The Obama administration's latest climate regulations have sent supporters of anthropogenic greenhouse gas emission cuts on the prowl for conclusive evidence that making such emission cuts will not substantially damage the economy.

Arguments are being made that "historically, the demand for electricity was closely tied to growth in the economy; only recently have the two decoupled."  It is far too early to be making conclusive claims that American economic growth has become decoupled from electricity demand.

The following plot shows normalized real per capita GDP and per capita electric power consumption in the United States since 1960.

From 1960 to 2000, increasing electricity consumption closely tracked increasing economic growth.  But per capita electricity consumption peaked in 2000 and has been approximately constant since then.  The climate activists get excited when they see what happened between 2000 and 2006: economic growth was still taking place even though per capita electricity consumption was static.  Ergo, the U.S. economy decoupled from energy consumption.

Not so fast.  Look what has happened since 2006.  Both economic growth and electricity consumption are in almost perfect correlation, and both have flatlined.  That isn't decoupling.  Instead, the 2006-to-present data suggests that energy consumption and economic growth remain tightly coupled in the USA, and that neither of them is going anywhere fast.

There is a two-way causation on electricity consumption and economic growth.  Slower economic growth will generally result in less electricity consumption, and less electricity consumption will lead to slower economic growth.  Starting in 2005, electricity prices in the U.S. began their rapid increase up to the current astronomical values.  Between 1985 and 2004, the nominal average price of residential electricity increased by only 15 percent.  Since 2004, the price has increased by 40 percent.  In other words, the rate at which electricity prices have increased over the past decade is fivefold higher than the rate of increase during the preceding three decades.

From 1961 to 1973, the price of electricity was essentially constant in nominal terms.  Over this time-frame, average annual real per capita GDP growth was 3.3 percent.  Between 1974 and 1982, the price of electricity skyrocketed, and average annual real per capita GDP growth was only 1.0 percent.  Another stable period of power prices existed from 1983 to 2004, and annual per capita economic growth averaged 2.4 percent.  Electricity prices started rising quickly in 2005, and since then annual real per capita GDP growth has averaged only 0.6 percent.  Periods of higher economic growth have correlated with stable electricity prices (declining in inflation-adjusted terms) for more than half a century, and the data suggests that there is no reason to believe that this two-way correlation-causation has changed.

Electricity use does not perfectly track economic growth.  There are lag periods, as we would expect.  The economy does not immediately flip a switch on and off; sometimes it takes a few years for shocks to the system on one end to become evident at the other end.  This may be what has taken place with the relationship between electricity consumption and economic growth.  The short-term appearance of decoupling from 2000 to 2006 has been superseded by a period of recoupling from 2006 to the present.  It is foolhardy to use this six-year anomaly from the first half of the 2000s to make multi-decadal economic and energy policies based on the assumption (which, at present, appears to be incorrect) that the American economy has permanently decoupled from electricity use.

The quantity of carbon dioxide emissions per unit of GDP continues to decline in the U.S., but at an ever slowing rate.  In 1980, it took 0.20 kg less CO₂ to produce a dollar of GDP (in constant 2005 terms) than it did in 1970.  By 1990, the decline was a further 0.22 kg of CO₂ per dollar of GDP.  But during the 1990s and 2000s, the rate of decline slowed to 0.08-0.09 kg of CO₂ per dollar of GDP per decade, or less than half the rate of decline seen during the 1970s and 1980s.  It is getting harder to further lower carbon dioxide emissions per unit of GDP.

The "northeastern cap-and-trade program known as the Regional Greenhouse Gas Initiative (RGGI), which first put in a carbon cap in 2009" is also being touted as evidence that such agreements have reduced carbon dioxide emissions in some regions.  The following nine states are in the RGGI: Connecticut, Delaware, Maine, Maryland, Massachusetts, New Hampshire, New York, Rhode Island, and Vermont.  At the New York Times, this graph was recently shown as evidence that the RGGI has reduced carbon dioxide emissions among its member-states.

