Why the Phrase “Energy Leapfrogging” Is Misleading
I have seen a number of blog posts, panel discussions and news articles that extol the idea of energy leapfrogging. A recent Business Week column on India described, “leapfrogging the nation’s ailing power-distribution infrastructure with solar-powered local networks — the same way mobile-phones have enabled people in poor, remote places to bypass landlines.”
The dramatic increase in cell phone coverage in the developing world is astounding and definitely worth admiring. But, the supposed analogy between cell phones and distributed solar is misplaced for several reasons.
- Distributed solar is a much bigger investment for a household than a cell phone. Distributed solar takes on different forms, ranging from a single solar lantern, to a solar home system that will power several lights and a cell phone charger to a solar microgrid, covering at most around 50 households.
Let’s take a mid-sized solar home system. In Kenya right now, an 8W mKopa solar home system requires a $33 down payment and then about $0.45 a day for a year, for a total of almost $200. The system allows the owner to power 2 LED lights and a phone charger. By contrast, Kenyans can buy an inexpensive cell phone for around $20 and pay no monthly fee. Sending texts costs ~1 cent.
I suspect that part of the solar versus cell phone price difference relates to my next point.
- Cell phones still take advantage of a centralized network but distributed solar does not. Yes, cell phones are nearly ubiquitous in many parts of the developing world, but so are cell towers. Centralized networks provide dramatic cost advantages that cell phones exploit but that distributed solar, like solar home systems, do not.
Imagine if everyone in the US tried to make a cell phone call at exactly the same moment in time. This would completely overload the existing systems, as they are currently built to meet actual cell phone usage, where the millions of users are making independent decisions about when to use the network.
When we start to make decisions at the same time, the system overloads and drops calls or fails to connect them. That’s what happened right after the Boston Marathon bombings, as everyone simultaneously tried to reach loved ones in Boston. Except in rare cases like disasters, though, the basic network economics allow us all to connect our calls at many times lower cost than if we each had dedicated network services. Roughly speaking, if every cell phone user requires 1 unit of capacity when they make a call, we do not need to build 1,000 units of capacity to supply each additional 1,000 users because those users are essentially never all making calls at exactly the same moment.
Distributed solar misses out on these basic economics. Taken to an extreme, providing every unconnected household with a solar home system would require many, many times more generating capacity than supplying them with grid power. In the US, EPRI has calculated that a typical household would pay four to eight times as much for equivalent services from an off-grid system.
The beauty of network economies is the more people on the system, the more we can rely on the powers of averaging. I would be curious to see a calculation similar to EPRI’s that compared a 50-household microgrid to grid power. I suspect that 50 people living in the same village will have correlated demand, so averaging doesn’t do as much good.
- Distributed solar for every household should not be the ultimate development goal. Related to the above point, unless the economics of local solar generation change dramatically, it will not make sense to power every home, hospital, school, office building and factory with an unconnected distributed solar system.
I sense that the proponents of energy leapfrogging are implicitly assuming that just getting a household “electricity” means we can stop thinking about electrification and move on to the next development challenge. Getting households enough electricity to power a light and a cell phone charger is a step, but not the eventual development goal. We will need much more generating capacity to power fans, refrigerators and home production, like sewing, not to mention powering health clinics, hospitals, schools, and local food stores. As I have argued before, both health care and education are consumed locally, so if we are going to use grid power for that, we might as well use it for homes.
- A cell phone provides higher service quality than a landline phone, while that’s not necessarily the case with distributed solar. I realize this is obvious: cell phones are portable but landlines aren’t. The expression, “wait by the phone” is going to go out of style soon. Cell phones also let you send texts, and, in some cases, surf the web. About one third of US households have dropped their landlines in favor of cell phones.
But, we are not seeing US households disconnecting from the grid to get a solar home system. Yes, a small number of US homes have rooftop solar, but they’re still drawing from the grid when their solar panels aren’t producing electricity.
True, solar can be more reliable than the grid in some developing countries where daily outages lasting several hours are the norm. But, we don’t yet know how newly electrified households value that reliability. Clearly, those of us living in the US, who are used to less than 3 hours of outages per year, value reliability. In our hyper connected lives, living without internet, lighting, refrigeration, etc., is a real drag. But, for households who don’t yet have electricity, we don’t really know whether they’d rather have highly reliable electricity or enough electricity to occasionally operate a TV, fan or home sewing machine. And, in the US, without extensive backup capacity or a grid connection, distributed solar is much less reliable than the grid.
