Authorised and regulated by the UK’s FCA to provide investment accounts, we are bound by CASS rules to segregate and protect client assets.
Electricity – cheap, reliable, clean, abundant – and forgettable?
The better the infrastructure, the less you notice it. For most of us, electricity is a perfect example. As we have settled into our lockdowns, we take for granted that the lights will come on, the refrigerator will keep our food fresh, our computers and televisions will switch on and so on. Moreover, we assume that the electricity will still be there as we introduce new technologies and processes to decarbonise and digitise our lives.
When we think of electricity, we are probably focused on getting the best price, perhaps whether to use a smart meter and maybe if we should sell a bit of power back into the grid. Every now and again, we think about who is going to switch it back on when it shuts down. If we ever think about where our electricity comes from, we probably focus on the environmental or resource security aspects e.g. fossil fuels, nuclear or renewables. While it may not be as exciting, the easily overlooked electricity transmission and distribution system – the “grid”— may be the keystone to the edifice of our modern lives.
We have become accustomed to cheap, reliable, clean and abundant sources of power.
Globally, the grid will require substantial investment over the next twenty years. According to work by Oxford Economics, we will spend $25tn at 2015 prices and exchange rates globally on electricity infrastructure between 2016 and 2040. Nonetheless, this extrapolation of trend expenditure on electricity infrastructure will still leave us short of global needs by $4tn.
We divide future grid investment into three categories. The first is “catching up” by emerging markets which still lack the infrastructure to provide adequate access to electricity. The second is “keeping up” in developing markets which need to maintain and upgrade aging transmission networks. The third is “moving forward” globally, in particular to allow decarbonisation and digitalisation to continue.
Nothing happens without the grid
The combustion engine literally powers the industrial economy, but at a price. In essence, fuel is burned in order to create pressure that is converted into mechanical energy, and often into electricity. The petrol based internal combustion engines used today are incredibly efficient, make use of a highly stable and transportable fuel source, and can be easily moved themselves. However, the price of convenience is idle capacity (the engine in the car in the driveway when we aren’t driving), lack of scale (one big engine is more efficient then lots of little engines) and limitations on fuel used (petrol is a very effective fuel source but is also environmentally costly).
A more efficient solution is to have large scale stationary plants produce the power, ideally from clean fuel sources, which then gets distributed to the point of consumption as and when it is needed. This isn’t as convenient as having many small, mobile power plants (like car engines), but it can be vastly cleaner and more efficient. Of course, these large, stationary plants are completely useless without the grid to transmit and distribute the electricity from point of production to point of consumption.
The grid can be very delicate and complex.
For example, in the absence of practical storage (batteries are prohibitively expensive), the power demanded must equal the power produced at every moment, and vice versa – the load must be balanced. Moreover, as neither supply nor demand for power is constant over time, it follows that production capacity must be at least as great as potential peak demand, and demand must be no less than baseline production. For example, nuclear power plants don’t easily switch on and off, so whatever power they produce must be consumed and any additional power demand must be met by flexible generators like combined-cycle gas turbines. The grid’s capacity must be at least as great as maximum production/consumption, and because there are multiple paths electricity can take through the grid, flow (as well as phase) must be actively managed to maintain balance.
Finally, an efficient grid can have a significant impact on available supply. As electricity is transmitted, particularly over long distances, the charge naturally dissipates. Improvements to the grid over the past century have reduced this loss from about 16 per cent to 6 per cent in the US, and loss rates are broadly similar in Western Europe.
The US grid (which is popularly called “the largest machine on Earth”) has 170,000 miles of high-voltage line and 6 million miles of local distribution cable. A quick back-of-the-envelope extrapolation of the energy infrastructure intensity in the US to the rest of the world indicates that bringing the everybody else’s grid up to the level of the US would require an additional 3 million miles of high-voltage line and 100 million miles of low-voltage cable.
In 2018, the average American consumed 12.2 MWh of electricity, about four times the global average. To bring the entire world up to the level of current American consumption, it would require an additional 70 thousand TWh of electricity supply.
