by Vaclav Smil
Given steel’s key role in modern economic development, it has been unavoidable that the metal’s output has reflected rather accurately the macroeconomic trends: pre–WW I expansion and WW I peak followed by sharp decline and recovery of the 1920s; crisis of the 1930s followed by rising prewar and WW II demand; post-1950 expansion checked only by the economic downturn brought by crude oil price rises of 1973–1974; then 15 years of fluctuating output before the global steel production reached a new all-time high in 1989, before a decade of slow growth or stagnation; and, most recently, the China-driven expansion that was strong enough to only slightly dent the total for 2009 (down by about 8%) and bring another record performance in 2010—and in all subsequent years until 2014.
The industry’s financial footprint is much lighter because steel products are relative bargains: in 2014 output of the global steel industry was worth about $2 trillion, while the 2014 gross world product (GWP) reached about $108 trillion in terms of purchasing power parity (IMF, 2015). Consequently, steel output was worth less than 2% of the world’s economic product—but that was about six times as much as the sales of semiconductors: in 2013 they reached a new record of just over $300 billion or less than 0.3% of GWP (SIA, 2014). As steel prices fell in 2015 the total worth of the industry’s output decreased commensurably.
This was just the latest of recurrent changes: steel industry has always been cyclical, responding to the changes in economic activity, and the first 15 years of the twenty-first century have seen large price fluctuations. In the United States FOB prices of hot-rolled steel more than doubled between January 2000 and the summer of 2004 (from about $350/t to a peak just above $800/t), the next summer they fell below $500/t, and then they rose to a record level of $1,200/t in July 2008; a year later they were back to $400/t and have since fluctuated mostly between $600 and $800/t. At the beginning of 2015 they stood at just above $500/t and were $100 (ex-works prices) lower in Western Europe and $200 lower in China, reflecting the worldwide surplus of the metal (Steel Benchmarker, 2015).
As in every major industry, steelmaking has seen a great deal of consolidation and the emergence of large multinational companies. The largest one, Arcelor Mittal (headquartered in Luxembourg) employs more than 200,000 people in many EU countries, North America (where, among other operations, it owns the assets of the defunct Bethlehem Steel), Asia, Latin America, and Africa. South Korean POSCO (headquartered in Pohang) has plants in China, Vietnam, and Mexico; Russia’s Severstal (headquartered in Cherepovets) operates also in Poland, Latvia, Ukraine, and Italy. At the same time, steel production is nowhere near as highly concentrated as it is in many other important industrial and extractive sectors.
Most notably, only two companies (Boeing and Airbus) supply the global market with large jetliners, only four companies (CFM International, GE, Rolls Royce, and Pratt & Whitney) make their engines (gas turbines), and the top 10 iron ore producers (including China’s Ansteel Mining, Brazil’s Vale and Samarco, and Australia’s Rio Tinto, BHP Billiton, and Fortescue Metal Group) extract more than 90% of the world’s supply (Mining-technology.com, 2014). In contrast, in 1990 the top 10 steelmakers produced only 18% of the world’s steel, and by 2013, despite the intervening intranational mergers and the rise of new large multinational companies (ArcelorMittal operates in 20 countries), that share rose only to 27%, and the largest company, ArcelorMittal, had only 6% of the global market (WSA, 2015).
And while Boeing and GE Aerodivison have been profitable, ArcelorMittal, the world’s largest steelmaker, lost $1.1 billion in 2014, $2.5 billion in 2013, and $3.8 billion in 2012 (ArcelorMittal, 2015)—and its loss-making performance has had plenty of company both among private companies and among China’s state-owned enterprises (whose opaque accounting hides the real extent of their unprofitability). Between 2009 and 2013 US Steel posted 5 years of consecutive losses (O’Hara, 2014). Data collected by McKinsey show that during the first decade of the twenty-first century about 40% of large steel companies had negative cash flow (McKinsey, 2013b).
