It’s time for Curation 2.0. The curator model was a big advancement, but there are some commonsense enhancements that can result in a better product for users.

Tranching is back in vogue lately. I think I’ve had at least three teams reach out to me about tranching design. Where there are three, there are likely more, so I’m going to share my advice publicly.
First, let’s understand how challenging it is to tranche an evergreen vehicle like a Morpho or Euler vault. People have to be able to withdraw at some point - there’s no winding down date where all the books are balanced and gains/losses distributed.
This means you have to prevent junior tranche participants from collecting their outsized yield, then leaving ahead of senior creditors, upending the intended distribution of losses. MakerDAO/Sky ran into this very problem with the 2024 ConsolFreight default where the junior creditor was monitoring the investment better and exited ahead of a default, leaving Maker holding the bag. Eventually it was written down after a residual recovery.
So adding a tranche where a curator is in the junior position is helpful, but the mechanics matter a lot. Some commonsense options to add tranches to curator vaults thoughtfully:
* Permanent capital from the curator is locked in the vault. This cannot be withdrawn until the vault winds down. This prevents early exit when the curator (who is better placed to spot problems than anyone else) sees things going sideways on a Sunday evening.
* Subordination ratios should be enforced. For example, if there must be 20% of the pool in the junior tranche, no new deposits are accepted into the senior position until more junior can be found. Similarly, junior cannot withdraw at all unless there is at least 20% of the pool in the junior position. This ensures that junior creditors cannot exit before bearing losses they were paid to take.
* Junior withdrawal limits should be in place. These could be a maximum amount or a time lock on junior withdrawals (even if in the intended subordination range)
* Adding new collaterals requires the consent of a majority of the deposit base. This prevents the risk profile from migrating unexpectedly. For example, if you deposited into a vault that’s lending only to blue chips, you should not wake up to find it now allocates to loans against fartcoin.
* Curator fee cannot be increased without the consent of a majority of the deposit base. This prevents curators from growing their AUM with a low or on-market rate and then raising it abruptly with minimal communication.
There are other, more complex things that could be added, but some combination of these features overlaid onto Morpho and Euler and even Aave and Compound (who are themselves really just curators with different parameter sets) would be an easy way to build more robust lending and products for HNWIs and institutions. It also helps ensure vaults behave in a way that depositors and borrowers expect.
If anyone wants this built, or is even building this now, my DMs are open. But moving full speed ahead with better credit and credit-like products is really important so that DeFi can continue to grow and become more competitive with alternatives.