Skip to content

Financial Inclusion and DeFi: A Comprehensive Analysis of DeFi’s Impact on Fintech Transformation

By Shreemoy Mishra, Senior Researcher. Coauthor: Alamira Jouman Hajjar, Senior Research & Editorial Manager.

Financial Inclusion and DeFi: A Comprehensive Analysis of DeFi’s Impact on Fintech Transformation

This article is a comprehensive analysis of the top use cases of DeFi focusing on payments, borrowing, savings, and insurance. Drawing insights from “FinTech for Billions,” the analysis contrasts traditional Fintech approaches with innovative crypto and blockchain solutions and extracts lessons from the fintech industry which can be applied with novel technologies to help create a more inclusive financial ecosystem.

Financial inclusion is an effort to make everyday financial services available to more of the world’s population at a reasonable cost. – Investopedia

Economic Models and The Impact on Financial Inclusion

As in any industry, financial service providers have a structure to accrue different types of fixed and variable fees on the services they provide. For example, US banks have a fee structure that ensures banks earn at least 4% annual returns on deposits to cover their financial costs.

To further explain this economic model, several US banks require a minimum account balance of $1500. Falling below that threshold, account owners will be charged a fee of ~$5 per month. The figures are somewhat different for banks in Argentina, Turkey, or India, but the economic model is similar.

A $5 monthly fee is hardly an impediment to financial inclusion in high-income economies. However, such fees and minimum balances can be a barrier for those with low incomes, particularly in developing economies with lower financial inclusion rates. For example, there are millions of accounts at Indian banks with near zero balances. They exist because of directives coming from government officials implementing economic inclusion drives. The account owners are counted in official measures as “financially included”. But the banks don’t really want them as customers, such accounts are not profitable. Another example would be Latin America, where ~65% of the population is unbanked or underbanked with no or limited access to financial services such as credit cards, loans, or global payment options.

In recent years, there has been massive growth and adoption of cheap, instant payments through mobile phones. But financial inclusion is more than bank accounts and mobile payments – it must include access to simple credit and insurance products. Many of these remain too complex and inaccessible to those at the bottom of the pyramid.

How can we get closer to expanding access to the masses? That is the quest of Bhagwan Chowdhry and Anas Ahmed in their book “FinTech for Billions”. The context is set in India, but the issues and market frictions are relevant for all developing economies. The book has nothing to do with cryptocurrencies. However, there are lessons for creating everyday financial services using blockchains. In this article we present some takeaways and adjacent ideas.

Categories of Financial Services

There are four main categories of everyday financial services: Payments, Lending & Borrowing, Saving and Insurance. In this chapter, we dive into each use case and how blockchain-based solutions can impact its accessibility by financially excluded populations.


Digital payments have taken off worldwide. The first phase of growth accompanied the adoption of mobile phones, rise of social networks, online retail, and the sharing economy. Mobile phone numbers became a unique identifier for most services. This expansion was further accelerated by the global pandemic starting 2020.

The use case of paying for everyday expenses using cryptocurrency has not seen much traction. For one, the relative ease and adoption of mobile payments set the bar quite high.

The crypto payments use case has also been held back by scaling limitations (cost and speed). This is where scaling solutions such as the lightning network and other layer 2 solutions such as optimistic or zero knowledge rollups tried to make some gains – initially. The rollup projects have since shifted attention away from payments to scaling decentralized apps such as trading using Uniswap.

While high-frequency crypto-payments have not materialized, the case for low-frequency crypto payments remains strong. The use of stablecoins for payments is attractive in hyperinflationaly economies where the motivation is as much inflation protection as is the ease of payments. These types of payments remain at low frequency because cost and speed can still be limiting factors.

Another low frequency, but high-impact use case is the cross-border setting i.e. remittances to support family members from abroad. This area continues to attract investments to build stablecoin-based remittance corridors to Latin America, Africa and Asia.

There are some hybrid options offered by so-called neo banks such as LemonCashBuenbit, Uala and Belo. Their users gain inflation protection from stablecoins and also engage in high-frequency payments using the regular banking system through these apps. Some of these applications have tie ups with card processors such as MasterCard and use their network. The neo banks automatically sell the minimal amount of a user’s crypto holdings needed to make the payment in local currency. This happens in a just in time manner at the moment of payment.

Learn more about the payments landscape in this in-depth report.

Lending & Borrowing

As mentioned at the start, the economics of financial services can sometimes make it harder to serve marginalized users. For instance, there are fixed costs to underwriting a loan such as to verify a borrower’s income, conduct risk analysis etc. Due to such costs, smaller loans must offer a higher rate of return for a lender to break even. For example, if the fixed cost is $100, then a $500 loan must return 20% over the life of the loan ($600 including interest). By contrast, a loan of $5000 would only need to return 2% over its lifetime for the lender to break even. Lenders may decide not to offer small unsecured loans at all or they may charge high interest rates. Those with lower income and assets may not qualify for cheaper loans and remain underserved.