I worry that the idea of energy leapfrogging lulls us into ignoring the difficult development versus environment tradeoffs involved with bringing electricity and other improved energy services, like motorized vehicles, to people who do not currently have them.
I am not claiming that solar has no role in electrifying the 1.3 billion people who live without it. To the contrary, grid-scale renewables can take advantage of network economics. If solar panel or storage costs are coming down, why not centralize them and connect them to the grid? And, there certainly may be some areas that are so far from existing grid infrastructure, that solar is cheaper than extending the grid. I’d call that an energy stepping stone, though.
Modern energy can transform people’s lives, so it’s unfair to insist that households who do not currently have electricity use the high cost, zero-carbon alternative. Language matters, so let’s stop talking about energy leapfrogging and keep our eyes on the goal of achieving cost-effective, low-carbon solutions.
Categories
Catherine Wolfram View All
Catherine Wolfram is the William F. Pounds Professor of Energy Economics at the MIT Sloan School of Management. She previously served as the Cora Jane Flood Professor of Business Administration at the Haas School of Business at UC Berkeley. From March 2021 to October 2022, she served as the Deputy Assistant Secretary for Climate and Energy Economics at the U.S. Treasury, while on leave from UC Berkeley. Before leaving for government service, she was the Program Director of the National Bureau of Economic Research’s Environment and Energy Economics Program, Faculty Affiliate of the Energy Institute at Haas from 2000 to 2023, as well as Faculty Director of the Energy Institute from 2009 to 2018. Before joining the faculty at UC Berkeley, she was an Assistant Professor of Economics at Harvard. Wolfram has published extensively on the economics of energy markets. Her work has analyzed rural electrification programs in the developing world, energy efficiency programs in the US, the effects of environmental regulation on energy markets and the impact of privatization and restructuring in the US and UK. She is currently working on several projects at the intersection of climate and trade. She received a PhD in Economics from MIT in 1996 and an AB from Harvard in 1989.
Energy Institute Blog 
Dear Catherine,
You give four principal reasons for why the analogy between cell-phones and distributed solar is misplaced. Putting aside the fact that cell-phones are not the only exemplar of leap-frogging, and hence not necessarily the appropriate unit of comparison for whether “leapfrogging” is happening or possible, I see a couple flaws in your reasoning:
Reason 1: Distributed solar is a much bigger investment for a household than a cell phone. This is probably your strongest argument. Obviously, given the facts you lay out participating in distributed solar is a bigger lift than getting a cell phone, and this will invariably slow adoption. Yet people in developing countries (well at least some people) much more expensive items like cars, as evinced by the clogged streets of Bangkok and New Delhi. There is clearly a question here of exactly which customers in developing countries will do the leap-frogging, and how much value they feel they can derive from investing in distributed solar. For some, solar lanterns pay off over kerosene lights if they can spread the investment over a year or two. For others, a battery + PV system will serve them better and more cheaply in a spotty grid connection with 2-6 hour blackouts then would a home inverter and lead battery or a diesel generator. It will boil down to whether they can see value in investing in distributed solar, and have the means to capture that value.
Reason 2: Cell phones still take advantage of a centralized network but distributed solar does not. This seems to me to be your weakest argument. You talk about a “network effect” that is really not that relevant for electric power. For telecommunication, “network effects” come in two positive varieties: one from economies of scale and better asset utilization (think transatlantic cables and relay satellites) and the other from the ability to connect more an more people together (the more people are connect to the network, the number of possible two-way communications expands more than exponentially). The first effect has to do with cost, the other with function. For electric power, however, you don’t care if your electrons come from nearby or far away so the functional network effect goes away. For my wireless phone, I do care if I talk to my mother on the east coast as opposed to any random woman her age nearby. So a central “network effect” benefit for cell phones from being connected to a grid is just not prsent for electric customers wether they are connected to the macro grid or a micro-grid. What the grid does potentially offer is: cheaper energy from a central power station, episodic peak instantaneous power amortized over many customers, energy storage, and insurance against unusual conditions at the cost of some resilience when the whole network (or some local part) goes down. It is not a given that all these services will be cheaper from the grid. Even if some of these services are cheaper from the grid, customers may only find one or two of them attractive on costs (not the whole bundle). I am working with a startup looking at selling premium solar+battery systems in India that only partly use the grid (see the related Chadbourne and Parke’s Project Finance Newswire article: “Biggest Change to Power Markets Since Edison http://www.chadbourne.com/files/Publication/a92d70c1-4d71-4984-aed2-0d8f3925e51a/Presentation/PublicationAttachment/d77e681e-d47a-4a7d-a149-0dbf574776fd/pfn_1114.pdf“); this is too is part of leapfrogging I think.
