Variations - Githubissues

SocialEconomyDataLab / spec

The Social Investment Data Lab Specification is being developed as a draft data specification for describing social investment.

http://spec.socialeconomydatalab.org

Other

5 stars 4 forks source link

Variations #34

Open ScatteredInk opened 6 years ago

ScatteredInk commented 6 years ago

Social finance deals are not fixed.

Finance elements can be adjusted based on business stress, chances of recovering a loan and so on.

This might involve:

writing off part of the principle;
a payment holiday;
interest rate reduction;
conversion of loan to equity;

We will need a way to model these adjustments and, in particular:

the period in which they occurred;
the component of the deal they apply to; and,
any new deal components created as a result.

ScatteredInk commented 6 years ago

Variations could be used as a way of recording the difference between planned loans/equity investments and actual disbursements. The use case for this is to understand:

how well funders are planning for their capital needs;
how well funders are assessing the requirements/prospects of funded businesses; and
how well capital-raising businesses are meeting expectations.

So the top-level investment object might contain the amount loaned at disbursement, and we might have variations before and after this moment:

identifier	variations/0/date	variations/0/status	variations/0/description	variations/0/credit/0/identifier	variations/0/credit/0/principal/before	variations/0/credit/0/principal/after	variations/0/credit/0/principal/description
1	2018-01-01	Offered	Panel decision made	a
2	2018-03-01	Active	Loan disbursed	b	1000	100	Receiving organisation lost proposed lease and loan downgraded.

timgdavies commented 6 years ago

This makes good sense. Modelling wise, I've two questions:

Do we need description at the level of each change, or could we have reason and description at the level of the variation, and just before and after at the individual field level? (This is a question for @ScatteredInk)
Are we better to model variations/0/credit/0/amount/before and variations/0/credit/0/amount/after (which means we need to create a schema programatically, replacing each leaf with a before/after object) or could we use variations/0/before/credit/0/amount / variations/0/after/credit/0/amount - where we could just re-use the overall schema as a child of before and after (I think this is going to be mainly a question for @kindly thinking about how we might import this data into our data storage layer)

ScatteredInk commented 6 years ago

Do we need description at the level of each change, or could we have reason and description at the level of the variation, and just before and after at the individual field level? (This is a question for @ScatteredInk)

Reason and description at the level of variation sounds cleaner to me.