The seaborne iron ore cost curve

Here’s a chart from UBS to raise a few eyebrows:

iron roe break evens

Any comments on the notion that FMG is cheaper than Vale are welcome! UBS argues it’s because of the higher shipping costs despite Vale’s C1 costs of $23wmt.

Comments

  1. ResearchtimeMEMBER

    I know I have posted this link before, its a Metalytics cost curve (look at page 28 http://www.stocknessmonster.com/news-item?S=FMG&E=ASX&N=668109). Their cost curves are (as far as I know) the most accurate globally – UBS base their curves of them, as do a bunch of others. In regards to cost, just to reiterate, Brazilian itabrite ores grade 37-55% Fe, so it takes substantially more to process them and up-grade the product. The benefit comes when you sell it – Rio spec is 62.5% Fe, Vale spec starts at 64.5% (mostly >65.5% Fe), with some shipments as high as 66-67% Fe. Obviously they get better prices too…

      • ResearchtimeMEMBER

        Not really – given there are a variety of iron products, a plethora of contaminants, such as Si, P, S, etc – even a variety of iron ore sizes, loosely termed lump and fines with need to be sintered (an extra cost) before they enter the blast furnace (otherwise they would be blown away via thermal turbulence). Even substantial differences between hematite (majority of producers) and magnetite (covers a lot of Chinese production), which is actually exothermic, which is beneficial when you smelt, meaning that you can conserve on the amount of coking coal required.

        That’s why its best just to use the representation of a generic C1 cost curve – and then break it down from there. For instance, the bottom blob on the cost curve (see above link) is Rio, which lend their name to Rio Spec, averaging 61.5% (getting more difficult to reach these days in the Pilbara– average is now down to 60.8-61.2% Fe) within certain metallurgical parameters (e.g. P has to be less that 0.05%). The often price people quote is Rio Spec; by comparison FMG sells around 59-60% Fe, but it receives discounts greater than 11% of offical Rio Spec, substantially greater than the difference accounting just for the iron ore content. Primarily because the majority of the ore are fines, and in particular, elevated contaminant levels. Iron ore marketing is actually quite complicated and subjective.

      • ResearchtimeMEMBER

        Not really – given there are a variety of iron products, a plethora of contaminants, such as Si, P, S, etc – even a variety of iron ore sizes, loosely termed lump and fines with need to be sintered (an extra cost) before they enter the blast furnace (otherwise they would be blown away via thermal turbulence). Even substantial differences between hematite (majority of producers) and magnetite (covers a lot of Chinese production), which is actually exothermic, which is beneficial when you smelt, meaning that you can conserve on the amount of coking coal required.

        That’s why its best just to use the representation of a generic C1 cost curve – and then break it down from there. For instance, the bottom blob on the cost curve (see above link) is Rio, which lend their name to Rio Spec, averaging 61.5% (getting more difficult to reach these days in the Pilbara– average is now down to 60.8-61.2% Fe) within certain metallurgical parameters (e.g. P has to be less that 0.05%). The often price people quote is Rio Spec; by comparison FMG sells around 59-60% Fe, but it receives discounts greater than 11% of offical Rio Spec, substantially greater than the difference accounting just for the iron ore content. Primarily because the majority of the ore are fines, and in particular, elevated contaminant levels. Iron ore marketing is actually quite complicated and unfortunately, subjective.

    • ResearchtimeMEMBER

      Add onto that processing, very long distances for rail, demurrage, shipping – if comparing on a CFR basis.