6
The Privatization of Land

As suggested in the last chapter, the switch from indigo to coffee production was a response to changing conditions in the international market. Indeed, artificial dyes began appearing on the market in 1856, and uncertainty existed about the future of indigo; international markets expanded thanks to lower transportation costs and to the rapid internal growth in the European economies, and the Salvadoran economy rapidly adapted to the new situation. This explanation gives the impression that the decision to switch from indigo to coffee was based exclusively on economic grounds, and that it was efficient to change the structure of agricultural production. In the long run, this is a sensible interpretation. In the short run, however, the increase in coffee production cannot be separated from a political and economic decision of great importance: the privatization of ejidos and communal lands. This event is known in Salvadoran history as the liberal reforms.

Current Interpretations of the Problem

The understanding of this problem is central to every interpretation of the economic history of El Salvador. Torres Rivas thinks that the liberal reforms facilitated "more than in any other region of Central America the concentration of land ownership."^[1] This remains the prevalent view, although it has been qualified by others. For David Alejandro Luna, large properties were strengthened by the agrarian policies of the

nineteenth century in general, but the specific consequence of the liberal reforms was the appearance of "anti-economic small landholdings," a feature of the Salvadoran countryside ever since.^[2] The main interpretation of the liberal reforms has been articulated by David Browning in a pioneering book on changes in Salvadoran landscape and society.^[3]

As Browning's interpretation is by far the most elaborate, it is convenient to analyze it in detail. He discusses with great insight the enormous gap in the concepts of land use and land property used by the white elite and by the Indian communities. According to Browning, the Salvadoran elite perceived two obstacles to economic growth: excessive dependence on indigo production, and confusion with regards to land titles.^[4] Their solution to the problem illustrates their peculiar point of view—the problem of dependence on indigo was solved with a new dependence on coffee; the problem of land titles was solved by depriving the Indian communities of their land.

Browning suggests that the factor that moved the elite to desire changes in land tenure was the introduction of coffee. The Indian and ladino concepts of land use were incompatible, and the expansion of coffee made the confrontation unavoidable. "Above all else," says Browning, "it was the introduction of coffee that persuaded those in control of the nation's affairs to reform the use and ownership of the land of El Salvador."^[5] Ejidos and communal lands were thought to be an obstacle to the development of coffee production. Commercial agriculture was a difficult concept for Indians to understand: "In many [rural] communities," says Browning, "explanations about the value of certain crops in international markets have little meaning or importance." Thus, "it is not a matter of saying to a villager that he should be prepared to grow export crops alongside maize and beans."^[6] The problem became more complicated with the introduction of coffee, which was

... an alien plant of which the individual cultivator had little knowledge and to which he attached no importance, either cultural or economic. In addition a plant that required five years' growth before it gave fruit was not easy to accept into a cycle of cultivation that was attuned to the annual change of the seasons, and where because of economic necessity and the cultural heritage it was assumed that each year brought rain, growth and food.^[7]

Indians thought that land had to be used to produce subsistence crops, whereas big landowners gave priority to commercial crops that, it can be inferred, were more profitable. Following this line of reasoning, it can be concluded that for the agricultural sector to grow it was necessary to eliminate traditional forms of landholding. Those who cultivated

ejidos and communal land had such a narrow concept of the role of land that they could not take advantage of the better prices of cash crops.

This concept presents a problem since indigo, a commercial crop, had prospered since colonial times when communal and ejido lands were of great importance. But indigo, according to Browning, was different since

... the slow expansion of indigo production developed alongside indigenous subsistence farming and, though competition between the two for land did exist, both systems were given time to adjust to the requirements of the other within a legislative framework that was designed to preserve both new and indigenous structures.^[8]

The pace of expansion permitted an adaptation without conflicts. That was not the case with coffee. Moreover, Browning asserts that the elite changed its attitude toward land. This change was due to the different production methods and economic organization involved in coffee cultivation. He found the following differences:

- The crops were introduced in different ways. Indigo was introduced slowly whereas coffee expanded rapidly.

- The methods of cultivation of both crops were different.

- Both crops had different labor requirements. Coffee needed more year-round labor.

- The geographical location of both crops was different because of the narrower ecological conditions under which coffee can be cultivated with profit.^[9]

Browning completes his interpretation by saying that the rapid development of transportation facilities contributed to the rapid growth of coffee production. The main reason for the national elite to promote the reform of land ownership was the desire to increase the production of coffee. There were substantial differences in the cultivation of coffee and indigo that demanded a different attitude toward land. Unfortunately, communal and ejido lands were in the hands of people who were interested only in subsistence crops. The economic project of the national elite was incompatible with traditional forms of landholding.

In fact, there is much to learn from this interpretation, and the main lines of the story seem to confirm it. As far back as the period between 1859 and 1863 Gerardo Barrios transferred public land to the private sector with the sole condition that it should be used to produce coffee. As described above, the process culminated in the abolition of commu-

nal and ejido lands with decrees issued in 1881 and 1882. The prologue of the 1881 law stated that,

The land owned by communities impedes agricultural development, is an obstacle to the circulation of wealth, and weakens family ties and the independence of the individual. Its existence is against the economic and social principles adopted by the Republic.^[10]

A specific sector of Salvadoran society thus identified its group interests with those of the country.

The most recent interpretation of the liberal reform is Rafael Menjívar's. He also gives great importance to the introduction of coffee, but he approaches the problem from a Marxist perspective. His main concern is not why the reforms took place but what the role of the reforms was in the development of capitalism in El Salvador. To him the liberal reforms played a key role in the process of capital accumulation. According to his interpretation ejidos and communal lands "would be the lands that, under the process of originary accumulation, would become the constant part of capital."^[11] The beauty of the process was that at the same time that the elite appropriated land, labor was freed to work in the coffee plantations. Key to his interpretation is the amount of land involved. Undoubtedly the more land changed hands, the stronger the argument become. He estimates that 40 percent of the territory of the country was directly affected by the reforms.