DeFi lending is typically secured (collateralized) and the process is typically automated using smart contracts. In this case, the fixed costs can be dropped to zero – though, in practice, protocol fees are never zero. But many users who are marginalized in traditional finance may not have adequate crypto-collateral to put up for DeFi loans. This can change only if they receive some income in the form of crypto-currencies.

Loan size, duration and interest rate are some obvious key considerations when thinking about financial inclusion. But for many people, the most crucial factor may be the timing – i.e. the ability to borrow when a need arises without warning. Although lower-income individuals can plan for anticipated expenses such as weddings and school tuition payments, it is the unanticipated hospitalizations, funerals, or sudden losses of income where the inability to borrow magnifies the misery.

The likelihood of facing unpredictable expenses creates value for credit cards and overdraft facilities on bank accounts. The underbanked lack formal lines of credit. These are situations where local moneylenders come through for those who have no other option.

Though much derided, not all moneylenders are predatory. They fill a gap. Through repeated interactions, long term relationships, and social networks, they are able to reduce the fixed cost of risk analysis and monitoring outstanding loans. Unsecured lending in DeFi is not practical in most cases. However, there may be value in working with moneylenders and converting them into intermediaries. Crypto startups with workers in the field could perhaps target collateralized stablecoin loans to moneylenders. The moneylenders can use these funds to expand their business and serve more customers.

Learn about the lending and borrowing landscape in this in-depth report.

Overdrafts and Automatic Credit Lines

Some banks extend overdraft facilities to their customers. This allows account holders to spend more than their balance – up to some specified overdraft limit. Overdraft limits depend on a customer’s profile and banks typically charge fees for exceeding limits. They were particularly useful in the past when people wrote physical checks and accidentally exceeded their balances. Their continued availability is still valuable to those using checks, and also debit card users who do not check their balances frequently.

An exchange-based crypto wallet can offer credit lines similar to overdraft based on users’ previous transaction history. These accounts are usually KYC-ed, and operate without self custody, so the wallet provider can perform risk analysis, monitoring and loan recovery.

Quantity Discounts, Store Credit and Installment Plans

In developing countries such as India, many households purchase items like shampoo or instant coffee in single-use packets. The per unit cost of these packets is much higher compared to buying them in large bottles and jars. Quantity discounts are a universal phenomenon. Even in the US, per unit costs are much lower at large discount stores such as Costco and Walmart. Lower budgets and a lack of access to credit lines prevent many people from taking advantage of buying things in bulk.

To quote Dr. Mishra: “Growing up in India, I recall picking up groceries from the local store. I wouldn’t take any money with me, just a small notebook. On each visit, the store owner made identical entries in our tiny notebook and his own huge ledger. This is just a primitive version of a distributed but centralized ledger. My dad would pay the store owner later – at which point the old entries would be scratched out – a sort of regenesis. Store credit not only reduced the need to pay up front, it also allowed buying things in bulk. Even for lower income households, these can become opportunities for friends and neighbors to split up large purchases and share the quantity discounts.”

A blockchain based approach can easily replicate the dual entry system of store credit. But what it cannot do by itself is reduce the monitoring and recovery costs for store owners. We would still need humans in the loop to make it work.

Installment payments are a common version of credit. Buy now pay later (BNPL) using equal periodic installments (EPI) are popular in many countries. There are DeFi versions of these too. In regions with rapid inflation, installment schemes are attractive even without the credit aspect. Suppose a buyer does not really need to purchase on credit, then what matters is the gap between the inflation expectations of the buyer and the seller.

For example, consider an item priced at 1000. Suppose a seller offers the item on an installment plan of 12 monthly payments of 100 each. Then, the seller expects inflation to be (roughly) around 20%. For this installment plan to be attractive to a buyer, either the buyer expects inflation to be higher, or the buyer is currently cash constrained (needs to purchase on credit), or a combination of both. Quoting Hajjar: “Turkey in 2023 witnessed a very fast and very high inflation surge, so installment schemes played a crucial role in helping communities sustain certain lifestyles. I bought a plane ticket that was priced at 20,000 TL (at the time of purchase that was 830$), with a 5-months installment scheme, the final amount I ended up paying was 750$. Imagine the potential impact on families, aiding in covering expenses such as school tuitions, mortgages, or even acquiring new computers.”

As the authors of Fintech for Billions point out, there may be significant impact of using the same installment model to enable bulk purchases of consumables. This is another way of expanding economic inclusion.


