# 1 | Decentralised Banking

Decentralising central banking and MakerDAO

This is another issue of Dirt Roads. No financial reports or marketing material dress downs, no investment advice, just deep reflections on the important stuff happening at the back end of banking. The time you dedicate to read, think, and share, is precious to me, I won’t make abuse of it.

About this Issue

This issue is the second of a series about what banks might become in the next decade or two. I have argued that, along the continuum that goes from absolute centralisation to total atomistic decentralisation, the local equilibrium of banking might soon shift towards a more decentralised model. Following my last post I have received a lot of backfire from seasoned experts and non-professionals. I hope that this issue will cast some light on what I see happening. My intention is to describe, rather than predict.

But before starting, let me clear the air. I have followed with utmost respect Marc Rubinstein’s Net Interest Substack. Marc is a seasoned financial analyst and former hedge fund manager. Like me, he has spent his entire career studying financial services companies. Unlike me, he has looked at them ‘from the outside’ as a public investor, rather than ‘from the inside’ as an advisor or controlling shareholder. I have started dedicating serious time to the DeFi (Decentralised Finance) r-evolution in the recent past while, it seems, Marc was doing the same. I was working on a post on the evolution of banking and here you go Marc published a piece on DeFi a couple of days before I intended to publish mine. I went on writing a post on MakerDAO and DeFi (this post!) and OBVIOUSLY Marc published a great piece on Maker a week before. Nevertheless, I will publish mine, and yes, I will steal some of Marc’s ideas and incorporate them within my framework.

Decentralising (a Bit More of) Central Banking

Let me break the news for some of you: we live in a decentralised banking world. In this world, the Central Bank — which is typically responsible for monetary policy and some level of prudential banking regulation, outsources a lot of decision making to authorised banks. While complying with some generic qualitative and quantitative guidelines, authorised banks manage relationships with clients and take the ultimate decisions on where to deploy (clients’) liquidity. There are currently more than 350 authorised banks (or MFI — Monetary Financial Institutions) in the UK alone. It is actually more than that. If we follow the Bank of England’s definition, as Marc quotes in his post, authorised banks do not merely custody money but actually create more of it via issuing loans and contextually depositing those loans in people’s accounts. They are, in other words, delegated ultimate authority on how to efficiently deploy fresh printed money in a system overseen by a Central Bank. The Central Bank acts merely as a parametrical authority — in this context. This, to me, amounts to a lot of decentralisation.

As I mentioned in my previous post, and I am ready to challenge anybody on this, authorised banks have been suboptimal, or more often than not cumbersome, in dealing internally with those client management and allocation responsibilities. An infinite series of controlling or merely bureaucratic overlays often created more opacity than benefits. Rather than managing risk, they hide it beyond the parameters monitored by the regulators, where they sit on waiting to come back and bite. Again, I am generalising and, again, I am ready to challenge any honest practitioner on my statement.

His Holinesses John von Neumann and Oskar Morgenstern taught us that designing the incentives of a closed system is the best way to predict and steer the outcome of that system. I am not the first one to have noticed that the incentives of the banking system are currently producing a suboptimal outcome, one that is more in favour of insiders than shareholders, clients, or lenders of last resort.

Is it possible to redesign the incentive system of banking for a better alignment and a more efficient use of resources? This is the key question that the DeFi community, including Rune Christensen and the crew at MakerDAO, is trying to answer.

What is MakerDAO

MakerDAO was born in 2014 as a Decentralised Autonomous Organisation (DAO). Based on the Ethereum blockchain, it simply enables users to create (crypto)currency. However, as any connoisseur of banking (rather than tech VC) noticed, Maker became much more than a crypto-vault. Going back to our banking framework jargon:

The MakerDAO protocol (the authorised bank)…

…allows users (borrowers)…

…to create (crypto)currency (currency in the form of deposits)…

…in exchange of selected tokens (to finance certain projects on credit)…

…placed as collateral (in exchange of specific guarantees)

…through the use of smart contracts on the Ethereum blockchain (bank’s infrastructure)

In Maker’s words (from Maker’s whitepaper):

1. Users create a Vault and fund it with a specific type of collateral

2. Vault owners demand a certain amount of Dai in exchange of keeping the collateral locked in the Vault

3. To retrieve a portion (or all) the collateral, Vault owners must pay back a portion (or all) the Dai outstanding (including a Stability Fee)

4. While the collateral is locked in the Vault, any Vault deemed too risky (according to parameters established by Maker’s governance and market conditions) is liquidated through a Liquidation Mechanism, automatically,by the Maker protocol

Glossary for the layman:

Vault: a smart contract (an automated piece of software living in the Ethereum blockchain) that custodies a certain type of collateral until certain conditions are met, without the need of external supervision.

Dai: cryptocurrency issued by the Maker protocol, soft-pegged to 1 USD — the ‘soft’ pegging means that there is no physical dollar set aside for each Dai issued but rather that a set of mechanisms and incentives embedded within the protocol supports the peg.

Stability Fee: fee that accrues on the Dai outstanding and needs to be repaid by the Vault owner to access the collateral in the Vault.

Liquidation Mechanism: mechanism designed to (a) maximise the liquidation proceeds of a risky Vault, (b) create and protect a central buffer to cover extraordinary losses, (c) guarantee governance stability, (d) avoid economic dilution of the community.

