Kiva Data Collection

[Anqi Fu]

Hi, my name is Anqi Fu, and I am an economist at Stanford GSB. I am
conducting a research study on social lending behavior on Kiva. I was
wondering if it was possible to access data on the specific amount
each lender contributed to a particular loan. For example, if a
borrower asks for total funds of $100 for a project, I would like to
know if lender X contributed $50 on a specific date to this project.
Right now, using the Kiva API, I have only been able to find the
aggregate loan amount over all lenders.

Thanks for your time!

[Paul Ericksen]

Unfortunately, the amount a given lender has loaned to a borrower is not public information. For currently fundraising loans, it is possible to infer this information from the API http://api.kivaws.org/v1/lending_actions/recent.xml -- if you keep track of what the loaned amount was prior to the current lending action, and you see a new lending action and the new loaned amount of that loan is up by $50 but there's only one new lender, you can infer that new lender gave $50, rather than the default $25. But, for past loans, this information is not public.

--
You received this message because you are subscribed to the Google Groups "build-kiva" group.
To post to this group, send email to build...@googlegroups.com.
To unsubscribe from this group, send email to build-kiva+...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/build-kiva?hl=en.

--
Sent from my computer. No excuses for misspellings or missing words.

[Pim Schaaf]

Hi Anqi,

Like Paul says, there is no direct way of getting that data. I must warn you that once in a while the lending_actions/recent.xml method won't contain overlapping data (even when checked with only 1 min. intervals), so that might not be as reliable as you want it to be.

Liu et al. (2012) have, for similar reasons, used the loan amount divided by the amount of lenders as a proxy for the specific amount each lender contributed to that loan. Bigger loans are found when less people lend to projects with higher total funds required. See the following quote:

In addition to lending frequency, we are also interested in the effects
of motivation categories and team affiliation on the amount
lent. However, to protect lender privacy, individual loan amount is
not available through Kiva data API. Therefore, for this analysis,
we employ a proxy variable for the amount lent. We know the list
of projects that each lender lends to, as well as the total amount
lent to each project. We therefore assume that each lender to a
project lends an equal amount. Once we apply this assumption to
all projects, we have a proxy for the total amount lent by each user.

and the results they find:

Table 9 presents four OLS regressions using the proxy lending
amount as the dependent variable. Independent variables in each
regression are the same as those in Table 8. While the significance
and direction of motivation categories and team effects remain the
same as those in Table 8, it is informative to highlight the size of
some of these effects. Specifically, a lender motivated by general or
group-specific altruism lends $6 less per month than others, while
those motivated by external reasons lend approximately $7 less. By
contrast, a lender who sees Kiva as an effective development tool
lends $5 more per month, while one motivated by religious duty
lends $9 more. Again, when controlling for team affiliation (column
2), we find that a lender belonging to any team(s) lends $31
more per month than those without any team affiliation, while each
additional team joined is associated with $16 more lent per month.
Overall, the effects of motivation categories and team affiliation on
amount lent is consistent with those on lending frequency.

Check out the article through http://adf.ly/8YlDA (skip the ad).

Good luck!