Resend: Loan payments vs financial calculator mismatch

[Geert Janssens]

On Monday 19 October 2009, Mike or Penny Novack wrote:
> >The small excess in monthly payments is probably due to a different
> >method of interest calculation and I do not have time to investigate. You
> >may be able to produce the correct payment by adjusting the rate. In
> >particular, I see that entering a rate of 5.6% produces a payment of
> >343.53
>
> That is likely to be an unsolvable nuisance, a situation where the
> differences between different ways to construct an amortization table
> and different choices in how/where rounding takes place and different
> choices for dealing with the reality that equal payments (all of them)
> rarely going to be an exact solution mean that while discrepencies may
> be made small, they can't be eliminated.
>
> In my day I have written a number of programs to "generate an
> amortizaton table" for x "rents" as well as the other sorts of things
> like accumulation tables, etc. There are a number of different ways the
> calculation can be done even if totally agreed as to how the interest is
> to be calculated calculated (and that's not certain).
>
> Michael D Novack, FLMI

I suppose the "unsolvable" part you are referring to is that there may be a 
rounding error when using different algorithms to calculate an amortization 
table. I do agree there.

However, when I have to change the rate from 6.6% to 5.6% to get the correct 
result (or close to correct), this is no rounding error anymore.

And more even: the GnuCash' built-in financial calculator give me the correct 
result (+/- a rounding error) with the unmodified parameters.

I'd expect that inside one program, the same algorithm would be used in the 
different interfaces to calculate amortization tables. However in GnuCash this 
is not the case. Using the same set of parameters gives me very different 
results via the financial calculator and the loan payments druid.

I asked my question to verify if this difference was on purpose, for example 
to cater for different accouning rules or something. But this doesn't seem to 
be it.

So I think this is a bug, and I'll write a bugreport for this in the next 
couple of days.

Geert