Following some links on health care last night, I found an interesting libertarian, I think, blog. One of the authors had a post up on private vs. public health care costs. Here follows the written part of the post. Please see the chart above.
“…since 2000, so the administration has stated we absolutely must increase the government’s role in providing coverage to our citizens. Simply study this comparison of the effective cost controls of Medicare vs. private health insurance, and I’m sure that you’ll be forced to agree (from Table 13):”
Having made the error of not looking at the raw data closely enough, or forgetting where I saw the raw data and posting only a graph or chart, I went to look at the numbers. You can find the Table 13 the author cites here. Looking at the graph, it appears that government spending on health care has gone up faster than private spending. First, let me say, that it is totally inappropriate to compare such disparate groups. There are way too many variables. When one does studies, it is important to try to compare groups that are as similar as possible. That aside, let us look at the raw numbers.
Average Annual Change by Period –Medicare– Private Insurance
1970-2007– ————————- 9.2———-10.4
1970-1993——– ——————- 11.0———12.8
1993-1997——– ——————– 7.3———-4.4
1997-1999– ————————– -0.3———-5.8
1992-2002- ————————— 6.4———-9.7
2002-2007—————————– 8.4 ———-6.9
Looking at the graph it looks as though medicare spending is going up faster than private. Looking at the raw numbers, that is not is what we see. Notice in our first line that Medicare spending over the last 27 years is 9.2% vs. private of 10.4%. Ok, the author has decided to only graph those numbers from 1990 on, which is his prerogative, but having this data available, he should have indicated why he chose this period. If one looks at the raw numbers in Table 13, one sees that between 1990 and 2007 Medicare increased by a factor of 3.06 while private insurance increased by a factor of 3.15. What gives? Looking more closely at the graph, it becomes more clear that Medicare is starting at a much higher baseline. This fools the eye into thinking that Medicare is increasing faster.
A few other items of note. As I have been worrying over the recent 26% increase in insurance costs proposed by our carrier, I realized I missed an obvious point. Insurance companies, like banks, make much of their money by investing their premiums. Like many of the rest of us, insurance companies have taken a hit in their portfolios. This suggests that as the data comes in for 2009, we should see private premiums go up even more than they have been. Lastly, as I was thinking about this (presumably) young man’s post, it occurred to me that one of the easiest ways to confirm his idea, would be to find out what a private carrier would want to insure someone over 64. I intend to ask our carrier this question on Monday. I hope to report back the results here soon.
Steve, I added the chart to our Media Library, in case you want to “Add an Image” next to the Upload/Insert prompt in Edit Post.
Ok, the author has decided to only graph those numbers from 1990 on, which is his prerogative, but having this data available, he should have indicated why he chose this period.
I think I explained it quite clearly – in fact you quoted it:
“since 2000″
I included the data since 1990 just to show the past trends.
Why 2000? Because that’s the period of cost increase trends that have been cited by the administration.
…this (presumably) young man’s post,
Not so (unfortunately) much.
one of the easiest ways to confirm his idea, would be to find out what a private carrier would want to insure someone over 64.
You would also have to compare the relative qualities of care, and to be truly relevant, you’d have to compare the two approaches with similar demographics of coverage.
No, you did not explain why you chose 2000, you simply stated “since 2000″. I have your entire quote above. Since you are here, why not include all of the data. Why did you choose to graph from 1990 and not also add in the raw numbers (with the calculations)? God knows I make posting boo boos, so I hope that is all this is, but it does not look so good.
“I included the data since 1990 just to show the past trends.
Why 2000? Because that’s the period of cost increase trends that have been cited by the administration.”
I am afraid I missed that. I will look for their quotes unless you have them handy. Odd that they would only cite 2000 and on. It is interesting that when you calculate 2000 to 2007 Medicare goes up by a factor of 1.76 and Private insurance by 1.71, supporting your case. However, if you go back even one year to 1999 you then get Medicare increasing by a factor of 1.80 and private insurance at 1.84. You are clearly a numbers guy, and probably much brighter than I am, but this still does not make much sense to just choose this particular 7 years and then graph from 1990 on. Maybe add in the hard data when you can in the future?
“You would also have to compare the relative qualities of care, and to be truly relevant, you’d have to compare the two approaches with similar demographics of coverage.”
