Bulmer-Thomas V., Coatsworth J., Cortes-Conde R. The Cambridge Economic History of Latin America: Volume 2, The Long Twentieth Century

Подождите немного. Документ загружается.

Cambridge Histories Online © Cambridge University Press, 2008

P1: GDZ

0521812909c08 CB929-Bulmer 0 521 81290 9 October 6, 2005 12:32

312 William R. Summerhill

analytical insights. To assess its economic impact, capital must be disag-

gregated into its component parts because the contribution of investment

varied considerably across individual sectors and even speciﬁc investment

projects. On average, railways absorbed far more capital than any other

activity in this era. The consequences of railway capital formation have

warranted focused investigation.

IMPACTS OF RAILWAY INVESTMENT

The extent to which transport improvements promoted longer-term eco-

nomic growth in Latin America depended on several interrelated effects.

Railways altered existing production relationships, sometimes profoundly.

The lower costs of transport they offered inﬂuenced the decisions of con-

sumers of transport services, who were often agricultural producers. The

leading examples involve regions such as the Argentine pampa and the fer-

tile interior regions of south-central Brazil, which had been beyond feasible

exploitation before the introduction of the railroad. These experienced a

wave of new settlement, investment, and commercial activity once infras-

tructure made them accessible. Transportation cost reductions widened

the area of proﬁtable cultivation, stimulating the rise of new agricultural

enterprises, new settlement, new crops, and new investments in farming.

Rapidly rising agricultural production and incomes were typically the result.

Assessing the impact of the transport infrastructure that was implanted

in Latin America in the late nineteenth century requires the consideration

of an array of beneﬁts and costs and not just private proﬁts to infrastructure

investment. Conventionally, there were four channels by which transport

infrastructure registered its impact on the economy. First was the direct cost

savings on transporting goods and people by rail, or the “social saving.”

Reductions in transport costs freed up resources for other uses, creating a

net savings to the economy as a whole. The second channel was the “forward

linkage” from railways to other sectors of the economy. Railways provided

an impetus to new activities and often fostered increases in the stock of

resources available to ﬁrms and sectors that relied on transport services.

When such forward linkages transpired, they usually took the form of new

investment in agriculture, mining, and industry and, sometimes, new set-

tlement. The third channel involved the demand for inputs necessary for

the construction and operation of railways (“backward linkages”), stimu-

lating industries that in turn further raised output per capita. The fourth

Cambridge Histories Online © Cambridge University Press, 2008

P1: GDZ

0521812909c08 CB929-Bulmer 0 521 81290 9 October 6, 2005 12:32

The Development of Infrastructure 313

channel is a residual category encompassing various consequences fostered

in an indirect fashion by the interaction of railway development with legal

structures, government, and the like (“institutional externalities”). In Latin

America, the impulse to growth via these different channels was highly

unbalanced. The social savings and the forward linkages were generally of

much greater magnitude, and greater importance, than the other categories

of linkage.

The ﬁrst and most obvious impact of investments in railways in Latin

America was the direct effect of the reduction of transport costs. Cost

reductions per unit of freight or per passenger trip included lower fares and

charges, savings on time, and reductions in travel hazards that the railroad

provided. Railroads reduced the resource requirements for producing a

given amount of transport services. The magnitude of these direct gains

depended on the availability of affordable substitutes for the railroad and

the share of the economy’s output that required transportation. Where

either substitutes were affordable or the share of total output carried by rail

was small, the income gain from railway infrastructure was modest. Even

where gains were large, they must still be considered in terms of the cost of

securing them – that is, relative to the cost of providing the infrastructure.

A railway that was too costly, in terms of the capital required to establish

it, might not have been economical at all.

The poor pre-rail conditions in Brazil, Colombia, Argentina, and Mexico

meant that the cost of carrying freight and moving passengers was quite

high, and the unit cost savings provided by the railway were large. When

added to the costs of other inputs to production, transport costs were high

enough in many regions to eliminate any chance of proﬁt for potential pro-

ducers of a wide variety of crops and manufactures over very large areas of

Latin America. Goods that were otherwise affordable at the farm or factory

gate became prohibitively expensive once the transport bill was tacked on.

