Chinese Accounting Firms (CAFs)

Published: 26 November 2024| Version 1 | DOI: 10.17632/h5x2yc2wbh.1
Contributor:
Tariq H. Malik

Description

This study investigates the relationship between Board Gender Diversity (BGD) and firm performance in Chinese Accounting Firms (CAFs), focusing on the mediating role of concrete versus abstract business narratives. Below is a summary of the key elements, findings, and implications: Research Hypotheses Hypothesis 1: BGD positively correlates with CAF performance as measured by Return on Human Capital (ROHC). Hypothesis 2: The concreteness of a firm's narrative positively correlates with ROHC. Hypothesis 3: Concrete narratives mediate the effects of BGD on ROHC. Data and Methodology Sample Size: Data from 2,222 Chinese Accounting Firms (CAFs) with at least 10 employees, located across 286 cities in China. Data Source: Orbis database by Bureau van Dijk, covering NACE 6920 (accounting/auditing activities). Performance Metric: Return on Human Capital (ROHC), calculated as revenue per employee. Narrative Analysis: Firm narratives rated on a 10-point scale from abstract (1) to concrete (10). Key Models: Multilevel mixed-effects models and Ordinary Least Squares (OLS) regression, incorporating interaction terms and robustness tests. Notable Findings Direct Effects: BGD significantly improves ROHC (β = 0.3794, p < 0.01). Concrete narratives positively impact ROHC (β = 0.4058, p < 0.01). Interaction Effects: The interaction between BGD and concrete narratives significantly enhances ROHC (β = 0.0197, p < 0.01). This suggests a multiplicative relationship, where BGD's effect on ROHC is amplified in the presence of concrete narratives. Sensitivity Analysis: Variations in BGD and narrative concreteness (e.g., increases by 10%, 20%, or 50%) consistently show positive effects on ROHC, maintaining significance. Robustness Checks: The models explain a substantial variance in ROHC (R² = 0.677; Adjusted R² = 0.675). Lower Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) values indicate strong model fit. Implications Theoretical Advances Construal Level Theory (CLT) by empirically linking communication styles (concrete vs. abstract narratives) to firm performance. Highlights the role of female directors in promoting concrete communication styles, which enhance operational efficiency and client relations. Practical Encourages firms to integrate BGD into their governance structures. Recommends prioritizing concrete narrative styles to strengthen trust, transparency, and client retention.

Files

Steps to reproduce

Steps for Data Collection, Organization, and Analysis 1. Data Collection Source: Orbis database by Bureau van Dijk, a global dataset providing detailed firm-level information. Filter firms using NACE Code 6920, which pertains to accounting, auditing, and tax consultancy. Focus on Chinese Accounting Firms (CAFs). Firm Selection Criteria: Firms with at least 10 employees to ensure inclusion of structured organizations. Cross-sectional data collected for the year 2020, covering financial and operational metrics. Variables Collected: Dependent Variable: Return on Human Capital (ROHC): Calculated as revenue divided by the number of employees. Independent Variables: Board Gender Diversity (BGD): Proportion of female directors on the board. Narrative Concreteness: Text-based rating of business descriptions (scale of 1-10). Control Variables: Firm age, board size, revenue, number of employees, geographic location, ownership (state or private), and client type. Firm Narrative Data: Extract textual descriptions of firms’ operations from the Orbis database. Translate non-English descriptions into English for consistency. Rate narrative concreteness based on detail level, using a scale: 1 (Abstract): Vague, high-level descriptions. 10 (Concrete): Detailed, specific, and actionable descriptions. 2. Data Organization Variable Construction: Dependent Variable: ROHC = Revenue ÷ Employees. Independent Variables: BGD, Narrative Concreteness. Interaction Variable: BGD × Narrative Concreteness. Data Cleaning: Standardize text size for narrative analysis. Data Categorization: Control Variables: Include categorical (e.g., geographic location) and continuous (e.g., revenue, age) variables. Group firms by geographic diversity across 286 cities. 3. Data Analysis Descriptive Statistics Summarize key metrics (means, standard deviations, ranges). Examine correlations between variables using a correlation matrix. Statistical Models Base Model: Test the relationship between BGD and ROHC. Narrative Model: Assess the effect of narrative concreteness on ROHC. Full Model: Include interaction effects (BGD × Narrative Concreteness) to test for synergy. Control Inclusion: Add control variables for robustness. Regression Techniques OLS Regression: Test the linear effects of BGD and narrative concreteness on ROHC. Interaction Analysis: Examine combined effects of BGD and narrative concreteness. Multicollinearity Check: Use Variance Inflation Factor (VIF) to ensure no high correlation among predictors (threshold < 10). Model Fit: Evaluate model performance using R², Adjusted R², Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC). Graphical Analysis Correlation Plots: Show linear/nonlinear relationships. Interaction Plots: Depict effects of combined predictors (BGD × Narrative Concreteness) on ROHC. Distribution Plots: Validate normality and log transformations of variables.

Categories

Management, Firm Organization, Effect of Gender, Gender, Accounting Profession, Firm Performance

Licence