Harima lives in eastern Uganda, about an hour’s drive from the nearest town. She purchased a solar home system and was quite happy with the product. It provided illumination well into the evening, and an outside light gave her an added sense of security. Or at least it did until she missed a payment to the company that financed her system, and her lights were turned off. Her darkened house communicated a message, and each of her neighbors understood it: Harima does not have any money, not today.
On the nights she spent in the dark, Harima felt embarrassed and ashamed. Nobody had lied to her. She had understood and agreed to the terms of her solar loan, which said that if she failed to make a payment, her system would be remotely locked. Her provider offered her additional time to repay, but it could extend that grace period for only a few days. And without the lockout technology embedded in her solar home system, how could it be sure of recovering value on small-asset loans in remote areas, where repossessions are expensive, if they’re even possible?
Pay-as-you-go (PAYGo) solar loans like Harima’s are part of a widespread experiment with alternative forms of collateral that have far-reaching implications for how low-income households access credit. In the United States, 70 percent of subprime auto borrowers now have “starter interrupt” devices in their cars. In Mexico and the United States, consumers can now purchase, on credit, smartphones that are remotely disabled in the event of nonpayment. The rapid emergence of remote lockout technology in lending raises some important and difficult questions: When is it appropriate to disable a financed asset? What benefit does lockout technology provide to a lender? What balance must be struck between borrowers’ dignity and lenders’ need for security?
From a PAYGo creditor’s perspective, lockout technology prevents unscrupulous users from stealing or reselling assets. But the greatest benefit of this technology is linking people’s use of their solar systems to their loan payments. Lockout technology turns borrowers’ lighting (or lack thereof) into an immediate and tangible reminder to repay their debt, making clean energy a manageable, as-you-go experience, akin to a prepaid mobile phone plan.
But this may not be how it feels to customers. Low-Income Financial Transformation (L-IFT) is a research firm that has been exploring links between financial services and energy consumption in low-income households. From August 2016 to June 2017, L-IFT conducted Financial and Energy Diaries Uganda in 12 districts across Uganda. In its April/June 2017 end-line survey of 1,327 customers, it included questions about respondents’ awareness of PAYGo and their willingness to take a PAYGo loan, before and after the terms were explained to them.
The results were illuminating. Only 45 percent of respondents had heard of PAYGo, and only 19 percent were confident they knew what it was. L-IFT researchers then explained the basics of PAYGo solar (that the customer acquires a solar device while paying little bits over time). People were enthusiastic, with 63 percent saying they would be more interested in acquiring solar if they could do it via PAYGo. But when the lockout nature of the loan was explained to them, interest dropped. Half of all respondents did not like the idea, and interest in acquiring solar via PAYGo dropped to 40 percent.
Some of the respondents said that lockouts were counter-productive. Switching off their lights during hard times would force them to borrow or buy kerosene, making it more difficult to resume paying off a solar loan. Others indicated it showed a lack of trust. Many wondered what restitution they would receive for the days their lights were turned off, if and when they completed payment on time. As one customer said: “There is pressure to pay on time, and yet sometimes money is not available. ... It’s not good to switch off someone; after all, they will eventually pay.”
Respondents also considered solar systems to be very different from prepaid electricity or airtime. The latter are delivered by an external provider, relying on widespread, shared infrastructure. But the solar system is in your house; its electricity is not going anywhere else.
The issue of over-indebtedness is a growing one in the PAYGo space, and customers’ concerns about being locked out should be taken seriously. CGAP/FIBR research shows that PAYGo customers do sometimes struggle to make payments and must cut back on consumption to make solar payments, albeit not in ways they considered severe. Turning off someone’s lights imposes an additional expense, which may cause them to fall further behind.
There is conflicting evidence on whether lockout technology prevents defaults, despite lockouts resonating with creditors. “I have not seen any data to indicate that lockouts have a significant impact on payment rates. They are by no means a magic bullet for collections," said Daniel Goldfarb, CEO of Lendable, a risk analytics firm in the PAYGo sector. Borrowers still default on PAYGo loans — at higher rates than in other credit sectors — and providers still must carry out repossessions. Moreover, the model may not translate; we suspect that lockout technology may be inappropriate for income-generation loans, when a customer is depending on the device for revenue to service their loan. Turning off a motorcycle taxi or solar irrigation system may be even more counter-productive if your goal is to recover capital.
