Layoffs Cast a Long Shadow

Key findings:

While considering the option of a potential layoff, one HR executive at a Fortune 500 company told me that “layoffs are a culture killer” and should be seen as a last resort - yet the labor market since 2021 has been characterized by mass layoffs. IO Psychologist Wayne Cascio (2005) includes the failure to manage layoff survivors effectively as one of ten mistakes to avoid when restructuring:

Employee morale is often the first casualty of downsizing, as survivors become narrow-minded, self-absorbed, and risk averse. Many firms underestimate the emotional damage that survivors suffer by watching others lose their jobs. In fact, a great deal of research shows that survivors often suffer from heightened levels of stress, burnout, uncertainty about their own roles in the new organization, and an overall sense of betrayal. …[S]urvivors are looking for signals such as the following. Were departing employees treated fairly, and with dignity and respect? Why should I stay? What new opportunities will be available to me if I choose to do so? Is there a new business strategy to help us do a better job of competing in the marketplace?

This research uses Glassdoor reviews to analyze the impact of layoffs in the post-COVID world of work. It fits into a broader body of research that tracks the impact of layoffs on voluntary attrition and employee disengagement. Our research finds that layoffs have a significant and lasting impact on worker satisfaction, extending beyond the two years we analyze after a layoff event.  

Identifying layoffs

Layoffs.fyi tracks and cites news articles related to layoffs, neatly identifying the dates when these layoffs became public information. We begin our layoffs dataset with all U.S.-based layoffs identified on the site with at least 250 employees impacted. Since these layoffs skew towards the tech sector, we supplement with WARN Act notices for companies that are not included in the layoffs.fyi dataset. The 1988 WARN Act requires employers above a certain size to provide at least 60 days public notice before a mass layoff or location closure.

WARN Act notices pose two challenges for this research:

First, the notices occur at the location level, meaning that a large layoff from a single employer can be spread out across multiple locations and dates. To address this issue, we look at 90-day windows within an employer. If an employer files WARN Act notices impacting more than 250 employees in a 90 day window, we consider this a layoff event.

Second, companies frequently choose to announce layoffs before the WARN notice. Various actors monitor WARN Act notices, including journalists, so by announcing beforehand companies can break news on their own terms. To address this issue, we conducted time-limited web searches for each WARN Act layoff wave in our dataset to identify pre-layoff announcements. This aligns the layoff date with when the information became public, as in the layoffs.fyi data.

This case illustrates both potential issues: Activision-Blizzard filed six WARN Act notices on February 2, 2024 for six California locations, with the layoffs effective on March 30, 2024. The number of employees impacted ranged from 49 in the Woodland Hills location to 479 in Irvine, with a total of 899 positions eliminated across all locations. However, the Microsoft Gaming CEO announced the layoff a week earlier on January 25th, as reported in The Verge. In this case, we identify this as a single layoff impacting 899 employees, announced on January 25. In cases where we could not identify an earlier news article, we use the date of the earliest layoff in the wave as the layoff announcement date.

The result is a dataset of 304 layoff events from 197 companies between March 2021 and April 2025. Figure 1 presents layoffs by month, and the rolling 6-month mean of events. As seen in the figure, major layoffs were rare in 2021, increased in the second half of 2022, peaked in early 2023, and remained elevated above the 2021 level through early 2025.

Glassdoor reviews

We combine the layoffs data with Glassdoor reviews at the employer level. Employees may leave anonymous reviews of their employers, which include one to five star ratings on the company overall and six workplace factors (career opportunities, compensation & benefits, culture & values, diversity & inclusion, senior management, and work/life balance), some binary or ternary ratings (e.g. CEO approval), and open text fields where reviewers can list pros, cons, and advice to management. The review also captures information about the employee, such as job title, job location, and tenure.

Our analysis pairs each review to a layoff event, and reviews are included in the final dataset if they were left between 360 days before the layoff and 720 days afterwards. The result is a dataset of 298,314 reviews from the 197 employers in our dataset, centered around the layoff event.

For the companies with multiple layoff events in the time period, reviews are considered “pre-layoff” If they occur before a company’s first layoff, and then the counter restarts for each subsequent layoff wave.

