Financial Services Background
The origins of finance are as old as civilization itself, with the earliest historical evidence dating back to 3000 BC. Banking originated in the Babylonian empire where temples & palaces were used as safe places for the storage of valuables. Over the past 5,000 years there have been countless evolutions in how we pay, earn, save, spend, lend, borrow, invest & insure; yet we might be in the midst of the greatest concentrated period of innovation yet, due to the human & financial capital gravitating towards the broader FinTech industry.
There is no shortage of reasons to target financial services as an end-market for both entrepreneurs and investors with trillions of dollars in annual revenue and market capitalization, coupled with hundreds of billions of dollars in net income, few industries have the size & scale. Looking at some of the sub-sectors of financial services a handful of data points depicts the opportunity set as nearly every market is measured in trillions:
- Alternative Assets- $14.0T+
- Banking / Digital Banks- The market cap of global banks closed 1Q21 at ~$9.0T.
- Insurance- Trillions of dollars of life / non-life premiums written per year with trillions of dollars of market cap.
- Payments- Payments industry revenue is expected to grow from $2.0T to $3.0T over the next 5-years
- Real Estate- Commercial RE is a $17.0T market in the US alone.
- Trading- There are trillions of dollars of equities, fixed income, and derivatives traded per year.
The companies that continue to dominate these end-markets are largely incumbents founded 50+ years ago, and in some cases 100’s of years ago, built on legacy technology stacks that have grown through M&A. Mainframe computers, programs written in COBOL, and green screens are still common throughout these organizations
Due to the economies of scale across many of these end-markets (e.g., Banking, Exchanges, Insurance, Payments, etc…), coupled with regulation that often disproportionately benefits better capitalized incumbents, the financial service industry has been one of the most acquisitive over time.
Take one look at the consolidation that’s occurred from 1990-until today where 36 different banks have become four, giving rise to labels such as Systemically Important Financial Institution’s (SIFI) aka “Too Big To Fail.”
Given the state of the “technology providers,” coupled with the complex regulatory environment, it should come as no surprise that financial markets are riddled with inefficiencies. These inefficiencies were only exacerbated by Dodd Frank which was a double down on centralization and many of the problems that led to the financial crisis. We firmly believe that market participants and regulators should look to technology, not regulation, to solve problems; as it wasn’t regulation that failed us during the crisis, but the inability to understand market microstructure, counterparty risk and leverage that ultimately failed the system. As opposed to increased centralization, and therefore risks concentrated within and between these SIFI’s, decentralization is a mechanism to spread risk apart and return markets to their natural state.
While there has been progress on the technology front since the GFC, basic blocking & tackling still results in 8–11 figure mistakes. In 2013 when David Murdock took Dole Food private there was a discrepancy between shareholders covering 36,793,758 shares & claims from shareholders owning a total of 49,164,415 shares. In 2014 when Vista Equity Partners took Tibco private there was a $100.0mn error based on a wrong share count used by Tibco’s advisor. In a May 2018 tender offer, AbbVie initially told shareholders the clearing price was $105/share based on the preliminary share coun, but the final price was actually $103/share due to “a failure by Computershare to properly account for all validly tendered shares that had been received the prior evening by electronic mail.” The inability to properly account for “electronic mail,” in 2018 is causing investors hundreds of millions of dollars of losses.
It’s not just share counts that Wall Street continues to get wrong, data propagation, dissemination, and subsequent investment activity also results in monetary mistakes. In March of ’20 a simple S&P rebalance error cost Invesco $105M:
In March 2020, due to volatility in the equity markets, S&P Dow Jones Indices communicated the decision to delay, and ultimately to separate, the rebalancing dates for its indices and noted some indices would be rebalanced in April and others in June. The Company noted this delay but not the separation of rebalance dates and omitted rebalancing the Funds on April 24, 2020 when S&P rebalanced the Index. The Company discovered this omission and rebalanced the Funds on April 29, 2020. The Company will make a contribution to the Funds of approximately $105 million to compensate them for the performance difference that arose from market movements between April 24 and April 29.
The Robinhood / $GME saga in January had endless media coverage, unfortunately very little of it dealt with the true core issue; financial market infrastructure. Robinhood formed a new entity called “Robinhood Securities” in 2016 & spent the next 2 years building their own clearing system to self clear & migrate off APEX, which they publicly unveiled in October of ’18. US equities still settle T+2 this means that when a user buys / sells a stock the cash / securities aren’t settled for 2 business days. During that time the clearing firm (in this case RH) needs to fund that trade with DTC (as illustrated below). Robinhood’s “clearinghouse-mandated deposit requirements related to equities increased ten-fold. And that’s what led [them] to put temporary buying restrictions in place on a small number of securities that the clearinghouses had raised their deposit requirements on.”
When CEO Vlad Tanev testified in front of Congress we were able to see a detailed account of what happened:
- Their collateral obligation on 1/25 was $124M the same day the VIX spiked 62% (3rd largest daily gain since 1990).
- On 1/28 NSCC told them they had a deposit deficit of $3B. ($1.3B of VaR based)
- In addition to the VaR based requirements, they had an “excess capital premium charge” of $2.2B. The NSCC indicated RH owed $3.7B versus $696M already on deposit with NSCC, so the net amount due was approximately $3B
While some argue the T+2 settlement of public equities as a feature not a bug, as it enables funds to utilize capital more “efficiently,” it really just helps to build potential leverage and risk throughout the system. There’s no reason today we couldn’t have true T+0 DvP (delivery versus payment) in US public equities, the technology is here today and we’re seeing it throughout the crypto world.
The biggest value destructor of all is leverage, particularly hidden leverage. Due to the way many of these unilateral ISDA agreements work between investors & banks, banks are largely ignorant of the leverage their clients have away from them. Take the famous example of Long-Term Capital Management (“LTCM”). In ’98 LTCM borrowed $125B on a $4.8B capital base (26x) to build 60,000 trading positions with a gross notional value of $1.4T. US regulators had to orchestrate a bailout (from a consortium of banks) as there were concerns that their collapse could cause systemic damage. Raoul Pal has told this story on LTCM countless times, where a bunch of sales folks at different banks were all bragging about the size of their business with LTCM, and came to the sudden realization that if they were all remotely telling the truth there was a big problem, as each of the numbers far outpaced the size of business a $4.8B fund should have been able to put on.
