An attractive alternative in a cheap industry.

Reasonable debt structure under the hands of debt-conscious management.

Interesting valuation that offers protection even for depressed cycle performance.

Why Steel

As of the end of May 2020, steel, together with banks, O&G drilling, some insurance businesses and fertilizers, is one of the cheapest industries in the market. Its performance in the last year has been negative, losing 20% of its market value.

The steel industry is trading at 10x earnings, a 20% discount to its book value and 6x 2019 EBITDA or 10x its 2020-21 estimated EBITDA.

The 20 largest steel companies as of that date are listed below. As we can see, there are significant differences in historical growth, margins, return on capital and current valuation, as measured by EV-to-EBITDA-LTM.

ArcelorMittal (MT) seems to be the cheapest company in terms of EV multiple to LTM’s EBITDA, which was the main reason to start looking into this company. Besides, it is clearly an industry leader able to capitalise into the consolidation that a crisis like the present one might bring.

Steel industry snapshot

The steel industry is not particularly profitable. There can be particularly good years, but the bulk of the returns on capital are going to be in a 5-15% range, and they can be volatile from one year to another. The main driver of return on capital is margin, rather than turnover on capital.

The ROCE of the industry, represented by those 20 largest companies (and MT following a similar path in the last 15 years), was clearly in a negative trend until 2010, when it reached less than 5%, and then had a gradual recovery until 2019, when it reached 15% just before the COVID induced crisis on production first and on demand afterwards.

MT is clearly not the best performer, especially with regard to EBITDA margin, but is able to keep up with its return on capital because is more efficient in turnover and in the control of maintenance capital expenditures.

Steel production is a certainly a complicated business:

• It is close to a perfect commodity for many segments of the finished product. Besides, no matter how old is the article you read on steel or on what country, excess of capacity is always the main concern. Overinvestment and overcapacity cycles have developed in Europe and USA in the 18^th and 19^th centuries, and more recently in China. When overcapacity has been created, producers tend to demand protection from their governments, which only makes the problem persist, creating incentives to perpetuate it.
• Cost production curve is almost flat, especially from $400-450 per ton. Cheaper producers are based in CIS, which is attributed (McKinsey) to low-cost captive raw materials access, a key reason for differences in cost of production per ton.

• The main steel consumers – auto and construction – are themselves very cyclical, so steel demand is cyclical by nature. This adds volatility in the output volumes and prices.
• The business is capital-intensive. It is estimated that the average investment per ton of raw steel production is $250-300. Since the OCDE estimates that global capacity as of end of 2018 was 2,234M tons, that means the total invested capital is c. $650 billion, and that is only in basic raw steel production.
• Gains in productivity from technology are not huge, but they can be critical. This is an old industry, and while new little changes happen constantly, there have not been big game-changer technologies that have made all previous invested capital obsolete. There are two main production technologies that are selected based on the development of the producer country. When there is enough development, there is enough scrap product that can be re-melted using EAF, which is cheaper, cleaner and more efficient than BOF, which the classic technology used when iron ore needs to be used to produce raw steel.
• The input costs – iron ore, energy either in the form of coke coal, electricity or natural gas, scrap and different alloys, etc. – are all also volatile and subject to their own independent cycles, and can suffer from local imbalances that add to global cycles.
• The industry, many times voluntarily, is the focus of political intervention, regulation and protectionism practices that make management of steel companies even harder.

My view is that there is no way I will be able to have a view on what will be the short-, medium- or long-term evolution of steel prices, production, input prices, etc. Everything is too volatile and too fragmented to think of any investment in MT based on a reasonable estimate of the basic economic parameters for the industry.

On the other side of its sheer complexity, there is some inherent stability to steel. It is a product that will not disappear in the foreseeable future, as it is an essential product for everyday life products, and the range of uses and applications is enormous.

Therefore, my purpose will not be to make a guess on future steel price or future cost input prices. I find that impossible for my limited experience. My purpose will be to determine what is an average low point in the industry cycle and to build from there what could be a conservative valuation of the company.

A recap on ArcelorMittal’s recent history

Mr. Mittal started his own business history in Indonesia in 1975 when he changed his father’s plan to sell an unnecessary land plot and decided to build a steel plant on the site. Later, he bought a plant in Mexico that the government had built for more than $2 billion but that was losing money year after year. The next big one was in Kazakhstan. In 1997, the family floated a 20% stake in part of the international steel business and kept the Indian and the Kazakh business within the family.