Dai Savings Rate: parameter (to be fully automatised) that determines the amount Dai holders can earn by locking their Dai into dedicated Vaults — the parameter is used by Maker to keep the peg to the USD in periods of instability.

The MakerDAO Way vs. the Traditional Way

Maker reinvented a decentralised and trustless (subject to hacks) way of doing banking, stripping it of all the bells and whistles characterising the gargantuan modern universal banking. But was it all worth it? What are the advantages of the Maker way, how is it performing, and what are its main threats and missing bits?

The Bright Side

Transparency. Running on the Ethereum blockchain, Maker’s algorithmic logic, remuneration, and performance are transparent and auditable real time (!), and its financial results are available only few days after the closing of monthly accounts. DAO (read the Board) discussions are also fully auditable and searchable.

Efficiency. At the end of May, Maker’s 2021 YTD net banking income amounted to c. USD 52.1m, vs. operating costs of c. USD 807k, this is 1.6% cost-to-income ratio. No provision for losses was recorded, as Maker actually made profits rather than losses during the April-May crypto-market turmoil.

Profitability.Recorded profits have been used to burn (destroy) MKR tokens (the tokens giving holders governance and implied financial rights over the Maker protocol — the thing as close as possible to shares in the crypto-world). Maker estimates USD 151.6m net profits for 2021, corresponding to a burn (buyback) rate of c. 50k tokens over the year at the current c. $3k MKR price. This is c. 5.6% of the total market cap, or a P/E ratio of 18. Considering Maker’s outstanding growth (see below) it doesn’t seem like an aggressive valuation. Assuming that the value of the assets is a fair representation of reality (big assumption) and discounting for any other risk.

Scalability. Arguably, a significantly larger MakerDAO wouldn’t mean a significantly larger cost base or more complex infrastructure, assuming the protocol is solid.

From MakerDAO

The Dark Side

Protocol risk. MakerDAO is a DeFi application running on the Ethereum blockchain, and operating through the use of several bits of blockchain infrastructure (including external actors such as oracles, keepers, and development teams). In addition, the protocol is designed to ensure security and efficacy. This means that Dai owners (and Maker’s MKR token holders) inherit Ethereum’s fragilities as well as Maker’s. There are operating risks in traditional banking too, but the model has been far more tested.

Credit risk. We argued in the previous post that, in its purest existential qualities, a bank is simply a trustable and (hopefully) knowledgable group of people supported by a bit (a lot) of infrastructure. Let’s assume that a working blockchain-based protocol solves the trust problem and the infrastructural problem; what about the knowledgable part? At Maker, risk parameters are set by a decentralised governance committee, do we trust that those parameters are solid enough to guarantee the bank’s funds? At the end of May 2021 Maker’s surplus buffer amounted to USD 38.6m, vs. total loans of USD 2,361m — that is only 1.7% of outstanding assets. In the case that losses from liquidation would exceed this amount, the event would result into a dilution of both governance and remuneration for token holders with potential impacts on protocol stability (the only way Maker replenishes the rainy day buffer is by issuing new MKR tokens — read rights issue).

From MakerDAO

Traditional banking analysts should look differently at credit risk in a decentralised world — I will dedicate a future post entirely to this topic.

Currency stability risk. Dai has sustained its soft-peg with the USD quite well in the past, but what would happen if the Maker protocol would show internal fragilities resulting in a significant depreciation of Dai vs. USD (and consequently collateralised tokens)? Vault owners need to pay back their outstanding debt in Dai in order to access their collateral value — a sudden decrease in Dai value (relative to USD and collateralised tokens) could trigger immediate repayments with a currency that holds a much lower value relative to its real economy counterpart (the USD). In the hybrid world we live in, it matters.

Regulatory risk. By soft-pegging Dai to USD, Maker remains an automated lending and saving protocol that outsources some of its trust-making to a centralised authority, and particularly to the FED, which might not be happy to anchor an unregulated pseudo-bank. Whether this will result in some form of blockchain infrastructural regulation (difficult) or in limitations to the exchangeability of Dai (more likely) remains to be seen. The FED, like many other Central Banks, is sharpening its focus on the development of a Central Bank Digital Currency (CBDC), which will be a de facto competitor for Dai.

Interestingly, the development of CBDCs and the consequent internalisation of risk taking decision making within the Central Bank would result in the most centralised banking system the world has ever seen.

What’s Missing

The future of Maker is still uncertain, but its merits in trying to revolutionise lending, currency creation, and deposits holding, are undeniable. More than anything, Maker is forcing banking analysts to look at banking in its purest core qualities, stripping it of all unnecessary ancillary services.

Asking whether Maker is a good bank equals to asking whether:

• The protocol facilitates a good alignment of incentives among borrowers (Vault owners) and equity providers (MKR token holders)

• The Maker community (or the required majority) is able to take sound decisions regarding credit underwriting via setting the right risk parameters

• The Maker protocol is successful in decentralising and automatising the price stability mechanisms that keep the value of Dai stable (relative to the USD) — traditionally the task has been dealt with craftsmanship by the Central Bank

A sound, stable, and working money exchange protocol is only part of the picture. Hidden by the overwhelming speculative activity we are currently experiencing, the merit of each supported collateral (i.e. token) is linked to the financial viability of its underlying project. Assessing this merit, which is the essence of banking, is what a good protocol alone cannot guarantee.