Solving extreme poverty with solar panels

The new energy system offers an unprecedented chance to lift 660 million people out of energy poverty by giving them cheaper and ubiquitously available energy. Let's make it happen.

666 million people live beyond the reach of any electricity grid, 85% of them in sub-Saharan Africa (Tracking SDG7, 2025). So far that means they have no access to electricity which is increasingly the type of energy people need to flourish.

I was reminded by that again when I saw this blog of Max Rosen from Our World In data. Notice where the people in blue live…
(To discuss please use the posts on Bluesky or Twitter.)

The textbook assumption is that these poor cannot pay for energy. The numbers say something stranger: they already pay, at the worst rates on Earth. Households without electricity spend $38 billion a year on fuel-based light, and per unit of useful light a kerosene wick costs up to 150 times more than an efficient bulb (Mills, 2005). A family burning kerosene pays $12 to $18 a year to light one room badly and fill it with smoke (CEIBS, 2024). The poorest customers in the world’s energy market pay its highest prices. Nobody designed that; it is simply what happens when the only light you can afford is the kind you buy by the bottle, week by week. Or maybe when nobody offered you an LED light on solar yet.

Lighting is the smaller part of the problem. 2.1 billion people cook on open fires or rudimentary stoves, and the smoke killed an estimated 2.9 million of them in 2021, including 309,000 children under five (WHO, 2025). The burden disproportionally falls on the women and girls who gather the fuel and stir the pots. Count the sickness, the gathering hours and the climate damage and the bill reaches $2.4 trillion a year (World Bank, 2020). We are, collectively, paying a fortune for the privilege of poisoning the cooks.

Nonetheless we keep failing at delivering improvement: free electricity blocks, fuel subsidies that mostly warmed the already-comfortable (Komives et al./World Bank, 2005), giveaway stoves... Enough attempts failed, in enough different ways, that some serious people (without exception comfortably above the floor) now conclude floors are a mistake. I think these failed attempts again and again show us most people are too fair, kind, empathic and decent to give up on so many of their fellow humans. We just haven't found the right formula yet. But the new energy system is about to change that. And what was considered to be the hardest part, now looks like the easiest part.

The people with no access to a grid at all were considered to be the hardest part. The good news is that in our new energy system they are now the easiest part! Over 95% of those without a grid connection live in rural areas close to the equator. (Tracking SDG7, 2025) (ESMAP, 2025) (World Bank Access, 2023) (World Bank Population, 2023) (IEA, 2025) (Our World in Data, 2025) That means they have two things in common: plenty of sun and plenty of room for solar panels.

One solar panel and a battery the size of a beer crate is enough to give a household there 1 kWh a day even in the rainy season. Enough for phones/computers, lights, refrigeration and a fan. Add two more solar panels (not more batteries) and induction cooking becomes possible. No more searching for wood and dung hours a day. No more destroying any nature left. No more poisoning lungs.

Currently three solar panels would cost about $200. The cheapest 2 kW off-grid inverters costs around $100. A five kWh battery will cost around $200 in a few years. $500 in total. Include streamlined financing and distribution and you are looking at around $700 for a family of on average seven people. $100 per person.

Such a setup could easily last 20 years and if people have to pay for it themselves they will take care of it and sell it if they don’t need it anymore or upgrade. So that’s $5 per person per year. Times 660 million people we are talking about 3 billion per year. Compare that to the $2.4 trillion in health damage alone and it is a truly paltry sum. If you care about nature and climate change it’s also the most effective intervention possible per dollar. And it aligns with our feeling of fairness. Very few people will say: these families don’t deserve $5 per person per year to solve their energy problems.

But it gets even better! Because poor households in sub Saharan Africa already pay around $10-$50 per person per year for kerosene lights and $10-$50 per year for cooking fuel. So a solar kit would only around a tenth of that! We don’t have to present this as charity. We are only providing an organisational upgrade that gives them direct access to a cheaper universal energy source that is already directly available to them. And on top of this less sickness and enormous productivity gains because you can read at night and no longer need to go foraging for wood hours a day.

What we need to do is create competition (companies are much better and cheaper at delivering this than NGO’s and governments). Standardize development, create open standards and open source software and reference designs. Provide cheap financial backing that makes risk taking cheap. In short, provide capabilities to entrepreneurs willing to tackle this while avoiding extraction. One of the simplest and most effective distribution mechanisms is a kill switch that is de-activated after the loan is finished. So when households don’t provide mobile payments, their electricity stops working. As soon as the payments start, the electricity works again. Brutal but effective and cheap. And once the loan is paid off the installation is theirs forever.

That last is important because this could become an extremely extractive scheme targeting a vulnerable population. Some firms keep ownership deliberately vague, so the paying never quite ends (Riedke & Adelmann, 2022). So we have to make predatory lending schemes unattractive by withholding anything that was created open source, withholding any financial support, and by mandating against it if you are a government. What we need is something that people pay off in a maximum of five years, and then it’s theirs. No strings attached. No kill switch. No lock in to a software vendor who might not provide over-the-air updates.

