Saturday, May 3, 2014

Weekly Indicators for April 28 - May 2 at

 - by New Deal democrat

There was a real schizophrenic read on Gallup consumer spending vs. the ICSC and Johnson Redbook this past week.

International Week in Review: The US' Weak Growth Head Fake Edition

This is over at

Friday, May 2, 2014

Is it really different this time? historical recoveries for low vs. high paid jobs

 - by New Deal democrat

In the last week there has been another spate of articles on the order of The low wage job explosion from CNN Money.  On Monday Prof. Mark Thoma picked up the issue.

There is no doubt that what the articles are saying about the post-recession jobs situation is correct.  Unfortunately, none of these articles compare this recovery with prior recoveries.  The question is, is this jobs recovery producing a different mix of jobs from earlier recoveries, or is it producing a similar mix, but taking a longer time to fully recover?  Put another way, is this jobs recovery different in kind, or just different in intensity compared with prior recoveries?

This is an issue I spent some time researching a while ago, but never wrote about.  Since I left a comment at Thoma's site earlier this week and got a plaudit from the well-known Anne in response, I thought I would repost the research I've done so far here.

Here's the hypothesis I am testing:  the typical pattern in recoveries is that low wage jobs recover faster than high wage jobs.  Thus, earlier in a recovery it will be the case that the economy seems to be producing only low wage jobs.  Since most post-World War 2 recoveries restored all jobs relatively quickly, the pattern was not long-lasting.  This recovery, however, is from a much deeper hole and is taking much longer, so the period in which it is producing mainly low wage jobs is taking longer and has been profoundly noticed.

If my hypothesis is correct, then as the unemployment rate drops, a bigger proportion of higher wage jobs ought to be created.  Dean Baker produced a graph earlier this week comparing the state by state unemployment rate with low wage food service jobs, that lends support to this proposition:

Notice that, the lower the unemployment rate, the lower the relative rate of low wage food service jobs.

So I examined two types of jobs: I compared recoveries in construction jobs (relatively high paying) with retail jobs (low paying), measuring how long it took for each to recover fully from their pre-recession highs for each recession since World War 2. Here's a chart of what I found:

constructionretaillag in
1953 11/19542/1955-3
1957 11/19641/19592
1970 4/1971 9/19707
1974 4/19788/197544
1981 12/19841/198323
1990 3/19962/199425
2008 n/an/an/a

With the exception of the recovery from the 2001 recession, in every recovery beginning with 1957, retail jobs fully recovered before construction jobs.  This supports my hypothesis.

In essence, I am going to extend the research by the National Employment Law Project, shown in the below graph, which compared only job recoveries from 2001 and 2008, to prior job recoveries.

In addition to retail, note that the two biggest categories of low paying jobs are food service and drinking places, and administrative and waste services.  Data for these two categories of jobs only goes back to the 1990's, so they are of very limited value for the comparisons I need to make. For the record, food service and drinking place jobs returned to their 1990's peak in December 2003 (3 months before construction jobs); waste services returned to their 1990's peak in September 2005 (18 months after construction). Food service has not regained its pre-Great Recession peak. Waste services exceeded its pre-Great Recession peak in November 2013.

In addition to government, I need to test the hypothesis against some of the other high paying private sector job categories besides construction. I imagine this will take a few weeks.  I'll post results as I get them, and we'll see if the hypothesis pans out.

Prime working age employment population ratio improves in 2014

 - by New Deal democrat

As I pointed out below, that the employment population ratio has increased while the percentage of the population in the labor force has decreased seems like a good thing, suggesting that the reason for the decrease is moving more and more towards Boomer retirements.  Part of that may be that Boomers who want to retire can now enroll under the ACA, and don't have to wait for their Medicare eligibility.

I thought I'd look at a metric Paul Krugman has examined from time to time, in order to avoid the confoundment of the data by Boomer retirements; namely, the employment to population ratio of 25 to 54 year olds.  And I got a surprise:

Since last October, the employment to population ratio in this age group has increased from 75.5% to 76.5% (in March, it was actually higher at 76.7%).

It is now higher than it was at the beginning of 1985.  It has made up about 1/3 of its Great Recession decline.

This is good news.

