Work from Home Will Have a Bigger Impact on the Economy than AI

No, it’s not because AI will have a small impact.

Thank you for reading our work! If you haven’t yet subscribed, please subscribe below:

If you would like to support us further with reaching our 1,000 subscriber goal by year end, please consider sharing and liking this article!

As Nominal News grows larger, we will be able to make this a full-time project, and provide more content. If you would like to suggest a topic for us to cover, please leave it in the comment section.

---

Photo by vadim kaipov on Unsplash

We are now 4 years removed from the start of the Covid pandemic, during which many workers transitioned to remote work, which became a prominent form of work organization. Recent research has shown that the work from home (WFH) and hybrid work models have had large positive impacts on the economy. WFH will, in my opinion, have a greater influence on economic development in the next few decades than AI.

Work from Home Research

During the pandemic, many firms changed their work model from a primarily office-centric model to WFH. Both firms and workers discovered that WFH is a feasible working method, and to this day, there is still a significant number of people working either fully remote or in hybrid work set-ups. As can be seen in the following chart, it appears this new form of work organization is here to say.

The addition of WFH as a form of work organization has naturally led to many questions about whether it is beneficial. We will focus on 4 key issues that have been present in the WFH discussions:

• WFH and productivity;

• WFH and innovation;

• WFH and labor opportunities; and

• WFH and its impact on geography.

Productivity

There have been a lot of concerns around worker productivity regarding WFH. When referring to productivity, we are interested in the question of whether WFH workers generate more or less ‘output’ per hour. There has been an influx of new research papers on this question and we have touched upon it in our discussion of the misinterpretation of the research by the Economist.

Many articles and commenters point to headlines that show that worker productivity is lower when working from home. However, the research on this topic often focuses on very specific, repetitive tasks such as ‘number of calls per hour’, ‘number of typed words’ or ‘number of lines of code’. In most cases, even if research finds that workers are ‘less productive’, the total output (i.e. number of calls made) in a day is the same as when people worked in an office setting. 

To investigate WFH productivity, Gibbs, Mengel and Siemroth (2021) studied 10,000 employees in an IT company. Gibbs et al. were able to collect data on exactly what tools and software the employees were using on their computers, and thus were able to track when these employees were working. Gibbs et al. found that WFH workers generally met all their targeted deadlines, but spent more hours working. This suggests that ‘productivity’ when working from home is lower. 

Interestingly, lower productivity depended on the type of worker. Workers that had children saw lower productivity than workers without children. Women also saw larger productivity drops. However, workers that had been at the company for a longer time were actually more productive in a WFH setting.

In contrast to Gibbs et al., research by Bloom, Liang, Roberts and Ying (2013) showed that call center workers that went remote did 13% more work, of which 9 percentage points was due to working more minutes per shift, as people were sick less often and took shorter breaks.

Another study by Emanuel and Harrington (2023) looked at the performance of call center workers at one of the largest companies in the US. At this firm, prior to the Covid pandemic, some call center workers worked from an office while others worked from home. Using call log data and workers’ schedules, Emanuel and Harrington were able to track how many calls per hour workers were responding to, while controlling for factors such as spending time on work training or other company tasks.

During the Covid pandemic, all workers were forced to work from home. The workers that used to work in the office, were now also working from home. Compared to their office level, the number of calls they took per hour dropped by 4%. However, one of the reasons for this fall in productivity was attributed to an operational reason – it was tougher for workers to connect to their managers or receive assistance from others in dealing with challenging calls, resulting in more time spent per call. Nonetheless, overall customer satisfaction did not change when workers moved to remote work.

Separately, Dutcher and Saral (2022) studied how team productivity may be impacted by work from home and found unexpected results. Using an experimental set up, participants were randomly assigned to either work from a location of their choice or in the laboratory. In the first round of the experiment, each participant performed a series of transcription tasks for five minutes, for which they got paid a per completed task rate (piece rate pay). In the next two rounds, each participant was assigned to a team of three workers. Prior to the beginning of each of the rounds, the participants were told that they would be working with two remote workers in one round and two ‘office’ (i.e. in the laboratory) workers in the other round. 

Each participant would also choose how many of their tasks to contribute to their ‘own account’ or to a ‘team account’. For the tasks in their ‘own account’, they would be paid the same piece rate as before. For the tasks in the ‘team account’, the pay per task was higher. The caveat, however, was that in the team account the pay would be based on the participant that contributed the least amount of tasks. Thus, if you do 10 tasks, and contribute all 10 to the team account, while one of your team members who also does 10 tasks, but contributes 5 tasks, you would receive pay only for the 5 tasks. Thus, how many tasks you contribute to the ‘team account’ depends on how many tasks your fellow team members will complete and contribute to the team account. 