Review volume & layoff events

The combined dataset of layoff events & reviews shows two clear trends: there is a spike in review volume immediately after the layoff, and the proportion of reviews explicitly mentioning layoffs increases dramatically. Figure 2.1 shows the increase in review volume by a 7-day trailing average of reviews, indexed to the period 1–2 months before the layoff. Review volume surges immediately after the layoff, and is 44% higher the week immediately following a layoff event.

Figure 2.2 demonstrates the salience of the reviews, showing the percentage of reviews that explicitly mention layoffs using terms like “layoffs,” “laid off,” or “laying off.” Mentions spike immediately following the layoff, but remain a common feature of reviews even after the initial spike, appearing in roughly 5% of reviews and indicating the durable impact layoffs have on the employee psyche.

Methodology

Our core question is how layoffs impact an employer’s brand, and for how long. To answer this question, we utilize a semi-parametric approach using a three-way fixed effects linear regression, with employer, month-year, and high-level occupation fixed effects. Since company performance should be correlated with both Glassdoor ratings and layoffs, we also control for medium-term stock performance.

Our variable of interest is a series of “months since layoff” dummy variables, which allow flexibility in how the data can show how ratings react to and recover from the layoff event. Fixed effects allow us to ensure that firm- and time-specific variables are not confounding the observed impact. Since ratings and the impact of layoffs can be very different for frontline employees, we interact each of these fixed effect variables and the stock performance control with a frontline employee dummy variable. We also include a fixed effect for the wave of a layoff, which allows additional waves to have additional impacts on ratings.

rating_itk = months–since–layoff_i + frontline_i(1+month–year_t+firm_k+ stock performance_kt+layoff wave_i)+e_it

This approach allows us to see how Glassdoor ratings react to layoffs while controlling for company performance, and unobservable differences over time and between companies that are not related to layoffs. We allow all those differences to vary for frontline workers. Table 2.1 shows the summary statistics for the variables included in the base regression. We also fit an ordered logit regression with an expanded set of controls reflecting the significant predictors explored in the following section, and found similar results: current employees were 19.4% more likely to leave a one-point lower review in the six months following a layoff relative to the six months preceding one, and results were statistically significant in the first year and started to recover in the second with mixed significance.

Caveats

We cannot definitively say what would have happened to these ratings if these companies had not engaged in layoffs. The decision to lay off employees is clearly related to other factors, like poor business performance, that may impact an employee’s review, and maybe reviews would have fallen due to these other factors even without the layoff. We address this concern by controlling for within-company changes in stock performance. It could also be the case that announcing a layoff provides new information about the company, which shocks reviews. As such, we should interpret the results as being the combination of these two related effects: what is the direct impact of losing colleagues to a layoff, and how does the announcement of a layoff change the way you perceive your employer?

Finally, since Glassdoor reviews are user-initiated, they may not be representative of the entire population. Independent research has shown a close correlation (0.69) between internal employee engagement measures and Glassdoor ratings for companies with more than 100 reviews (which is 96% of the employers in our sample; seven layoff employers have between 77-99 reviews.) We find the best evidence for causality to be how clearly the layoff announcement coincides with the change in review content & volume, and the robustness of these results when including controls. 

As demonstrated in the above chart, layoffs trigger an immediate impact on Glassdoor ratings, with the lowest average ratings occurring the month of the layoff. While ratings rebound slightly, in month two, the monthly estimates remain significantly lower than the benchmark of six-months prior to the layoff for at least two years.

Ratings show a small but statistically significant drop preceding the layoff. This may indicate that an employer’s brand begins to suffer before the layoff, but it also could be due to unidentified pre-layoff announcements for layoffs among WARN companies - meaning some employers may have announced their layoffs internally before day zero, triggering some negative reviews. This pre-layoff decline does not appear if we restrict the sample to the layoffs.fyi dataset, which are exclusively identified by breaking news of layoffs.

Figure 3.3 presents the estimated impacts of layoffs, allowing the monthly estimates to vary between current and former employees. The shading shows the statistical significance of the estimates for current employees, while former employees estimates are represented in purple. Generally, former employees show no statistically significant difference in their ratings pre- and post-layoffs.

