My P2P and Social Lending Experience

Home » Lending Strategy » How Many Notes Do I Need to Be Diversified on Lending Club?

Jan 06

Having a diversified account is the single most important thing you can do to have a positive experience with peer to peer lending. But what does it really mean to have a diversified account? Usually when people say diversified, they mean a portfolio that spreads out the risk enough so that it can weather the risk of defaults without a catastrophic impact to overall returns, and also perhaps a portfolio that is spread across a variety of types of notes and risk profiles as to give the investor exposure to lots of different trends in the market. For example, it seems to make sense that you wouldn’t necessarily want to invest in only A notes (the safest) or only G notes (the riskiest), and it only seems to make sense that if you had $1,000 to invest, you wouldn’t put it all in one note if you had the option to spread it over many notes.

One of the questions I’ve also struggled with a bit, is just how to construct a diversified Lending Club portfolio – how many notes do I need to buy? And not just how many notes do I have to buy to generate a return greater than zero, but how many do I have to buy to generate average returns? The answer depends on your tolerance for risk, and the size of your wallet. To start, let’s take a statistical approach to the problem – I’m not a statistician by training, but I understand the basic concepts, and I’ve got enough of the math in my blood to work out things like statistical sample sizes of a population, which is what we’re going to do here.

Below, I’ve calculated the necessary sample size to reach various confidence levels – naturally, the more notes you buy, the more representative your portfolio is likely to be, and therefore the closer it should mirror the average performance rates. This is the goal – before we think about how to beat the average with any kind of tricky filters or brilliant investing strategies, we want to think about how we just don’t do *worse* than the average.

Note Grade | Population | Sample Size Required - 90% Confidence Level | Investment Required | Sample Size Required - 95% Confidence Level | Investment Required |
---|---|---|---|---|---|

A | 29,148 | 1,598 | $39,964 | 2,218 | $55,458 |

B | 61,843 | 1,646 | $41,156 | 2,311 | $57,782 |

C | 48,624 | 1,634 | $40,861 | 2,288 | $57,201 |

D | 27,278 | 1,592 | $39,814 | 2,206 | $55,170 |

E | 12,439 | 1,488 | $37,223 | 2,012 | $50,316 |

F | 5,672 | 1,302 | $32,574 | 1,687 | $42,178 |

G | 1,155 | 686 | $17,163 | 780 | $19,501 |

If it’s been awhile since your last stats class, here are a couple terms and definitions to reacquaint yourself with:

*Confidence Level*– a measure of how sure you can be that your sample size mirrors the trend in the total population. Think of it this way: if you randomly built thousands of different variations of the same size portfolio, the confidence interval defines what % of the time you’d get the same result. Usually, statisticians shoot for a 95 – 99% confidence level.*Confidence Interval*– this is your margin of error. That is, if you run thousands of different tests with the same size portfolio, you’ll find your result to be within a X % of the average rate of return, with X being your confidence interval. This means our end portfolio should return the average, + or – our confidence interval. For example, if we use a confidence interval of 4%, and the average ROI for a note grade is 10%, we can be confident our portfolio will return between 6% – 14%. The lower the confidence interval we use, the more accurate we can be, but the more notes we will need in our portfolio.*Population*– the total number of items that you can measure; the universe. In our case, this means all the notes issued in a particular grade, on which the average return and default rates are based upon.*Sample Size*– this is what we’re trying to solve for – based on assumptions we make to the confidence level and confidence interval, and the number of notes we know to be in the population, what sample size – that is, number of notes in our portfolio – do we need to feel good that we have a statistically representative group?

In this case, the population is the total number of notes at each grade. For our purposes, the confidence level and confidence interval is basically how fast and loose we’re willing to play with our portfolio risk. The greater the confidence level and / or lower the confidence interval, the less sure we can be our sample portfolio will represent the total population, and therefore adhere to the averages.

At first glance, it’s pretty depressing – you need to buy a lot of notes to meet a 95% confidence level and be within 2% of the average ROI! I mean, holy cow, even at a rather crappy 90% confidence level, you still need to invest nearly about $40,000 *per grade* to have a statistically significant sample size, and at a 95% confidence level it’s nearly $60,000 per grade. When I first ran these calculations, they were pretty sobering to me, because I’m a long way from being able to put this kind of money to work. But wait a second – Lending Club says that 99% of investors with more than 100 notes have positive returns, and 100% of investors with 800+ notes make money. What is our analysis missing?

For one, Lending Club is making a distinction between an average rate of return, and a positive return. If we want to do the same, we can raise our confidence level to 99% and also increase our margin of error to the point where the low end of our returns for any note grade equals zero. And you can see from the table below, using these figures we can go up to a maximum confidence interval of 5.06% (the Adjusted NAR of the A note grade) until we hit zero. But even then, that would be the risk for portfolios that just invested in A notes. Assuming we invest across more than that grade, our risk drops as the returns of other notes makes up for any potential losses. At this point though, I’m sadly at the limit of my statistical skills. I can say though, that I think Lending Club has a combination of cautious investors and luck to get the results they have based on the tables above. That is, lenders tend to construct larger, broader portfolios, and there just haven’t been enough investors over enough time to create an outlier of bad decisions and circumstances of a large portfolio that loses money.

Note Grade | Population | Sample Size Required | Investment Required |
---|---|---|---|

A | 29,148 | 635 | $15,894 |

B | 61,843 | 643 | $16,079 |

C | 48,624 | 641 | $16,034 |

D | 27,278 | 634 | $15,871 |

E | 12,439 | 617 | $15,443 |

F | 5,672 | 583 | $14,580 |

G | 1,155 | 416 | $10,403 |

Since our goal isn’t to avoid losing money, but to make the average return, let’s get back to our original question – what does it take to be diversified? Looking around the blogosphere, there are some other valid analyses we can review. For example, Orchard Platform has an interesting article on portfolio simulation using Monte Carlo analysis (running lots of simulations), which shows that after 200 notes, the standard deviation (volatility) of a portfolio starts to decline, which they recommend as the minimum investment. In this case, the platform they looked at was Prosper, but given the similar platform and market of notes, I think we can extend the conclusion to Lending Club as well. At 750 notes, the volatility basically flatlines.

On LendAcademy, Peter Renton concludes that it takes 500 notes to be fully diversified – again, mainly with the goal of ensuring a positive return vs. hitting the average return. Also interesting about the analysis on that post though, is that there seems to be little value in buying more than 750 notes, as the ROI range looks to stabilize, and there’s no benefit to be had by adding more notes, at least in terms of reducing portfolio risk and the impact of defaults.

The conclusion I’m reaching with my own portfolio is two-fold – first, I don’t have nearly enough money to build a portfolio that statistically represents and ensures my returns fall near the average. But secondly, I can pretty much guarantee positive returns with a portfolio as small as 200 notes, and my target of notes to hold should be 750 before I increase the average investment.

As a side note, investors that are new to the p2p lending space should pay particular attention to the results and volatility of smaller portfolios on Peter’s table – at 50 – 100 notes, there are lots of investors who look to have lost money, and some have faced some really painful losses. Orchard’s analysis is similar and basically says don’t invest in the p2p lending space unless you can fund 200 notes, or $5,000.

Copyright © 2017. All Rights Reserved. Powered by WordPress. Designed by MageeWP Themes