Time: D H M S

UPDATE 10/19/16: As you can see, I've locked in the approved weighted average rate hikes for 40 states plus DC, leaving 10 states to go. I do plan on filling in the remaining approved rate hikes as the data for those 10 states comes in, but at this point it's quite clear that 25% is the magic number. The weighted average has been hovering between the 23-26% range since the first few approvals started being publicized in mid-August, and has stabilized in the 24-25% range for the past month. Over 77% of the total U.S. population is represented by these 40 states (+DC); unless there's some dramatic final rate changes in the remaining 10 states, that national 25% average isn't likely to budge by more than a rounding error.

As proof of this, I assumed that the requested rate hikes are approved exactly as is for all 10 states.

Result? The national, weighted average rate hikes went from 25.36%.

At the beginning of 2016, the Kaiser Family Foundation ran an analysis of just who the remaining uninsured Americans are by income, employment status, legal residency status and so on. They came to some interesting and important conclusions, many of which I incorporated into my own massive "SHOW YOUR WORK" pie chart project back in April, which attempted to break out the entire U.S. population by type of healthcare coverage (or the lack thereof, in the case of the uninsured).

Yesterday, Kaiser published a completely updated version of their analysis. In addition to having more recent data to work with (including the increased number of ACA exchange enrollees, the increased number of people covered by Medicaid thanks to ongoing ACA expansion progress, etc), they also made some changes in their methodology. The most obvious change is that their estimate of the total uninsured (non-elderly only; you'd have to bump the number up by perhaps 1 million more if seniors were included) has dropped from 32.3 million to 27.2 million nationally.

In addition, however, the proportion of people falling into different categories has shifted as well. Here's how they break it out:

OK, now that both HHS and I agree that the ACA exchange target for the 2017 Open Enrollment Period is around 13.8 million QHP selections, it's time to dust off The Graph and reset it for 2017.

It was fairly easy to do so this time because this is the first year that the starting and ending dates match (November 1st to January 31st). All I really had to do was bump up the 1/31/17 projection from 12.7 million to 13.8 million, and the rest of the projection line adjusted accordingly.

With that in mind, here's how I expect things to play out for Year 4 (click below for a higher-res version):

Some important notes to keep in mind:

Unfortunately, due to an ongoing side project of mine, I haven't had a chance to write up a full analysis/projection for OE4 as I usually do around this time.

However, I've been informed that the HHS Dept. plans on issuing their 2017 Open Enrollment projections in the near future, so I'm throwing this out there quickly:

I expect somewhere between 13.5 - 14.0 million people are likely to select QHPs via the ACA exchanges during the 2017 Open Enrollment Period (which runs from 11/01/16 - 1/31/17).

As always, it's important to remember that QHP selections are not the same as effectuated policies...there's usually around 10% or so of enrollees who never bother paying their first premium and thus are never actually enrolled, and of course there's attrition after that as people drop their policies after 1, 2, 3 months or more for various reasons (many replaced by new enrollees via SEPs and so forth).

UPDATE: Oh for heaven's sake. Turns out HHS is projecting 13.8 million...but no one will ever believe that my own expectations were around the same range because they announced it half an hour ago and I didn't notice.

(sigh) OK, believe me or don't. So be it.

Due to an ongoing external committment, I'm not able to write up a full analysis of this, but CMS has issued their "Mid-Year Effectuation Report" which updates at least some of the ACA exchange stats for the first half of 2016:

October 19, 2016

First Half of 2016 Effectuated Enrollment Snapshot

For the first half of 2016, an average of 10.4 million consumers had effectuated Health Insurance Marketplace coverage – which means those individuals, paid their premiums and had an active policy through one of the Health Insurance Marketplaces nationwide as of that date. [i] Effectuated enrollment is generally lower in January and February, since coverage purchased in the weeks before the final Open Enrollment deadline does not begin until March. June effectuated enrollment was slightly higher than the average for the first half of the year, about 10.5 million.  These amounts do not include individuals enrolled in coverage through New York and Minnesota’s Basic Health Programs, which currently enroll about 650,000 people.

The ACA exchange in Arizona has had some pretty dramatic turns over the past month or so. When the dust settled, every county in the state will still have at least one carrier offering plans on the exchange...although only one. Anyway, today the AZ DOI joined Pennsylvania and Michigan in releasing their final approved rate hikes for both the individual and small group markets:

Right on top of Pennsylvania, the Michigan Dept. of Insurance has issued their final approvals for 2017 individual and small group market rate increases. As has been pretty typical this year, the final approved rates aren't all that different from what was requested; a little nip/tuck here and there, and the 17.2% average requested has been slightly trimmed to 16.7% approved for the indy market. Meanwhile, the small group average is barely noticeable: 2.6% requested, 2.5% approved. Unlike most states, the MI DOI has already done most of the heavy lifting for me, so I don't even have to use my own spreadsheet to calculate the weighted average.

(thanks to commenters "M E" and "joe" for the heads up).

