The price of spot iron ore has sunk to $75.50 per tonne this week, its lowest level since 2009. The scale of the price collapse has been breath-taking.

Iron ore has dropped by over 35 per cent since the start of the year, a significantly worse performance than any other industrial metal. But what's really shocking is that the price is now at a level that until recently was collectively deemed impossibly low.

It was only in April that José Carlos Martins, executive officer of ferrous and strategy at Vale, the world's largest producer of iron ore, told analysts that "one thing is for sure, the price will not go below $110 on a sustainable basis".

This was not irrational producer exuberance. Martins was only voicing the prevailing consensus view when he went on to argue that "we have many times seen the price going below this level but recovering very fast".

Well, here we are with the price trading not just below $110 but a lot lower still. And sustainably so.

That tells you that something has gone very wrong with the iron ore narrative. This market is in a place it was not supposed to be.

And while big producers such as Vale, Rio Tinto and BHP Billiton are sticking to that narrative, they are now facing the unpredictable consequences of a pricing war they collectively started.

NICE THEORY...

The "big three", which have some of the lowest-cost operations in the world, are bringing an unprecedented amount of new supply to the market.

Between them they lifted production by almost 12 per cent over the first nine months of this year, and the ramp-ups and expansions are continuing.

They all knew that there would be an impact on price, but the theory, as expounded back in April by Vale, was that it would be limited.

After all, they could argue, the market for iron ore will still expand for many more years as the world's biggest buyer, China, pumps out ever more steel to build infrastructure and new houses.

And lower-cost production from Brazil and Australia's Pilbara will displace higher-cost production, not least in China itself.

Or as Sam Walsh, chief executive at Rio expressed it, "now is the time for others to really feel the consequences of the price against their operating costs and for them to make decisions".

There's plenty of anecdotal evidence that small higher-cost iron ore mines in China are indeed closing en masse, even if corroboration from its notoriously unreliable iron ore statistics is still sadly lacking.

As Australia's share of China's imports inexorably rises, it is clear that other marginal suppliers must be suffering too. But the price still "shouldn't" be as low as it is. So what's gone wrong?

...SHAME ABOUT REALITY

Quite a lot, it seems.

Firstly, cost-curve arguments are just fine in theory, but reality often turns out to be a messier affair. Commodity prices have an annoying habit in periods of oversupply of not only falling below the consensus equilibrium price but also staying there longer than expected.

Iron ore is proving no exception. A host of small, unmechanised, low-grade mines in China may well have exited the marketplace, just as they have done in the past during periods of price weakness.

The scale of the current supply surge, however, means that cost-curve displacement must move well beyond such easy targets.

And others may be in no mood to surrender so easily. Chinese producers are no different from those anywhere else. Faced with low prices, they will try to cut costs.

Macquarie Bank's annual commodities conference in China heard how the whole Chinese iron ore sector is now lowering strip ratios, trimming labour costs and seeking tax reductions in a collective effort to move down the cost curve.

Some will soldier on even if they're losing money, a counterintuitive but ultimately rational producer response, according to analysts at Goldman Sachs. Closures cost money too and come with a loss of option value.

Supply, in other words, can take a lot longer to adjust than the theory says. Secondly and much more surprisingly, it turns out that Chinese demand growth is not the given it was assumed to be.

Its steel juggernaut has shuddered to a standstill. Official figures show year-on-year production growth of exactly zero in September. Alternative figures from the China Iron and Steel Association show its members' output actually fell over the last couple of months.

Analysts will hotly contest the accuracy of both data series, but the trend is a more reliable friend when it comes to Chinese steel sector statistics, and the trend is flat to down.

The big iron ore producers still talk of peak Chinese steel production coming only in the next decade, but the really uncomfortable truth is that a mini-peak has already arrived. Not exactly the best backdrop against which to be engaged in a pricing war.

LAST MAN STANDING?

"War does not determine who is right, only who is left," British philosopher Bertrand Russell said. Which pretty much sums up the attitude of the world's big three iron ore producers.

They may have already been proved wrong in their assumptions about how the market would absorb their extra supply. At issue now is whether they have bet correctly that they will be the last men standing, however low the price gets.

The problem with wars is that they tend to take unpredictable turns. One such turn was Glencore's approach to Rio Tinto to see if it might be interested in "some form of merger", a word that seems to have a different meaning for the Swiss commodities giant than for the rest of us, judging by its previous "merger of equals" with Xstrata.

The approach was quickly and entirely predictably rebuffed.

But Glencore's chief executive, Ivan Glasenberg, a man with strong views on producers' past failures to recognise market realities, has thrown down a far marker.

Rio may well be able to keep producing iron ore at a price below the pain threshold of most others, but at what cost to its own margins and its own shareholders? The company, with over 90 percent of profits coming from the sector, has turned itself into a play on the iron ore price.

Which is now trading at a level no-one predicted even a few months ago. Well, at least it can't go any lower, right? That would be ... well ... impossible. - Reuters
* The opinions expressed here are those of the author, a columnist for Reuters.