17. Case study: Is gig work decent work?

Bama Athreya
Open Society Foundations and Fellow, Just Jobs Network
  • Digital platforms commonly fragment available work and encourage an oversupply of labour, a form of arbitrage that undermines wages and working conditions.

  • Limited access to worker data sets creates information asymmetries that increase platform control over work and reduce worker agency.

  • Development co-operation actors and national policy makers should focus on measuring the macro effects of platforms on labour markets to determine how they affect overall employment and working conditions.

  • Projects that enhance collective rights for gig workers outside of traditional union structures are a necessary complement to “worker voice” technologies.

Since early 2020, the COVID-19 pandemic has accelerated a global transition to digital mediation in the world of work. Many of the big winners in the economic shifts – Amazon, DoorDash and Instacart, to name but three – are global companies that enable web-based platform interface between buyers and sellers of goods and services. They are now an important source of work worldwide.

The transition to platform work has been a novelty in countries where formal employment is the norm. It is less so in low- and middle-income countries where informal service work is already prevalent. The long-term resilience of economies may rely on the ability of workers, including low-wage and precarious workers, to negotiate for decent work in digitally mediated markets.

Over the past decade there has been a steady increase in development co-operation investments in digital technology for decent work. These have included interventions to smooth labour markets by connecting workers with jobs or short-term tasks (gigs) and interventions using technology to collect and curate information about workplaces for worker-management relations. It is not surprising that development co-operation providers have invested in platforms that promised to correct for labour market information gaps (USAID, 2019[1]). Developing country policy makers, too, consider platforms a possible solution to long-standing and seemingly intractable unemployment and underemployment. It is appealing to believe new technologies can address these problems. Yet, digitialisation is no panacea for persistent systemic barriers to decent work.

Imperfect labour markets, and in particular the information asymmetries that make it easy to exploit workers, are an important development challenge. The promise of digitalisation was that it would open new opportunities for workers and empower them. Yet, the limited evidence available suggests that platforms that match workers with tasks or jobs may not create more work for more people and that purport to enhance communication may not reduce information asymmetries that leave workers with little control over the data they share with employers.

In certain contexts, web-enabled technology has helped connect economic actors. Platforms such as Etsy, which harnessed growing enthusiasm for a peer-to-peer sharing economy, have played a useful role in addressing information asymmetries across geographies. Some job-matching platforms such as the International Labour Organization’s Employment Counselling System in Jordan have intentionally targeted refugee populations, correctly identifying this group as facing significant barriers to employment. Development co-operation providers such as United States Agency for International Development (USAID) have invested in digital platforms such as Babajob (India) and Bong Pheak (Cambodia), to name just two, on the assumption that more and better information will reduce search costs and other frictions, enable more jobseekers to find work, and ultimately reduce unemployment (Athreya, 2020[2]).

But projects have generally measured success in terms of engagement metrics, i.e. number of users or “hits”, rather than a platform’s broader labour market effects on unemployment or underemployment. While this may help evaluate effectiveness among the target group, such projects have not shared evidence regarding possible displacement effects in local labor markets. In short, there is little to suggest that such platforms create employment.

To date, there is little systematic evidence of the global effects of platforms on people in low-wage and precarious work, despite the growing number of platforms catering to this population. The International Labour Organization conducted the first comprehensive global survey, interviewing 12 000 platform workers worldwide for its recent flagship report (ILO, 2021[3]). The report charts trends indicating increased penetration of digital labour platforms in every region but, notably, the data were insufficient to project actual estimates of the worldwide platform labour force.

Where evidence does exist, it suggests that platforms are designed to draw in very large numbers of users and then engage in labour arbitrage – the practice of shifting existing jobs away from higher paid and more secure workers to lower paid and more precarious workers – both within countries and across borders. The International Labour Organization found evidence throughout its survey that digital platforms cultivate and benefit from excess labour supply, which leads to greater competition among workers for tasks and lowers per-task prices.

Platforms are designed to draw in very large numbers of users and then engage in labour arbitrage – the practice of shifting existing jobs away from higher paid and more secure workers to lower paid and more precarious workers – both within countries and across borders  

This can be seen in the case of service platforms such as Uber and Grab, which have disrupted local taxi and transportation options in many locales and flooded the market with unregulated providers. Across all such platforms, according to recent studies, more than 80% of work is performed by approximately 20% of the available workforce. Without these full-time workers, platforms could not fulfill the demand for services (Gray and Suri, 2019[4]). At the same time, a vast reserve pool of part-time or occasional workers is extremely important for continued labour arbitrage. By creating a situation of labour surplus, they ensure a continuous downward pressure on prices or wages for those who are engaged in full-time work. In location-based sectors such as transportation, this pressure is on traditional as well as platform providers, and evidence suggests that conditions for transportation providers in many low- and middle-income countries have deteriorated (Rest of World, 2021[5]).

