5
The Symbiosis of Exports and Regional Trade

Most Latin American economies went through a period of rapid growth between the 1850s and 1920 as the extension of steamship services and the construction of rail lines from major ports to agricultural and mineral production zones allowed raw materials from the region to supply the burgeoning industrial economies of Western Europe and North America. Peru certainly participated in this experience of rapid growth, but more than any other nation in Latin America, its economic development during these seven or eight decades was punctured by a devastating depression lasting from 1873 to the late 1880s. The depression was brought on by the collapse of the lucrative guano export business, the destruction and loss of resources inflicted by Chilean troops during the War of the Pacific (1879–83), and the particularly severe cyclical crises simultaneously wrecking the European economies.^[1] Growth of foreign trade, capital investment, and government revenues was impressive in the decades after the crisis.^[2] But growth came with considerable costs: increasing dependence on the performance of key export sectors; a decline in "autonomous development" and in the growth of domestic industrial output; increasing foreign ownership of enterprises; intensifying regional and social income disparities; and the concentration of the modern sector of the economy in Lima and a few enclaves along the north coast and in the central highlands.^[3]

In this chapter I examine the degree to which southern Peru shared the paradoxical national experience of strong growth with few lasting benefits. I argue that growth of foreign trade, although significant by regional standards, was less impressive in the south than it was in the north and center. Before 1920 the trade cycles were not as disruptive for the region as they were for Peru as a whole, and the growth of wool exports was associated with an expansion of domestic regional commerce. In great parts

of southern Peru, especially in the altiplano, trade continued to be shaped by low per capita incomes, the strength of the peasant household economy, and a neocolonial, clientalistic sociopolitical structure.

Southern Peru's Export Performance

By the 1850s the export of wool to Europe had come to constitute the altiplano's most important source of income. After the slow demise of the colonial mining-supply economy, producers and traders adapted to the new commercialization system. Yet it is difficult to speak unequivocally of an overpowering wool export boom for southern Peru during the following sixty-five years. Evaluation of wool export performance during this period depends on specific variables: the volume or value of exports; sheep or alpaca wool exports; and, most important, the currency used as a basis for measurement. The task is further complicated by the fact that, in spite of wool's strategic importance for the region, Peru never became a major source of supply in the world markets: during the first quarter of the twentieth century the country's wool exports constituted between one-third and one-half of one percent of world wool production.^[4] Most international trade publications did not bother to list data on Peruvian exports, so that on such matters as the FOB price of wools in a Peruvian harbor, the disposition of the wools in the consumer nations, and even the volume of exports, statistics for Peru remain less reliable than they do for major producers such as Argentina or Australia. Although the data for the years between 1855 and 1920 are more reliable than those for the immediate postindependence period, the statistics discussed below should be viewed with caution, particularly regarding year-to-year fluctuations.^[5]

Five countries with high land/labor ratios—Australia, New Zealand, South Africa, Argentina, and Uruguay—rose to dominate the world trade in wools during the second half of the nineteenth century. As the modern woolen industries grew rapidly, first in England between 1850 and the late 1870s and thereafter in other European countries and the United States, these overseas suppliers controlled a growing share of world trade in animal fibers. The shifting pattern of the wool trade was accelerated by the diminution of sheep herds in most of the major West European wool consuming nations.^[6]

Peru also benefited from this rapid growth of demand for overseas fibers. But it could not keep pace with the five major wool exporters because its wool-producing regions, notably the altiplano, lacked the one condition allowing them to augment production for export in close correlation to rising demand: plentiful land not employed in production for domestic

markets or for peasant household subsistence. While countries such as South Africa and Australia increased their sheep wool exports more than tenfold between the late 1850s and 1911–15, Peru merely doubled its exports. As early as the mid-1860s Peruvian sheep wool exports reached a peak volume not surpassed until the boom years between 1916 and 1919. The volume of national alpaca wool exports rose somewhat more steadily, growing by about 150 percent between the late 1850s and 1910–14 (tables 5.1, 5.2). Yet, although the elasticity of wool supplies for export markets was much lower in Peru than it was in the major export nations, it was not altogether absent. In the short run, merchants, large landholders, and peasants adjusted their wool remittances to foreign demand. Furthermore, for many decades Peruvian wool production increased more rapidly than did exports. Thus, growing domestic consumption must be considered as one factor that limited export expansion.

Sheep wool exports from Islay, the port on Arequipa's coast that handled most of southern Peru's maritime trade until 1870, expanded vigorously from the mid-1850s until 1867 in terms of both volume and value (table 5.1). In the latter year nearly two thousand metric tons were exported, close to double the amount of average yearly exports during the quinquennium 1855–59, a period of record wool exports itself. As prices rose vigorously in reaction to the cotton famine and the rapid growth of Britain's woolen industries, total value of sheep wool exports climbed by nearly 200 percent between 1855 and 1867 (figs. 5.1, 5.2).^[7] Expansion was interrupted in individual years, such as 1857 and 1865, because of civil wars, when the road to the port was blocked and transport animals became scarce.^[8]

Sheep wool exports contracted briefly between 1868 and 1870, with prices declining by up to 30 percent and total export value plummeting to below half of its mid-1860s peak level (from 1,151,318 to 413,241 soles

Figure 5.1
Annual Volume of Sheep and Alpaca Wool Exports from Islay and Mollendo, 1855–1929
(Five-Year Averages)

Figure 5.2
Value of Sheep and Alpaca Wool Exports from Islay and Mollendo, in British Pounds Sterling and Soles m.n., 1855–1919
(Five-Year Averages; 1855–59 = 100)

m.n.). When trade recovered again during most of the 1870s, sheep wool prices in Britain (CIF prices [that is, the price including cost, insurance, and freight]) for the first time showed the effects of overproduction of wools in Australasia and the Río de la Plata.^[9] After 1873 the British woolen industry, the all-important consumer of Peruvian sheep wools, was hit hard by shrinking export markets in continental Europe and the United States, the effect of both a severe cyclical crisis there and a change in fashion, when consumers began to abandon Bradford worsted goods in favor of soft, all-woolen fabrics, often made with merino wools; this branch of the industry was more developed in France, where Peru sold little wool.^[10] Even in the best year of the 1870s (1877) total value of sheep wool exports (Peruvian currency) lay about 25 percent below the record value of 1867.

During the 1880s the double blow of an industrial crisis in Europe and the War of the Pacific in Peru produced the most severe depression of southern Peru's wool export for the whole period under consideration. As early as 1879 British CIF prices had dropped 20 percent below the 1855–59 average. In 1880 the Chileans blockaded the port of Mollendo and destroyed port installations. Although there was no fighting in southern Peru's wool-producing zone, pack animals required to get the clip to railroad stations became scarce, and the army placed large orders for uniforms with shops and the one existing woolen mill in Urcos (department of Cuzco), diminishing exportable wool surpluses.^[11] By 1882 the quantity of sheep wool exports through Mollendo had declined to just over half of the average amount for the years 1855–59. Although the restoration of peace in 1883 permitted the sale of stocks accumulated during the war years—sheep wool exports nearly tripled in 1884 from the preceding year—the slump continued, perhaps because of the civil war between Mariano Iglesias and Andrés Avelino Cáceres, until 1887.^[12] Even so, southern Peru suffered less from the war and its aftermath than did central and northern Peru, where the 1880s produced the most severe economic and social crisis between the Wars of Independence and the Great Depression of 1929–32, causing the bankruptcy of many hacendados and endemic social unrest.^[13]

The recovery of southern Peru's sheep wool exports coincided with a difficult period for international trade. The volume of exports from Mollendo increased moderately from the late 1880s until 1897, reaching a peak 25 percent higher than the average for the years 1855–59 yet still substantially below the level of the boom years of the mid-1860s. But through the early 1890s prices in Liverpool and London continued their long decline, begun in 1873, to less than two-thirds of the average price during the late 1850s.^[14] World production of wools increased by more than a third

between 1887 and 1895, affected especially by the dizzying growth of sheep herds in Australia.^[15] This surge of supply coincided with a severe slump of demand during the depression of 1890–95. The cyclical crisis intensified the structural crisis of the British woolen industry—still by far the most important customer for Peruvian wools—which saw its markets eroded by the growing success of foreign competition.^[16] The total value of sheep wool exports from southern Peru in terms of pound sterling at British ports thus recovered only slightly from the nadir of the 1880s.

After 1895 international wool markets began a long period of prosperity. As global wool production stagnated, prices increased steadily until the end of World War I, except for brief recessions in 1907–8 and 1911–12.^[17] Peruvian sheep wool exports did not immediately benefit from this price increase. Southern Peru exported mostly coarse wools consumed by the mills around Bradford, which turned out traditional worsted fabrics whose market share continued to decline.^[18] Thus, in contrast to average wool prices on the international markets, CIF prices for Peruvian sheep wool began to recover only after 1904; 1915 was the first year since 1878 during which they surpassed the average level for the years 1855–59. Export volume may have reflected these specific market conditions for Peruvian sheep wool. It stagnated through 1902 at between 900 and 1,200 tons, about the same level as during the second half of the 1850s. But after 1903 exports expanded vigorously, reaching over 2,000 tons in 1911, a volume slightly above the previous peak of 1867. Shipments from Mollendo dipped briefly but sharply in the recession years of 1907–8 and 1912.

After a brief crisis in southern Peru's wool business during the early months of World War I, brought on by the country's financial woes, Europe's wartime scarcities produced a wool export boom of unprecedented dimensions between 1915 and the first postwar year of 1919. At more than 2,500 tons, the volume of sheep wool exports in the peak year of 1917 was more than two and a half times larger than the average for the years 1855–59. Although only a modest increase over previous peak years such as 1867 or 1911, this growth of exports demonstrates the responsiveness of merchants and producers to favorable market conditions.^[19] By 1918 a kilogram of average Peruvian sheep wool sold for 79.4 pence at an English port, over three times the average for 1855–59. The combination of extremely high prices and record export volumes brought a bonanza of earnings for the five-year period 1915–1919. In the following year, 1920, prices and export volumes plunged precipitously, with severe repercussions for southern Peru's economy (discussed further in chapter 9).

Overall, the record of southern Peruvian sheep wool exports between the mid-1850s and 1920 is not one of impressive growth. Prices and total

value in terms of British currency went through two growth periods, 1855 to 1867 and 1903–4 to 1919–20. Although World War I briefly led to an export bonanza of much greater proportions than that during the first cycle of expansion, export volume had grown only modestly from the first to the second peak. In between there lay a period of decline, from 1867–72 to 1882–95 (depending on whether one looks at price or volume), and another of halting recovery, from the late 1880s to the early 1900s.

