bg
Member of DD Central
Posts: 1,368 
Likes: 1,929
|
Post by bg on Aug 3, 2017 7:50:07 GMT
Newbie Report after *2* months D-class loanbook, diversified as requested (100 loans) - I went for yield, with risk. Manually selected (?!) based mostly on their financial account (allegedly profitable business, no one-man-bands, no obvious balance sheet lunacy) 2 in liquidation, one notified late payer two months late (downgraded). That's an annual default rate approx 15%, obviously unsustainable. The question I'd ask is - how old were the loans when you bought them? Getting into 100 D loans in 2 months sounds like you bought a lot of them on the SM, just as they are coming up to peak risk. I believe that FC's averages are based on an assumption of buying new and then holding: but the risk of failure for a given loan is clearly not constant, and if investor A is shifting his share of the risk downwards, there must be an investor B whose risk has been shifted upwards. Do you want to be A or B? When is peak risk?
|
|
bg
Member of DD Central
Posts: 1,368
Likes: 1,929
|
Post by bg on Aug 3, 2017 8:14:29 GMT
Could it be that there was higher chance of loans defaulting straight after you bought them because someone knew there will be an issue with loans they sold and it will be better from now on? There seems to be a presumtion that with any loan that is bought on the SM that defaults/goes late, the seller knew the company was running into issues. Not everything is a conspiracy! For a band D loan, FC estimate the ANNUAL bad debt rate is 5% and for band E its 8%. That means with 45% expected recovery rate that every year you would expect 9% of D banded loans to default and 14.5% of E banded loans to default. That is every year. So in a 2 month period on a portfolio of 100 D's you would expect 1.5 defaults.....so 2 is hardly out the ball park. I would expect on a number of portfolio's of 100 random loans that there would be a distribution of defaults between 0 and about 6 or 7 (for very unlucky people) in a 2 month period. Your performance is roughly around expectations. I have been running a large FC portfolio since 2012. I have 35 loans that have defaulted/downgraded. Yesterday I went through all of them and googled the company name as well as looking up looking them up in public records. For my 35 loans there was only that recent £250k D (god knows why that wasn't downgraded earlier) that I could possibly have found out any information that would have let me sell the loan before it was downgraded. On top of that its just not practical to do this for all the loans in my portfolio everyday. It would take hours of boring work for a minimal saving (as people have said before it would be like working for way under the minimum wage). You just have to accept that there are risks...returns are not smooth but if you run your portfolio for a long time (3 years+) then general downturn aside you will on balance get the returns FC advertise as average (but there will be a distribution around this, some will do worse, some better - the more loans in your book the closer you will be to average). It's never nice to lose money on a loan and you always think, ah i wish i had never bought that, what could I have done to avoid it but you can't think that way. Look at the broad performance over a long period of time (just like any long term investment).
|
|
|
|
Post by GSV3MIaC on Aug 3, 2017 8:41:59 GMT
'when is peak risk' .. if you look at the slope of the graphs on FCs statistics page (assuming they are still there) you can see .. or analyse the loan book, which I quit doing some time ago. Definitely an inverted bathtub curve though. 6 month old loans are more likely to go bad than fresh ones (and ones in their last months are safer again) .. plus, with amortising loans, you are more likely to lose 100% of the amount you have left (the 'only 40-50% loss' was based on buying new loans, which
have usually repaid something before they go bad .. and if they die very early, FC has sometimes repaid all the capital.).
As I said up-thread, sellers usually sell for a reason, even if the reason is just better understanding of the default risk curve. But yes, it is statistics, and there is considerable variation (again, FCs graphs show what the spread can be, even for someone in 100 separate loans).
