Prices and cycles: Booms, super-cycles, money and the dollar

5.1 Business cycles, booms and super-cycles are different claims

A business cycle lasts a few years. A boom, for Radetzki and Wårell, is a sharp, simultaneous rise in real prices across a broad group of commodities. They count three since 1945, beginning in 1949, 1973 and 2004, each triggered by unusually fast economic growth that producers could not meet. The first, linked to the Korean War, was over by 1952. The second, fed by crop failures, oil cuts, inflation and the end of Bretton Woods, peaked in early 1974. The third lasted far longer.

A super-cycle is a larger claim: an upswing of 10 to 35 years and a full cycle of 20 to 70, driven by the industrialisation of a major economy. Erten and Ocampo filtered real prices from 1865 to 2010 and found four super-cycles of 30 to 40 years, with amplitudes 20-40% above or below the long-run trend. For non-fuel commodities they ran from 1894 to 1932 (peak 1917), 1932 to 1971 (peak 1951), 1971 to 1999 (peak 1973), and from 1999 onwards. Dates vary by series: for metals alone they find 1885-1921, 1921-1945, 1945-1999 and 1999 onwards. Their tests found world output leading non-oil prices, so those cycles are demand-driven. For oil the causality ran the other way, from prices to output.

Radetzki and Wårell are sceptical. They see the third boom as an extended investment cycle: prices stay high until enough capacity arrives, which in a simple model caps a boom at 12 to 15 years. They date the end of the third boom to about 2015, when large new capacity, notably in iron ore, came on stream.

• Radetzki & Wårell (2021), A Handbook of Primary Commodities · pdf pp. 127-140
• Erten & Ocampo (2012), Super-cycles of commodity prices, UN DESA WP 110 web ↗ · pdf pp. 3-4, 7-12, 24-25

5.2 The China super-cycle and its unwinding

In early 2005 Alan Heap at Citigroup argued that metals had entered a multi-decade super-cycle, driven by materials-intensive growth in an urbanising China and by high-cost new supply. Iron ore contract prices rose 71.5% for the year from April 2005, when the largest previous rise had been 19.5% in 1980; coking coal rose 119%. Between September 2005 and May 2006 the monthly copper price more than doubled, from $3,857 to $8,044 a tonne.

The scale of Chinese demand is hard to overstate. Radetzki and Wårell put China at 17.1% of world GDP in 2015 but 124% of global copper demand growth between 2005 and 2016, 116% for nickel, 92% for aluminium and 37% for petroleum. Shares above 100% mean demand elsewhere shrank. Supply was slow: a large mine can take eight to ten years, and a tonne of annual copper capacity, long reckoned at about $5,000, cost $10,000-15,000 by 2010.

Then the investment cycle turned. Humphreys shows mining capital spending peaking in 2011 and 2012, a year behind prices. Copper mine output, which grew 1.7% a year over 2001-2011, grew 4% in 2012 and 8.5% in 2013; seaborne iron ore rose almost 10% in 2013. China's growth slowed from 10.3% in 2010 to 7.4% in 2014, and its steel consumption fell in 2014 for the first time since 1995. Profits of the 40 largest miners fell 49% in 2012 and 72% in 2013, and they wrote off $97 billion of assets. Humphreys still suspects a lasting step up, perhaps from cheaper manufactures, dearer energy and a weaker dollar.

• Humphreys (2015), The Remaking of the Mining Industry · pdf pp. 104-113, 146-150, 220-222
• Radetzki & Wårell (2021), A Handbook of Primary Commodities · pdf pp. 136-138

5.3 Financialisation: did index investors inflate the 2008 spike?

In the 2000s commodities were sold to pension funds as an asset class, helped by Gorton and Rouwenhorst's finding that futures had matched equity returns and moved against them. Money poured into passive index funds such as the S&P GSCI, which Humphreys puts at about 70% energy. He counts index-fund money rising from about $9 billion at end-2002 to $200 billion by mid-2008; Radetzki and Wårell's wider measure of financial investors' commodity assets went from about $13 billion in 2003 to $450 billion in 2011.

The returns came from three sources: rising spot prices, the roll yield, and interest on the collateral not needed for margin. Iwarson split the GSCI's 15% annual return over 1970-2005, measured in Swedish kronor, into about 6% from spot prices, mostly rising oil, 2% from roll and 7% from collateral. Radetzki and Wårell argue the comparison with equities was fallacious, since stock futures would earn the same collateral return.

Did the money move prices? The hedge-fund manager Michael Masters told Congress it did, and the CFTC's chairman agreed that index funds contributed to the 2008 bubble. The mechanism, though, is clear only in contango, where arbitrage passes futures buying through to spot. Much of the boom was in backwardation. Most studies, including IMF reviews in 2006 and 2010 and a survey of 100 papers, found little lasting effect; some found positions followed prices rather than leading them. Gilbert found brief bubbles, on 21 of 753 days in crude, and evidence that index buying inflated food prices in 2008. The strongest counter-evidence is simple: iron ore and minor metals, untraded on futures exchanges at the time, rose just as dramatically.

• Radetzki & Wårell (2021), A Handbook of Primary Commodities · pdf pp. 138, 162-170
• Humphreys (2015), The Remaking of the Mining Industry · pdf pp. 114-120
• Tilton & Guzmán (2016), Mineral Economics and Policy · pdf pp. 120-124
• Blas & Farchy (2021), The World for Sale · pdf pp. 233-235

5.4 Exchange rates: the dollar's weight in the world decides how much it moves prices

Most internationally traded commodities are quoted in US dollars, so a common claim is that a weaker dollar lifts commodity prices. Radetzki and Wårell show why the answer depends on the dollar's economic weight. Imagine two currencies, the dollar and the euro, and let the dollar double in value. If 99% of world activity is in the dollar area, supply and demand barely change and the dollar price stays put. If only 1% is, then nearly all producers and consumers think in euros, the euro price must return to its old level, and the dollar price halves. Reality lies between. Because the United States' share of the world economy has shrunk since the 1950s, dollar commodity prices have become more sensitive to the dollar's exchange rate.

For freely traded prices the currency of quotation does not otherwise matter. It matters where prices are sticky: in bilateral contracts set after long negotiations, a seller who quoted in an appreciating currency gains.

Currency also works through costs. Humphreys notes that rising commodity prices lift the currencies of commodity-exporting countries, which raises their producers' dollar costs and steepens the cost curve. He suggests part of the 2000s step-up in dollar prices reflected the dollar itself: the trade-weighted dollar peaked in February 2002, slid for most of the decade, and in 2007 broke below its 30-year range. Geman goes further: for metals, currency movements are the main source of long-term price volatility. Copper priced in dollars is relatively stable because many costs are in dollars, while copper in euros inherits the volatility of the euro-dollar rate.

• Radetzki & Wårell (2021), A Handbook of Primary Commodities · pdf pp. 125-126, 133
• Humphreys (2015), The Remaking of the Mining Industry · pdf pp. 109, 113-114
• Geman (2005), Commodities and Commodity Derivatives · pdf pp. 188-189

See it in the table

Check your understanding

Erten and Ocampo find that world output leads non-oil commodity prices. What do they find for oil?

Why is the rise of iron ore and minor-metal prices in the 2000s evidence against the view that index funds drove the boom?

In Radetzki and Wårell's two-currency thought experiment, when does a stronger dollar push dollar commodity prices down most?