As this sketchy review of the literature suggests, there are four questions worth exploring in order to understand the implications of the liberal reforms. First, what was the amount of land involved? Second, what was the political climate that made them possible? Third, what was the role of the introduction of coffee? And finally, how did the reforms alter the distribution of land?

The Importance of Ejidos and Communal Lands

As the authorities confirmed with a land-tenure survey carried out in 1879, what was at stake was a good portion of the territory. Although the survey was incomplete (it provided no information for three rich provinces, Ahuachapán, San Miguel, and La Paz, and its accuracy was affected by the great confusion over land tenure that prevailed at the time), it had enough information to provide an idea of the order of magnitudes involved. The governors' reports on which the survey was based originally appeared in the Diario Oficial , and were summarized

by Rafael Menjívar in his book Acumulación originaria (see table 21).^[12] Menjívar's table compares the area of the ejidos to the total land area and to the arable land of each province. The data on arable land included in the table was taken from an agricultural census carried out in 1950; therefore, they are a high estimate of what was considered to be arable land in the 1870s. According to the figures, 12.9 percent of the total territory of the country and 17.7 percent of the arable land was ejido land (assuming that the three provinces that have not been accounted for followed a similar pattern). Of course, the last percentage is a low estimate because of the assumption made about arable land. Based on the same information, Browning asserts that "well over a quarter of the country ... was considered by the villages to be their own domain."^[13] Menjívar believes that this estimate does not include communal lands and, after making some assumptions, reaches the conclusion that by 1879 "no less than 40 percent of the territory was covered by ejidos and communal land."^[14] Both writers have the same point of departure but arrive at substantially different estimates.

Menjívar's figure is a very high estimate. He starts with Browning's estimate of 25 percent but, based on the fact that there was great confusion as to what was an ejido and what was communal land, asserts that most communal lands were not included in the survey. He tries to arrive

at an estimate by assuming that the area covered by the communal lands can be obtained by subtracting ejido land, haciendas, and nonagricultural lands from the total area of the country. It is a reasonable assumption, much more reasonable than the actual selection of data to make the estimation. The area of the country's territory is not in question, and the area covered by ejido lands, although known to be imprecise, at least is based on a contemporary survey. The amount of land covered by haciendas presentes a trickier problem. Menjívar, following Browning, asserts that the extensions of the haciendas can be estimated from the report written in 1807 by the intendente Gutiérrez y Ulloa.^[15] This approach has two problems: the intendente's report does not have enough data to yield anything more than a general impression of the importance of haciendas, and land tenure had experienced important changes in seventy years. The economy had experienced transformations, the private land market had been very active, land transactions involving ejidos had taken place, and the government had sold terrenos baldíos. Even if an accurate estimate of the land covered by haciendas in 1807 was available it would be irrelevant to estimate hacienda land in the 1880s. By contrast, the area of nonagricultural lands is taken from the 1950 agricultural census. This presents another serious problem: new agricultural techniques and higher demographic pressure almost certainly made the area of 1950's nonagricultural land significantly smaller than 1879's. Menjívar does not account for this. A sample of land titles registered in the land registry of the western provinces suggests that by the 1880s common lands were a small percentage of the total. Only 3.74 percent of the land in the sample was common land. (The evidence is not conclusive since there were allegations that Indian communities were slow in privatizing their lands. Nonetheless, the western provinces were the area where Indian communities were strongest and more organized, and one would expect that they would have made a good effort to claim their lands.)^[16] Menjívar's estimate can be considered, at best, as a very high upper bound. Browning is more cautious because he believes that, as the distinction between ejidos and communal lands was always blurred, it is not certain whether the latter were included in the survey in any consistent fashion. His estimate of 25 percent of the territory is already almost twice the estimate found in the 1879 survey, making a very generous allowance for the undercounting of communal land.

It is not necessary to agree with Menjívar's exaggerated estimate to argue that the size of communal and ejido lands was far from negligible. Not all ejido and communal land was cultivated, and private land was not fully cultivated either, but there is ample evidence, both in the 1858 census and in the 1879 survey, of the importance of ejidos and commu-

nal lands in the production of foodstuffs and export crops. There is no question about the importance of these forms of land tenure in terms of their size and of their contribution to the Salvadoran economy. Moreover, ejidos and communal lands were closely identified with colonial institutions and with one significant group in Salvadoran society: the Indian population.

Political Context

The substantial amount of land involved suggests that the move to privatize all land cannot be seen as an exclusively economic decision. A large segment of the population was negatively affected by this decision, which, therefore, could not be taken without considering political realities. Ejidos and communal lands were not an obstacle to the expansion of coffee, and coffee, by itself, was not the main reason for the liberal reforms. Foreign demand for Salvadoran agricultural products had made land relatively scarce and more desirable. Under those circumstances a very small group that had been able to achieve control of the apparatus of the state did not shy away from using its power, moving to change land legislation to its advantage.

Changes in the balance of power between Indian communities and ladino landowners interested in producing for the international markets are crucial to understand the liberal reforms. They were two groups who had never seen eye to eye. Before independence, colonial institutions had mediated their conflicts, but, with Spanish authorities gone, conflict broke out regularly. The old tradition of endowing towns with ejidos and Indian villages with common land was threatened by the new and radical ideas of liberals inspired by Adam Smith. Two factors helped to keep the landowners away from ejidos and communal lands: the existence of abundant uncultivated land and the power of conservatives in neighboring Guatemala. Rafael Carrera, the Guatemalan conservative caudillo, had come to power after a rebellion provoked by liberal excesses and with the strong support of the Catholic church and the Indian communities. Moreover, he wanted like-minded governments in Central America and used force to get his wishes. During the first three quarters of the century the risks of confronting Indian communities were higher than the rewards. Indian rebellions were always supported or used by conservatives to advance their political goals. As long as Indian groups could obtain support in Guatemala, they represented a real threat. Indians had less to fear from conservatives, who wanted to protect their power base, than from liberals, who wanted to free every market (including land) from colonial restrictions. Moreover,

there was great confusion about the difference between ejidos and common lands.

