Strong Jobless Claims

The jobless claims report was strong again as claims fell from 219,000 to 213,000. Quite frankly, if you asked economists 3 years ago if such a report was possible, most would have said “no”. This cycle has seen record lows in the ratio between the labor force and the jobless claims. The August claims are a few thousand above the cycle low in July. I think comparing the trough in this cycle to current claims is similar to looking at the recent changes to the Q3 earnings estimates.

To be clear, both stats were amazing, which means it’s tough to improve upon them. Earnings growth estimates usually fall heading into the quarter, but they have been rising this year. A slight increase in the amazing Q3 estimates is worth applauding. The trough in the jobless claims can be considered a recession indicator. However, if the claims stay near where they are now, there doesn’t need to be a new low for me to discount the possibility of a recession.

Recession Timing Indicators

The point about using the trough in the jobless claims as a recession indicator was made because the table below looks at the troughs in a few indicators to determine the current recession risk. The S&P 500 and unemployment claims are in the leading indicator series, so there is some overlap in those metrics.

(Click on image to enlarge)

Jobless Claims

First, let’s look at the unemployment claims. To be clear, the data used in this table is the 4-week moving average. As you can see, the 4-week moving average of the unemployment claims bottoms an average of 12 months before recessions when you look at the past 7 cycles. The last bottom was May 12th which was 3 months ago. However, don’t think that the economy is about to go into a recession in 9 months.

The 4-week average was 213,500 on May 12th and is 214,250 now. It’s spurious to worry about a recession based on a 750 difference in the 4- week average unemployment claims. While quantitative data is important, let’s not lose sight of reality. A 750 difference is like the difference between a 420 foot home run and a 418.5 foot home run. It is immaterial. If the past data is revised in the next report, this could be a new low. Also, another strong report next week could cause there to be a new low.

The jobless claims peaked after the stock market in 4 of the past 7 cycles. This makes sense because we’re looking at a 4-week average which includes 3 weeks of old data. I would be interested to see how the week by week reports forecast stock market returns. The biggest problem with using that is it gives off even more false warning signs than the 4-week moving average.

S&P 500

The next indicator is the S&P 500. Since the S&P 500 topped in late January, the top is about 7.5 months ago. If the average is correct, the economy should already be in a recession. However, this table doesn’t mention the plethora of times the stock market stayed below its record high for a few months without a recession following soon afterward. There have been many false positive warning signs. This indicator is just like the jobless claims in that the market is very close to the previous peak. The S&P 500 is within 1% of the January record. Calling for a recession based on the stock market being less than 1% away from its record is spurious.

Yield Curve

The other indicators don’t show a recession is coming anytime soon. The yield curve inverts an average of 19 months before recessions. The curve hasn’t inverted yet as the difference between the 10-year yield and the 2-year yield is 28 basis points. Don’t fall into the trap that states the yield curve will definitely invert after 2 more hikes. It’s possible, but not a guarantee as the next hike is already priced in. It’s not as if the curve flattens 25 basis points after a hike. To be clear, I am calling for an inversion by the end of the year. I’m making the point that this is far from a lock. If the curve inverts in December, the next recession will start in June 2020.

New Home Sales

New home sales peaked in November 2017. Since the latest data is from June, it has been 7 months since the peak. The average recession occurs 32 months from the peak which means a recession could be coming in 24 months. It’s difficult to predict where the economy will be that far in advance. It potentially lines up with the yield curve’s prediction if it inverts in the next few months. The housing market’s weakness is starting to get disconcerting. I expect the November 2017 peak will hold for at least the next 2 months because the latest data hasn’t been strong.

The MBA mortgage applications index fell 3% week-over-week which was below the 2.6% decline in the previous week. The purchase index was down 2% and the refinance index was down 5%. It’s not surprising to see the refinance index fall since mortgage rates have been increasing. The 30-year mortgage interest rate was 4.59% as of August 9th. This is near the 5 year high which means it doesn’t make sense for anyone who bought a house in the past 5 years to refinance if they’re only looking to lower their interest rate. You’d need to go back to 2011 to find higher rates than the recent peak.

Leading Indicators

The leading indicators peaked in the last report in June, so there are no serious worries about a recession warning being triggered by this index. The leading indicators peak 12 months before recessions, meaning the earliest a recession can occur is the 2nd half of 2019 if the average holds. The stock market has peaked before the leading indicators in 3 of the past 7 business cycles.