Should I Invest in LendingClub or Prosper?

Lending Club Booth

A legit investing opportunity?

“The surest way to ruin a man who doesn’t know how to handle money is to give him some.”
–George Bernard Shaw

I recently had someone ask me if it was wise to invest in LendingClub or Prosper. For those of you who do not know what these are, they are peer-to-peer lending institutions that usually allow people to borrow money for less than they could at a traditional bank. It’s, generally, because these people have abnormal profiles (read: lower credit scores), so they go the route of peer-to-peer lending instead. These online lenders also claim to be able to offer lower interest rates because they don’t have nice, big, shiny bank buildings and the requisite VP parking spots to have to pay for.

From an individual lender point of view, it is appealing because you get to diversify your risk across lots of small loans. Usually, the loans are in increments starting at $25, and you get to read the dossier on the borrower. The borrower has to provide their FICO information to the platform (e.g. LendingClub or Prosper), and then they get a chance to provide a normative explanation for why they need the money, what they’re going to do, and anything else they can think of to tug on your heart strings or satisfy your inner green visored bean counter that this is a safe loan.

I have a LendingClub investor account, although I’m no longer actively funding – the reasoning which I’ll explain later. I’ll be walking you through my numbers and my results shortly.

First, let me explain why I thought that I could be a good investor in LendingClub:

  • I’m very analytical. I’m not a SAS-wielding PhD statistician, but I know my way around analytics, and there’s even a business school case study which I authored centering on Markov chain modeling that is floating around the Internet somewhere.
  • I actually worked in the credit card industry. While I didn’t do the credit modeling, I was close enough to the process to think I had a pretty good idea of how to look at profiles and determine, reasonably well, the likelihood of getting my money back.
  • I had the money to invest. While this wasn’t exactly swinging for the fences, it was an alternative risk investment which, if things didn’t work out, wasn’t going to kill me or cause me endless sleepless nights.

How has it worked out so far?

I started investing in LendingClub accounts on March 26, 2010 and stopped funding them on May 26, 2011, investing in a total of 86 notes.

There were times when that number was in the mid-teens, but what the number shows is the projected returns as if people paid their loans on an expected basis. That’s why I expect that number to dip lower over time until I’ve pulled out all of my money.

If you have a large enough sample size, you’ll hit the averages. However, with 86 loans out of a total of 106,138, I’m likely to be skewed by outliers.

This is exactly what has happened. In total, the chargeoff rate for LendingClub accounts is right at 4% (well, 3.99544%). Yet, I skewed in the wrong direction.

I’m not exactly surprised, since I was bitten primarily from dipping my toes into the deep end of the risk pool. Below are the profiles of my charged off loans.

Thus, I wasn’t as smart as I thought I was when it came to picking winners and losers. The reality is that I don’t have enough of a sample size to determine if I was truly skillful or I’d just been unlucky. To have a confidence interval of 95% (2 standard deviations) and be at +/- 1% on my ROI, I’d need 8,807 loans. I was at 1% of the required number, meaning that I can have almost NO idea if I was good or not (there’s a 10.6% confidence interval for those probability and statistics fans out there).

However, it wasn’t the statistical evaluation which made me decide to not invest in LendingClub. It was a moral decision.

By the time June 2011 rolled around, I was putting the final touches on the sale of my company. I knew at that point that I wanted to be a financial planner. I figured it’d be pretty hypocritical if I was going to be running around preaching that there is no such thing as “good debt”, yet be lending people money. Since you can’t just call in the loans, I did the best thing that I could – stop investing in the loans and withdraw money as it became available. That’s what I’ve been doing since June, 2011 and will continue to do so until the account is empty.

“But Jason,” you think to yourself, “I’m not a financial planner! I’m not tied by your constraints! Should I invest in LendingClub?”

Of course, my stock answer is to see your financial planner (ahem…contact me?), since I don’t know your particular situation and, therefore, can’t give personalized advice.

However, I will say this based on my experience. Assuming that I invested $25 per loan, I’d need to invest $220,175 ($25 X 8,807 loans) just to have a 95% confidence interval in my ROI. So, let’s say that I kept that 8.35% return. I could say that 95% of the time, my return would be somewhere between 7.35% and 9.35%, and 5% of the time, it could be goodness knows what. Furthermore, what if I wasn’t as good as I thought, and by the time the results from loan number 8,807 came in, I was down 20%? That’s a lot of money to invest and a long time to wait (up to 60 months) to determine that my particular investing strategy sucked.

