Weekly Indicators for May 20 - 24 at Seeking Alpha

 - by New Deal democrat

My Weekly Indicators post is up at Seeking Alpha.

The big contradiction between what the yield curve is forecasting, and what most of the rest of the long leading indicators are forecasting, continues.  Meanwhile Trump’s tariff  “policies” are creating chaos in other sectors.

As usual, clicking over and reading should not only bring you up to date, but helps reward me with a penny or two for my work.

Economic indicator death match update: either bonds or housing sales are giving a false signal

 - by New Deal democrat

Yesterday featured the week’s sole important economic release: new home sales. But first, let’s update the inverted bond yield curve, which got more dramatic yesterday. The below graph compares the depth of the inversion yesterday (bottom) with March of 2007 (top):

With the sole exception of the 2 vs. 10 year spread, the yield curve is virtually screaming oncoming recession at this point.

Now let’s turn to new home sales. As a refresher, new home sales are the most leading of any housing series, but they are extremely volatile and heavily revised. Yesterday they were reported down -6.9% m/m for April, but that wasn’t the important news, because March was revised higher to a new expansion high. As the below graph shows, only once in the past 50+ years have new home sales made their cycle low *before* the onset of recession - in 2000:

So, while it’s only one data series, just going by past history, yesterday’s report very likely negatives any recession in the near future, completely contradicting the bond yield curve.

Of course, mortgage applications, and new home sales, have rebounded this year because interest rates have fallen from a peak of close to 5% last November to 4.1% this week. So in the below two graphs, I show new home sales (blue) vs. mortgage rates (red, right scale) since the latter series was started in 1971:

Mortgage rates didn’t decline at all prior to the 1980 recession, and declined only trivially in the month before the onset of the 1970 and 1981 recessions. They declined more significantly — -0.6%, -1.7%, and -0.8% — before the onset of the 1991, 2001, and 2008 recessions. In the cases of both the 1991 and 2008 recessions, housing still declined in part due to oil price shocks and in 2008 the unwinding of leverage in housing and the mortgage markets. The 2001 recession, by contrast, was a producer-led downturn, as the dotcom bubble imploded, so the mild upturn in housing was not enough to overcome it. Housing did turn down *during* the 2001 recession, as people got laid off, but did not match its 2000 low.

Now here is an update of the same data for this expansion:

The bounce in new home sales in the past several months has been much stronger than that of 2000.

The big fundamental question in the economy right now is whether the upturn in housing will be enough to overcome the downturn in corporate profits for the past two quarters.  In short, one of these two powerful economic indicators is giving a false signal.

Initial claims, temporary staffing point to weaker May jobs report

 - by New Deal democrat

As I’ve noted a few times recently, I’m paying additional attention to the weekly jobless claims numbers, partly because I suspected that the late Easter this year resulted in some residual seasonality (which I think has been demonstrated), and partly because if my slowdown forecast is correct, it ought to start showing up there.

The initial claims report this morning covered the week during which the BLS surveyed employers for the May jobs report coming out in two weeks. In the four weeks that coincided with the April report, initial claims made new 49 year lows, averaging 201,500. In the past five weeks that will coincide with the May report, the average has been 221,500.

So, while this isn’t precisely on point, here’s a graph of the four week moving average of claims (blue, left scale) vs. the unemployment rate (red, right scale):

I don’t think it was a coincidence that the unemployment rate fell to a 50 year low during the four week period that initial claims made 49 year lows. And I strongly suspect that the June report will take that back, with an unemployment rate of 3.8% +/-0.1%.

If that happens, it will be significant, because a 3.8% unemployment rate would be exactly what the rate was 12 months previously. And a YoY unemployment rate that has not improved has only happened once during this expansion (September 2016, right before the election).

Further, at the moment initial claims are higher YoY:

Even if they continue at the 210-212,000 range for the next few weeks, that’s still only about a 4% improvement from a year ago. 

Bottom line: unless jobless claims continue to fall in the weeks ahead, they are consistent with a slowdown.

And while I’m at it, another leading sector - temporary jobs - continues to show weakness as reported in the weekly American Staffing Association’s Index:

The four week average of the index is -2.9% YoY, the worst showing since the 2015-16 slowdown.

I was surprised by the strong +12,000 temporary jobs number in the April jobs report. I am expecting either that to get revised downward, or a poorer comparison when May’s report comes out, or both.

In short, the weekly data so far this month is consistent with an oncoming slowdown in employment gains. We’ll see in two weeks.

