Cattle and crisis: The genesis of unsustainable development in Central America

Cattle and crisis: The genesis of unsustainable development in Central America

Patricia Howard-Borjas

Professor

Chair, Department of Gender Studies in Agriculture

University of Wageningen, the Netherlands

In this article, the impacts of ranching expansion are described and analysed through an examination of the interrelations between cattle ranching, land use, forests and the environment; between ranching, rural employment and migration; and between cattle, food security and nutrition. Each of the aforementioned processes is defined more precisely and related to the cases of Honduras and Nicaragua, beginning with the relations between pasture expansion and land use, concentration and value in Honduras. The relations giving rise to agricultural unemployment and population flows are explored, as are the determinants of the capital-labour ratio in ranching, and migratory flows in Honduras and Nicaragua are discussed in relation to agricultural frontier expansion, urbanization, deforestation and environmental degradation. The impacts of all these relations on purchasing power, food consumption and food security are then traced.

Latin Americans use the term ganaderización (literally, "cattleization") to characterize the process of ranching expansion. It denotes a process of takeover, of total domination. It captures the massive changes in land use that occur as livestock and pasture encroach on virgin forest and areas settled by tillers. But. as peasants in Central America know, ganaderización also has broad negative socio-economic connotations. While ranching provided high foreign exchange earnings through beef exports, it also produced a deepening of the most important environmental and social problems of the age: deforestation, environmental degradation, declining crop yields, unemployment, lack of access to land, social conflict, rural out-migration and increasing malnutrition and food insecurity. Ranching also served to block other, more productive investments in agriculture, which is the most important economic sector in the region. In short, cattle ranching and crisis in Central America are inextricably linked.

Ranching began to expand rapidly in the region in response to the opening of the United States market for imported beef at the beginning of the 1960s, but ranching had been the region's economic mainstay throughout the colonial and post-colonial periods. The first cattle appeared on the isthmus in 1520, within 15 years of the conquistadors' first landing. They were part of Spain's cultural heritage, and along with them came the latifundia, which in effect were large cattle ranches. Because ranching requires very little labour, it could thrive even after the indigenous labour force had been largely decimated. Cattle production in turn provided the material basis for the social, economic and political structures of the colonial and post-colonial period. It permitted the latifundia to exist as relatively autonomous units of production by providing them with food and cash. In turn, the relative isolation and independence of the latifundia laid the foundation for the intense regionalism and conservatism of the era.

Smallholder crop production and extensive cattle ranching dominated the countryside over most of the region up to the end of the First World War. Most farmers produced cattle and crops for subsistence. Land was readily accessible and, although not all was highly productive. the traditional swidden production systems with their long fallow periods kept yields stable. Even after the Second World War, the transition from a post-colonial order to a "modern" economy did not provoke drastic changes in the rural productive structure. While there were technical revolutions in some agricultural export sectors such as banana and cotton, there were no such significant changes in cattle or grain production systems. These technical revolutions were to cause ripples in the rural economies but, by comparison, the expansion of traditional ranching systems, which began in earnest in the 1950s, was to provoke a tidal wave. It was not of the kind foreseen by those who proclaimed that beef exports would become a bonanza for the region. Rather, it became the basis of the region's wholly unsustainable form of development.

The economic weight of other agricultural export sectors increased after the Second World War but their global importance in the rural productive structure diminished as ranching expanded. In the 1950s, the great United States banana companies held a monopoly control over much of the most productive land on the fertile coastal plains and became the focus of widespread social conflict. They began to rid themselves of vast extensions of land, sometimes converting it to pasture and sometimes selling it to ranchers. Since those who demanded access to this land were banana company workers, the companies made efforts to raise productivity and were able to decrease their labour force dramatically.

The net result was that the preponderance of the banana companies in local markets for labour and land and, therefore, their position as centres of social conflict had diminished greatly by the 1960s (Alionso and Slutsky, 1982).

With the opening of the United States market for industrial-grade beef, ranching expanded over most of Central America, provoking major changes in agricultural land use which, in turn, set off an unprecedented series of negative impacts. Ranching continued to be extensive. Each head of cattle required 1 ha or more of land, so the process of ranching expansion necessarily implied great increases in pasture area (Table 1). Cattle ranching came to occupy land of every agricultural potential: large extensions of land not suitable for agriculture, as well as high-quality land, were converted to pasture.

While cattle production in Central America requires a lot of land, the use of labour is more or less six person-days per hectare per year, meaning cattle production employs the lowest amount of labour of all agricultural activities. This is to say that the capital-labour ratio in cattle production is very high. But, in this case, a high capital labour ratio does not translate into high productivity, given that capital is used very extensively over a long period of time: it takes from four to six years to produce a cow or steer of slaughter weight. Given the high capital-labour ratio, ranching expansion gave rise to increasing unemployment and underemployment, both in rural and urban areas, and therefore depressed incomes and effective demand throughout the national economies.

TABLE 1

Changes in pasture area in Central America, 1961/65-1980

Country	1961/65 ('000 ha)	Percentage of area 1	1980 ('000 ha)	Percentage of area 1 1960-80	Percentage change
Costa Rica	969	65.52	010	76.3	107.4
El Salvador	606	47.4	610	42.2
Guatemala	1 039	41.2	1 334	42.3	28.4
Honduras	2 000	56.2	3 400	64.9	70.0
Nicaragua	3 384	71.4	4 880	78.6	44.2
Panama	910	61.2	1 161	65.9	27.6

¹ Total agricultural land area.

Source: CEPAL, 1985.

The two processes - conversion of land to pasture and increasing rural unemployment - together resulted in migratory Rows, deforestation and environmental degradation. Cattle production can occur even on very poor-quality land, which triggers competition for this type of land, increases incentives to incorporate frontier land and induces environmental degradation that exacerbates competition for land access. It also has another important effect: it tends to increase the concentration of land in the hands of large landowners, since more land must be incorporated in order to expand production, and it is basically larger ranchers who have the financial and political power to increase their herds and their land holdings. Many small producers practiced semi-nomadic swidden agriculture, leaving land to lie fallow for long periods before returning to sow crops. As large ranchers expanded their herds, they claimed this fallow land and converted it to pasture. Since small producers usually did not have land titles, they had little recourse other than to attempt to farm on the limited land they had left, continuing to use the same swidden systems. The result has been declining soil fertility and crop yields and an end to the sustainability of swidden agriculture.

What triggered these processes in Central America was a new investment opportunity: the export of live cattle and beef. Beef production and exports expanded continuously over the period 1950-1980. International beef prices guaranteed high profits to anyone who could export the "red gold" (Howard Ballard, 1987). International lenders and local governments provided credit for the expansion of beef production. Large indigenous herds and vast tracts of "idle" land provided the basic conditions for increased production. National and foreign industries opened export beef slaughterhouses authorized by the United States Department of Agriculture and consolidated United States market channels. Agreements between the new vertically integrated firms permitted them to divide markets and form monopoly prices both for cattle purchases and wholesale beef sales in the internal market. At the same time, the internal demand for beef grew and entered into competition with exports, with the result that prices to Central American consumers were driven up to levels higher than those in the United States wholesale market.

The low productivity and production of basic food crops found in Central America, together with their relatively high cost to consumers, are in part attributable to cattle production. First, land suitable for agriculture was converted to pasture, displacing much basic grain production to land with low or no agricultural potential. This depressed productivity and incomes that producers could derive from grain production for the internal market while. at the same time, retail prices for these goods were increasing. Second, rural unemployment and land lessness grew, implying that a larger proportion of the population came to depend on food purchases instead of production for its own consumption. Since incomes were low owing to high unemployment and underemployment, purchasing power decreased. With overall yields stagnating, the population expanding and purchasing power decreasing, the consumption of basic foodstuffs fell below minimum levels.

To satisfy the minimum food needs of the population in Central America in the year 2000. it has been estimated that basic grain production will have to double over the 1985 level and the area in grain production will have to increase at a rate of 3.9 percent per year, in comparison with the 1 percent annual increase that occurred between 1950 and 1984. In 1980, only 9 percent of the land on farms was used to produce maize and beans in the region (CEPAL, 1985). Nevertheless it is worth noting that. in spite of the tremendous increase in cattle production over the period 1950- 1980, there was little change in per caput consumption of beef and milk and, in some countries. there were even declines.

In this article, the impacts of ranching expansion are described and analysed through an examination of the interrelations between cattle ranching, land use, forests and the environment; between ranching, rural employment and migration; and between cattle, food security and nutrition. These are not discreet relations: each set of relations is in turn interlinked, and only by understanding these links and seeking to affect their interactions can sustainable development be achieved.

In the following sections, each of the aforementioned processes is defined more precisely and related to the Honduran and Nicaraguan experiences. beginning with the relations between pasture expansion and land use, concentration and value in Honduras. The relations giving rise to agricultural unemployment and population flows are, explored, as are the determinants of the capital-labour ratio in ranching, and migratory flows in Honduras and Nicaragua are discussed in relation to agricultural frontier expansion, urbanization, deforestation and environmental degradation. The impacts of all of these relations on purchasing power, food consumption and food security are traced.

Both Honduras and Nicaragua have experienced different degrees and forms of deforestation, environmental degradation and rural population displacement attributable to ranching expansion. The Honduran case is far better documented statistically but the Nicaraguan case has points of particular interest. The causes and consequences of the conversion of land to pasture in terms of population displacement are very similar, with only one important exception: Nicaragua's agrarian frontier is more sharply deemed and consists of the largest reserve of tropical humid forest north of Colombia. In Honduras, ranching expansion led to the deforestation of pine and broadleaf forest within already populated areas, and migration to tropical forest areas is recent.

In Nicaragua, the displaced population migrated steadily away from the rich volcanic soils of the Pacific towards the drier and less fertile interior and spilled into the tropical humid forest. Focusing on this last process can delineate the ongoing cycle of deforestation and environmental degradation which is triggered or sharply exacerbated by pasture expansion.

CATTLE AND LAND

In 1952. Honduras was a country with great ranching potential. About one-third of all land on farms was already in pasture, another half was idle and only one-fifth was cultivated semi-intensively with annual or permanent crops. Agricultural production was limited by the restricted size of the internal market and by the fact that the banana companies already supplied the limited export market. The massive conversion of land for ranching signalled the first major expansionary phase of endogenous investment in agriculture in the modern era. With the opening of the export market, beef became the biggest business in Honduras. For this to occur, production had to expand through the incorporation of thousands of producers into newly formed market channels. In fact, the spread of ranching became the major force in the transition to a market economy in rural areas. Production for the market increased through the creation of new ranches and the conversion of existing farms into ranches. Ily 1974, more than 80 percent of all farms of 20 ha or more produced cattle for the market.

