Hi Adam,

Thanks for confirming. I've created a ticket for this: FINERACT-3049
https://issues.apache.org/jira/browse/FINERACT-3049

I did some small tests with both engines and noticed a difference. For the
period 2024-01-01 to 2024-02-01 (12%, 10,000, daily, ACTUAL), the
cumulative engine gives 101.92 (31/365), while the progressive engine gives
101.64 (31/366).

I also tested a period crossing the year end (15 Dec to 15 Jan). The
progressive engine splits the calculation by calendar year, while the
cumulative engine doesn't.

I propose to keep this ticket focused on cumulative declining balance
loans, covering the leap-year calculation and the year-boundary split, so
that it behaves consistently with the progressive engine.

Flat interest and AprCalculator need more investigation, especially since
APR uses a fixed 365 when calculating the stored annual rate. I think it
would be better to handle those separately.

Some existing integration/e2e tests with daily interest and 2024 dates may
assert the old 365-based numbers. I expect to update those expected values
in the same PR. Is that acceptable?

Does this scope sound okay to you? I've set the ticket type to Bug, but I'm
happy to change it to Improvement if you think that's more appropriate.

I will start with the leap-year part while I wait for your reply.

Regards,
Senesh Perera
GitHub: https://github.com/ImalkaPerera

Reply via email to