Original research · BORME · Provisional edition

Who holds the seats: concentration and fragmentation in Spain’s corporate administration

987,436 companies with a BORME filing in the last five years and 1,426,380 company–director pairs in 2025, reconstructed as one national graph. The 1% of directors with the most seats gain ground, from 8.44% to 9.46%; in Madrid the rise is three times larger, from 8.44% to 11.62%, while Barcelona does not move. The network of shared directors stays fragmented everywhere.

9.46%
of seats in Spain held by the top 1% of directors in 2025 (8.44% in 2013)
11.62%
in Madrid, the province with the largest rise (8.44% in 2013)
8.5%
of Spanish companies belong to the largest connected component, against 24.9% expected at random
01 · Concentration

A larger share at the top, above all in Madrid

Each company–director relationship counts once. The charts follow reconstructed year-end snapshots of recently filing companies; Spain is one graph, not the sum of its provinces.

In Spain, the 1% of directors with the most seats held 8.44% of observed seats in 2013 and 9.46% in 2025, 1.02 points more. The Gini index of seats per director moves from 0.257 to 0.275.

Madrid holds 23% of the pairs and accounts for most of the movement: there the top 1% goes from 8.44% to 11.62% (3.18 points; Gini from 0.262 to 0.308). Excluding seats whose start was never observed, the national direction holds: 7.80% to 9.41%.

2013 20252025
Spain · BMadrid · B
Share of seats held by the 1% of directors with the most seats, recently filing companies (variant B), 2013–2025. Solid line: Spain as one graph. Dashed line: Madrid. Move over the chart or use the slider to pick a year.
What this does not establish. Concentration alone cannot distinguish professional service providers from business groups, ownership vehicles or changes in governance structure. Without the tail (corporate officers and directors holding more than 20 seats nationwide at that date) the national Gini moves from 0.233 to 0.239: most of the movement is in the tail.
02 · Provinces

The same lens, province by province

Every province is reconstructed with the same rules and the same year. A director who links a Madrid company to a Barcelona company counts in the national graph and in neither province, which is why Spain is not the sum of its rows.

Madrid gains 3.2 points between 2013 and 2025; Barcelona is flat and 20 of the 44 provinces with at least 2,000 companies fall. The table recomputes the comparison for the selected year.

Population B, selected year. * Fewer than 2,000 companies: three random draws only, read with care. Click a heading to sort.
Province comparison for the selected year. The top 1% share and the largest component scale with population size: compare the observed ÷ random ratio and the Gini across provinces, not the raw component share. The administrative-closure rule is published very unevenly across registries (last column).
03 · Connectivity

Connectivity has limits

In 2025, 47.6% of Spanish companies share a director with another company, but only 8.5% belong to the largest connected component (9.0% in 2013). Seats per company fell from 1.51 to 1.45.

The observed component is 36% of its random benchmark in 2013 and 34% in 2025. The network is persistently fragmented relative to that benchmark; it is not becoming more fragmented. Madrid's component is larger, 10.3% against 34.1% at random, with the same ratio.

The largest connected component

Companies connected through chains of shared directors, as a share of all companies in the selected population.

0%10%20%30%40%
The benchmark keeps every company’s and every director’s number of seats and rewires who sits where. The range shows the draws (five for Spain, ten for Madrid); it is not a confidence interval.
04 · Explorer

Change the lens

Two populations, thirteen annual snapshots, five population definitions and nine measures, all from the same dated panel.

2013
2025
Change since 2013
YearPairsCompanies¹OfficersPairs / companyTop 1%GiniInterlockedLargest component
¹ Companies retaining at least one pair. For TC and TL, graph percentages use all B companies, including companies left with no retained pairs. Officers may be natural persons or corporate administrators.
05 · Method

The evidence, with its limits

What is a seat, and who is included?

A seat is one distinct company–officer pair at a year-end snapshot, reconstructed from the appointments, re-elections, cessations and revocations published in BORME since 2009. Several governance roles held by the same person at the same company count once. Officers include both natural persons and corporate administrators.

The extraction targets director-role prefixes (administrators, board members, chairs, managing directors). Powers of attorney, auditors and liquidators are out of scope. A cessation naming a role with no matching open seat closes the most recently opened seat of that pair; a cessation with no prior history is treated as a seat that predates the window and flagged unknown-start.

How do the five population variants differ?

A: all open pairs, excluding dissolved, extinct or deregistered companies. B: A plus a filing in the current or preceding four years and no unlifted administrative closure. BX: B without unknown-start seats. TC: B without corporate officers and without directors holding more than 20 seats nationwide at that date. TL: as TC, but using the lifetime count of companies per director (uses future information).

For TC and TL, companies left with no pairs stay as isolated graph nodes. The national graph and the provinces use the same rules as the Madrid panel.

How much history is missing?

Records start in 2009. In the national B population, 18.2% of pairs in 2013 and 2.3% in 2025 are seats whose start was never observed. Starting the series in 2013 does not eliminate that left-censoring; BX is a sensitivity check, not a correction.

The five-year recency rule can exclude functioning companies with unchanged governance. Recent filing is not proof of economic activity.

How should the charts be read?

The top 1% share is the fraction of pairs held by the 1% of directors with the most seats. Gini measures inequality in pairs per director. Interlocked companies share at least one director with another; the largest connected component also includes indirect chains. These links do not establish ownership, common control, family ties or misconduct.

The random benchmark fixes both degree sequences (pairs per company and per director), rewires the edges without duplicates and applies one full round of degree-preserving swaps. Ten draws for Madrid, five for the national graph and the provinces.

What remains open?

Administrative closures (tax and NIF) are published very unevenly across registries: they touch 0.2% of Madrid companies against more than 30% in Las Palmas or Tarragona. The rule is close to inert in Madrid and not comparable across provinces. The province is the company’s registry province; links between provinces appear only in the national graph.

Validation samples against source filings, an appointment-flow analysis (where new seats go) and incorporation cohorts are planned. No “professional administration” mechanism has yet been established.

Data and limitations

  • BORME records registered legal acts, not a company’s economic situation. Everything above is descriptive: no causal claim is made.
  • Person identity is reconstructed from the published name; BORME carries no personal identifiers. Namesakes can inflate large portfolios and connectivity, more so in a graph of a million companies than in one province.
  • The incomplete year 2026 is excluded. Values before 2013 are not shown because the observation window was still filling.

How to cite this study

Mapa Societario (2026). Who holds the seats: concentration and fragmentation in Spain’s corporate administration. Study based on Spanish commercial-registry (BORME) data, September 2026. https://mapasocietario.es/en/studies/spain-directorships/

You may reproduce these figures with attribution.

Download the data (CSV)

Explore any Spanish company

The study looks at the whole. The tool does the same for a single company: search a company or a director and explore their links in an interactive graph.

Open the search →