The Disconnect Between GDP And Energy In The United States

On Friday I put up the following chart, and invited readers to comment on what's wrong with it.

From Bloomberg's Sustainable Energy In America 2013 Factbook. Due to the massive government stimulus of 2009, both GDP and energy consumption grew together in 2009-2010. As the stimulus petered out, energy consumption fell off dramatically but GDP continued to grow (black circle).

For your convenience, I have circled the suspicious anomaly—economic growth has radically departed from growth in energy consumption since 2010. That trend flies in the face of everything we know about economic expansion and energy consumption—they are inseparable, at least on a global scale. I am going to assume you have read and understood my descriptions of Tim Garrett's work, which I most recently explained in Is Global Economic Growth Persistent?

The Bloomberg report chalks the divergence up to great energy efficiency in the United States. That's a happy story, but the evidence which supports it ranges from thin to imaginary. I will not argue today that the weak efficiency story is false. (I'll comment on that later this week.) Instead, it behooves us to explain the divergence in another, more satisfactory way.

In Garrett's work, the consistent relationship between economic expansion and the increasing energy consumption which supports that expansion hold on the global scale. We might expect there to be some small local exceptions to the rule in places like British Columbia or Sweden, but the sheer size of the United States economy suggests that the constant relationship between the two trends should hold (more or less) in America as well. (We have to consider trends like the offshoring of manufacturing to energing markets, which effectively moves energy consumption to other countries.)

If we look at the long chart, we see that there has been a GDP-energy dependency since 1845.

Note the contraction in both GDP and energy consumption during the Great Depression (condensed scale on the left). Click to enlarge. Source

Thus we see that the divergence between GDP and energy over the last few years is unprecedented in the United States. That departure strongly suggests that GDP is artificially inflated. And we don't have far to look to see how this has been accomplished. Consider the next two graphs.

From Mike Shedlock's Government Spending as Percentage of GDP (August 28, 2012). Chart by Doug Short of Adviser Perspectives

From the Washington Post's A scary graph from Goldman Sachs, Ezra Klein, the WonkBlog (February 5, 2013).

Both graphs illustrate the simple, uncontroversial conclusion that the American economy (and thus the GDP measurement) has remained highly dependent on extravagant deficit spending, even after the American Recovery and Reinvestment Act (2009) ran its course. This suggests that GDP growth in the absence of growth in energy consumption is largely a monetary phenomenon, for the government basically (outside of some defense spending) writes checks, which they make good with borrowed money. The Federal goverment does not manufacture things, it typically does not distribute goods, it does very little which is physical, as opposed to monetary. By contrast, there was a large physical ("shovel ready") component in the 2009 stimulus.

You might read my recent post Look At The GDP Fraud! We might go so far as to say that every GDP print in the last 2 years has been fraudulent in the sense described.

If government spending follows the dashed line in the Goldman Sachs graph—this is called "the sequester"—then we can reasonably expect GDP to follow energy consumption, as it has every year since 1845. GDP would thus decline in a most alarming way. I'll quote from Ezra Klein's post, the source of the 4th graph above—

“This is a large decline in historical context, but it is not without precedent,” [Alec] Phillips [of Goldman Sachs] continues. “This would mark the third decline in the last 50 years; the first occurred around 1970, after Vietnam War spending had peaked, followed by another around 1990 as military spending declined following the end of the Cold War and multiple rounds of spending caps were enacted.”

Even if there’s precedent for this kind of a drop, there are at least three reasons to be particularly concerned about it happening now. First, the previous periods of austerity that Phillips mentions were primarily driven by the end of major wars. That meant that a lot of that spending was not directly raising living standards in the United States, and so its absence didn’t have the kinds of consequences of, say, the sequester.

Second, this graph misses the significant tax-side austerity we’re engaged in, with the expiration of the payroll tax cut and the high-income tax increases from the fiscal cliff deal serving as an additional drag on growth. And third, the economy remains unusually weak by historical standards, and this kind of fiscal drag is going to make it that much harder to recover.

Indeed, if we look at energy consumption (in quadrillions of BTUs), we see just how weak the private American economy truly is.  It is total nonsense to use an "improved efficiency" argument to legitimize the disconnect between GDP and energy consumption in the United States. And if the arguments I've made today are correct, and the spending cuts are put in place, we can reasonably expect GDP to track energy consumption in the future, all other things being equal, just as it did during the Great Depression. In short, we can expect GDP to decline, just as those who fear "the sequester" say it will, unless a miracle occurs.

I shall comment further on "efficiency" and declining energy consumption some other time.

Comments

[Alexander Ač]

Dave,

I think your observation is correct. It is also 100% given that "efficiency happy story" will be pushed be all kinds of propaganda for a long time... just as Germany is touted as "green-hero (follow-US) economy" in Europe these days by few confused environmentalists and politicians alike...

Alex

[Oliver]

Dave, many thanks for the explication. I am ill-equipped to fully understand the true vs. fictional functionality of these graphs, mostly because I have managed at long last to cleanse my mind of the filthy pollution I ingested studying Economics several eons ago. So I can only make a general observation regarding the proclamation of fantasy GDP figures by so-called officials.

