Location, Location, Location: The Persistence of Fortune

Yves here. This intriguing paper curiously fails to consider how climate change can (will) affect the patterns they describe. And it also fails to give clues as to why some cities keep gaining population and income and others fall by the wayside. For instance, Atlanta and Birmingham, Alabama were of comparable size in the 1970s, yet Atlanta has grown while Birmingham stagnated. And Birmingham has some advantages, for instance, having the best medical school in the South.

By William F. Maloney,Chief Economist for Equitable Growth, Finance and Institutions, World Bank Group, and Felipe Valencia Caicedo, Assistant Professor, Department of Economics, Bonn University. Originally published at VosEU

e persistence of economic fortune over the long run has been the subject of intense research. This column investigates the persistence of patterns of economic activity in the Americas at the sub-national level over the last half millennium. The location of today’s prosperous cities and regions within each country is closely correlated with the location of indigenous population centres before the arrival of Christopher Columbus in 1492. Policymakers seeking to make radical changes in the spatial distribution of economic activity should be mindful of the centuries-old, even pre-colonial, forces working against them.

Does economic fortune persist in the long run? This question has been the object of intense research. While some studies find a reversal at the national level for colonised countries (Acemoglu et al. 2002), others show that the geographical patterns of population and prosperity have largely persisted even for thousands of years (Comin et al. 2010, Ashraf and Galor 2013, Davis and Weinstein 2002, Chanda et al. 2014).

The question of persistence is important as it speaks to the role of geography, agglomeration, institutions, and other deep-rooted factors of comparative economic development. It also contributes to recent discussions about the role of pre-colonial versus colonial determinants of economic development (for the African case, see Gennaioli and Rainer 2007, Michalopoulos and Papaioannou 2013). Ultimately, this type of research helps to inform the debate about path dependence and place-based policies (see, for example, Ades and Glaeser 1995, Bleakley and Lin 2002). In our research, we provide new evidence of persistence from pre-colonial times using data at the sub-national level for the Americas (Maloney and Valencia Caicedo 2016).

For modern-day policymakers, the reality of such long-term continuity implies first, that there may not be anything ‘optimal’ about the present distribution of economic activity – agglomeration effects may simply be driven by economic considerations millennia old. Second, this same consideration may prove challenging to any grand-scale efforts to alter long-existing spatial and urban patterns.


We focus on the Americas, and follow the literature in taking the continent as a large historical natural experiment. We are also aided by the availability of precise anthropological and archaeological data, the product of a long tradition of research summarised in William Denevan’s 1992 book and used recently by Bruhn and Gallego (2012). These historical measures of population density can be seen as what they are, but also, following Malthus, as a proxy for prosperity.

We test for the continued influence of this variable as we control for standard geographical and weather controls, suitability for agriculture, river network density, ruggedness, and prevalence of malaria. We use country fixed effects to net out the effect of national level variables such as policies and institutions, as well as different estimators to assess the potential role of outliers.

Empirical Findings

We reach three main conclusions in our research. First, and most surprisingly, we document that settlements today are located where there used to be pre-colonial settlements. This persistence is far from obvious given that in many cases the original indigenous populations were practically eradicated during the colonisation process.

We also show that certain geographical and weather characteristics – most notably agricultural suitability, distance to the coast and malaria prevalence – were important in determining where indigenous people settled in the first place.

Lastly, we show that places that used to be rich also tend to be richer today. Massachusetts and California, for example, had the highest density of native Americans – roughly the same average level as Chile and Argentina – in the territory that would become the United States, and they are among the richest regions today.

That is largely true too among Spain and Portugal’s colonies in the hemisphere. In Colombia, there is a strong correlation of pre-colonial population density and both population and income per capita (see Figure 1). Mexico City, the richest city in Latin America, was founded on top of Tenochtitlan, the ancient capital of the Aztecs, itself the most densely populated pre-colonial area in the hemisphere.

Figure 1 Sub-national persistence in Colombia


Potential Mechanisms

The persistence of demographic and economic activity could work through different channels. Geographical determinants aside, agglomeration effects arise from path dependence à la Krugman. This in turn might be the product of externalities and complementarities in production, as hypothesised by Marshall.

Geographical factors appear to be important in explaining initial indigenous settlements. California, for example, had and retains rich farmland, a productive coast, and a temperate climate. But ‘locational fundamentals’ are not enough to explain the patterns we see; controlling for a rich set of geographical fundamentals does not eliminate the importance of pre-colonial density for present population density and income.