This transport-cost wedge between producers and ﬁnal users limited large

portions of the population to purchasing only the most basic of necessities.

When confronted with long, arduous journeys by foot, animal, or expen-

sive trips by coach, many persons never ventured far from their immediate

locale. Such features sharply limited the opportunities for material progress.

Economies marked by high transport costs had vastly lower levels of eco-

nomic activity, and thus lower levels of income, than those that enjoyed

easy access to navigable waterways or manmade forms of infrastructure.

With the relative price of shipment so high in so much of Latin America,

it is little wonder that farmers, mine owners, nascent industrialists, and

Cambridge Histories Online © Cambridge University Press, 2008

P1: GDZ

0521812909c08 CB929-Bulmer 0 521 81290 9 October 6, 2005 12:32

314 William R. Summerhill

governments sought to improve transport conditions and lower costs by

implanting a more efﬁcient technology from abroad. Railways, by provid-

ing low-cost shipment, directly improved allocative efﬁciency and raised

the level of output and income. In the case of the labor market, railways

could enhance the geographic mobility of free workers, making it possible

for them to go to jobs over distances previously considered unbridgeable.

The estimation of the magnitude of these direct effects is an exercise

in applied beneﬁt–cost analysis. The direct impact of the railway can be

partitioned into the resource savings produced by the two main categories

of transport services: the savings on freight shipment and the beneﬁts of

passenger travel. The freight social saving of railways is conventionally

deﬁned as the difference between the actual cost of shipping goods in

one year by rail and the cost of shipping the same goods over the same

distances by an alternative means of transport in the absence of railways.

Passenger social savings are analogous, being the difference between the

costs of passenger rail travel and the cost of traveling by nonrail modes of

transport.

Including the cost of passenger travel time, the cost savings that railways

created on these two types of service comprised an addition to the economy’s

total output. Because total output in the economy with railways was equal to

the ﬁnal output of the transport sector and the output of the sectors that

produced all other goods and services, dependence on wagon or mule

transport to produce the output of the transport sector meant that labor and

capital involved in transport services could not be used in the production

of other goods. Producing freight transport service with railways would

release labor and capital involved in transport to other activities in the

economy. The freight social saving achieved by shifting from costly nonrail

modes of shipment to railways is equivalent to the loss in national income

incurred by hypothetically switching from railroads to the next-best mode

of transport. Passenger social savings arise in a very similar fashion and are

equal to the total amount of resources saved by transporting people by rail

rather than the alternatives (walking, riding, coaching).

Throughout Latin America, railway expansion resulted in large unit sav-

ings on freight services. The unit cost of shipping merchandise fell consid-

erably in the decades following the opening of the ﬁrst railways. The decline

in the unit-cost in Mexico between 1878 and 1910 was at least an order of

magnitude and was nearly that much in Argentina from the 1860sto1913.

Coatsworth, Growth Against Development, 96–104;Summerhill, “Economic Consequences of

Argentine Railroad Development.”

Cambridge Histories Online © Cambridge University Press, 2008

P1: GDZ

0521812909c08 CB929-Bulmer 0 521 81290 9 October 6, 2005 12:32

The Development of Infrastructure 315

Brazil’s decline in the real transport charges over the same interval of time

was similarly on the order of tenfold.

With the growth of each nation’s

rail system during the second half of the nineteenth century accompanied

by agrowing volume of freight services, social savings became quite large by

the end of the nineteenth century. In 1900,freight social savings accounted

for as much as one third of Brazilian and Mexican national income, and

in Argentina they were around one quarter of the economy’s total out-

put. Colombian freight social savings were less than in these cases. For the

late 1920s, an upper-bound measure runs between 3.4 and 7.9 percent of

gross domestic product (GDP). That the social savings for Colombia were

smaller than these other cases can be attributed to two features: Colombia’s

relatively low share of railroad transport output in GDP (which is also indi-

cated by its relatively low level of track per capita), and the availability of

some water substitutes for a portion of railroad trafﬁc.

