Spanish Banker Database (SBDb)

The Spanish Banker Database (SBDb), version 2.0, documents the governing bodies of Spanish deposit-taking institutions between 1974 and 2018. Institutions are identified using their code in the Bank of Spain's Register of Institutions together with a legal-entity identifier, which defines the population on a registry basis rather than by membership of a sector association. This makes it possible to distinguish institutional entities, follow changes in legal form and name, and, through the shared register code, trace institutional continuity across mergers and conversions. Linkage information derived from external business and political records is provided to demonstrate that individuals in SBDb can be connected to other data universes. These linkages illustrate that capability and do not constitute complete populations of business or political relationships. Coverage - Commercial banks (AEB): since 1974 - Savings banks (CECA): since 1983. From the restructuring of the sector after 2010, this category also covers institutions that transferred their financial activity to commercial banks while remaining affiliated to CECA; they are recorded under CECA throughout. - Credit cooperatives (UNACC): since 1995 - 663 institutions, of which 648 have at least one observed governing body - 21,160 individuals - 135,883 governance records, including 914 exact duplicate rows arising from the construction process - 134,969 unique individual–institution–year–position observations The starting years reflect the first year of systematic publication of the corresponding yearbook, rather than the first year in which the institutions existed. The 1997 UNACC yearbook could not be consulted; that year was reconstructed from adjacent years and is less complete than the rest of the credit-cooperative series. Sources and methodology Governance data were compiled from yearbooks published by the AEB, CECA and UNACC through manual data entry, direct retrieval from digital records and OCR digitisation. Names were harmonised using a systematic disambiguation protocol and linked across sources through exact and distance-based string matching: Damerau–Levenshtein distance for political records and Jaro similarity for business records. External linkages draw on records for 455,683 individuals appointed or elected to legislative and representative office between 1943 and 2018 and on 4,241,726 company directors recorded in SABI (INFORMA; Moody's–Bureau van Dijk), based on information from the Spanish Commercial Registry. Data protection Individuals are pseudonymised: the deposit contains stable internal identifiers but no personal names. This preserves the principal analytical uses for which the database was designed while limiting the disclosure of directly identifying information. File structure The deposit consists of six related data tables in semicolon-separated CSV format, encoded in UTF-8 with BOM, together with a codebook documenting the variables, reference values, known limitations and linkage rules. Each table has an explicit level of observation: - bank_v2.csv — one row per institution and year. It contains the Bank of Spain register code (idbde), legal-entity identifier (idcompany), composite institution identifier (idbankunique), sector association, institution type and name as reported for that year. - bankrelationship_v2.csv — intended level of observation: individual × institution × year × position. It links idbankunique, IdBanker and IdBankPosition. The file contains 914 documented exact duplicate rows; users producing counts should deduplicate on the complete set of four columns. - bankexternalrel_v2.csv — one row per individual, year and linked company. It contains binary indicators of business (has_business_rel) and political (has_political_rel) linkages, linkage type (relation_type), harmonised company name, NACE Rev. 2 code and one source name prior to harmonisation (companyname_raw). Where several source records resolve to the same individual–year–company key, one source string is retained and alternative raw variants are not preserved. - banktype_v2.csv, bankassociation_v2.csv and bankposition_v2.csv — reference tables for institution type, sector association and position category. The three main tables have different levels of observation. Joining them directly without first aggregating to the required analytical level will multiply rows. The accompanying codebook explains the appropriate linkage procedures.