Invoice discounting using kelly criterion by automated market makers-like implementations

TL;DR

This paper proposes a decentralized invoice discounting market using an Automated Market Maker and Kelly criterion to optimize funding, demonstrating resilience and profitability across various invoice scenarios.

Contribution

It introduces a novel decentralized approach for invoice discounting using liquidity pools and Kelly-based heuristics, avoiding traditional intermediaries.

Findings

Resilient performance with partially collateralized invoices

Effective profit generation for liquidity providers

Robustness against bogus and nonpayable invoices

Abstract

There is a persistent lack of funding, especially for SMEs, that cyclically worsens. The factoring and invoice discounting market appears to address delays in paying commercial invoices: sellers bring still-to-be-paid invoices to financial organizations, intermediaries, typically banks that provide an advance payment. This article contains research on novel decentralized approaches to said lending services without intermediaries by using liquidity pools and its associated heuristics, creating an Automated Market Maker. In our approach, the contributed collateral and the invoice trades with risk is measured with a formula: The Kelly criterion is used to calculate the optimal premium to be contributed to a liquidity pool in the funding of the said invoices. The behavior of the algorithm is studied in several scenarios of streams of invoices with representative amounts, collaterals, payment delays, and nonpayments rates or mora. We completed the study with hack scenarios with bogus, nonpayable invoices. As a result, we have created a resilient solution that performs the best with partially collateralized invoices. The outcome is decentralized market developed with the Kelly criterion that is reasonably resilient to a wide variety of the invoicing cases that provides sound profit to liquidity providers, and several premium distribution policies were checked that contributed with extra resilience to the performance of the algorithm.