Defi Needs Better Primitives

Originally published in the Defiant

I was looking at this table comparing PC, web, and blockchain platforms the other day and couldn’t help but think, DeFi needs better primitives — that is to say, DeFi needs better base-monies.

This familiar table presents today’s blockchain development stack alongside the PC and web application stacks of yesteryear, painting the compelling picture of limitless possibilities. Encoded within it, the promise of a new and decentralized finance infrastructure — beyond the reach of politics.

Just as HTTP and FTP were the primitive building blocks of the web, today’s decentralized assets, BTC and ETH, are the primitive building blocks of the new DeFi economy.

These assets can be arranged to construct instruments for lending, debt, synthetic equities, and more. But today their limitations lie in their volatility and high correlation to one another — hence the requirement for extreme over-collateralization and looming risk of auto-liquidation.

What about DAI?

Often when I mention this to others operating in decentralized finance, they ask: “What about DAI, isn’t that a stable and decentralized primitive?

Not quite,” I respond, “DAI is derived from other primitives like ETH.

In Economics, there are: theories of banking, theories of money, combined theories of banking and money — but banks and monies are never mistaken as interchangeable without consideration.

Banks have balance-sheets, and at some level of depth, they hold in reserve a type of collateral that lives outside the realm of credit and debt, called base-monies (what we’ve been calling primitives).

Base-monies don’t have balance-sheets, they are not collateralized, they are atomic units within a given sector — and the quest for the ideal base money, has long preoccupied monetary economists.

Today, decentralized banks like MakerDAO compensate for the high volatility inherent in decentralized base-monies by over-collateralizing, often locking up several times the amount they aim to represent. Even still, risks of auto-liquidation loom ever-present.

One promising attempt to break free of this problem, was the introduction of Multi-Collateral-Dai (MCD). The thought being that: perhaps by leveraging a portfolio of collateral assets, their combined volatility could be reduced.

Unfortunately, the myriad new cryptocurrencies that followed BTC and ETH, have not helped. Instead, today’s floating-price cryptocurrencies mimic their movement pattern with supernatural closeness. With such hypercorrelations, the risk cannot be diversified away.

At present, the MakerDAO project faces increased pressures to introduce centralized collateral assets like Digix Gold. But in a more ideal world, the set of decentralized base-monies would be less correlated, or more stable — this would allow decentralized banks to avoid introducing centralized points of failure.

Said differently, uncorrelated defi primitives (base-monies) are the next line of defense against DeFi becoming, just Fi.

Still, there’s more to consider than hypercorrelation — and more to learn from the history of base-monies.

Lessons From Primitives-Past

Not long ago the dollar was redeemable by foreign governments for the base-money, gold, under Bretton Woods. Thriving as the global reserve currency after World War II, demand for US dollars exceeded the rate at which the state could obtain enough gold necessary to continue backing it. Naturally, this imbalance in supply and demand caused the price of gold to soar, further restricting it from circulation in a self-reinforcing cycle.

Then, as it is now, the global economy depended on US dollars to function — and the currency was at serious risk of entering a deflationary spiral.

The state needed the flexibility to increase its money-supply countercyclically. That is to say — in order to stimulate the circulation of money — the money-supply would need to move against the blind-forces of nature.

And since gold was neither sufficiently countercyclical nor elastic, in 1971 US president Richard Nixon was forced to cancel the dollar’s redeemability into gold altogether.

Thereafter the money-supply, and thus the value of the dollar, would be determined entirely at the discretion of human policymakers.

Bretton Woods Tip #1: The supply elasticity of base-monies matters, particularly when they are part of a broader feedback loop in support of a functioning economy.

Bretton Woods Tip #2: Countercyclical economic policies are executed in response to economic shocks to stimulate liquidity, when market forces would otherwise encourage users to hold.

Flawed as it was, the Bretton Woods standard enforced a simple set of rules: either the state had enough gold to back the money-supply, or it didn’t — and the world would simply have to deal with it.

This straightforward constraint had the benefit of eliminating all possibility of runaway inflation, preventing political tampering, and forcing judicious long-term considerations, around the certainty of said constraint.

And the Nobel goes to … “Rules Rather than Discretion”

Eventually this dichotomy between absolute rules and human discretion would captivate the minds of monetary-economists Kydland and Prescott — well into the era of pure fiat monies.

The two economists redefined this problem, as the distinction between time-consistent (rules-based) and time-inconsistent (discretion-based) economic policies — and they would eventually share the 2004 Nobel Memorial Prize in Economics for their discovery.

Kydland and Prescott, concluded that economic policies made at the discretion of people, however aligned in social objectives, would never result in the objectives being maximized — because discretionary economic planning is not a game played against nature, it’s a game played against profit-maximizing economic agents.

The agents’ knowledge of fallible humans behind the curtain, would inevitably distort incentives to encourage the abuse of any discretionary system, forcing undesirable tradeoffs between near-term individual gain and long-term collective well-being.

Their discovery left many in the world of Economics missing the simpler days of immutable commodity-reserve-currencies, free from human governance — but with the memory of Bretton Woods still fresh, vividly aware of their costs.

Kydland-Prescott Tip: Human discretion (or governance) be it individual or collective, results in suboptimal outcomes, so long as economic actors understand the system can be changed by the individuals in charge.

Looking Forward

Ultimately, if Nixon had replaced gold with BTC or ETH at the end of Bretton Woods, we wouldn’t have been any better off. But could a new and different base-money now be designed to fit the bill?

We certainly think so, and it is with this great macroeconomic puzzle in mind, that we created a new defi primitive: the AMPL.

At a high level, decentralized banks like MakerDAO can be viewed as directionally reducing the role of centralized banks. We felt — along similar lines — it was time to gently expand the role of base-monies.

In Economic vernacular: the AMPL is a countercyclical synthetic commodity-money, with perfect supply elasticity.

It is also an ERC20 token and series of smart contracts — that accepts 24hr volume-weighted-average price-information from oracles — and proportionally increases or decreases the quantity of AMPL’s held in every user’s wallet.

The Ampleforth protocol adjusts supply by updating a global scalar coefficient of expansion, in response to deviations from a price-target, once everyday.

The AMPL does not have a balance-sheet, it does not retake custody of tokens or airdrop new tokens to adjust supply — and no, it’s not a stablecoin — at least not by the use-cases assigned to the role of stablecoins today.

It is a new decentralized base-money.

But it does have another unique property. Due to the fact that gains and losses for AMPL holders are reflected in both the quantity of units held — and the price per unit — the Ampleforth protocol introduces a fundamentally different set of incentives.

As a result, profit-maximizing actors responding to these unique incentives, produce a step-function-like movement pattern that is expected to be meaningfully less correlated with today’s decentralized assets.

Our intent is to fill the near-term void of an uncorrelated primitive and grow into filling the long-term void, of a macro-economically friendly base-money.

Thanks for taking the time to hear our thoughts, we would love nothing more than to hear yours!

Evan Kuo,

You should follow me on Twitter: @evankuio

Defi Needs Better Primitives was originally published in Ampleforth Blog on Medium, where people are continuing the conversation by highlighting and responding to this story.