Dec 082014

In discussing one of the main themes in this blog – Intelligent Investing – I have chosen to focus on concepts and methods rather than present specific investment cases. Exceptions have been Meyer Burger (here, here, here and here), Barratt Developments (here, here and here) and Elan (here and here). These have been successful investments. But since another major theme in the blog is Experts, and their more or less deliberate tricks aimed at trumpeting good calls and obfuscating bad ones, here is my take on Ferrexpo, which so far has been a spectacular dud.

I presented my investment case in Ferrexpo about a year ago at the European Investment Summit. Here is an excerpt from the presentation. Since then, there have been two major developments: the Ukrainian crisis and a steep drop in the price of iron ore. On the first issue, I took the view that, short of an outright Russian invasion of Ukraine, which I regarded as very unlikely, Ferrexpo’s operations were not going to be significantly affected by the turmoil. This turned out to be right. On the second issue, however, I was wrong. I thought that the big three iron ore producers – Rio Tinto, Vale and BHP Billiton – would choose price over volume and limit the expansion of their production capacity. They didn’t. Their expansion plans went ahead, on the assumption that if they didn’t increase production others would, and with the intent of squeezing out the high cost producers, concentrated in China and subsidised by the local government. As a result, the iron ore price fell from 130 dollars per ton to 70. In this new environment, Ferrexpo decided to postpone the capital expenditure plans required to realise its own capacity expansion.

My investment case, however, was not predicated on the assumption of price stability. Of course, a big price drop can’t be good news. But, as long as it is not permanent, it should not a have a big impact on valuation. As shown in the graph, Mr Market sees all price moves as permanent:

The high correlation between the stock price of Ferrexpo and the iron ore price (the same is true for the other iron ore producers) means that the market has no idea of where the iron ore price will go. If the price is at 50, as it was between 2008 and 2009, it assumes that it will stay down there forever, and prices Ferrexpo below 1. If it is at 190, as it was in early 2011, it also assumes that it will stay up there forever, and prices the stock at 5.

But the graph also shows very clearly that the iron ore price is highly cyclical, and therefore there is no reason to expect that price levels – high or low – will persist. My valuation shows that Ferrexpo is still cash positive at low iron ore prices. Its production costs are not as low as they are for the big three – returns to scale are very important in the mining business – but the company is certainly not one of the high cost producers that would go out of business in a protracted low price environment. My numbers show that, even at a prudent normalised iron ore price of 100, Ferrexpo’s correct valuation would be above 3. At a price of around 1.8 in late 2013, there was – I thought – a sufficiently ample margin of safety.

Boy, was I wrong. In 2014, as the iron ore price began its steep descent and, to add insult to injury, the Ukrainian crisis intensified, the stock price dropped from 1.8 to 1.3 from January to September. Then in the last two months – shortly after I reiterated my valuation case at the Value Investing Seminar in Italy – the drop accelerated, taking Ferrexpo all the way down to below 0.7. So much for safety.

The move was preceded by steep downgrades from a number of brokerage analysts. No surprise there: as we have seen with Meyer Burger, analysts’ target prices (dashed line) just tread along market prices (solid line). In 2011, when the price was 5, Ferrexpo was a Buy for 69% of analysts. Now, with the price below 1, Buys have dwindled to 28%. Rather than accurate price forecasts, target prices are mere sentiment indicators.

Be as it may, as an observant Blackstonian investor I proceeded to try to prove myself wrong. So I talked to some analysts, starting with those who had carried out the most aggressive target cuts, in some case moving straight from Buy to Sell. How could they possibly value at less than 1 a company that I thought was worth at least 3? Ukraine and the iron ore price, was the predictable answer. Yes, I replied, but how does that exactly translate into your valuation? I had to get their spreadsheets – which some of them kindly provided.

The first thing I checked, of course, was the cost of capital in the terminal value. Jack up the cost of capital and you can get any valuation you want. In the most interesting case, however, the WACC was a reasonable 10%. Minus 3% terminal growth, ok. But what’s that? Terminal FCF/(WACC-G) times 0.5? That’s got to be a mistake. I corrected it and, lo and behold, the valuation more than doubled. So I wrote back to the analyst and pointed that out. “Eeek. Well spotted” was his reply “but all I need to do is to take the terminal growth rate down by a smidgeon!” Which of course was not the case: even if he had halved the growth rate, he would have obtained a much larger valuation. But that’s not the end. Going deeper into the spreadsheet, I saw that, in the ‘Extraordinary Items’ line, the -14.6 million reported in the first half of this year was carried over into the future. I corrected that mistake and the valuation, based on the analyst’s own model, increased by a further third, to 2.6 times his price target. I pointed that out as well. Any qualms? No. The target was reconfirmed a few days ago.