Further, history provides important counterexamples. The Aztec capital Tenochtitlan was founded on a small, swampy island whose chief attraction appears to be that it was not coveted by the neighbouring tribes. Its weather was unreliable, its growing season was short and famine was not uncommon. Moreover, frequent flooding inundated the city with its own filth, breeding epidemics. But in founding Mexico City on top of Tenochtitlan, Hernan Cortés was clear about what was attractive:

This great city of Tenochtitlán is as big as Seville or Córdoba […] It has many plazas where commerce abounds, one of which is twice as large as the city of Salamanca […] and where there are usually more than 60,000 souls buying and selling every type of merchandise from every land […] There are as many as forty towers, all of which are so high that in the case of the largest there are fifty steps leading up to the main part of it and the most important of these towers is higher than that of the cathedral of Seville. The quality of their construction, both in masonry and woodwork, is unsurpassed anywhere. (La Gran Tenochtitlán, Segunda Carta de Relación, 1522).

In short, he valued not only the available labour force, but also the area’s pool of skilled artisans, dense commercial networks, organisational structures, and collective knowledge – attributes that we associate with dynamic cities and dense population agglomerations. Similar phenomena are present in North America, where British, French, and Dutch explorers used indigenous knowledge, maps and trade routes to survive and establish colonies, trading posts, and fur-trading networks.

Discussion and Conclusion

The finding of persistence is also somewhat surprising, as other research has stressed a ‘reversal of fortune’ at the national level – a finding that we replicate in the aggregate for the Americas. This reversal is attributed to the emergence of growth-inhibiting exclusionary institutions arising from the need to control larger indigenous populations. Although we find some evidence for this – we confirm that slavery has a depressive effect on current social outcomes – the forces leading to persistence appear to dominate at the sub-national level. We do not find, for example, any effect of pre-colonial density on income inequality.

Argentina offers an interesting exception to the pattern of sub-national persistence. Its reversal was mostly due to the removal of royal prohibitions on trade with Spain through the Atlantic port of Buenos Aires. That policy change meant that goods no longer had to be hauled over the Andes to Lima, Peru on the Pacific Ocean, and then back overland to the Atlantic Ocean. Releasing this once-suppressed geographical comparative advantage was a powerful shock to geographical fundamentals.

But this seems to be the exception to the rule, and policymakers attempting to make radical changes in the spatial distribution of economic activity should be mindful of the centuries-old, even pre-colonial, forces working against them.

See original post for references

Print Friendly, PDF & Email


  1. Disturbed Voter

    There will always be people and circumstances resisting the usual process. What is that? That the capital city will become gigantic and suck all the resources and money out of the provinces. As true today in Mexico City, as in ancient Rome. How does this analysis challenge the process whereby the people with political power will exploit the hinterlands? How does this analysis challenge the process where it takes money to make money, so small differences in economic advantage, cumulate over time, so that all the small towns in each US state are shrinking, in favor of the state capitals? That “sucking sound” of Ross Perot wasn’t just between countries, but within. But it isn’t the only game in town, some locations are superior to the state capitals … and so provide natural barriers to entry of capital city exploitation.

  2. DJG

    As always when the so-called experts write about “Latin America,” Brazil drops into a black hole of information. So far as I know, Sao Paulo and Rio de Janeiro were not founded on major pre-Columbian settlements. So this article appears to be an analysis of Mexico City, and that’s about it.

    Last night, I finished The Masters of Empire, about the People of the Three Fires, especially the Odawa, who dominated Lakes Superior, Huron, and Michigan for centuries. The book only partially supports the hypothesis of this article. The major population centers of the Odawa were towns along the Straits of Mackinac.

    So I remain unpersuaded, and I do hope that the authors can get out in the world more often.

    1. Gio Bruno

      Yes. Many folks think America was empty of native people when the white folks arrived. Ain’t so.

      Not only were the Great Lakes region heavily populated, but the area around St’ Louis (now known as Cahokia (Mounds) State Park) included a city with a population larger than London, England (at the same period in history). While California had a relatively large native population (300,000 pre-Columbian), the whole of the US lower 48 is estimated by some to have reached 20 million natives.

      The article seems to have a Colonialist bias. Try reading “One Vast Winter Count” for a better view. And J. Diamonds “Guns, Germs, and Steel” on how North America appeared “empty” to post-Columbian settlers. (Small bullets and small pox disappeared the natives rapidly.)

      1. Steve Gunderson

        Small Pox did remove a lot of the natives. There are accounts from Cortez’s men describing bodies of Aztecs heaped on the sides of the road.

        Small Pox may have killed up to 90% of the natives.

  3. Carolinian

    Just a quibble: I’m a bit dubious that Birmingham was ever a true rival to Atlanta. Back in the late 70s I was told by a transplanted Brit that the Atlanta metropolitan region was the same size as London. The Atlanta city population is quite small because of constricted city boundaries. Atlanta is one of the original “edge cities.” Most of the action is in the suburbs. Indeed the Braves have decided to move their stadium there.

