, however — notwithstanding that they must put their name and professional identifier, the “National Producer Number” (NPN) on the application to get paid. When reporting the unauthorized switch, says Shelli Quenga of the nonprofit brokerage at the South Carolina-based Palmetto Project, a broker can’t make a consumer complaint — the consumer has to do that. And the target population, Quenga points out, is “mostly lower-income people, maybe working multiple part-time jobs. The last thing they want to do is to sit down and write a complaint.”

Nolan says that HAFA is working to streamline the complaint process. The trade group will send members a list of fraudulent agents, their NPNs, and the state in which they’re operating; set up a hotline to report fraud; and provide a template cease-and-desist letter. The HAFA site currently provides detailed instructions for reporting fraud.* These note that CMS requests “that agents/brokers conduct a 3-way call with the consumer and the Marketplace Call Center at 1-800-318-2596 to report the unauthorized changes” — not, as Quenga stresses, the easiest ask of a harried consumer.

According to Nolan, the fraud appears to be centered in call centers in South Florida that recruit multiple brokers selling in multiple states. Importantly, agents are not required to declare who they’re affiliated with — so the apparent responsibility for the fraud is diffused among multiple individual agents. Enforcement, moreover, devolves to the states — CMS will refer fraud complaints to the relevant state Department of Insurance.

An executive at a web broker says that many of the bad actors who switch enrollees’ plans without their consent get their “clients” from lead generators — companies that post the misleading ads and sell respondents’ contact information to brokers — and that many of the victims of unauthorized plan switching were also initially enrolled by brokers using lead generators. Those who are susceptible are targeted serially. The lead brokers are hard to shut down, as many of them operate overseas and hide their tracks.

The harm can be redressed, but it takes time and diligence. Shelli Quenga of the Palmetto Project says that her agency learns of new unauthorized plan switches “every single day.” Often the client learns about the switch when they go to a doctor’s office and are told that they’re no longer covered. Can they be switched back? Quenga says that they can — “We’ve been able to get them all fixed. But it’s time consuming, and it’s stressful for the consumer. If they’re a person going to the doctor and that’s how they’re notified about the switch, it’s a good bet they’ve got health issues, and this creates a whole new bureaucratic nightmare.”

Typically, says Quenga, it takes two hours to fix a plan switch. For brokers, time is money: “in that time, a typical agent could be enrolling eight people.” Sometimes, far more time has to be invested. “If the marketplace can’t fix it, you have to go through [CMS’s] Complex Case Office. They’re very responsive — it’s terrific how quickly they assist -- but you have to get them on the phone when a client is available. “

Web-brokers are a key tool for fraud detection. As I’ve noted before, most marketplace enrollments in the 32 states using HealthCare.gov are broker-assisted, and most brokers enroll clients not on the HealthCare.gov site itself but rather via commercial web brokers deploying Direct Enrollment or Enhanced Direct Enrollment (DE or EDE). These sites streamline enrollment and provide various client management tools. (According to a CMS presentation, 81% of active broker-assisted enrollments are via DE or EDE.) Both Quenga and Nolan affirmed that the web brokers are key to detecting fraudulent plan switching, as they notify brokers any time there’s a change in the broker of record for one of their (now former) clients.

I had wondered whether the widespread use of DE/EDE platforms in HealthCare.gov states might be in some way facilitating the fraudulent plan-switching, as plan-switching is mainly a HealthCare.gov phenomenon, and there is as yet no DE/EDE in state-based marketplaces. The answer is no. Any broker or agent registered with HealthCare.gov can get access to the account of any enrollee for whom they have name, date of birth and state enrolled, regardless of the enrollment platform they use, including HealthCare.gov. HealthSherpa, the dominant web broker, has a fraud prevention team as well as fraud controls built into the application — flagging, for example, a broker for whom a large percentage of enrollments are plan switches. [Note, 4/18: See update 2 below: the call centers alleged in a lawsuit to have perpetuated this fraud have their own proprietary EDE web brokers, and those are alleged to be key to large-scale fraud, as 1) enrollments and enrollment switches can be processed more quickly on an EDE than on HealthCare.gov, and 2) use of a proprietary EDE impedes other web brokers’ ability to flag changes in the designated broker. UPDATE/CORRECTION, 6/3/24: Re the strike-through above: HealthCare.gov has no agent/broker portal; unauthorized plan-switching or enrollment on HealthCare.gov itself would require effectively impersonating the enrollee. See this post for more. High-volume fraud probably really does require use of an EDE platform. I regret the error.