The picture looks different if transport costs and currency exchange rates are considered. The dramatic decrease in maritime freight charges between the 1850s and the early decades of the twentieth century—brought about first by steel-hulled sailing clippers and, beginning in the 1870s, by regular steamship lines—contributed, as Berrick Saul has put it, "to the steady decline of import prices for Britain during the 'Great Depression' [of 1873–95], whilst reducing the impact of unfavorable terms of trade for primary producers."^[20] In other words, an increasing share of CIF prices accrued to exporters. The reduction in transport costs varied greatly according to specific commodities and routes, with high-bulk/low-value goods hauled over the longest distances generally benefiting most.^[21] Rory Miller has suggested that between 1863 and the late 1880s shipping rates for wool from Peru to Liverpool declined by at least 50 percent. By my own calculations (table 5.3) nearly half of the precipitous decline in the international wool prices between the early 1870s and mid-1890s was absorbed by the reduction in the cost of overseas shipping, insurance, and other incidental charges. For Peruvian producers the effects of declining CIF prices were much less drastic and more short-lived than it appears at first sight.^[22]

A compensatory effect of similar proportions was achieved through currency devaluation. As silver became demonetized in one nation after another beginning in the early 1870s, it lost 50 percent of its value in terms

of gold until the mid-1890s. Peru's adherence to the silver standard led to the devaluation of the silver sol by the same ratio from the 1860s, an expression of the agricultural exporters' power, especially between 1887 and 1895.^[23] When the Piérola administration switched Peru to the gold standard between 1897 and 1900, the exchange rate stabilized at about twenty-four pence per sol.^[24] This parity was maintained with only slight fluctuations until Peru's export boom during and immediately after World War I led to a brief appreciation of her currency by about 20 percent, only to slide below the gold parity during the early 1920s.

This currency devaluation made the slump of sheep wool prices during the last quarter of the nineteenth century less severe and shorter in terms of soles (fig. 5.2). As early as 1889 the price lay above the average for 1855–59, and in 1894 it surpassed the previous peak of the mid-1860s. In 1897 the CIF value of sheep wool exports through Mollendo lay only 21 percent below the best year of the 1860s in terms of soles, and by 1905 it had surpassed the previous peak of 1867 by nearly 30 percent.^[25] If we take FOB prices into consideration, the whole decade between 1892 and 1902 (before the renewed expansion of export volumes) appears as one of growing affluence for southern Peru's sheep wool trade.^[26]

But the role of currency transactions in southern Peru was more complex than this analysis suggests. At least until 1910 the department of Puno relied primarily on Bolivian coins as a medium of circulation, whereas Peruvian currency remained scarce. By one calculation, between the first emission of debased coins under President Santa Cruz in 1829 and the termination of this practice after the overthrow of the Melgarejo regime in 1869, some thirty-four million pesos were minted in Bolivia, of which ten million circulated in Peru.^[27] Repeated attempts by both the Peruvian and Bolivian governments to withdraw this debased coinage from circulation proved insufficient, and the debased quintos and arañas , coins valued at one-fifth and one-fourth of a Bolivian peso, were still the most common currency in the Peruvian altiplano decades after their last coinage.^[28] In 1890 the Peruvian government once again undertook to convert all "bad money" circulating in southern Peru; this effort was opposed by merchants and the general public, burned by losses sustained in the conversion of paper money during the late 1880s. Although the operation seemed to succeed in Cuzco and Arequipa, the minister of finance and trade had to admit that "since [in Puno] the Bolivian pesetas called arañas are the only circulating money, it will be somewhat more difficult to let it disappear soon."^[29] As Indian livestock herders insisted on being paid in coins for their wool and hides, the withdrawal of Bolivian currency without replacing it with Peruvian coinage of small denominations would have led to a collapse of the wool export business.

Monetary confusion in Peru's altiplano was greatest between the mid-1880s and 1905. Debased Bolivian peso denominations and coins of the more recent decimal boliviano currency, dubbed soles moneda boliviana in Azángaro, predominated, but gradually circulation of Peruvian soles moneda nacional (soles m.n.) increased. As late as the 1890s some "strong" pesos continued to change hands, probably official Peruvian coinage from before the introduction of soles in 1863. Bolivian currencies constituted a widespread medium of payment in the Peruvian altiplano as late as 1920. At that time many stores in towns such as Puno, Juliaca, and Ayaviri and even farther north in Sicuani and Cuzco considered money exchange an important part of their business, prominently displaying the service in advertisements.^[30] Throughout the wool-producing area Bolivian currency was the lubricant of trade at least until 1910.

The predominance of Bolivian currencies is important for calculating regional earnings from exports because their exchange rate fluctuated vis-à-vis Peruvian currency. Bolivia fully adopted the gold standard only in 1908, eleven years after Peru had taken this step.^[31] During this period Bolivian silver currencies depreciated against Peruvian soles m.n. proportionately to the declining value of silver on international markets, just as the Peruvian currency had done vis-à-vis the pound sterling up to 1897 (fig. 5.3).^[32] Because producers were paid overwhelmingly in debased Bolivian pesos or bolivianos, their devaluation resulted in higher earnings from wool exports for several years between 1898 and 1910.

But fluctuations in the value of Bolivian coins in the Peruvian altiplano show another influence: that of demand fluctuations for wool. "Coinage is becoming scarcer every day and the prices [for wool] in silver are reduced, as is natural," one trader in Ayaviri noted at the beginning of the World War I boom in mid-1915.^[33] The demand for coinage—and this meant mostly Bolivian coins until the 1910s—in the altiplano was determined primarily by the export conjuncture: an export boom led to appreciation of the Bolivian coins vis-à-vis soles m.n., and trade crises, such as those of 1901–2 and 1907–8, brought the value of the coins down. Such currency fluctuations attenuated the price paid to producers in boom periods and diminished the rate of decline during crisis years while making imported goods more expensive. The free currency market of the altiplano worked as a buffer—especially for Indian peasants, who relied most on Bolivian currency—against the gyrations of the export commodity cycle.

Alpaca wool exports from Islay and Mollendo developed differently than did those of sheep wool (see table 5.2, figs. 5.1, 5.2). During the late 1850s export volumes for both fibers, at about one thousand tons, were nearly equal. But between 1869 and 1920 alpaca exports exceeded those of sheep wools in every year except one. During the first decade of the twentieth

Figure 5.3
Exchange Rates between Bolivian Currencies and Soles m.n. in Azángaro Transactions, 1895–1910
Sources : REPA and REPP, 1895–1910.

century, with exports of up to 2,600 tons, alpaca export volumes often were double those of sheep wool. The former increased rather steadily during the whole period from 1855–59. Only the averages for two five-year periods, 1880–84 and 1910–14, show moderate declines. However, the steady growth in the volume of alpaca wool exports failed to translate into an equally strong growth in their value, as prices went through a longer and deeper slump than did sheep wool prices. Alpaca wool was produced by only three Andean countries (with the lion's share coming from southern Peru); until quite late in the nineteenth century the totality of industrial consumers consisted of a small number of manufacturers located mostly in the West Riding, and demand for the precious fiber was highly dependent on fashions. Thus, market conditions for alpaca wools differed greatly from those for sheep wools.

Because production of alpaca wool was widely dispersed among a large number of Indian smallholders with no organization and little information about market conditions, Andean producers had little control over prices. On the contrary, for many years a few processors were able to impose low prices. The three manufacturing companies that were "overwhelmingly dominant" as consumers of British alpaca wool imports until the early 1870s—Fosters, Titus Salt, and G. and I. Turner—frequently colluded to keep down the price of their raw materials.^[34] In the United States, a market of increasing importance since the closing decades of the nineteenth century, the preeminent consumer, the Farr Alpaca Company, successfully sought to discourage the entry of new competitors into the market.^[35]

Such evidence for oligopsonistic control over the alpaca wool market, in addition to cyclical and structural weaknesses of demand, must be taken into consideration as a factor underlying the particularly severe and long-term depression of alpaca wool prices. The much finer and longer alpaca fiber had always fetched considerably higher prices than sheep wool had. The premium for alpaca peaked during the 1850s and 1860s, with prices of two-and-one-half times those for sheep wool. When international wool prices began their long-term decline in the mid-1870s, alpaca wools were hit considerably harder and longer than sheep wools were. A fashion change in the 1870s exacerbated the cyclical crises over the following fifteen years, sending CIF alpaca prices from a peak of 96.8 pence per kilogram in 1864 to a nadir of 22.3 pence in 1888.^[36] In terms of pounds sterling the price did not reach the average level of 1855–59 in a single year between 1876 and 1917, and during most years of this period it was less than half the average price for the late 1850s. In spite of continued growth in export volumes five-year averages of total value remained below or barely above the par value for the 1855–59 period until the boom years of World War

I (sheep wool export values in pounds sterling had witnessed a more steady recovery since 1905). Alpaca wool prices profited from the boom of World War I, but less so than sheep wool prices did.^[37]

Of course, because of the Peruvian currency devaluation, the slump of alpaca wool export values in terms of soles m.n. during the 1880s was also less severe than if it were measured in pounds sterling. As early as 1895 the total value of alpaca wool exports in soles m.n. surpassed the previous peak of 1866. By this measure, for the remainder of the period alpaca export values grew parallel to those for sheep wool, in spite of the stronger growth of volume. However, the alpaca trade benefited less from reductions in overseas transport costs than did sheep wool.^[38]

Although the longer-term trends for wool prices and export volumes (table 5.4) underscore the responsiveness of southern Peruvian producers and traders to external demand, two problems need special consideration:^[39] (1) Why did sheep wool export volumes stagnate between 1892 and 1902 when, as early as 1893–94, CIF prices in soles m.n. had reached levels comparable to the previous peak of the mid-1860s? (2) By contrast, why did alpaca wool export volumes continue to grow so vigorously between the mid-1880s and 1905–6 when prices in pounds sterling were depressed throughout this period and even prices in soles m.n. failed to return to the level of the mid-1860s? To answer these questions, we need to consider the development of domestic production of both fibers, changing ratios of domestic consumption to exports, and the impact of foreign demand on different groups of wool producers.

Information on these issues is fragmentary. It seems unlikely, however, that a strongly differentiated development in the production of both fibers caused these phases of countercyclical export volumes. During the century after independence sheep populations in the altiplano and neighboring livestock zones grew approximately threefold while alpaca herds may have grown somewhat faster. As noted in chapter 2, around 1840 a higher percentage of total alpaca wool production than of sheep wool production was exported. Given the trends for export volumes and total production of both fibers, this situation did not change over the next eighty years. All estimates from the early twentieth century suggest a considerably larger production of sheep wool than of alpaca wool, although alpaca wool exports continued to outpace those of sheep wool.^[40] Production ceilings thus cannot explain the stagnation of sheep wool exports during the years of improving market conditions between 1892 and 1902. If anything, the expansion of alpaca wool exports might have been hampered by short supply, especially in years of rapidly growing demand. The British engineer A. J. Duffield, who claimed to have studied Peruvian wool production methods for four

years, wrote in 1877 that "all the wool of the alpaca, the llama and the vicuña is sent to England. No Peruvian of any social standing has had the pluck or the sense to do anything towards extending the cultivation of alpaca wool." Had its production been expanded, Peru might have derived "a net annual income of £20,000,000."^[41]

Although this figure certainly was poetic exaggeration, designed to underscore the Peruvian elite's failure during the "age of bird dung," other evidence also suggests that Indian peasants routinely sold most or all of their

alpaca wool crop during the second half of the nineteenth century.^[42] As late as 1927 the Indian peasantry in Chumbivilcas, a livestock-raising province in the department of Cuzco, owned about 72 percent of domestic cameloids; during the preceding sixty years this percentage would have been even higher.^[43] As hacendados remained reluctant to enter the alpaca-raising business, supply of this fiber was highly dependent on peasant production. With the exception of a few items of ceremonial clothing, peasants appear to have substituted sheep and llama wools for their home consumption of alpaca wool. This substitution was prompted both by the price premium paid for alpaca and by the strong pressures wool traders exerted on them. Traders could still earn handsome profits even during the long phase of declining demand for alpaca wools by passing on the lion's share of price reductions to the peasant producers. Thus, the role that alpaca wools played in peasants' household economy and the exploitive commercialization practices into which they were tied help explain the failure of alpaca wool exports to respond to the long-term downturn in international demand.