Dutcher and Saral found that ‘office’ workers believed that the remote workers would contribute fewer tasks, and, therefore, the ‘office’ workers chose to contribute less of their tasks to the team account. The overall team performance was dragged down not because remote workers were not contributing their effort (their tasks), but by ‘office’ workers that put in less ‘team’ effort, because they thought remote workers would be ‘lazier’. This suggests that team productivity can be influenced by presumptions of the work effort of remote workers.

From the academic research on this topic of productivity and WFH, we can so far conclude a few things:

1. Overall output per worker has stayed the same when working from home or from an office.

2. Productivity may or may not decrease. However, whether productivity is a relevant measure can be questioned as workers still produce the same output. 

3. Productivity may also be influenced by the perceptions between remote and office-based workers. 

4. Better knowledge of the firm processes and access to colleagues appears to matter in making WFH more productive.

It is important to reiterate that the option of working from home on a large scale has only been available for several years, while the office-based model has been around for many decades. Thus, without any significant adaptation of work places to this new reality, it appears that the WFH option has no adverse effects on output. Once we improve the WFH model, it is not inconceivable that productivity could unambiguously become higher than in offices.

Innovation

Frey and Presidente (2023) analyzed how remote work impacts the development of disruptive discoveries. Looking at research teams that used to work from the same location, but were split into working away from each other, Frey and Presidente found that, on average, the probability of the team producing disruptive research¹ fell by 10%.² However, when Frey and Presidente looked at the data only after 2010, it turns out there was no difference in the probability of creating disruptive research whether working in person or remotely. There actually may have even been an increase in innovation when working remotely. The intuition as to why remote work became just as effective as colocation after 2010 is quite simple: we have developed many technologies such as Zoom, Teams, Dropbox that significantly improved remote work productivity.³

Labor Impacts

Remote work has dramatically changed the job landscape for two groups: people with disabilities and people that take care of dependents, especially mothers. As established in a report by the Economic Innovation Group, people with disabilities were significantly more likely (by 3.5 percentage points) to find employment in 2022 (after the Covid pandemic) than in 2019 (before the pandemic). At the same time, people without disabilities, at the time of the report, had a lower likelihood of being employed in 2022 than in 2019.⁴

Regarding mothers, Harrington and Kahn (2023) studied the employment outcomes of WFH on working mothers. In industries where a large share of workers were WFH, the gap in employment between working mothers and working women without children reduced.⁵ Moreover, the incomes of mothers increased by 1.3 percentage points relative to other women in the workforce.

Harrington and Kahn hypothesize the reason that we may be observing this pattern is due to the change in the behavior of firms towards mothers requesting WFH. Prior to the pandemic, WFH  may have sent a signal to the firm that the mother will commit less effort to the firm, and thus firms chose to invest less in mothers’ skill development (such as training or giving them few opportunities). In jobs where this investment is crucial to performance, such as finance and marketing industries, mothers had to basically choose to either commit to going to the office or exiting their job. With the recent increase in WFH due to the pandemic, it turns out firms were wrong in believing working mothers would not put in the effort. Thus, firms started employing mothers that prefer a WFH setup. 

The labor developments of the option of WFH appear to have two major contributions to the wider economy:

1. It increases the available opportunities to individuals that were previously unfairly disadvantaged in the labor force. This also increases total available labor and increases labor’s diversity which is unambiguously a good thing for the economy (and the individuals involved).

2. The option for mothers (and fathers) to WFH, without completely sacrificing their careers, allows them to commit more time to their children. Research has shown that parental time is an important input in the development of children (Del Boca, Flinn and Wiswall, 2013; Del Bono, Fracesconi, Kelly and Sacker, 2016). Moreover, since mothers are now able to earn more, this is another source of additional monetary investment in their children.

Altering Geography

WFH has also started to transform our urban organization. As we extensively suggested in our previous article on determinants of house prices, WFH may improve urban affordability. Brueckner, Kahn and Lin (2023) have shown both theoretically and empirically that high productivity cities (high income) saw price and rent drops due to WFH, while low productivity cities saw price increases, as people could move to these ‘cheaper’ cities without foregoing their higher income of the more ‘expensive’ city. Moreover, prices within cities flattened – there was less price dispersion based on the distance from the city center (which is where jobs are typically located) due to WFH, with suburbs becoming more expensive and city cores becoming cheaper.

Gill, Hensvik and Skans (2023) looked at the impact of WFH on restaurants in residential areas and found that after Covid pandemic, their sales grew relatively more than that of restaurants in the city core. Moreover, restaurant workers in residential areas saw their pay grow by $117 per month more than the income of restaurant workers in the commercial-office areas. This was driven by the fact that they worked longer hours actually, implying there is more business for these residential restaurants. 

