Uber hours, and some Labor Day housecleaning

September 5th, 2016 at 11:00 am

Noam Scheiber has an interesting piece up at the NYT on choices made by Uber drivers as to how many hours they work (Scheiber’s stuff on the evolving workplace is must-read journalism, IMHO). Apparently, some drivers work more when fares are low and less when they’re high (Uber boosts fares in periods of high demand). That choice drew this commentary (my bold):

“In one camp is a group of so-called behavioral economists who have found evidence that many taxi drivers work longer hours on days when business is slow and shorter hours when business is brisk — the opposite of what economic rationality, to say nothing of common sense, would seem to dictate.”

I don’t get is why this is such a theoretical or common-sense puzzle, i.e., why isn’t this simply a case of the “income effect” dominating the “substitution effect?” Microecon 101 posits that there are two countervailing forces affecting labor supply when your (after-tax) wage goes up. One, the substitution effect, says work more as the opportunity cost of not working just went up (i.e., the cost of an hour of chilling = your hourly wage). The other, the income effect, says work less because you can now afford to chill more without lowering your income.

In fact, the article does suggest that this latter effect is in play as drivers appear to be roughly hitting their income targets sooner when fares are up (though there’s also something in there about how more experienced Uber drivers don’t drive less at peak fares).

At any rate [sic], I may be missing something and it’s a small point, but I wonder if what’s going on here is emblematic of a more serious problem I’ve noticed in the tax debate: income effects are heavily discounted, if not ignored. And this makes supply-side tax cuts look more pro-work than they probably are.

To be fair, the empirical literature supports the view that the substitution effect dominates—when they simulate labor supply responses to tax changes, the CBO uses a substitution elasticity that is multiples of that for the income effect: 0.24 versus -0.05 (see Table 4 here). Moreover, the assumption that most workers can freely set their hours as they see fit belies most workers’ reality, which is one reason why the Uber drivers and other gig economy workers are a cool group to study in this space, as they’ve inherently got more flexibility.

But we shouldn’t be at all surprised if those who can do so work less when their pay goes up. For many, it’s an unquestionably rational choice!


On other matters, a few links to pieces you may have missed and might enjoy.

My (pretty long) Labor Day piece over at WaPo. Yes, we need more unionization, a critical yet diminished balancing force. But is such a revival plausible? That’s where centralized bargaining comes in.

Also at WaPo, a plea to the Fed to do what jazz musicians do when they get lost in the tune: when in doubt, lay out!

Finally, Arthur Levitt asks me many topical questions and gives me some time to answer them!

Print Friendly, PDF & Email

11 comments in reply to "Uber hours, and some Labor Day housecleaning"

  1. Evan Grantham-Brown says:

    No, it isn’t.

    We aren’t talking about a permanent raise here. We’re talking about a situation where the worker’s wage varies by the day. On busy days, an Uber driver will have a high hourly wage. On slow days, the same driver will have a low hourly wage. Working long hours at a low wage and few hours at a high wage is strictly sub-optimal. You sacrifice time *and* money.

    To illustrate, let’s say we have two drivers who work Mondays and Tuesdays. Mondays are busy, generating $14/hour, and Tuesdays are slow, generating $8/hour.

    Driver Alice works 6 hours on Monday and 2 hours on Tuesday. She makes a total of $100 in 8 hours of work.
    Driver Betty works 2 hours on Monday and 8 hours on Tuesday. She makes a total of $92 in 10 hours of work.

    Betty is worse off in every way. She works more hours for less money.

    • Jared Bernstein says:

      That makes sense, for sure, but what if they’re optimizing day-by-day? Then they’d hit an income target early on a high wage day and later on a low-wage day–that would look “irrational” from a weekly perspective (or even a two-day perspective!) but not a daily one. I knew someone who did this (an artist/musician) so N of one and not representative…but what do you make of that?

      • Evan Grantham-Brown says:

        I guess that would make sense if a person with some other source of income is using Uber as a way to plug a hole. An unexpected expense comes up, you drive for Uber until you’ve made enough money to pay it, then quit and go back to relying on your regular job. If the expense crops up during a busy time for Uber, you can pay off the expense relatively quickly; if it’s a slow period, you might be at it a while; but you need the money ASAP, so you don’t have the luxury of waiting for the opportune moment.

