The Hacking of Employment Law.

• Author: Alexander, Charlotte S.

1. INTRODUCTION

   The workplace is increasingly mediated, governed, and constrained by software. (1) Companies use software in hiring and making work assignments, setting workers' schedules and tracking their hours, assessing performance, and managing payroll. Software allows firms to automate and routinize management tasks, scale those operations across an entire workforce, and track, monitor, and mine data to make predictions and informed decisions. In doing so, software holds the promise of increasing efficiency, decreasing waste, and even guarding against biases inherent in human judgment. (2)

   However, software's gradual accession in the workplace is not all positive. Using software, firms can implement systems that are largely consistent with existing laws but avoid or evade rules on the margin. For example, timekeeping programs can be set to "round" hourly employees' timecards to the nearest quarter-hour. (3) Though any given employee might lose only a few minutes per day, software's ability to automate this function and apply it across an entire workforce can result in significant losses--and wage and hour law violations--in the aggregate. (4) Software can also create information asymmetries. Employees who lose pay as a result of rounding may not know about the software functionality or their lost wages, unless they keep meticulous parallel time records of their own (an unlikely assumption). Moreover, even when an employee discovers the employer's practices, outdated or ill-fitting regulations can provide employers with some measure of cover. Software's newness allows employers to convince courts that they have engaged in something other than old-fashioned noncompliance. As a result, employment law erodes.

   In other situations, firms use software lawfully to avoid, rather than violate, legal rules. Here, firms structure their relationship with workers to avoid triggering some or all of the costs of complying with employment law. Such tactics can cause harm beyond the loss of the specific workers' rights or protections being avoided. In particular, software-enabled avoidance can create new norms that cause harms not contemplated by employment law as written.

   Scheduling software is a case in point. (5) Employers use scheduling algorithms to identify the employees whose hours are nearing some cost-triggering threshold--forty hours per week and the overtime entitlement, for instance. Those workers are sent home and replaced by others at the last minute. Affected workers lose overtime pay, but neither overtime pay nor any specific number of work hours is guaranteed by the law. This avoidance strategy creates a new set of harms to the workers who are called in as replacements. Though these workers may appreciate extra hours, last-minute call-ins create significant instability, requiring workers to find dependent care on a moment's notice, disrupting family routines, and interfering with workers' ability to attend school or hold a second job. In addition, employers sometimes demand that employees remain available for unexpected call-ins. If they refuse, employers deny them access to future shifts. Employers' use of software to reduce overtime costs thus creates a new norm in which workers' schedules are unstable and unpredictable, but the harm that results is beyond the reach of employment law.

   Finally, firms can use software in ways that mix noncompliance and avoidance. Here again, outdated or ill-fitting regulations allow firms to test the line between breaking the law and merely side-stepping it. The use of work distribution platforms by "gig" firms, such as ride-sharing company Uber, offers an example. (6) Uber uses its app to connect drivers with customers. The company classifies these drivers as independent contractors and disclaims any employer-employee relationship. (7) Indeed, the rules that distinguish employees from independent contractors assume the existence of some human who controls aspects of the work relationship. (8) Once the human intermediary is replaced by an app, the analysis becomes more difficult. In this way, firms like Uber use software to engage in a form of regulatory arbitrage, (9) claiming independent contractor status for their workers and opting out of employment law coverage. Consequently, more workers operate in spaces beyond the law's reach, and employment rights are left only to the privileged few who are able to claim employee status.

   In using software to avoid and violate employment laws, employers accomplish a regulatory hack. We use the term "hack" in two senses. The term's technical definition is to "gain access to a computer illegally." (10) It can also mean something closer to a "workaround," i.e., "a clever solution to a tricky problem." (11) The latter usage connotes ingenuity (in the eyes of the hacker, at least) but not necessarily law-breaking; the former explicitly contemplates violating legal rules. We use the term "regulatory hack" (12) to refer to unlawful noncompliance, lawful avoidance strategies, and conduct that falls somewhere in between. The hacked target is not a computer but the system of laws that regulate the work relationship. Software makes it possible.