But wait! you think. “Why don’t I just invest in one class?

We can look at the extremes of population. On the low end (both in population and quality) is the G class. They have 795 loans total, of which 16.2% charged off. To get a statistically significant sample with the abovementioned parameters, you’ll have to invest in 734 of the loans. That’s $18,350.

On the high end is the B class. They have 34,981 loans. You’ll have to invest in 7,535 loans to get a good reading. That’s $188,375.

Therefore, based on the amount required, which, unless your net worth is greater than $4.4 million (assuming you’re trying to achieve a diversified portfolio), you can’t invest in LendingClub or Propser with sufficient rigor to determine if you’re using an appropriate investing strategy and meet my guidelines for taking a shot with your money.

Because of this, you truly are gambling. If you want to gamble and try to earn some money, then be my guest. Just understand that you’re gambling, not investing. I’m not going to beat you up as long as you stay in the guidelines, but, from my point of view, you’d be better off investing that money in one of two ways:

  1. Trying to start your own side gig to turn into a business. 41% of the world’s millionaires are entrepreneurs.
  2. Invest in yourself to make yourself more valuable in your job. If you can get a 10% payraise, then that will compound returns year over year; whereas, with LendingClub or Prosper, you have to continue to find loan opportunities which perform at that return rate. I know which one I think is riskier.

What do you think? Am I too harsh? Have you invested in LendingClub or Prosper? What were your results? Did this article cause you to rethink your strategy? Tell us about it in the comments below!

Online Comprehensive Financial Planning

Get your comprehensive financial plan now at!

About Jason Hull, CFP®

Jason Hull, CFP®, is the Chief Technology officer of myFinancialAnswers, an online comprehensive financial planning service.


  1. Great post.

    It’s very tempting to chase yield with this asset class, but it seems difficult to scale up without tedious labor. I understand that the larger P2P lenders will help automate the process for their larger investors, but now this math shows it also takes a significant financial commitment to avoid a nasty negative skew.

    I also wonder what seductive psychological phenomena are at work here. It must offer a sense of real control to be able to screen the borrower’s criteria, to say nothing of the power to grant their wishes. Sort of like sitting in the skybox of the Roman Coliseum as the gladiators gaze up at you for a thumbs up– or get a thumbs down.

    Junk bond funds and mortgage bond funds seem a lot less risky by comparison… and a lot less labor, too. Heck, even crowdfunding and angel investing seem easier.

    • Nords–

      Thanks for the usual thought-provoking comment!

      I certainly jumped into this without running the numbers first to see the level of commitment I’d need to truly understand my success (or lack thereof). There’s probably a meta-lesson to be learned about investing without truly knowing what you’re investing in.

      You raise a point I’d not considered regarding the psychological underpinnings of being the lender rather than the borrower. In a sense, it’s a power play. If you read some of the questions and commentary, there are some real a$$holes rigorous questioners, which, given the size of the loan they’re probably making, seems all out of whack. I think I asked one question in the entire process, and it was to encourage the prospective borrower to actually write something about what was going on rather than to have a blank profile and ask for money.

      The process itself is, indeed, time-consuming. I estimate it took me, between screening and reading the profiles, about 10 minutes per loan that I actually made. Given that almost all of my loans were $25, using my ROI, I’d make $2.0875 per loan, or $12.53 per hour. Not a great use of my time relative to some other things that I could have done or had utility for.

      I had a few heuristics that I used which could have scaled as a model, but a lot of it came down to (despite my claims of the capability of using analytical rigor) gut feeling. I’m sure someone could build an amazing model out of the performance of successful loans to date, but it’d be a heck of a lot of data crunching (hello Hadoop and MapReduce) and even then, with the wrong assumptions or modeling, could blow up like Long Term Capital Management. I just read recently that the London Whale error was based on a bad formula in Excel, so even automating the process might mean that you automated the wrong process. I’d not want to be the one to raise and manage a $100 million fund to invest in LendingClub or Prosper, even if I could build the model to 99.99% confidence. I’d not sleep well (or at all) at night.

  2. Thanks for this article Jason. As I mentioned to you, I was curious what you thought about these types of investment websites so it’s nice to hear your point of view. It definitely makes me think twice about putting any money here.

