London, 9 April 2010

Substantial resource slack will continue to restrain price rises and ensure U.S. inflation remains subdued for some time, according to the rate-setting Federal Open Market Committee (FOMC).

Janet Yellen, president of the Federal Reserve Bank of San Francisco, and widely tipped as the front-runner to be next Fed vice-chairman, has been the most vocal advocate of this idea. She argues huge amounts of spare capacity and the big difference between current output and the economy's full, non-inflationary potential (the "output gap") will ensure inflation remains low for a long time.

Yellen leads a wing of the Fed more worried about renewed weakness in the housing market, the greater than expected slowdown in "core" inflation, and the gradual wind down of monetary and fiscal stimulus, fearing it could tip the economy back into recession, or at least a prolonged period of sub-par growth.

Backing this view, researchers at the San Francisco and New York Feds recently published an essay showing inflation has continued to decline across a broad range of goods and services, even when the drag from housing is excluded - see www.frbsf.org/publications/economics/letter/2010/el2010-11.html

The San Francisco and New York Fed researchers argue underlying inflation is decelerating in line with the rise in unemployment, while the large amount of cyclical slack continues to deprive most businesses of pricing power.

Inflation may actually have slowed more than usual, notwithstanding the sharp rise in oil and metals prices. The last FOMC minutes record "Participants saw recent inflation readings as suggesting a slightly greater deceleration in consumer prices than had been expected".

Cyclical vs Structural

How far surplus capacity in the economy, and associated unemployment, restrain price increases in the early and mid stages of the expansion will depend on a number of factors, and is surrounded by a significant degree of uncertainty.

One unresolved question is how much of the apparent "slack" is cyclical (caused by lack of demand, which should be brought back into use as the economy improves) and how much is structural (capacity that has been permanently idled and will never return, as the industries have moved offshore to China and other parts of Asia).

While cyclical slack should help restrain pricing power and dampen inflation until it has been fully re-absorbed, structural slack is not really slack at all, representing capacity that has in fact disappeared forever, and should really be subtracted from measured capacity. The question of how much capacity has been lost is crucial. While the Fed does adjust its figures to take account of capacity that has become obsolete, the rate of obsolescence is impossible to observe directly and must be estimated.

There are reasons to think structural capacity losses might have accelerated in the last five years. Part of this is the accelerated trend towards offshoring more production to Asia. Part of it may reflect capacity that has been rendered uneconomic by the steep rise in energy prices. It is possible (even likely) published measures overstate true manufacturing capacity, and therefore the real degree of cyclical slack. The Fed has cut its estimates for manufacturing capacity slightly over the last fourteen months, but the adjustment may not be enough (Chart 0 in the attachment at the bottom of this piece).

Uneven Distribution

The other issue is where precisely the slack is located. Slack is not evenly distributed. Utilisation rates vary quite widely between different industries and different industrial sectors. The attached charts show actual utilisation rates compared with their long-term averages for industries at three separate stages along the production chain:

* Crude processing industries are resource extractors (miners, oil and gas producers, loggers) or industries that perform basic transformations (pulp, fertilisers, chemicals).

* Primary and semifinished processors include makers of steel, glass, cement, refined petroleum products, refined metals, yarn and electronic components.

* Finished processors make the finished items used by households (food, drinks, vehicles, clothes) and businesses (agricultural and industrial machinery, aircraft, railroad cars etc).

The full categorisation is attached (SOP Classification) together with the share of each industry in total manufacturing value-added. Chart 1 shows utilisation rates over the very long term (1972-2010) while chart 2 shows rates in more detail over the last two full cycles (starting in 1990).

The first point to note is that slack varies sharply between different stages of processing. Utilisation rates plunged at all stages between September 2008 and the middle of 2009. But the deterioration was much worse for primary and semifinished processors (what I am going to term the midstream) than for crude processors (upstream) and makers of final consumer and business items (downstream). Upstream processors have subsequently witnessed a much faster improvement in their fortunes than businesses at other processing stages.

As a result, there is twice as much slack in midstream primary and secondary processing (where utilisation rates are 12 percentage points below the average) as in downstream finished processing (where utilisation is only off 6 points) let alone crude processing (where utilisation is almost back to the long-run mean).

Structural Problem

Part of the explanation for varying utilisation rates is that capacity has grown far more slowly in downstream crude processing than in the other sectors over the last 40 years as investment in domestic extractive industries has withered while productivity in other sectors has soared. With much less capacity growth in the good years there is now much less slack in the downturn.

But long-term trends cannot account for all the recent variation in utilisation rates. Until the end of 2005, utilisation rates at all three processing stages were similar and broadly synchronised.

In fact, utilisation rates in the midstream have been falling since January 2006, well before rates turned down in the finished goods category (September 2007) or in crude processing (August 2008). Problems afflicting midstream processors started well before the recession and are only partly cyclical in origin.

Whether globalisation, offshoring, China, yuan undervaluation, or another factor is the cause, there is clearly a big structural component, which will not be fully cured by low interest rates, however long they are sustained.

If yuan undervaluation is the root cause of structural problems for midstream U.S. manufacturers (something that is not clear) an adjustment in the bilateral exchange rate might halt or even reverse some of the structural loss of demand. But it is far from certain. Capacity stranded by changing production patterns may not become unstranded just because the exchange rate alters more favourably. Some losses may be more or less irreversible.

More important are the lessons for monetary policymakers and investors. Huge amounts of apparent slack in midstream manufacturing companies and utilities might in fact reflect stranded investments that will exert much less downward pressure on prices going forward, and be masking smaller capacity gaps in other parts of the system. Even if midstream producers struggle to raise prices, crude processors and finished goods manufacturers may be able to wield more pricing power, sooner.

And with pricing power concentrated at either end of the processing chain, it is businesses in the middle (from steel and cement makers to component manufacturers and oil refiners) that look set to have their margins squeezed hardest.

Ends --

By John Kemp, Reuters columnist - for Commodities Now

The views expressed are his own.

Attachments:

US Slack

615 Kb