Skip to main content

Health Insurance Costs and CPI

By November 22, 2023Commentary

Other people have noticed that when it come to measuring inflation, the Bureau of Labor Statistics does some goofy things, especially in regard to medical services and health insurance.  If you want to take a closer look at how changes in prices are calculated, here is one explanation, at the BLS site.  (Methods Handbook)   You might notice a couple of things right away.  One is that this is a survey of businesses and consumers.  What, it is too hard to go actually collect representative prices?  In this day and age of online shopping, etc?  Of massive databases of credit and debit card charges and electronic payments?  You can’t collect actual data on rents paid (not quotes for current rent on vacant apartments), on food prices, on gas prices, on clothing prices?  You can’t find out what payers actually pay for medical services?

The second thing you would notice is that these surveys are only in urban areas, where many but not all Americans live.  Again, with the kind of databases on what people are actually paying for different services and goods, why is this how inflation is determined?  A third thing is that major categories of consumer spending are just missing.  Here is a big one–taxes, which to some extent represent the cost of government services.  Taxes of all types continue to rise rapidly for most consumers.  Interest is not included, but interest costs are also a major component of spending.  Interest rates rising is inflation in the cost of buying money.  Excluding these ignores important contributions to the inflation consumers actually feel.

But my own area of focus, health care, is particularly interesting.  According to the chart of categories, I don’t even see health insurance listed, although it clearly is considered in the overall inflation index.  In regard to medical services, you would note if you look at response rates that they hover in low 30% range, which is very poor, but may reflect the fact that most people don’t pay most of their medical costs themselves.  Most medical care is paid for by Medicare, Medicaid or private health insurance issued to employers or individuals.  All of these payers have large claims databases that tell you exactly what is paid for every service and they tell you how much the consumer paid as a copay or deductible.  Most providers are now in large systems that have billing software that also will tell exactly who paid what for which services.  That is what should be used to determine inflation in medical services.

Health insurance, as this release describes, is just goofy.  Since health insurance pays for a lot of medical services, the BLS is concerned about in essence double counting, so it tries to ignore most of the health insurance premium and just measure changes in the cost of administration and in profits.  This hasn’t work very well, so BLS has to make adjustments, the most recent of which is described here.  (BLS Memo)   I am not sure I understand what the hell they are doing, but whether it is medical care or health insurance, BLS has at least over the last two years grossly understated actual price increases.  The Kaiser Family Foundation work on health insurance premiums demonstrates this clearly.  And other work on medical care costs also shows much greater increases than BLS does.  This is 20% of the economy, so screwing it up is a big issue.  A Zero Hedge post also discusses the problem.  (ZH Post)

Unfortunately, if I looked at closely at other major categories as I do at health care, I think I would see the same problems.  And I come back to my basic question.  In a world awash in actual payment data, why is BLS using surveys and adjustments and other dubious methods?  Could it be that it wants to be able to manipulate the data?

Join the discussion One Comment

  • D says:

    Thx for pursuing this. The legacy media used to cover this, but their time helping the public are over. I’m glad truth is still sought. I’m thankful for your work and voice.

Leave a comment