Industrial production is the key manufacturing indicator

17 Mar 2011 by Jim Fickett.

The growth trend in manufacturing continues. The key manufacturing indicator to watch is industrial production. Factory orders are commonly thought to lead production, but do not. The ISM PMI is very noisy, but it is a useful leading indicator to consult in ambiguous situations.

Today the Federal Reserve released the Index of Industrial Production through February. There is considerable month-to-month noise, but the overall trend still seems to be healthy growth:

In the continual quest to concentrate attention where it most aids investing, let's look in a little more detail at measures of manufacturing.

When I buy a stock there are many considerations that go into the decision. These include many questions specific to the company, such as whether the company is over- or undervalued, and whether management is adding real value. Other considerations are broader, such as whether current trends benefit the relevant sector, and whether any overseas markets are growing. I also consider the overall market cycle, and apply much more stringent criteria to a purchase when (as now) I think the market is primed for a correction, than when (like after the fall of Lehman) almost everything is undervalued.

Stock market cycles are to some extent just a manifestation of mob psychology, unconnected to fundamentals, but they are at least correlated with the business cycle, which is closely tied to the most volatile sectors, manufacturing and construction. So the importance of following the manufacturing sector is one piece of conventional wisdom that I completely agree with.

The three most commonly watched indicators for the health of the manufacturing sector are

1. The Index of Industrial Production (and the closely parallel manufacturing sub-index)
2. The Institute for Supply Mangement's manufacturing index, the PMI
3. Factory orders, usually thought to be a leading indicator

Is it really necessary to watch all three? Or is this one more case of an excuse for noisy headlines and itchy trading fingers? Here is a plot of all three together:

First consider the value of factory orders as a leading indicator. (Note that I am looking at the data from the full report. The advance report for durable goods orders comes out one week earlier and is revised more heavily.)

• In late 1997 and early 1998 factory orders dropped precipitously; this was a false alarm
• In June of 2000 factory orders peaked and then began a long decline; this was a real turning point, but it is unclear how much information was added – industrial production hit a double peak in June and September of 2000 and then declined
• In January of 2002 factory orders bottomed; industrial production bottomed in November 2001 – in this case factory orders were lagging, not leading
• In December of 2007 and June of 2008 factory orders hit a double peak; industrial production hit a high plateau from July to December of 2007 and then fell – here again it would appear that factory orders lagged rather than led
• Factory orders bottomed in March of 2009, while industrial production bottomed in June 2009 – here factory orders did lead

At turning points one always wants every bit of data available. But the record for factory orders as a leading indicator is not encouraging – usually little or no information is added.

Probably everyone believes that factory orders lead production because the physical act of ordering obviously occurs first. However orders do get changed or canceled, and factories, like wholesalers and retailers, also manage a changing inventory. So perhaps it is not so surprising that factory orders do not really lead.

How about the PMI? The PMI is meant to be a directional indicator – one expects it to be above 50 when the production is growing, and below 50 when production is shrinking. As one can see quite clearly above, this is often true, but also often false. The PMI does offer, if you are willing to accept the noise, some value as a leading indicator. In the above graph most peaks/troughs in industrial production are preceded by peaks/troughs in the PMI. In particular, most recently

• The PMI peaked in January 2004 and decayed gradually over the next few years; this was one of many indicators that the boom of 2004-7 was very superficial
• The PMI hit bottom in December 2008, the stock market bottomed in March of 2009, and industrial production bottomed in June of 2009; here the PMI might have been some help in deciding that there was not too much danger in grabbing the low stock market valuations of early 2009.

My conclusions are:

• Manufacturing is key to the business cycle and worth tracking
• Actual production is the most fundamental measure, and the best way overall to track the sector
• Factory orders add little information, and time is better spent elsewhere
• The PMI is noisy, but is worth looking at from time to time for additional insight into ambiguous situations