#31 First we must build Infrastructure

What Fintech Infrastructure must be built before the African Fintech Revolution takes off

Hi all - This is the 31st edition of Frontier Fintech. A big thanks to my regular readers and subscribers. To those who are yet to subscribe, hit the subscribe button below and share with your colleagues and friends. 🚀

Sponsored Content:

This week’s post is sponsored by Hyperverge. As onboarding and KYC goes digital, fintechs, banks, telcos and any other consumer facing company will need to develop capabilities to digitally onboard and verify their customers. Hyperverge is an industry leading digital KYC platform that uses advanced Artificial Intelligence and Machine Learning to help customers with their onboarding. With Hyperverge, companies can minimise fraud, minimise operational costs and provide a seamless customer experience.

I liked Hyperverge because they have proven KYC capabilities in a large market such as India which shares numerous similarities with Africa such as; limited bandwidth in some areas, diverse facial features, low quality smartphones and high rates of attempted fraud. Hyperverge has offered digital KYC to the State Bank of India, Jio and Bajaj Finserve. In Vietnam, they have partnered with FE credit. Hyperverge has performed over 400 million checks over the last two years and are currently in advanced talks with two prominent Fintechs in Nigeria.

As a proof of their AI capabilities. Hyperverge has been Ranked #2 in the US, among top 10% global Face Recognition vendors as per NIST benchmarking and is 1 among the only 3 providers in the world to have a single-image based passive liveness detection solution, certified for Level-1 PAD as per the ISO 30107 standard.

I am bullish about this solution in Africa largely due to cost and reliability. Customers could include Fintechs, Banks, Insurance companies, Asset management companies, hospitals, schools and vendors in the financial space such as CBS and DBP vendors who want to add them onto their market place. For more information, reach out to me on samora.kariuki@frontierfintech.io

Introduction

African Consumer Fintechs have been raising significant amounts of money over the last few years. The promise of a young and growing population coupled with low levels of financial inclusion have driven investor interest across the continent. This year, Africa is on track to almost double the VC investments received in 2020. What a year it will be and the promise for 2022 looks even brighter.

What becomes clear whenever you’re building a fintech company in Africa is the lack of infrastructure. My experience has not necessarily been from a start up perspective, but rather intrapreneurship within an organisation. Whenever you build something, it always becomes clear that other operators may need your service more than your target customers do. In my case, other savings associations and micro-finance institutions needed our payments platform more than our target customers. It soon dawns on you that sometimes the demand for the underlying infrastructure is the bigger business.

Similarly, a number of African consumer fintechs are pivoting to offering B2B APIs that have been built out from scratch. 

I’m reminded of a situation in 2019 when the former president of the United States, Donald Trump was on the warpath with Jeff Bezos over the US Postal Services. Jeff Bezos got himself in trouble by saying "I didn't have to build a transportation network to deliver the packages,... It existed: It was called the post office." The idea was that the success of Amazon was directly tied to the existence of the US postal service. If it hadn’t existed, Amazon would have had to build its own delivery network, probably costing billions of VC dollars whilst shifting attention from its core product and mission. Potentially, we wouldn’t have AWS now. Trump argued that the US Postal service is subsidising Amazon, but it could equally be argued that it’s doing exactly what it’s meant to be doing - enabling parcels to be delivered thus driving commerce.

In Africa, Amazon-like services such as Jumia have failed to get similar traction due to the lack of cheap last mile infrastructure such as the US Postal Service as well as well structured address systems. 

For consumer fintech to explode in Africa, there will have to be a period of infrastructure development. Just in the same way some cities quickly grow due to the finalisation of a nearby port, road network, train station or canal. 

Therefore what sort of infrastructure needs to be built, which countries are further ahead in their infrastructure journeys and what needs to be done? To answer this, one must first start with some stated principles.