Sheep wool exports, by contrast, responded more closely to international demand fluctuations. A higher share—probably above 50 percent by the 1890s—was produced on estates. The stagnation of exports during the mid and late 1890s corresponded to a phase in which, because of the devaluation of the currency and increasing levels of industrial protection, the terms of trade favored domestic processing over exports. A number of woolen mills were founded both in the south and in Lima.^[44] Wool producers and traders must have sold a growing albeit still minor share of the clip to these new mills. By 1902 the output of the Cuzco mills may have begun to affect peasant demand for cheap imported cloth in the south, as the British vice consul in Mollendo reported a sharp drop in sales of Bradford serges over previous years.^[45] But in the following years the terms of trade again favored raw material exports, a result of declining tariffs for textiles, currency stabilization, and the stagnation in international wool supplies. Industrial wool processing in Peru seems to have stagnated while exports increased again.^[46]

Even if wool dominated the region's foreign trade, southern Peru never became a single-export economy between the 1850s and 1920. The long-term share of wool in the total value of exports from Islay and Mollendo hovered around 60 to 65 percent (table 5.5). It briefly increased to as much as 80 or 90 percent in boom years during the 1860s and 1910s and may have fallen below 60 percent in the decades of declining wool prices between the mid-1870s and mid-1890s.^[47] After the turn of the century other livestock products, especially cowhides, contributed up to 7 or 8 percent to southern Peru's exports.

A wide range of mineral and agricultural commodities complemented southern Peru's wool exports. Most of these contributed to the region's foreign trade in appreciable amounts only for a decade or two, but as one product was displaced from international markets others grew. Cascarilla (chinchona bark) remained the region's second export commodity through the 1880s but dwindled to insignificance during the next decade as aspirin replaced quinine as a cure-all. During the 1890s coca leaves, rubber, and borax began to be exported from Mollendo in large amounts, each contributing briefly between 7 and 20 percent to total southern Peruvian exports. But by the early 1910s these trades had withered, displaced by substitute products, more efficient foreign producers, or conflicts with the Peruvian government.^[48] Beginning in the mid-1910s cotton and, more briefly, sugar, both produced in the valleys between Arequipa and the coast, became important export commodities for southern Peru.

Minerals such as silver, gold, and copper were, of course, the other major exports from the region. However, although as late as 1840 about one-half of regional exports came from mineral ores, concentrates, and bars (especially silver) or specie, mining in the southern sierra appears to have undergone a long decay between the 1850s (after the boomlet in gold exploration in Puno's montaña) and the 1880s, perhaps because of lack of capital, scarcity of labor, and especially the remoteness of many deposits, resulting in exorbitant transport costs.^[49] The extension of the rail link from the port of Mollendo to the northern altiplano by the late 1870s, greater political stability after 1886, and the rapid fall of exchange rates between 1890 and 1893 led to an upswing of mining activity in the southern sierra. Between the early 1890s and 1907 British, American, and native capitalists made substantial investments in machinery for the extraction and processing of silver and gold ores in Arequipa's Cailloma province and in Lampa, Carabaya, and Sandia provinces in Puno.^[50] The high share of minerals in Mollendo's global exports during the 1890s was due largely to increasing Bolivian transshipments of silver, copper, and tin. But during the years 1902–4, when statistics for Peruvian and Bolivian exports through Mollendo are separated for the first time, gold, silver, and copper still contributed 17 percent to Peruvian exports from that port. Yet during the 1910s mineral exports declined. Gold was marketed mostly within Peru; a temporary turn to the domestic market might also account for the temporary decline of silver exports.^[51]

This broad range of secondary exports regularly contributed about one-third of foreign trade earnings. Not only did they buffer cyclical fluctuations in the wool trade, but they also extended the geographic reach of monetarized trade to diverse and distant districts, from the valleys of

Arequipa to its highland mining zones and the broad eastern escarpment of the Andes. This diverse range of exports strengthened the network of regional trade conducted by hispanized and foreign traders. Although regional commerce, outside of the circuits maintained by the peasantry, had atrophied during the long decades of declining markets between the late eighteenth century and 1850, it now rose to new vitality in close but complicated interrelationship with the export trades.

Imports, Domestic Production, and Regional Trading Circuits

Southern Peru's balance of trade with foreign countries conducted through Islay and Mollendo stayed positive for most years between the mid-1850s and 1919.^[52] Trade surpluses were especially impressive during the mid-1860s and the years of World War I but were large in most years since the 1890s. In part such surpluses flowed out of the region to Lima in the form of import duties and direct and indirect taxes. The south, like other provincial regions of Peru, helped to pay with its balance-of-trade surplus for the higher level of imports consumed in the capital, whose port, Callao, consistently garnered a share of imports much beyond its hinterland's share of national population and its export capacity. But part of these surpluses must have been retained in the south and contributed to the growth of expenditures for consumption or investment satisfied with regionally produced commodities.

"At least 60 percent of the population of Peru is practically negligible so far as the purchase of foreign goods is concerned," wrote William E. Dunn, the commercial attaché of the United States embassy in Lima, in 1925. In his opinion the Indian peasants, living in their "bare and desolate mud or stone huts in the Andean heights" and lacking "the remotest ideas of comfort in life," limited their purchases of manufactured goods to "a few cheap tools or an occasional novelty that strikes their fancy." They spent the "petty sums" they earned from the sale of wool and other products on the purchase of alcohol and coca leaves. Dunn viewed "mestizos of the lower class," with their "dilapidated houses, . . . furnished with only a few indispensable household articles," as only slightly better customers of imported goods. In fact, he thought that "the combined purchasing power of the Peruvian people might well be compared with that of the average American city of about 650,000." (Peru had about five million inhabitants at the time.) Demand for a broad range of upscale consumer goods, especially important for U.S. foreign trade by the early twentieth century, was concentrated in Peru's major cities, especially Lima.^[53]

Although Dunn's cultural prejudices are striking, there is no reason to accuse him of underestimating the Indian peasants' propensity to consume imported goods. It was, after all, his job to identify potential markets for U.S. products. Southern Peru, with its great concentration of Indian peasants, consumed relatively few imports in relation to its population. During the first two decades of the twentieth century the region, with close to one-third of the country's population, accounted for 10 to 15 percent of Peru's total imports.^[54] Without major urban centers the disproportion between population and consumption of imports was, of course, greater yet in the altiplano. As early as 1850, a few years after the terminal crisis of southern Peru's textile manufactories, the important trade fair at Vilque saw wholesale purchases of regional products for export of 490,000 pesos, while the sale of imported wares to altiplano traders and consumers amounted to only 300,000 pesos.^[55] And for the period between 1920 and 1935 the Arequipa export and import house of Guillermo Ricketts y Compañía "bought much [primarily wools for export] and sold little [imports]" in Puno.^[56]

Among the fifty-two different articles of wearing apparel imported into Peru in 1913, imports through Mollendo accounted for 10 percent or more of national imports for only fifteen articles.^[57] Ready-made imported apparel did not sell much in southern Peru, either because its high price limited demand to two or three retail stores in Arequipa, or because the cheapest grades of standard clothing items were already being produced in small domestic clothing manufactories, or because most of the "better" families in the provinces continued to rely on seamstresses and tailors for major items of their wardrobes such as dresses and suits.^[58] Similar problems were faced by a wide range of imported consumer goods in southern Peru. Domestic and artisanal production continued to supply many of the traditional items of consumption. At the same time, many items in the broadening range of commodities newly added to consumption patterns in southern Peru, especially since the 1890s, were produced by a growing number of small factories and artisanal shops in Arequipa and Cuzco or in Lima and other coastal towns.

To be sure, an ever-broadening range of imported goods did circulate throughout the southern highlands, integrating the remotest hamlet and the humblest peasant family into a commercial chain that had its other end in mighty factories in Bradford, Limoges, or Essen. As early as the 1830s and 1840s certain imported textile materials, hardware, glassware, and foodstuffs were consumed in altiplano provinces. Consumption patterns changed considerably after the coming of the railroad, and especially after the late 1880s.^[59] The range of goods available in altiplano stores and

markets expanded rapidly, and many of the newly offered items were imports. The dry-goods store of José Pantigoso Chavez on the Plaza Mayor of Puno town in 1858 carried a total of 74 items, of which 21 can be identified as likely imports, 36 as domestically produced, and 15 as either imported or domestic goods.^[60]

By comparison, in 1890 the Casa de Comercio de Efectos Ultramarinos y de Abarrote Moller and Compañía, also located in Puno town, carried 241 different articles, of which 148 were probable imports, 35 were domestic goods, and a further 58 were of uncertain origin.^[61] The more affluent of Azángaro's hacendados purchased imported furniture, household goods, apparel, and luxury food items. For example, in February 1873 the hacendado Manuel E. Paredes from Azángaro received special foods, beverages, and glass and china were from a general and dry-goods store in Puno. Of the bill, which came to 142 pesos, 5 1/2 reales, items amounting to at least 56 1/2 pesos corresponded to imported goods, among them Norwegian beer and Spanish canned fish.^[62] In the 1909 will of the hacendado Mariano Wenceslao Enríquez, long-time parish priest of Azángaro, appear such imported items as "one new French piano," a typewriter, and a sewing machine.^[63]

Indian peasants also spent some money on imported goods. In the early 1860s, probably the apex of the relative strength of imports in southern Peruvian markets, the English traveler Clements Markham claimed, with considerable exaggeration, that "almost all the woolen clothing of the Peruvian Indians is now imported from Yorkshire, and their shirtings from Lowell."^[64] Beginning no later than the last decade of the nineteenth century, peasants and poorer urban folk in the altiplano purchased such imported wares as basic tools (e.g., scissors and plowshares), needles, mirrors, and aniline dyes.^[65] For southern Peru as a whole, the composition of imports underwent a major shift between the 1860s and the 1910s. In 1863 at least 52.3 percent of total imports through Islay were textiles, but their share had declined to an average of 34.9 percent between 1913 and 1916. By the mid-1910s, 40.5 percent of imports through Mollendo consisted of metal goods, ceramics, glass, cement, timber, paints, oil, rubber goods, tools, and machines, a broadened range of consumer and investment goods difficult to disentangle.^[66]