    Also while Atlanta has always enjoyed the advantage of location–a transportation hub–it’s also quite likely that Birmingham was held back because Alabama, along with MIssissippi, was the beating heart of segregation. This doesn’t mean that Georgia wasn’t also thoroughly Jim Crow–as Howard Zinn describes–but the New South corridor running from Atlanta up to Charlotte was in many ways in the vanguard of Southern businessmen deciding racism was bad for business.

    1. Steve Gunderson

      I believe up until the 1950’s, states like Alabama worked to keep factories out so that they would have a cheap workforce work the crops.

    2. Dwight

      Georgia obviously had a geographical advantage over Alabama. We should have a high-speed train from Boston to Atlanta, but we don’t. I guess it’s because of those dumb racist Southerners.

      1. John Zelnicker

        Painting with a rather broad brush there, Dwight?

        We should also have a high-speed train from Boston to Washington, DC, but we don’t. Guess it’s because of those dumb racist Yankees. (Sounds just as wrong as your statement, doesn’t it?)

        (And yes, Northerners were and are just as racist as Southerners; they just didn’t have the Jim Crow laws requiring it).

  4. Dave

    Doesn’t Atlanta have the nation’s biggest and busiest airport? Plus it’s closer to Europe. It’s on the way to Florida from the North East and thus was a crossroads with important rail links that were part of the Civil War.

    As to Birmingham, a reliance on steel is great while a country is building itself, not so much later. Cast iron pipe is great for noise control, but is a royal pain to work with.

  5. beth

    Checking with population figures, Atlanta metro was approximately one million in 1970 and London was declining in that period to about 7 million.

    Certainly during the 1960’s Atlanta was not an edge city. The white flight may have led to that later, I do not know. Sports teams move out of the city to where voters will subsidize their stadiums. I see that as a plus for the city itself, not a negative. That takes money out of government for private individuals.

    Also, Atlanta developed into a mecca for black education early in the last century. Education there was valued by blacks much like in Jewish culture. Many blacks got good educations locally and were given scholarships to elite schools in the Northeast and Midwest and some came back to Atlanta. That was not the case in Birmingham. The civil rights movement changed the mindset of both blacks and whites so that the entire region benefited.

    As far as population concentration, the many hillocks and winding road layout that follows old cow paths does not make it easy to concentrate the population, but also makes Atlanta beautiful.

    1. Carolinian

      My friend meant the same geographical area, not population. Perhaps I didn’t make that clear.

  6. Alex morfesis

    Atlanta has been feeding of the nipple of florida for the last 25 years…florida is the only major state without a federal reserve bank…the st louis fed needs to be moved to tampa…atlanta is the fed bank seat for florida….

  7. Jim Haygood

    ‘Controlling for a rich set of geographical fundamentals does not eliminate the importance of pre-colonial density for present population density and income.’

    How can one possibly “control for fundamentals” without making prejudicial assumptions that beg the question?

    We proved it with SAS, so it must be true.

    Maybe unseen ley lines govern where folks settle.

    *reaches for his dowsing stick*

  8. Wade Riddick

    Water is the world’s first cheap highway. Human settlements have clustered near water not simply because of the need for seafood and clean air (and water, of course) but also because rivers and seas grant easy access to distant trade routes for a fraction of the cost of travel by land. Prior to 1500, the only real systematized economic expansion in any modern sense came from lateralized extension of trade and that required low transportation costs which itself required proximity to water (a grim historic irony now that ocean levels are rising and punishing cities that were once advantaged by this proximity).

    Prior to 1500, there wasn’t much systematic in the modern economic sense about the way people cleared land for farming, the process of mining or colonization or the routine improvement of and application of new knowledge to business problems (aka science and technology – this begins with steam power, by the way). Growth depended on trade and that trade depended on cheap transportation, which depended on water.

    More geographic economists than you’d think miss this point because they simply pipe data on waterways into a database. Not every stream or river reaches a good destination for trade, so simple correlation fishing doesn’t always reveal the relationship. You can see this in Russia, where the rivers don’t really run in the right directions for wealth. This simple fact hampered its development up until modern railways lowered internal transportation costs. Walt Rostow used to make this remark in seminar.

    I’d be willing to wager the single most important factor in the historic geographical advantage these settlements enjoy boils down to lower transportation costs viz. that paleolithic UPS, shipping on actual ships.

    If true, the new advantages in modern development will have more to do with actual highways and air routes (and Atlanta was in a much better location for both of these than Birmingham*). It will involve methods of exchange in the less tangible bases of our new economy, information – like internet connections but also informal TED talks and faculty mixers and other hubs where smart people congregate. As energy costs drop and sea levels rise, you should begin to see a geographical diversification in cities that was previously unthinkable.

    * Rail might even be a factor too. It’s why Sherman targeted the old Empire City of the South.

  9. Winston

    This is true for some countries. It true for countries that do not try to create balanced urban development.

Comments are closed.