For most families, saving has traditionally taken the forms of bank deposits, home ownership, and precious metals such as gold. Most of the long term saving was automated through employer supported pensions, or compulsory contributions to retirement accounts, social security, etc. Retail investing in financial markets is still a new thing in the history of financial services.

The unbanked typically have few assets, thus, expanding their access to bank accounts is an important first step to access saving solutions such as annual interest on their deposits. But even with the magic of compounding, the interest earned from small deposits will be too little to have any meaningful near term impact on savings. Instead, Chowdhry and Ahmed suggest diverting some of the interest income into premium payments for bundled insurance products. Such settings can help create default demand for basic insurance services – which the customers may not otherwise purchase at all.

Saving in crypto and DeFi has a mixed record with high profile failures such as TerraUSD and Celsius. There have also been successes like Compound and liquidity provider (LP) fees on Decentralized Exchanges such as Uniswap. However, earnings from protocol fees and yield farming are for sophisticated users, not for the financially marginalized.

For those who are underbanked, the initial benefits from onboarding them to DeFi may lie in inflation protection through dollar denominated stablecoins and payments. Of course, there are several stablecoins they can choose from. And there are micro and macro forces at play for both the issuers and the users.

Inflation-plagued economies sometimes also have capital controls on foreign exchange leading to a parallel, underground market in foreign currencies such as US dollars. A case in point is Argentina where such controls introduce a wedge between the official exchange rate and the market rate on the street (called the “Blue Dollar”). That gap is so wide (currently around three times the official rate), and the fact so widely known, that even foreign visitors use local P2P exchanges to convert their fiat money, crypto or stablecoins to pesos.

Another interesting way to combat hyperinflation is changing the “units of counting”; a colleague in Argentina mentions that in their residential complex people have switched to quoting prices for services in liters of petrol instead. Inflation protection offered by stablecoins can be very attractive in such economies.

Learn about saving and inflation-protection solutions in this in-depth report.

Stablecoins and access to banking

Bloomberg columnist Matt Levine notes an interesting relationship between the market structure of stablecoins and interest rates in the US. A stablecoin is a type of cryptocurrency that is pegged to the US dollar (or some other fiat currency).

The two largest issuers of stablecoins are Tether and Circle. Circle issues USDC. It is registered in the US and its finances are easy to audit. Its larger rival, Tether, issues USDT. Tether is not registered in the US and its financial structure is not as transparent. Holding stablecoins does not offer any income. Traders use stablecoins to get in and out of crypto investments, just like parking undeployed capital in money market funds. In recent months, as US interest rates have gone up, Circle has been losing market share to Tether. Levine speculates that most people who hold USDC also have access to US financial markets, while most holders of USDT do not. So even if Tether does nothing, it gains market share as Circle’s users liquidate their USDC holdings to sock them away at Citibank, Discover, etc. This may explain why Coinbase recently increased its rewards offering to over 5.2% APY on USDC (up from 2% a month ago). It is crucial to know that Coinbase and Circle are partners, and thus have a joint interest in promoting USDC.

There are several stablecoins that users can choose from on various networks. If the adopters of a particular stablecoin are truly underbanked and economically marginalized, then their use of that stablecoin should be consistent over time and largely uncorrelated with bank interest rates. On the other hand, if users do have access to the formal banking system, then there can be a lot of volatility in the demand for that stablecoin. Higher volatility can show up as larger deviations from the peg. This can be particularly impactful for stablecoins that are crypto-collateralized since their design is economically inefficient to begin with.


Two well known information problems arise in credit and insurance markets. One is adverse selection which refers to the inability to distinguish good risks from bad ones. If everyone is offered a loan or insurance on equal terms, then that will attract a disproportionate amount of riskier applicants. The other is moral hazard, which refers to the possibility that a customer’s own private behavior can influence their riskiness for a bank or insurer.

In traditional finance, banks and insurers respond to information problems by restricting coverage levels or terms of loans. Insurers try to do this via deductibles, which is the portion of a loss a claimant must pay by themselves before insurance kicks in. They match higher deductibles with lower premiums. Customers self select into various levels of coverage based on their own perceptions of exposure. Lenders try to attract responsible borrowers by requiring larger down payments for cars or homes.

It is very common to structure loans and insurance products specifically for individuals based on their historical behavior such as credit ratings, driving records, employment history, even demographics. Such historical records are always tied to unique identities. This is hard to do in blockchains because a person with an unfavorable record can always create a new identity (address) at no cost. Another obvious problem is the lack of monitoring. In traditional insurance, the insurer can verify the ground truth – that a loss actually happened. Such monitoring is impossible to do on-chain, since such data about loss is necessarily off chain.

From a DeFi perspective, of the four basic financial services, insurance is the hardest to operationalize. All forms of insurance applications in blockchains rely on trusted third parties for monitoring and submitting off-chain data – which greatly reduces the value in using blockchains for such purposes. Even without the blockchain aspect, previous attempts at peer-to-peer “insurtech” have all fizzled out.

