A Deep Dive into our Study on California's AB5
We compare the employment outcomes of occupations that were affected by AB5 in California to those same occupations in other states before and after AB5 was enacted
Last week, we published our California AB5 study, which provided the first empirical investigation of reclassification policies in the United States by analyzing the effects of California Assembly Bill 5 (AB5), the country’s strictest law for working as an independent contractor. The intention of AB5 was to encourage organizations and businesses, large or small, to hire workers as traditional (W-2) employees and to reduce worker misclassification.
A summary of our results is here. In this post, I will unpack how we did the study, what our results mean, and address some comments from a Bloomberg Law article which highlighted our study.
As a way of background, in September 2019 AB5 codified into law the 2018 California Supreme Court decision in the Dynamex case, which imposed the “ABC” test to determine a worker’s legal classification status. AB5 expanded the application of the ABC test to all purposes under the Labor Code, Unemployment Insurance Code, and IWC Wage Orders, and to nearly all occupations and industries. AB5, along with AB2257, gave over 100 occupations an exemption from the law. And, California voters also passed Proposition 22—a ballot measure that additionally exempted platform-based transportation and delivery workers from AB5. Here is the timeline of the changes:
How did we do the study?
Our goal was to empirically examine the effects of California’s AB5 for those workers who were not exempt from the law.
We used the Census Bureau’s Current Population Survey (CPS) from January 2011 to September 2023. We measure overall employment, labor force participation, self-employment, and traditional employment for each state, occupation, and month.
We matched the legal exemptions in AB5 and AB2257 to the occupations in CPS and we excluded from our study the occupations that were legally exempt from AB5.
Our analysis uses a difference-in-difference approach to compare the employment outcomes of occupations that were affected by AB5 in California to those same occupations in other states before and after AB5 was enacted.
So for example, let’s take farmers who were not exempt from California’s AB5. Our analysis follows the difference in monthly employment outcomes among farmers within California vs farmers in our control states, and it then measures the change in this difference once AB5 is enacted. The difference-in-difference approach thereby allows us to identify the relative changes in labor market outcomes among non-exempt occupations within California compared to such outcomes in our control states.
What do our results mean?
Contrary to the goals of the lawmakers, we find no consistent evidence that traditional (W-2) employment increased for non-exempt occupations in California post-AB5.
We find that AB5 is associated with a significant decline in self-employment and overall employment for non-exempt occupations in California. Specifically:
Self-employment fell by 10.5 percent on average for non-exempt occupations, while overall employment fell by 4.4 percent on average for non-exempt occupations
Occupations with a greater prevalence of self-employed workers saw greater reductions in both self-employment and overall employment
IN SHORT: On average, 1 in 10 self-employed individuals may have lost self-employment opportunities in California among occupations not exempt from AB5, while there is no evidence of an accompanying increase in traditional employment opportunities among workers in non-exempt occupations.
Only in one instance do we see a statistically significant increase in traditional employment. For occupations in the top quartile of self-employment as a share of total employment, we observe a statistically significant increase in traditional employment. This finding could be consistent with the idea that stricter legal classification would reduce worker misclassification, and it is likely that worker misclassification could be more prevalent among occupations with high rates of self-employment. Meaning that if AB5 was going to have a positive impact on traditional employment, we'd expect it precisely among occupations with relatively numerous independent contractors.
However, even in this case, the magnitude of the increase in traditional employment (10.2 percent) was smaller than the significant decline in self-employment (17.7 percent).
Addressing the comments in a Bloomberg Law article
Recent reporting from a Bloomberg Law article featuring our study stated that “California’s strict worker classification law—known as AB5—resulted in a 4% decline in employment in the state” (emphasis added).
Our data and empirical strategy do not permit us to estimate the effect of AB5 on employment in the entire state of California. Our findings refer only to employment outcomes among occupations that were not exempt from AB5.
The article also cites economist Michael Reich who states the study doesn’t take into account whether economic trends in other states were similar to California’s. Our study, in fact, discusses the limitations of the data and our empirical strategy, including the observed inconsistencies in pretreatment parallel trends. As we state in the paper, our empirical strategy cannot yield estimates that can be causally interpreted due to the noisy nature of the data available, and thus our findings primarily indicate robust statistical associations (as robust as the nature of the data and the nature of studying policies such as AB5 can permit given the timeline, implementation, etc.).
Reich also comments that, in contrast with our results, other evidence shows an increase in self-employment over the last few years. However, our findings are consistent with a general increase in self-employment at the national level or within California as a whole. Instead, our findings suggest a relative decrease in self-employment only among non-exempt occupations when compared to other states.
The article also cites Heidi Sherholz, President of the Economic Policy Institute stating that “[t]he high number of job losses for full employees indicated by the study also calls into question the validity of the findings” and that “[t]here’s just really no story where traditional employment would fall.” Our main results, however, are not that traditional employment declines, but that we fail to find consistent evidence that AB5 significantly increased traditional employment, as intended by lawmakers. Additionally, as we discuss in the paper, self-employment and traditional employment could both decline as a result of AB5 if some small businesses who had workers on payroll and also relied heavily on independent contractors were forced to shut down. These stories are indeed highlighted in the anecdotal evidence that has emerged in the wake of AB5.
Conclusion
Our full study is here: “Assessing the Impact of Worker Reclassification: Employment Outcomes Post–California AB5” Authors: Liya Palagshvili, Paola Suarez, Christopher Kaiser, and Vitor Melo
Our study on AB5 compares the employment outcomes of occupations that were not exempt from AB5 in California to those same occupations outside of California before and after AB5 was enacted.
We find that AB5 did not simply alter the composition of the workforce as intended by policymakers. Instead, AB5 was associated with a significant decline in self-employment and overall employment for non-exempt occupations in California compared to our control states, and we fail to find consistent evidence that traditional employment significantly increased as a result of AB5
What do our findings imply for the new Department of Labor rule on independent contracting? Check out my post on “New Study on California AB5 and Implications for the Department of Labor Independent Contractor Rule”