In the 23 years since, we didn’t learn our lesson. Estimates of what Bill Hwang was able to do at Archegos are still coming out, with some speculating he was able to build up notional exposure of $100B on a <$10B capital base. Somehow this $10B fund, caused nearly $10B in losses across Wall Street. Hwang entered into swap agreements with numerous counterparties, these were almost exclusively bilateral ISDA agreements that don’t speak to each other, so none of the banks were aware of the other banks exposure.
Market participants may not always be economically rational, but they are always economically self-interested. If there was real time visibility into the underlying performance of CDO’s, CMBS, and RMBS, rest assured the big banks would not have maintained the level of exposure that they did on their balance sheet during the GFC, rating agencies would have provided more accurate ratings, and funds would not have trusted them blindly regardless. This would have curtailed the capacity for originators, regardless of their incentives. If CS knew Archegos position sizes, they would have stopped extending leverage, turns ago.
We’re constantly searching for solutions to improve market infrastructure, but this concept of decentralization has the ability to re-architect it completely. Decentralized Finance or “DeFi” consists of numerous protocols that are enabling technology to provide services across financial services, where you have multiple entities that need to confidentially share data & logic, model complex financial instruments, with a single source of truth & no central authority, in a permission-less and / or non-governed way.
Imagine the re-architecting of the lifecycle of a mortgage where you create a new instrument leveraging smart contracts from the loan origination system, throughout the securitization process and secondary trading. As opposed to opaque baskets where investors didn’t know what was in the securitization until it failed, we can allow smart contracts to inform owners of the CMBS or RMBS with up to the minute performance on mortgages and key borrower characteristics so a change in value can be measured on a payment-by-payment basis. This will enable the creation of dynamic portfolios where baskets can be updated if FICO scores drop below a certain threshold, or delinquent payments reach a certain date.
The Invesco, S&P index data issue at its core is a federated trust problem. What if you could create an ecosystem for secure & real-time data sharing? A unified protocol that allows for data distribution, entitlements control, subscription/licensing, error corrections, derivative data product creation and a host of other workflow management capabilities that have not yet been realized within existing centralized offerings.
What if there was a way for funds / banks to be able to cryptographically assert their positions in total didn’t exceed any regulatory threshold (and for regulators to look at this in real time)? Similarly what if the banks could ensure there wasn’t hidden leverage in the system to a single counterparty?
Evolution of FinTech
The re-architecture of financial market infrastructure is one of the biggest opportunities within FinTech. Given the market opportunity, it’s no surprise that FinTech is everywhere today, yet this wasn’t always the case. As recently as several years ago only a mid-single-digit percentage of VC funding was going to the FinTech sector; now every bank CEO, management team, & public / private market investor is forced to have a thesis.
In 2019 Bain’s Matt Harris claimed FinTech is the fourth technology platform writing, “Rather than constituting a segment on its own, fintech joins the internet, cloud and mobile as the fourth major platform technology.” In discussing the evolution of these platforms over time he writes:
The pattern is clear. The internet comes along and people rebuild old stuff on it. Then they build brand new stuff with connectivity as a key ingredient. We go from talking about internet companies to presuming that all companies use the internet as a component. Cloud software was the same thing — at first, companies just make their software accessible via the internet, but ultimately they rebuild the applications as web native from scratch, and thus change the game.
Next, mobile comes along (on top of connectivity and intelligence), and at first people do their old stuff on it. Then it too becomes a function in the stack, and they build new stuff with it, as an ingredient. Now these three platforms constitute a stack of functionality: persistently connected individuals and businesses with access to nearly infinite compute and storage, with highly intelligent software at the edges, defining the user experience. Connectivity with intelligence and ubiquity.
This stack (internet, cloud, mobile) has enabled us to do old school financial services in a digital way. While the first generation of fintech companies created billions of dollars of value, because of new enablers like Plaid, Cross River Bank, Finix and Wisetack, we’re now moving past that phase to one where fintech moves from being a business model unto itself, to being the fourth layer in the stack or the “fourth platform,” wherein financial functions like payments, lending and insurance join connectivity, intelligence and ubiquity as layers of the stack upon which new companies can be built.
During this relatively short period of time we’ve seen distinct phases of FinTech as it starts to expand into mainstream vernacular.
FinTech 1.0 companies have generated hundreds of billions of dollars in value by taking legacy processes & doing them in a natively digital way. As opposed to completing endless paperwork to open a bank account, firms like Chime, Monzo, Revolut, Square, and SoFi have allowed you to open an account in minutes, entirely via a mobile device. The same thing has happened with brokerage accounts, as companies like Robinhood & eToro not only digitized the process but also eliminated transaction fees, causing the incumbents to be fast followers. Companies like Lemonade have created a “21st century insurer” by offering a more modern UI / UX, with better marketing and a comprehendible suite of products across renters, homeowners, pet, and term life insurance. On the B2B side enablers such as Plaid have allowed companies to access data and create new products that previously weren’t feasible, while companies like APEX & DriveWealth gave anyone the ability to offer equity trading, and firms like Marqeta / Galileo gave anyone the ability to offer a debit or credit card.
FinTech 1.0 exits have been a long-time coming. PayPal was one of the OG FinTech companies with an IPO in Feb of 2002 at $800M, but it was only public for 5 months before EBAY announced the acquisition of it for $1.5B, and it took another 13 years before PayPal was a standalone independent public company again, when it was spun-off from EBAY with an initial valuation of $49B. Square IPO’d in November of ’15 at a $2.9B valuation (which was less than ½ its last private round valuation of $6B in Oct ‘14). Today PYPL & SQ boasts market caps of $300B / $98.5B respectively. Despite their public market success, and the increased private market activity, there was a vocal contingent of FinTech skeptics pointing to the lack of exits.
That’s finally starting to change over the past three-years, really accelerating over the last 12 months, with exits of FinTech 1.0 companies with IPO’s from Adyen ($8.3B), Affirm ($12B), Coinbase ($50B), Marqeta ($12B), and Robinhood (TBD but rumored to be $30-$40B), SPAC transactions with eToro ($10B), MoneyLion ($2.4B), Acorns ($2.2B), APEX Clearing ($4.7B), OpenDoor ($6.3B), and SoFi ($8.6B), and M&A with the likes of Divvy ($2.5B), Galileo ($1.2B), Credit Karma ($7.1B), Finicity ($935M), Kabbage ($850M).