        However, if you don’t have an immediate need for cash, I’m not seeing how it’s rational behavior. (People don’t always behave rationally, of course.)

    • Smith says:

      Hey, they’re only talking about new drivers, until they learn the ropes, as per this quote:

      “Over all, there was little evidence that drivers were driving less when they could make more per hour than usual. But that was not true for a large portion of new drivers.”

  2. Smith says:

    This is what I wrote April 30th 2015 on this blog:
    “This is intuitive and I don’t have data, but I don’t see it anywhere. Raising the minimum raises those previously making the minimum. After a year or two you get a raise. Guess what, it’s above the minimum. And so on and so forth. There is a ripple effect. Business knows this and that’s another reason why they fight it. It’s similar to when unions made up 30% of the workforce, and that critical mass threshold meant there wage agreements wound up indirectly covering everyone, from non-union workers in the same wage scale to the white collar workers in the same industry. Businesses needed to pay workers in a manner comparable to union wages and benefits or everyone would join a union.
    Today businesses are able to make well paid union workers like auto workers and teachers an object of scorn instead of envy, higher wages are seen as killing jobs as factories move south and overseas, teachers are seen as privileged over compensated causes of higher taxes. That narrative and the causes of such needs to change for workers to regain lost power and respect.”

    This deserves it’s own comment, so I’ll leave the restoration of labor power to a separate post. I would add that the idea of the minimum’s ripple effect and union contracts ripple effects are a) intuitive and common sense, b) not something that wasn’t observed and written about previously, especially when unions were powerful, and c) are particularly obvious for anyone who has ever worked in a union environment. Observe how NYU tried to buy off the striking grad students.
    The sad part is that unions are unable to get this story out, that everyone gains from their progress. Instead they are painted as overpaid and privileged objects of scorn. Yes, many union contracts enforce onerous work rules, job security clauses, and strict seniority pay, that leads to inefficiency, not to mention notorious corruption by union leaders. Improving how unions operate must be a goal in restoring some measure of their influence.
    The recent strike in France was partly about exactly the issue raised in the Post post, which was centralized bargaining. Business with the collusion of a Socialist government led by soon to be ex President Hollande tried to destroy their power. Union membership in France is actually very small, but they retain and exert an outsized influence that affects all the country’s workers. But neither France nor Germany rely on unions to protect worker’s basic rights. It’s legislated instead, as it is only partly in the U.S.

  3. Smith says:

    In reviewing how to revive unions and gain from their benefit, a few important, nay critical, points.
    1) Much if not most labor reform is legislated and doesn’t spring forth directly from unions. Unions do however play a critical role in passing legislation. It becomes “agree to this law or a union will form to fight for these conditions and more, causing even more headaches” It also becomes “pass this law so it affects all businesses and not just the unfortunate few (in the eyes of business) that become unionized” The biggest benefits to workers are government mandated, not the result of union contracts. Minimum wage, 40 hour work week with overtime beyond, end of child labor, guaranteed pension benefits (we call it social security), guaranteed healthcare benefits (because in the U.S. it’s an employer benefit until medicare coverage begins). In much of the rest of the developed world, especially western Europe, healthcare, guaranteed paid sick leave, maternity leave, paid vacation time, childcare and many work rules and corporate governance provisions that one only finds in union contracts are in fact legislated.
    The push for minimum wage is a great example, entirely legislated, but with very important support from organized labor.
    Look at the Fair Labor Standards Act. Business figured out how to label everyone a professional or managerial and avoid it’s provisions. End this abuse by eliminating “exempt” status. Today’s office worker is the modern factory worker. Walmart promotes workers to low level management positions to put them on salaries instead of paying them hourly. Tech workers have a special exempt status tied to low wage levels (low for tech workers that is). Quality child care should be a universal right instead of a privilege of the rich or a corporate benefit affordable only to large businesses.
    Failure from mismanagement where unions still prevail (again the example of auto and teaching) should not be blamed on unions, though they must be open to reform. See Germany for shared corporate governance.
    Finally, individual workers will have leverage even without unions, when economists decide not to control inflation by slowing the economy, causing unemployment, depressing wages, and instead realize the way to control prices is by, get this, taking measures to control prices.