Land availability conditions producers' specialization within ranching. At the beginning of the expansionary period, most production was integral (breeding, development and fattening all took place on the same farm). As the market expanded, producers began to specialize in one phase or another of the production process, depending on the amount of land they had available and their market access. Smallholders could not easily specialize in the more profitable fattening operations, since this requires a large amount of land, so they tended to specialize in breeding, which is more land- and labour-intensive. The largest holders specialized in fattening, and the intensity with which they use the land is the lowest of all farm strata. This specialization and differential market access was to determine both how land and cattle would be distributed and to whom the riches would fall. To expand cattle operations, farmers had to obtain more land to accommodate larger herds especially the larger producers who use land more extensively. Land pressure began and, with it, land values rose, land concentration and Endlessness increased and land struggles began in earnest.

While the struggle for land in Central America has drawn worldwide attention, not many have associated it with cattle ranching. Prior to the 1950s, when land had little value, the rural population had relatively free access. The expansion of ranching permitted producers to use land to generate revenue, which of course had to give an enormous boost to land values. While there are no statistics available to show changes in land values, a series of hypotheses can be formed based on the patterns of land-use change and new land incorporation.

Land conversion

Two phases in the postwar development of ranching in Honduras are related to land values and market expansion. The first, which extended from the 1950s to the mid1960s, was based on the growing export market for live cattle and entailed the conversion of farmland in fallow, forest and brush to pasture. The second phase, which covered the period from 1965 to at least 1980, corresponded to the expansion of the United States market for imported beef and entailed the incorporation of virgin land, usually state-owned.

In the first phase, the total area in farms decreased, whereas in the second it increased substantially (Table 2). The decline in area between 1952 and 1965 is possibly related to the process of pasture conversion. As land began to be capitalized. degraded or fallow land could be abandoned. The incorporation of new land in the second phase of expansion reflects the relative saturation of land already in farms and land values and prices which were high enough to bring new, less accessible land into production.

TABLE 2

Evolution of land use on farms in Honduras, 1952-1974

Land use	1952		1965		1974
	Hectares	Percentage	Hectares	Percentage	Hectares	Percentages
Annual crops	296411	12	342267	14	366341	14
Permanent crops	174653	7	190556	8	212011	8
Fallow	424 797	17	255 020	11	140 291	5
Pasture	822 562	33	1 131 198	47	1 347 177	51
Forest and bush	727 365	29	460 472	19	533 046	20
Other	61646	371186	3	30390	1
Total area	2 507 404	100	2 420 649	100	22 629 859	100
	Change 1952-1965			Change 1965-1974
	Absolute		Percentage	Absolute		Percentage
Annual crops	445 856		16	24 077		7
Permanent crops	15903		9	21455		11
Fallow	- 169 747		-40	-114 729		-45
Pasture	308 636		38	215 979		19
Forest and bush	-266 893		-37	72 574		16
Other	9 540		16	-40 796		-57
Total area	-86 755		-4	209 210		9

Source: Calculated on the basis of data from the agricultural and livestock censuses of 1952, 1965 and 1974.

Although total farmland area varied only slightly between 1952 and 1974, there were vast changes occurring in land use. In 1952, some 800 000 ha were in pasture. By 1974, there were 1.3 million ha, representing around 51 percent of all farmland. The increase of half a million hectares in pasture was five times as great as the increase in the area of annual and permanent crops (+100 000 ha) and was accompanied by a reduction in fallow and forest that was nearly equal to the increase in pasture, annual and permanent crops combined (-600 000 ha). More than half of the growth in pasture land took place in the first period. Land conversion was stimulated by the opening of a market for live cattle in neighbouring El Salvador. Pasture increased by some 300 000 ha, compared with an increase in annual and permanent crops of 60 000 ha, and a decrease in fallow and forest of 300 000 ha. When Honduras began selling in the United States beef market, in only nine years (1965-1974) pasture increased by 200 000 ha, again five times the increase in annual and permanent crops (+46 000 ha). The area in brush also increased, signalling the abandonment of degraded pasture. But, in the eight years between 1974 and 1982, pasture expansion reached unprecedented dimensions. In 1982, 2.2 million ha of farmland were used for grazing, which represents a 900 000 ha increase over 1974. Had the total area in farms remained constant, pasture would have occupied 84 percent of all farmland in 1982.

During the first phase, there was a process whereby farms were integrated into market channels and land capitalization. Pasture came from land previously in fallow. forest and brush in all departments except those bordering El Salvador, where pasture expansion was most intense and the availability of idle land was restricted. These departments not only had reductions in forest, fallow and brush, but also saw a decrease in annual crops. Over the entire country, the 400 000 ha decrease in other land uses was explained by the decrease in the area in farms (-87 000 ha), plus the increase in pasture (300 000 ha). Of this decrease, 270 000 ha were in pine forest and brush and 85 000 ha were in fallow, which accounts for 89 percent of the change in land use.

The loss of fallow land signalled the end of sustainable swidden agriculture practiced by small farmers. In this system, land must lie fallow so that soil fertility can be naturally regenerated. The destruction of this traditional system implied that small farmers were forced to change management practices or purchase inputs to maintain soil fertility. or rapid degradation of land resources would result, which of course was the ultimate effect (DeWalt, 1982; Howard Ballard, 1987). The loss of forest land also represents the incorporation of soils of low or no agricultural potential.

In the second phase, from 1965 to 1974, pasture expansion occurred exclusively through the incorporation of new farmland in Atlantida, Colon, Yoro, Olancho and El Paraíso, which constitute a geographical belt extending from the southeast to east and northeast, most of which was agricultural frontier land, and in La Paz to the west.

Cattle rent

As long as there was no productive use for land, it effectively had little value. As ranching expanded, idle land was brought into production. Producers probably first incorporated higher-quality land and, as land became scarcer, proceeded on to land with less or no agricultural potential. During the first phase, several things must have happened. First, to the degree that producers gained access to land by paying for it, the market price for cattle had to be high enough to permit payment of that land price and still allow the producer to break even or permit a profit. The price of cattle would therefore also yield a value to those ranchers who already owned their land. Second, the price of cattle also permitted poorer-quality land to be brought into production, which would have increased the value of higher-quality land as a result of higher relative productivity levels. Third, the growing predominance of pasture in land use gave way to competition with alternative land uses, creating a situation where, not only did ranching produce land rent on ranches, but it must have increased rent on all other agricultural land and therefore must have increased all other agricultural prices on a par with cattle prices.

The conversion of idle land on farms was nearly complete by 1964, which meant that further investments in ranching would either be oriented towards increasing productivity or towards increasing the size of the herd and the amount of land available, or a combination of the two. Technological improvements were restricted by the cost of imported inputs, which were prohibitively high, given existing cattle prices. The rational alternative was to incorporate new land, virtually free of cost. The incorporation of land on the agricultural frontier should have served to inhibit the increase in land values over much of the country. Nevertheless, the basic reason for frontier expansion in Central America was that ranching saturated existing land, land values increased and ranchers encountered serious obstacles to increasing productivity.

The capacity of cattle production to outcompete grain production is clear in the Honduran case. A 1982 study, sponsored by the United States Agency for International Development (USAID), showed that: "... it is notable that only a third of the land classified as adequate for cultivation is utilized in this form. In contrast, the majority of the land used for annual and permanent crops has been classified for forest or extensive uses. These apparent paradoxes are caused principally by the existence of large properties which use fertile soils extensively for cattle or which maintain land in reserve, while small parcels on steep slopes or poor soils are used intensively by a dense population" (Campanella et al., 1982, p. 57-58).

The basic explanation for these phenomena was anticipated in 1776 by Adam Smith in The wealth of nations. First, Smith assumed that the production of basic grains would be the reference point for determining the magnitude of agricultural rent, grains being the main source of sustenance for the population and generally the type of production which predominates in agricultural land use. Second, Smith noted that: "The cattle bred on the most uncultivated moors, when brought to the same market, are, in proportion to their weight or goodness, sold at the same price as those which are reared upon the most improved land. The proprietors of the moors profit by it, and raise the rent of their land in proportion to the price of their cattle."

In Honduras, there are two tendencies evident in the relative prices of cattle and maize: the first is for producers' prices for cattle compared with maize to fluctuate slightly depending on increases or decreases in world beef prices; and the second is for consumer grain prices to rise faster than beef prices. This tends to support the hypotheses regarding land conversion and land values, since the increase in beef prices permits an increase in land value and, if these values become generalized in agriculture, they must do so through grain prices. Grain prices have to rise either at the same rate or faster than beef prices. While it is not feasible to make empirical comparisons between grain rent and cattle rent, the hypotheses developed here, building on Smith, are not contradicted by the general pattern of landuse change which must be explained by referring to relative prices, including absolute and relative land values.

If the expansion of ranching to a position of predominance in land use had no other effect than to increase land values, and if the relative prices that spurred this expansion put land into pasture rather than into grain production, then it is logical to conclude that: i) a major source of income in cattle production was derived from land rent; ii) the expansion of ranching increased land values not only directly but also indirectly through all other production that was in competition with ranching for land. iii) ranching expansion increased income transfers from other productive sectors towards agriculture through higher agricultural prices. which would have tended to reduce the overall profitability of other sectors and depress the final demand: and iv) grain production would have been pushed continuously on to land which was not appropriate for agriculture. Cattle ranching was the principal means by which landowners realized the value of their land. and therefore landownership and land rent were principal incentives for pasture expansion. The tact that a profit could be realized even on land with no agricultural potential any that cattle ranching was more profitable than virtually any other activity. including rational forest exploitation. basically explains the massive land conversion and hence deforestation and land degradation in Central America.

Land concentration

In 1952. there were 156 000 farms in Honduras. By 1974, the number had increased by one-quarter. But the area in farms had risen by only 5 percent, show rig that the average farm size was on the decline. The number of minfundia (less than 10 ha) increased rapidly, but this increase was not accompanied by a proportional rise in area. Minifundia represented 75 percent of all farms in 1952. rising to 78 percent by 1974 (only about one-quarter of these farms raised cattle in 1974). The area in minifundia increased by only 9 percent so that, in 1974, on average each farm in this sine group had only about I ha of land. Farms with 5 to 50 ha accounted for almost one-third of all farms in 1952 and for less than 20 percent in 1974, but the area in this group stayed about the same. The only significant change in the distribution of land came from a partial dissolution of the latifundia, which gave way to the formation of new medium and large farms (50 to 1000 ha).

