Digital Lending in Kenya - General Notes and Outlook;

Thoughts on the Digital Lending Space and a Q&A With Kevin Mutiso the Chair of DLAK

Introduction to the Digital Lending Space

The digital lending space in Kenya has in the recent past been mired in controversy with proposed regulations to curb digital lending being passed by parliament, at the same time the Central Bank has been mulling increased regulation of the space. The increased scrutiny comes at a time when there has been an increase in defaults whilst a number of digital lenders have taken on more unconventional approaches to recovering bad loans. An additional bone of contention has been the seemingly usurious rates applied by digital lenders with some Annual Percentage Rates (APR) a measure of the total cost of credit, reaching over 100% per annum. 

The Cambridge Centre for Alternative Finance (CCAF) 2018 report showed that there was a total volume of US$ 304.5 billion in alternative finance globally - mostly digital lending issued in 2018. Of this amount, China accounted for over two-thirds of total volume with a volume of US$ 215.4 billion. The largest financing model was a Peer to Peer (P2P) consumer Marketplace model which was worth US$ 195.2 billion or 64% of total alternative finance raised. P2P business funding was second with 50.3 billion worth of alternative finance. Over half of total funding for onward lending was raised from institutional investors such as banks, pension funds and mutual funds. In Africa and the Middle East only 17% and 12% of funding respectively was raised by institutional investors.

In Africa, over 58% of total alternative finance is marketplace consumer lending, with Balance sheet-based lending i.e. secured accounting for 22% with peer to peer marketplace business lending accounting for only 9%. In Kenya, the 2019 FinAccess Household survey showed that digital loan apps had an uptake of 8.1% amongst the sample surveyed. This was almost double the uptake rate of personal loans from financial institutions. The uptake rate amongst the sample surveyed of overdrafts was 0.2% and credit cards at 0.5%. The most popular credit product by uptake rate i.e. percentage of sample surveyed who have taken that product was shopkeeper credit at 29.7% followed by borrowing from family or a friend at 10.1%. Digital loans are thus the third most popular means of accessing credit in Kenya behind shopkeepers and family/friends.

In the Kenyan market space, Banks and MNO’s pre-dominantly the tie-ups between KCB, NCBA and M-Pesa dominate the market with over 97% market share according to a report by Microsave. The bulk of digital borrowing is done to manage day to day expenses, emergencies as well as acquiring or maintaining inventory for businesses. Multiple borrowing is prevalent with over 62% of borrowers having more than one digital loan.

I wanted to understand what the opportunity is and how the space will evolve with time. The recent actions of the People’s Bank of China to rein in online lending after the failed Ant Group IPO shows that globally regulators are waking up to the digital lending space and grappling with how to regulate the industry. The same regulatory grappling has been evident in the UK with the FCA being given powers to regulate the “Buy Now Pay Later” industry.

Evolution of Digital Lending in Kenya;

The Digital Lending space was officially born in my view in 2012 with the launch of M-Shwari. M-Shwari is a banking service in which Safaricom collaborated with a bank in this case NCBA to offer micro-savings and lending capabilities to its massive M-Pesa client base. The thinking was simple, we have all this data on M-pesa transactions, airtime usage, data usage and other data points; we can use this to create credit scoring that will enable micro-credit to our M-Pesa clients. We can also offer a savings account since there is an observation that a number of our clients are using M-Pesa to save their money. The same product was also extended to KCB with its KCB-Mpesa offering which has scaled to over 10m customers.

This was later augmented by the Fuliza product by Safaricom which is essentially an overdraft facility that is activated on your M-Pesa wallet. The thinking was that Safaricom noticed many transactions failing due to insufficient balances despite the wallet being replenished at a later date. If you could just enable the customer to complete the transaction, then he would be able to payback once he received some funds. Fuliza has gone on to be very successful and it has become a part of Kenya’s Lexicon.

Stoner254🍁🇰🇪 @Stoner254K

Me receiving money in my mpesa but I know fuliza will take all of it

In essence, the emergence of M-Pesa as the de-facto payments layer of Kenya’s financial system, a maturing of digital financial client information and lastly the emergence of the Credit Reference Bureau model in 2013 were the kindling required to ignite the digital lending space in Kenya. Since this period, numerous players have emerged in the market with the main ones being M-Shwari, Fuliza, Branch, Tala, Equitel, Timiza by Barclays amongst others.

There are a number of business models that exist within the digital lending space;

Direct Consumer Lending

This is the primary lending model used by the major players. In this model, digital lenders use proprietary credit scoring algorithms to make decisions on who to lend to and what amount to lend. The lenders are also in charge of on-boarding and disbursements with the latter happening primarily on M-Pesa. The main players are M-Shwari, Fuliza,KCB M-Pesa, Branch, Tala, Equitel and Timiza. Nonetheless there are plenty of players. There are also digital salary advance products where customers borrow against their salary. In this case, the lender has to plug-in to the company’s payroll data.

