The New ACA Score And The Perils Of Letting Cost Estimates Drive Policymaking

by Stuart Butler

Health Affairs Blog, July 24, 2012

If you were expecting the Congressional Budget Office (CBO) recalculation of the Affordable Care Act (ACA) to rival the drama of the Supreme Court decision, you will have been disappointed.  But the new CBO re-estimate underscores the dangers of basing major policy changes on such forecasts.

While the Court caused political shockwaves by declaring the ACA’s “penalty” to be a “tax”, that in itself did not have major implications for the new score.  CBO projects a net reduction in total federal outlays of $84 billion over 10 years (the new total is $1,168 billion) and an increase in the uninsured of about 3 million, when compared with its earlier estimates.  Significant numbers, but not earth-shattering.  The Joint Committee on Taxation provided revenue estimates for the CBO forecast.

But CBO faced enormous challenges in estimating the impact of the ruling, and its seemingly precise estimate clouds the significant guesses it had to make.

The Challenge Facing CBO

Rather than the tax-not-penalty decision, the part of the Court’s ruling with the greatest budgetary implications concerned Medicaid.  It also gave CBO its biggest headache.  The Court struck down the provision requiring states to expand Medicaid coverage to households up to 133 percent of the poverty level.  So CBO had to figure how many states – indeed which states, given different Medicaid coverage pattern today in the states – would decline short-term federal funding for the expansion (what the Wall Street Journal called the “teaser rate”), knowing they would have to pay a portion of the cost in future years.  The governors of several states, including Texas, Florida and South Carolina, have already declared they would refuse the money.

So what could CBO do?  Take such declarations at face value?  Factor in a bit of politics, consult a political crystal ball, and predict what might actually happen in state houses next year? Give a range to show, say, the different effects of all states declining or all agreeing to expand Medicaid?

To compound the scoring challenge, CBO had to project what would happen to newly eligible Medicaid households in states that decline federal money to expand Medicaid.  Some, but not all, would be eligible for subsidies in the new ACA exchanges, depending on household income.  But how many would sign up (adding to subsidy costs, even though they would not have added to federal Medicaid costs)?

According to former CBO Director Douglas Holtz-Eakin, depending on your guess the potential range of budgetary effects would be large.  If the six states currently saying they will not expand Medicaid follow through with that threat, Holz-Eakin estimates the net budget impact would be $22-80 billion between 2014 (when the expansion occurs) and 2021.  But if all states decide the expansion is an offer they must refuse because of the long-term state cost, the impact could be as high $627 billion.

A False Sense Of Certainty

Faced with the uncertainty about state action, CBO declined to give a range (a bad thing).  Instead it avoided second-guessing states (arguably a good thing for people who wear green eyeshades rather than appear on political talk shows) and projected a number based on “the middle of the distribution of possible outcomes.” That’s defensible, in the same sense as a weather forecaster splitting the difference on the chances of thunderstorms next Friday, but it results in a seemingly precise number that is almost certainly wrong and yet gives a false sense of certainty to policymakers.

This case of CBO scoring highlights once again why it is so important for policy to be driven a bit less by budget estimators generating apparently precise numbers.  That leads to policies based on guesses needed for a computer algorithm rather than on judgment and public discourse about the nature and purpose of government.  But alas, Congress insists that CBO come up with one number for its deliberations, not a range or a zone of probability.  And so even though CBO wisely warns that its number “should not be viewed as representing a single definitive interpretation of how the ACA should or will be implemented in light of the Court’s decision,” regrettably it will be.

Posted response by Don McCanne:

The policies inherent in the Affordable Care Act do create uncertainty as to how many will remain uninsured. Nevertheless, the CBO projection of 30 million uninsured will be the number quoted in policy discussions, since it is the best we have, and it does provide a rough estimate of how short ACA falls in the goal of covering everyone.

To provide a perspective of the enormity of this policy failure, let’s assume that we chose a different set of policies that would cover the same total number of individuals, still leaving 30 million uninsured. Let’s say that we want to have 100 percent coverage in each state, beginning with the largest states since they would have the greatest health care burden in terms of the sheer numbers of patients. Let’s continue to cover 100 percent of each state until we run out of funds to pay for coverage for the last 30 million people – the residents of the least populated states.