Thus, advancing the state of inclusion for formal insurance at the individual level is challenging. In their book, Chowdhary and Ahmed state that formal insurance can work on a larger, group scale. The group could be churches, temples, cooperatives, local banks or even local governments. The handling of premium collection and claims distributions for losses within the group remains informal. The authors say “Markets work well on large scales and from a distance. Informal mechanisms work better at smaller scales and locally”

In the DeFi setting, the natural way to experiment with group insurance schemes is by setting up Decentralized Autonomous Organizations (DAOs). Some elements could be entirely on-chain, such as premium collections and claim awards within the group. However, such DAO’s will still require the reporting of off-chain data for claim validation so DAO membership would need to be controlled. More importantly, the DAO may need to be organized as some type of real-world, legal entity with designated officers and administrators, otherwise it would not be able to enter formal insurance contracts on behalf of its members.

Liquidity constraints can limit ability to pay premiums. There is an aversion to making regular upfront payments. Biases make it hard to quantify or evaluate the benefits from an insurance policy. The likelihood of losses or adverse situations are underestimated. The benefits of (even large) payouts in rare situations is also underestimated. Any innovation that helps avoid such underestimation, or make it easier to make regular payments, can lead to significant uptick in adoption of formal insurance. One suggestion from the book is to bundle insurance on a per-transaction basis – just like bundling travel insurance with airline tickets.

Avoiding Mistakes and Redressal

In their book on financial inclusion, Chowdhry and Ahmed stress on the need to protect users in two ways: the first in helping prevent errors, and the second the availability of redressal options to help fix mistakes.

For example, new users of ATM cards would sometimes enter an additional 0, such as entering 10,000 when they meant to enter 1,000. For those who do not have sufficient balance in their accounts, the attempt would fail. This can scare some users into believing they have less money than they thought. But users can also get shocked when the withdrawal succeeds and they get much more cash than they wanted.

Not all ATMs accept cash deposits, and even if they do, new users may not know what to do.

Better user interfaces with voice prompts and confirmation screens can help reduce errors. Redressal mechanisms, even when available, can be challenging for new users – many of whom feel intimidated by the thought of having to deal with bank officials.

People familiar with blockchain and crypto wallets are keenly aware that the possibility of errors is very high in crypto. The consequences can be severe and redressal for incorrect transactions may be literally impossible. We are reminded to manage our seed phrases and private keys (i.e. passwords and login information) carefully. We are warned to be careful when sending coins to contracts, to use the approve and transferfrom pattern – otherwise the coins could get trapped and unspendable. We are reminded to verify chainIDs and address checksums when interacting with applications or assets that exist on multiple blockchain networks. Traders are advised to use small amounts or take advantage of slippage controls. We know there are scammers in the ecosystem.

Some guardrails are available, such as using multi-sigs wallets and “safes”. There are new options for wallet seed recovery. Modern crypto wallets try to protect users from common mistakes, but it is still a very steep learning curve. Redressal in crypto may remain extremely challenging. A high degree of redressal would require trusted intermediaries – which goes against some of the ethos of decentralized money and finance – however such tradeoffs can help improve financial inclusion. Suppose we make these trade-offs and redressal becomes feasible with some trusted intermediary (human in loop). A crucial bit of advice from the book is that our systems must be designed with real incentives for the intermediaries to actually act.


“Having a checking account is a privilege, not a right ..” – New York Times (November 2023)

This story from the NYT is not about the financially marginalized! Most readers of this article are very likely to have access to advanced financial services. We take our accounts and services for granted. It is easy to forget that this state of financial inclusion may not be permanent.

The story highlights cases where american banks close some accounts without warning. Some of these sudden closures are triggered due to suspicious activities related to money laundering. But a lot of innocent people have their accounts closed without warning – they discover something is wrong when their debit or credit cards stop working suddenly. While such cases are rare, there are other ways in which access to services, borrowing in particular, can be reduced. For example, we may have excellent credit in our home countries. But when we move to a new country, for work or education, then our credit records are not transferable, and we need to start over. Losing a job can also lead to a reduction in the ability to borrow. In some regions, adverse changes in a consumer’s credit ratings can lead to denial of insurance coverage for cars or homes.

Public blockchains are unaware of geographical boundaries. Credit records tied to blockchain addresses (decentralized identities) can be used globally. Since users can interact with smart contracts directly, it is also harder to exclude people from using DeFi services through geofencing. However, as with traditional finance, continued access to DeFi based services can change over time and should not be taken for granted. We have to work towards not only expanding access but also preserving it.

Acknowledgement: I am grateful to Patricio Gallardo and Javier Álvarez Cid-Fuentes for suggestions on previous drafts.