jlend
Member of DD Central
Posts: 1,817
Likes: 1,444
|
Post by jlend on Jun 16, 2015 15:44:11 GMT
There is mention of money at YR being stranded,albeit not often. How often does this occur? What is the value of this? Are there stats on how many times currently YR is set below MR ? Has any modelling shown what the impact of the change might be? The change to MR. What impact have you assumed this would have had on the MR,if it had been in place over the last few months ? What impact do you think the change to YR and MR will have on the MR combined? I assume this will be difficult to assess given the two changes will happen at the same time. Have you thought about staggering the changes,perhaps doing the MR one first ? Around 20% of money on the markets has a YR set. This doesn't nessarily mean it's stranded as when we speak to customers they are happy to leave the money "resting" at the higher rate. Although this is a strange response (in my opinion) because if RateSetter had done more loans on the day their money would have gone. Around 2,000 lenders have a YR set below the MR We've done no modelling on the impact, remembering we are doing these as the right thing to do rather than looking for an outcome. Either combined (as they are being changed for independent reasons) or individually. I do not know if the MR will stay flat (probably) or increase/decrease. I personally think it could lower the MR slightly as the average matched should be lower than that offered. However the previous method was dependent on how many active lenders set a low offer at an ungodly hour vs customers completing on loans. This strange situation could cause unwarranted movements upwards or downwards. Hopefully the new SONIA style will bring stability and accountability Brand/Marketing wanted to stagger the change, but it was easier technically to make the changes in one go as we have scheduled releases to IT (we've grown, it isn't as immediate as it was in the golden days). Besides, big bang is always better in the long run.... @ westonkevRSwww.linkedin.com/profile/view?id=19236219Thanks for the prompt reply as always. Very interesting, some responses were not what i expected. I have been a great fan and lender on ratesetter since i signed up in 2010. At the moment i do feel i will miss the current YR. Setting a relatively high YR rate, having the money sit on the market for a few days sometimes,but then getting lent out automatically during the peak times/days. With the re assurance that the money will be placed at the MR if i happen to set YR too low for the peak rate days of the month. I am sure i will get use to the new ways of working as i like the provision fund above all else. I would have liked the option of keeping the old YR. But i do appreciate the drive to keep things simple. At the moment it feels like i will end up spending more time manually setting rates every day until i see how the markets behave for a month, either up or down, where the rates settle, and how much volatily there is. Then after a month i will have forgotten the old YR ever existed ☺
|
|
pikestaff
Member of DD Central
Posts: 2,136
Likes: 1,484
|
Post by pikestaff on Jun 16, 2015 15:51:47 GMT
Thanks westonkevRS, I look forward to an explanation of how the 2dp MR works. [...] westonkevRS's answer is mostly right but let me add a bit more based on my experience and some detective work by others. p2pindependentforum.com/thread/1940/3yr-market-4-on-sides. NORMALLY, both YR and MR offers are offers made by lenders, which borrowers may or may not accept. Where this is so, then except in VERY unusual circumstances which I shall describe later, all such offers correspond to a single borrower rate which is the rate that, compounded monthly, gives the stated lender rate. For example, for loans entered into since lender fees were abolished: - all 5.9% loans have a contract rate of 5.75%, which is the minimum rate needed to round down, after compounding, to 5.90%. - all 6.0% loans have a contract rate of 5.85%, which is the minimum rate needed to round down, after compounding, to 6.00%. This means that, normally, only a subset of all possible 2dp rates will arise. This relationship also holds true for older loans with a 10% fee, if the fee is deducted before compounding. HOWEVER, it works differently if you choose to match an existing request in the borrower queue. In this case the borrower request is not limited to the subset of rates described above. Back in January, I tested this by matching a borrower request that was displaying at 5.6% in the 5 year market and it came back with a contract rate of 5.54%, which corresponds to an annualised rate of (1+5.54%/12)^12-1 = 5.68%. I believe that if you match a borrower request that is displayed as x.x% you will never get less than the annualised rate you have asked for, but you may get up to 0.09% more. A further complication arises if you offer more than the borrower wants. For example, suppose the borrower request in my last example had been for £1,000 and was from a single borrower. If I had offered (say) £10,000 to "match" this request then it appears that the system would have treated this as an offer of £10,000 at 5.68% (matching the rate requested by the borrower). The £9,000 excess would have gone into the lending queue at 5.68%, but displayed as 5.6%. In my view this is a bug and the unmatched portion of the lender's offer should go into the queue at the exact rate stated (5.6% in this case). VERY rarely, it appears that such an unmatched offer has fed into the determination of the next day's MR, with the result that the MR has not been limited to the subset of rates that I described earlier. I can't say for sure how rare this is, but I doubt that it's more than once or twice a year. In a sample of 50 of my loans (made on different days) I did not find one. westonkevRS, With the new system coming in, which calculates MR as an average, it will be important to avoid this kind of problem. Unless RS moves to displaying rates to 2dp (which I do not advocate) the developers need to make sure that the day's MR is always rounded to x.x0%, otherwise there will be a lot more apparent queue jumping. For example if the MR was calculated at 6.16% and not rounded, but displayed at 6.1% (the current display behaviour) then subsequent offers made at 6.1% would appear to jump the queue. I would round the calcuated MR to the nearest 0.1% to avoid any systematic bias, so that in this example the MR for the day would be set at exactly 6.2%.