The partial dissolution of the latifundia in Honduras was due to an agrarian reform process. In 1962, with the enactment of the first Agrarian Reform Law. and once again in 1974 with the revision of this law. Iatifundistas were compelled to divide or sell off part of their property. Rant has represented the majority of all large farms in the country (about 90 percent). and they were not protected by any of the law's exclusions. Ranching is the activity with the lowest intensity of land use. and land serving a low-intensity use was specifically targeted by the 1974 law. In some cases. part of the land was sold: in others. it c hanged formal owner while the original owner maintained control; and in still others, the land was simply divided among family members. Yet the agrarian reform law did not really discourage cattle ranching. Indeed, people with farms of 200 to I 000 ha were not very concerned by the law's provisions with regard to intensity of use: in 1952, 77 percent of the farms in this category raised cattle. and this proportion rose to 92 percent in 1974.

Looking only at land in pasture (Table 3). in 1952 and 1974, 2 percent was in minifundia: by 1982, this had fallen to 1 percent. The small producer category (10 to 50 ha) maintained a nearly constant proportion of pasture land over the entire period 1952-1982. at around 28 percent. Therefore, it was principally the medium and large ranches that benefited from the partial dissolution of the latifundia. where the proportion of land in pasture in these categories increased from 46 percent in 1952 to 54 percent in 1982. I he proportion of pasture land held by latifundia fell from 26 percent to 16 percent and finally to 8 percent over the same period.

What was the role of ranching expansion in the global processes of concentration in Honduran agriculture? To answer this question, it must first be determined to what degree the concentration of pasture land contributes to global land concentration anti. secound, to what degree and in which categories ranching expansion was important for the formation of new farms.

Ranching came to dominate production on all farms except the minifundia. and it also accounted for the formation al all new medium and large farms over the period 1952-1974. Although the proportion of farms that were ranches diminished in global terms (from 50 percent to 42 percent), this was wholly attributable to the drop in {he rate of formation of minifundia producing cattle. given that ranches increased as a proportion of farms in every category from 20 ha and more to the degree that the increase in the number of ranches explains all of the increase in farms for every category between 20 and 1 000 ha. Data on c hate c s in land use by sine group show that pasture land represented a much greater proportion of total farmland in 1974 than in 1952, especially on farms with more than 20 ha.

TABLE 3

Distribution of farms, cattle and pasture by Arm size in Honduras, 1952, 1974 and 1982

Farm size (ha)	1952	11974	1982
Percentage of farms
<5	42.2	40.3	23.1
5-10	21.1	19.8	18.1
10-20	16.3	16.2	21.8
20-50	13.4	14.9	21.2
50100	4.1	4.8	8.8
100-200	1.6	2.2	4.4
200 1 000	1.1	1.5	2.4
>1000	0.2	0.2	0.2
Percentage of cattle
<5	18.3	9.6	4.9
5-10	13.0	7.7	5.9
10 20	13.4	9.9	11.6
20-50	17.7	17.2	20.3
50-100	11.6	12.5	17.5
100-200	7.9	11.1	15.6
200-1 000	10.8	21.9	20.5
>1 000	7.4	10.2	3.7
Percentage of pasture
<5	1.9	1.9	1.2
5-10	4.0	4.0	3.0
10-20	6.8	7.3	7.3
20-50	15.4	16.8	18.7
50-100	14.1	13.6	18.2
100-200	11.4	13.7	18.0
200-1 000	20.7	26.3	21.4
>1 000	25.7	16.3	12.2

Source: Calculated on the basis of data from the agricultural and livestock censuses of 1952 and 1974; and SRN, 1984

The Gini coefficients show that here was a high degree of land concentration in 1952 (0.75) that did not change over the period (0.76 in 1974). However, regional patterns are quite diverse. Gini coefficients departmental for the departmental level show a sharp process of land concentration in those areas which are agricultural frontiers or dominated by ranching or both. with {he exception of Yoro. which is the centre of foreign and national agribusiness and agro-industry. But. in contrast to what was occurring with farmland in general. there was a slight process of deconcentration of pasture land, which is explained by the partial dissolution of the latifundia.

CATTLE AND LABOUR

Ranching expansion in Honduras and Nicaragua not only entailed increasing latch concentration it also led to an increase in the capital-labour ratio in agriculture (the ratio between capital and labour rose while the relative demand for labour declined). The effects of this were rising rural and urban unemployment and poverty: rural labour force displacement, manifest in rural-urban and rural-rural migratory flows: and greater pressure on ecologically fragile agricultural frontier regions. In this section, the dynamics of employment generation and the relation between employment, poverty and migration are treated briefly. after which all of the relations are examined empirically using, the case of Honduras.

The overall employment effects of ranching expansion in the region are determined in par by the demand for labour in primary production (dairy and beef), marketing (of live cattle and of beef and dairy products - wholesale and retail) and processing (agro-industries and cottage industries linked to slaughtering and dairy production). The demand for labour in any given activity is determined by the productivity of laborer which in turn is dependent mainly on the technological conditions of production, expressed as the capital-labour ratio. The total demand for labour in a sector will depend on the average productivity of labour and the total amount of production.

Increasing the use of capital with respect to labour reduces the relative amount of labour needed while the total amount of labour employed may continue to increase as a result of expanded production. Sectoral unemployment results when increases in demand for labour are not sufficiently high to offset labour displacement, and it can result in overall unemployment if other sectors are not generating compensatory levels of demand. Unemployed or underemployed workers exert pressure on the employed population through competition, which can depress wage levels either overall or in particular sectors. All of the elements that come into play in the determination of the capital labour ratio therefore affect labour demand and supply, and all the factors that condition variations in the capital-labour ratio between businesses and sectors of production also condition the distribution of the labour force. Two of these elements are of special importance in agriculture: production time versus labour time, and the scale of production. Production time versus labour time is important as a determinant of the relative productivity of labour, whereas the scale of production is often an important determinant of the capital-labour ratio, since the use of technology generally increases with the scale of production. In agriculture. the tendency over the long term is to expel labour. This is a result of: i) a tendency to increase the scale of production and incorporate more capital per worker; and ii) the fact that, after frontiers are saturated, additional land cannot be produced (production cannot continue to expand except through productivity increases, which reduces the amount of labour required). Whether this expulsion will proceed rapidly or slowly is the basic question: the main concern is whether or not the "excess" labour in agriculture will be absorbed by other sectors.

It must also be pointed out that, while the capital-labour ratio may be very high, production times may be very long, counteracting the productivity increases to be expected with a high capital-labour ratio. Although there is a very high capital-labour ratio in ranching, production times are long and hence labour productivity is low, which reduces profits. The low profit rate resulting from a lengthy production time is partially counteracted by land rent and partly by specialization.

Permanent migration flows are essentially labour flows from points of expulsion (low demand) to points of attraction (high demand). Further, labour moves in relatively predictable directions and magnitudes: i) from regions (or sectors) with high capital-labour ratios to regions (or sectors) with low ratios where small-scale production or unproductive activities dominate; ii) from regions where little investment is being made to regions where investment is expanding rapidly; and iii) from regions where agriculture predominates to regions where industrial or tertiary sector activities predominate. Investment is also attracted to regions with low wage rates (where underemployment is high), all other conditions being equal. Workers are expelled from one rural area only to reappear in another where they can transform idle land into productive land. This is the basis of many permanent rural-rural migration flows.

Also, virgin lands will continue to attract labour until, relative to the capital incorporated, they are saturated and become either geographical deposits of the unemployed or areas of expulsion.

Ranching expansion in Honduras and Nicaragua entailed increasing land concentration, an increase in the capital labour ratio and the imposition of a very long production process over most rural areas. The combined effects of these changes gave rise to rural labour force displacement, manifest in rural-urban and rural-rural migratory flows, and contributed to a substantial increase in rural unemployement and the opening of agricultural frontiers to penetration and concomitant deforestation.

Global labour productivity in Honduran agriculture increased substantially over the period 1952- 1974, especially in agricultural exports (cotton 125 percent, tobacco 124 percent and coffee 97 percent), which are the most labour-intensive and depend mainly on wages. Increases were also above 30 percent in maize and first-crop beans. Growth rates in production were higher in cotton and coffee, but increases were greater than 30 percent for all selected crops, except second-crop beans, rice and sorghum. These changes reflect the transition from a subsistence-based agrarian economy to a market-based economy, which was accompanied by capitalization and the intensification of production. Dividing total production of major crops by the total active population in agriculture, average productivity per agricultural worker increased from 2.43 tonnes to 2.86 tonnes between 1961 and 1974. This helps in part to explain the slow growth in the agricultural workforce, which rose by only 14 percent over the period in comparison with an increase in the total workforce of 21.5 percent. In general, with area in production increasing only very slowly, a strong process of farm formation and even a weak process of concentration occurring, the increase in productivity resulting from a high capital-labour ratio resulted in a decrease in the relative demand for agricultural labour over time. This occurred even though a significant part of production took place on small farms, where productivity is low relative to larger farms, and despite the fact that agricultural productivity was low globally, even in comparison with other Central American countries.

The use of technology was uniformly low in Honduran as well as Nicaraguan ranching, as was productivity. Productivity varied more in accordance with differential land fertility than with the differential use of technology. In Nicaragua, Latino-consult (1975) reported that most production systems tended to be traditional and extensive no matter what type of land was available. When analysing the economic performance of farms in three categories in Honduras, the government derived estimates that approximated technical coefficients. The weight of land and cattle in overall investment are of primary interest, since these are the components that overwhelmingly determine labour demand.

Table 4 shows that land alone accounted for 47 to 49 percent of total investment among all farm size groups analysed in 1984. Other fixed capital (buildings and corrals) averaged 8 to 19 percent of total investment, including land, whereas operating capital (cattle and inputs for cattle and pasture production) averaged 30 to 34 percent of investment including land. Cattle alone represented from 91 to 98.4 percent, a relationship that held among all regions and all farm categories. Labour costs accounted for 6 to 8 percent of total investment including land, and 11 to 16 percent excluding land. The average ratio of labour to operating capital was between 2.1:10 and 2.4:10. These estimates indicate the overwhelming weight of land and cattle in the coefficients and the very small proportion that is labour cost.

Global labour demand in ranching in Honduras was estimated using data on labour use per head of cattle and per hectare of pasture, considering that the use of other inputs would not cause distortions because of their low incidence. Labour use in ranching on land with a high agricultural potential is probably not superior to that on land of a lesser quality since, on poorquality land, grazing is more extensive and weed invasion is higher, so more land must be cleared and cleaned more frequently per head. In any case, to account for differences based on ecological factors, estimates of labour demand per head were weighed by animal load (head per hectare) to derive regional estimates of labour use.

