Resources and updates
Downloadable resources and updates to the BBCE database are given here. This gives all the background definitions used for BBCE and the Atlas of Entrepreneurship, downloads, and additions in the light of experience by the BBCE research team and other users.
Feedback and suggestions are welcomed, which (with their permission) will be added to this website to give continuing updating. Updates to the User Guide are also welcomed. Please use the contact page to provide information or email Bob Bennett directly.
- Documentation in Working Papers and supplementary data downloads
- Coding to industry sector aggregates (EA17, EA51, SIC; and HISCO, PST)
- Extension of sector classification to Feinstein's sectors
- Urban and town definitions
The downloads here include the total population as published, and the total number of the economically active from the BBCE broken down by sex, for employers, own-account proprietors, and workers. All data here and the Atlas are weighted. The downloads are available by RSD and Scottish parish (from the Atlas), and for county and urban areas and towns (defined below), with breakouts by sector. These downloads aim to help understanding, support use by students and other users to allow experiments with various analysis, and encourage replication and extension of previous work. This should also facilitate use by researchers who want to join up historical trends with modern data for specific geographical units. Research users wanting to develop other spatial units and their own selections of individuals will need to go to the BBCE and I-CeM at the individual level.
Full definitions of how the data are supplied in these downloads and the Atlas are given in a Supplement to the User Guide
Atlas of Entrepreneurship data downloads by RSD and parish
The Atlas of Entrepreneurship page includes downloadable variables for spatial units based on RSDs (E&W) and parishes (Scotland) with linked GIS mappable files will be available shortly. This will allow research users to use much of the spatial codes in BBCE at local level.
Data downloads for urban areas and towns
The BBCE and Atlas uses urban definitions based on research by Chris Law and Brian Robson for the population of each town over 2,500 in size in England and Wales. See: Law, C. M. (1967) The growth of urban population in England and Wales, 1801-1911, Transactions, Institute of British Geographers, 41, 125-43; Robson, B.T. (1973) Urban Growth: An approach, Methuen, London. The Law-Robson original population data are available as a UKDS data download to which has been added John Langton's towns data from the 1680s. BBCE gives entrepreneurs aligned with Law-Robson for towns with a population of 10,000 and over in England and Wales, and additionally gives all Scottish Burghs. The towns and urban areas are displayed in the Atlas. A summary of the major entrepreneurship patterns for towns is given in Urban History and The Age of Entrepreneurship Chapter 9. The town definitions are given in:
England and Wales urban classification and towns
The BBCE data downloads for towns and Burghs are given by sex, totals, and EA17 business sectors:
Downloadable BBCE data for E&W towns: TOTALS
Downloadable BBCE data for E&W towns: by SECTOR
Downloadable BBCE data for Scottish Burghs: TOTALS
Downloadable BBCE data for Scottish Burghs: by SECTOR
Data downloads for counties and regions
The census data in BBCE and I-CeM is given for county downloads, which can also be used to define regions to allow regional economic analysis and linkage to other data sources. The BBCE data for counties are given by sex, totals, and EA 17 business sectors:
Downloadable BBCE data for E&W counties: TOTALS
Downloadable BBCE data for E&W counties: by SECTOR
Downloadable BBCE data for Scottish counties: TOTALS
Downloadable BBCE data for Scottish counties: by SECTOR
Definitions of regions (in the census termed Divisions) and various other regional definitions common used (such as NUTS 1, government Economic Planning Regions, and Clive Lee's analysis) are given as a download based on census counties from which users can construct different regional BBCE data.
Download definitions of regions
This download allows an approximation to common regional boundaries over time. The main difficulty is minor parts of Middlesex which became small parts of counties marked * after 1965, when most of Middlesex transferred to London. If London is aggregated with the SE then the historic counties align with modern regions closely, as used by Clive Lee. Economic Planning and Standard Statistical Regions. Alignment with NUTS is more difficult because of the creation of the Eastern region to include Essex separate from the SE. A more detailed alignment would require reallocation of sections of each county at different time points. The main changes are summarised below.
Economic Planning Regions 1945-94 (EPRs) were almost identical to Lee, C. H. (1971) Regional Economic Growth in the UK since the 1880s, McGraw Hill, London; hence not shown. The main difference was Derbyshire district of High Peak in the NW EPR and not East Midlands EPR; and Furness moved from Lancashire to Cumbria in 1974. Historic counties aggregate to Lee and EPRs fairly closely.