|
|
|
Post by wibble on Jun 16, 2015 16:07:05 GMT
Westonkev
I'm sure analysis has been performed to compare how the old Market Rate system compares to the new Market Rate system over the past x months - any chance you can share that? (for example, a graph depicting the last x months, of old system vs new system) I hope that makes sense.
|
|
jlend
Member of DD Central
Posts: 1,817
Likes: 1,444
|
Post by jlend on Jun 16, 2015 16:36:37 GMT
Westonkev I'm sure analysis has been performed to compare how the old Market Rate system compares to the new Market Rate system over the past x months - any chance you can share that? (for example, a graph depicting the last x months, of old system vs new system) I hope that makes sense. Good question. I think they have not done any modelling based on a reply westonkev posted this afternoon unless i miss understood the reply to my post. But perhaps they could for us so we can get an idea about what might happen. Albeit we can't really guess what the long term effect will be.
|
|
jlend
Member of DD Central
Posts: 1,817
Likes: 1,444
|
Post by jlend on Jun 16, 2015 16:39:45 GMT
Thanks westonkevRS, I look forward to an explanation of how the 2dp MR works. [...] westonkevRS's answer is mostly right but let me add a bit more based on my experience and some detective work by others. p2pindependentforum.com/thread/1940/3yr-market-4-on-sides. NORMALLY, both YR and MR offers are offers made by lenders, which borrowers may or may not accept. Where this is so, then except in VERY unusual circumstances which I shall describe later, all such offers correspond to a single borrower rate which is the rate that, compounded monthly, gives the stated lender rate. For example, for loans entered into since lender fees were abolished: - all 5.9% loans have a contract rate of 5.75%, which is the minimum rate needed to round down, after compounding, to 5.90%. - all 6.0% loans have a contract rate of 5.85%, which is the minimum rate needed to round down, after compounding, to 6.00%. This means that, normally, only a subset of all possible 2dp rates will arise. This relationship also holds true for older loans with a 10% fee, if the fee is deducted before compounding. HOWEVER, it works differently if you choose to match an existing request in the borrower queue. In this case the borrower request is not limited to the subset of rates described above. Back in January, I tested this by matching a borrower request that was displaying at 5.6% in the 5 year market and it came back with a contract rate of 5.54%, which corresponds to an annualised rate of (1+5.54%/12)^12-1 = 5.68%. I believe that if you match a borrower request that is displayed as x.x% you will never get less than the annualised rate you have asked for, but you may get up to 0.09% more. A further complication arises if you offer more than the borrower wants. For example, suppose the borrower request in my last example had been for £1,000 and was from a single borrower. If I had offered (say) £10,000 to "match" this request then it appears that the system would have treated this as an offer of £10,000 at 5.68% (matching the rate requested by the borrower). The £9,000 excess would have gone into the lending queue at 5.68%, but displayed as 5.6%. In my view this is a bug and the unmatched portion of the lender's offer should go into the queue at the exact rate stated (5.6% in this case). VERY rarely, it appears that such an unmatched offer has fed into the determination of the next day's MR, with the result that the MR has not been limited to the subset of rates that I described earlier. I can't say for sure how rare this is, but I doubt that it's more than once or twice a year. In a sample of 50 of my loans (made on different days) I did not find one. westonkevRS, With the new system coming in, which calculates MR as an average, it will be important to avoid this kind of problem. Unless RS moves to displaying rates to 2dp (which I do not advocate) the developers need to make sure that the day's MR is always rounded to x.x0%, otherwise there will be a lot more apparent queue jumping. For example if the MR was calculated at 6.16% and not rounded, but displayed at 6.1% (the current display behaviour) then subsequent offers made at 6.1% would appear to jump the queue. I would round the calcuated MR to the nearest 0.1% to avoid any systematic bias, so that in this example the MR for the day would be set at exactly 6.2%. Seriously... i am glad there are lenders like pikestaff who look at these sort of things. Thank you
|
|
jlend
Member of DD Central
Posts: 1,817
Likes: 1,444
|
Post by jlend on Jun 16, 2015 16:45:22 GMT
Around 20% of money on the markets has a YR set. Around 2,000 lenders have a YR set below the MR Kev, thanks for sharing the facts. If I understand this correctly, this means about 4500 lenders have a YR set and of those, nearly half have it set at below MR or are you including accrued repayments that may not have gone out in months as you refered to "money" as opposed to "lenders"? Good question. We get annoyed when one lender puts money on the market at what seems like a low rate. Imagine what will happen when up to 2000 lenders do that. Hopefully most will reset their YR over the next week.
