|
Post by ablrate on May 27, 2015 11:35:10 GMT
We are usually good at communicating on here, so feedback of not updating regularly has been taken on board and I will now be giving a daily update of where we are. 1. Secondary Market - looking good. Waiting on the accrued interest issue on a certain type of trade to be resolved, but we know the solution it is now just being implemented. 2. Data migration - all came into the platform nicely - a little bug that is over crediting interest to some (I am sure you would like us to leave that but .. er .. no ) 3. Some minor CSS issues. 4. Auto - Investment is due to be released on the next version due a few weeks after launch - On this one we will be rolling it out in stages. There are a number of elements to it and it can do some petty cool things but it is complex, so we will be rolling it out in three stages, initially it will be just auto-buying, after that the really awesome stuff will be rolled out after extensive testing on our test server. As we are at this stage with auto-invest I would be very glad to hear any ideas that we could look at... if we implement a brolly and pen are yours! - Do not say we are not generous! ... and no cheating 'automatically invest our money' ... is not an idea.... well it is... but you know what I mean.... We will be emailing everyone on the platform today - but just in case you were wondering about the interest payment due on the 26th - this all went ahead as planned and is all up to date, but we did it on the new platform to test that element, so it will be within the data on your new dashboard, hence why no email was received saying it had been paid. Regards Ablrate
|
|
huxs
Member of DD Central
Posts: 300
Likes: 218
|
Post by huxs on May 27, 2015 12:33:14 GMT
Hi,
Here are some ideas on how I think Auto-invest could work (saying this some may not be relevant depending on the types, throughput and structure of deals on the new platform).
Risk appetite: To be able to specify which type of loans (Aircraft, Machinery etc:) to invest in Maybe take this further and also allow to be able to select sectors (or even countries) that you don't want to auto invest in To be able to set a maximum LTV To be able to set a minimum rate To allow the choice of auto invest in new loans only or SM as well
Diversification: To be able to set a Maximum amount, or percentage of total invested for each loan To be able to limit investing with the same company or further tranche of loans (so I don't end up putting all my investment in say 10 loans but all to the same company) To auto rebalance portfolio if I increase or decrease the funds invested that the system will look to rebalance my loans (by either selling or buying on the SM) to allow me to achieve my diversity targets (I am sure this is very complex but would be pretty cool). The same is true if a new loan came along if I am currently invested in 10 loans but want more diversity then when a new loan comes along you sell parts of my 10 loans to buy into the 11th loan. (Maybe this is more like a manages portfolio where you manage the investment across as wide a set of loans as possible??)
Re-Investment: (require the same as above but also:) To be able to specify to re-invest capital, interest or both automatically
The big thing with auto investing is that you want to know that your money is going to be invested quickly (normally across a diverse set of loans) and that normally requires a steady flow of loans otherwise it maybe better to appraise each loan as it comes up and only move money into Ablrate when loans are listed.
|
|
indy
Member of DD Central
Posts: 57
Likes: 18
|
Post by indy on May 27, 2015 15:47:00 GMT
I personaly always prefer to evaluate each loan and invest manually. Can you make sure there is enough cake left for us before the robots gobble all the loan up. If the loans are large enough then this may not be a problem or maybe we could comit funds pre drawdown before the robots distribute the balance.
Maybe a better way to do it would be to put the loan details on the platform pre drawdown to give us all a chance to evaluate the loan, we could then pay in funds and tell the robot the maximum we would like to invest. As soon as the loan goes live the robot would automatically invest our money, if the loan was over subscribed the system could work out how to divi the loan up in a fair way eg everybody gets 80% of what they wanted. I am sure this is more complicated than it sounds but would be a great system as the platform grows. People soon lose interest if they keep missing out on loans.
