A Three Token Stablecoin Model Based on the Fed

There are many ways to go about creating a stablecoin, but what would it look like if you modeled your coin explicitly after the Federal Reserve? That’s what I am going to lay out here. Note that this model is meant to be purely hypothetical, and any resemblance to an existing coin is purely coincidental.

The first thing to point out is that the US Dollar is a banknote issued by the Federal Reserve. That is, the Fed is like a bank, and USD are the liabilities of the bank. The Fed also holds assets, mostly in the form of notes from the US Treasury that pay out in US Dollars after some period of time (i.e., Treasury bills).

Thus, to model a coin after the Fed, you really also need a Treasury. So our model starts with two coins: bank coin and treasury coin. Bank coin is issued by the bank contract, and is a liability of the bank contract. Treasury coin is issued by the treasury contract, and is a liability of the treasury contract. The treasury coin represents a promise (by the treasury) to pay some amount of bank coins after some amount of time.

Note that the treasury contract is the one obligated to pay bank coins, but it can’t mint bank coins. All it can do is mint more treasury coins and give them to the bank in exchange for bank coins, which it then uses to pay the interest (or principle) of the treasury coins.

The bank contract and the treasury contract exchange some coins. So then the treasury contract owns bank coins (which are assets for the treasury and liabilities for the bank) and the bank owns some treasury coins (which are assets for the bank and liabilities for the treasury).

Now, in the case of the US fed and the US Treasury, the Treasury is also the owner of the Fed. So our treasury contract will own the bank contract in the sense that any profits from the bank will be sent to the treasury.

But who owns the Treasury? In the US, the Treasury is owned by the government. So we will also have a government, which will be represented by a third coin, the governance coin. So any profits from the Treasury will be distributed to owners of the governance coin. So here are the three coins:

• bank coin (liabilities of the bank contract)
• treasury coin (liabilities of the treasury contract)
• governance coin (owners of the treasury — and indirectly, the bank too)

The bank is charged with maintaining stability of the bank coin. To do this it first needs to know what it means to be stable. So the bank must regularly consult with an oracle to determine whether it the value of bank coin is stable.

If the oracle says that the value of the bank coin is going down (inflation), the bank needs to reduce the supply of outstanding bank coins. It does this by selling some of the treasury coins it owns. This increases the relative value of bank coins.

Since treasury coins are denominated in bank coins, their value (T) depends on three things: 1) the value of bank coins (B), and 2) the inflation rate (I), and 3) the real interest rate (r). For simplicity, let’s assume that the treasury coins are perpetuities (pay out a constant amount over time, forever), the equation is:

T = B/(I+r), where R = r + I is the nominal interest rate

So, if you decrease inflation, you increase the value of T. Seemingly, if we increase the value of B by reducing supply, which also reduces inflation, I, we should increase the value of T. But we are also increasing the supply of treasury coins, which we expect should decrease the value of T. Furthermore, if the oracle is all-knowing, it would already know that the bank is going to reduce supply, and therefore the predicted inflation rate should already take into account expectations about what the bank is going to do! This sounds like a theoretical problem, but it has real implications because actual financial markets (which are the best oracle we have) respond (in advance!) to what they think the Fed is going to do.

One possible effect of Fed interactions in the market is that it ends up changing r instead of I. This is the source of a pretty significant debate between different schools of economists. For example, the diagram below shows an example of a possible (and somewhat complicated) result of the Fed acting to increase the nominal interest rate. Initially, the real interest rate (r) goes up but eventually inflation goes up (π).

Also, for a bank coin that isn’t a primary economic reserve currency, it is not clear whether transactions in bank coin can impact the real interest rate (which is usually considered an economy-wide variable).

If it isn’t clear by now, a stablecoin modeled on the US Treasury and the US Federal Reserve is somewhat mind-bending and it’s not clear if it would actually work. However, it works pretty well for stabilizing the US dollar so we shouldn’t dismiss it outright.

At the end of the day, a fiat currency like the dollar is a social construct. It has value because billions of people across the world accept it as payment for transactions. This also has implications for algorithmic stablecoins in that they may be subject to a chicken-and-egg problem if they can’t really be stable until they are widely adopted…and they can’t be widely adopted until they are stable.