SolarAid

SolarAid is a London-based international development charity that works to bring solar lighting and power to rural communities in sub-Saharan Africa. It was founded in 2006 by Jeremy Leggett, the founder of the British solar company Solarcentury, and it wholly owns a social enterprise called SunnyMoney, which for several years was one of the largest sellers of solar lights in Africa. Its stated goal is to "light up every home, school and clinic in Africa by 2030, using safe, clean, solar power."

SolarAid's story is useful well beyond the charity itself. It is a case study in a question that runs through modern development work: should an aid organization give things away, or help build a market in which people buy them? SolarAid chose to sell. It became, by its own account and by figures from the World Bank's Lighting Africa programme, a dominant seller of small solar lights in parts of East Africa, helped create markets that commercial companies then took over, and then deliberately pulled out of those countries. This article covers the problem SolarAid set out to address, its history, how its sales model worked, its finances, the evidence on what solar lights actually achieve, the criticisms and limits of its approach, and practical points for anyone assessing it.

The problem: lighting without electricity

Scale of energy poverty

Hundreds of millions of people in sub-Saharan Africa live without a connection to an electricity grid. SolarAid's own materials put the figure at almost 600 million people across the region, and its 2023/24 annual report cites a global figure of 685 million people without electricity as of 2022, about 80% of them in sub-Saharan Africa. The Ashden judges' assessment of SolarAid in 2013 used a similar figure of about 590 million Africans off the grid.

For households in that position, lighting after dark has traditionally meant kerosene (called paraffin in much of Africa and Britain) lamps, candles, wood fires or, increasingly, cheap battery-powered torches.

The cost of kerosene

Kerosene lighting is expensive relative to rural incomes. A 2015 Climate News Network article on SolarAid reported that a kerosene lamp can cost families about 15% of their income. In a survey of 85 Tanzanian households reported by Ashden in 2013, weekly kerosene spending among households that bought solar lights fell from US$1.60 to US$0.50.

Health and climate damage

Simple kerosene wick lamps are also harmful. They are a fire risk and produce smoke indoors. A 2012 study by Nicholas Lam and colleagues, published in the journal Environmental Science and Technology, found that 7% to 9% of the kerosene burned in simple wick lamps is converted into particulate matter that is nearly pure black carbon. The authors estimated global black carbon emissions from kerosene lighting at about 270,000 tonnes a year (with an uncertainty range of 110,000 to 590,000 tonnes), about 20 times higher than earlier estimates. Black carbon is both an air pollutant and a strong short-lived contributor to warming, and the authors argued that replacing kerosene wick lamps deserved strong consideration in programmes targeting short-lived climate pollutants. A University of California, Berkeley release about the study also pointed to separate research in Nepal in which women who reported using kerosene lamps at home had 9.4 times the rate of tuberculosis of those who did not.

Why small solar lights

Small, self-contained solar lights combine a little photovoltaic panel, a rechargeable battery and an LED. They need no grid, no fuel and little maintenance. By the mid-2010s, prices had fallen far enough that basic models were within reach of many rural households. The 2015 Climate News Network article reported solar kits costing as little as US$10, with the cheapest study light giving four hours of bright light after a day's charge, and larger models offering up to 100 hours of light, phone charging and radio charging, at up to about US$140 for products aimed at small businesses.

Founding and early years

Solarcentury's 5% pledge

SolarAid was created in 2006 by Jeremy Leggett, who had founded the British solar company Solarcentury. The charity was funded from the start by a pledge from Solarcentury to donate 5% of its annual profits. According to Climate News Network, the charity received £28,000 in its first year from that pledge, a figure that later grew towards £500,000 as the company's profits rose. Leggett has argued that the practice is easy for businesses to afford: "Most companies would not miss 5% of their profits, and the gains are enormous," he told the network, adding that the donations had helped staff morale and retention. He later promoted a "5% club" of businesses willing to make similar commitments to climate and poverty charities.

SolarAid is registered with the Charity Commission for England and Wales as charity number 1115960. The register records it as a charitable company, with registration dated 30 August 2006.

Macro solar: what did not work

SolarAid's own ten-year review, published in December 2016, is candid about its early experiments. In 2006 it ran pilot programmes with deaf and disabled people in Tanzania, followed by pilots in Malawi and Zambia in 2007. For about four years it experimented with what it called "macro solar": fitting larger photovoltaic systems to schools and health centres.

