Author Topic: Can more filtering improve upon LC's expected returns in the long run?  (Read 14334 times)

MoneyTree

  • Newbie
  • *
  • Posts: 41
    • View Profile
    • Email
I suspect this topic has been raised before, but I haven't been able to find it.

My question is simple: Assuming LC (and Prosper) are trying to create an efficient market, isn't it true that opportunities for higher-than-expected returns through filtering (beyond simply choosing an interest rate) will:

1) eventually disappear (as LC discovers these inefficiencies and adjusts to eliminate them), and
2) become fewer in number and less significant in scale (due to #1)

Even if the answer to the above is "yes", it doesn't necessarily mean that it's not worth filtering at this stage of the game, but it would seem to suggest that that one shouldn't expect the detected opportunities to pan out well as the results of the analysis.

BTW, I'm not referring to fields that LC is not legally allowed to filter on, such as employer or state, (at least that's my understanding). Nor am I referring to the ability to filter based on comments the borrower makes.

A simple example of what I'm talking about would be minimum income. It seems that simply filtering for that one factor can improve one's expected IRR. This should not be the case in an efficient market, should it? And if not, won't LC eventually make adjustments to eliminate it?

I hoping you guys can show me that I'm wrong in my analysis, since it doesn't bode well for the long-term efficacy of filtering.

rawraw

  • Hero Member
  • *****
  • Posts: 2756
    • View Profile
Why do you think they are trying to "create an efficient market" ?  I don't really understand what you are talking about.

Fred

  • Hero Member
  • *****
  • Posts: 1421
    • View Profile
I see LC & Prosper more as a manufacturer than a market.

They manufacture Notes and investors buys them.  Investors are simply price takers, no market microstructure is happening -- where highest bid is met with lowest ask.

However, Notes are not manufactured with equal quality -- even those of the same interest rates.    So investors want to make sure to inspect the goods before purchasing them.

FOLIOfn, on the hand, acts more like a market.

brycemason

  • Hero Member
  • *****
  • Posts: 801
    • View Profile
    • P2P-Picks.com
    • Email
The P2P platforms objective is to maximize profit, which it does primarily through origination fees. Thus, they will do what it takes to maximize originations. If a portfolio is returning 6-7% as LC is, they can make more money by mixing in some additional weaker loans (at what the weaker borrowers perceive as competitive interest rates--less than their true risk would command). As long as the overall portfolio returns don't fall too low, they keep their credibility and make more money. Thus, there will always be opportunities to note pick.

MoneyTree

  • Newbie
  • *
  • Posts: 41
    • View Profile
    • Email
An "efficient market" is one where the cost of things and the value of things are closely correlated. In the context of P2P lending, an efficient market is one in which the notes have an interest rate that is appropriate to the level of risk. LendingClub has very good motivation to assign interest rates correctly. If they set the rate too low for the risk they will lose investors (to other investment opportunities that offer better risk/reward); if they set the rate too high for the risk, they will lose borrowers (to other less expensive borrowing opportunities).

LendingClub acknowledges that they adjust their rating system from time to time. The reason they do this is to attempt to improve upon the accuracy of the interest rates they assign notes.

Investors who do additional filtering are attempting to find notes that have been over-priced relative to their likelihood of defaulting (risk). If such notes exist, they represent situations where borrowers are being over-charged to borrow money. LendingClub would be foolish to do this intentionally, and I'm sure they don't. I believe they are actively doing everything in their power to have all borrowers treated equally relative to each borrower's risk level. To do anything else makes absolutely no sense.

So, my question remains: is there any reason to believe the opportunities we are (perhaps) finding to get higher than LC's expected returns are going to continue to exist? Or will most of them be discovered and eliminated over time by LC's underwriting team?


MoneyTree

  • Newbie
  • *
  • Posts: 41
    • View Profile
    • Email
The P2P platforms objective is to maximize profit, which it does primarily through origination fees. Thus, they will do what it takes to maximize originations. If a portfolio is returning 6-7% as LC is, they can make more money by mixing in some additional weaker loans (at what the weaker borrowers perceive as competitive interest rates--less than their true risk would command). As long as the overall portfolio returns don't fall too low, they keep their credibility and make more money. Thus, there will always be opportunities to note pick.

