An estimated 41% of full-time employees in the US work either fully or partly remotely, sparking discussion about the value employees place on remote work and its impact on wage-setting. This column estimates the value US tech industry employees assign to remote work. On average, individuals are willing to forgo 25% of their total compensation for a job that offers partially or fully remote work but is otherwise identical to a fully in-person position. Tech employers offer slightly more pay for remote positions than otherwise identical in-person positions.
Remote work has become increasingly common in recent years (Hansen et al. 2023). It is estimated that 11.8% of full-time employees in the US work fully remotely, with an additional 29% working partly remotely (Barrero et al. 2023). This shift has led to much discussion (Stanton et al. 2020) about the value employees place on remote work and how it impacts wage-setting practices. For example, a group of studies uses survey data to estimate how much of a pay cut individuals would accept for remote work (e.g. Barrero et al. 2021, Bartik et al. 2024, Mas and Pallais 2017).
In a recent paper (Cullen et al. 2025a), we provide new insights into the debate. We estimate the value employees assign to remote work by analysing revealed preferences in a high-stakes, real-world setting. Additionally, we examine compensation data to assess how preferences for remote work have influenced employer wage-setting practices.
Our study utilises a unique dataset collected for another study (Cullen et al. 2025b). This dataset tracks the job offers received by 1,396 workers and the offers they ultimately chose. The survey respondents belong to the US tech industry, and they were recruited in collaboration between the researchers and Levels.fyi, a leading platform that provides comprehensive wage data for professionals in the tech sector. The tech sector provides a particularly relevant setting for analysing remote work, given its high prevalence of flexible work arrangements and its status as a high-paying, innovative industry (Barrero et al. 2023).
The average participant is 32 years old, has seven years of work experience, and is being offered $239,000 in total annual compensation. Common roles include software engineer, product manager, and data scientist, with top employers such as Google, Meta, and Apple. About 18.3% of job offers are fully in-person, while 81.7% are remote. Women make up 16.3% of the sample, which is slightly lower than some estimates for the tech industry.
The survey gathered rich data on job-offer characteristics such as total cash compensation, job location, and whether the position was remote. We also incorporated additional data, such as employer ratings from Glassdoor.com. To estimate the amenity value of remote work, we used a conditional logit model. The model allowed us to analyse the choices individuals made when selecting job offers, while controlling for various offer characteristics such as employer ratings, company type, benefits, and location. Our key finding is that, on average, individuals are willing to forgo approximately 25% of their total compensation for a job that is otherwise identical but offers partially or fully remote work instead of being fully in-person.
Previous studies have explored variations of a survey question asking how much of a pay cut individuals would accept to work remotely. Our estimate of the value of remote work is three to five times as high as the estimates from these previous studies. For example, Barrero et al. (2021) found that employees were willing to accept a 7% pay cut to work from home 2–3 days per week. Bartik et al. (2024) reported that the median worker would accept a pay cut of around 5% for remote work. And Mas and Pallais (2017) estimated that the average worker would accept a 10% pay cut for remote work.
One possible reason for the discrepancy is the difference in samples, as our study focuses on tech workers. Remote work may be more appealing to high earners, and tech jobs may be particularly suited for it. This likely explains part of the gap but not all of it. Using the latest data from the Survey of Working Arrangements and Attitudes, we find that the average respondent would take a 7.0% pay cut for remote work. Among a more comparable group – IT workers earning over $150,000 – that figure rises to 15.8%, but is still below our 25% estimate. Thus, we attribute part of the discrepancy to methodological differences. In other words, existing survey methods may underestimate the strength of preferences for remote work. For example, when asked how much of a pay cut they would accept to work remotely, employees may underreport the true amount due to concerns that their employers could use their responses to justify salary reductions
In the second part of the study, we test the compensating wage differential hypothesis – i.e. given the strong preference for remote work, we would expect remote positions to pay less on average than otherwise identical in-person positions. We use novel data on salaries for tech jobs collected by Levels.fyi. This data allows us to compare individuals who started their jobs at about the same time (between June 2023 and June 2024), for the same company (e.g. Google), in the same company-specific position title (e.g. Level-3 Software Engineer in Back-End API Development), in the same location (e.g. Bay Area), and with a similar experience level (e.g. 0–3 years). We reject the hypothesis of compensating wage differentials. If anything, remote positions pay on average slightly more (1.1%) than otherwise identical in-person positions.
We propose potential explanations for this puzzle. For example, companies may hesitate to create pay gaps between remote and in-office workers to maintain fairness and avoid legal risks. Since women are more likely to work remotely, reducing remote pay could unintentionally widen the gender wage gap. Another possible explanation is worker sorting: to attract top talent, firms may offer remote work as a perk alongside – or instead of – higher wages. Another explanation could be optimisation frictions: as remote work is still relatively new (Shah et al. 2024), companies may be in the process of adapting and have yet to refine their wage-setting practices.
source: cepr.org