Liberals, in particular, identified all the supposed evils of communally owned land with the alleged backwardness of the Indian population, even though many ladinos benefited from the ejidos. Interest in Indian lands had been explicit since the early years of the national period. Juan Manuel Rodríguez, one of the leaders of the independence movement in El Salvador, wrote in 1824:

Interior land is badly distributed, because Indians own most of it as common land and the so-called ladinos have none. The former do not enjoy true ownership, therefore they do not value their land, and they only impede its cultivation by others.^[17]

Rodríguez's message may have been too blunt. Other liberal leaders couched the same concepts with an appeal to loftier principles of equality:

In the old system class differences impeded the use of land, and in the so-called Indian towns, they [the Indians] decided how to use their land, cultivating part for neighbors of the same towns who did not belong to their class. The law of equality established by our current system must abolish every privilege since there is only one class of citizen, every one with the same options, it is very just to level all to the same rights and obligations.^[18]

There was a conflict, and it had clear ethnic connotations. In fact, both conservative and liberal leaders were wary of the Indian population. Needless to say, all positions of leadership were held by criollos or ladinos. When there was a danger of Indian revolt they put their differences aside. Anastasio Aquino's revolt in 1833 "had the salutary effect of causing all conflicting parties to unite against [the Indians]."^[19] But liberal policies were more likely than conservative ones to exacerbate Indian grievances.

In the unstable environment of the federal period, criollo and ladino landowners were far from being in full control. The Indian population sometimes used the new political avenues created by independence. In 1835 Tomás Espinosa, a man of Indian extraction, was elected chief of the state of San Salvador. When ladinos and criollos opposed his election, the Indian population fought to support him. Gral. Morazán, a liberal and president of the federation, sided with the opponents of Espinosa, who was out of power in less than a year.^[20] A few months later, in May 1837, an Indian rebellion

... rose upon the white and mulatto population murdering them and plundering their houses, under a supposition that the origin of the devastating cholera was attributable to poison thrown into the rivers and that even the medicines supplied to them were impregnated with deleterious substances.^[21]

A cholera epidemic had helped to trigger resentments against landowners. It was a disorganized affair that, unlike Carrera's similar rebellion in Guatemala, completely failed. However unsuccessful, these Indian uprisings reminded liberals like Juan Manuel Rodríguez that there were limits to their programs.

External pressures further helped to check the liberal program. After the breakup of the federation in 1839 the Guatemalan conservative leaders, first Carrera and then Cerna, were always available as allies to fight against liberal excesses in El Salvador. Liberals were in power from 1845 to 1851, but they had to be tactful in advancing their programs. Nonetheless, there were two Indian uprisings under their watch.^[22] When conservatives had a chance to rule, they, in turn, had to realize that the Salvadoran elite was, by and large, liberal.^[23] A good example of this was Francisco Dueñas, El Salvador's foremost conservative and the whipping boy of liberal historians who, thanks to the help of Carrera, replaced liberal Gerardo Barrios in 1863. Unseating Barrios took a fullfledged war between Guatemala and El Salvador, helped by Indian revolts against the liberal leader.^[24] Dueñnas's conservative instincts were reflected mainly in his respect for the Catholic church and for Indian institutions. He never attempted to role back the clock. He passed legislation giving representation to Indian communities but did not reinstate laws endowing towns with ejidos. He protected the church as a spiritual force but upheld an 1832 law declaring the capellanías property of the state.^[25] Many of his economic policies were a continuation of those of his arch-enemy Barrios: improvement of roads and ports, pro-fessionalization of the army, installation of telegraphs, sale of public lands, and the like. His main achievement was to provide a measure of much needed stability. Exports grew faster under Dueñas than under Barrios, and coffee took off during his period. Indians gave him their full support, and he did not touch their lands. When liberals tried to stage a revolt in San Miguel, Dueñas defended himself with troops recruited in Indian towns.^[26]

Things changed, however, and liberals came to power in Guatemala. One of Dueñas's main sources of support was gone; his government was toppled in 1871. Without him Indian revolts returned; their institutions were again under siege. The first revolt took place in 1872 when 400 Indians armed with machetes rose in Cojutepeque under the battle cry

"Dueñas and religion," but both Dueñas and the bishops were on the losing side.^[27] Indian unrest returned three years later. In March, 1875, in the predominantly Indian town of Izalco, people "believing that their rights were being threatened by the sale of an ejido, instead of going to the courts were led into disobedience and rebellion."^[28] A few months later priests and Indians joined forces in San Miguel. A priest by the name of Juan Manuel Palacios led a revolt that started with the opposition of street vendors to using the new market facilities built by the government. Between "3 to 4,000 male and female fiends robbed, burnt, and devastated all that they did not take away." As many as 300 principal families were threatened with death, and the military governor of San Miguel province was castrated and killed. His body received so much abuse that it was difficult to lift it from the ground to give it a decent burial.^[29]

The conflict between old and new ways was clear. The idea of selling their produce in an enclosed marketplace and having to pay rent for that privilege was completely alien to the Indians. Open markets in the central square had been the custom since before the arrival of the Spanish. The only thing needed to sell a product was to bring it to the central plaza, sit on the ground, and wait for customers. For the local authorities the new marketplace represented an effort to modernize the city, making it more European—how could anybody oppose it? In their view only culturally inferior Indians could not grasp the convenience of a clearner and orderly marketplace.^[30] The Indians lost that fight; their cause was hopeless. The San Miguel revolt was the last real threat to the triumphant liberal republic. The civilian leaders of the revolt were executed, and the main authorities of the Catholic church were sent into exile.^[31] Neither the church nor the Indian communities could gather resources to fight against an elite that was eager to expand its wealth and now enjoyed access to modern communications to coordinate troop movements. More importantly, the end of strong conservative governments in Guatemala removed the main obstacle to an attack against communal forms of land ownership.