|
|
|
Post by GSV3MIaC on Jun 16, 2015 20:47:49 GMT
Will be interesting to model (or see) how MRs, under the new system, perpetuate themselves through time. If the MR is being set as an average, and the average is largely determined by all the money stuck it at the MR, then it may become a sort of self-fulfilling prophecy, and very difficult to move in any direction (up, OR down) without a =lot= of manual money being pitched in (and accepted) at much higher or lower rates .. and if MR is busy being rounded to 0.1% I suspect it would require a nuclear blast to move it up/down by even one notch?!
e.g. half a million at X% requires something rather exceptional to shift it .. adding £480k at 'one step up' (X+0.1%) just means a 1dp rounded average of .. yep.. X% for the next day's MR .. rinse and repeat. Or maybe the MR is going to be rounded to 0.01%, but only displayed to 0.1%? (Or maybe you discount MR money when you figure out what the day's average match rate was?).
|
|
|
Post by westonkevRS on Jun 16, 2015 20:51:24 GMT
Thanks westonkevRS, I look forward to an explanation of how the 2dp MR works. [...] westonkevRS, With the new system coming in, which calculates MR as an average, it will be important to avoid this kind of problem. Unless RS moves to displaying rates to 2dp (which I do not advocate) the developers need to make sure that the day's MR is always rounded to x.x0%, otherwise there will be a lot more apparent queue jumping. WOW! Do you want a job! I think you know the matching process better than 99% of RateSetter employees, including (sadly) myself. I have been advised by our CTO that the new MR will definitely only be set at 1 decimal place, albeit it could be derived from a base lender interest return (i.e. the non AER reinvestment number) of 2 decimal places.
|
|
|
Post by westonkevRS on Jun 16, 2015 20:58:14 GMT
Westonkev I'm sure analysis has been performed to compare how the old Market Rate system compares to the new Market Rate system over the past x months - any chance you can share that? (for example, a graph depicting the last x months, of old system vs new system) I hope that makes sense. Actually no deep statistical analysis had been done comparing old MR to new MR, it was just the right thing to do. However tonight (as I haven't got anything better to do) I did perform some analysis to determine the impact. It wasn't perfect as I looked at the previous days contracts ove the full 24 hours, rather than just the 6am to 10pm slot for the new MR calculation. I won't share the full analysis because I only did it tonight and haven't fully shared internally, however the net effect of the change was...... virtually zero.My calculation had the monthly market rate reducing by a tiny negligible percentage and the 5-year money being consistent. However this cannot be used to estimate the future impact, as there are so many other factors at play (how many borrowers we have, new lenders, impact of advertising, point in month) and also the new YR which I didn't model. So DO NOT make any decision based on this analysis. I'm sure if the rates do shift materially you'll all repeat this thread as a hoist to my own petard. westonkevRSwww.linkedin.com/profile/view?id=19236219
|
|
|
Post by westonkevRS on Jun 16, 2015 21:05:25 GMT
Will be interesting to model (or see) how MRs, under the new system, perpetuate themselves through time. If the MR is being set as an average, and the average is largely determined by all the money stuck it at the MR, then it may become a sort of self-fulfilling prophecy, and very difficult to move in any direction (up, OR down) without a =lot= of manual money being pitched in (and accepted) at much higher or lower rates .. and if MR is busy being rounded to 0.1% I suspect it would require a nuclear blast to move it up/down by even one notch?! I think you are right. I'm sure Marketing hope the the advertising works and we have a nuclear last of money piling in..... ..... Which I wouldn't mind too much because we get a better class of borrower with lower rates. Just as well we are especially efficient and therefore our final APRs aren't high (http://www.altfi.com/article/0989_comparing_the_big_three_a_dive_into_the_loan_books_of_zopa_funding_circle_and_ratesetter " Those of you who are familiar with the rates that .... and RateSetter offer their lenders may be surprised to see ....’s mean lending rate higher than RateSetter’s") westonkevRSwww.linkedin.com/profile/view?id=19236219
|
|
|
Post by westonkevRS on Jun 16, 2015 21:15:54 GMT
Around 20% of money on the markets has a YR set. Around 2,000 lenders have a YR set below the MR Kev, thanks for sharing the facts. If I understand this correctly, this means about 4500 lenders have a YR set and of those, nearly half have it set at below MR or are you including accrued repayments that may not have gone out in months as you refered to "money" as opposed to "lenders"? Simply 2,000 number of lenders that have a YR setting lower that their MR, so most will be reinvesting but it wasn't segmented by monies actually on the market.
