“Obedient to constraint, I was compelled to submit.” – Mikhail Bulgakov
In Eliyahu Goldratt’s extraordinarily successful business management novel The Goal, we learn of a hapless and somewhat overweight Boy Scout named Herbie. The protagonist of the book, a factory manager named Alex Rogo, unexpectedly finds himself leading a 15-member group of Scouts, including Herbie, along a 10-mile-long trail in the woods. The trail is narrow and the boys mostly walk in single file. Rogo struggles with keeping the group together while simultaneously trying to maintain a good pace. Darkness is approaching fast and, as the responsible parent, he has to make sure he safely delivers the entire group to the designated camping site before nightfall.
Before proceeding further with the story, we should warn you that this novel was published some 37 years ago – the language used to describe Herbie and the problems he is causing is a little jarring by today’s standards. It seems Herbie’s weight is the root cause of the troop’s issues. At first, as each boy walks at their natural pace, the line between the fastest boy and Herbie stretches more than mile. When Rogo collects the group and places Herbie in the front, they stay together as a close pack, but the boys behind Herbie begin to complain at the slow pace he is setting.
And they have reason to complain. They are making terrible time and Rogo begins to doubt they will make it before sunset. One of the other boys asks Herbie what he is carrying in his outsized backpack. A little embarrassed, Herbie refuses to answer. Rogo asks Herbie to hand the backpack to him, and the bag is so heavy he nearly drops it. When Rogo inspects the contents of the bag, he discovers – and don’t say we didn’t warn you – it is stuffed food, including sodas, candy bars, spaghetti, tuna, and pickles. Along with his food and camping gear, Herbie also packed a large iron skillet. A growing boy’s gotta eat, after all.
You can probably guess where the story goes from here. Having identified Herbie as the constraint, Rogo relieves him of the excess weight he was carrying, spreads it across the other boys and himself, and places Herbie again at the front of the line. Herbie goes from zero to hero and leads the group at a much faster pace. The boys reach their destination on time, and all’s well that ends well.
The entire point of The Goal is to teach Goldratt’s theory of constraints, which he developed during his career as a management consultant. While the presentation of Herbie’s story seems wildly out of fashion today, it is a timeless lesson. Identifying the constraint in any process and working to relieve it is a simple but powerful management approach. There’s always a new constraint to replace the last one, and so constraint-based management principles can lead to steady and continuous improvement.
Constraint-based thinking also works well when analyzing interconnected industries and supply chains in search of investment ideas. The astute industry observer can predict trends in the economy, work backward to understand how the existing supply chains interact, identify potential future pinch points, and place bets (either in the stock market or privately in the real economy) on likely future outcomes. If the government decrees that Product A must be produced at scale, and you need Input B to make Product A, and there are no alternatives to and few suppliers of Input B, it would be wise to gain exposure to the price of Input B, because it’s likely going way up. Front-running pinch points just before they materialize can be incredibly lucrative.
We were reminded of Herbie when a friend passed along a report from an innovation think tank making bold claims about the feasibility of generating 100% of America’s electricity needs from solar, wind, and batteries. The report was produced by an outfit called RethinkX, and here’s how they describe themselves:
“RethinkX is an independent think tank that analyzes and forecasts the speed and scale of technology-driven disruption and its implications across society. We produce impartial, data-driven analyses that identify pivotal choices to be made by investors, business, policy and civic leaders.”
So far, so good. The analysis our friend sent was titled Rethinking Energy 2020-2030: 100% Solar, Wind, and Batteries is Just the Beginning. You can download the complete work here (you don’t actually need to give your email address to get it, you can simply click “Instant Download”). The YouTube video that accompanies the report was published here. You can tell this is a very important intellectual contribution to the energy policy debate because the man in the video is wearing a tie.
After reading the executive summary and watching the first few minutes of the YouTube video, we had some bad news for our erstwhile buddy: this is the stuff of adolescent scribblings in a spiral notebook.
Let’s start with a few choice quotes from the executive summary (emphasis added throughout this piece):
“We are on the cusp of the fastest, deepest, most profound disruption of the energy sector in over a century. Like most disruptions, this one is being driven by the convergence of several key technologies whose costs and capabilities have been improving on consistent and predictable trajectories – namely, solar photovoltaic power, wind power, and lithium-ion battery energy storage. Our analysis shows that 100% clean electricity from the combination of solar, wind, and batteries (SWB) is both physically possible and economically affordable across the entire continental United States as well as the overwhelming majority of other populated regions of the world by 2030.
“One of the most counterintuitive and extraordinary properties of the new system is that it will produce a much larger amount of energy overall, and that this superabundance of clean energy output – which we call super power – will be available at near-zero marginal cost throughout much of the year in nearly all populated locations. The SWB disruption of energy will closely parallel the digital disruption of information technology. Just as computers and the Internet slashed the marginal cost of information and opened the door to hundreds of new business models that collectively have had a transformative impact upon the global economy, so too will SWB slash the marginal cost of electricity and create a plethora of opportunities for innovation and entrepreneurship. What happened in the world of bits is now poised to happen in the world of electrons.”
