Sustainability 2011 , 3
1915
If I-O tables were redone to distribute consumption costs to the businesses paying for them, patterns
likely to persist are (1) high intensity sectors would still have the most varied energy intensity, (2) the
largest sectors would be close to, but below average and (3) the energy producing sectors would have
somewhat below average energy intensities. The energy sectors do consume lots of energy, but it also
has a high value added, and so the ratio of the energy used to the value produced is low.
Another problem for applying I-O table data to estimating particular energy needs is how the data
reflects only national energy accounts. Most products and services have substantial global content. For
example, a great deal of the high energy using production for products consumed in the US is now
performed overseas, particularly in Asia. EIA data shows energy use in the US beginning to level off
starting in the 1970’s, even as US GDP and consumption continued to grow [22] (see Figures 3,4).
That rapid divergence between US energy use and GDP is the complete opposite of the consistent
world GDP and energy use relationship over the same period. The global trend has been of smoothly
growing GDP and GDP/btu [22] (see Figure 1) in constant proportion. Explaining why the global data
shows such smooth global trends, and national accounts do not, has been argued as a statistical fluke or
the averaging of random variation. Our expectation is that it is a result of the global economy working
smoothly, to allocate resources according to the comparative advantage of productive differences for
individual business communities around the world, as free market theory has always suggested
it should.
1.5. Background on EROI for Wind Turbines
Kubiszewski et al. [14] performed a meta-analysis to summarize the net energy of wind turbines
based upon a suite of previous studies of 114 calculated values for EROI (see Figure 2). The wide
spread of the data shows evidence of large inconsistencies in the methods of defining which energy
inputs to count. The variation is over an order of magnitude with reported values from over 50:1 to
near 1:1. The average EROI for all studies was reported at 25:1 although the average for operational
LCAs (those based upon actual performance of a turbine) was lower at 20:1. There is as yet no
national account data for a wind energy sector industry group to compare. Wind utilization estimates
varying from 15% to 50% also add to the inconsistency in assumptions presented.
Kubiszewski et al. described process analysis methods, including LCA, and compared them with
studies using I-O table data. The former showed an average EROI of 24:1 while the latter had an
average EROI of 12:1, a difference some attributed to how process analysis involve a greater degree of
subjective decisions [14]. The differences in capacity factors and from omitting the untraceable energy
needs of labor and business services required [23,24] would seem to account for the variation. These
results are comparable to the method presented here only as studies of business-scale energy use for
which there are no industry group studies.