This post is the second part in a series. The first part is here.
Techno-pessimists are the loyal opposition of techno-optimists. They keep us honest by forcing us to be rigorous in our optimism, while joining us in our call for policies to accelerate the rate of progress. In response to a wave of techno-optimist posts, the blogger at Applied Divinity Studies expressed skepticism and demanded rigor, arguing that stagnation should still be our default outlook. I plan to address four of the blogger’s basic arguments (paraphrasing heavily):
We’ve picked the low-hanging fruit of science
Productivity has been slowing down, why should it accelerate now?
Solar, batteries, and other green energy tech isn’t for real
Life expectancy is stagnating
In an earlier post I explained why life expectancy A) isn’t stagnating, and B) isn’t a good measure of technological progress in the first place. Now I’ll explain why solar, batteries, and other green energy technologies are a big deal, not just technologically but also economically.
Slowly, then all at once
The Applied Divinity Studies blogger posts a chart showing the amazing decline in solar power costs during the 2010s:
Observing this chart, the blogger asks: If solar cost declines are going to increase productivity growth in the 2020s, why didn’t they do so in the 2010s?
Can you show me where The Great Stagnation was? Maybe there was one from 1990 to 2006, but you can’t reasonably look at this graph and infer that 2006-2017 was a bad decade.
And yet somehow, the Stagnation Hypothesis was convincing in 2012 as the price dropped from $1.68 $/W to 0.88 $/W. And it remained convincing in 2017 as they dropped further from 0.55 $/W to 0.46 $/W.
What exactly about the last year feels compelling in a way that previous progress was not?…How is this “Optimism for the 2020s” as Noah writes or a “Crack in the Great Stagnation” as Watney suggests if it’s been going on for the last 10 years?
It’s a good question. In fact, the decades since 1976 have ALL been amazing years for progress in solar technology (though the 90s and 2000s less so). But before we can start seeing that progress in the aggregate statistics, we need two things:
We need solar to be installed at large scale, and
We need solar to be not just cheaper than fossil fuels, but substantially cheaper.
Progress in solar technology has gone through several phases. A 2018 paper by Goksin Kavlaka, James McNerneya, and Jessika E. Trancik chronicles the first two of these. In the first phase — roughly, the mid-1970s through 2000 — solar power was mostly confined to the laboratory, since it was too expensive to use as a power source for industrial or commercial applications. During this time period, solar cost drops were driven by government research efforts.
Around 2001, things shifted, when solar got cheap enough that companies started to deploy it at a big enough scale for learning curves to start having an effect. As more solar was installed, companies learned how to do it more cheaply, and took advantage of more economies of scale. That kept the cost declines going. And as cost declined further, more solar was installed. This process is still going to this day.
As solar got cheaper, the amount of solar power installed in the world increased more-or-less exponentially (remember that an exponential is a straight line on a log scale):
The third phase of solar was when it started to become an appreciable fraction of the world’s electric power generation. This was sometime in the 2010s, depending on how large you think “appreciable” is:
But notice that even as of 2019, solar was still just a small fraction of total power generation — about 3% of the global total. You can’t make an impact in the aggregate productivity statistics by changing 3% of global electricity consumption to a different source. That is why the amazing technological progress in solar hasn’t yet translated to faster economic progress. Technology is embodied. It doesn’t automatically change economics just because you invent it; you also have to build it.
But now it’s getting built. As many COVID denialists learned to their great dismay in 2020, exponential curves move fast. Bloomberg New Energy Finance’s graph above is probably way too conservative in its prediction of how much of our electricity we’ll get from solar, as Ramez Naam explains. The real growth rate of solar has tended to be more than 3 times as fast as BNEF forecast in the past.
An energy source that is 3% of electricity generation will not show up in the productivity statistics, but one that is 30% of electricity generation definitely will. And THAT is why solar cost drops didn’t show up in aggregate productivity statistics in the 2010s, but will definitely show up in the future. (Note: For those who are concerned about intermittency, realize that we’ll just overbuild for winter/clouds/morning/evening, and use batteries at night.)
