Remote working, bargaining power and productivity

Lena Anayi

Remote working soared during the Covid-19 (Covid) pandemic. Over half of British workers worked from home during the initial Covid lockdowns (first panel in Chart 1). And by February this year, nearly a third of workers were still doing so at least some of the time. But will this last? In this blog post, I explore firms’ and workers’ attitudes to remote working, the extent to which these may differ, and factors that might affect negotiations between them on future remote working arrangements.

Chart 1: Proportion of workers working from home, full-time or hybrid(a)(b)

Sources: BBC/YouGov, Decision Maker Panel (DMP), Labour Force Survey, ONS Business Insights and Conditions Survey (BICS), ONS Opinion and Lifestyle Survey, Prolific, and Understanding Society.

(a) BICS results relate to employees only, and DMP results relate to full-time employees only.
(b) The second panel presents the range of results (minimum and maximum) within each category.

Of course, as the Covid pandemic continues, many workers and firms will still be figuring out their long-term remote working preferences. And lack of matched worker-firm data containing information on both sets of preferences means researchers may struggle to accurately estimate the degree to which individual workers’ preferences are misaligned with those of their employers.

Misalignment in workers’ and firms’ preferences around remote working

Still, various survey results show that future remote working preferences differ across workers and firms. The second panel in Chart 1 shows that, post-pandemic, workers desire more remote working than they expect will be available to them, and their expectations are slightly higher than those of firms. Many workers may therefore prefer more remote working post-pandemic than their individual employers will be willing to accommodate.

It may also be possible to gauge some degree of misaligned preferences by investigating the drivers of these. For instance, what if workers and firms were all motivated by the same thing, such as maximising productivity? If so, assuming mutual agreement on how remote working affects this, there may not be much disagreement around future arrangements after all.

Starting with the drivers of workers’ preferences, I run a probit regression using the Understanding Society household survey micro-data, with a dependent dummy variable for whether the worker desires more remote working post-pandemic relative to pre-pandemic. As explanatory variables, I include worker demographics such as gender, age, level of education, region, household size and whether they are a parent of dependent children. I include employer industry (A38 group), firm size and the firm’s flexible working policy. I also include job-related characteristics such as mode of employment (employee versus self-employed), pre-Covid earnings, weekly hours worked during Covid (total, overtime hours, and change relative to pre-Covid), office commuting time, job tenure, occupation (NS-SEC group), and whether the individual is a ‘key worker‘, as well as remote working status pre and during Covid. And I include workers’ subjective assessments around the impact of remote working on their hourly productivity during Covid, pre-Covid work-life balance, pre-Covid job satisfaction, and whether they experienced feelings of loneliness during Covid.

The regression results (with statistically significant regressors shown in Chart 2) reveal several things.

First, workers’ remote working preferences are unrelated to employer characteristics, with the exception of their flexible working policy. This suggests that workers’ preferences cannot be predicted based on their employers’ characteristics, implying that workers and firms may often disagree.

Second, preferences are also unrelated to worker demographic characteristics.

Third, worker preferences are positively associated with whether the worker had been working from home during Covid (with this making them 19% more likely to favour more of it post-pandemic, all else equal), and whether they found it to be productivity-enhancing (with each 10 percentage point boost to productivity making them 4% more likely). Higher daily commuting time to the office also raises workers’ propensity to favour more remote working (with each additional hour making workers 4% more likely), as does dissatisfaction with work-life balance (8% more likely) and general job dissatisfaction (8% more likely). Meanwhile, feelings of loneliness during Covid lower workers’ propensity to favour more remote working post-pandemic (18% less likely), as does being a ‘key worker’ (16% less likely).

Perhaps surprisingly, workers that had already been working remotely pre-pandemic are less likely to favour more of it post-pandemic, suggesting a possible limit to how much remote working workers ultimately desire.

Chart 2: Drivers of worker preferences around increased remote working post-pandemic(a)(b)

Source: Understanding Society.

(a) Average marginal effects are reported due to the non-linearity of the probit link function, such that the estimated impact of any regressor varies as its quantity increases.
(b) N = 1,979. MacFadden R-squared = 0.27. Log-likelihood = -879.51.

Turning to the drivers of firms’ preferences, I run a similar probit regression using the BICS firm-level data. As explanatory variables, I include firm demographics such as firm age, size, region, industry (A38 group) and ownership origin (UK/EU/US), and I also include pre-Covid information on (log) productivity, level of intangible assets, and ratio of office rental costs to revenues. These information are drawn from firms’ pre-Covid responses to the Annual Business Survey (ABS). Furthermore, I include information on whether firms raised remote working levels during Covid, peak remote working during Covid (percentage of workers), and their subjective assessment of how remote working affects productivity (which I interpret as referring to worker efficiency).