Principles of African Fintech;

Hopefully these will be called the “Samora Fintech Declaration” ten years from now. To arrive at a goal, first you must state it, otherwise you may end up doing something completely different. In my view, financial services in Africa should drive the following principles;

1. Enabling easy inter continental money movement;

2. Drive financial inclusion;

3. Enable credit to get to whoever deserves it;

4. Drive trade and commerce via easy payments;

5. Africans should be able to transact anywhere;

6. Lastly, just like Visa does, infrastructure should enable transactional identities so as to allow customisation and personalisation;

That list is not exhaustive but it does a decent job of highlighting the key areas that need to be targeted.

The Underlying Stack

Different elements of infrastructure need to be built and each piece can either be built by the private sector or by the government. The core consideration is whether the element is non-exclusive or non-rivalrous with the latter being the definition of a pure public good. Public goods are thus best built by governments. 

Identity

Underpinning any financial service is the ability to prove an identity and link that identity to a payment request or store of value. There are many approaches to identity but ultimately, the most reliable is a sovereign identity issued by a national government. ID systems in Africa lie within a spectrum with countries such as Kenya and South Africa having robust ID systems whilst countries such as Nigeria have multiple ID systems. The gold standard is an Estonia type digital Identity that is built around open data sharing through permissioned access. 

The Aadhar system is the most ambitious global ID system covering over 1 billion people and enabling full identification and life-cycle management through biometric verification. Bank based systems such as Bank-ID in Sweden only work if there are high levels of financial inclusion. Nigeria has the Bank Verification Number (BVN) that enables unique identification of everyone in the banking system independently of your bank issued account number. It has enabled fintechs to do onboarding, but it's contingent on having a bank account and thus may not be useful for driving inclusion. The Nigerian government is introducing a new identity system based on having a National Identity Number (NIN) issued by the National Identity Management Commission. This should enable a single ID system thus improving identification.

Kenya and South Africa have already embarked on modern ID systems with the former issuing a new system branded “Huduma Namba” that is more service oriented. South Africa has already issued over 13 million ID cards with advanced security features. 

Identity should be a public good particularly in a digital world where ID usage should be non-rivalrous i.e. more than one entity should be able to verify my identity at the same time. In countries such as DRC, Burundi and Central African Republic, ID systems are often manual, decentralised and paper based. These systems hamper the delivery of banking services even within the formal banking system let alone digital apps.

Payments

Once you fix identification, the next thing should be instant payments and settlements to any store of value account or wallet. The gold standard currently is the United Payments Interface (UPI) of India which in July processed over 3.2 billion transactions worth around US$ 82 billion in July. UPI is part of the India Stack, a set of core infrastructures that cover payments, identity, KYC and document management. A well working payments system essentially enables real time accounts settlement instructions at high volume and high availability. In a recent article on Visa, I mentioned how Visa processes over 1,700 transactions per second with high availability. UPI on the other hand has an uptime of 99.9% in July 2021. 

In Europe, the Single Euro Payments Area (SEPA) enables real time credit or debit payments across the Euro area. In the United Kingdom, Faster Payments is the real time payments option. The EU and UK also have a legacy of card based payments.

The payments landscape in Africa is quite varied. South Africa has a card led model where payments volumes were anticipated to hit just below US$ 90 billion in 2021 with over 50 million cards issued. In Kenya, M-Pesa is the biggest retail payment system with over 25 million active users. Nigeria has a bank-led model and has built an instant payment system that handled approximately US$ 7 billion dollars in value in 2020. Ghana has a hybrid system with an advanced instant payments system handled by Ghanaian Interbank Payments and Settlements Systems Ltd (GhIPSS) as well as a robust mobile money market.

Countries that have a mobile money led system have managed to drive financial inclusion further than bank led systems. Nonetheless, given that a robust instant payments system hasn’t developed particularly in places like Kenya and Rwanda, financial innovation has somewhat been stifled. At the same time, a dilemma has occurred. Post-covid, both the Rwandese and Kenyan Central Banks have implemented price controls on telcos and banks by capping or removing fees on specific transactions. In my view this has occurred because the governments have failed to see the strategic value of payments as a public good and thus rode on private infrastructure for too long. Second order concerns around whether these governments will go further down the route of price controls in other vectors could lead to investors developing cold feet. 

