Multiple linear regression for predicting payment time in factoring

Sammanfattning

This project proposal stems from the experience of working with factoring services in a small Swedish financial company. Factoring involves purchasing invoices from companies at a discount, providing immediate liquidity while transferring the risk of collection to the finance company. Currently, factoring fees are typically fixed for a predetermined period, with adjustments rarely made. This study aims to explore the feasibility of using in house historical data to predict when an invoice will be paid. Information about when a particular customer is expected to pay an invoice helps assess and manage credit risk. If a finance company can anticipate late payments, steps can be taken to minimize losses. This can be done by taking action against the customer or the finance company can adjust its credit assessments. Financial companies often use third parties who provide financial data for a fee. The study investigates how companies can instead use historically collected data that is available in-house to predict when an invoice will be paid by using a multiple linear regression model.