A Breakdown of the Aggressive and Conservative Slider

[Note: The numbers in these tables are fictitious and are used only as examples for this article.]

One of LendingRobot’s most useful settings is it’s “Fully Automated” mode. For a full description of the Fully Automated mode, see our article here. For now, we’ll be discussing the LendingRobot scoring mechanism, or how we select notes for a portfolio. The Fully Automated mode automatically buys the best notes for a portfolio based on a client’s risk tolerance. The risk tolerance is selected by moving a button on a slider that ranges from “conservative” on the left side to “aggressive” on the right. But what constitutes a “conservative” note versus an “aggressive” note? Today we are going to delve into the mechanics of the LendingRobot scoring mechanism to share what it really means when we say ‘aggressive’ or ‘conservative’. 

Each marketplace releases loans at specific times throughout the day for the primary market. Shortly after each release, LendingRobot’s aggressive algorithm is run and each loan is assigned a score (from highest to low). Just like the scoring in Olympic diving, higher scores are better. Unlike Olympic diving, negative scores are possible. As one would imagine, loans with negative scores are automatically rejected from being selected, regardless of the quantity of loans released in any particular round. Certain grades get penalized depending on what type of strategy you decide on.

We’ll start with an aggressive strategy for our first example.

Loan #1 Loan #2 Loan #3
Grade A3 G5 D2
Aggressive LendingRobot Score 3 -3 9
Possible Selection Yes/No No Yes

We see that three loans are released in this round. LendingRobot selects the top 25% of loans in any given round. As mentioned above, Loan #2 would be rejected from the selection pool because of its negative score. This leaves two loans available for selection. Between Loan #1 and #3, #3 has a higher score and would be selected for investment.

Even though LendingRobot typically selects only the top 25% of notes, there are some instances in which Loan #1 would also be purchased. The prime example is if a client has significant cash to invest, but the platform is releasing very few notes. The algorithm calculates the effect of diversification versus cash drag (the effect of non-invested cash on overall portfolio return) and allocates appropriately.

Our second example is a conservative strategy. 

Loan #1 Loan #2 Loan #3
Grade A3 G5 D2
Aggressive LendingRobot Score 3 -3 9
Conservative LendingRobot Score 3 -3 -2
Possible Selection Yes No No

Loan #2 is still weeded out by our algorithm, while Loan #3 is penalized for its grade, which results in a negative score and causes the loan to be rejected from consideration.

And finally, our last example is a mixed conservative/aggressive strategy. As most of our clients are neither fully aggressive nor conservative, but a combination of the two. By selecting an expected return, you are effectively selecting a portfolio that mixes “x” percent of aggressive loans with “1-x” percent of conservative loans to hit the target return.

Loan #1 Loan #2 Loan #3
Grade A3 G5 D2
Aggressive LendingRobot Score 3 -3 9
Conservative LendingRobot Score 3 -3 -2
Possible Selection Yes No Yes

Just like the previous tables Loan #2 gets weeded out by our algorithm as it has a negative score. For example, an targeted expected return of 8% may be a mix of 40% conservative loans and 60% aggressive loans. The LendingRobot algorithm will allocate \$400 towards the conservative strategy and \$600 towards the aggressive strategy. While the portfolio composition may change in any given round (a marketplace platform may not release enough loans to evenly match all strategies), the algorithm will shift its purchasing habits to attempt to maintain the portfolio’s overall target composition.

In regards to understanding the estimate below your Expected Return, it is our confidence intervals in which we take into account a portfolio of 10 loans. With the 95% confidence interval the lower bound is the worst performing 2.5% of 10 loan portfolios and the upper bound is the best performing 97.5% 10 loan portfolios. It is much more likely to have a positive return if you are diversified in more loans.

2 Comments

  1. Ben says:

    Why is loan #2 assigned a negative score in all of the scenarios?

    1. Vanessa Hoying Vanessa Hoying says:

      Loan #2 is assigned a negative score in all the scenarios because our algorithm sees it as a bad return.

Leave a Reply to Ben
Cancel Reply