|
|
seabbs
P2P Blogger
Posts: 41
Likes: 22
|
Post by seabbs on Aug 3, 2017 8:42:12 GMT
|
|
seabbs
P2P Blogger
Posts: 41
Likes: 22
|
Post by seabbs on Aug 3, 2017 8:52:10 GMT
|
|
bg
Member of DD Central
Posts: 1,368
Likes: 1,929
|
Post by bg on Aug 3, 2017 9:12:53 GMT
'when is peak risk' .. if you look at the slope of the graphs on FCs statistics page (assuming they are still there) you can see .. or analyse the loan book, which I quit doing some time ago. Definitely an inverted bathtub curve though. 6 month old loans are more likely to go bad than fresh ones (and ones in their last months are safer again) .. plus, with amortising loans, you are more likely to lose 100% of the amount you have left (the 'only 40-50% loss' was based on buying new loans, which have usually repaid something before they go bad .. and if they die very early, FC has sometimes repaid all the capital.). As I said up-thread, sellers usually sell for a reason, even if the reason is just better understanding of the default risk curve. But yes, it is statistics, and there is considerable variation (again, FCs graphs show what the spread can be, even for someone in 100 separate loans). Graphs gone I think (and the scale was always too narrow for me to draw any conclusions) but I have looked at the loan book and I haven't seen any real pattern to it. I think many sellers sell to try and make premium and if its not selling they dump at par to reduce any outsize risk to that particular loan.
|
|
|
|
Post by GSV3MIaC on Aug 3, 2017 10:47:59 GMT
Yep, FC seem to have mislaid their graphs (shows how often I no longer log in), so you'd have to download/analyse the loan book (and it is not easy to tell from that exactly when a load went bad, unless you have tracked them regularly). Looking at seabbs 's graphs, (albeit they are not up to date) seems to confirm what was the case last time I ran a proper analysis (a good while ago now) .. there really was no statistically significant correlation of loans going bad vs risk band (which is not to say that there might be better recovery prospects on A+). And, struggling to recall the details, I don't think term had much effect either .. lots of loans died at 6-9 months whether they were 24, 36 or 60 month loans (there were no very short ones back then I think). The 'sell after 6 months' strategy worked reasonably well, although as mentioned on another thread you can get caught out by defaults even in month 1, so best not to have too many eggs in one basket (for D/Es it might be 'sell after 27 days'). I think I'll stick with secured loans though (in the early days FC had some secured on major capital equipment, like cranes and suchlike, which were quite appealing at the (variable) rates then on offer). |
|
seabbs
P2P Blogger
Posts: 41
Likes: 22
|
Post by seabbs on Aug 3, 2017 11:32:25 GMT
Hi there, The plots exclude recent data as loan volume has increased with time, if you don't account for this then you get a massive bias towards defaults in the first 12 months from origination. If you fancy a play for yourself all plots were generated using this: www.seabbs.co.uk/shiny/fcdashboardFrom evaluating the plots there is definitely evidence that risk of default increases based on risk band (roughly in line with FC's predictions) - the repayments over time graphs are normalised within credit band. I agree the risk of default does increase after 3-6 months, there does appear to be an ongoing relatively stable risk of default after this point but it gets difficult to estimate due to a lack of data. It isn't clear to me that a single 48 month loan is worse/better than two 24 month loans for example. Something to look at further. I am planning on looking at feature importance for defaults shortly so I will get back to you with some more hard data. |
|
|
|
Post by wayne12 on Aug 5, 2017 20:04:10 GMT
Hey ratrace, any advice for a newbie who has yet to reach his all important 100 loans lent mark? Hi wayne Yes if you wish to invest in FC for the long term l would offer the following advice. 1 Manual bidding and been willing to put in the time and effort to understand about investing in FC is a must. 2 Over the long term look to get your max holding in any loan down to around 0.5%. 3 lf you hope to get a much better return then the average,then you will have to be willing to put most/all of your money into B to E loans. 4 With D and E loans its wiser to buy the shorter term loans. 5 Don't overlook the older loans on the SM, because at times there is some very good value on offer. Thanks for the advice, I'll try and stick to it.
|
|
|
|
Post by ratrace on Aug 6, 2017 13:49:55 GMT
Hi wayne Yes if you wish to invest in FC for the long term l would offer the following advice. 1 Manual bidding and been willing to put in the time and effort to understand about investing in FC is a must. 2 Over the long term look to get your max holding in any loan down to around 0.5%. 3 lf you hope to get a much better return then the average,then you will have to be willing to put most/all of your money into B to E loans. 4 With D and E loans its wiser to buy the shorter term loans. 5 Don't overlook the older loans on the SM, because at times there is some very good value on offer. Thanks for the advice, I'll try and stick to it. Thanks Lets us know how you are getting on when you reach your 100 loan part.
|
|
