|
|
Post by chris on Jun 4, 2014 10:13:49 GMT
Chris, Would you ever consider revealing the size the of the queue? i.e. how much is in front of me? Some might want to chose loans with shorter queues. From an IT perspective I have no issue (other than on screen clutter) but that would be a business decision.
|
|
|
|
Post by jackpease on Jun 4, 2014 10:16:01 GMT
Am i missing the point here? The reason there is excess demand over supply is because of Assetz fixed interest rates - just as communist price controls on staple foods etc leads to shortages, queues and rationing, so do 'interest rate controls' lead to rationing and the ensuing bloodbath no-win discussion about 'fairness' which will be a poison chalice for Chris etc.
Why not auction as most others? (And the shortage of loan parts is an obvious side effect to not being able to sell at a premium)
I've said before that borrowers on Assetz seeing the scramble to lend must see that the excess demand suggests the interest rate could be lower. And if there were lower interest rates, there might be more loans coming through.
Jack
|
|
|
|
Post by chris on Jun 4, 2014 10:23:02 GMT
You don't say if when the purchaser is selected as the "winner" they will be able to buy everything available if their mandate allows. I still think it would be fairer to restrict the maximum that can be bought in one transaction if the supply is scarce. You either have to have a permanent restriction or no restriction otherwise the system becomes unfair. For example you have a loan listed on the AM and there is just £2000 available I reach the front of the but because of the 'scarcity' instead of the £1000 share I want I only get £100 worth and I am sent to the back of the queue. However just as this happens an underwriter releases £500,000 onto the AM meaning that the loan has ceased to be scarce and the next person in the queue gets the full £10,000 worth they are looking for. I perceive this has highly unfair as I get less than someone who has been in the queue for a shorter time than me and I have to work my way round to the front of the queue to get the remaining £900 I want potentially for the whole scenario to repeat itself. That's partly why I think a queue is unfair. But if it's £100 chunks with a weighted random lottery then your absolute chance with any one loan unit never drops to zero the market is just skewed to try and give some semblance of fairness with factors that constantly try and rebalance. The problem with full fulfilment of any one persons order is that there will be lenders on the platform looking to invest >£100k per loan, some even into the millions per loan. We already have several lenders with multi-million pound portfolios and that number will go up not down in time. Should a single lender be able to snap up that entire £500k release you talk about before anyone else gets a shot just because they're at the head of the queue? I'd much prefer large investors like that to get a bit of a weighting in their favour to help them deploy their cash but that it is more balanced amongst all our lenders. Don't forget there'll always be someone with deeper pockets than you so it's in your interests that I find a fair system!
|
|
|
|
Post by chris on Jun 4, 2014 10:34:09 GMT
Am i missing the point here? The reason there is excess demand over supply is because of Assetz fixed interest rates - just as communist price controls on staple foods etc leads to shortages, queues and rationing, so do 'interest rate controls' lead to rationing and the ensuing bloodbath no-win discussion about 'fairness' which will be a poison chalice for Chris etc. Why not auction as most others? (And the shortage of loan parts is an obvious side effect to not being able to sell at a premium) I've said before that borrowers on Assetz seeing the scramble to lend must see that the excess demand suggests the interest rate could be lower. And if there were lower interest rates, there might be more loans coming through. Jack We fundamentally disagree with other platforms policy of pricing to market liquidity rather than risk. FC have already had to impose minimum rates on their markets due to the mess that pricing to liquidity creates, and it's fundamentally unfair to expect all our lenders to create their own credit models and understand what a fair minimum rate should be for their loans. A provision fund can help address that to a degree but again we feel they are inadequate in their protection of lender capital. None of the P2P provision funds are anywhere near the size (percentage of loan book wise) even held by the banks at the beginning of the credit crunch and we all know how that worked out. Every lender on our platform will get a fair rate of return for the risk they are taking, without pressure to take undue risks just to get their money deployed. Being IT I probably haven't explained that as well as Andy Holgate could but hopefully you get my point. Edit: I should add that there will be an adjustment of rates over time as our current loans are fairly niche. As we broaden our appeal to more mainstream loans then inevitably rates will fall accordingly as the risks vs security will be different. The market balance between supply and demand will dictate where we focus our efforts, so lenders will be able to choose their rates in that respect.
|
|
|
|
Post by Come_on_Grandad on Jun 4, 2014 10:35:46 GMT
I agree that a queue system, however you construct it, could leave someone perceiving that they are being disadvantaged. I think it may be easier to construct a weighted lottery that is widely perceived to be fair. The weighting should be the minimum of - your available cash balance at that instant
- the amount your mandate for this loan says that you want to buy
If you hold the winning ticket, you buy all that you want. Any remainder is offered to the runner-up ticket, etc The problem with a weighted system particularly if the weighting includes investment activity is that it becomes a disincentive to investing in the more unpopular loans with the potential result that Underwriters are left holding large chunks of loans that nobody wants. After all why would I want to purchase £20 worth of a loan that I perceive to have a high risk if it is going to put me toward the back of the queue for a loan which I perceive to have a low risk and that I potentially want to invest £1000 in. You are still talking queue. What I mean is that, as each seller appears, a new lottery is held. Your previous history of success or failure is not relevant. Your £20 purchase in loan A only affects your chances in loan B, moments later, to the extent that your cash balance is reduced by £20.
