Male occupations in parish registers

Before 1841, census data are either of insufficient quality as a basis for occupational estimates or, before 1801, entirely non-existent. Fortunately, information about the gainful activities of individuals, particularly when they were men, was quite regularly recorded because local dignitaries were interested in such information, because it could be used to distinguish individuals with the same name and place of domicile, or simply because it was perceived as an integral element of a person's identity. The best systematic records of such information can undoubtedly be found in parish registers, particularly in the records of Anglican baptisms. This is especially true after so-called Rose's Act, which came into effect in January 1813. Officially called the 'Act for the Better Regulating, and Preserving of Parish and Other Registers of Births, Baptisms, Marriages and Burials in England', the express purpose of this piece of national legislation was to 'greatly facilitate the proof of pedigrees claiming to be entitled to real or personal property, and be otherwise of great public benefit and advantage'.¹ It introduced standardised forms for the registration of marriages, baptisms, and burials in Anglican parish registers. The new standard form for baptisms included a space for the 'Quality, Trade or Profession' of the father of the baptised infant. Private as well as public baptisms were, from now on, to be recorded which, it was concluded, had 'evidently added to the number of Registered Baptisms'.² Rose's Act ensured that, from 1813, historians have access to data on male occupations that cover all English and Welsh parishes. The Cambridge Group has painstakingly collected and transcribed these data, and used them as the basis for what has rightfully been called an occupational 'quasi-census' for the 1813-20 period.³

The data for this quasi-census are not perfect, but their imperfections have either merely negligible impact on the male occupational structure derived from them, or can be fairly easily corrected, as Peter Kitson et al have shown.³ Combined with the 1841 and 1851 censuses, the 1813-20 quasi-census presents historians with occupational data of unprecedented quality for the first half of the nineteenth century. The 1813-20 baptism data are vastly more occupationally precise than the social tables on which historians have relied in the past; they provide reliable information not merely on the main sectors (primary, secondary, tertiary) but also on their subsectors (textiles, retail, etcetera) and even individual occupations (weaver, grocer, etcetera). They are, like the census data, carefully codified using Wrigley's Primary-Secondary-Tertiary (PST) system, a flexible hierarchical taxonomy of occupations, allowing one to easily switch between different levels of occupational abstraction.

Furthermore, their coverage is essentially universal and so geographically fine-grained that it allows for occupational analyses not only at the level of the entire country, but also for individual counties, hundreds, census registration districts, or even parishes and chapelries. Much of the discussion on long-run economic developments and the industrial revolution has been focused on the national level, since that is the level at which the national accounts approach delivers its insights. Quantitative occupational data at sub-national levels offer the opportunity to put spatial relationships back in the consideration of long-run economic development and the industrial revolution. Crafts and Harley rightly noted that 'regional development varied considerably and that exploring this diversity offers the potential of a set of quite different and valuable insights into the experience of the industrial revolution.'⁴ Using occupational information to generate a quantitative understanding of regional developments is therefore one of the great promises of the Occupational Structure project. For that reason, all of the Group's occupational data – including those from parish registers – have been meticulously linked to a geographical information system (GIS), developed by members of the Group and in particular by Max Satchel, building on the work of Roger Kain and Richard Oliver and an earlier historical GIS by Humphrey Southall and Nick Burton.⁵

Prior to Rose's Act, baptism records also occasionally recorded the father's occupation, if the local rector or bishop felt such information was valuable. All 11,400 Anglican parish registers were meticulously searched for such records by research assistants employed by the Cambridge Group, providing the basis for the Group's pre-1813 occupational structure estimates. Only parishes in which occupations were recorded in at least ninety-five per cent of baptisms for which an occupation could be expected – so, for example, excluding illegitimate children – were included, to ensure the data are reliable. As with the census data and the 1813-20 data, these earlier parish register data were codified geographically and occupationally.

This resulted in three subsets of parish register data, one for the early eighteenth century (c.1710), one for the middle (c.1755), and one for the end of that century (c.1785).⁶ In their recent chapter in the Cambridge Economic History of Modern Britain (CEHMB), Shaw-Taylor and Wrigley used the data from the first of these subsets to generate estimates for the early eighteenth century male occupational structure for England and Wales and for two regional clusters of counties, at the sectoral and sub-sectoral level.⁷

That is not to say that parish registers do not suffer from limitations. An obvious weakness of the information provided by Anglican baptism registers is that the occupations of mothers were not recorded. They therefore only serve as a basis for documenting the male occupational structure. But even then, baptism registers only provide occupational information for a sample of parishes before Rose's Act. Coverage was particularly low for the mid and late eighteenth century, as shown in Map 1, and close to non-existent outside London before 1700.

Map 1. Registration of male occupations in English and Welsh baptism registers over the course of the eighteenth century
Note: Parishes and chapelries in which male occupations were reliably recorded are indicated in red.
Data source: Parish register occupational database, created by the Cambridge Group.

