Leverage Model

The Leverage Model determines how much capital the Senior Pool allocates toward each Borrower Pool, based on how much it trusts each Borrower Pool. Currently, the automated Leverage Model as described here is not yet live, but it was included in the community-approved roadmap for near-term development linked here.

Trust Through Consensus

In order to determine how to allocate capital from the Senior Pool, the protocol uses the novel principle of "trust through consensus." This means that while the protocol doesn't trust any individual Backer or Auditor, it does trust the collective actions of many of them. At a high level: when more Backers supply capital to a given Borrower Pool, the Senior Pool increases the ratio with which it adds leverage.

Because this approach relies on counting individual Backers, the protocol must ensure they are in fact represented by different people. Therefore, all Backers, Borrowers, and Auditors require a unique entity check to participate (see the Unique Entity Check section).

Leverage Model Formula

The leverage amount, AA, that the Senior Pool allocates is determined by the formula, A=SDLA = S * D * Lwhere:

  • SS is the total capital supplied by Backers.

  • DD is the distribution adjustment on a scale of 00 to 11, which accounts for how evenly distributed the Backers are. DD is closer to 00 when the distribution is skewed and closer to 11 when the Backers are more equally distributed. This ensures no single Backer has an outsized influence. The formula for DD uses the percent supplied by each Backer, sns_{n} , and is based on the Herfindahl-Hirschman Index:

D=1i=1nsn2D = 1-\sum_{i=1}^n s_n^2
  • L is the leverage ratio on a scale of 00 to the maximum potential leverage ratio. Based on the number of Backers, bb, the leverage ratio increases linearly from BminB_{min} , the minimum number of Backers necessary for leverage, to BmaxB_{max} , the maximum number of Backers necessary to achieve the maximum potential leverage, LmaxL_{max}:

L=Lmaxmax(0,bBmin)BmaxBminL=L_{max}*\frac{max(0, b-B_{min})}{B_{max}-B_{min}}

Last updated