Passenger beneﬁts

were much less than freight savings in these countries, which contrasts with

the results found in more advanced economies, where passenger beneﬁts

comprised a large share of the total direct impact of railroads.

Passenger

beneﬁts were, in part, proportional to the value of travel time that rail-

roads saved, which differed a good deal between backward and advanced

economies because of the higher wages of the latter.

The historical and economic signiﬁcance of the social savings can be

better understood in a broader comparative context. Table 8.5 presents

direct savings estimates on railways in a variety of cases at various points in

time. Direct comparisons across countries are more difﬁcult to make than

implied by the table, because of the differing assumptions underpinning

the estimates in each case. Note, however, that in contrast to the results

derived from studies of railways in North Atlantic economies, and even

from some backward economies in which pre-rail transport was relatively

efﬁcient, the freight social savings from railways in Latin America were

often large by global standards.

Even with large social savings, infrastructure investments could be uneco-

nomical if the returns to the economy that they generated were not large

enough to justify the capital dedicated to the railroad sectors. Yet, based on

William Summerhill, Order Against Progress: Government, Foreign Investment, and Railroads in Brazil,

1854–1913 (Stanford, CA, 2003), 74–83.

Mar

ıa Teresa Ram

ırez, “Los ferrocarriles y su impacto sobre la econom

ıa colombiana,” Revista de

Historia Econ

omica 15:1 (2001): 81–91.

Gary Hawke, Railways and Economic Growth in England and Wales, 1840–1870 (London, 1970), 40–54;

Albert Fishlow, American Railroads and the Transformation of the Antebellum Economy (Cambridge,

1966), 90–2;J.Hayden Boyd and Gary M. Walton, “The Social Saving from Nineteenth-Century

Rail Passenger Services,” Explorations in Economic History 9 (Spring 1972): 233–54.

Cambridge Histories Online © Cambridge University Press, 2008

P1: GDZ

0521812909c08 CB929-Bulmer 0 521 81290 9 October 6, 2005 12:32

316 William R. Summerhill

Table 8.5. Estimates of railway freight social savings for various

countries

Country Year

Social saving as a share of

GDP (percent)

England and Wales 1865 4.1

England and Wales 1890 11.0

France 1872 5.8

Germany 1890s 5.0

Spain 1912 23.7

Belgium 1865 2.5

Belgium 1912 4.5

Russia 1907 4.6

China 1933 0.5

United States 1859 3.7

United States 1890 8.9

Argentina 1913 26.0

Colombia 1927 7.9

Mexico 1910 38.5

Brazil 1913 22.0

Source: Summerhill, Order Against Progress, 98.

what is known about the social rate of return for at least three Latin America

cases, it is clear that investment in railway infrastructure was a wise one in

the aggregate (though such investments could still be considerably unfruit-

ful in speciﬁc instances). For several major lines (including the largest and

most heavily traveled) in Mexico, Argentina, and Brazil, the social rates

of return to railroad infrastructure were quite large, easily high enough to

justify investment, and even subsidy. The social return to investment in

the Mexican Central railway system exceeded 20 percent per year by the

1890s, the social returns implied by the beneﬁt–cost ratios on Colombia’s

larger railroads in the 1920swere impressive, and a conservative estimate

for the social rate of return for the railway sector in Brazil in 1913 ranged

between 18 and 23 percent.

In terms of their direct beneﬁts, outlays on

railway systems were an investment in Latin American infrastructure with

handsome aggregate returns.

Coatsworth, “Impact of Railroads on the Economic Development of Mexico” (Ph.D. dissertation,

University of Wisconsin, Madison, 1972), 140–2;McGreevey, Economic History of Colombia, 264–9;

Summerhill, Order Against Progress, 99–105.

