Bondora API: analysis, experience, options

Since the beginning of December Bondora has allowed investors to use third party API applications to invest into loans. In theory this is an amazing opportunity – people who are capable of analysing the dataset to create investment models can use this opportunity to achieve higher returns and manage their risks in a way a passive investor is just not capable of doing. However, the reality at this moment is a bit different.

Lack of available tools

So far the only API client that has been publicly released is the Beeplus client created by Jarmo. Hopes were that there would be multiple clients competing, but this has proven to not be the case. (I tried to contact LendTower who talked about releasing a client, but I didn’t get an answer from them.) I can only guess what the reason for lack of public clients is, but what probably impacts it is:

  • programming a usable client takes time/effort/skill
  • the process of how the API works isn’t final (sandbox issues, documentation issues), so developers are still waiting to work on it
  • people would rather not be responsible for other people’s money (mistakes the application make could be blamed on the developer)
  • there is no chance of monetisation at this moment (either due to people not being willing to pay or lack of usability for API at this point)
  • people are only sharing the clients with close friends

I asked Jarmo, who created Beeplus a few questions, since I know a lot of people are using Beeplus to invest via Bondora.

A bit about yourself (who are you, how long have you been investing?)

I’m an experienced software developer with some spare time in my hands (happens sometimes, rarely). I have also been investing small amounts at Bondora for a few years now.

What motivated you to create Beeplus?

When I saw an article about Bondora creating an API at their blog, I wanted to create something for myself to invest better at Bondora because their Portfolio Manager is very limited. After I started playing with the API in the sandbox environment I saw an opportunity to not just help myself, but also others like me and created the BeePlus. This is how it all got started essentially.

Who is the typical Beeplus user? (Do you have any user stats to share?)

I haven’t conducted any interview processes yet among BeePlus users, but deriving from some communication I’ve had with them, I’d say that currently big part of them are early adopters and more technical users. I haven’t made (almost) any PR to promote BeePlus and all current users have somehow found BeePlus. To be honest, I don’t know how. Currently BeePlus has 50+ active investors (meaning they participate daily bidding on Bondora’s auctions) and from the end of November (time when BeePlus went live) BeePlus has made 5000+ bids on behalf of investors through Bondora’s API.

Is Beeplus a hobby project or will you monetize it at some point?

Currently BeePlus has been a hobby project, but I’ve had some thoughts about making some money with it as well by introducing a freemium model (meaning that core functionality is free, but premium functionality will cost a little bit of money). It’s not going to happen in near-future if ever though.

What is best about using Beeplus? What are you planning to add in the future (or is Beeplus now ‘ready’)?

Best of BeePlus is that it is user-friendly (hopefully), because UX (usability) has been one of the goals since from the beginning. Security and privacy is also very important to me (for example many users ask for functionality, which would require BeePlus to query about investors available balance at Bondora’s account, but I’ve not done these yet due to privacy reasons – it might change in the future though). BeePlus is far from ready – I have a pretty long TODO list for it. I’m going to add some investment limits (monthly and/or daily) to the rules, secondary market support and so on. Functionality list is endless and I’m always hearing out all the ideas users might have as well, but at the same time I want to keep BeePlus simple and easy to use (UX, right?), which means that not everything will be implemented in the end.

Any wishes in terms of usability from Bondora’s side?

Bondora’s API has not been live long and that’s why there might have been many problems among API documentation and bugs in the system. Hopefully it will get better. However, there is a large problem of Bondora having lower priority for all API users compared to their own Portfolio Manager, which they are not planning to change. For now.

As you can see, luckily there is one option on the market written by someone who personally also invests and has been kind enough to share their work on the open market. However, at this point there are a few problems with how the API functionality works overall, that I’d like to address that I hope will get fixed in the future.

Prioritisation of bids

At this moment, the priority for bids is as follows:
1. Investors (likely to be institutionals) who bid via API to fill up whole loan
2. Passive investors using Bondora portfolio manager
3. Investors who bid via API

There are many key issues here. Firstly, at some point when institutional investors come into play, then we will be seeing more and more (or well, not seeing) loans not even hit the primary market. If we talk about a fund that wants to invest 2million/year then they are definitely not going to be investing into small pieces of loans.

At this point we do not have institutional investors on the market yet (as far as I know), but it will be interesting to see how it will play out. At one point, I’d say it might be a good idea to gather up your friends, pile up your investment money and start buying whole loans. Provided that you can pile up about 100K euros with your group of buddies this would probably work reasonably well in terms of diversification.

The second problem is the low priority of API bids, which is actually making the API rather useless at the moment. The whole point of the API being the idea that you can bid on specific loans that fit your investment model – however, since you are currently mostly bidding into loans that portfolio managers do not managed bid full, then trying to implement any strict filters for your loans doesn’t really work since you just don’t have the volume to choose from. This is especially true when we start thinking about a more conservative portfolio which at this point is considered by most to be Estonian loans AA-C.

