COLUMN-The decline and fall of South American aluminium production: Andy Home

(The opinions expressed here are those of the author, a columnist for Reuters.)

* Brazil production graphic: http://link.reuters.com/vev44w

* Brazil, US physical premiums: http://link.reuters.com/buv44w

By Andy Home

LONDON, April 1 (Reuters) - U.S. aluminium producer Alcoa has this week announced the full shuttering of the Alumar smelter in Brazil.

Alcoa (NYSE: AA - news) and minority partner BHP Billiton (NYSE: BBL - news) curtailed two potlines at the 440,000-tonnes-per-year plant over the course of 2013 and 2014. Now (NYSE: DNOW - news) the third and last will also be mothballed.

It's the fifth, and largest, Brazilian smelter to be shuttered since 2009 - a major retreat for what was once one of the world's top producing countries.

And it's symptomatic of a broader decline in primary aluminium production in South America.

A region that was once a major exporter is now becoming increasingly dependent on imports, a tectonic change in flows with major implications for aluminium pricing.

ALUMAR - A BRIEF HISTORY

The first, 127,000-tonnes-per-year, potline at the Alumar smelter in the Brazilian state of Maranhao was fired up in 1984.

A second line, at 156,000 tonnes per year, followed in 1986 with a third launched in 1990.

This was a period of rapid expansion of primary aluminium capacity in Brazil.

Operators were attracted by the country's seemingly abundant hydropower and the resulting cheap electricity prices, a key cost input in aluminium smelting.

True, a warning of things to come occurred when Alumar was forced to reduce output in 2001 due to compulsory power rationing. But the joint owners were still sufficiently confident in the plant's future to expand the third potline in 2006.

Now, though, the smelter is one of the highest-cost in Alcoa's portfolio, which is why the company has been incrementally trimming its output since 2013.

The final cut comes after the company said on March 6 it was placing 500,000 tonnes of annual smelting capacity under review for mothballing, closure and sale.

Alumar is the first, and by inference the most economically compelling, casualty of that review.

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Graphic on Brazilian primary aluminium production: http://link.reuters.com/vev44w ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

BRAZILIAN DECLINE

Alumar's story is symptomatic of the broader decline in Brazil's primary aluminium sector.

Alcoa's Pocos de Caldas smelter in the state of Minas Gerais was the company's first foray into Brazil back in 1970. It has been fully shuttered since May of last year.

The Valesul smelter was closed in 2009, while Novelis closed its Aratu plant in 2010 and its Ouro Preto plant in 2014.

There are now only two primary smelters left operating in Brazil: Hydro's 460,000-tonnes-per-year Albras and CBA's 475,000-tonnes-per-year Sorocaba.

National production has fallen from a peak of 1.67 million tonnes in 2007 to 962,000 tonnes in 2014.

Annualised production was running at just 828,000 tonnes in February, according to the latest figures from Brazilian aluminium association ABAL.

With the full curtailment of Alumar, it will fall even further.

The sector's shrinkage is largely attributable to rising power costs. Hydropower, which seemed so plentiful back in the 1990s, has turned out to be an increasingly scarce resource after years of chronic drought.

Faced with historically low reservoir levels, Brazil will have to import power from Argentina and Uruguay this year. Even that may not be enough to avert power rationing if the rainy season fails to live up to expectations.

Such constraints represent a major challenge for power-hungry smelters, particularly at a time of low prices and compressed producer margins.

VENEZUELAN FALL

Venezuela's two smelters, both owned by the government, have so far avoided closure but are doing no more than limping along.

Here the prime cause of steadily dwindling production is the combination of years of underinvestment and labour unrest.

Venalum has a capacity of 430,000 tonnes per year but it hasn't produced anything close to that since 2009.

A management report, quoted by local newspaper El Universal in January 2014, lamented the fact that only 262 production cells out of 905 were operating at the end of 2013 due to a lack of raw materials, supplies, equipment and spare parts.

Production last year declined further to just 110,000 tonnes and the company has warned customers it will no longer be able to meet industry-standard purity levels due to its "difficult financial situation".

Some of those customers, a Japanese consortium, have been trying to negotiate the sale of their stake for years.

If anything, Venezuela's other primary smelter, the 170,000-tonnes-per-year Alcasa plant, is faring even worse with output last year running at just 17 percent of capacity.

Production from Venezuela, like Brazil once a smelting powerhouse and a major exporter to world markets, hit a 30-year low in 2014, according to the industry ministry's annual report.

The only other primary smelter in South America is Argentina's Aluar, which seems to be still operating, having narrowly avoided a cut-off of its power supplies early last year.

TRADE FLOW REVERSAL

The deteriorating health of the South American smelter sector is chronicled with each month's global production figures from the International Aluminium Institute.

February's regional run-rate was 1.33 million tonnes. A year ago it was 1.80 million. At its peak in 2008 it was 2.70 million tonnes.

South America was once a major exporter of metal to world markets.

In 1994, for example, Brazil was the third-largest and Venezuela the fourth-largest source of aluminium imports into the United States, according to Lloyd O'Carroll, analyst at Northcoast Research. ("Notes on aluminium pricing for investors", July 25, 2014)

Both are now importers, particularly Brazil with its extensive aluminium fabrication sector. The country imported 306,400 tonnes of primary metal in the first nine months of 2014, up from just 78,100 tonnes in the same period of 2013, according to ABAL.

That supply gap is only going to widen as Alumar joins the list of smelter casualties.

This shift in trade flows has significant implications for physical premiums.

A couple of years ago there was no concept of a delivery premium into Brazil because the country didn't need imports.

Now it does and the likes of Platts, a leading global energy, metals and petrochemicals information provider, have launched Brazilian premium assessments to reflect this new trade.

As of last week the Platts premium for Brazilian delivery stood at $590 per tonne over LME cash, significantly higher than either U.S. or European premiums.

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Graphic on Brazilian and U.S. physical premiums http://link.reuters.com/buv44w ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

It has also held relatively steady as other regional premiums slide.

And that is where the decline and fall of South American aluminium production is also a key factor.

As the market ponders the outlook for sliding physical premiums, there is a compelling argument that they will not simply revert to historical norms as finance deals unwind and load-out queues at the London Metal Exchange shorten.

The U.S. premium once reflected the cost of attracting metal from South America. No more.

The marginal import tonne is now more likely to come from the Middle East, where smelter capacity has risen in counterpoint to the decline in Brazil and Venezuela.

For that to change will take either a significant improvement in the aluminium price or a long-lasting change in rainfall patterns in Brazil.

Actually, it would probably take a combination of both to halt what looks like an inexorable decline and fall in South American aluminium production.

(Editing by Dale Hudson)