Online outsourcing and the future of work

business talk

Civil society will need to acknowledge the benefits of free trade in services. The current wave of nationalism spreading throughout the world has led many governments to retreat from the principles of free trade. Although such policies may seem to benefit a nation, generations of economists know they actually hurt the nation’s long-term economics. Ultimately, free trade benefits both the producer and consumer surplus. Computations of welfare that defend protectionist trade policies usually count producer welfare but not consumer welfare. NAFTA may indeed have cost some manufacturing jobs in the U.S., but American consumers are better off as a whole, due to the subsequent lower prices on commodity products imported from low cost countries.

Just as faith in markets versus governments waxes and wanes over time, so too will this debate on nationalism versus globalism. When the benefits of nationalism fail to materialize, the world will return to free trade. The next major debate will center on trade in services, rather than trade in goods. Outsourcing will take center stage once again and will change the very nature of multiple policy debates, such as immigration. A Filipino must wait years before being allowed to live in the U.S. but can now register online to work in the U.S. through online labor markets in seconds. As the debate on immigration shifts to productivity of human capital, the location of that capital becomes irrelevant once most work shifts online.

Government policy should, at minimum, not interfere with the development and growth of online labor markets and, at best, encourage their development. Examples of policies that could harm online labor markets are net neutrality rules that prevent pricing the Internet and restrictions on the buying or selling of work online through a form of labor tariff. Currently, such tariffs do not exist, but it is entirely conceivable that a protectionist nation state could impose wage tariffs on all online work sourced from different nations.

To further push the analogy with free trade, we know trade creates winners and losers, even though it makes the economy better off as a whole. The traditional response to those hurt by trade has been Trade Adjustment Assistance (TAA), a federal program in place since 1962 that provides benefits to both workers and firms (Hornbeck 2013). Workers displaced through trade receive retraining, relocation allowances, and extended unemployment benefits – while firms hiring displaced workers receive loans, loan guarantees, technical assistance, and tax benefits. In the name of fairness, TAA seeks to compensate the losers by redistributing wealth from the winners (the population of taxpayers). The budget for TAA as of 2015 was $710.6 million,5 still a small sliver of the total $3.68 trillion federal budget in 2015.6

In large measure, the TAA program is fairly uncontroversial. Most of the debate centers on the size and scope of the programs, and it does not follow traditional party lines. One problem with TAA is its exclusive focus on trade-related economic shocks. For example, the TAA website tells the story of Mr. Bustamante, a machine operator who lost his job because his company shut down his factory and moved it overseas. The TAA program paid for his retraining as a bus driver, and he soon earned a full-time job as an HVAC mechanic (United States Department of Labor, Employment, and Training Administration 2016).

In regards to this, two immediate concerns come to mind. First, had the company kept the plant in place and simply replaced Mr. Bustamante with a robot who could operate the machine, Mr. Bustamante would not have been eligible for TAA benefits. Thus, it seems somewhat arbitrary that the government is compensating individuals from one kind of economic shock (trade) but not another (technology). Second, robots may eventually replace all mechanics, thus eventually displacing Mr. Bustamante from his new job as an HVAC mechanic. Obviously, it is not the government’s role to pick and choose between industries, as it is no better at forecasting future technological change than others. However, that is exactly what it is doing through the TAA program when it pays for retraining into jobs that may eventually become obsolete anyway. This begs the question of whether this is the best use of taxpayer funds.


(United States of America. Department of Labor. Congressional Budget Justification: Employment and Training Administration. 11. 6 United States of America. Congressional Budget Office. Monthly Budget Review: Summary for Fiscal Year 2015.)

