Survey Says?

by Brian Pretti CFA, Contrary Investor. May 21, 2010

I will not belabor the point here. Over the recent past we have seen some very strong manufacturing surveys. Of course the headline ISM survey has been strong for some time now, but a number of the smaller regional surveys are catching up, and fast. Most regional manufacturing surveys now rest at multi-year highs. Moreover, the Richmond Fed manufacturing survey just hit an all time new high in the recent month. Clearly manufacturing is very strong at the moment. Isn't it? This is exactly what the ISM number reflected this week, right?

First, we importantly need to remember that most of the manufacturing surveys we see are diffusion indices. They measure a preponderance of either positive or negative responses in these surveys. But virtually none of these measure absolute levels of nominal economic activity, and as such we need to be very careful about historical comparatives. One of the key indicators a lot of analysts look at in China is electric utility sales/usage to corroborate, or otherwise, many a headline manufacturing number that is reported. Should we be doing the same domestically? We believe we should, and that's what lies below. As you'll see, utility stats do not portray the same picture as do the headline manufacturing surveys. Again, and as always, neither good nor bad. As investors we simply need to make sure we are in touch with the fundamental reality of the moment.

First up and very simplistic is a quick look at the history of utility industry capacity utilization. Without question there is a certain seasonality to these numbers. Additionally, at least over the last few years, utilities have actually been adding to capacity, effectively depressing an already depressed utilization number relative to historical context. But at least as of the April 2010 reading, we just are not that far off of three decade lows. After a spike up in late 2009 and early this year, utilization dropped quite meaningfully over the last few months. Before suggesting the manufacturing revival is over, we again need to address calendar seasonality with respect to maintenance related shutdowns. We've looked back through the years and have seen sporadic seasonality, not necessarily the same pattern we are seeing now. So we need to see what the months ahead bring us, but we suggest watching these numbers for an anecdotal read on the reality of a US manufacturing recovery. Unless manufacturers have found a way to increase production without the need to increase usage of electricity and gas, we believe this is important data to review alongside the very strong survey results of the moment.

The reason I’m bringing this issue up right now is in response to a number of company specific earnings reports that have come our way as of late. I was particularly struck by recent comments just a month back by the Chairman and CEO of American Electric Power. This is no backwater utility and has operations in States with a decent amount of manufacturing. Listen in:

“We had strong results for the first quarter because of successful rate proceedings, increased off-system sales and favorable weather,” said Michael G. Morris, AEP chairman, president and chief executive officer. “For several quarters we’ve been seeing some signs of economic improvement in the states that we serve, but the signs have yet to be followed by a broad increase in electricity demand that typically accompanies economic improvement. Residential electricity sales have improved year on year. Industrial sales are down slightly from first quarter last year, but we are no longer seeing the steep declines experienced throughout 2009. Commercial sales continue to struggle.

“In light of the slow recovery of electricity demand and depressed wholesale prices, we are taking further meaningful steps to reduce our costs,” Morris said. “These steps include non-labor and labor reductions such as an offer of a voluntary severance package.

What? Personally, I thought residential would be flat on its back to down relative to the first quarter of 2009 given the continued increase in foreclosure activity across the US. And this was the only segment of AEP's business that was up year over year? Moreover, with the touted macro manufacturing recovery Stateside, just how the heck could industrial sales be down year over year? Especially set against the comparison of the Armageddon first quarter of 2009? In March of 2009 the ISM survey rested at 36.3. Today it's quite the different story as its reaching for the starts, no? And AEP has felt none of this? The following table documents electric utility capacity utilization nine months after each recession over the last four decades. Of course the implicit assumption is that the latest recession ended in June of last year, as per the rhythm of reported GDP.

Nine Months After Each Recession End	Elec. Utility Capacity Utilization
August 1971	93.6%
December 1975	84.5
April 1981	84.9
August 1983	85.5
December 1991	88.0
August 2002	88.1
March 2010	78.7

I’m scratching my head and simply raising my hand to point out what is a clear dichotomy, it seems, between current perceptions of a manufacturing recovery and what may be a bit different reality in the industrial sector via the data point anecdote that is capacity utilization for electric utilities. The CEO of AEP is directly telling us above that he sees an atypical recovery as per the rhythm of utility sales. Lastly, if the folks at AEP believed industrial demand was indeed about to recover over the near to intermediate term, then why would they be looking to take "meaningful" steps to further reduce costs? Again, I’m just asking the question and pointing out the clear divergence between headline manufacturing surveys and the tone of electricity sales to these very same manufacturers as we'll see below.

