How To Read the Evasive October Health Exchange Numbers
So the October exchange data is out and the report is a bit hard to follow, in part because it dodges and weaves trying to put the best face on things. Fortunately, I have years of experience as a corporate planner digging into numbers from division heads trying to disguise what a train wreck their results are. So here are the numbers in a simple graphical form (click to enlarge)
Here is a simple narrative following these numbers: The exchange web sites had 26,876,527 visitors representing 47,840,217 estimated potential insured persons. Of these, 846,184 applications have been completed covering 1,509,883 persons, of which applications covering 1,477,853 persons have been processed by the government to test eligibility. Of these reviewed applications, 396,261 persons were eligible for Medicare or some other free program while the rest needed private coverage. Of these, about a third were deemed eligible for a subsidy. About 10% of those people eligible for private coverage have put a plan in their shopping cart, though it is unknown how many are subsidized and how many are not. An unknown number have actually purchased insurance. An unknown number of Medicaid eligible people actually have enrolled.
There are some real problems with the report's presentation. Here are the worse issues:
- They switch back and forth between applications and persons covered by applications (which is about 1.78 persons per app.) This is presumably a bid to make the numbers as large as possible. All the numbers above the first one in the chart above are persons covered by applications, not applications. As you can see, I have converted the web site visitors to this same basis so we can get an apples to apples sales funnel. Note that this means the 106,185 number for people who have "chosen a plan" is actually a lot fewer applications, perhaps less than 60,000.
- They leave out the three numbers any reasonable person would most want to see. How many people actually signed up and (if appropriate) paid for coverage? Those numbers are completely missing. How many Medicaid eligible people actually enrolled? How many of the 106,185 people covered by a plan in a shopping cart actually paid (the shopping cart abandonment rate at private websites is about 2/3, if I remember correctly)? And how much did these enrollments cost the taxpayer in terms of subsidies?
As bad as this report is for the administration, the truth is actually worse, as they have assiduously avoided including the numbers a reasonable person would want to see. Without any other evidence, I have to assume that these obvious numbers were left on on purpose because they were awful.




