Primer: Key drivers of success for digital lenders in emerging markets
Access to capital in our markets is a fundamental issue faced by private businesses and individuals. This issue is characterized by large problem sets that need to be solved at scale to reach the promised land of sticky and scalable revenues, positive unit economics, defensibility, durability, and operating profitability. Whilst all the above presents a significant opportunity, there are some prerequisites for startup success that must be understood (startups who fail to understand these precepts are setting themselves up to fail). This blog post will outline these tenets, and we will elaborate on this topic in future posts.
Thanks for reading Terra Incognita! Subscribe for free to receive new posts and support my work.
There are problems, and then there are money problems
As the title suggests, “money problems” are complex, differentiated, situational and sensitive to economic cycles. What makes these problems complex are three core root causes listed below:
1) Businesses and consumers in emerging markets come in many shapes and sizes.
This is a fundamental concept in lending that is unforgiving when abandoned to chance and is characterized by heterogeneous consumer and business behaviour not just from country to country but city to city and even neighbourhood to neighbourhood.
2) In the face of inflation, Central Banks will always raise rates first and ask questions later.
Emerging markets tend to suffer from supply-chain shortage-driven inflation versus demand-driven inflation because of administrative and capital controls, which in turn are driven by Current Account Deficits that have become harder to finance. Depreciation and supply-chain shortages are hyper-inflationary and inevitably lead to contractionary monetary policy, i.e., raising interest rates. The effects of a contractionary monetary policy are a double-edged sword for lending companies, where higher profitability goes hand in hand with higher impairment risk.
3) Effective lending requires an ever-increasing proficiency in collections.
Businesses and individuals in emerging markets have been borrowing informally for centuries, if not millennia, but through a complex social network of guarantors, credibility reports and penalty systems. Democratizing lending builds a greater level of social anonymity within the system, which is further reinforced through an impersonal relationship with the lender. This disassociation and anonymity often give rise to moral hazard, which can severely affect your collection performance.
The three commandments of digital lending
Having looked at the underlying root causes of the problem set; it is time we look at some simple ways to make sure you are hedging against them. Together they form the three commandments of digital lending in our markets and have been collected through our experience investing in digital lending companies (Zood, Abhi, Trukkr, Billz, Finja and Oasis) as well as companies that we have not invested in, such as Kaspi, Kredivo, Akulaku, Mercado Libre, LendingKart, PayFazz and more.
The first commandment of digital lending is to build scalable distribution channels
In our experience, this is a product of the borrower you are servicing, the data you have access to, and how optimized your service is for the vertical your borrower operates in. Some approaches we have seen work well are highlighted below:
Embedded distribution works well when there is a smartphone app, website, POS machine, or an ERP product a borrower already uses. This is a partnerships-based distribution strategy which allows one to generate maximum network effects and multiple clusters for data harvesting. It is a cost-efficient way to acquire customers but can be haphazard when it comes to conversion and engagement if your partner’s product is sub-par. If, however, you find a partner with a super-focused product serving a large borrower base in a specific vertical, then an API-led model combined with a seamless origination and collection process can do wonders. An important consideration here is the ticket size and tenor length of your loans, where most models have demonstrated that lower ticket size and shorter tenor-based loans, i.e., 3 months or less, tend to perform better.
Case-Study: Enabling credit at checkout to leap-frog credit cards in Indonesia. Akalaku and Kredivo have started disrupting the credit card industry in Indonesia, where 20% of all transactions are paid through credit cards on major e-commerce websites. Kredivo estimates BNPL now accounts for 3-4% of all transactions on its partners platforms but, more importantly, has provided access to credit to 10 million customers in less than 8 years. This is equivalent to the total number of unique credit card holders in Indonesia; comparatively, credit cards have been around for at least 20 years. Additional benefits of the model for BNPL providers are the ability to access purchase history, merchant quality and a charge-back mechanism using the marketplaces escrow and settlement systems through an API-led approach. When combined, they form an effective mechanism to finance e-commerce transactions effectively and at scale.
Sturgeon Portfolio Companies implementing embedded distribution as a model. Abhi and Finja have been able to utilize this model for Earned Wage Access and FMCG-based lending, respectively, with hundreds of thousands of loans successfully deployed. The approach remains similar in utilizing enormous amounts of data held by distributors and employers to score prospective borrowers accurately while having a valuable partner that can facilitate collections.