Glaring insouciance apart, other models had 13% WACC and more (which has the same effect as halving TFCF/(WACC-G)), justified, as one analyst put it, by Ferrexpo’s ‘tough postcode’. And in all of them the last Free Cash Flow before the terminal period was calculated based on the current iron ore price, or thereabouts. This gives a value, in five to ten years from now, which is more than 50% lower than the 2014 level. No wonder: with the iron price at the current level, not for the next couple of years but for ever, iron ore mining would become a very tough business, viable, perhaps, only for the big three. This seems to me a very unlikely scenario.

So it is hard to convince myself that I am the one who’s got the valuation wrong. Analysts are lowering their valuations because the stock price is going down. And why is the stock price going down? Because the iron price is going down, of course – just look at the chart! You can’t but admire the exquisite circularity of the argument. Which, by the way, ends up sanctioning the analyst’s own irrelevance: know where the iron ore price is heading and you know where the stock price will go.

In a more basic sense, however, there is no question that I have been proven wrong: I thought it was very unlikely that Ferrexpo’s stock price could reach such a low level. Clearly, the margin of safety was not as wide as I thought. And if the iron price stays at current levels for a while, as everybody is expecting, the risk is that the stock price will stay there as well, as the chart suggests.

But I have been there before. If the reasons why the price is where it is are those above, Mr Market will catch up, sooner or later. As Andrew Harding, the CEO of Rio Tinto Iron Ore, said at last week’s Investor Seminar:

Our view remains that the developing world will continue to drive demand for iron ore, through urbanization, industrialization and increasing domestic consumption patterns. On the supply side, we have already seen significant curtailments of iron ore supply from the Chinese domestic sector, as well as reductions from non-traditional suppliers such as Indonesia and Iran. We expect around 125 million tonnes to leave the market this year in response to lower prices. Yes, the present price compared to recent prices is depressed, but the value proposition of our iron ore business runs over decades, not today and not tomorrow.

I wish Ferrexpo’s management could be as forthcoming, rather than taking it on the chin and keeping a very low profile – supposedly to earn the respect of unreciprocating City analysts.

Print Friendly
  • Thanks Danilo.

    Well put. I am curious about your model: I would think that forecasting 100 years EPS in an automated manner is equivalent to assuming a terminal growth rate, isn’t it?

    Anyway, the key assumption, as you say, is about normalised margins at the end of the forecast period. Most analysts are now carelessly assuming very low margins forever for iron ore companies. And so are fund managers who are, even more carelessly, dumping the stocks. We’ll see what happens next year!

  • Danilo Santiago

    Hello Massimo:

    Just saw your posting on LinkedIn and really liked it. The comment “Mr Market sees all price moves as permanent” really resonates with the core of our process at Rational.

    See below a picture (figure #1) of a page we included in our presentation recently. The page also explains another observation you made: “And why is the stock price going down? Because the iron price is going down, of course”. I would add that prices usually follow “NTM EPS” – which of course, on the case of a commodity producer, will strongly correlate with the price of the commodity the company sells. Quite amazing to observe this correlation of current EPS and share price since “The Market” (or sell-side analysts) have a lot of information nowadays and should know better – but they will disregard facts and get extremely biased by current EPS.

    Last, to avoid the temptation of “torturing” WACC or any other driver of a perpetuity – which will give you any value you want – we forecast EPS for 10 years – after that, the “model” becomes a perpetuity: not with a “simple formula” – we just keep forecasting EPS for “100 years” (on an automated manner, of course) and discount that cash flow back. We also can’t change WACC of our companies – it is fixed.

    Of course the intention is not to perfectly forecast what will happen in 10 or 100 years – after bringing margins/sales to more normal levels (for instance in 5 years), the model already goes to “average margins/EPS” at levels consistent with the company/industry. To calculate “margin of safety” for a long/short we use scenarios – see example below (figure #2) for CNW’s EPS “path” under different scenarios – I presented that company in Trani a few years ago, when we exchanged some emails.

    At the end however, we should be happy that the market have such a short-sighted view of valuations: if they were all rational investors, value-investing wouldn’t work.

    We are seeing now a big over-reaction on oil related stocks – for sure some of those companies will be great investments going forward.



    Rational Asset Management