Cambridge Histories Online © Cambridge University Press, 2008

P1: GDZ

0521812909c08 CB929-Bulmer 0 521 81290 9 October 6, 2005 12:32

The Development of Infrastructure 317

LINKAGE EFFECTS

The notion of linkages serves to categorize the various potential ties between

one economic activity and others. In the case of railroads, there were many

possible channels of such effects. Conventionally, historians have focused

heavily onforward and backwardlinkages fromrailroad development,along

with other externalities that do not ﬁt cleanly into any single category.

Forward linkages from Latin American railways were of the conventional

type, in the sense that they created incentives for crop expansion, new

investment, and, in the relatively more labor-scarce regions, immigration.

As the railway networks grew, cheap transport became available on an

increasingly broader scale, pushing beyond the boundaries of cultivation

that prevailed at the moment of the railway’s ﬁrst appearance. Moreover,

railways themselves became more efﬁcient over time, producing output at

alower unit cost as they operated at higher levels of capacity utilization.

Hence, the forward linkages were dynamic in nature.

The main elements comprising the forward linkage generally reduce

to the increase in the stocks of factors of production. Railways in some

areas reinforced traditional patterns of commerce; in others, they rear-

ranged things completely. These linkages varied from nation to nation and

even across regions within nations. The rather dramatic reduction in trans-

port charges created marked regional realignments in crop specialization

and the output mix. In the settler economies of Argentina and southern

Brazil, it was railways that fostered heavy ﬂows of immigrants from abroad

to the interior. They may well have encouraged higher savings and new

investments, though the evidence on such points is so scant as to justify

excluding any detailed discussion of them. By bringing more land, and fre-

quently higher-productivity lands, into use, railway extensions raised the

average products of labor and capital in the affected areas. And by shifting

out the demand curve of labor, immigrants were attracted by higher wages,

leading to a mutually reinforcing process of capital and labor importation.

The strength of these effects depended on the conditions in each setting.

In Mexico, railways revitalized the mining industry, which generated large

capital inﬂows but led to virtually no immigration. In Argentina, by way

of contrast, huge quantities of capital and labor poured into the country as

the railroad opened highly fertile farmland to new cultivation.

In short, forward linkages generated a unique boost in the level of output

per capita, beyond that registered by the direct social savings, though to

date there are few quantitative measures of their size. Instead, the more

Cambridge Histories Online © Cambridge University Press, 2008

P1: GDZ

0521812909c08 CB929-Bulmer 0 521 81290 9 October 6, 2005 12:32

318 William R. Summerhill

qualitative effects have occupied historians’ discussions of the railway for-

ward linkage in Latin America. Railways have long been viewed as having

played a pivotal role in Latin America, less through their role in raising

incomes and more in terms of increasing the export orientation of the

region’s economies. In virtually all of Latin America, railroads boosted

export activities and, in Mexico and Argentina, the effect was so strong

that the export sectors grew faster than the economy as a whole. In Mexico,

market conditions, mineral endowments, and railway transport rate policies

favored rapid export-sector growth during the Porﬁriato. By one estimate,

in 1910 more than 50 percent of the tonnage on Mexico’s most impor-

tant rail line was export-sector produce, more than double the ratio that

obtained in 1885.

Mexico’s exports were especially railway-intensive, using

a large share of rail transport capacity. The consequences of this for alter-

ing the output mix of the economy were dramatic. In 1910, the share of

exports in GDP had nearly doubled over their share in 1877, rising to 17.5

percent.

In Argentina, the overall output mix similarly responded to rail-

way development, accelerating the shift from pastoral activities to intensive

agriculture and making for rapid increases in the volume of exports. Prob-

ably even more so than in the case of Mexico, the surge of late-nineteenth

century export growth in Argentina rested almost entirely on the arrival of

railways.

In an interesting departure from the Argentine and Mexican pattern of

a railway-induced growth, Brazilian railways differed with regard to their

forward linkage effects and their impact on the output mix of the economy.

Railway expansion in Brazil was actually accompanied by a declining share

of export-sector freight and a slowly declining ratio of exports to GDP.

The share of exports in Brazilian GDP declined at an annual rate of some-

where around 1 percent from 1870 to 1913.

