While I Wasn't Paying Attention, First-Quarter Activity Came Down to Earth
…But it looks like a deferral into a stronger economy for the rest of the year
A week ago was another lost weekend of all work and no play and no NO-BAH-DI-NOZ newsletter. My day job has been too all-consuming. So I was surprised today when I looked around at the economic indicators and saw this:
This is the Federal Reserve Bank's "nowcast" that I've written about frequently since the winter. The mechanical exercise of taking inputs from recent news and crunching them into the output of a 1st-quarter GDP growth figure had yielded a number ranging from 9.5% to 10.0% to 8.3% for about four weeks before taking a tumble in mid-March.
Clearly, the big hit came on Mar 16, the day after I posted my last newsletter. Too busy to pay attention, I had not seen any particular news that week or in the subsequent week to tell me the first-quarter GDP outlook was softening. So today, I looked up the reasons and present them to you now.
One of that Tuesday's reports was the monthly retail sales numbers from the Commerce Department. Instead of the expected +0.4% increase—expected, that is, by a regular survey of economists—retail sales dropped -3%, a huge miss by the estimators. However, the previous reading for January was revised upward to a very strong +7.6% increase. The net effect of the revision up and the new data down seems to be a net negative for the sake of the GDPNow model. My evidence is the "PCE" column of this table of details from the Atlanta Fed.
PCE = Personal Consumption Expenditures, i.e., the "C" for Consumption part of the C+I+G+NX equation for GDP that maybe some people learned in school. It's about 2/3 or 70% of U.S. GDP. The next four columns are different parts of Investment spending (business spending), the "I" part of the equation. G = Government spending, which gets a column. And NX = Net Exports = Exports – Imports. As you see, Atlanta's model breaks them all down each day they make a revision. The big revision of Mar 16 came pretty heavily in all but one of the major categories, all downward. The exception is Government spending, which is obviously and unrevisedly up.
Retail sales is very strongly tied to PCExpenditures, so you can see in that column why I deduce that the Mar 16 revision was due in part to the new retail sales info being a net negative. Now look at the large 3/16 drop in Equipment. That takes us to the next data release of that date: Industrial Production. These numbers come monthly from the central office of the Federal Reserve. The Marketwatch news service tells about them under the headline "Severe winter weather chills U.S. industrial output in February"
Production down 2.2% last month, first decline after four straight gains
The numbers: Industrial production fell 2.2% in February, the Federal Reserve reported Tuesday. The gain was well below Wall Street expectations of a 0.5% gain, according to a survey by the Wall Street Journal.
The bulk of the decline was due to severe winter storms last month, the Fed said. Texas was hit with subfreezing temperatures that overwhelmed the state’s electricity grid and led to 57 deaths.
What happened: Manufacturing output fell 3.1% in February after a 1.2% gain in the prior month. The output of motor vehicles and parts slumped 8.3%, the fourth decline in the past six months. Mining output dropped 5.4% after a 2.1% gain in January. Utility output rose 7.4% on higher demand for heating.
Big picture: The severe weather managed to do what COVID-19 couldn’t – slow the manufacturing sector. Economists think that the factory sector remains on firm footing. All eyes are on how the service sector recovers from the pandemic.
That weather-related explanation probably applies to the retail sales story, too. If so, then the setback in cold February will be reversed and compensated for in subsequent months—the spending and production will just be deferred: no net loss. The weaker-looking first quarter will become a stronger-looking second quarter with no major effect on the bigger story of the whole year.
About services, we've been hearing every night on the news that airline traffic as measured very precisely by Transportation Security Administration (TSA) screenings have recovered to almost 100% of pre-COVID levels. [Usually, the network news reports this in a tsk-tsk, naughty-naughty tone, as they've mostly long since crossed the line from reporting what happens to telling us what we should think about it.] Maybe next weekend I'll dive deeper into a broad array of high-frequency indicators that might shed a lot of light on exactly where we are now. Having missed 3 of the previous 4 weekends due to my day job, I'll try next weekend to be particularly illuminating.
For now, I'll just add some bits of anecdotal information. In my little corner of western Massachusetts, I see from what little I go out of the house that conditions are strong. The downtown of the university-hub town of Amherst retains almost none of the strongly creepy post-apocalyptic ghost-town vibe that prevailed one year ago, all summer, and I'd say all through 2020 at least.
