University of Twente Student Theses

Login

Estimating Dutch SME EBITDA from public data: A predictive framework for early-stage lead sourcing

Jeurissen, S.G. (2025) Estimating Dutch SME EBITDA from public data: A predictive framework for early-stage lead sourcing.

[img] PDF
6MB
Abstract:Valuing Dutch Small and Medium-sized Enterprises (SMEs) in early-stage M&A processes is challenging due to limited financial disclosures, which typically only include balance sheet data. This thesis, conducted in collaboration with Marktlink, develops a predictive framework to estimate EBITDA solely from publicly available financial data. A sector-specific approach is applied to business services, wholesale, and construction, recognizing that financial structures differ significantly between industries. Using a dataset of Dutch and Belgian SMEs from the Orbis database, multiple regression and machine learning models (Multiple Linear Regression (MLR), Random Forest, and XGBoost) are trained and evaluated. The best-performing model, XGBoost on log-transformed features, achieves a Symmetric Mean Absolute Percentage Error (SMAPE) as low as 4.35% and an accuracy of 85.93%. SHAP analysis highlights total assets, equity, and tangible fixed assets as key drivers of EBITDA. The models significantly outperform common rule-of-thumb benchmarks and are validated through cross-validation and economic consistency via a stylized DCF framework. This research provides M&A advisors with a scalable, data-driven tool to support lead sourcing and prioritize acquisition targets, demonstrating that reliable EBITDA estimations are feasible without full income statements.
Item Type:Essay (Master)
Faculty:BMS: Behavioural, Management and Social Sciences
Subject:83 economics
Programme:Industrial Engineering and Management MSc (60029)
Link to this item:https://purl.utwente.nl/essays/106491
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page