Again, as stated, capacity utilization numbers for the industry do have some seasonality and are influenced by maintenance activities that can take capacity off line for a time. So let's take this just one step further and check in with our friends at the Department of Energy. The data below is electricity sales/usage for the industrial sector (as measured in kilowatt hours) going back close to four decades. The chart speaks for itself. Yes, there is calendar based seasonality here as is plainly visible in the non-seasonally adjusted data used to generate the graph. Although not universally true in every single year, trough periods usually are seen in the first quarter of each year. The data in the table below is through January of 2010 (the latest available). But calendar seasonality or otherwise, you can see that recent levels of usage are lower than any trough recession period dating all the way back to the mid-to-late 1980's at least. Again, in a very simplistic sense, does this square with the apparent strength we are witnessing in manufacturing surveys of the last three to six months?

As with utilization data, let's repeat the exercise of looking nine months out after and official recession and calculate the growth in the use of electricity by the manufacturing sector.

Nine Months After Each Recession End	Growth In Electricity Usage By The Industrial Sector Nine Months After An Official Recession Conclusion
December 1975	5.3%
April 1981	4.0
August 1983	14.8
December 1991	2.5
August 2002	10.5
March 2010	0.2

We've never seen this slow a pick up in usage in a post recessionary period, not even in what was ultimately to be the double dip environment of the early 1980's.

Clearly an additional factor in the current cycle is the continued off shoring of US manufacturing that has been an ongoing trend for quite some time now, really accelerating over the last decade. So what that means to us is that the surveys we see today, and this really goes for many industries in addition to industrial/manufacturing, do have a certain survivorship bias. In other words, companies that have gone out of business are no longer being reported in the current surveys. This speaks to a certain non-comparability of historical data to the present, but it's really always this way in every cycle to be honest. The Darwinian survivors are the strongest of manufacturing companies and certainly push survey results higher with any type of recovery as business flows to a smaller group of surviving companies. We're not debating a manufacturing recovery of the moment. It's clear that is occurring. The question is magnitude and sustainability relative to prior cycles. All the data above tell us is that as per electric and gas utility input in the manufacturing process, we're still well off of prior highs in actual industrial usage and utility industry capacity utilization, and in many senses much nearer multi-decade lows. We'll be watching this data ahead and briefly updating as the inventory rebuild winds down in the quarters ahead, as was exactly the case with the inventory contribution for 1Q 2010.

Before leaving the wonderful world of the utility sector a few final comments regarding the equities themselves. Although I personally continue to believe the electric utility industry may have a brighter longer term future than not as potentially being a big part of the total US energy solution (electric cars for starters?), near term headwinds remain intact. The data above tell us as much. Growth in usage/sales is going to be very hard to come by on all fronts - commercial, industrial and residential. And if the rate of change in quarterly GDP growth slows as we move into the end of this year, which I expect, the road will become a bit more bumpy. The following chart is a three decade view of the relative performance of the utility sector set against the S&P, as representing the broad equity market itself. Utilities were a pretty dramatic price under performer throughout the secular equity bull of 1980-2000. Importantly, though, what is not picked up below is the contribution of utility dividends. In truth, that is also missing from S&P return as we are only looking at price. It should not be new news to anyone that utilities are considered defensive in character. They have been big time under performers off of the March 2009 equity market lows, as is clear in the chart below. But in terms of rhythmic trend, utilities have been a price out performer since the market highs of early 2000. And this is despite the Enron blow up (yes, Enron was in the utility index at one time) and the issue of merchant energy overbuilding earlier in the prior decade that crushed many a utility stock for a time.

You know and I know that one of the key macro questions for investors right now is whether we are still working within the environment of a secular equity bear market. The jury remains out. And, of course, the extraordinary hyper charged liquidity environment of the moment only confuses the issue all the more as price is reflective of both this liquidity needing to find an expression or outlet, and ongoing fundamentals of the industry and companies specifically. Can the chart above help guide us in deciphering this secular bull/bear issue as we move forward? We hope so, but we'll simply have to see how it all goes. The utility industry is clearly "telling us" as per the data we have reviewed that excess capacity is an issue. Not a good thing for forward pricing. The carbon issue looms large if indeed the Administration pushes hard on a green agenda. At some point rising interest rates will be competition for utility dividend yields. And as we all know, dividend related taxation change lies dead ahead of us. As you can see above, the relative performance "channel" for utilities has been pretty darn well defined since 2000. We're at the bottom of that channel right now. A break to the downside in terms of relative performance will be a big negative, so keep an eye out here. If indeed we bottom in terms of relative performance and turn higher, that will be saying something about potential change in the risk appetite of investors broadly. Time to monitor a bit more closely near term.

Brian Pretti

© 2010 Brian Pretti

Contact Information

Brian Pretti CFA | Editor and Publisher, Contrary Investor
P. O. Box 4402 | Walnut Creek, CA 94596 | Observations | E-mail