Reason 3: Distributed solar for every household should not be the ultimate development goal. Of course, the ultimate development goal should be energy services for people in developing countries. That being said, I am not sure I see why this is relevant. You seem to assume that distributed PV is sufficient for ” to power a light and a cell phone charger” yet not sufficient to ” power fans, refrigerators and home production, like sewing, not to mention powering health clinics, hospitals, schools, and local food stores.” I really don’t think that makes sense. With enough solar resource and local storage you can do all these things. The question is at what cost? For far flung communities providing the wires and other supporting equipment to connect to the macro grid can be prohibitively expensive, as you admit. One of the distribution utilities in Queensland Australia has insane average T&D costs per customer of 20c/kWh. Clearly some seriously bad gold-plating must have happened there, but a lot of these cost come from serving distant communities that might have better been served by a local microgrid with no or much smaller connections to the grid. Even people closer to the grid (as many un-served customers are, you told us in one of your previous pieces) can face very expensive connection charges to benefit from a grid mostly built to support influential elites and political blocks. There is also a false presumption, I believe, in your piece that leap-frogging entails a complete disconnect from the grid. In my mind this is to narrow a definition. A much more interesting version of leapfrog entails the gradual interconnection of many micro-grids at different scales leaning on each other for much more limited combined benefits then a centralized grid. This is something akin to what Lorenzo Kistof at CAISO calls a “fractal grid” and might evolve much more organically in a place with much less legacy infrastructure.
Reason 4: A cell phone provides higher service quality than a landline phone, while that’s not necessarily the case with distributed solar. Strangely enough, I was just remarking to a friend that phone service used to be so much better when we used landlines and could “hear a pin drop.” Today we use inferior IP telephony, Skype and cell-phones whose connection quality varies widely and are just getting used to worse quality communications. But, I think I kind of get your point. I am always irritated at the analogy between a dumb grid and a “smart” grid being like going from phones to broadband. At the end, the only service most of us want is reliable power without having to think about it. Ultimately, distributed solar either on its own or in tandem with the grid has to supply the same thing. Still, if/when total costs are at grid parity, there are two extra “apps” distributed solar (with storage) can supply: better resilience (akin to having a backup generator) and more control over one’s energy destiny (e.g. if your utility provider goes off the deep end and starts charging you insane amounts to connect or you just don’t like their service you have an alternative, a competing source of supply.). So yes, distributed solar (especially if integrated in a grid optimized for it) can potentially offer you cheaper, better service. HECO is already losing customers in Hawaii because they failed to appreciate this.
Ultimately, I don’t think our outlook is necessarily that different, but our frames differ. Unlike the stance you take, I believe that a leapfrog is likely and desirable in developing countries. I just spent an hour today with a German network engineer talking about all the challenges that distributed generation creates for their medium voltage grid. Some may think that this means that distributed power is an expensive option just for the rich that can afford it. I believe this means that we in the developed countries are unfortunate to have sunk so much money already into a grid optimized around a central power paradigm, and that other countries may have the opportunity to develop “fractal” grids that allow them to provide energy services to their citizens at much lower financial and environmental costs. Granted there is no “manifest destiny” for this new paradigm for energy development, as yet fully realized anywhere, but there is enough information at hand to take risk and seize the opportunity for those ambitious enough, or who have yet to be served by the old paradigm. I find it amazing how energy experts and people like Bill Gates underestimate the pace at which solar is being deployed and its costs are coming down (I routinely see 15 year cost projections for policy-makers beat in a matter of a few years if not months). I reject the idea that advocates for leapfrogging “insist that households who do not currently have electricity use the high cost, zero-carbon alternative” but rather believe that we can help these households achieve a cleaner, cheaper, more resilient, more empowering (excuse the double meaning) energy future with distributed renewables connected to each other in newer and better ways. It doesn’t need to be just in developing countries, either, but rather a co-evolution with developed countries and regions that are already headed in a new direction. Let’s stop pretending that the old ways of industrializing are the only way, and keep an open mind for new solutions.