   We explore regulatory hacks through four case studies: timekeeping programs, screening and selection algorithms, scheduling software, and work distribution platforms. In each instance, we identify the harms that software can cause and examine whether the software enables noncompliance, lawful avoidance, or some combination. We then consider the implications of these hacks for employment law as a whole and offer some possible regulatory responses.

   This Article proceeds as follows: Part II defines the term "regulatory hack" in more detail and tracks the development of the concept and terminology in the work of earlier scholars and writers. Parts III through VI present the four case studies, and Part VII turns to possible regulatory responses.

2. "REGULATORY HACK" DEFINED

   This Part traces the origins of the "regulatory hack" concept and term in cyberlaw, copyright, and employment law scholarship.

   1. Roots in Cyberlaw and Copyright Scholarship

      Software-enabled noncompliance and avoidance have been a central focus of cyberlaw and copyright scholarship for many years. Drawing upon copyright controversies over peer-to-peer file sharing and the rapid expansion of the internet in the late 1990s and early 2000s, scholars focused considerable energy on the relationship between software and law. (13) From this well-developed literature, we find Tim Wu's theoretical framework most instructive. In a 2003 article, When Code Isn't Law, Wu describes three ways in which software can erode, challenge, or alter legal rules: evasion, "avoision," and change. (14)

      Evasion, according to Wu, consists of strategies to "decreas[e] the odds of being punished for violating a law," such as a bank robber's wearing a mask to disguise his or her identity. (15) Wu views software as enabling both the underlying legal violation and measures to evade detection. (16) "[L]aw-busting code," in the file sharing context, allowed millions of users to trade digital music files freely, without regard for copyright limitations. (17) By creating the conditions for noncompliance on such a large scale, software allowed everyday people to violate copyright law, giving "cover" to end users whose very number would seem to lower the odds of detection and punishment. (18)

      The term "avoision," which Wu borrows from Leo Katz and Ronald Turner, refers to "the avoidance of laws in ways that evade the law's intent or purpose but do not actually constitute unlawful behavior." (19) Avoidance can also be characterized as regulatory arbitrage, a term that originally referred to multinational companies' forum shopping in search of countries with the most favorable tax rates and regulatory environment. (20) Regulatory arbitrage has been used to describe many forms of legal avoidance and gamesmanship--both domestic and cross-national--such as workarounds "to avoid taxes, accounting rules, securities disclosure, and other regulatory costs." (21) Avoidance tactics almost always involve behavior that is technically legal, but that "exploit[s] the differences between a law's goals and its self defined limits." (22) Thus, as Wu points out, peer-to-peer file sharing software was designed in a decentralized way to avoid vicarious liability for software companies as intermediaries under the copyright regime. (23)

      Change, finally, comprises direct efforts by regulated entities to alter legal rules to their benefit. (24) This could consist of lobbying, "a tool that delivers... legal change for a price." (25) Change can also come about via litigation, by which "interest groups determine or radically influence the regulatory agendas of agencies." (26)

      In Wu's view, the three options are interrelated and can act as substitutes for one another. Regulated entities might choose to invest time and resources in evasion or avoision, rather than direct change efforts. In other words, when the law generates compliance costs, regulated entities may try to change the rules through the front door (presumably while continuing to comply), use software to violate the law through the back door, or opt out of the law's requirements entirely.

      Further, in what Wu calls a "reaction to the reaction," (27) regulatory change might occur as a result of widespread evasion or avoision, as the law loses its grip on the behavior of regulated entities. Such change could cut in either direction: regulation could be loosened, consistent with the interests of the regulated, or it could be revised and tightened, in the interests of the regulators, to "restore the balance" that existed pre-evasion or avoision. (28)

   2. Renewed Attention with the Rise of the Gig Economy

      The problem of legal avoidance has long been a concern for employment law scholars. (29) Miriam Cherry, for example, has written...