The review says this worked in some respects but revealed two problems. Most of the systems were being used mainly to charge lights and phones, which smaller devices could do. And servicing was difficult: when inverters tripped and shut systems down, local engineers often struggled to get them running again, leaving systems idle while the charity struggled to pay for repairs.

Micro solar and the birth of SunnyMoney

By contrast, what the charity called its "micro solar" work, distributing small, portable solar lights through schools and sales agents, was growing quickly. These devices needed no servicing and delivered light where it was needed most, inside homes.

In response, SolarAid set up SunnyMoney as a social enterprise to concentrate on this work. Sources give slightly different dates. SolarAid's 2016 review says SunnyMoney was established in late 2008. A SolarAid blog post from 2015 says SunnyMoney began selling lights in early 2009. Ashden's 2013 profile dates the setting up of SunnyMoney to 2010, with sales beginning in 2007 under SolarAid itself. The differences probably reflect the gap between SunnyMoney's formal creation, its first sales and the launch of its school campaigns.

SunnyMoney's stated objective, in SolarAid's words, was to distribute solar lights in "last mile communities" where commercial organisations did not go because it was too hard to make a profit. Crucially, it sold lights rather than giving them away, so that local distributors could build businesses and earn an income.

Leggett later described the arrangement as running SunnyMoney as though it were a for-profit company, but using philanthropic funds as equity, with SolarAid not expecting a financial return.

The schools model

The breakthrough on Mafia Island

According to SolarAid's ten-year review, the breakthrough came on Mafia Island, off the coast of Tanzania, where the SunnyMoney team and local partners sold 3,000 lights in three days. Success did not follow immediately elsewhere, but evaluation of what had worked on Mafia led to a key discovery: sales would not take off without the support and profile of respected community networks. That led to a strategy built around head teachers.

How a school campaign worked

Ashden's 2013 assessment describes the model in detail.

1. Meeting. A SunnyMoney field team invited groups of 20 to 80 head teachers to a meeting at a central venue, often a school. The team explained the benefits of solar lights and the products available.
2. Orders. Head teachers returned to their schools and collected orders from pupils' families.
3. Incentives. Schools received a free study light for every 40 orders, or a larger light for every 80.
4. Delivery and sales. About two weeks later, the team returned to collect orders and money and to hold a sales meeting, followed by a second sales meeting a few weeks after that.
5. Follow-on network. Once demand was established, SunnyMoney set up a more permanent distribution network of local agents and shops.

The approach exploited a strength that was not about technology at all. Head teachers are trusted figures in rural communities, have a direct line to parents, and have a reason to care about whether children can study after dark. Ashden reported that in late 2012, over 38 days in two Tanzanian regions, 93% of invited head teachers attended meetings, representing schools with 96% of pupils, and over 5% of pupils bought a light. Tanzania has more than 16,000 schools, which gave the model a very large potential reach.

A 2014 Renewable Energy World article summarized the logic: a major obstacle to spreading solar lamps in Africa is the lack of retail channels in remote regions, and SolarAid addressed it by training locally respected head teachers and giving them sample lamps for demonstrations.

Products and prices

SolarAid does not manufacture lights. It sources from manufacturers, including d.light and Greenlight Planet, the maker of the Sun King range. Ashden reported campaign prices of about US$7.50 for d.light's S2 study light, up to US$40 for larger lights, with subsidies applied in some cases. Direct sales outside campaigns cost 15% to 20% more because they were unsubsidised and included transport.

Warranties ran from six months to two years depending on the product. Ashden reported a return rate of about 1%, with SunnyMoney estimating that up to another 1% might fail without being returned. Earlier lights used nickel-cadmium batteries lasting about two years; newer lithium iron phosphate models were expected to last five years or more.

Growth and recognition

Sales milestones

SolarAid's sales grew quickly in the early 2010s.

SolarAid's website now cites 2.5 million solar lights distributed and 13.3 million people in sub-Saharan Africa reached. Its estimates of people reached are based on multiplying lights by household size; Ashden's 2013 calculation, for example, assumed about five people per household.

Awards

In 2013 SolarAid received three notable awards:

• A Google Global Impact Award, from Google's philanthropic arm.
• A Guardian Sustainable Business Award.
• An Ashden Gold Award, one of the best-known prizes for sustainable energy in development.

