Print

[mothandrust]

It's too bad we can't cash out while we're ahead.

Well, if the class action is successful, we can -- 100% return of principal plus 10% annual interest.

minus lawyer fees.... 40% of that   

Attorneys are asking the judge to award attorneys fees (separately).

What's next after that, recovering $54/share in cash plus lost dividends for Citigroup stockholders? 

[bamalucky]

$27,758,496 has been charged off


http://www.prosper.com/lend/performance.aspx

8.75% are late

Fast forward a little over a month

Charge-offs    $30,330,060  20.11%

Plus 8.28% are late

1.70% Late (16-30 days)

So the recent late are getting slower,but thats just due to the loan pool getting smaller

[Investar]

Fast forward a little over a month ... the recent late are getting slower, but thats just due to the loan pool getting smaller

I would suggest it's not the "size" but the "age" -- as Eric's Lates By Payment Report shows, the majority of wheels fall of the carts during the first 12 months. During the most recent 12 months, we are coming up on 6 months "Quiet" with no new carts starting down the path. The percentage of total loans on the bumpiest part of the ride is going lower and lower each month. Maybe that's what you meant. 

[Fred93]

I would suggest it's not the "size" but the "age" -- as Eric's Lates By Payment Report shows, the majority of wheels fall of the carts during the first 12 months. During the most recent 12 months, we are coming up on 6 months "Quiet" with no new carts starting down the path. The percentage of total loans on the bumpiest part of the ride is going lower and lower each month. Maybe that's what you meant. 

Eric's chart misinforms.  The portfolio of loans contains loans of different ages.  They are in fact highly skewed toward younger loans.  His denominator is the total of all late payments.  More "1st" payments have come due than "12th" payments, so of course a higher fraction of the total of late payments are "1st" payments than "12th" payments.

This isn't a probability of going late.  This is a probability that if you are late, what month are you.  Nobody needs to know that.  It is highly skewed by the distribution of loan ages.

If you want to know when loans go late, see my charts, or the one recently posted by somebody (was it researchpro?) showing the probability that an individual payment is late.

[ira01]

If you want to know when loans go late, see my charts,

Yes, exactly right.  http://www.prospers.org/forum/fred93s_blog_030109_late_loan_stats_update-t12320.0.html

Unfortunately, as can easily be seen, even very old loans continue to go late faster than lates cure, since the slope of most of the lines continues to be positive.  Put another way, the percentage of bad loans continues to increase almost every month, even for the very old loans.  It appears likely that each monthly cohort will exceed 40% bad.

[onthefence]

I would suggest it's not the "size" but the "age" -- as Eric's Lates By Payment Report shows, the majority of wheels fall of the carts during the first 12 months. During the most recent 12 months, we are coming up on 6 months "Quiet" with no new carts starting down the path. The percentage of total loans on the bumpiest part of the ride is going lower and lower each month. Maybe that's what you meant.

I agree.  That chart produced by Eric does not provide an useful information as far as I can tell.  All his chart means is that more loans were produced recently than when prosper began.

"% This Cycle" refers to total lates as a percentage of all lates.  What would be more meaningful is if the number was "% Lates of Total Loans This Cycle"

ETA: On 2nd review, it might be me who is misinterpreting the chart.  I PMed Epetrol to see if he has any insight to provide on the issue.

[Investar]

I would suggest it's not the "size" but the "age" -- as Eric's Lates By Payment Report shows...

That chart produced by Eric does not provide an useful information as far as I can tell.
ETA: On 2nd review, it might be me who is misinterpreting the chart.  I PMed Epetrol to see if he has any insight to provide on the issue.

The chart and graph focuses on "when" in the cycle, not how many. In that respect I find it a handy kwik-reference, albeit for limited purposes. The introduction to the data describes Eric's intention and the data limitations pretty well. As Ira01 pointed out , "the percentage of bad loans continues to increase almost every month, even for the very old loans" and Eric's study confirms that. And the cumulative effect is well known, thanks to Fred's graphs in particular. Eric's specific study is ancillary to that data, certainly, but IMHO adds clarity. 

[Investar]

I would suggest it's not the "size" but the "age" -- as Eric's Lates By Payment Report shows...

