From budget projections released by the Congressional Budget Office (CBO) last week:
CBO now expects that output will fall slightly short of its potential, on average, even after the economy has largely recovered from the recent economic downturn.
We’ve thrown in the towel on our long-time assumption that the economy never again falls into recession.
Shocker: the business cycle lives!
We grossed up the CBO’s projections partly with an estimate of the average effects of automatic stabilizers. Lo and behold, Appendix E of the new report links budget projections to the average effects of automatic stabilizers.
What’s more, the CBO slashed its figures for potential output. The two changes together added over $1 trillion to projected debt in 2024.
We’ll be presumptuous and pretend that someone at the CBO read our research before implementing the new approach. Our blog may not be the main reason (or even part of the reason) for the change, but please don’t burst our bubble. Let’s say we’re 99% responsible for the CBO’s discovery of the business cycle.
We’ll even give our mole in the Ford House a name: “Jefferson Smith,” after the Jimmy Stewart character in “Mr. Smith Goes to Washington.” As you’ll guess if you’ve seen the movie, we’re assuming Mr. Smith is earnest, hard-working and wants to do well by the taxpayer.
Even more importantly, he wants us to acknowledge the tentative steps toward reality in last week’s report.
So, geeky bloggers full of unsolicited advice, you should be happy now, right?
Well, actually no.
Sadly, the extra $1+ trillion in debt is dwarfed by adjustments that still need to be made.
To show why the CBO remains way too optimistic, we’ll start with the new unemployment rate projections:
The CBO’s long-term assumption of 5.5% is where the new recession “allowance” comes into play. If not for the nod to the effects of automatic stabilizers, we’re told that the long-term rate would be about 0.25% lower.
Questions for Mr. Smith:
Why such a tiny change? Consider that the new long-term rate is still below the “Great Moderation” average from 1984 to 2007. Didn’t we learn that the economy’s performance over that period was illusory?
Next, we compare unemployment rate projections to the CBO’s assumptions for 3 month Treasury bills:
Call us cynical, but it’s hard to imagine unemployment falling continuously through an interest rate jump to 3.7% in 2018. There are good reasons to expect a poor economy after interest rates rise, and especially after a trend that persists for several years. But don’t just take our word for it. To gauge the effects of large, sustained rate changes, we compared the change in interest rates for every 12 quarter period since 1945 to the change in the unemployment rate over the following year:
Here’s an interpretation of these results, lifted from “M.C. Escher and the Impossibility of the Establishment Economic View” (where we shared a similar analysis with the same conclusions):
While the results speak for themselves, I’d be remiss if I didn’t add qualifiers. For one, the sample sizes fall as you move from left to right across the charts. [See technical notes post.] Moreover, history doesn’t always foretell the future; this time could be different.
But the thing is: the data makes perfect sense. Higher interest rates have obvious effects on risk taking and debt service costs. It stands to reason that the economy won’t just sail through the large rate hikes needed to restore historic norms.
If anything, the charts likely understate the future effects of rising rates, because today’s debt levels are far higher than average historic levels. Any normalization must also include a wind-down of unconventional measures such as quantitative easing, which presents additional challenges.
More questions for Mr. Smith:
Considering the data in Chart 3, why does the CBO expect the unemployment rate to fall in 2019 after an assumed 3.4% interest rate jump over the prior three years? Why is it expected to fall in 2018 and 2020 after interest rate increases of 2.9% and 2.2% in the preceding three year periods?
Before you answer, though, understand our perspective on the rote explanation that projections beyond 2017 are based on “trends in the factors that underlie potential output” rather than “forecasts of cyclical movements in the economy.”
The CBO might as well write: “We only forecast ‘good’ cyclical movements – hence our long-standing prediction for a robust recovery – not ‘bad’ cyclical movements.”
The aggressive recovery assumption for the next four years just doesn’t jive with a meager recession allowance for the remainder of the projection. As shown in Charts 2 and 3, the worms crawl out when you look at the effects of interest rate changes, while Chart 1 shows that the long-term unemployment rate of 5.5% is too low.
Another question for Mr. Smith:
Why not balance the assumed recovery with the payback that invariably occurs after the economy heats up and interest rates rise?
Here’s how that might look:
Here are the key assumptions behind the recession scenario:
- From 2014 to 2017, it’s exactly the same as the CBO projection. (Hence, optimistic.)
- The unemployment rate then jumps by 0.8% in 2018, 1.1% in 2019 and 0.8% in 2020, based on CBO interest rate projections and the history shown in Chart 3.
- By 2024, the unemployment rate falls back to the post-1970 average of 6.4%. (For discussion of reasons to exclude the ‘50s and ‘60s from the average, see “7 Fallacies About the Lengths of Things.” Or, just poke around economics chatter for a Beveridge curve chart as well as the increasingly mainstream view that jobs market fundamentals are nothing like they were fifty years ago.)
Back to Mr. Smith:
Remember those “find the next number in this progression” problems from grade school? Well, let’s say you’re looking at the last 60 years of unemployment rates and using the same “find the next number” approach to guess the next 10 (as in Chart 4 above). Isn’t the chart’s green line a better answer than the red line?
In our last chart below, we show that the CBO’s budget outlook depends quite a lot on Mr. Smith’s answers to our questions. More precisely, the unemployment rate assumption has a huge effect on budget deficits:
Notice that we split the red line in two to show the effects of the CBO’s recession allowance, with the dashed red line representing the old, “recessions don’t exist” approach. The recession allowance has little effect on its own. Although the CBO attributes over $1 trillion of extra deficits in this year’s report to changes in economic assumptions, other changes had greater impact.
The bigger issue is what might happen in a genuine recession. As shown by the green line, we’re projecting the recession scenario to add $2.2 trillion to total deficits through 2024. (Again, see our technical notes post for details.) Needless to say, it’s a budget killer.
We could go on to show effects on public debt, but we’ll save those for our next update of “The Chart That Every Taxpayer Deserves To See,” which will show that other ways of making CBO projections more realistic have even larger effects than the recession scenario.
In the meantime, maybe the CBO would consider a staff screening of “Mr. Smith Goes to Washington,” and in particular, the Lincoln Memorial scene?
Here’s the setting: Stewart’s Mr. Smith takes a seat on the floor near Lincoln’s statue, dejected and determined to abandon his stand against Washington’s dishonest ways. The beautiful Jean Arthur, playing Smith’s assistant, lurks behind the columns. She finally walks over to him, sits down, and delivers the classic, Hollywood kick-in-the-butt speech.
She convinces him that he can’t quit, because he has “plain, decent, everyday, common rightness, and this country could use some of that… so could the whole cock-eyed world…”
“It’s a forty foot dive into a tub of water,” she says, “but I think you can do it.”
Could it be that the CBO’s Mr. Smith has the same “common rightness”? That he’ll make the “forty foot dive” to battle Washington’s resistance to inconvenient truths?
Those are our final questions, and we look forward to answers in future CBO reports.