Thesis of Wisnu Uriawan


Subject:
Trustworthiness for Personal Lending on Blockchain

Defense date: 02/12/2022

Advisor: Lionel Brunie
Coadvisor: Youakim Badr, Omar Hasan

Summary:

Individual borrowers request loans for personal projects or urgent requirements. Small and medium-sized enterprises (SMEs) need loans to scale up their businesses [1]. Most loans given by traditional banks are secured loans, which have terms and conditions that are sometimes not easy for borrowers to fulfill. Banks require collateral or guarantors to guarantee that borrowers return their loans. Collateral can be in the form of assets (i.e., houses, vehicles, savings, deposits, and securities) [2]–[4]. A guarantor is a person who gives some guarantee for the person or SME applying for loans [5], [6]. With a secured loan, the lender can take over an asset provided as collateral if the customer cannot make the loan payments. In addition, many documents may be needed during the loan application process. Administrative costs may be required at the time of application submission as well. There is also little visibility in the centralized process, so the borrower does not know the clear reasons for being accepted or rejected. Moreover, the traditional loan application process is time-consuming.

On the other hand, Blockchain technology is emerging and is successfully applied in many financial applications [7], including Peer-to-Peer lending [8], [9]. In 2020, peer-topeer consumer lending in Europe (excluding the U.K.) reached 2.9 billion U.S. dollars [10]. P2P lending platforms allow lenders and borrowers to connect without going through a bank. Some of the problems of traditional lending such as burdensome administrative processes can be avoided in P2P lending. A significant increase in P2P lending is estimated to be close to one trillion U.S. dollars by 2050. Today many lending platforms are available supported by Blockchain technology (such as, SALT, BlockFi, ETHLend, etc.), however, they still require collateral [11], which does not help borrowers avoid this burden.

In this thesis, we propose that borrowers’ trustworthiness be used as an alternative in lending applications so that borrowers are no longer burdened with collateral or guarantors. The problem is how to reliably calculate trustworthiness. Another problem is how to enforce the conditions of lending and borrowing in a P2P environment. In this thesis, we make several contributions in these areas, which are summarized below.

As a first contribution, we build a borrower trustworthiness scoring model called LAPS. The LAPS formula computes the borrowers’ trustworthiness score based on variables such as loan risk, user activity, user profile, and social recommendations. In the LAPS model, we introduce social recommendations in support of a borrower. The advantage of the LAPS score is that it may help to convince the lenders/investors to provide loans with minimal or no collateral. They get assurance that the borrowers are able to pay back the loans, and the recommender bridges the gap between the borrower and the lender.

Our second contribution is to develop smart contracts that enforce the conditions of lending and borrowing in a P2P environment. The smart contracts are program codes powered by Solidity and Ethereum blockchain that can be realized in a decentralized P2P manner. They can replace the intermediary process between users (borrowers, recommenders, and lenders/ investors) and eliminate the third party. We use these smart contracts to enable interactions between users (borrowers, recommenders, and lenders/investors) in a secure way where all agreed upon rules are enforced. The smart contracts implement the LAPS trustworthiness score formula.

As the third contribution, we introduce TrustLend as a personal lending platform and present its design and implementation details. We develop a fully functional prototype of TrustLend as a decentralized application (DApp). The prototype shows the features required by borrowers, recommenders, and lenders/investors to rely on trustworthiness by implementing trustworthiness scores with Ethereum smart contracts. TrustLend deploys the smart contracts and applies the LAPS formula to compute the borrowers’ trustworthiness scores. We describe the prototype architecture and conduct various personal loan simulations.         

TrustLend presents architectural designs that meet the needs of borrowers, recommenders, and lenders/investors, such as permissionless access, integrity, and security. In addition, the platform aims for several other properties: Transparency: all system transactions are traceable and accountable. Automatic enforcement of terms: autonomous transactions by smart contracts are binding for all participants. Reduced costs: we attempt to minimize gas costs for feasible economics. Keywords: Blockchain, Collateral, DApp, Decentralization, Ethereum, Lending, Smart Contracts, Social Recommendation, Trustworthiness, P2P.


Jury:
M. Kosch HaraldProfesseur(e)Université de PassauRapporteur(e)
M. Mostefaoui AhmedMaître de conférenceUniversité de Franche-ComtéRapporteur(e)
Mme Seba HamidaMaître de conférenceLIRIS Université Lyon 1Examinateur​(trice)
M. Brunie LionelProfesseur(e)LIRIS INSA LyonCo-encadrant(e)
M. Hasan OmarMaître de conférenceLIRIS INSA LyonCo-encadrant(e)
M. Badr YouakimProfesseur(e)Pennsylvania State UniversityCo-encadrant(e)
Mme Sonigo-RehnMaître de conférenceIUT Besancon-Vesoul, Université de Franche-ComtéExaminateur​(trice)
M. Gianini GabrieleProfesseur(e)Université des études de MilanExaminateur​(trice)