Closed-Loop distribution works well when you own the ecosystem you are lending to. This model comes with significant upfront customer acquisition costs and higher short-term non-performing loan provisions, especially when building in a high-competition environment. Once the portfolio and customer base matures, however, there are significant advantages around profiling the borrower and consequently minimizing non-performing loans long-term. A typical business model this works well for is a vertically controlled marketplace that can control all touchpoints the borrower encounters, such as purchase history, logistics, warehousing, and payments. Enhancements to the model are based on the ability of the business to consume and integrate other sources of data concerning the borrower in their origination strategy, which provides an advantage that ages like a fine wine.
Case-Study: Capturing the credit market through building a Super App for Kazakhstan. Kaspi has proven the merits of patiently building closed-loop models to deliver outsized profitability, engagement, and growth. For those unfamiliar with the company, Kaspi has 11 million monthly active users, against a population of 19 million and has become ubiquitous through layering almost any service a person in Kazakhstan requires on top of a bank account; claims of it being the operating system for Kazakhstan are not too far from the truth with almost any digital service available through Kaspi. Their outlier approach has enabled them to build a deposit-led balance sheet, and offer payments, e-commerce, and lending across many verticals. The combination of these verticals has enabled a truly engaging lending flywheel while maximizing customer utility driven by their “one-stop-shop” approach. This has culminated in them earning a multi-billion-dollar IPO on the London Stock Exchange.
Sturgeon Portfolio Companies implementing closed-loop distribution as a model. Zood is an example of how a vertically controlled marketplace can successfully enhance its model through BNPL-led lending programs to generate double-digit increases in average order values, sub-12-month customer acquisition cost payback periods, and one of the best gross margins to NPLs ratio in the industry. Today, Zood is the market leader for e-commerce and BNPL in Uzbekistan and has recently expanded to Pakistan to capture an extremely large and underpenetrated market. Trukkr is a logistics aggregation and financing platform for Pakistani truck drivers. Trukkr has built a marketplace model to match truck drivers with large corporate partners, and successful orders are financed to ensure both high job completion rates and a high APR factoring business. The closed-loop system allows Trukkr to have significant collections proficiency and incentivizes truck drivers to maintain a high willingness to pay to retain access to future contracts.
Being agile means embracing multiple paths often. One thing to remember as one looks at these models is that your customer profiling is what drives which path you dedicate more resources to, but there will always be times when you must pursue both.
The second commandment of digital lending is building effective risk controls in your portfolio supported by data science and quality tested vigorously.
Most lending companies that are good at distribution but do not build a risk management function that is data/technology-led over time fail horribly. The only way to be non-tech enabled and survive in lending is to do lower volumes and chase bigger ticket sizes, but as a strategy that builds in long-term value accretion, this is a non-starter. Therefore, as a digital lender, your IP is your risk model and how that model generates a coordination mechanism that is a category leader in profitability and online volume once the company reaches a sustainable growth state.
The use of Big Data, AI and effective data harvesting integrations should be a priority for digital lenders, and failure to use this unfair advantage is a cardinal sin. Risk management is a day-zero activity and starts with building a team of professionals that understand risk from both a conventional and experimental sense.
I urge you to consider the following age-old advice whenever you want to bet the farm on an opportunity to be a digital lender: “Bulls make money, bears make money. Pigs get slaughtered.”
Every successful digital lending company in the world has played the long game, and those who have not to tend to be consigned to the history books as either crooks or fools.
The third commandment of digital lenders is building a supply of capital, and if you are a founding CEO reading this, make no mistake, this is your most important job. Period.
So, you have distribution and a world-class risk management system in place, but you are a lending company without a supply of ready-to-deploy capital. This is potentially the worse possible situation to be in as a Digital lender because what you have now created is a customer churn and repayments delinquency scenario. The former will lead to your strongest competitors gobbling up your best customers, and the latter will make your business insolvent. Delinquency risk is especially important, true, and prevalent because of the absence of widespread credit-reporting systems and the use of credit scores in other transaction types, such as renting for example, in emerging markets, that serve as deterrents in developed markets. This is understandable from the point of view of the customer when their returns on capital borrowed outweigh any past-due payment penalties imposed, and there are very few alternatives available to replace your company as a source of capital.
Furthermore, the cost of recovery may exceed the exposure at risk, and the availability of law enforcement resources to enforce small recoveries may not be adequate. Therefore, you should always feel compelled enough to arrange capital, and your supply planning needs clear long-term thinking around it.
Fear not, in emerging markets, what has worked for our portfolio companies and the greater universe is three fundamental processes:
Equity-led balance sheets in skeleton-style lending business models
Bank partnerships to generate credit lines and/or revenue share
Securitization if there is a large and sophisticated enough money market
Early-stage success is built upon raising your way to relevancy
Equity-led balance sheets can provide a quick go-to-market solution but serious dilution for the company even when using venture debt, which is backed by generous equity warrants. However, when there is no appetite for other models in sight, then you must default to this model, but with a strict focus on building in maximum dilution limits into the round and with unit-economics level profitability as a minimum requirement. Many businesses like LendingKart in India and PayFazz in Indonesia have raised efficient equity rounds to demonstrate proof of concept before engaging non-dilutive sources of funding meaningfully.