|
|
ilmoro
Member of DD Central
'Wondering which of the bu***rs to blame, and watching for pigs on the wing.' - Pink Floyd
Posts: 10,864
Likes: 11,092
|
Post by ilmoro on May 27, 2015 16:35:59 GMT
I personaly always prefer to evaluate each loan and invest manually. Can you make sure there is enough cake left for us before the robots gobble all the loan up. If the loans are large enough then this may not be a problem or maybe we could comit funds pre drawdown before the robots distribute the balance. Maybe a better way to do it would be to put the loan details on the platform pre drawdown to give us all a chance to evaluate the loan, we could then pay in funds and tell the robot the maximum we would like to invest. As soon as the loan goes live the robot would automatically invest our money, if the loan was over subscribed the system could work out how to divi the loan up in a fair way eg everybody gets 80% of what they wanted. I am sure this is more complicated than it sounds but would be a great system as the platform grows. People soon lose interest if they keep missing out on loans. So basically the AC system then
|
|
indy
Member of DD Central
Posts: 57
Likes: 18
|
Post by indy on May 27, 2015 17:11:37 GMT
I personaly always prefer to evaluate each loan and invest manually. Can you make sure there is enough cake left for us before the robots gobble all the loan up. If the loans are large enough then this may not be a problem or maybe we could comit funds pre drawdown before the robots distribute the balance. Maybe a better way to do it would be to put the loan details on the platform pre drawdown to give us all a chance to evaluate the loan, we could then pay in funds and tell the robot the maximum we would like to invest. As soon as the loan goes live the robot would automatically invest our money, if the loan was over subscribed the system could work out how to divi the loan up in a fair way eg everybody gets 80% of what they wanted. I am sure this is more complicated than it sounds but would be a great system as the platform grows. People soon lose interest if they keep missing out on loans. So basically the AC system then I have not used their system, if that is what they do then yes. At least it shows it can be done.
|
|
duck
Member of DD Central
Posts: 2,583
Likes: 5,714
|
Post by duck on May 28, 2015 6:22:46 GMT
I too like to evaluate loans on an individual basis and I make notes on my spreadsheets regarding my 'views' of the risks and if to buy more, keep a small investment, sell at some day in the future etc. That way I 'take responsibility' - which from a platforms position is probably advantageous if a loan turns! With Auto-invest it can be perceived that the decision is more with the platform (not true if the 'system' works correctly but the way the mind works - basically the difference between 'buying' and 'being sold')
That said, I am not against Auto Invest (I use it daily on AC) but it would be useful to have a way to 'reserve' funds in your individual account. When transferring in new funds to cover a new loan Auto Invests can eat into this cash before the new loan is drawn down so you don't end up with the investment you want in the new loan. Perhaps 'reserve' cash until loan draw down and then when the loan has drawn Auto Invest can take over for additions/sales?
|
|
SteveT
Member of DD Central
Posts: 6,873
Likes: 7,918
|
Post by SteveT on May 28, 2015 7:34:43 GMT
I look forward to the flow of new Ablrate loans being great enough to warrant Auto-Invest functionality!
A capability I would like to see (and is missing from the otherwise quite sophisticated AC system) is "Auto-Diversify". This would permit Lenders automatically to invest in new loans by reducing slightly their holdings in existing loans. A simple check-box on each loan could give Lenders the option to "Opt In" or "Opt Out" of Auto-Diversify for that loan, either to ring-fence an existing holding of a loan they particularly like or to prevent investment in a new loan they wish to avoid.
|
|
|
Post by ablrateandy on May 28, 2015 8:30:36 GMT
Thanks for all of the thoughts so far and keep them coming! Pretty much all of the above is already in The Plan. To get it right will take a lot of work, but one of the reasons for the slight delay now is making sure that the building blocks were in place to start off with - which hopefully they now are. A lot of the Autoinvest stuff revolves around the secondary market, which is going to be different on our platform from anyone else's. Making sure that the allocations are fair is of course the most difficult part, but we have the algorithms worked out and it is just a matter of applying them correctly within the application. There's a lot of debate in the P2B and P2P sector around how investors are treated and AI has to be agnostic in its treatment.
I am hoping that robots won't benefit anyone anyway if we build this correctly, and we would block them where we can. AI is designed so that it does exactly what you tell it to do within a set of parameters and I'm pretty confident that AI would beat any manual investor or robot in achieving its goals.
