Initial Claims Illustrate Fed Dependency on Undependable, Ignorable Data

The headline, fictional, seasonally adjusted (SA) number of initial unemployment claims for last week came in at 271,000. The Wall Street economist crowd consensus guess was right on the money. That happens on occasion.

Instead of the seasonally manipulated headline number expectations game, we focus on the actual trend of the actual data. Facts and reality are much more useful than the Wall Street captured media’s fantasy numbers. Actual claims were 263,221, which is another record low for this calendar week, continuing a nearly uninterrupted string of record lows that began in September 2013.

Employers in some sectors are hoarding workers. Similar behavior in the past has been associated with bubbles, and has led to massive retrenchment, usually within 18 months or so. In the housing bubble, similar behavior continued well beyond the peak of that bubble in 2005-06. Employers seem to take their cues from stock prices. The current string is now 3 months beyond the point at which other major bubbles have begun to deflate. Is the bungee cord simply longer this time, or is this the new paradigm?

The Department of Labor (DoL) reports the unmanipulated numbers that state unemployment offices actually count and report to the DoL each week. This week it said, “The advance number of actual initial claims under state programs, unadjusted, totaled 263,221 in the week ending June 20, an increase of 4,457 (or 1.7 percent) from the previous week. The seasonal factors had expected an increase of 1,607 (or 0.6 percent) from the previous week. There were 305,029 initial claims in the comparable week in 2014”

Initial Claims and Annual Rate of Change- Click to enlarge

Click to view chart

This week of June is a swing week, with claims sometimes up, sometimes down, with a wide variance. There’s no evidence of seasonality. The actual change this week was an increase of 5,000 (rounded). That compared with an increase of -4,000 for that week last year and the 10 year average for that week of a decrease of -5,000 (rounded).

Week to week changes are noisy. The trend is what’s important and it remains on track. Actual claims were 13.7% lower than the same week a year ago. Since 2010 the annual change rate has mostly fluctuated between -5% and -15%. This week’s data was again on the strong side of that range. That unusual degree of strength has now been persistent for 2 months. There’s no sign of an uptick in the trend of firings and layoffs.

There were 1,861 claims per million of nonfarm payroll employees in the current week. This was a record low for that week of June, well below the 2007 previous record of 2,147. The 2007 extreme occurred just a few months before the carnage of mass layoffs that was to begin later that year. Employers were still clueless that the bubble had ended and that that would have devastating effects.

Because employers apparently tend to take their cues from stock prices, we cannot depend on this data for advance warning of a decline in stock prices, although there should at least be concurrent confirmation.

Initial Claims and Annual Rate of Change- Click to enlarge

Click to view chart.

I look at an analysis of individual state claims as a kind of advance decline line for confirmation of the trend in the total numbers. The impact of the oil price collapse started to show up in state claims data in the November-January period. While most states show the level of initial claims well below the levels of a year ago, in the oil producing states of Texas, North Dakota, Louisiana, and Oklahoma, claims have been consistently above year ago levels since the turn of the year. North Dakota and Louisiana claims first increased above the year ago level in November of last year. Texas reversed in late January. Oklahoma joined the wake shortly after that.

Data for the June 13 week:


These numbers have varied widely, week to week but the trend of claims being significantly higher than the same week last year has been persistent. Texas, with a huge and somewhat more diversified economy has improved since April as the price of oil rebounded and stabilized, and the state showed a year to year decline in the week ended May 30. But the following week Texas again had more claims than last year.

In the June 13 week, 11 states had more claims than in the same week in 2014. That was down from 12 the prior week, but up from 10, 4 weeks earlier. This number fluctuates widely week to week with many states near even. At the end of the third quarter of 2014 just 5 states showed an increase in claims year to year. At the end of 2014 that had increased to 8. In early April this year the number had risen to 22. So there has been some moderation in this trend as the oil collapse has leveled off.

The 22 states that were higher in early April gives us a benchmark to watch, similar to an advance decline line in the stock market. If the number of states showing a year to year increase in claims should exceed 22, it should be an indication that the national trend of decreasing claims is reversing. That could be an advance warning of a big stock market decline as well.

I track the daily real time Federal Withholding Tax data in the Wall Street Examiner Professional Edition. The year to year growth rate in withholding taxes in real time is now running +5.8% in nominal terms. The growth rate has been remarkably consistent around 6% over the past couple of months.

The June 12 week was the reference week for the May payrolls survey. The numbers for that week showed a year to year gain of +6.4%, which compares with a year to year gain of +5.4% in the May reference week. The daily data suggests that withheld taxes for that week did not grow as much, but it would still mean that jobs growth was about the same in June as in May. Whether the cockamamie seasonally adjusted headline number reflects that reality or not is a crapshoot. It takes the BLS 7 revisions of the SA data over 5 years to fit it to the actual trend. The first release is hit or miss.

Assuming that the SA headline number is a reasonable representation of the actual trend, it would move the Fed closer to attempting a rate increase, especially now that CPI data has begun to catch up to reality.

The Fed’s favored measure of inflation, PCE, just released, suppresses the measurement of inflation even more than CPI. That would put the Fed even further behind the curve in recognizing that inflation is running much hotter than the official measures show. The Fed knows that, and has inserted weasel words into its various propaganda releases that it will raise rates as long as the Fed thinks that inflation is moving toward the 2% target. It does not actually need to be at the target. The Fed is prepared to ignore the official measures because the members realize that they’re bogus.

The Fed will use or ignore whatever stats it wants depending on whether they fit its preconceived narrative, which is “We’re gonna try to raise rates at least once this year, and if that doesn’t work, we’ll think of an excuse not to do it again, because raising rates is really data dependant depending on which data dependably supports our narrative, and which data we will ignore, because it all depends on dependably dependant official data, none of which is dependable.”

The actual claims data, and actual withholding data, show the financial engineering bubble economy is still at full boil. This will continue to encourage the Fed to engage in the charade of pretending to raise interest rates sooner rather than later, but only because they have conditioned the market to expect it, a conditioning that they now regret they had undertaken. So now the Fed is saying, “just once and then we’ll see.”  They’re walking back expectations now because they know they will have problems getting rates to go up. I cover that subject in depth in your weekly Money-Liquidity Pro reports.