.comment-link {margin-left:.6em;}
Visit Freedom's Zone Donate To Project Valour

Friday, May 04, 2007

Employment April And Trends

Someone asked if the economy is growing. There is something of a manufacturing expansion, clearly, but the consumer side is now clearly in trouble. Services is up in the air. NACM CMI was mixed; service sector looks relatively weak while manufacturing is taking an uptick. From 2006 to 2007, services dropped from 58.3 to 55.4, while manufacturing increased from 56.6 to 57.9. From March to April, services held steady whereas manufacturing rose from 54.6 to 57.9.

The details on both series are a bit troubling; there is some underlying financial weakness showing. For manufacturing it is particularly obvious; accounts placed for collection shot up from 50.0 in March to 58.3. Another way to look at this is that for manufacturing, the index of unfavorable factors rose considerably more than the index of unfavorable factors. The positive index was achieved on the strength of more credit granted, and it looks like credit standards might be loosening to achieve that.

The amount of credit extended dropped for both indices. Services looked healthier overall, but not sterling. The general picture is of an economy struggling to adapt to pressures. I cannot emphasize enough that a commercial credit bubble has existed in tandem with the consumer credit bubble, and that what happens to the commercial bubble will mean more for the economic outlook than in the consumer credit bubble. Consumers can and will declare bankruptcy and continue living. Some businesses are going to go under and some will stop expanding. If jobs stay relatively strong, consumer defaults may hurt badly, but they pass over relatively quickly. If businesses have to cut much, the consumer impact will be doubled and long term growth will be affected.

Today's monthly employment release confirms the iffy economic situation. The headline (establishment) number is 88,000 new jobs for April, but 25,000 of those were government (mostly local), so the more significant number is 63,000.

Real earnings are going to fall again. That figure comes out in the middle of May, but average weekly earnings fell in this release (due to a shorter work week), so we know real earnings will fall. Needs inflation for consumers isn't slowing much, if any.

Sectors losing jobs according to the establishment survey were construction, manufacturing and retail. Not so good. Professional and business services gained, education and health gained, and leisure and hospitality gained. Out of the net positive 88,000 jobs added, 25,000 were in government and 53,000 were in education and health. Both of these categories include a lot of government paid services, so this report does not look much like an expansion of the private economy (which ultimately must pay for our government expenditures).

We'd better hope the manufacturing sector continues to expand. The weak dollar should help there. Because we import so many of our consumer goods, weak consumer spending doesn't necessarily hurt the manufacturing base.

The household survey didn't look too hot. Overall employment declined 468,000. The reason the unemployment rate only edged up to 4.5% is that the workforce is reported as having declined by 392,000. Unemployed workers are reported as having risen by 77,000. This pretty much summarizes the pattern so far this year - the household survey shows one thing, and the establishment survey shows another.

Probably due to the weakness in retail, the unemployment rate for workers aged 16-19 moved up to 15.3%. A year ago it was at 14.6%, and in March it was at 14.5%. This was largely female unemployment; the unemployment rates for men in this age bracket rose .3% from the prior month but dropped .1% from a year ago. The unemployment rate for construction workers has increased YoY from 6.9% to 8.6%.

The one table that everyone should read each month out of this report is Employed Persons by Class of Worker and Part-time Status. YoY, persons working part time for economic reasons such as slack work or only being able to find part-time work has risen by over 400,000 persons. Seasonally adjusted figures show the rate of increase in this category picking up.

Revisions to Establishment Non-Farm:
Original Jan: p137,258...Latest: 137,329 (r+71)..[Int. 137,313]
Original Feb: p137,410..Latest: 137,419 (r+9)....[Int. 137,442]
Original Mar: p137,622..Latest: p137,596 (r-26)
Original Apr: p137,684
There is no clear pattern to the revisions, which may indicate a shift in employment trend. Or it just may be noise. Much of the excitement about these month-to-month releases is overdone, because on revision the pattern often looks quite different! If the March number had not been revised down, numbers for April would be causing panic in the streets. But the March number will be revised once more, and the April figures will be revised twice.

Once you get the final number, it's often wise to average over two months to figure a trend. Having said all that, it is hard to see anything really good in the matched releases (original to original, final to final). If next month came in well, the picture would change sharply. Employment indices for both ISM Svcs and Mfg came in so strong that I am hoping to see better numbers for May.