But first, let's get the base application out there for everyone to have a look at. I need to get my video tutorials done before the boss gets in and starts making noise in the background!
|
|
SteveT
Member of DD Central
Posts: 6,873
Likes: 7,918
|
Post by SteveT on May 28, 2015 9:24:30 GMT
Thanks for all of the thoughts so far and keep them coming! Pretty much all of the above is already in The Plan. To get it right will take a lot of work, but one of the reasons for the slight delay now is making sure that the building blocks were in place to start off with - which hopefully they now are. A lot of the Autoinvest stuff revolves around the secondary market, which is going to be different on our platform from anyone else's. Making sure that the allocations are fair is of course the most difficult part, but we have the algorithms worked out and it is just a matter of applying them correctly within the application. There's a lot of debate in the P2B and P2P sector around how investors are treated and AI has to be agnostic in its treatment. I am hoping that robots won't benefit anyone anyway if we build this correctly, and we would block them where we can. AI is designed so that it does exactly what you tell it to do within a set of parameters and I'm pretty confident that AI would beat any manual investor or robot in achieving its goals. But first, let's get the base application out there for everyone to have a look at. I need to get my video tutorials done before the boss gets in and starts making noise in the background! That sounds encouraging. The fundamental first step is to ensure that the new secondary market is highly liquid and dynamic which, almost by definition, means that loan parts need to be offered and purchased at par, without premium or transaction fee. A highly liquid SM encourages lenders to take bigger initial stakes (acting as surrogate underwriters) in the confidence they'll be able to release / reallocate funds later. Only FC to date (as far as I'm aware) have achieved an active SM that supports seller premiums and charges a selling fee but this is down to: a) a very high deal flow (typically 10+ new loans per day) b) having many thousands of active users c) the dominance of "flippers" who invest their personal time (and/or programming expertise, for those who use bots) to steal a march on the average investor and then sell loan parts to them for a margin Unless Ablrate is confident of a) and b) and happy to encourage c), I would say that the models of AC and SS are the ones to take a lead from. Both have modest deal flows (1 - 2 deals per week), smaller user bases but enjoy highly liquid SMs. This allows new loans to be filled and drawn-down very quickly whilst permitting lenders to acquire loan parts in existing loans without cost or penalty. Examples where the opposite is the case include ReBS where, despite a decent flow of new loans, the fundamental illiquidity of their SM prevents most lenders investing more than they're happy to hold to term. This means that larger loans fail to fill (or fill very slowly at a high average rate) which presumably must be a major impediment to growing the business. If ReBS did away with seller premiums and eliminated their selling fee I suspect they'd find it easy to attract much more money into the platform and compete effectively for new loans.
|
|
TFTO
Member of DD Central
Posts: 143
Likes: 71
|
Post by TFTO on May 28, 2015 9:57:25 GMT
I would prefer not to have anything to do with robots as I like to have a good look at the loan before investing. I never use the secondary market unless it is to sell at a profit as on FC.
As for Assetz, I have been slowly selling my holdings on there as I cannot get along with a business model which only allows investment using a robot on the SM.
Each to their own but as long as you allow me to invest in the primary market then that's fine.
|
|
|
Post by ablrate on May 28, 2015 12:35:37 GMT
Hi All
Thanks for your responses so far. Some great ideas and some validation of what we are already doing. To allay some concerns about AI; you will essentially have two separate parts of your account, an AI account and a manual account. Anything in AI will beaver away doing what you have programmed it to do (with opt outs) and all others will be manually invested. If you don't want to do any AI investing, no problem, leave your money in the manual account and all is good. We'll keep some of the other functions close to our chest right now..
Update on the new site later today when we receive today's reports from developers.