The Ashden judges highlighted the distribution model: "One of the biggest challenges in getting sustainable energy to the poor is reaching the 'last mile'," they wrote.

Organisation at its peak

Ashden reported that in the 2012/13 financial year SolarAid had 62 employees and income of £3.58 million, of which 68% came from sales and 32% from grants and donations. SunnyMoney had 49 staff, 35 of them local. Renewable Energy World listed funders including the UK Department for International Development, the U.S. Agency for International Development, Solarcentury and Google, and said the organisation intended to become self-sustaining by 2017.

Catalysing a market, then leaving

Tanzania's tipping point

The clearest measure of SolarAid's impact in Tanzania comes from surveys by SNV, the Netherlands Development Organisation, cited in SolarAid's ten-year review. In 2012, an SNV baseline study in Tanzania's Lake Zone found that only 3% of households were using solar products. By 2015, follow-up research found that more than 50% of households in the six regions of Mwanza, Mara, Shinyanga, Simiyu, Geita and Kagera were using solar lighting. By late 2015, 41% of rural respondents described solar as the most affordable and easy-to-use energy source, compared with 17% for grid electricity.

SunnyMoney was a large part of that change. SolarAid reported that its sales accounted for 52% of all solar light sales in Tanzania in the first half of 2014, 36% in the second half of 2014, and 44% in the first half of 2015. Based on data from the International Finance Corporation, it believed Tanzania's market moved from "early adopters" to "mainstream" in 2014. SNV's Tanzania programme manager, Josh Sebastian, was quoted saying the work had been "transformational in putting kerosene lanterns on the pathway to extinction."

A SolarAid blog post from September 2015 added context. Tanzania had become the continent's most thriving solar light market, accounting for 12% of Lighting Global-tracked sales in Africa. SunnyMoney had sold over 900,000 lights there, reaching about 15% of the off-grid population, but for every light it sold, about four were sold by other organisations.

The price of success

That growth brought competition. SolarAid's September 2015 post, titled "Tanzania and the price of success," described lower prices, more choice, wider availability and greater awareness, but also counterfeit and poor-quality imitation lights appearing in markets and shops. Revenue fell, and the drop threatened SunnyMoney's operations in other countries. "For the time being we've scaled back our operations and reduced SunnyMoney's sales capacity," the charity wrote. A July 2015 restructuring announcement said the chief executive had decided to reduce the size of the Tanzanian office.

SolarAid's review says that by November 2015 SunnyMoney had pulled out of Tanzania, and by May 2016 out of Kenya, having "largely completed our mission of catalysing the market to a point where other social enterprises and small businesses could take over." It described the exit from Kenya positively: SunnyMoney no longer worked there "because we have helped develop the market for solar lighting, in some areas, to the point that local businesses are operating successfully."

The review quoted Erica Mackey, then chief operating officer of Off Grid Electric, a Tanzanian solar-as-a-service company, describing SunnyMoney as a "market primer": "We have found it much easier to penetrate markets where customers have already had a positive experience with solar products."

Lessons drawn

SolarAid's own conclusion was that it needed to adapt faster. "When we next catalyse a market," it wrote in 2015, "we will need to ensure we are in a better position to adapt quickly." The episode illustrates a tension built into the market-building model: a subsidised or philanthropically backed seller that succeeds in creating a market is likely to be undercut by the commercial competitors it helped attract, and must then either compete, change role or leave.

The wider off-grid solar sector

SolarAid did not operate in isolation. Its growth coincided with a rapid expansion of solar power worldwide and of companies specifically serving off-grid households. Its ten-year review, citing REN21's 2016 Global Status Report, noted that global installed solar capacity grew from 6.7 gigawatts in 2006 to more than 227 gigawatts in 2015, and that in 2015 investment in renewables in developing countries exceeded that in developed countries for the first time.

The same review profiled several of the businesses that grew up in East African off-grid markets, which shows how crowded the field became:

Two developments stand out. First, pay-as-you-go models, in which customers pay small amounts by mobile phone to unlock a device, made larger solar home systems affordable without a large upfront payment. Second, as these companies grew, a charity-backed distributor of small lights became less necessary in markets that commercial firms could reach. Both trends help explain why SolarAid withdrew from Tanzania and Kenya and moved towards services in harder-to-reach places.

