I know this is about bonds but since it’s so heavily about dex liquidity, I want to share some thoughts on dexes as they stand.
Fundamentally, LPing is not really about providing swap liquidity for a token pair. LP positions being used for swapping is an emergent property of a volatility-neutral market making position. What I mean by that is, when someone enters an LP position, they are taking the position that the two assets in this pair will not deviate from their current price ratio. Any deviation from the current price ratio when the position was entered is a loss - this is the impermanent loss. As it happens, the constant product formula is useful for both a volatility-neutral automatic portfolio balancing algorithm and a “bottomless”, price-agnostic token swapping pool, but the true origins of the constant product formula lie in the early days of algorithmic market making and very specific portfolio construction techniques for very specific outcomes.
So, what this means is that for liquidity providers, it’s a losing proposition to provide liquidity on a token pair where the underlying assets will change price. Does that make sense to anybody? Can you tell me any crypto, or really… any equity that obeys that rule in the long term? Stablecoins do, and, well, unsurprisingly, they attract orders of magnitude more liquidity than any other liquidity pool. So what that means is that it’s literally bad investing to LP basically any token except stables for any reasonable amount of time with the expectation of profit unless those yields are subsidized by artificial token incentives.
The protocol can “circumvent” this problem with bonds, but that just means the protocol is going to take on the “incorrect” position of LPing volatile assets, distributing those losses among the rest of the token holders over time.
One partial solution is what Uniswap V3 has done, which allows LPers to select the liquidity range they want their funds to apply to. This is still fundamentally a volatility neutral position, but you are allowed to select the range you are expecting volatility to remain neutral within, and the worst case scenario is that you just don’t make any money from fees when it trades outside of that range, as well as suffer a bit of impermanent loss (but significantly less than you would have if you LPed ‘normally’)
Another solution is the central limit order book (CLOB), which is well documented to be by far the most capital efficient exchange format (fun fact, AMMs have existed for decades! there’s a reason nobody uses them for exchanges outside of crypto). The downside here of course is the complexity involved in LPing in this style of exchange. This can be abstracted slightly with a “market making” protocol but, really, at that point you’re just building a slower Uniswap V3 with extra steps.
All this to say that I think the long term solution to the liquidity problem is more capital efficient markets. The objectively correct answer is all exchanges in the future will tend toward CLOB once crypto has enough active power users (specifically, high net worth individuals and competent market making firms to LP on CLOBs). But that doesn’t help us right now, I guess. So what conclusions do I think we can draw from all of this?
Ask oureslves this question: if we go the route of using bonds to build liquidity for our dex, operating under the assumption that we simply have no choice but to use the Uniswap V2 style pool for now for the sake of user experience, are we okay with the protocol accepting the losses involved with holding LP positions long term?
The advantage of “renting” LP is short term capital efficiency - $1,000,000 of invested capital in the form of emissions will attract more liquidity immediately than market selling $1,000,000 worth of our token in low-liquidity pools and using that money to buy our own LP. If we use bonds, we do avoid that problem, but we are also kind of trading the short term pain (in the form of depressed token value) for sustained long term pain (in the form of decreasing treasury value, as well as limited dex liquidity duing the critical growth phase), so I think we will still need some kind of long term strategy for solving liquidity issues that doesn’t rely on either renting liquidity or buying liquidity with bonds.
One thing is definitely true with the bond approach - the cost of taking this “losing” position is equal to the impermanent loss incurred from holding the position + the upfront equity cost of buying the LP, which will likely cost much less than renting liquidity with emissions will long term, so it’s a better immediate solution, but I don’t think it’s the solution.
The last point I’d like to make about dex liquidity is that the problem we’re really talking about here is not really liquidity, but swap depth. $1m of liquidity in a well managed uniswap V3 pool is multiple orders of magnitude better than $1m of liquidity in a uniswap V2 pool. Per the uniswap V3 docs, in a uniswap V2 stablecoin pool, like USDC-DAI, only 0.5% of the liquidity in that pool is allocated to trading around $0.99 and $1.01. So if there is $1m of liquidity in a balanced USDC-DAI pool, you can only make a $5000 swap before you exceed 1% slippage. That’s completely asinine! But I think it highlights just how god awful uniswap V2 pools are for facilitating exchanges, and I think that means there’s just no good solution to the liquidity problem that isn’t just using an entirely different method of exchange.