Regulatory changes have eased the path to fraud. In the marketplace’s lean years, insurers cut back on broker commissions, which reached a nadir in 2018 but have probably doubled since then, particularly after the subsidy boosts provided by American Rescue Plan Act (ARPA) in March 2021 triggered an enrollment surge. ARPA also increased by millions the number of enrollees paying zero premium, and the number of zero-premium plans such enrollees could be switched into without requiring payment for effectuation. In June 2022 CMS issued guidance stipulating that insurers could not pay lesser commissions for enrollments outside of Open Enrollment, enabled by Special Enrollment Periods (SEPs) for which applicants could qualify by citing various life changes, such as loss of employer-sponsored insurance, divorce, etc. That came on the heels of CMS finalizing a new continuous SEP for enrollees with income up to 150% FPL - effectively allowing year-round enrollment with unlimited plan switching for this income group, which now accounts for more than half of enrollees in HealthCare.gov states. ARPA rendered the benchmark (second cheapest) silver plan free for enrollees with income below that threshold, and in most markets, multiple bronze plans (as well as two silver plans) are available to such enrollees for zero premium.

Prevalence? Quenga, in South Carolina, says that the Palmetto Project has flagged 34 cases that have agent-of-record/permission issues, out of about 3,500 enrollments — roughly 1%. A web broker executive reports that about 0.5% of applications processed on their platform result in a complaint — and agents are incentivized to complain if an account is switched away from them. Not every unauthorized account switching is reported, and not every complaint reflects actual malfeasance, but these snapshots perhaps give some sense of scale in a marketplace of 21 million enrollees.

The problem should be solvable. Appleby points out that the state-based marketplaces generally require more information before a broker can access a policy they’re not already named on — generally by requiring two-factor authorization. In Colorado, Louise Norris tells me, the enrollee is sent a code that the presumptive new broker must enter to access the account. Three brokers working in HealthCare.gov states whom I spoke to, however - -including Quenga from the nonprofit Palmetto Project — agreed that requiring client approval in real time could seriously inhibit enrollment. Brokers often do their client work at one time — say, during the day — and then execute the applications at a separate time. Toward the end of Open Enrollment in particular, time is of the essence, and clients can be hard to reach quickly. “Any protection you put in against people who are scammy is going to hurt those who are doing good work,” a web broker executive told me.

Quenga suggests that requiring two-factor authorization only outside the annual Open Enrollment Period (OEP) could limit the friction to cases of actual plan switching, which outside of OEP requires a SEP. Client poaching also occurs during OEP, however — in fact, plan-switching (usually without switching brokers) during OEP is routine and indeed encouraged, as premiums and benchmarks change every year. Thus, the lock that discount insurer Ambetter placed on broker designations as of December 31, in an attempt to block unauthorized plan-switching, could backfire in cases where the switching occurred before that deadline.

Requiring that a broker provide the enrollee’s social security number to access the account could be an effective means of fraud control. At present, all a broker certified by HealthCare.gov needs to access an account is the person’s name, date of birth, and state. CMS has apparently raised the objection that not all enrollees have a social security number; many legally present noncitizens, eligible for marketplace coverage, do not. Most documents showing immigration status have a number that presumably could be substituted. But Quenga referenced at least one exception: the J-2 visa, issued to spouses and dependents of participants in exchange programs.

Another possibility is requiring the client’s direct signoff not when the application is submitted but afterward, when the broker seeks payment. That would impose some administrative burden, but not during the enrollment process, when time is of the essence.

States that have refused to enact the ACA Medicaid expansion have inflated the target market for scams

The higher prevalence of plan switching in the federal marketplace (HealthCare.gov) than in state-based marketplaces (SBMs) is related to the much faster enrollment growth in the 32 HealthCare.gov states — possibly because of differences in enrollment procedure, but almost certainly because of differences in the client base. All of the 10 states that had refused to enact the ACA Medicaid expansion as of Nov. 1, 2023 use HealthCare.gov, and those states account for three quarters of HealthCare.gov enrollment and five sixths of enrollment growth in HealthCare.gov states from 2020 to 2024. In fact, the nonexpansion states account for three quarters of enrollment growth in all states in that period (7.3 million out of 9.9 million).

The lowest-hanging fruit for poacher-brokers is probably in the lowest subsidy-eligible income brackets, both because multiple zero-premium plans are available to low-income enrollees and because they tend to be lower-information. Low income enrollees are most concentrated in HealthCare.gov states, where 55% of enrollees in 2024 had income below 150% of the Federal Poverty Level (FPL), qualifying them for zero-premium silver plans with the highest level of Cost Sharing Reduction. The percentage is highest in the ten nonexpansion states, where eligibility for marketplace subsidies begins at 100% FPL, rather than the 138% FPL threshold in expansion states (below which enrollees are eligible for Medicaid). In Florida, Georgia, and Texas, the three states Appleby sites as ground zero for plan switching, 65% of enrollees have income below 150% FPL. In those three states, enrollment has almost tripled since 2020, from 3.5 million to 9.0 million in 2024.