We’re in the midst of FinTech 2.0 where “embedded finance” is all the rage, and financial functionality is becoming a native component of both the technology & business model stack for a wide array of businesses. Financial products are becoming more ubiquitous & accessible, while at the same time moving to the background of a product offering. There’s been a lot of ink spilled over the concept of Embedded Finance over the past 3 years, including this post from last summer, as well as Matt Harris, A16z, Google’s Nik Milanovic, Anthemis, FinTech Today, Ron Shevlin, and countless others.
On the B2C side we’ve seen a rampant rebundling trend where nearly all companies are starting to expand their initial product suite, and offer some combination of a hybrid checking / savings account, brokerage account, crypto trading, PFM tools, P2P payments etc…. (we’ve written about this here).
We think we’re still in the middle innings of what financial super-apps should look like with a larger focus on self-driving money / autonomous finance, coupled with the automation of the liability side of the personal balance sheet, coming to fruition over the next couple of years. Financial data will be portable and no longer serve as a moat for certain lenders.
Due to the work that some of these B2B companies like Plaid, Finicity, Galileo, Marqeta, Apex, DriveWealth, Alloy, Bond, Moov, Rize, Modern Treasury, Qolo, Nivelo, Treasury Prime, etc… have done we’re starting to see early signs of finance becoming part of the stack, where we’ve evolved beyond simply the digitization of legacy processes.
Technology companies & large consumer brands are slowly becoming on-ramps into financial services. Companies like Uber, Lyft, DoorDash, and Instacart have all rolled out fairly complex financial service products for their various independent contractors, for the benefit of both their customers & [not] employees. Apple launched the Apple Card, while Amazon & Shopify have their own integrated payment offerings.
While at first it seems like companies such as Divvy, Ramp, Brex, Clara, and Jeeves may all appear like yet another “corporate card,” they each have built slightly different budgeting / insights, and expense management software that goes hand-in-hand with the core corporate card offering taking aim not just as AmEx but the PAYX & Expensify’s of the world with a more tightly integrated offering. A16z’s recent investment in Jeeves tackles the pain point facing many companies in a post-pandemic world, where they have a distributed work-force across different countries, which requires paying for expenses across different regions, in multiple currencies, using different payment rails. Each country has their own bank transfer system (e.g., ACH vs. SPEI), issuance rules with BIN Sponsors, payment processors, bank accounts, all of which affect back-end reconciliation.
SaaS companies are embedding payments, we’re seeing integrated lending at the payroll level, or PoS, with integrated insurance at the point of purchase of an online good, or even a home. We are starting to interact with financial products without knowing it, or dealing with financial intermediaries. But what if financial services could become an even more embedded part of the technology stack?
FinTech 3.0 | DeFi- ReArchitecting
We’re still a long way before decentralized finance or “DeFi” becomes part of the institutional fabric of financial market infrastructure; and to get there a number of things need to happen. We’ll talk a bit about where we are today and where we think this can go over time.
What Is DeFi:
While there’s some debate about the history of DeFi, many credit the origin of the term to an August 2018 conversation between ETH developers & DeFi entrepreneurs from 0x, Dharma, and Set Protocol. While the term hadn’t yet been coined, MakerDao launched in Dec ’17 as an Ethereum protocol that enabled users to issue a dollar-pegged stablecoin collateralized against digital assets. This enabled anyone to take out a loan without relying on a centralized entity. In less than 3-years since the term was coined & under 4-years since we saw the first native DeFi launch we now have ~$67.5B “locked” in DeFi (which is down from the ATH of ~$90B last month).
DeFi as a concept is still evolving considerably and if you ask 100 different people, most will have different definitions of what it is today & where it can go. DeFi is not a single product or company, but is instead a set of protocols that are enabling technology to provide services across financial services, ranging from payments, lending & borrowing, trading, insurance, derivatives, and security permutations. These services are offered via decentralized applications (“Dapps”), these DeFi Dapps enable both builders & users to combine their services, offering various degrees of composability that’s just not feasible in a centralized financial ecosystem.
Both interest & usage have grown exponentially over the past 3 years & it seems as if the industry is at an inflection point looking to cross the chasm to mainstream adoption.
In terms of the value proposition there are several areas where DeFi should be able to provide the same if not better functionality than the legacy financial system with advantages across:
- Accessibility- The unbanked population is still estimated to be ~1.7B people globally. DeFi opens up financial services to anyone with an internet connection not just offering banking functionality, but removing the concept of “underbanked” altogether as users will no longer be dependent upon the same underwriting criteria with exclusionary practices they are exposed to today. In developing countries, the financial institutions are less robust than in America, and often face corruption/incompetence.
- Cost Reductions- DeFi has the ability to remove layers of intermediaries that charge multiple fees at high margins.
- Composability- Because the vast majority of DeFi is built on open-source code, developers are able to build on top of others’ applications. This accelerates innovation creating the concept of “money legos.”
- Programmability- DeFi allows for money to become programmable. What if in the next round of fiscal stimulus checks the government restricted what the money could be used for (e.g., essential goods & services not $GME or $AMC calls), and when it could be used (e.g., inside of 90 days). This is technically feasible today with modern card issuers but becomes more feature-rich with DeFi. We’ll be able to program what money can do for you, such as conditional payments, tax assessments, etc… as opposed to having to send specific instructions to a bank.
- Speed- While critics like to compare V/MA transactions per second vs. BTC / ETH they fail to account for the time it takes for money to actually move, settle & become usable in the CeFi world. Although speed varies based on what protocol / service is being used, most DeFi protocols / smart-contracts are self-executing & have true DVP settlement, which is much faster than CeFi counterparts.
- Transparency- By design DeFi is more transparent than CeFi allowing anyone to audit the history. This allows for insight on leverage / collateralization that’s not feasible in the CeFi world, and provides a layer of security & trust not applicable in much of the developing world. Governance is also decentralized which has optics of being more democratic (although ownership remains concentrated).
We’re going to compare & contrast the CeFi world with where DeFi is today & where it can go in the future across payments, banking, borrowing, lending trading, insurance and more.
