The “car-sharing” services Uber and Lyft continue to generate no end of controversy. A recent strike by French taxi drivers protesting Uber turned violent at times, and French authorities subsequently arrested two of Uber’s local executives for continuing to operate its low-cost UberPop service despite a national ban. Uber and Lyft have for the most part been able to strike deals with regulators to continue operating in the U.S., but both companies now face potential jury trials over whether their drivers are employees or independent contractors under California law, and the California Labor Commissioner has held that at least one Uber driver is, in fact, an employee.

In a recent essay for the University of Chicago Law Review Online, I considered the likely effects of Uber’s meteoric rise on various important social goods. Those include passengers’ and others’ safety; passengers’ and drivers’ rights against discrimination; and drivers’ working conditions more generally. While the essay focused on Uber, much of its analysis applies as well to Lyft.

Part of my motivation for writing the piece was a sense that existing debate around Uber was caught up in debates about inequality more generally, and that many were reacting to Uber based on their political priors. For example, after an Uber driver struck and killed a young girl and the company disclaimed responsibility, many argued that the company should be banned as inherently unsafe. But that argument begs a basic question in the design of tort regimes: how to minimize the joint costs of accidents and precautions.

To know whether Uber ought to be banned on safety grounds, one would want to know whether Uber drivers cause more accidents than traditional taxi companies. I have seen no data on this question, nor do I expect it exists given the company’s relative youth. One would also want to know whether insurance companies could develop new products that compensate those hurt by car-sharing service drivers. I argue in the essay that they probably could, and have now heard anecdotally auto insurance companies are requiring policyholders to confirm, when renewing, that they do not drive for a car-sharing service.

Some high-profile crimes by Uber drivers have also led debate astray. Now, the rate of passenger assaults or kidnappings by drivers would ideally be zero. But how to get there? Critics of Uber have argued for more stringent criminal background checks to ensure passenger safety. But as the EEOC and various courts have noted, such background checks tend to screen out racial minorities at a disproportionate rate. So before imposing such obligations on Uber, policymakers might want to determine how to balance passenger’s safety interests with drivers’ equality interests.

Some critics of Uber also seem to misunderstand the company’s business model. Yes, Uber is undercutting traditional taxis in part by avoiding the extensive regulations of the taxi sector, and that is unfair. But as I argue in the paper, the company’s business model relies not just on that regulatory arbitrage, but also on its dramatic reduction of search costs and exploitation of network effects among users and drivers. Because Uber drivers are linked more-or-less instantly to potential riders, it is plausible that their drivers will have riders more often than traditional cabs that rely on radio dispatch or street hails, allowing Uber to have similar gross receipts despite lower fares.

To be clear, it is not apparent which of these means of cost cutting now predominates. Nor is it clear how important Uber’s occasional driver subsidies are to its ability to expand or maintain its market share. But it is possible that Uber is creating a market with a new equilibrium, one that delivers lower prices and higher quality service for passengers alongside decent rewards for drivers.

Where lawmakers and regulators do need to step in to promote social goods, moreover, Uber’s rise will probably make their task easier based on two features of the company’s model: first, that it is consolidating the sector, and second, that it is gathering gigabytes of data on driver and passenger behavior.

Regarding the first, in many cities the taxi sector is highly fragmented: medallion owners lease medallions to management companies, who lease them to cab companies, who lease cabs to drivers. Fragmented industries with multiple intermediaries between owners and workers pose serious regulatory challenges, as I have argued elsewhere, because they give rise to enormous monitoring costs and collective action problems. Not so Uber. It is General Motors in comparison—an industrial behemoth that can unilaterally set terms for hundreds of thousands of drivers.

Regarding the second, regulators could push Uber to use its enormous data collection and management capacity to shape drivers’ behavior. For example, Uber could deactivate the accounts of drivers who consistently refuse to take fares to low-income and minority neighborhoods, or from passengers with names that appear African-American or Hispanic. When a traditional taxi does not stop for an African-American rider, in contrast, nobody but the driver and victim know. My point here echoes Columbia Law’s own William Simon, who has argued that regulators and privacy advocates ought to pay more heed to “the potential contributions of new technologies to both order and justice.”

The multi-billion dollar question going forward will be whether Uber in fact employs its drivers. I am happy to admit I got this one slightly wrong in my essay, as two recent federal courts denied Uber’s and Lyft’s motions to dismiss putative class actions, holding that juries must decide whether those drivers are employees. If that happens, drivers would be entitled to various benefits such as minimum wages, overtime pay, workers compensation, unemployment insurance, and even collective bargaining rights. That would force Uber and Lyft to fundamentally change their business models, and they will surely appeal.

Judge Vince Chhabria’s order denying Lyft’s motion to dismiss is worth a read, because it recognizes two aspects of the employment question that often elude judges and commentators. First, Judge Chhabria notes that the jury “will be handed a square peg and asked to choose between two round holes.” Lyft drivers aren’t classic employees, he reasons, because they faced basically no regulations about their appearance or their vehicles’ appearance, their work schedules, or their work locations. Yet, the Judge reasoned, Lyft drivers also are not classic independent contractors—think of a plumber or carpenter—because they have no special skills, and because their activities are a central part of Lyft’s business. The law simply doesn’t have a category for this sort of work.

Second, Judge Chhabria’s opinion recognizes that employment status is fundamentally a value judgment based on considerations of power and fairness. For example, he wrote that minimum wage laws target the “‘weakest and most helpless class’ of workers,” because such workers need state protection “against the bargaining advantage employers have over employees.”

In good legal realist fashion, I would push further and pose the question this way: are Uber and Lyft’s practices causing such significant social costs that it is appropriate to hold them to the duties traditionally associated with employment? There is no easy answer. Surely many drivers are dissatisfied with Uber and Lyft, especially insofar as the companies led them to believe they would earn more after lease costs, taxes, and depreciation. But just as surely, many drivers are working part-time, in cars they already own or lease, and are quite happy with the arrangement.

Which brings me back to more basic questions about this new business model, and about the net social costs of existing taxi regulations. Can Uber and Lyft really ensure high-quality and low-cost service while also enabling drivers to earn a decent living? If so, then they could leverage their goodwill with drivers and passengers to press for new laws clarifying drivers’ status rather than taking their chances with juries. In my mind, the right balance might involve holding Uber and Lyft to some employment duties while exempting them from others, and imposing antidiscrimination duties and insurance requirements. Properly regulated, such companies could significantly increase welfare while respecting core public values.

The preceding post comes to us from Brishen Rogers, Associate Professor of Law at Temple University Beasley School of Law.  The post is based on his recent article, which is entitled “The Social Costs of Uber” and is available here.