|
|
|
|
Post by chris on Jun 4, 2014 10:43:00 GMT
I agree that a queue system, however you construct it, could leave someone perceiving that they are being disadvantaged. I think it may be easier to construct a weighted lottery that is widely perceived to be fair. The weighting should be the minimum of - your available cash balance at that instant
- the amount your mandate for this loan says that you want to buy
If you hold the winning ticket, you buy all that you want. Any remainder is offered to the runner-up ticket, etc The problem with a weighted system particularly if the weighting includes investment activity is that it becomes a disincentive to investing in the more unpopular loans with the potential result that Underwriters are left holding large chunks of loans that nobody wants. After all why would I want to purchase £20 worth of a loan that I perceive to have a high risk if it is going to put me toward the back of the queue for a loan which I perceive to have a low risk and that I potentially want to invest £1000 in. I can see your point but again your chances will never reduce to zero. If this behaviour does become prevalent then we'd need to consider alternatives to rebalance the market as that's certainly not desired behaviour but I don't think it will pan out that way.
|
|
markr
Member of DD Central
Posts: 766 
Likes: 426
|
Post by markr on Jun 4, 2014 11:05:52 GMT
You don't say if when the purchaser is selected as the "winner" they will be able to buy everything available if their mandate allows. I still think it would be fairer to restrict the maximum that can be bought in one transaction if the supply is scarce. You either have to have a permanent restriction or no restriction otherwise the system becomes unfair. Yes, sorry, badly worded. The restriction would be permanent but only relevant when supply is scarce. When supply is plentiful the queue would zip round at breakneck speed filling everyone's requests one chunk at a time (conceptually, anyway, what would really happen is that the queue would be locked and the computer would calculate how the parts would be distributed according to the queuing algorithm and then dish out the allocations. From the users point of view, they'd have asked for, say, £10000 and got it or, if there's not enough for everyone, they got, say, £3700 which is their "share" of what's available).
|
|
j
Member of DD Central
Penguins are very misunderstood!
Posts: 2,188
Likes: 540
|
Post by j on Jun 4, 2014 11:19:42 GMT
You either have to have a permanent restriction or no restriction otherwise the system becomes unfair. Yes, sorry, badly worded. The restriction would be permanent but only relevant when supply is scarce. When supply is plentiful the queue would zip round at breakneck speed filling everyone's requests one chunk at a time (conceptually, anyway, what would really happen is that the queue would be locked and the computer would calculate how the parts would be distributed according to the queuing algorithm and then dish out the allocations. From the users point of view, they'd have asked for, say, £10000 and got it or, if there's not enough for everyone, they got, say, £3700 which is their "share" of what's available). Are you planning to do a test run live, rather then behind the scenes, to see how well/unwell the new system would perform? I am 'very' slowly warming up to the idea as long as there is no discrimination that if I got hold of units in loan A, I still have the same chance in loan B 2 hours later when it comes on AM (i.e. obtaining a part in one loan should have no material effect on the same member's chances of getting part/s in another, even if they ran simultaneously or very close, timing-wise, to each other)
|
|
|
|
Post by chris on Jun 4, 2014 11:47:14 GMT
Yes, sorry, badly worded. The restriction would be permanent but only relevant when supply is scarce. When supply is plentiful the queue would zip round at breakneck speed filling everyone's requests one chunk at a time (conceptually, anyway, what would really happen is that the queue would be locked and the computer would calculate how the parts would be distributed according to the queuing algorithm and then dish out the allocations. From the users point of view, they'd have asked for, say, £10000 and got it or, if there's not enough for everyone, they got, say, £3700 which is their "share" of what's available). Are you planning to do a test run live, rather then behind the scenes, to see how well/unwell the new system would perform? I am 'very' slowly warming up to the idea as long as there is no discrimination that if I got hold of units in loan A, I still have the same chance in loan B 2 hours later when it comes on AM (i.e. obtaining a part in one loan should have no material effect on the same member's chances of getting part/s in another, even if they ran simultaneously or very close, timing-wise, to each other) Currently considering the options but one option I'm considering is coding the new queue rules sooner rather than later and putting them on the live site before the big release we're planning. We'd have a week or so of live running that way with which to see how it's working out, giving us the opportunity to fine tune any weightings etc. As planned though there is an effect if you get loan units in loan A and then some in loan B come up. There has to be some kind of balance between those choosing to invest all their cash in 3 or 4 loans vs those that stick funds into 100 loans. The person putting funds into 100 loans with less discretion (a perfectly valid strategy) should not be able to deploy their funds 25 odd times more quickly than someone following a different strategy - using a generalisation that all loans are sold with equal regularity. I'd like both strategies to be as equally valid on the platform as possible.
