Population Growth and Technological Change — One Million B.C. to 1990 — Michael Kremer

AuthorAthenarium / Posted on

Population Growth and Technological Change — One Million B.C. to 1990 — Michael Kremer

The population-technology dynamic

One salient feature in the long run of human history is the relationship between population level, population growth, and technological growth. You can imagine, for example, how limitations in farming technology might constrain the populations of pre-industrial farming societies. A smaller population might imply fewer inventors and commercial exchange, inhibiting the rate of technological progress in a chicken-and-egg way.

So, if we expect a society’s population to influence its technology growth, and its technology to influence its population levels, then we would expect to see some relationship between population levels and population growth in historical data (see Figure 1 and Figure 2 below). Indeed, this is the very model and empirical test that Michael Kremer explores in his famous paper Population Growth and Technological Change.

Figure 1. Population growth — 1,000,000 B.C. to Present
Figure 1. Population growth — 1,000,000 B.C. to Present
Source: Adapted from Michael Kremer. (1993). Population Growth and Technological Change — One Million B.C. to 1990; and World Bank World Population Totals.
Figure 2. Average World GDP per capita — 1,000,000 B.C. to Present
Figure 2. Average World GDP per capita — 1,000,000 B.C. to Present
Source: Adapted from ‘preferred series’ in J. Bradford De Long. (1998). Estimates of World GDP, One Million B.C. — Present. 

Nonrivalrous nature of technological change

Kremer argues that innovation tends to rise with population levels because of the “nonrivalrous” nature of technology. Norms and culture permitting, people tend to replicate and adapt good ideas once discovered. Kremer says that larger populations can help by way of greater “intellectual contact”, “specialization”, and “size of the market” — more throws at the innovation dartboard, if you will.

People often perceive the times before industrialization as an economic and technological backwater. Many macroeconomic models, likewise, focus on “discontinuous breaks” or “multiple equilibria” to describe global development. What Kremer sees instead is a slow, haphazard process of gradualism, in which take-off is just around the corner.

We know that while pre-industrial aggregate statistics look quiet, or stagnant even, humanity has been accumulating knowledge for a very long-time. And as Kremer writes of the population-technology dynamic:

“Through most of history the growth rate of world population has been approximately proportional to the level of population. Moreover, among societies with no opportunity for technological contact, those with greater initial population attained higher technology levels and population densities. These facts are difficult to reconcile with prevailing growth models in which technological change is independent of population.”

Michael Kremer. (1993). Population Growth and Technological Change — One Million B.C. to 1990. 

A natural experiment

Towards the close of the Last Glacial Period, much of humanity consisted of hunter-gatherer tribes and societies. This point, Kremer notes, provides us with a “natural experiment”, as the end of the last Ice Age around 12,000 years ago separated the Old World, the Americas, and Australia for some time. As Kremer describes, several millennia later:

“[By] 1500, just after Columbus’ voyage reestablished technological contact, the region with the greatest land area, the Old World, had the highest technological level. The Americas followed, with the agriculture, cities, and elaborate calendars of the Aztec and Mayan civilizations. Mainland Australia was third, with a population of hunters and gatherers. Tasmania, an island slightly smaller than Ireland, lacked even such mainland Australian technologies as the boomerang, fire-making, … and bone tools, such as needles.

Michael Kremer. (1993). Population Growth and Technological Change — One Million B.C. to 1990. 

For a regional example, Kremer notes that “when the land bridge between ancient Britain and Europe was cut off, around 5500 B.C., Britain fell technologically behind Europe”. To cut a very long, complex story short, more land and food means more people. This results in more trade, invention and food; which leads to more people and so on over the long run.

I am, of course, grossly oversimplifying the matter. Many economic, political, technological, environmental and contingent factors are at work. But I hope this illustrates, in brief, one of the autocatalytic processes that seems to shape the course of human development.

The diversity of life

These relations are not dissimilar to the phenomenon we find in biology. As Edward Wilson describes in The Diversity of Life, biodiversity, for instance, depends partially on habitat size. It is a proxy for the availability of resources, environmental heterogeneity, and potential niches for species to fill.

The population growth of a species or community, likewise, is a function of its current population level and the population of its prey and predators. Overpopulation, for example, may lead to decline as the species competes for a dwindling pool of prey, while fuelling a boom in predator populations.

Large populations, however, may not constrain human societies in the same way it does to other animals. But it assumes that the returns to scale and productivity from invention allows us to transcend our environmental limits. This is not a guarantee, no matter how smart, hard-working and well-intentioned we think we are. 

Junctures, errors, and circuit breakers

Indeed, history is far from a smooth geometric process. Kremer’s empirical tests of population growth relations, for example, are statistically insignificant during outlier events that decimated human numbers. This includes the Black Death (1300s), Mongol invasions (1200-1300s), The Thirty Years’ War (1600s), and the Ming Dynasty’s collapse (1600s). Long run population estimates, likewise, are also sensitive to measurement error, unmodelled shocks, and systematic bias.

What’s more, complex, positive feedback loops may contain their own circuit breakers. Fertility rates, for example, have fallen over the last century. This is due in part to many social and economic forces, from rising income levels to improved access to contraception and higher labor force participation rates among women. Few economists, I suspect, could have predicted this long before the fact. The same applies to us today in regard to a fuzzy future.

Occam’s Razor

While Kremer acknowledges that it might be “possible to explain the [population-technology] data through some other variable”, he wonders “why one would want to [seek] an alternative explanation”, “given that a simple model, based on economic theory of technology as a nonrival good, is consistent with the data over such a long period” of time. 

Herbert Simon, however, reminds us in Rational Decision-Making that Occam’s Razor is double-edged when there are trade-offs between model fit and practical description. Kremer describes to us only the approximate factors behind human change. Curious folks want to understand the mechanisms that give rise to such phenomena, both in the short and long run.

Guns, germs, and steel

As Jared Diamond writes in Guns, Germs, and Steel, Eurasia got its head start in part because agriculture emerged first in the Fertile Crescent — providing an early runway for population growth, trade networks, institution building, and technological progress to follow.

Mainland Australia, by contrast, had no major domesticable crops or animals until European colonials arrived with foreign species. This is one of the starkest examples, I think, of sensitivity to initial conditions. Differences in natural endowment greatly altered our continental trajectories.

If we rearranged the starting conditions, from populations to endowments to topography, might we predict the same results? Indeed, population growth and technological change are likely to remain interrelated. But in the example above, even if Australia had a head start in population size following the last Ice Age, it might still lag behind over the long run if it could not transition towards food production.

We might think about markets and institutions similarly, too. Are they a naturally emerging product of time and size? Would humanity have stumbled upon modern day democracy sooner or later, or not at all, if the initial conditions were slightly different? With a small sample size and only one global history, it is not so easy to say.

Finite resources and externalities

Finally, while the population-technological change dynamic is compelling, we should not take this as evidence in support of policies that seek population growth. As Kremer notes, it does not account for “exhaustible natural resources” and other negative externalities that might harm our livelihoods.

Indeed, around two hundred years ago, economists like Thomas Malthus were troubled that exploding populations would someday outpace growth in food production. Today, we contend with our insatiable appetite for trinkets and the ultimate burden it puts on Mother Earth. Kremer’s paper, I think, puts much of this in a fascinating, long-range light. 

Sources and further reading

Recent posts