The new liberal authorities had boundless faith in the free market. The example of the United States inspired them. President Valle delivered a message in 1876 summarizing these feelings; in his view the principle of laissez-faire, "observed with very few limitations in the United States of North America, has transformed deserts which until recently had been inaccessible into very beautiful cities; has criss-crossed the territory of the union with railroad tracks, and has taught and proven to the entire world the value and potential of free institutions."^[32] How could one apply those ideals in El Salvador? Certainly not by freeing the political system. The Salvadoran elite was not

ready to give up its privileges; before attempting to do anything so rash it was better to free the land market in order to bring prosperity.

Talk about the "evils" of communal land (a code phrase for Indian property) became commonplace. When in 1878 the governor of San Vicente tried to explain the stagnation of agriculture in his province, his explanation amounted to a threat. According to him, the reason for the wasteful use of land was "the lack of laws to regulate the cultivation of ejido lands, since it is not yet possible to turn them into private property."^[33] It was only a matter of patience; by the end of the year the mayor of the village of Mejicanos was proposing the distribution of commercial plants such as cacao, rubber, agave, and, of course, coffee. Those who devoted at least one-quarter of their ejidos to the cultivation of those crops were promised a title to that land. It was the kind of measure that received lavish praise from the government. Moreover, it deserved emulation: the municipalities of Soyapango, Ilopango, Tecolu-ca, and Villa Verapaz followed the progressive example of the mayor of Mejicanos.^[34] In his 1879 report to the legislature, the minister of government praised the action of those forward-looking towns and stressed the damage done by ejidos and communal lands, which, in his opinion, belonged to groups "removed from all progress and activity." "Let us cut, then, with a firm hand, the chains that enslave agriculture," he concluded.^[35] The day before, the president had signed legislation emulating at a national scale the example of Mejicanos. The door was open to the privatization of land. Governors were instructed to carry out a survey of land tenure to find out how much land was not privately owned. The process culminated with laws passed in 1881 and 1882 abolishing communal lands and ejidos. The export sector needed room to expand, and it was clear that communal lands and ejidos could be privatized without major political consequences.

Yes, the Indian communities did have a different way of doing things, but they were not an obstacle to the growth of the coffee industry or to economic growth in general. They participated in the market and were quite capable of playing a constructive role in the economy. In fact, after the 1879 legislation they made substantial efforts to cultivate coffee, and they had always engaged in the production of commercial crops.^[36] The liberal reforms were the culmination of a struggle for power. The local elite had strengthened its grip on power to such a degree that it could reorganize land tenure without fear of the reaction. There was resistance, but the social unrest that followed the reforms, although sometimes violent, was easily controlled. Indian communities could not match the weapons, the numbers, the organization, or the mobility of the army. Moreover, they had nobody to turn to for support. No other group in Salvadoran society had a community of interests with

them. By the end of the 1870s Salvadoran conservatives were not eager to defend traditional institutions; they had become coffee planters. (The leader of the conservatives was Dr. Manuel Gallardo, the successful planter mentioned earlier.)^[37] Guatemala had ceased to be a conservative stronghold. Indian revolts were bound to fail. There were revolts in 1884 in Izalco and Atiquizaya, in 1885 and in 1889 in Cojutepeque, and in 1898 in Santa Ana province. In the last revolt the peasants cut off the hands of the Jueces Repartidores Ejidales .^[38] Revolts were "a delicate problem," but there was an army to deal with them, and it did.^[39]

Statistical Analysis

Even though there was a political opportunity, it is still necessary to understand the economic incentives to take advantage of that opportunity. If it was necessary to turn to coffee because it became more profitable than other crops, then changes in coffee and indigo prices were important elements in the equation. To achieve growth, decisions have to be made as to how to allocate resources between a wide range of economic activities. The allocation of resources between coffee and indigo could not have been the same in 1830, when the price of a cwt. of coffee was four times the price of a pound of indigo, as in 1889 when the price of a cwt. of coffee was twenty-four times the price of a pound of indigo (see fig. 1). The behavior of relative prices across time is therefore important in the discussion of the profitability of coffee production.^[40]

But before starting, it will be convenient to describe the data used. Fortunately, coffee and indigo price series for the English market are readily available. These series have been deflated with the English industrial products price index. The industrial price index is preferred over the general price index because the value of the pound sterling for the Salvadoran exporter was a function of the amount of industrial goods (mainly textiles) that he could buy with the product of his exports. In other words, the deflated prices reflect monetary units used to purchase industrial products at constant prices. This procedure assumes that the contribution of indigo and coffee to the growth of the Salvadoran economy stemmed from their contribution to the country's capacity to import industrial products, which seems a reasonable assumption.

The value of coffee and indigo exports was not deflated because the official value for these exports was computed with the tarifa de aforos , which in the case of indigo remained at the same level during the entire nineteenth century and in the case of coffee changed only sporadically. The basic data set includes the following data series:^[41]

[Full Size]

Figure 1. 
Relative Price

— Coffee prices in England (1864–1889);

— Indigo prices in England (1864–1889);

— Coffee exports (1864–1889);

— Indigo exports (1864–1889); and

— English industrial price index (1869 = 100).

These series were used to compute the following:

— Deflated coffee prices (1869 = 100);

— Deflated indigo prices (1869 = 100); and

— Relative prices (coffee prices over indigo prices).

