Using financial data to enhance labour market statistics in Scotland

Study design and governance

The project progressed from understanding the source data, through to developing and testing a classification method, to considering how the resulting outputs might be used in practice. Review points were built into the work so that findings and limitations from each phase could inform the next phase.

The feasibility study was a research exercise rather than the implementation of new statistics. Its purpose was to test what could be learned from the available financial data, assess the quality and limitations of the approach, and identify any additional work that would be needed before moving to ongoing data delivery. The detailed evidence from the first three phases and the proposed roadmap from the fourth phase were evaluated by an expert steering group and are published as supporting documents.

Data used in the feasibility study

The study used a pseudonymised and aggregated consumer transaction dataset sourced from NatWest Banking Group and analysed by Smart Data Foundry. The data included weekly summaries of account balances, income and expenditure, alongside age band, sex and geography. The Scotland-specific dataset contained approximately 575,000 customer records across all 32 local authorities.

The analytical work focused on customers aged 18 to 64. This differs from official labour market estimates, which focus on those aged 16 to 64 for headline employment and inactivity rates and those aged 16 and over for headline unemployment rates.

Coverage was dependent on a customer being within the source data and having sufficient geographic information. Therefore, the data was not a random sample of the Scottish population. The study needed to first assess the representativeness of the data and later apply weighting to adjust for uneven coverage across local authority, age band and sex.

Design of the study

The study was designed as a proof of concept broken into four phases as outlined in Table 1.

The source data did not contain a record of whether a customer was employed, unemployed or economically inactive. The study developed proxy classifications from transaction patterns and assessed whether these produced potentially useful estimates. This highlights an important distinction between the labour market status gathered from respondents to existing labour market surveys and the labour market status estimated from the financial behaviour of customers.

The project considered the potential value of using financial data as an additional source of labour market indicators, particularly where a large dataset and frequent observations might support more timely or granular analysis. It also considered the practical, methodological and governance requirements that would apply if the work were developed further.

The study used an iterative approach. Methods and assumptions were refined as evidence became available, and the outputs of each phase informed the scope and priorities of the next phase of the feasibility study. This allowed issues such as uneven population coverage, differences from official concepts and classification uncertainty to be identified and considered prior to the analysis and recommendation phases.

Table 1: Structure of the feasibility study

Steering group and quality assurance

A short-life expert steering group was established to provide quality assurance and constructive challenge throughout the study. The role of the steering group was to review the proposed methods, assumptions, evidence and interpretation, and to help ensure that the conclusions accurately reflected what the analysis could support. The group included statistical and analytical expertise from the Scottish Government, National Records of Scotland and the Welsh Government.

The steering group considered the work at planned points during the first three phases. These discussions provided an opportunity to test the analytical approach, identify areas requiring clarification or further work, and review the outputs before each phase was signed off. Actions and recommendations arising from the group were used to refine the analysis and reporting.

Steering group sign-off indicated that the work completed for a phase was sufficient for the feasibility study to progress. It did not provide assurance that the method was ready for ongoing use. Those decisions would require further independent review, operational testing and formal statistical governance.

The first three phases produced technical reports that were provided to the steering group for review and sign-off. The fourth phase consolidated the findings from the three phases and set out a proposed roadmap to ongoing data delivery. The roadmap was a plan for possible future development and delivery; the activities within it were not part of the completed feasibility analysis.

The four supporting documents provide the detailed technical and operational evidence to support further exploration of financial transaction data as an additional source of timely and detailed labour market insight. This report provides an overview of the feasibility study documents, summarising the study’s purpose, approach, findings, limitations and implications for a wider audience.