Saturday, November 20, 2021

Weekly Indicators for November 15 - 19 at Seeking Alpha


 - by New Deal democrat

My Weekly Indicators post is up at Seeking Alpha.

From time to time here, I put up a post including “you’re reading the right blog” in the title.  This week was similar at Seeking Alpha. In the past week or so, there have been a number of articles in the mainstream financial media about how the shipping bottleneck is at least beginning to ease, and gas and oil prices have begun to decline.

But that is something I started to write about in my “Weekly Indicators” columns almost in real time beginning 5 weeks ago.

This is exactly the kind of thing tracking high frequency indicators is designed to accomplish; a heads-up alert about a change in trend almost as soon as it starts, and well before monthly or quarterly data shows it, let alone the mass media reporting about it.

And reading today’s article will continue to provide you with that information, and reward me just a little for being ahead of the pack.

Friday, November 19, 2021

Coronavirus dashboard: As the winter wave takes hold, how much will resistance from prior infections hold numbers down?


 - by New Deal democrat

In Europe and North America, the winter wave is underway. While vaccinations clearly work, in most countries of the West there is a reservoir of defiant antivaxxers, who are not going to get vaccinated unless they are absolutely forced to. What does that mean for the present, immediate, and longer term future of endemic COVID?

In the US, nationwide average cases have increased from just over 70,000 3 1/2 weeks ago to nearly 95,000 yesterday, but about 5,000 of that is probably because Veterans Day went out of the weekly average. Deaths increased slightly from their roughly 1150 plateau  as well:

In the graph above, it is interesting to contrast the current situation with that of one year ago cases. With similar weather, but no vaccinations, and several subsequent huge waves of infections, in the first 2 weeks of November US cases had more than doubled to roughly 170,000. How much of that is vaccinations, and how much if any is increased seroprevalence among the population due to having already been infected?

Domestically among US States, one year ago there was the worst outbreak of the entire disease that was centered on the Dakotas, and started with the Sturgis motorcycle rally in early August. South Dakota went from 9 cases/100,000 to 165 per 100,000 at the peak in mid-November. This year it went from 4 cases/100,000 just before the rally to a peak of 64 per 100,000 5 weeks later, and has since declined to as low as about 35 per 100,000. As of today with very noisy data it is between 45 to 50/100,000. 

Even now only 54% of South Dakotans are fully vaccinated. In North Dakota, only 48% are. By contrast, a similarly populated rural State, Vermont, is 72% fully vaccinated.

So let’s compare both North and South Dakota with Vermont. First, here are average weekly cases for all three States since the start of the pandemic:

Vermont had been a star, up until the last 45 days, with the lowest % of infections of all States throughout 2020. More particularly, since their massive outbreaks one year ago, it is particularly noteworthy that the Dakotas and Vermont tracked virtually identically up until the Sturgis rally this August, despite their radically different vaccination levels.

Here is the cumulative number of confirmed cases in all 3 States:

Since February 1 of this year, 5.5% of Vermonters have had confirmed infections. Only slightly more, 6.1%, of South Dakotans have, while in North Dakota the numbers are worse, at 7.9% of the population newly infected since then.

In short, it certainly does seem, at least in the case of South Dakota, and maybe in North Dakota as well, that the massive outbreak last autumn has acted to depress cases ever since - although Delta still made a substantial impact.

This also seems to be borne out by the MIdwestern numbers as a whole, as States immediately adjacent to the Dakotas are seeing much bigger increases - among the biggest of all States, while the Dakotas by contrast are like the calm in the eye of a hurricane:

Next, let’s look at the the US States in the lower Mississippi valley and Deep South, where Delta first struck:

MO and AR have increasing trends, while FL, LA, MS, and AL look flat except for the Veterans Day comparison. Florida, with fewer than 7 new cases per 100,000, now has the lowest rate of all 50 US States (although Puerto Rico, with 4 cases per 100,000, is doing even better). I emphasize that this is not to praise Florida in the slightest, as it has among the highest total infection and death rates among the entire country, but rather to point out that, Delta having burned through the dry tinder, and with weather now more favorable to outdoor activities, there are fewer susceptible people left.