Why sell rather than give?

SolarAid's decision to sell lights rather than distribute them free was its defining choice. Its reasoning, as described in its own materials and in press coverage, rests on several arguments.

• Money stays local. Climate News Network reported that selling rather than donating keeps money in local communities, creates jobs for local agents and lets profits fund further expansion.
• Businesses outlast projects. A free distribution ends when the funding ends. A network of local agents who earn a living from selling and servicing lights can continue.
• Paying signals value. Households that pay for a product are more likely to want it and to use it. The Zambian randomised study offers an illustration of the risk with free distribution: only about 15% of pupils given free solar lights in that trial reported actually using them.
• Free goods can undermine markets. Giving products away can make it harder for local businesses to sell the same products.

There are counterarguments. Selling, even at subsidised prices, excludes the poorest households, who may not be able to afford even a $7.50 light. SolarAid's Light Libraries, where pupils borrow lights from school, and its newer service models are partly an answer to that problem. And because SunnyMoney itself relied on philanthropic capital and subsidies, it was not a pure market actor either. In practice, SolarAid's model sits between traditional aid and a commercial business.

SolarAid today

Where it works

SolarAid now concentrates on Malawi and Zambia, where SunnyMoney operates through local subsidiaries. Wikipedia also notes work through partners in Uganda and Senegal. Its 2023/24 annual report lists partnership testing in Senegal (with a company called Moon) and Sierra Leone (with Easy Solar), and Light Library installations in Madagascar through a consortium called Solar United, while its Kenyan and Tanzanian entities are listed as dormant.

Programmes

The charity's programmes have broadened from selling lights.

The shift to Energy-as-a-Service is a notable change. In effect, SolarAid is returning to some of the territory of its early "macro solar" experiments, but with the lesson that systems must be maintained, built into the business model.

Leadership

The 2023/24 annual report names John Keane as chief executive and Lionel Marumahoko as chair, appointed in September 2024. Mirjana Škrba, the outgoing chair, remained as treasurer.

Timeline at a glance

Finances

The Charity Commission register provides SolarAid's reported income and expenditure for recent years.

For the year ending March 2025, the register shows donations and legacies of £1.88 million, charitable expenditure of about £3.31 million, and spending on raising funds and other expenditure of about £639,610. The charity reported about 82 employees, down from 90 the previous year.

The 2023/24 annual report gives more detail on that year. Group income was about £2.5 million, down 11% from £2.8 million. Donations were £1.03 million, programme grants £847,000, and a single funder, the Norwegian energy company Statkraft, contributed £500,000. Income from solar light sales fell from £1,015,000 to £644,000. The annual report gives total expenditure of £3.8 million, which differs from the register figure for that year; the difference likely reflects how group and charity accounts are presented, and readers should consult the full accounts for detail. Free reserves were £560,000, against a stated minimum of £400,000, while the Malawi and Zambia subsidiaries had reserve deficits of £735,000 and £548,000 respectively, supported by loans.

The annual report lists principal risks including a difficult fundraising environment, cost-of-living pressures, delays in institutional funding, unpredictable stock purchases, dependence on a single major donor, currency movements, political and economic change, climate events and cybersecurity.

Two trends are visible. Sales income, once the majority of SolarAid's revenue, has fallen sharply as the organisation has moved away from mass light sales. And spending exceeded income in several recent years, drawing down reserves.

What do solar lights actually achieve?

SolarAid's claims of impact are based largely on its own surveys and standard estimating formulas. It is worth separating what is well established from what is less certain.

Kerosene displacement: strong evidence

The most consistent finding, from both SolarAid's own surveys and independent research, is that solar lights reduce kerosene use. Ashden's 2013 report found weekly kerosene spending among surveyed households fell 71%. Independent studies reviewed by researchers generally agree that solar lanterns reliably reduce kerosene consumption. A study by Kudo and colleagues in Bangladesh, for instance, found a significant decrease in kerosene use.

Emissions: estimated, not measured

SolarAid's emissions estimates rely on a standard default. Ashden reported that SolarAid's 2013 figures of about 36,000 tonnes of CO2-equivalent avoided a year and about 15 million litres of kerosene a year were based on a UN Clean Development Mechanism default value of 0.092 tonnes of CO2-equivalent per light per year, or about 38 litres of kerosene. SolarAid suggested actual savings may be higher. These are reasonable estimates, but they are calculations from a default factor, not measurements. They also do not include black carbon, which, as Lam and colleagues showed, may make the climate case for replacing wick lamps stronger than CO2 figures alone suggest.

