Going back as far as reasonably comparable Census data will take us, we can see how challenging life in 1870 was compared to today. More than half of the population subsisted on farms and lived at the whim of the weather. The U.S. was still recovering from the Civil War, the deadliest in its history, and slavery had only recently been abolished. Overall life expectancy was just shy of 44.
And yet there were some technological and social breakthroughs just around the corner that would set in motion significant changes to everyday life. Many of the big projects that had been put on hold due to the war among other issues, like the first fully functioning transatlantic telegraph cable and the transcontinental railroad, were completed in the final years of the 1860s. The land grants issued by the government to incentivize the construction of railroads created new ways of raising large amounts of money from investors for big infrastructure projects, and the shortened time it took for information to travel from Europe (from around 9 days to mere seconds), stabilized prices and encouraged additional investment to flow in from abroad, laying the groundwork for the Second Industrial Revolution.
Early demands for shortening the workweek
While there were significant changes afoot in the type of work that was done, most people had long been accustomed to a 60-hour workweek or more. It was typical practice to work from dawn until dusk on farms in the 19th century, and this was reflected in the normal working hours in manufacturing and other fields. Yet some dreamed for better.
The nascent labor movement of the time, inspired by the abolition of slavery was led by the National Labor Union (NLU). The NLU had begun organizing "Grand Eight Hour Leagues" across the country advancing the idea of eight hour days for workers. These leagues succeeded in getting maximum hours laws passed in many places. But little change in the length of the workweek was actually seen. Many of these laws were ultimately unenforceable, and where enforceable ones did exist, they generally stipulated 10 hour days and were limited in scope.
Why didn't we see bigger changes during this time? Another challenge faced by this early push for a shortened workweek was the erratic boom and bust cycles of the era. Nowhere is this clearer than in one of the largest industries of the time, the construction of railroads.
The prospects of early railroads excited both investors and the public, but the complexity and scale of the projects meant that the returns from such long-term investments often fell short of expectations. The Panic of 1873, which is sometimes considered the world's first truly global financial crisis, exemplified the economic challenges that persisted into the 1890s: investor confidence was frequently shaken by corrupt or insolvent companies, and the return to the gold standard set by Germany in the aftermath of the Franco-Prussian War forced most other countries to follow suit. This tightened the supply of money, leading yet more investors and individuals to hoard gold rather than spend or invest it, which ultimately lowered prices and wages everywhere. The economic conditions that this created worldwide-- often called The Long Depression-- were particularly dire in parts of Europe and a wave of immigration to the U.S. soon followed. Falling prices and these new arrivals left U.S. workers in a very difficult bargaining position, and most people took work whenever they could find it.
Nevertheless, the Knights of Labor arose from the ashes of the NLU, and led a firmer push for an eight-hour workday with a promise to stop working beyond 8 hours starting on May 1st, 1886. With slogans like “Whether you work by the piece or work by the day, decreasing the hours increases the pay”, the leaders of this movement clearly saw how a shortened workday would benefit workers beyond the additional free time. But different ideas for strategy existed between the leaders of this movement, reflecting the difficult realities of the time. The violence that transpired later in the year at Haymarket and elsewhere caused further fractures. In spite of this, targeted action in certain localities and trades led to a slight decrease in hours over the following decade which we can see in the surveys of average hours worked.
The tide starts to turn
Looking closely, we can see how average hours dropped even further in the 1900s. A closer look at the demographics of workers reveals some key shifts that were happening behind the scenes.
Changes were slow, but one of the more striking features of this chart is the number of children under 15 who were employed-- in 1880, these children were 8.2% of the overall workforce!1 Until 1900, increasing numbers of children worked full-time: either on farms or in industry. While compulsory schooling laws had existed in some states as early as the 1850s, other states, particularly those in the South, did not pass similar laws until the 1900s and 1910s, and enforcement was not always consistent. As more states passed and enforced these laws, we can see a clear jump in school attendance for all ages over the 1900s.
Taking a closer look at the amount of work done in each field reveals some other clues: the percentage of work being done on farms steadily decreased from 1870 to 1910. The biggest shift in magnitude was into work in the construction, extraction, and transportation sectors, which is exactly what one might expect from this golden age of railroads.2
However, running close behind were administration, finance, and legal jobs. The early information economy of the Industrial era required many more educated workers-- nearly 10 times as many in 1910 as in 1870 (or 3.66x as many after accounting for growth of the adult population) and this was only growing. By 1906, when a bill restricting child labor was first introduced to Congress, a survey found that only 20% of employers wished to hire children. Between local child labor and compulsory schooling laws, and the unmistakable growth in white-collar jobs requiring an education, the direction that the country was heading in was increasingly clear.
The labor market began to feel a squeeze at the other end of the age spectrum as well. The 1910 Census shows a sharp drop in the percentage of men (and to a lesser extent women) over age 65 who reported working a job:
Why were so many older people working before, and what caused them to start retiring in the 1900s? Looking back 40 years to when these people were young gives us a clue: many of them were veterans of the Civil War. And in 1907, Union Civil War pensions became available to all veterans based on age alone with no need to prove a disability. The system for distributing these pensions became a model for the Social Security system which followed less than 30 years later.
Between fewer children and seniors working, and the fastest-growing jobs requiring more education, the stage was set for some even greater changes in the following decades.
Some more details on the author’s adjustments.
Census samples from 1870-1930 include occupation details for children age 15 and under, but do not count them as part of the labor force. My adjustments here do count all children counted in 1930 and earlier who have a professional occupation (an OCC1950 code between 0 and 970) as members of the labor force.
Census samples from 1910-1970 count individuals whose listed occupation is “Farm laborers, unpaid family workers” as members of the labor force, but this occupation is not available for previous and subsequent years. An extrapolation for the number of individuals, including children, who likely performed this work is made for 1870-1900 based on the number of individuals who reported themselves as “Farmers (owners and tenants)” in each year and the ratio between the number of these individuals and unpaid family workers from 1910, and these individuals are counted as members of the labor force to more accurately capture the amount of person-hours worked on farms. A similar extrapolation is done for numbers after 1970, but this is a much less significant share of the total.
Lastly, there is a sight discrepancy between the count of members of the military between the Census and Department of Defense official figures. This is because prior to 1940, many service members reported their more specific occupation within the military (e.g. pilot, ship captain, cook, etc.). An extrapolation is done to re-categorize these individuals as members of the armed forces based on the 1940 Census, when both aspects were captured, with adjustments made based on the prevalence of each of these occupations in preceding years. Starting in 1920, service members stationed abroad or those who had recently moved may not be counted by the Census. These uncounted members are added to the labor force as if they were completely uncounted members of the population, and their ages are assumed to be similar to the sample of service members who were enumerated by the Census.
The inspiration for using IPUMS data in this way comes from Joel Elvery of the Cleveland Fed’s analysis of changes in the occupational structure from 1860 to 2015. My analysis here however is slightly different— I have aligned some jobs to other categories, and used larger Census samples when they were available (not just 1% samples), in addition to the adjustments described above.