As one of those who prepared forecasts that underestimated the cost progression for solar (in 2009 for the CEC), I agree wholeheartedly with your comments here. The “network effect” for electricity are simply the combination of economies of scale and portfolio effects. The cost advantage is disappearing, and the portfolio advantage has been overblown (in the U.S. we are 15 times more likely to have a locally-caused outage than a system outage) and technology may be solving that problem as well. Yes, the developing world can avoid our dependence on an overinvested grid.
We addressed the cost forecasting problem in the 2013 version of our analysis by including plausible ranges for future costs. Policymakers should act to consider these ranges rather than leaning on single point forecasts to justify their decisions. That means leaving more decisions to many individuals rather than leaving them in the hands of a single utility or regulatory commission.
What leads you to think the portfolio advantage has been overblown? I would be very curious to see more numbers on this point.
Catherine, in my testimony before the CPUC on behalf of the Agricultural Energy Consumers Association, I calculated that distribution level outages were about 15 times more likely than system level outages. It appears that the reliability problem, at least in California, is driven by issues in the grid connections. That seems to call out for an analysis comparing the reliability gain from reducing grid reliance versus a generation portfolio.
Eric-
Thanks for the detailed comments. A couple replies:
On reason 2, there’s a third aspect of network economics that your summary missed, but that you mention in the end of your paragraph: “episodic peak instantaneous power amortized over many customers, energy storage, and insurance against unusual conditions at the cost of some resilience when the whole network (or some local part) goes down.” These services require consumers to be linked by a network, and are what EPRI estimates would cost 4-8 times as much if provided on a distributed basis in the US. I would love to see a similar comparison for the developing world.
On reason 3 and your reference to my previous post about high connection charges, you have to be careful to separate what households are paying (price) from what it costs the utility, and particularly, what it could cost the utility under a mass connections program, where we suspect the costs per connection would be much lower. Also, in Kenya, the recent grid has been built out to supply schools and health centers, so not just elites.
On your last sentence, I agree, I’m happy to keep an open-mind, but I also think it’s important to bring the good parts of the old ways into the future and suspect that the leapfrogging discussion has undervalued the grid.
I agree on:
bringing the good parts of the old ways into the future
So that means rethinking some of the old ways; for me reliability is at the top of the list.
On point 4:
“Clearly, those of us living in the US, who are used to less than 3 hours of outages per year, value reliability”
We value reliability, yes, but there are limits to everything. Here, we have cost/reliability tradeoffs administratively decided for us. Perhaps this is an area where differentiation of service quality can emerge, regardless of how old or young the grid is.
On point 2, EPRI probably was making this assessment assuming traditional responses to those requirements. The rapid evolution of what’s available may be redefining this entire category.
On point 3, at least in the U.S. the connection costs for a new customer are approaching the average connection costs. In other words, the economies of scale are disappearing.
Thanks, Catherine. I wonder how you see the concept of leapfrogging playing out in healthcare? Mobile health is an enabling trend for distributed, patient-centric care, which is part and parcel of the growth of smart phones, but what are the main mitigating factors in emerging markets like India leapfrogging our model of brick-and-mortar hospitals in every town?