    • Hi Courtney–

      Glad to help out! You helped me too, as you gave me fodder for a blog post! I can always use more of those! Feel free to shoot me a note if there were any unanswered questions that came out of my post.

  3. I have been an investor in Lending Club since 2009 and Prosper since 2010. I disagree with many of your assumptions here. I share my returns publicly on my blog and in 2012 I received a real world return (which is different than what Lending Club and Prosper tell you) of 10.77% on a portfolio with around 5,000 notes.

    While p2p lending is certainly not without risk it is maturing now into an asset class that is providing consistent returns for well diversified investors. Now, I am not a statistician but I take issue with your assumption that you need many thousands of loans to be confident in your portfolio returns. As of today no investor with at least 800 notes has lost money at Lending Club. None. The vast majority are earning returns in excess of 8%.

    There is a reason this asset class is growing at over 100% a year and closely approaching $2 billion in total loans issued. It is great deal for investors and borrowers. I will continue to put new money to work there.

    • Peter–congratulations on your successes and thanks for commenting! LendingClub and Prosper are probably becoming more mainstream platforms and, therefore, lowering the overall risk profile of the borrowing community.

      You have 5,000 notes. I don’t know what your actual distribution is, but you’re getting closer to a statistically significant profile within the guidelines I set forth than I with my piddly 80 something notes would ever hope to achieve. You also put over $110k of funding into the accounts to seed them (totally concur with the XIRR returns, by the way). That investment would require, for my guidelines, a net worth of $2MM to get there.

      Remember, I didn’t say that you couldn’t invest in Lending Club or Prosper, but, rather…

      …you can’t invest in LendingClub or Propser with sufficient rigor to determine if you’re using an appropriate investing strategy and meet my guidelines for taking a shot with your money.

      Are there worse things in the world to do with your money? Certainly. Is it the sure shot that a lot of people think it is? No.

      Perhaps I was being too harsh. If I lower the confidence intervals down to +/- 9.64% (the average return for LendingClub), then it only requires 110 loans. If you’re dead average, then you’ll know in that many whether or not you’ll at least break even. However, you’d have to let all of the loans run to completion before you could determine that significance, as, up until a loan is paid off in full, it runs the risk of default. Therefore, a truly prudent approach would require waiting until you’ve cleared out the portfolio before reinvesting. That’s either 36 months or 60 months of waiting for the polls to close.

      This also assumes that you’re maintaining all of the variables (e.g. your decision framework) exactly the same for each decision point. Each time you change a parameter, you throw off the measurement.

      There’s also a tax difference between the ordinary income and long-term capital gains. If you’re above the 20% marginal tax bracket, you’ll need a higher return to get the same after-tax benefits as a long-term capital gain.

      So, again, if you’re a SAS genius and can build the model, then you probably don’t require the tight parameters that I set forth in the article. However, for the rest of us, I still stick to a similar (albeit probably not as strictly adhered to thanks to your thought provoking comment) conclusion:

      You may (and probably will) profit from investing in those platforms, but a) you’ll have no true idea why you did, and b) you’ll have no idea if you can sustainably reinvest with the same level of success.

      Because of that, I’d be more in support of a Propser/Lending Club index fund; that actually might be a viable investment (if such a creature existed) that I’d feel comfortable suggesting as part of an income portfolio. The parallels between individual stock picking and individual loan picking are pretty striking to me.

      Thanks again for commenting. I do appreciate it. I wish you continued success!

      • While I still disagree with your requirements for a high confidence level in your return I think the idea of an index fund for Lending Club and Prosper is a good one. There is currently such a fund for Lending Club but it has a $500K minimum and Prosper offers no such opportunity. But I expect we will see innovations in this area in the near future.

  4. Curious, if you used limited funding to purchase 86 notes….why sink twice the minimum $25 into D and E grade loans? Seems you would have been better off to stay with the minimum $25 until you could fund at least a few hundred notes, at least more than 86 with the same dollar amount invested. Haven’t you also skewed the probability by causing a heavier weighting to certain notes rather than spreading the risk?

    • Actually, I was funding with $25 for about the first 40 and then upped it to $50 for the remainder, so it looks like I upped the risk twofold (raised amount, higher risk categories) when the $50 investments were also across the board.

      You do raise a good point about my poor choice to up the funding amount per note and I agree with you on it. I should have gone for more, smaller shots.

      Thanks for dropping by and for the comment!

