How learning curves give a glimpse of the future
Change is a fact of life. While change is most often perceived by us when some aspect of our lives is altered, the economy and the universe are in a constant state of flux. Money changes hands and stars disappear forever. One change has been felt particularly strongly and is heralded as the quintessential challenge of our times: climate change. Extreme weather events and higher global temperatures are disrupting the fabric of many ecosystems on our planet. Anthropogenic factors, mainly the exploitation of increasing quantities of fossil fuel (exhuming ancestral organic matter) for our energy needs, since the Industrial Revolution have unraveled an imbalance in the carbon cycle.
We created the problem (for us and our fellow species) and we better solve it. I, for one, believe in human ingenuity in the face of adversity. Humans were on our way to starvation with existing agricultural techniques when the Haber-Bosch miracle of synthetic fertilizer extended our food supply almost indefinitely. Similar transformations are needed in our energy habits to start to undo the ecological damage. We are taking baby steps, slowly but surely, progressing towards a more sustainable energy future. The times are surely a changing — the energy transformation is underway.
Replacing energy dense fossil fuels with renewable alternatives (which are usually less dense) makes most practical sense in stationary applications such as electricity generation, where size and volume are not of the highest concern. Renewable electricity from solar photovoltaics and wind turbines are now the de facto choice of new electricity in both the rich and the developing countries. The rise of the electric vehicle pioneered by Elon Musk’s Tesla is seeking to similarly disrupt the transportation sector away from the conventional gas guzzlers.
Any large scale disruption or transformation within a sector is driven by the cold science of economics, not the emotional appeal for philanthropic action. Renewable electricity is gaining momentum because they are cheaper today than conventional fossil based electricity. This was not the case a decade ago. Price of solar electricity declined by 89% between 2009–19 and continues to do so. How do we explain this trend, and how do we predict the future? Learning curves come to the rescue.
Renewable energy technologies follow learning curves, which means that with each doubling of the cumulative installed capacity their price declines by the same fraction. The ‘learning effect’ in solar technology has held almost immaculately since their inception. The price of solar modules declined exponentially as the installed capacity increased exponentially (as seen in the logarithmic chart). Learning curves are also called experience curves since the X-axis represents the experience gained from increasing capacities. The relative price decline associated with each doubling of experience is the learning rate of a technology. The learning rate of solar modules is a whopping 20.2%, which means that their price dropped by a fifth every time their capacity doubled. This sustained learning rate has brought the price of a solar module from $100 per watt to less than $0.5 per watt in 33 years.
While solar modules follow the learning curve, ultimately we are interested in solar energy, not just the modules. From the chart below, the learning rate for solar energy is even higher at 36%, even before accounting for government subsidies like the investment tax credit and feed-in-tariffs. This shows the other factors such as balance of plant equipment likely have a higher learning rate. Onshore wind power had a lower learning rate of 23% and offshore wind had an even lower 10% learning rate.
Note that fossil fuel energy technologies do not follow the learning curves. Neither the price of coal nor energy from coal has historically followed a learning curve. There is reason to believe that coal would not get any cheaper: a large portion of their energy cost is fuel commodity cost which sets a price floor for energy produced using coal, and the technology to generate electricity from coal is mature (and therefore unlikely to see large gains in efficiency).
While learning curves explain the decline of price with experience and predict possible future prices, it still remains to be explained why they work in the first place. The EIA reports four principal ways in which costs come down with time:
- Research and development. While uncertain in outcomes, it can create potential spillover benefits for other sectors (e.g., fuel cell technologies can be recycled as electrolyzers). Typically these occur over long time horizons.
- Learning by doing. Experience with manufacturing and deploying products helps develop cost-cutting improvements across each step of the supply chain including installation, operation, sales and financing.
- Economies of scale. As companies produce larger volumes, they can get better deals with suppliers or set up automation to reduce costs. Another way to achieve economies of scale is through division of labor across the sector, i.e., by specializing a part of the entire production supply chain.
- “Learning by waiting.” Some companies especially in poorer countries can choose to wait instead of investing in their own research and development to avoid reinventing the wheel and save costs. Globalization is a key enabler of faster adoption across the world.
Any company looking to disrupt an existing industry would invariably start off with a more expensive product. That should not dishearten them as long as they see a clear path for cost reduction. Learning curves are a useful tool to estimate future economics and therefore current value of a company. However, learning curves do not manifest by themselves. They need to streamline operations every step of the way.