From the “public” platform, they made the first step in the US in 1998 with the acquisition of Inland Steel. By then, they controlled 2% of world steel production.

With the start of the century, the Chinese expansion brought excellent years for the industry, coupled with a consolidation effort in which Mr. Mittal had a clear leading role. Finally, the steel industry was putting more focus on profitability rather than volume.

According to this study done in 2008, the strategies pursued by different steel companies to survive and grow in what certainly is a difficult industry can be grouped into eight different categories, from capacity growth with competitive acquisitions and JVs to location strategies, vertical integration, niche focus, etc.

According to the analysis, Mr. Mittal showed in the company’s ascent to the global steel pinnacle that it was possible to be active in most of those strategies at the same time. He picked a mixture of strategic (meaning expensive) acquisitions mixed with discounter strategies by picking near-to-bankrupt steel companies at bargain prices. The only strategy he did not seem to exercise was the joint venture route.

The company’s expansion to become the world steel leader reached its peak with the $34 billion deal for the control of Arcelor, the second-largest global producer coming from the French historical Usinor.

It was like mixing water and oil. Arcelor – representing the engineering tradition with its high-end product range, heavy old-school management structures and global planning departments, so full of history and know-how as well as costs and lack of productivity – having to be merged with the localist, low-end cheap steel producer with no apparent organization other than Mr. Mittal jumping from one bargain to another.

The resulting 130-million-ton steel giant, with 10% of global steel production, was about to face the toughest test from the global financing crisis that started in April 2007 and exploded in 2008, just a few months after Mr. Mittal became MT’s CEO. By the closing of 2007, the year the merger took place, the company had $147 billion in assets and $34 billion in financial debt and$62 billion in equity.

As the crisis hit, shareholders and investors realized that valuations had been pushed up in the pre-crisis years. In addition to the global financial crisis, the steel sector would have to deal with a huge problem: the transition from a Chinese demand-driven market into a Chinese oversupply-driven market.

The company would reach its lowest point in 2015-16 with 20% of 2007’s EBITDA, 50% of the starting equity and only 10% of the initial market capitalization. The end of 2016 left behind a period of write-offs, lay-offs, asset selloffs, capital increases, debt restructuring, etc.

Mr. Mittal and MT survived the financial crisis and the Chinese oversupply cycle that had driven EBITDA margin from more than $200 per ton in 2008 to less than $50 per ton in 2016, while going through a $20 billion deleveraging process just after the integration with of an old-school elephant as Arcelor.

For Mr. Mittal, that restructuring process just after going from the distressed opportunistic buyer he had always been to a top-price cycle-peak buyer of one of the biggest targets must have been an experience as enlightening as painful.

The Company Today

MT is the world’s leading integrated steel and mining company. Its strengths focus in global scale and scope, technical capabilities, and a diverse portfolio of steel and related businesses, one of which is mining.

The company is the largest steel producer in the Americas, Africa and Europe and is the fifth largest steel producer in the CIS region. ArcelorMittal has steel-making operations in 18 countries on four continents, including 46 integrated and mini-mill steel-making facilities.

MT has approximately 191,000 employees and 130 million tons of steel products production capacity. Approximately 37% of its crude steel is produced in the Americas, 49% in Europe and 14% in other countries, such as Kazakhstan, South Africa and Ukraine.

The company produces a broad range of high-quality finished and semi-finished steel products and sells them in local markets through its centralized marketing organization to a diverse range of customers in the automotive, appliance, engineering, construction and machinery industries across 160 countries.

MT has a leading market share in its core markets in the automotive steel business and is a leader in the fast-growing advanced high-strength steels segment. In 2016, ArcelorMittal introduced a new generation of advanced high-strength steels, including new press hardenable steels and martensitic steels. In 2017, the company launched the second generation of its electrical steels that play a central role in the construction of electric motors.

MT’s mining operations are integrated with its global steel-making facilities and are important producers of iron ore and coal. In 2019, approximately 52% of ArcelorMittal’s iron ore requirements and approximately 12% of its PCI and coal requirements came from them, and it is also a significant producer of coke, satisfying 95% of its coke needs through its own production facilities.

Its production sites have good access to shipping facilities, including through its owned, or partially owned, 15 deep-water port facilities and linked railway sidings. The company has its own downstream steel distribution business, primarily run through its Europe segment.