But the weight of such imports in the overall expenditures of the altiplano's population, from community peasants to large, hispanized landholders, should not be exaggerated. With the exception of certain textiles, purchases of imports by hacendados or Indian peasants constituted occasional, extraordinary expenditures and did not belong to their day-to-day consumption. The items of the normal diet not only of peasants but also

of hacendados, when they did not stem from the livestock growers' own production, originated mostly in regions adjacent to the altiplano.^[67] Outlays for domestically produced maize, rice, noodles, flour, salt, fresh and dried fruits, and sugar occasioned considerable expenditures, particularly for the peasants. A great deal of income was spent on stimulants such as coca leaves and alcohol (aguardiente de caña for the peasants, pisco [brandy from Peru's coastal vineyards] and wine for the hacendados). Candles and fuel (if the dung of one's animals was insufficient), locally produced pots and silverware, and other domestically produced household items had to be purchased on a regular basis. Although the construction of a peasant's adobe hut required the purchase only of timber—not a cheap item in the woodless altiplano—the building of a hacendado's urban residence and, more generally speaking, investments in urban real estate swallowed up much money. Both hispanized hacendados and indigenous peasants invested savings from the sale of wool or other livestock products in land and livestock. There were expenditures for transportation, for the education of children in the case of more affluent landholders, and, last but not least, for numerous national, municipal, and church taxes and fees. In short, altiplano wool producers, after selling their wool to merchants, did not immediately turn around and spend all their returns on imported goods. This fact was obvious to Clements Markham, who wondered what alpaca herders did "with the enormous sums of money thus received." He suggested that they routinely buried such cash income.^[68]

Among the altiplano peasantry, burial may indeed have been a common method of saving money for large, special expenditures (e.g., baptisms, marriages, and funerals), a type of deferred consumption. More generally, the increasing revenue brought into the region by wool exports and associated activities stimulated regional trade with domestically produced goods. Since the 1850s the export of wools had become the lead sector of the southern Peruvian economy. As export earnings grew, demand for regionally produced goods rose. Wools were now fueling the economies of the southern highlands just as silver had during much of the colonial era. As Manuel Burga and Wilson Reátegui observe, wool exports helped to form an economic region that became dynamic through that trade.^[69]

What had changed since the mid-eighteenth century was the spatial definition of the economic region, the social composition of commercial networks, and the distribution of benefits from trade. The symbiotic relationship between the conjuncture for the major export commodity and the conjunctures for a broad range of regionally exchanged goods did not change, however. The strengthened foreign trade nexus had undermined or even destroyed specific processing activities and commercial flows by the

ERRATUM

In the title of table 5.6 (page 173) the units of measurement should read "metric tons," not "1,000 metric tons."

mid-nineteenth century, but over the next seventy years it did not have the strength or the explosive impact to completely eliminate trade in artisanal goods and other regionally produced processed commodities. On the contrary, indications are that the share of regionally or nationally produced goods among the total goods consumed in southern Peru rose after about 1870.

Take the crucial case of wools. As late as 1840 close to 60 percent of sheep wool and over a third of alpaca wool produced in southern Peru was not exported but was processed in the region, either on looms in peasant households or in the remaining obrajes. Most obrajes finally collapsed during the mid-1840s, a new wave of cheaper imported textiles entered the region, and the price for wool rose dramatically during the 1850s and 1860s. Because of these developments, the share of wools that was exported appears to have risen sharply. By 1867 only about 42 percent of sheep wool and 20 percent of alpaca wool may have been retained domestically (table 5.6), confirming the frequent complaints by British consuls that the exports of alpaca wools were limited by tight supply. But over the following fifty years this trend was reversed as livestock herds grew faster than wool exports. For the whole period 1830–1917, southern Peru's sheep population may have grown nearly three and a half times, whereas sheep wool

exports grew 2.8 times between 1840 and 1917. Alpaca herds may have grown more than four and a half times, with exports expanding by a factor of 3.7 between 1840 and 1917.^[70] Between 1867, when the ratio of wool exports to wool production probably reached its highest point, and 1917, sheep wool production may be estimated to have grown by nearly 70 percent, while exports increased by only 30 percent. The share of domestically retained wool thus was elevated to at least 55.9 percent in the case of sheep wool and a more modest 25.9 percent in the case of alpaca wool.^[71]

Two reasons account for the growing domestic consumption of raw wools. One is the opening of modern woolen mills in southern Peru that processed the regionally produced raw materials. The first opened in 1861 on Hacienda Lucre, site of an old obraje close to Cuzco, by the Garmendia family, a prominent family of large landholders since colonial times. After a mere twenty years of nearly complete prostration of elite-controlled textile production, this was the first step in the reversal of that trend. But southern Peru's modern textile industry grew very slowly. Between 1895 and 1910 two more mills began operations in Maranganí and Urcos, in the colonial centers of Cuzco's textile production in Canchis and Quispicanchis provinces, and one opened in Arequipa. The Maranganí factory, "one of the most progressive and up to date enterprises in South America," had modern English, German, and Belgian machinery installed. It relied on Indian community peasants and colonos as workers, something that remained true for all textile factories in Cuzco before 1920.^[72]

These factories were small in terms of capital, installed capacity, and work force in comparison with the woolen mills in Lima and especially with the cotton mills in the capital, owned by powerful foreign enterprises such as W. R. Grace and Duncan Fox and Company. The south Peruvian market for manufactured woolens was now disputed by three sets of producers: factories located in the region itself, in Lima, and in Europe. By the second decade of this century more nationally produced textiles were sold in the region than imports. And although the large Lima factories dominated the market for cotton goods, the southern factories were relatively strong in woolens up to the end of World War I, especially in highland departments such as Cuzco, Apurimac, and Puno, but briefly also in Arequipa. The Cuzco factories specialized in baizes, cashmeres, flannels, and blankets aimed at peasants and other downscale consumers, a market segment for which their lower transport costs and intimate knowledge of regional styles and marketing arrangements gave them the edge.^[73]

Nevertheless, according to one report, by 1918 Peru's factories still absorbed only some 680 metric tons of sheep wool, below 10 percent of national production and just under 20 percent of exports.^[74] Although this

industrial use of wools clearly contributed to the shifting balance between exports and domestic consumption, the surprising fact is the continued weight of wool processing in peasant households. Even during the height of the wool export boom toward the end of World War I, nearly one hundred years after the opening of direct foreign trade with England and more than fifty years after the installation of the first modern woolen mill in the country, cloth woven on the simple looms of the peasant households must have consumed between 45 and 55 percent of all sheep wool and close to one-fourth of all alpaca wool produced in southern Peru.^[75] Estimates are too rough to indicate with confidence whether the relative share of peasant household production in the disposition of southern Peru's total wool clip had begun to decline during the first century after independence, when consumption of wool by obrajes was replaced by wool exports and modern factory processing. But it is clear that with the strong growth of livestock populations the absolute quantity of wools processed in peasant households continued to increase until 1920. Although southern Peru's rural folks and the poorer strata of the towns had purchased certain imported textile items since the 1830s or 1840s and had begun to use domestically manufactured woolens since the 1860s, the attendant reduction in the per capita consumption of household-produced textiles was more than offset by the increase in the region's population that continued to rely on some homewoven fabrics. To reiterate this crucial point: the absolute amount of woolens produced in peasant households continued to grow during the century after independence, even if per capita consumption began to decline.

This increased output of homespun woolens would not have been unusual as a relatively brief transitional phenomenon accompanying the formation of an integrated domestic market in conjunction with the emerging dominance of the capitalist mode of production, as Emilio Sereni has shown for the Italian case. For a period of some thirty years following the onset of the formation of a large-scale textile industry in the 1860s, the processing of linen and hemp in rural households in Italy continued on a high level, and the number of looms in the countryside continued to increase even until the early 1890s. But the opening of Italy to massive foreign trade, the liberal policies of the resorgimiento , and, after 1880, a policy of industrial protection created a national market in which rural producers increasingly succumbed to the large factories in the northern cities, industry and agriculture became thoroughly separated, and the mezzogiorno was made into a "dependent territory" of the northern industrialists.^[76]

In the comparative perspective of Western Europe, Sereni considered the Italian transition toward a capitalist national market excruciatingly slow, held back by "feudal remnants." But the Peruvian case was quite different.

Before 1930 no national market emerged; foreign trade and modern industry, rather than functioning as battering rams bringing down the walls of southern Peru's traditional modes of production and exchange, accommodated themselves to regional interests; agriculture and industry remained highly linked, and domestic household production grew along with foreign trade and modern industry, whose capacity to expand thus remained limited.^[77] Low productivity in agriculture and artisanal production, as well as the neocolonial structure of the society, made the southern Peruvian highlands resilient to the forces of change.

The money circulating in southern Peru through export activities thus stimulated a broad range of domestic production and processing for regional trade without drastically changing the mode by which these commodities were produced. The demand for many of these commodities moved together with the export economy. Maize in Cuzco became scarce in 1917 when demand from the core livestock area between Sicuani and Lake Titicaca rose in conjunction with the wool export boom.^[78] Sheep and cattle on the hoof from Puno and maize from Cuzco briefly encountered strong markets in Chilean-occupied Tarapacá during the 1910s, when that region's nitrate export peaked.^[79] Since the 1850s the consumption of grape alcohol from the valleys of Moquegua and around Arequipa had increased in the rural areas of the altiplano. After the railroad facilitated access from the coast to the southern highlands, cane alcohol from the rapidly modernizing north coast sugar estates pushed out the southern grape alcohols, and peasant consumption of aguardiente de caña grew in close relation to conjunctures of the wool market.^[80]

Besides the various woolen goods, a broad range of artisanal products found strong demand throughout southern Peru. Artisans from the altiplano and workshops in Arequipa or Cuzco did not merely continue to produce standard items of long-standing demand but adjusted their production to shifting urban consumption patterns. Potters in Santiago de Pupuja and Pucará diversified from the plain jars and pots used for cooking in most of the altiplano, creating vases, ashtrays, and ornamental pieces sought after by urban middle-class families. Leather workshops, especially in Arequipa, also adjusted their production to new urban demands in footwear, clothing accessories, and household furnishings.^[81]

Southern Peru took on its contours as a distinct trading region during the second half of the nineteenth century as a consequence of this articulation between foreign and regional trade. From Desaguadero to Abancay, from Huancané to Camaná, and from Quillabamba to Moquegua, dozens of trading circuits transmitted the impulses from the export activities to remote valleys and mountain slopes that produced commodities for the

regional market. While regional interchanges intensified along with foreign trade, commerce with Bolivia, which had been the major pole of the erstwhile colonial pattern of exchange, lost in relative weight. Southern Peru became a distinct region for which the links with Liverpool were more important than those with Lima. Since the early 1870s the railroad line connecting the port of Mollendo and the entrepôt Arequipa with the highland zone had become the backbone of this trading region, erecting new commercial hierarchies, fostering new urban centers, and relegating others to a marginal position within those hierarchies. The creation of a new spatial hierarchy in southern Peru's commerce as a result of the rise of the wool exports and the construction of a modern transport funnel went hand in hand with the renovation and intensification of social hierarchy in trade: the growing centrality of the import and export merchants, benefiting most from the advantages of improved means of transportation and communication; the establishment or, in some towns, the expansion of several layers of middlemen, from wholesalers, regional bulkers, and owners of well-stocked general stores in important centers along the rail line to itinerant peddlers and wool purchasers; and the dependent incorporation of mestizo and Indian traders into the bottom rungs of these hierarchies. As Gordon Appleby has argued persuasively, trade in southern Peru became organized into a "dendritic" model, akin to a tree in which the life of even the remotest twig in the crown depends ultimately on the main stem.^[82]

But this was not the whole story. Much of the regional trade continued to flow outside of the channels of this "dendritic" system, even if it did not escape its influence in terms of demand fluctuations. Many points of conflict and tension marked the attempts by merchants entrenched in the export- import hierarchy to impose their trade routes, their prices, and their commercial intermediation on peasants, muleteers, and others who continued to ply older, more autonomous trade circuits based on intraregional complementarity. The conflicts erupting in southern Peruvian trade since the late nineteenth century were not primarily between groups attempting to impose foreign trade against the maintenance of intraregional exchanges or vice versa. Especially in the altiplano there were few traders who did not have a foothold in both types of activity. At stake was the social distribution of benefits within a regional trading system made up of heterogeneous elements.

