July 17, 2018
Online Outsourcing and The Future of Work
The push towards global connectivity and the worldwide expansion of the Internet, combined with the ongoing decline in the cost of technology, will change global labor markets and the nature of employment. Future technology will coordinate the assignment, distribution, and measurement of tasks to a billion-person network across multiple countries and dozens of skill vectors. The fixed-salary, full-time employee will be replaced with a network of online contractors who receive and deliver their work through computer-mediated auctions, assignments, and other economic mechanisms. I document these trends, explain the underlying economics, investigate the current labor laws and benefits for independent contractors, and suggest policy proposals to prepare for the transformation.
The text you are now reading began as an audio file recorded in Texas: it was sent to the IBM Watson artificial intelligence engine that converted the audio into text, edited by a worker in South Africa who competed in an online auction against a pool of workers around the world, reviewed by a graduate research assistant in Washington D.C., printed and marked up by the author in Texas, and edited once more by another worker in the Philippines. All of these steps occurred entirely through the Internet, and they represent the modern-day equivalent of the assembly line. The Internet has expanded its reach around the world through strategic investments by technology giant like Facebook and Google. These mostly poor populations coming online for the first time will serve as producers, rather than consumers, bringing their own labor to the global marketplace. The Internet will serve as a massive clearinghouse for matching buyers and sellers of work, and this will have deep ramifications for the global distribution of wages, skills, and productivity.
This transformation, currently in its early innings, is not inevitable. The recent push toward nationalism and retreat from global economic integration can threaten this new global market. It is incumbent on entrepreneurs, financiers, intellectuals, and members of civil society to commit to free trade in services, especially that which will take place over the Internet. In what follows, I assemble preliminary evidence on this economic transformation, anchor the analysis in economic principles, speculate on long-term economic impacts, and recommend possible strategies to encourage online labor markets through actions of the public and private sector.
Evolution of Outsourcing
The first wave of outsourcing was offshoring, in which manufacturing jobs in the 1960s-1970s moved abroad because of free trade (Mankiw and Swagel 2006; Gereffi 2005). The second wave of outsourcing took place during the first Internet bubble of the late 1990s and early 2000s. During this time, companies set up operations abroad to outsource services handled by their back office, such as call centers in India. The third wave will be online outsourcing. Technology platforms will allow buyers and sellers to contract for work through dynamic and continuous auctions. Corporations will no longer only hire full-time employees who conduct all the tasks themselves, but rather these employees will interface with a global network of workers available on demand. This third wave of outsourcing will open up whole new worlds of productivity, expand economic opportunity across the globe, and change the nature of work. Forbes showed that in 2014, nearly 25% of the United States’ workforce were freelance or contingent workers. It is projected that 40% of the workforce, or 60 million U.S. workers, will be freelancers by 2020 (Pofeldt 2015).
Outsourcing’s initial wave brought with it passionate claims that foreign workers would steal American jobs. We now know much of the rhetoric was overstated. Global wages reflect productivity, and American factories’ productivity is also high, given high levels of capital investment. Many of the benefits from offshoring failed to materialize because capital-enabled factories in the U.S. could not compete with the much lower capital-to-labor ratios of factories abroad. This same logic falls apart in the online world. The spread of information technology has made IT more uniform across the world. For the purposes of services, a desktop computer in New York versus a laptop in the Philippines are roughly equivalent; therefore, it is likely that outsourcing will shift more jobs outside of the firm than offshoring did in the past.
The notion of distributed computing (which forms the basis of cloud computing) has inspired the development of distributed work. In the future, occupations as varied as architecture, financial services, software development, writing, customer service, and marketing will occur partly over an international network of contractors through a continuously adaptive mix of auctions and assignment algorithms.
To fix ideas, let me distinguish between online outsourcing and the “gig” economy. Companies like Uber, TaskRabbit, Thumbtack, and Airbnb all match buyers and sellers together over Internet-based technology platforms which result in a local transaction. Though Uber is a global company operating in cities around the world, every transaction takes place by matching a buyer (passenger) and a seller (driver) together in a local trade because transportation is a physical transaction. Alternatively, websites like Upwork and Amazon Mechanical Turk (AMT) match buyers and sellers of labor in global transactions. Upwork serves as a directory for companies and individuals to hire a wide variety of freelancers to perform computer programming, typing, copyediting, virtual assistance, and other Internet-based tasks. In fact, due to global disparity in skills and wages, these transactions are almost exclusively non-local, often with buyers residing in high-wage countries and procuring services from workers in low-wage countries.
