The Central Role of Cheap Energy in American Growth
Are the best days of the American economy behind us?
Some politicians, both actual and aspiring, seem to think so. This view is also supported by economists such as Robert Gordon, who argues in his new book "The Rise and Fall of American Growth", that the extraordinary economic and productivity growth experienced for much of the 20th century was largely due to a unique set of one-time technological changes that are not repeatable.
Many others, particularly from the tech sector, challenge this pessimistic view, arguing that future innovation will continue to expand productivity and power economic growth. Bill Gates emphatically rejects Gordon's conclusions pointing to dramatic gains in agricultural, medical, and material science that he believes will be as transformative as any prior technological change. Other optimists, such as Hal Varian, Google's chief economist, question productivity measures, implying that the standard statistics don't adequately capture how new communication and other technologies actually improve people's lives.
This article argues that both sides in this debate may be missing the mark by failing to adequately consider how energy and energy pricing fundamentally drive productivity and economic growth. When energy prices are low and stable, key decision makers substitute energy for labor (and capital) in ways that significantly improve productivity growth. Conversely, at times when energy prices are high and volatile, businesses and homeowners are have an incentive to invest in approaches that use less energy and drive down productivity growth. From this perspective, both sides generally ignore the real technology changes - involving horizontal drilling and hydraulic fracking - that have already led to dramatically lower energy prices and are likely to power both productivity and economic growth in the future.
The debate over productivity often seems to fall back on an economic concept called Total Factor Productivity (TFP) as a proxy to describe the underlying qualitative impact of innovation and technological change on productivity growth. For example, Gordon evaluates TFP in four separate time periods as follows: 1921-70 where average annual TPF growth was 1.9%; 1971-1993 where it fell sharply to .6%; 1994-2003 where it spiked back up to 1%; and 2004-2014 when it went down to .4%. Except for the brief period from 1994-2004, Gordon finds that productivity growth has not come close to matching the levels achieved from 1920-1970. Gordon argues that the spread of specific and unrepeatable innovations (such as electrification, cars, and sanitation) led to the rapid growth seen from 1920-70, but he projects that the current more narrow qualitative nature of technological change (involving computing, communication, and entertainment) will result in far slower future growth.
This article presents a very different perspective on the rise and fall of American growth, suggesting that changes in average TFP growth were driven primarily by substitution toward and away from energy in response to dramatic shifts in real energy prices and volatility, and not primarily due to qualitative differences in the type of technological change occurring in any one period. This linkage becomes clear by comparing average real oil prices (as a proxy for overall energy costs) in each time period and comparing them against average TFP growth.
For example, during the time period from 1920-1970, oil and energy prices were incredibly low and stable, averaging under $17/barrel over fifty years, and never rising above $20/barrel after 1926. This extraordinary low pricing and stability enabled businesses, homeowners, and other economic entities to substitute energy for capital and labor in profound ways that seem to have led to sustained TFP growth averaging 1.9%.
For the period 1971-1993, however, real oil prices roughly tripled averaging $53/BBL, with a high of $106/barrel in 1980. This sharp increase in real energy prices and volatility convinced businesses and others to substitute capital and labor for energy to manage energy bills in ways that likely help to significantly reduce average TFP growth (ultimately down to .6%). Government data also shows that oil consumption fell by 15% during the 1970s and into the early 1980s confirming this substitution effect.
Upon the conclusion of the Gulf War in 1991-92, real energy and oil prices declined significantly, averaging about $33/BBL from 1994-2003 with a peak of $39/BBL in 2000. Once again, lower and stable real energy prices led businesses, homeowners and other decision makers to start substituting energy for labor and capital through approaches that helped raise TFP growth to an average of 1%, with oil consumption increasing significantly after 1992 through 2003.
Rising energy demand from China and elsewhere after 2003 and through 2014, increased real energy and oil prices such that oil prices surged to an average of $104/BBL during this period. Energy decision makers again began substituting away from energy, reducing consumption by as much as 10%, and slowing down TFP growth to an average of .4%.
Academic work over the past ten years has tended to confirm the linkage between energy pricing, energy consumption, and economic growth. A 2010 book shows statistically that much of the past growth in the US economy has been driven by ever cheaper energy and not by technological progress in some general and undefined sense. The author argues that energy should be considered as a third factor of production (in addition to labor and capital) and that statistically US economic growth can be explained fairly well by these three factors, without resorting to largely undefined concepts like TFP that depend on qualitative analyses of technological change and innovation.
The energy-centric view of American economic and productivity growth allows for a more optimistic view of the future. It suggests that pessimists like Gordon have failed to grapple with one of the key technological changes of our time - the impacts on energy prices associated with the combination of horizontal drilling and hydraulic fracking. These new drilling technologies have greatly expanded the domestic supply of both natural gas and oil and have helped lead to reductions of energy prices on world-wide markets, with oil now trading in the $30-40/barrel range.
Based on historical experience, this alternative view of American growth that explicitly considers energy pricing and innovation suggests that both productivity and economic growth may ultimately be more likely to replicate the experience of 1994-2004 rather than the lower levels from 1970-1994 and 2004-14. Recent technological innovations in the energy sector are now producing relatively low oil and energy prices, which should once again allow key economic actors to substitute energy for labor and capital in ways that increase productivity growth.
In summary, this analysis suggests that movements in productivity growth are likely to be driven more by shifts in real energy pricing than by qualitative differences in the nature of technological change, which seem to be at the heart of the debate between pessimists like Gordon and the optimists coming out of the tech sector. This finding, when combined with the recent advances in energy drilling and recovery technologies, suggests a reasonably optimistic view of future American productivity gains and overall economic growth, but for reasons somewhat different than those put forward by either Gordon or the tech optimists.