Management Accounting Reports Supporting Executive Performance Evaluation

Abstract

This research introduces a novel, cross-disciplinary framework for executive performance evaluation by integrating principles from computational linguistics, network
theory, and behavioral economics into traditional management accounting report structures. Moving beyond conventional financial metrics and balanced scorecards, we propose the Dynamic Semantic Performance Evaluation (DSPE) model, which analyzes
the qualitative narratives within management accounting reports—such as managerial
commentary, risk assessments, and strategic initiative descriptions—to extract latent
performance signals. The methodology employs a hybrid approach combining semantic
role labeling adapted from natural language processing, sentiment trajectory analysis,
and coherence network mapping to quantify executive strategic thinking, communication clarity, and alignment with organizational objectives. We developed a proprietary corpus of 150 anonymized management reports from diverse industries spanning
1998-2004 and applied the DSPE model to evaluate the reports’ supporting role in
performance assessment. Results demonstrate that DSPE-derived metrics, particularly ’strategic coherence density’ and ’narrative foresight alignment,’ provide significant explanatory power for subsequent organizational outcomes, explaining 42% of
the variance in long-term firm adaptability measures, a substantial improvement over
traditional variance analysis (15%). The model successfully identified high-performing
executives who were overlooked by purely quantitative metrics and flagged potential
strategic myopia in cases where financials appeared strong. This research contributes
originality by fundamentally re-conceptualizing management accounting reports not as
passive repositories of financial data but as active, structured narratives that encode
executive cognitive patterns and strategic capability. It offers a novel methodological
bridge between accounting practice and computational text analysis, providing a more
holistic, dynamic, and forward-looking tool for governance and evaluation.