Mixed Week of Economic News Leaves December Looking a Little Stronger and Q4 a Little Weaker
... and leads to a more general principle about media distortion of economic perceptions
Key Points:
Atlanta Fed model now shows 4th quarter GDP growing 10.4%, downshifted from 11%
The reason mostly concerns weak news from Nov on personal income and consumer spending
Every measure of the housing market is so hot that it's too good: buyers can't find homes
U of Michigan consumer sentiment improves (also retail sales and factory orders)
… but survey highlights a "Covid Gap" between our experience and what we're told of others' experience
After two weeks that I described last time as "soft," the economy did better in the week ended December 19, according to the Weekly Economic Index (WEI) from the Federal Reserve Bank of New York. The latest reading of the blue line below shows the general level of economic activity was -2.21% lower than it had been one year earlier.
The source offers one paragraph of Commentary to identify the nature of the good news:
The increase in the WEI for the week of December 19 is due to a decrease in initial unemployment insurance claims and rises in fuel sales and electricity output, which more than offset declines in tax withholding and rail traffic.
Reading backward, on Halloween the index was -3.18% below year-earlier, -5.08 around Labor Day, -7.19 on Independence Day, and -9.55 around Memorial Day. So the improvement—with interruptions in July, September, and December—has brought us more than 80% of the way back from the bottom of the hole in April toward the baseline full-sized economy of a year ago.
However, in the most recent week the Fed bank of Atlanta saw signs of weakness that caused their GDPNow nowcast model to produce a slower rate of growth for the fourth quarter. After yielding Q4 growth estimates near 11% since Thanksgiving, the model last week downshifted to 10.4%. Coming so late in the quarter when so much of the econ info is already in the books, the revision was quite large; so we'll look carefully at what specific data the Atlanta Fed's model reacted to last week.
First, on Dec 22 we got news that the 3rd quarter GDP figure, already an eye-popping record coming off the catastrophe of the 2nd-quarter COVID crash, was revised upward. Previously reported at +33.1% average annual rate of growth, we now think the economy actually grew at a +33.4% pace. That's nice, but not really significant for the sake of what we think is happening in the present quarter and beyond.
On the same day, the National Association of Realtors came out with their monthly read on existing home sales. Their press release is detailed and interesting enough to bear quoting:
Existing-home sales fell in November, snapping a five-month streak of month-over-month gains, according to the National Association of Realtors®. All major regions either took a step back or held steady in terms of their respective month-over-month status, but each of the four areas experienced significant year-over-year growth.
Total existing-home sales, completed transactions that include single-family homes, townhomes, condominiums and co-ops, decreased 2.5% from October to a seasonally-adjusted annual rate of 6.69 million in November. However, sales in total rose year-over-year, up 25.8% from a year ago (5.32 million in November 2019).
"Home sales in November took a marginal step back, but sales for all of 2020 are already on pace to surpass last year's levels," said Lawrence Yun, NAR's chief economist. "Given the COVID-19 pandemic, it's amazing that the housing sector is outperforming expectations."
Yun notes that job recoveries have stalled in the past few months, and fast-rising coronavirus cases along with stricter lockdowns have weakened consumer confidence.
"Circumstances are far from being back to the pre-pandemic normal," he said. "However, the latest stimulus package and with the vaccine distribution underway, and a very strong demand for homeownership still prevalent, robust growth is forthcoming for 2021."
The median existing-home price for all housing types in November was $310,800, up 14.6% from November 2019 ($271,300), as prices increased in every region. November's national price increase marks 105 straight months of year-over-year gains.
Total housing inventory at the end of November totaled 1.28 million units, down 9.9% from October and down 22% from one year ago (1.64 million). Unsold inventory sits at an all-time low 2.3-month supply at the current sales pace, down from 2.5 months in October and down from the 3.7-month figure recorded in November 2019.
Properties typically remained on the market for 21 days in November, seasonally even with October and down from 38 days in November 2019. Seventy-three percent of homes sold in November 2020 were on the market for less than a month.
"The positive momentum that home sellers are seeing will carry on well into the new year," Yun predicted, citing low mortgage rates and remote-work flexibilities.
Yun's projections of a continued housing market rebound were the consensus among economic and housing experts during NAR's Real Estate Forecast Summit, held earlier this month. Industry insiders in attendance agreed that mortgage rates will hover around 3% in the coming year, and said they expect an annual median home price increase of 8.0%. …
Distressed sales – foreclosures and short sales – represented less than 1% of sales in November, equal to October's percentage but down from 2% in November 2019.