Moreover, the steel industry would require an average margin of 16% (earnings before interest, taxes, depreciation, and amortization) in order to be economically sustainable in the long term, but the mean margin between 2000 and 2014 was less than 14%, and it is not expected to recover significantly (McKinsey, 2013b). This means that the majority of steel companies cannot play an active role in the industry’s consolidation and that the necessary capital will have to come from outside investors. The contrast has been stark: modern steelmaking industry is indispensable and it has become much more efficient than in the past—but it has not been sufficiently profitable, and the future of this critical industrial sector cannot be seen with assured confidence.
Massive governmental intervention (financial, legislative, directly targeted to the sector, or indirectly supportive) has been the norm for all major steel producers, and its unhappy outcomes have included deteriorating profitability, distorted markets, trade disputes, and a near-chronic excess capacity (McKinsey, 2013b). The average global capacity utilization rate was about 78% between 2005 and 2015, during the latest great economic retreat the rate dipped to just below 60% by December 2008, it was barely above 70% by the end of 2012, a year later it was still below 75%, and at the beginning of 2015 it was 72.5% (WSA, 2015). This overcapacity exists in the EU, Russia, Ukraine, and Japan, but it has been particularly troublesome in China, leading to repeated calls for the industry’s restructuring and closure of up to 80 Mt of capacity by 2018, but this has been resisted by provinces and cities where steelmaking is a leading source of employment.
Although highly mechanized and equipped with numerous electronic process controls, the global steel industry still directly employs more than 2 million people (and twice as many indirectly), and that total should be doubled again by including all related suppliers. Its importance remains critical even in the countries that pioneered its modern development. Perhaps most notably, Germany—with the United Kingdom and the United States one of the three great industrial powers of the pre–WW I world—remains a manufacturing leader with steady export surpluses. Schmidt and Döhrn (2014) see steel as an indispensable cornerstone of German industry, and there are readily available statistics to confirm this judgment. Stahlinstitut VDEh (2014) estimated that in 2013 about 4 million German jobs were in steel-intensive industries (led by carmaking, machinery, and electro industries), and it calculated that in 2013 about 72% of the country’s foreign trade surplus could be traced to these industries.
The two other important challenges have been the industry’s aging assets (now the norm in all affluent economies) and a pronounced shift of margins to mining (a worldwide phenomenon brought by rapidly rising costs of raw materials). Steel plants in industrialized countries are getting old: in the EU 95% of all capacity was older than 25 years in 2013, in the countries of the former USSR the rate was 81%, and in the United States all of the operating plants were older than a quarter of a century (VDEh, 2013). And value creation in steel has shifted upstream, away from the steel industry to the suppliers of raw materials in general and to iron ore in particular.
In 1995 81% of profits came from steelmaking, 11% from coking coal, and 8% from the mining of iron ore, but by 2011 the shares were 26% for steelmaking, 28% for coke, and an astonishing 46% for the extraction of iron ore (McKinsey, 2013). In the year 2000 the coking coal and iron ore needed to make a tonne of hot-rolled coil were about $80 and by 2010 they surpassed $400 (McKinsey, 2010), but by 2014 the end of the expansive stage of Chinese ironmaking reduced the price of imported ore by nearly two-thirds compared to its peak in early 2011, and coking coal prices were more than halved, restoring a more acceptable division among the costs of steelmaking and its raw materials. In the closing chapter I will appraise the major factors that will either promote or counteract the reasons behind the current shifts in the global steel industry.
Flows and Consumption Rates
Given the magnitude of material flows involved
in the global iron and steel industry, a variety of finished products containing steel, their very different lifetimes, and the increasing importance on using the anthropogenic stores of the metal, it would be not only interesting but also quite useful to have a fairly reliable understanding of the movement and stocks of steel, inputs required to make it, and by-products of its smelting. Obviously, producing such flow-and-stock analyses is easier on a national basis and for the countries with good, long-term statistics, but (given the globalization of resource trade and the high level of manufactured exports) global appraisals would also be welcome.
Studies of anthropogenic cycles of elements in general and metals in particular have become more common since the year 2000 (Chen & Graedel, 2012)—but, as with all such large-scale, and relatively complex, flow-and-stock assessments, the results will have significant margins of error. Many flows are closely monitored (e.g., production and trade of iron ore, production of pig iron and crude steel, recovery of commercial scrap metal, steel use in automotive sector) and others can be estimated with a fair degree of accuracy (generation of iron-containing blast-furnace slag), but quantification of some flows (particularly the generation of obsolete scrap and the mass of metal abandoned in landfills or simply thrown away) requires more detailed (and often missing) information about steel content of structures and consumer products, and assumption regarding their average durability.
