How Congress Manipulates CBO Projections

A. Fulin Baldwin

One of the key parts of pretty much every Republican’s campaign platform in the 2016 election was a promise to repeal (and usually replace) the Affordable Care Act (“Obamacare”). As a plan to win the nomination and eventually the Presidency, that makes perfect sense: politically, the ACA has been an abject catastrophe for Democrats that has helped to decimate their ranks of elected officials, even while the plan’s namesake remained quite popular.

Now that the GOP holds the Presidency (in the form of a man who is already a month and a half late on his promise to repeal and replace the ACA on “day 1”) and both houses of Congress, the process of repealing the ACA should be pretty easy. It should also be a political slam dunk: the program is unpopular among moderates, hated by the GOP base, and repealing it would fulfill a major campaign promise.

But they haven’t repealed it yet, much to the consternation of Republican voters. What’s more, they really don’t seem like they are that close to repealing it, and the possible replacements are looking more and more like the existing program.

So, what happened? Well, two things. First, the ACA is actually quite a bit more popular than is often assumed. While the concept polls very poorly and cost a lot of Democrats their jobs over it, there are several provisions that remain extremely popular, most notably coverage for pre-existing conditions and allowing children to remain on their parents’ insurance until age 26. The level of unpopularity stems from things that are necessary to make those popular provisions work: the individual mandate and the overall cost.

The second reason for the delay in repeal is a little bit more mechanical. Stories have begun to circulate that the biggest sticking point in replacing Obamacare is in putting together a program that will get the required Congressional Budget Office score. This, I thought, was worthy of a little bit more discussion on what the CBO is, what it does, and how it can be manipulated.

First, the reason for the importance of the CBO score is The Byrd Rule - a Senate rule that amends the Congressional Budget Act of 1974 to allow Senators, during the Reconciliation Process, to block a piece of legislation if the CBO says that it will increase the federal deficit over a ten-year period. Since reconciliation is the only hope of passing a repeal (Democrats would filibuster any bill they could filibuster), it is imperative that the CBO determine that any repeal be, at worst, “deficit neutral” (someday I will give you 10,000 words on that phrase, and roughly 4,000 will be expletives).

The CBO is a non-partisan, unbiased analytical office that calculates the budget impact of proposed legislation exactly as it is written. This last distinction is very important: the CBO takes the written word literally, even when those written words are obviously misleading.

As an example, we can look to the Alternative Minimum Tax and its impact on Federal Deficit projections. Enacted in 1982, the AMT was intended to block the ultra-wealthy from claiming “too many” deductions and reducing their income tax level below an arbitrary amount deemed to be fair (in practice, it punishes upper-middle class workers in areas with high state and local taxes, expensive real estate and multiple children, but that is a piece for another day). The problem with the AMT is that Congress seemingly forgot to index the tax for inflation, so by the late 1990’s, it was beginning to catch huge numbers of middle-class taxpayers who were never intended to be targets of the tax. Congress got into the habit of passing a “patch” every year to exempt these taxpayers and limit the tax to only those people who it was intended to impact.

The reason they didn’t permanently fix the bill, though, is largely a function of the CBO and the way it works. When the CBO published budget projections during all of those years, they based their projections on the law as written, which means they were assuming large revenues from the AMT that no thinking person really expected the Treasury to collect. By patching the bill every year rather than making an easy permanent fix, Congress insured that the CBO’s deficit projections were artificially low (at least as compared to the real scenario as understood by reasonable observers).

And why was the AMT permanently fixed in 2012? Well, because Congress and the President were looking to extend the Bush tax cuts, and wanted to get the maximum top-line headline effect of the bill, so they included a permanent AMT fix as a means of upping the overall size of the tax cut. When the CBO scored the tax cut, they used as a baseline the 1982 AMT, despite there being no realistic chance that the law would play out as written.

There is another way that the CBO gets used to mislead Americans on the budget impact of legislation, and it has to do with the 10-year time-frame established by The Byrd Rule. Any time a politician proposes a new program, they do everything they can to minimize the 10-year cost (or maximize the 10 year benefit) of the proposal. The easiest way to do this is to phase in the benefits very slowly while establishing new revenues more quickly.

