Silo Finance: Interesting, but…

The DeFi ecosystem needs lending markets for the long tail of protocol tokens.

• Being able to borrow and sell a token (short) is good for price discovery. One reason so many tokens acquire insane valuations in a bull market is there is no way to short them. I could find nowhere to short LUNA with size back in May.
• Perpetuals aren’t available on these long tail tokens because there’s no way for a market maker to be delta neutral during negative funding rates (long perpetual, short spot) without a lending market. If no market maker is willing to long the perp, it makes no sense to allow a trader to short the perp. Hence, no perp market.

So I was initially very excited by the idea of Silo Finance. It has an elegant design of risk isolation such that blow-ups of one token only affect the tokens in that silo (typically that long tail token and ETH) without any spread of contagion to other Silos. Intuitively, it seems the interest rate on deposited ETH in each silo would be different based on the risk of borrowing that ETH using the long tail token in that Silo as collateral.

If I want to use CRV as collateral to borrow CVX, I perform a hop in between: I use CRV first as collateral to borrow the ETH in that silo, and then use that ETH as collateral to borrow CVX in the second silo. The interest rate is netted out as deposit rate of CRV - borrow rate of ETH from Silo 1 + deposit rate of ETH is Silo 2 - borrow rate of CVX. ETH is the “bridge” asset.

Nice design.

However, the protocol has added a stable coin (”XAI”) layer on top which can be used as a bridge asset as well. Each silo can be extended credit in the form of XAI by governance.

I’m not sure I like this. I assume the reason for adding the stable coin layer is that it “lubricates” the markets without having to wait for ETH liquidity to be deposited into each Silo. Or maybe having a stable coin is the thing to do these days. Regardless, it seems to me to add a lot of governance (”How much credit do we extend to this silo or that silo?”).

Taking a step back, looking at the lending markets landscape, we know AAVE is the biggest player in money markets. People can borrow any asset against any other asset in the shared pool, and they are even doing experiments with long tail assets in a limited way, and of course, they have to be very careful about what assets they allow as collateral so they have a somewhat complicated design with concentration limits and so forth, but it’s very governance heavy.

On the other hand, Compound Finance, the second-biggest money market, recently changed to a model in which only USDC deposits can earn a yield, and the only asset that can be borrowed is USDC. All other deposited assets can only act as collateral for borrowing USDC. In other words, Compound Finance is a lot like MakerDao with USDC obtained from depositors instead of DAI created by minting.  In other words, Compound Finance has greatly simplified their protocol and diminished the required governance.

In my eyes, Silo Finance has another niche to capture: money markets with governance minimization. Like Uniswap, the ideal long tail protocol should allow anyone to activate factory pools and wait for the markets to sprout. Maybe a lending protocol needs a bit more governance to set collateralization ratios in, say, three different buckets, but aside from that, it should be “set-and-forget”. Most of the risk should be handled by the market in the form of interest rates, not by governance. Yet, with the stable coin credit lines to each silo, governance becomes a bigger factor.

I’m likely over-simplifying the complexity of standing up a long tail lending platform, so take everything I say with a grain of salt.

At the time I’m writing this, it has a ~$92M FDV, so it’s at a fairly low relative valuation. I like the lending silos; I’m not thrilled about the governance-set stable coin credit lines. That said, at some point, valuation may get low enough for the gm Portolio to grab a bag as a speculative position.