TABLE 4

Percentage composition of investment in ranching in Honduras, 1982

Size group (No. of head)	I. Structure, including land rent (as % total investment)
	Land rent	Fixed capital	Operating capital	Labour
10-29	47,7	19.1	27.3	5.9
30-99	47.3	11.3	33.5	8.0
100-299	49.2	8.4	34.6	7.9
	II. Structure, including land rent (as % total investment)
10-29	36.5	52.2	11.3	2.2:10
30-99	21.4	63.5	15.1	2.410
100-299	16.5	67.9	15.5	2.3:10

Source: Howard Ballard, 1987, Tables 5.3 and 5.4

Estimates from Nicaragua (Warnken, 1975) show that the labour required to produce each head of cattle is around five person-days per head per year. This translates into approximately seven person days per hectare per year. The ratio of labour to cattle and land in 1982 in the Honduran case was estimated by indirect means on the basis of cattle survey results, presented by the Natural Resources Secretariat (SRN): the ratio is 6.25 person days per hectare per year. This can be compared with the other activities in Honduran agriculture. The 1987 Continuous Labour Force Survey of Rural Areas provided other estimates of the labour force absorbed in "livestock activities". It showed that livestock activities together absorbed an average of 40 500 person-days per month, with a minimum of 35 000 and a maximum of 50 000 in different months. The average represents 8.2 percent of the agricultural workforce and 5.5 percent of the total rural workforce, which are substantially below the estimates for 1982, even though the latter include all livestock activities. Comparing livestock production with coffee production, which is the activity that absorbs the highest amount of labour. coffee employed 63 000 workers per year, with a minimum of 30 000 per month and a maximum of 122 000 per month, which represents an average of 12.7 percent of the agricultural workforce. It is estimated that coffee occupied about 125 000 ha of land (Aguilar, Reyes and Vigil, 1980, p. 54), so that, in comparison with ranching, this activity employed nearly twice the labour on one-thirtieth of the land.

TABLE 5

Labour time employed in different agricultural sectors in Honduras, 1970s

Crop	Day per ha per year
Tobacco	482.6
Banana	390.0
Coffe	202.0
Sugar cane	99.0
Maize	73.1
Rice	22.9
Beans	63.1
Cotton	76.7
Sesame	43.0
Sorghum	61.7
Plantain	83.2
Cattle	6.3

Source: Aguilar, Reyes and Vigil, 1980, and author's own estimate.

Estimating that annual crops occupied some 405 000 ha in 1982, which was equivalent to 15 percent of the area in pasture, it can be seen that maize, beans, sorghum. soybean and rice, activities which employed 52 percent of the agricultural labour force, absorbed more than six times the labour force on less than one-sixth of the area. In 1987, to generate one job in annual crop production, about 1.6 ha of land were required. To generate one job in ranching, about 65 ha were required. Applying these coefficients to the amount of labour absorbed in annual crops in 1987, if all the area in pasture were converted to annual crops, it would be possible to employ an estimated 1 600 000 additional workers, that is 3.3 times the actual quantity of labour employed.

The ratio between cattle and land in 1974 can be used to develop regional estimates of labour force absorption in ranching for the purpose of determining total labour time needed in each department, relating this to the quantum of labour displaced that year by cattle ranching. Table 6 presents estimates of total labour that would have been absorbed if pasture land had been employed in maize production. Ranching absorbed an estimated 4.2 percent of the total labour force and 7 percent of the agricultural labour force in 1974, on 51 percent of the farmland. If all pasture land had been used for crop production, an additional 35 percent of the agricultural labour force, or 25 percent of the total labour force, could have been employed. Some 130 000 potential full-time agricultural jobs were displaced by ranching in 1974. Between 1974 and 1982, the area in pasture increased by 95 percent, that is by 1 238 000 ha, which could have employed some 731 000 workers in maize production, according to the coefficient applied in 1974 of 73.1 person-days per year in maize. Adding together the 1974 estimate of 130 000 displaced workers and the 731 000 jobs in the areas where ranching expanded between 1974 and 1982, the total would be 861 000 additional jobs that could have maintained rural families (i.e. if it is supposed that each employed worker maintains at least three dependents, the total population supported would have amounted to some 2 600 000 persons).

While it is certain that not all the land in cattle production was fit for growing maize, various studies show that a large proportion of it probably was while that which was not is land that is frequently not fit for raising cattle either (see, for example, Feder, 1980; Campanella et al., 1982). The World Bank estimated that some 500 000 ha in pasture are of high enough potential to produce even more labour-intensive crops. If this amount of land, equal to 20 percent of the land in pasture, were to be used for maize production, approximately 148 000 jobs would be created. This would increase rural agricultural employment by one-third at a cost of less than 10 500 jobs in ranching. It is evident that ranching displaced at least 21 times the amount of labour that it absorbed.

The demand for labour in the beef and milk agro-industries in Honduras is also relatively low. The amount of labour incorporated in marketing could not be measured but it probably outweighs the amount used in industrial processing, especially if the large number of retail vendors in the informal sector are included. In 1975, in formal milk and beef agro-industries, 1 615 workers were employed by the 16 firms that were classified as 'livestock, slaughter and meat preparation and conservation" activities. This represented 12.5 percent of all workers in food agro-industries and 4.4 percent of the industrial labour force. The figures for the milk agro-industry were 4.9 and 1.7 percent, respectively, with a grand total of 17.3 and 6.1 percent for both activities with respect to the totals (DGEC, 1977, p. 14). If an annual growth rate of 2.85 percent is assumed in both industries since 1975, there would have been 2 733 workers in both industries in 1982, which would represent 5.5 percent of the labour force in ranching, with a ratio of one industrial worker for every 20 agricultural workers engaged in cattle and milk production. Nevertheless, this figure probably overestimates current employment in the slaughter industry because of the employment losses resulting from the closure of the majority of the export slaughterhouses. Labour absorption by agro-industries is significantly restricted by the seasonally of these industries. which is in turn related to the seasonally of production and an excess of installed capacity that was apparent even before the beef export crisis: the minimum capacity utilized for individual plants in 1982 was 38 percent while the maximum was 66 percent. Comparing the demand for labour in beef and milk agro-industries with the expulsion of labour attributable to ranching expansion, there is virtually no attenuation of labour expulsion owing to agro-industrial growth.

TABLE 6

Total labour use if pasture were in maize production in Honduras, 1974

Department	Days/year in maize ('000)	Worker-year in maize	Maize as % of agric. EAP	Maize as % of agric. EAP	Workers displaced
Atlántida	2 237	7 990	31.9	19.5	7 270
Choluteca	4030	14390	41.2	29.4	13140
Colon	1 832	6 540	31.1	25.9	5860
Comayagua	1 981	7070	26.5	18.9	6400
Copán	2 582	9 220	24.7	19.5	8 490
Cortés	4025	14370	38.7	13.1	12710
El Paraíso	3 104	11 090	34.9	26.8	10290
Francisco Morazán	2578	9210	23.0	6.5	8110
Intibucá	1 028	3 670	20.2	16.1	3 320
La Paz	1 033	3690	24.2	19.5	3320
Las Islas	115	410	30.9	11 0	380
Lempira	1 801	6430	19.2	169	5810
Ocotepeque	946	3 380	27.4	22.0	3 200
Olancho	4 688	16 740	52.0	41.7	15 020
Santa Bárbara	3721	13290	32.7	22.9	12390
Valle	1 295	4630	28.9	20.4	3 970
Yoro	3 652	13 040	35.2	25 5	11 800
Honduras	40648	145170	31.5	19.0	131460

EAP = economical active population

Source: Autor's own calculations based on Howard Ballard, 1987, Table 5.6

MIGRATION AND ENVIRONMENTAL DEGRADATION

The ensuing broad process of rural population displacement accompanying pasture expansion in Central America has been hypothesized here and in other studies: as pasture encroaches on cropland and virgin lands, rural families see their land access so reduced as to make survival impossible. This occurs principally as ranchers take over fallow land held in usufruct. Together with those peasant farmers who find their labour becoming redundant owing to ranching expansion, they migrate either to the agricultural frontier or to urban areas. In the agricultural frontier zone, peasant labour is frequently used by ranchers to clear new land for pasture: peasant farmers are given access to land for a year or two to clear and cultivate, after which pasture is sown and the land reverts to the rancher. Peasant farmers are forced to search for new land, thereby perpetuating the ranching expansion/deforestation/migration cycle.

It is also widely held that the expulsion of the peasantry and their resettlement in frontier regions results in the deterioration of the land resources available to them since these are areas of inferior land quality. On poor-quality land, or where dry and rainy seasons are marked, pasture is often overgrazed, resulting in land degradation (depletion of soil nutrients, weed invasion, erosion, soil compaction) and a reduced capacity to support livestock. This is particularly pronounced in tropical humid forest regions with soils deficient in phosphorus (Hecht, 1982; Romanini, n.d.). As recent studies have shown, the end result is often land abandonment and the opening up of new land, processes that imply continued human migration and deforestation (Place, 1981; Hecht, 1982). To summarize, cattle expansion triggers population redistribution through the following mechanisms: i) conversion of land in agricultural production or fallow to pasture, displacing labour; ii) absorption of land reserves, prohibiting the expansion of other agricultural or forestry activities and productive employment of the labour force; iii) restriction of urbanization and labour absorption within cattle regions owing to the lack of regional employment multiplier effects; iv) displacement of small and medium producers on to marginal land where production is more extensive, quickly triggering land competition in marginal and frontier areas; v) promotion of land abandonment owing to degradation from the over exploitation of resources, further compelling migratory flows and continued deforestation.

Ranching has not expanded to the same degree everywhere, nor has it had the same employment effects. The degree to which ranching gives rise to structural changes in the labour force - its composition and distribution among geographical areas and, economic sectors - depends on a possible alternative employment of labour within a particular geographical area. The alternatives in turn depend on the possibilities for incorporating more land into production. the expansion of more labour-intensive crops and the industrial and commercial growth within the region. Three global scenarios are predictable: i) in regions where pasture expansion is relatively restricted, where urbanization and industrial expansion are occurring or where other agricultural activities are expanding, regional labour displacement will be low; ii) where pasture is expanding rapidly and industrialization is occurring, new land cannot easily be brought into production, so the rural labour force should become urbanized; iii) where little industrial expansion is occurring, possibilities for the incorporation of new land are limited and other, more labour-intensive crops are not expanding, labour resources should be expelled from the region and should seek employment in the rural or urban wage labour market or land in the agricultural frontier regions.