I find the continuing unemployment claims number and part-time work numbers to be more helpful than anything else. The continuing claims number shows that employment has picked up a bit recently. It also shows that employment trends are weaker this year than last. I am pretty sure that the involuntary part-time work increase is due mostly to weakness in construction.

In terms of a quarterly outlook, it is hard to believe that the economy will pick up overall. If anything, weakness rather than strength seems to be diffusing. Why so many economic pundits are claiming that income growth will shore up consumer spending to produce a pickup in growth later this year is beyond me. I don't see that in the numbers; I see real incomes for the majority of Americans dropping. You've got your working-age people, and the non-managerial real wages for them are dropping. Then you have retired individuals, who are on fixed incomes with yearly CPI-based adjustments (usually). Both groups are getting hit hard by inflation.

It is at times like this that consumer debt overload begins to assume more weight. Take that 400,000 unwillingly part-time workers. If they are carrying much in the way of consumer debt, they are going to feel its pressure far more than last year. The wave of mortgage resets rolling through the summer is going to throw a little more water on the embers of consumer spending.

Until needs inflation drops, consumer spending is going to be a constrained affair. The economy does not run on sales of $300 purses. With any luck, exports for durables will rise on the weakened dollar. In the short term, an expansion in manufacturing won't help that much with unemployment, but in the long run it may amount to much healthier capital flows and underlying investment which will lead to a much healthier economy.

Productivity numbers are due to keep dragging in the dust, given the pattern of real employment drop we are seeing. Generally accepted economic theory is that rises in productivity are what permit a country to experience strong GDP growth with low inflation.

The difference between the real numbers (absolute declines in employed workers and absolute rises in unemployed workers) and the relative numbers (unemployment rate only rising to 4.5% and steady since Jan overall) are probably being produced by demographics such as the retirement of the leading edge of the baby boomers and the exodus of illegals. GDP is a total measure of growth, and to expect GDP to expand much if this condition continues is more than a bit delusional, IMO.

The Fed would be taking a terrible risk in lowering rates. Only the idea that the Fed will control inflation over the long term is preventing a rise in mortgage rates, and only that will help to mitigate the mortgage mess. If they do anything, they should raise rates.

I believed early this year that the Fed would cut rates to prevent the early implosion of HELOCs and seconds, which is where the majority of the losses in foreclosures are being incurred. But it didn't, and now it has no room to cut them.


Maybe you could clear something up: what is the relationship between the employment situation report and the hours worked component of the productivity report? After seeing the 1q decline in hours worked, I thought this might mean bigger revisions (down) in the job growth for the quarter.

How do you get a decline in hours worked on the one hand and an increase in jobs and the average work week on the other?

I agree with you that the divergence between incomes and real wages/fixed income is becoming significant. I wonder why the AARP doesn't pick up on this as its members are particularly hurt by needs inflation (and they don't tend to by hedonistically-deflating plasma screens!).

David Pearson
David, BLS monthly employment and the productivity series use different data compilations. Since both are subject to much revision, eventually one would hope to see a convergence. Wage and salary compensation comes from BLS and BEA.

The very preliminary productivity report.

If you look at this they are showing drops in productivity, but also in real compensation YoY. For the 1st quarter, business real wages overall dropped 1.9 and non-farm business dropped 1.5. Manufacturing durable goods rose 2.6 but nondurables dropped 0.2. The net (because nondurables are larger) was a 1.6 rise.

All of this data is imperfect, but I would say that jobs are a bit weaker than the establishment data is showing.

According to the productivity report, aside from durable manufacturing, real wages per hour for businesses dropped in Q12007. GDP personal income includes broader measures of income, so one can understand the discrepancy there.

I have a hunch that the executives at AARP might be a bit too well paid to realize what's happening to the average retiree! Give them a few more months.

I never bother with productivity until 3 months later. It gets revised so much that it means nothing to me on first release. I also don't think it says anything about fundamental economic trends. It's really a measure of business cycle IMO. We do seem to be at the end of the last and beginning of the next!

Preliminary productivity for the entire business sector was 1.3. Sure looks recessionary to me, but I don't trust the numbers.

My simulations for the third/fourth quarter intersection have my mouth hanging open on the consumer side. Real buying power for the bottom 65% of population down @5% over the first three quarters? Yikes!!!! If you look just at hourly you ignore insurance increases and benefit cuts, and those are proliferating.

