Scott Sumner  

The CBO's forecast, with and without "austerity"

On Greece... The Jevons Fallacy...

After my previous post, commenters James (in London) and Jeff directed me to the official CBO forecasts of August 2012, with and without austerity:

CBO's Baseline: Taking into account the policy changes listed above and others contained in current law, under CBO's baseline projections:

• The deficit will shrink to an estimated $641 billion in fiscal year 2013 (or 4.0 percent of GDP), almost $500 billion less than the shortfall in 2012.

• Such fiscal tightening will lead to economic conditions in 2013 that will probably be considered a recession, with real GDP declining by 0.5 percent between the fourth quarter of 2012 and the fourth quarter of 2013 and the unemployment rate rising to about 9 percent in the second half of calendar year 2013.

• Because of the large amount of unused resources in the economy and other factors, the rate of inflation (as measured by the personal consumption expenditures, or PCE, price index) will remain low in 2013. In addition, interest rates on Treasury securities are expected to be very low next year.

An Alternative Fiscal Scenario: To illustrate the consequences of possible changes to current law, CBO has produced projections under an alternative fiscal scenario that incorporates the following assumptions: that all expiring tax provisions are extended indefinitely (except the payroll tax reduction in effect in calendar years 2011 and 2012); that the AMT is indexed for inflation after 2011; that Medicare's payment rates for physicians' services are held constant at their current level; and that the automatic spending reductions required by the Budget Control Act, which are set to take effect in January 2013, do not occur (although the law's original caps on discretionary appropriations are assumed to remain in place).

That set of alternative policies would lead to budgetary and economic outcomes that would differ significantly, both in the near term and in later years, from those in CBO's baseline:

• In 2013, the deficit would total $1.0 trillion, almost $400 billion (or 2.5 percent of GDP) more than the deficit projected to occur under current law.
• The economy would be stronger in 2013: Real GDP would grow by 1.7 percent between the fourth quarter of 2012 and the fourth quarter of 2013, and the unemployment rate would be about 8 percent by the end of 2013, CBO projects.

Now in fairness the actual austerity was not quite as bad as it might have been. The deficit fell roughly $400 billion to $680 billion in fiscal 2013, a reduction about 10% less than the predicted amount, which would have produced an estimated $641 billion deficit with the entire fiscal cliff enacted. But that's certainly a lot of austerity compared to keeping the deficit at a trillion dollars.

Obviously the economy did much better than the negative 0.5% RGDP growth and 9% unemployment in late 2013 that were expected by the (Keynesian) CBO. But even more strikingly it did far better than the 1.7% RGDP growth and 8% unemployment at yearend in the alternative scenario.

In fact, RGDP grew 3.12%, and unemployment fell to 6.7%.

Here we have not only the economy reacting to the Great Austerity Experiment better than predicted by the CBO, but even far better than predicted if there were no austerity.

Does this ring a bell? Do you remember the Great Stimulus Experiment of 2009? The time that the unemployment rate didn't just rise much more than expected in response to the stimulus, it rose far more than expected under the alternative scenario of no stimulus!

In that case the Keynesians said something to the effect that the crisis turned out to be far worse than expected. I'm actually not sure that's true; we pretty much knew the size of the financial crisis when the forecasts were made. But I'll cut them some slack; the severity of the GDP decline in late 2008 was not fully understood.

But that excuse hardly applies to 2013, which was a run of the mill recovery year much like all the others of the past 5 years---with no notable shocks.

Except of course for the Fed's stimulus, which the market monetarists predicted would offset the effects of austerity, and did so.

Two grand Keynesian experiments and two abject failures. Followed by two times where the Keynesians started crowing about how they'd been right about everything. You can't make this stuff up.

PS. Some people ask me; "If the Keynesian model is so bad then why do experts like the CBO use that model?" Good question.

Comments and Sharing


COMMENTS (23 to date)
Nick writes:

How about the CBOs forecasts on how much US debt will cost to finance for the next ten years? Interest rates are heading back up to where?
Personally, I think the official government expert projection that stands out the most at the moment is the long run neutral fed funds rate--one of the few numbers even more stultifying to the public than a CBO forecast. Now that the QE taper is over, will we start tapering that laugher down to a more reasonable level?

Kevin Erdmann writes:

Absolutely brilliant.

Note, too, the interest rate projection. Not only did interest rates not go down - they shot up in 2013. So, demand for Treasuries declined because of austerity, and supply of Treasuries increased because of QE3. Loose money is supposed to push down yields, they say. And rates went up!

joemac1985 writes:


Check out Mark Zandi's forecasts as well. See, what he forecasted.

Ray Lopez writes:

Sorry but Dr. Sumner I think is wrong. Here is what I found from the World Bank website:

1/ go to USA

2/ See the comment by the World Bank for 2013: "“2013 2.2% [growth rate]… ‘Slow growth thanks to sequestration. Low nominal GDP growth thanks to low inflation.’"