Regards Ablrate
|
|
james
Posts: 2,205
Likes: 955
|
Post by james on May 28, 2015 13:34:37 GMT
As we are at this stage with auto-invest I would be very glad to hear any ideas that we could look at... if we implement a brolly and pen are yours! - Do not say we are not generous! ... and no cheating 'automatically invest our money' ... is not an idea.... well it is... but you know what I mean.... One thing I haven't seen elsewhere is auto-invest with an undo capability, say capped at a maximum of 48 hours from the automated buy order in "round one". Then a "final round" with no undo to fill the loan after anyone has backed out who wants to. Auto-invest then settable to round one only for those who want to review everything. The 48 hour time limit and final round" with no backout provide a definite end time for the deal. Maybe discuss time limits a bit with those who do want to review everything to see what they need, while also bearing in mind the borrower's desire for rapid turnaround. "Fast" loans with no backout time could be offered but I hope that a limit like 48 hours would be short enough to discourage that. The undo capability should help to deal with an issue I've seen elsewhere where those who want to review get no loans because automatic bidding fills them all. Lets those who want to review participate more fully. Criteria for auto-invest would probably need to allow for things like "no SPV lending" and "no non-UK investing" for those who don't like those things. Of course people doing this would see lower deal flow and slower lending but it beats not lending at all. At some point you're going to see lender demand exceeding loan supply and need to come up with some way to ration lending. The best sort of approach I've seen described for this is something along the lines of a queue and £5 maximum first round, then £5 more on repeated subsequent rounds until all auto-bid limits have been reached, if ever. A queue position having a value of perhaps £100 or £250 worth of lending, position not lost until that much has been lent. This amount is set to cause those with shallower pockets to lend at something reasonably close to the levels of those with deep pockets and high per-loan amounts, otherwise those lending a lot per loan can do far more lending than those doing less. A queue position value of something around the mean amount per lender on all loans could be about right. The lowish per-round limit is to maximise lender diversification and also keep lending speed up - people are happier with regular small amounts than longer time gaps and bigger amounts. £5 is chosen to maximise the "I did something" rate, it might be a bit small for this platform, the important point is multiple rounds and relatively low amounts per round so there's regular action and good diversification. £50 is probably the maximum desirable per round for a cautious person, even that's a bit on the high side for a person who is looking to invest only a few thousand Pounds and wants to do it with no more than 5% per loan exposure. Next up is keeping money invested. At some point there will be lenders with enough money being returned that they can't lend it out fast enough using the normal limits. This will affect the bigger lenders. Allocating the first 25% of a loan to a queue for those with "old money" and running through that queue can help to address this, while not unduly slowing things down for newcomers, I hope. The 25% can be adjusted based on the rate of old money and new money inflow to accounts, so that overa month it's expected that all old money will be invested, provided say at least 50% of borrower demand still goes to the all money market. This is a sort of halway house between all old money first and a higher impact on new lenders and no priority at all for old money that will frustrate those who want to keep their money invested. Loans have different properties, like duration and estimated credit risk. Queues need to be independent for each of these. It they aren't what happens is that those who are willing to lend to say an unpopular 60 month term end up getting almost only the 60 month loans and nothing shorter because they always lose their queue position the moment a 60 month loan comes in and are at the back of it by the time shorter loans come in. This means lots of independent queues for you to manage but keeps the diversification between terms and credit quality for the lenders. Few people would be happy being willing to lend for 12-60 months but having the queue system in effect limit them to just 60 months. Yes, it really does happen this way if one queue is used - it's based on experience with the Bondora queue system. One queue has a manipulation risk as well. Compare a person who sets terms or risk levels, say allowing 12-60 months or low to high credit risk and range of interest rates with another person who sets lowest credit risk and highest interest rate with most desirable term. Because so few loans match the "perfect" criteria of the second person they will always be at the front of the queue for them and can use the queue system to stuff their portfolio with only the best loans, without paying much of a price in lending speed. With multiple queues they can still use the narrow criteria but will pay the price in reduced speed. You probably already guessed, this is based on experience with the Bondora systems and how to use and abuse them. A cap on reselling within 90 days is desirable as well. Say no more than 5% of total loan value acquired each month, counting only loans that are on time or still in their time before initial payment towards the limit. This is to restrict flipping and favour longer-term investing. In a shortage of borrower situation the premium for loans on a secondary market can be significant so flipping can become quite attractive. Based on experience of flipping and the Bondora markets. The time limit increases the time for which loans must be held before they can be flipped, increasing the capital required to do it. 90 days means you're exposed to credit risk and can't just do low risk buy and flip quickly before first payment flipping. Unlimited reselling if a person then accepts no more buys for 90 days, so people can exit the platform if desired without circumventing the limits. Those with the capital may still engage in flipping so more controls might be needed, but these should do as a start. Limits per borrower, per loan, per SPV, per ultimate customer (lessee of a plane, say) and per middle customer (lessor, say). For diversification and exposure control. Do those things and you should avoid the pain points that I've seen happen on the assorted platforms I've looked at.
