Akademik Veri Yönetim Sistemi
Sustainability and Family Business from the Past to the Present, İstanbul, Türkiye, 11 Kasım 2025, ss.24-25, (Özet Bildiri)
This study investigates the causal impact of
family ownership on firm performance in the Turkish energy sector. The energy
industry is characterised by high capital intensity, strategic importance, and
increasing sustainability pressures, thus rendering it one of the most
competitive markets. In this context, family businesses function not only as
economic entities but also as strategic actors that shape corporate governance
dynamics. It is imperative to accurately identify the causal effect of family ownership
on firm performance to facilitate a comprehensive understanding of competitive
market structures within the sector.
In contradistinction to conventional panel data
techniques, the present study utilises the Double Machine Learning (DML)
framework, an advanced econometric method designed to produce unbiased causal
effect estimates in the presence of high-dimensional covariates. The method
utilises flexible machine learning models to orthogonalize the influence of
complex control variables, including firm size, leverage, investment ratio,
market share, risk indicators, and market concentration. This enables the
estimation of the net causal effect of family ownership on firm performance.
The estimation procedure relies on Neyman-orthogonal scores and cross-fitting
to reduce regularization bias and overfitting, thereby ensuring statistically
valid confidence intervals.
The DML approach's inherent flexibility further
facilitates the analysis of treatment effect heterogeneity, examining whether
the effect of family ownership varies with structural factors such as
competition intensity, firm size, and leverage levels. This provides a more
nuanced understanding of firm behaviour, extending beyond the confines of
average treatment effects, and captures variations in performance across
diverse competitive environments.
The present study makes a novel methodological
contribution by applying DML within the context of family-owned firms in the
Turkish energy sector. In comparison with conventional econometric
methodologies, DML provides more robust causal estimates in settings
characterised by complex, high-dimensional financial data. The results provide
valuable insights for policymakers, investors, and researchers by elucidating
the causal mechanisms through which family ownership influences firm
performance under varying competitive conditions.