Tuesday, October 1, 2019

California wants more monopsony in the labor market.

The case of AB5 in California is an interesting clarifying case regarding the motivations and goals of labor regulation.  AB5 redefines the distinction between contractors and employees and is mostly an attempt to force Uber and Lyft to treat their drivers as employees rather than contractors.  This will entitle them to benefits, workplace protections, and the minimum wage.  My experience with these firms is that the contractor status is a key component of the benefits of their model, and that in most markets, changing to an employee-based model will make it difficult for them to continue.

Normally, one argument in favor of higher minimum wages is that firms have monopsony power over unskilled laborers, so they hire fewer workers and pay them less than if the market were more purely competitive.  Thus, raising the minimum wage does not lead to much unemployment.  Firms can afford to pay more.  The minimum wage just transfers some of the monopsonist gains back to the workers.

But, the interesting thing about this particular market is that you would be hard-pressed to find a market that was a closer approximation of pure competition.  On the customer side, Uber & Lyft are basically commodities.  Riders can check on both services, and will generally go with the one that has the shortest wait at the lowest price.  Many drivers drive for both, so there is little the firms can do to differentiate their service.

On the driver side, the firms must pay enough to entice drivers to be available.  In fact, Uber and Lyft pay more than the market clearing price for drivers that have riders in their cars because in order to win more passengers, they need to pay enough to induce drivers to be available, which in this industry, inevitably means idle time.

In fact, Uber & Lyft have very little control over what their drivers earn.  Since this is a competitive industry with free entry and exit, and since the firms must accept as many drivers as they can, within reason, so that they can offer customers a shorter wait time than the other firm does, drivers determine their earnings by entering or exiting the market.  If Uber & Lyft pay more than is necessary, more drivers will enter the market, and they will spend more idle time without riders in their cars.  This will happen even within the existing pool of drivers. If Uber decides to raise the payment they make to drivers in a market, that will induce more drivers to Uber and away from Lyft.  If you ask drivers what happens in markets where one of the firms changes their pay rates, you will find that the total weekly earnings don't change much.  If Uber raised their pay rates, then a driver who drives for both will find that they get more rides from Lyft because drivers will have substituted between the two firms until the net total pay (idle time plus paid time) roughly evens out.

This is a classic case of queuing.  And you can see the queue adjusting in real time to changes on the ground.  In fact, that is the beauty of the contractor model.  There are hundreds or thousands of drivers in a city, and drivers are constantly adjusting between Lyft and Uber, between times of day or location.  Each driver is in a constant chess match to find the most lucrative way of driving that matches their needs and constraints, and the key variable at the center of those tactics is minimizing idle time.  Each driver is increasing or decreasing their willingness to queue depending on the opportunities available to them as drivers or outside the rideshare industry.

Compare this to the minimum wage debate.  Effectively what minimum wage opponents argue is that those markets are generally competitive, so that a high minimum wage will increase unemployment.  Unemployment is a queue.  The minimum wage is set above the market clearing rate, so workers queue to supply the limited demand for employment.

In the minimum wage debate, monopsony is treated as a preexisting condition which the minimum wage is meant to cure.  Here, there clearly is no monopsony.  In fact, these firms are so lacking in market power that even the proponents of AB5 sometimes express doubt that their business model is sustainable. In reality, AB5 is meant to create monopsony.  But, queuing is already a natural part of this model.  So, what AB5 would do is make Uber & Lyft gatekeepers reducing the quantity of labor supplied in the market.  Since drivers would be employees, and the firms would be responsible for their total earnings from both idle and active time, the firms would have an incentive to minimize idle time.  They would have an incentive to limit the number of drivers.

This would not necessarily change the total amount of queuing time.  It would simply segregate it so that the riders who are now chosen by the gatekeepers to be employed would have less idle time, and the riders who are not chosen would be in the queue known as unemployment.

I think this would be tragic.  The beauty of the contractor model is that workers who have been turned away by the gatekeepers in other industries that have employee models can enter this business without dealing with gatekeepers.

One aspect of this industry that would be interesting to study is that there is a great amount of variation in driver earnings.  Even this MIT study which found low earnings levels on average (which I think have been revised up) shows a tremendous range in driver earnings.

What's interesting is that this is a completely open marketplace.  There is little that drivers can do to keep other drivers from horning in on their driving strategy.  There are few barriers to entry.  (Even the car isn't much of a barrier.  There are companies that partner with Uber and Lyft that will rent you a car for less than $5/day.)