The easiest way to fix this would be to reserve a certain percentage of loan total value to API bids. The reason why API bids are still getting AA loans at the moment for example is because the portfolio manager bids are mostly small enough that they do not fill up loans with higher values (you can also see fluctuations at points of the month when portfolio managers are “empty”). However, this will not last long since with every 200 loans in a portfolio the bid size increases – meaning in a few months more loans would likely be filled up by the autobidders. When visiting Bondora Pärtel mentioned that they would be looking into this, I hope if the situation does get problematic some sort of a fix would be made to the priority list (especially if other entities enter the bidding priority list). Currently as it stands, you cannot use the API for much other than a glorified country picker unless you have very strong models and are interested into investing into higher risk loans.

My bidding experience

Screen Shot 2016-01-15 at 14.49.30

Ever since the start of December I have invested 95% on my money via an API application, and a few loans I have picked manually. I am using an application that my husband built and as mentioned before, I am sharing it with a few close friends who have asked for it (since I don’t really want to be responsible for errors/support to an extent a public application would need). I am not using any complex model since I haven’t had time to develop anything but I am not ruling it out in the future, provided I have some time to analyse the dataset with the help of some analysis programs (not seeing this happening any time soon).

As it stands, for the whole month of December both my business & private portfolio used the application to bid. For my business account I managed to invest into 100+ Estonian AA-C loans. Mostly of course the result being B&C loans, since they are generally big enough, and there are enough of them to go around for everyone.

Starting January I set my private portfolio to only invest into Estonian AA & A loans, as I am slowly starting to withdraw money but I am still planning to keep a small conservative portfolio running. So far, it seems like I get an AA or A loan every 3-5 days. This is about the pace that I expected and will help me keep the pace of withdrawing all deposits into my private account within two years an reducing the portfolio size steadily.

For my business account, at the moment it seems like there is enough loans to go around if you are not investing at insane rates. Getting into at least 50 EST AA-C loans seems entirely possible, and for most private investors that is a perfectly reasonable level of diversification. As I said before, as the portfolio managers start bidding more or as institutional investors enter the market the situation might change, but if you don’t want to use very complex models you can get the money out. However, using very complicated predictive models at this point is rather impossible, however at least the clients are being built (if only secretly at this point), that in the future they will hopefully be of much more use.

5 thoughts on “Bondora API: analysis, experience, options

  1. Hi.

    I also bid via the API and i was wondering one thing: have you had success in investing into loans that are below 1100-1200€? It seems as if those are always filled by just one user or 4-5 PM-s. The users of those 4-5 PM-s are always/mostly the same.

    To me it looks like something fishy is going on with those loans. Reserved to certain privileged users perhaps. But then again I might be just paranoid 😀 .


    1. I think it might just seem so, since there probably are just a few investors investing such big amounts that they always stand out (plus, they probably also always have money while other PMs might not, at least one of those big bidders is an institutional investor though). If someone bids for a full loan (regardless of being instititional or not) they get priority, otherwise all PMs should be equal.

  2. Well it’s probably institutional or API user with set bid size 1-1,2k, which is first in the order to bid.

  3. Oktaedter also seems to have a plugin that has some sort of API support built-in through javascript, but that’s not a free option and I personally haven’t had any experience with that one.

    In general though API in today’s format is relatively useless indeed for any more meaningful filters than country and Rating for several reason:
    a) You have no idea what the actual sample of loans is that you can invest into. In other words, you analyze the whole dataset, and come to some conclusion that loans matching a certain criteria are a great investment.

    However, when bidding, the good portion of the loans that made this segment look great, get funded before you even have a chance to bid, and your actual results end up being totally different from what the analysis showed, since, in reality, you can only invest into a small portion of that segment.

    There’s also no way to figure out from the data, what loans are even available through API, so you can’t exclude the irrelevant portion from your analysis.

    b) Even if you get into this specific segment, it will happen so rarely, that the sample of loans you’ll get is so small, that the risk of your portfolio is seriously higher than that in the analysis. Simply because of the randomness of outcomes from such small samples.

    In other words, best way to use the API today, is to use the best historical filter of country EST to end up with an above average returns at below average risk levels.

    For example, according to Bondora’s XIRR (which is not the best way to look at return of relatively fresh loans, of course, but will serve for this purpose), the returns for EST vs the rest would be following based on time when loans were issued:
    – 2013:
    EST: 23.2%
    weighted average: 21.0% (the weight of non-EST loans was small)
    EST: 23.9%
    weighted average: 16.5%
    EST: 21.8%
    weighted average: 19.5%

    For that, however, Bondora could’ve saved the time, money and resources of everyone and just add that filter to their Portfolio Manager 😉

    Perhaps we could even see the DCA fees by now in the UI then.

  4. In case someone reading this article starts to use BeePlus and is experiencing problems then feel free to contact me directly regarding them. Also, if you feel that something really important is missing from the application then I’d happily hear about it as well. Happy investing!

Comments are closed.