A look at the data gives an idea of the main changes that took place in the structure of agricultural production. In 1864 the value of coffee exports was only 80,605 pesos; twenty years later it was above 2 million pesos. At the same time, the value of indigo exports declined slowly but steadily after a peak of 2,768,576 pesos reached in 1873 (see table 19). Part of the explanation of this phenomenon can be found in the changes in international prices of these products. Coffee prices in the English market increased after the mid-1870s.^[42] The average price in the period 1865–1870 was 57.48 shillings per cwt. and in the period 1876–1880 had increased to 95.36 shillings (adjusted for inflation). The average indigo prices for the same periods were 5.47 shillings and 4.66 shillings per pound. But to have a better understanding of the role played by prices in the change from indigo to coffee production it is necessary to analyze the data more closely.

A number of statistical manipulations were performed to understand this problem. These manipulations include regressions of coffee exports (dependent variable) on coffee and indigo prices, relative prices (price of coffee over price of indigo), and indigo exports. The results of the regressions are shown in table 22. The determination coefficient of the regression on coffee prices was 0.82, and it improved marginally when prices were entered with a lag of one year (in this case the determination coefficient was 0.85). This result is consistent with the interpretation of the liberal reforms as necessary for the growth of the export sector (which is not the same as saying that that was the case). Looking only at the figures one could say that coffee production expanded because prices went up; planters took advantage of the opportunities offered by the world market.

This tentative conclusion is reinforced by the results of the regression on relative prices. Presumably producers based their decisions not only on coffee prices but also on the alternatives available to them. Since the main alternative was indigo they considered the ratio of coffee to indigo

prices. Not surprisingly the regression in relative prices is significant: the determination coefficient is 0.79, and the slope of the regression line is quite steep. By contrast, coffee exports do not seem to be closely correlated to the amount of indigo exports. If part of the resources that were originally employed in indigo production were transferred to coffee, the correlation between coffee exports and indigo exports could be expected to be negative and strong. However, the correlation is not significant. One explanation for this puzzle is to be found in the strong fluctuations in the series of indigo exports. It has been discussed above how indigo production was very sensitive to disruptions created by wars, locusts, and droughts. This accounts for strong fluctuations that are not due to price variations or reallocation of resources but which affect the statistical analysis.

Moreover, these results point to the distinct possibility that coffee expanded instead by using resources previously used for food production rather than for indigo production. Although there are no direct data on food production for the 1880s, it is clear that despite its fluctuations indigo production fell very slowly during the second half of the century. This implies that most of the resources used for the expansion of coffee were diverted from food production. Resources were very scarce, and the rapid growth of coffee could not possibly come only from net growth. The liberal reforms, which affected Indian communities and ejidos traditionally linked to the production of foodstuffs, insured that coffee could expand at the expense of food.

The fact that statistical manipulations were consistent with the interpretation that the liberal reforms were necessary for the growth of the export sector can be substantially modified when one looks more closely at the policy of incentives followed by the government. The incentives

given by the government to stimulate coffee production greatly reinforced market signals. If the benefits to be obtained from coffee production included not only the income from the sales of the beans but also the land where they were produced, land had no opportunity cost for the producer. Instead, land was counted among the benefits. Under such circumstances it would have been hard to find attractive any other productive alternative; as a result, coffee production expanded more and faster than what would have been suggested by market signals. Coffee planters took advantage of the opportunities opened up by changes in the international markets and, also, by changes in the rules of the game dictated by governments run by planters. A policy of incentives for one product led to a reallocation of resources that had a very direct impact on welfare. The liberal reforms affected the allocation of resources between alternative products and also between different social groups. In the 1850s changes in transportation costs had made production of export crops more attractive than food staples; in the 1880s that trend was dramatically strengthened by adding land to the benefits. As a result Indian communities lost their land and had to pay higher prices for food.

It is clear that government policies tipped the scale in favor of coffee production, but what kind of case did the government have to justify its policies? When the legislation was passed in 1881 and 1882 the behavior of prices did not offer an altogether clear message. Figure 1 shows a decrease of coffee prices relative to indigo prices from 1848 to 1869, then an increase until 1874, a plateau lasting five years, a decline from 1879 to 1885 and, from then on, a rapid increase. The price situation faced by the producers at the time of the new land laws was not clear. The effects of the invention of synthetic dyes on indigo prices were slow to appear. The most important breakthrough in the production of artificial dyes took place in 1883 with the discovery of the chemical structure of indigo, but it took over fifteen years to find an economically feasible substitute for indigo.

Despite the uncertainty about prices that was perceived in 1881, the liberal legislators imposed a program of very radical reforms. In the final analysis their decision was correct in one sense—coffee prices soared and, therefore, it made sense to specialize in its production. But it is difficult to argue that the data available to the legislators at the time that they made their decision supported the need for such radical changes in property holding. A posteriori, the decision made sense. A priori, it was a big risk for the country as a whole, although it was a huge success from the point of view of the planters. The change had economic costs. Government policies distorted market signals and encouraged the economy to move away from food production. The social costs of the

new legislation were far reaching: the privatization of land unleashed market forces that put Indian communities and small peasants at a disadvantage, great numbers of peasants ended up dispossessed of their land, and wealth distribution was radically changed. The possibilities of a more harmonious growth of the economy and society were eliminated.

It seems clear that changes in relative prices did call for a reallocation of the factors of production, but the timing and the means used to carry out such goals are highly questionable. Such extraordinary government measures were unnecessary to promote growth. Indeed, when the economic stimulus was powerful enough, as in the case of the cotton boom earlier in the century, the different economic units could adapt rapidly to the new conditions without having to change the structure of land tenure.

