– by a New Deal Democrat
As usual, the first data for the new month is for manufacturing and construction. Here’s a look at each.
The ISM manufacturing index and especially its new orders sub-index is an important short leading indicator for the manufacturing sector. In August, after two months of slight contraction, the leading new orders sub-index improved to 51.3, indicating expansion. The general index also continued to show expansion, reporting 52.8 for the second consecutive month:
This index has a very long and reliable history. Going back almost 75 years, the new orders index always fell below 50 within 6 months before a recession, and in three cases didn’t actually cross the line until the first month of the recession itself – even though the recession didn’t start until after the general index fell below 50, but actually usually under 48.
This stamp means we’re not out of the woods, but on the other hand, we’re not in a recession now.
As for construction for July, the report shows a a nominal decline of -0.4%, although June’s initial decline of 1.1% was revised down to just -0.5%. The leading housing sector fell -1.5%, although June’s initial decline of -1.6% was revised down to -1.1%:
Adjusting for changes in construction materials prices, which fell -0.6% for the month, “real” construction spending actually rose +0.1% month-on-month, and housing spending fell -0.9 % on a monthly basis. Here’s what “real” construction costs have looked like over the past few years:
The decline in housebuilding spending, while significant, was smaller than the decline in 2018-19 and nowhere near the -40.1% decline experienced before the end of 2007.
While housing spending lags behind other housing data such as permits, starts and sales, it has the advantage of being much less noisy. Over a period of several months, this is an almost pure signal. And what that tells us is — which is not surprising at this point — the decline in housing is real and is likely to be reflected as a negative component of GDP for the third quarter.