Reading and assessing governance proposals
A monitoring process that needs no opinion on the proposal — and still triggers in time.
This lesson has had no expert review. It was written for this platform and against the evidence it cites; nobody has gone through it independently.
Learning objectives
- You can classify proposals by their effect on existing positions.
- You can design a monitoring process that does not depend on a judgment.
Check your prior knowledge
Answer these for yourself before reading on. Wherever you hesitate is where this lesson pays off.
- Which decisions affect positions already entered?
- How do you learn of a proposal before it is decided?
- What separates a parameter change from an upgrade?
Core concept
Three classes, sorted by effect
First, proposals with no effect on existing positions: use of funds, partnerships, personnel. Second, parameter changes taking immediate effect on existing positions: loan-to-value, reserve factor, rate curve, accepted collateral. Third, changes to the code itself. For monitoring, all that counts is whether a proposal falls into class two or three — the substantive assessment comes after.
The process must not presuppose an opinion
A monitoring process that only fires when someone judges a proposal bad fails exactly when it matters — on a proposal that sounds reasonable. What holds is a process firing on the class rather than on the judgment: class two or three means a review inside the window, regardless of how the proposal is argued.
The window starts at notice, not at the decision
A timelock runs from the decision. Your own usable window starts only once someone has noticed it. Checking weekly against a 48-hour timelock normally loses the entire window. Monitoring frequency is therefore not a matter of diligence but follows arithmetically from the shortest timelock in the book.
Definitions
- Parameter change
- A decision changing market quantities without replacing the code.
- Notice
- The moment a decision is noticed and assigned internally.
Model
Proposal submitted — watch the source, do not wait for the decision
Determine the class — does it affect existing positions?
For class two or three: trigger a review, regardless of the argument
Hold it against the pre-agreed exit threshold
Decide before the timelock expires
Formulas
Required monitoring frequency
Check interval < shortest timelock − time to unwind the position- shortest timelock
- smallest delay across all protocols used
- time to unwind
- time closing the position actually takes
Limit: Holds only for decisions subject to a timelock at all. For emergency powers the right-hand side is zero or negative — frequency then replaces no protection, it only bounds the damage.
Worked example
Weekly checking is not enough
- Shortest timelock in the book
- 48 hours
- Position unwind duration
- 6 hours
- Current check frequency
- weekly
Required: check interval < 48 − 6 = 42 hours. Actual: 168 hours.
The documented 48 hours of warning are, under weekly checking, practically zero.
Reading: The protocol's figure is right; it simply does not take effect. A report naming the timelock as protection without naming your own check frequency describes a protection that does not exist.
Retrieval
Exercise on real data
Assign the case study's proposal to one of the three classes and compute the required check frequency from the figures given there.
Case study: parameter change →Application
Design the monitoring process in four sentences.
Related case studies
Institutional reading
- Bank
- Does the check frequency follow arithmetically from the shortest timelock in the book?
- Asset management
- Is the exit threshold fixed before a proposal is submitted?
Metrics in this lesson
Key takeaways
- Only two of the three proposal classes affect existing positions — monitoring sorts by that.
- The process fires on the class, not on the judgment.
- Check frequency follows arithmetically from the shortest timelock.