Labour arbitrage also takes place at a regional and global level. Global platforms for cloud-based work such as Amazon’s Mechanical Turk, Rev and Upwork are designed so that work is performed virtually, thus pitting workers in less developed countries against those in OECD countries as they bid for tasks. This includes tasks requiring specialised skills such as editing, dubbing and design work (Hill, 2017[6]). Some types of digital piece work, such as geo-tagging, have from the outset been outsourced to countries where informal work is the norm, and here, the competition may occur between workers in different low- and middle-income countries.

Digital platforms also appear to be leading a fragmentation of available work. One widespread trend, even in countries with large formal economies, seems to be the fragmenting of formerly salaried or long-term contract positions into piece work (De Stefano, 2016[7]). In all countries, workers also face the fragmentation of piece-work assignments into ever smaller micro-tasks. There are insufficient data to determine whether this has increased either overall work available or average incomes for informal workers.

Whether platforms are employers is a heavily contested question in OECD countries (International Lawyers Assisting Workers, 2021[8]). In countries where informality is widespread, however, workers performing platform-enabled tasks such as delivery, transportation and even cloud-based task work were in most cases already working outside of formal employment relationships, with fluctuating availability of gigs or piece-work assignments.

A number of platforms established in recent years provide digital tools for workers to provide direct feedback to employers on workplace conditions – what has been called worker voice technology. There has also been substantial investment, including through development co-operation, in experiments using platforms to connect workers with one another. The promise of improved conditions for workers is not always realised. Evidence to date suggests that replacing offline social networks with such online tools is problematic, with platforms able to amass and potentially monetise workers’ data for other purposes.

Development co-operation providers and private philanthropic donors have invested in technology intended to provide management with information about the conditions of the people they employ, prompted by their growing fascination with information and communications technology as an enabler of social justice and what is commonly called worker voice. Ulula and Labor Link typify such investments.

Most such platforms have followed a data-extractive model and target improved business solutions (e.g. lower turnover and heightened workplace productivity) (Rende Taylor and Shih, 2019[9]). They typically extract information from workers via push-pull methods such as sending polling messages to capture data sets regarding common workplace issues, though individual workers have limited means to follow up on the results of these polls. While the underlying assumption is that employers will use results to improve conditions for workers, project outcomes generally are not measured in terms of actual workplace improvements, but in the level of worker engagement with the platform.

Some of the projects aim to use digital platforms to connect workers with one another to foster collective information sharing and possible collective action (Farbenblum, Berg and Kintominas, 2018[10]). These projects built on observations that low-wage and precarious workers such as migrant domestic workers in the Gulf states, though hindered by limited access to social media, were nevertheless finding and connecting with one another in organic ways on common messaging platforms such as WhatsApp and Facebook Messenger. Some organisations, inspired by this model, have created targeted apps such as Just Good Work (Fifty Eight, United Kingdom) and Golden Dreams (Issara Institute, Thailand) to attempt to reach and provide means for workers to share information.

Worker and citizen data sets can be a valuable asset for governments and societies, and worker voice platforms enable clients to apply the data they amass about their workers to internal business solutions. But monetisation of client and worker data sets is also fundamental to platform business models (Lee, 2018[11]), and data sets may not always be handled in ways that protect workers’ interests and privacy. Development co-operation agencies investing in worker voice and other platforms that capture data tend to have strong guidelines to protect individual privacy. Some agencies have open data policies that enable data sets to be accessed by other public entities such as academic researchers. But workers themselves, and their representative organisations, have lacked rights to access these data sets or to control their further use.

This imbalance is increasingly salient as more such apps are developed and deployed. Recently, private sector actors backed by venture capital have developed apps that target workers’ mutual interest in connecting with one another (Gurley, 2021[12]). Any value recouped from such activity will surely be in the data sets accumulated over time regarding worker behaviour.

A common feature of platforms is their tendency to treat workers as individual data points rather than as members of a group who are capable of acting collectively. Platforms to assign gig work use individual data to optimise work assignments, disrupting the traditional social networks that play a major role in informal labour markets. While this may create opportunities for some workers, it may displace others (ILO, 2021[3]). Furthermore, the data harvesting that is integral to the business model of platforms also enables labour market manipulation.

