by David Ramsey
**DHS issued a correction the day after this post. The $5,200 and $5,975 numbers provided to reporters were incorrect. The percentage differential and the total difference in cost are unchanged. You can read a detailed explanation of the correction here. I have highlighted sections in the post below that use the inaccurate numbers. Again, broadly speaking the conclusions and questions remain the same.
The Arkansas Department of Human Services told reporters today that the cost of offering folks health coverage through the "private option" in the state will be about 13 to 14 percent higher than offering coverage via traditional Medicaid expansion. When other factors are included, the number gets close to zero. They project that the incremental costs of the private option for the feds would be $120 million per year at the high end, but with some additional offsets not enumerated that would bring it down to a rounding error.
Read a summary of the DHS findings here. Also worth noting — I asked whether they had any concerns whatsoever of DHS walking back the "private option" offer. I received a one-word answer: "No."
DHS commissioned a study from legal consultants, actuarial consultants and the Arkansas Insurance Department. They found Arkansas-specific factors — as well as cost benefits of moving a large volume of beneficiaries to the private market on the exchange — that they believe will mitigate many of the cost concerns raised by previous estimates about an approach using private insurance companies.
Probably the single biggest factor is that they found that the differential between private health insurance premiums and the cost of Medicaid was significantly lower in Arkansas than other places, about 25 percent. They did not speculate on why this is so, but their finding gives the state a much lower baseline.
The rosier cost projections also depend on predicted benefits from competition. DHS and the Insurance Department believe that doubling the pool of folks on the exchange will bring a significant number of additional carriers into the fold. "Just since this has been on the table, we’ve already seen companies that have inquired to the Department about coming in and participating," Insurance Commissioner Jay Bradford said. "This concept has really sparked interest over the last two or three weeks."
The new study focused on the impact on the feds. "The question that continued to come up was that even if this looks good for constituents and good for the state, what does it do for cost in general, particularly for federal cost," DHS Director John Selig said. "There has been a CBO estimate out there…the good news is that…the added cost to the federal government, instead of a 50 percent increase, it looks like it’s short of 15 percent and could be as little as zero."
Those skeptical of this conclusion, which represents a dramatic shift from previous experiences with premium assistance, will likely be left with questions. While DHS provided a summary of their findings, linked above, and spent an hour with reporters today explaining their general framework, we have only a bare-bones understanding of the methodology and little in the way of hard numbers. At this point, we have assertions, not equations.
When I suggested that the projections seemed to be optimistic, Medicaid Director Andy Allison begged to differ. "I don't believe that this is an optimistic estimate, I believe it's a conservative estimate," he said. "It's possible that that rate relief could be greater....I think that the possibility that this marketplace could be significantly more efficient and competitive than the marketplace that we have now is real. We're not assuming some sort of Pollyanna-ish level of competition. This is from actuaries who've seen, for example, Medicaid managed care for similar populations in other states that cover pretty much this level of income."
After the jump, a closer look at the DHS findings, some questions, a skeptical economist, some thoughts from lawmakers, and more. Wade in to the weeds!
Using projections that were completely new from previous DHS studies, they predict that the average cost of insuring someone from the expansion pool via Medicaid would be $5,200 in 2014 (this is on low end according to DHS, for reasons I'll explain below...but looks kind of high based on everything else I've seen). The average cost of insuring someone in the expansion pool via the "private option," they project, would be $5,975.
Wait, you ask! What about those $6,000 and $9,000 numbers from the CBO? As we've reported before those numbers, even aside from inflation (those are 2022), are too high for Arkansas. But the key question has always been whether the differential — 50 percent — was applicable. DHS says no, partly because Arkansas is unique and partly because they believe that the "private option" plan will achieve more to bring down costs of private insurance than previous premium assistance plans.
So how do they get all the way down from 50 percent to the 13-15 percent range?
1) DHS says that the 50 percent figure that the CBO used is "is significantly higher than the actual differential in Arkansas." Based on an actuarial review, they found that the rate difference between the state's private market and its Medicaid program is less than 25 percent. I've speculated about this before — is there something about Arkansas that would make not just costs lower but the difference between public and private spending lower? I couldn't think of any reason that would be so! The DHS folks said that "why?" was a good question, but didn't have an answer. As an empirical matter, though, they said Arkansas has an unusually low differential.
2) Starting at a baseline of 25 percent, they knock off 5 percent because they project that introducing the expansion pool "will increase competition among carriers and generate some price pressure on providers, since they would be compensated at competitive rates for all clients."
Here's Arkansas Medicaid Director Andy Allison:
Adding as many as a quarter million adults under 138 FPL to the exchange has a radical, very positive impact on that marketplace. It makes that market far more attractive to private carriers. We’ll likely see more competition.