As noted previously, steel consumption is an order of magnitude (nearly 20-fold) larger than the combined use of the other four leading metals—but annual iron ore extraction and steel use amount to a small fraction of global material consumption whose three most massive components are minerals used in construction (sand, crushed and dimensional rock, clays), fossil fuels, and biomass (Smil, 2013). Recent annual extraction of more than 3 Gt of iron ore is surpassed by coal mining and oil production—respectively, 7.8 Gt and more than 4 Gt, with natural gas flow remaining below 2.5 Gt/year (BP, 2015)—and the global consumption of bulk construction materials, sand, clay, gravel, and stone (used in natural form or processed to make cut stone, bricks, tiles, and, above all, cement) is in the order of 40 Gt.
Starting with the flows that are known with high reliability, the global annual output of steel passed the 1 Mt mark in 1875, and by 1900 it was just above 28 Mt; the WW I peak in 1917 was 82 Mt, the pre–WW II record was reached in 1939 at 137.1 Mt, and the long post–WW II expansion peaked in 1974 at 708 Mt. After a quarter century of stagnation and slow growth the century ended with an annual output of 850 Mt, and just 2 years later began the China-led period of extraordinary expansion that brought the output to 1.43 Gt by 2010 and to nearly 1.7 Gt by 2015. These comprehensive (and fairly reliable) production statistics make it easy to calculate the aggregate global steel output. During the second half of the nineteenth century it amounted to about 400 Mt; during the twentieth century nearly 31 Gt of steel were made, with half of it after 1980; during the first 10 years of the twenty-first century the global steel industry added 11.6 Gt; and the 5-year aggregate between 2011 and 2015 was almost 8 Gt.
During the twentieth century, when the worldwide output of steel rose 30-fold, average annual exponential growth was 3.4%, but 80% of that gain came after 1950, when the annual exponential growth averaged 3%. And, largely because of China’s extraordinary production increase, global output of crude steel nearly doubled during the first 15 years of the twenty-first century, to about 1.65 Gt, with the annual exponential growth averaging 4.4%. Of course, the twentieth century was also an unprecedented period of worldwide population growth, with the total rising from 1.65 billion in 1900 to 6.07 billion in the year 2000 (a nearly 3.7-fold increase and average annual exponential growth rate of 1.3%) and to 7.3 billion in 2015. This means that during the twentieth century average per capita supply of crude steel increased from just 17 kg to 140 kg (8.2-fold increase), and that it reached 226 kg in 2015 (1.6-fold increase in 15 years).
Obviously, this global average conceals enormous national disparities because steel consumption is closely associated with the rapidly ascending phase of economic growth caused by large-scale industrialization and urbanization and with associated expansion of infrastructures and rising ownership of a greater range of consumer products. During the twentieth century these processes had run their course only for a minority of the world’s fast-growing population (concentrated in affluent countries of Europe and North America and in Japan, Australia, New Zealand, South Korea, and Taiwan). China, the world’s most populous nation, has unequivocally joined the trend (following decades of Maoist mismanagement) since the 1990s, and India, the second most populous nation, has been at least two decades behind China.
As a result, there has been no obvious long-term link between the growth of average global per capita steel consumption and average per capita economic product: the latter rate (expressed in constant, inflation-adjusted, monies) rose 4.8 times during the twentieth century compared to a 8.2-fold increase of per capita steel consumption—but since the year 2000 the two multiples (each at 1.6) have been virtually identical. In contrast, national per capita rates generally confirm the expected association between the consumption level of steel and GDP, and some notable departures from these expectations have obvious explanations.
Historically, steel consumption can be used as a close proxy of economic overall accomplishments: average annual per capita output of steel is an excellent surrogate measure of economic advancement, and tracing this rate informs us about the fortunes of major economies and about the distance between the modal use and the leading means. In 1913, when the global mean was only about 40 kg/capita, the United States averaged about 255 kg, Germany’s mean was nearly 265 kg, and the United Kingdom’s rate was less than 190 kg/capita, clear indicators of the global economic leadership achieved by those three nations. A century later the global mean of crude steel consumption rose to 236 kg/capita; the US rate (at about 334 kg) was about 40% higher, and the German rate was 507 kg, but the British mean was just 153 kg/capita (WSA, 2014).