A common feature of platforms is their tendency to treat workers as individual data points rather than as members of a group who are capable of acting collectively. Platforms to assign gig work use individual data to optimise work assignments, disrupting the traditional social networks that play a major role in informal labour markets.  

Increased use of algorithmic management is another significant labour market disruptor. Algorithmic management uses artificial intelligence both for data collection and continuous surveillance of workers. Researchers have documented a number of harms resulting from the lack of guardrails on algorithmic management and have noted that codes are set to exert downward pressure on wages (Mateescu and Nguyen, 2019[13]). Design features created to manage workers through client ratings, or to de-platform workers for minor infractions, can leave workers with little choice but to perform work under exploitative conditions for fear of negative ratings. Mateescu and Nguyen (2019[13]) found these management features may deter workers from reporting harassment or abuse.

Other researchers note that continuous algorithmic cues to perform more work may lead workers to ignore their own well-being (Kellogg, Valentine and Christin, 2020[14]). Indeed, working through extreme fatigue has been a documented problem in the ride-hailing sector, and as accident rates became known, some platforms created features that force drivers to log out after a certain number of hours (Scheiber, 2017[15]).

Platform companies possess a sophisticated ability to penetrate labour markets and substantially direct economic activity. This power can be used for good or for ill. When workers have themselves attempted to collect their own data and then reverse-engineer the code, they have gained important insights into overall labour markets that enabled them to negotiate better terms of work (van Doorn, 2020[16]). Indeed, platform data sets on workers, utilised properly, could enable policy makers to work with employers and worker organisations to truly optimise labour market outcomes for all actors.

Some labour rights organisations such as the Centre for Migrant Rights (Centro de los Derechos del Migrante) are developing platforms that work directly with unions. For example, the centre has developed and launched its own platform to connect workers with employers and with each other. They work directly with a union that represents the workers and therefore gives them a collective say in how the platform is governed. These investments are worth supporting and expanding. So, too, are experiments to enhance collective rights for gig workers outside of traditional union structures. Examples include data platforms owned and controlled by worker organisations such as WeClock, Worker Info Exchange and Driver’s Seat.

It is tempting to believe there is a technological fix for every difficult problem. In the case of platform work, relying on anecdotes or even overall engagement metrics to judge the success of digitally enabled interventions may lead to the conclusion that gig work is, indeed, decent work. Individual workers, constrained by unpaid care burdens and offered a choice of flexible work, may report that gig work has improved their income. However, when an entire community or class of such workers begins collectively to rely on platform gatekeepers, opportunities for decent work may fade. Policy makers seeking to expand opportunities for decent work should be wary of investments that measure success in terms of short-term and individualised outcomes.

Development co-operation providers and other investors should help ensure that technology in digitally mediated markets lives up to its promise to expand work opportunities and empower workers. The following three recommendations are important to promote and protect decent work:

  1. 1. Measure labour market investments based on their macro, not micro effects. As a starting point, governments and development co-operation providers will need to measure deficits and gains in decent work in broader terms than individual success. If some individuals benefit but labour markets as a whole weaken, policy makers must consider how platforms may be contributing structurally and systemically to the erosion of decent work.

  2. 2. Enable workers to negotiate over their data and its use. In situations where workers willingly provide certain data to companies, they need visibility into the data set and the right to contest the programming logic behind automated decision making. This is true regardless of whether those companies are using workers’ data for job matching or receive the data as feedback on workers’ issues. Moreover, workers have a right not only to the raw data they provide. They also have a right to know how companies are using these data. This entails obliging companies, including those providing a human resource function, to share code with workers where that code is directly relevant to their work; see, for example, such a case in the United Kingdom (International Employment Lawyer, 2021[17]). Spain has been an early mover in mandating that platform workers have transparent access to algorithmic decision making (Ortiz, 2021[18]).

  3. 3. Keep worker organisations in the loop. Too often, interventions aimed at empowering workers, particularly those who have been restricted or excluded from traditional labour markets, are designed without consultation with relevant worker and civil society organisations that represent their collective interests. This failure gives rise to simplistic notions of worker voice that conflate aggregation with collective agency. Taking the time to consult these groups and consider the consequences of new tools is crucial. Technological interventions can reduce friction and increase the speed and ease of certain transactions. However, only when humans are firmly in the loop can they make sure that decisions do not sacrifice too much humanity for the sake of efficiency.


[2] Athreya, B. (2020), “Slaves to Technology: Worker control in the surveillance economy”, Anti-Trafficking Review 15, pp. 82-101, http://dx.doi.org/10.14197/atr.201220155.