Political economy justification for trade assistance programs exists namely as support for trade agreements. Trade assistance is a way to compensate the losers from free trade so they will support (or at least not oppose) trade agreements. However, a natural analog to trade agreements with respect to technology does not exist. When Amazon replaces factory workers with a fleet of robots, this does not impinge on any specific national economic policy because technological change is a secular productivity shock. Thus, it is not clear that the government needs to have any kind of formal TAA for technology. There are no similar ways for workers to politically oppose technological change through government policy. Therefore, political economy arguments do not work for making a persuasive argument to compensate losers from technological change. Rather, the most the government can do is prepare workers through broad-based approaches.

Skill Development and Education

Skill development and education

A better approach is for the government to acknowledge online labor markets as inevitable and prepare workers for them. However, rather than structuring an ex-post corrective policy like TAA, a more sensible approach is to consider ex-ante investments in human capital that prepare people to participate in online labor markets. Primarily, this can occur through broad-based education and skill development. It is risky to build government policy around technological innovations because those innovations change much quicker than Washington can move, so a natural alternative is to leave the details to the states; they can compete amongst themselves in offering training programs for those seeking to work online. The federal government can play an effective role as a disseminator of information, making these online labor markets fully transparent through uniform disclosure laws and light forms of market-based certification.

Online labor markets already have a wide spectrum of skills necessary for their participation: ranging from data entry at the low end to advanced software development at the high end. A nationwide emphasis on STEM education would help with both the development of the market itself as well as American participation in the market once it is developed. Yet, existing proposals for STEM education are vague at the moment. Both the federal and state governments are scrambling to make investments in STEM education to respond to the secular shortage of computer science majors. It may require a more targeted approach, with careful understanding of the needs of online labor markets. This targeted form of educational investment can be optimal because it avoids picking specific winners and losers and provides enough flexibility for individuals to tailor their own educational needs. However, getting the balance right between broad-based and targeted education will be a challenge.

As a concrete example, the government could solve collective action problems of disparate private actors by suggesting investments that promote skill development in addition to the usual investments in public education that states routinely make. Simply identifying areas of need and requiring public disclosure about online labor markets could induce the private sector to make the requisite investments in such areas. For example, if learning to code in Python becomes the dominant technical skill of the future, the government could use some of its National Science Foundation research funds to support Python education and research. It could also encourage national dialogue about developing companies that provide broad-based Python training to a wide class of workers.

Nonetheless, there may be a limit to upskilling. The government needs to watch the development of the online labor market to ensure any transfers to the skill development and education sector do not distort the behavior of market participants’ ex-ante. For example, careful monitoring of online labor markets may show that American workers suffer from poor English grammar relative to their Filipino counterparts. This signal can inspire greater federal and state investments in English language training. Again, the key is the government leaning on signals from the market; in this case, from the emerging online labor market.


The main investment needs to come from the private sector. First, the labor markets themselves have room to grow, and the existing platforms like AMT and Upwork need much broader participation and engagement. Corporate America need not be ashamed to add online workers to their staff of full-time equivalent (FTE) employees. Finally, financers like venture capitalists play outsized roles in the future evolution of industry, and they would be wise to take part in the future profit-making opportunities of online markets.

Online labor markets will require expertise in building technology platforms closely calibrated to market conditions, such as finding the best way to measure human performance and integrate said data into the economic mechanisms that run the platform. Participation in the online labor market will require skills that emerge from the demand side of the market itself. As the market grows, the platform will reflect the needs of employers and entrepreneurs around the world, thereby providing a window to the different skills necessary to power the global economy. For example, ten years ago, onsite network management was a key function of most businesses because of the growing pool of desktop computers connected on company networks. However, as technology shifts and more of these resources move to the cloud, the key skill in high demand will be the ability to manage multiple Amazon Web Services (AWS) systems distributed remotely. It is imperative for the online labor market to reflect timely, and even prospective, signals on labor demand.

Worker Benefits

Worker benefits

A beginning question is how to classify workers. In the U.S., traditional full-time employees are classified and protected by the Fair Labor Standards Act (FLSA), which guarantees minimum wage, overtime pay, and explicit rules for calculating hours. If a worker is not covered by FLSA, then they are classified as an independent contractor. Therefore, there is an ongoing debate on how to classify gig economy workers and whether they should have the same benefits as full- time employees.