"While we still face economic and health challenges ahead, I have zero doubt that the nation will continue to recover from this pandemic," said NAR President Charlie Oppler.
The standard interpretation of November's declining new home sales, which I share, hinges on the highlighted information in the middle of the excerpt: "Unsold inventory sits at an all-time low 2.3-month supply at the current sales pace." That is, people bought fewer homes in Nov than in Oct for the simple reason that they couldn't find many to choose from. I've said it before in these newsletters: the housing market is so hot that available supply is not keeping up with desired demand.
One angle the excerpt didn't mention is this: imagine what happens in spring 2021 when significant numbers of people are vaccinated. At that point, sellers will appear in the market, willing for the first time in over a year to let bunches of people traipse through their open homes. It's going to be a great time to be a realtor (and a home inspector, and a furniture seller, and a home remodeler, and a landscaper, and …).
In spite of the top-line number on housing sounding negative, the combined effect of the Dec 22 news was for the Atlanta Fed's GDPNow-cast to be raised one tick to +11.2% projected growth for the 4th quarter.
Then over the next two days came several other data dumps that caused the drop down 8 ticks. One of those was the report on new home sales from the U.S. Census Bureau. Similar to the story for existing homes above, November's seasonally-adjusted sales of newly built homes fell from Oct to Nov, but rose from Nov '19 to Nov '20. The respective percentages were -11% and +20.8%.
Once again, the explanation for the month's dip was, quoting from a housingwire.com article: "the comparative lack of new homes available for purchase, rather than a significant slowdown in demand, was what primarily drove the new home sales decline." Specifically, there are 43% fewer completed homes for sale than there were a year ago.
Another way to tell that the monthly drop of 11% results from fundamental strength in the housing market rather than reflecting a weakness comes from the accompanying statistic about prices. "The median sales price for new houses at the end of November was $335,300, over $7000 higher year-over-year."
It's possible that the double-digit drop in November was worse than analysts' projections, accounting for part of the downward revision to the GDPNow-cast. But clearly the bigger factor was Wednesday's news about consumer spending and personal incomes. These are monthly components of GDP and always follow the GDP numbers by one day. Being monthly, they are much more 'granular' (to resort to a tired, overused adjective) than the quarterly nature of GDP. The numbers for November were poor: consumer spending down 0.4% and personal incomes down 1.1%. The latter is the third drop in four months, and the spending drop is the first since April's catastrophe.
Some excerpts from the standard Associated Press story:
Consumer spending accounts for around 70% of economic activity so even small declines can weigh heavily on economic growth.
[E]conomists say the pullback on spending by Americans last month could be a harbinger of worse days to come as a resurgent coronavirus leads to more shutdowns.
“A rapidly worsening health situation, weakening income, depleted savings for lower income families and cooler weather ed consumers to slam their wallets shut in November,” said Gregory Daco, chief economist at Oxford Economics. “The economy is entering 2021 with very little dynamism.”
I've pointed out before that Mr. Daco has been accentuating the negative (and basically wrong) for a number of months. My interpretation most weeks since this project began in September has been more in line with what I read in Prof. Tim Duy's most recent (Dec 21) blog post: "That’s really not much softness all things considered." Still, my headline from last week is consistent with the Daco view of the most recent observations: December is soft, but …"
Because these downer income and spending numbers feed directly into GDP calculations, the Atlanta Fed's GDPNow model spit back a lower rate of Q4 growth.
On top of the negative incomes/spending news comes this report of consumer sentiment from the University of Michigan:
Wait, what? Is it "deteriorates"/"Slips" or is it "Improves"? Let's consult the text of the university's own press release:
ANN ARBOR—Consumer sentiment improved in the December 2020 survey despite the ongoing surge in covid infections and deaths, according to the University of Michigan Surveys of Consumers.
The improvement was due to Democrats becoming much more positive and Republicans much more negative about long-term prospects for the national economy, said U-M economist Richard Curtin, director of the surveys.
Twice as many Democrats as three months ago expected a continuous expansion over the next five years (54% up from 27%), while that same favorable expectation was nearly cut in half among Republicans (32% down from 60%). Importantly, the near-term outlook for the national economy remained unfavorable due to an expected rise in stay-at-home orders and business shutdowns due to the surging covid virus. …
Consumer Sentiment Index
The Consumer Sentiment Index was 80.7 in December, up from 76.9 in November, but well below last December’s 99.3. The gains reflect improvements in the Expectations Index (74.6, up from last month’s 70.5 but below last year’s 88.9). The Current Conditions Index rose to 90.0 in December, between last month’s 87.0 and last year’s 115.5.