April jobs report: a blowout (almost)

- by New Deal democrat

In April 288,000 jobs were added to the US economy.  The unemployment rate fell from 6.7% to 6.3%, a new post-recession low.  January and February were revised upward by 36,000, with both now over the 200,000 threshhold as well.

As usual, first, let's look at the more leading numbers in the report which tell us about where the economy is likely to be a few months from now. These were mixed but with a tilt towards positive.

  • the average manufacturing workweek fell from 41.1 hours to 40.8 .This is one of the 10 components of the LEI, and will contribute significantly towards a negattive number.

  • construction jobs increased by 32.000. YoY 189,000 construction jobs have been added.

  • manufacturing jobs  rose by 12,000.

  • temporary jobs - a leading indicator for jobs overall - increased by 24,000.

  • the number of people unemployed for 5 weeks or less - a better leading indicator than initial jobless claims - declined by 14,000 to 2,447,000, compared with December's 2,255,000 low.

Now here are some of the other important coincident indicators filling out our view of where we are now:

  • The average workweek for all nonsupervisory workers was unchanged at 33.7 hours.

  • Overtime hours were also unchanged at 3.5 hours.

  • the index of aggregate hours worked in the economy rose by 0.3 from 108.1.  This is a new record.

  • The broad U-6 unemployment rate, that includes discouraged workers decreased from 12.7% to 12.3%, also a post-recession low.

  • The workforce declined by 806,000. Part time jobs decreased by 276,000.
Other news included:
  • the alternate jobs number contained in the more volatile household survey decreased by 73,000 jobs.  The household survey jobs numbers had been lagging the establishment survey numbers, but as expected this difference has now been entirely made up, with the household survey showing a 1,993,000 increase in jobs YoY.

  • Government jobs increased by -15,000.

  • February was revised upward from 197,000 to 222,000.  March was also revised upward by 11,000 to 203,000  Upward revisions happen in expansions, and after a weak spot in late 2013, these revisions in the last several months have all been positive.

  • average hourly earnings were unchanged at $24.31. The YoY change is +1.9%.  As a result, YoY average real wages probably fell slightly in April, given the expected slight rise in consumer prices due to the weakening of the Oil choke collar.

  • the employment to population ratio was unchanged at 58.9%, and has risen +0.3% YoY. The labor force participation rate declined from 63.2% to 62.8%, and has fallen by 0.6% YoY.  The usual  caveats about discouraged workers and Boomer retirements apply, but the comparison of the two measures, i.e., a greater percentage of the population is employed, but a smaller percentage of the population is in the labor force, looks very much like Boomer retirements are moving to the fore.
  • the number of people who are not in the labor force but want a job now (the best measure of long time discouragement) declined 245,000 and now totals 3,125,000.
This was a blowout report by the standards of the last 7 years.  This would not be seen as a blowout by any standards prior to 2000, however.  With one exception, all of the metrics moved substantially and positively.  We may have sub-6% unemployment, finally, by the end of this year.  The Doomer memes of part-time employment and discouraged workers both took a hit, as it appears that, excepting Boomer retirements, participation is increasing and discouragement is decreasing.

The big decrease in the civilian labor force will probably get lots of attention this month, but no obvious reason for the decline besides monthly noise appears on the surface.

The one big caveat to this report is wages. They are stalling and in real terms are declining again slightly. The YoY comparison, which had been as high as 2.5%, has faded back under 2%.  This is not good, especially over 4 years into a jobs recovery.  What happens to wages the next time there is an economic downturn.  That is the problem that worries me the most.

From Bonddad:

These are my thoughts, in no order of importance:

Let me begin with this curmudgeonly caveat: I really don't like the monthly employment number report.  The US labor market is wide and deep, and has many nooks and crannies.  And those various nuances are still pretty negative: under-utilization is high, confidence is low and utilization is weak.  However, the market loves this number and report, so let's dive into the details.

Wow.  288,000.  That's a great headline number.  Overall growth was far more tilted to the service sector as 220/288 (or about 76%) of the jobs were added in that sector.  Considering the harsh winter weather, that's to be expected.  Manufacturing only added 12,000 jobs, but considering the high degree of automation involved with US manufacturing now, that's really not a bad number.  And construction only added 32,000.  My guess is contractors are still trying to regroup after the winter.  Hopefully we'll see a stronger increase in this number as the weather warms up and housing starts, well, start.

