Economists estimate wage theft costs workers more than $15 billion a year. Journalists at the Center for Public Integrity investigated how effectively the Department of Labor’s Wage and Hour Division (WHD) combats this systemic problem.
Our analysis of 15 years of data received from a Freedom of Information Act request, combined with other data and documents, and supplemented by extensive interviews of human sources, found:
* Companies that repeatedly steal from employees are rarely fined extra;
* USPS is one of the worst offenders;
* the higher an industry’s share of immgrant workers, the greater the rate of wage theft.
After our main USPS investigation piece was published, Fernández Campbell and Public Integrity received emails from more than 100 current and former USPS mail carriers, who shared their experiences with wage theft at the agency.
Many of them believed they were the only ones struggling to get paid for all their work. Our story allowed them to realize that they were not alone.
In addition, Public Integrity hosted a virtual Q&A with legal experts to help USPS workers and provide advice about the proper steps to take when they suspect or believe that they have been cheated out of pay.
Our primary source of data for this series was an export of the Department of Labor Wage and Hour Division’s Wage and Hour Investigative Support and Reporting Database (WHISARD). We got this data through a FOIA request. The data came as a series of dozens of different tables exported as csv files. We essentially reconstructed the WHISARD database (or a portion of it) by importing those csv files into an R environment and linking the tables. We also brought additional data into that database, primarily American Community Survey microdata on employment by nativity status and industry downloaded from IPUMS USA.
Data used and sources:
15 years worth of Department of Labor Wage and Hour Division cases via the DoL WHD through a FOIA request.
American Community Survey microdata on employment by nativity status and industry via download from IPUMS USA.
10 years of private arbitration data leaked by a member of the National Association of Letter Carriers, one of the unions representing USPS employees whose wages were stolen.
In terms of specific tools and technologies, we primarily used R and RStudio for the data analysis. We also used Google Sheets to share our data with partners. We used Datawrapper to produce visualizations for the stories.
What was the hardest part of this project?
Analyzing this data presented some challenges.
For the story focused on companies we classified as “repeat offenders” because of repeat violations, we had to determine a way to group employers. We settled on grouping by their employer identification number (EIN), commonly used for tax purposes. However, EINs were either missing or withheld in about one third of cases involving minimum wage or overtime violations. We thus excluded those cases from portions of the analysis that required identifying specific companies. We used the entire universe of cases to calculate other figures (such as the total money improperly withheld per year).
We faced another challenge as we began working on the focused on immigrant workers: How to quantify the number of immigrant workers and specifically the industries that tend to cheat immigrant workers and get in trouble for that? This became a problem particularly because WHD does not document immigration status when conducting investigations. The best we could do to overcome this challenge was to determine if a correlation existed between the proportion of a given industry’s workforce that was foreign-born and the rate of wage theft in that industry. To conduct an analysis, we combined the WHD wage theft data with U.S. Census Bureau American Community Survey microdata downloaded from IPUMS USA, which allowed us to calculate the proportion of foreign-born workers in each industry. The industry codes in the census data differed from the industry codes attached to the employers in the wage theft data so we had to use a crosswalk table to convert them. Once we had successfully crosswalked the codes, we were able to join the two sets of data together and calculate the rate of wage theft cases per 100,000 workers for 95 industries along with the proportion of each industry’s workforce that was foreign-born.
What can others learn from this project?
First, don’t be satisfied with the information government agencies put up on their websites. The Department of Labor’s Wage and Hour Division puts statistics up on their website that can be downloaded as spreadsheets. But missing were crucial details of each case such as employer name. That’s why reporter Alexia Fernández Campbell ended up filing a FOIA request for the database that agency personnel actually use to track cases. We could not have produced a story nearly as in-depth otherwise.
Likewise, always be thinking about alternative sources to get the information you need. Center reporters filed many FOIA requests with the Department of Labor for specific wage theft case narratives, but the agency redacted workers’ names for privacy reasons. So reporters ended up identifing victims of wage theft through other means: reaching out to labor centers in several states and contacting employment attorneys who represent workers in wage and hour lawsuits discovered in court records.
Another example of that comes from our USPS story. Fernández Campbell was unable to get help from the national postal worker unions. Union leaders did not want to give Fernández Campbell arbitration records related to wage theft or put her in touch with employees. A spokesperson for the National Association of Letter Carriers tried to persuade Fernández Campbell that wage theft at USPS was not a problem worth writing about. To reach postal employees, Fernández Campbell posted messages on six online postal employee forums and messaged USPS mail carriers on LinkedIn. USPS workers who saw her messages on the postal forums reached out to her via email to set up interviews. Two of those mail carriers, who are union members, sent Fernández Campbell copies of private arbitration records.