When Obamacare was passed, it included large expenditures in the form of health insurance subsidies, and substantial revenue sources (a supplemental income tax on wealthy taxpayers, new capital gains tax, tax on high-priced employer-provided plans, etc.)

One other thing the CBO can’t do is ask “Hey, is there any reason that we should consider capital gains taxes to be at all related to health care expenditures?” There is no ability for the CBO to opine on the appropriateness of a “funding source,” they just do the math…it therefore doesn’t matter that the ACA just included a straight un-earmarked tax hike, it only matters that the various unrelated provisions of the bill were all totaled up in the final score.

The ACA was passed in 2010, but included a slow implementation schedule that phased benefits in over 3-5 years {all of the numbers coming up come from old CBO reports}. For the years 2010-2013, Net Expenses of the ACA were expected to be under $5 billion. In 2014, that was to jump to about $55 billion, and then $90 billion in 2015 before settling into a norm of near $150 billion annually in 2016 and beyond (when, surprise surprise, the author was leaving office!).

The revenue sources, however (which, again, were in some cases wholly unrelated to health care) were implemented much more quickly – rising steadily from about $10 billion in 2010 to $50 billion in 2014. All of this led to an initial 10-year projection of something like $700 billion in costs (all of which was to be incurred in the years 2015-2019) and almost as much in new revenues and “other cost savings.”

The clear implication, obviously, is that this law costs about $70 billion per year, which seems to be maybe not so outrageous to keep people from LITERALLY DYING IN THE STREEETS, and that cost is offset mostly by new taxes on a couple of rich people.

Then a funny thing happened…the law actually cost somewhat less than originally intended because fewer people signed up and healthcare costs rose less rapidly than the CBO was anticipating (while Candidate Obama told you that your premiums would fall by $2,500, independent analysts were telling a very different story) and by 2015, the cost projections for the years 2015-2019 had gone down from $700 billion to somewhere around $500 billion. Great news right? Well, sure…

Now, however, when the CBO did its ten year projections, the ACA didn’t get the benefit of having four years with no benefits being paid…the projection included ten full years of full implementation. Even with the new, lower costs baseline allowing for a $142 billion cut to the projection for 2015 to 2024, the new 10 year projection was $1.2 Trillion. What was sold as a plan that implied a cost of $70 billion annually had come in under budget and still managed to turn out to be a program that costs nearly twice that much.

{These two facts may seem contradictory…how can the law cost so much, but repealing it also cost a lot? The answer is that the GOP doesn’t want to repeal it; they want to replace it with something that eliminates the new taxes that offset the cost of the law but preserves the subsidies to buy insurance for the poor and coverage for people with pre-existing conditions. They want to continue to deliver many of the benefits without requiring any costs, so they are waiting, I believe, until one of their members comes forward with magic.}

The ACA is far from a unique example, and in fact is a pretty common means by which legislators write new bills. The 2003 Medicare Part D expansion, for example, passed with a proposed 10 year (2004-2013) cost of $396 billion. The clear implication, again, is “Oh, OK…$40 billion per year to give prescription drugs to everyone on Medicare. That’s not so bad! This is a good use of the ‘surplus’!” One obvious problem…the program wasn’t to be fully implemented until 2006, so that ten year window included three years of not actually paying much money out. (A second obvious problem is that they were lying and expected it to be $530 billion even while saying $396 billion.)

So, by 2008, the program was widely touted as having come in under budget…and yet the next ten years were now projected to cost over $800 billion. Only in the world of government accounting can something cost less than expected and still twice as much as you (kinda) promised.

All of this is not to criticize the CBO, which is an extremely useful public institution that remains our best unbiased source for projecting long-term federal budget issues. It is merely a reminder that the CBO is bound by some pretty simple rules, and Congress is neither bound by the same rules, nor unskilled in manipulating those rules. CBO scores and projections can be extremely useful, but we always need to take care to understand exactly what goes into those projections.