Vox published a report from Jay Angoff, a former Obama administration official who is now at a law firm. The Angoff report examines the federal cost of the health exchange websites and then calculates the cost per enrollee in each state. The headline is deliberately enticing: “Hawaii’s Obamacare cost nearly $24,000 per enrollee.”

There are two problems with the Angoff’s analysis, one major and one minor but still significant:

1. The report takes the $2.7B cost of the federal website and divides it equally between the 36 states that used the federal website.
2. The report does not include state spending or spending by non-government groups.

Major Flaw: Federal Spending is Allocated by State

The Vox article does not mention the math where the $2.7B federal price tag is divided equally by 36 states ($75M per state). This is a significant omission. This assumption is explained in the report:

Administrative costs are allocated 1/36 to each of the 36 states, regardless of population, because the technology necessary to operate the Exchange is substantially similar in each state regardless of its population. To the extent that non-technology costs are higher in larger states, this allocation overstates cost-per-enrollee in smaller states and understates it in larger states.

No, no, no! This is a poor assumption! I would not allocate the cost equally to each state. That would be like taking an F-22 fighter jet and allocating its cost by state. After all, the technology to operate it is the same in every state!

I took the analysis and revised it by two methods. First, it would be better to allocate cost by enrollee. This makes sense as the enrollees should “pay” for the site. The second run-through, I calculated cost by population which is a sensible way to allocate costs on something that is theoretically available to all. It may be best to allocate costs by the number of taxpayers per state but that gets a bit complicated.

The spreadsheet here shows the states that were included in the federal website, sorted by the cost per enrollee as in the Angoff report, then followed by two columns: 1) cost by enrollee if the costs were allocated by enrollee and not state and 2) cost per enrollee if the costs were allocated by population.
As you can see, there is a massive difference in the numbers. Because of Angoff’s methodology of dividing costs by state, the “most expensive” enrollees are all from states with small populations.

Below is my entire analysis. As a side note, I’m depressed impressed surprised about the amount of federal money that some states (Hawaii, New Mexico) were able to obtain.

Minor Flaw: Lack of State Spending in the Analysis

The report also suffers by lack of state spending and lack of spending data from non-government groups. While it may be difficult, or impossible, to get spending from non-government groups, state spending is easier to obtain. As an example, Maryland received $171 Million in federal dollars but added an additional $20M of its own money. The report underestimates Maryland’s cost by 15% – and Maryland’s costs are born solely by Maryland’s taxpayers (not equally by county if the report chose to make the same assumption as above).

Overall, the Angoff report is fatally flawed by making a poor decision to allocate federal costs by state and not including state spending. To its credit, Vox noted the missing data but it did not mention the assumption that grossly distorts the spending data for the health exchanges.