“Let’s look at the bot­tom line. The bot­tom line is this: 10 mil­lion Amer­i­cans have health insur­ance today who would not have had it with­out the Afford­able Care Act. Ten mil­lion.”  — Sen. Richard Durbin (D‑IL)

Oba­macare crit­ics have tak­en con­sid­er­able plea­sure in point­ing out the law’s short­com­ings — can­celed poli­cies, high­er pre­mi­ums and cost-shar­ing, nar­row net­works — the list goes on.

Now that the web­site’s worst prob­lems have been resolved and enroll­ment num­bers look more respectable, you can hard­ly blame the law’s embat­tled advo­cates for using favor­able num­bers a bit loose­ly. After all, if those who find fault with the law have some­times behaved irre­spon­si­bly, why should­n’t its defend­ers stretch the truth a bit?

There’s just one prob­lem: the claim that the law has result­ed in 10 mil­lion new­ly insured Amer­i­cans is more than a stretch.  It’s a whopper.

Where did the 10 mil­lion fig­ure come from?  Sen­a­tor Durbin appears to have added the num­ber of peo­ple who were report­ed to have select­ed plans in the exchanges through Feb­ru­ary 1 (3.3 mil­lion) to the num­ber of peo­ple who were deter­mined by state agen­cies to be eli­gi­ble for Med­ic­aid over the last three months of 2013 (6.3 mil­lion), and then round­ed 9.6 mil­lion up to an even 10 million.

Sim­ple math that is sim­ply wrong.  The topline num­bers from those two gov­ern­ment reports are quite mis­lead­ing.  You have to dig through their details to ascer­tain more accu­rate num­bers and con­cede that pre­cise esti­mates of how many peo­ple have gained cov­er­age through the law are sim­ply not possible.

In this post, I will look at the most recent CMS report on exchange enroll­ment, doing my best to hew as close­ly as pos­si­ble to the facts and min­i­mize spec­u­la­tion.  In a sub­se­quent post, I will look at Med­ic­aid enrollment.

For those who hate charts and num­bers, here are two key take­aways, so you don’t have to read on.

• 3.3 mil­lion peo­ple had select­ed plans through Feb­ru­ary 1, the most recent date for which com­plete data are avail­able; that num­ber eclipsed 4 mil­lion late last month, although we have no infor­ma­tion yet about those rough­ly 700,000 addi­tion­al people.
• That does not mean that 4 mil­lion unin­sured peo­ple now have obtained cov­er­age.  The num­ber is almost cer­tain­ly much low­er than that — prob­a­bly in the range of 500,000 to 800,000 — although the Admin­is­tra­tion has­n’t col­lect­ed ade­quate data to pro­vide a more pre­cise estimate.

If that is all of the num­bers you care to think about, no need to read fur­ther.  But if you want to know more, start with this chart, which you can click to enlarge.

The chart shows how many peo­ple “select­ed plans” (which is dif­fer­ent from actu­al­ly buy­ing insur­ance, as I will explain lat­er) through the exchanges in each of the first four months of open sea­son, which began last Octo­ber 1.  Per­haps the most strik­ing fea­ture is how bad­ly things were going last fall when the web­site was work­ing so poor­ly and how well things went in Decem­ber, when an esti­mat­ed 1.79 mil­lion peo­ple select­ed plans.  Near­ly 10 times as many peo­ple select­ed plans in the most recent two months (Decem­ber-Jan­u­ary) than in the pre­vi­ous two months (Octo­ber-Novem­ber), a tes­ta­ment to the Admin­is­tra­tion offi­cials and gov­ern­ment con­trac­tors who pulled the pro­gram back from disaster.

A clos­er look shows that enroll­ments aren’t trend­ing in a favor­able direc­tion.  Note that 1.15 mil­lion select­ed a plan in Jan­u­ary, a 36 per­cent drop from the Decem­ber enroll­ment fig­ure.  (The fact that total enroll­ments did­n’t reach 4 mil­lion until late Feb­ru­ary sug­gests that month­ly enroll­ments fell fur­ther last month, but that won’t become clear until CMS releas­es more com­plete Feb­ru­ary data.)