The New Times (Rwanda) @NewTimesRwanda

No charges and fees will henceforth be incurred when transferring funds from one’s bank account to their mobile money wallet or vice versa, the central bank has announced in a new directive.

buff.lyCentral Bank scraps ‘push and pull’ charges

Despite this, a number of African countries, Kenya included, have plans to launch retail instant payment services as well as open banking regulations. By 2025, most countries in Africa should have instant payments systems.

Cash-in Cash Out (CICO) Infrastructure

Over 95% of payments in Africa are still cash based and financial services in Africa need on and off ramps from cash to digital money. Mobile money players such as M-Pesa have done an amazing job in this regard. M-Pesa has over 250,000 agents spread across Kenya. In Nigeria, a shared agent network system has been promoted by the Central Bank meaning that the agents system in Nigeria is more interoperable than the Kenyan one. South Africa is riding on bank infrastructure such as ATMs and branches whilst Ghana has a robust mobile money infrastructure.

I always say that M-Pesa is unique from a global perspective, one of the key ingredients for its success in my view was an already dominant telco business in terms of market share. M-Pesa could then build its agent network almost exclusively. A lot goes into recruiting and growing agents and one of the key areas that doesn’t get attention is agent training on things like AML, KYC and other compliance issues. It doesn’t make sense to train and nurture agents who will most likely work for your competition. In markets where agents transact for multiple partners, I’ve noted that they are less rigorous around compliance and transaction processing.

Of course on the other hand, the issue with mobile money players is their lack of interoperability. Although most will argue that this is not their concern and rightly so. One of the core jobs to be done to build interoperable on and off ramps is to build an instant settlements system and an open banking infrastructure. Nonetheless this in itself is not sufficient, thereafter a “Chaordic” organisation as the one Dee Hock envisioned for Visa is required. In this case, an organisation with shared ownership by the participants that enables not only agency banking but open branch infrastructure that enables high value transactions to be done at any branch. 

This will be difficult to implement given that most banks are still not keen on sharing their infrastructure. Additionally, the political ramifications of making so many branch staff redundant may be unpalatable. Nonetheless, open CICO infrastructure is a core ingredient. We’ve seen countless times from the article on SWIFT to the article on VISA that the only way to build robust financial infrastructure is through collaboration and member owned organisations. Banks in Africa just don’t like to collaborate.

Banking and Payments as a Service

We’ve spoken in detail before about the separation of distribution from manufacturing in banking where banks do the manufacturing including KYC, deposit protection and payments (licensed functions) whilst Fintechs do the distribution by building out the intuitive user journeys for customers. This in the long-term enables niche financial products that are tailored for user groups such as farmers, university students, doctors etc. 

In the UK, USA and Europe, Banking as a Service has emerged as a business model where Fintechs partner with a bank to offer financial services to their customers. This is an API driven business model where the bank benefits from deposits as well as payments commissions. There are three approaches to BaaS;

1. Pure API plays - Examples are Railsbank and Bankable where these companies focus on building out BaaS APIs that connect both the Fintech and the partner bank;

2. Hybrid BaaS - An example being SolarisBank and Cross River where the bank offers both the license and the technology. Recently, I did an article on Vodeno and Aion who are offering a comprehensive pan-European BaaS product.

3. Pure License plays - Where the bank just offers its license and has to partner with a technology provider - Examples include Evolve Bank and Trust and Lincoln Financial.

In Nigeria, players such as Wema Bank and Providus have enabled Fintechs to launch savings and wallet products whilst offering the license and virtual accounts. In Kenya, Equity Bank has attempted to offer BaaS through their Jenga APIs. Nonetheless, I understand that developers complain about cost and reliability. Additionally, it doesn’t seem like there are store of value APIs and its mostly payments and data APIs for Equity account holders.