There were some interesting developments in the household survey.  The civilian non-institutional population increased 181,000 but the labor force dropped by 806,000.  That's a really big drop.  On the good side, the number of unemployed decreased by 733,000 but the total number of employed dropped by 73,000.  These rather interesting internals explain the sharp drop in the unemployment rate and participation rate.  These internal numbers are a bit odd; I'd like to see a hard-core statistician explain them. 

Total hours worked and total wages paid were stagnant.  While this is usually a bad thing, I would offer the following explanation: rather than increase hours and pay, maybe employers started to add to payrolls instead?  That's just a thought.

February and March were revised a higher by a combined amount of 36,000.  This tells us the initial weakness from those readings is still intact; but we weren't as weak as we thought initially. 

I agree with NDD that, when judged from the perspective of this expansion, this is a good report.  However, when we broaden that base of comparison, it's not that great. 

Thursday, May 1, 2014

More shallow DOOM about jobs at Daily Kos

 - by New Deal democrat

I haven't had enough time to clean up some draft posts this week, one of which is about the relative strength in this recovery in terms of low wage jobs vs. high wage jobs. Is this recovery unique in that regard, or is it similar to past recoveries?

I figured there would be a Doomer post or two at the usual suspects,and I had hoped to already have the main research done before they weighed in, but I haven't had the time.  So, earlier this week Mark Thoma had a brief piece up, and I had an exchange with commenter Anne on the topic.  I'm just going to repost the comments here for now:
NDD: Is this recovery actually any different than other post World War 2 recoveries in terms of creating low wage jobs before high wage jobs? I looked at a few post-war recoveries and it looked like the same pattern was followed, but full employment returned quickly so the initial shortfall in higher paying jobs wasn't noticed. Dean Baker seems to be making a similar point in response to the NYT article; namely, it's the shortfall from fulll employment that is responsible for the prevalance of lower paying jobs.
I'm asking the above as a real question, rather than making a statement. If somebody can compare this recovery vs. other post WW2 recoveries in terms of the order of the return of low vs. high paying jobs, I'd be very interested in what they find.

NDD: OK, I've compared construction vs. retail employment as proxies for high and low wage jobs. I used construction since, unlike manufacturing, it can't all be offshored.
Here's what I found. Beginning with the 1957 recession, in every recession except the 2001 recession, retail employment returned to its previous high sooner than construction employment, by as little as 2 months and as much as 62 months! The exception was that after 2001, construction returned to its previous high 7 months ahead of retail.
Needless to say, neither has returned to its pre-Great Recession high, although retail is closer. If these are fair proxies, this supports the idea that the current lag in high paying jobs is a question of intensiy, and is not different in kind from most post WW2 recessions.
Anne:  Having now looked carefully at the data from this perspective, I agree. Again, well done. The problem is evidently general demand, not the structure of demand, which makes sense. 
I'll have a lot more to say on this issue in the next few weeks or so, because I have a lot of number crunching to do.
For now, the question to Meteor Blades and all of the Doomers at Daily Kos is,
How does the structure of demand (low wage vs. high wage jobs) stack up since 2009 ...
  • vs. 1948?
  • vs. 1954?
  • vs. 1957?
  • vs. 1960?
  • vs. 1970?
  • vs. 1974?
  • vs. 1980?
  • vs. 1982?
  • vs. 1990?
  • vs. 2001?
Has this recovery uniquely produced more low paying jobs? Or, like prior recoveries (as to construction vs. retail jobs at least), is the recovery in low paying jobs simply taking place more quickly than the recovery in high paying jobs?   Put another way, is this recovery different in kind or just different in intensity? Until you can answer that question, you really don't have a clue about whether this recovery is different from any other recovery in terms of how quickly it has produced low paying jobs vs. high paying jobs.

But don't disturb the echo chamber by forcing them to do some actual comparisons.  Spoils all the fun in DOOOOMMM.