Two studies on other Andean regions have emphasized that between the 1870s and 1880s regional commercial circuits were disrupted through the corrosive effects of economic liberalism and war. Tristan Platt has pointed to the "internal market" for maize and wheat in Upper Peru and Bolivia up to the 1860s. Cochabamba and, after independence, Chayanta supplied

the northern Bolivian altiplano and adjacent regions in southern Peru with these staples. The community peasantry of Chayanta played a prominent role in the "internal market" as long as the Bolivian state pursued a protectionist trade policy and maintained its alliance with the Indians symbolized by the tribute nexus. This "internal market" was destroyed within little more than a decade, when between the late 1860s and early 1880s the Bolivian national elite turned to a free-trade policy, allowing Chilean grains to flood the national market, and simultaneously attempted to institute a liberal land policy, which, had it been successful, would have destroyed the Indian ayllu and replaced it throughout the republic with large estates.^[83]

The other case concerns the central Peruvian sierra, analyzed by Nelson Manrique. Since the 1840s a dynamic, relatively autonomous regional market had developed, based on the rising demand for livestock products in Lima and an intensive trade in locally produced cane alcohol. This regional market conjuncture, in contrast to the picture drawn by Platt for Chayanta, primarily benefited large landholders and merchants. It entered into crisis through the destruction and social mobilization wrought on the region by the War of the Pacific. As most families of the central sierra's regional oligarchy saw their properties destroyed, sources of credit withdrawn, and markets challenged, they lost control over the regional economy first to better-capitalized businessmen from Lima and, after 1900, to "imperialist capital" taking over the copper-mining industry. In Manrique's view the formation and demise of the central sierra's regional market were inevitable steps toward the formation of Peru's national market.^[84]

Southern Peru underwent a different development. After painful adjustments to the displacements of certain domestic goods by imports between the 1780s and 1850s, the regional market became tied to the growth of foreign trade. The liberalization of imports during the 1850s and 1860s helped to establish foreign trade as the strategic, lead sector of southern Peru's economy, favored the formation of new mercantile hierarchies dominated by the foreign houses of Arequipa, and in this way fostered the growth of the "dendritic" trading pattern that defined the south as a region in and of itself. But this trade liberalization did not lead to a general crisis of intraregional trade, as Platt has described for Chayanta. The crucial difference between the two cases concerns the commodities involved and the producers. In Chayanta the strategic products for peasant participation in the "internal market"—cereals—could easily be replaced by imports once the tariff and transport conditions favored such replacement. In southern Peru the strategic export commodity was produced to a considerable part by peasants. The same activity of livestock raising that involved the peasantry in the export nexus simultaneously produced a broad range of goods for regional trade, from wool and hides to meat, tallow, butter,

and cheese. Wool bulkers could hardly have wanted the peasants to cease barter and trade in such commodities, as it would have made them totally dependent on income from wool sales for export, inevitably leading to demands for higher prices for peasants' wool.

Other commodities from different parts of southern Peru, such as coca leaves, tea and coffee, cane and grape alcohols, chile peppers, raisins, and olives, had little to fear from overseas imports. Between 1888 and 1897 higher import tariffs and exchange rate devaluation led to increased levels of effective protection, but this situation did not automatically benefit regional trade controlled by peasants; rather, it led to modest import substitutions through a few new factories, such as the woolen mills discussed above, iron foundries, and breweries, which merely shifted the origin of manufactured goods confronting some of the peasants' domestic goods. Industrialists and merchants were either identical or the former were highly dependent on the latter before 1920. For this reason the national debate over import tariffs during the 1890s does not appear to have split the region's business elite. Until the end of the World War I export boom most members of that elite, in Cuzco and Puno as much as in Arequipa, remained free traders.^[85]

The Altiplano's Commercial System Between the 1850s and 1920

Until the late 1850s the export of wool through Islay was controlled by only four companies of foreign origin operating from Arequipa—two British, one French, and one German.^[86] For them, the import of manufactured goods from Europe was as important as their export business. In both activities the foreign houses still operated mostly in Arequipa itself. As importers they worked as wholesalers; as exporters they bought wool from middlemen directly in Arequipa or made a yearly venture to one of the annual fairs on the altiplano.^[87]

In these circumstances independent Peruvian merchants had a considerable role to play in all stages of wool bulking prior to final exportation. Affluent men such as José Mariano Escobedo, an Azangarino resident in Arequipa, or the Areqipeño José María Peña often entered the wool trade only as one line of business among many.^[88] Escobedo owned a number of haciendas in Azángaro and received large public works contracts from the government. In 1851 he began the wool trade as a partner of the German merchant Guillermo Harmsen.^[89] Peña had owned gold mines in the Cordillera de Carabaya since before 1850s.^[90] Sometime before 1865 he had formed a company with Eusebio Prudencio, a Bolivian, "for the purchase of chinchona bark and alpaca wool and the sale of merchandise in Soraicho,"

located in Huancané province from where the trade with the Bolivian lakeshore as well as the ceja de la selva could easily be organized. Prudencio was to take care of business in Soraicho while Peña arranged for the sale in England as well as the purchase of Peruvian and imported goods finding demand on the eastern rim of the altiplano.^[91] Peña probably held similar contracts with traders from other wool-producing regions, and his mercantile operation spanned the whole altiplano.

Merchants with the necessary capital joined together with persons who could ensure a supply of wool from their own extensive estates. This was the case in the company founded by Manuel and José María Costas and Antonio Fernández, all of Puno, on April 15, 1853.^[92] Fernández saw himself incapable of paying his share of the company's capital, so it was paid in by the Costas brothers as credit. The Costas brothers found it advantageous to join in a common enterprise with Fernández because he was owner and renter of numerous livestock estates and promised "to cede to the Society all sheep wools proceeding from his haciendas." These haciendas happened to be located west of Puno in the geographic funnel of the altiplano leading to the main mule road to Arequipa and thus could well serve as warehousing, washing, and packing points for the wools purchased by the company prior to their transport across the western cordillera.^[93]

Only six years after the founding of the company Fernández admitted owing his partners the amazing sum of 53,832 pesos, 4 reales, a sum corresponding to more than 10 percent of the total value of average annual sheep wool exports from Islay during the 1850s and more than the value of the largest altiplano estates during those years. This debt represented cash withdrawals from the company's funds as well as sales on Fernández's own account of sizable quantities of company sheep wool as payment of private debts to the Arequipa export houses.^[94] The size of such transactions suggests that a relatively small number of Peruvian wool traders, either individuals or companies, must have supplied a large share of all wool subsequently exported by the Arequipa houses. Despite the heavy debts incurred by their early partner Fernández, the Costas family continued its wool-trading business at least until 1925.^[95]

Before the 1870s the tendency of the foreign houses to limit their mercantile operation largely to Arequipa and the concomitant strong position of Peruvian merchants such as Escobedo, Peña, and Costas also had consequences for the structure of southern Peru's credit network. During the 1850s and 1860s altiplano wool traders and hacendados depended on the Arequipa export houses for credit—for example, for their wool purchases or investments in real estate—considerably less than they would toward the end of the century. The large regional wool bulkers apparently

possessed sufficient working capital not to require advances from the exporters for purchasing large amounts of wool; thus, they could operate independently and decide when and to whom they would sell their wools. It seems to have been these regional bulkers who extended credit to their suppliers.

Another important source of credit existed in the altiplano during the third quarter of the past century. A small group of merchants, who partially overlapped with the regional wool bulkers, had become something like specialized bankers extending credit to dozens of shopkeepers, landholders, and magistrates. Antonio Amenábar was one such "merchant banker." Born in Córdoba, Argentina, between 1824 and 1826, he pursued a great variety of business ventures in Puno and Tacna, including transportation (he invested in the first metal cargo ship operating on Lake Titicaca) and collection of government taxes.^[96] In 1865 his Puno stores contained goods, including imports, worth 40,839 pesos. At the same time he had outstanding credits, extended since 1860, amounting to 46,675 pesos (including interest). Among his fifty-four debtors, who owed him between 70 and 6,000 pesos principal besides interest, were members of the department's wealthiest and most powerful families—the Macedos, the Pinos, the San Romans, the Núñez, the Aguirres, the Tovars, and the Arésteguis. Some of them owned haciendas in Azángaro province.^[97] Nine years later, in 1874, his will no longer listed merchandise from a store, which by then was possibly the property of his wife through a legal "division of goods."^[98] But he still had outstanding credits amounting to 36,500 pesos, extended to twelve clients since 1871.^[99]

On a smaller scale, some provincial traders also seem to have attempted a measure of specialization into such credit operations. A case in point is Pedro Palazuelos, hacendado and trader from Putina in Azángaro province. After his death, sometime before May 1865, his wife and son attempted to collect some 5,800 pesos in outstanding credits extended to nine clients, most of whom were prominent Azangarinos. Besides further, probably minimal debts by another nine clients, Palazuelos's heirs also tried to recover rented livestock and 6,000 pesos given as advances to Indian peasants for purchases of wool and slaughter animals.^[100] The amount of general-purpose monetary credits extended by Palazuelos and the number of his debtors justify calling him a specialized creditor. But the borderline to the more general phenomenon of small money loans, practiced by many landholders and traders in the altiplano, seems blurred.

The scarcity of cash threatened vital routine commercial transactions for many families. A dense network of small, reciprocal credit transactions, organized through ties of kinship and trust among friends, sprang up in

the towns and hamlets of the altiplano in an era of increasing trade. Such petty credit exchanges could shift imperceptibly into more hierarchical, asymmetric credit relationships, constructed by provincial traders who had somewhat more cash and were eager to profit directly from the cash scarcity or to build up a dependent clientele for mercantile endeavors.

