Blue Elephant Capital Management

A quick thought on big data

Peer-to-peer lending blogs are all over the news that Lending Club will cut back on the fields available to investors that allow them to screen loans for a number of important characteristics. This will make it more difficult for investors to dive as deeply into the credit model that Lending Club is running.

There are a few dozen institutional investors and countless individuals that are trying to use data in order to screen out potential bad loans. A quick look at Lending Club’s website suggests that 80% of investors have achieved returns between 6.7% and 10.2% over an 18-month seasoning period. Now, we cannot tell from this data whether the performance difference is based on individual loan selection or quality bias. With default-rates near all-time lows, the likely differentiator is quality – the higher the overall risk of a portfolio, the better the returns have been.

There isn’t a day that goes by without us meeting someone who claims that they have a unique, proprietary process by which they can pick out better loans than anyone else. Let’s think about that just for a moment. To start, we have two major underwriters between Prosper and Lending Club who must constantly upgrade their models to compete with each other. Then, you have the few institutions and hundreds of investors doing their own screening. Out of each 100 loans made, somewhere around 8 on average will default (according to the underwriter’s themselves). How can everyone do it? It seems pretty unlikely to us.

To be clear, we do run an API to screen loans on the Prosper platform and we’d prefer that they keep the data as open source as possible. However, we think most investors in this space have lost for the forest for the trees.  Peer-to-peer investing is not magic sauce – there will be cycles and there will be losses. Prosper, Lending Club and all other lenders past and present will make mistakes – getting too aggressive at the wrong time and pulling back hard after defaults are well underway. What do we do about this?

At Blue Elephant, our claim has been and will continue to be that investors need to treat p2p much like other asset classes. There are times to be up in quality (now, for example) and times to be lending more aggressively. We’ve arranged access to multiple underwriters differentiated by style, grade and geography. There may even be times where we pull away from the space for a time. Sure, we will try to pick out the worst loans and leave them to others – but this matters much less than knowing where in the quality spectrum to be and when not to be involved at all.

That’s what the partners at Blue Elephant have spent the previous 15 years doing – creating value for investors through business cycles by understanding markets, interest rates and risk.

Lending Club’s decision to reduce the amount of data available is disappointing, but not surprising. It gives us a good chance to put everything in perspective. It is impossible for hundreds of investors to all correctly cherry pick the winning loans. In fact, until the next 2 or 3 default cycles play out, we’re not sure how anyone can even make the claim. While we’re biased, we think it highlights that the more important investing expertise comes from the macro side – understanding the broader markets and the correlations of the space to those markets.