University:
University College London
Course:
LAWS0021 | Company Law
Academic year:

2024
Views:

140

Pages:

3
Author:

Nam B.

An Interdisciplinary Empirical Project Presented by: Dr. Hardman and Dr. Rowell Date: 28 April 2022 Introduction - Interdisciplinary Empirical Research: Law & Data Science. - Social Ties. - Initial aim was to see what we can do. - Started project December 2019, now nearly finished. INEDS - The independent non-executive director is a mainstay of international corporate governance for listed companies. - - Advantages: • Check & Balance on executive directors (prevent opportunism) • Reduces "agency costs" suffered by shareholders • Provide strategic guidance • Signaling that willing to be monitored (& can find monitors) Disadvantages: • Cronyism • In hoc to the system Getting the Data - - Companies House: • Public Facing Platform • Free Company Date Product • Companies House API London Stock Exchange: • Monthly Issuer List (Main Market Only) - 767 Companies, 4986 directorships - Director independence: • - Company publications, websites, regulatory announcements Closed Ended Investment Funds: • AIC members Companies House API - Data resources have unique URLs that are accessible via HTTP requests • Data is provided in JSON format - A Java application was developed to communicate with the Companies House - database via the API - Several types of record are available that are of relevance to this study: • Company 1  Company records o Indexed by unique company number o Records company type, status, name, o Supplemented by AIC membership list and LSE issuers list • Officer 1, Officer 2, Officer 3  Officer records o Records officer name, DoB, address,... inconsistently o Multiplicity was a problem o Resolved ambiguous cases by matching names & DoB where possible • Appointments records  Records appointment of one officer at one company  Records officer role: only directorships were selected  Supplemented by directorship status dataset to distinguish independent / non-independent directorships Method: Graph Theory - The company/officer network has the structure of a mathematical graph - A graph is an object composed of vertices (companies, officer) and edges (appointments) and is used to model pairwise relationships between objects - Representing the company/officer network as a graph allows us to leverage powerful methods from graph theory to inspect the structure - For this purpose the graph theory software library JGraphT was used Results - The complete graph is not fully connected and consists of numerous isolated groups of companies and officers referred to as connected sets - The length of the shortest path between two directors is a useful statistic - Path length = number of intermediate companies between the two directors - The distribution of shortest path length over all pairs of directors quantifies how densely connected the overall graph is - Median path length between two random directors is 6: actually quite long Implications - Data hard to find case for simplification - No evidence of cronyism. But: - • We only included current appointments • We only included director appointments • We did not (could not?) include personal relationships A Director Daisy Chain: • Independent from management, but not from system.