An Addendum to Recent Data Points – The Message from Lumber
We have recently discussed the manufacturing sector’s problems several times, as well as the trends in employment data. As a general remark to this, it seems that the sectors that have lately contributed especially strongly to employment growth were health care & social services, leisure & hospitality and construction.
The echo boom in construction may be running out of steam though. Interestingly, this is once again happening even before a rate hike cycle has begun – in a sector that traditionally benefits disproportionately from low interest rates. Data on housing starts, permits and construction are admittedly quite erratic from month to month. However, there are a number of signs that the sector may be about to cool down. Said erratic data on starts and permits have just nosedived again from what remains a historically low base.
Lumber futures prices (weekly) over the past five years – click to enlarge.
A friend has also pointed us to a recent article by Tom McClellan, which discusses the message from the decline in lumber prices. We were trading bond futures and options in the 1980s, and back then everybody in the trading community seemed to be well aware of the connection between lumber prices, interest rates and the economy. However, as Mr. McClellan notes, the Federal Reserve may actually not be aware of lumber’s strong record as a leading indicator. He shows the correlation by means of a chart aligning lumber prices with housing starts, with lumber prices set forward by 10 months:
Lumber prices set forward by 10 months vs. housing starts.
Incidentally, lumber prices are also a long leading indicator of initial unemployment claims, although it seems possible that the historically still low overall level of construction employment may make this correlation less pronounced this time around. We have little doubt though that lumber prices remain a relevant leading indicator for the home building industry.
Employment and Manufacturing Activity
In addition, we want to show two chart updates sent to us by our friend Michael Pollaro, which compare the rate of change of employment survey data with business sales and inventories. The caveat here is that payrolls data are eventually going to be adjusted out of recognition, but the current situation is still interesting – once again from the perspective that business sales tend to be a leading indicator of employment.
Annual rate of change of business sales, vs. annual rate of change of employment according to the establishment and household surveys – click to enlarge.
Not surprisingly, a large gap has opened up between the two series recently. This is not least so because manufacturing, while representing the largest share of gross economic output, only provides about 13% of total employment. It takes a while from time the economy’s foundation begins to crumble to its effects spreading to other (mainly service) industries and the labor market.
The next chart shows the sharp rise in the business inventories-to-sales ratio compared to the growth rate of two employment surveys.
Inventories to sales ratio vs. annualized growth in total employment according to the establishment and household surveys – click to enlarge.
As you can see, business sales have a clear tendency to lead employment, while the inventories-to-sales ratio usually exhibits a slightly smaller lead time and begins to spike once a recession actually gets underway. It is interesting that it has risen so sharply already, well before an official recession seems in train or even imminent (although one could probably argue that manufacturing is already in a recession, while the mining and oil sectors are of course in a severe recession by now).
Conclusion
The diverging trends between the performance of manufacturing sector on the one hand and services and construction sectors on the other hand are unlikely to last. Once the sectors of the economy that have recently contributed the most to employment growth begin to struggle as well, we should begin to see an increase in initial claims and payrolls data should begin to weaken.
From the perspective of stock market investors the trend in initial claims is quite important, as small divergences between claims and stock prices tend to occur close to turning points and the two data series exhibit an empirically very well established negative correlation.
Initial claims (inverted) vs. the S&P 500 Index. Turning points in the two series as well as their trends are very closely aligned, with small leads and lags evident near the turning points – click to enlarge.
Charts by: BarCharts, Tom McClellan, Michael Pollaro, StockCharts
