Initial Jobless Claims Fall Again
Before reviewing the BLS non-farm payrolls report, let's look at the jobless claims report from Thursday since it is generating a lot of discussion. These discussions draw from the spike in continuing claims versus the decline in initial claims.
It’s worth noting that the initial claims data is from the week of January 3rd and the continuing claims data is from December 28th. However, they still differ if you look at one week old data of initial claims. It’s also notable that the spike in continuing claims was the week of Christmas which makes it less trustworthy.
Specifically, initial claims fell from 223,000 to 214,000 which was 5,000 below estimates. Notice how last week’s reading showed a low number of claims as well. Because the spike in the week after Thanksgiving wasn’t included in the 4 week average, it fell from 233,500 to 224,000.
A 38,000 decline from the December 7th peak was the largest 4 week decline since September 2007. Even though that’s a positive stat, it sounds a bit scary since the last recession started in December 2007.
Takeaway is big declines in initial claims don’t mean a recession can’t come a few months later. Point is a recession hasn’t started in January. It has little to do with how the economy does this April. Bears love to say initial claims have bottomed for the cycle. I agree, but don’t find that to be important.
Just because claims won’t fall below 193,000 again doesn’t mean a recession is coming. There hasn’t been a new bottom since last April, yet the economy hasn’t had a big decline.
Biggest news the bears are harping on is seen in the chart above, continuing claims spiked from 1.728 million to 1.803 million which is the highest reading since April 2018. There was a similar spike after last Thanksgiving last year, but it wasn’t as large as this one. The cause is seasonal unless claims stay high the next few reports.
Supporting this point, non-seasonally adjusted claims are up 3.3% from last year. Supporting the bears’ point, the quarterly average of the yearly growth in continuing claims has increased 4 quarters in a row and is the highest since Q4 2009.
Slightly Disappointing Jobs Report
December jobs report slightly missed estimates instead of slightly beating them as predicted. However, the prediction that manufacturing job creation would be weak was correct. We didn’t expect a large negative revision to November job creation which was fairly accurate.
Specifically, job creation in October was revised lower by 4,000 to 152,000. Job creation in November was revised down 10,000 to 256,000. November report was still strong which means it held up the 3 month average.
Job growth in January was 145,000 which missed estimates for 158,000 and the guess of 170,000. This was the 111th straight month of positive job creation which is the longest streak ever by a long-shot. No one needed to worry about wage growth driving inflation since it fell and missed estimates. 3 month average of job creation was 184,000 which is very strong because of the November report. While it’s not the best to see an estimate miss and negative revisions, neither were large.
12 month average of job creation is 176,000. To be clear, population growth is 104,000. Because people are rejoining the labor force, to keep up with its growth, 157,000 jobs need to be added per month. This report missed that threshold, but the 3 and 12 month averages are above it.
As you can see from the table below, manufacturing lost 12,000 jobs which missed estimates for a 1,000 decline. November’s reading was revised up by 4,000 to 58,000. My inkling that the jobs created because the GM strike ended wouldn’t be revised away was correct.
As you can see, there was notable weakness in professional and business services as it added only 10,000 jobs. This is worrisome since it’s a key cyclical industry that pays workers well. This is widely different from the ADP report which showed the industry added 61,000 jobs. That’s a big source of the 63,000 difference between the two reports.
Trade, transportation, and utilities created 40,000 jobs which is very solid. But it was well below the 78,000 shown in the ADP report which was the 2nd most ever. It's not surprising this industry didn’t have one of its best months ever. That seemed like an anomaly when I reviewed it.
On the positive side, information added 3,000 jobs and leisure and hospitality created 40,000 jobs. They were shown to have lost 14,000 and 21,000 jobs in the ADP report. Government created 6,000 jobs instead of the 8,000 that was expected. Meaning, private sector job creation missed estimates by 2,000 less.
Unemployment rate stayed at 3.5% which makes sense since job creation was slightly above population growth. U6 underemployment rate fell from 6.9% to 6.7% which is a new cycle low. It’s also a new record low going back to 1994 as the previous record low was 6.8% in October 2000.
Accuracy of Manufacturing PMIs
Since the ISM manufacturing PMI has been much lower than the Markit PMI, their relative accuracy has become a hot topic. BLS report gives us additional information on this. The chart below shows the employment readings of both reports along with BLS job creation in the manufacturing industry.
As you can see, the Markit employment index has a 91% correlation with the BLS reading, while the ISM index only has a 78% correlation. Also, the ISM reports were too optimistic in 2018.
Conclusion
Jobless claims were low, but continuing claims spiked in the last week of December. I think the latter was caused by seasonality. December jobs report was light, but still signaled the labor market is in fine shape. Both the 3 month and 12 month averages are above the number of people added to the labor force.
Manufacturing job creation was weak. Markit manufacturing report is more accurate which is good news because it has recently been much more positive than the ISM reading.
Recent Comments