March personal spending rebounds from winter doldrums as reflected in yesterday's GDP report

 - by New Deal democrat

I have a new post up at

The very positive March personal spending report is consistent with the idea that we are rebounding from the near-recessionary consumer spending of January and February that were reflected in yesterday's flatlining Q1 GDP report.

Wednesday, April 30, 2014

The decline in real private residential spending raises a yellow flag for US economic growth in 2015

 - by New Deal democrat

I discuss this aspect of today's first quarter GDP report in a post over at

1Q GDP: Blame the Weather

This is up over at

1Q GDP: Blame the Weather:

Don't freak out about the poor 1st quarter GDP report (except for one thing)

 - by New Deal democrat

So GDP grew by the smallest amount possible, +0.1% annualized, in the first quarter of 2014.

If you've been reading my "Weekly Indicator" series religiously, you know that all through January and February I was reporting that the economy, as measured by the high frequency weekly data, had hit an air pocket.  It recovered in the latter part of March.

This was one of those rare cases where the weather really was a valid excuse, and that's what the late March and April data have been confirming.

So, yes, this was a bad positive number.  But a bad positive number with a perfectly reasonable explanation, that need not carry forward at all.

With one exception.  And that is the line that reads, "Real residential fixed investment decreased 5.7 percent, compared with a decrease of 7.9 percent" in the fourth quarter of 2013.

According to UCLA Prof. Edward Leamer, a decline in real residential fixed investment as a share of GDP is the first warning sign of a recession about 5 quarters out.  With a decline of two quarters in a row, the outlook for 2015 has become more problematic.

I'll have more up at later, and I'll update with a link here.

Monday, April 28, 2014

My housing bet with Calculated Risk: March results

 - by New Deal democrat

As most readers know Bill McBride a/k/a Calculated Risk and I have a charitable bet about the direction of housing in 2014.

In his forecast for 2014 residential investment, CR said, "I expect growth for new home sales and housing starts in the 20% range in 2014 compared to 2013."

By contrast, several months ago, in a post at, I said that "If the typical past pattern is followed, we will shortly see permits running 100,000 less than one year previously."

Here are the terms of our bet:  If starts or sales are up at least 20% YoY in any month in 2014, I will make a $100 donation to the charity of Bill's choice, which he has designated as the Memorial Fund in honor of his late co-blogger, Tanta.    If housing permits or starts are down 100,000 YoY at least once in 2014, he make a $100 donation to the charity of my choice, which is the Alzheimer's Association. 

This morning the final monthly report on housing for March, pending home sales, was reported by the NAR.  The index was up 3.4% m/m (its best showing in 4 months) but remains down 9.7% YoY:

(h/t Mortgage News Daily)

So, with the first three months of 2014 in the books, how do we stand?  Below are two graphs.

First, here is a graph of the change, in thousands, YoY of starts (blue), permits (red), new home sales (green), and existing home sales (orange) (note that the St. Louis FRED does not track pending home sales):

Next, here is the YoY% change in the same four statistics:

Both of these graphs show the clear deceleration in the housing market through 2013 and further into March 2014, to the point where 4 of the 5 monthly reports have turned negative YoY.  The strongest metric is housing permits, which has rebounded over the last two months from a +3.3% YoY reading in January to a +12% YoY reading in March.  Nevertheless, the graphs make clear that except for permits, deceleration has turned into outright YoY decline.

More specifically, through March 2014:

  • Permits are down -4% from their October 2013 high
  • Starts are down -9% from their December 2013 high
  • New home sales are down -16% from their January 2013 high
  • Existing home sales are down -15% from their July 2013 high
  • Pending home sales are down  -12% from their June 2013 high
Bill's forecast of 20% annual growth will take a real reversal of momentum of most of the metrics (housing starts and sales will have to be up about 25% on average YoY for the next 9 months), although there are some individual months where there are some easy YoY comparisons.  Obviously I am expecting some further deterioration, concentrated between now and mid-year.

Initially after I made this forecast, most people seem to be either surprised by or ignoring the housing slowdown.  While the horrible March new homes number in part reflects the volatility of that series, it seems to have served to concentrate economic minds on the sector now.

Sunday, April 27, 2014

International Week in Review: US Housing Market Starting to Cause Concern Edition

This is up over at