DeFi's Original Sin
Banking vs Central Banking
Crypto’s root use case as a hard money substitute has fixated the community’s attention around central banking and money
History teaches us that the arc of financial development is to do banking before central banking
The Federal Reserve was established in 1913 or ~140 years after independence
In that time the U.S. successfully financed railroads, real estate, wartime weapons, ships via banking channels without a central bank
In the U.S. Civil War era – banks operated as private businesses that issued their own money native to the community
The notion of a central bank was underpinned by Walter Bagehot in the 1873 classic Lombard Street
Bagehot centered the central bank use case as “lending freely against good collateral at a penalty rate”
The modern derivation of this is the so called “lender of last resort”
Central banks were meant to be a backstop – not a faciliatory of economic affairs
Defi Today
Right now there is limited real world utility across any of the crypto stablecoins
Most DeFi is leverage folding chasing its own tail
This can be shown w/ the velocity of money in ETH over time
DeFi yields come from arbitrage price differences / demand for speculation
DeFi is a prime beneficiary of the liquidity lottery played by central banks
But in bear markets with dampened volatility borrow utlization shrinks
We see this today as TVL and USDC/USDT market caps have shrunk dramatically
As the cost of money soars and liquidity tightens utility will need to be centered around real world applications
Banking Framework
Banks borrow short and lend long
This transformation of time is the key use case in earning a fair liquidity premium
The DeFi community talks about overcollateralized / undercollateralized / algorithmic models
But few seriously discuss actual banking
Let’s use an example with DAI
I have $100mm in deposits and I make a $100mm loan
If my cost of liquidity is 4% and my loan is 8% I make 4%
In one year I book $4mm in equity – in 2 years I book $8mm
Loans in this construct creates a profitable, collateralized bank
Overcollateralized lending is just an on-chain pawn shop
Undercollateralized lending is subject to confidence, panic bank runs that are highly visible on-chain
Implicit Debt Ceilings
The biggest hurdle to creating an on-chain bank is the implicit debt ceiling from liquidity pool drainage
Let’s say MakerDAO lends $100mm DAI to Walmart
Walmart has no use for DAI given its payroll / purchase orders are denominated in USD
So Walmart would swap DAI into USDC and transfer it to Coinbase and redeem USDC for USD
Walmart may just to borrow DAI because it has a lower rate than a USD alternative and has an implicit fragility option
Therefore an on-chain bank’s loan book can only scale with it’s liquidity pools
Nobody would borrow $100mm of DAI if there is only $50mm of exit-to-USD liquidity
The debt ceiling problem can be managed by
Creating use cases to hold DAI
Issuing a form of DAI that is not convertible over a period of time
Locking in defined liquidity over a period of time
Confidence / resiliency growth whereby DAI is equivalent to savings
Developing actual use cases for someone to hold DAI means MKR can mint DAI and create loans without creating liquidity pool strains
The debt ceiling problem means banking cannot grow without sturdy deposits
With mercenary, yield seeking capital we cannot build an efficient on-chain bank
Savings-As-A-Service
The prime use case for a stablecoin is something you wouldn’t appreciated unless you’ve lived in a non-OECD country
Much of the non-OECD world experience degradation of savings denominated in local currency
This is due to the political failures in reducing fiscal/current account deficits
Countries use inflation as a coordinated playbook to devalue savings and lower borrowing costs
The largest borrower in any country is the government itself
That’s why we see robust dollarization in emerging markets
Having the ability to save in USD in Turkey gives you access to labor-time destroyed by inflation and bank failures
But access dollars is not easy and has high friction costs
Official USD rates are manipulated by governments
DeFi in this lense has a use case as a global savings account that can exist in countries without the presence of branches or institutions
There is no JP Morgan in Liberia
DeFi becoming a savings technology for non-OECD countries allows on-chain banks to have sticky deposits
Sticky deposits mean less mercenary liquidity and higher interest margins on loans
On-chain banking has a captive liquidity layer
Some Ideas for On-Chain Banking
Building an on-chain bank requires sturdy liquidity which I believe can be tapped into from onboarding hard currency demand from non-OECD countries
But there are other areas in DeFi that need significant improvement
DeFi needs a yield curve - almost all activity is spot
On-chain banking may mean significant liquidity needs to be locked/non-redeemable primitives
Borrow/lend tools need to be advanced via loan servicing and on-chain indentures
Credit originiation in real markets like strcutured credit, commerical real estate
On-Chain Banking End Game
The end game of on-chain banking is simple
On-chain banking is cheaper given cost structure
Immediate payment / settlement
Tokenization allows for borrow/lend incentives not possible today
Here is a relative comparison of key KPIs of JP Morgan vs MakerDAO
Excellent post, thanks!