I tell you something else - the problem with state and local public debt and unfunded pensions is going to plop public employees into the same boat with private in a lot of areas this go-round.

Here's the kicker that no one ever looks at - the table that shows YoY changes. Since last year, real non-farm compensation is unchanged, while total business real compensation is down -0.3. Productivity up for the business sector only 0.9, for nonfarm 1.1.

Do you simulate consumer spending by quartile/decile? I've not seen this kind of thing out there, and with the income/wage dynamics I'd think it would be interesting. Maybe you could incorporate it in a future post?

It kills me that so many out there are focused on employment as the be-all-end-all of spending, credit experience, etc. They made the claim that subprime defaults were most correlated with employment until just two months ago. In reality the dynamics of falling real wages and over-burdened consumers are sufficient to cause spending to fall and defaults to rise. I get the feeling that for a lot of black-box econometric models out there, including ones used to determine provisioning levels, this will be a "fat tail" event.

David Pearson
David, I do it by household characteristics and demographics.

Obviously the spending patterns of an area with a population biased toward 20s/30s are going to be different than a population with a bulge in the 50s/60s. Homeownership rates, etc all make a big difference.

For an example, look at this post on Orange County, CA.

I think workable simulations have to use a detailed analysis of a sample of areas with different characteristics. That gives you benchmarks and high and low value ranges. Then you can extrapolate from that.

Crude oil sold off last week but the gasoline costs seem destined to go higher because of refinery problems.
David, I am putting a lot of estimation modules into Java so that I can release them under a GNU license with a web-based front end.

I have been developing and testing financial analysis programs that help both individuals and community banks plug holes in their finances and seek ways to generate income and investment. For individuals, the focus is usually on generating enough savings to prevent having to go into short-term debt for items like car repairs, appliances, etc. There have been some near-spectacular successes.

I am currently working with a bank to develop a pretty comprehensive bootstrap program using all this stuff. I think it is going to work well. They are very community-oriented. They seem excited by the possibilities not just because they are decent people but because they believe that the plan will work to greatly enhance their long-term prospects.

What I'm getting at with the above is that this is not pie-in-the-sky stuff; it seems to work on the ground. But the success rate demonstrated from making marginal changes that accumulate over time has caused me to be acutely sensitive to the nature of regressive economic circumstances. If developing a program for an individual family that allows them to save $10 a week will take them from bankruptcy to prime credit/low debt/homeowner/business owner in 7-8 years, the logical conclusion is that economic and taxation changes that strip quite small proportions of spendable income from households can produce a shocking degree of economic stagnation.

Needs inflation combined with high debt ratios and stagnant real incomes is about as regressive as you can get. So yes, I agree with you that we may be facing a fat-tail event. I can hardly see how we will not unless current conditions shift quite a bit.

What most people don't realize is that prime/conforming delinquency experience is at abnormally low levels. This has to do with refinancing options that allow borrowers to weather long periods of negative household cashflows even before discretionary spending. Take away those refinancing options and the negative cashflows quickly lead to normalized levels of defaults for those products. The same is true of credit card and auto loans.

Thus far the negative cashflow dynamic has only shown itself in subprime. I built a simple model of CA subprime households and determined that the MEDIAN was likely in the red by about $2,000/mo (again, before discretionary).

It may be that the resulting racheting back of alt-a CLTV's and doc requirements may, by itself, be the trigger event that uncovers a much larger population of these negative cashflow households. The sign would be a spike auto and credit card loan delinquencies.

The real question is what proportion of households, compared with previous recessions, are entering the recession with negative cashflows? I think the number is unprecendented, and I wish economists spent more time sampling populations to gauge this effect, and less running backward-looking econometric models that show FICO scores and unemployment as the dominant variables.

David, I can't disagree with a word you wrote, except that you seem to be implying that the problem rests solely in the subprime originations. It is not; inflated home values, a belief that the market would continue to inflate and a sense of invulnerability caused a lot of borrowers to buy multiple houses or homes with no money down. Many of the at-risk transactions do not meet the definition of subprime.

It's not the FICO, it's the DTI combined with CLTV. Alt-A is hardly immune. It's almost impossible to believe that we are seeing defaults in some of these originations as high as they are when they are so recent, interest rates are so low, and the economy is so good.

The killer is the geographic concentrations of the bad loans; that's going to produce even more defaults.
Post a Comment

Links to this post:

Create a Link

<< Home

This page is powered by Blogger. Isn't yours?