The main considerations in this space are distribution, ability to verify and carry out fraud checks, ability to report default and importantly, ability to lock out a defaulting customer. In this case, it’s easy to see how M-Shwari, Fuliza and KCB M-Pesa have the biggest market share. The distribution is easy due to M-Pesa’s scale. In fact, Ant Financial in China and Amazon in the USA have managed to convert their significant user bases particularly third party suppliers into loan customers by partnering with financial institutions

.Fraud checks and identity checks are done also on the back of Safaricom and being locked out of M-Pesa comes at a significant cost and thus there is a strong incentive to repay. Despite this, we are starting to witness people converting to cash due to these digital loans. It’s quite common to hear someone telling you “don’t pay me on M-pesa because I have “Fulizad”. This is a useful read on what’s happening on the ground.

Buy Now Pay Later

Within the consumer space, a new smaller niche has emerged of Buy Now Pay Later financiers. The main players here are Aspira and Lipa Later. The model is based upon international players like Affirm, Afterpay and Klarna. In BNPL, the company in this case Apira and Lipa Later sign up retailers who will accept them as part of the checkout option both online and off-line. Customers then download the app, get verified and qualified for financing. Upon check-out, if you select the BNPL option, you’re offered the equivalent of a term-loan that can range from three months to six months. The retailer gets liquidity from the BNPL provider and the customer pays the BNPL company equal monthly installments.

The attraction of this model is that it tends to be cheaper than credit cards but offers debit card functionality but with delayed payments. For the merchant the advantage is that you addressable market grows as more of your goods become “affordable”. 

Here is a video that’s a good primer on BNPL;

Supply-Chain Finance

Supply-Chain Finance is business lending in which value-chains are digitised and financing is embedded within the value chain. An example is a manufacturer having a value chain that includes on one side suppliers and vendors and on the other side wholesalers and retailers on the distribution side. There are commercial interactions between the firm and all the stakeholders that bring about claims and counter-claims through the issuance of invoices and purchase orders.

As a financier, if you plug yourself digitally into this ecosystem, you can help smooth cash flows by helping suppliers for instance with invoice discounting/factoring as well as LPO finance. On the distribution side you can finance inventory. 

In this model, banks are partnering with Fintechs with banks providing balance sheets and the Fintechs providing the technology. 

Q&A with the Chairman of the Digital Lenders Association of Kenya

To delve deeper into this area, I did a short Q&A with Kevin Mutiso who’s a friend and the chairman of the Digital Lenders Association of Kenya.

Kevin Mutiso @KevinMutiso

In Kenya, the Credit Reference Bureau (CRB) has approx. 12 million individuals listed on the bureau. These 12 million individuals have 110 million credit records. Of those 110 million records, 14 million records are negative records. Not 14 million individuals. #Reality #Truth

Kevin has been involved in the digital lending space in Kenya for many years and runs Alternative Circle which started its life as digital loan provider and has gradually evolved into a digital lending solutions provider.

Samora: Could you give a broad background of the evolution of digital lending in Kenya - who were the first players in the industry, how they did distribution and how the industry looks like?

Kevin: Digital lending is a channel, what we are known for is digital-first credit experiences. Meaning we do KYC, Scoring, and disbursement usually with very little physical interaction with the customer. In 2013, the CRB system enabled digital-first as we could get credit data on customers. By 2015, we had several players and fraud was the main risk. By 2018, various business models had emerged (1-day loans to trade finance loans to Buy now pay later loans etc). In DLAK for example our products are so varied, we have leasing companies, micro-finance, nano-lending, by now pay later, salary advance, and many more). Most models started getting profitable after their 3rd year as the scoring tools improved over time.

Samora: What is the opportunity and business case behind digital lending businesses? Moreover, at a steady state what would a viable business look like in terms of three main metrics - NPL, Net Profit Margin and LTV/CAC?

Kevin: The market size is the opportunity. It is estimated that there is a Kes 200bn credit deficit for MSMEs and SMEs. Currently, including banks, we serve Kes 30bn. The main challenges are systemic, such as the cost of transactions, fraud, the first-loss risk is high and cost of capital. We believe regulation should help de-risk the business models. Profitability is varied. Nano-loans are now breaking even, buy now pay later not yet, trade finance is extremely profitable. Our interest rates vary from 1% a day to 7.5% a month. NPLs vary too...BNPL with the highest and trade finance with the lowest.

Samora: Recently there has been a lot of regulatory push to control pricing and conduct of the digital lenders in Kenya - what's your stance on all this?