And it’s actu­al­ly worse than that.  For rea­sons that aren’t entire­ly clear, CMS stopped count­ing Decem­ber enroll­ments on Decem­ber 28, mean­ing that CMS’s “Decem­ber” last­ed just four weeks.  CMS’s month of “Jan­u­ary” actu­al­ly began on Decem­ber 29 and end­ed Feb­ru­ary 1.  So it is com­par­ing a five-week “Jan­u­ary” with a four-week “Decem­ber” and still show­ing a 36 per­cent decline in enrollments.

Chances are, peo­ple whose poli­cies had been can­celed were scram­bling to sign up for cov­er­age before the end of the year and that lots of activ­i­ty occurred dur­ing those last 3 days of Decem­ber.  If we com­pared 31-day months, I sus­pect that more than 2 mil­lion peo­ple enrolled dur­ing Decem­ber and few­er than 1 mil­lion select­ed plans in Jan­u­ary and that the actu­al enroll­ment drop was more than 50 percent.

In any event, there was no enroll­ment “surge” in Jan­u­ary, though that is how the press gen­er­al­ly report­ed it, and pre­lim­i­nary indi­ca­tions are that there was none in Feb­ru­ary either.  That’s some­what trou­bling, although it’s rea­son­able to expect that the March 31 dead­line for open sea­son enroll­ment will make March a very big month.

The next chart exam­ines the ages of peo­ple who have select­ed plans through the exchanges.

The chart shows enroll­ment by age group, with the blue bars rep­re­sent­ing enroll­ment through Decem­ber 28 and the red bars rep­re­sent­ing enroll­ment through Feb­ru­ary 1.  Over­all, the num­bers are not very encour­ag­ing.  The largest group of enrollees (30.7 per­cent as of Feb­ru­ary 1) is 55–64 year-olds, and more than half (53 per­cent) are 45 years of age or old­er.  Peo­ple over 45 out­num­ber those aged 18–34 by more than two to one.

It’s true that the per­cent­age of peo­ple in the 18–25 and 26–34 age brack­ets inched up by a frac­tion of a per­cent between Decem­ber 28 and Feb­ru­ary 1 and that the per­cent­age of enrollees over 55 has fall­en (to 30.7 per­cent from 32.8 per­cent).  But it’s still rather an old bunch and it will take quite a surge of young and healthy peo­ple sign­ing up this month to make any real change in this.

That’s bad news for insur­ers, since it makes it more like­ly that the pre­mi­ums they’ll col­lect won’t cov­er their costs.  The good news for insur­ers is this: the Admin­is­tra­tion has promised them bil­lions in cor­po­rate wel­fare pay­ments to ease their pain, some­thing I’ll quan­ti­fy lat­er in the post.

The next chart shows anoth­er puz­zling trend: most peo­ple who shopped for a plan in the exchange decid­ed not to enroll.

This chart shows that 7.3 mil­lion peo­ple com­plet­ed the enroll­ment process, pro­vid­ing a slew of per­son­al infor­ma­tion, includ­ing data about their finances, and were deter­mined eli­gi­ble to enroll in a plan.  That’s not a par­tic­u­lar­ly quick or easy process.  And yet of 7.3 mil­lion who slogged through the process and could have tak­en the final step of select­ing a plan, only 3.3 mil­lion did so.

While there has been lots of media atten­tion to the 3.3 mil­lion peo­ple who did select a plan, there’s been almost noth­ing writ­ten about the 4 mil­lion who did not.  The fol­low­ing chart takes a clos­er look.