Very true. As a practitioner, I essentially never see denial of services for Medicare. I believe Medicare is a generally broader coverage program, but that is just a belief. I will see if I can get our insurance broker to give me equivalent quotes. He may also know offhand if there is any service that private insurance covers that Medicare does not. It is a somewhat difficult comparison, as I am sure you know, since private coverage varies so much. I will also try to find out what expenses would be when you add in deductibles. The column heading simply states “private health insurance”. How many people do not have deductibles anymore?
Thanks for stopping by, BTW. Nifty blog you guys have. It is now on my bookmarks bar.
Scott- Thanks. I am such a techno-tard.
Steve
No, you did not explain why you chose 2000, you simply stated “since 2000″.
“since 2000″ is a continuation of the headline. In full, the quote should read:
“Health care costs have skyrocketed since 2000, so the administration…”
Odd that they would only cite 2000 and on.
The acceleration in Medicare cost increases from 2000 has been used to create a sense of crisis.
It is interesting that when you calculate 2000 to 2007 Medicare goes up by a factor of 1.76 and Private insurance by 1.71, supporting your case. However, if you go back even one year to 1999 you then get Medicare increasing by a factor of 1.80 and private insurance at 1.84.
Those are trivial differences – the point is that the administration has laid the blame for increasing costs of insurance at private industry’s feet, when the government program has not performed any better in controlling costs per beneficiary.
I included data from 1990 to 2000 because I like to give at least a decade run-up to the period of time in question, so that people can see the behavior of the curves in some context. I didn’t go farther back because the older Medicare and private health insurance programs start looking less and less like the current programs, and because I didn’t want to pull too much attention away from the period I was talking about – 2000 to 2007.
Ok, fair enough. I think your summation here is correct, for the most part. It would be fair to say, IMO, that private insurance and Medicare have both been poor at controlling costs. My caveat being my point about how insurance companies make their money. As president of my small corporation, I am dealing with very large increases in health insurance rates. I would expect others to be reporting the same soon, if they have not already. This is speculation on my part, but informed speculation. Will be interesting to see the real data come out. Depending on how bad it is, it is possible private insurance will look much worse. OTOH, at the site linked below, public insurance costs seem to rise during recessions. (Chart page 2)
http://www.pwc.com/Extweb/home.nsf/docid/5D9A1874EA16285C8525751C0067D98D
Thanks for coming by our little blog to explain. Your willingness to explain yourself is commendable. I still disagree a bit with your methods, e.g. I think there are too many variables to really compare the two groups in a completely meaningful way. But then, my buddies claim I am too fussy about studies. I also now try to add in some of the raw data when time allows, as graphs can be distorting. Alas, I fall short of my own standards, so I cannot throw stones. Keep up the good work. I lurk at a lot of blogs and yours is one of a very few that actually bother to use data rather than all opinion all the time.
Steve
We’ve had a substantial increase in our premiums too, at our organization of about 100 people. No serious medical issues since last year.
There are basic flaws with “Table 13″ linked and cited in the original blog and this one.
“Per capita” is a meaningless calculation, because it is a straight-line arithmetic calculation that has no correlation to the actual populations covered under each category.
A simple(istic) illustration:
I spend $100 on candy for 100 people. 75% of the people are adults, 25% are children. According to the method used in Table 13, I spent $1 per person, but what you don’t know is that the children consumed 50% of the candy. Sorta makes your “per capita” a ridiculous number to look at, let alone use as a basis for comparison.
So, the first question is this: in public and private spending, what was the actual population on which the money was spent in each category?
This flaw, with respect to all of you promoting this discussion — and it is a very important discussion — is egregious. Another simple illustration:
If I spend $100 on 50 fire fighters, and I spend $100 on 100 widget makers, my per capita in each is $2 and $1 respectively. If the following year I spend $150 for each group, my per capita are $3 and $1.50 respectively. If, however, I “combine” my numbers, they look very different: $200 for 150 is $1.33 per capita, $300 is $2 per capita.