The results of these two studies are not surprising. De-linking jobs from land will have profound impacts on how we decide where to live. Cities will change because of WFH, as we already see many offices closing or standing vacant. Office vacancy rates in the US are at 20%, which was last seen in 1979. This will increase residential housing supply in the long run, as land that was previously used for offices can now be converted into housing. Moreover, people may now be able to optimally choose whether they prefer a city or a different lifestyle without having to consider the local job market. This may result in economic growth for new regions in countries, potentially at an expense to the current economically-developed areas.⁶

The WFH Revolution

We are just at the beginning of the WFH work organization revolution. Research has already shown that WFH has profound impacts on the economy:

• It has enabled more people to find work and given them more work flexibility, including less time spent on commuting;

• It has reduced the ‘motherhood’ penalty, allowing women to both work and raise children, which will have long run positive economic outcomes;

• It may have opened new avenues for collaboration and innovation;

• It has transformed our decisions of where to live and where businesses open (residential areas vs city cores).

As our WFH-related technologies improve and work-organization continues to integrate WFH, we should expect to see continued progress in the above listed dimensions. Younger generations that grow up with the WFH option will also be able to better adapt to this reality. This does not mean that WFH will, or should, replace other forms of work organization. Office-based and colocation modes will still be valuable. The right balance between WFH and office-based work, along with improved worker training, will be struck in the future, increasing both output and worker well-being. This will result in a better allocation of resources which will translate into economic growth. 

My prediction that WFH will have a bigger impact than current AI is not based on the fact that I don’t believe AI will have significant benefits.⁷ It’s more that WFH is more akin to the economic shifts we saw with women entering the labor force due to both improvement in women’s rights and the development of home appliances. Women entering the labor force made a profound change on the economy, as the labor force doubled in a short span of time. The changes to the economy that’ll be brought up by WFH will be similar in nature, albeit likely on a smaller scale. 

There is a lot of excitement around AI, and how it might change the future. We need to get as excited about the changes WFH will bring to the global economy.

---

Interesting Reads from the Week

1. Article: James Bryant explains why markets and CEOs might be over-estimating the current state of AI. Also, check out his Substack if you’re interested in tech, AI, art and poetry!

2. Articles: We’ve had a lot of new data on the labor market. Check out Guy Berger‘s update here and Pascal Michaillat ‘s update here.

3. Article: We’ve also had new inflation data come in for the month of March that was hotter than ‘expected’. American Economy Daily describes this release and the concerns it raises. One thing worth adding – there is still a lot of lagged shelter inflation, which we discussed might be ‘artificially’ raising current inflation.

---

If you enjoyed this article, you may also enjoy the following ones from Nominal News:

The Work from Home Illusion: Is It? (July 15, 2023) – the Economist covered several work from home studies. The conclusions the Economist drew from them, however, were quite wrong. What did these studies actually say, and why work from home might be the optimal form of work.

ChatGPT and AI - Predictions on Labor and the Macroeconomy (April 3, 2023) – with many predicting AI technology will revolutionize work, how can we use economic models and frameworks to think through the implications of this new technology, and why many predictions in the media are contradictory.

Survivorship Bias and Opinions (September 24, 2023) – we are typically unaware of how survivorship bias can influence our perception of the world. Not only do we often see perspectives that ‘survived’, but we also make assumptions from our own experiences.

1. Disruptive research was measured by looking at citation patterns. A research paper typically cites other work. If a research paper (call it Research Paper A) cited a set of other papers (Research Paper B), but then future research papers (Research Paper C), cited Research Paper A but not Research Paper B, Research Paper A would be considered ‘disruptive’.

2. This result, that innovation dropped due to remote work, was heavily discussed in the media when it was reported in another paper.

3. It’s not surprising that this improvement in disruptive innovation has occurred. As highlighted by WFH researcher Nick Bloom, the average office distance between research team members has increased from 100km to a 1000km over the last 60 years.

4. This difference suggests that it was not an economy wide effect, i.e. it was easier for everyone to become employed regardless of disability status, but rather the changes between 2019 and 2022 gave people with disabilities a higher probability to become employed.

5. For a 0.6 percentage point increase in the share of workers (with a particular college degree) that work from home, the gap in employment level between working mothers and working women without children reduced by almost 0.8 percentage points.

6. The impacts could even be felt across countries if firms allow employees to work from abroad. In the Fray and Presidente study on disruptive innovations, many researchers already work across borders.

7. Some of my thoughts on how to predict the impact of AI can be found here. Interestingly, AI might even make WFH better, as AI might make WFH workers more productive.