The bar chart on the left shows that 4.2 mil­lion of the 7.3 mil­lion that were eli­gi­ble to enroll in cov­er­age through the exchanges qual­i­fied for a sub­sidy.  Of those, 2.6 mil­lion — around 62 per­cent — select­ed a plan. Not bad.

But now look at the bar chart on the right.  Of the 3.1 mil­lion who were eli­gi­ble to enroll but did not qual­i­fy for sub­si­dies, few­er than 600,000 select­ed a plan.  More than 4 out of 5 peo­ple who were told they had to pay full price for cov­er­age said no thanks.  That’s a bit trou­bling and actu­al­ly a lit­tle puz­zling, since plans are using very tight provider net­works and high deductibles to keep pre­mi­ums about 15 per­cent low­er than CBO expect­ed.  But even these pre­mi­ums are high­er than peo­ple who don’t get sub­si­dies are will­ing to pay.

One result of this con­sumer aver­sion to pay full price for Oba­macare cov­er­age is that 82 per­cent of those that have select­ed a plan qual­i­fy for sub­si­dies.  That fig­ure has been wide­ly report­ed though one of its major caus­es is poor­ly under­stood: the over­whelm­ing major­i­ty of enrollees will get sub­si­dies because so few con­sumers will buy insur­ance with­out them.

This next chart pro­vides an esti­mate of how many of the 3.3 mil­lion who had select­ed a plan as of Feb­ru­ary 1 actu­al­ly have insur­ance coverage.

This cas­cade chart shows the num­ber of peo­ple who were eli­gi­ble to enroll (7.3 mil­lion), the num­ber who select­ed a plan (3.3 mil­lion) and the esti­mat­ed num­ber of peo­ple who paid their first pre­mi­um and who, con­se­quent­ly, are actu­al­ly insured (2.6 mil­lion).  That last num­ber is based on indus­try esti­mates and sur­vey data sug­gest­ing that 1 in 5 peo­ple who select­ed a plan nev­er did pay their first premium.

But even that does not mean that 2.6 mil­lion peo­ple were new­ly insured. The McK­in­sey con­sul­tan­cy this week report­ed, based on its own sur­vey data, that only 27 per­cent of those who had enrolled in cov­er­age as of Feb­ru­ary 13 were pre­vi­ous­ly unin­sured (that’s up from 11 per­cent in a pre­vi­ous McK­in­sey sur­vey).  It also found that only 53 per­cent of pre­vi­ous­ly unin­sured peo­ple who select­ed a plan actu­al­ly paid their first pre­mi­um, com­pared with 86 per­cent of those who had cov­er­age before. So only 27 per­cent of enrollees were pre­vi­ous­ly unin­sured and near­ly half of them are still unin­sured because they did­n’t pay their first premium.

Piec­ing togeth­er these sur­vey results with the CMS data, it seems rea­son­able to esti­mate that some­thing in the range of 500,000 to 800,000 pre­vi­ous­ly unin­sured peo­ple have so far obtained cov­er­age through the exchanges.

That num­ber is noth­ing more than an edu­cat­ed guess, of course. More­over, it rep­re­sents a snap­shot.  With the enroll­ment dead­line less than four weeks away, it is pos­si­ble that a wave of young, healthy and unin­sured peo­ple are flock­ing to healthcare.gov to obtain cov­er­age to avoid the new fed­er­al tax penal­ty on the unin­sured.  We may well be look­ing at a very dif­fer­ent enroll­ment pic­ture when the final open sea­son fig­ures become avail­able next month.

But what if it does­n’t turn out that way?  What if insur­ers end up with a pool of sub­scribers who are old­er, sick­er and few­er in num­ber than they’d expect­ed?  Does that mean that they will seek exor­bi­tant pre­mi­um increas­es or stopped sell­ing through the exchanges entirely?

Per­haps.  But I don’t think so.  Insur­ers sell­ing through the exchanges do not have to rely exclu­sive­ly on the pre­mi­ums they col­lect to cov­er their costs.  They also can count on bil­lions in direct pay­ments from the fed­er­al gov­ern­ment designed to relieve them of most of the bur­den of cost­ly claims and to allow them to turn prof­its even if their rev­enues fall short of projections.

