Abstract
Although air and water may be more critical short run needs, commodities underpin our very existence. They provide sustenance, shelter, and material goods. They are the backbone of industrial prowess and essential drivers in our war on global energy poverty and essential in the provision of a sustainable energy future. Of late, pestilence, energy transition, war, deglobalization, and other factors are again focusing our attention on these essentials. Fears of fossil fuel shortages may be replaced by rising mineral demands and fears of critical mineral shortages. The studies in this book seek to educate us on commodity markets, including energy, metals, and agricultural products. The book consists of an executive summary and introduction along with four chapters that consider the history of these markets, econometric estimates of demand for these commodities, what drives commodity price cycles, and the cause and effects of commodity price shocks. It can be read sequentially, chapter by chapter, or used as a resource for individual topics.
Commodities have seen dramatic growth over the last century with shifts in relative shares and geography. Chapter 1 by John Baffes, Wee Chian Koh, and Peter Nagle is my favorite chapter. It takes a high-level look at the evolution of commodity markets, in many cases for over a century. Aggregates as well as representative products are included for energy, metals, and agricultural products. The 24 figures in the chapter, each with 3 to 6 subfigures, provide a wealth of visual information on these markets. The narrative highlights important features in the figures as well as technological, institutional, trade, main uses, policies, and economic factors that might be driving the market changes.
The authors outline some of the policies related to commodity shocks that have developed and note that recent countercyclical macro policies have tended to be more successful than other policies such as subsidies and trade restrictions. They divide the policies into three categories: “policies to manage the macroeconomic impact of commodity price fluctuations, “structural policies to reduce vulnerability to commodity price fluctuations,” and “policies to moderate boom-bust cycles.” Their discussion includes pros and cons of a variety of policies including fiscal and monetary policy, capital flow management, market mechanisms such as futures and options contracts, sovereign wealth funds, economic diversification, and the environment.
Their graphs show the stated policies fall far short of the net zero by 2050 (NZE) needs for reductions of fossil fuel use and increases in biofuels and renewables. Of interest to many will be their discussion of the increased needs for metals needed for the energy transition. For example, natural gas requires around 1000 kg metals/MW while solar needs more than 2000 and wind around 3000. They cite the key metals needed for the transitions including aluminum, chromium, cobalt, copper, iron and steel, lithium, manganese, molybdenum, nickel, platinum group metals, rare earth metals especially cadmium, indium, and neodymium), silver, titanium, and zinc. The main influence of the energy transition on agricultural commodities will be biofuels (ethanol from corn and sugar cane and biodiesel from canola and palm oil).
They discuss energy transition challenges including the fear of stranded assets along with insufficient investment. For example, such fears may lead to insufficient investment in abundant fossil energy to support the transition. At the same time, there may be insufficient investments in renewables. The authors show the investment needs in both oil and gas and in renewables under various IEA scenarios for the transition including NZE. Under NZE, they show carbon emissions by sector, biofuel production, and hydrogen use for hard-to-decarbonize sectors—shipping, heavy vehicle road transport, aviation, chemicals, cement, and iron and steel. They note other difficult hurdles including opposition to removal of existing subsidies for fossil fuels and challenges of reducing agricultural emissions such as methane from animals and rice paddies.
For the newcomer, there is a short but interesting discussion on the history and use of price benchmarking in commodities as well as a section on supply management agreements. They show prices when agreements were in place and not, and generally conclude such agreements likely influenced the price.
John Baffes and Peter Nagle consider the drivers of commodity demand by major group (energy, metal, and food) in chapter 2. Their discussion supplemented by numerous graphs gives us an historical snapshot of growth by major commodity. I found the high-level discussion and citation of other drivers including industrialization, urbanization, technical change, and policy interesting. Most historical examples are well known. (e.g., sails to steam and horses to internal combustion engines). However, newer examples of changing commodity needs, especially relating to the energy transition, will be a nice introduction for some readers.
They follow their overview of market changes with econometric estimates for demand per capita for the three commodity groupings along with some forecasts. They use a pooled mean group autoregressive distributed lag estimation technique (Pesaran, Shin, and Smith (1999)). They aggregate 62 countries for energy, 43 countries for metals, and 77 countries for food on data from 1970–2019. In this technique, long run estimates in the equation are constrained to be the same across countries, while those in the short run are allowed to vary. They use log variables but allow income (GDP per capita) elasticities to vary with income by including income squared. They find income and income squared highly significant, with the income elasticity falling at higher levels. At lower incomes, metals have the highest income elasticity, while food crops have the lowest. Interestingly, the rankings change at higher incomes with both metals and food income elasticities becoming negative. They include fixed effects, a time trend, and make vague reference to various included country specific characteristics. They mention proxies included for urbanization and industrialization but do not include any results.
Although they find some interesting results and trends, I would be inclined to treat the results on such aggregated data with caution. This recommendation for caution is underscored by the positive and highly significant positive price elasticity.
Commodities are notable for their price volatility. Whether the price changes are temporary and cyclical or permanent are of interest to policymakers whether in importing or exporting countries. The policy response is likely to vary depending on the source of the price shocks: global vs. commodity-specific events, driven by demand or supply shifts, temporary or permanent. Alain Kabundi, Garima Vasishtha, and Hamza Zahid seek to provide such information by considering commodity price cycles in chapter 3. They set the stage by presenting price indices for major commodity groups with a discussion of a variety of commodity price shocks ranging from weather and policy shifts to Covid 19.