  5. Guys,

    Thank you for an excellent thread and followup questions. As the proverbial SAS yielding PhD Physicist, I have been playing around the historical loan portfolio that Lending Club published and trying to use machine learning methods to derive probabilities of defaults.

    There is one underlying problem with LendingClub: their loan portfolio only goes back to 2007 and hence only loans originated in 2007, 2008,2009 and some loans from 2010 have reached maturity. That is a pitifully small sample to build a statistical model on. To build a robust model you would need 20-30 years if history with hundreds of thousands of loans or millions of loans to start constraining your confidence levels on your own credit risk ratings.

    So I concur with the author that at this point investing into Lending Club is fun, mostly harmless but gambling nevertheless.

    • Pavel–

      Thanks for hopping on there! I took a look at what you have so far; it’ll be interesting to see the final conclusions of your sampling!

    • Jon Pildis says:

      With all due respect to Pavel and PhDs everywhere, it was SAS-wielding PhDs that wrecked Wall Street with their sophisticated risk modeling. They believed that mortgage bonds were risk free because they had way too much data and not enough common sense.

      • Fair point, although, the story of the quant model, from Long Term Capital Management through to the mortgage tranche collapsing was one of waving away risk. I believe that what Pavel is modeling (and what I’m emphasizing) is that there’s too much risk, and by the time you actually acquired enough data of your own, you’d have spent a long, long, long time and a lot of money getting to your conclusions…and that’s only if the conditions remained exactly the same as they were when you were building your model.

        I daresay that both you and Pavel are coming to the same conclusions by different paths, though I could be unintentionally putting words into either mouth (or onto either keyboard?).

        The bottom line is that there’s no way I’d recommend that any of my clients spend anything but their at risk capital in this endeavor, and even then, there are higher multiple risk/reward alternatives to take a crack at.

        • Jon Pildis says:

          The point I’m trying to make is that there is no investment that passes the threshold to which P2P is being held. Stock could have another lost decade, interest rates could spike to 18%, housing could crash… there are ZERO investments that provide a risk-free return over t-bills. The key question investors have to ask is whether they are being compensated for both the know and unknown risks of any particular investment. I think P2P is has one of the better risk/reward ratios… especially if you define risk as the loss of capital.

          • I agree up to the point of the risk-reward ratios. You fail to paint the other gloom scenario – you pick a batch of borrowers who tell you that they’ve changed their ways, turned the corner, got religion, and are going to pay the debt unfailingly, only to have their best intentions run smack dab into a layoff and you get to write off your loans. As I argued in the article, if you could have a true index of loans, you’d diversify away the individual risk just as index funds in the market do, but there’s no true animal like that in the P2P industry. There’s too much money chasing too few borrowers at this point. Maybe at some point in the future this equation changes, but, as of now, unless you have a significantly high net worth, there’s no way to appropriately diversify away the risk of your lack of knowledge of the borrower market.

            Let’s be practical about this. How many people are truly, justifiably going to get the return on time invested in picking individual loans compared to other investments they could make? We’re discussing the outside thinnest sliced edge of people who would have the knowledge and ability to create statistical models to actually potentially segregate skill from luck. That’s not the average person who’s going to hop onto a P2P platform thinking that he can read a story about any given borrower and be able to discern the probability of failure better than the next person can.

            As there is no “index fund” for P2P lending and no discernible edge that the average person trying to sort through applications is going to have in picking the true 10% return from the “appears to be 10% but really isn’t because there’s some risk in there that I can’t gauge,” I stand by my original premise that it’s inappropriate for the average investor.

            I do appreciate the discussion; don’t get me wrong! I am glad you’ve jumped in and offered your thoughts!

  6. That is a really great post, i have been thinking about investing on prosper for quite a while now but i have not decided. I got a loan from there when i started my business (which i paid on time and in full) and it was a lot easier to get money from that site than it was from a bank. They do not require worthless documentation like a business plan and stuff. needless to say I really enjoyed my experience as a borrower and im sure there are other like me out there.

    • Hi Jaran–

      I imagine that there’s a segment of people out there for whom borrowing from LendingClub or Prosper is indeed easier, and it may be be a better route for debt consolidation than trying to go through banks or (God forbid) a debt consolidation agency. If you feel like paying it forward and don’t expect a return, then I see nothing wrong with providing money for others to borrow, but expecting super high returns, particularly if the first few succeed, will lead you into disappointment and may cause you to become overly aggressive in trying to chase returns.