The gov­ern­ment released fig­ures this week that quan­ti­fies the so-called “pre­mi­um sta­bi­liza­tion” pay­ments that the gov­ern­ment expects to make direct­ly to insur­ers.  The Pres­i­den­t’s FY 2015 bud­get esti­mates that the gov­ern­ment will pay insur­ance com­pa­nies near­ly $15.5 bil­lion for the cur­rent plan year, including:

• $10.02 bil­lion in “rein­sur­ance” pay­ments, in which the gov­ern­ment will pay 80 per­cent of the cost of claims in excess of $45,000; and
• $5.45 bil­lion in “risk cor­ri­dor” pay­ments to com­pen­sate insur­ers whose actu­al costs exceed their pro­ject­ed costs by more than 3 percent.

This is an enor­mous amount of mon­ey that is in addi­tion to rev­enues from pre­mi­ums which, of course, also are sub­si­dized by the gov­ern­ment.  Here is a chart that com­pares gov­ern­ment sub­si­dies to indi­vid­u­als with gov­ern­ment sub­si­dies for the cur­rent plan year that will be paid direct­ly to insur­ance companies.

CBO esti­mates that sub­si­dies to indi­vid­u­als for pre­mi­ums will cost the gov­ern­ment $13 bil­lion this year and that cost-shar­ing sub­si­dies — pay­ments that reduce out-of-pock­et pay­ments by low-income enrollees for med­ical goods and ser­vices — will cost an addi­tion­al $3 bil­lion, for a total of around $16 bil­lion.  That’s only slight­ly more than the near­ly $15.5 bil­lion that insur­ers can expect in direct pay­ments from the government.

The gov­ern­ment will pro­vide almost as much mon­ey to cor­po­ra­tions that sell insur­ance as to indi­vid­u­als who buy it.

The Admin­is­tra­tion notes that it is extract­ing the $10 bil­lion in “rein­sur­ance” pay­ments from peo­ple in group health plans.  That is true. If you have employ­er-spon­sored cov­er­age, the IRS is tak­ing mon­ey from your plan and HHS is dis­trib­ut­ing that mon­ey to insur­ers that sell through the exchanges.  It is, in effect, a spe­cial tax tar­get­ed on peo­ple who obtain insur­ance through their employ­ers to ben­e­fit com­pa­nies that sell through the exchanges.

They also argue that the the gov­ern­ment will recov­er the entire $5.45 bil­lion in “risk cor­ri­dor” out­lays, a prospect that even the Kaiser Fam­i­ly Foun­da­tion’s Lar­ry Levitt finds dubi­ous.  At best, those out­lays will net to some­thing less than $5.45 bil­lion but still con­sti­tute a con­sid­er­able subsidy.

And that is why enroll­ment fig­ures, how­ev­er they turn out this year and in the future, are less impor­tant than many have assumed.  In a nor­mal mar­ket, pre­mi­ums must exceed an insur­er’s costs, which requires an insur­ance pool broad enough to be finan­cial­ly viable.  The exchanges are not nor­mal mar­kets, but gov­ern­ment con­trivances that dis­tort mar­ket dynam­ics beyond recog­ni­tion. As a result, the crit­i­cal deter­mi­nant of exchange via­bil­i­ty is not the num­ber and health of enrollees.  Finan­cial sus­tain­abil­i­ty is rather deter­mined by the com­bi­na­tion of pre­mi­um rev­enue and cor­po­rate wel­fare pay­ments.  Those pay­ments should be suf­fi­cient to keep future pre­mi­um increas­es tol­er­a­ble and insur­ers prof­itable regard­less of how many peo­ple obtain cov­er­age and how healthy they are.

A com­fort­ing thought for insur­ers; less so for taxpayers.