The top line in the above chart clearly illustrates the first driver of lower ratings: an immediate fall in average ratings from current employees. Ratings from current employees do recover, and after 21 months the results tend to be statistically insignificant though they remain directionally lower than pre-layoff levels. However, the results here are consistent with the view of gradual recovery in ratings from current employees, beginning roughly 12 months after the layoff event but not fully recovering even 24 months after the layoff. With the pace of rating recovery present in the second year after the layoff, ratings will reach average pre-layoff levels 32 months after the layoff.

Understanding the second driver of lower ratings requires another piece of information. Surprisingly, the average of ratings from former employees do not change significantly post-layoff - that is, employees impacted by the layoff are not harsher in their reviews than other former employees leaving reviews before the layoff. In other words, employees who quit or were fired pre-layoff are just as unhappy on average as employees who are laid off. Former employees still drive the overall rating down because the layoff triggers a wave of new reviews from former employees - presented in figure 3.4.

Before the layoff, former employees represent 26% of all reviews, but 33% afterwards - particularly in the first six months. This increase in the proportion of reviews coming from this lower-rating group also pushes ratings down. 

For the balance of the report, we analyze only the impacts on current employees. While terminated employees obviously experience the brunt of the layoff, we are focused on the long shadow layoffs cast within the organization. We use the change in ratings for current employees to measure the substantial and durable internal impact of a layoff.

Layoffs are most dramatically associated with:

On average in the first six months after the layoff, companies in the top tercile see their ratings from current employees drop 0.22 points, compared to 0.18 points for the middle tercile and an insignificant 0.02 in the bottom tercile. Companies who built successful employer brands over years are the ones who stand to lose the most in a layoff. This theme will re-surface in other results: the best ratings are the most sensitive to layoffs, because this is where employers have invested heavily to build goodwill. On the flip side, layoffs don't change the views of employees who already have a negative opinion of their employers.

Comparing self-reported salaries to our predictions, we divide employees into three categories: those whose pay is aligned with predictions (±15%, defined as core talent, representing 55% of workers in the dataset), those whose pay is much higher than predicted (≥15% above prediction, defined as key talent, representing 24% of the dataset), and those whose pay is much lower than predicted (≥15% below prediction, defined as marginal talent, representing 21% of the dataset). 

While pay does not always accurately reflect a firm’s assessment of talent, many firms espouse a pay-for-performance culture and so the two are likely to be correlated. The data does show that, before the layoff, employee satisfaction ratings are highest for key talent and lowest for marginal talent with a highly statistically significant gap between them.

We see from this chart that ratings are highest for key talent before a layoff, but that they take the largest drop: 0.16 points on average in the first six months. Core talent ratings start somewhat lower and drop only slightly less (0.15 points in the first six months), while marginal talent starts with the lowest ratings, dropping only 0.08 points in the first six months. Employers go to great lengths to attract, retain, and motivate key talent - and these efforts work, as evidenced by higher overall ratings from key talent before relative to other groups. Layoffs set these efforts back, with post-layoff ratings from key talent stabilizing similar to average pre-layoff ratings from core talent.

The impact of subsequent layoff waves is roughly twice as severe for the first few months after the layoff. Estimates become much noisier in the second year with the restricted sample, but we can see that by the end of the first year, average ratings between initial and subsequent layoff waves are roughly aligned. In the short term, subsequent waves of layoff trigger an even sharper decline in reviews from current employees.

Baseline ratings are similar between management and non-managers before the layoff. In the six months after the layoff, ratings drop 0.12 points for individual contributors and 0.17 for managers and above. Ratings tend to remain lower into the second year, but the trends become statistically indistinguishable 18-24 months after the layoff.

All tenure groups see statistically significant drops in ratings relative to six months pre-layoff, and the ordinality of ratings is consistent post-layoff, with the lowest ratings coming from the longest-tenured employees. All tenure groups see a recovery of ratings within the two-year window.

The estimated impact on ratings may not sound large at first blush, but the vast majority of employer ratings fall in the 3 to 4.5 star ratings: only 1.2% of all employers with at least 50 reviews on Glassdoor had an average rating below 3 stars in 2021, and only 6.2% had ratings higher than 4.5 stars. The interquartile range is only 0.5 points wide, ranging from 3.7 to 4.2, and 90% of employers’ average ratings fall between 3.3 and 4.5.

The median employer in the layoffs dataset was slightly above average among all companies with at least 50 reviews in 2021: the overall rating of 3.99 put them in the 56th percentile. The 0.13 point fall in ratings drops them to the 41st percentile, well below average.

