An Interesting Look At Inflation

[Independent]

Originally Posted by RockOn

On Williams, you'd have to do what John Williams (supposedly) does, find out how it was calculated then compare it to how it is done now. Not an easy task. You might find Williams is actually accurate, more than you give him credit for. How can you dismiss his data without knowing if he is correct? If he is correct, how can that be explained? Is the world different now than it was in the 70's, 80's or 90's?

Well, I'll punt on trying to disprove Williams. I did give you a link to a BLS paper in which they look backwards and calculate the impact their Boskin related changes. It wasn't 3% per year, but, of course, that paper was produced by the BLS.

I look at how transparent the BLS is, and how many outside "auditors" look at their methods. Then I look at how opaque Williams is. I know which approach makes me feel more comfortable. But, that's just the way I am, it doesn't prove anything.

[RockOn]

Originally Posted by Independent

Thanks for the link. It certainly shows different numbers than the CPI, and it covers a pretty broad group of people.

If you follow the link in the footnote, you’ll find that the data is for 2001 and 2005. So the 30% increase over four years equates to an average of 6.7%, compounded. But the employee share of the premium also went up over that time, so the average employee cost went from $1,921 to $2,185. That’s a 35% increase and 7.7% compounded.

Over the same period, the average CPI increase was 2.5%, and 4.8% for the “Medical” category. That’s nearly 3% lower than the increase in employee insurance costs. What the … ?

Here’s a clue, over the same period the total per capita US spending on health care went up 6.5%. The total cost should include any heavier use of services and entirely new services. Neither of those factors gets into the CPI, but they show up in both total spending and insurance premiums.

For example, one of the fastest growing medical categories was prescription drugs. The CPI has them increasing at 3.8% per year, even lower than total medical costs. But total US spending was up 8.5%. Much of the difference probably comes from the fact that between 1994 and 2005, the number of prescriptions per person went up at a rate of 3.8% per year. Using that average for 2001-5, we get prices+more scripts = 7.6% annual growth. But some of those new prescriptions were for entirely new drugs. If they average a higher price than older drugs, it’s easy to close the remaining gap. The doctors who prescribed those drugs would probably say that their patients got better medical care from those additional prescriptions.

I expect this is the issue with the other components. Some of the increase in insurance premiums is higher prices, some is heavier usage, some is new stuff (higher quality). The CPI is only supposed to catch the higher prices, so we see a gap.

I think this quality/usage thing is extremely hard to measure in a fast moving field like medical care. I’m not going to try to convince either myself or you that the CPI has got this one nailed. I think that the 7.7% growth in insurance premiums is clearly “too high” for a price index, but I don’t know if the 4.8% in the CPI is right. I’m not saying that total expenses aren’t going up by 7.7%, nor that people don’t feel stressed by this. I am saying that total expenses are going up faster than prices.

---

The other point you raise is the “weight” that the CPI gives to the medical component. You doubt the 6%. Looking at the data you provided, families with group health pay $2,185 or 6.1% of income for their insurance premiums alone. But, only 71% of families under 65 have private health insurance of any kind.

We know that Medicare, Medicaid, and other gov’t pay 46% of total health costs. The CPI includes these families that use gov’t to pay for much of their costs, not just those of us who use only private sources. That doesn’t make the rest of us feel better, but it’s all that the BLS is claiming to do.

If you look at the aggregates, only 12% of total health costs are paid directly by individuals, 34% are paid by private insurance, and much of the insurance cost is paid by employers. If you know that total health costs are 16% of GDP, or 21% of personal income, it’s pretty easy to get a number far closer to the CPI's 6% than to the 20%+ that you were suggesting.

I used these sources for the numbers above: Historical, (the ".... source of funds" download).
http://www.kff.org/rxdrugs/upload/3057_06.pdf
Bureau of Labor Statistics Data

I'll just surrender at this point. Your explanations are all well thought out and logical if you choose to believe the logic. When you read your post, it sounds like what I think is going on, they include this group, exclude that group, include this service, exclude that service, include this product, exclude that product, etc. Statistically it's possible to come up with any result you want, i.e. my belief they are "lying with statistics".

I think the CPI should be based on the "average" Americans fuel costs, the health insurance premiums, grocery costs, mortgage costs, the utility bills, education costs, entertainment costs, vacation costs, etc., etc., etc., I don't think it would be that hard to come up with an average American (my opinion is that it should be a family of four) and track what their inflation really is. CPI should equal the average Americans cost of living increase. Keeping track of total expenditures like "other linens" and basing CPI on that, isn't valid IMO. That method is currently reducing reported CPI because of all the low cost products coming out of countries like China. Many of those products are not essentials. I suspect that John Williams old calculations reflect a measurement of CPI which is much closer to my thinking. I am fairly sure the average American family today does not believe the CPI is correct for them. In the 70's we believed the CPI numbers were correct because they reflected the reality people were experiencing. Not true now for most of us.

[mb]

Originally Posted by clifp

In contrast, I just priced sheets at Costco $80 for good queen sheets, throw in beach towel (for us West coast folks) that is $100. .2% of average family making $50K = $100. So a set of sheets a year and few towels and there you go. Only a small percentage of the households buys fuel oil every year, .

So we therefore need the fuel oil inflation index for New England and the beach towel inflation index for CA and HI.

I'm happy to be subject to the beach towel inflation index

MB

[RockOn]

Originally Posted by mb

So we therefore need the fuel oil inflation index for New England and the beach towel inflation index for CA and HI.

I'm happy to be subject to the beach towel inflation index

MB

Maybe what is happening is that most of us are getting hit with a few of the large outliers (college, healthcare, fuel, and groceries for me, fuel oil for others) drastically impacting the CPI-P's for a large number of us, while CPI-U averaging across tons of catagories is canceling things out across the board. Just a thought.

Sort of like the theory of gains coming from a "highly diversified portfolio of non-correlated assets" if you get what I mean.

[Independent]

Originally Posted by RockOn

I think the CPI should be based on the "average" Americans fuel costs, the health insurance premiums, grocery costs, mortgage costs, the utility bills, education costs, entertainment costs, vacation costs, etc., etc., etc., I don't think it would be that hard to come up with an average American (my opinion is that it should be a family of four) and track what their inflation really is. CPI should equal the average Americans cost of living increase. Keeping track of total expenditures like "other linens" and basing CPI on that, isn't valid IMO. That method is currently reducing reported CPI because of all the low cost products coming out of countries like China. Many of those products are not essentials.

I think this sums it up pretty well. The pie chart that calmloki posted is really good for this. I can scroll around to the blue areas and find things that are fun to buy and use - new TV, camera, car, or clothes, furniture, "decorations". When I hit the brown I see stuff that I buy because I have to, not because it's fun to use - hospitals, gasoline, fuel oil, "water and sewer", tuition, milk, eggs.

We know, if money gets tight, we could postpone much of the blue is stuff for a long time. But we feel that we have to keep buying the brown. The problem with prices isn't merely that they are going up a little faster than wages, it's that the things that are going up are the items that we feel we can't postpone and don't enjoy buying. I expect that someone who re-weighted the BLS prices by simply deleting all the "enjoyable and postponable" things would get higher inflation that would reflect our feelings a lot better. (sorry, but I'd drop entertainment and vacations, too)

[RonBoyd]

Here is an interesting take on this... if you follow all the links.

Calculated Risk: Tim Duy: Misunderstanding the CPI

Including this:

http://www.bls.gov/opub/mlr/2006/05/art2full.pdf