There was a political opportunity and an economic incentive, but even if it made sense to switch to coffee production, land redistribution was not the only alternative open to the liberal leaders. There were three markets that could respond to government intervention: land, labor, and credit. The liberal reforms affected the first two and left the last one intact. It was an ironic choice since the main difficulty in the cultivation of coffee was the high startup cost of coffee plantations.

Further insight into the problem can be gained by analyzing the alternatives available to the government. The land market was a clear choice. The very importance and the geographical location of much of the ejidos and communal lands made them tempting, but the evidence suggests that even without government intervention the land market was changing rapidly and so was the balance between the different factors of production. Even before the full implementation of the liberal reforms in the early 1880s the pressure of export agriculture on land was becoming apparent. Land has always been a symbol of wealth and very desirable, but land is only one of many forms of holding wealth. It is desirable as long as it can be put to productive use or can provide capital gains. In the years after independence land had been abundant relative to capital and labor, and sales of terrenos baldíos kept prices at roughly the same levels. There was no incentive to hold land for capital gains since its price was not rising. The shortage of labor made it impossible to expand cultivation beyond certain limits, and the pressure of demand was not felt in full force. By the 1880s the picture had changed. Demographic and economic growth had altered the relative position of land and labor. The population had grown and, although it cannot be said that labor was abundant, it was enough to satisfy the needs of the growing demand. At the same time the growth of international trade made it possible to exploit more land.

(A fine point of interpretation is in order. In El Salvador coffee pro

duction was the new element that made land more attractive and, thus, scarce in the economic sense. Hence, the introduction of coffee is considered the key to understanding the liberal reforms. Although the demand for coffee was instrumental in making land scarce, by stressing the concept of land scarcity the problem can be related to similar events in other parts of the American continent, thus highlighting both the specific features of the Salvadoran process and its participation in a broader phenomenon. The extension of railroad networks in Mexico and Argentina and the Indian wars in the latter country were efforts to alleviate the pressure on land. In El Salvador, demand for land increased but no more public land was available; the competition for it became more urgent, and the methods to acquire it were ruthless.)

The impact of export agriculture on traditional forms of land tenure prior to the reforms can be illuminated by an analysis of the wide variation in the percentage of ejido land between provinces (see table 23, col. 7). Although data limitations make it difficult to draw definite conclusions, it is useful to compare the degree of privatization in the different provinces with their share of the total of indigo production. There are two possible hypotheses with regards to the relationship between concentration of indigo production and ejido lands.^[43] One is that land was more attractive in the areas where indigo was cultivated more intensively and, therefore, haciendas expanded at the expense of ejidos and communal land. In those areas indigo planters took advantage of every opportunity (legal or otherwise) to acquire new land and, after a while, the percentage of private land increased. For this hypothesis to make sense land scarcity had to be felt well before 1880.

A different possibility is that the scarce resource was labor and not land. In that case indigo haciendas would have had an incentive to be close to population centers where labor was available either for hire or, in colonial times, in the form of repartimiento. Ejidos and communal lands were a reserve of low-cost labor, and the haciendas wanted to profit from that reserve. If this was the case, it can be expected that the areas where indigo was cultivated intensively also had the highest concentration of ejidos and communal lands.

Table 23 compares indigo production with the amount of ejido land in each province. The data on this table were used to test the correlation between the percentage of ejido land and indigo production in each province. As the provinces were of different sizes, the total amount of indigo production is not the relevant measure. Instead, it seems that the ratio of indigo production to arable land gives a better idea of the importance of indigo in each province.^[44] The latter ratio was used to rank the provinces according to intensity of indigo production and percentages of ejido land. A Spearman rank-order correlation was used to

compare these ranks. The value of Spearman correlation coefficient was-0.66, which indicates the ejido land was less prevalent in provinces where indigo was important. This result favors the scarce-land hypothesis. The rapid expansion of export agriculture made land more attractive, and planters found ways to acquire ejidos and communal lands.

Since the data are incomplete, the test just performed, by itself, is not enough to be conclusive. Among the provinces not included in the computation were some, like San Miguel, which were important producers of indigo. Nonetheless, the same rank test was performed with the totals (as opposed to percentages) of ejido land and indigo production, and with combinations of total values and percentages, and in all cases Spearman's r was negative and had an absolute value above 0.50. This indicates, at least, that the pressures of commercial agriculture were already eroding the institutions of ejido and communal land.

Changes in land tenure were directly linked to changes in the labor market. Authors such as Rafael Menjívar give great importance to this link. In his book Acumulación originaria he argues that ejidos and communal lands were eliminated due to "the need to integrate land to capital in coffee activities and the need for free labor."^[45] That is, one of the reasons for the liberal reforms was to increase the availability of labor by denying access to land to an important sector of the population. This interpretation assumes that labor was becoming scarce because of the increased demand created by the specific characteristics of coffee production. It is difficult to be conclusive about his point. The results of the comparison between concentration of indigo production and concentration of ejido land and most of the qualitative evidence suggest that labor was scarce even for the production of indigo. If this is the case, the new element in the process is land scarcity, and labor scarcity can be considered to be a constant. That is, the liberal reforms were triggered by the pressure on land created by the new opportunities created by the opening of the economy. This stress on land scarcity does not deny the fact that it was very convenient for plantation owners to have a greater pool of landless peasants to recruit labor from. What is important to emphasize is that land scarcity was already eroding the ejidos and communal lands without any change in legislation.

Despite the importance of land, the efforts to modernize the markets could have focused on the credit market rather than on the land market. It can be argued that the backwardness of financial institutions was a greater obstacle to the expansion of coffee than the scarcity of land. There were precedents. Credit mechanisms, however rudimentary, had existed since colonial times. In the 1800s Guatemalan merchants advanced money and goods to the big planters, who in turn gave credit to the smaller planters, and the Montepío de cosecheros de añil had been

an important credit institution. But the system was so informal and the legal system so inadequate that risks were high and transaction costs onerous. The realities of coffee production demanded a more modern system. New credit institutions could have solved the problem of investment costs. Even private credit with high interest rates would have been a more efficient solution than the elimination of ejidos and communal lands.

