1. Overview

This dataset provides annual board-of-directors composition data for Community Interest Companies (CICs) registered with Companies House. It captures the size, turnover, diversity, and tenure characteristics of CIC boards across the full history of the CIC legal form.

The dataset is produced by the UK Third and Civil Society Sector Database project, which assembles open administrative data into research-ready panel datasets covering charities, community interest companies, and other civil society organisations.

Each row represents one organisation in one financial year, capturing board size, director turnover, gender diversity, nationality diversity, age composition, tenure, and founding board retention. The data is derived from Companies House officer records and aligned to financial years using the XBRL accounts filing history.

2. What are Board of Directors Records?

Community Interest Companies (CICs) are a legal form designed for social enterprises that use their assets and profits for the benefit of the community. CICs are subject to a statutory community interest test, an asset lock that prevents private extraction of value, and regulatory oversight by the CIC Regulator. Directors of CICs bear responsibility for maintaining the company’s community purpose, making board composition a key indicator of organisational governance and capacity.

Like all UK limited companies, CICs are legally required to register their officers (directors and secretaries) with Companies House. This public register records each officer’s appointment date, resignation date (if applicable), nationality, date of birth (month and year), and occupation. These records form the basis of this dataset.

Companies House Officers Data

The Companies House officers API provides structured data on every person who has served as a director or secretary of a registered company. Key fields include the date of appointment, date of resignation (where applicable), the officer’s nationality, their date of birth (year and month only), and their stated occupation. This information is publicly available and updated as companies file changes.

From Officers to Board Panels

To produce this dataset, individual officer records are mapped to financial years using an inclusive overlap approach: a director is counted as serving in a given financial year if any part of their tenure overlaps with any part of that financial year. Officers are then aggregated to produce board-level summary metrics for each organisation in each year.

This transformation converts person-level appointment records into an organisation–year panel, where each row summarises the composition of the board during a specific financial year. Metrics include the number of directors, rates of appointment and resignation, gender and nationality proportions, age statistics, tenure measures, and founding board retention.

3. Dataset Contents

Field Descriptions

Sample Data

The table below shows the first five rows of the dataset. Numeric values are rounded to two decimal places for display.

Field Coverage

The following fields have less than 100% coverage. Missing values typically reflect organisations that had no resignations, appointments, or officers with identifiable characteristics in a given financial year.

4. Coverage & Completeness

The dataset covers 41,401 unique CICs across financial years 1992–2026, with 186,338 organisation–year records in total. The coverage of each field is shown below.

Figure 1. Number of organisation-year records by financial year. Source: UK Third Sector Database, Board of Directors dataset.

5. What Can You Learn?

This dataset enables a wide range of research into the governance and leadership of community interest companies. By providing annual board composition metrics across the full history of the CIC legal form, it supports both cross-sectional and longitudinal analyses of board dynamics.

Example research questions include:

• How has CIC board size changed over time?
• What is the gender composition of CIC boards, and has it improved?
• How long do founding directors remain on CIC boards?
• What is the relationship between board turnover and organisational age?
• How does nationality diversity on CIC boards compare across regions or sectors?

Figure 2. Mean board size by financial year. Source: UK Third Sector Database, Board of Directors dataset.

Figure 3. Mean proportion of female directors by financial year. Source: UK Third Sector Database, Board of Directors dataset.

Figure 4. Mean proportion of founding board members still serving, by financial year. Source: UK Third Sector Database, Board of Directors dataset.

Figure 5. Distribution of board sizes across all organisation-year records (values above 20 grouped). Source: UK Third Sector Database, Board of Directors dataset.

6. Limitations & Caveats

Gender Classification

Gender is inferred using a three-tier classification system: first by title (e.g. Mr, Mrs, Ms), then by forename matching against national name databases, and finally cross-validated using the UK Social Security Administration (SSA) name data. The proportion of directors with unknown gender is excluded from the denominator when computing prop_female. This approach may misclassify directors with non-Western names or gender-neutral forenames.

Age Approximation

Companies House records only the year of birth for company officers (month is also available but not used here). Age is therefore calculated as the financial year end year minus the birth year, which can be off by ±1 year depending on the officer’s actual birthday relative to the financial year end date.

Nationality

Nationality is self-reported by officers to Companies House and may not reflect current citizenship status. Many officer records have missing nationality data. For the purpose of computing prop_non_british_irish, British and Irish nationality variants are grouped as “domestic” and officers with unknown nationality are excluded from the denominator.