National means of affluent countries ranged between 153 kg for the United Kingdom and 1105 kg for South Korea, while China’s average consumption reached about 570 kg (nearly four times the British mean), India, with 64 kg, remained an order of magnitude behind China, and the chronic underdevelopment of sub-Saharan Africa was clearly indicated by very low steel consumption rates of 14 kg in Nigeria and less than 10 kg in Ethiopia and Zimbabwe: only the relatively developed South Africa averaged above 120 kg/capita. The relationship between per capita steel consumption and per capita GDP thus shows a high correlation for low and medium rates.
A log–log plot of the national averages of the two variables shows that for GDP up to about $2000/capita and for average per capita steel consumption of 150–200 kg there is a relatively narrow scatter along a straight diagonal—but the link breaks down as countries get richer, and the plot shows considerable scatter for middle-income and high-income countries: their steel per capita consumption can range from as little as 120 kg to more than 1000 kg. Consequently, we can find countries with nearly identical steel consumption but with very different per capita GDP: both Iran and France now consume about 220 kg of finished steel products per capita but their 2013 GDPs (expressed in purchasing power parities) were, respectively, $16,000 and $40,000. Conversely, countries with nearly identical GDP have very different per capita consumption of finished steel products: in 2013 GDPs in Japan and the United Kingdom were, respectively $37,000 and about $36,000, but Japan consumed per capita nearly 520 kg of steel, while the United Kingdom consumed only about 150 kg.
This finding is not surprising as it replicates many other relationships between per capita GDP and material and life-quality indicators. Above certain levels required to satisfy what is generally perceived as a high level of economic development or satisfactory quality of life, there could be significant differences explainable by peculiarities dictated by natural conditions or specific demands of economic production. Obviously, Japan and Germa
ny, the two leading exporters of automobiles as well as heavy machinery and entire industrial (electricity-generating, petrochemical) plants, will have a higher per capita steel consumption than France or Spain, and very high steel consumption rates for Taiwan and South Korea (respectively about 950 kg and 1105 kg per capita in 2013) are anomalies caused by the extensive use of the metal in building large oceangoing vessels (oil and LNG tankers and dry bulk carriers), and in South Korea’s case also automobiles and major household appliances for export.
But obvious differences would persist even if we were to account accurately for the net steel used solely for domestic infrastructural, industrial, and commercial investment and for purchases of consumer products: again, different sectoral compositions of national economies (such as China’s high share of industrial production in contrast to continuing deindustrialization of many Western economies) and differences in ownership of vehicles (and their size) and appliances will matter. And a simple quantitative measure (be it apparent per capita consumption of crude steel or finished steel products) does not capture key qualitative aspects of the metal’s use, namely the increasing share of higher-value products (high-strength steel, special alloys, stainless steel) that are now increasingly required for many exacting applications.
Steel Stocks
The world’s aggregate steel output between 1850 and 2015 was about 51 Gt, and an obvious question to ask is: what has been the fate of this large mass of metal? Looking at the twentieth century I estimated that by its end a tenth of 31 Gt of steel produced between 1901 and 2000 was oxidized or destroyed in wars and industrial and construction demolitions, about a quarter was recycled, and 15% was embedded in aboveground structures and underground (or underwater), with the remainder being the accumulated steel stock of about 15 Gt, or about 2.5 t/capita, available for potential conversion to new metal (Smil, 2005). My estimate was confirmed by the most detailed attempt to quantify the global steel stock published by Hatayama et al. (2010). That analysis put the total 2005 steel stock in 42 major countries (after doubling since 1980) at 12.7 Gt, with 60% embedded in structures and about 10% in vehicles (and with US stocks at 2.7 Gt and Japan’s stores at 1.1 Gt). Raising the global total by about 20% in order to account for stocks in countries not counted in the analyses would bring it to 15.2 Gt, almost identical to my estimate.