What you find if you ask drivers about their work is that there is a tremendous amount of variety among drivers regarding what they need from their work and what strategies they use to get what they need.  In the minimum wage debate, opponents often point out that employers will make non-wage adjustments to counter regulated wage gains - less flexibility, fewer benefits, etc.  What we can see here is that the drivers themselves, in the unregulated rideshare market are actively engaged in some massive rebalancing between pecuniary and non-pecuniary benefits.  The variance in earnings might be partly explained by skill, or location.  I'm sure in Phoenix it's easier for a driver that lives in old-town Scottsdale to roll out of bed and turn the apps on and get rides immediately than it is for one on the far west side who might need to drive downtown to get to a busy area.  But, surely those factors can't explain that much variation.  Drivers are making choices about when they want to work, what types of riders they want to pick up, etc.  The 2am bar scene is a sure-fire earnings winner, but many drivers happily sit it out.

So, from a public policy point of view, those who would regulate this market aren't trying to fix a market failure.  There is no market failure.  AB5 creates monopsony power by imposing a wage floor and a regulatory framework in this market, with the hope that the economic rents will be claimed by the drivers.

This is telling.  I think it's a bit of a misunderstanding to think that Progressive, egalitarian political policies are intended to make up for economic rents claimed in imperfect markets.  Egalitarian policies require economic rents.  You can't divvy up the spoils in your preferred way if you don't have spoils.

In this particular case, engineering corporate power and then trying to transfer the gains to the workers will be a huge loss.  First, I just don't think the business model can work that way.  There are countless ways that drivers now manage their queuing in a way that is productive which simply couldn't be managed centrally, including being simultaneously available for both Uber & Lyft. But, furthermore, this is basically a classic labor market.  This is not much different than, say commission sales work.  In the same way, sales jobs frequently have highly variable earnings distribution that comes from hard-to-quantify skills.  Many workers try out sales, fail miserably, and then quit.  So, there are some real winners, but also high turnover, and many workers that just don't do sales well and don't make much money doing it.  This market isn't much different than that.  If there are some drivers who are only making $5/hour, then they shouldn't drive.  Or, maybe they are retired and they just like to have an excuse to get out of the house and meet people.  Creating a market that drives this vast sea of diversity out and turns it into a cookie cutter job where you go where you're told, everyone makes a similar, lowish wage, with much less flexibility for the drivers will mean that a lot of drivers will lose things they value.  And, many of the drivers that are making $20/hour or more will either make a lot less or will be driven out of the market altogether because being contractors is a key element to their driving strategy.

And, this will likely fail at its own goals.  The loss of productivity and the loss of a potential chance to earn income without gatekeepers making the hire/no hire decision will leave a lot of drivers out.  In the current competitive rideshare market, it is other opportunities that determine what drivers earn.  If similar work can get you $12/hour in other jobs, and a driver in that city can earn $13, then that worker, on the margin, will drive, adding to the queue time for all drivers as more drivers must divvy up the same number of rides, until similar drivers are only making $12 after factoring in idle time.  Regulatory impositions like this do nothing to improve those other opportunities.

The rhetoric on this issue tends to be anti-corporate, as if this regulation will force the firms to treat their workers better.  But, the firms are powerless to significantly increase the pay to their drivers.  The regulation requires a playing field that engineers more corporate power.  The idea is to use that corporate power to lessen wage inequality.  It will only lessen wage inequality within the rideshare industry, and it will do so at the expense of some of the better paid drivers and at the expense of potential drivers who will now not get hired.  And it will lower the value added from the rideshare industry.

AB5 is crony capitalism.  It has to be.  It can't do what it purports to do without creating a framework that gives the firms power to limit access to the market.  As I mentioned in the previous post, this might be a generalized point.  Maybe more powerful firms are correlated with less variance in wages.  The egalitarian project requires powerful firms so they can be directed by the state to distribute the gains from that power.  But, trying to engineer that outcome with policies like AB5 is fraught with potential downsides.  I haven't seen evidence that AB5 proponents have attempted to fully understand those downsides.  It would probably be impossible to fully understand the potential downsides.  In the end, driver incomes are determined by the available alternatives.  This applies generally to all workers, really.  It is unlikely that the fates of workers in general will improve by imposing regulations meant to take available alternatives away.

The fact that the rideshare industry is such a decent approximation of textbook competitive markets makes it a great example for understanding which complaints about our present economy are complaints about information being conveyed by functional markets about the state of the world and which complaints are about market failures.  To my eye, there is a lot of confusion on this distinction.


  1. I agree with this highly intelligent post.

    My only caveat would be that any developed nation should strive for "tight" labor markets. If labor "scarcity" is the norm, then Uber drivers will do well enough.

    Of course, this also assumes property zoning is abolished and housing costs go down instead of up. In California, Uber drivers are earning market wages but paying bloated, non-market housing costs.