Kevin: DLAK wants regulation. The regulation will standardize the market and the bad actors will have to pay for their sins. Secondly and most importantly, it will de-risk our business models significantly as regulation ensures certain standards are adhered too.

Samora: What success factors enable the top players to distinguish themselves from the rest - would like you to approach this from a tech perspective, commercial and strategy perspective - lastly from an institutional perspective i.e. composition of shareholders, funding model etc;

Kevin: The most successful lenders will differentiate themselves by being the best at lending to specific value chains. I personally do not think that being the largest lender should be the objective. The objective is to be the most relevant lender to a value chain. 

Samora: What pivot avenues do you see for some of the players in the industry now - Also how do you see the industry evolving;

Kevin: I see a future where banks partner with fintech companies to provide balance sheet as the banks cannot focus on every value chain. 

Samora: Lastly, let's talk about identity and e-KYC - how is your industry working towards enabling e-KYC and digital onboarding;

Kevin: We have created a Credit Fraud Prevention System (CFPS) at DLAK and are working on a strategy to have it regulated. This tool is a real-time tool that helps lenders manage fraud and over-indebtedness without breaching privacy rules and or replacing the CRB system. The idea is that we as lenders would collaborate on this matter to reduce risk significantly and the system would be a not-for-profit solution as its impact would improve profitability for all.

Closing Thoughts

Digital lending is rapidly evolving and the nature of lending is that you never feel the consequences of reckless lending until much later on. My overall thoughts can be summarised below;

Fraud and Digital Identity;

Digital identity and KYC are foundational elements within the Fintech industry. In Kenya, we have the Integrated Population Registration System IPRS which is underpinned by the traditional ID system. It offers a basic system for verifying identity but it has still faced issues being prone to fraud with identity theft. The issue with identity is that there are two types of identity; explicit identity i.e. who you say you are and implied identity i.e. what your digital trail actually says you are. The latter tends to be a more reliable guide to identity and underpins most of the fraud models being built by digital lenders. Explicit identity i.e. who you claim to be is prone to fraud and without a reliable central data repository can be unreliable.

I think the magic would be to have a way to integrate the two i.e. verify explicit identity and map it to your digital trail to really identify you. This would be in addition to what Kevin alluded to i.e. Credit Fraud Prevention System and an integrated government database that also provides information on good conduct. The idea would be to have such a framework without sliding into an Orwellian reality.

Listen to this for a good breakdown on digital identity in Kenya

Ultimately, work needs to be done cross multiple stakeholders to create a universally acceptable centralised KYC and data registry for financial services.

Data and Data Protection

One of the major issues within digital lending has been the fact that some of the digital lenders use aggressive recovery practices such as calling your loved ones or messaging your friends when you haven’t repaid your loan. This is achieved through data scraping algorithms that scour your phone for all kinds of data from call logs, messages, social media data, location and any other useful data point. This is enabled by clients accepting the terms and conditions of these loans.

Kenya just recently passed a Data Protection Act in 2019 that is meant to govern such practices. However the relevant bodies envisaged in the Act are yet to be actually instituted. One of my major concerns is that digital loans could actually act as a loss leader for a foreign intelligence agency or any other actor that wants to trawl through the phones of Kenyas citizens. You could gain a pulse on what people are talking about, how they behave and how they communicate. This is an area that needs strict laws. Such data in the wrong hands is a threat.

Open Banking and Instant Payments

One of the main opportunities that Open Banking can create in the digital lending space is furnishing digital lenders with better and more robust data sets from which to build out their credit scores. I see third-party credit scores emerging that are linked to repayment of loans, cash flows and other data that is accessible via open banking. Some of the existing players in the market can pivot into third party credit score platforms such as Experian.

Additonally, an Instant Payment system such as UPI of India can help reduce costs for digital lenders. Given that most of the disbursements happen on M-Pesa, the high costs for disbursing on the platform are passed on to borrowers. For lower amounts, this ends up increasing the APR.

Plaid recently launched an income verification service via API that would really enable digital lenders, particularly BNPL lenders make better credit decisions.

Adverse Selection and Moral Hazard

Adverse selection and moral hazard are two very key concepts in finance. Adverse selection occurs where there is asymmetric information in a market i.e. one side knows more than the other in a transaction. The classical lemons and peaches conundrum in used car markets is the best example. Moral Hazard occurs typically when there is an imbalance in risk-reward mechanisms. For instance, in the global financial crisis, moral hazard was present because the banks didn’t bear the full costs of the risks they took. 