Anxiety keyword prevalence increases 64% among current employees in the six months after a layoff. Unlike other outcomes explored in this report, this impact worsens in the second year following the layoff, so post-layoff anxiety becomes a more common feature of reviews after the layoff.

Thematically, these reviews reinforce the findings on the review factors that take the biggest hits: ratings of senior management, career ratings, and culture/values.

Layoffs clearly negatively impact the lived experience of employees, including those who were not laid off.

Generally, we find that the likelihood of a reviewer to click on or apply to a job is consistent before and after a layoff event, but the layoff event motivates a massive increase in the number of current employees going to Glassdoor to leave reviews and thus also a proportionate increase in the number applying for jobs after a layoff. 

Figure 4.3 shows the normalized number of current employees who are starting new job applications on Glassdoor within 30 days after leaving a review. We find a 40% increase in the number of current employees applying for jobs on Glassdoor in the 12 months after a layoff. This is in the ballpark of other research, which estimates that a 1% reduction in force increases voluntary attrition by 31%. More concerning for the organizations, this increase is disproportionately key talent (p = 0.02 for 𝝌² test), suggesting that this voluntary attrition includes more people the organization does not want to lose.

Decreased Employee Engagement: 1.2% of payroll

There are multiple mechanisms through which decreased employee engagement translates to increased costs or decreased revenue. Previous research has even tied these directly to Glassdoor review ratings, e.g. customer satisfaction rates are correlated with changes in a company’s Glassdoor ratings.

Some of these measures are overlapping, so we favor identifying a broad measure of the cost of disengagement. Gallup has estimated that actively disengaged employees cost employers 34% of their salary. They also estimate that 17% of employees were actively disengaged in 2024.

Combining this with Glassdoor data, we find a similar proportion of employees providing 1- or 2-star ratings. Specifically at the layoff companies analyzed here, we find 14.0% of current employees give 1- or 2-star overall ratings before layoffs, and 17.6% do so in the first year after a layoff, marking a 26% increase in employee disengagement. Using this 3.6 percentage point change in active disengagement multiplied by Gallup’s 34% of payroll as the cost, we estimate that increases in active disengagement costs employers 1.2% of post-layoff payroll.

Increased Voluntary Attrition: 4.1% of payroll

Similar to employee engagement, prior research has made a direct link between Glassdoor ratings and the cost to fill a position, e.g. higher Glassdoor ratings lead to more job search behavior on other job posting sites. Similar to our cost estimate of employee disengagement, we favor a broad measure of the cost of attrition rather than estimating each potentially overlapping mechanism.

We assume that the increase in job search behavior is proportionate to the increase in voluntary attrition the year after a layoff, implying a 40% increase in voluntary attrition. We combine this with a baseline voluntary attrition rate of 13.5% from Mercer. Estimates for the cost of backfilling a role vary widely, likely reflecting both the difficulty of measurement and variability across roles. One commonly cited range is that costs fall between half and twice an employees’ salary. We use 75% of an employee’s salary as a moderate estimate of the cost. Putting these together, we calculate that voluntary attrition costs an additional 4.1% of payroll in the year following a layoff.

Total Cost Estimate

In total, we estimate that layoffs cost an additional 5.3% of total payroll through higher rates of active disengagement and voluntary turnover. Frequently, layoff announcements include the percentage of the overall workforce that is impacted (e.g. company X lays off 6,000 employees, or 3% of its workforce.) We use this information when available to infer the number of employees who were not impacted by the layoff, and then scale that number up to cover missing information, estimating 3.8 million layoff survivors. We then use the average wage index of $66,622 from the Social Security Administration to calculate the total payroll of these workers. Finally, some employees survived three or more layoffs over the time period. Since there were 197 initial layoffs and 107 subsequent layoffs, we scale up the total cost estimate by 54% accordingly.

The final estimate is that these layoffs costed these companies an additional $20.8 billion dollars in lost productivity and increased attrition between April 2021 and April 2025.

Employers should weigh these impacts carefully when considering a layoff, and recognize that their decision will impact their employer brand for years to come. It takes years to build a successful employer brand, and the companies with the strongest employer brands have the most to lose in a layoff.