By Jonathan Movroydis
Stephen Haber is the Peter and Helen Bing Senior Fellow at the Hoover Institution and the A. A. Jeanne Welch Milligan Professor at Stanford University.
In this interview, Haber describes the extensive research behind his recent paper “The Ecological Origins of Economic and Political Systems.”
This ten-year effort began with his interest in a series of long-standing puzzles in the social sciences. Why do wealthy countries tend to be stable democracies? Why do levels of per capita income and democracy also correlate with a range of other development indicators, such as investments in human capital, the strength of intellectual property systems, freedom from government expropriation, and financial development? Why are the prosperous democracies not randomly distributed across the planet but, rather, geographically clustered? Why did these patterns only emerge over the past two centuries?
Haber and his coauthors, Roy Elis and Jordan Horrillo, offer a theory that explains these puzzles and then provide a battery of tests designed to falsify it. The core of the theory is that societies make costly investments in legal systems, political institutions, human capital, and the like when there is a return to doing so—that is, when there is a fundamental challenge that must be addressed. Until the mid-nineteenth century, when the combination of railroads, steamships, and refrigeration made it economically feasible to move staple foods over long distances, the fundamental challenge facing all societies was insuring against starvation under the constraints imposed by local climates and geographies. In a world in which starvation always loomed, and in which solutions had to be found close at hand, human beings experimented. Some of the things they did—the choice of one crop over another, a deal struck with a neighbor, an appeal to a powerful individual—kept hunger away until the next harvest. People could not, however, experiment just any way they liked—at least not if they wanted to live to see the next harvest. Their choices were bound by basic factors rooted in climate and geography: for instance, what could be grown, how much of it could be grown, how long could it be stored, how far could it be moved, and how frequently would its production fail because of severe weather events? The hard constraints imposed by nature yielded very different incentives toward the investments listed above. The result was the formation of very different forms of social organization.
Haber, Elis, and Horrillo refer to the resulting forms of social organization as “transactional,” “risk-pooling,” “self-sufficient,” and “pastoral” ecologies, in which an ecology is understood to mean the physical environment, the living organisms, and the social adaptations made by human beings to survive. Transactional ecologies emerged where factor endowments (the resources a country has at its disposal) favored insuring against starvation through the trade of easy-to-store—and hence easy-to-trade—low-moisture annual crops. Their core characteristic was decentralized decision making coordinated through markets. Risk-pooling ecologies emerged where it was easy to grow and store low-moisture annual crops, but where spatially and temporally correlated droughts ruled out local trade as an insurance mechanism, instead favoring forced sharing. Their core characteristic was centralized decision making. Self-sufficient ecologies emerged where factor endowments favored insuring against starvation through the production and consumption of difficult-to-store—and hence difficult-to-trade—high-moisture perennial crops. Their core characteristic was decentralized, uncoordinated decision making by households. Pastoral ecologies emerged where factor endowments precluded crop production of any type and instead favored insuring against starvation through the herding of large herbivores that could convert the cellulose in wild grass into meat and milk. Their core characteristic was decentralized, uncoordinated decision making by mobile bands. Haber, Elis, and Horrillo stress that these ecologies should not be thought of as hierarchically ordered; rather each was an effective solution to the problem of insuring against starvation under constraints imposed by nature.
They hypothesize, however, that these ecologies were differentially suited to addressing a new great challenge of survival that emerged in the nineteenth century: the need to rapidly absorb the broad suite of mutually dependent legal, financial, educational, organizational, political, mechanical, chemical, and electrical technologies that historians refer to as modernity. Societies that were able to absorb the new technologies quickly built fast-growing economies and powerful nation-states, capable of conducting industrialized warfare. Societies that were unable to do so were vulnerable to being colonized, dominated, or absorbed by those that had moved more quickly. Modernity was, in short, characterized by vicious geopolitical competition—and transactional ecologies were better positioned than the others to absorb its technologies, because absorption was coordinated through decentralized markets.