Of course, it is unlikely that the world would (or could or should) consume electricity on the scale of the US. Nonetheless, it is clear that meeting the emerging markets demand for power is going to require a continued and substantial infrastructure build. Oxford Economics estimates that low and middle income countries will require $3.9tn of investment between 2016 and 2030 simply to keep pace with their own needs, much less catch up with developed markets. Thus, significant investment is needed simply to bring the rest of the world up to today’s developed world level of access to electricity.
Even in the most developed economies, substantial additional investment is required simply to keep the grid functional. The first electric grid was built in 1882 in New York City by Thomas Edison. At 100v and it was just enough to provide electric light to 59 customers on Wall Street. One hundred and forty years later, the grid in the US and throughout the developed world has grown enormously in scale and complexity, but much of the installed infrastructure is itself many decades old.
Power failures are a common occurrence even in the most technologically advanced cities. We can use the example of just three cities and one summer. In July of last year there was a blackout in New York City because of a mechanical failure at a single substation which then cascaded through the city. In August, London had a blackout as a gas turbine failed just as supply from wind turbines declined. In October, San Francisco suffered a blackout as a combination of high winds and high demand caused the utility to shut the grid to avoid wildfires.
In the US and Europe, consumers can expect one to three power outages per year.
The common denominators are typically high demand (usually exaggerated by air conditioning during heat waves), hot weather, and older equipment. As a result, load shedding, or rolling blackouts, are common when demand exceeds generation and/or distribution capacity, and the only solution is to maintain additional peaking – and thus often idle – capacity.
The increased need for capacity that usually remains idle is substantial. For example, in 1980 to 1984 in New York and New England, electricity demand was greater than eighty percent of peak demand only about ten percent of the time. In other words, twenty percent of electricity capacity was idle for ninety percent of the time. By 2005 to 2009 demand had changed so that demand was greater than seventy percent of peak demand around ten percent of the time, leaving thirty percent of electricity capacity idle for ninety percent of the time. Thus, the amount of electricity capacity needed only to serve demand in the high decile of demand had increased by fifty percent in twenty-five years. Thus, this increase in required infrastructure is a function only in the changing distribution of demand in the day and across the seasons.
As we move forward into the 21st century, there are at least three major goals which depend deeply upon the smooth functioning of the grid. These include the shift toward clean sources of power, electric vehicles, and the explosion of data-driven technologies.
Connecting the grid to clean energy
In 2018, about 80 percent of total energy demand was met by hydrocarbons – oil, gas and coal. This in part reflects the fact that these hydrocarbons can be easily transported, and thus that power plants can be placed next to load centres, like cities. Accordingly, there are substantial savings on power loss by avoiding long distance transmission.
The cleanest sources of renewable energy are not transportable.
Solar power is efficiently gathered in the deserts of Arabia, wind power from the plains in the American Midwest, tidal power from the coasts of Scotland, geothermal in Iceland. To make use of these power sources, we need to generate electricity near the fuel source and then transfer the power (often over vast distances) via the grid. Moreover, as these fuel sources are variable (the sun doesn’t always shine, the wind doesn’t always blow), extra generation and distribution capacity must be in place accommodate fluctuations.
Power transmission across the long distances from power source to load centre will require the widespread introduction of ultra-high voltage direct current (UHVDC) systems. This allows three times as much power as similar voltage alternating current systems to be moved up to several thousand miles rather than several hundred. But, it will also require substantial new cabling and DC convertors, and this is expensive. For example, the Desertec initiative – which has stalled – looked to bring power (particularly solar) to load centres via UHVDC. The intense sunshine in the Sahara, for example, could easily power Europe – but the political and logistical complexities, along with a price tag of EUR 500bn, have thus far proved insurmountable.
Ending the gasoline age
Petrol is extremely energy dense, easily transported, and relatively simple to convert into useful energy, as discussed above, and thus the 20th century was “the gasoline age”. However, extracting and consuming petrol has a severe environmental impact. There are two ways to replace petroleum in the energy chain. One is to develop another portable fuel that is at least as safe and efficient, but less environmentally costly. Hydrogen is promising, but the technology is still emerging.