SSRs: The Standard Statistical Regions used for government office statistical reporting until 1994 (when replaced by NUTS 1) differed from EPRs, sometimes including London in the SE and sometimes not; modern Cumbria (Cumberland and Westmorland) was moved from the NW to Northern Region, including the Furness part of Lancashire after 1974. + The E. Midlands excluded part of S. Humberside (Lindsey, Lincs.) which was taken into Y&H (E. Riding).
Inter-census record linkage between years can be readily developed from BBCE/I-CEM. The BBCE is unique as a digital census resource that covers all years for England and Wales 1851-1911, since it gives the main entrepreneur population for 1871. Developing record linkage is a large scale research task which requires appropriate algorithms and relies heavily on the quality of the original data collection and transcription of the census records. Results of a pilot study developed for the BBCE records of entrepreneurs in the census are promising and demonstrate some important features of entrepreneurship not previously well understood in the 19th century: the dynamics of switches between employer, own account, worker and unoccupied status; and the diversity of growth experiences.
The methodology of record linkage is given in WP 24 and the database definitions in WP 25. The database itself is available for other users to allow tracking between years; it gives RecIDs for over 468,000 record-linked individuals who, for at least one census between 1851 and 1891, were employers or own account proprietors. A non-entrepreneur stratified quota full population and random sample covers 21 occupational sectors containing 32,174 linked people in 1851-61, and 61,167 in 1881-91.
Downloadable database record-linked Entrepreneurs 1851-91
Downloadable database sample record-linked Non-entrepreneurs 1851-91
Users may want the names of the linked individuals. These can be accessed by applying for an I-CeM special licence.
A methodological paper in Information Processing and Management shows how machine learning, deep learning and other methods can be used in addition to traditional logistic regression for classifying and supplementing census responses to overcome non-responses and changes in the design of census questions. These methods are extendable to record linkage.
Further development of record linkage methods is a very fruitful area for future research. A paper assessing linked career changes of entrepreneurs is to appear in Social Science History. Contact is welcomed from others developing methods or applications.
The BBCE is primarily a source for non-corporate proprietors, without equal for 1851-1911 in its size and breadth of coverage from the largest to the smallest firms, and including own-account sole proprietors that are often omitted from business history studies ('the ranks of the anonymous' as Peter Payne called them). BBCE includes company directors, where they can be identified from either their census descriptor or from record linkage with the Directory of Directors (DoD). These are identified in BBCE (as the variable DIRECT) and available to other researchers, with separate codes for N of companies they direct (DIRNO), the location of these companies (DIRGEOG1, DIRGEOG2, etc.), and their company sectors (DIR1, DIR2, etc.) in BBCE, as defined in WP 14.
The census is a partial record of directors. Others that are contained in the DoD cannot be definitively identified with a census record, either because the person was foreign, absent or not recorded on census night, or because the information is too imprecise to allow confident identification of the census individual. Because of this the full database of the DoD keyed for the years that are believed to best cover the census years (1882, 1892, 1902 and 1912) will be made available as an additional database deposit. This identifies and codes all individuals in the DoD, their multiple roles in different companies, and the companies themselves.
Many businesses continued as partnerships throughout the 19th century, as today, without becoming limited companies. These can be compared with limited company proprietors and sole traders. They are coded in BBCE where identifiable from census descriptors, as PART with the number of partners and gender mix. An analysis of the 1881 data for partners is given in an Economic History Review paper.
Large and other firms
The largest firms often have significant archives or other information available that allows their identification within census records and provides additional scope for research. It is a major task to link all these data sources with the BBCE and this has not been attempted. However, collaboration with other researchers offers great scope to develop this potential. One such initiative by Prof Leslie Hannah involves constructing a full record of all manufacturing firms with 1,000 or more employees using BBCE and other sources (about 410 in 1881). This fills out BBCE, extends its coverage, and checks its accuracy. It is intended to make this dataset fully available. Several publications derive from this. An analysis of the firms of 1,000 or more employees in 1881 with the data for that years is to appear in Economic History Review http://onlinelibrary.wiley.com/doi/epdf/10.1111/ehr.13118 (WP version).
