I'm not sure how soon the design and then, implementation for HBD time locks will come. I imagine things won't start... today.
But since the topic is hot, I gave it some more thought and identified a number of elements that will probably need to be addressed or handled carefully.
Interest After the Time Lock Expires
Let's say Alice wants to lock 100 HBD for 90 days for x% annual interest.
How would this process work (assuming it is done on the base layer, no bonds system at the second layer)?
This is a scenario, we will all see how it will actually work when it will be implemented.
Alice must have 100 HBD liquid in her account, then choose from her favorite interface the duration she wants to lock it for (probably a dropdown). In this case, she will choose 90 days. She will make sure the interest is acceptable for her. Probably she will have to confirm details. She will have to use her Private Active Key to confirm the transaction.
For 90 days (under normal conditions - we will discuss more later), she will receive the fixed x% on her 100 HBD.
The question is what happens after 90 days?
Scenario 1: Time lock expires, but the interest remains fixed at x% until Alice withdraws her HBD.
Problem: This defies the purpose of the time locks after the initial period passes.
Scenario 2: Interest remains fixed at x%, but time lock is automatically extended by the same period after a grace period to withdraw funds.
Problem: Too complicated. Can generate frustration and anger if funds get re-locked before someone realizes (notifications can be posted) they need to be withdrawn.
Scenario 3: Time lock expires, but interest drops to the interest for the shortest lock period or no interest at all is paid until funds are withdrawn or time lock is extended.
Outcome: Other than potential complications coding it, this seems like the best solution.
Note: Many of the problems above are generated by the fact the user (Alice) needs to use her Active Key to authorize whatever operation is decided at the end of 90 days. With a posting key, this could have been automated by a public service. But since this is a financial transaction, nothing can be done to improve the user experience, as far as I can think of.
Fixed Interest Rate and the Haircut Rule
The discussion was, at least during the last CTT, that funds entering HBD time locks should offer a fixed interest rate, so that the investor knows what he or she is going to receive, even if witnesses decide to change those rates the next day.
This makes sense because of the time lock. You can't lock away stablecoins if you don't offer one of two things (or both):
- predictable APR for the locked period of time
- the option to withdraw/liquidate if the investor is no longer pleased (in exchange of paying some penalties or receiving no or lower interest)
But, when we say fixed APR for a HBD "deposit", we have to keep in mind there is no such guarantee. If the haircut rule is applied, then no interest is paid to anyone, fixed or not.
The acceptable solution I see to this is that the interest is paid regularly throughout the lock period instead of once at the end. The investment still remains locked for the whole period.
This way, if or more likely when we will hit another haircut event, at least the investment pays the interest up to the point the haircut is reached (and potentially after restabilizing, if the time lock is long enough).
3-day Withdrawal Delay or 3-day Time Lock?
There is no guarantee we will continue to have the 3-day period in either form once the HBD time locks mechanism replaces what we currently have, but we all assume we will still have an incentivized short term option, and because we are used to this duration, I'll continue to use the 3-day period.
I talked in my post from yesterday about the difference between the two. A 3-day delay allows withdrawals at any time, but they are pending for 3 days (or until cancelled), during which no interest is received. A 3-day lockup period would not allow the account holder to withdraw HBD for 3 days, but probably the withdrawal, once initiated, would be instant.
The question is: does it make sense to build short-tem time locks, or we should use something like we already have implemented for the HBD savings account?
I think the answer to this question relies heavily on what we discussed in the first section, what happens after the time lock expires...
But I don't see why we wouldn't leave this mechanism we already have in place, on the very likely scenario it will pay a much lower APR compared to longer time locks.
Conclusion
I'm pretty sure the HBD time locks mechanism will have enough edge cases that will make it something not trivial to build and test. So we might not see it in the next hard fork in my opinion.
This is something I came up with after 2 days, sometimes as the topic came up in the podcast recording (fixed interest rate, for example).
I think this is going to be the last post on this topic from me, for a while. More when it'll be better defined so we can have something to talk about rather than imagining scenarios.