Financial Year Alignment

An inclusive overlap approach is used to assign officers to financial years: an officer is counted as serving in a financial year if any part of their tenure overlaps with any part of that financial year. This means an officer appointed on the last day of a financial year, or who resigned on the first day, is still counted as serving in that year.

Founding Board Definition

The founding board is defined as all directors appointed on the earliest appointment date for a given company. This date serves as a proxy for the incorporation date.

What is NOT in the Data

• Individual officer names or identifiers — data is aggregated to organisation level for disclosure protection.
• Charities without Companies House registration — The trustees of unincorporated charities can be downloaded from the CCEW data download portal, however this is a snapshot, not a longitudinal record: register-of-charities.charitycommission.gov.uk.
• Non-CIC companies — this dataset covers Community Interest Companies only.
• Companies without XBRL accounts filings — 29,824 CICs with officer data but no matched financial year records are excluded. See the Coverage section for details.

7. Citation & Licence

Suggested Citation

McDonnell et al. (2026). TCSS CIC Board of Directors Dataset. UK Third and Civil Society Sector Database. Available at: https://uk-third-sector-database.github.io/data/. Licensed under CC BY 4.0.

If you would like to learn more about this dataset and how it can be applied to your project or research programme, please contact research@brawdata.com.

8. Changelog

A1. Pipeline Architecture

The Board of Directors dataset is produced through a three-stage pipeline that transforms individual officer records from Companies House into an organisation–year panel of board composition metrics.

1. Officer Data Preparation — individual officer records are cleaned, standardised, and enriched with gender classifications and founding board flags
2. Financial Year Alignment — officers are mapped to the financial years during which they served, using an inclusive overlap rule against XBRL accounts data
3. Board-Level Aggregation — officer–year records are grouped by organisation and financial year to produce board-level summary metrics

A2. Officer Data Preparation

Date Standardisation

• Company numbers are zero-padded to 8 digits (preserving letter prefixes like SC, NI, OC)
• Appointment and resignation dates are parsed to standard format; appointed_before is used as a fallback when appointed_on is missing

Gender Classification

A three-tier priority system is used:

1. Title-based (Tier 1): Mr → male; Mrs/Ms/Miss/Lady/Dame/Baroness → female; Dr/Prof/Rev/Sir/Lord → ambiguous (falls through)
2. Forename-based (Tier 2): Uses the gender_guesser Python package; male/mostly_male → male; female/mostly_female → female; unknown/androgynous → unknown
3. SSA Cross-Validation (Tier 2b): US Social Security Administration baby-name data provides an independent robustness check; stored in a separate column but does not override the primary classification

Priority: title wins over forename; the gender_final column reflects the title-then-name cascade.

Founding Board Identification

• The earliest appointed_on date per company serves as a proxy for incorporation date
• All officers appointed on that date are classified as founding board members
• founding_board_size is computed as the count of founding members per company

Forename Extraction

• Forenames are extracted from the Companies House appointments dataset by joining on officer_id
• Where multiple appointments exist for the same officer, the first non-null forename is retained
• Corporate officers (those flagged as is_corporate_officer) are identified and flagged

A3. Financial Year Alignment

Building the Organisation–Year Scaffold

• Financial year records are drawn from the XBRL accounts dataset (Companies House iXBRL filings)
• Deduplicated to unique company-number / financial-year-end pairs, retaining uid, coyno, fy, fys (FY start), and fye (FY end)

Overlap-Based Matching

Officers are matched to financial years using a vectorised inner join on company number, followed by an inclusive overlap filter:

• appointed_on <= fye AND (resigned_on is missing OR resigned_on >= fys)

This means an officer is counted as active in a financial year if any part of their tenure overlaps any part of the financial year period. For each officer–year match, the pipeline computes:

• tenure_at_fye_years: (fye − appointed_on) ÷ 365.25
• appointed_in_fy: True if appointed_on falls within [fys, fye]
• resigned_in_fy: True if resigned_on falls within [fys, fye]
• age_at_fye: fye.year − dob_year (approximate)

Unmatched Companies

A4. Board-Level Aggregation

The final stage groups officer–year records by organisation and financial year to produce one summary row per organisation per year.

A5. Linking to Other TCSS Datasets

The uid field provides a direct link to all other datasets in the TCSS ecosystem. All records in this dataset use the GB-COH- prefix format.

A6. Reproducibility

• Pipeline code is available in the project repository at github.com/uk-third-sector-database/tso-database-builder, in the boards-of-directors subdirectory
• Requires Python 3.12+, dependencies managed with uv
• Key dependencies: pandas, gender_guesser
• All paths and parameters are defined in a central config.py file