Finally, Haber, Elis, and Horrillo hypothesize that, because local factor endowments conditioned pre-1800 forms of social organization and because those forms of social organization conditioned rates of technological absorption during the nineteenth and twentieth centuries, the distributions of economic development, human capital, democratic consolidation, and other indicia of prosperity that we observe across countries today are strongly correlated with those factor endowments. Geography and climate mattered for long-run development but indirectly, by shaping the organization of societies.
Haber, Elis, and Horrillo then test their three hypotheses by developing geospatial datasets that approximate, as closely as they can, the ecological characteristics that bound the production, transport, and storage of calorie- and nutrient-dense foods circa 1800 in the densely populated nuclei from which nation-states later emerged. They find that a vector of exogenous factors—which were binding constraints on food production, transport, and storage within these predecessor societies—account for roughly two-thirds of the cross-country variance in per capita GDP today. Importantly, these factors account for progressively less of the variance in economic development going back in time; before 1800 they account for almost none of it. They also find that countries that emerged from transactional ecologies are associated with faster economic growth from 1800 to 2000, faster democratic consolidation from 1850 to 2017, faster development of markets from 1500 to 1800, higher investment in trade-related human capital circa 1800, faster absorption of nineteenth-century technologies, and a lower probability of being colonized.
Will you talk about the origins of this project?
Stephen Haber: I have been interested in the question of why there are rich countries and poor countries since I was an undergraduate. I became involved with this line of research about ten years ago, following a coauthored paper I published with Victor Menaldo, a former graduate student who is now on the faculty at the University of Washington. Our paper, “Do Natural Resources Fuel Authoritarianism? A Reappraisal of the Resource Curse,” countered a lot of academic literature arguing that an abundance of oil and gas in an economy distorts incentives and therefore discourages democratization and economic development. We demonstrated that this argument was a fallacy based on errors in thinking and in statistical inference.
This prompted a colleague of mine to ask: “If it isn’t oil, then what accounts for the variance in economic development across the globe?” I had no ready answer to the question, and so I started reading about countries that have been poor for a very long time. What I found was that these countries’ climates tended to be very wet or very dry. This led me down the path of studying the long-run effects of climates on how human beings prior to the nineteenth century organized themselves to survive against starvation. The dynamics of each form of social organization in turn ultimately impacted the way the populations living in each environment absorbed modern technologies and political and economic innovations, such as railroads and steamships, as well as universal suffrage and patent systems.
A large part of this ten-year effort was understanding the conditions under which different crops grow, the conditions under which they can be stored, the distance crops could be transported given eighteenth-century technologies, and the like. It also required that that we master Geographic Information Systems (GIS) in order to recreate the conditions of the eighteenth century, and that we learn the machine learning methods necessary to carry out statistical analyses. We were fortunate in that we were able to consult experts in several disciplines across Stanford—including biologists, economists, and political scientists. Having access to these resources is one of the advantages of Hoover being located on the campus of a major university. I don’t think this type of study would have been possible if I were based at another think tank.
Can you describe your methodology?
Stephen Haber: My team and I constructed two big datasets to test the theory. One dataset was designed to re-create the ecological conditions of the densely populated nuclei from which modern nation-states later emerged. Think about modern Germany for example. Until 1871, it was a congeries of more than three hundred independent principalities, duchies, free republican cities, and kingdoms. Those entities were politically and culturally subordinated to the region around Berlin, the Margraviate of Brandenburg, over the course of the seventeenth, eighteenth, and nineteenth centuries. If you look across the world, you’ll see that all modern states emerge directly from a civilization’s largest and most powerful city around the beginning of the nineteenth century.
We then used geographic information systems to estimate the size and ecological characteristics of these densely populated nuclei of today’s nation-states in 1800. One can think of it as using Google Maps, but instead of measuring travel time by car, train, or bus we calculated energy expended using three eighteenth-century transportation technologies: a Conestoga wagon, the boat that Lewis and Clark rowed and poled up the Missouri River, and a human porter using a tumpline. We had to work out the physics of how much energy it took to move a ton of bulk goods using those technologies over flat terrain or navigable water.