There is no reason to believe that the Indian communities were an obstacle to growth. They were perfectly efficient within the limits imposed by their agricultural knowledge, and ejidos and communal lands had produced commercial crops for a long time before the reforms. The more "backward" elements of society had a long tradition of successfully competing in the marketplace. They had the land and the willingness to work, and faced the same obstacles as everyone else: difficult access to credit and to new agricultural techniques. But under the new realities imposed by the expansion of coffee they would lose their lands, and no effort was made to provide them with credit or education. Ejidos and communal lands had been used for commercial crops for a long time before the reforms. The more "backward" elements of society had not shied away from participation in the market economy. When the government pushed them to produce coffee they were ready to try in order to avoid losing their lands; they understood the incentives perfectly well, even though the fact that they had not had access to education put them at a disadvantage. Browning asserts that the weight of evidence suggests that by the 1880s the villages had made positive efforts to increase coffee production and had considerable success in doing so.^[46] Maybe, but even if they had rapidly learned the new cultivation techniques, the lack of credit would have placed an enormous obstacle to their efforts. Thus, the failure to modernize the credit system was an important element to insure that only the elite would be able to profit from the coffee business. They could have done it; other countries of Latin America, such as Colombia and Costa Rica, with organizational problems similar to El Salvador's, but with different allocation of resources and political options, did find ways to finance investment in coffee plantations. In those countries coffee plantations prospered without abrupt changes in land tenure, and credit systems developed to satisfy the needs of the planters. Mortgage credit, future sales, and direct credit from foreign houses solved the problem.^[47] In fact, El Salvador was developing the same mechanisms quite rapidly. There is no question that it was a difficult task. The initial cost of updating financial laws and setting up a banking system would be followed by the difficulties of managing complex operations. The transaction costs of lending to a great number of small producers would have been high, and enforcing repayment

would have been a major undertaking and a source of constant tensions. The lack of people with knowledge or expertise in financial matters made the problem worse.

The practical dimension tilted the scale against a sweeping credit reform. It was easier for a poorly organized and underfunded central government with little expertise in financial matters to unleash the land market than to organize the credit market. In the end the enforcement of the new arrangements was ultimately left in the hands of the new landlords, and the state did not commit itself to a long-term involvement beyond its capabilities.

If the state was to play any role in modernizing the economy, the land tenure system was the easiest target, but by no means the only one. The goal of modernizing the credit market was within reach, and it was quite possible to leave ejidos and communal lands alone and still produce coffee. Every choice, be it to modernize the markets of land, credit, or labor, involved technical as well as political difficulties, but none was impossible. The final result, the liberal reforms, provides a snapshot of the economic and political balance in the 1880s. After all, not every country would have found it easy to change the rules of the game for an important segment of the population without fearing a major social upheaval. Choices were made not only in terms of growth but also in terms of allocation of resources between different social groups.

Extent and Timing of Land Transfer

The extent of the land transfer is not entirely clear. Some communal lands were successfully claimed by their occupants. But changes in land tenure followed the pattern imposed by the profitability of land. Browning makes this point when he states:

Villages away from the coffee areas were less affected by abolition: their more dispersed population meant that the many villagers who lived and worked on isolated plots were able to claim these by right of possession, as poseedores actuales ; there was less competition for the common lands and less intrusion into them by private estates; the character of the private estates, generally cattle haciendas with large areas of uncultivated land, allowed continued unregulated settlement by squatters, unlike the coffee plantations.^[48]

These variations in the transfer of land tenure strengthen the interpretation that the liberal reforms were the product of competition for

scarce land and the unequal distribution of power between the elite and the Indian communities and towns. Where land was less attractive and, therefore, there were less incentives to compete, the villagers were allowed to claim their land.

The transfer of land took more time than the liberal leaders had expected. The administration of the process of issuing new titles for all the ejidos and communal lands proved to be difficult. The law of 1882 gave six months to occupants to claim ejido land. They were supposed to go to the mayor of the municipality where the land was located, prove that they were cultivating the land, pay a fee, and obtain a title to the land. If they failed to claim the land, after the six-month period they would lose their rights and the land could be sold in public auction. It was most unreasonable to assume that the process could take so little time. Before long it was realized that six months was too short a period; the municipalities were inundated with claims and the period was extended. Many mayors were illiterate, did not understand the law, or found it to their advantage not to understand it. Indian communities, in turn, showed "little will to divide their property."^[49] The problem was complicated further when the government required that titles should be registered in the land registry. In March 1884 the executive branch was still extending the period for issuing titles for ejidos. The extensions were always short, between three and six months, and seem to have been enacted to ease administrative problems rather than to make it easier for the villagers to claim their lands.

It is not difficult to imagine the obstacles in granting land titles. The efforts of powerful individuals to take undue advantage of the laws, the number of claims, and the limited ability of new institutions to handle the problem contributed to lengthen and complicate the process. The legal system was slow, and issuing land titles was a "delicate problem" that opened up opportunities to the unscrupulous. Skilled lawyers hired by rich landowners with friends in high office had an edge over illiterate peasants. That natural advantage was magnified by the lack of distinction between public and private interests. An illustration of this problem is given by the ordeals of a British merchant who, frustrated with Salvadoran political and judicial systems, asked his government to help him. Moses Levy, who had businesses in El Salvador, complained to the Foreign Office in 1881 that El Salvador was a small country "where the natives are bound together by ties of consanguinity, intermarriage or similarity of interests and where it often happens that men of unprincipled character are raised to positions of influence, whence they can use their power to defeat the ends of justice; it is easy for persons to arrange together to bring about this wholesale system of robbery."^[50] Poor Mr. Levy was talking about his personal battle to settle financial claims and

not about the liberal reforms, but the situation of illiterate members of Indian communities could not have been better.