Here goes. When a prostitute passes her prime, she can only hang on to her profession by ladling on more and more make-up.

I sincerely apologize to prostitutes for conflating them with government officials.

[Alexander Ač]

Also small comment - apart from declining energy consumption, there was also increasing shale oil/gas production with relatively low EROEI, so the NET energy picture is a bit worse than shown... but I think everybody here understands that,

Alex

[Jim]

Hi Dave,

Firstly, thank you. This is stuff you just can't find these days. You correctly inferred that I am new to your blog, but I also believe sources like this should be supported. I'll be donating this week.

I do think the recent divergence can be largely explained by governmental spending, but I also suspect we've greatly exceeded what we normally would have achieved in GDP since 1980 because of the combination of public and private debt. On the IER chart, look at the growth rate from WWII to 1985, look at the rise from 1985 to now, and compare it to charts of public and private debt during those periods. It is literally stealing from the future.

I'm looking forward to the energy efficiency debunk. Thanks again.

[xavier durant]

There is a simple and obvious explanation IMO. All energy consumption is not related to GDP. Some energy consumption is frivolous and produces nothing. I'm thinking of recreational driving, heating homes warmer than needed and driving a gas guzzler SUV or empty pickup.

Evidently the U.S. has reduced some of this kind of nonsense. The result is entirely consistent with the chart. Some of the saved energy went to increase GDP and the remainder reduced overall energy consumption.

Simple really.

[Dave Cohen]

@xavier

You simply made that nonsense up off the top of your head, just before and as you were typing your comment, right?

And you don't have a shred of evidence to back it up. And there's 165 years of data which shows you are wrong. But you don't need any evidence, do you?

And you didn't really care what I wrote, am I right?

I am issuing a Monkey With A Keyboard alert.

-- Dave

[Alexander Ač]

Xavier,

"All energy consumption is not related to GDP." - yes, it is, since GDP by definition *IS* consumption (spending of money). *Saved* energy cannot go into GDP, precisely because it is saved. If you have money on your account, and do not spend them, they are not contributing to GDP... (if your money in bank is not used for some kind of ponzi-scheme, of course...)

Alex

[Jim]

Oof. There is something wrong, though. Compare the dates on on the Bloomberg chart to the IER chart. The IER chart only goes to 2001 - why does it show the same divergence as 2010 on the Bloomberg chart?

[Jim]

I see it. It's a combination of things. The IER chart counts wood, but shows it dropping to zero in 2001 - that explains a large part of the energy drop. It's possible the just stopped counting it as an energy source that year. The two charts look so much the same because the Bloomberg chart compresses the GDP (both measured a real instead of nominal) scale - it goes to 14 instead of 10. And 2001, even though it had a recession in the middle of the year, technically grew in GDP by a tiny fraction for the year as a whole. Sorry for the multiple comments, will try to avoid.

[JohnWDB]

@Xavier,
A couple of points.

1) While certain complex things may seem conceptually simple once they are understood, never mistake your perceived simplicity as evidence you understand them. It doesn't work both ways. One can't solve a problem simply by declaring it to be facile. Be especially wary of this when you have done minimal research. Also look up "The Dunning Kruger Effect" to understand why we tend to think we understand things that we don't.

2) We can see from your comment that you don't understand what GDP is. It isn't "productivity"; it is spending. Do some background reading. GDP doesn't care whether spending is wasteful or not, it only measures spending, and the more spending, the higher the GDP. Frugality/prudence/saving will always lower GDP, all things equal. That doesn't mean frugality is bad; quite the opposite--it is essential. What it does mean is that your explanation, though hopeful, cannot be correct.

[Mike Roberts]

The disconnect has certainly been marked in the last few years, for the reasons you mention, but I think the divergence started in the early to mid-eighties, from looking at how the correlation shifted then. If I recall correctly, that's roughly the time that you think the empire's decline started. Seems about right. It would be interesting to see a time line of when various fiddles were added to both the inflation calculation and the GDP calculation, and perhaps an adjusted GDP line. I don't know if ShadowStats has it right, but their chart of GDP growth (GDP data requires subscription) since 1980 seems to gel with the divergence.

[Makati1]

GDP is increased when you get divorced, die, have a child, pay a prostitute, get married, total your car, buy a gun, etc. If money is involved, it increases GDP. (No, your house being underwater does not decrease GDP. That was money already spent and the sale will only incur more money changing hands.)

[Cab]

I don't see that 165 years of prior history is 'proof' that energy consumption is, and always will be, tied to GDP. We live in strange times where software is replacing real hardware and young people would rather spend their evenings typing at keyboards than 'cruising in T Birds' (or whatever they used to do in the US - I'm from the UK). If the entire population gradually shifts to eschewing travel in favour of Skyping, and spends its entire existence in small apartments living off vitamin pills (who says it can't happen?), I would expect energy consumption to drop, but I would also expect to see a re-definition of money and what it can buy. In nominal terms GDP could well grow, but what the tokens were being spent on would be different from the old days. As an example, I understand that these days, real money is actually spent on virtual goods by people 'living' in virtual worlds, eating only processed food cooked in microwave ovens. Anything is possible.