    1. Exactly. If labor scarcity is the norm, then Uber drivers will do well enough in an unregulated contractors model.

    2. That's basically how Singapore works. We don't have a minimum wage here and not much restrictions on firing employees, but we do have a generally strong labour market. (Housing is mostly run by the government, but at least they are building enough houses again, so rents have gone down in the last few years.)

      One problem with a strongly regulated labour market is experienced growing up in Germany is that the economic rents labourers on the inside are supposed to enjoy, are often eaten up by obnoxious bosses. Add to that the misery of the queue of unemployed, and it's not nice.

      France recently had a strong demonstration: France Télécom rebranded as Orange and wants to restructure. But they aren't allowed to just fire people. So managers systematically made employees' lives miserable. Many employees didn't have better options than being bullied at Orange and stuck it out. Some of them, and that made the news, chose suicide over quitting.

      If France had stronger labour markets, workers wouldn't need to put up with the bullshit. But also, employees wouldn't need to torture employees in the first place, they could just fire them.

  2. https://blogs.imf.org/2019/10/02/to-tackle-housing-affordability-in-canada-build-more-houses/#more-27254

    Well, in some ways "duh," but when someones agrees, embrace....

  3. > But, the firms are powerless to significantly increase the pay to their drivers.

    There's basically only two things the firms could do: decrease their own takings (so that drivers keep more), and improve efficiencies with eg introducing things like uber pool, or giving drivers more information for forecasting demand etc.

    But yes, you brilliantly line out that a single small company decreasing their own takings wouldn't do very much. It's only via the effect on the general equilibrium that wages increase.

    I remember a similar post on Econlib (or so) analyzing the impact of sweatshops in poor countries that came to similar conclusions. Working for a sweatshop itself isn't better than other local shops---it's just a different point on the trade-off curve between income, stability, perks etc---but having lots of sweatshops around still tends to make an economy take off.

    1. Yes. You're right. The great thing about this is that it is such a transparent market that anyone can see these classic market functions by driving a little or by talking to the smart, tactical drivers when we're using the service.

    2. Although, one problem is that with labor, there is such a corrosive tendency to move from an analyst mindset to an advocate mindset. Even smart drivers will frequently complain about the rates Uber and Lyft are paying. That's perfectly fine. We'd all like to make more money. But, that's sort of like how farmers are always complaining about the price of wheat or something. It's not a sign of market failure or exploitation. That's just reality. So, even talking to drivers, you have to process their complaints in a way that doesn't just devolve to "They are unhappy that rates aren't higher, so they must be exploited."

  4. Another interesting corollary: just as Uber can't increase effective hourly wages, even if they wanted to; they can not decrease effective hourly wages either.

    (I take it as a strength of your post that you leave that conclusion for the reader to make.)

    1. In fact, I think a misunderstanding here is at the heart of some driver complaints. Generally, Uber & Lyft enter a market with highly subsidized driver rates which attract drivers to the market when demand is still developing. So, a developing market will have very high $/mile driver rates with a lot of idle time. Over time, the firms keep decreasing the $/mile rate as the market develops, and they also can increase the % of revenues they retain. With each decrease in driver pay rates, the existing drivers will complain about how U&L just keep screwing the drivers, lowering pay and keeping more of the revenue. But, it looks to me like driver earnings tend to remain about the same over time.

      Now, it is still true that some of those drivers will see declining revenues. The change in market structure over time will benefit some drivers and hurt others. So, if you had some tactic for avoiding idle time or being productive during idle time under the earlier market context, you will probably not make as much as idle time declines and driver rates decline. But, that doesn't necessarily generalize to all drivers. Some drivers will make more, and some drivers will be attracted into the market that wouldn't have been able to profitably drive before.

      To me this is typical of discourse on labor vs. capital issues. The labor advocates (and many driver groups) obsess over how much of the revenue the firms keep vs. the drivers. But, over time, there is little correlation between the % taken by the firms and driver earnings, and I would bet that when the market development teams at U&L are looking at how to develop a market to be profitable, the % of revenue that drivers keep is a much less important factor to them than other issues, like developing enough demand to increase the potential density of drivers in an area.

  5. This is telling. I think it's a bit of a misunderstanding to think that Progressive, egalitarian political policies are intended to make up for economic rents claimed in imperfect markets. Egalitarian policies require economic rents. You can't divvy up the spoils in your preferred way if you don't have spoils. Read more at hoc nghe spa

  6. Yes i am totally agreed with this article and i just want to say that this article is very nice and vey informative. More blogs please. 경마사이트

  7. That is a very good tip especially to those new to the blogosphere.
    Short but very accurate info… Appreciate your sharing this one. A must read post.

    My web site - 바카라사이트