With the increase in southern Peru's exports of wool and other raw materials, demand for transport animals rose during the middle decades of the nineteenth century. At least until the late 1860s a lively trade in mules from Salta kept up the supply of these animals. In June 1857, for example, the merchants Juan Bautista Coret and Telesforo Padilla from Salta and Simón de Oteira from La Paz passed through Puno with some sixteen hundred mules, which they hoped to sell on their march to Lima, a good part of them probably in the vicinity of Arequipa.^[101] In Puno, de Oteira purchased another 540 mules from Fernando del Valle, a local hacendado and "merchant banker."^[102] It appears that del Valle regularly acquired mules in order to let them regain their strength on his estates in Acora after their long trek from Argentina. He resold some of these animals to passing traders and kept enough for his own trains, with which he transported pisco from Moquegua to La Paz.^[103]

During the third quarter of the nineteenth century the transport business concentrated in Arequipa and adjacent valleys as this city became the entrepôt for southern Peru's exports. Markham considered Arequipa's muleteers, whose increasingly numerous mule herds had taken over much of the city's fertile campiña from food crops, to be a "wealthy class of men" during the 1860s.^[104] Of course, these mule trains did not enjoy a monopoly over southern Peru's transport requirements. Peasant landholders and hacendados transported much of their products on their own llamas. Specialized transport entrepreneurs in the altiplano working with llamas also continued to ply the trans-Andean routes.^[105]

One of the reasons why the foreign merchants limited their activities largely to transactions in Arequipa itself lay in the extreme difficulties of transport between the coast and the altiplano. In the words of one foreign traveler, "Most of the roads are merely mule tracks and they are taken over passes of the Andes from 14,000 to 17,000 feet above the sea amidst snow and ice."^[106] The regularly established mail covered the distance from Arequipa to Puno in three days, and pack trains needed at least five days for the same route.^[107]

Between about 1860 and 1875 changes in the commercial structure of southern Peru began to affect the balance between the various social groups involved. The boom of wool exports and generally favorable commercial conditions during the 1860s led to the establishment of many new export

houses.^[108] As early as 1862 the mercantile community of Arequipa and Islay was clamoring for a railroad connection between the Ciudad Blanca and its port.^[109] On December 18, 1869, the Peruvian government commissioned Henry Meiggs, an American engineer involved in the nitrate business, to construct a railroad line from the Pacific coast via Arequipa to Puno. Choosing a termination point a few kilometers south of the established port of Islay, the line connecting Arequipa with Mollendo was opened for traffic on January 1, 1871. By 1873 the whole line of 351 kilometers to Puno, crossing the western cordillera at an altitude of about forty-five hundred meters at Crucero Alto, was completed.^[110] From Puno regular boat services across Lake Titicaca connected this line with Bolivia, and by the early years of this century, a number of cargo ships had established regular links with a whole string of small ports, including Moho and Huancané in the north and Ilave, Juli, Yunguyo, and Desaguadero in the south.^[111]

The northern section of the altiplano began to be integrated into this modern transportation funnel to and from the coastal port with the construction of the railroad line ultimately linking Cuzco to the Puno-Arequipa line. Meiggs received the government contract to build this line in December 1871. Branching off from the previously constructed line at Juliaca, about forty kilometers north of Puno, it pursues a northerly course and for some sixty kilometers straddles the border between Lampa and Azángaro provinces. This course, necessitated when powerful citizens of Lampa rejected the track's passing through their town,^[112] gave rise to four railroad stations—Calapuja, Laro, Estación de Pucará, and Tirapata—conveniently located for large parts of Azángaro's livestock-raising zones. By the mid-1870s, before the financial crisis halted further construction, the railroad line had been completed for 131 kilometers to the town of Santa Rosa, close to the northwestern rim of the altiplano. All of the Peruvian altiplano had now been put into reasonable proximity of rail or boat transport; except for isolated hamlets in both cordilleras most locations lay within two days' journey on muleback from railroad stations or ports.^[113]

Traffic on the line remained thin until the 1920s. One passenger train a day traveled in each direction between Arequipa and Mollendo, and only two per week ran between the entrepôt city and Puno, covering the distance of 350 kilometers in ten to twelve hours. Freight trains were no more frequent, and schedules changed often in accordance with traffic conditions. These frequencies remained basically unchanged until the mid-1920s.^[114] The railroad operated at a loss during its first years. It was built not to accommodate already existing trade but to produce its own business through a major reduction in transport costs and an anticipated expansion

of trade. Such expectations by the government and southern Peru's business elite proved optimistic at best. In contrast to Mexico, where the shift from a slow and inefficient transport system to railroads contributed to strong economic growth,^[115] southern Peru's growth of trade and production remained relatively modest. The modernization of the transport system failed to transform the relations of production in the countryside or to deepen the market. Railroad freightage grew primarily by concentrating commercial flows through the consolidation of the dendritic system, drawing business from muleteers. Railroads accelerated economic growth in Latin America most when the commodities to be moved were bulky items, such as minerals or grains, and when there were virgin resources (land, major untapped mineral deposits) in the production zone accessed by the rail link. Neither of these conditions prevailed in the case of the altiplano. Moreover, since transport costs represented only a small part of the FOB values of wools—no more than 15 percent by one estimate—a reduction in freight rates through the switch from mules and llamas to rail could produce only small savings for the producer.^[116]

Paradoxically, when trade flourished, the modern transportation funnel rapidly reached its full capacity. British consuls repeatedly complained about bottlenecks caused by shortages of boxcars, the inefficient steamer service across Lake Titicaca, and insufficient warehouses in the port of Mollendo.^[117] The Peruvian Corporation, the British proprietor of the Ferrocarril del Sur and the lake steamers since 1890, apparently adapted to southern Peru's business environment by keeping transport space scarce.^[118] Yet railroad freight rates appear to have been relatively low compared to those of other lines in Peru.^[119]

Nevertheless, traders and livestock growers in the altiplano frequently complained that the Peruvian Corporation's transport rates favored imports and Bolivian transit commodities over goods from the Cuzco and Puno highlands shipped to the coast.^[120] During the first forty years the railroad, to be marginally profitable, depended on cargo to and from Bolivia, especially ores but also wool from the departments of La Paz and Oruro as well as imports for the urban market of La Paz. Despite wool's great importance for the regional economy, shipments of it from the Peruvian production areas, the core zone of the railroad's operation, never brought enough business for the down trips to break even. When the new La Paz–Arica line opened in 1914, Bolivian business on the Ferrocarril del Sur was sharply reduced, with "catastrophic effects" for the line's profitability. To compensate, in 1919 the Peruvian Corporation began a series of rate increases in excess of rising costs. In the following decade the company attempted to increase the tonnage of wool shipped by railroad from the altiplano to Mollendo. Until the mid-1920s it unsuccessfully

sought to establish a huge wool-production syndicate under its control in Puno and Cuzco, an undertaking meant to ensure increased wool shipments on the rail line. As late as 1932 the Ferrocarril del Sur drew only 7.4 percent of its total freight income from the transportation of wool.^[121]

Llama and mule transport was not replaced by the railroad overnight. Pack trains were still needed to take wool and other livestock products from the estates and offline urban bulking centers to the warehouses at railroad stations and lakeside ports and bring back domestic and imported supplies, hauled up to the altiplano from the coast or Arequipa by railroad. Nor did transport animals disappear immediately from the old mule paths crossing the western cordillera. As late as the 1920s as much as 20 to 25 percent of exported altiplano wool was transported to Arequipa by llama or mule trains.^[122] In particular, many owners of small and medium-sized haciendas, who wanted to benefit from direct sales to the export houses in Arequipa but were in no position to profit from day-to-day fluctuations of wool prices, preferred to use their own or their shepherds' llamas as inexpensive means of transport.

The new extractive enterprises that developed in the department of Puno beginning in the 1890s—gold mining in the cordillera and piedmont of Carabaya and rubber collection further down in the rain forest—brought muleteers and llama drivers additional business on the roads and trails that fed into the railroad line. The supply of mining and rubber companies in Carabaya and Sandia provinces proved profitable both for altiplano estate owners, who dispatched their own transport llamas or those of their colonos, and for professional muleteers, mostly from Arequipa, who plied these new circuits with strings of up to a hundred mules.^[123]

After the mid-1890s, construction of improved roads, passable by four-wheeled vehicles, began in the Peruvian altiplano. The foreign-owned Inca Mining Company constructed a road from the Tirapata railroad station toward its gold mines at Santo Domingo on Río Inambari. For over ninety kilometers this road passed through Azángaro province, connecting its northern and eastern districts (Asillo, San Antón, and Potoni) more comfortably with the rail line.^[124] Public road construction projects connected the capital, Azángaro, to the railroad station Estación de Pucará (a distance of some thirty kilometers) as early as 1896, and Azángaro with Asillo (about twenty kilometers) by 1916. During the 1920s the eastern part of the province around Muñani and Putina was incorporated into the road network both by a link to Azángaro town and through a more southerly direct connection to Juliaca via Huancané (see map 1.1).

By the late 1920s Puno's road network, with about two thousand kilometers of improved roads completed, was the most extensive in the whole republic. One observer explained this comparatively rapid progress as the

consequence both of the favorable terrain—the broad pampas of the altiplano—and the abundance of Indian labor pressed into the heavy chores of road building by local authorities and estate owners. These factors were important, but the new roads also underscored the departmental elite's influence in national politics during the first third of this century. Paradoxically, although the condition of altiplano roads was described as "excellent" as early as 1913, there still hardly existed any motorized vehicles. Only during the 1920s did trucks begin to assume an important role in transport to and from railroad stations. Just as in the case of railroads, construction of roads largely preceded effective demand.^[125]

In spite of its modest impact on the scale of trade and on relations of production, the modernization of the transport system has to be considered the most important motor behind changes in the spatial patterning and the social and economic structure of trade in the altiplano.^[126] Previously the role of towns in the bulking of wool and other commodities for export, as well as in the distribution of imported goods, had been relatively small since in a "landscape of uniform, primitive transportation" much wool could be transported from the producer to the entrepôt city, Arequipa, without being handled in towns, even though it passed through several layers of middlemen.^[127] The switch of the predominant transport mode from llama and mule train to railroad and lake steamers brought about the rise of commercial centers, strategically located on the rail line or the lakeshore, where merchants handled the transshipment of goods flowing in both directions through the funnel. Puno, Juliaca, and Ayaviri benefited most from this new spatial pattern. Other important commercial centers owed their very birth to the railroad, among them Estación de Pucará and Tirapata in Azángaro province.^[128] Some urban centers that did not share the advantage of being located on the rail line or the lakeshore managed to hold on to and even intensify commercial activity. This was true of Azángaro town and Putina in the eastern part of the province, which remained secondary centers of wool bulking because of their distance from railroad stations or ports.^[129]

A growing number of traders, shopkeepers, and commercial establishments now set up business, and the penetration of trade by the Arequipa export houses intensified. These changes were already becoming noticeable by the mid-1870s, though the crisis and dislocation caused by the War of the Pacific retarded their full development until about 1890. Some of the new traders and shopkeepers had long been residents of the department of Puno; many others came from Arequipa, Cuzco, Tacna, and other parts of Peru, and quite a few came from overseas. Except for the entrepreneurs and engineers at the newly opened mines, foreigners settled in the largest

towns, principally Puno and—after the turn of the century—Juliaca, rapidly becoming the major wool-bulking center in the department because of its strategic location at the hub of the new transport lines. Combining wholesaling and retailing activities, many foreigners established well-stocked stores with adjacent warehouses where they sold imported and Peruvian merchandise and purchased "country products," mostly wool and hides but also gold, rubber, and coffee.^[130]

Rodolfo Möller may stand as an example of foreign retailers in Puno. His Casa de Comercio de Efectos Ultramarinos y de Abarrote Möller y Compañía carried a great variety of foreign and domestic merchandise: ten bottles of "eau de Cologne, best grade," some hundred cans of different aniline dyes, 26 cases of Lion brand Norwegian beer, 27 varas of white narrow cotton flannel, 221 packages of white thread for "Chardio" sewing machines, kerosene from Tumbes, paper, silverware, scissors, and knives.^[131] Möller had bought most items on credit from the import and export house of Enrique W. Gibson y Compañía in Arequipa.