A comment about the economy and the 2020 election

 - by New Deal democrat

Recently I’ve seen a bunch of takes to the effect that “the economy is doing great, and therefore it is likely that Donald Trump will be re-elected.” In my opinion that fear is overblown for three important reasons.

The fist, least noteworthy reason, is that there is still a lot of time between now and the election. As I noted Monday, many - but not all - models of the economy indicate that a recession is likely between now and then, for reasons having nothing to do with the age of the expansion. Needless to say, a recession in 2020 would not bode well for either Trump or the GOP. 

Secondly, consider what economic interventions Trump and the GOP have made since they inherited the economy from Obama. There have been three: 

1. They passed a tax cut that lopsidedly favored the wealthy and corporations, that has generated zero acclaim from the middle and working classes - and with the decrease in tax refunds, may have generated net negative feelings. 

2. Trump has started several trade wars that are proving unpopular, partly because they mainly have hurt portions of his own base, partly because they are  resulting in net higher prices to consumers that may be getting noticed, and partly because negatively affected businesses may start laying off workers.

3. Trump is held responsible for the government shutdown that resulted in a mini-recession.

In short, it’s not clear to say the least that the public at large would give Trump credit for an economy that he mainly inherited from Obama and as to which his known interventions have been received negatively.

Finally, and most notably, the example of the Bush vs. Gore 2000 election strongly cuts against Trump. As I wrote in 2016, all of the fundamentals-based election models, such as the “bread and peace” model, or models based on the unemployment rate or on consumer income and spending, indicated that Gore should have won by nearly a landslide, on the order of 55%-45%, as shown in the graph below:

Instead, Gore won the popular vote by only 0.5%, despite being able to run on both peace and prosperity - the biggest outlier of the entire series going back to 1952. 

Two big factors held Gore back: first, the economic expansion had gone on for nearly 10 years, and at some point the public takes it for granted, or in other words, “so what have you done for me lately?” Second, as his Vice President, Gore was stained by Bill Clinton’s slimy personal life. 

Both of the factors that worked against Gore in 2000 are likely to work against Trump in 2020: if the economy remains in expansion, the public will probably take it for granted; and Trump’s pervasive sliminess, both public and private, will work against him. In short, Trump is likely to underperform compared with the fundamentals even more than did Gore.

While the example of 2016 certainly means that the 2020 election is another “all hands on deck” moment for Democrats, and nothing should be taken for granted, even if the economy remains in expansion as it is now I do not think that means Trump wins the election.

San Francisco Fed: ease of finding a new job is driving improved labor force participation

 - by New Deal democrat

This is a surprising result that is worth noting: the San Francisco Fed found that the increase in prime age labor force participation in the past five years has not been due to new people being drawn into the labor force, but rather by a very large decrease in people leaving it: 

[Note: keep in mind that prior to the early 1990s, both inflows and outflows are increasing due to the secular trend of women entering the workforce.]

Why is this surprising? Because you would think that increased wages would draw people on the sidelines into the workforce. This is something I’ve looked at a few times in the past several years, and the pattern has been clear:

1. The unemployment rate declines

2. Once the unemployment rate declines enough, the decline in labor force participation decelerates, but nevertheless continues.

3. Average hourly wage growth starts to improve.

4. Labor force participation starts to increase.

Here’s a graph showing this relationship since 1994:

The San Francisco Fed says that the reason for the big decline in outflows has been the ease of finding a new job, although that appears to be speculation. It might be that improved wage growth is something that is noticeable to people already in the labor force, rather than those presently outside of it.

Anyway, a counter-intuitive result worth noting.

Yes, Virginia, the government shutdown really did cause a mini-recession

 - by New Deal democrat

For the past several months, I have been pounding on the idea that the government shutdown, during which 800,000 jobholders were temporarily laid off without pay, had a much bigger impact on the economy than was originally thought.

This morning we get the following graph from Bank of America Merrill Lynch, which speaks for itself:

One of the most important insights from behavioral economics is that losses have an outsized effect on behavior compared to gains, usually on the order of 2 to 1. In the case of the government shutdown, about 0.5% of the workforce went without pay for about 45 days. Using the 2:1 ratio, that would translate into a -1% deadweight loss to the economy during that time. 

Of course, the workers got back pay when the government reopened - but if the 2:1 ratio holds, there wouldn’t be an equivalent “kick” from renewed spending. Which seems to have been the case, since the March +1.3% rebound in real retail sales didn’t make up for the -1.6% decline in December.