It seems like MT has a good business set-up that it has taken time to create and develop. This is a truly global diversified footprint steel producer with the correct links and bonds to the top global steel consumers in most of the geographical markets it operates.

ArcelorMittal KPIs

What I think are the company’s KPIs are summarized in the following table.

As we can see, production capacity has been slowly reducing from 130 million tons (metric) after the merger of Arcelor and Mittal to the current 112 million tons, despite the different punctual acquisitions made by the company in the US, India and Italy in 2014, 2018 and 2019.

Average utilization of installed capacity for the whole period has been 74.5%, but we can see it has been higher than the average since 2015. Something similar happens with turnover, which on average has been 1.6x, but it has performed better since 2015.

EBITDA has averaged $9.6 billion equivalent to 10.8% of revenues, or $107 per ton. EBITDA is quite volatile going as high as 18%, close to $200 per ton, and as low as 6%, or $50 per ton. In dollars per ton, it can perfectly double, like in 2010, or halve, like in 2019, from one year to another.

A 5-6% EBITDA barely covers maintenance capex and debt interest service, so despite the fact that it looks positive, it is the very minimum to keep this company, and probably any other steel company, alive.

McKinsey estimates that the EBITDA margin required for long-term sustainability (cover variable, fixed and capital costs) is 17%, which compares with a 10% average in the 2004-16 period shown by McKinsey in the analysis linked to above.

In 2009, the decline in EBITDA margin was clearly caused by the fall in demand that led to a utilization rate of 55%, the lowest point in the series. But in 2015 and 2019, it had more to do with a fall in steel prices, not sufficiently compensated by a fall in input costs.

In line with EBITDA fluctuations, the normalized return on capital employed, ROCEn, moves significantly from one year to another. The average ROCEn is 9.2%, but it can be twice that, as in 2007, or three times less, like in 2019. Normalized ROCE means EBITDA net of maintenance capex and constant tax rate over long-term assets and working capital net of goodwill. There is no effect from write-downs, impairments, etc., and no leverage effect. I also do not deduct changes in working capital, as they tend to put noise without changing the average.

How the business is valued by the market, I think, is probably best tracked by EV-to-EBITDA and EV-to-installed production capacity or to actual production in tons.

All three metrics tell a consistent story, as they have constantly decreased with time, especially since the 2008 crisis. In 2006-07, the EV/EBITDA was at 10x-12x, and in the last five years, the range has been 4x-6x. EV per installed steel production capacity was c. $1.650 per ton in 2007 and is now at $290 per ton, close to 6 times less. Same thing if we look at EV per ton of produced steel.

What is interesting is that the average EBITDA per ton in 2017-19 (to smooth out the low figure of 2019) is 50% of the maximum reached in 2007, while the EV per installed production capacity has divided almost by 6x. In other words, the valuation of the business, before debt is considered, has fallen almost 3 times more than the actual fall in its margin capacity.

Normalized valuation estimate

For this valuation estimate:

• I will use 2019 revenues of $77.7 billion that come from 86 million tons of steel shipments at an average price of $928 per ton. To put this figure in perspective, the amount of steel sold is at the 40^th percentile of last 14 years, while selling price per ton is at the 30^th percentile. The normalized revenue figure should probably be higher, but since 2020 will be certainly lower, I think it is more prudent to use that reference.
• I will use the average EBITDA margin for the whole period of 14 years, which comes out at 10.8% rounded down to 10%.
• I will use a 4% maintenance capex rate, also in line with the historical average and equivalent to $3.1 billion p.a.
• With those assumptions, normalized EBITDA comes out to $7.77 billion p.a. and normalized FCF comes out to $3.95 billion p.a.

At current market values, those normalized figures imply a 3.2x EV-to-EBITDAn and a 6.3x EV-to-FCFn. A 6.3x multiple on FCFn implies a 15.9% unlevered FCFn yield for us, the investors.

Looking at it from another perspective, applying a 10x multiple to FCFn, the resulting EV is $39.5 billion that net of debt, pensions, cash and JVs (with a 30% haircut), which implies a $25 billion equity value equivalent to $20.6 per share, already taking into account the last capital increase and the deferred conversion of recently issued mandatorily convertible notes (at minimum conversion price).