Razing the hillsides: pasture expansion and migration in Honduras

In Honduras. the major historical development leading to population redistribution was the growth of the labour-intensive banana industry on the northern coastal plains at the beginning of the century. Labour demand was high enough to provoke immigration from neighbouring countries as well as from rural areas. However, the area that could be planted to bananas was limited by agro-climatological factors and the overall influence of banana production on the rest of the rural economy was reduced because of this sector's high degree of dependence on imported inputs. By the 1950s, the northern coastal departments had a large wage labour force with a small industrial base and even less possibility for growth in the area of farmland owing to the concentration of land under banana company ownership. In the 1950s, the banana companies began to break up their holdings owing to banana plagues and flooding, but the nationwide banana strike of 1954 also stimulated the companies to increase the productivity of labour. The outcome was a reduction in the number of banana plantation workers from 35 000 in 1953 to 15 500 in 1960. Displaced workers initiated a struggle to obtain idle banana company land and many eventually gained access (Alonso and Slutsky, 1982).

Coffee production was slow to take hold in Honduras and has its basis in small-scale production. It is limited to a few departments in the northwest and south where it has had important effects by creating predominantly seasonal labour markets. Cotton production expanded in the 1950s and 1960s, but had relatively little impact on the labour force owing to its geographical confinement. The decline in cotton production in the 1970s considerably diminished employment in this sector. During the 1950s, 1960s and 1970s, idle and virgin land was largely absorbed by ranching.

Kramer (1986) investigated the relation between pasture expansion and migration within three departments in western Honduras on the Salvadorian border (Ocotepeque, Copán and Lempira). The results obtained, which mainly coincide with those presented here, caused him to reject his original Malthusian assumptions. He initially supposed that four variables would be related to net out-migration: pasture expansion (positively), population pressure (indicated by population density and positively related), labour-intensive cash crop production (negatively related) and land distribution (indicated by Gini coefficients). Of these variables, pasture expansion was in all cases highly positively correlated with net emigration, whereas population pressure was not correlated with emigration but rather with immigration (areas of low population density attract population but areas of high population density do not necessarily expel population). The inequality of land distribution was also insignificant, while the cash crops variable was significant.

The single most important indicator of rural labour force absorption and distribution is land use and, in the framework presented here, land use serves as a proxy for the capital-labour ratio. Island in annual or permanent crops represents a lower capital-labour ratio than land in pasture; land in fallow, brush or forest represents idle capital. A rather complex model of determinants of labour force migration must be constructed at the regional level, taking into account the following: pasture expansion, expansion of other crops, farm formation, new land incorporation and processes of industrialization and "tertiarization".

Labour force mobility between 1950 and 1974 may be examined through estimates of interregional migration. Both the magnitude and flows of migration are of interest, so direct methods of estimation art employed based on place-of-birth data from the 1961 and 1974 censuses, and place of residence five years prior to the census in the 1974 census. To adjust for inter-census mortality in the migrating population and, more precisely, approximate real intercensus migration, global survival coefficients based on mortality tables were applied, one for the period 1950-1960, and the other for the period 1960-1970, as recommended by the United Nations when age-specific data are not available (UN, 1972, p. 8).

At the departmental level, labour force expulsion and migration can be explained on the basis of indicators that reflect the possibilities for employment (Table 7). With respect to agricultural and livestock production, the availability of new land for farming is a critical determinant of employment generation. In Honduras, new land incorporation is influenced by the area in valleys. since the country is very mountainous. Together with the percentage of total land area in farms at the beginning of the period, these variables indicate the gross restrictions on the availability of new land for farming. Next, the percentage of farmland in pasture at the beginning of the period and the absolute increase in pasture compared with the total change in farmland give gross indications of change in the capital-labour ratio in agriculture and, consequently, of a region's ability to employ rural workers. Finally, an indicator of new employment generation (total, agricultural and non-agricultural) is used where the change in the economically active population (EAP) (new employment) is divided by the population of working age (in Honduras, ten years or older) at the beginning of the period:

EAP₂ - EAP₁ / Working age population₁ x 1 000

This gave the generation of employment per 1 000 persons of working age. This indicator has the advantage of homogenizing the population base between regions, and differences in activity rates do not enter into the equation. It supposes that the causal direction is that of employment generation influencing migration and not the reverse (migration does not determine employment generation). For reasons mentioned previously, this indicator can only be constructed for the period 1961 - 1974.

Areas of labour force attraction. Four departments were experiencing substantial net immigration over the period 1950-1974: in two cases it was due principally to processes of industrialization and '`tertiarization" (Cortés and Francisco Morazán), and in the other two to the existence of large tracts of idle or virgin land (Atlántida and Colón/Gracias a Dios). Of the latter two, Colón/Gracias a Dios is the only true agricultural frontier zone, consisting of largely inaccessible tropical moist forest lowland.

TABLE 7 Major indicators of migration in Honduras

Department	Net migration rate 1961-1974 (%)	LAND USE							EMPLOYMENT
		Percentage of total farm area in 1974			Absolute change, 1952-1974		%change in area, 1952-1974		Total change in EAP, 1961-1974		New employment by sector, 1961-1974
		Valley	Farms	Pasture	Farms	Pasture	Annual	Perman.	Absolute	Percentage	Total	Agric.	Non-agric.
Attraction
Cortés	32.6	39.6	45.0	63.7	19 118	29 989	48 7	25.0	45 768	71.4	348	69	279
Colón	27.8	24.4	4.5	44.7	70 374	39 518	151.5	25.1	11584	84.7	351	341	9
Atlántida	18.0	28.0	32.8	47.7	42763	43973	114.7	1.5	13081	47.0	218	152	66
Fco.Morazan	13.3	14.2	23.2	39.7	-98732	-13327	-9.2	-19.5	48909	53.2	256	29	227
Equilibrium
Yoro	3.6	21.7	31.4	42.5	44342	53461	27.8	10.2	13547	36.0	164	124	40
Comayagua	-4.6	11.6	24.1	46.6	-24311	3949	15.9	20.9	9465	33.9	153	87	65
Expulsion
El Paraíso	-4.7	9.5	337	54.4	46795	50623	42.5	83.9	8640	26.5	125	78	46
Olancho	-8.3	10.0	9.4	55.6	120355	93935	101.9	-3.3	9491	30.0	136	113	22
Choluteca	-10.0	25.8	56.6	65.7	-62938	13409	24.9	-0.9	4983	11.3	53	18	34
Copán	-11.1	15.3	57.2	54.8	7005	51 838	17.0	89.2	7045	17.4	86	60	58
Santa Bárbara	124	6.8	520	56.7	34382	92837	-3.6	63.2	11 176	23.9	118	60	58
Lempira	153	1 5	313	44.2	-31308	25443	0.7	11.3	5650	17.5	78	68	11
La Paz	-16.7	1.1	33.6	40.9	-14506	5565	-5.9	45.1	1 383	7.9	35	22	13
Intibucá	-17.3	4.4	35.8	31.6	2357	827	10.2	9.8	1 964	94	42	4	46
Valle	23 5	20 6	42 4	44.7	7 528	17 692	8.2	-42.2	1 235	5.7	24	-9	34
Ocotepoque	-31.8	5.7	47.6	65.6	-3084	15810	-20.9	16.9	-485	-3.1	-14	-6	-8
Honduras	-	15.9	23.5	51.3	525215	525542	23.6	21.4	194807	34.3	160	67	93

With the exception of Francisco Morazán, each of these departments has a relatively favourable proportion of area in valleys relative to total farmland area in comparison with the rest of the country. In Atlántida and Colón, growth in the area in farms was high and either equal or superior to the increase in the area in pasture. This process of farm formation is the source of substantial immigration. Accompanying this growth was an increase in the area in annual and permanent crops, reflecting the conversion of idle banana company lands (Alonso and Slutsky, 1982, p. 29-30). In both Atlántida and Colón, pasture expansion was high over the period, but the area in staple crops also increased. The agricultural labour force increased at a far higher rate than the national average, showing that the attraction in the northeastern region is towards rural areas. The magnitude of employment generation is much lower than in urban departments of attraction but comparable with that of the Department of Yoro, where agriculture is capital-intensive. Atlántida and Colón attract substantial immigration, although this has been limited by the almost complete lack of infrastructure development in the latter department; nevertheless, in the near future Colón's tropical moist forest will almost certainly be a major refuge for the rural unemployed.

Cortés and Francisco Morazán contain the two largest cities in the country (San Pedro Sula and Tegucigalpa, respectively) but differ in that Cortés has a dynamic agricultural export-oriented hinterland and a large industrial base, whereas Francisco

Morazan's hinterland is generally depressed, similar to the situation in the surrounding rural departments, and urban employment is confined to the state and tertiary sectors. In Francisco Morazán, area was lost for almost all crops and. at the same time. the area in pasture decreased by almost 30 percent, which signals a process of land abandonment. Nevertheless, Cortes showed little dynamism with respect to agricultural employment generation, which was barely superior to the national average. In both cases, the generation of non-agricultural employment is the highest in the country, both in absolute and relative terms. In Cortés. the fact that 64 percent of farmland was in pasture, a proportion equivalent to that of the departments most highly specialized in ranching- Choluteca and Ocotepeque - goes far towards explaining why agricultural employment generation in Cortés. a centre of agricultural export production with very fertile valleys, was inferior to that of four other departments.

Areas of labour force expulsion. A total of ten departments, all of which are predominantly rural, experienced net labour force emigration over most or all of the period 1950-1974. At least three of these departments were centres of early ranching development dating back to the colonial era: Choluteca, El Paraiso and Olancho. Five departments (Ocotepeque, Lempira, Intibucá. La Paz and Valle) represent a misery belt along the border with El Salvador and are more integrated economically with that country than with Honduras. Ranching expanded in this region as a function of the growing demand for live cattle in El Salvador before the United States market opened in the 1960s, and it was the purpose of Kramer's study (Kramer. 1986) to show that population expulsion from these departments and from neighbouring Copán is strictly related to ranching expansion, as both proceeded along major cattle market routes.

Ranching expansion and population expulsion in Copán, Santa Barbara, El Paraiso and Olancho intensified with the opening of the United States market in the 1960s and the construction of export slaughterhouses. Copán and Santa Bárbara were initially departments of attraction for labour force, receiving migrants from the departments bordering with El Salvador but, with the increase in ranching, they began to expel population. Choluteca, El Paraiso and Olancho had constantly expelled population and, while the latter two experienced very high levels of pasture expansion over the entire period, by the early 1950s Choluteca was almost completely specialized in ranching and unable to expand much further.