Education: mixed evidence

SolarAid's early surveys reported that 75% of respondents named studying as the main use of the light and that study time rose by about an hour a night. Independent randomised trials paint a more complicated picture.

The Zambian study is especially relevant because it was conducted in one of SolarAid's current countries. Its authors suggested two explanations for their null result: battery-powered flashlights had already become the dominant light source in rural Zambia, and energy access appeared to be a relatively unimportant input to education in settings like Zimba compared with other factors. Overall, the evidence supports the idea that solar lights improve the quality and cost of lighting; it is much weaker on whether they improve school results.

Household money: plausible, self-reported

Savings on kerosene are real for households that previously used it, as Ashden's survey suggests. SolarAid's website cites a figure of US$315 million saved by families a year. That figure is SolarAid's own estimate and depends on assumptions about how much kerosene each light displaces and for how long lights remain in use.

Worked example: payback on a solar light

The following uses the figures reported by Ashden in 2013. It is an illustration based on those survey averages, not a guarantee for any household.

• Weekly kerosene spending before buying a light: US$1.60.
• Weekly kerosene spending after: US$0.50.
• Weekly saving: $1.60 - $0.50 = $1.10.
• Price of a basic study light at campaign prices: about $7.50.
• Weeks to recover the cost: $7.50 / $1.10 = about 6.8 weeks.

On those numbers, a household would recover the price of the light in under two months. Over a year (52 weeks), the saving would be about 52 x $1.10 = $57.20, several times the price of the light. Over two years, the minimum warranty period for some products, it would be about $114.40.

The calculation shows why selling lights, rather than giving them away, can work for poor households: the product pays for itself quickly if it displaces kerosene. It also shows the sensitivity to assumptions. A household that used little kerosene, or already relied on cheap torches, would save much less, which is consistent with the Zambian study's observation about flashlights.

Criticisms and limits

Market building versus market distortion

SolarAid's model relies on subsidies, philanthropic capital and donated incentives such as free lights for head teachers. That helped create demand, but the same subsidies can undercut purely commercial sellers. SolarAid itself acknowledged this tension in Kenya, expressing the hope that local businesses would expand "without the need for further subsidised solar." Its decision to exit Tanzania and Kenya once commercial markets emerged is consistent with treating subsidy as a temporary market-building tool.

Quality and counterfeits

Cheap, poor-quality solar products can damage consumer trust. Leggett warned in 2015 that some "ghastly cheap products" did not last and could harm solar's reputation. SolarAid's 2015 post described counterfeit and poor-quality imitations appearing in Tanzanian markets. In Furukawa's Ugandan study, in which test scores fell, the author suggested that flickering from lamps that were not fully charged may have hurt children's productivity. Quality-assured products, warranties and after-sales service are therefore central, not optional.

Repair and waste

Small solar products contain batteries and electronics with limited lifespans. As millions of lights reach the end of their lives, repair and disposal become real issues. SolarAid's current repair work is a response to this, but the scale of the problem is large across the sector.

Evidence based on self-reported data

Many of SolarAid's headline impact figures, such as people reached and money saved, are its own estimates built from sales numbers and survey averages. They are plausible but not independently audited impact measurements. Independent trials have found smaller or more mixed effects on outcomes such as education.

Financial sustainability

The goal of financial self-sufficiency by 2017, reported in 2014, was not reached. Sales income has fallen, and the 2023/24 annual report shows dependence on a small number of large funders and reserve deficits in the African subsidiaries.

Unmet targets

SolarAid's stated aims have moved. Ashden reported in 2013 a goal of eliminating the kerosene lamp from Africa by 2020, and Renewable Energy World in 2014 reported a goal of replacing all kerosene lamps with solar lights by 2020. Those targets were not met; the current target is to light every home, school and clinic in Africa by 2030. Hundreds of millions of people still lack electricity, which is a reflection of the scale of the problem rather than any single organisation's failure.

Common misconceptions

"SolarAid gives away solar lights." For most of its history its main model has been selling lights, at subsidised or market prices, through SunnyMoney and local agents. Donations pay for programmes, subsidies and market-building, and some programmes such as Light Libraries provide shared access to lights.