The regression results (with statistically significant regressors shown in Chart 3) indicate that firms finding remote working to be productivity enhancing are 44% more likely to favour more of it post-pandemic (versus those finding it to be productivity neutral), all else equal, while those finding it productivity reducing are 28% less likely to do so.

However, productivity isn’t the only consideration for firms. Those with higher office-related overhead costs also prefer more remote working post-pandemic, with each additional percentage point increase in the ratio of overhead costs to revenues making a firm 4% more likely to do so. These firms may desire cost savings through reductions in office rent, and as renters they may also have more flexibility to make such adjustments. Cost reductions can also contribute towards improved firm productivity through higher profit margins.

Interestingly, US-owned firms are 14% less likely to favour more remote working post-pandemic (versus UK owned), suggesting that cultural factors could also be at play.

Chart 3: Drivers of employer preferences around increased remote working post-pandemic(a)(b)

Source: BICS data matched with ABS.

(a) Average marginal effects are reported due to the non-linearity of the probit link function, such that the estimated impact of any regressor varies as its quantity increases.
(b) N = 2,659. MacFadden R-squared = 0.48. Log-likelihood = -5691.55.

For both workers and firms, therefore, productivity considerations are important. But there are other important factors too, some of which firms and workers may overlook. Workers may desire work-life balance, workplace camaraderie or shorter commutes, while firms may desire cost savings arising through lower office rents.

Negotiations between workers and employers: a bargaining power story?

Competing preferences between workers and firms may create opportunities for negotiation, if employers agree to this. BICS survey respondents were asked in September 2021 about their ‘main consideration when deciding who can return to their normal place of work’. Chart 4 shows that around half of firms (52%) indicated a willingness to negotiate with workers, with around a third (32%) unwilling to do so. During that same period, firms who were willing to negotiate over working arrangements had double the proportion of workers working remotely (26% versus 13%).

Chart 4: Whether the employer or employee determines future working arrangements

Source: BICS (wave 39).

Although the survey question asks specifically about the timing of workers’ return to offices, firms’ decision-making around this may be assumed to correlate closely with their broader remote working arrangements for the future.

Are firms that are willing to negotiate simply more mindful of employee preferences? Or perhaps they have relatively weak bargaining power? Maybe they are better able to accommodate employee preferences, due to being more profitable?

There are various measures of both employer bargaining power and firm profitability that allow us to test some of these possibilities. I run another probit regression using the BICS firm-level data (again matched with pre-Covid ABS responses), this time with propensity to negotiate as the binary outcome of interest. Controlling for firm size, productivity and industry, I simultaneously include various indicators of weaker firm bargaining power as explanatory variables, each of which are individually positively associated with propensity to negotiate. These include unionisation at workplace, reliance on migrant labour, whether the firm is currently struggling to hire, facing labour shortages or high staff turnover, or has recently raised workers’ wages, and firm-level labour tightness (ratio of vacancies to employment). I also include profit margins (pre-Covid) and reported Covid impact on profits as profitability metrics, both of which are individually positively associated with propensity to negotiate. And I include firms’ reported remote working productivity impacts, and overhead costs relative to revenues.

I find that reported productivity impacts of remote working best explains firms’ willingness to negotiate. Neither bargaining power nor profitability measures matter when all factors are considered simultaneously. Employers are 14% more likely to negotiate if they deem remote working to be productivity-enhancing (relative to finding it productivity-neutral). This may be because they perceive fewer productivity risks around potentially allowing workers to opt for continued remote working. In turn, employers are 21% less likely to negotiate if they find it to be productivity-diminishing.


Increased remote working is likely to remain a permanent feature of the post-pandemic British economy. But the extent of this will depend on both firms’ and workers’ preferences, and these may not be aligned. Firms may favour lower overhead costs, while workers may seek better work-life balance or shorter commutes. And even if they both seek to promote productivity, they may disagree over the impact of remote working on this, for instance if workers fail to internalise its effects on team cohesion or ideas generation, or if firms ignore its impact on worker engagement. When workers’ and firms’ preferences differ, an employer’s willingness to negotiate with its employees is likely to depend largely on its own assessment of the productivity impacts of remote working.

Lena Anayi works in the Bank’s Structural Economics Division.

If you want to get in touch, please email us at or leave a comment below.

Comments will only appear once approved by a moderator, and are only published where a full name is supplied. Bank Underground is a blog for Bank of England staff to share views that challenge – or support – prevailing policy orthodoxies. The views expressed here are those of the authors, and are not necessarily those of the Bank of England, or its policy committees.

One thought on “Remote working, bargaining power and productivity

  1. This is an important and interesting piece of research which should be more widely publicised.

Comments are closed.