Copán, Santa Bárbara, El Paraíso and Olancho present more complexities in their forms of development and possibilities for labour absorption. Santa Bárbara and Copán had a low proportion of land in valleys (7 percent) compared with land in farms in 1974 (52 percent), but the region is the country's largest coffee producer. While there was a net increase of 41 000 ha in farm area, there was also a net increase of 145 000 ha in pasture, so the proportion of total farmland in pasture jumped between 1952 and 1974, indicating large-scale land conversion. The increase in permanent crops (including coffee) represents only slightly more than 10 percent of the increase of land in pasture over the period. All of this increase in crops occurred between 1952 and 1965, since absolute reductions were evident between 1965 and 1974. The sharp rise in coffee production increased labour demand sufficiently to permit: net immigration over the first period, 1950-1961, but the increase in net agricultural employment was well below the national average between 1961 and 1974. Even though Santa Bárbara had a volume of new employment equal to that of rural departments of attraction, it was not sufficient to retain the growing working age population. Immigration and emigration both intensified over the 1950s, as cattle, coffee and sugar production expanded. By the 1960s, net migration turned negative, involving the largest gross outflow in the country between 1961 and 1974.

El Paraiso and Olancho, in the southeast, also experienced a very sharp process of pasture expansion, but the effects of this on the labour force were ameliorated by three factors: first, the relatively favourable relation in 1974 between land in valleys and land in farms, indicating a high potential for new land incorporation over the period; second, more labour-intensive crops were expanding: and third, an incipient process of industrialization was occurring. That land was available for expanding production is indicated by the change in total farmland (+ 167 000 ha), although most of this change went to pasture (+143 000 ha) which absorbed 55 percent of all agricultural land in 1974. El Paraiso became the third largest producer of coffee after Santa Bárbara and Comayagua, followed by Copán. Tobacco and cotton production also expanded, thus increasing the demand for labour. Part of the new agricultural employment was attributable to forestry activities, Olancho being one of only three departments where the farm area in forest increased, but this growth was not complemented by nonagricultural employment generation: the zone is almost exclusively dedicated to the production of raw materials. Between 1965 and 1974, the area in coffee production increased in Olancho and El Paraiso even more than in Santa Bárbara and Copán (6 000 ha); the area in cotton increased in Olancho whereas it diminished in other departments. The area in basic grains also increased substantially, especially in maize (17 000 ha) and first-crop beans. Nevertheless, these increases are very small in comparison with the 91 000 ha increase in pasture, with the majority of this expansion occurring in Olancho. This factor determined the overall low level of employment generation compared with the area in production.

Net migration rates were negative over the entire period in both departments, although immigration and emigration rates were increasing, indicating that there were many who sought to improve their lot by finding land or employment in a region with a high rate of farmland expansion. This was not the type of expansion that would provide a permanent base of sustenance for the population. In spite of the very large total land area available in Olancho, most of this land is either not suitable for agricultural production or is inaccessible.

Choluteca, Valle, La Paz, Intibuca, Lempira, and Ocotepoque are all economically depressed, with very low infrastructure development. Only Choluteca and Valle have much valley land (20 to 26 percent); the rest are very mountainous. Choluteca had 26 percent of its area in valleys and twice that area in farms in 1974 (57 percent). Total farm area decreased by 98 000 ha but pasture continued to increase, rising from 143 000 to 156 000 ha so that, by 1974, 65 percent of all farmland was in pasture. The decrease in land in farms was grossly attributable to the abandonment of degraded forest and fallow land. Choluteca lost area in coffee production but this was compensated by increases in cotton and tobacco, where Choluteca and Valle were the principal areas of cotton production in the country. Valle's farm area increased over the period, but pasture expanded by 148 percent, incorporating 10 000 more ha than were taken on as new farmland.

Over the period 1965 - 1974, both departments lost almost 10 000 ha in cotton, which definitely meant a loss of employment. Choluteca also lost some 1 000 ha in coffee. This, together with the increase in pasture, was devastating to the labour force. Employment generation per 1 000 caput was 53 in Choluteca and 24 in Valle, in comparison with a national average of 160. Agricultural employment increased by only 5 percent in Choluteca and decreased absolutely in Valle by 3 percent. where employment generation stood at 18 and -9, respectively, in comparison with the national average of 67. In both departments, only 1 200 net new agricultural jobs were created in 13 years. The generation of non-agricultural employment was 34 per 1 000 caput in both departments, higher than agriculture but among the lowest in the country, totalling only 4 200 new jobs. Net migration rates were correspondingly negative, high and on the rise. Already Choluteca and Valle experienced net out-migration in 1950, between 20 and 25 percent of those born in these departments left the region between 1961 and 1974.

La Paz, Lempira and Intibucá have the smallest proportion of land in valleys in the country at 1 to 4 percent. Only 31 to 36 percent of the land area was in farms in 1974. Of this scanty land base, 31 to 45 percent was utilized extensively in pasture and, although La Paz and Lempira lost farmland over the period, the area in pasture rose. In La Paz, 58 percent of the area in fallow was lost, as was 61 percent of the area in forest. In Intibuca, the decrease in fallow and forest went mainly to an increase in abandoned pasture land, which indicates environmental degradation. In Lempira, massive areas were deforested and land in fallow was converted to pasture.

Between 1965 and 1974, the net change in area in basic grain production in these three departments was negative. The only permanent crop that increased in area was coffee, principally in La Paz and Lempira. The total workforce increased by a rate that was two-thirds below the national average in Intibucá and La Paz, and by half the national average in Lempira. Agricultural employment generation was low and, in the case of Intibucá, negative, while the generation of non-agricultural employment was greater in Lempira and very low in La Paz and Intibuca. The inability of these departments to absorb labour is founded entirely on the poor resource base and the highly extensive use made of it by ranching. Net migration rates were negative over the entire period and immigration rates were very low.

Ocotepeque presents the most extreme case of all. since it is the most highly specialized department in ranching. Valleys account for around 6 percent of land area, and the proportion of land in farms reached 48 percent, for the lowest ratio of valleys to farms in the country. Nevertheless, pasture represented 66 percent of all farmland. The increase in pasture came at the direct expense of the area in annual crops, forest and land in fallow. Employment generation was a negative 14 per 1 000 caput. Net migration rates were also the highest negative rates in the country over the entire period. Between 1950 and 1961, the rate was a negative 26 percent, rising to a negative 33 percent between 1961 and 1974.

With respect to migration streams, the heaviest net inter-census flows between 1961 and 1974 came from the northwestern departments of Santa Bárbara and Copán and extended to Atlántida, Yoro and Colón in the northeast, although the largest single flows were to the urbanized department of Cortés on the north central coast.

Cortés attracts migrants from every department except Choluteca and El Paraíso in the south, which are proximate to Francisco Morazán, indicating that rural urban migration predominates over rural-rural flows. All significant rural-rural flows move towards Atlántida, Colon and Yoro, and the departments which contribute are many. The search for employment outweighs the search for land. The magnitude of flows is greatest when these proceed from those departments where cattle expansion is greatest: Ocotepoque, Copan and Santa Bárbara, Lempira, Valle, Choluteca, El Paraiso and Olancho.

Nicaragua: destruction of the tropical humid forest

Changes in Nicaragua's agrarian social structure have typically been attributed to the development of agricultural exports. The major changes studied concern the expansion of coffee in the early twentieth century and the increase in cotton production between 1950 and 1970. With coffee expansion, landlords ceded plots on the peripheries of their plantations to 6`tie down" peasant labour. This system attracted peasant farmers from the Pacific coast to the central and northern highlands where most early coffee production was located (Espinales, 1974). Later, the expansion of cotton production was considered to be a major factor in the displacement of the peasant population from the northern Pacific coastal region and in the opening of the agricultural frontier (Weeks, 1985; Jarquin, 19751. In the 1950s, the frontier included nearly one-half of Nicaragua's land area. By the early 1970s, only about one-third of that area could still be considered to be largely unsettled.

The expansion of cotton production resulted both in the conversion of peasant farmers into agricultural wage labourers and in emigration. It appears, however, that more peasant farmers expelled from their land stayed on to work the plantations than those who outmigrated (Baumeister and Havens, 1983). Nevertheless, although cotton production expanded in one of the most densely populated regions of the country and gave rise to a large-scale expulsion of the peasantry, its global effects on the rural social structure are paled by the expansion of cattle production. The rate of growth in cattle production was higher than that of any other agricultural export crop. The number of cattle increased by 90 percent between 1963 and 1974 and once again between 1974 and 1979. Nicaragua was the largest beef exporter in Central America (total exports increased from 8 300 tonnes in 1961 to a peak of 51 000 tonnes in 1978, or by 602 percent). This substantial expansion in beef production was accompanied by internal migration rates that ranked among the highest in Latin America. and over the period these rates accelerated. The increase in livestock numbers triggered slow but intense migrations from the rich coastal plains and the drier interior regions towards the agricultural frontier where ranching came to predominate, and it also contributed significantly to rural-urban migratory flows. Large population flows, accompanied by colonization schemes and further ranching expansion in the tropical humid forests of the agricultural frontier, spelled massive deforestation.

The Nicaraguan case is particularly difficult to analyse because of the lack of reliable statistics. The 1952 agricultural and livestock survey may have overestimated production, the 1963 census was fairly accurate, but the 1971 census (the last available) undercounted population and failed to cover many rural areas entirely, which particularly affects the analysis of frontier expansion and cattle regions since these had the most serious problems of coverage (CELADE, 1971). There are no aerial photographic surveys or even complete maps of comparable quality that permit the evaluation of changes in pasture as a component of land use over the period, especially in the frontier regions, although LANDSAT images have recently provided a clearer picture of the extent of deforestation and pasture along the agricultural frontier.

In 1950, ranching occupied the same territories that had been dedicated to that activity since the colonial era. Cattle were introduced to the Pacific lowlands in the fifteenth century, and from there cattle and settlements spread to the interior (Ryan et al., 1970, p. 31). The largest concentrations of cattle (slightly more than one-third of the national herd) were found in the interior departments of Matagalpa, Chontales and Boaco, whose borders roughly coincided with the agricultural frontier; another 28 percent was to be found in the northern and southern Pacific region (Leon, Chinandega and Rivas). Between 1950 and 1963, the national herd experienced low growth (6 percent) with a significant geographical redistribution. Ranching tended to be displaced from the northern Pacific coast towards the central interior owing to the expansion of cotton production and increasing regional specialization in ranching, where breeding and milk production became concentrated in the humid highland areas which provide better year-round grazing conditions. As ranching expanded in the Pacific and interior regions, ranchers often obtained land along the humid agricultural frontier to graze herds during the dry season, which meant that twice the land normally employed to raise cattle was brought into production and that the principal activity in the tropical humid frontier region was ranching. Over the eight-year period 19631974, cattle numbers increased by 90 percent. All departments participated in this growth, but many of the trends evident in the redistribution patterns of the 1950s were accentuated - the most pronounced being the trend towards the concentration of ranching in Chontales, Boaco and Matagalpa, which increased their share of the national herd to 45 percent in 1971, more than doubling cattle numbers.