For the reasons I gave above, I don't believe this to be the case, but perhaps I'm wrong. Do you have any evidence supporting your notion that LC knowingly and intentionally foists a limited number of improperly rated notes onto investors in order to increase origination fees? It would seem to be a very short-sighted strategy, as long as they have competition, which they do (e.g Prosper).

brycemason

  • Hero Member
  • *****
  • Posts: 801
    • View Profile
    • P2P-Picks.com
    • Email
It's just an inductive argument. I have no evidence, just analyzing incentives. Why wouldn't Prosper be doing the same thing? If you want to maximize originations, you have the incentive to cut loans with some underpriced risk like drywall in cocaine.
« Last Edit: April 23, 2013, 12:54:52 AM by brycemason »

MoneyTree

  • Newbie
  • *
  • Posts: 41
    • View Profile
    • Email
It's just an inductive argument. I have no evidence, just analyzing incentives. Why wouldn't Prosper be doing the same thing? If you want to maximize originations, you have the incentive to cut loans with some underpriced risk like drywall in cocaine.

But your analogy doesn't quite work, since drywall cannot be separated from the cocaine once it's mixed in, whereas filtering can separate the good notes from the bad. Why would it be a more successful strategy for LC to purposely rate the notes inconsistently, instead of simply setting all notes' interest rates lower than the risk would warrant? It would seem that the later strategy is more likely to work, since it can't be gotten around through filtering.

Fred

  • Hero Member
  • *****
  • Posts: 1421
    • View Profile
An "efficient market" is one where the cost of things and the value of things are closely correlated....

For some reasons, when I read about "efficient market", I was thinking more about Efficient-market hypothesis (EMH) which deals primarily with information and prices.

Peter

  • Administrator
  • Hero Member
  • *****
  • Posts: 754
    • View Profile
    • Lend Academy
    • Email
I am not so sure that LC makes most of their profits on origination fees. If I had to guess I would say investor service fees provides more profit. They certainly make most of their revenue from service fees but the costs involved in generating that revenue (acquisition and processing costs) would be high also. Now, I have no inside information here and I don't think LC or Prosper will ever share this with me but the cost to bring an investor on to the platform is very small so it leads me to believe that this side of their business is indeed the profit center.

But getting back to the original question. It is an interesting hypothesis and one I may explore in a blog post at some point. If the platforms were truly efficient there would be no advantage in selecting loans - the returns would be uniform. But in the real world where investors can select very tiny slices of the loan platform it stands to reason that some slices will perform better than others - even those slices priced identically.
Publisher of the Lend Academy blog

See my returns here: http://www.lendacademy.com/returns

brycemason

  • Hero Member
  • *****
  • Posts: 801
    • View Profile
    • P2P-Picks.com
    • Email
Re: Can more filtering improve upon LC's expected returns in the long run?
« Reply #10 on: April 24, 2013, 09:39:14 AM »
We definitely know their revenues are higher from origination fees per their 10Q

Fred

  • Hero Member
  • *****
  • Posts: 1421
    • View Profile
Re: Can more filtering improve upon LC's expected returns in the long run?
« Reply #11 on: April 24, 2013, 11:12:42 AM »
I am not so sure that LC makes most of their profits on origination fees.

Per Form 10-K, for the "Nine Months Ended December 31, 2012":

Origination fees                      $26,013
Servicing fees and other revenue      $ 2,914
...
---------------------------------------------
Total Net Revenue                     $28,593

Fred

  • Hero Member
  • *****
  • Posts: 1421
    • View Profile
Re: Can more filtering improve upon LC's expected returns in the long run?
« Reply #12 on: April 24, 2013, 11:16:03 AM »
In thousands.  :)

MoneyTree

  • Newbie
  • *
  • Posts: 41
    • View Profile
    • Email
Re: Can more filtering improve upon LC's expected returns in the long run?
« Reply #13 on: April 24, 2013, 01:06:05 PM »
But getting back to the original question. It is an interesting hypothesis and one I may explore in a blog post at some point. If the platforms were truly efficient there would be no advantage in selecting loans - the returns would be uniform. But in the real world where investors can select very tiny slices of the loan platform it stands to reason that some slices will perform better than others - even those slices priced identically.
I hope you will! We've already seen at least one example of what I'm talking about, as when Prosper lowered the rates they give repeat borrowers, thus effectively removing what had been an effective filtering opportunity for investors. Given that, I see no reason why they (and LC) won't do the same in the future as they discover other such patterns.

(Just to clarify, Peter: when you say "there would be no advantage in selecting loans - the returns would be uniform" I assume you mean selecting on criteria other than interest rates, since there is a legitimate reason why the returns for higher interest rate loans should be higher than for low interest loans, namely risk.)

Zach

  • Administrator
  • Hero Member
  • *****
  • Posts: 622
    • View Profile
    • Email
Re: Can more filtering improve upon LC's expected returns in the long run?
« Reply #14 on: April 24, 2013, 05:26:56 PM »
We definitely know their revenues are higher from origination fees per their 10Q

I think what Peter is trying to get across is that while the revenue may be higher, the gross margin on  origination fees is likely lower than the gross margin on investor service fees.