Satoshi entitled the Bitcoin White Paper, “Bitcoin: A Peer-to-Peer Electronic Cash System.” The first lines of the abstract read, “A purely peer-to-peer version of electronic cash would allow online payments to be sent directly from one party to another without going through a financial institution.” While there’s been a lot of debate outside the scope of this post about the true intent of the white paper, and what Bitcoin’s vision is today re: payments (if any), the concept of cheaper digital payments has been something entrepreneurs have been targeting for what seems like forever.
Payment revenue is ~$2.0T globally with hundreds of trillions of dollars of payment volume conducted across the world, often times multiples of GDP. Take China for instance where mobile payments are +18x in 5 years, up from ~$2.0T in 2015 to an estimated ~$36T in 2020, nearly ~2.5x GDP.
We’ve written at length about the market opportunity for payments broadly (nearly all of which have applicable corollaries in the DeFi world).
The most common payment in the US today is the four-party payment system which has been largely unchanged for decades. In the four-party model you have: (i) Issuing Bank (ii) Network (iii) Merchant & (iv) Merchant Acquirer. In this model issuers (typically banks) look to put general purpose credit / debit cards in the hands of consumers in order to buy things. In the Marqeta S-1 they highlight this model:
“In 2019, consumers and businesses worldwide made over 440 billion purchase transactions on global network cards, aided by approximately 24 billion payment cards in circulation. Since the advent of card-based payments in the 1940s and 1950s, card payments have become the backbone of commercial activity due to their ease of use and widespread acceptance. A complex ecosystem underpins these transactions, consisting of Issuing Banks and Acquiring Banks, Acquirer Processors, Issuer Processors, and the Card Networks that facilitate the exchange of information and funds behind each transaction.”
While there are billions of transactions per year few understand the movement of money that occurs when you swipe a debit / credit card.
In this four-party model assume you spend $100 at a merchant with a merchant discount rate specified by the merchant acquirer of 1.99% + $0.25, which in this case would equate to $2.24. Therefore the merchant would receive $97.76 from the acquirer & the acquirer has $2.24 to distribute between the issuer and card network.
The biggest fee here is “interchange” which consists of various different rates depending upon whether or not the card was present, the type of goods or service being purchased, whether it was a debit or credit card, a corporate card, international transaction, etc…. In a basic debit card not present e-commerce transaction the issuer is entitled to 1.65% + $0.15; so on our $100 transaction that’s $1.80. In this scenario the issuer will debit the consumer’s account by $100 & pay $98.20 to Visa (netting the $1.80 interchange fee for themselves). There is now $0.44 to allocate between the network (e.g., V) and the acquirer.
Similar to interchange fees, V/MA earn what are called assessment fees, charged on a per transaction basis, that can range based on the type of transaction, risk, debit / credit, pin-Authorization etc…. In this example let’s assume its 0.11% + $0.0195 to V when the transaction is authorized & another $0.003 when it’s settled. This means Visa charges $0.13 in this example so there is $98.07 ($98.20-$0.13) less the merchants $97.76, resulting in $0.31 to pass on to the acquirer.
This is a highly simplified example (not bifurcating between front-end processors, back-end acquiring processor, issuer processor, etc…) but still shows the complexity of a simple payment in a four-party payment system.
Stablecoins have become one of the most fundamental building blocks of DeFi & serve as a critical component of the ecosystem to facilitate the balance of these financial service offerings without having to take incremental price risk by quoting borrowing / lending / trading etc… in BTC or ETH. Today two of the top 10 cryptocurrencies by market cap are stablecoins (Tether / USDC )
There are several types of stable coins including (i) Centralized Fiat-Collateralized (ii) Decentralized Crypto-Collateralized (some of which may be algorithmically determined), etc… The original “stablecoin” was Tether, a centralized fiat-collateralized stablecoin launched in 2015 and today (despite much controversy) boasts a market cap of $62B. Tether started off relatively slowly, with little activity in its first year. But in 2017, as a result of the move in BTC & ETH, Tether started to take off. Tether supply passed 1M for the first time in January 2016. By January 2017 it was a little under 10M. By January 2018, as Bitcoin’s price was peaking at close to $20K, the Tether supply had grown to over $1.4B.
Given the trust issues involved in any centralized stablecoin (exacerbated by a lack of transparency around Tether), a consortium that included Coinbase & Circle launched USDC in September of ’18, as an alternative fiat-collateralized stablecoin, which in under 3 years boasts at market cap of $22.6B and is completing ~$3.3B in on-chain volume per day YTD. Just like Tether initially positioned itself, this stablecoin is backed 1:1 by US dollars in a bank account.
It’s not just USDC that has seen explosive growth over the past 24 months. In addition to USDC, Paxos has its own centralized stablecoin which powers Binance USD ($9.7B market cap), PAX ($1.0B market cap), HUSD ($785M), etc… as does Gemini, TUSD, etc…
Decentralized crypto collateralized stablecoins such as DAI have lagged their centralized peers, but have still seen their own exponential growth over this time period. Maker is a system built on Ethereum that governs a decentralized stablecoin called DAI. DAI aims to be pegged to the USD dollar for dollar. DAI can be “minted” by anyone within the Maker system by locking up crypto as collateral (predominantly ETH) and taking out a loan of DAI. The collateral provided needs to be greater than the amount borrowed so that the loan is overcollateralized (e.g., lock up $200 of ETH to borrow $100 of Dai). We now have ~$100B in stablecoins outstanding.
The main use cases for stablecoins today are (i) Trading (ii) Inter-Exchange Settlement (iii) Cross-Border Payments and (iv) Remittance. If we look at the breakdown of account sizes the vast majority of accounts are smaller (<$100USD) which tends to give some credence to the individual cross-border payments / remittance category. The larger transactions are largely trading / inter-exchange settlement driven today.
A digital dollar doesn’t seem to be groundbreaking innovation, after all in the US, when you deal with PayPal, Venmo, Cash App, Zelle, etc… dollars for all intent & purposes are digital. Within these closed-loop ecosystems a stablecoin & their digital dollar will operate in a similar fashion (e.g., no transaction fees, near instantaneous settlement).
The biggest difference with a stablecoin is that this dollar can be sent anywhere in the world with the same result. This has increasing importance in a world with distributed work forces where employees can live & work from anywhere which can become a payments nightmare (e.g., recent payment IPO DLocal enables 600+ local payment methods cross 29 countries). Stablecoins are also self-custody (which has both pros / cons, but more cons from the lens of an institutional investor), transferable 24x7x365, open, interoperable and most importantly programmable through the use of smart contracts, leading to “programmable money” which has the potential to be the most valuable component of the stack.