The International Energy Agency estimates that, under its assumption of 30 per cent electric vehicle penetration by 2030, the energy equivalent of about 4.3m barrels per day of oil product demand – roughly 5 per cent of total oil demand – would be saved by various efficiencies in the switch to electric vehicles. However, there would also be the need to generate and distribute an additional 1,110 TWh of electricity. Goldman Sachs recently estimated that, assuming full electric vehicle penetration, as much as $2.8tn of grid investment would be required globally.
Alternatively, we could centralise the energy production with more attractive fuels and then distribute the power to the vehicle in the form of electricity.
Moreover, the grid would need to increase capacities for an even larger peak load. A recent survey by PWC found that the largest portion of British battery electric vehicle owners charge their vehicles at home in a three-hour window between 5pm and 8pm. These electric vehicles all charging at the same time would add dramatically to the intraday peak in electricity demand.
Finally, a second infrastructure constraint to penetration of electric vehicle is the development of a charging network. Goldman Sachs assumes that 80 percent of charging will be done at home, but that even so another $2.6tn will need to be invested in infrastructure to provide charging points.
Keeping the digital age alive
Computers use a surprising amount of power. The Semiconductor Industry Association predicted in 2015 that, at then current rates, computing related energy demand in 2040 would exceed global energy production. Of course, we will surely develop more efficient computers, but there is little expectation that we will be using computers less.
There are three main drivers of computer related demand for electricity – blockchain, the internet of things, and data centres. It’s extremely difficult to project how much these rapidly evolving technologies and activities will grow, and thus how much power they will demand, but it is sure to be substantial.
For example, by one estimate, cryptocurrency alone – which is only one blockchain related activity – currently uses about 80-100 TWh of electricity, or roughly twice that of Singapore.
As blockchain penetrates further into supply chains and digital transactions, it seems likely that this demand for computation and power will grow sharply.
The internet of things will help make some things more efficient but adds its own power footprint. Machine-to-machine communication requires power, generally over wireless systems that are far more energy hungry than wired systems. As 5G is rolled out and enables a sharp increase in the ability to connect devices, the power demand will increase, even if not linearly.
The data produced from the internet of things will further help to feed the growth in massive data centres. In 2017, data centres used about 416TWh, roughly 3 per cent of total electricity demand globally, and the growth is exponential. To a great extent, these data centres can be located next to stranded power sources, ideally in cold climates – thus, Iceland being the perfect location. However, the need to minimise data transfer over long distances and across borders will pull many of these data centres away from the cheap energy sources and back onto the shared grid.
How to invest in electricity grid growth
The companies that own and operate the grid are almost always heavily regulated, if not themselves owned by governments. These companies seem to have a stable and albeit unspectacular return profile.
We prefer obtaining exposure to those companies which will build the grid of the future. At the moment, three engineering firms in particular seem to stand out: ABB, Schneider and Siemens all of which we see as being well positioned with a deep technology base and global footprint.
 Oxford Economics. July, 2017. Global Infrastructure Outlook.
 US Energy Information Agency.
 Oxford Economics. July, 2017. Global Infrastructure Outlook.
 Electric Power Research Institute. Distribution Reliability Indices Tracking Within the United States. Palo Alto, CA, 2003.
 Massachusetts Institute of Technology. 2011. The Future of the Electric Grid.
 International Energy Agency. May, 2019. Global EV Outlook 2019.
 Goldman Sachs. October, 2018. From Pump to Plug.
 PWC. April, 2018. Charging Ahead! The Need to Upscale the UK Electric Vehicle Charging Infrastructure.
 Goldman Sachs. October, 2018. From Pump to Plug.
 Semiconductor Industry Association. 2016. International Technology Roadmap for Semiconductors 2.0.
Past performance is not a reliable indicator of future returns. Forecasts are not a reliable indicator of future returns. If the information is not listed in your base currency, then the result may increase or decrease due to currency fluctuations.
If not otherwise indicated, all graphs are sourced from Dolfin research, April 2020.
For more information please read our disclaimer.