Astute readers will see the interpretation problems surrounding this figure.  Carbon emissions from electricity in the "nine northeast cap-and-trade states" were already in steep decline prior to the RGGI starting up in 2009 – with the decline starting in 2005.  Actually, the rate of decline between 2005-2008 among the RGGI states looks about the same as the rate from 2008-2013, as do the pre-/post-2008 decline rates for the "other 41 states" without a cap-and-trade system.  Wouldn't this be evidence that the RGGI has failed to reduce carbon emissions from electricity relative to the non-RGGI states, rather than being supportive of the program's effectiveness?

British Columbia also has a carbon tax, and I've seen the same type of problematic arguments in support of its effectiveness.  If regions already have declining greenhouse gas emission trends before carbon pricing comes into force, it is almost impossible to use various approaches (e.g., regional comparisons, econometrics, etc.) to prove that carbon pricing had an effect if the trendings for the pre- and post-carbon pricing periods are similar – as we see above among the RGGI states.

The NYT figure above also includes data for 2012 and 2013, and cites the "Energy Information Administration" for its source.  However, the EIA appears to include data only up to 2011 in its most recent data release from February 25, 2014.  I sent the author an inquiry, and she indicated that the data comes from ENE, "a research and advocacy group based in Boston."  Thus, I consider the 2012 and 2013 data from the NYT article unverified and speculative until published directly on the U.S. Energy Information Administration website.

Comparing total emissions between the RGGI and non-RGGI regions is also not an ideal way to examine whether carbon pricing has had any impact on individual or collective behavior, since this approach fails to account for differential population growth between the areas.  The non-RGGI regions saw population growth of 2.6 percent between 2008 and 2011, whereas the cap-and-trade RGGI region experienced much lower population growth of only 1.6 percent during this time-frame.  If you don't normalize emissions trends for these two different rates, the carbon pricing conclusions will be flawed.  Perhaps population growth is lower in the cap-and-trade region because people are fleeing – and staying away from – the carbon pricing system?  We could call them carbon pricing refugees.

When carbon emissions from electricity are normalized to population, it appears the RGGI region's per capita emissions trend didn't change for the 2005-2008 versus 2008-2011 periods (see the dashed lines in the figure below; solid lines are actual year-to-year data).  In other words, this suggests the cap-and-trade program had negligible impact on behavior.  In contrast, the per capita emissions trend for the non-RGGI region declined much more rapidly during the 2008-2011 period relative to 2005-2008, a finding also not in favor of the RGGI's supposed impacts.

If the RGGI were truly successful at reducing emissions in its region, we might expect to see a greater rate of reduction in per capita emissions after the program came into force when compared to the pre-program trend, and also see a greater post-versus-pre-program difference relative to the trends outside the RGGI zone.  Instead, we see neither of these criteria being met using the available data.

One might conclude that – to date – the RGGI has just put a price on carbon (i.e., added a new tax) without substantially altering the desired behavior.  Not surprising.  It is simplistic to assume that taxing an activity/entity will automatically reduce its occurrence, particularly for products as fundamental to modern society as energy.

Stepping back and looking at per capita carbon emissions from electricity in the RGGI and non-RGGI regions since 1980, we see that emissions peaked in the non-cap-and-trade region in 2000 and have began a steep decline of late down to levels last seen in the early 1980s.  And this reduction occurred absent any carbon pricing.  In the RGGI region, wide variability in emissions occurred between 1980 and 2005, but the steep descent beginning in 2005 – and perhaps as early as the late 1990s – didn't require any carbon taxation, either.

Based on what I am seeing, the RGGI looks like a waste of money.  It is attempting to fight a problem whose real risk is poorly defined and potentially trivial (aka anthropogenic climate change), and it doesn't appear to be overly effective at contributing towards its stated goals of reducing carbon emissions in the region it applies to.  As best I can tell, carbon emissions were already on the steep decline in the RGGI region, and this program hasn't increased the rate of decline.  All it appears to be doing is artificially increasing the cost of energy for residents and businesses.

The proponents of carbon pricing also claim – as the NYT article's title exemplifies – that we can have the "Best of Both Worlds," namely rapidly declining carbon dioxide emissions and unhindered economic growth.  This sounds suspiciously like getting a free lunch, which the Second Law of Thermodynamics continually reminds us is impossible.  One wonders how much better the economy of the RGGI region would have performed since 2009 had it not been encumbered by unnecessarily high energy prices.