The CBO-Reliable?

Great Debates
1

Since the CBO has been in service (1981 I think), How often have predicted results come to fruition as stated (for the purposes of this discussion, close is good enough)?

The CBO was used as a tool by Bush to proclaim a “surplus” in order to get tax cuts.

The CBO was, proclaimed by Dems to get a health care bill passed, to be revenue neutral.

In it’s past history have many of their predictions been predicated on true numbers vs projections that never come or are they always based on the most positive projection instead of a realistic one?

2

The CBO itself analyzed its accuracy at forecasting economic conditions in a 2009 report. You can find it here: http://www.cbo.gov/doc.cfm?index=10484

As far as scoring bills goes, they score them using the guidelines provided by Congress. Their deficit projections have also been analyzed, and are generally pretty accurate after adjusting for policy changes (for example, the Clinton surpluses were wiped out by the Bush tax cuts and increased spending after 9/11).

As to your last question, the CBO has been using “fan chart” showing a range of possible outcomes based on varying economic performance. I don’t know when that chart came into vogue. You can see one of them at this 2002 report page: http://www.cbo.gov/doc.cfm?index=3277&type=0&sequence=6 That link also includes an analysis of past reports’ accuracy.

3

The CBO and GAO are admired for their professionalism. The quality of the OMB and CEA varies.

Congressional Budget Office
General Accountability Office
Office of Management and Budget
Council of Economic Advisers

4

It’s probably worth noting that most of their predictions will go untested. Whenever someone proposes something, they release some predictions of what the likely results would be. Based on those predictions, the proposal might be dropped or changed, resulting in the need for new predictions. Similarly, any one prediction might give X different predictions, based on different assumptions. Most of these will also go untested, and sometimes even all of them. For example, if I predict that the next year will be the most profitable year for the US Government in all recorded history, but instead an asteroid hits the planet and kills all life before the year has even begun, then is my prediction actually wrong? Minus the asteroid, the prediction may well have been correct. Beyond the question of the quality of the predictions, there’s the question of how many scenarios and assumptions were described. In general, they may choose illustrative examples rather than real world ones so that legislators see the range of effects. None of those simulations will ever be enacted in the real world.

5

The one problem with the CBO is that it must accept any propositions put to it for the scoring system. That is, if one wanted to make another healthcare bill, and assume that after ten years suddenly we’ll have money falling from the sky and aliens bringing the Universal Enlightment and naked ladies dancing in the streets, the CBO can’t argue. It must score the bill based on those projections.

OK, OK. Congress uses slightly more plausible propositions than that. But what they expect is often improbably, like claiming that for the purposes of Bill “X” we’re going to save X billions of dollars… 95% of which will come 10 years down the line. It’s an old trick adapted to a newer situation: instead of generous gifts to the public now and pain for the next office-holder, you promise generous gifts now and improbably bad choices to the next office holder.

6

I’m pretty sure this is incorrect.

The CBO does not score something based on, “pretend a zillion dollars magically appears”.

It is my understanding that the CBO must use the laws as they currently stand when scoring something. I suppose they could answer the question of what happens if a zillion dollars magically appears but the actual scoring assumes current laws/policy that are in effect when the scoring happens.

Otherwise you could get into endless “what if” scenarios.

7

The CBO does score things based on projected population/economic growth rates and things like that provided by Congress, though. I think that’s what sb is referring to.

8

What if Congress tells them, "we promised a bunch of cuts (to whatever) and we will tax people at x so use those numbers. When in fact, the cuts don’t come and Congress fails to get whatever new tax they wanted.

Has this/does this happen currently?

I guess that is what it comes down to for me as to how reliable they are. Yes, if we assume this then we get that.
But if it’s all just accounting games, why are they taken seriously by either side of the aisle (when the other side uses them)?

Whack A Mole: I think that Bush did indeed use “predictions” of the Clinton surplus to get his tax cuts passed, when in fact the Clinton surplus never actually surfaced as predicted.

9

Yes, Kearson, that does indeed happen. In fact, for quite a time (and perhaps even now), the CBO projected using two different models - the current-law model and the “alternative baseline”.

One is based entirely on the current law (Bush tax cuts expire, AMT isn’t fixed, Docfix isn’t repealed) and the other is based on a set of “expected” outcomes which basically amounted to things staying they way the currently are (Bush tax cuts extended indefinitely, AMT being fixed like it always is, Docfix being done as usual). These are all clearly explained in the documents produced by the CBO.

It is an accounting “game” only to the extent that all projections are a game unless you have the ability to see the future (in which case, who needs projections…).

As far as projections go, however, given the assumptions they make (and disclose), they do a pretty damn good job.

10

In my opinion, the CBO is one of the best, least-partisan organizations in Washington.

But you do have to understand their limitations. They are number crunchers, not policy analysts. They are directed by Congress, and congress has a lot of power to skew the reports by setting up the assumptions and constraints the CBO must operate under.