So the latest number for Dec was up since Nov, as the ABA Banking Journal's headline portrayed it. From where did Marketwatch and Morningstar get their downer headlines? It comes from the fact that UMich publishes a preliminary Dec number earlier in the month (called a "flash" estimate), and compared to that level, the final Dec number was not as high. Since the flash estimate is by definition based on incomplete info, it's clearly more meaningful to compare the final monthly stat to the previous month's final stat—as acknowledged in the university's own write-up, above. Again, the improvement was from 76.9 to 80.7 ~ enough to write home about.
As an index number, the value has no particular meaning: it's just defined to be 100 in some arbitrary base year in the past. We gain an interpretation of "80.7" only by comparing it to other values in the time series, such as 115.5 one year ago. Here, for full context, is the graph prepared by Banking Journal, breaking down the "Headline index" into its two components covering the present and the future:
Whether looking at current or future, the recovery from the March-April Catastrophe has been conspicuously partial—much more partial than any other indicator I have tracked these several months: manufacturing activity, retail sales, (un)employment, and as we have seen this week, both housing and GDP.
UMich highlights one interesting finding from its survey, and I want to spend some time and space doing the same. When people are asked about what they've experienced in their own lives and pocketbooks, their answers put us back to the level of 2014-2016, i/o/w 6-8 recovery years after the bottom of the Great Recession. In sharp contrast, when people are asked about their impression—based primarily on media reports—of the general state of the economy (i.e., how they think other people are doing), their answers put us back to the bottom of the G.R. itself.
Notice in the graph that during the G.R. around 2008, there was essentially no divergence between perceptions of one's own personal finances and of the economy at large. Indeed, the UMich press release notes the same thing:
The typical pattern over the past half century has been that consumers’ assessments of their personal finances and the national economy followed similar trends. Currently, assessments of the economy have fallen more than twice as much as consumers’ overall judgments about their current finances.
So, why has the typical pattern broken down this time, such that now people are more than twice as gloomy about other people's or the nation's status as they are about their own? One potential answer could be that the UMich survey systematically over-samples privileged people, and ones who are also especially aware that their own 2014-16-like status is more advantaged than the mass of other people's 2008-like status. But I see no way to support this explanation; if UMich believed their survey were so starkly un-representative, they would probably feel compelled suspend the project, or at least to note the problem in their literature. Instead, they assert that their survey is "based on a nationally representative sample that gives each household in the coterminous U.S. an equal probability of being selected" and stand by the published results without asterisk nor disclaimer.
The graph-makers at UMich see a large and unusual divergence between people's own experience and what they think they know about others' experiences and call it the "Covid Gap" of 2020. I believe the correct explanation for this is the one I already hinted at: People are the world's foremost authorities with perfect information about their own personal finances, while they are relying on second-hand hearsay for their impressions of remote concepts like the national average of the economy. And from where do we get this second-hand hearsay? It must be the case that the news media have been telling a much gloomier story than what is true of people's "lived experiences"(another horribly overused and unnecessary expression: of course any experience is 'lived' and anything lived is an experience).
The graph above might as well be re-labeled to have the solid line tagged as "lived experience" and the dotted line tagged as "media spin." Let's face it: people answering UMich's ‘the economy’ question are not consulting press releases and databases from the Bureau of Labor Statistics, IHS Markit, the Federal Reserve Bank of Atlanta, the National Association of Realtors, and their favorite economist on Substack (where my double digits of subscribers amount to nothing).
In fact I would broaden the principle to say that whenever the news-derived sense of a large abstract concept (like "the economy") is found to be very different from the self-derived sense of a large and representative sample of people about their own expert experience with the same concept, the former must be wrong, and likely due to a deliberate distortion in the dissemination of information.
Just two additional notes, the first being just this link whose URL tells the story:
The second is that Thu Dec 24 saw another data point summarized in this headline:
US: Durable Goods Orders rise by 0.9% in November vs. 0.6% expected
This good news did not outweigh the personal income and consumer spending disappointment when it came to the GDPNow-cast for the 4th quarter But since orders for big machinery take multiple months to be filled, this news has more bearing on the coming quarters of 2021, which continue to look good. (See last week's newsletter.)