The Peruvians, from both inside and outside the department of Puno, who had swelled the ranks of altiplano traders and shopkeepers since the 1870s spread more evenly through the towns of the region. In Azángaro newcomers settled especially in the capital of the province, some of the larger district capitals such as Asillo and Putina, and, of course, at the railroad stations Tirapata and Estación de Pucará, where warehouses and shops sprang up around the railroad tracks on what previously had been pastures.

Many of the newcomers first appeared in the altiplano as muleteers or itinerant traders from Arequipa or other surrounding regions with which Puno traditionally had maintained an active commercial interchange. One such man was Manuel Sixto Mostajo Yáñez. His father, an itinerant peddler from Arequipa, had frequently passed through Azángaro on his sales trips to Sandia, the piedmont region northeast of the province. Manuel continued this business. Sometime during the 1890s he married Victoria Enríquez, who belonged to an old hacendado family from Azángaro and was the niece (or even daughter) of the town's long-time parish priest, Mariano Wenceslao Enríquez. Mostajo now established his residence in the provincial capital, built—apparently with financial aid from Father Enríquez—a large house on the Plaza de Armas, and established a store, where he sold such goods as cloth, window glass, and the like.^[132] He became wool-buying agent for the Arequipa export house of Ricketts and in 1903 received a contract from the Peruvian postal service to transport baggage and parcels from Azángaro to Sandia.^[133] As early as 1901 Mostajo was so well integrated into Azángaro's provincial society that he became treasurer of the town's municipal council.^[134] But he maintained his itin-

erant trading expeditions to the ceja de la selva and to the annual fairs at Pucará, Rosaspata, and Cojata until the end of his business career. In 1930 Mostajo ordered for such a peddling circuit items including woolen shawls, cheap felt skirts, cheap hats, Italian borsalino hats, cotton cloth, thread, aniline dyes, and even unassembled bicycles.^[135]

Among long-time residents of Azángaro a growing number of hacendados diversified into wool trading and shopkeeping just as the Paredes family had done since the 1840s. Around 1890 the half brothers Bernardino Arias Echenique and José Sebastián Urquiaga inherited Hacienda Sollocota, a small livestock estate of colonial origin, around which they built a vast landholding complex. At the same time they went into business as alcohol merchants in Azángaro town, and at least Urquiaga also operated as a wool trader.^[136]

Owners of strategically located haciendas were able to use their locations to purchase the production of smallholders in the surrounding area. This tendency was already observed by the Italian naturalist Antonio Raimondi in 1864 regarding an estate in Azángaro's district of Potoni, on the slope of the Cordillera de Carabaya: "Hacienda Potoni has as its object the collection of the sheep wool of the surrounding countryside and of the much more valuable alpaca wool produced in the immediately neighboring province of Carabaya."^[137] In 1904 Mariano C. Rodríguez, a landholder and trader from Rosaspata, Huancané province, offered to sell altiplano products to Ricketts y Compañía in Arequipa. He explained that his Hacienda Huaranca Chico was located close to the Bolivian border and that there, on the estate, "one can easily buy all these products [wool, rubber, and chinchona bark]."^[138] Access to smallholder production was a more crucial qualification for altiplano traders than was formal knowledge of commercial operations.^[139] Rodríguez explained to Ricketts that his family used the income from livestock ranching "to comfortably supply our needs." For his business he spent no money on rent and little on wages because "we have Indians who work for us nearly free of charge. In this way I can utilize the profit from the business to capitalize [it] and to attend to unforeseen business losses."^[140] While the operation of livestock haciendas was to provide the income for a comfortable life-style, Rodríguez hoped to use his commercial business as a source of capital accumulation.

Below the ranks of mestizo merchants and specialized itinerant traders flourished a complex and fluid world of peasant barter and trade. Exchange between altiplano peasant herders and agricultural producers of different ecological levels had been an important part of the structure of Andean society since prehispanic times. Trips by owners of livestock estancias from Azángaro to the ceja de la selva regions of Carabaya, Sandia, or the adjacent

valleys of Bolivia ordinarily served the purpose of exchanging wool, dried meat, hides, and other altiplano products for coca leaves, maize, and medicinal herbs for the herders' own consumption. But it was only a small step to trade more of such goods than one needed for family subsistence, particularly when the terms of trade developed favorably for livestock products.^[141] If such a trip went well, the peasant from Azángaro could profit from the barter or sale of his or her products in the ceja de la selva and again on return to the altiplano from the sale of the surplus maize or coca leaves.

Tomás Lipa, a peasant from Putina, seems to have engaged in this kind of trade, frequently making trips to Caupolicán province in Bolivia. In 1888 he rented one sector of estate pastures of Hacienda Chamacca, district Azángaro, and promised the owners, Josefa González and her husband Lorenzo Aparicio, to take one mule loaded with goods on their account with him on each trip to Bolivia.^[142] Nineteen years later, in 1907, his son Pablo Lipa extended a credit of 1,200 soles to the notary Filiberto Aparicio González, son of Lorenzo, who then handed the whole estate over to Lipa as security for a contractual period of four obligatory and five voluntary years.^[143] Trade had permitted the younger Lipa to accumulate this large amount of cash and, with it, the temporary possession of a hacienda. Other peasants found an avenue to trade by opening tiny stores in their rural communities—selling a few pounds of coca leaves, sugar, and so on—or by selling to urban bulkers not only their own wool clip but also that of relatives and neighbors. As Benjamin Orlove has noted, such bartering and trading peasants "shade[d] imperceptibly into the traveling buyers."^[144]

As trading networks in the altiplano became denser, the Arequipa export and import houses strengthened their position vis-à-vis the other groups involved.^[145] The marketing of Peruvian wools overseas continued in the same manner for the whole period under consideration. Sheep wool was sold for the Arequipa export houses at auctions in London and Liverpool, whereas alpaca wool was disposed of "through the even more archaic system of private deals between handlers and manufacturers" in England.^[146] In contrast, by the turn of the century Australia and New Zealand had established national wool auctions, and Argentina had developed an intermediate system in which export houses shipping directly to French manufacturers, purchasing agents of European houses, and consignment agents competed with each other.^[147] In the Peruvian system prices were least responsive to local supply conditions, whereas the position of the exporters vis-à-vis the producer was strengthened.

Improved transport conditions and fast communication links (the telegraph) between the coast and the altiplano permitted the Arequipa houses to bring to bear their advantages—greater capital resources, links with the

European wool importers, and up-to-date information on prices and market conditions. Since the 1870s and particularly during the economic recovery after the War of the Pacific, companies such as Gibson, Stafford, Ricketts, and Braillard had established purchasing and sales outlets in the commercial centers and at key railroad stations of southern Peru's wool production zone, from Desaguadero in the south to Sicuani in the north and from Conima (Huancané province) in the east to Santa Lucía (Lampa province) in the west.^[148]

Specific arrangements of market penetration varied greatly. In the most important centers the export houses opened branch offices. Some traders in the altiplano became their exclusive agents, handling all their sales and purchases through one Arequipa house. Among independent traders some concluded long-term contracts with the Arequipa houses, whereas others concluded contracts with varying export firms in Arequipa. Branch offices, agents, and large independent wool bulkers depended on three sources for their fiber purchases. Owners of small and medium-sized haciendas frequently offered their clip to the agents on a yearly basis and delivered it to agents' shops and warehouses on their own pack animals. Many peasant smallholders and colonos brought small amounts of wool to traders' shops on an irregular basis. And last, hundreds of itinerant traders, usually working firmly established circuits, combed the weekly markets in district capitals and communities as well as the yearly fairs for as much wool as they could find or could afford to buy and then sold it to the agents. Many small traders worked for only one agent and one export house.^[149]

The owners of the largest haciendas preferred to deal directly with the Arequipa companies. José Guillermo de Castresana, an Arequipeño businessman and, from 1906, owner of Hacienda Picotani in Muñani district, regularly instructed the administrator of his estate when to expedite the wool clip to the railroad station Estación de Pucará and on to Arequipa.^[150] Through social ties with one or the other export house, Castresana possessed information on wool price fluctuations and adjusted the transport schedule of the wools accordingly.

The Arequipa houses regularly extended credit for wool purchases to their agents and independent traders, who did likewise with their suppliers. In 1920 the roughly ten wool-buying agencies in Santa Rosa every week received between six thousand and fifteen thousand soles in coins from their respective export houses in Arequipa.^[151] By advancing a large share of the total value of wool, the traders hoped to secure supply. Wool buying was highly competitive on each level, between the Arequipa exporters, between agents, and between itinerant traders. The Arequipa houses attempted to gain complete control of local producing zones and were willing

to have their agents pay considerably higher prices for wool to the itinerant traders during an interim phase in order to ward off competition.^[152]

The Arequipa houses exercised the predominant influence over short-term wool prices in the altiplano. Every week they cabled price quotations to the traders there, refusing to purchase any wool above that figure. The trader then calculated the cost of handling and freight as well as his profit margin to determine the price that he could afford to pay his suppliers. Although price quotations generally had to follow the prices established at the London wool auctions, the exporters could afford to vary the margins between the world market quotation and the price at which they purchased wool, particularly in a declining market.^[153] In 1921, for example, the Sociedad Ganadera del Departamento de Puno, organized by hacendados at the height of the postwar crisis, complained that the prices for wool paid to them by the Arequipa merchants not only had declined precipitously since the boom years of the war but were even low compared with the prewar levels—for old clients, eleven pesos per quintal compared with thirty-five pesos before the war.^[154] Yet prices paid for average Peruvian sheep wool at a British port stood at nearly the same level in 1921 as they had in 1913. The key to the exporters' strong position vis-à-vis their suppliers lay in their relatively plentiful working capital, which permitted them to accumulate stocks in their warehouses. Thus, as Appleby has observed, "In a falling market the Arequipa house might either suspend purchases or set extremely low prices until it could dispose of its stocks of high-priced wools."^[155]

Producers and small traders were not totally defenseless against the price dictates of the exporters. Lively competition between houses such as Gibson, Ricketts, and Stafford permitted them to sell their wool to the highest bidder.^[156] Nevertheless, many hacendados and wool traders preferred to maintain lasting commercial ties with one house because the exporters could bestow a number of advantages on long-time clients, including better credit terms, banking services, preferred treatment in the supply of imported goods, and guaranteed prices for wools during the three to six months between the contract and actual delivery.^[157]

With the penetration of the wool trade in its primary bulking stages by the Arequipa export houses, direct credit links between these houses and the owners of larger haciendas, traders, and shopkeepers in the department of Puno became more frequent. Hacendados selling their livestock products directly to the exporters maintained current accounts with them by which they purchased imported goods on credit against the security of their next wool clip. Wills of hacendados often listed debts to the import and export houses.^[158] Between the mid-1870s and early 1880s this growing credit

dependence led to a number of sales of urban real estate in the department of Puno by local traders and hacendados to the export houses in payment of debt, probably a consequence of falling wool prices.^[159] However, prior to the 1920s merchants in Arequipa, Juliaca, and Puno rarely gained control of important landholdings in Azángaro province through debt foreclosures.

