Gig Economy Is Piecework. But This Isn't Dickens.

Amazon Flex, the latest Uber-esque tool, is more of an opportunity for workers than a threat.

This is a workhouse. Uber is not one.

Source: Hulton Archive/Getty Images

Earlier I wrote about the way tech giants strive to ensure that they will own a piece of major emerging technology markets. Exhibit A: Amazon Flex, a newly announced service which will pay people $18-25 an hour to deliver packages in their own vehicles. If there’s going to be a sharing economy, then Amazon wants its share.

All this has Kevin Drum concerned: “It now seems as though the 'sharing economy' is any job that's somehow related to a scheduling app and provides workers only with odd bits and pieces of work at the employer's whim. In other words, sort of like manual laborers in the Victorian era, but with smartphones and better pay.” At Interfluidity, Steve Randy Waldman muses that we may need to force these contractors into employee status as a sort of antitrust action, to make sure that none of these app platforms gets too powerful.

I confess, I’m not clear on what the problem is. Manual labor in the Victorian era was not primarily awful because it involved short-term contracts; it was awful because the jobs were grim, the pay was low, and injured workers frequently ended up destitute. Getting paid $25 an hour for doing something much more pleasant than scrubbing floors with caustic chemicals does not tug at my heartstrings in the same way.

I also find it hard to worry that Uber, or Amazon Flex, is going to develop dangerous dominance in the market for people driving stuff around in their cars. “People driving stuff around in their cars (including passengers)” is a market with very low barriers to entry, which is why taxi firms have invested so much in lobbying for laws, and powerful regulators, to protect them from competition.  Network effects can, of course, create barriers to entry -- but in a business where the only requirement is a personal vehicle and a skill that almost every American adult possesses, those effects are unlikely to be enough to allow any company to abuse either customers or workers for very long. 

Network effects are most powerful where switching costs are very high. If I’ve paid $300 for a VCR, and a bunch more for VHS tapes, I’m unlikely to be interested in converting to your Betamax format. If I’ve paid several thousand for a PC and assorted software, and stored a lot of work in formats specialized to those programs, it’s going to be tough to convince me to go over to an entirely new system. By contrast, what’s the switching cost for an Uber user? I might sprain my thumb trying to open a new app? Even the switching costs for drivers are relatively minimal. Which means that no firm in this business is going to be able to enjoy substantial monopoly rents, because if they do, some bright entrepreneur with a modicum of capital at their disposal is going to dive in and compete them away.

The real concern, I think, is that these jobs will become substitutes for better jobs: more stable, better paid. This is obviously going to concern left-wing commentators, many of whom have already expressed worry that the “gig economy” is bad for American workers.

That may be. But that rests on the hidden, and so far unproven, assumption that the gig economy is in fact displacing workers by driving down the value of the work they do, rather than creating new economic activity that simply wouldn’t exist if it weren’t for these apps -- and possibly in the process providing work for workers displaced from other industries, for reasons that have nothing to do with Uber or Instacart.

Kevin is, of course, right that these apps somewhat mimic the casual labor market of the 19th century. But there’s a difference beyond the labor conditions: The apps radically drive down the transaction costs to providing these sorts of services. A Victorian laborer might have had to tromp from house to house looking for work, or stand outside for hours waiting for someone to come by looking for workers. No longer.

If that sounds merely nice, think again. It’s enormous. Transaction costs shape markets in fundamental ways. They are the reason that firms exist, argued Nobel Laureate Ronald Coase; when it is too costly to keep going to the market to buy goods and services, firms bring those functions in-house to allow more efficient coordination.

Lowering transaction costs can even generate additional supply and demand. A transaction cost is a part of the price of a good or service, and often a quite substantial part. And when you lower the price, you make it possible for people to consume more, while suppliers supply more.

Imagine that I am sitting at home one night, and I would love to go meet some friends for a drink. Unfortunately, it’s hard to get cabs to come to my neighborhood, it’s not quite safe to walk alone at night, and driving to the outing would defeat the purpose, since I couldn’t drink once I got there. So instead, I sit at home and watch reruns.

Lowering the transaction costs makes it easier for me to find someone who is willing to drive me over to the bar. This hasn’t displaced someone else’s job; it’s an entirely new piece of economic activity that simply wouldn’t have happened if the transaction costs had remained high. We get additional employment, additional consumption and additional happy hours with our friends, drinking margaritas and arguing about "Game of Thrones."

Of course, that’s not to say that no one is ever displaced -- owners of taxi medallions have, for example, seen a dramatic decline in the business of “owning the right to drive in New York City.” But proving that someone got hurt is not the same as proving a net decline in the position of workers.

So what happens if we force firms to treat these contingent workers as employees? We add back transaction costs, of course. Benefits, legal compliance, various payroll taxes. “And good!” the left might say. “They ought to be paying those things.”

But doing those things internally, rather than buying services in the market, is expensive. It makes the employer less flexible, and raises its fixed costs. It has to recover those costs somehow, which can happen in one of three ways:

  1. Charge customers more, which will reduce demand and result in fewer hours of labor being consumed
  2. Lower the wage paid to the workers
  3. Exert a lot more control over worker schedules and conditions to maximize the ratio of work to overhead. Workers will probably not be allowed to show up whenever they feel like and work as long as they want. They will, for example, probably have their hours capped to prevent them from qualifying for health insurance.

Of course, some workers would rather be in that situation. But what about people who use these jobs to “moonlight” around other obligations, like day jobs, school or child care? Worker supply, as well as consumer demand, seems likely to fall. And we haven’t even discussed what happens to consumers who benefit from these services; we’re just talking about what will happen to the workers.

I think there are real concerns about what has happened to the wages of American workers. But so far, I don’t see much evidence that Uber or Airbnb is the cause. Most of it seems to come from automation of low- to medium-skilled work in the manufacturing and clerical sectors, or the outsourcing of those jobs abroad. Trying to fix those problems by destroying Uber’s business model is like trying to cure your headache by taking a hammer to the bottle of aspirin.

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.

    To contact the author of this story:
    Megan McArdle at

    To contact the editor responsible for this story:
    Philip Gray at

    Before it's here, it's on the Bloomberg Terminal.