We also had to factor in that these regions were not flat. So, we estimated the size and shape of what are called hinterlands, the regions surrounding the largest cities, in 1800. Our team then applied our GIS techniques to datasets from the National Oceanic and Atmospheric Administration and the Food and Agricultural Organization of the United Nations so that we could calculate the quantity of twenty-two different crops that could be grown in these hinterlands and how long they could be stored.
We also looked at how frequently agriculture in each hinterland could be devastated by weather shocks, mainly droughts, as well as how they could be impacted by endemic malaria and Tsetse flies, which kill horses and oxen. Ultimately, we tried to recreate as closely as we could the conditions that human beings faced when producing and storing calories needed for survival.
A large factor that drives much of the variance in hinterland sizes is access to rivers. The rivers in today’s world are not what they were in 1800 because humans, since then, have dynamited rapids, built canals around waterfalls, and dredged sandbars. This meant that we had to reconstruct the world’s rivers as they ran before 1800. This effort involved extensive work by roughly thirty research assistants, who acquired this information from reading primary sources written by eighteenth- and nineteenth-century travelers. We then entered this historical data into our GIS-based model.
One of the main questions we needed to answer was, how long could low-moisture, annual crops be stored in these hinterlands? We calculated this through a two-step algorithm, which required finding out the temperature and moisture content before and after the food was placed in storage. Throughout this process, we had to immerse ourselves in the literature of crop storage.
Will you describe some of the ecologies that emerged just prior to 1800?
Stephen Haber: Over the last five centuries, humans have faced two big challenges. One was the challenge of insuring against starvation, which humans faced up until the advent of railroads and steamships because it was very expensive to move food over long distances. There was of course trade in all kinds of luxury goods, but moving staple crops was cost prohibitive until the nineteenth century. That meant that societies had to solve the problem of preventing starvation based on their local environmental characteristics. For example, the way in which humans survive in a rain forest is very different than in a desert.
What we demonstrate in the paper is that there were basically four forms of social organization that emerged from various groups of human beings adapting to their respective environments. By adaptation, we mean cultural adaptation—the way people behave and plan and expect others to behave and plan, not adaptation in the Darwinian sense. Human beings are the same everywhere.
The first type of ecology we discuss in the paper is called “transactional.” An example of this that I like to use is the Mid-Atlantic United States. It is a place where crops ripen only seasonally, are rarely destroyed by weather shocks such as droughts and floods, and can be stored for a long period of time. Stored food in the Mid-Atlantic could also easily be transported, because the terrain is mostly flat and the rivers are navigable. What emerged was an ecosystem where people prevented against starvation by trading with one another.
If you look at the geography of the Mid-Atlantic even in 1800 you’ll notice that there are many small cities and large towns. These were loci of trade, places where farmers traded with one another to mitigate idiosyncratic droughts and other production shortfalls.
One of the interesting features about the colonial history of the Mid-Atlantic United States is the absence of any attempt at central planning or centralized administration. Rather, independent economic agents made decisions coordinated by the market. The lack of centralization translated into a political culture that pushed back against the British elites who governed the colonies starting in the seventeenth century. It’s no wonder that the US was the first colony in the world to throw off colonialism.
Transactional ecologies work from the bottom up. That bottom-up nature meant that when confronted by the next great challenge—the shock of modernity in the nineteenth century—the Mid-Atlantic states not only quickly absorbed technologies that were developed elsewhere but even invented some of their own that other societies had to absorb. This was more than just industrialization and railroads. Societies had to figure out how to absorb a broad suite of advancements in law, finance, military planning, university-based research, and production technologies.
In contrast to the climate and geography of the Mid-Atlantic American colonies that fostered transactional ecologies were places like northern China, where weather shocks are large, severe, widespread, and lasted for very long periods of time. Ecologies such as that of northern China were proficient at growing low-moisture, annual crops that are very high in calories. But because of their aggregate weather shocks, they needed to prevent starvation by constructing compulsory insurance systems. These have taken various forms across the globe. In Mexico, it took the form of the colonial hacienda, and in China, a state-run granary system.
