Male occupational descriptors in testamentary documents

Millions of testamentary documents (wills and testaments, inventories, letters of administration, bonds, and codicils) have survived for England and Wales, covering much of the two countries from the sixteenth century (and sometimes even earlier) up to the census period. When the decedent was male, his occupational title was often recorded in these documents. This occupational information is relatively easily accessible because it is often contained in the indexes which have been created to provide access to the original documents. Many such indexes were created in printed form in the nineteenth and twentieth centuries, which can be converted into digital data, using optical character recognition software. Better still, many county record offices have done this already, or have recreated digital indexes from scratch. These indexes are often highly structured, splitting off occupational from other items of information such as the decedent's name and domicile and the date at which the document was created. Even when the latter has not yet been done, the information is often presented in text strings of a fairly consistent nature, making it relatively easy to extract the required information. Using such printed and digital indexes, a database of testamentary occupational information was created, covering the vast majority of English and Welsh counties, and listing the occupations of over 800,000 men for the 1600-1850 period.

Probate documents thus constitute a rich source of male occupational information. However, they are also a problematic source, as they are not representative of the contemporary population. It can be calculated that roughly four out of five English male householders who died in the 1600-1850 period left no trace in the probate record. The men that were probated were not a random subset of the adult male population. The church courts could charge for grants of probate if the decedent's estate was valued at five pounds or more, and therefore had a financial incentive to encourage application for probate in these instances. But they could not force anyone to make a will and/or inventory. It is therefore not surprising that many decided to avoid the expense and bother of applying for probate. Since the trade-off between, on the one hand, the cost and effort of the probate process and, on the other hand, its value in case of disputes over the estate, was more likely to be positive for high-value than low-value estates, the former are overrepresented in the probate record. It is unsurprising, then, that men who were wealthy and/or engaged in activities that required much capital – such as farmers, merchants, tanners, and brewers – were more likely to make probate than men who were poor or whose occupation required little or no capital – such as tailors, weavers, domestic servants, or labourers. The occupational bias of the probate record is clear from Figure 1, in which the relative share of men making probate has been calculated for a sample of occupations by comparing probate data with parish register data from the same area and time period. As the figure shows, the probability of an early-nineteenth-century farmer in Cheshire leaving a probate document was four times higher than a butcher, twelve times higher than a weaver, and twenty-five times higher than a labourer.

Figure 1. The relative chance of being probated for a sample of occupations in Cheshire, c.1817
(relative to farmer = 100%)

As a result, male occupational structures derived directly from probate data are highly unreliable. Fortunately, this problem has proven to be surmountable. An extensive description of the solution is available as a working paper [link] but, in short, the method works by calibrating the probate data using parish register data from the same time period and locations. This approach uses the strengths of each of the two main data sources – parish registers and probate documents – to remove the weaknesses in the other. Parish register data are reliable but, before 1813, only available in select years and parishes, and hardly at all before 1700. Testamentary data are available much more widely in both temporal and geographical terms, but are severely biased. By using the parish register data to remove this bias, the (calibrated) probate data can be used to interpolate and extrapolate the parish register data. The results are reliable male occupational estimates for England and Wales, at twenty-year time intervals, at national, regional, and local scales, covering the period from c.1600 right up to 1813 when, as a result of Rose's Act, parish register coverage is complete for England and Wales, and probate data are no longer needed to fill in the gaps.