The macroeconomic forecasting business has become quite unhinged in the current situation, with existing models seeming to have their wheels coming off as old relationships simply do not hold and reported data seems unreliable and going in all sorts of directions. We have already seen this happen regarding forecasts that were made for the May employment numbers, with most forecasters projecting employment declines that would have been more than 10%, some of them by a lot more than that, although none more than 20%. But in the end employment was estimated to have grown by over 2%, a situation of the forecasters simply being wildly wrong.
As it is, with the month of June now over and thus the second quarter over, it looks increasingly to me like most of the forecasters have not learned their lessons from that May employment fiasco. I suspect that in many organizations they find it hard to revise their models, especially on short notice, even when it is clear their models are not working. We see a lot of the forecasters making predictions of a large second quarter decline in GDP, but more numbers have come out for May, and most of them have been positive, some of them very positive, and if June continues to be positive, even if at a lesser rate than May given renewed shutdowns occurring due to the uptick in Covid-19 infections as June proceeded, this may further make some of these strongly negative forecasts even further off.
So what are some of these forecasts and what do the latest reported numbers look like? First, we must note the first quarter outcome. It seems that GDP declined by -4.8% or 5.0% for the first quarter at an annualized rate. All of the decline occurred in March, more than offsetting modest growth in both January and February.
But the forecasts for annualized rates of decline for the second quarter are awesome, at least most of them. I am getting these from a post by Menzie Chinn over on Econbrowser and are from less than a week ago on June 26. Here are some:
GDPNow -39.5%
New York Fed -16.3%
St. Louis Fed -38.16%
IHS Markit -35.3%.Clearly, the only one not showing a massive decline is the New York Fed.
Menzie also showed how the forecasts have evolved for two other forecasters. A blue-chip group’s forecast was for a -25% rate on April 30, but slid to -35% by June 5.
The Atlanta Fed has had its forecast make large movements, starting out at a relatively modest -12% as of April 30, but then plunging to a whopping -54% as of June 5, but then in the face of more recently improving data by June 26 this forecast had moved to not a not quite as whopping -40%.So what does estimated data look like? One estimate I have seen for the month of April had the actual decline of GDP being -11.4%, which translates to about a -42% annualized rate. But we already see here the danger for all of those forecasts listed above except for that of the New York Fed. It is near certain indeed that the economy has been growing in both May and June. If so that annualized rate of -42% looks to be a definite outer bound.
Now if there has been barely any growth in May and June we might still see numbers in the 30s for the annualized rate of decline. But at least for May the numbers do not look like that. We have already seen employment grow at over 2%, which is the actual growth, not the annualized, which is much higher. We now have an estimate for consumption, which grew at over 17% in May. Given that consumption is on the order of 70% of GDP that is pretty much the ballgame right there, barring some sharp turnaround in June. This is especially the case as estimates of construction also seem to show sharp growth in May, a major component of investment, although if inventories fall sharply, that might offset the construction increase. State and local governments were almost surely declining and probably still are, but probably not a massive rates. Trade is especially hard to predict, with indeed net exports appearing to decline in April by somewhere between -7 and -16%. But exports might actually be rising now as much of the world economy appears to be growing again.
The apparently steep decline of GDP in April will be hard to overcome during these past two months. But it is not out of the question that we might actually see a slightly positive figure for the quarter, especially if it turns out that growth in June continued to be as strong as it looks like May was. But even if it flattened out some as I suggested might happen in a recent post, it looks to me that the sharply negative predietions still being held to by so many forecasters simply look to be way off. Even the much less negative New York Fed may prove to have been too negative, even if indeed the quarter outcome is still negative overall for GDP growth, with it probably going to be more than a month before we shall know at all reasonable.
I conclude by noting that even if the second quarter is positive or only mildly negative, growth prospects going forward for the near future look less promising. This is not only because of the recent spiking of Covid-19 cases with associated shutdowns but also because portions of the large fiscal stimulus that has been going on and has probably aided the recent growth will have disappeared or will do so unless Congress acts to keep them going. In particular, the individual stimulus checks have ceased, and the expanded unemployment benefits are scheduled to cease at the end of July. What is more certain is that we are truly profoundly uncertain about what will transpire in the next few months.
Barkley Rosser
They are dancing in the streets about the employment numbers, but forcing people back to work early is directly responsible for the virus’ resurgence. Fucking idiots!