This was not a bottom-up solution that relied on markets. This was a top-down solution in which everyone was forced to participate. Such a system more naturally translates into centralized political authority. It was much more difficult for these “risk-pooling” ecologies to absorb modern technologies in the nineteenth century. One consequence was that countries that emerged from risk-pooling ecologies tended to become dominated by countries that emerged from transactional ecologies. A classic example of this is China’s defeat in the Opium Wars.
Following these conflicts, China realized it needed to catch up with the West. But it could not do so by relying on markets. Rather, it relied on a centralized state-administered program known as the “Self-Strengthening Movement.” An example of how badly this worked was the way it addressed investing in human capital. The Qing Dynasty sent 120 students to the United States to study in American schools. After three years, the political leadership brought the students back to China because they were learning “bad ideas” about democracy.
“Self-sufficient” ecologies emerged in regions where it was too rainy to grow most low-moisture annual crops and where it was too hot and humid to store those that could be grown. People therefore insured against starvation by growing high-moisture perennial crops that, while difficult to store, ripen year-round or can be left in the ground until they are needed. This meant, however, that there was little to trade and tax. The Congo is a quintessential example of these conditions. Human beings certainly survived, and even thrived. But because there was neither a market that could coordinate a response to modernity from the bottom up, nor a centralizing authority that could engineer a response from the top down, it was extremely difficult to absorb the technologies of the modern world rapidly and as a broad suite.
A somewhat similar situation emerged in what we call “pastoral” ecologies, places where it was too dry and cold to grow crops of any type, but where the conditions favored grasslands. In this kind of environment, people herded herbivores. These historically low-population ecologies have weak decentralized states because it’s very hard to tax mobile animals.
Both subsistence and pastoral ecologies fell behind in the technological race of the nineteenth and twentieth centuries and were usually overrun by risk-pooling or transactional ecologies.
Modern Russia is a powerful state that covers a considerable amount of the world’s landmass. What type of ecology emerged there?
Stephen Haber: Modern Russia emerged from the area around Moscow, which has two ecological characteristics: the growing seasons are very short; and the propensity for weather shocks that wipe out crops is very high. As in northern China, the dominant form of social ecology that emerged in Russia was risk-pooling.
People in eighteenth-century Russia were not independent farmers trading in urban markets, such as in northeastern colonial America. They were dependent peasants. The landlords would provide food to the peasantry during weather shocks, but the rest of the time they extracted every kopek from the peasants.
Do these risk-pooling societies ultimately absorb modern technologies at some point?
Stephen Haber: Centralized political systems tended to engineer the absorption of new technologies from the top down. When these societies absorb technologies, they do so in fits and starts, and ultimately fall behind. The further they fall behind, the harder it is to successfully transition to a modernized economy.
Is this research predictive of the future? Does it have an application to public policy?
Stephen Haber: The goal of the research was to understand why things are the way they are today, not to predict the future. I want to be very clear here. We are not claiming that we can explain 100 percent of the variation in levels of development based on our ecological variables from the past.
What we’re saying is that we can account for about 60 percent of the variance. There is another 40 percent of the variance that is the result of human agency, idiosyncratic events, and other factors outside of our model. That said, explaining 60 percent of the variance in levels of development around the globe today is far more than anybody else has been able to achieve in this literature.
Do we think that this framework predicts the future? I want to be very careful here. The great challenge facing societies now is not that of insuring against starvation. It’s now possible to move food around the globe at very low costs and at rapid speeds. I don’t want anyone to think that this is a framework that predicts the future based on the organization of agriculture in the past.
However, I think there are insights that can be drawn from this research of the past about the questions that societies face today—mainly how ecological factors impacted the way people organized themselves and, in turn, how they adapted to their next great challenge.
One could glean insights, for example, on the challenges the United States faces with a rising China. This is essentially a contest between a decentralized market-oriented system versus a politically centralized system. The way in which American society is organized is going to affect how quickly we can make the next technological leap in areas of artificial intelligence, digitization, quantum computing, and the like. There’s an open question as to whether we or the Chinese are going to be better at absorbing those new technologies.