The Growth Mystery (Warning: Graphs)

The following is a picture that has been haunting me for the past few days:

Another blog shows a graph that extends the phenomenon another 80 years into the past (although take this with a grain of salt since it shows no variation before 1890):

So what is it?  Basically, the graphs show US Real GDP (on a logarithmic scale) over time from the 19th century.  Surprisingly, despite all of the technological progress, drama, wars, and changes to our government during that time period, the growth rate hasn't changed.  The Great Depression? No problem, after some wild fluctuations, the economy returns to its regularly scheduled trend, as if it had never happened. It's as if history doesn't matter.

The first graph comes from the President's most recent budget, where he argues that the economy will return to normal trend just like it always has.  Of course, by this reasoning, it will return to trend regardless of what the President, or congress, or the Fed try to do.  It all doesn't matter in the long run.  GDP growth is driven by some invisible constant.

I am skeptical for two reasons.  First, it seems too magical.  Why would GDP growth remain constant?  Second, Real GDP is actually fake.  That is, nominal GDP (the amount in dollars) represents an actual amount of money.  But real GDP (RGPD)  represents nominal GDP (NGDP) corrected for inflation.  But I have little faith in measures of inflation.  How can you compare the value of a house with air conditioning to one without air conditioning, or this:

  with this: .

In an age of stagnant technology, real GDP might actually mean something.  But it is difficult to compare apples and oranges, and I think our life today is an orange compared the apple of 1890.  In fact, things are so different that the measure of inflation is even forced to change from time to time.  The standard measure of inflation is the Consumer Price Index (CPI) as measured by the Bureau of Labor and Statistics (BLS).  It is based on the prices of a "basket of goods".  But this basket includes things like airfare, computer software, and medical services that didn't even exist back then.  So how is that a realistic comparison?

I don't think that inflation is a totally useless concept, I just don't think it makes much sense for long term comparisons.  But the mystery of growth still remains.  If inflation and hence RGDP are made up, why do they end up giving us such a perfectly predictable growth rate?  Perhaps the way we calculate RGPD ends up reflecting something entirely different than overall productivity.  It might not make sense to say that the amount of value produced by the economy grows at 1.8%, but there must be something.

So, check out this next graph:

Population!  US population has grown at a constant rate of 1.3% annually since Europeans first arrived (I assume this graph doesn't include native populations living in the current US boundaries).  RGDP, as currently measured, seems to be perfectly correlated with population growth.  But wait, aren't the GDP graphs PER CAPITA graphs?  That is, GDP per person is growing at 1.8% and population is growing at 1.3%, so total GDP is actually growing at over 3% annually.  This is much more than the rate of population growth.  That means we can't simply say that productivity remains constant.  Still, it is something to think about.  Perhaps as the number of people connected by a single social system grows, the productivity of each person increases as well.  In any case, we now have two perfectly exponential graphs to deal with.  And how about a third?

Total energy use in the US is also growing at an exponential rate, about 2.9%.  This is quite close to the growth in total RGDP.  Also, notice that there is a drop off at the end in both energy and GDP graphs.  And it is not due to the recent recession.  The drop off in energy use started in 1980.  Plus, it is pretty obvious that energy use can't keep growing at 3% per year forever without some serious world changing technological advance.  If we took energy growth at only 2.3% (to make the numbers work out better), the US alone would be consuming all of the sun's energy within 400 years, and all of the energy in the galaxy within 2500 years:

That's a lot of energy.  But the fact remains that RGDP continued to grow at a healthy pace for a few decades after energy use started to drop off.  Maybe GDP used to be perfectly correlated with energy use until things started to get more energy efficient?  The efficiency of cars and trucks certainly went up over that time period:

Another thing to think about (are you sick of graphs yet?) is labor force participation.  It seems odd that per capita GDP growth would remain constant if population is growing at a constant rate, but the labor participation is not:

Notice the spike that happened starting in the 1970's.  Perhaps RGDP is correlated with energ use, but just before energy use started to decline the labor force participation rate started to go up.  This may have temporarily masked the effects of reduced energy use.  Now let me remind you.  If this story is true it implies that RGDP growth will almost certainly not follow the trajectory we have been on for the past several hundred years.  It will track energy use.  But this doesn't mean that we can't become more productive or lead better lives.  As I mentioned earlier, RGDP growth is a phony concept anyway.  We can't compare the value of items in an environment where technology is changing so fast.  Plus, if I am right about RGDP growth tracking something like energy usage + labor force participation, then advances in energy efficiency would show up as negative productivity growth.

In any case, it is likely that both energy use and population growth will follow something called a logistic curve.  These may appear to grow exponentially for a while, but eventually they start to flatten out:

So what do I make of all this?  Well, I think the conclusion the President draws from his graph is unrealistic.  Just because RGDP growth has been constant for the last 400 years doesn't mean it will keep growing that way forever.  It sounds funny to think about it that way, but that is how a logistic curve works.  Second, it may not be such a terrible thing for RGDP growth to level out because we can still enjoy any improvements that come from improved efficiency and technology.  RGDP doesn't seem to do a good job of measuring these things unless they are correlated with increased energy use.  Finally, my whole analysis could be way off.  I just started thinking about this seriously last night.  So I welcome your thoughts.