Technological Change and Work: Opportunity and Inequality
This is part two of a summary of my PhD thesis. You can find part one here and more information about the thesis here.
Does technology make jobs better? Optimistic visions of the age of automation suggest that new inventions will relieve workers of the most burdensome and unpleasant aspects of our work — tedium, heavy lifting, and dangerous parts of jobs.
However, evidence from historical instances of innovation cautions against this optimism. Indeed, one of the biggest debates in economic history is about whether the downsides of the Industrial Revolution, such as working in dangerous, smoky factories, with a foreman or overseer demanding workers stick to their tasks for very long hours, offset higher wages for factory labor.
In this post, I will look at one type of innovations, what economic historians call “macroinventions” (big changes, like steam engines or factories) and explain what some historical examples suggest about their impact on work. In the next post, I’ll describe how smaller changes (“microinventions”) and business practices change jobs after companies have adopted macroinventions.
First, let’s take a step backward. What is a job? The definition I use is that a job is a set, or “bundle”, of tasks put together by an employer. HR Departments compile these for modern job advertisements, but employers have always needed to have an idea of what they want workers to do, whether it is stacking bricks, plowing a field, or cleaning a factory machine.
In the examples that I analyzed, macroinventions tended to increase the range of different tasks that are required to produce or good or provide a service. This happens even though the new technology increases productivity. How can these two outcomes fit together? Industrial technology replaced human or animal dexterity and strength with machines that could maintain accuracy for longer than the human body, and apply more force. In manufacturing, for example, this meant it was possible to produce more and bigger products. In a similar way, a lot of modern automation aims to replace our brains with computer processes that make accurate calculations faster and for much longer.
Replacing humans and animals added tasks that had been implicit or internalized when people and animals worked without industrial machinery. In pre-industrial production, muscles usually provided power. In a factory the power source — let’s say a steam engine — required specialized employees to provide fuel, make adjustments, clean, and maintain it. The range of tasks to power production increased, and employers responded by separating those tasks into discrete bundles, or jobs.
Here are two more detailed examples. First, in the textile industry before the Industrial Revolution, yarn for cloth was produced by a single spinner turning a spinning wheel. She had a few tasks that she repeated hundreds or thousands of times per day. She controlled the fineness of the yarn with her fingers and by turning the spinning wheel, and wound it on to a bobbin that she would give to a weaver.
Spinning machines, located in factories using steam or water power, had many spindles (each of which produced one strand of yarn) and rotated them much faster than the spindle of a spinning wheel. The spindle speed and yarn fineness were taken out of the worker’s hands; instead they were controlled by fast-moving gears and straps. Instead, workers were employed to clean and oil the machinery, to patch (or “piece”) broken threads back together, to adjust the machinery, repair it, oversee other workers, and to monitor the water wheel or steam engine. Machines produced many more pounds of yarn per worker, but they also widened the range of tasks.
In response, managers bundled tasks into new jobs: cleaning, oiling, and piecing was one task bundle, and one new job (“piecer”); machinery adjustments and repairs were another (“engineer”); labor oversight was a third (“overseer”). In many large factories, managers differentiated within these bundles, assigning some children to remove bobbins of spun yarn and carry them to a warehouse, while other children only cleaned, and a third group only pieced.
Off the shopfloor, there was a growing array of administrative tasks and jobs. Factories could produce far more yarn than hand spinning. As a result, they needed employees to arrange purchases of raw materials (cotton or wool) and machinery, administer the company’s operation (hiring, firing, marketing and business strategy), and oversee expenditure (sales and wage payments). For these task bundles — new jobs — they hired professional managers, clerks, and accountants.
These jobs had distinct functions, with workers completing different tasks, but there were also large differences in pay, occupational risk, and work intensity. Overseers or foremen were paid three or four times the wage of the workers they supervised, and managers could earn 20 times the wage of ordinary workers. Supervisors and white-collar workers were rarely injured in accidents, and managers and clerks avoided the choking, dust-filled air on the factory floor. The highest paid jobs were superior on dimensions in addition to income.
This process of technology expanding the range of tasks and therefore the number of separate occupations also occurred outside of manufacturing. In 1750, transportation in North America was mostly by horse and cart or in small boats. There were few occupations, which was the result of a limited range of tasks. A cart or carriage driver directed their vehicle, fed and cared for the horses, and loaded the baggage and goods. Maintenance of roads or waterways was primarily carried out by unskilled laborers, sometimes aided by teamsters and masons or carpenters to repair bridges and culverts. These workers shoveled dirt or gravel, cut back bushes, and placed or replaced stones, bricks, and wooden components of structures.
By 1850, the railroad had dramatically expanded the range of tasks involved in moving goods and people. Instead of one driver, each train had an engineman (who we now call an engineer) to control the locomotive, a fireman to fuel it, a conductor to oversee the operation of the freight or passenger cars, and several brakemen to stop the train, change switches in remote locations, and act as flagmen when needed. That was just the on-train crew: railroads also had a wide variety of workers in stations to serve passengers and handle ticketing, baggage, and freight; an array of mechanics and laborers to repair and maintain their locomotives and cars; train control personnel such as switchmen, dispatchers, and telegraphers to direct and communicate with trains; maintenance workers to repair track, bridges, viaducts, and tunnels; and all of this work was overseen by many new managers and a corporate bureaucracy that included lawyers, accountants, and civil engineers.
This huge expansion in the division of labor to provide the same service — moving goods and people — was the railroad companies’ response to the vast increase in the range of tasks. The technology of the steam-powered railroad demanded that managers put together many new task bundles, which meant new occupations, to cope with the complexity of operations.
What were the impacts on the differences between jobs? Like in the textile industry, there was wage inequality, with master mechanics in charge of a roundhouse earning multiples of the wages of machine shop laborers, conductors making several times the wages of the brakemen they oversaw, and railroad managers reaping the biggest gains. Again, though, job conditions became more unequal: working as a brakeman was not only poorly-paid compared to white-collar or supervisory work, but it was also exceptionally dangerous. Brakemen were frequently thrown from train cars or crushed between them, and they lost fingers or hands in regular service. All workers on trains, as well as the modestly-paid track laborers, had high-intensity, exhausting jobs — a sharp contrast with the relatively comfortable positions of railroad lawyers or accountants. The white-collar employees also had much greater job security. Across the board, differentiation of work into distinct occupations led to inequality on dimensions beyond wages.
On the other hand, these were new jobs, some of which were very good by 19th century standards. There were more options available for men in 1850 with formal education and managerial skills, or advanced manual skills, than there had been a century earlier. The broader range of tasks allowed for new work opportunities as well as greater inequality.
Adam Smith famously described the division of labor in a pin factory in The Wealth of Nations. He explained how separating tasks without changing technology could increase productivity and encourage individual workers to develop labor-saving inventions. This process is different: in this case, the division of labor increased in response to a change in technology.
It is also distinct from a phenomenon observed by economists in rich countries during the last few decades. Many good, middle-income jobs have been replaced by automation and offshoring, in a process that is sometimes described as “job polarization” or “hollowing out”. Rather than polarizing work, these industrial technologies stratified work, creating a series of new occupations, some of which were good jobs. However, technology did little to improve the worst jobs, and thereby increased inequality.
Innovation that widens the range of tasks, creates new jobs, and leads to job stratification is one possible outcome of macroinventions. In the next post I will discuss what happened after companies adopted major innovations but still sought to increase output per worker.