The vast proliferation of health plans in the marketplace, particularly in HealthCare.gov states, also facilitates plan switching. The Miami marketplace currently confronts an applicant with 172 available plans. For a 40 year-old in Miami with an income of $30,000 (slightly over 200% FPL), twelve zero-premium plans are available. In Houston, six zero-premium plans are available to the an applicant of that age and income, and in Atlanta, four (with another three costing less than $1/month). An enrollee from whom no premium is demanded could go months without noticing that his plan had been switched.

While the prevalence of the kind of fraud discussed here is probably low in percentage terms, its existence and recent proliferation highlights the mismatch between the needs of low-income people and a complex commercial health insurance market with legions of profit-seeking players. Most brokers provide good and essential service - -but in a market where complexity benefits only the sellers, not the buyers. That said, the ACA marketplace provides coverage of varying degrees of adequacy to millions who would have lacked access prior to its establishment and expansion under ARPA. This particular form of market dysfunction should be easy to eradicate.

The ACA marketplace depends on brokers and agents, who now execute more than two thirds of active enrollments in HealthCare.gov states (passive auto-re-enrollees, who accounted for 22% of enrollees in HealthCare.gov states in 2024, may or may not have engaged a broker to originate their enrollment). Brokers are professionals licensed by their state; to sell ACA marketplace plans, they must also register with HealthCare.gov and/or a state-based marketplace, sign agreements, and complete required training. At present, more than 83,000 brokers and agents are registered with HealthCare.gov, up from 49,000 in 2018. The relationship is based on trust: as noted above, an agent needs only an existing enrollee’s name, date of birth and state to access her application and do anything executable in it — and that includes completing the client’s required signoff on the choice of authorized representative. The percentage of bad actors may be small, but they can do a high volume of transactions and wreak a lot of havoc. New controls are plainly needed.

- - -

* Update, 4/11/24: A HAFA fraud reporting tool is now live.

* Clarification (in progress), 4/13/2024: I used published data about marketplace broker commissions averages to suggest that marketplace brokerage commissions have roughly doubled since their low point in 2018. That is (as laid out in this prior post) , a KFF tracker using data from Mark Farrah and Associates shows a national avg. of $9.79/month in 2018 and $15.16/month in 2021, the last year tracked, and Agility Insurance Services shows premiums generally in the $20-25/month range for select insurers in select states in 2024. In the comments below, Ronnell Nolan of HAFA tells me that the alleged doubling doesn't match the experience of members she's heard from. A current average in the $20/month range seems accurate; it’s the alleged average of less than $10/month in 2018 that seems to be in question. See comments below from Nolan and broker Guy Furay, and my responses. I’m seeking further information from KFF (per my last note below, it’s not clear to me how insurers offering zero commissions in 2018 figure into the average), and I’d be happy to hear from other brokers who come across this post.

UPDATE/CORRECTION 4/16/24: The KFF/Farrah data confirms brokers’ contention that commissions have not doubled on a per member/per month basis since 2018. The PMPM averages cited by KFF divide all enrollment by total commissions paid in a given market — including enrollments that were not broker-assisted, as Cynthia Cox of KFF explained to me (TBH, the data note in the KFF tracker says as much, but I assumed that the total enrollments referred to there were broker-assisted enrollments. The percentage of enrollment assisted by agents and brokers in states using HealthCare.gov has risen from 41.8% in 2018 to 78.4% in 2024,, according to CMS. That near-doubling could account for most of the apparent near-doubling in average commissions.

Update 2, 4/16/24: Two new developments regarding unauthorized plan-switching and enrollment: 1) CMS has put out a statement promising to make victims of this fraud whole and vowing, “CMS is evaluating all regulatory, operational, and technological options to prevent unauthorized activity.” 2) Julie Appleby, who broke this story, reports today that a lawsuit seeking class action status has been filed against two call centers in Florida, their owner, and a lead generator alleged to be perpetrating the unauthorized plan switching and enrollment fraud. Two allegations in the suit touch on issues addressed in this post: 1) the suit alleges that the fraud was stimulated by the continuous SEP for enrollees with income up to 150% FPL implemented in early 2022; and 2) the suit alleges that the fraudsters used an EDE platform developed by one of the defendants (parent company to one of the defendant call centers, TrueCoverage) called BenefitAlign, which is listed by CMS as an approved EDE web broker here. The suit further alleges that the other defendant call center, Enhance Health, bought its own proprietary EDE, Jet Health Solutions (also on CMS’s approved EDE list). A large majority of EDE enrollments are executed by HealthSherpa (I estimated about 80% here; HealthSherpa claimed 90% of EDE enrollments back in 2021). Switching enrollees from a less-used platforms would likely evade detection by HealthSherpa, which notifies brokers when one of their clients switch to a different broker, or by other web brokers that handle a significant number of enrollments.