David R. Henderson  

This is Misleading CNN--and CBO

France's NIRA did just as poor... I'm becoming increasingly worr...

On the way home from work this evening, I heard an incorrect statement of the latest CBO report's findings on the Republican-proposed health care bill. (I'll have more to say about the bill in a future post; I'm still digesting the opinions of others who seem to have looked at it more closely than I have.)

The CNN news report stated that the CBO estimated that the bill would save $337 billion over the 2017-2026 period. A quick check of the CBO report shows that the CBO was as misleading as CNN. Here's the CBO:

CBO and JCT [Joint Committee on Taxation] estimate that enacting the legislation would reduce federal deficits by $337 billion over the 2017-2026 period. That total consists of $323 billion in on-budget savings and $13 billion in off-budget savings. Outlays would be reduced by $1.2 trillion over the period, and revenues would be reduced by $0.9 trillion.

Notice that the CBO, like CNN, counts $337 billion, not only as a cumulative deficit reduction, which it is, but also as a "savings."

But then the CBO goes on to say:

The largest savings would come from reductions in outlays for Medicaid and from the elimination of the Affordable Care Act's (ACA's) subsidies for nongroup health insurance. The largest costs would come from repealing many of the changes the ACA made to the Internal Revenue Code--including an increase in the Hospital Insurance payroll tax rate for high-income taxpayers, a surtax on those taxpayers' net investment income, and annual fees imposed on health insurers--and from the establishment of a new tax credit for health insurance.

So the largest "costs" come from a reduction in taxes. Admittedly, $361 billion of this tax cut is a new tax credit. Tax credits are, in many ways, like subsidies. So the actual tax cut is not as big as it looks.

Let's assume the CBO's numbers are correct. Then the saving is not $337 billion, but, rather, $1.2 trillion. (I'm accepting, which I really shouldn't, the bad CBO habit of treating a dollar in 2017 the same as a dollar in 2026. For this to be true, nominal interest rates would have to be zero. They aren't. They're low, but they're not that low.)

So what's true is that the proposal would save $1.2 trillion, that $0.9 trillion of this saving would go to a tax cut, and that $0.3 trillion of this saving would go to reduce the cumulative deficit. A $0.9 trillion tax cut is not a cost, although, as noted above, one could count $361 billion of this tax cut as a cost because a tax credit is like a subsidy.

Comments and Sharing

CATEGORIES: Fiscal Policy , Taxation

COMMENTS (15 to date)
Jerry Brown writes:

You talk a lot about 'savings' here, and I am guessing you are using it under the same definition as the CBO is using it? In any case, it doesn't seem that the CBO is using the term 'savings' in the same way economists use the term in National Accounting. It gets confusing.

But anyways, if we can't nail down what it means, can we use the term 'savings' as in 'savings of lives'? I imagine that any honest Obamacare opponent would still admit that of the millions who gained coverage, through either the Medicaid expansion or through the exchanges, there were at least some hundreds of people who received care they wouldn't have received and that such care could have saved their life.

So how does that use of the word savings tie into the way the CBO uses the word or the way economists in general use it?

Kevin Dick writes:

@Jerry Brown. In fact, the findings of the RAND and Oregon Health Insurance Experiments call into question your assumption that increased access to insurance actually leads to lives saved. So it's hard to definitively credit the ACA with any actuarial benefits from the value of statistical lives it may have saved.

AlanG writes:

@Kevin Dick - have you read the New England Journal Of Medicine article on the Oregon experiment? It was rather limited in terms of the medical outcomes studied and only focused on some short term metrics. It did find "... increased access to and utilization of health care, substantial improvements in mental health, and reductions in financial strain..." The metrics measured where they did not see a difference between the Medicaid-Insurance group and control group were "...measured blood-pressure, cholesterol, or glycated hemoglobin levels..."

We have no idea whether there would be any long term benefits within the Medicaid community. One only need to look at the long term Framingham Heart Study to show the length of time that is required to gauge long term risk/benefit issues. Your statement "...it's hard to definitively credit the ACA with any actuarial benefits from the value of statistical lives it may have saved..." while true is not what the Oregon study could possibly demonstrate.

Of course if the AHCA is passed we may never know the answer to that. One other thing that was in Obama care that will also be lost is the requirement for the collection of quality data. I fail to understand how conservatives or libertarians continue to object to this type of spending. If the goal is to make reliable individual decisions, one needs both price and quality information. This exists in almost any other area of commerce but not healthcare.

liberty writes:

"I'm accepting, which I really shouldn't, the bad CBO habit of treating a dollar in 2017 the same as a dollar in 2026"

-- is that right? I don't have time to check right now, but I thought that they did use CPI to inflate dollars out over the years. They certainly do for budget and economic outlook estimates.

Kevin Dick writes:

@AlanG. Yes, I've read it.

But picking at the lack of detail and power of a study whose results are in line with a previous study with better detail and power (RHIE) violates standard probabilistic reasoning. And it's not just the RHIE, there's a fair amount of evidence that more health care in general on the margin does not improve long term outcomes. See Robin Hanson.

The fact is that we have to make decisions with the data we have now and that data points to there being little evidence increased access to health care should be credited with actuarial benefits of saving lives.

Which is what the OP was arguing for.

