Implement decrease in staking rewards

AIP-30 - Implement decrease in staking rewards

Summary

In Aptos tokenomics overview, Aptos Foundation presented the anticipated token supply changes across time. Currently, the maximum staking reward rate is a constant annualized rate of 7%. This AIP proposes a 1.5% yearly decrease in staking rewards to align with Aptos tokenomics:

• The maximum reward rate declines by 1.5% yearly until a lower bound of 3.25% annually (expected to take over 50 years).

For example:

• Maximum reward rate in the 1st year(year starts from genesis timestam 2023/10/12): $7\%$
• Maximum reward rate in the 2nd year: $7\% * (100\%-1.5\%) = 6.895\%$
• Maximum reward rate in the 3rd year: $7\% * (100\%-1.5\%)^2 = 6.791575\%$
• …
• Maximum reward rate in the 51rd year: $7\% * (100\%-1.5\%)^50 \approx 3.28783\%$
• Maximum reward rate in the 52nd year: $max(3.25\%, 7\% * (100\%-1.5\%)^{51}) = 3.25\%$

Motivation

To fully align with Aptos tokenomics overview.

Rationale

Considerations:

1. Year starts from date of genesis: timestamp based (10/12)
2. 1.5% decrease happens at the end of every year. This means that at the current rewards rate (7%), the effective rewards rate at the end of the year would be 6.895% (7%-1.5%*7%)

Alternative solutions:

1. We can compute gradual decreases throughout the year(e.g. every 30 days), but this would make rewards calculations more complex

Reference Implementation

https://github.com/aptos-labs/aptos-core/pull/7867

Future Potential

All rewards and reward mechanisms are also modifiable via on-chain governance

Suggested implementation timeline

Targeting end of Q3 2023