Compare that to this set of numbers for 1960 and 1970, and ask the question at the end:
Billions of dollars, millions of people
Year — Public — Private — Total — Population
1960 — 6.8 — 20.7 — 27.5 — 186
1970 — 28.1 — 46.8 — 74.9 — 210
Percentage increases in each number
10 years — 313% — 126% — 172% — 13%
For how many people was each sum of money spent? Until you (general, specific) can answer that question, you have said absolutely nothing about health care expenditures in the US. And unless you track those covered population numbers in every cohort in every measurement period, you have wasted your time and offered nothing of value for analysis.
““Per capita” is a meaningless calculation, because it is a straight-line arithmetic calculation that has no correlation to the actual populations covered under each category.”
Lost me here Franklin. Is not per capita a pretty simple calculation. Medicare in particular has pretty good data. Assuming they know how many enrollees they have, and what their expenditures are, why can’t they figure out per capita costs? (Disclaimer. Beer was consumed with July 4th dinner tonight)
Wordadvocate-Thanks. 2009 data should be interesting to see if I am correct.
Steve
I’ve only had one beer tonight, Steve. The per capita numbers in the table are total expenditures for the entire population of the US. No breakdowns on how many people covered in public or private plans. Just a straight “I have this many” and “I spent this much.”
Look at it from a covered group POV instead. If two companies have 100 employees each, and they go to two different providers, and company A pays $1,000 per employee per month (your per capita) but company B pays $1,500, have you learned anything? Do you know how much was paid back to employees in claims? Do you know how much each group spends for co-pays? Do you know how much each group spent in out-of-pocket expenses, and for what services?
Table 13 can answer no questions that could provide a valid comparison. Not one.
Ok, up and sober so one last try. The heading at the top says per enrollee. The Table 7 heading notes per capita expenditures, but 13 says per enrollee. Here is the link so you can look at it faster. (I should have put it in my original post, duh.)
http://www.cms.hhs.gov/NationalHealthExpendData/downloads/tables.pdf
Steve
My broker had a baby (our labor floor) Tuesday and I am going out of town tomorrow for a week or so. It may take a while to report on the private insurance costs for someone in the Medicare age bracket. What initially struck me as a clever idea may be more difficult to sort out. I am pretty sure he will quote me some sky high quote, but that will be for an individual policy. Probably the best I can do would be to ask for an individual quote for a person with the same level of health, but at a younger age and then compare. Actually, I wonder if insurance companies even keep actuarial tables for people of Medicare age anymore?
Steve
Thanks, Steve. I will examine the full panoply of tables to see if they answer the questions I pose, and whether they confirm or contradict my complaint.
This is not good science, people. First, I can’t tell if this is a typical ideological (on both sides) device of looking at some data and making it tell the story you’d like. It’s a lot better to formulate a hypothesis (like, for the following reasons I’d expect Medicare per enrollee spending to be less than commercial spending per enrollee, all other things being equal) and then find some data which shows that. I don’t think that’s what’s happening here.
Second, this per enrollee phrase is not clear–does it mean per subscriber (could be more than one person) or per member?
Third, Table 13 shows two comparisons–one overall, and one holding benefits comparable. Is this a chart of the former or the latter?
Back in 2003, a peer-reviewed paper showed that Medicare had controlled costs better than the private sector. (http://content.healthaffairs.org/cgi/content/full/22/2/230) It also said “To compare spending growth between these two payers, we present four measurement principles that should be implemented when drawing such comparisons, and we apply them to the National Health Accounts data files.” Have those principles been followed here?
Then the authors say ” We attribute Medicare’s ability to equal—and using our measures, actually exceed—the private sector in controlling the rate of health spending growth to Medicare’s ability to price aggressively for the services it covers.” Still true, and is at the heart of the argument that a public plan would be less costly than a private plan. So far, I haven’t seen any hint of how a public plan (or any private plans) intend to change the culture of medicine to be like Green Bay rather than Miami.
Finally, all that other data in the National Health Expenditures report needs to be properly analyzed to get some idea of the drivers of the increases. Otherwise, you can’t make valid comparisons. Trying to say anything meaningful at high levels like “Medicare” or “commercial insurance” is useless–first prove that you have two apples of similar species; then tell me similarities and differences between them. I don’t see that happening here–so I think we can safely ignore any conclusions anyone might want you to draw from this graph. The debate about how to control costs will have to have something far more substantial for its evidential kitbag.