Online outsourcing refers to the latter kind of trade, whose development is still in its infancy. Nonetheless, this kind of outsourcing technically falls under the gig economy because workers are acting as independent contractors, rather than full-time employees, which has consequences for benefits discussed later.
As aforementioned, the two major gig economy platforms are AMT and Upwork (Ipeirotis 2012). Upwork registers eight million freelancers on its two main sites: Elance and oDesk. The company claims that 53 million Americans are freelancing at any point, contributing $700 billion to the economy (Upwork n.d.). This indicates that Americans participate in online labor markets both as workers and as employers. Elance posted over one million jobs in the first quarter of 2011, and almost 3.5 million by the end of 2013. Freelancers earned $285 million in total earnings by the end of 2013 (Elance 2013).
Scope of the Transformation
The scope of this shift is significant. Consider this simple fact: the growth of online marketplaces as a business model has ballooned in recent years, which has enabled buyers and sellers to contract through technology platforms. Indeed, the world’s top five firms in the last five years with regards to market value have come from the top five United States technology companies alone; Facebook, Google, Amazon, Apple, and Microsoft all run immense marketplaces. For example, Amazon has created well-defined procedures for sellers to list their goods online, advertise to buyers, close a sale, evaluate transactions, and establish long-term reputation mechanisms to bring in future business. What has already occurred in the goods marketplace will expand to the services markets.
This argument relies on the extensive, rather than the intensive, margin of labor supply. The intensive margin refers to increasing the hours of work from those already in the labor force, whereas the extensive margin refers to bringing new populations to the labor market altogether. The intensive margin will count workers who leave full-time, face-to-face jobs in favor of online work, whereas the extensive margin will count all of the population who will come online in the future to use the Internet as a means of work. My thesis is that the extensive margin will dominate the intensive margin, thus growing more rapidly in the future than today. Of the world’s population of 7.36 billion people, 4.20 billion people do not have any kind of Internet access that would allow for the performance of work online. If one-eighth of these people are brought online and enter the workforce through online opportunities, then the total global labor market would increase by 15%.
These calculations are admittedly speculative, but so too were the early calculations on the growth of the Internet. For example, as Jeff Bezos has stated in public speeches, the key statistic that led him to leave his comfortable, high-paying job at the hedge fund D.E. Shaw in favor of starting Amazon.com was the Internet’s 2300% annual rate of growth in 1994. With technology giants like Facebook disclosing spending of $860 million per year on Internet infrastructure, the droves of people coming online are virtually inevitable.
While this transformation will be significant, it is unlikely to completely change all of current work practices. For example, work requiring tight coordination, creative input, and face-to-face feedback from multiple employees will always provide a need for full-time, in-person employment. As the trend towards agglomeration grows, the countervailing force will be the increased efficiency of urban centers, which will still make it efficient for some work to occur inside the corporation face-to-face. Thus, traditional corporations will not vanish, but rather adapt to this new reality in which online workers are a non-trivial component. Full-time employees will interface with the network of on-demand workers, thereby shifting some work out of the firm.
The Underlying Economics
The academic literature on the economics of organization can provide some guidance on the boundaries firm’s and, most importantly, the kind of work taking place within the firm versus the market. Nobel Laureates like Oliver Williamson and Ronald Coase established their careers by developing theories of transaction costs that help determine the location of work and the structure of the firm (Coase 1937). Broadly speaking, when transaction costs are high, organizations like corporations are efficient at conducting work because the control structure of a hierarchy is an effective way to produce output. In contrast, when transaction costs are low, individuals can contract with each other at low cost and can do so through price-mediated markets.