[7] De Stefano, V. (2016), “The rise of the “just-in-time workforce”: On-demand work, crowdwork and labour protection in the “gig economy””, Conditions of Work and Employment Series, No. 71, International Labour Organization, Geneva, https://www.ilo.org/wcmsp5/groups/public/---ed_protect/---protrav/---travail/documents/publication/wcms_443267.pdf.

[10] Farbenblum, B., L. Berg and A. Kintominas (2018), Transformative Technology for Migrant Workers: Opportunities, Challenges and Risks, Open Society Foundations, New York, NY, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3295430 (accessed on 13 October 2021).

[4] Gray, M. and S. Suri (2019), Ghost Work: How to Stop Silicon Valley from Building a New Global Underclass, Mariner Books, Boston, https://ghostwork.info/ghost-work (accessed on 13 October 2021).

[12] Gurley, L. (2021), “A new app is taking labor unions out of union organizing”, Motherboard/VICE Media, https://www.vice.com/en/article/epn87m/a-new-app-is-taking-labor-unions-out-of-union-organizing.

[6] Hill, S. (2017), Raw Deal: How the “Uber Economy” and Runaway Capitalism are Screwing American Workers, St. Martin’s Griffin, New York, NY, https://us.macmillan.com/books/9781250135087/rawdeal.

[3] ILO (2021), World Employment and Social Outlook 2021: The Role of Digital Labour Platforms in Transforming the World of Work, International Labour Organization, Geneva, https://www.ilo.org/wcmsp5/groups/public/---dgreports/---dcomm/---publ/documents/publication/wcms_771749.pdf.

[17] International Employment Lawyer (2021), “The battle for transparency in AI decision-making”, International Employment Lawyer, https://www.internationalemploymentlawyer.com/news/battle-transparency-ai-decision-making.

[8] International Lawyers Assisting Workers (2021), Taken for a Ride: Litigating the Digital Platform Model, International Lawyers Assisting Workers, https://www.ilawnetwork.com/wp-content/uploads/2021/03/Issue-Brief-TAKEN-FOR-A-RIDE-English.pdf.

[14] Kellogg, K., M. Valentine and A. Christin (2020), “Algorithms at work: The new contested terrain of control”, Academy of Management Annals, Vol. 14/1, pp. 366-410, https://doi.org/10.5465/annals.2018.0174.

[11] Lee, K. (2018), AI Superpowers: China, Silicon Valley, and the New World Order, Houghton Mifflin Harcourt, Boston.

[13] Mateescu, A. and A. Nguyen (2019), Explainer: Algorithmic Management in the Workplace, Data & Society, Washington, DC, https://datasociety.net/wp-content/uploads/2019/02/DS_Algorithmic_Management_Explainer.pdf.

[18] Ortiz, M. (2021), “Spain is about to shatter the gig economy’s algorithmic black box”, Wired, https://www.wired.co.uk/article/spain-gig-economy-algorithms (accessed on 13 October 2021).

[9] Rende Taylor, L. and E. Shih (2019), “Worker feedback technologies and combatting modern slavery in global supply chains”, Journal of the British Academy, Vol. 7/1, pp. 131-165, https://www.thebritishacademy.ac.uk/publishing/journal-british-academy/7s1/worker-feedback-technologies-and-combatting-modern-slavery-in-global-supply-chains.

[5] Rest of World (2021), “Gig workers are uncertain, scared, and barely scraping by”, https://restofworld.org/2021/the-global-gig-workers (accessed on 13 October 2021).

[15] Scheiber, N. (2017), “How Uber uses psychological tricks to push its drivers’ buttons”, The New York Times, https://www.nytimes.com/interactive/2017/04/02/technology/uber-drivers-psychological-tricks.html (accessed on 13 October 2021).

[1] USAID (2019), Gig Work on Digital Platforms: Executive Summary – Learning, Evaluation and Research Activity II (LER II), United States Agency for International Development, https://pdf.usaid.gov/pdf_docs/PA00WHJ7.pdf.

[16] van Doorn, N. (2020), “At what price? Labour politics and calculative power struggles in on-demand food delivery”, Work Organisation, Labour & Globalisation, Vol. 14/1, pp. 136-149, https://doi.org/10.13169/workorgalaboglob.14.1.0136.

Metadata, Legal and Rights

This document, as well as any data and map included herein, are without prejudice to the status of or sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the name of any territory, city or area. Extracts from publications may be subject to additional disclaimers, which are set out in the complete version of the publication, available at the link provided.

© OECD 2021

The use of this work, whether digital or print, is governed by the Terms and Conditions to be found at http://www.oecd.org/termsandconditions.