Working By The Task

AThe arrival of platforms that sell labour in single tasks has been described, with equal confidence, as the future of employment and as a marginal curiosity. Both descriptions draw on the same surveys, which is possible because those surveys ask whether a respondent has earned anything through such a platform in the previous twelve months. A person who drove for two weekends in March and a person who delivers forty hours every week are counted identically. When the Northgate Institute for Labour Studies re-analysed a national survey in 2021 and separated the two, it found that 14 per cent of adults had used a platform to earn money at some point in the year, but that 2.3 per cent had earned more than half their income that way. Neither figure is wrong. The gap between the two is not a detail. They answer different questions, and the larger one, being the more striking, is the one that circulates. The habit of quoting it has consequences beyond arithmetic, because a rule written for a workforce of that size will be resisted with a force proportionate to the number it appears to cover.

BYusuf Kirimi, who directs the Institute's labour data programme, argues that the disagreement about size conceals a more useful fact about duration. Working from the internal records of four platforms between 2016 and 2021, some nineteen million accounts in total, his team found that 63 per cent of those who completed a first task had stopped entirely within eleven months, and that the median active period was under twenty weeks. A small group behaved quite differently: the tenth of accounts responsible for most of the hours worked stayed an average of three years and treated the platform as a principal occupation. Kirimi's conclusion is that two populations are being described by a single word, and he is careful not to call the larger of them a workforce. For the majority the platform functions as a temporary supplement, entered and left according to circumstances elsewhere in a household's finances; for the minority it functions as a job, and it is the second group whose conditions any regulation is presumably meant to address.

CThe reason people accept the arrangement is usually given as flexibility, and there is now a price attached to that word. In a field experiment run with a courier company in 2019, workers were offered guaranteed shifts at a wage 14 per cent above their average platform earnings; fewer than a third accepted, and take-up was lowest among those with caring responsibilities at home, precisely the group that fixed schedules are commonly assumed to help. Acceptance rose above half only among couriers who had no other employment at the time, a group the trial had not been designed to isolate. Later interviews suggested that the guarantee was less attractive than it looked because what it removed was the ability to stop at short notice rather than the ability to choose a starting time. The premium, in other words, is paid for interruption rather than for planning, and on these figures it is worth about a seventh of a wage. It is worth noting that the experiment ran for six weeks in one city, and that a shift offered at twice the platform average would presumably have found more takers.

DCamila Restrepo, an economist at the Aldenhoff School of Public Policy, is sceptical that the flexibility being purchased is real. Platforms, she points out, do not merely permit a worker to choose hours; they price those hours, raising rates during periods of heavy demand and offering bonuses conditional on completing a set number of trips within a window. The windows themselves are chosen by the platform. Analysing satellite positioning records from 900 riders in a single city over three months, she found that 71 per cent of hours were worked inside the same four daily periods, and that riders who declined the bonus structure earned 22 per cent less an hour than those who followed it. Freedom that is expensive to exercise, Restrepo maintains, is freedom in a formal sense only. She accepts two limits on her evidence: the recordings come from one platform in one labour market, and she cannot distinguish riders who follow the incentives from riders who would have chosen those hours in any case.

EThe practical problem is that most proposals for reform assume the population Kirimi describes as the minority. Minimum earnings guarantees, holiday entitlements and pension contributions are all designed for continuous attachment to a single employer, and they deliver little to someone whose entire history on a platform is eleven weeks. Systems that attach entitlements to the person rather than to the post, funded by a levy on each transaction and accumulating across platforms, fit the observed pattern rather better; two European jurisdictions have adopted a version of this since 2022, though it is too early to say what it costs, and a levy of this kind works only where platforms are obliged to share a register of hours. Restrepo would go further and regulate the incentives themselves. Kirimi doubts that a single instrument can serve both populations, and suspects that any uniform rule will be either too heavy for the weekend earner or too light for the full-time one. Both mechanisms, genuine choice and engineered choice, appear to be at work, in proportions that differ by city and by platform.