Adverse selection presents itself in digital lending in the sense that typically the most desperate customers are the ones who will sign up for some of these digital loans. Good customers i.e. those with healthy income profiles typically already have overdraft facilities or credit cards within the financial system. Essentially when your business is ‘digital lending’ then naturally you will attract the worse types. There is a case to be made of shifting the business model towards embedding finance into verticals in which customers don’t self-identify as needing cash. Probably this is why as Kevin notes, trade finance has the lowest NPL and the highest returns. In this instance, customers self-identify as having good cashflows and existing verifiable incomes. in the African context when you say you’re issuing loans, people often translate this to “you’re giving us money”.

The counter argument of course is that within the African/Kenyan context, most of the borrowers are borrowing for their first time and this is their first time engaging with the financial system. Therefore, the argument of adverse selection doesn’t necessarily apply as it would in the West where pay-day lending has a severe adverse selection concern.

Moral Hazard is one of the reasons why the initial vintage of peer to peer lenders in Europe and UK failed. Lacking ‘skin in the game’ makes such marketplaces drive for volume rather than credit growth which leads to NPL problems down the road. This has been the crux of two interesting developments. On one hand, the regulation of Ant Financial i.e. the requirement to hold more capital and on the other hand, the murmurs that M-Pesa would want to open up its lending business to more players a la Ant Financial. Regulators would want a situation in which there is a balance between volume growth and credit quality. A purely market-place approach is not in their best interest. In the case of M-Pesa, it would make a lot of commercial sense to open up its platform to all forms of lenders and get a facility fee for each loan issued. Similarly, it would be in its best interest not to tie up default with being locked out of M-Pesa. However, from a regulatory perspective, it could create a moral hazard issue.

Balance Sheets and the Hidden Threat of Regulation;

Banks don’t always work well and the financial system is often suboptimal. Nonetheless, the banking system is a marvel of human ingenuity in my view. A system that is designed to allocate resources within society and enable humans to achieve their ambitions at an individual and societal level through the clever allocation of risk. One of the pivotal elements behind this is balance sheet management, in essence the idea that you can put up some capital and lever up by accepting deposits often up to 10x your capital level and lend this money. 

When interest margins are good and you manage risk properly, this is an amazing business. The issue a number of fintech lenders face is financing via capital and in some cases institutional debt that is often expensive. This exposes you to significant losses when NPL’s rise as your hurdle rates are often very high. 

Nonetheless, the biggest issue with the funding model is the regulatory threat when the regulators decide to act. If regulators decide to hike up capital requirements then a number of the digital lending models would simply collapse because balance sheet management would get much more difficult. This is why bankers keep calling for similar capital requirements in digital lending. This point can also tie up to the adverse selection problem.

Is Digital Lending a Panacea

An interesting fact is that over 62% of Kenyans have taken up multiple loans from digital loan providers. Additionally, the bulk of these funds at a consumer level are used to manage day to day expenses and emergencies. One of the issues that pop up is whether digital lending is just the modern-day Grameen experiment. Do people need digital loans or do people actually need a stronger economy that provides stable jobs, good incomes and increased economic opportunity.

There’s a useful quote from Ory Okollo that captures this conundrum;

“I’m concerned about what I see is the fetishization around entrepreneurship in Africa. It’s almost like it’s the next new liberal thing. Like, don’t worry that there’s no power because hey, you’re going to do solar and innovate around that. Your schools suck, but hey there’s this new model of schooling. Your roads are terrible, but hey, Uber works in Nairobi and that’s innovation.

During the Greek bail out, no one was telling young Greek people to go and be entrepreneurs. Europe has been stuck at 2% or 1% growth. I don’t see any any entrepreneurship summit in Europe telling them you know, go out there and be entrepreneurs. I feel that there’s a sense that oh, resilience and you know, innovate around things—it’s distracting us from dealing with fundamental problems that we cannot develop.

We can’t entrepreneur our way around bad leadership. We can’t entrepreneur our way around bad policies. Those of us who have managed to entrepreneur ourselves out of it are living in a very false security in Africa. There is growth in Africa, but Africans are not growing. And we have to questions why is there this big push for us to innovate ourselves around problems that our leaders, our taxes, our policymakers, ourselves, to be quite frankly, should be grappling with.

… I think sometimes we are running away from dealing with the really hard things. And the same people who are pushing this entrepreneurship and innovation thing are coming from places where your roads work, your electricity works, your teachers are well paid. I didn’t see anyone entrepreneur-ing around public schooling in the US. You all went to public schools, you know, and then made it to Harvard or whatever. You turned on your light and it came on. No one is trying to innovate around your electricity power company. So why are we being made to do that? Our systems need to work and we need to figure our shit out.”

Could this be applied to significant chunks of the digital lending space in Kenya? If people are borrowing to make ends meet, is digital lending a solution or a symptom of the problem?

It would be interesting to watch how this space evolves over time.