A key concern for regulators is the workplace safety net. Ever since the Occupational Safety and Health Act (OSHA) of 1970, a raft of employment protections come with full-time jobs. Indeed, OSHA first attached workplace disability insurance to jobs, while worker’s compensation has been attached to jobs since 1911 at a state level and 1906 at a federal level (Guyton 1999). These benefits are deeply entrenched in both law and American culture, and the race is on to think of how to represent these benefits in an online world. While campaigning in the 2016 presidential election, Hillary Clinton stated that the “on demand, or so called ‘gig’, economy is creating exciting economies and unleashing innovation. However, it is also raising hard questions about workplace protections and what a good job will look like in the future” (NASSCOM 2014).

This captures a particular concern Democrats have with the online economy, namely the erosion of workplace benefits. In the current regime (determined through labor law), employers are required to provide extensive benefits regardless of the skills or responsibilities of the worker (Department of Labor 2016). This effectively makes the worker a fixed cost to the firm, and therefore difficult for the firm to scale up or down as business needs change.

From a distributional perspective, those who are hired receive both benefits and a wage, whereas those who are not hired receive nothing. In this sense, the current regime of attaching benefits to the job increases inequality between workers for the same reason minimum wage does. Benefit exchanges would certainly be an improvement over this current regime because they allow more choice and flexibility for both the worker and the employer. The worker can choose the level of benefits they seek, and the employer can dial up or down the quality of benefits they need to pay. Every employer will face the same economic tradeoff: dollars spent toward wages can substitute for dollars spent toward benefits. However, the firm cannot unilaterally increase both without cost. The benefits attached to jobs are not “free”; they are a clear cost to the firm. On the margin, firms forced to pay large benefits through mandated labor laws will, in turn, offer lower wages. This emerges from the simple fact that the firm is subject to a budget constraint.

A more fundamental question is whether benefits should be attached to jobs at all. In a free market, workers can simply purchase whatever benefits they need; these benefits do not need to run through their employer. Unless a clear case for economies of scale can be made, workers would be better off simply receiving higher wages and buying what they need a-la-carte in a marketplace. Of course, this would require a more developed market for benefits. If legislators successfully rolled back some of the current onerous labor laws, it is entirely conceivable this market for benefits would emerge. Indeed, no theoretical or conceptual reason explains why benefits should run through employers and be such large fixed costs as they are today.

The chief argument for employers purchasing benefits themselves is the reality of economies of scale. If the firm acts as a single buyer on behalf of its workers, it may negotiate better prices and better terms than workers purchasing benefits themselves (Employee Benefit Research Institute 2002). In this sense, the firm can pass these bargaining profits down to workers, making workers better off. The benefits provider is able to lock in a large pool of customers at once, which may be advantageous against contracting individually with workers. Even before employee-sponsored insurance became widespread, there is some evidence that private groups still provided health insurance. For example, Beito (2000) documents that fraternal organizations provided healthcare for its members before World War II.

Some solutions that are market-oriented have circulated in the policy space to address the problem of providing benefits to flexible online workers. For example, in National Affairs, Eli Lehrer discusses worker-controlled benefit exchanges that would substitute for the traditional package of benefits offered to full-time employees through employers. Workers can shop around from different vendors and assemble a package of benefits that provides a “replacement income stream,” which effectively serves the role of disability insurance, worker’s compensation, and sick leave. The key with these exchanges is their flexibility: workers can mix and match across different vendors, and employers do not need to commit to using the same exchange. This flexibility will lower costs for workers because vendors will compete with each other to offer the best benefit at the lowest price. Employers would pay into the exchanges, just as they pay into the benefit vendors under current law. This notion of market-based exchanges is promising because it injects a level of flexibility and creativity in the benefits space that does not exist today, as workers would have choices among vendors.

Related outsourcing resources

    Transform your business