With the foundation of the Banco de Arequipa in 1871, modern bank credit made its appearance in southern Peru. Based on the capital of the regional oligarchy, the bank extended its operations to Cuzco and Puno.^[160] Together with the increasing importance of the large import and export houses, banks slowly supplanted merchant bankers such as Amenabar or del Valle as credit institutions in the altiplano. This process is symbolized in a credit contract from 1875 in which a Puno shopkeeper received a credit of 1,700 pesos from the Banco de Puno—presumably a branch of the Banco de Arequipa—but relied on the merchant banker Fernando del Valle as cosigner for the loan, who in turn received a mortgage on real estate as security for his potential obligations against the bank.^[161]

The Banco de Arequipa and its Puno branch fell victim to Peru's financial collapse of 1876. Only in the late 1880s did another bank begin to operate again in Arequipa, and Puno had to wait until the early 1920s before a branch of the Banco de Perú y Londres opened its doors.^[162] Between the late 1890s and World War I Peru's reconstituted banking system remained cool toward agro-exporters; the banks pursued a stable money policy in opposition to export interests and considered agricultural mortgages a bad risk. In 1902 Wenceslao Molina, a professor of animal husbandry at San Marcos University in Lima and the heir to Hacienda Churura in Putina, outlined a program of measures necessary for the establishment of a modern livestock industry in Peru. Reflecting a general frustration of large agro-exporters with the credit system after the adoption of the gold standard, Molina demanded the establishment of agricultural mortgage banks, "which could fill the void left by the presently [existing banking institutions], which grant credits only to the merchants and keep them from the hard-working rancher."^[163]

Alfonso Quiroz's recent work on Peru's financial institutions confirms Molina's critique. Yet despite their unwillingness to finance agricultural improvements, banks did play an increasing role in export financing. With the help of the banks, bills of exchange finally became common in the wool export business, a first step in alleviating the chronic cash shortage in the altiplano. But only large traders and producers benefited from the introduction of this financial tool.^[164] The hacendados' reliance on credit from the export houses increased their multifaceted dependence on these merchants, and they inevitably resented their relative weakness.^[165] Whenever

earnings from wool exports declined, hacendados experienced the uneven distribution of the benefits from the trade particularly keenly and—as after the brief slump of 1901–2—called for measures "to throw off the yoke of the export houses."^[166] This conflict of interest intensified in the period of the sharp slump of the early 1920s.^[167] Prior to this date, however, long periods of improving market conditions muted the conflict and left the hegemony of the export houses untouched. In the words of the owner of a small hacienda in Azángaro, "the wholesalers were in charge."^[168]

If hacendados had reason to complain about the uneven distribution of benefits from the wool trade, the indigenous peasants entering the market with small amounts of wool found themselves in a much more disadvantageous position. Although competition sometimes led wool exporters and their agents in the altiplano to outbid each other to secure the production of a large hacienda, the indigenous smallholders rarely received this benefit.^[169] This was just one of the economic consequences of the sociocultural domination of the Indian smallholder by the social groups controlling the commercialization of wool.

Wool traders in the altiplano automatically classified sheep wool into "estate wool" and "common wool." Estate wool was presumed to have longer and finer fibers, to be of a uniform white color, and to contain less hay and dirt. The wool bought from Indian estancia owners was automatically downgraded as being dirtier and having shorter fibers and a greater admixture of black wool.^[170] Thus, peasants received considerably less for their wool than did estate owners. In 1920, for example, woolbuying agents in Santa Rosa paid five soles less per quintal of sheep wool to peasants than they did to hacendados, at a current price of fifty-five soles per quintal, a discount of nearly 10 percent.^[171] When buying wool from Indians, the traders at times used rigged scales. They discounted one pound in every quintal for dirt, wetness, and the weight of the rope holding the bales together, although such weight losses were calculated in the basic price for unwashed wool. When peasants wanted to sell wool or hides to merchants, they were often met at the outskirts of town by alcanzadores , who sought to persuade the approaching Indians to sell their goods to a particular merchant, a persuasion that could take the form of money advances, alcohol, or brute force.^[172] Once in the store, they were obliged to buy alcohol, sugar, or maize at inflated prices.^[173]

In some parts of the southern highlands, the bulking of wools produced by peasants could be organized in a fashion similar to the colonial repartos de bienes. In a village of Chumbivilcas province in Cuzco department, peasants protested in 1882 that "in certain seasons of the year [wool traders] from different towns come to our huts . . . and force an exces-

sively small price for this commodity on us. At the time of the wool clip they capriciously snatch [the wool] from us, weighing a quintal as an arroba. When for this very reason we are incapable of paying the whole debt which they force us to contract, they double our losses by charging usurious interests, and end up secretly taking all our livestock from us."^[174] In Puno's Chucuito province the authorities themselves practiced a forced system of wool bulking as late as 1920. A few weeks before the shearing season, in December or January, the district governor would distribute money, lent to him by wool traders, to the Indian livestock herders, obliging them to deliver a specified amount of wool. If they refused to accept the conditions, the governor employed the communal authorities to deposit the money for the wool at the peasants' hut, and the latter knew that they "had to come up with the equivalent amount of wool."^[175] Mayors and subprefects entered the wool business precisely because they had power over Indian herders and thus could guarantee supplies to exporters.^[176]

Imposing low prices and securing supplies for specific traders and authorities were the goals of these methods of deceit and force. Such schemes were practiced again and again by those with leverage over Indian peasant producers. They subverted an intrinsically competitive market into a myriad of monopolistic relations of appropriation. But they were no longer the very precondition of peasant market participation, as repartos de bienes still might have been during the eighteenth century.

Between the 1850s and 1920 Indian livestock herders increasingly came to view market transactions with hispanized traders as important regular parts of their household subsistence economy. Their margin of autonomy in exchange relations was diminishing as the dendritic system matured after the War of the Pacific. As late as the 1850s and 1860s a legend flourished among Puno's elite that Indian peasants had buried about ten million Bolivian pesos, their receipts from increasing wool sales, money that thus "vanished from circulation."^[177] In other words, there existed a sphere of monetarized circulation among the peasants that lay outside the control of the altiplano traders.

Over the next sixty years both pull and push factors brought the peasants into increasing dependence on hispanized traders. The everdenser network of itinerant traders and wool-buying agents, especially after 1890, made it more difficult to escape their purchasing pressures. The diminishing land base in the parcialidades, caused by hacienda expansion and population increase, led to smaller average livestock herds for most peasant families. Fiscal measures sought to capture more of the peasants' monetary income from wool sales. Between 1867 and the 1890s governments in Lima undertook repeated attempts to collect the contribución

personal , a new head tax that replaced the contribución de indígenas. A new alcohol consumption tax established in 1887, excise taxes on alcohol, sugar, and tobacco introduced in 1904, and the broadened collection of the contribución de predios rústicos after 1902, which in Azángaro primarily affected peasants, were at once means and expression of a more effective control over the peasantry, especially after the 1890s. The railroad and telegraph and the establishment of rural police posts in each province after 1895 made it easier to quell uprisings or discourage them before they were undertaken.

But this greater reliance on market transactions did not mean the abandonment of the traditional goals of the peasant economy: subsistence of the family in the context of communal solidarity. Cash received for livestock products from traders paid for the considerable expenses of festivities, such as those for a community's patron saint, a baptism, a wedding, or a funeral. Indian livestock herders from the northern altiplano perhaps made cash purchases of commodities such as salt, pottery, or alcohol with which to maintain long-standing barter relationships in neighboring areas as the montaña in Bolivia's Larecaja province. As the parity of exchange values between bartered goods remained constant for longer periods of time, during phases of increasing prices for wool and hides it was advantageous to purchase barter goods with cash received for livestock products.^[178]

Indian peasants held back much wool from the export trade even during boom years. Other livestock products, including tallow, sheepskins, and dried meat, entered the dendritic system of monetarized trade in even smaller proportion. Besides direct consumption in the peasant household, these goods continued to serve as means of exchange in traditional barter relationships. For example, each May after the shearing season the colonos of Hacienda Picotani abandoned their estate and went down to the valleys of Sandia to provision themselves with maize, coca leaves, and other foodstuffs in exchange for livestock products and homespun baizes.^[179] During the World War I boom peasants from around Juliaca refused to sell any serge to the merchants to whom they sold raw wool, as they were taking increasing amounts of this home-woven cloth 150 kilometers farther north to Sicuani, their traditional spot for provisioning with maize.^[180]

Indian livestock producers did not object to being integrated into the market, as they had come to rely on cash as part of their family subsistence strategy. What they objected to was the force and deceit that traders and authorities routinely imposed in the "marketplace." However, they did not accept such exploitation fatalistically. Indians "give life to business in this region," as wool trader Francisco Rodríguez of Santa Rosa wrote to Ricketts in 1918, and this market position allowed them to use ruses, tricks, and

plain common sense to countermand and limit their exploitation at the hands of unscrupulous traders.^[181] Indians mixed hay and dirt into the wool, moistened it, and even poured sugar water over it to increase the weight.^[182] In 1932, when Bolivian customs agents began to collect export duties on wool, alpaca herders from south of the border ceased to sell their wool to Peruvian traders in border towns such as Cojata and Moho, instead establishing relations with merchants in Puerto Acosta on the Bolivian side.^[183] Peasant herders sought to adjust their sales both to the rhythms and requirements of their household economy and to market fluctuations.

During the 1920s, a period of unstable wool prices, traders again and again complained about "the absence of the Indians" from weekly markets or annual fairs. "The Indians held back [their alpaca wool] in expectation of price improvements, they only sold amounts indispensable for satisfying their most pressing needs," wrote the Ricketts's agent Hipólito Sanchez from Moho in Huancané province in September of 1926.^[184] They tried to sell the wool during the season of highest prices, September through December, coinciding with the months of the agricultural cycle when last year's harvested food began to get scarce. Before celebrations such as carnival or patron saints' festivities, wool sales increased, as did alcohol purchases.^[185]

Peasants considered trading a complex skill entrusted only to the most experienced and honored family members, an activity in which sons were trained from a tender age, when they accompanied their fathers on expeditions to distant market towns or subtropical valleys. To give some stability and predictability to trading in those alien environments, the peasants sought long-lasting trading relationships with compadres (their children's godparents, whom they sought out as protectors). A trader might have as many as six hundred compadres among peasants selling him wool.^[186]

But resistance always proved fragile. The very compadre to whom the peasant had entrusted himself over many years for his wool sales might exploit that dependence. And the attempt to hold the wool clip until prices were high collapsed when there was a prolonged price slump or when food was scarce because of crop failures. Then the reverse pattern was set in motion: peasants had to market their wool as fast as possible, whatever the losses. They sheared the animals before term, bringing shorter fibers to market; the price they received would be doubly low on account of the inopportune season of sale and the low quality of their wool.^[187] Of course, poor peasants, with scarce resources of land and animals, faced these problems more frequently than did affluent comuneros and colonos. And alpaca herders, such as those high up in the Cordillera de Carabaya, with

their undisputed control over the know-how of producing the cameloids' precious fiber, may have had a more stable market position than did sheep-wool producing comuneros from the altiplano proper.

Through the tremendous intensification of the competition for land, the wool market may have driven forward this type of social differentiation among the peasantry. But comuneros and colonos had a number of countervailing strategies at their disposal that assuaged such effects of the wool market before 1920. For better or for worse, the prosperity of the altiplano's indigenous livestock herders—as well as of hacendados, transport entrepreneurs, traders, and administrators—had become linked to the vagaries of international demand for wool, just as they had depended on the fortunes of the mining centers in Upper Peru until the end of the colonial era.