Recently I noticed two changes on p2p lender’s Lending Club website.
On the statistics page the link to download the loan data was removed. Before it was possible to download the complete loan data since inception of the service. Furthermore the predefined setting for the parameter “Loans issued from” is set on March 1,2008 now. That means, if you look on the page and do not change that parameter manually you see how loans performed that were issued between March 1, 2008 and today. Older loans issued between June 1, 2007 and Feb 29, 2008 are not included in the displayed results.
When I noticed that, I was reminded of what Prosper did with it’s statistics. Prosper segmented it’s loans (e.g. prosper select index) and cited only results for better performing segments in press releases. Furthermore the predefined values on Prosper’s statistic page, were set in a way that lowered the late payments and default ratios compared to an average over all Prosper loans.
But Lending Club had successfully positioned itself with transparency a core value in the past, so I asked Lending Club to comment on the reasons for the changes.
Rob Garcia, Director Product Strategy told P2P-Banking.com:
This is a temporary situation. We chose to take down the files due to privacy concerns raised by our customers. We are working to address these concerns in a way that continues to provide full transparency to platform data, while protecting the privacy of our customers….
On the setting of the parameter he stated:
The default setting for the statistics page is a year. So since we are now in March, the “From” date is defaulted to March 2008. This is to show the most relevant annualized indicators for the last year. Users can then change the “From” and “To” dates to explore the indicators for a specific time frame they may be interested in, including from inception (June 1, 2007). We did this based on numerous email inquiries from lenders asking for annual default rates instead of a general default rate since inception (so that they can compare annual defaults to annual interest rates to get actual net returns). We’re looking at tools to make that calculation easier…
Yesterday Lazy Man wrote about his observations on how Lending Club reports risk. The posted screenshots show that interpretation of the risk figures is not obvious under certain circumstances.
I trust Rob, however I need to see it back up there as I do use the data myself for analysis and trending. Lending has been paused for me as this is slightly ‘Prosperish’ (fishy and unacceptable) to me.
-F
Wiseclerk:
Thanks for your consistent and dependable coverage of the P2P lending space globally.
I read your concerns about transparency. Let me assure you that it is Lending Club’s commitment to provide as much information as possible to our community while protecting the identities and privacy of our members.
It caught my attention that you compared our statistic page to that of Prosper. We do not think the comparison is adequate: Prosper drew criticism by implementing loan filtering criteria that included better performing loans and excluded others that were not performing as well. Those pre-filtering criteria were not visible, explained clearly or easily accessible on the statistics page (you had to click on a link to make the criteria apparent). What’s more, they advertised higher returns and lower default rates in their communications based on these pre-filtering criteria. I believe they have, while in their quiet period, disposed of all pre-filtering criteria, but at the time, this was clearly misleading.
In our case, we are not pre-filtering the loans to those that are performing better. The default time period was set to the last year as per lenders’ feedback and is prominently displayed on the page. There has to be a default time period, and the last year seemed the most relevant.
As always, we are open to feedback from the community. I would love to hear from lenders on this topic either here or through feedback@lendingclub.com.
Best regards,
Rob Garcia from Lending Club
I’d like to see them detail what changes they made and why directly on the statistics page. Rob’s rationale for the changes seems perfectly reasonable, so the best way to minimize questions about transparency is to detail what you’ve done and why.
I’m happy to announce that the loan data is now available to download.
We have addressed the privacy concerns raised by our members by removing a few columns from the file: Current employer, Tenure, Hometown and Previous employer.
Enjoy!
Rob Garcia from Lending Club