In February 1956, US economist Robert Solow published a model to understand economic growth, which was followed by Australian economist Trevor Swan in November that year. The nine-month delay and Swan’s inhibitions led the model to be dubbed as the Solow model. On its 70th anniversary, it remains central to current debates around the impact of artificial intelligence (AI) on the global economy. It is a remarkable coincidence that the term AI was coined by the computer scientist, John McCarthy, in 1956.
The question of growth has occupied attention of most economists since Adam Smith published a book in 1776 with the title An Inquiry into the Nature and Causes of the Wealth of Nations. Smith’s wealth of nations was driven by division of labour and freer markets. In 1798, Thomas Robert Malthus’s pessimistic prognosis overshadowed everything. Malthus opined that as land is limited, growth in agriculture will be lesser than growth in population, leading to widespread hunger and poverty.
Malthusian thought, even though highly despised, was not wrong as there was barely any growth for most part of human history. Growth engines started sputtering after the industrial revolution in the 18th century and gained pace in the 19th century. High growth led to the rise of large industries, and it was characterised by unequal growth. The debates shifted to Marxism which argued that large-scale-driven capitalism is doomed to fail. Marx was later joined by Joseph Schumpeter who argued that creative destruction will destroy capitalism.
The above mentioned growth descriptions were based on a grand narrative combining history, politics, society, and economics.
In the inter-war period, the narrative approach gradually gave way to building mathematical models to estimate growth. In the 1930s, economist Simon Kuznets developed a measure to estimate national income. After World War II, Keynesian economics became central to economic policy with questions of reviving growth becoming central. Paul Samuelson had joined the Massachusetts Institute of Technology in 1940 and was writing textbooks based on Keynesian economics using the mathematical approach.
The first growth model was developed by two economists, Roy Harrod and Evsey Domar, independently of each other in 1939 and 1946 respectively. The model based on Keynesian economics links economic growth directly to savings and capital. It states that the rate of growth depends on the savings ratio divided by the capital-output ratio.
The 1956 Solow model built on Harrod-Domar by adding that like capital, labour is also an important factor of growth. So, if a country had higher labour resources it could substitute labour for capital and vice versa. However unlike Harrod-Domar, the Solow model says that capital accumulation faces diminishing returns. So by simply adding more capital a country cannot generate higher growth. Instead, a country has to make technological progress — a major breakthrough for the model. Until then, while it was understood that technological progress was important for growth, it was not formally a part of the economists’ toolkit. Solow was duly awarded the Nobel Prize in economics in 1987. The model gave three major insights on the wealth of nations.
First, capital accumulation cannot sustain long-term growth because of diminishing returns. As the capital grows, new investment replaces depreciated capital until the economy reaches a steady state where capital per worker stops growing. So higher savings and investments can lead to higher income but not long-term growth.
Second, to break out of the steady state, the economy needs a new technological shock. The model did not explain where the technology comes from and assumed it is an exogenous factor.
Third, the model explains that if both poor and rich countries deploy the same amount of capital, it is the poor ones that will grow faster than the rich nations. This is because rich countries face diminishing returns on capital accumulation whereas poor countries do not face diminishing returns at the start of their growth journey. It also cautions that rich countries which do not strive for technological progress will stagnate, as is the case with most developed countries.
Future economists worked on the limitation of the Solow model — assuming technological shock is manna from heaven. In the 1980s, Paul Romer (2018 Nobel Prize winner) said growth is driven by internal factors such as education and research and development. According to him, even if capital and land diminish, new ideas to organise economic activity remains unlimited. Robert Lucas, the 1995 Nobel Laureate, emphasised the importance of human capital in an economy’s growth, enriching the models. William Nordhaus, a co-winner of the Nobel with Romer, built on the Solow model to add environmental costs of production. The history of growth models from a depressing Malthus to cheering Solow is, in essence, the economic history of humanity as well.
To conclude, how does the Solow model help us assess the impact of AI? Current discussions hover around whether AI is a technology, a capital-augmentation investment, or a labour-replacement shock. Once debates settle which of the three is better suited to AI, we can think about the impact of AI on economic growth. As of now, we find AI everywhere but not in growth. To paraphrase Solow’s famous quip on computers in 1987, “You can see the AI age everywhere but in the productivity statistics.”
The author teaches at the National Institute of Securities Markets
Disclaimer: The views expressed are the author’s own and do not reflect the official policy or position of Financial Express.