Under such a policy, everyone would be covered except all of the residents of Alaska, Arkansas, Delaware, District of Columbia, Hawaii, Idaho, Kansas, Maine, Mississippi, Montana, Nebraska, Nevada, New Hampshire, New Mexico, North Dakota, Rhode Island, South Dakota, Vermont, West Virginia, and Wyoming.

Who could ever support such a ridiculous policy scheme? Yet since we have buried other ridiculous policies into a highly dysfunctional, fragmented, administratively wasteful ACA financing scheme, we are going to accept those policies as the best we can do? It’s okay to have 30 million uninsured as long as we don’t concentrate them in two-fifths of our states?

Stuart Butler implies that we need policies based on “judgment and public discourse about the nature and purpose of government.” If we were all sincere in such a discourse, and laid partisanship aside, we would end up with a decision to enact a single payer national health program that included everyone.


By Don McCanne, MD

Stuart Butler has a valid point that the variables in the Affordable Care Act and its implementation are great enough that we cannot accurately predict the precise numbers of individuals who will remain uninsured. The information that the Congressional Budget Office relied on to come up with 30 million uninsured does give us a very rough estimate of the total, and enough confidence that we can say with certainty that at least tens of millions will remain uninsured.

So what is Stuart Butler’s point? He indicates that instead of using this precise number to establish policy, we should choose policies based on “judgment and public discourse about the nature and purpose of government.” Well, let’s use the less precise estimate of tens of millions uninsured. What then should be the role of government?

It is clear that the private sector has been incapable of insuring everyone, and that government involvement is required if universal coverage and access are to be our goals. It is also clear that the Affordable Care Act is not an adequate response on the part of government since that is what led to this estimate of the large numbers who will remain uninsured.

So what would Butler propose? In 1989, we thought we knew the answer when he coauthored the report from the Heritage Foundation which called for the principle that, “Every resident of the U.S. must, by law, be enrolled in an adequate health care plan to cover major health care costs,” and “The requirement to obtain basic insurance would have to be enforced,” and “If the family did not enroll… a fine might be imposed.” Yes, an individual mandate.

This did prove to be partisan since the concept was used in proposed Republican legislation designed to counter the health care reform proposal of Bill and Hillary Clinton. The Clinton effort fizzled, and the Republican proposals were never enacted. Little did we realize then that a couple decades later the partisan support for the individual mandate would flip from the Republicans to the Democrats.

In a USA TODAY op-ed this year, Butler indicated that his views had changed based on newer policy research, and that today the individual mandate “means the government makes people buy comprehensive benefits for their own good, rather than our original emphasis on protecting society from the heavy medical costs of free riders.” Sort of, you take care of your own needs but don’t turn to me for help when you can’t meet them – the perpetual conflict between individual responsibility and social solidarity.

He wrote, “health research and advances in economic analysis have convinced people like me that an insurance mandate isn’t needed to achieve stable, near-universal coverage. For example, the new field of behavioral economics taught me that default auto-enrollment in employer or nonemployer insurance plans can lead many people to buy coverage without a requirement.”

We know that default auto-enrollment does result in higher rates of participation, but we also know that an entirely voluntary program for purchasing our often unaffordable private health plans would leave far more uninsured than would the policies contained in the Affordable Care Act.

We should have that public discourse that Butler calls for – the one about what we, as a nation, want. Do we want absolutely everyone to be included in our health care system? Or do we want to continue with policies that provide the equivalent of covering everyone in the most populous states, but none of the residents of Alaska, Arkansas, Delaware, District of Columbia, Hawaii, Idaho, Kansas, Maine, Mississippi, Montana, Nebraska, Nevada, New Hampshire, New Mexico, North Dakota, Rhode Island, South Dakota, Vermont, West Virginia, and Wyoming?

Heritage Foundation: “A National Health System for America” 1/2/89

USA TODAY: “Don’t blame Heritage for ObamaCare mandate” by Stuart Butler

Re-posted with permission from