3/ Note that today, Friday, Jan. 31, 2015, the Commerce Dept. just announced the GDP for 2014, as (screen scape from today's Google News): "“The Commerce Department reported Friday [1/31/15] that US gross domestic product grew at an annual 2.4 percent pace last year, up from 2.2 percent in 2013″

4/ Conclusion: aside from all the other points Dr. Sumner makes, which I have not reviewed, I think it's clear to say that 2013 in the USA, with a 2.2 % growth rate in GDP, had a smaller growth rate than the 2014 GDP growth rate of 2.4% (which may or may not be statistically significant due to noise, but that's another matter). From this one could reasonably conclude that the 2013 GDP growth was lower than possible due to sequestration, which is exactly the point the Open Letter Keynesians were making in late 2012, and what the World Bank says (see above). Granted, this is not the only reason 2014 might be greater than 2013, but it's a reasonable interpretation. Note Sumner makes various points involving real GDP vs nominal GDP, but with low inflation the two are very close to one another and for all intents pretty much the same.

Steve Y. writes:

Related: Greg Mankiw points to a research paper that concludes most of 2014's job growth was due to cessation of unemployment benefits.

Less government spending produced more job growth. Who would have believed it?

ThomasH writes:

Getting somewhere.

Next we need to know the forecasts of the other, non-fiscal policy changes, including monetary policy that went into the CBO model to see if the model is bad or the forecasts are bad.

Ditto the MM model whose results were better.

Maybe the models are equally good but market monetarists are better at forecasting what the Fed will do. :)

Scott Sumner writes:

Nick, The markets expect rates to stay low, so I do as well.

Thanks Rick.

Kevin, That's a good example.

Ray, You are using the wrong data, I've discussed this issue in many recent posts.

The CBO uses Q4 to Q4, and so do I.

Steve, Yes, I did a post on that a few days ago.

Thomas, I'm not good at forecasting what the Fed will do, and I rely on market forecasts for the economy.

Andrew_FL writes:
PS. Some people ask me; "If the Keynesian model is so bad then why do experts like the CBO use that model?" Good question.

Maybe because it's functioning just as Al Ullman intended it to.

Daniel Kuehn writes:
"Two grand Keynesian experiments and two abject failures."

Weren't you suggesting just yesterday that these weren't great experiments?

Daniel Kuehn writes:

I wouldn't keep seeking out clarity on this if it weren't for the fact that you seem to want to have your cake and eat it too.

If you want to bash Krugman and Konczal bravado, great - it deserves a little bashing.

But you're also treating this like a real failure of Keynesianism - and if you don't think this is a good test and if you think Krugman and Konczal are being brash then you really can't get any mileage out of this in weighing the merits of fiscal policy.

liberty writes:

"Except of course for the Fed's stimulus, which the market monetarists predicted would offset the effects of austerity, and did so."

Where is the evidence that monetary policy provided any kind of offset, or the evidence that austerity even needed an offset??

Otherwise, great post!

Andrew_FL writes:

@Daniel Kuehn- Grand in scale =/= Great in quality.

Scott Sumner writes:

Daniel, I meant grand like the Grand Canyon. Really big. And again, fiscal stimulus was not a significant part of the new Keynesian toolbox as recently as 2007, so they really needed some big wins. And they got two big losses. Perhaps not statistically significant, but a big blow.

liberty, The only evidence I have is that things turned out about as I expected, and much differently from what Keynesians expected. This is why we need good, subsidized NGDP prediction markets, so we can settle these debates once and for all.

Of course we also have a tiny bit of evidence from asset market reactions to Fed initiatives, but not enough to establish how big the effect is.

Andrew, Ah, we think alike.

Daniel Kuehn writes:

Scott -

"And they got two big losses."

So it is a valid test?

I am genuinely confused about what you think of the quality of the test at this point.

Bob Murphy writes:


I love this running set of posts, and I'm glad you made the mirror-image point about the Obama stimulus sparing us from an economy worse than we realized.

However, there's a little hitch in your own offset narrative I think. If the Fed had had an NGDP (or even P) level targeting regime in place, I would totally see your point.

But they didn't, they instead had committed to a fixed monthly amount of asset purchases.

So if the CBO, Krugman, Scott Sumner, and the markets already knew exactly how much stimulus the Fed would be providing, and based on that the Keynesians in the group said "austerity will cause a deviation from baseline of such-and-such," how can you see they failed to anticipate that the Fed would fill the gap through offset? Didn't the Fed provide just as much stimulus as everyone expected?

(Sorry if my timeline is off.)

David R. Henderson writes:

@Bob Murphy,
I don’t understand your point.

Kevin Erdmann writes:

Scott, the interest rate projection I referred to is in the excerpt you posted. I wasn't sure if you caught that.

AndrewFL, that's what's so fascinating about EMH. It is only as good as human cognition. And, in complex systems, especially political ones, especially ones with counterintuitive outcomes, group cognitive inertia is very strong.