The key assumption that underpins RethinkX’s work is their belief that solar, wind, and battery technologies are undergoing a classic S-curve adoption cycle because of relentless innovation breakthroughs that are bringing down costs. Solar is by far the most important technology to their analysis, so we’ll focus our attention on it. Here’s how they frame the issue in their introduction:
“For solar PV, the capital costs per kilowatt of installed capacity have declined by a factor of nearly 1,000x since they were first introduced in the late-1970s. In the United States, capital costs have fallen at an average rate of 16.1% each year over the last decade, and when viewed correctly on a logarithmic plot instead of a linear plot the consistency of the trend is unmistakable. Our analysis conservatively assumes that solar PV capital costs will continue to decline throughout the 2020s at an average annual rate of 12%.”
One significant and well-known challenge to wide-scale implementation of solar technology is its low capacity factor. Capacity factor is a measure of the actual electricity a system produces, as opposed to its nameplate or rated capacity. By convention, rated capacity measures how much electricity a system could theoretically produce under ideal conditions, while capacity factors correct for real world performance. As shown in the figure below, because the sun only shines during the day and clouds can impact a solar panel’s performance, solar has a low capacity factor of roughly 25%.
RethinkX proposes a solution to this dilemma: the US should simply install anywhere from 5 to 10 times as much rated solar capacity as the entire current US electricity grid! No seriously, that’s their solution. Once the needed batteries are included and the grid is further topped off with wind, RethinkX sees a smooth path to an energy utopia:
“The construction of a 100% SWB system in the continental United States would cost less than $2 trillion over the course of the 2020s – just 1% of GDP – and would support millions of new jobs during that time.”
The report goes on to develop proof cases by analyzing what it would take to install 100% SWB systems in California, Texas, and New England. In their most aggressive scenario, RethinkX sees 328 GW of new solar in California, 583 GW in Texas, and 197 GW in New England, all by 2030. As a benchmark, the entire US installed approximately 19 GW of new solar in all of 2020.
Our critique of this RethinkX report is threefold. First, it is based on a fatally flawed assumption about why solar capital costs had been decreasing over much of the past decade (as we’ll see shortly, solar costs are going up in 2021). Second, it misses a nearly insurmountable constraint that makes most of their projections grossly irresponsible. Finally, it ignores the substantial energy input needed to achieve their plan, making it virtually indistinguishable from a perpetual motion machine. We’ll take all three in order.
While there certainly has been significant innovation in the solar industry, much of the dramatic decrease in the cost of solar seen in the past 15 years can be attributed to a simpler explanation: China decided it wanted to dominate the industry and it flooded the world with artificially low-cost supply in effort to put everybody else out of business. That effort has largely succeeded. Here’s how Scientific American described it back in 2016:
“Between 2008 and 2013, China’s fledgling solar-electric panel industry dropped world prices by 80 percent, a stunning achievement in a fiercely competitive high-tech market. China had leapfrogged from nursing a tiny, rural-oriented solar program in the 1990s to become the globe’s leader in what may soon be the world’s largest renewable energy source.”
The article goes on to describe the devastating impact China’s move has had on the US solar industry:
“China’s move eclipsed the leadership of the U.S. solar industry, which invented the technology, still holds many of the world’s patents and led the industry for more than three decades. Just how China accomplished that and why it did is still a matter of concern and debate among U.S. experts.
One clear result is that the U.S. solar industry was hit hard by plunging prices and can no longer supply more than a third of rapidly growing U.S. appetite for solar panels, according to a recent Department of Energy report exploring ‘opportunities and challenges’ of solar manufacturing.”
Having actively participated in the solar industry during this period, we are under no illusions about how China pulled off this “miracle.” They leveraged cheap labor, had minimal environmental restrictions at solar production factories, blatantly stole intellectual property developed by Western companies, poured tens of billions of illegal subsides into the industry to support domestic suppliers, and – ironically – used ready access to cheap coal to produce the energy-hungry inputs needed to make solar panels. None of this has anything to do with S-curves and innovation, nor is it particularly sustainable. In fact, China’s behavior has likely stunted the pace of innovation in the industry. Who wants to invest valuable R&D dollars only to watch China steal the work without recourse?
Which leads us to the most critical constraint RethinkX overlooks – and the second major flaw in their analysis: the production of solar-grade polysilicon. Polysilicon is the foundation of the solar industry, and one simply cannot make solar at scale without it. The process by which polysilicon is manufactured is incredibly complex and energy intensive. It involves transforming high-grade quartz (i.e., sand) – itself a thermodynamic sink – into pure polysilicon that is ready for further processing and use.
China’s behavior has so devasted the US solar industry, a full-blown trade war has erupted. Only three polysilicon manufacturers remain in operation here: REC, Hemlock Semiconductor, and Wacker. As recently as 2019, REC was shutting down domestic polysilicon production, although there are new moves afoot to consider its reopening. This is from their 2020 Annual Report:
“Impacts of Chinese tariffs on polysilicon manufactured in the United States, uncertain market conditions, and reduced demand for the Company’s solar grade polysilicon resulted in the shutdown of the FBR polysilicon plant in Moses Lake, Washington on May 15, 2019. After this date, all polysilicon produced by REC Silicon was manufactured in the Semiconductor Materials segment from its plant in Butte, Montana.”