Their first statistical analysis uses a frequency domain approach (Corbae, Ouliaris, and Phillips (2002) and Corbae and Ouliaris (2006)) to separate price changes into three cyclical components: < 2 years, 2–8 years, 8–20 years, and a long run permanent component apparently represented by a time trend. They catalogue the number of cycles as well. Their data is monthly from 1970 to 2019 for 27 commodity price series and for the following six groupings (3 fossil fuels, 6 base metal, 3 precious metals, 6 annual agricultural products, 5 perennial agricultural products and 4 fertilizer products). They find transitory and permanent components each tend to on average explain about half of the price changes. However, there is considerable heterogeneity on which cycles tend to have the largest influence for each commodity or grouping. Micro-economists may be somewhat uncomfortable with the lack of a structural model and readers will need more explanation and more familiarity with the methodology than I possess to determine their quality.
The second set of statistical analyses is a factor-augmented vector autoregressive (FAVAR) model on aggregate global monthly data (January 1970-September 2021). This model is used to determine whether various shocks across time are demand shocks, supply shocks or shocks from a common global factor. The three global variables in the FAVAR are growth in industrial production, inflation, and an unclear common commodity factor. Identification of the shocks comes from the signs of an inverted impact multiplier matrix. Although this analysis likely has more appeal to economists because it investigates drivers of price shocks, again familiarity with the methodology will be required to understand the derivation and quality of the results.
In chapter 4, Alain Kabundi, Peter Nagle, Franziska Ohnsorge, and Takefumi Yamazaki carry on the theme of causes of industrial commodity price shocks but also consider the consequences especially for emerging markets and developing economies (EMDEs). Their analysis is limited to energy and metals. They begin their discussion with a cross country comparison of dependence on commodities for export earnings and government revenues. Out of 153 countries considered, in 62 oil exports account for 5% or more of export revenues, and in 58 industrial metals and ores account for 5% or more. For such countries, commodity price movements and terms of trade, especially for the oil exporters, likely contribute to business cycles.
The authors do two sets of statistical analysis - structural vector auto regressions (SVARs) on global data to identify commodity demand and supply shocks and a local projection estimation technique at the country level to measure the effect of shocks on local GDP. The seven SVARs estimated, one for oil and one for each of aluminum, copper, lead, nickel, tin, and zinc, are on monthly data from February, 1996 to July, 2021. The industries are represented by the global growth in the product’s price and the global growth in production, whereas the global economy is represented by growth in industrial production. They found demand shocks for all seven commodities tended to move together, whereas supply shocks were much more commodity specific.
Using their impulse response function from the SVARs, they trace the effect of a 1% supply or demand shock for the 18 months after the shock. They found demand shocks tended to be larger in their effect on price variability, but effects tended to die out sooner than for supply shocks. Forecast error variance analysis showed price variability largely dominated by demand shocks.
Another set of statistical analyses considers the effects of commodity price changes for oil (estimated from the average price of West Texas Intermediate, Brent and Dubai), for aggregate metal prices (measured as a trade weighted export average of aluminum, copper, lead, nickel, tin and zinc), and for aluminum prices on country and global economic activity. To set the stage, they identify and discuss periods of price jumps and collapses for aggregate metals and oil defined as price changes that are higher than 20% over a six month period. Such jumps tended to happen before global recessions and during recovery. Price shocks for metals happened more often but were on average smaller in magnitude.
Their statistical analysis applies a local projection estimation technique developed by Jordà (2005) with some later modifications to measure response of a country’s GDP growth to commodity price changes. They use annual data from 1970–2019 for 153 countries. Growth in a country’s GDP is a function of the growth in the commodity price, lags in the country’s growth in GDP, global industrial production, and the country’s inflation rate. Subscripts in the methodology appendix suggest they run separate country estimates to derive impulse response functions (IRFs) for each country and commodity group (oil and metals) as well as separate estimates for aluminum importers and exporters. They use the IRFs to show the effect of a 10% increase in commodity price on inflation for the next 5 years for six aggregate groups: EMDE oil exporters, oil importers, metal importers, metal exporters, copper exporters, and copper importers. It is not clear how they aggregate the individual country responses or compute the confidence intervals shown. They find output gains for the EMDE commodity exporters from price increases were shorter and smaller than output losses from price decreases. The effects on importers tended to be smaller and more delayed.
The chapter ends with some policy implications. Government fiscal policy of spending the windfalls during booms and cutting back during price declines exacerbates commodity price cycles. Sovereign wealth funds could be a more stabilizing alternative as would moves towards more diversified economies.
The descriptive material requires no prior familiarity with energy or metal markets and provides a broad overview or introduction to these markets and their evolution over time. For Energy Economists, the discussion of oil many be familiar. What I found most appealing was the historical overview of the commodity markets, the discussion of needed materials for the energy transition along with the policy discussion relating to commodities. The statistical analysis presented interesting questions and results. However, familiarity with the methodology will likely be needed for their full understanding.