In the early eighteenth century, only eleven per cent of all parishes recorded occupational information for a period of one or more years. For the mid and late eighteenth century, the figures are even lower, at only three and four per cent respectively. For occupational estimates at larger geographic scales, such as for England and Wales as a whole, small samples are not necessarily a problem, as long as the sample of parishes is representative of the population from which it is taken. However, as Map 1 shows, the samples are geographically non-random. Even in the relatively well-covered early-eighteenth century, urban areas are overrepresented, as are certain regions like industrialising Lancashire and the West Riding of Yorkshire, with other areas such as Wales, the South-West and East of England, the North, and several counties in the West Midlands covered hardly at all. Average coverage across the totality of England and Wales may have been eleven per cent in the early eighteenth century, but in two out of every three counties, coverage was below ten per cent. In the mid and late eighteenth century, lack of coverage was even more pronounced, with eight out of every nine counties below the ten per cent mark, whilst not a single parish in which occupations were registered could be found in four out of every five counties.

Even within counties, occupational structures were very far from uniform, so at this lower geographical level the sampled parishes cannot simply be presumed representative either. Well aware of these issues, Shaw-Taylor and Wrigley did not simply base their c.1710 national estimate directly on the parish register sample. Instead, they divided the pre-Rose's Act parishes along two axes: urban versus rural, and north-west England versus the rest. The underlying assumption is that the covered parishes, known not to be representative of all parishes, are much more likely to be representative of parishes on the same side of these divides. More reliable occupational structures can therefore be calculated for the rural and urban subsets of parishes, and for parishes inside and outside north-west England. These partial occupational structures are subsequently recombined to create a national one.

On a national level, and provided that, as in c.1710, the parish data have a reasonable geographic spread across the England and Wales, the reweighting method is likely to generate good results. Potential regional biases within the urban and rural subsets will be much diluted on a national scale. The approach is less likely to generate reliable national results for the mid and late eighteenth century, as the required spread of parish register data across England and Wales is not available. For this reason, Shaw-Taylor and Wrigley refrained from including national estimates for these periods in the CEHMB chapter. The approach is also unreliable for smaller geographical levels, and Shaw-Taylor and Wrigley have therefore only used it nationally and on two large geographic regions. Had it been applied to a smaller region, such as a single county, the results might have been seriously misleading. For example, the Lancashire urban sample in c.1725 contains transport-dominated Liverpool but not manufacturing-dominated Manchester. This bias cannot be remedied by urban-rural reweighting.

To create reliable regional estimates before Rose's Act, and any estimates before 1700 parish registers need to be supplemented with other sources of occupational data. That is were testamentary documents come in.

A talk by Leigh Shaw-Taylor, given at the ENCHPOGOS 2017 conference held at Robinson college, Cambridge in September 2017, provides an introduction to our work using parish registers and the censuses. Shaw-Taylor, L., Parish Registers and Census Data 1695-1911, ENCHPOPGOS Conference, Cambridge September 2017.

¹ Referenced in Basten, 'From Rose's Act to Rose's Bill: a reappraisal of the 1812 Parish Register Act', Local Population Studies, 76 (2006), p. 43.

² Ibid, p. 44.

³ Kitson et al, 'The creation of a "census" of adult male employment for England and Wales for 1817', Cambridge Working Papers in Economic and Social History, 4 (2012), http://www.econsoc.hist.cam.ac.uk/docs/CWPESH%20number%204%20March%202012.pdf.

⁴ Crafts and Harley, 'Output growth and the British Industrial Revolution: a restatement of the Crafts-Harley view', The Economic History Review, 45:4 (1992), p. 721.

⁵ Satchell et al, 1851 England and Wales census parishes, townships and places (Colchester: UK Data Service, 2016), http://reshare.ukdataservice.ac.uk/852232/; Kain and Oliver, The historic parishes of England and Wales: an electronic map of boundaries before 1850 with a gazetteer and metadata (Colchester: UK Data Service, 2001), http://dx.doi.org/10.5255/UKDA-SN-4348-1; Southall and Burton, GIS of the ancient parishes of England and Wales, 1500-1850 (Colchester: UK Data Service, 2004), http://dx.doi.org/10.5255/UKDA-SN-4828-1.

⁶ The precise periods for which occupational data were recorded in parish registers differed by diocese and, indeed, individual parish. The 1710, 1755, and 1785 dates therefore only represent the weighted average midpoints for each period, taking their data from quite a wide range of years. For example, the early eighteenth century sample, centred around the year 1710, contains parish register data from as early as 1695 and as late as 1729.

⁷ Shaw-Taylor and Wrigley, 'Occupational structure and population change' in Floud, Humphries, and Johnson (eds), The Cambridge Economic History of Modern Britain. Volume 1. Industrialisation, 1700-1870 (Cambridge: Cambridge University Press, 2014), pp. 53-88.