Driving around on Saturday 27 March, the town seemed alive, people were lounging around in the 'Common' or central park, a new playground was under construction in a different park, a conspicuous restaurant location was being rebuilt inside for its next tenant, pedestrians were strongly in evidence all around, and there were enough students circulating on the flagship university campus that I actually had to stop at a crosswalk as I passed through the heart of the campus on North Pleasant Street. That definitely had not happened any other time in the past 54 weeks. I picked up a local newspaper and saw that the police blotter page looked as completely full as ever with citations for loud music and similar disruptive undergrad behavior.
Out on the more suburban-style commercial strip of Amherst/Hadley, economic activity has been at completely full strength, to my eyes, since sometime last summer or fall. The lines at the Starbucks drive-through are insane (but still no one allowed to sit inside – same as at McDonald's & Dunkin [but not Wendy's, which does allow seating], [and I still puzzle that McDonald's and Wendy's restrooms are open, but Dunkin's never are]). People are getting their oil changed, circulating in and out of every kind of strip-mall location, and are very active at Home Depot.
Even JCPenney, which had been all but deserted almost every time I was ever there in 2016-2019, this Saturday had dozens of cars in the lot on Saturday. I also got my drivers license renewed last week, and apart from the guards blocking my entrance until I could prove I had business there, that place was in full operation, with people going out on road tests (but only in frequently sanitized RMV-owned vehicles).
I hear that other parts of the country have been re-opening more publicly and earlier than we have. Here, people are very compliant: I haven’t seen another live person’s face for a year, except for the one I live with and some (some) of the people I see walking alone outside. To people elsewhere, many of the anecdotes I reported here might seem “so 2020,” but for this place, what I’ve reported from this weekend is pretty new, and refreshing.
Speaking of a broad array of high-frequency indicators, consider this Weekly Economic Index (WEI) of 10 such indicators compiled by the Federal Reserve Bank of New York:
The following paragraph I composed two weeks ago and pulled it from that newsletter in a late edit, due to being ridiculously over my target word count:
One last note about the WEI is that right now is when it will cease being useful for us to look at it, if I understand correctly how it is constructed. WEI is meant to show a comparison to year-ago levels. A year ago all these numbers fell of the COVID cliff, so I think starting this next week, the blue WEI line will start showing big gains from those depressed levels. On that graph it will look will look like it shot up since last week, but in fact it will only reflect that 52 weeks ago 'shot down' from 53 weeks ago. Actually, my understanding is that the NY Fed does all this as a 13-week moving average, such that we'll see only 1/13 of the COVID-cliff effect in any one new week's point on the blue line. In that event, it will take a few more weeks of March into April before we see 3/13 or 5/13 of last year's abrupt cliff showing up in the form of a big rebound this year, by comparison. If I'm right about that, the WEI graph will soon become meaningless to us for reasons that go way beyond the unemployment fraud issues.
Indeed, the NY Fed adds a new second sentence to their usual very brief Commentary that accompanies the chart:
The increase in the WEI for the week of March 20 reflects increases in electricity output, tax withholding, and fuel sales (relative to the same time last year) and a decrease in initial unemployment insurance claims, which more than offset a decline in rail traffic. Because the WEI measures changes over a 52-week period, the strong increase also reflects the sharp deterioration in economic conditions during the same time last year.
…thus confirming my warning that "it will cease being useful for us to look at." There is already nothing to learn from the fact that the little Commentary above cited four of 10 indicators as having risen significantly and only one as having fallen: measured from the cliff of 52+ weeks ago, it means nothing. Don't expect any more screenshots of WEI until late 2021, perhaps.
I close with just a brief story from Yahoo! News that reminds us where we've been, and how that compares to where it looked like—one year ago—where we were headed.
One year later: The crisis that wasn't, and the crisis that was
Felix Salmon
Thu, March 25, 2021
The economic crash was bad, but it turned out not to be nearly as bad as many economists feared.
Flashback: A year ago this week, economist Nouriel Roubini published a column in which he speculated that we could be entering a "Greater Depression" even harsher than the Great Depression of the 1930s.
We wouldn't have a V-shaped recovery, he wrote: "Rather, it looks like an I: a vertical line representing financial markets and the real economy plummeting."
St. Louis Fed president James Bullard was also highly bearish, predicting that the U.S. unemployment rate would hit 30% in the second quarter.
The reality was much less dire. Unemployment spiked to 14.8% in April and then rapidly fell back; it's now 6.2%, which is where it was in July 2014.
Total employment stands at 143 million, well below pre-pandemic levels but still just back to where we were at the beginning of 2016.
U.S. GDP now looks set to reclaim its pre-pandemic levels either this quarter or next.