Measuring economic growth is difficult, especially for periods for which little information is available. However, harmonized national accounts were set up in most countries after World War II, and they provide different ways to measure aggregate production, either through summing the added values of all resident production sectors, or through summing all the incomes distributed by those sectors. Based on a broad set of historical studies, Maddison (2003) reconstructed income per capita data over the past two centuries, and added some point estimates for earlier periods (in 1 Common Era (CE), 1000 CE, 1500 CE, 1600 CE and 1700 CE). Such estimates often require educated guesses on unobservable trends, but they nonetheless show the best information given what is known at one point in time.
Recently, Bolt and van Zanden, (2013) have revised and complemented Maddison’s work with the “Maddison project”), and some of the findings are striking.
Over the past millennium, income per capita in the selected countries has increased 32-fold, from 717 US dollars per person per year around the year 1000 to 23,086 dollars in 2010. This contrasts sharply with the previous millennia, when there was almost no advance in income per capita. The figure shows that it started rising and accelerating around the year 1820 and it has sustained a steady rate of increase over the last two centuries. One of the main challenges for growth theory is to understand this transition from stagnation to growth and in particular to identify the main factor(s) that triggered the take-off.
Is the finding that there was stagnation in the standard of living until 1820 truly robust? This claim is particularly important given that mankind experienced significant technological improvements that would have been expected to increase productivity and income per person, from the Neolithic revolution to the invention of the printing press.
Two facts corroborate the idea that there was indeed stagnation over the most part of human history: first, estimates of longevity computed on specific groups across time and space do not display any trend before 1700 CE. For example, De la Croix and Licandro (2015) show using a long-running database of 300,000 famous people that there was no trend in mortality during most of human history, confirming the existence of a Malthusian stagnation epoch.
Have you read?
Second, body height computed from skeletal remains does not display any trend either, while height is known to depend very much on nutrition when young (Koepke and Baten, 2005). This indicates that there was no systematic improvement in nutrition over time. One has to wait until the 19th century to observe a trend in height, as witnessed by the data of the Swedish army.
The three measures of standard of living proposed here – GDP per capita, height and lifespan – are therefore in the same direction: that of stagnation for most of human history. The economic growth that we now enjoy, with its positive effects on the standard of living but also its negative effects on the environment, is therefore an unprecedented and recent phenomenon on a historical scale.