  4. Kaleberg says:

    To start with, Uber is not transparent to drivers. I had an interesting ride with an MBA between jobs who is using Uber to see if it will pay for her car ownership. She said that Lyft – most drivers use more than one dispatch service – has a fixed percentage rate, but it is impossible to tell the value of providing an Uber ride with any precision. There’s a ten percent range, and a driver never knows what he or she has earned until at least the next day. This paper does not seem to address this asymmetry.

    In addition, this paper doesn’t discuss the strategies that drivers use to decide whether to take an Uber gig or wait for a gig from another service. According to a recent Seattle Times article and discussions with many drivers, Lyft rides tend to pay better overall than Uber rides and Lyft encourages tipping as well. It is possible that during busy times drivers would quite rationally wait for a Lyft ride, so they are less likely to respond to Uber.

    This paper also ignores the fact that there are two types of Uber drivers, career drivers and amateurs filling gaps in their budget. A lot of Uber drivers not only drive for Lyft but also drive taxis and black cars when it makes sense. Taxis, of course, require signing up and buying a shift, but that’s a whole different story. The amateur drivers drive to pay auto expenses, cover desired purchases, make up for shortfalls or as a way of maintaining one’s life style during a job search. The amateurs usually have a desired income, so it is likely that this paper is finding career drivers optimizing their overall incomes while amateurs are getting the short run cash they need.

  5. Michael says:

    Wouldn’t there be other factors involved in those hours besides simply the monetary factor? Wouldn’t factors like traffic congestion (peak hours might correspond to peak traffic flow, making it harder to actually transport more passengers at the higher rate, somewhat negating the economic advantage) influence when drivers might actually want to drive?

    Imagine working in a plant. If you work the weekdays, when its cost-efficient for the company to run the A/C, you might get paid $7 an hour. However, weekends or after normal hours might make the company want to cut back on costs, perhaps by allowing the temperature to be raised 10 degrees. Simply being paid $12 an hour wouldn’t necessarily compensate for being in a totally different environment (especially if you were sensitive to such a 10 degree difference) and might fail to be accounted for in such a study. In addition, the discomfort factor added to the extra wage might actually cause greater stress and impact the driver’s overall ability to engage in such behavior over the long term. A driver might see the simple cost-benefit analysis and initially spend the bulk of their driving during peak hours, hoping to make more money, but if it brings added discomfort, their initial enthusiasm could be eroded to the point that the economic factor loses priority. Being rational creatures means we often take all factors into account when deciding what our maximum efficiency is, not just the economic factors. The only irrationality is not accounting for said factors to begin with.

    • Scott F says:

      I would point out that at non-peak times, the Uber drivers may not be “working” at all during much of their shift. They could be parked and watching to the ball game on their phone. During peak times, it is rush, rush, rush!

      That difference in effort may make any monetary analysis uninsightful.

  6. Person says:

    I don’t think you’ve really captured the real cause and effect here. This is what I believe is happening:

    Uber changes rates according to demand for services, so when there’s a shortage of drivers the wage goes up to attract them. The cause is the preferred hours of the workers, the effect is the shortages and the higher wages.

    All I see here is that the drivers are failing to respond to the higher wages. This is good in my view, and it is a sign that people are getting their priorities straight. Spend time with family, work when it is convenient, and if you can pay the bills money means very little.

    You’re reversing cause and effect. It would be really strange if drivers worked more when wages were higher. It would be a sign that Uber management needs to be replaced.

    • Smith says:

      Yes, this makes sense, especially because the very definition of an Uber driver is a measure of independence, hence it’s a gig not a job, choose when you want to work. I know someone who gave up a well paying regular 9 to 5 job to Uber, partly for that reason.
      Also, anyone who has ever scheduled employees or worked at a job with variable schedules, especially younger workers, knows the routine, especially in service positions like wait staff. Do you give up your Friday night or Saturday night out to make more money? For young Uber drivers in urban areas, the best times for wages are often also the best times for no wages. Thus the higher rates may be offered not just because of higher demand, but also lower supply.

      This still doesn’t negate point the study only found sub optimal behavior in new drivers.

      Also, how is all the analysis in all the comments, much of it common sense, missed by the study and the news article?