Ranching expansion triggered migration principally from those areas that already had a significant proportion of agricultural land in production. Chontales and Boaco became the principal departments of expulsion (relative emigration from Chontales stood at 36 percent, and Boaco lost 28 percent by 1971) (Ballard, 1985). The main departments of attraction were experiencing large-scale urbanization (Managua) or were frontier departments Zelaya to the east, Nueva Segovia to the north, Rio San Juan in the central southern interior). Zelaya, which constitutes more than a third of the country and consists principally of tropical humid forest, had a very large increase in the volume of immigrants, who came principally from the neighbouring ranching expansion departments of Matagalpa, Chontales and Boaco: by 1966-1971, it received the equivalent of half of the volume migrating to Managua, and relative immigration exceeded that of the latter, urbanized department.

Zelaya's immigration was almost totally destined for rural areas and probably originated almost totally in rural areas. The attraction to virgin land in the frontier zones increased substantially over the period: immigration grew from 1.8 percent annually between 1950 and 1963 to 11 percent between 1963 and 1971. Chontales, Matagalpa and Boaco contributed between 77 and 79 percent of all immigrants to Zelaya over the period 1950- 1971, where the number of immigrants increased in volume by 163 percent between 1963 and 1971. These flows are significantly underestimated owing to the undercounted population in 1971: CIDCA's aerial reconnaissance survey executed in the early 1980s estimated a population figure for the department of Zelaya that is almost twice as large as that estimated from previous censuses (CIDCA, 1982).

Other than faulty census data, there is one small source of information for the analysis of the migratory process triggered by pasture expansion: a survey that covered 271 farmers and was conducted in the area of the Rigoberto Cabezas colonization project in Zelaya in 1965. Results of this survey give strong indications of the reasons underlying the migratory movements towards the tropical humid forest zone.

Taylor (1967) concluded that most migrants came from the central region where pasture expansion was forcing both landowning and non-landowning peasants to search for new land.

Taylor's conclusions emphasize the problems of environmental degradation where agricultural production takes place on the poor soils of the central region. While there have been no thorough studies of environmental degradation in the central or Atlantic regions, there is secondary information that serves to support the hypotheses on structural change in agriculture and continued human migration as a result of pasture expansion, deforestation and land degradation. Data mentioned earlier indicate underutilization of prime land, gross over utilization of poor land, and large-scale use of nonagricultural land. Pasture expansion, as Taylor pointed out, absorbed land in fallow in the early 1960s in those departments where a large percentage of the land was already in production. When that land was absorbed, pasture was carved from remaining forests anti began extending into the agricultural frontier zone. According to Myers (1980, p. 135): "Although Nicaragua is considered to feature the largest tract of TMF Tropical moist forest remaining in Central America, this is only a small portion of what is reputed to have existed half a century ago. It is being eliminated at a rate of at least 400 km² per year (conceivably twice as fast), partly due to shifting cultivation and partly due to settlement agriculture. The principal cause... is the encroachment of man-established pasture...."

Ugarte (1981) has written about the processes occurring in Zelaya. While he noted that colonization of the frontier began in earnest only 20 years before, "today one only observes small areas of forest in almost inaccessible places... that represent no more than 20 percent of the area". Fallows are increasingly shorter owing to "increasing demographic pressure", and this contributes directly to dramatic decreases in crop yields in two to four years. Decreasing yields lead to the abandonment of land to pasture, a process that is normally accompanied by a change in landowners who purchase the best available land. The displaced peasant farmers move further into the tropical humid forest. Pasture that replaces crops slowly loses its initial levels of productivity: "Therefore, due to inadequate animal loads given the dynamic decrease of pasture yields year after year, the usual overgrazing of pasture accelerates drastically the process of pasture deterioration and degradation in these hot-humid regions" (Ugarte, 1981, p. 6). On a parallel, there is a gradual weed invasion, phosphorus is lacking, and protein and energy production is very low. Cattle productivity is consequently abysmal: cows produce on the average one calf every two to two and a half years and milk production averages 2.5 to 3.5 litres per day. It is possible to conclude that, even if cattle markets ceased to expand and cattle populations stabilized, human migration, deforestation and environmental degradation would continue indefinitely.

BEEF EXPORTS AND FOOD SECURITY IN CENTRAL AMERICA

The growth in Central American beef exports which accompanied the opening of the United States market was propelled by a developmentalist vision. USAID added other reasons for promoting ranching: "to provide high-quality proteins which are terribly scarce in the human diet in tropical and subtropical countries" (cited in Feder, 1980, p. 78). Initial prognostics indicated that, with an increase in beef production in the Central American region, the price of beef in the internal market would decrease, per caput consumption would rise and food deficiencies would decrease. Incomes would rise with the continuous development propelled by beef exports, while the law of supply and demand, functioning freely in a competitive market, would assure that a greater number of consumers would have access to beef which would be increasingly cheaper. If this were not the case, population growth would be to blame.

The beef supply had to increase at a rate slightly lower than 2.8 percent per year over the period 1960- 1980 just to maintain per caput consumption in Central America, but this rate was by far exceeded. There are other, more significant determinants of supply and demand in the region which were not contemplated in the prognostications: the internal price for beef in Central America is determined by the Chicago Yellow Sheet, given that internal supply depends on the amount of beef available after subtracting exports, and almost all beef was potentially exportable. Further, internal market prices are not normally determined by the free play of market forces. The global tendencies were: i) while beef production increased, exports increased at an even faster rate; ii) the wholesale price in the internal markets increased, not only in proportion to the increase in United States wholesale prices, but even more; iii) per caput consumption decreased, stagnated or increased only very slightly in the different countries of the region; and iv) the distribution of beef consumption worsened, as did nutritional deficiencies in the majority of the Central American population.

The demand curve for beef in Central America increased with the increase in incomes of the middle and upper classes, - and inflation was an additional element which contributed to price increases on the internal market (Simpson and Farris, 1982; Feder, 1980). Data from Nicaragua and Honduras show the relation between prices - in the export and internal markets: internal market supplies stagnate or increase when export prices decrease. Nevertheless, in these cases the internal retail price rises above the export market price instead of falling below it. The greatest increases in internal market prices in the history of the region have occurred precisely over the last decade, at the moment of the world beef market crisis, when internal supplies increased. Therefore, per caput consumption of beef has diminished in Central America since the beginning of the crisis in 1980, from 11.2 to 10.2 kg per year.

Total production of beef in Guatemala, Nicaragua, Honduras and Costa Rica grew at an average annual rate of 5.6 percent over the expansionary period, increasing from 102 600 tonnes in 1961 to 263 000 tonnes in 1981. The comparable rate of growth in the United States is 1.7 percent per year. The total change in production over the period is closely correlated with the change in exports. The size of the national herds was not a limiting factor for production after 1967, so total production in any given year depended on current export prices, expectations with respect to export prices in the following year and feeding conditions.

Since the beginning of the global beef market crisis in 1980, Central American exports declined by 138 000 tonnes to 61 000 tonnes in 1987. The value of exports decreased more than proportionately owing to the decrease in prices. By 1984, the Nicaraguan Ministry of Agriculture and Agrarian Reform (MIDINRA) had revised its estimates of net foreign exchange earnings from beef to reflect the increase in the costs of inputs and the decrease in export prices. The net amount of foreign exchange generated per US dollar invested decreased from about $ó.00 to between $1.50 and $2.10, depending on the technology employed in primary production. Honduran estimates show that, in 1984, the net generation of foreign exchange had become negative (SRN, 1984). Nevertheless, the United States continues to be the preferred market for Central American beef exporters, given that the price there is still above that of alternative markets (Ballard, 1985).

Per caput consumption of beef in Central America had also declined since the beginning of the crisis. Beef consumption is related directly to income levels and prices, and income levels had either stagnated or diminished in real terms since the mid1970s. The price elasticity of beef, estimated at -0.40 for Mexico and Central America has little meaning for those sectors of the population whose income levels virtually preclude access to this luxury consumer good. Income elasticity is very high at 0.70, but this indicator does not adequately reflect the fact that beef is consumed in significant quantities only by those groups that can easily fulfill other caloric needs.

In Honduras, a consumer can buy 6.4 kg of maize or 5 litres of milk for the price of less than 0.5 kg of beef. One day's salary in the manufacturing sector, where the minimum wage in 1984 was US$2.75 per day, permitted a worker to purchase about 0.7 kg of beef, or 8 litres of milk, or 12 kg of maize. With incomes at these levels, it is not surprising that a study by USAID in Nicaragua in 1976 concluded that "neither the urban nor the rural majority had sufficient incomes to obtain a nutritionally adequate, "minimum cost" diet, even if they used all of their incomes to buy food" (cited in Barraclough, 1982, p. 37).

Per caput consumption of beef varies considerably within the region, from a high of about 24 kg per year in Panama (less than 0.5 kg per person per week) to a low of 5.7 kg per year for El Salvador (about 0.5 kg per person per month). In Honduras, where beef exports rose more rapidly than in any other country in the region, per caput consumption was 11 kg per year lower than the minimum recommended and did not change over the entire expansionary period. Nevertheless, even these figures are deceiving. Data from Nicaragua and Honduras reveal the extreme inequalities in beef consumption by income group which are typical of Central American countries. In Nicaragua, 50 percent of the poorest population group consumed one-third of the beef that was consumed by the highest income group (0.5 kg per month versus 0.5 kg per week). In Guatemala, at the end of the 1960s a study showed that "17 percent of the rural population consumed 0 to 2.5 g of animal protein (per day) and the remaining majority only from 5 to 15 g" (Suchand, 1973, p. 7). The data presented by Cardona show that per caput consumption varies between 1.23 and 14.24 kg per year, depending on the department, where eight departments had a consumption level inferior to 2 kg per year.

It is widely recognized that protein deficiencies are the major nutritional problem in Latin America and the Caribbean, and the situation in Central America is particularly grave. Central American countries occupy the lowest positions in Latin America, both in terms of calories and proteins. It was estimated in 1971 that 57.4 percent of all children between one and four years of age in Costa Rica were malnourished (Suchand' 1973, p. 28). In Nicaragua, this figure was 56.6 percent with 15 percent suffering from severe malnutrition. In Honduras, 48 percent of all children under five years of age are malnourished, according to weight/age indicators, and there are 80 000 severe malnutrition cases each year.