Paul Krugman @paulkrugman
The real economic story: reopening brought a substantial bounce in jobs, but also a surge in Covid-19 cases, which already appears to have stalled recovery. This is probably about as good as it gets. And we still have double-digit unemployment 1/
9:34 AM · Jul 2, 2020
But Trump and his sycophants will read this report as vindication; not only will they fight measures to control the pandemic, but they’ll lose interest in continuing to help workers, state & local govts, etc. 2/
By the time reality sets in in the form of health care crisis and sagging economic numbers, it will be too late to avoid both a huge death toll and a long period of stagnation or worse. If Trump loses — likely but not certain — Biden will face one hell of a mess 3/
Guv said that the biggest source of infections was family gatherings. Might be, but right up there would have to be those who have no choice but to work at jobs with little if any worker protection. Penny-wise and pound foolish.
July 2, 2020
Economy Brings Back 4.8 Million Jobs in June, Unemployment at 11.1 Percent
By DEAN BAKER
Since February, employment rates for Black men and Hispanic men have dropped 9.4 percentage points and 10.3 percentage points, respectively.
The June employment report showed the economy adding back 4.8 million jobs, following a gain of 2.7 million (revised up by 200,000) in May. This two-month gain leaves the economy down by just under 14.7 million jobs from its pre-pandemic level in February.
The unemployment rate fell 2.2 percentage points to 11.1 percent. However, this understates the actual drop since the number of people misclassified as employed rather than unemployed fell from roughly 3.0 percent of the labor force in May to 1.0 percent in June. It is important to remember that these data reflect the state of the labor market in the second week of June before the coronavirus infection rate began to soar again.
The job gains were primarily in the sectors that had been hardest hit in the shutdown. Restaurants added 1,483,000 jobs in June, accounting for more than one-quarter of the total gain. However, employment in the sector is still down by more than 3.1 million from its February level. Retail added 739,800 jobs, with more than a quarter of these (201,600 jobs) being in clothing. Retail employment is now down by a bit less than 1.3 million jobs from February.
[Graph]
Employment in health care jumped by 358,000 in June, with the 190,400 gain in dentists’ offices accounting for more than one-half of the rise. Health care employment is still down by more than 900,000 since February. Manufacturing added 356,000 jobs and construction added 158,000 in June. They are now down 757,000 and 472,000 jobs, respectively, from pre-pandemic levels.
One sector that has not recovered is motion pictures. While it added 2,600 jobs in June, its current employment level of 212,600 is less than half of its February level. Mining also continues to be hard hit, losing another 9,800 jobs in June. The hardest-hit sector within mining is support activities, which are mostly jobs associated with opening new sites. These are down by 71,600, more than 20 percent, since February.
The local government sector added 57,000 jobs in June, while the state sector shed 25,000 jobs. The two sectors together are down 1,487,000 jobs since February. This figure will grow sharply as governments enter their new fiscal year with massive shortfalls unless Congress appropriates a large amount of assistance.
On the household side, the picture is supportive of a strong but very uneven bounce back. First and foremost, it seems that people who reported being on temporary layoffs are getting their jobs back. This number stood at 18,630,000 in April, but is now down to 10,565,000, accounting for 59.5 percent of the unemployed. The number of unemployed who are reentrants also has risen sharply from 1,477,000 in April to 2,356,000 in June. This reflects greater confidence in the labor market.
On the other side, more than 4.6 million people have dropped out of the labor force since February. Just under 1 million of these workers have been prime-age (between 25 and 54) women. Workers over age 65 account for more than 900,000 of those leaving the labor force. More than 1.9 million young workers (under age 25) left the labor market.
While the overall employment-to-population ratio (EPOP) is down by 6.5 percentage points since February, it is down by 8.6 percentage points for Blacks and 9.4 percentage points for Black men. The EPOP is down 9.1 percentage points for Hispanics and 10.3 percentage points for Hispanic men. For Asians, the EPOP is down by 9.9 percentage points.
These differences show up clearly in looking at employment by education level. The EPOP for college grads is down 4.0 percentage points since February, compared to a drop of 7.6 percentage points for those with just a high school degree and 9.4 percentage points for those without a high school degree. This also shows up in the loss of part-time jobs. While involuntary part-time employment has more than doubled, from 4.3 million to 9.1 million, voluntary part-time employment has fallen by more than 20 percent, from 22.2 million to 17.1 million. This reflects the loss of jobs in restaurants and hotels, many of which are part-time.
This is a mostly positive report, indicating that the economy was bouncing back, although it still has very far to go, with pretty much all the numbers looking worse in June than at their very worst point in the Great Recession. However, this bounce back is virtually certain to slow, especially if Congress does not come through with additional funding for the unemployed and for state and local governments. With the pandemic spreading out of control, there will certainly be more shutdowns, and people will be reluctant to use services even in areas without shutdowns. In short, we are very far from a situation that is anything close to normal.
