Interpreting the Census Data
At the time of the 2021 Census on 21
st March 2021, the UK was beginning to emerge from a third national Covid-19 lockdown. The Government had issued the ‘Roadmap out of Lockdown’ at the start of the month, reopening childcare facilities, schools and colleges
[1]. However, the ‘stay at home’ rule was still in effect for everyone except for key workers and construction workers on Census Day.
Despite these extraordinary circumstances, the ONS conducted the 2021 Census as originally planned with only minor clarifications to the questionnaire that had been tested between 2016 and 2020.
Rather than asking individuals about where they would have been working on collection day if lockdown restrictions were not in place, the ONS instead asked respondents where they were ‘ordinarily working’ on 21st March 2021, which may have affected how people responded to the question. This in turn may have inflated the figure for the number of homeworkers and masked the actual picture of usual commuting flows.
As shown in Figure 1, overall levels of commuting flows were much reduced across England in 2021 compared to the earlier 2011 Census, as would be expected given the lockdown restrictions in place at the time.
Figure 1 Census 2011 and Census 2021 Origin-Destination travel to work routes across England (mapped to MSOA population weighted centroids)
Source: ONS Census 2011 and 2021 | Lichfields analysis
When isolating individual regions, a number of trends emerge. Firstly, the reduction in commuting flows is most noticeable in major cities, while more rural areas are less affected. For example, this is shown in Figure 2 by the significant reduction in commuters to London from the East of England. This is likely a reflection of the prevalence of office-based occupations within the London workplace economy, as well as greater reliance on public transport connections – many of which were significantly reduced during this period – which workers may have chosen to avoid during the pandemic for reasons of social distancing.
Figure 2 Census 2011 and Census 2021 Origin-Destination travel to work routes in the East of England (mapped to MSOA population weighted centroids)
Source: ONS Census 2011 and 2021 | Lichfields analysis
By contrast, the decline in commuting flow was less pronounced in northern England. However, there is a noticeable trend of decreased intercity travel during the pandemic, perhaps also a reflection of the desire to avoid public transport. Figure 3 shows the change in commuting flows between Huddersfield, Bradford, Halifax, Leeds, and York between 2011 and 2021.
Figure 3 Census 2011 and Census 2021 Origin-Destination travel to work routes in Yorkshire (mapped to MSOA population weighted centroids)
Source: ONS Census 2011 and 2021 | Lichfields analysis
This data suggests that any long-term shift in commuting patterns is unlikely to be consistent across the country, with a rural to urban variation, and major office centres such as London being disproportionately affected.
Commuting in the Post-Pandemic World
The fall in the number of commuters between the two Censuses was fuelled largely by an additional 9 million people shown to be working from home (a 300% increase). However, with people now returning to work in offices, it is difficult to establish to what extent the pattern observed in 2021 is still representative of travel patterns today.
The Department for Transport (‘DfT’) Annual Travel Survey indicates that the number of commuting trips has started to recover since the pandemic, with commuting trips per person per year increasing by around 20% between 2021 and 2022 as shown in Figure 4. This is reflected in public transport usage data; for example, DfT reports that as of October 2023, rail journeys in Great Britain excluding the Elizabeth Line had recovered to around 80-90% of the pre-pandemic baseline (albeit the relative recovery has been largest for leisure journeys at weekends). Taken together, this suggests that the travel patterns observed in the 2021 Census are already outdated and no longer provides an accurate reflection of the current commuting landscape.
Figure 4 Commuting trips per person per year
Source: DfT National Travel Survey (2023)
Meanwhile, the ONS’s Opinions and Lifestyle Survey revealed that in February 2022 (after the work from home guidance was lifted), 8 in 10 workers who had to work from home during the pandemic planned to hybrid-work going forward
[2]. Subsequently, between February and May 2022, the prevalence of hybrid working rose from 13% to 24%, while the proportion of those working exclusively from home declined from 22% to 14%.
Figure 5 highlights a clear trend in the prevalence of homeworking between different income groups; those more likely to have adopted home or hybrid working generally had higher personal incomes, while those on lower incomes were more likely to not be able to homework.
Figure 5 Location of work by personal annual income
Source: ONS Opinions and Lifestyle Survey (Great Britian, September 2022 to January 2023)
Over two-thirds of workers in managerial, professional and associate professional occupations either solely worked from home or hybrid worked. Conversely, those in elementary or sales and customer service occupations were less likely to be able to homework and therefore the mainly travelled to work. Younger people (aged 16 to 24) were also more likely to travel to work compared to other age groups, with only 6% choosing to solely homework. This may have reflected the constraints to regularly working from home from smaller or rented accommodation.
Finally, there was regional variation in travel to work habits across the country. In London up to 60% of workers were either home or hybrid working, and in Wales 19% of employees solely worked from home. This contrasted to the North East where only 13% of employees solely worked from home and 62% travelled to work every day.
Figure 6 Location of work by country and English regions
Source: ONS Opinions and Lifestyle Survey (Great Britian, September 2022 to January 2023)
The ONS has recently published experimental data estimating a revised commuting position for 2021 during the pandemic, as well as what commuting patterns might have looked like in 2021 had there been no pandemic
[3]. This project also aims to provide annual estimates of travel-to-work behaviour for the period 2011-2021 and beyond in future updates.
While these experimental results are not intended to be used for decision-making purposes, they do show that the 2021 Census may, as expected, represent an underestimate of the true level of commuting patterns
[4]. Figure 7 shows that while there was a decline in travel-to-work during the pandemic compared to the ‘no-pandemic’ scenario, this is less marked than the picture of decline indicated by the comparison of the 2011 and 2021 Censuses shown in Figure 1.
The ONS’ interest in developing these experimental data sets is an acknowledgement regarding the accuracy of the 2021 Census commuting data, and may represent a step towards addressing the challenge presented by planning in the context of a ‘snapshot’ of data published once every ten years.
Figure 7 Estimated ‘no-Pandemic’ and ‘mid-Pandemic’ travel to work patterns across England, 2021
Source: ONS Data Science Campus | Lichfields analysis