We’re evaluating technological trends in Indian medtech, and this article is a great contribution to our framework: http://www.globalhealth.care
Catherine,
I see your point that there are some dissimilarities between the leapfrogging that has occurred in the telecommunications industry and the challenges in energy. However, the notion of energy leapfrogging still seems appropriate, even if a good portion involves utility scale development. In developing countries, where there is a massive build out of infrastructure taking place, there is an opportunity to make a better choice at the moment of investment. The opportunity created by large scale investment contrasts with places where infrastructure is much more developed. In those places, the low cost opportunities are more constrained by the natural rate of capital stock turnover. The phrase “energy leapfrogging” appropriately and helpfully coveys some of the optimism that is justified due to impressive cost gains in cleaner energy supply and demand technologies. The latest figures from Lazard — known as a middle of the road, authoritative source — show that most inexpensive natural gas and coal options are priced at 6.1 cents and 6.6 cents per kilowatt-hour, respectively. Before subsidies are taking into account, Lazard finds the least expensive solar options to cost about 7.2 cents and wind to cost 3.7 cents per kilowatt-hour at the low end. Yes, there is a need to think about how to best integrate variable renewables, but practice and analysis is demonstrating the feasibility of doing this for modest cost.
“If solar panel or storage costs are coming down, why not centralize them and connect them to the grid?”
This assumes that there is a grid to connect them to! By all means let’s compare the costs of distributed RE with the build out and operation of a centralized grid. That’s a conversation worth having, but that’s not what I see here, which is basically that electricity services are more expensive to provide than communications services…
It’s one thing to run a few lights, a computer and a cell phone charger. If you have a very simple lifestyle and are content with that, then leapfrogging might make sense. However Catherine’s point about coincident loads is pretty important – so much so that Sam Insull devoted a lot of time and effort trying to understand it and then take advantage of it to reduce the level of investment required to serve customers. If everyone had their own solar PV supply and their own storage, there’s still the matter of coordinating the operation of distributed generation and storage to maximize the benefits and minimize the costs. No one has demonstrated how to do that yet i an environment where there are many customers (campus microgrids are not comparable and do not count in my view).
I would think that campus microgrids are much more energy intensive with variable and coincidental loads (think of hourly class changes!) than any developing world village. I rather think the campus microgrids represent the extreme bounding case that tests the resiliency of these systems.
The important difference with the world of Insull is that the economics of scale for generation are disappearing, and the “last mile” costs are escalating faster than any other cost component. That would seem to imply that grid economics are being flipped.
Great post. I wonder how robust the LNG boom will be, esp. given the huge capital investments now arriving in fleets, etc. Don’t dump your diesels! (I ride a bike 🙂
I’ve seen it happen. In 1992, when a guy in a small african village had a small solar panel. That was before the phones, but it allowed him to charge batteries for a radio, and some electric light. It was a leap forward compared to candles or an oil lamp. And very significant numbers of people are in that off-grid situation. The same village still has no grid, but it did get a school, and it has a few solar panels, so they can have a computer room and electricity. The village has collective big solar panels+waterpump and a water tower, to water the vegetable gardens, and there are taps near every house. There is a phone tower, also running on solar power, linking the village to the big towns and the market women coming to buy the vegetables and other produce. No more little ‘cheap’ noisy stinky unreliable diesel generators here. This is leapfrogging. It happens.
And even in the larger towns, the grid power is often not very reliable. So many people have some backup, for which solar power with a battery is great. Solar can compete quite well with that stinky noisy diesel generator. Leapfrogging is great for development. And it is also good for the environment. Less noisy stinky diesel generators. And yes, also less CO2 emissions.
Looking how energy economies of scale are changing rapidly, I’m not sure that connecting to a larger grid is a necessary step on the way to development. Independent household and microgrid systems are approaching cost effectiveness in the developed world despite the huge cost advantage the grid has in sunk and depreciated costs. In countries that will need to develop a grid, it seems that these microgrid technologies narrow the gap even further.
Excellent post that every politician and every environmental advocate should read (and understand). I’m looking forward to reading other feedback.
There will no doubt be a few readers who argue that rooftop solar is now cheaper than retail electricity on a cost per kWh basis. That may be true but it does not account for the storage that is currently provided free via net metering or the cost of the network infrastructure that makes net metering possible, both of which entail significant costs. Maybe someday, but not right now.
Well said.
It might be economically more attractive to just get charging systems where human or animal power charges a battery.
The main advantage of this leapfrogging/ microgrid approach is in the initial capital requirement being also modulat, ie smaller. But in the long run it would not only be more expensive, but likely also less reliable.
Whenever this type of claims are made i go by the rule ‘follow the money’. In this case there are numerous researchers pushing for the microgrid among the academic/ research and the NGO community – grants/ funding.