Marqeta will be one of the largest payment IPO’s in recent years coming out at ~$12B+. Its biggest unlock as a business was the separation of the card issuer from the issuing bank. Marqeta’s “modern issuing” platform enables a new class of issuers (non-banks) to distribute card credentials (physical, virtual, tokenized) to users with customized/programmable features. This allowed Marqeta to enable a number of new payment flows on-demand service, including Just-In-Time (“JIT”), dynamic spend controls with provisioning of credentials for only approved users / merchants / amounts / frequencies / time, etc…
In no means are we trivializing what Marqeta has built but, through the use of smart contracts money can be programmable at the base layer. When you order food using DoorDash or groceries using Instacart, stablecoins / smart contracts can be working in the background as money moves from the app to the delivery driver’s payment card, allowing the driver to pay for exactly what you ordered, and nothing else.
We’ll still need a software layer that sits on top of these public blockchains for merchants, payment companies, digital wallets etc… to interact with. There will need to be Account API’s, Payment API’s, Payout API’s, etc… to manage flow of funds, fraud management, operations, interoperability with legacy rails, automated workflows, etc….
This type of programmable money has the potential to open countless if/then payment scenarios for C2C & C2B, and B2C payments. In prediction markets / gambling if an event occurs (that is validated by an oracle) a payment is automatically sent, the same thing can happen for parametric insurance, derivative contracts, and accounts payable / receivable contracts.
The team at Modern Treasury is innovating in the B2B payment space. In a recent blog post they highlight that B2B payments sum to ~$18.5 trillion, which is over 76% of payments in the United States. At that size, nearly half of B2B payments happen by paper check, and each check is estimated to cost $39 to process. The dominance of checks is forecast to dramatically drop in the next three years as paper checks are overtaken by electronic payments.
The most common form of “electronic payments” for B2B is ACH payments. Approximately 92% of value transferred over ACH is for B2B payments, which is growing at a rate of 6.9% / year. The ACH network is built on archaic infrastructure from the 1970s, and despite the ~$55T per year in transaction volume, it is riddled with fraud as nearly ~5% of all transactions are fraudulent.
The electronification of B2B payments + fraud reduction, is a prime opportunity for programmable money & is closely interconnected with the digitization of contracts more broadly. DocuSign was one of the clear winners of the COVID-19 pandemic as the stock is ~+200% since last March, with revenue growing at a 45% CAGR over the past 3-years, due to the digitization of paper contracts. In August of ’19 DocuSign brought Clause’s “Smart Clause” technology into their agreement cloud. DocuSign describes the partnership as follows:
Clause is the leading provider of Connected Contracting® technology, enabling users to add Smart Clause® provisions to turn their existing legal agreements into “living documents” that integrate with enterprise software systems and blockchain networks. Companies in industries as diverse as manufacturing, construction, supply chain, insurance, telecom, financial services, retail, and others can use Clause to automate compliance with legally binding contract obligations, reducing operating costs and minimizing revenue leakage. The Clause platform also provides real-time visibility about the contract performance of partners, suppliers, customers, and other counterparties.
Clause has published several case studies including a parametric insurance contract with Clyde & Co insuring solar energy producers against the risk of a shortfall in expected energy generation due to unfavorable weather in the form of lower than anticipated solar activity. They’ve also written about improving commercial supply agreements, accounts receivable insurance, milestone payments, smart fuel surcharges, supplier data, vendor compliance, & address verification.
As agreements become digitized & contracts transform into “living documents” integrating with a more connected ERP system, IoT devices, edge computing, etc… the ability to create programmable money tied not just to a legal contract itself, but all the requisite data involved in a contract, should significantly reduce fraud as well as reconciliation time & money. Programmable money allows you to truly “set and forget” payment terms, which is the ultimate manifestation of “self-driving money,” “autonomous finance,” and “embedded finance.”
In terms of where value will accrue with programmable money today people are looking at companies like Circle & Paxos who earn interest on the float on deposits. If stablecoins become part of the institutional fabric of financial market infrastructure we will see come combination of Central Bank Digital Currencies (CBDC) as well as interoperable bank coins (e.g., JPM coin) which will shift value accrual away from deposits, to value added technology services & API’s that allow developers to build functionality on top of programmable money to fit business use cases.
Banking / Borrowing & Lending:
CeFi Banking / Borrowing & Lending
How do banks make money? We wrote about that in the context of Neobanks here, but simplistically the two biggest drivers are (i) Net Interest Margin (“NIM”) & (ii) Fees.
NIM is simply the difference between the interest a bank earns on loans, compared to the amount it is paying on deposits. The below chart shows the average NIM for all U.S. Banks from 1984–3Q20 when the series was discontinued.
NIM has two main components to it:
- What banks earn on Loans (Mortgages, Personal Loans, Student Loans, etc…)
- Cost of Capital- What banks pay on deposits
Even though there are ~5,001 commercial banks & saving institutions and another 6,293 credit unions, network effects are prevalent with the Top 15 Banks in the U.S. controlling in excess of 50% of total deposits. In order to compete in a CeFi world they need to have the lowest cost of capital and largest balance sheet.
In order for a bank to lend to a potential customer there’s a lengthy onboarding process, a robust credit check, ongoing servicing of the loan, and an alphabet soup of regulators involved.
DeFi Banking / Borrowing & Lending
In DeFi it’s no surprise that two of the most popular initial use cases have been lending & borrowing. Prior to DeFi the concept of lending & borrowing has always required both some form of trust & a third-party intermediary. The system is maintained via a convoluted system of credit, whereby the borrower must exhibit the ability to repay the loan in order to be qualified to borrow, among a laundry list of other qualifications and requirements by the banks. While this existing system works well, it has its shortcomings which has led to the exclusion of large pockets of the population from the traditional financial system.
In DeFi these barriers don’t exist allowing anyone with collateral to be both a lender & a borrower, removing the lengthy onboarding process & at times exclusionary practices. In terms of lending, everyone can pool capital into a decentralized liquidity pool, of which borrowers can take from & pay back at an algorithmically determined interest rate. To borrow you only need to provide collateral to take a loan in DeFi, after which point in time you are free to do what you want with the proceeds of that loan (which today most everyone uses to speculate on crypto).