What would happen to provider rates in the exchange with so many more paying customers included? The expectation is that provider rates wouldn’t need to be as high as they are now in order to secure access for that new population. The estimate that we’ve included, based on actuarial advice, experience around the country, and knowing Arkansas’s market, is that there’d be as much as 5 percent rate relief—and we think that’s conservative—in the exchange.
3) They believe that private health insurance companies will be more nimble and effective at health-plan management and allowable cost-sharing to impact consumer behavior. This knocks off another 5 percent.
4) Now we're down to 15 percent. A small portion of patients within the expansion pool will remain on the state Medicaid program. The feds would not be subject to the rate differential for those folks so that knocks it down to the 13-14 percent range.
Okay, but we're still talking about somewhere around $190 million more per year. That's a lot. But there are three additional factors outside of a Medicaid/private rate comparison for the expansion pool:
1) Because DHS believes that the "private option" will drive down prices on the exchange, this would also save the feds money on the folks between 138 and 400 FPL that they'll be subsidizing. They project this to save the feds around $70 million per year.
2) They think that traditional Medicaid expansion would drive the cost of providing Medicaid coverage up (not accounted for in this or previous Medicaid cost estimates). In order for Medicaid to provide sufficient access for the expansion pool, they think reimbursement rates would have to go up. "Arkansas' existing Medicaid program appears to provide sufficient access," they write, "but adding 250,000 participants would require significant new commitments by providers to meet the increased demand for medical services." Achieving that access would require spending more, and DHS suggests that traditional Medicaid expansion would pull Medicaid reimbursement rates up toward...well, toward what reimbursement rates would be under the "private option."
3) Finally, to the extent that rates under the "private option" would be higher, this would increase tax revenues. More spending means more offsetting taxes; if Dr. Joe gets a little bit better reimbursements (private v. Medicaid), he'll have slightly more taxable income.
DHS does not provide numbers for #2 and #3 above, but they believe they will produce significant additional offsets. So according to their projections, the additional cost of the "private option" as opposed to Medicaid would be between 0 and $120 million per year.
Lots of questions! Here are some to start.
• I asked how competition would reduce prices if consumers in the expansion pool weren't paying anything out of pocket for premiums. Wouldn't that drastically limit price sensitivity?
DHS's response is that the subsidies will be designed to encourage price competition between carriers. If, say, the second-lowest priced plan in a particular tier is the cutoff line, then multiple carriers will be competing to offer low prices in the bidding process. (If your bid is just above the line, that's potentially a giant pool of customers that would have to pay out of pocket to buy your plan, which they'll be exceedingly unlikely to do.)
"The tax subsidies are pegged relative to other plans," Selig said. "We’re expecting a great deal of price competition among carriers that want this business. We need to make sure the rules assure price competition among carriers."
I asked Boston University health economist Austin Frakt, who has argued persuasively that "private option" style plans will cost the feds more, about this. He said:
There will be competition in exchanges. But Medicaid-eligible consumers won't face the full marginal cost of care because nearly all the cost sharing is picked up for them. That's likely to lead to greater utilization and higher costs. Nevertheless, the bidding structure will help. It's is a lot like Medicare Part D. There are some differences, but the key similarity is that the subsidy level is tied to bids. Even with that, do the Part D plans offer a product cheaper than Medicaid or the VA can get? Not even close.
• The D-G's Charlie Frago, noting that the price projections are heavily dependent on competition, asked about how many companies they were expecting to participate and how many would be necessary to achieve the projected downward impact on rates.
"We’ve had six inquiries but no formal applications and three stand-alone dental companies that have inquired," Bradford said. "I really think we’ll attract even more once this thing becomes official."
We'll know what companies are participating by the end of June. Bradford said they hoped to have 10 to 12 choices. Would that be enough?
“I think the assumptions are valid with significantly fewer choices than that," Allison said. "We’re comfortable that the interest levels are high enough, the appeal of this market great enough and the existing players and letters of intent sufficient enough."
“The actuaries made an assumption that there would be enough for it to be competitive without saying that there would be a specific number,” Selig said.
Just how many is enough?
“That is an open question,” Allison said.
“It depends on their appetite," Bradford said. "If a company comes in that’s very competitive, that’s going to solve a lot of problems for our consumers.”
• Are there additional costs, administrative or otherwise, associated with wraparound coverage?
Allison: "Expected admin costs are lower under this option than they would be. You're not paying a per claim charge, etc. For those individuals that need services that go beyond the essential benefits package offered in the qualified health plans of the exchange, we would need to provide those. We're assessing those benefits now and don't believe that for the population that would participate that there would be very many of those. We don't think the administrative costs of doing so would be significant."
