What is enterprise DeFi? It depends on who you ask.
If you ask Paul Brody, at EY, it’s about enterprises using DeFi (decentralized finance).
If you ask me, it’s about creating new, decentralized web-scale enterprise structures and effective internet-based coordination mechanisms among known or unknown legal and natural persons. In other words, governance is everything.
Paul’s and Richard’s perspectives are entirely valid and certainly part of the picture. From my perspective, enterprise DeFi must decentralize the enterprise, not centralize DeFi software in the hands of legacy corporate structures.
The operative word is decentralized, not finance
Thanks to the army of mathematics and physics PhDs whom Wall Street hired between 1970 to 2008, it is incredibly difficult to invent new finance. With the rare exception of, say, flash loans, any competent quant can map financial innovations in DeFi 1:1 to a previous financial innovation on Wall Street.
Ajit Tripathi, a CoinDesk columnist, is the head of Institutional Business at Aave. Previously, he served as a fintech partner at ConsenSys and was a co-founder of PwC’s U.K. Blockchain Practice.
That means the real innovation in DeFi is decentralization, and even that has roots and parallels in the history of money and commerce. Decision-making by token-holder voting is not entirely different from proxy voting, and traders guilds in ancient India and medieval Italy ran mutualized lending, trading, insurance, money and asset management services using similar decision-making systems.
Public blockchain technology enables decentralization to work at internet-scale among pseudonymous participants who don’t need to trust or even know each other, and that’s a pretty big deal. Ergo, if becoming enterprise means losing this core aspect of decentralization at web-scale, then it’s just CeFi (centralized finance) or legacy-Fi on the blockchain, not DeFi.
So what exactly is decentralization? I have written about the topic over the years. The key question to ask in assessing if something is decentralized is the governance question, i.e. who has control over decision-making and to what extent? That has two dimensions – technical decentralization and economic decentralization.
Technical decentralization means that anyone can use or contribute to the software or technical infrastructure created for the decentralized product or service. In DeFi, the deployment, enhancement and operation of the software is not controlled by any one legal or natural person. The open-source nature and public blockchain deployment of DeFi is therefore non-negotiable.
Using a public permissionless blockchain helps because it exposes the functionality to everyone who cares to examine or reuse it, but it’s not sufficient. If a crypto exchange builds and runs smart contract-based applications on a public permissionless blockchain, but 95% of the commits to the blockchain come from employees or contractors of the exchange, the blockchain is not very decentralized.
Also, if the exchange can reorganize the blockchain at will, then such DeFi applications are certainly not decentralized. In fact, Linux and MySQL are more decentralized than such dapps. Similarly, if the founding team of a DeFi protocol controls the admin key that can be used to make arbitrary code changes, then such a protocol is not very decentralized.
As it turns out, not all open-source software is the same, and some open-source software ships with licensing that gives special rights or even control to the creators of the software. The fewer such rights and restrictions embedded in DeFi software, the more decentralized such software is. In essence, I believe no license is more decentralized than GNU licensing, which in turn is more decentralized than MIT licensing, which again is more decentralized than Apache licensing. The legal rights and restrictions embedded in the software licensing are much more important.
Taking technical decentralization to enterprise requires some compromise because enterprise needs to be commercially incentivized to use the software and because enterprise entities often assume liability for non-performance of the software or the service. That is why Geth (which allows anyone to run an Ethereum node) ships under GNU licensing and Hyperledger Besu ships under Apache 2.0 licensing.
Economic decentralization has two core aspects: control and liability. The less concentrated each of these is, the more decentralized the protocol.
Enterprise structures, whether sole trader businesses, partnerships or corporations, are defined by who makes the decisions that govern the operation of the service offered and who has the liability for nonperformance. A limited liability corporation, which is the most popular enterprise structure, assigns limited liabilities to principals, i.e. owners/shareholders.
The liabilities for nonperformance often sit with agents, i.e. managers whom shareholders hire to make decisions on their behalf. When the agents’ incentives are not aligned with the principals’, it leads to principal-agent conflict. A classic example: When Wall Street execs make decisions to maximize their bonuses, rather than long-term shareholder value.
Indeed, there are even aspects of decentralization within centralized corporations. The level of economic decentralization depends on the level of concentration in stakeholder voting. Before the existence of blockchain technology, internet-scale coordination of stakeholder votes was practically nonviable. Shareholders had to elect a board of directors to represent them. In such a structure, if the CEO of a corporation can push any decision through on a whim without any pushback from a board looking out for long-term value, such a corporation is more centralized.
Similarly, there are elements of centralization even within DeFi protocols, and as Securities and Exchange Commission commissioner Hester Peirce’s safe harbor proposal recognizes, nearly everything starts centralized. Bitcoin, the first ever DeFi protocol, started with one guy writing a white paper on an obscure internet channel and then a few developers, like Gavin Andresen, writing the bulk of the code.
Over the next few years, bitcoin (BTC) became widely distributed, and the number of contributors to bitcoin core grew from three people to a few hundred people. The fact that bitcoin was practically worthless but practically brilliant in the early days meant people weren’t hoarding as many bitcoins as they could or running giant mining pools next to hydropower dams. In fact, one could even argue that after the initial years of rapid decentralization, bitcoin has become more economically and technically centralized as its value has gone up hundreds of times.
In the case of a DeFi protocol, the exam question is how much more decentralized it is than when it started and what’s the rate of increase in decentralization? Are tokens becoming less concentrated in the hands of whales? Is the team using treasury to force decisions? Is the team controlling an admin key to force decisions through? Are a few large whales making decisions? Does a participant need to buy tokens on an exchange before they can vote, or if they have a good idea that benefits the community, can they just ask the community to delegate votes to them?
So what is enterprise DeFi?
DeFi and enterprise have a lot to learn from each other. The most important thing enterprises can learn from DeFi is how to use public permissionless blockchain technology to become more technically and economically decentralized to enable much greater fairness and inclusion in economic activity. For example, corporations can employ public blockchain technology for shareholder voting and limit the role of the board of directors to guardians rather than often conflicted intermediaries. Corporations can further enable shareholders to delegate votes to other stakeholders in the society and thus alleviate the externalities and ill effects of shareholder capitalism.
Then there is a lot of technical innovation and practices to be shared. Specifically in enterprise blockchain, a lot of work has been done on security, data privacy and software engineering practices. In DeFi, there’s an infinite library of robust, battle-tested smart contracts built by highly motivated and talented engineers that enterprises can simply adapt or use. DeFi protocols can also benefit from robust software development, testing and deployment practices established in enterprise software.
And yes, of course, enterprises can simply plug into DeFi and benefit from the growing asset liquidity, the return generation opportunities and the entire category of peer-to-peer and peer-to-protocol web 3 commerce that is coming online with non-fungible tokens.
Most of all, forward-thinking enterprises can facilitate the legal innovation that’s required to provide a firm legal foundation for decentralized protocols. A recent and extremely pertinent example is the Wyoming DAO (decentralized autonomous organizations) Law. As DeFi goes to enterprise, we will use more of such legal innovation to decentralize the enterprise rather than building centralized, hierarchical, rent-seeking entities on decentralized technology.