Looking at the Top 10 DeFi protocols based on total value locked in them, the top 3 are all lending protocols, with 5 of the top 10 falling into that category accounting for ~$38B.
Today borrowers have to put up significant collateral to take a loan with collateral ratios often well in excess of 100%. If we look at Aave or Compound for instance there is a liquidity pool where lenders deposit assets & receive interest, while borrowers take loans & pay interest. Each pool sets assets aside as reserves to hedge against volatility. These reserves also help ensure that lenders can withdraw their funds when they’re ready to exit the protocol.
Interest rates are determined by supply and demand for the asset, and lenders/borrowers can exit their loan and take their money out at any time. This way, lenders and borrowers can interact directly and immediately based on interest rates vs. working out loan terms (counterparty, collateral, maturity, rate).
These rates have changed over time, but have ranged from low-to-mid-single digit APY (e.g., 2.5–3.0%) to as high as 40–50%+. Even at the high end it is more competitive than lending rates in parts of LatAm where they can range from 30–70%+ APY at any given point in time. On the lending side some have been as low as inside 25bps (Maker) and as high as 40%. These rates are often times variable, changing during the duration of the loan period.
Aave has recently surpassed Compound in terms of TVL despite a similar construct (and what some would argue is a worse UI / UX). Aave offers a product called “flash loans” where users don’t need any collateral to access them. In lieu of collateral, there are time requirements; users must pay the loan back within the same transaction. Flash loans are a native DeFi creation that couldn’t exist in the legacy world. While the use cases today are limited, you can see the potential for them to be used as a way to refinance existing loans, for arbitrageurs to take advantage of mispricing's, or swap collateral. Aave also supports ~20 different collateral types versus 9 for Compound.
To earn interest there are native protocol tokens such as CTokens in Compound & aTokens in Aave, where interest accrues & can be redeemed on a 1:1 basis. They also have governance tokens; in Aave’s example LEND tokens (which also entitle them to fractional fees of interest paid when staked, as well as discounted fees), as well as Compound’s COMP token
Today due to the overcollateralization requirements, DeFi borrowing / lending is largely insular to the crypto ecosystem, where market participants view this as yet another form of leverage to speculate on crypto prices. Yield farming has been a big contributor, where people look to lend to protocols, with the purpose of accumulating governance tokens, constantly searching for the best rates & ROI. Given the price appreciation in some of these governance tokens, there are times where it was profitable to lend to a protocol such as Compound, borrow from it, deposit what you borrowed from it, rinse & repeat several times over with the value of governance tokens exceeding the spread between the borrow / lending rates. There were some cases of YieldFarming in 1Q21 that resulted in 4 figure + annualized yields.
In the future if we are able to build out on-chain identity tied to borrower history data, real-time collateral movement capabilities, etc… these collateralization requirements should decrease over time and open it up to the broader user base. With the ability to access global liquidity how does this change what a modern bank looks like? What does it do to the network effects of scale from a cost of capital perspective?
Trading / Derivatives:
Exchanges were one of the most attractive business models in the CeFi world with significant moats due to network effects & regulatory scarcity. CBOE, CME, ICE, & NDAQ combined have a $180B market cap, and will generate $16.5B in revenue, $10.3B in EBITDA, and $6.9B in net income this year. They have grown over time via a series of M&A as liquidity begets more liquidity, and their data becomes increasingly valuable.
In crypto, traditional CeFi exchanges were some of the initial recipients of VC funding & have driven a lot of the venture value-to-date. Coinbase became the seminal IPO and now trades with a ~$48B market cap itself, FTX has raised money at a $20B valuation, while Kraken has raised at $15B. Binance Coin, which isn’t a true equity claim, implies a $57B “market cap” for Binance, while Huobi Token implies $2.6B for Huobi. BitMex never raised outside money but at the peak there were rumors of $1.0B+ per year in net income between the three co-founders & team.
In crypto these CEX’s became honeypots for hackers and we saw hundreds of millions of dollars lost due to hacks ranging from Mt Gox, to Bitcoinica, Poloniex, Bitstamp, Bitfinex, Bithumb, QuadrigaCX, Binance, GateHub, etc… As the asset class has matured, the quality of the infrastructure providers has matured as well, as has their security policies & procedures, but the prize has also gotten bigger.
One of the early solutions to this was the advent of decentralized exchanges (“DEX”). Similar to other dApp, DEX’s work by utilizing smart contracts & on-chain transactions to reduce or eliminate the need for an intermediary. There are two main types of DEX’s
- Order Book-Based DEX- Very similar to a CEX, users submit orders (market or limit), while buyers & Sellers are paired beginning from the top of the book (highest bid / lowest ask) down. As opposed to assets being held in a CEX wallet they are held in a users own wallet, which resulted in latency & a sub-optimal user experience, creating Liquidity Pool DEX’s.
- Liquidity Pool DEX- Liquidity pools are essentially reserves of tokens in smart contracts, where users can buy or sell tokens instantly from the available tokens in the liquidity pool, with execution determined algorithmically. Liquidity pools can be shared across multiple DEX platforms and include protocols such as Uniswap, Sushiswap, Bancor, and Serum.
- Automated Market Makers (“AMM”)- we’ve seen liquidity pool DEX’s introduce AMM’s which algorithmically determine the price of the assets in the liquidity pool using a constant product for reserves (as opposed to having market makers like Citadel on platform).
Kyle Samani at Multicoin wrote a good blog post recently entitled, Technical Scalability Creates Social Scalability. In it he writes, “This essay assumes that the most important function of a blockchain is not non-sovereign money in and of itself. Rather, the most important function of a blockchain is DeFi…Blockchains should therefore be designed and managed over time first and foremost as DeFi development platforms.” He points to Solana as the only L1 solution that can support DeFi at scale as it “provides a predictable path to scale indefinitely.”
When Sam Bankman-Fried (CEO/Founder of FTX / Alameda & ex-Jane Street) was looking for a way to partake in the DeFi ecosystem he created the Serum Foundation and “launched Serum, a complete, non-custodial DEX running an on-chain central limit order book (CLOB) on the Solana blockchain. In addition to order books, it has cross-chain support, stablecoins, wrapped coins, and the ability to create custom, complex, and novel financial products. The combination of Solana and Serum DEX presents the speed, cost, and UX that users expect from centralized exchanges, all while being fully decentralized and trustless. With sub-second trading and settlement times, and lower transaction costs than any other DEX at $0.00001 per transaction, Serum is well positioned to help DeFi scale to billions.”
