Print

[Fred93]

By the way (1,868,557/19,994,575 = 0.0934 or 9.34% of total Principal lent prior to April 01, 2009 has charged-off or defaulted. This data does not include any lates.

That's not a rate.  It doesn't involve time.  Whenever you compute any statistic on a big pile of loans of different ages, you get soup.  The resulting numbers will depend on the age distribution of the loans.  Loans default over time.  Just like interest is paid over time.  To capture this time dimension, it is useful to talk about how much interest is paid per unit time, or how much defaulting occurs per unit time.

Your calculations are something.  I'm not quite sure what.  They certainly aren't rates.

If they are useful to you, great.  But be very careful that you aren't fooled by the fact that you're computing numbers that are very sensitive to the age distribution of the portfolio.

[cowdog]

If it can be said that 40% of loans (the defaults) return only 30% of principal, those are some hard numbers to dodge.

[baeventures]

By the way (1,868,557/19,994,575 = 0.0934 or 9.34% of total Principal lent prior to April 01, 2009 has charged-off or defaulted. This data does not include any lates.

That's not a rate.  It doesn't involve time.  Whenever you compute any statistic on a big pile of loans of different ages, you get soup.  The resulting numbers will depend on the age distribution of the loans.  Loans default over time.  Just like interest is paid over time.  To capture this time dimension, it is useful to talk about how much interest is paid per unit time, or how much defaulting occurs per unit time.

Your calculations are something.  I'm not quite sure what.  They certainly aren't rates.

If they are useful to you, great.  But be very careful that you aren't fooled by the fact that you're computing numbers that are very sensitive to the age distribution of the portfolio.

Ah, now I understand. Thanks much for the explanation.

I still think these numbers are helpful, as while they don't find the value over X time, they do indicate what your fixed loss is although it is not a "rate". Once you determine the average time frame prior to loans defaulting, what would the equation look like to determine the annualized default rate? I have started to work on this data and should have some time this weekend to compile.

BAEVentures

[bankomatic]

I would find this hard to believe. Prosper has a default rate in the area of 40% give or take 5%. How can Lending Club have a default rate of 2.3 to 2.5%? This looks like BS to me. I am also calling BS that a typical Lending Club investor earns 10% pre taxes.

About 40% of prosper loans default over a 3 year period.  That's an annualized rate of about 17% if memory serves me.  The newer prosper loans, after they tightened the requirements will do a bit better.

Lending club has invented some funny calculation I believe, and is underrepresenting their own default rate in the number they present.  My guess is that you need to double their number.  I haven't done the work to detail that out, but may do so at some point. 

Yea, you are right the figures weren't directly comparable, and I agree with you that they are doing some creative accounting to hide the real default rate.