Everything I’ve just said goes for batteries as well. Electric cars are just now hitting the point where they can compete with internal combustion cars in terms of usability and affordability, and the market for electric cars is just starting to take off. Notably, some carmakers are starting to cancel internal combustion models in anticipation of the switch.
Like I said: Slowly, then all at once.
But in order for solar and batteries (and other green energy technologies like wind and hydrogen) to really make a big impact in the aggregate productivity statistics, they can’t just replace fossil fuels — they need to be substantially better than fossil fuels ever were.
That looks likely to happen soon.
Cheap energy is what boosts productivity
Clean energy isn’t just clean; it also has the potential to be very cheap. We’ve had hundreds of years to improve and refine fossil fuel technologies; they’re not getting much cheaper. Fracking was an exception, but it’s a pretty modest improvement compared to what’s happening in renewables:
There’s really no end in sight yet for improvements in solar and batteries. Cost drops are continuing simply from scaling up, and new materials and technologies are on the horizon that could generate continued price declines per unit of energy.
In other words, solar and batteries have a long way to go.
And that’s important, because simply replacing fossil fuels with solar doesn’t automatically improve productivity by a noticeable amount. If you replace a coal plant with a solar plant that generates energy for 97% of the coal plant’s price, that’s good for the environment, and it creates some jobs, but it doesn’t really boost productivity much because the energy generated is only a tiny amount cheaper than before.
The real productivity gains happen when new energy technologies get much cheaper than old ones. If solar plants can produce electricity at half or a third of the (levelized) cost that we got with coal or gas plants, that’s a big deal economically. The productivity slowdown of the 1970s happened in large part because oil stopped being cheap; likewise, much cheaper energy would enable a productivity boom.
In fact, there’s a hypothesis that physical technologies are more effective than digital or other technologies at goosing the productivity numbers. I laid that hypothesis out in a previous post, which I will now quote from:
Another possibility is that “atoms” innovation, unlike “bits” innovation, unlocks the potential for more extensive growth; instead of simply making us more efficient or more creative at using resources, physical innovation may allow us to gobble up more of the world around us. (As Tyler Cowen and Ben Southwood put it, “many scientific advances work through enabling a greater supply of labor, capital, and land, and those advances will be undervalued by a TFP metric”). A 2013 retrospective on productivity growth by Robert Shackleton of the Congressional Budget Office suggested:
Finally, the sweep of the 20th century underlines the extent to which long-term TFP growth and economic growth in general have been influenced by the development of energy and transportation infrastructure suited to the expansion of suburbs.
Of course, this is a hypothesis; I can’t prove things will work this way. But even if this hypothesis is wrong, energy is a very important input into pretty much all production processes, and so much cheaper energy will improve productivity growth — as soon as solar and batteries get much cheaper than fossil fuels and are installed at sufficient scale!
A final thought: Sustainability is also real progress
So the two reasons solar and batteries haven’t yet produced higher aggregate productivity growth are that they haven’t been installed in large volumes yet, and they haven’t gotten so cheap that they dramatically lower the price of energy yet. But current trends give us every reason to believe that both of these things will happen.
But in addition to measured productivity growth, green energy has another benefit: Sustainability. They help prevent catastrophic climate change, for one thing. And also, because the supply of sunlight is not limited, solar frees us from the need to constantly invent new and better extraction technologies just to hold energy costs constant.
Neither of these things adds to current productivity, but they both add to expected future productivity — because future productivity won’t be nearly as high if we experience climate catastrophe and fuel keeps getting harder to extract. As Tyler Cowen writes in his book Stubborn Attachments, the future is very long, and if we value future humans, sustaining productivity over the long haul is more important than bumping it up for a moment. So sustainability really can be viewed as a form of progress, even if it doesn’t juice the statistics in the present.
Hi Noah. I have only one point of disagreement after reading this: Under your "Sustainability" section you state, "the supply of sunlight is not limited".
I must reluctantly disagree. Our supply of sunlight is most certainly limited ... to about 3 billion years of the Sun shining tops.
Of course, our problems are a lot bigger than sustainable energy supply long before the Sun becomes a red giant.
Guy who won't be around by then and who is OK with that.
Not to point attention to horrid comments, but Substack should have a word filter on comments, right? Probably should enable that feature 😬