July 2, 2020
Coronavirus
US
Cases ( 2,800,147)
Deaths ( 131,003)
Anne:
These are all recent jumps in cases. I would expect an increase in deaths in the next few weeks.
I heard on the radio that this job increase was twice what was being forecast by the leading forecasters. So my post remains correct, with this more evidence of the forecasters being wildly off.
That said, note near the end I say we are heading into a slowdown due to the respiking of the coronavirus.
The mining support category is largely the oil service industry and is driven by oil drilling.
https://fred.stlouisfed.org/graph/?g=syql
January 15, 2018
Employment in Oil and Gas Extraction, 2014-2018
https://fred.stlouisfed.org/graph/?g=syq0
January 15, 2018
Employment in Oil and Gas Extraction, 2007-2018
Run:
These are all recent jumps in cases. I would expect an increase in deaths in the next few weeks.
[ Worldwide cases are increasing, deaths having fallen for a time are level. American cases regularly have unfortunate extended effects:
https://www.nytimes.com/2020/07/01/health/coronavirus-recovery-survivors.html
July 1, 2020
Here’s What Recovery From Covid-19 Looks Like for Many Survivors
Continuing shortness of breath, muscle weakness, flashbacks, mental fogginess and other symptoms may plague patients for a long time.
By Pam Belluck ]
You do realize this is over in July. It’s reopened as much as it can. Long term job losses surged in June as expected.
Also, the bls stopped counting furloughed workers as unemployed which massaged the data. You can see the divide between this and labor department data. About a 13 million persons divide.
July 2, 2020
Coronavirus
US
Cases ( 2,814,375)
Deaths ( 131,234)
July 2, 2020
Coronavirus
US
Cases ( 2,830,448)
Deaths ( 131,413)
Bert,
I agree that reopenings will peter out, but they are still happening. Virginia just moved to Phase 3 of its reopenings yesterday.
Barkley:
Michigan took a step backwards to regroup. No in house drinking if 70% of your revenue is not the result of food sales. Three hundred and 50-something new cases. Whitmer wants a stranglehold on Covid in MI. Chrysler workers are refusing to go back to work.
July 2, 2020
Coronavirus
US
Cases ( 2,837,189)
Deaths ( 131,485)
India
Cases ( 627,168)
Deaths ( 18,225)
UK
Cases ( 283,757)
Deaths ( 43,995)
Mexico
Cases ( 231,770)
Deaths ( 28,510)
Germany
Cases ( 196,717)
Deaths ( 9,064)
Canada
Cases ( 104,772)
Deaths ( 8,642)
Sweden
Cases ( 70,639)
Deaths ( 5,411)
China
Cases ( 83,537)
Deaths ( 4,634)
Notice that the deaths to confirmed coronavirus cases ratio for the UK is 15.5%, which is astonishingly high to me, but which healthcare analysts in the UK have offered no public explanation for.
As for the spread of infections in the US, that is beyond disheartening or excuse to me.
anne:
New study out by Henry Ford Hospital ACO in Michigan. Take a guess . . .
You really can’t expect economists to make good predictions when the economy is being battered by a novel epidemic and a lunatic political party. Epidemiologists have just started getting a handle on how the virus spreads, its effects, potential treatments and so on. Meanwhile, we have our national leadership in total denial and a major political party that sees the epidemic as a political issue rather than a governance issue.
How could an economist predict, even approximately, the death count or case count necessary for a Republican governor to allow cities to require masks in private venues? I’m rarely impressed with economists ability to predict, but we’d need some kind of oracle here.
Since most Americans have yet to recover from the previous economic crisis and have been slowly bled since the 1980s, I don’t see us recovering in the next decade unless there is a Democratic sweep this fall and the epidemic is still raging in early 2021. Even then, it would be touch and go. I’m reconciled to the New Normal which, even from my privileged vantage point, kind of sucks. (I really hope I am wrong.)
Kalesberg
Is a good prediction good until it appears to be bad? Good prediction is kind of an oxymoron, yes?
Kaleberg:
If it helps, I take solace in actuarial analysis instead of predictive ones. Having well thought out plans for as many contingencies makes me feel much better because I at least know what I am going to be doing when the eventualities play out and all the predictions turn out to be wrong in their own way.
So to Run’s question: a prediction’s goodness is not predicated on its accuracy but whether it engenders a good response to a real possibility should that possibility be actuated.