Thaomas writes:

Thanks. I was surprised that the massive changes reduced benefits by only 337 billions. Both the reductions in taxes on high-income people and reductions in benefits to poor people are bad things and should not be offset.

Jerry Brown writes:

Kevin Dick, you argue we have to make decisions with the data we have now. That is apparently reasonable but not necessarily true. Even if the data is collected fairly, it is being collected under an existing set of conditions for a limited time period.

Imagine collecting data on the arm movements of a man in a strait-jacket for some time period. The data would show very limited arm movements. It might lead policy-makers to the conclusion that the man cant move his arms. That data would be completely useless if the strait-jacket was removed.

If medical care has any life saving ability whatsoever, and everything we know about it says it does, wouldn't increasing access to it by the millions result in at least several hundreds of life savings? Honest critics have to address that in any analysis of savings. In my opinion.

John Eleazarian writes:

In addition to the misleading quote on savings, the news sources generally mislead when they present the CBO's estimates as persons "losing coverage." The report deals with fewer people purchasing coverage because the law repeals the tax penalty for doing without. That isn't "losing" coverage, it's choosing not to purchase it.

Sad to say, I have no expectation that the media will clarify or correct either misleading item.

Thomas Sewell writes:


If medical care has any life saving ability whatsoever, and everything we know about it says it does, wouldn't increasing access to it by the millions result in at least several hundreds of life savings?

The logical flaw comes when you conflate health insurance spending with "access to medical care". The ACA increased some of the first, but very little (and may have even reduced through crowding out and regulatory effects) the second.

How many new Doctors were created by the ACA vs. how many were driven out of practice?

AlanG writes:

@Thomas Sewell - you trust a survey where the return rate based on the total # of MDs in this country was 2.5%??? I don't.

@Jerry Brown and Kevin Dick - getting and aggregating health care data is very difficult. We have very little inter-operability and a real lack of standards in medical records. When I was still working at PhRMA I staffed an effort to develop and Observational Medical Outcomes Partnership that would develop new ways of studying large databases to better detect rare adverse drug events and would eventually be useful in studying drug benefits as well. I can tell you first hand that this was a very difficult project. We spent almost two years in getting all the specifications set up (working with research experts within the pharma industry, academia and the FDA). I was able to secure the first $20M to get the pilot up and running and recruited the executive director. The project sponsored competitions in algorithm development, set up guidance for inter-operabiity and a lot of other stuff. The bottom line is that one cannot rely on short term data to predict long term outcomes other than the acute injury suffered from certain adverse drug reactions.

I'm always skeptical when I read reports such as the Oregon Medicaid paper and some of the others as well as the time periods are too short. As I noted in my earlier post it took a lot of years to confirm the Framingham Heart Study findings. We still rely on some European data sets in the drug safety area as they have national medical care systems and electronic data records that go back in some cases decades. Much of the vaccine safety information came out of these data sets.

This stuff is hard work and until we get a handle on how to better use the information identifying quality of care will continue to be difficult.

Jerry Brown writes:

Thomas Sewell, that is an interesting point. I had not considered that the ACA may have restricted supply.

In general, I would have thought the ACA constituted an upward shift in demand for doctor services. And in general, I imagine that increases in demand for a product tend to result in either increased production or increased prices or both. I will have to give your paper a more thorough reading. I would note that the survey was published in Sept. 2014 and ACA didn't really go into effect until 2013, so the time period size of the data sample is still questionable.

Jerry Brown writes:

Here is a very short post with an estimate of 'life' savings between the two plans. Apparently, the Social Security Administration expects to pay out less under the new proposal because of the increased number of deaths. That's a 'savings' I would never have thought of...


pyroseed13 writes:


This argument that the Oregon Medicaid Experiment is unreliable because "it did not measure long term outcomes" is faulty. Why would you need to wait say 10 years to confirm that "access to health insurance" leads to reductions in blood pressure or cholesterol? Doctors who provide medication and lifestyle advice for these issues generally don't say, "Come see me in 10 years and we will see how everything worked out." Two years is plenty of time to determine if their were any meaningful reductions in those quality-of-life outcomes. And in general, the study confirms what many critics of ACA noted: Lifestyle choices are more important for your own health than insurance, and there is not much the government can do to make us eat better, exercise, or quit smoking.

Thomas Sewell writes:


I trust it better than the source you didn't cite... please, rather than just reflexively criticizing what your priors disagree with, provide additional data from a different source.

The idea that additional requirements, regulations and changes caused at least some Doctors to decide to retire early shouldn't be very controversial when we know it caused other Doctors to stop accepting insurance at all and go to an all cash practice.

The obvious issue being that the supply of new Doctors in the U.S. is severely limited, with a long pipeline effect. If 10x as many people this year decided they wanted to become Doctors as last year, with limited medical school spots and the years it takes to become a Doctor, we'd barely see a tiny increase a decade from now. There are some substitutes (Existing Doctors deciding to work longer hours, etc...), but it's not exactly a free market in health care, now is it?

David S writes:

I agree with your "savings" terminology, but would take it further:

The proposed bill saves $1.2 Trillion, and refunds $0.9 Trillion of that savings to the taxpayers.

(I also agree with the above posters that any health care plan that does not increase the number of medical practitioners cannot have a large benefit to society, only distortion costs - but that is orthogonal to your point)

Comments for this entry have been closed
Return to top