It's certainly a better idea than Tether's false accounting, but I'd rather have my dollars undercollateralized by a bank backed by reserve lending and depositor protection schemes than overcollateralized by much riskier assets like ETH.

I agree. But eventually they plan to allow for many forms of collateral, and peg the price to a basket of goods. If convenience were the same, I would rather hold an overcollateralized currency with stocks, bonds, gold, ETH, etc. underneath. Such a thing does not exist yet, but I'm excited that it will, if MakerDAO continues to follow through on their plans.

Trouble is, I don't see how smart contracts validate the actual presence of assets like stocks, bonds and gold underneath. There's obviously going to be centralization and trust necessary, in which case I might as well rely on regulated financial intermediaries exchanging real dollars rather than hold a synthetic dollar with the downside counterparty and residual value risk associated with holding derivatives linked to a basket of goods but no upside.

I'm also struggling to understand the incentive for people to overcollateralize to receive tokens notionally equivalent to dollars and how that's convenient

To your first point, the vision is to use other Ethereum tokens tied to various assets and securities. How those asset-backed tokens are set up and how trustworthy they are depends on the individual token. I know there are a few asset-backed tokens (gold, real estate) but I have no knowledge as to how legit they all are.

To your second point, the collateralized loan allows for very very low interest margin investing. You can use the DAI to invest in something, sell it at a higher price, repay your loan, and make money on the difference.

DigixDAO works on tokenized gold. If there is demand, we will see every possible good as a token. It shouldn't be that difficult to make an virtual product like an ETF into a token on a blockchain.

There are financial incentives to create and destroy DAI. A user of DAI doesn't need to collateralize, he just buys DAI.

DAI demand will grow, if the use cases grow. It is also a good safe heaven for crypto traders.

> the idea behond DAI is that it is over collaterized by Ethereum coins

I love how we've come full circle to uninsured fractional reserve banking.

It's actually the reverse of fractional reserve banking. Any Dai that is created must be over-collateralized by 150%. It's not capital efficient, but that's a different argument than what's being discussed.

> It's actually the reverse of fractional reserve banking

It's analogous. (More accurately, it's a carry trade.) If Ethereum crashes more than 33% (given 150% overcapitalisation) the token breaks the buck. Betting your less-liquid assets won't fall below your more-liquid liabilities is maturity transformation and presents the classic risks of fractional-reserve banking.

There are mechanisms in place for this scenario. Incentives for people to "break the contract" of contracts that come close to undercollaterization.

This has happened, and AFAIK, DAI is still stable. Maybe that's the result of some extra manipulation by the project or whatever. I would love to see an in depth analysis of DAI over the past 6 months.

No, it's part of the protocol. When a collateralized position falls below the 150% (or whatever) mark, that position gets automatically liquidated and auctioned off to pay for the debt.

And the positions are set up to where you can borrow less than the maximum amount of DAI per ETH, to make your position more resistant to price drops.

> When a collateralized position falls below the 150% (or whatever) mark, that position gets automatically liquidated and auctioned off to pay for the debt

The same thing happens with banks. It turns into a run when this forced liquidation drives prices down further which in turn fuels further redemption requests. It’s very possible to start the day 150% capitalised and end 20% because you sold 10% which tanked the market.

Parent's example uses a 200% reserve. How is that fractional?