Though Brazilian railways were

intended to foster export growth from the outset, they, in fact, generated

even more rapid growth in the domestic-use sector of the economy. In

Colombia, the rise of the coffee sector in the late nineteenth century owed

Coatsworth, Growth Against Development, 130–4.However, the intensity of the railroads’ effect on

Mexico’s export sector has been disputed; compare Coatsworth, Growth Against Development, 122–33,

and Kuntz Ficker, Empresa extranjera y mercado interno: el Ferrocarril Central Mexicano, 1880–1907

(Mexico City, 1995), 211–348.

John H. Coatsworth, “The Decline of the Mexican Economy, 1800–1860,” in Reinhard Liehr, ed.,

America Latina en la

epoca de Sim

on Bol

ıvar: La formaci

on de las econom

ıas nacionales y los intereses

econ

omicos europeos, 1800–1850 (Berlin, 1989).

Summerhill, Order Against Progress, 135–47.

P1: GDZ

0521812909c08 CB929-Bulmer 0 521 81290 9 October 6, 2005 12:32

The Development of Infrastructure 319

much to railway development, and most of the lines built there before 1914

were established with an eye to boosting the coffee trade. But even there,

the export linkage turned out to be weaker than one might have expected.

Railways in Latin America did not have powerful domestic backward

linkages, despite the fact that railway construction and operation required

large quantities of iron and steel, coal, and sophisticated capital equipment.

Absolute weakness was apparent in the very limited demands for domestic

manufactures, minerals, and ﬁnancial inputs derived from railway con-

struction in the region. Backward linkages from Latin American railways

were not especially weak in comparison with the advanced economies of

the North Atlantic. That Latin American railways do not appear to have

been uniquely deﬁcient in this regard can be gleaned from a careful exami-

nation of the nature of backward linkages in the industrializing economies.

The advanced economies of the North Atlantic enjoyed their ascendance

because of changes across a broad front of activities that collectively com-

prised sustained economic growth. There, the successful articulation of even

modest backward linkages depended on a process of industrial advance that

predated the diffusion of steam locomotion.

Though Latin American nations did not come to provision the bulk

of their own railway needs before 1930, there was, in a number of cases, a

modest magniﬁcation of industries that supplied the railways. Nearly every-

where in Latin America, domestic railway shops over time came to provide

repair services and basic fabrication of rolling stock in the form of passenger

cars and freight wagons. Assembly activities still required a high portion

of imported content in many nations because domestic manufacturing

was not able to provide basic components. Porﬁrian Mexico succeeded in

manufacturing some of its own rails and also provided some of its own

fuel coal. But there were strong limits to these possibilities for most of the

period under study; by the turn of the century, Mexico possessed only a

single modern steel mill.

In Brazil, no domestic iron and steel manufac-

ture developed in response to railways, coal mining in the south received

little stimulus from railway operation, and only items of basic rolling stock

came to be manufactured domestically during this period. Chile, like most

Ram

ırez, “Ferrocarilles y su Impacto,” 102–8.

Germany stands out as the principal exception in this regard; see Rainer Fremdling, “Railroads and

German Economic Growth: A Leading Sector Analysis with a Comparison to the United States and

Great Britain,” Journal of Economic History 37:3 (1977): 583–604.

Haber, Industry and Underdevelopment, 30–7.

P1: GDZ

0521812909c08 CB929-Bulmer 0 521 81290 9 October 6, 2005 12:32

320 William R. Summerhill

other nations, provided some of the inputs for railroads and established

assembly and maintenance facilities for rolling stock. But it did not enjoy

an industrial boom as a result of railway development.

In the absence of domestic backward linkages, Latin American railways

created leakages from the income stream that Latin America would not have

had if it had been able to tap its own derived demands for railway inputs. In

Mexico, it is estimated that around 60 percent of railway revenues in 1910

were spent abroad on imported railway inputs.

In Brazil, the ratio was

even higher and the leakage from foreign provisioning ran to as much as

3 percent of GDP.