Jump Trading has dozens of engineers building on Solana / Serum, and last week one of the largest DMM’s on NYSE (GTS) announced it would be joining the Pyth Network, a decentralized financial market data distribution platform. Pyth was launched by Jump & Alameda built on top of Solana, with the objective of allowing precise, high performance market data to be consumed by smart contracts using Solana & other L1 protocols.
One of the biggest areas of interest in CEX’s has been margin trading with the average leverage on exchanges such as Binance / Bitmex at ~7–8x. We are starting to see decentralized offerings of margin trading such as dYdX, a decentralized exchange protocol for lending, borrowing and margin/leveraged trading. Today it offers margin trading with up to ~5x leverage using either DAI or USDC as collateral. Funds are automatically borrowed from lenders on the platform & you are liquidated as your position falls below the collateral threshold.
If we think about Dark Pools in the legacy market & the advent of ATS’, DeFi can follow a very similar path to CeFi to gain institutional adoption. We need to see continued innovation like that within the Solana ecosystem, and solutions around on-chain identity, custodial workflows, market data, scalability, and leverage.
Other Areas of Interest:
There are countless other areas of interest including Insurance (e.g. NXM), prediction markets, asset / fund management etc…. all of which are in their own nascent stages of development and can be augmented by the building blocks that is DeFi.
In Insurance we’re starting to see some insurers look to replace traditional re-insurance with smart contracts where they offer investors access to non-correlated yield, with double-digit returns, in exchange for taking idiosyncratic risk (e.g., hurricanes).
In prediction markets we’ve seen generation 1.0 platforms such as Augur offer sports & current events betting, with centralized entities built on top of decentralized protocols.
For asset / fund management we’ve seen early attempts at DAO’s offering crowdsourced / decentralized / auto re-balanced investment opportunities within stated bands of target position sizes etc… (with varying degrees of success).
Other areas include:
What Else Needs to Happen:
For DeFi to truly cross the institutional chasm there are a few things that need to happen beyond just continued technological innovation & sophistication, including (i) Regulatory Clarity, (ii) On-Chain Identity, (iii) Permissioned vs. Permissionless delineation, and (iv) Interoperability with legacy infrastructure.
The financial service industry is one of the most regulated industries in the world, and for good reason, as it has an impact on every single person on this planet. In the U.S. alone, there’s an alphabet soup of regulators including the SEC, CFTC, FINRA, the Fed, the OCC, FDIC, CFPB, State Bank Regulators, State Insurance Regulators, State Securities Regulators, etc… This isn’t going away any time soon.
For FinTech companies if they want to provide financial service products alongside technology there’s what appears to be an insurmountable amount of regulatory effort that needs to go upfront, prior to knowing if you have any product-market fit. Want to offer US listed equity trading? Either need to be a broker-dealer or work alongside an APEX or DriveWealth. If you want to launch an “exchange” to trade private company shares, collectables, or alternative assets more broadly, you need to first procure a broker-dealer license, and then file a form ATS in an attempt to become an “Alternative Trading System” in a process that could easily take 12–24 months (good luck trying to get an exchange medallion & becoming an SRO). What about a predictions market? You either need a no action letter such as PredictIt, or receive CFTC approval for an authorized Designated Contract Market (DCM), such as Kalshi. Want to launch a bank? You’ll need to either buy an existing bank & get Fed / OCC approval to take over their charter (e.g., Jiko), or decide on the various different paths between state trust, state charters, or OCC charter application processes (e.g, Paxos, Square, SoFi). Want to lend? Need to have your MTL’s & understand what states may or may not have reciprocity or blue sky laws. The list goes on and on.
Utilizing DeFi if you want to launch an exchange, a lending protocol, a predictions market, or a “bank” you can near instantaneously*; with the one big caveat that the regulators have yet to opine on where they see any of this falling within the current construct. Until 1Q21 this has been too small to matter, but most everything that people are building on DeFi would be illegal if tried in the CeFi world (e.g., Uniswap’s exchange, Makers lending, Curve’s MSB business, etc…). For DeFi to be a part of truly re-architecting financial market infrastructure it will need to have regulatory clarity. The largest financial institutions don’t have any other choice.
The industry is aware of this and Uniswap has recently put forth a $40M governance proposal to fund a DeFi policy organization. They highlight how “in the AML/CFT arena, policy changes are now being considered that could significantly harm DeFi. For example, the FATF’s March 2021 draft guidance suggested that governance token holders could qualify as regulated VASPs, or that crypto exchanges could be prohibited from allowing withdrawals to non-custodial wallets. Policymakers are also concerned about the risks posed by stablecoins — a critical piece of the DeFi ecosystem — to monetary sovereignty and financial stability. As time goes on, we can expect to hear more questions and concerns about numerous other subjects as well, including the US federal securities, commodities, and tax laws.”
If it can find a way for the technological benefits to exist within a regulatory framework there’s significant potential ahead, but without it, it will always live on the fringe.
Banks & financial institutions must adhere to strict KYC / AML policies. In todays’ DeFi ecosystem this is largely impossible. We need to find ways for individuals / firms to cryptographically attest that they are who they say they are, and are permissioned to do what they intend to do, and interact with others that have gone through a similar process, in a way that’s consistent with banking regulations.
This is the antithesis of what many of the DeFi entrepreneurs would like to believe. This will likely result in segregation of assets where certain liquidity pools are “ring-fenced” as “clean” versus not clean, but that is part of the benefits of the composability of the DeFi lego-blocks.
Permissioned vs. Permissionless Blockchains
This is a third-rail issue for most in the crypto ecosystem who convulse at the thought of a “permissioned blockchain” and believe it’s no different than a shared database. Enterprise blockchains such as Axoni, Chain, Figure, Paxos, R3, and Symbiont have raised $100’s of millions of dollars but have not found the same level of product-market-fit as their permissionless peers. There are some instances where having a sovereign-grade censorship resistant protocol is needed (e.g., Store of Value), however when dealing with the traditional finance world regulators & FI’s need the ability to audit, freeze assets, unwind transactions, etc… which may lead to some use-cases migrating to a permissioned environment. While these permissioned use cases have struggled by & large, there are a few use cases that have slowly gained traction that are worth monitoring:
Figure- Figure was founded by Mike Cagney (SoFi founder) in 2018 to “transform financial services using blockchain.” Since that time they’ve leveraged blockchain for loan origination, equity management, private fund services, banking and payments sectors. They have originated $100’s of millions of dollars of HELOC’s on-chain and are making a concreted push to vertically integrate the mortgage life-cycle while expanding into other use-cases such as alternative assets, now that they’ve procured an ATS designation.
