Why do I own a car when the bus is cheaper?
I was thinking about that question after reading a terrific twitter thread from the Economist’s Africa editor. Jonathan Rosenthal hits on an amazing 22 billion rand coincidence: South Africa’s troubled utility is on track to spend about the same amount of cash this year (ZAR22bn or about US$1.3bn) on diesel for its peaker plants than a private mining company spent building a wind farm that will provide a similar output of electricity. Of course, the diesel will last one year, while the wind farm should last for twenty. Jonathan rightly concludes: “that’s insane.”
So that raises a good question: why would anyone spend more on diesel than wind if wind is cheaper? I believe the answer lies in why I still own a car, even when the bus is far less expensive.
Some throat clearing first:
South Africa’s utility Eskom is a catastrophic mess and it’s clear that many of its decisions are based on something other than cost-efficiency or sound long term planning. The government made an epic long term blunder nearly two decades ago when it doubled down on mega coal projects and then stifled its own massive renewables potential, a series of mistakes that the US and other donors are now trying to partially help mitigate.
I’m unclear why Eskom is running its open cycle gas plants on (more expensive and dirtier) diesel but no one would do that if they had a better option. I also won’t pretend to understand South Africa’s diesel procurement system, but I share many others’ suspicions that someone is getting richer from this option.
Many other of South Africa’s power choices, like giving 20-year contracts to floating barges usually used as an emergency last resort, are suboptimal (less diplomatic adjectives also valid here).
All these South African specifics aside, there is widespread confusion about how different types of electricity generation can be compared. I often hear throwaway lines about “X is cheaper” so anything else must be the result of corruption, corporate propaganda, or some other act of bad faith. It is correct that the average global cost per kilowatt-hour of wind-generated power is cheaper than the same thing from diesel. The latest levelized cost of electricity (LCOE) estimates from Lazard report that global prices are about 2.6-8.3 US cents/kWh for wind versus 15.1-19.6 cents/kWh for peaker gas (diesel would be even higher). It would indeed be insane to spend many multiples more for the same electrons. Yet are they the same? Is wind always really cheaper? It depends.
Who pays “average”?
“Average global” cost is just that, an average. Actual power costs depend a lot on location and how a plant is used in the power grid. Wind power is obviously cheaper in windier places (and ditto for solar in sunnier places) because better resources means more time producing, allowing the investor to spread the initial fixed capital investment over more kilowatt-hours. Similarly, power costs from diesel or gas will decrease with more usage (that’s why open cycle peaker plants are costlier per kWh than baseload-like combined cycle plants). And of course fuel prices, the main driver of cost in fossil power plants, also fluctuate both geographically and over time.
Final electricity costs are greatly affected by interest rates too. Places like South Africa, where borrowing rates are 6-8 percentage points higher than in the US, will pass those higher costs onto the electricity coming out the other end. (Spreads in Ghana, Kenya, and other frontier markets are even greater.) And since wind is relatively more capital expenditure-intensive than peaker gas plants, higher interest rates will affect wind more than gas/diesel.
Cost of power is also a lot about reliability
Most of all, simple cost comparisons obscure the main difference: dispatchability – or what non-energy nerds might call “can you get it when you need it.” Of course wind power is available when the (often random) local weather conditions allow it. Adding storage can help smooth the fluctuations a bit over the short-term, but they can’t guarantee 24/7 because you cannot know with certainty for how long you’ll need backup. Storage costs quickly rise (and relative benefits diminish) as you add more capacity to cover longer durations (that you will, by definition, use less often). That’s why in practice, most “storage” cost estimates are for 4 hours. That might work a lot of the time. But is “a lot of the time” good enough? How much extra should a country pay for those potential extra hours/days/weeks of available power? LCOE can’t tell us.
All of this brings me back to my idle car sitting in the garage. If my job depends on getting to work on time every single day, it doesn’t matter how cheap the bus is if the schedule is random. I will need to own a car for those days when the bus doesn’t show up and I still need to get to work. My neighbor berating me for owning a car because “the bus is cheaper” would not make much sense. And it would be annoying if they persisted after I explained why.
South Africa’s round-the-clock rolling blackouts mean the country desperately needs all kinds of new electricity generation. But it's not as simple as just picking the cheapest kilowatt-hours – or replacing all its diesel power with new wind. In a rational world, Eskom would be investing heavily in cheap wind & solar, but also building up a healthy reserve of more costly backups to use when needed. And not using expensive backups when cheaper alternatives are available. Doing anything else would be, well, insane.
Is cheaper wind always cheaper?
It’s almost universal for energy analysts to start by citing Lazard LCOE as proof that wind and solar are now cheaper than fossil fuels. All sorts of policy conclusions then flow from this starting point. Yet, the fine print on Lazard’s website states, “Analysis excludes integration (e.g., grid and conventional generation investment to overcome system intermittency) costs for intermittent technologies.” Citing Lazard LCOE is often a marker of biased analysis to follow.
1. LCOE are a fraudulent measure. It's literally meaningless in the process of delivering real energy to real consumers. Wind and solar advocates use LCOE because the real world measures (the things that help inform rational decisions about energy policy) show that, outside of niche needs, wind and solar are not a reasonable solution for providing energy to populations of people.
2. Your analogy about the bus vs. the car is welcome. Now, apply that concept to not just your individual desire to go somewhere at a particular time. Apply the concept to the hospital that needs power to ensure that emergency surgery can be done; the manufacturing plant that needs to be able to operate in a stable and predictable manner; the food distribution center that needs to ensure that it will have power for it's refrigeration needs; etc.
In a rational world, one would see that wind and solar are not additive to energy infrastructure. They are parasitic. They add needless complexity and costs to existing, stable, baseload energy needs that increase the risk of outage while adding nothing of value to the job of providing energy to customers in a rational manner.