|
|
|
|
Post by batchoy on Jun 4, 2014 11:50:36 GMT
The problem with a weighted system particularly if the weighting includes investment activity is that it becomes a disincentive to investing in the more unpopular loans with the potential result that Underwriters are left holding large chunks of loans that nobody wants. After all why would I want to purchase £20 worth of a loan that I perceive to have a high risk if it is going to put me toward the back of the queue for a loan which I perceive to have a low risk and that I potentially want to invest £1000 in. You are still talking queue. What I mean is that, as each seller appears, a new lottery is held. Your previous history of success or failure is not relevant. Your £20 purchase in loan A only affects your chances in loan B, moments later, to the extent that your cash balance is reduced by £20. The problem with having a lottery every time something is put up for sale is that you are generating a awful lot of lotteries and your chance of picking up anything becomes totally random. At least with a queue against each loan you stand chance of moving forward with some chance of picking up something at some point and if the queues are published in the same way as pre-bids you have some visualisation of where you are in the queue and the level of funds that are in front of you in the queue be it 5000 lenders each with £100 requirements or one lender with a £500,000 requirement. As previously stated for a system wide AI system rather than the current loan by loan system, I have no problem with the use of a lottery to randomly generate a new queue for a new loan from those lenders whose AI configuration matches the loan but what I am against is weighting that queue since as soon as you start weighting you start creating distortions which can be perceived as unfairnesses, and which people can start to play in order give themselves an advantage. One example of this is the on the Bondora system where they have changed the way the queues are generated for Profiles (read AI settings for AC). As described the weighting system was supposed to favour older over new profiles, however experience seems to suggest that they may have got it wrong and newer profiles are being favoured over older ones thus people can potentially manipulate their position in the queue by chopping and changing their profiles on a daily basis.
|
|
|
|
Post by chris on Jun 4, 2014 12:04:31 GMT
batchoy - hopefully bugs in other systems don't disprove the model or overall approach. With a system wide weighted lottery you are moving forward in the queue each time anyone on any loan successfully buys a loan unit as your odds of winning the next lottery will be greater than theirs. There are too many problems with queues that simply haven't been addressed by anyone on here - I can list them all again if needed.
|
|
|
|
Post by bracknellboy on Jun 4, 2014 12:06:38 GMT
Am i missing the point here? The reason there is excess demand over supply is because of Assetz fixed interest rates - just as communist price controls on staple foods etc leads to shortages, queues and rationing, so do 'interest rate controls' lead to rationing and the ensuing bloodbath no-win discussion about 'fairness' which will be a poison chalice for Chris etc. Why not auction as most others? (And the shortage of loan parts is an obvious side effect to not being able to sell at a premium) I've said before that borrowers on Assetz seeing the scramble to lend must see that the excess demand suggests the interest rate could be lower. And if there were lower interest rates, there might be more loans coming through. Jack Personally I like Assetz approach of using their skills to price to risk rather than liquidity. Of course they may have their judgement wrong, but I expect its a lot better than mine.
|
|
merlin
Minor shareholder in Assetz and many other companies.
Posts: 902
Likes: 302
|
Post by merlin on Jun 4, 2014 14:33:41 GMT
I have been fascinated watching the debate between Chris and few of the more enlightened posters on the shape and functionality of the future AC system. However I wonder how many existing and potential investors in AC have even a basic working knowledge of systems theory, queuing theory and game theory, all of which seem to have been under discussion on this thread. My guess is not very many and perhaps not many more than I have fingers on my two hands. This then may also be the sum total of investors in AC if such knowledge will be required to be able to invest wisely in AC loans in the future. KISS
|
|
|
|
Post by Ton ⓉⓞⓃ on Jun 4, 2014 20:37:36 GMT
batchoy - hopefully bugs in other systems don't disprove the model or overall approach. With a system wide weighted lottery you are moving forward in the queue each time anyone on any loan successfully buys a loan unit as your odds of winning the next lottery will be greater than theirs. There are too many problems with queues that simply haven't been addressed by anyone on here - I can list them all again if needed. In effect I'm sure that we've been having a test of AI since it launched, and chris has been noting how it's worked. So yes keep doing tests. I've the impression that for certain people, despite having mandates set up, the present system hasn't gotten them any stock to speak of at all. It's just about failed for them, Are there certain Lenders it's failed for and why has it failed for them and how will the new system be better?
|
|
|
|
Post by Ton ⓉⓞⓃ on Jun 4, 2014 20:47:36 GMT
I've not had much sleep, so some of this might seem to come from a dream-world.
Would it help if the Primary Market worked in a way that benefited one group way and the secondary worked to benefit those who were put at a disadvantaged by that.
Would it be a silly idea to have the primary market was slanted to bigger deals and the AM aimed at smaller deals.
Or have it that for particular loans they are aimed at those invested with below 50K in AC, and other loans aimed at other types of Lender. So that no particular game would win out overall in AC but that all games are allowed, but that some games were better suited to different loans. Just so long as it was clear what game would work best in which loan.
At that point I woke up and can't remember anything else...
|
|