For its part, Hemlock Semiconductor operates only one polysilicon plant in Michigan (the largest plant of its kind in the US), and its ill-fated adventures in Tennessee are a cautionary tale for the industry. Here’s how Wikipedia describes it:
“The company expanded with the Japanese joint venture partners Shin-Etsu Chemical and Mitsubishi Materials, for a new $1.2 billion plant opening near Clarksville, Tennessee. Though it officially opened in 2012, chemicals were never inventoried and no product was made. The plant was under negotiations in 2011 for a further $3 billion expansion, to keep pace with manufacturing competition from China.
In December 2014, Hemlock Semiconductor Corporation announced the permanent closure of the $1.2 billion Tennessee plant, due to adverse conditions from industry oversupply and ongoing challenges from global trade disputes.”
Notice a pattern here?
From Wacker’s website, we discover how expensive it was and how long it took to build the newest polysilicon plant in the country, its facility in Charlestown, Tennessee:
“The polysilicon production site in Charleston, Tennessee, is WACKER’s largest single investment ever, totaling some US$2.5 billion. WACKER began starting up individual plant sections in Charleston in December after a construction period of just under five years. In the coming months, we will gradually ramp up production and expect to reach Charleston’s full capacity in the third quarter of 2016.”
It takes about 5,000 tons of polysilicon to manufacture 1 GW of solar capacity. As a benchmark, Hemlock can produce about 36,000 tons per year and Wacker’s new Tennessee plant churns out about 20,000 tons annually. That’s roughly enough for 11 GW of new solar per year. RethinkX would have you believe the US can add thousands of GW of new capacity – all by 2030, mind you – and that the cost of solar will keep falling. Left unstated in their analysis is where all the polysilicon will come from, how the new capacity will be constructed in time, or how trying to procure it from the market won’t drive prices sky high. Quite the contrary, the price of polysilicon has nearly tripled (for reasons we’ll get into next) since they published their thesis in October 2020, proving our point that polysilicon is one fat Herbie.
Why so much focus on US production, silly chicken? Can’t we just continue to source supply from China? If only it were that simple. The reason we can’t flows directly from our third major critique of the RethinkX work. Installing solar capacity is quite energy intense, and that energy must come from somewhere. The energy payback period on solar is a heavily politicized number with significant sensitivities to wildly varying assumptions, but nobody denies it takes multiple years to harvest more energy from a solar installation than the energy input required to produce and install it. Today, that energy comes from fossil fuels, the very same industries RethinkX thinks deserve zero incremental investment dollars going forward.
The following irony is not lost on us. The European energy crisis, caused in part by the faith put into superficial and dangerously unserious studies such as this, spilled over into China. One of the first victims of the energy spike in China was polysilicon production, thereby proving definitively that the rosy projections of ever-decreasing costs as laid out in the RethinkX report are laughable. When given the choice between the fun of staying cold today and a nebulous, handy-wavy utopian perpetual motion dream of the future, China chose rationally. Here’s how Bloomberg reported it:
“Solar power’s rare year of rising costs may get worse thanks to China’s power crisis. Prices of silicon metal, used to make the material that comprises solar panels, have surged about 300% since the start of August after a top-producing province ordered production be slashed amid a power crunch. China dominates global solar production, with its coal generators powering many of the factories that make clean energy equipment.”
We’ve poked a lot of fun at the RethinkX report in this piece but work such as theirs is no laughing matter. It is consumed by equally uninformed progressive politicians and Wall Street asset managers who conclude the only thing preventing the US from a zero-carbon future is political will – basic physics and fundamental economics be damned. As for public perception, take a moment to look at the comments left on that YouTube video to get a taste of its complete conviction in “the possible.”
We’ve outlined sound and rational approaches to improving our energy mix, which includes healthy investments in the very technologies RethinkX promotes, but better isn’t better when only zero will do, even if zero is impossible. Maybe if we put on a tie, more people might hear us.
If you enjoy Doomberg, subscribe and share a link with your most paranoid friend!
For that video… Tell me you don’t understand economics without telling me you don’t understand economics. Lol
Trying to math this out. Avg solar panel is 2Msq in size. 165W/m2 panel = 330w/day = 120.45 kwh/yr (est)
one panel crucially requires 20g (.643015 troy oz) of silver with 55,942M troy oz worldwide (est)
I think that means world can make 87B panels supplying: 10.5T kWh = 10.5B mWh = 10.5M gWh = 10.5K tWh
Except the world generated approx 17k tWh in 2020 alone. So (if I'm doing this right) there isn't remotely enough silver in the world to replace all energy sources with solar.
Those panels will also require at least 174B sq meters of land.
Meanwhile, the auto industry itself needs 55M oz of silver annually