Similar scope exists to expand the detail available on smaller firms, especially by working with data from case studies locally, sectors, and family histories. This is a major task and researchers interested in developing these approaches, developing collaboration, or have information available, are welcomed to make contact. Initially in BBCE project follow-ups this is being done through local and sector case studies.
A pilot study using tax data for 1861-81 has been developed for a sample of firms linked to the BBCE data. This assesses variations in profit levels and profitability (measured as profit per worker). The paper expands Marshall's contemporary insights, showing heterogeneity to be dominated by sector differences at both '1-digit' and '5-digit' levels. Localisation effects also occur, reflecting urban/rural differences, accessibility to railways, and to a lesser extent access to waterways. Differences in firm-level organisation (portfolio diversification and partnerships) are less significant, and demographic effects have little significance. Also large unexplained elements of heterogeneity indicate important elements of entrepreneurial agency: (Cambridge Journal of Economics, 2022)
BBCE identifies many businesses that have multiple activities (about 10% of all employers in each year), as shown in the Atlas. A portfolio exists if the ENTOCCODE of the sector of the first business activity differs from the next, and so on. These are ranked in order of listing in the census, to give the BBCE variable PORT1, PORT2, etc. (up to 8 multi-businesses occur). An analysis of portfolios is given in The Age of Entrepreneurship Chapter 11, and for farm portfolios is given in a Journal of Rural Studies paper. Where census respondents gave details of the employees in each of their businesses this has now been extracted for 1881 and ranked in order of size, available as a separate download for portfolio businesses (Group 1 stating employees); they are identified by GSU ID and occode as in Schürer, K. and Woollard, M. (University of Essex) (2000) 1881 Census for England and Wales, the Channel Islands and the Isle of Man (Enhanced Version) [computer file] UKDA, SN-4177:
Download of portfolio workforces 1881
Example of portfolio analysis for farmers in J. Rural Studies paper (2016)
Demography and gender
Because BBCE is a census-based resource it contains a wealth of information on individuals by age, sex, marital status, relationships to others in the same household, their birthplace, and place of residence. Their occupation not only allows entrepreneurial or worker status to be established, but is also a measure of social class. This gives major opportunities for individual level analysis to understand demographic and social structures. The Age of Entrepreneurship Chapters 7 and 8 begin to develop this, as do other publications such as examination of gendered entrepreneurship and international comparisons; Chapter 10 and other papers begins the analysis of migration. A Continuity and Change paper extends analysis of how all individuals, but especially entrepreneurs, responded to economic opportunities. All these dimensions will reward much further analysis.
Tests of data coverage
BBCE is the largest digital dataset ever assembled on 19th and early 20th century UK entrepreneurs. But it relies on the self-identification by individuals of their entrepreneurial status in census responses, which is imperfect, and the success of modern efforts to supplement and interpret incomplete responses. BBCE is more complete for large firms than the few tables that were published by the historic censuses for 1851-81, but appear to have fewer smaller firms than census publications recorded; see HMJ paper. However, some entrepreneurs were frustratingly poor at recording their business characteristics in the census.
Tests of representativeness of employer response giving their workforce size have been developed through comparison with the Hannah data for large firms, and for firms of 10 employees and over through record linkage (HMJ paper forthcoming).
Feedback form other researchers on gaps in coverage and issues of representativeness is welcome.
Many other avenues for research exist using BBCE/I-CEM, especially when combined with new research questions. The BBCE website will attempt to keep up with these developments and post links here to help the research community of users who can benefit from exchange of experiences with the database.
Early developments are:
- Firm size: Gibrat and lognormal distribution; with anonymised test data set: Mendeley Data
- Deep Learning demonstration which has a supplementary anonymised test data set: Mendeley Data
- Record linkage data: WP 24, WP 25, and data downloads attached to these papers
- Directors' data 1881-1911: (information to follow)
- Populations Past. Mapping the demography of England and Wales 1851-1911
- Atlas of Fertility Decline in England and Wales
- I-CeM Integrated Census Microdata, the digital census microdata to which BBCE is linked
- I-CeM User Guide
- The Cambridge Group for the History of Population and Social Structure
- Campop project on Occupational Structure of Britain