Paxos- Paxos has a number of businesses including a stablecoin business mentioned earlier (as a result of a NY Trust license), an exchange in ItBit, crypto-as-a-service where they power other digital wallets such as Revolut & PayPal, as well as an enterprise blockchain product focused on US equity settlement.
- In April, Paxos announced, “Instinet and Credit Suisse settled US listed equities trades on a same-day settlement cycle (known as T+0) via Paxos Settlement Service. The trades occurred at 11 AM ET and 3 PM ET and were settled at 4:30 PM ET, demonstrating the platform’s ability to enable same-day settlement for trades conducted throughout the day. In the legacy system, settlement can only occur the same day if trades are completed before 11 AM ET, and therefore is rarely utilized. Paxos Settlement Service is a private, permissioned blockchain solution designed to allow two parties to bilaterally settle securities trades directly with each other. The platform is interoperable with the legacy clearing system and can facilitate settlement on any time cycle. It is also the first live application of blockchain technology for the US equities market and enables the simultaneous exchange of cash and securities to settle trades.”
Symbiont- Symbiont is a permissioned blockchain designed for enterprise finance targeting specific pain points within existing market infrastructure across four core end-markets including (i) Data (ii) Fixed Income, Currencies & Commodities (“FICC”), (iii) Mortgages, & (iv) Alternative Assets.
- Data- Today Vanguard manages $1.3T worth of index funds utilizing Symbiont Assembly. Distribution of index data comes in bespoke formats to subscribers in a number of different ways including secure FTP sites, similar batch-process based means, augmented by manual intraday updates via secure email, and even physical website screen scraping. Subscribers currently need to maintain historical databases of each index from which they receive data to conduct index roll-backs, recreate snapshots of the index for past dates / times, and conduct research on time-series constituents and changes. Reconciliation of subscriber index database copies to provider sources is incredibly difficult and in some cases, not possible. Data delivery via Assembly expedites the process, eliminating the need for manual updates, and reduces risk providing the data vendor a unified mechanisms to manage access, controls & distribution.
- FICC- Symbiont is working with Vanguard, Bank of New York Mellon, and State Street to digitize FX forwards & swaps by using Assembly to manage the creation and execution of complex OTC derivative agreements, effectively reducing risk and leveling the playing field for financial institutions (asset managers/corporations/banks/NBFIs) that rely on OTC derivatives to run their businesses. Prior to binding an agreement, counterparties propose and accept the business terms. Using those terms, the smart contract populates controlling legal documentation (ISDA and CSA) that is signed and stored on the blockchain. Based on the terms of the agreement and market data published to the network, collateral moves back and forth between the counterparties to fulfill their contractual obligation. Rapid movement of collateral throughout the life of the agreement greatly reduces counterparty risk. Automation also reduces the reporting and administration costs associated with trading a derivative book. The smart contracts on the network can manage the lifecycle of these agreements without the need to use calculation agents to ensure the integrity of the agreement. The frequent calculation of P&L and movement of collateral supported by our platform greatly reduces counterparty risk, limiting exposure to the P&L generated between calculation periods. Clients can also avoid information leakage, which is the cost of revealing positional information to market participants who might use that information to their advantage.
- Mortgages- Symbiont & Lew Ranieri are creating a decentralized mortgage record that securely shares data among servicers, investors and other key parties. The Ranieri and Symbiont partnership is converting legacy mortgages into auditable, immutable, and verifiable Symbiont Smart Securities where the complete lifecycle of the loan can be recorded and shared automatically with authorized parties. While there’s been innovation across the mortgage supply chain, servicers have predominantly relied upon large, monolithic legacy systems. developed in the 1960s (BlackKnight) that are not true servicing platforms, but mainframe-based accounting platforms without adequate workflow capabilities, largely built in COBOL. Over the last 10 years, new loss mitigation processes and regulatory controls mandated by the CFPB and other bodies have increased complexity in mortgage servicing. Due to the changing regulation, servicers have been unable to find a single system that meets all of their needs, which often times this results in servicers building “add-on” technologies in order to provide a quick fix as cheaply as possible. This patchwork of disparate systems creates process gaps and data integrity issues. As a result, mortgage servicers are faced with the escalating costs of running multiple systems.
Some of these solutions might have staying power, others might be replicated on permissionless protocols where they can be iterated by a wider group of developers, and others may be a solution looking for a problem, but this is still an area that’s important to keep an eye on, despite the lack of ideological alignment with the core DeFi proponents.
Interoperability with Legacy Infrastructure
Just as Paxos is interoperable with the existing DTC system, USDC announced interoperability with Visa, and Symbiont with SWIFT, Bank of New York Mellon, & State Street, for DeFi to really take off, during what will inevitably be a long migration period there needs to be a semblance of interoperability with legacy infrastructure, to provide institutional participants time to transition. While self-custody is a panacea for many of the crypto purists, it’s just not feasible for the largest pools of capital which means there needs to be a place for the State Streets & BNY’s of the world.
The amount of human & financial capital flocking to DeFi is growing at exponential rates, and it’s an exciting time to be an entrepreneur, builder, investor, and user in the industry. In less than 3-years since the term was initially coined, and four-years since the first product was released, we have $100’s of billions of dollars of capital in the ecosystem, with billions of dollars of value being transferred every day on rails that didn’t exist 5 years ago. Whether it’s programmable money, DEX’s, on-chain identity, democratization of borrowing / lending, access to global liquidity, a truly global, interoperable & instantaneous payment system, the potential permutations are limitless.
DeFi Pulse has an extensive list of leading protocols across lending, trading, derivatives, payments, wallets, asset management, infrastructure, insurance, UI/UX, stablecoins, education, newsletters & communities that’s well worth the read to go down the proverbial DeFi rabbit hole.