The key economic feature of the Internet has been the massive amount of disintermediation, which has led to a marked decrease in transaction costs. Before, a buyer and seller contracting with one another usually required a face-to-face meeting, possibly a long history of a business relationship, or even the use of a lawyer. However, the Internet has replaced many of these requirements as buyers and sellers no longer need to be in the same physical space, nor do they require local services like banking or legal contracts. Instead, reputation mechanisms online can serve the same function to enforce behavior. As such, the secular decrease in transaction costs brought on by the Internet has led to a shift in the boundaries of the firm, as more and more work will move from within corporations into the marketplace through networks of contracts.
There exists a vast number of examples of this widespread disintermediation in transactions. Previously, an entrepreneur in the U.S. would hire a graphics company to design its website. This company served the role of providing office space, recruiting talent, and securing payment. Those designers can now register as independent freelancers over the Internet, and entrepreneurs can contract directly with them. The Internet can serve many of the same roles that were once performed by the company: online search can substitute for business development, Paypal can provide a vehicle for payments, and the platform itself can provide performance evaluation and monitoring. Disintermediation cuts across a wide swath of industries, ranging from architecture to financial services, and the Internet has enabled the removal of the middleman to provide more direct contact between buyers and sellers. Chicago economist Ronald Coase proved transaction costs were the only barriers to efficiency between two bargaining parties. As transaction costs fall, negotiated outcomes move closer to efficient solutions. The growth of digital technology and its concomitant performance measurement, search, and payment functions all help eliminate intermediaries and reduce transaction costs to facilitate greater trade between buyers and sellers.
The second major trend is the rise of global Internet connectivity. Facebook and Google have made large investments in expanding the reach of the Internet, especially to poor, rural, underserved areas. As of 2015, only 3.2 billion (43% of the total world population) individuals have access to the Internet (Kende 2015; Broadband Commission 2014, 2015). The specific technology used to deploy the Internet will vary depending on the area, ranging from airborne drones, balloons, low-level satellites, towers, etc. While it is in the economic interest of these companies to bring new users to its platform, the positive externalities on society will be large. With the Google Chromebook available for a mere $150, cheap laptops will spread across the developing world. These newcomers are unlikely to be large consumers given their low incomes and, therefore, are not the usual audience Facebook and Google target for advertisement. Instead, they will bring their labor to a global market, and new companies will form to harness this human capital. Bringing this human capital online will undoubtedly be the largest economic transformation the world has seen.
The third economic trend is the shift from input-based pay to output-based pay. Input-based pay describes most forms of compensation in place today with physical, face-to-face employment. Workers are paid a salary, possibly with some discretionary bonuses. Output-based pay refers to paying based on the performance of the worker. This can take the form of stock options for the CEO or a piece rate for every unit produced by a worker on the assembly line. The remote and distributed environment of the Internet makes it a natural candidate for output-based pay, which will lead to significant efficiencies.
The academic support for output-based pay is long and deep. Output-based pay, also referred to as performance pay, has three primary benefits for productivity. First, it aligns incentives between the employer and the worker; second, it attracts better workers to the firm; third, it retains these talented workers over time. These three factors are called the incentive, selection, and retention effects of performance pay.
Notable labor economists like Edward Lazear have shown performance pay can lead to a 44% increase in productivity over fixed salaries and other input-based pay (Lazear 2000). This increase splits evenly between incentive effects and selection effects. The economic benefits of performance pay explain the widespread use of equity compensation among executives and directors of major corporations around the world. It also explains the use of performance pay throughout financial services and entrepreneurship, where founders of new companies receive large equity grants as incentives to grow the firm. These benefits will transfer to online workers.
The measurement of human output is only increasing over time, since sensor technology is becoming ever more portable and cheap. The self-quantification movement, which first established itself in fitness culture and health tracking, will eventually permeate most jobs. The spread of cheap sensor technology will improve the monitoring and measurement of human output, leading to Big Data opportunities for better, more refined, and more dynamic compensation schemes.
Performance compensation schemes are called “mechanisms” in academic literature. They are complex systems of payment to workers based on a vector of environmental variables, prior productivity measures, and other economic covariates. Surge pricing by Uber is one example of a mechanism that dynamically adjusts prices to match supply and demand for drivers and passengers. Other similar mechanisms will improve the productivity of online workers.