The “normalized valuation” implies a value per ton of production capacity of $350, which compares favorably with the historical accumulated gross investment made by the company of $660 per ton in Plant and Equipment.

Besides, we are not taking in account intangible assets built or paid ($5 billion) by the company along its existence and across its global footprint. Assets like client relationships, global presence and global service for its clients, specialized product knowledge or service, etc. – all of this you are buying for free.

Leverage and debt structure

A multiple on FCF is a simplified way to make a discounted cash flow. So, when I discount un-levered FCF and deduct nominal value of debt from the resulting value, I am being conservative, since I am not applying any time value adjustment to debt cash flows nor taking in account the positive tax effect of interest expenses. I do that because I prefer to undervalue those with higher debt.

I like to think more on what the right conservative debt for each business should have instead of plugging in the valuation whatever debt structure the company has. In the case of MT, I think the company should have no more debt that it can service on the more complicated years, which come with:

• An average EBITDA margin of 6%, or $4.7 billion p.a.
• An average maintenance capex of 3.6% of revenues, or $2.8 billion p.a.
• FCF, net of 15% taxes and maintenance capex, would be $1.6 billion p.a.
• If we assume a 4% cost of debt and a 10% annual amortization rate, the maximum debt the business would have to hold is $11.4 billion, equivalent to 2.4x debt to the depressed EBITDA ratio and 1.5x debt to a normalized $7.7 billion EBITDA.

By 2008, before the crisis hit the sector, debt was as high as $37 billion, which paired with an historically high EBITDA ($18.8 billion) made for an apparently low ND/EBITDA ratio of 1.5x. The problem appeared next year, in 2009, when EBITDA went to $4 billion and made the ratio jump to 5.3x.

The suffering that was to come was enormous. From 2008 to 2018, gross debt had to be reduced by almost $24 billion, down to $13.7 billion ($10 billion net of cash). As of the end of Q1 2020, the company reported gross debt of $15.5 billion, still above our reference, but coupled with $4.7 billion of cash in the balance sheet and $9.3 billion of unused credit lines, it makes for a reasonable scenario.

Comparing the current net debt to 2019 EBITDA would yield a 2.2x ratio, which seems reasonable, especially if we take in account that 2019 EBITDA was in the lower part of the historical range.

Also relevant is the maturity schedule that I show on the table, together with the un-covered maturities that current cash or lines would allow:

As we can see, only using its current cash, MT can re-finance all its maturities until the end of 2021, which is a comfortable position. That would leave all current unused lines available.

But even with this comfortable position, in May 2020 the company made a $2 billion equity placement in the form of:

• $750 million of equity, representing approximately 80.9 million common shares at an offering price of $9.27 per share.
• Deferred equity in the form of mandatory convertibles for $1,250 million that at maturity (3 years) will be converted into common shares. The notes pay a coupon of 5.50% per annum, payable quarterly. The minimum conversion price will be equal to $9.27, and the maximum conversion price will be 117.5% ($10.89 per share) of the minimum conversion price.

The Mittal family trust participated in the offerings by placing an order in an aggregate amount of $200 million (half in each offering), equivalent to 13% of the equity offering and 8% of the deferred equity (10% of the combined amount), which is below their 37.4% reported stake in the company.

I think the company went a bit ahead of itself, but the safe side is probably the side you went to err on. Probably all the suffering that went on from 2008 to 2016 has played its part in the decision making.

Starting with the normalized FCF we have estimated before, we can get to the leveraged normalized FCF we would have just after the common equity issued by the company in May 2020 and the resulting equity-debt that the company would have after the mandatory notes conversion. I’ve also made the example with the ideal debt I calculated before:

With the current leverage, when the company recovers a normalized level of FCF, the effect of debt will be to reduce FCF by $688 million from interest expenses that would bring $103 million in tax savings, leaving the levered FCF at $3.3 billion, which would yield 31% on market cap as of end of May 2020 (€9 per share, or $9.9 per share).

If equity had to be increased in order to bring down debt to its ideal conservative level of 11.4 billion, then the levered FCF would offer a 24% yield on adjusted market cap, which shows that the market is either valuing MT too cheaply or that it is assuming that the average normalized 10% EBITDA of the past is never to be recovered.

At current share prices, even the depressed EBITDA margin scenario would yield a reasonable yield to the investor, since it would be around 9%.

I think MT offers an interesting enough risk-reward opportunity to invest in.