8 ms·

Wow. Source? That would be cool to look at.

by tkahn6 16y ago

seancron 16y ago
Mailing list thread: http://mail.python.org/pipermail/python-list/2010-April/1241578.html http://mail.python.org/pipermail/python-list/2010-April/1241... Blog post about it: http://jrvarma.wordpress.com/2010/04/16/the-sec-and-the-python/ http://jrvarma.wordpress.com/2010/04/16/the-sec-and-the-pyth... Proposal (pdf): http://www.sec.gov/rules/proposed/2010/33-9117.pdf http://www.sec.gov/rules/proposed/2010/33-9117.pdf
anigbrowl 16y ago
http://www.sec.gov/rules/proposed/2010/33-9117.pdf http://www.sec.gov/rules/proposed/2010/33-9117.pdf Warning - huge document, you should save it rather than try to view in the browser. See page 1 and section III.B.1 (page ~200 & surrounding). What's being proposed is not so much writing regulations in Python (the sort of idea I have been fascinated with for some time), but requiring sellers of complex securities to supply Python code as part of their disclosure requirements. Calculations of potential returns and so forth can then be undertaken by potential buyers using the same methodology used by the sellers. Currently, information about cash flow related to a security are described verbally in a special financial dialect of legalese, or sometimes in mathematical terms. Calculating the actual profits or obligations in $ is left as an exercise for the reader, and their calculation is only as good as their underlying assumptions about the meaning of specific terms. Should a dispute arise later on, the result is a long technical argument over how these calculations should be performed and whether the accounting methods employed are standard in the industry or not. Of course, if this argument takes place in a court it can become very expensive and last for years, which is why buyers of securities perform all their due diligence before purchasing the security. As this due diligence can easily become expensive and time consuming, a lot of buyers try to reduce the up-front costs by considering the rating assigned to the security by an agency like Moody's or Fitch, who are presumed to be specialists in evaluating these things. Unfortunately, excessive reliance on such ratings has resulted in significant loss for a lot of fund managers and investors in recent years. So what the SEC is proposing (and which you have until August 2 on which to comment), is that in future some securities offerings should come with the cash flow calculations pre-documented in Python, which would be the equivalent of a mortgage calculator offered by a bank to people who are thinking of buying a house and which is supposed to give an accurate picture of what the monthly numbers will be for a given set of input conditions, eg interest rates or whatever.

Wow. Source? That would be cool to look at.

by tkahn6 16y ago

Wow. Source? That would be cool to look at.

Mailing list thread: http://mail.python.org/pipermail/python-list/2010-April/1241578.html http://mail.python.org/pipermail/python-list/2010-April/1241... Blog post about it: http://jrvarma.wordpress.com/2010/04/16/the-sec-and-the-python/ http://jrvarma.wordpress.com/2010/04/16/the-sec-and-the-pyth... Proposal (pdf): http://www.sec.gov/rules/proposed/2010/33-9117.pdf http://www.sec.gov/rules/proposed/2010/33-9117.pdf

anigbrowl 16y ago

http://www.sec.gov/rules/proposed/2010/33-9117.pdf http://www.sec.gov/rules/proposed/2010/33-9117.pdf Warning - huge document, you should save it rather than try to view in the browser. See page 1 and section III.B.1 (page ~200 & surrounding). What's being proposed is not so much writing regulations in Python (the sort of idea I have been fascinated with for some time), but requiring sellers of complex securities to supply Python code as part of their disclosure requirements. Calculations of potential returns and so forth can then be undertaken by potential buyers using the same methodology used by the sellers. Currently, information about cash flow related to a security are described verbally in a special financial dialect of legalese, or sometimes in mathematical terms. Calculating the actual profits or obligations in $ is left as an exercise for the reader, and their calculation is only as good as their underlying assumptions about the meaning of specific terms. Should a dispute arise later on, the result is a long technical argument over how these calculations should be performed and whether the accounting methods employed are standard in the industry or not. Of course, if this argument takes place in a court it can become very expensive and last for years, which is why buyers of securities perform all their due diligence before purchasing the security. As this due diligence can easily become expensive and time consuming, a lot of buyers try to reduce the up-front costs by considering the rating assigned to the security by an agency like Moody's or Fitch, who are presumed to be specialists in evaluating these things. Unfortunately, excessive reliance on such ratings has resulted in significant loss for a lot of fund managers and investors in recent years. So what the SEC is proposing (and which you have until August 2 on which to comment), is that in future some securities offerings should come with the cash flow calculations pre-documented in Python, which would be the equivalent of a mortgage calculator offered by a bank to people who are thinking of buying a house and which is supposed to give an accurate picture of what the monthly numbers will be for a given set of input conditions, eg interest rates or whatever.