The state Dept. of Insurance has released their approved rate hikes for 2017, and it's bad news in two different ways. First, the overall full-price average rate increase looks like it'll be roughly 32.5%...over 8 points higher than the original rates requested by the carriers. Secondly, even with those higher increases, two more indy market carriers (Keystone Health Plan and Geisinger Quality Options) are pulling off the exchange, although both will continue to offer off-exchange plans.

It's important to be careful with the full carrier names here, because they often operate under several different very similar ones (Keystone Health Plans vs. Keystone Health Plan East, for instance, which is not pulling off the exchange).

So here's what's happening with each:

Friday, September 30, 2016
Organization US Congress Joint Economic Committee (JEC)

Dear Mr. Gaba,

First off, let me say that I appreciate the work that you're doing here on the ACA Individual Market premium numbers. The data you provide is both comprehensive and useful, as well as possibly unique; I haven't come across another consistently-updated source for state-by-state premium figures.

I work as a Research Assistant for the Joint Economic Committee. Our committee is charged with keeping members of Congress up-to-date on the latest economic data. As you might expect, our members are keenly interested in information about the premium changes that their constituents can expect in the 2017 open enrollment period, and the JEC had hoped to cite back to your state-by-state premium increase data. Would you have any issues if we did so?


(name withheld)

Joint Economic Committee – Republicans

Tuesday, October 04, 2016

Dear (name withheld)—

When I originally calculated the average requested rate hike for New Hampshire, I came up with a weighted estimate of around 13.1%. A month later, the average dropped a few points...but not for a good reason: One of the remaining ACA-created Co-Ops, Community Health Options, decided to pull out of New Hampshire (they started out as a Maine-only operation, expaneded into nearby NH for the 2nd and 3rd year, but are pulling back to Maine-only again). Since CHO would otherwise have been requesting a more than 40% increase, them dropping out actually lowered the average increase for everyone else. This obviously illustrates a major caveat with my "average rate increase" methodology: It only applies to those who are able to renew their existing plans. The moment a carrier pulls out of parts/all of a state, or drops PPOs (while keeping HMOs), etc, I have to remove a portion of the existing enrollees from the equation completely.

I noted about a month ago that Blue Cross Blue Shield of Alabama had lowered their original 39% rate hike request a bit to 36.1%. As has been the norm this year, the state regulators went ahead and approved the final request. Kudos again to Louise Norris:

For 2017, only Blue Cross Blue Shield of Alabama will participate in the exchange. In August 2016, the carrier filed rate increases for 2017 that average 36.1 percent (with a range from 20.6 percent to 38.3 percent). This was a revised rate filing, and was slightly lower than the average rate increase proposal of 39.3 percent that the carrier initially filed in June.

The Alabama Department of Insurance approved the 36.1 percent average rate increase in October 2016, and the new rates will take effect in January 2017. reports that pre-subsidy rates for Bronze plans will increase between 20 percent and 23 percent, while Silver and Gold plans will increase in price between 32 percent and 38 percent.

Blue Cross Blue Shield of North Carolina originally requested an 18.8% rate hike back in June, but after the Aetna pullout, they revised their request upwards to 24.3%. Cigna, which is expanding onto the ACA exchange next year, followed suit by bumping up their request from 7% to 15%.

I haven't seen any formal announcement from the NC Dept. of Insurance yet, but BCBSNC just posted the following blog entry announcing their 2017 rates...and it certainly looks like the 24.3% request was indeed granted as is:

Blue Cross and Blue Shield of North Carolina customers purchasing ACA plans on the individual market will see an average increase of 24.3 percent in their premiums for 2017, compared to this year’s rates. That’s higher than our original rate filing back in May (an 18.8 percent increase).

When I plugged the numbers in for Utah way back in June, I came up with a weighted average request of around 30.7%. 

Louise Norris gave me a heads up that the approved rates were in for UT, and sure enough she's correct. Not a whole lot to report, however; most of the requests were approved as is, with only minor modifications; the approved average is slightly higher:

Regular readers may be a bit confused here, as Oregon's insurance dept. had already approved 2017 rates back in early August, for a statewide average of around 23.8%.

However, as reported by Louise Norris:

But when four carriers (Atrio, BridgeSpan, Providence, and — off-exchange — Regence) agreed in August to cover a broader service area than they had originally intended for 2017, state regulators also allowed them to further increase their premiums due to the increased risk they would be shouldering. Final approved average rate increases for Oregon’s exchange carriers are as follows:

  • Atrio Health Plans: 29 percent (up from the originally-approved 20.8 percent).

  • BridgeSpan: 23 percent (up from the originally-approved 18.9 percent).

  • Kaiser Permanente: 14.5 percent (range is from 10.9 to 22 percent).

I originally estimated the requested average rate hike for Wisconsin indy market carriers back in August. I came up with a weighted average of around 20%...but this was questionable due to my not being able to come up with the actual enrollment figures for 4 of the 15 carriers in the state (note: several of these have more than one entry for different types of plans):

A couple of days ago, the Wisconsin Insurance Dept. announced the approved rate increases. Unfortunately, the articles about it don't provide hard numbers for either the rate change or enrollment figures for each carrier either, but they did provide the overall weighted average increase, which is really what I'm trying to calculate anyway, so there you go: