Productivity in economics refers to the amount of labor and other inputs that are required to produce a good or service. In its simplest form, labor productivity, it measures the amount of labor required to produce something. For example, it might have taken 80 hours of labor to produce a car in 1965, and 20 hours to produce a car in 2018 implying a fourfold increase in labor productivity. Unlike the example of a sudden increase in productivity we used in The Island thought experiment, labor productivity increases typically happen incrementally over time, even if in the end they can create some of the same problems we saw in that example.
What can cause labor productivity to increase? The obvious answer is that technology improves over time. Clearly it takes more time to hand wash clothes than it does to dump them in the washing machine and press a button. It is also clear that productivity can improve through adding more capital. The more washing machines there are, the less total labor will be spent in washing clothes. Finally, the use of energy can change labor productivity. If you don’t have electricity in your house, the washing machine isn’t much help.
In general, in economics, it is assumed that there are declining returns to increasing a factor of production. If you have an automobile plant with a certain number of machines which require a certain number of workers, adding workers (or work hours) until the required number is reached will greatly increase output. But once the plant is fully staffed, additional workers won’t produce much more output. The same applies to adding more machines of the same type without adding more workers. Of course, if you add workers, machines, and energy in the proper ratio you can grow output at the same productivity level. You can, for example, simply replicate the plant to produce twice as much output. But simply increasing capital and labor in a fixed proportion using the same technology and abilities doesn’t increase productivity.
To increase productivity, you can do a number of things. Let’s suppose that at the car factory the cars are painted by hand. If you install painting robots you will reduce the amount of labor per car and increase labor productivity. This is a new type of machine and will increase the amount of capital in the plant. You might already have welding machines in the plant but as they wear out you could replace them with automated machines that cost the same but don’t require as much labor. That is a technology change that increases productivity without requiring more capital[1].
What happens to worker pay when labor productivity increases? It seems intuitive that if it takes half as many labor hours to produce a car, the factory could pay each worker more. And indeed, it could if cars sold for the same amount and other costs remained the same. But the cost of cars is, under ideal free market conditions, going to be determined by supply and demand, and, under competition, the price of cars will go down until it just covers the cost of production (including reasonable return to capital, aka profit). Of course, sales of cars at the new lower price will go up. If the number of cars sold doesn’t go up as much as the labor productivity increase, the number of workers can be reduced meaning that workers may very well have to compete for jobs. And it is the supply and demand for workers of a specific type that determines their wages, so wages for these workers may go down. By the way, with changes in technology come changes in the type of workers needed. When painting robots are installed, there will be less demand for skilled painters and more demand for robot technicians.
When productivity goes up in the auto industry, workers affected by the change may not see an increase in wages, in fact some of them may lose their jobs, but the general public will effectively get a pay raise when the price of cars comes down.
If we look across the entire economy, if there is a general increase in productivity, the same amount of goods and services can be produced in a smaller number of work hours. Since people have to spend less of their incomes on these goods, there will be increased demand and more goods and services produced including perhaps new ones not available before. The extra demand and new production can absorb the workers displaced by automation. In short, when productivity increases, and employment stays the same, national income goes up: the quantity of goods and services per person increases. That doesn’t mean that everybody benefits equally, or even at all, that depends on the vagaries of the supply and demand for different kinds of labor.
Productivity increases as mentioned can be brought about by increasing the number of machines being used by workers, which is referred to as “capital deepening”, or by increasing the efficiency of technology. To try to quantify the role of technology versus increasing capital per worker, economists developed a measure called “total factor productivity” (also called multifactor productivity) which relates inputs of labor and capital to output. The increases in output that cannot be attributed to changes in labor and capital are inferred to be due to technological change[2]. Robert Solow who introduced total factor productivity in 1957 estimated that 80% of productivity increases over the period 1909 to 1949 were due to improvements in technology, not increases in capital per worker[3]. In a developing economy there is usually a need for more capital to increase productivity. In a fully developed economy increases in productivity are mostly due to technological advances.
In Part II we will look at the data on productivity increases and who the winners and losers are, but it’s no spoiler to say that the increases have been enormous. Despite producing more output than ever before, agriculture in the US went from employing 80% of the population in the early 1800’s to less than 2% now, and manufacturing went from employing over a third of workers in the US in 1960 to 7% now. In the US costs and prices, adjusted for inflation, have fallen in agriculture and manufacturing as total factor productivity has improved[4]. In rapidly developing countries, on the other hand, most of the improvement in labor productivity is due to increased use of capital[5].
It is interesting to note that increases in wages relative to the cost of capital can drive increases in labor productivity: as wages go up, investing in more machinery to reduce labor costs makes sense. The same can happen in reverse: if labor is abundant, wages may go down relative to the cost of capital and labor productivity increases may slow. However, if machines simply become more efficient as they are replaced this connection is less important, but still should be kept in mind when explaining productivity changes over time.
[1] There are many other ways productivity can be improved. For example, a large factory may be more able to use expensive machines than a smaller one. That is called “economy of scale”. Increasing worker retention can improve productivity since experienced workers can be more efficient. There can be synergies between businesses that are concentrated in a region (think Silicon Valley). Businesses are always striving to improve productivity.
[2] total factor productivity, is based on a model called a production function which is a highly simplified mathematical relationship between labor, capital, and output. Using multiple actual data observations on labor, capital, and output, it is possible to estimate how much of the growth in output is due to growth in labor and capital and an upward trend in overall output not attributable to either. That upward trend is the increase in productivity attributable to technology change. Economists have often refined MFP calculation to include intermediate products, human capital and energy.
[3] “Robert Solow’s Nobel Acceptance Speech 1987.” n.d. Accessed August 10, 2021. https://www.nobelprize.org/prizes/economic-sciences/1987/solow/lecture/.
[4] In recent years, perhaps surprisingly, capital deepening is responsible for more of the labor productivity improvement in the US. See Figure WW115.
[5]In developing economies where a lot of new capital is needed and wages are low, governments may intervene to capitalize some industries, for example steel production, even though doing so doesn’t reduce costs. The alternative free market process involves investors investing in labor intensive industries in low wage countries, eventually driving up wages and thus making capital investments in other industries more profitable. China has used both strategies.