BLS Data on Platform Work Reflects Challenges of Measurement
First BLS measure estimates 1 percent of the workforce uses online platforms to arrange work.
The Bureau of Labor Statistics (BLS) released findings from four questions about platform-mediated work Friday morning. These questions, added to the end of the May 2017 Contingent Worker Supplement (CWS), asked about tasks or jobs that had been arranged on a platform–an app or website that connects workers with customers and facilitates payment. After reviewing responses, BLS determined that the questions “did not work as intended.” They yielded a high number of false positives, or people saying they had engaged in platform-mediated work who in fact had not. For example, some respondents had said they were a police officer or bank executive, and then later indicated that they had arranged that job through an app–an indicator the respondent likely misunderstood the question. Rather than relying on the questions, then, BLS researchers listened to entire survey responses to individually identify those who had likely used an app or website to arrange work, based on their descriptions of their work.
This manual recoding process led to an estimate that about 1 percent of the workforce uses platforms to arrange work, roughly consistent with estimates from bank data, tax records, and surveys. Given that the survey only addressed work done in the past week, rather than the past month or year, this most recent estimate may reflect a higher rate of participation than prior measures. The platform workers identified by BLS were more concentrated in the transportation sector than the overall workforce, again aligning with other studies. Platform workers are a small subset of a broader group of alternative workers, those in arrangements outside of traditional, full-time, permanent jobs, including independent contractors, temp, contract-firm, and on-call workers. The CWS found that around 1 in 10 workers rely on these types of arrangements for their main job, although other sources suggest much of this work happens as a supplemental or side job.
Aware that measuring an emerging sector of the workforce would be a challenge, BLS took several steps to develop the questions, including consulting with stakeholders, taking public comment, conducting laboratory tests of questions, and, somewhat ironically, collecting sample responses on Amazon’s Mechanical Turk platform. Despite these efforts, the majority of question responses were deemed invalid.
Perhaps the toughest challenge for researchers moving forward is developing a meaningful measure of supplemental work. Surveys of primary employment show stability in alternative arrangements, but tax records show a sharp increase in independent contracting, suggesting that many people are earning supplemental income as independent workers. The CWS online platform questions intended to capture supplemental work by asking if platform work was primary, secondary, or supplemental. However, more than 90 percent of respondents indicated that their platform work was their primary job, many of which were the false positives recoded as non-platform workers. There is no way of knowing how many of these people may have done supplemental platform work in addition to their non-platform main job. A precise measure of supplemental platform income remains illusive.
Beyond supplemental work, these results highlight additional considerations for future researchers. For example, the reliance on the Current Population Survey format limits responses to those who indicated they were employed, yet some use platforms to earn income while out of more stable work. It is also unclear how selling and leasing platforms, like Etsy or AirBnb, might be reflected in this data, or how to best measure these types of activities. As the BLS states in their review of the data, considerable further work is needed to develop a meaningful survey measure of platform work. Researchers can also look beyond surveys. Administrative data, including tax records and privately held data, like the bank accounts analyzed by the JPMorgan Chase Institute, hold much potential. Additionally, qualitative research, including interviews with and observations of workers, can illuminate the experiences and motivations of workers.
Online platforms and side gigs are examples of types of work that do not neatly fit into the categories traditionally used to measure work. What these questions highlight most is the need to rethink how we measure work arrangements, and to invest in meaningful data in order to gain a comprehensive understanding and better design responsive policy. As the economy changes, we need consistent measures over time, along with new ways of thinking about and measuring work, in order to fully understand the experiences of workers and the challenges they face and develop policies responsive to their needs.