BUG: There may exist multplit roots in the solution of IRR function

[yatshunlee]

Describe the issue:

This was what I encountered:

What I got in Excel:

Reproduce the code example:

import numpy_financial as npf
print(npf.irr([-10000]+[327.24625]*16))

Error message:

NA

Runtime information:

NA

Context for the issue:

There may exist multplit roots in the solution of IRR function so that the irr() may give an incorrect result.

[Kai-Striega]

Hello @yatshunlee: firstly my apologies for taking so long to reply, I was on holiday over Christmas/New Years and didn't check my GitHub. I'm back and should be quicker to respond from now on.

What version of NumPy-Financial are you using?

irr has given us a lot of problems, and we are well aware that there is (usually) more than one solution. In the development version, it uses a completely different solver based on Newton's method and a guess parameter. guess can be changed to influence the solution from irr. Using a guess of 0.1 (10%) will give the same solution as Excel:

>>> import numpy_financial as npf
>>> npf.__version__
'1.1.0.dev0'
>>> npf.irr([-10000]+[327.24625]*16)
-1.7781470715009497
>>> npf.irr([-10000]+[327.24625]*16, guess=0.1)
-0.06765411344912187

I'm currently working on a grant to upgrade NumPy-Financial, after which I plan on releasing a 2.0 release which will contain this feature.

[yatshunlee]

Happy New Year~~ BTW

By referring to your iir function at v1.0, I have created a relatively stable (to me) version of iir which doesn't require me to give an inital guess. I can help to contribute and give an alternative way to solve. May I create a pull request so that I can share my version to you and see if my edit also works for you so that you don't have to spend extra time to solve it?

[Kai-Striega]

You're welcome to submit a PR and I'll review it.

However I think it's quite unlikely that we will move from the iterative solver for the reason that it is both simpler and faster.