Right after christmas, some spare time again in my holidays to update scripts and rerun analysis on Bondora’s PD and LGD-rates. My previous articles (here and here) explained in depth the details how the analysis is done, the method has not changed for the following analysis, so please read the previous articles about if you are interest in the calculations itself. Please note that all calculations are still done for all available loans and do not differentiate between rating grades or loan scores.
I have added Slowakia this time as an additionial full year of history became available. Numbers however are not really surprising, just consider your money invested in SK gone…
For the other 3 countries however the PD rates seem still to be very unsatisfactory from an investors point of view: Finnland increased from some 35 to now 40% and Spain increased from below 50 to above 50%. Just Estonia decreased slightly to some 17% now. Regarding the Days to default (“DTD”) there is a tendency that it increases fro all countries. This might be because the “B-secure” program has been introduced which however would just delay a default in most cases. This is however just my personal opinion as Bondora currently does not disclose any details about the impact of the B-secure program in their dataset.
Unfortunately there is not too much good news about the recovery rate trend: Significant rates are generated still only for EE investments and even there it takes still some ~3 years to recover a significant share of 80% and more of the defaulted amounts. Recovery rates for FI an ES are significantly lower.
As with now a little history it is also interesting to watch, if there is any improvement made in the relative recovery, i.e. we comare recovery rates relative to the analysis date. For this to compare I have put the results from the previous and the actual analysis into a single chart with “t – 4 quarter” (“-Q4”) to the most right point and then one quarter into history increasing to the left. Dark lines are from the current analysis, bright lines are coming from the previous analysis, also markes as “(prev)”.
Please note, that FI and ES for the respective first quarter are both with very low numbers of loans, so these numbers are not very representative. This chart also puts all countries into a nice comparison, just picking loans which defaulted 2 years ago: While EE loans recovered in average at 20%, ES is 5~10% and FI is ~10%. This shows that Bondora has quite obvious problems to recover non-EE loans at all. However, the relative recovery performance has improved a little: For both ES and FI the recovery after the same number of quarters has gained a little.
Bondora is still a very high risk investment in the broad – although they represent themselves in a different way. With default rates of 40 %(FI) and 50%(ES) and a recovery that might take 4~5 years to get even half the money back, the other investments need to account for almost 70~100% per year. And please note: This is a yearly default rate on portfolio that has invested into all rating grades. So for ES this might account to a total of 75% loss after 2 years! It is very obvious to see, there cannot be any win in these markets.
After last year’s analysis I have drastically changed my investment profile in two ways: I buy only the two best rating grades (A and AA) to reduce overall risk at all. And I use other platforms that offer buyback guarantees to all or portions of the loans (Mintos and Twino e.g.) as they seem to attract less risky loans by nature (otherwise the buyback would make no sense from their point of view).
It might be worth however to buy some 2nd market investments now with defaulted EE loans at very low rates (<30%) and hope they will recover within a few years. This is however very risky investement as Bondora has changed their policy to recover loans two times in the recent 18 months which for sure has a impact on the rates that is hard to predict. Also one needs to be patient as it will take 4~5 years to see significant gains in recovery.
What the impact of these changes is, well…we’ll see in another 2 or 3 years, when there is even more history data available…