In short, when we look domestically within the US, it certainly does appear that prior outbreaks, particularly more recent outbreaks of Delta, have provided substantial resistance to the spread of new infections.

Some interesting international numbers tell as similar tale. The UK, after a quick 2 week decline of 40% from Delta’s original peak, has had numbers see-sawing generally between that peak and trough for the past 3 months, and is currently averaging 60 cases/100,000. With a few exceptions the EU, by contrast, largely escaped Delta’s summer wave, but in the past 7 weeks has seen it really take hold, jumping from 10 to 45 cases/100,000. *If* present trends continue, the EU could overtake the UK in 10 to 14 days:

On the other hand Canada, immediately adjacent to some of the currently worst hit US States, has only had a slight increase from 6 to 7 cases/100,000 in the past 2 weeks. This is particularly noteworthy since 76% of Canadians are fully vaccinated, only a little above New England’s 70%+, and less than 10% higher than Germany, currently having a huge outbreak. Also, needless to say Canada’s weather is not particularly balmy compared with the northern tier of US States.

Here is what Canada’s numbers look like compared with the adjacent Northeastern US, which has similar if slightly lower vaccination numbers:

I suspect there are two explanations for this. One presumably is more universal compliance with public health measures in Canada. I suspect the other has to do with there being significant pockets of anti-vaxx resistance even in the heavily vaccinated States of New England. In other words, once you cross a certain threshhold - say, 70% - of vaccinations, maybe there is little additional benefit in terms of limited spread, while substantial pockets of 50% or 60% vaccinations still allow major outbreaks, with some leakage into the more heavily vaccinated regions.

I continue to think that this year’s winter wave will be substantially lower than last year’s - although it may exceed this summer’s Delta wave - and that each wave thereafter will echo, at increasingly lower levels, similar waves from one year previous, as both vaccinations increase, and resistance from prior infection among the unvaccinated population increases as well.

Thursday, November 18, 2021

Nobody is getting laid off: the continuing saga


 - by New Deal democrat

Initial claims declined another 1,000 this week to 268,000, and the 4 week average declined 5,250 to 272,750, both - yet again - new pandemic lows:

For the past 50 years, initial claims have only been at these levels for 2 months at the peak of the late 1990’s tech boom, and from late 2015 to just before the pandemic in 2020.

Continuing claims also declined 129,000 to a new pandemic low of 2,080,000:

Similarly, only a few weeks in the late 1980s, plus 2 months in 1999, plus the last 4 years of the last expansion were below this number:

I am a happily broken record: once again, the labor market remains extremely tight. For all intents and purposes, nobody is getting laid off. 

Wednesday, November 17, 2021

Housing construction continues to stabilize, but with record bottleneck in starts


 - by New Deal democrat

Last month I highlighted that housing constructions was stabilizing, following the stabilization in interest rates. This month continued that trend. 

In October, housing starts (green in the graphs below) decreased -0.7% m/m, while the more leading total permits (blue) increased 4.0%. The less volatile single family permits (red) increased 2.7%. As a result, the overall trend for all three metrics for the past several months is generally flat:

On a YoY% basis, starts are up 3.5%, permits barely up 0.4%, and single family permits down -6.3%:

The YoY increase in starts continues to be noteworthy because it highlights an unusual event which has taken place over the past year; namely, a record number of permits were issued for houses that were not promptly started. Here’s a graph of such housing for the past 3 years:

The current level is the highest since the 1970s (not shown).

In other words, the actual on-the-ground economic activity in housing construction hasn’t declined that much, presumably because housing materials at reasonable prices constrained the actual building of houses authorized by permits. On a rolling 3 month average basis, housing starts are only down -4.4% from their year end 2020 peak. This suggests much less of a real economic downdraft than would otherwise be the case, as typically it has taken a downturn of about -20% to be consistent with a recession.

As I have repeated many times, interest rates lead housing construction. And the evidence from mortgage rates is that housing should be (and is) stabilizing. In the past 6 months rates have stabilized between the 2.75%-3.15%:

As a result, I would expect stabilization or a moderate increasing trend in response. This is  shown when we compare the YoY% changes in mortgage rates (inverted) and single family housing permits over the past 10+ years:

In October, Mortgage rates only were increased 0.14% YoY, and for all intents and purposes have been flat YoY for the past 4 months. 