One point to note on how quickly you need to turn to non-dilutive financing is that equity-led models only work in markets where incumbents are tech-challenged and higher competition can be a death sentence. This is especially true in markets where incumbents can catch up quickly and you are engaged in a slow and low APR lending vertical.
Partnering for success is an incentive alignment exercise
Bank partnerships require immense levels of relationship building and maintenance where clear segmentation of markets and strategies is especially important. Beware the middle manager who feels you are in the way of their promotion and looks at you as an adversary. Strategic integrations with the bank are also hard to manage, with disastrous effects in terms of focus diversion and adhering to bank-led risk assessment criteria that are unreasonable/overly risk averse. Hiring the best and most networked-in talent possible from the banking world is a responsibility that you must take on early with a mandate built on diversification, clarity of goals, and executive-level sponsorship. Southeast Asia has taken the lead on bank-based financing lines, with Mizuho Bank providing a $100m credit line to Kredivo as part of their Series D.
We have also seen our portfolio companies raise mixed rounds earlier with strategic investors looking to provide a mix of equity and debt. We cannot encourage this enough because great partnerships are built on short-term and long-term incentives being aligned, and being a revenue centre for an established financial institution is never a bad place to be in.
Securitization is always the end-state
Securitization is an expensive process that requires one to prove that their debt product can deliver meaningful risk-adjusted returns to debt investors. It becomes a natural requirement to balance gearing ratios and adhere to optimal capital structure requirements. It also requires third-party verification and credit rating issuances and requires the business to be cash-flow positive.
Why local debt is always better
In our experience, companies that can leverage local debt markets and utilize localized instruments are best placed to win. The main reason is rooted in depreciation risk and its consequential unit-economic impact on your loan portfolio, which is likely to be more sensitive because of its shorter tenor. Another notable reason is that as an early-stage business, hedging programs for debt raises under $100m are not worth the administrative costs and premiums paid. Therefore, raising local debt is the right move, but how it needs to be accomplished is a function of localization.
Abhi has proven that local investors are hungry for debt
Most emerging markets have large and liquid existing pools of local capital invested into incumbent sectors that are poorly managed and lack innovation. Pakistan’s financial sector, for example, is under-rated and overly defined by a shallow equity market. The debt market is overlooked, and money market funds have been quietly growing at double-digit growth rates every year to the point of having hundreds of millions of dollars of liquidity now available. Knowing this, the team at Abhi was able to execute an over-subscribed multi-million-dollar Sukuk raise using a local investment management firm in less than 3 months.
History certainly rhymes because not too far in the past, we saw Nubank build a balance sheet using money market funds to support their credit card portfolio in Brazil. General Atlantic famously passed on Nubank in multiple rounds because they thought a bank without a traditional balance sheet just could not succeed, a decision that must hurt to this day.
Mercado Libre has re-defined the bar for digital lending success in emerging markets
Look no further than Mercado Libre and associated businesses to understand what a closed-loop distribution model with a sophisticated securitization program can achieve. As investors, we look for positive signals and match patterns, and we wanted to leave you with the following passage from Mercado Libre’s Q4’22 Investor Letter offers, which offers an insight into what we look for when assessing a Digital Lending company.
“Mercado Credito as a whole continued to deliver strong results in Q4’22 with a period-end portfolio of $2.8bn and IMAL spread of 48%. This was the highest spread achieved in 2022, a function of adjustments to our APRs, broadly flat originations (which means the slower formation of new provisions than earlier in the year), a larger mix of lower-risk cohorts in all markets, and better asset quality. The steps that we took in mid-2022 to mitigate the risks of a weaker lending environment - particularly in Brazil - have worked as intended, and, as a result, our early <90-day NPL improved sequentially to 10% in Q4’22 and was broadly stable year-on-year. Brazil made a notable contribution to this improvement in both the Consumer and Credit Card books. Looking at the year, we are pleased to have been able to strike a good balance between risk management, profitability, and growth at Mercado Credito. Nevertheless, we remain alert to the short-term headwinds that the business faces and we will maintain a cautious posture until we are confident that the cycle has turned.”
Thank you for tuning in and lots more to say on the topic; we will be following up with blog posts on “the hunt for a balance sheet” and “opportunities in fintech lending in South and Central Asia” next.
Thanks for reading Terra Incognita! Subscribe for free to receive new posts and support my work.