|
|
james
Posts: 2,205
Likes: 955
|
Post by james on May 28, 2015 13:38:37 GMT
For all, what sorts of things that you've seen in other places could be classed as abuse of systems, other lenders or borrowers? This is so that Ablrate can try to design in controls to make such things harder to do.
One anti-abuse tool is on a secondary market where a seller sets a price based on loan condition at the time. Then a payment comes in or something else happens to increase the value of a loan and the deal is accepted at the old price before the seller can change their offer. happens with impaired loans being offered at a capital discount to give a suitable return to the buyer, then it's abused by buyers getting the low price once the impairment is much reduced or gone. Options for automatically cancelling or suspending sell offers after a change or for review of deals by seller before final completion can limit the scope for this one, with say automatic completion if not confirmed or rejected within say 48 hours.
Just to give some idea of how bad this can get, one person analysed sales of discounted loans at Bondora and found that something over 50% of all such sales were made just after a payment had been made to change a loan from late to on time. The loans then often rapidly show up on the secondary market again, this time with a premium or discount reflecting the new facts.
|
|
james
Posts: 2,205
Likes: 955
|
Post by james on May 28, 2015 13:51:20 GMT
Only FC to date (as far as I'm aware) have achieved an active SM that supports seller premiums and charges a selling fee but this is down to: a) a very high deal flow (typically 10+ new loans per day) b) having many thousands of active users c) the dominance of "flippers" who invest their personal time (and/or programming expertise, for those who use bots) to steal a march on the average investor and then sell loan parts to them for a margin Bondora's is pretty active but flippers are a massive issue for those trying to set a fair price for impaired loans, with the flippers buying as soon as a payment is made. that makes it very hard to set a fair price for the impaired state without just giving more money tot he flippers and in turn that discourages byers who may not get the right premium for the impairment. The other major issue at Bondora is very poor secondary market search tools, in particular no way to weed out the chronically horrible sale offers that dominate the market. Few people will want to buy a loan at say a -20% expected return but you still have to wade through those looking for the worthwhile ones. Unless Ablrate is confident of a) and b) and happy to encourage c), I would say that the models of AC and SS are the ones to take a lead from. Both have modest deal flows (1 - 2 deals per week), smaller user bases but enjoy highly liquid SMs. This allows new loans to be filled and drawn-down very quickly whilst permitting lenders to acquire loan parts in existing loans without cost or penalty. I think it's possible to implement restrictions that can limit the scope for flipping and have just suggested some. More thoughts on how to achieve it would be welcome, to see if we can come up with a design that favours investing over flipping and allows investors to not have to be watching 24/7/365 to avoid being exploited. Examples where the opposite is the case include ReBS where, despite a decent flow of new loans, the fundamental illiquidity of their SM prevents most lenders investing more than they're happy to hold to term. This means that larger loans fail to fill (or fill very slowly at a high average rate) which presumably must be a major impediment to growing the business. If ReBS did away with seller premiums and eliminated their selling fee I suspect they'd find it easy to attract much more money into the platform and compete effectively for new loans. I think that in the case of Ablrate the institutional lenders acting as backup for retail will take care of this loans not filling issue. If it turns out that they don't, another round without restrictions on bids could be held before concluding that there is not lender demand for the loan at the terms offered.
|
|
|
Post by ablrate on May 28, 2015 14:24:38 GMT
wow... awesome guys... you really want that brolly and pen! lol..
Thanks guys I will review and respond to you all... thanks so much, keep them coming!
Regards Ablrate
|
|