Eric's chart misinforms.  The portfolio of loans contains loans of different ages.  They are in fact highly skewed toward younger loans.  His denominator is the total of all late payments.  More "1st" payments have come due than "12th" payments, so of course a higher fraction of the total of late payments are "1st" payments than "12th" payments.

This is certainly true. Eric's introduction says, "In order to get a perfectly accurate picture of things, we would have to look only at loans that have completed their full 36 month payment cycle. Since Prosper hasn't been around long enough to have a significant number of loans at that age, this chart may not be entirely accurate (although I suspect it will prove to be accurate enough for most purposes).
To compensate for this a bit, only loans originated prior to Jan 1, 2007 are included for now."

[epetroel]

Thanks to OnTheFence for filling me in about this discussion

I think Investar & Ira summed up the purpose of that chart exactly.  As for Fred93's comment about it being "misleading" or "highly skewed toward younger loans", I think that's a bit overstated.  The chart purposefully only includes loans originated before Jan, 2007 to minimize this effect.  So the denominator is in fact the same for the first 27 months (as you can see from the data below the chart). 

The data may be somewhat inaccurate because not all the loans have completed yet (as I mention in the summary), but I'm pretty confident that the overall trend shouldn't deviate much from what you see in that chart.  If Fred93 could explain to me his reasoning for thinking that the final results will be significantly different than what I currently show there, I would be interested to hear that.

The original purpose for that chart was to help with pricing risk in the secondary market (if one ever comes along).  The idea being that you could use that chart as a guide to help you figure out how much risk has been "baked out" of a loan at a certain age.  For example, a loan that has already made 12 on-time payments should be perceived as somewhat less risky than a brand new loan.  Using my chart, you would be able to quantify that risk - after 12 on time payments, the default risk should have decreased by about 70%.  If you used Fred93's charts, it would be around 65% instead after accounting for differences in methodology (My "Late" vs. Freds "1+ late", Fred's use of current loan status vs. my use of "first month late").

[bamalucky]

$27,758,496 has been charged off


http://www.prosper.com/lend/performance.aspx

8.75% are late

Fast forward a little over a month

Charge-offs    $30,330,060  20.11%

Plus 8.28% are late

1.70% Late (16-30 days)

So the recent late are getting slower,but thats just due to the loan pool getting smaller

Fast forward to april 7

Charge-offs    $30,953,285   20.49%

Plus 8.06% are late

1.66% Late (16-30 days)

ETF chargeoff percentage

[bamalucky]

7.01% of AA have been charged off
12.10% of A
15.49% of B
19.60% of C
21.85% of D
31.88% of E
40.11% of HR
54.46% of NC

Keep in mind that this isn't counting active late

[ira01]

$27,758,496 has been charged off

http://www.prosper.com/lend/performance.aspx

8.75% are late

Fast forward a little over a month

Charge-offs    $30,330,060  20.11%

Plus 8.28% are late

1.70% Late (16-30 days)

So the recent late are getting slower,but thats just due to the loan pool getting smaller

Fast forward to april 7

Charge-offs    $30,953,285   17.33%

Plus 8.06% are late

1.66% Late (16-30 days)

I think there is something wrong with your math -- how did charge-offs increase from $30.3M to $31M, yet the percentage dropped from 20.11% to 17.33% (considering that Prosper is not originating new loans, so the denominator of all loans funded wouldn't have changed)?

[ira01]

7.01% of AA have been charged off
12.10% of A

These are the shocking numbers, IMHO.  Wasn't the Experian default rate for AA's like 0.1%?     Lenders knew those numbers were very understated, but I doubt anyone expected the true numbers would be 70 times worse. 

[NewHorizon]

I think there is something wrong with your math --

Whew - I thought it was just me who was having trouble getting the numbers to come out right..
But it's not bama's math.  It's just cut/pasted from http://www.prosper.com/lend/performance.aspx

[lenderguy]

7.01% of AA have been charged off
12.10% of A

These are the shocking numbers, IMHO.  Wasn't the Experian default rate for AA's like 0.1%?     Lenders knew those numbers were very understated, but I doubt anyone expected the true numbers would be 70 times worse. 

I think it was 0.3%.  At the other end, HR was 19.1%, IIRC.