On the other hand, very few lenders have created a scalable solar financing operation without using lockout technology. Microfinance solar loans have existed for decades but have never reached scale, partially because lenders have been cautious in assessing borrower risk and diligent in requiring collateral. Lockouts may be painful for the customer, but PAYGo solar is growing because customers are willing to accept that tradeoff. Hundreds of thousands of people have completed PAYGo loans, and we should be cautious of disrupting a model that is working better than its predecessors. In addition, lockout technology is the key enabler of flexible repayment schedules, which was one of the reasons participants in the L-IFT survey responded positively to the idea of PAYGo in the first place.
Lockout technology is likely to be necessary for asset financing in Sub-Saharan Africa for the time being, but it remains surprisingly under-researched. Various providers are offering short-term liquidity loans, which are helpful to customers in emergencies. Providers such as PEG have experimented with insurance products that keep customers’ lights on during periods of financial difficulty. More experimentation is needed to understand the costs and benefits of this technology for providers and customers, and whether there is a better solution.
I think even the strongest advocates of PAYGo solar would agree with you. A purely commercial solution is just not feasible for the poorest 20-40% of households. Hopefully costs will continue to go down, but we need to find a sustainable way of subsidizing energy access that does not spoil the market for commercial providers. I'm quite excited by some of the work that is going on in Kenya to bundle energy subsidies with cash transfers.
Initial signals in India also are very similar, the remote lockout is a big deterrent and promotes the tendency to tamper with the system which leads to higher level of default. A flexible plan is required with emergency days of 7 days linked progressively with the regular payment.
I'm very interested to hear more about any testing that has been done in India, and certainly if remote lockouts are actually leading to higher defaults. Could not agree more on the flexibility/rewards approach, and I think a number of firms are leaning more this way, for example Mobisol.
When you rent a boat and missed the regular payments the boat shouldn't sink. The emergency days and rescheduled payments might be the answer for better PAYG systems.
Great blog. Where can I read the full data analysis of the L-ift Uganda diaries - their site only has that Methods booklet you link to? Have I missed a link in your blog to the data analysis report?
This is a good article and raises another question: How much time do these small solar units simply not work at all because it's too cloudy? At noon in central Africa eg Uganda, heavy clouds can reduce insolation (useful solar energy) by up to 80%. A small solar panel rated at 10W never makes that much even in direct strong sun, it might make 8W. Reduce that by 80% and you get 1W output, and you get that only for maybe 3 hours a day when sun is highest. Now imagine that it's rainy season and it's cloudy for weeks. How many thousands of people are paying for solar units that simply don't work (make enough power to be useful)? Since people need light, they will keep their kerosene lamps, and keep paying others for phone charging.
This problem affects hundreds of millions of potential users. The whole central Africa and Amazon region has heavy clouds for months, a lot of it is (was) rain forest! In South Asia the monsoons blanket the sky for months at a time. In the Himalaya and Andes mountains the winters blanket the skies.
CGAP should do a study about how much power these small solar units *actually* deliver over a year to users, and for those who can use them only intermittently, how much do they *actually* save on kerosene and phone charging?
Hi Simon, Everybody can actually get access to the data portal that holds practically all angles of our data. Feel free to send an email to aswinderen [at]l-ift [dot] com
It is a good article. I think after a lockout, there is a need for follow up to establish clear reasons. Sometimes, they may be beyond financial such as death or family migration etc
Really insightful article and has got us here chatting a lot about the pros and cons of lockouts. Have you seen or heard of any more of the experimentation that you advocated for here? Would you be interested in discussing further field research into this?
Dave - www.solarisoffgrid.com
Thank you for the comment. We have not yet seen detailed research into this issue, although we have heard from providers that are interested in pursuing it. I know that researchers from Berkeley are among those looking into the question. Happy to discuss further, please contact me at [email protected]