Doctors bled the ill for some time. Large portions of the world refuse to educate women. Most people consider it perfectly normal to cut off part of their baby's genitals. We've just lived through just about the worst 15 year period we've seen for equity holders and banks, and the popular view arising from this is that the Fed just manages the money supply for the sake of equity holders and bankers.

The problem is that for every Copernicus, there are a thousand witch doctors and alchemists. For every Sumner, there are a thousand blogs going on about the Rothschilds and gold, or the Koch's, or some other bogeyman that serves their identity - and they are usually much more confident than Scott. The question really isn't how could they be wrong, but how do we trust ourselves to pick the right corrective.

Andrew_FL writes:

@Kevin Erdmann-I'm confused what your response has to do with what I said?

Or which comment of mine you were responding to.

I don't think what you said is wrong I just found it to be kind of a non sequitur?

@Bob Murphy-If you change "level targeting" to "growth rate targeting," you could make a case that the Fed has been targeting a 4% growth rate for a few years now. Not explicitly, but de facto. Really though, that alone should be enough that a change in the trajectory of G should not alter the trajectory of NGDP. Less G would just mean more C and I.

Really I would prefer if "monetary offset" were called something else because I think it's a really unfortunate name. It's basically just describing the conditions necessary for "crowding out" to happen (and the inverse, for that matter)-that is, for changes in private spending and government spending to move in opposite directions.

Scott Sumner writes:

Daniel, What do you mean by "valid test?" Do you mean definitive test? There are none in economics. There were two big experiments. Both failed badly. When this sort of experiment goes well for Keynesians, they cite it to support their model. But you'll have to define "valid test." One can perform a specific test, which might be valid, but also argue that due to random shocks the test does not definitively establish a theoretical point. Thus flipping a coin 100 times is a "valid test" of whether it is a fair coin, but may or many not be definitive, depending on the context.

Here's another way of making my point. If we had had a recession in 2013, that would have substantially weakened the argument for monetary offset. But would it have definitively shown there is never monetary offset? No. I do not believe the 2013 experiment definitively disproves Keynesian economics, but it somewhat weakened the already incredibly weak argument for fiscal stimulus. That's far from nothing.

Oddly, I'd guess Krugman had something similar in mind when he spoke of a "test" in April 2013---important but not definitive.

Bob, Timeline is wrong. The Fed did QE3 and forward guidance partly to offset fiscal austerity.

Andrew, I agree that monetary offset is an unfortunate term as it implies the Fed must "do something" to offset fiscal austerity. But that's not necessarily the case. If they keep targeting inflation expectations at 2%, it will offset automatically.

Kevin Erdmann writes:


I was reacting to the PS sentence from your first comment. Not only does the CBO use it (and I liked your comment), but it seems to me that the Keynesian framing is generally accepted among portfolio managers. In fact, even though bond market prices eventually contradicted the CBO expectation, my guess is that you wouldn't figure out how that happened by reading the tactical documentation of bond portfolio managers. I think they were almost universally managing their portfolios based on the same expectations the CBO had. Markets can still be efficient even when everyone is consciously wrong, because second, third, etc. order effects end up influencing the price, even when we universally are too wrong about prices to consciously push them there. Housing in the 2000s is a good example. Everyone "knew" at the time that it was a bubble. (Even I did at the time.) But, those same people pushed home prices up through decisions we were making about asset allocation. Most of the time relative efficiency is achieved, but only through a very complex set of interactions among many different markets. Efficiency is highly likely, especially in something like bond markets that are so liquid and so interconnected with so many other markets. But, efficiency isn't guaranteed. And, in the unusual instance where some limits to arbitrage are placed just right to prevent efficiency, markets could become very inefficient in contexts where efficiency was happening in spite of widespread conscious errors. But, the whole system is frequently too complex to understand what is blocking efficiency, so a speculator can't really track down the practical cause of a particular inefficient price. And, if you do take a winning position its just like this issue with 2013 GDP. Maybe you found an inefficiency, and maybe you just got lucky.

Anyway, I'm blathering on. But, the issue of how seemingly everyone could be so wrong was what got me thinking.

Andrew_FL writes:

@Kevin Erdmann-That part was a quote. The part that I actually said was just implying that when Al Ullman sponsored the bill to create the CBO in the 70's, he intended it to provide essentially Keynesian analysis to justify government policy. And it still works that way to this day.

It's interesting to hear that Portfolio managers would work off the assumption the CBO's forecast is correct. It seems to me that people take "authoritative" sources of information being correct for granted. It's a lot easier than trying to figure out the truth for yourself and presumably the expected costs of bad information are tolerable compared to the expected costs of acquiring more accurate information.

Brian Donohue writes:

Great post, Scott. ZLB, fiscal tightening, strong QE3, growth. Easy-to-understand. Textbook material.

Comments for this entry have been closed
Return to top