Blog
Posted on March 20, 2017

It takes a village: Lessons from digitizing social school fees payments

Flexipay pic 3

Our work on the Kenya Financial Diaries made it painfully clear to us that school fees are incredibly expensive for low income families.  The lowest end public schools often ask about KSh 20,000 per year for a single student, when rural household incomes often average around KSh 6,000 per month.  Many families are able to make school fee payments by calling on help from family and friends outside the household—older siblings, aunts, uncles, grandparents, spouses working in other cities, and others—to help pay for fees.  But this social subsidy solution is imperfect.  Fifty-seven percent of Diaries families still had a child sent home from school at least once during our Diaries study year. 

We know finance can’t solve the biggest problem–making school fees cheaper—but, what if some payments improvements might make this social subsidy more efficient? 

In the third school term of 2016, BFA and FSD Kenya worked together to test a conceptual school-fees payments idea inspired by the Diaries.  We called the idea “Flexipay,” which is a digital billing and payments concept that could offer some blending financing options to help low income people manage big expenses.[1]  At the core, on a digital platform, a provider issues an SMS bill to the main party plus registered helpers.  All participants can pay using digital payments against a specific account.  Additional add-ons could include advances (small loans/credit) towards the payment (disbursed directly against the outstanding m-bill), wallets to build up funds for the payment, and financing for the biller based on receivables. 

We decided to first test only the core concept of issuing m-bills to primary account holders plus their registered helpers and enabling all parties to pay against their accounts from digital wallets.  We introduced a digital payment option at the school through M-PESA Paybill.  M-PESA users can pay the school directly making a payment to any student’s account.  This entails a fee, which in this case was paid by money senders.  We also directly engaged the social networks of payers with communications on bills, balances, and payments. We wanted to see whether these changes could add enough convenience and transparency to speed fee payments and reduce absences at a single low-cost rural, public day school.  And, if so, what would be the business case for introducing such a system?

We chose to test these ideas in a low-cost way before considering a larger investment of both funds and the time of a commercial partner, so we set up Paybill for the school and used a very simple—though time consuming—manual process to account for payments and conduct all the SMS communications to parents and helpers. 

Figure 1:  Schematic of experiment implementation.

flexipay1

Our very small experiment with one rural school has given us some reasons to be hopeful for the Flexipay value proposition.  We learned five key lessons, which you can read more about in this brief learning document.  In summary: 

  1. Parents and schools loved the introduction of Paybill as a fee payment option. In the first term after Paybill was introduced, 66% of student accounts had at least one Paybill payment.  The school was able to eliminate cash payments completely, which previously accounted for about 9% of all payments. Other payments were done through bank branches and agents, meaning the school received all payments digitally.  While Paybill was more expensive for parents and helpers than free bank deposits, it was preferable for many who could pay without leaving work, traveling to a bank or bank agent, and for avoiding problems with bank agents lacking sufficient float to complete payments. 
  1. Digital payment of fees can reduce student absences. While the average absentee rate at our test school did not decline, students whose accounts had at least one Paybill payment had fewer absences on average than their peers, a hopeful indication that Paybill may have played a role in making payments happen faster and reducing absences among this group.  Those with no Paybill payments missed an average of 6.5 days of instruction versus an average of 4.2 days for students with some Paybill use.  Given that there were only 40 days of instruction, every day matters.
  1. Parents and helpers loved SMS communications from the schools and want more. Parents and helpers loved being kept up to date on balances and payment deadlines. Helpers felt more engaged and willing to contribute knowing they were informed of balances directly and could pay the school directly on behalf of students.  In closing focus groups, parents in particular, asked if the platform could be expanded to inform them of student absences, behavior problems, and academic performance so that they could hear directly from the school and help correct problems at home.  Bulk SMS makes this relatively easy for schools, and research shows that such communication can have a real impact on learning outcomes.[2]
  1. There are hopeful indications that messaging can substitute for sending students home. Traditionally, parents and helpers trigger payments when a child is sent home.  That’s when they may pull money out of businesses or sell assets to address what’s become an urgent need.  Students were used as “exhibits” to employers and creditors, showing them that paying on outstanding debts and wages was also urgent.  Some parents told us the messages can substitute, even just before students are sent home. They can be easily forwarded to others’ phones.  We also saw that 67 payments (on an enrolment of ~400 students) were received on Paybill from numbers not registered as helpers, indicating some forwarding of billing messages.  Over time, this could further reduce absences by prompting payment before lessons are missed.
  1. The business case for Flexipay at schools will be driven by payments revenue. Our test school was very enthusiastic about receiving Paybill, but are less interested in investing in the kind of accounting system that would allow messaging on payments and balances to be automated.  They have a fixed cost in the form of a bursar, and saving his time is not a priority.  Messaging costs were low, about KSh 13 per student for the term.  But payments revenue for Safaricom was substantial, about KSh 60,000 for a year at the usage levels achieved in this experiment.  This would be more than KSh 500 million per year if all secondary schools in Kenya were covered and had similar average usage.  This offers Safaricom or another payments provider a substantial margin to cover things like software to account for payments and the costs of school acquiring.  At scale, margins on fee payments could even be reduced to invite platform stickiness or to generate income from float, especially for banks hosting school accounts.  But the revenue stream based on payments means that it would be much harder for a third party fintech player to make a business of Flexipay outside of a revenue share agreement with a payments provider. 

Flexipay can’t fix the high cost of education, but we are hopeful that this kind of engagement can help mobilize social fee payments and help students miss fewer classes, getting more value out of the substantial investments their families make in education. 

[1] Here, we were looking at school fees, but the concept could be equally applied to other large expenses like health insurance premiums, healthcare out-of-pocket costs, agricultural inputs, building materials, and motorbike purchases. 

[2] https://www.povertyactionlab.org/evaluation/facilitating-school-parents-communication-through-sms-technology-chile