For example, the CBO scored the health care bill with the assumption that various cuts to Medicare and other programs would happen, because that’s what the health care bill said would happen. In addition, they were directed to score the health care bill on a ten year timeline, and the bill was structured to front-load taxes to make the ten year score look better. But that doesn’t matter to the CBO - it has to score the actual bill as written, regardless of whether Congress has a history of violating or rewriting bills later. It is not allowed to editorialize or second-guess Congress’s actual performance - and that’s as it should be.

The result is that politicians on both sides of the aisle have managed to use the CBO as a political tool by crafting bills specifically designed to get a good CBO score, knowing full well that the features of the bil that allowed it to score well will be later removed in reconciliation or by amendment.

The bottom line is that you have to be careful to look at the bill itself and make your own determination as to the likelihood of the assumptions in it coming true. In practice, Democrats are good at doing that when Republicans get CBO scores they like, and Republicans are good at doing that for Democrat bills, but when it’s their own bill being scored and they get a score that makes their case, suddenly the CBO becomes the final arbiter whose judgment and numbers are solid fact that cannot be questioned in any way.

Sometimes the CBO will refuse to score a bill if they feel it does not contain enough detail. President Obama submitted a budget for CBO scoring, the CBO threw it back and said it couldn’t be done because it didn’t provide enough detail.

You can also get some insight into what the CBO really thinks by reading editorial comments by the Director, or by reading transcripts of hearings where the Director is questioned about a CBO report. For example, the CBO scored the health care bill as being deficit neutral, but the director later editorialized that the assumption of savings by cutting Medicare and by finding waste, fraud, and abuse was unlikely to occur, so in the real world the Affordable Care Act would likely not be deficit neutral.

11

Thank you all. This was exactly the information I was looking for.

12

Granted Congress may undo provisions at a later date that initially made a bill palatable.

Still, when they undo those provisions wouldn’t that get scored too?

So, if the CBO said the health care law would cost “X” over ten years and two years later when Congress tries to undo some promised funding (e.g. raise a tax) then won’t the CBO then say, under this new model, the cost is “Y” and accounts for the change in revenue?

13

I don’t think the CBO will re-run the score unless requested by Congress.

14

Maybe not but I am saying when the new bill is up which undoes some changes promised in the previous bill then the CBO will score the new bill which will account for congress backing off on something they previously committed to.

15

Only if a request is made for a rescoring. CBO assumes that all bills happen in a vaccuum. So when the request is made they’ll score what that request is and nothing more. Without a specific request to ‘score this bill and rescore this other bill from 2008’ it won’t be done.

16

So, say Bill-X was scored in 2009 and passed. It is now the law.

Bill-Y is now under consideration. Bill-Y, in part, removes some tax that was in Bill-Y.

They will not rescore Bill-X but Bill-Y’s score will include the change to the revenue which was previously promised in Bill-X.

In this fashion Congress cannot simply craft a bill any way they like to get a good score while secretly plotting to remove certain pieces in the future. I mean, they may be plotting to remove stuff in the future but when that future bill gets scored it will account for the changes.

17

What Congress can do is play “kick the can.” The budget rules only consider the budget effects of legislation for so many years. It was ten years in the House last time I looked, and “CBO score” referred to the 10 year impact on the budget. So a bill introduced today that gives free candy for everyone at a cost of $10 trillion through FY 2021 still gets a zero score if it includes $10 trillion in annual tax increases that start in 2021. (I’m simplifying here - they take things like interest and inflation and the effects of the increases and free candy on the economy into account). Pass that, then next year introduce a bill that extends the free candy and delays the taxes to 2022, and CBO will score that bill at zero too. (Or, more likely, wait until 2020 and extend the candy and delay the taxes until 2030). A dangerous game, and easy to see through, but they do try variations on that. This is not CBO’s fault - like others have said they work within the parameters their bosses set for them.

18

Exactly, which is why on a reliability scale, I’d say they were pretty damned worthless. Well, let me re-phrase: If it is being used as a tool to promote or decry something that Congress wants passed/repealed then it’s worthless.
Like all other tools, it’s the person using the tool that needs to be the responsible one.
Can anybody in Congress use the CBO or are there parameters to which offices (majority/minority) get to request it and if so, why are studies not brought up of past bills which had fatal flaws about the ‘Congress’ whom told them not to use certain aspects?

19

The CBO prepares a Budget and Economic Outlook twice a year for congress which will account for such shenanigans.

Congress may ignore them but the CBO will still put forth its best guess about the economy.

The CBO also, per a standing request from the Senate, provides an “Analysis of the President’s Budgetary Proposals”.

They use their own assumptions and not the administration’s assumptions for their analysis.

20

What total nonsense. If a bill increases spending within the 10 year window by $10 billion, and increases revenue outside the budget window by $10 billion, then it scores as a $10 billion increase to spending. If the next year there is a change to revenues that would affect the 10 year baseline estimate, the change would be scored.

In any case, even if spending were frontloaded in a bill, and revenues backloaded, that would be perfectly apparent in the CBO report. They include year-by-year estimates on the change to the budget baseline.