MacroScope

Jobless claims fell unexpectedly last week to 361,000. Analysts were particularly heartened by the improvement because the latest figures were finally “clean” of recent seasonal adjustment quirks related to auto factory shutdowns. That’s the good news.

Some lingering cause for hesitation: Eric Green at TD Securities reminds us that recent dips in claims have not necessarily translated into great bursts of new job creation.

Over past periods of this recovery claims at this level have been consistent with (monthly) job growth closer to 200,000. With claims back at these levels, one cannot presume that this will continue to hold given the level of uncertainty and slower growth momentum from which labor demand will lag.

The next employment report is not accelerating toward 200,000. A slowdown toward 120,000 is far more likely given the recent readings on PMI’s, Monster (online) advertising, and a sectoral breakdown which favors less job growth in leisure, business services, and information technology.

The economy created 163,000 jobs in July, much stronger than analysts had forecast. But the jobless rate climbed to 8.3 percent from 8.2 percent.

U.S. non-farm payrolls have come in below the Reuters Poll consensus for the past four months, the longest streak since an eight-month period in 2008-09 when the U.S. was in the depths of recession and, at one point, losing more than half a million jobs a month.

Compared with a few years ago when there was a very wild range of forecasts on a given jobs report — the widest spread polled since the financial crisis began was 575,000 for the May 2010 data — economists are now huddling together in a pessimistic pack.

For the July data, due out at 1230 GMT, the range of forecasts in the Reuters Poll (on a consensus of 100,000) has narrowed to a 107,000 spread between highest and lowest, compared with 132,000 for the June data.

Weekly data on applications for unemployment benefits have gained renewed importance since a weak March payrolls number left economists wondering whether a tentative labor market recovery was about to cave again. The last two weeks’ readings were just soft enough to leave investors thinking the country’s unemployment crisis may not be healing very quickly.

Daniel Silver at JP Morgan has dug deeper into the claims figures and found a curious trend: a repeated and distinctive tendency toward upward revisions in the numbers.

There has not been a downward revision to the initial claims data reported for the prior week since the start of March 2011, and this recent streak is not a new phenomenon—there have been upward revisions in about 90% of the weekly reports since the start of 2008, as well as going back even further to the start of 2000. These revisions are relatively minor (usually adding only a few thousand claims) and do not change the broader trends in the data, but they can lead to the weekly claims reports showing decreases to the more recent levels, whereas if the prior week had been unrevised, the reports would have shown increases in claims.

A look at the macroeconomic news and its impact on the mood of investors and the direction of the economy. Are we heading for a double-dip recession?

Jobless claims and trade data came in better than expected prompting some investor cheer today.

“We were expecting that things would slow down in the third quarter and start to pick up in the fourth quarter, but now it seems like the slowdown in the third quarter wasn’t as severe as we feared,” said David Sloan, an economist at 4CAST in New York.

In the last comparable recession, which we know wasn’t anywhere near as deep as the Great Recession just endured, U.S. jobless claims peaked at 695,000 in October 1982.

Weekly initial unemployment claims is an extremely reliable leading economic indicator because the figure is not derived from a survey. It’s an actual tally of real people without a job who are queuing up for the dole.

By the end of the following year, about 14 months later, weekly initial unemployment insurance claims had plunged by more than 300,000 to 372,000. They dipped even further to 333,000 in January 1984.

Economists seems to be having a difficult time sorting through the recent downward trend in jobless claims. On the surface, the news looks good. Benefit applications, which have been trending lower from a 26-year peak of 674k in March, fell to a nine-month low of 521k this week. Continuing claims also eased, and are now hovering just above 6 million.

On the one hand, this suggests the October employment survey could be a bit better following a September disappointment, particularly considering this is survey week for payrolls. At the same time, there are continued reasons to worry. JP Morgan’s Abiel Reinhart had this to say:

Claims are at their lowest level since the start of the year, although the level is obviously still high. To date, the four-week average of jobless claims is down 24,500 from the September payroll survey week, which is good news on the employment front. The downward trend in claims does suggest that payroll losses could moderate again in October. The insured unemployment rate fell to 4.5% from 4.6%. The fall in insured unemployment suggests that overall unemployment is approaching its peak.