"Particularly because the people that are the most frail and sickest will stay in the traditional Medicaid program," Selig added.
• Won't adding a pool of low-income folks drive premiums up?
"Understand that low income is very highly correlated with age," Allison said. "Many of these individuals will be young, many will have health benefits of being young, which are likely going to be far more powerful than the negative impacts of being poor."
But all else being equal 0-138 is riskier pool than 138-400, right?
"I think that remains to be seen," Allison said.
• Charlie asked whether the estimated 5 percent reduction in reimbursements was a concern for providers?
DHS believes that reimbursements will still be higher than Medicaid reimbursements (and certainly better than uncompensated care without expansion) so providers will still view this as a net positive, even if that 5 percent reduction also hits providers on folks outside the expansion pool.
• Does the unusually low amount of carrier competition in Arkansas offer more room for enhanced competition in a unique-to-Arkansas way?
Possibly, but, as Allison pointed out, Arkansas isn't that unique when it comes to this pool. "There aren't a lot of places now that have a healthy market for insurance coverage for these low-income adults," he said.
• So this projection is right and everyone saying that it will cost more is wrong?
We are directly challenging those assumptions. We do not believe [higher cost differential] will occur in Arkansas. We are rooting our analysis in standard actuarial assessments and rules of thumb, in Medicaid data from the state, and with a fair amount of input from the private sector and the private marketplace here. Taking all of that to estimate [Medicaid expansion] and [the "private option"]. And I think what you'll find if you look at competing analyses around the country is that they will have misjudged if not one, both. They're either not taking into account what would need to happen in the Medicaid program if it were just simply expanded or they're looking at this option to buy in to private plans and misjudging the nature of the competition, the nature of the prices that would be paid.
• Don't we have enough experience with premium assistance and using public money to buy private health insurance to know that "cost comparability" is hard/impossible?
"This is a brand new policy option," Allison said. "It's a context that hasn't existed before."
"A big part here is the volume," Selig said. "When you're taking a few Medicaid clients and putting them in to a few different plans, you just have no impact on rates, no impact on behavior. Bringing 250,000 people into this market, potentially bringing more carriers.."
One big question:
The "private option" per-person, per-year (PPPY) cost looks more or less like what you'd expect: $5,975. But the Medicaid PPPY number, $5,200, seems very high (this figure doesn't include the inflationary impact of expanding the pool mentioned above). For starters, remember that CBO estimate that everyone agrees is too high for Arkansas? That's $6,000 in 2022 dollars. Medicaid PPPY in Arkansas, with very low healthcare costs, is just $800 less in 2014 dollars? Plus, the old DHS estimate, which we were told numerous times intentionally estimated high on the cost side, had a PPPY of $3,900 (in 2012 dollars...confusing I know!).
Massachusetts, probably our best comp for Medicaid/exchange pricing, had PPPY costs less than $3,000 for individuals on Medicaid in 2009, and they have more expensive healthcare than Arkansas. (I think the "liberal grad students" will have more on this...). Didn't get very far on this today, but I did ask Allison why the number was so much higher than the PPPY cost of non-disabled adults in the current Medicaid program ($1,237 in 2009). Allison said that there were a small but significant number of high-cost outliers in the pool ("near disabled) that would drive up the average cost dramatically. Will follow up on this...
A few other notes:
• I know some liberals were concerned about this not being explicit in the DHS letter to the feds, but they once again confirmed that for folks below 100 FPL: all cost-sharing protections and benefits requirements under the January rules for the original Medicaid expansion remain in place.
• Parents 0-17 FPL will go to the exchange. The feds will continue to give the state the pre-ACA 70/30 match rates on that pool (boy, lots of opportunities for the state to save money if income line is close!).
"We're very excited," Matt DeCample, spokesman for Gov. Mike Beebe, said. “The past week or two the primary concern had been 'wait, isn’t this just going to hurt the federal deficit that much more if Arkansas does this private option?' The only numbers that anyone had were these CBO numbers. [Now] we’re able to look at it specifically for Arkansas, and those numbers look a whole lot better."
"These are initial actuarial estimates and we'll look at them," Sen. David Sanders said (Sanders has previously told us that he thought folks were overestimating the additional cost to the feds). "Actuarial analysis proves that even though an insurance exchange is a regulated market, it's still a market. And it functions differently than the standard Medicaid. It's a different animal.
"I think it's encouraging news about the process. This allows us to continue our work....I think when this is reported nationally, I think the whole ballgame just got changed."