Since housing construction is a long leading indicator, this series continues to suggest that the economy, after a period of cooling early next year, will also stabilize later on.

Tuesday, November 16, 2021

. . . And Industrial Production isn’t too shabby, either


 - by New Deal democrat

Industrial production, the King of Coincident Indicators, was reported for October this morning in addition to real retail sales, discussed already in my last post. - and it was also quite positive.

Total production increased 1.6% during the month, and manufacturing production increased 1.3%. This was the biggest monthly increase for total production since March of this year, and except for that month, also the biggest increase in over a year:

Both total and manufacturing production have returned to being higher than their immediate pre-pandemic levels, and also the highest since the onset of the pandemic:

The former is up 0.3% compared with February 2020, and the latter up 1.6%.

I wrote yesterday that, despite my real concern with inflation going forward, my array of indicators did not forecast a recession, and this morning’s reports on both retail sales and industrial production confirmed that point of view.

Now *that’s* good news: another blockbuster real retail sales report


 - by New Deal democrat

Yesterday I wrote that the financial and production sides of the economy still looked very positive, and that today’s retail sales number would be especially important.

Well, they were very positive, clocking in at up 1.7% month over month in October. Even after inflation, “real” retail sales were up 0.7%. September was unrevised. Although real retail sales are down -2.2% from their April peak, they are +13.5% higher than they were just before the pandemic hit, and 4.9% higher than January of this year: 

In September I wrote that ”while the recent decline from April is consistent with a slowing economy ahead, if sales stabilize here I don’t see this as a harbinger of an actual downturn.” That still looks correct, particularly as real retail sales are up over 2% since then, and 9.5% higher YoY. 

How extreme is that? The below graph subtracts 9.5% YoY growth from retail sales from 1948 through 2019:

With the exception of 2 months in 1983 and 1984, real retail sales haven’t been this strong since the early 1970s! That’s the last time the US had such a worker-favorable economy.

This also explains a great part of the supply chain bottleneck, since it is incapable of handling such a sudden jump in consumer demand. Here’s a graph I came across a couple of weeks ago showing activity at the two big California port facilities:

In other words, even though the ports are processing record volumes, they *still* can’t keep up with the increased import demand.

Now let’s turn to employment, because real retail sales are also a good short leading indicator for jobs.

As I have written many times over the past 10+ years, real retail sales YoY/2 has a good record of leading jobs YoY with a lead time of about 3 to 6 months. That’s because demand for goods and services leads for the need to hire employees to fill that demand.  The exceptions have been right after the 2001 and 2008 recessions, when it took jobs longer to catch up, as shown in the graph below, which takes us up to February 2020:

Now here is the same graph since just before the onset of the pandemic. Note the scale is much larger due to the huge downturn during the lockdowns and the comparisons one year later:

Last month I argued that, despite the lackluster initial jobs reports for August and September, this “argues that we can expect jobs reports in the next few months to average out about even with those from one year ago, which averaged about 500,000 per month.”

Well, two weeks ago those months were revised well higher, and October came in at 531,000 jobs added - which means that job growth indeed has continued to average about 500,000 per month. And should continue to do so in the next few months.

Monday, November 15, 2021

Dear Democrats: Yes, inflation is a problem


 - by New Deal democrat

[Update: I see where Larry Summers has obviously read my piece below, and says, “I agree with NDD!”  (Just teasing)]

In the past few days, I have seen a spate of articles and tweets from prominent partisans and economists telling Democrats not to worry about inflation, either because it is a transient supply chain issue, or else because Biden’s infrastructure and “Building Back Better” plans will not add to it. 

I dissent. I think the evidence is compelling that inflation *is* becoming a problem. Not because of Biden’s plans - which may not add to inflation - but rather because inflation in the three most important things that consumers either notice or care about - gasoline,  cars, and houses - is not so transient at all. In particular, I believe that elevated inflation numbers for housing, which is over 1/3rd of the entire metric, is going to persist for at least another year.

Let me take these in order.