By importing its railroad inputs, the region exported

potentially signiﬁcant employment and income effects overseas. That so

many inputs came from abroad may be interpreted to mean that railways

somehow caused the Latin American economies to miss out on a critical

opportunity to develop a manufacturing base and industrialize. Yet there

is little reason to expect railways in Latin America to have created around

them an entire manufacturing sector because they did not typically do this

elsewhere. In advanced industrializing economies, the backward linkages

from railways to industry were of limited importance because domestic

iron, steel, and coal activities were already large and well developed, as part

of a generalized process of industrialization predating the construction of

railways.

By contrast, such linkages were unimportant in countries like Brazil

because the domestic iron sector was so backward and domestic coal was

woefully inadequate in terms of quantity and quality. Identifying the spe-

ciﬁc reasons for the absence of stronger backward linkages from railways

in Latin America requires an analysis looking beyond railways to the con-

ditions of industry itself in the pre-railroad era.

Beyond their industrial

potential, railways also registered some – though often modest – impact on

domestic capital markets via lines that issued debt and equity at home.

Future research will, no doubt, better illuminate the full array of beneﬁts

and costs that railway infrastructure investment created. Studies that explic-

itly consider the relative merits and pitfalls of the rapid increase in foreign

Guillermo Guajardo, “Nuevos datos para un viejo debate: los v

ınculos entre ferrocarriles e industri-

alizaci

on en Chile y M

exico, 1860–1950,” El Timestre Econ

omico 65:258 (1998): 213–61.

Coatsworth, Growth Against Development, 143–4.

Summerhill, Order Against Progress, 133.

Robert W. Fogel, Railroads and American Economic Growth (Baltimore, MD, 1964), 190–210;Fishlow,

American Railroads, 141–3.

William Callaghan, “Obstacles to Industrialization: The Iron and Steel Industry in Brazil During

the Old Republic” (Ph.D. dissertation, University of Texas, Austin, 1981).

P1: GDZ

0521812909c08 CB929-Bulmer 0 521 81290 9 October 6, 2005 12:32

The Development of Infrastructure 321

investment, the differential transport cost reductions resulting from the

pattern of regulation, and the opportunity costs of subsidy policy, will aid

in reﬁning the characterizations historians establish with respect to infras-

tructure. The results to date for Mexico, Brazil, Colombia, and Argentina

suggest that these countries, like other economies that had especially back-

ward pre-rail transport systems, captured beneﬁts from infrastructure in

the form of railways that outstripped any other sources of material progress

in terms of their impact. Railway expansion in Latin America removed one

of the most important impediments to the process of economic growth –

high internal transport costs. By integrating markets and creating gains

from regional specialization and improved interregional terms of trade,

infrastructure embodied in railways spurred per capita income growth in

the Latin American economies and laid the groundwork for the forma-

tion of national markets that were required to initiate and sustain modern

industrialization.

PORT IMPROVEMENTS

The importance of steam in the transportation revolution was, of course,

not limited to locomotives on land. Even before railways began to rapidly

expand in Latin America, steam-powered shipping had already registered

an impact. The application of steam to freighters, and the creation of steel-

hulled ships, reduced the costs of shipping exports from Latin American

ports to overseas markets and the costs of importing key inputs, including

the very materials needed to construct railways. But these cost reductions

were,infact, modest in comparison to the impact of the railroad on inland

transport costs. Real overseas shipping rates to South America fell by as

much as one half between 1870 and 1930, but with considerable ﬂuctuation

around the trend.

The shipping component was but a small fraction of the

overall transport costs incurred by Latin American foreign trade. Indeed,

the development of new areas of settlement and production depended far

less on shipping than it did on the decline in inland transport costs, which

proved to be indispensable to the integration of markets for agriculture and

manufacturing.

Saif I. Shah Mohammed and Jeffrey Williamson, “Freight Rates and Productivity Gains in British

Tramp Shipping, 1869–1950”(National Bureau of Economic Research, Working Paper no. 9531,

2003), Appendix 2, 7–8.