And it’s already happening. Off-grid solar provided more than half of all new connections in sub-Saharan Africa and is already the cheapest route for roughly two in five of the people still waiting.(Tracking SDG7, 2025) In Bangladesh, the state-backed lender IDCOL financed 4.1 million solar home systems – the largest off-grid electrification program in history – with instalments deliberately calibrated to the kerosene spending each system replaced, and every contract ending in ownership (World Bank, 2021). The kerosene budget became a mortgage, the mortgage became a paid-off roof, and nobody resents a family that owns its panel.

You can argue about the bolting on of a temporary toll booth. Pay-as-you-go solar – the kit on credit, instalments by mobile money, a remote kill switch as collateral – carries financing premiums around 94% over the cash price at commercial terms, with roughly half of 2023’s customers written off or more than a month behind (Daggash/Energy for Growth Hub, 2025). I think we (people and governments in richer countries looking for a cheap way to counter climate change and do good) should lower these premiums by carrying the risks. And we could improve on some sense of dignity by giving people fair warning. Now it’s often: miss a payment and the kill switch turns your house dark in front of the neighbours; when one provider’s customers had the lockout clause explained to them, purchase interest fell from 63% to 40% (Waldron & Swinderen/CGAP, 2018).

But in fairness, the kill switch is also why an uncollateralized household gets credit at all: in a Ugandan trial, digital lockout cut default by 19 percentage points (Gertler, Green & Wolfram, 2024). So treat it like what it is – collateral. Let it do its job, then make it die. Any lease, lockout or remote-disable right should expire by law at the earlier of payoff or five years of active payments, after which the device is property, full stop. M-KOPA’s original kit reached full ownership in 365 daily payments (Arc Finance), so five years is not radical. It is a ceiling on how long the sun may be rented back to its owners.

Finishing the job is a campaign on several fronts at once, and none of them requires inventing anything. Researchers first: publish open reference designs – circuit boards, battery management, charge controllers, test protocols – that any competent workshop from Nairobi to La Paz can copy without paying for a licence. Laboratories have quietly run this experiment already: scientific equipment built from open-source designs costs roughly 87% less than its proprietary equivalents (Pearce, 2020). Pair the blueprints with the durability standards the field keeps asking for – replaceable batteries, available spare parts, repairability (Munro et al., 2022) (Samarakoon et al., 2022) – because a legal right to own hardware that dies in year three is a right to own junk. Chapter 8 made the general argument that open standards are how newcomers compete; here they are how a thousand local manufacturers come into existence.

Bankers second. Commercial pay-as-you-go needs effective interest around 30% just to survive its own write-offs (Waldron et al./CGAP, 2018), and at unsubsidised instalments only 22% of unconnected households can afford even the smallest real system (ESMAP/GOGLA, 2024). Cheap, patient capital changes that arithmetic: concessional loans and guarantees pull the rates down, and subsidies can ride payment rails that already exist instead of building bureaucracies that don’t. Togo pays $4 a month straight into the instalment account of a provider the household chooses – 134,000 kits and counting (GSMA, 2020); Rwanda’s results-based subsidies reached 370,000 households (GOGLA End-User Subsidy Lab). And at the very bottom, subsidy can safely go all the way to free: charging $0.60 instead of nothing for a bednet cut uptake by 60 percentage points, while free distribution produced no waste (Cohen & Dupas, 2010) and made people more willing, not less, to pay later (Dupas, 2014). Free at the bottom does not corrupt the market. The market was never going to reach the bottom on its own.

Appliance engineers third. Every watt an appliance does not draw is panel and battery nobody has to buy: a super-efficient solar vaccine refrigerator needs one tenth the panel and battery capacity of a conventional one (Efficiency for Access, 2023), and the same lever works through LED lights, DC fans and televisions. Efficiency is what makes the entry kit beat kerosene on price rather than on principle. The kit must also be something people want, because the poor refuse goods that mark them as poor and stretch for goods that signal a future – households living on a dollar a day buy televisions before the things experts call sensible (Banerjee & Duflo, 2007). The cautionary tale is the improved biomass cookstove: promoted for decades because experts approved of it, abandoned by year two because nobody wanted it, delivering no measured health gain (Hanna, Duflo & Greenstone, 2016). Induction stove are a completely different story but make them products people actually want. Maybe the first rung of a normal product line – same brand, bigger tiers above, the better-off buying up and the kit carrying no poverty mark (ESMAP MTF). Dignity counts.

Distribution fourth and last: sell the kit where airtime is sold, price it in mobile money, install it the same day, write the contract on one page. The kit must be the easiest purchase in the market, visible at every kiosk until nobody can fail to notice it exists. Indonesia showed what ubiquity plus a starter package does: a free initial kit and a built-out supply chain moved 50 million households from kerosene to LPG in five years (Thoday et al., 2018). Availability everywhere is not a logistics detail. It is how a floor announces itself.

Again: this is not charity. It’s cheaper energy and when rural KwaZulu-Natal was electrified, female employment rose nine percentage points within five years – from a baseline of about seven, a more-than-doubling – as electric light and cooking released hours into paid work (Dinkelman, 2011). Kenya’s M-Pesa lifted 194,000 households out of extreme poverty, the effect strongest where women ran the household – and the entire system presumes a charged phone (Suri & Jack, 2016).

The new energy system offers an unprecedented chance to lift 660 million people out of energy poverty by giving them cheaper and ubiquitously available energy. Let’s make it happen.