Notices published in the Diario Oficial provide a glimpse of the problems encountered. In July 1882, Higinio Calderón, administrator of the common lands of the town of San Ignacio, published a notice announcing that he was dividing the property up and giving an ultimatum to prospective claimants to voice their claims or lose them. Two weeks later Natividad Flores, the rightful administrator of the lands, challenged Sr. Calderón's rights to carry out the partition. Sr. Flores, it turned out, was illiterate. He was lucky that somebody warned him of the notice, but the trick was clever, and it is very likely that it was imitated by many. Most of the land was owned by illiterate people who could easily miss a newspaper notice. Many of the problems in registering lands had nothing to do with the complexity of the law but, as the governor of San Vicente pointed out, they arose because "some councilmen oppose their private interests to the common good."^[51]

Even under ideal circumstances the practical aspects of giving land titles for one-quarter of the national territory were overwhelming. A random sample of ejido lands that received title in the western provinces between 1882 and 1885 helps to illustrate the problem. The average size of each ejido plot claimed was 30.38 hectares at an average price of two pesos per hectare.^[52] The average size of the communal land plots claimed was even smaller, 23.625 hectares. This means that if ejidos and communal lands accounted for about 25 percent of the territory of the western region it would have been necessary to register more than 3,500 titles. During its first three years the Registro de la propiedad raiz de la sección de occidente registered 2,169 titles and transactions, 20 percent of which did not involve claims of ejidos or communal lands.

For those who wanted to take a more serious approach the practical problems were even greater. The Indian community of the village of Coatepeque hired a surveyor to divide its land among its four hundred members. It was a relatively rich community that owned 2,961 hectares of land. But imagine the problem if every community tried to be scientific about distributing its land. El Salvador simply did not have enough surveyors to do the job. As late as 1898 El Salvador had only nine individuals who called themselves civil engineers and who were just a little more than surveyors. The National University began teaching civil engineering in 1879. Conflicts about the partition of land were bound to arise.

The situation was further complicated by the different idea of land espoused by the Indian communities.^[53] Traditionally balsam growers owned balsam trees although the land on which they sat belonged to the community. With the privatization of land the rights of the tree owners

could not be ignored. When the land of Santa Isabel Ishuatán, a village located on the balsam coast, was distributed, the titles indicated that "if the land included balsam trees belonging to third parties, [the title holder] should allow the tree owners to cultivate and extract balsam from them. Both the title holder and the tree owners would have to pay a fine of ten pesos per tree if trees are destroyed."^[54] (The fine was substantial; the average market price of land was about eighteen pesos per hectare.) Similar provisions were included in titles issued by the village of Caluco, Sonsonate. Not surprisingly, it took more than six months to privatize all the land in the country. In order to obtain final title it was necessary to publish three notices in the Diario Oficial . In some instances the last notice was not published until 1899.^[55] After ten years the problem still was, in the words of the minister of government and war, "annoying in a thousand ways, of a serious character, and delicate."^[56] The last disputed lands became national property in 1891, and the last chapter of the "annoying" story was written in March 1897, when the Ley de titulación

de predios rústicos was passed. By that law the government gave away its rights on former ejidos and communal lands to those who occupied them in good faith.^[57]

Evidence from the land registry indicates that the privatized land was not equitably distributed. But this was not by any means a new feature in the Salvadoran landscape. A study of the size of private plots and ejidos in our sample shows a great degree of inequality in both categories. Most plots were under 5 hectares, but most of the agricultural land of the country was in larger estates of more than 100 hectares (see tables 24–27).^[58] In the case of the ejidos 73 percent of the land was given to 5.68 of the new owners whereas the 50 percent less fortunate scrambled for a share of 3.45 percent of the land. Private land was distributed in a similar fashion. The larger estates, 14.51 percent of the total, accounted for 81.9 percent of the land, whereas plots under 5 hectares, 54.83 percent of the total, accounted for 2.64 percent of the land. Multiple ownership was more likely to occur in the case of larger estates than in the case of smaller ones. This suggests that the reforms reinforced a trend

that was already very well established: they brought nonprivate land into line with the rest of the country. Ejidos and communal lands were already on their way out; the pressure of export agriculture was already eroding traditional forms of land tenure. The generalized inequality in land distribution may not have originated with the liberal reforms, but the reforms certainly helped to make the phenomenon lasting. By 1950 the situation had not improved. According to that year's census only 1.17 percent of all the agricultural properties in the country were greater than 100 hectares, but they accounted for 50 percent of the arable land. By contrast, 42.56 percent of the properties were under 5 hectares and accounted for only 11.62 percent of the arable land.^[59] By an overwhelming majority, the people who claimed land described themselves as agricultores or jornaleros (day laborers), but the reforms opened up opportunities for other people. Among those who claimed to have rights to ejidos and common lands one finds construction workers, blacksmiths, teachers, merchants, lawyers, physicians and, in the case of one of the largest claims, a field marshal.

The liberal reforms did not affect 40 percent of the Salvadoran territory, were not the beginning of land concentration or of minifundios , and were not indispensable for the expansion of coffee production. But no matter how much one may qualify the argument, they still stand out as a highly symbolic event in the economic and social history of the country. They epitomized and accelerated the changes that took place in the nineteenth century: the opening of the economy, the strengthening of an elite, the emergence of a numerous landless peasantry, and the increasing importance of coffee production. The reforms were not imposed to achieve economic efficiency, nor did radical redistribution of land achieve more growth than what a slower and more equitable process would have accomplished. The reforms arose from the struggle between different social groups to claim the benefits of the land, and not from the efforts of an enlightened minority to increase the efficiency of the economy.