Noticing changes in the price at the pump is the most visible manifestation of inflation. And in the past 18 months, since just after the lockdowns ended, on average the price of gas has increased from $1.77 per gallon to $3.41 one week ago. Not only is this a near-doubling in price, but it is the highest in the past 7 years:

Not only that, but energy prices tend to feed through into the wider economy with a 6 to 12 month lag. This is shown in the next two graphs, which show the YoY% change in energy prices (blue) vs. the YoY% change in all other prices (red). Here are the higher inflation 1970s and 1980s:

And here is our lower inflation era beginning in 1992:

(Note that I have divided energy prices by 2 and 6, respectively, in the two graphs for scale so that the relationship is better shown).

In 1974, 1979, 1991, 1999, the mid-2000s, and 2011, a substantial increase in gas prices was followed by an increase in the overall inflation rate for everything else. It certainly appears that the same pattern is happening now.

This suggests that, even if gas prices have peaked for now (quite possibly true), that inflation will continue to seep through into the overall inflation rate for the next 6 to 12 months.


Below is a graph of the YoY% change in prices since 1976 in new motor vehicles (blue) vs. the YoY% change in the number of motor vehicles (excluding commercial trucks) sold (red), averaged quarterly (note sales /10 for scale):

With the sole exception of the period right after the Great Recession, whenever there has been a surge in prices of new vehicles, YoY vehicle sales have turned negative.

Now here is a close-up of the past 5 years, monthly:

The big YoY spike in sales this spring is an artifact of the huge decline during the spring 2020 lockdowns. Since then as prices have spiraled higher, now up almost 10% in one year alone, sales have turned negative.


The pattern is similar for housing, as measured by the FHFA (dark blue) and Case Shiller (light blue) house price indexes. Big YoY changes in the price of houses, by 5% or more, have typically been followed by a decline in new houses sold as measured by building permits (red, /4 for scale) (note graph subtracts 5% for both house price indexes, so a 5% YoY increase shows as 0, better to show the relationship):

The above graph averages permits by quarter to cut down on noise. Below is the same data for the past 5 years, including monthly changes in permits:

Permits are now virtually unchanged from one year ago.

Consumer reaction

Unsurprisingly, consumers have decided that now is the worst time to buy a house, car, or major household appliances since the 1980-82 recessions:

They’re not wrong, and this isn’t good for Democrats.

Further, not only are the increase in gas prices likely to seep through into wider inflation in 2022, but so is housing, due to the unusual way that housing inflation as measured. I referred to this in an article last week, but want to spell it out further.

The CPI for housing is measured by something called “owners’ equivalent rent,” or, “how much could I rent my house out for if I wanted to?” It’s bizarre, but for this post just take it as a given.

Owner’s equivalent rent (blue in the graph below) typically tracks very close to the inflation measure for “rent of primary residence,” i.e., how much are renters paying each month? (gold):

Note the big YoY increases in rent, and owners’ equivalent rent, associated with 3 of the last 4 recessions. Note also the big increase in the YoY level of each in the past 6 months.

Now here  are the two house price indexes (blue and gray, /2 for scale) compared with owners’ equivalent rent (red):

While the relationship is somewhat noisy, it does appear that big increases in house prices have correlated with big increases in owners’ equivalent rent with a lag of 12 to 24 months (bottoms in the former tend to lead bottoms in the latter by about 6 to 12 months). That’s probably because big increases in house prices drive people out of that market, and into the market for apartment rentals. 

In other words, not only are higher gas prices likely to feed into higher rates of overall inflation in the next year, but “owners’ equivalent rent” also is likely to contribute elevated readings to CPI during that time frame as well.

Historically, big declines in sales of new houses and cars of over 20% has typically given rise to recessions over the past 40+ years:

New home sales, as measured by permits, is down over 15% but not 20% at present. Car sales, however, are depressed at levels typically associated with recessions before.

Needless to say, continued inflation at rates not seen since the 1970s in 2022 is not “transitory” and is not going to be good for Democrats.

I hasten to add that my array of long and short leading indicators are not pointing to recession at this time. Interest rates and bank lending are very accommodative, and producers are still going full-tilt. The key for consumers is going to be whether they pull back in overall purchases. We will get important information about that in tomorrow’s retail sales report for October.