BaaS is a full time offering that requires significant commitment from the ownership of a bank and not a side project. The bank that builds out regional BaaS capabilities will do well over the next 10 years. Key APIs will include regional and international payments to power trade. 

Credit Scoring Infrastructure

Africa has some of the least developed credit scoring mechanisms in the world. South Africa is unique in this regard with a long history of FICO scores provided by players such as TransUnion. This has created a robust market for consumer credit with players such as WoolWorths and Mr. Price offering their own credit cards to their customers. 

Kenya has a credit reference system but this is still based on negative information and not a FICO type rich positive information system. Nigeria has slowly evolved from a simple negative report system towards a richer scoring system. However, with the level of bank based financial inclusion being so low, the value of this system is tenuous. 

Brazil shifted towards a positive credit scoring system in 2011 and the top 5 banks even banded together to create Quod, a rich data company that provides credit scores as well as other products such as compliance, anti-fraud and billing solutions. I posit that Nubank could only build a credit card business on the back of an existing credit scoring system. The same can be said of Tinkoff.

In the future, existing credit bureaus should expand their data sources to include telecoms data as well as e-commerce data given that these are now generating significant amounts of data that could power a credit score. Of course, regulations also have to adapt to enable these data sources to be legally acceptable. A fintech in the future should be able to receive rich data that can enable them to make a quick credit decision.

I’ve spoken before about asymmetric information in banking and credit information sharing infrastructure can help with this. Shara, a new fintech company, is building data sharing infrastructure that enables community driven finance. The key insight is that at a mass market retail level, suppliers and fellow retailers are the biggest source of financing and thus collecting data and enabling a sort of “credit passporting” will enable retailers to expand their sources of credit. Pezesha is also building out credit scoring infrastructure. The question then becomes one of whether credit data should be non-exclusive and non-rivalrous. If so, regulated credit scoring should be the way to go.

Open Data

Open Banking or Open Data approaches are also required to power the next generation of consumer fintechs. Two approaches can be taken here. A market led model where a player like Plaid does the hard work or integrating to each financial institution across the continent. Alternatively, a European like regulator driven model where the regulator decrees that banks should share data through open APIs. Additionally, data should come from telcos, MFIs and Saccos to achieve the principle of open banking. Read here for a global perspective on open banking.

Regulations on open banking are currently in various stages from public information gathering to parliament across markets such as Kenya, Brazil, South Africa and Nigeria. I have a strong sense that by 2023 most of these markets will have open banking regulations with robust players in the market. In Africa, the main companies in the open banking space are Mono, PngMe and Okra.

Summing it all up

The diagram below summarises the whole discussion across 5 countries.

It is evident that India is leading the pack and potentially that’s why the Fintech ecosystem in India is so much more advanced than the rest of the world. Nigeria seems to be doing well in Africa and that’s why it is currently producing the most excitement in the African fintech space with start-ups such as Kuda Bank, Cowrywise and Piggyvest all riding on the above infrastructure.

Kenya is not that advanced and I feel that this is a legacy of Mobile Money where the financial system has moved to M-Pesa. In principle, customers in Kenya are included, can make payments easier than most other countries and can access credit through products such as M-Shwari, Fuliza and digital lenders that disburse through M-Pesa. Does the country need more innovation or is M-Pesa sufficient? 

In addition to all these factors, one necessary condition is proper and stable regulation. At the end of the day, doing business requires good infrastructure and stable regulation to thrive. Digital infrastructure can be built and many African countries are on their way to achieving this. Stable regulation is the harder bit as we’ve seen with the CBN of Nigeria. However, over the next 5 years, I think that the underlying infrastructure will be much more advanced and Fintechs will thrive thereafter.

As always thanks for reading and drop the comments below and let’s drive this conversation.

If you want a more detailed conversation on the above, kindly get in touch on samora.kariuki@frontierfintech.io