|
|
sl75
Posts: 2,092
Likes: 1,245
|
Post by sl75 on Jun 16, 2015 22:07:12 GMT
Around 20% of money on the markets has a YR set. This doesn't nessarily mean it's stranded as when we speak to customers they are happy to leave the money "resting" at the higher rate. Although this is a strange response (in my opinion) because if RateSetter had done more loans on the day their money would have gone. At which point it's no longer "resting", but usefully employed, earning interest at a rate the customer is happy with. Speaking personally, I'd happily "rest" funds in that manner until or unless a sufficient amount (of money or time) accumulates to make a withdrawal worth bothering with (whereupon the funds can be re-invested via another investment platform where an equivalent rate [adjusted for risk and personal preferences] is readily available, or simply left to "rest" in the bank account where it has more options for being immediately deployed elsewhere).
|
|
|
Post by westonkevRS on Jun 17, 2015 6:24:13 GMT
Around 20% of money on the markets has a YR set. This doesn't nessarily mean it's stranded as when we speak to customers they are happy to leave the money "resting" at the higher rate. Although this is a strange response (in my opinion) because if RateSetter had done more loans on the day their money would have gone. At which point it's no longer "resting", but usefully employed, earning interest at a rate the customer is happy with. Speaking personally, I'd happily "rest" funds in that manner until or unless a sufficient amount (of money or time) accumulates to make a withdrawal worth bothering with (whereupon the funds can be re-invested via another investment platform where an equivalent rate [adjusted for risk and personal preferences] is readily available, or simply left to "rest" in the bank account where it has more options for being immediately deployed elsewhere). I don't mean lent out into the loan pool, I mean there are 2,000 (approx) lenders with a YR setting lower than the prevailing MR. This was a snapshot, so any money on the market for these people is probably going to be lent shortly onto loans. We just want to make sure they know if they don't change the rate the money will be lent at their YR and no longer the higher MR from 24th June. Money sitting in the markets waiting to be lent does not earn interest. Money "resting" on our platform is not "usefully employed earning interest at a rate the customer is happy with" unless they are happy with 0%. It is sitting waiting to pounce if the right rate arrives, otherwise it would only be earning interest if it was withdrawn to the bank account. @ westonkevRS
|
|
spiral
Member of DD Central
Posts: 909
Likes: 456
|
Post by spiral on Jun 17, 2015 7:52:24 GMT
Money sitting in the markets waiting to be lent does not earn interest. Money "resting" on our platform is not "usefully employed earning interest at a rate the customer is happy with" unless they are happy with 0%. Although the bigger picture (if and when it gets lent out) is that about 15 days at 0% is interest well earned if it ultimately grabs you an additional 0.1% (in the 5yr market)
|
|
|
Post by westonkevRS on Jun 17, 2015 14:34:23 GMT
Money sitting in the markets waiting to be lent does not earn interest. Money "resting" on our platform is not "usefully employed earning interest at a rate the customer is happy with" unless they are happy with 0%. Although the bigger picture (if and when it gets lent out) is that about 15 days at 0% is interest well earned if it ultimately grabs you an additional 0.1% (in the 5yr market) "interest well earned if it ultimately grabs you an additional 0.1%" In the recent rising market everyone has been winning with the extra 0.1%. If the tide (rates) goes out, someone will be left with their pants down!
|
|