Portage and Path Dependence Hoyt Bleakley and Jeffrey

Portage and Path Dependence Hoyt Bleakley and Jeffrey Lin. 2012. Theresa Geyer, Economics B. Sc. , Universität Heidelberg 1

Portage Sites • During early settlement in North America, rivers were main transportation routes • Portage Sites: Places where boats/ cargo could be carried over land to avoid an obstacle on the waterway • Initial value: Completing trade routes Focal points of commerce in 1800 s • Became obsolete due to changes in technology 2

Paper examines why there is still a concentration of economic activity at portage sites in the US although the original natural advantages are nowadays obsolete Summary How is that question answered? The authors answer the question by comparing the predictions of economic models featuring different effects. Main result: The authors favor a model with strong increasing returns to scale in economic activity combined with historically sunk investments to explain the current distribution of economic activity. 3

Paper‘s contribution Path dependence Portage Sites Natural endowments Distribution of economic activity • Paper tells us by which mechanism obsolete natural endowments are able to explain a current distribution of economic activity • Disentangling effects of path dependence & natural endowments 4

The Fall Line & Areas of Analysis • Fall Line: Geomorphical feature dividing Piedmont and coastal plain many falls along this line Intersection fall line / river Portage Site Falls on 3 Mississippi tributaries Great Lakes & Mississippi River System 5

Further Steps and Background Information 1. Prove decline of portage 1. 2. Background information Quantitative evidence 2. Data proving that early portage sites are still densely populated 1. 2. Maps Statistical comparisons 3. Responses after obsolescence 4. Explanation 1. 2. 3. Historical Factors Contrast to neoclassical prediction Economic models of geography 6

1. 1 Short History and Background Information early 1800 s Transportation Infrastructure Change: (i) canals and locks (ii) railroads 1880 s 1930 • Portaging activities relatively important • employment at portage sites reaches a peak • average fall-line county contained 13. 1% of total watertransportation employment along an entire river • average fall-line county contained 2. 6% of total watertransportation employment along an entire river 7

1. 2. Quantitative Evidence: Employment Data Portaging activities have started to decline drastically after 1880 8

2. Data Contemporary distribution of economic activity across the southeastern United States (nighttime lights) FALL LINE Contemporary fall-line cities 9

In General: Population centers at the points where rivers cross the fall line. No population centers along the fall line, if a river is not present. 10

11

2. 2 Statistical Comparisons at Fall Line Explanatory variables: - Proximity to portage site - Size of watershed area upstream of the fall line regression Outcome variable: - Current population density Indicators: - census tracts/ counties - average intensity of nighttime lights 12

Results: a) Proximity to portage site as explanatory variable: - 10% farther away from a portage site 6% lower population density in the tract data and 2% lower density in the lights and county data. b) Including the size of watershed upstream: - 10% larger upstream watershed 4% higher density at the portage site c) Controlling for other effects no significant changes correlation between historical portage and the current distribution of economic activity Watershed size affects current density at portage sites 13

4. Responses after Obsolescence Both methods show: • Difference-in-difference no evidence for thatformal portagetest: cities approach became concentrated as • First less difference: portage original portaging activities advantage • Second difference: became obsolete 19 th/20 th century 14

5. Explanations for Persistence Decreasing returns to scale Congestion Costs An increase in all inputs (labour/capital) leads to a less than proportional increase in output Costs resulting because of negative externalities of high population density Sunk Costs 15

5. 1 Historical Factors? • Does historical density of specific factors/ combination of these at portage sites explain persistence? • Possible factors: railroad network, literacy, … • Result: Few differences between portage & nonportage sites if we condition on present-day population density 16

5. 1. Contrast to Neoclassical Prediction • Contrasts with predictions of models featuring locally decreasing returns to scale area with actual natural advantages relocation of production factors to this area Portage Sites should have been in decline 17

5. 2 Applying a model of economic geography. . . … and considering sunk costs • - No effect on long-run equilibrium One long-run equilibrium: utility Intersection ofshrinks V* and V(x) • Vertical distance Value of portage density - Natural will not • Due toadvantages decline in portage affect long-run equilibrium advantage& sunk capital -depriciation Alternate locations will (slows down convergence) have same V curve density • - In. Differences the mediuminrun: differences be persistent • Incan‘t the long run: convergence 18

. . . with strong economies of scale Increasing returns > congestion costs … and considering sunk costs - More possible long-run • Sunk investments could solve equilibria coordination problem - Natural advantages could • Vertical distance shrinks affect long-run equilibrium • Possibility for multiple equilibria - Coordination device at a site • Differences in density could persist for a long time 19

Implications Natural endowment CASE A Medium-run equilibrium oversupply of sunk capital in portage sites differences in prices for sunk capital Level of current population density Historical sunk investments CASE B long-run equilibrium 20

Results (I) • Although the advantages of the institution “portage site” became obsolete long time ago, the initial advantages still affect the current population density • path dependence can emerge if (i) historical advantages coordinate activity to a particular location (ii) returns to scale rise enough to sustain density there 21

Results (II): Main Economic Mechanism obsolescence Natural endowment • Water obstacle portage site Level of current population density Economic model with strong increasing returns to scale • one of the possible equilibria of density • Historical portages as coordination device Central hypothesis: Population density persists in places with obsolete natural advantages if there are strong economies of scale combined with historically 22 sunk investments which act as a coordination device.