"SolarAid makes solar lights." It does not manufacture products. It sources lights from manufacturers including d.light and Greenlight Planet.

"SolarAid still works across East Africa." SunnyMoney withdrew from Tanzania in 2015 and Kenya in 2016. Its main operations are now in Malawi and Zambia.

"Solar lights are proven to raise exam results." Randomised studies are mixed, and a large Zambian trial found no effect on national exam performance.

"The emissions savings are measured." They are estimated using a standard default value per light, not directly measured.

Practical points for donors and supporters

1. Look at the Charity Commission record. SolarAid is charity number 1115960. The register shows income, expenditure, trustees and annual accounts.
2. Read the annual report. It explains what share of spending goes to programmes, which countries are active, and what the main risks are.
3. Understand what your donation pays for. SolarAid's website says each £15 donation funds one solar light, but much of its current work is about services, clinics, repair and enterprise support rather than single lights.
4. Weigh the evidence. Kerosene displacement and household savings are well supported; education and broader economic effects are less certain.
5. Consider the market context. In places where commercial solar markets now function, the case for subsidy is weaker. SolarAid's focus on harder-to-reach communities, clinics and service models is aimed at areas where markets have not yet reached.

Frequently asked questions

When was SolarAid founded and by whom? In 2006, by Jeremy Leggett, founder of the solar company Solarcentury, which pledged 5% of its annual profits to the charity.

What is SunnyMoney? SolarAid's wholly owned social enterprise, which sells solar lights in rural areas through local networks, originally through head teacher-led school campaigns.

Where does SolarAid work? Mainly in Malawi and Zambia, with partner work in other countries including Uganda and Senegal, testing in Sierra Leone, and Light Library work in Madagascar. It previously operated in Tanzania and Kenya.

How many lights has SolarAid distributed? Its 2023/24 annual report gives 2,362,600 lights distributed to March 2024; its website cites 2.5 million.

What awards has it won? In 2013, a Google Global Impact Award, a Guardian Sustainable Business Award and an Ashden Gold Award.

Why did it leave Tanzania and Kenya? SolarAid says markets in those countries developed to the point that commercial and local businesses could serve them. Competition, including from low-quality imitations, also cut SunnyMoney's revenue.

How big is SolarAid financially? The Charity Commission register shows total income of about £3.95 million and similar expenditure in the year to March 2025.

Do solar lights help children learn? The evidence is mixed. They often increase study time, but several randomised trials, including one in Zambia, found no improvement in exam results.

Are solar lights better than kerosene? For lighting quality, cost over time, fire risk and indoor air pollution, yes. Kerosene wick lamps also emit large amounts of black carbon, according to Lam and colleagues' 2012 study.

What is Energy-as-a-Service? A model in which households or clinics pay for an energy service rather than buying equipment outright, with the provider responsible for keeping systems working. SolarAid is piloting it in Malawi and Zambia.

Sources

• SolarAid - Wikipedia
• SolarAid official website
• SOLAR AID, charity 1115960 - Charity Commission register
• SOLAR AID financial history - Charity Commission
• SolarAid Annual Report 2023-2024 (via OpenCharities)
• SolarAid 10-year market update, December 2016
• Tanzania and the price of success - SolarAid (2015)
• SolarAid restructured, July 2015
• SolarAid - Ashden Award Winner 2013
• SolarAid project benefits 9 million in Africa, 1.5 million lights sold - Climate News Network (2015)
• Has SolarAid brought solar to more Africans than any other organization or company? - Renewable Energy World
• SolarAid: providing a brighter future for African villages - The Guardian (2013)
• Household light makes global heat: high black carbon emissions from kerosene wick lamps - Lam et al. (2012)
• Let there be clean light: kerosene lamps spew black carbon - UC Berkeley
• Rural electrification, solar lanterns and education in Zambia - Stojanovski et al., World Bank Economic Review (2021)
• Powering education - Hassan and Lucchino, LSE Centre for Economic Performance
• Furukawa, Do Solar Lamps Help Children Study? Contrary Evidence from a Pilot Study in Uganda, Journal of Development Studies (2014)
• Kudo, Shonchoy and Takahashi, Impacts of Solar Lanterns in Geographically Challenged Locations: Experimental Evidence from Bangladesh - IDE Discussion Paper 502 (2015)