USAID has argued that beef exports increase the supply of viscera in the diet of the poor. Data from Nicaragua show that the poorest 50 percent of the population in reality consume less viscera in comparison with beef. Most of the Central American population, which is malnourished and landless, not only saw their agricultural land converted into pasture for the production of beef for export, but also saw the prices of beef rise continuously so as to be completely out of reach. The increases in the supply of beef benefited only the small middle and upper classes, and even they paid inflated prices. Nevertheless, protein deficiencies in Central America increased, above all in the period of ranching expansion, as a direct and indirect result of the growth in beef exports. Almost all the increase in beef production in the region was destined for the United States' consumption. In fact, in order to promote exports, the World Bank required governments to increase the price of beef in their internal markets and reduce, if not eliminate, export taxes if they wanted to qualify for sectoral loans (Jarvis, 1986, p. 142).

CONCLUSIONS

Beef cattle production for export caused the Central American region's first major wave of rural population displacement, a decline in relative agricultural employment and massive deforestation. Ranching combined two particular characteristics: a high capital-labour ratio and low labour productivity, which are attributable mainly to slow production times. These translate into very low levels of employment on a per hectare basis. Where pasture expansion took land out of other productive activities, existing agricultural jobs were eliminated. Where pasture expansion occurred through the incorporation of idle land, employment was created, but at levels that were far below those required to employ the growing workforce. Overall, pasture expansion displaced agricultural labour. But even more important, it created barriers to the absorption of additional agricultural labour owing to the particular form of investment in the sector. As pasture and land concentration increased, the growing rural population increasingly lost access to land. Peasant farmers could not obtain access to land to produce basic food crops. Agricultural workers found that they were searching for work in labour markets which were increasingly competitive and then virtually absent. Peasant farmers searched for land in a context where pasture expansion continually increased the value of even the worst soils. Now they must be taught to manage the poor soils they have come to occupy.

In Honduras, export crop production was experiencing relatively high increases in productivity per worker, which simultaneously expanded production and decreased relative employment. This is the principal reason why coffee- and banana producing areas could undergo a significant process of economic growth and still lose employment. Changes in the capital-labour ratio in the most labour-intensive crops explain part of the reduced capacity of agriculture to absorb labour: in Honduras, absolute increases were evident in employment owing to the incorporation of new land, but employment decreased on a per hectare basis. Still, the absolute increase in area in permanent crop production was so slight overall that these agricultural branches had very little impact on the overall structural changes in the labour force, including its geographical distribution.

The Nicaraguan case upholds the assertion that processes occurring elsewhere in Latin American tropical humid forest dominated by cattle production are likewise occurring in Central America. Undoubtedly' the expansion of the agricultural frontier and deforestation are direct results of migration stemming from pasture expansion in the rest of the country, and the particular conditions evident in the tropical humid forest region lend themselves to a continuation of a similar process within its domain. At least some of the migration coming from cattle producing departments in Honduras and Nicaragua is attributable to environmental degradation while probably a significant amount of migration within the agricultural frontier region in Nicaragua is due to the process of degradation resulting from agricultural or livestock production described earlier. Furthermore, agricultural producers bring cattle with them and increased production in this sector means increased herds and area under pasture.

The valorization of land in Central America is largely a function of the expansion of ranching. Because of its particular nature, ranching permits the valorization of even the worst soils and therefore permits them to be brought into production, which raises the rent on remaining soils. The reasons for the lack of intensification in the sector and the implications of this process for the distribution of income in the overall economy are discussed in detail elsewhere, but suffice it to say that cattle rent tended to serve as an obstacle to the development of other more labour- and capital-intensive agricultural development in the Central American region.

The expansion of ranching in agriculture over this period also tended to create barriers to rapid industrial expansion. Industrial growth is highly dependent on imported inputs, state protection and inefficiently produced agricultural inputs. This particular form of development in the industrial sector was not conducive to large increases in employment. With large-scale expulsion of the rural population and a form of industrial growth which lacked forward and backward linkages within the internal economy, growth in employment could not be sustained over the long term. The unemployed and underemployed population expanded dramatically and exercised severe downward pressure on wages, especially after 1974. In spite of the general downturn in investment and in international prices for beef, pasture expansion in the 1970s continued at even higher levels than those experienced during the preceding decade. All of the processes traced here accelerated after 1974.

The expansion of ranching not only had very important consequences for the agrarian social structure elsewhere; it had even more significant consequences for the entire social structure. In Honduras and Nicaragua, and more generally over much of Central America, the global process of economic growth did not increase the absolute demand for labour at the same rate at which it decreased the relative demand for labour. The impoverishment of the rural population became the impoverishment of the urban population, which translates into below-subsistence wages for most workers. The pressure exerted by the unemployed and underemployed is also felt by small producers who must lower their own remuneration to prevailing social levels. All through the productive apparatus, living standards were depressed, as was internal demand and. consequently, internally based economic growth.

BIBLIOGRAPHY

Aguilar, M.R., Reyes, G.M. & Vigil, C. 1980. Empleo rural. Estado y políticas públicas en Honduras. Tegucigalpa, CSUCA-IDRC.

Alonso, E. & Slutsky, D. 1982. Empresas transnacionales y agricultura: al cave del enclave bananero en Honduras. Tegucigalpa Editorial Universitaria.

Ballard, P. 1985. La inserción de Centroamérica en el mercado mundial de la carne de res. Paper presented at the First National Scientific Agriculture and Livestock Congress. Managua.

27-30 June.

Barraclough, S. 1982. A preliminary analysis of the Nicaraguan food system. Geneva Managua, UNRISD.

Baumeister, E. & Havens, A.E. 1983. Recruitment and retention of occasional workers in the export sector of agriculture in Nicaragua. 1981-82. Managua, CIERA. (mimeo)

BCH. 1986. Estructura económica y financiera de la industria 1980- 1984. Tegucigalpa, Central Bank of Honduras, Department of Economic Studies.

Campanella et al. 1982. Honduras - perfit ambiental del país - un estudio de campo. USAID

Contract No. AID/50D/PDC-C0247, August. McLean, Virginia, JRB Associates.

CELADE. 1971. Evaluación de la cobertura geográfica del censo nacional de 1971 de la República de Nicaragua. Serie AS. No. 12A. San José. Costa Rica, CELADE.

CEPAL. 1984. Statistical Yearbook for Latin America. Santiago, UN.

CEPAL. 1985. Statistical Yearbook lor Latin America. Santiago. UN.

CIDCA. 1982. Demografa costeña: Rotas sobre la historia demografía y población actual de los grupos etílicos de la Costa Atlántica nicaragüerise. Managua, Centro de Investigación y Documentación de la Costa Atlántica.

DeWalt, B. 1982. The big macro connection: population. grain and cattle in southern Honduras. Culture Agric., 14(winter): 1-12.

DGEC. 1963. Censo nacional de población y vivienda, abril 1961. Tegucigalpa, Government of Honduras, Ministry of Economy, General Statistics and Census Directorate (DGEC).

DGEC. 1974. Censo nacional de población y vivienda. 1974. Tegucigalpa.

DGEC. 1976. Censo nacional de población. marzo 1974 (3 vols). Tegucigalpa.

DGEC. 1977. Investigación industrial 1975. Tegucigalpa.

DGEC. 1982. Censo general de la población 1950. Tegucigalpa. (reprint)

DGEC. 1988. Encuesta continua de fuerza de trabajo - Area rural, 1987. Tegucigalpa.

Espinales, B. 1974. Dinámica de la población y estructura agraria en Nicaragua. Estudios Sociales Centroamericanos, 9: 165-208.

Feder, E. 1980. La irracional competencia entre el hombre y el animal por los recursos agrícolas de los países subdesarrollados. Trimestre Económico, 41(1): 49-84.

Government of El Salvador. 1954. Primer censo agropecuario 1952. San Salvador. Lud Dreikorn.

Hecht, S. 1982. Cattle ranching in the eastern Amazon basin: evaluation of a development strategy. University of California, Berkeley, USA. (unpublished Ph. D thesis)

Howard Ballard, P. 1987. From banana republic to cattle republic: agrarian roots of the crisis in Honduras. University of Wisconsin, Madison' USA. Ann Arbor. University Microfilms. (dissertation)

Jarquin, C.E. 1975. Migraciones rurales y estructura agraria en Nicaragua. Estadios Sociales Centroamericanos. 11: 87- 164.

Jarvis, L. 1986. Livestock development in Latin America. Washington, DC, World Bank.

Kramer, F. 1986. The impact of external markets on the structure of peasant agriculture in southern Honduras. Paper presented at the XIII International Congress of the Latin American Studies Association.. Mass., USA, 23 October 1986.

Latinoconsult. 1975. Mercado de ganado y de la carne bovina en Nicaragua. Síntesis (Tomo 1). Managua.

Myers, N. 1980. Conversion of tropical moist forest. Washington. DC, National Academy of Sciences.

Place, S. 1981. Ecological and social consequences of export beef production in Guanacaste Province, Costa Rica. University of California, Los Angeles, USA. Ann Arbor University Microfilms. (Ph.D. thesis)

Romanini, C. n.d. Agricultura tropical en sierras ganaderas. Mexico City. Instituto Nacional Indigenista.

Ryan, J. et al. 1970. Area handbook for Nicaragua. Washington, DC, United States Government Printing Office.

Simpson, J. & Farris, D. 1982. The world's beef business. Ames. USA. Iowa State University

Press.

Slutsky, D. 1978. La agroindustria de la came en Honduras. Estudios Sociales Centroamericanos. 8(22): 101-206.

"

SRN. 1984. Diagnóstico de la ganadería en Honduras (3 vole). Tegucigalpa, Government of Honduras, Natural Resources Institute.

Suchand, G.J. 1973. Decisional influences in peasant animal production: a Central American comparative case study. University of Oklahoma. USA, Ann Arbor University Microfilms. (Ph.D. dissertation)

Taylor, J.R. 1967. Colonización y desarrollo agrario en la zona del Atlántico. Managua. Instituto Agrario de Nicaragua.

Ugarte, P. 1981. Estudio de la ganadería bovina de una región representativa de la zona Atlántica de Nicaragua. Managua, Nicaraguan Ministry of Agriculture and Agrarian Reform (MIDINRA).

UN. 1972. Metodos de mediaciôn de la migración interna. Manual Vl, ST/SOA/Seria A/47. Estudios de Poblaciôn Num. 47. New York.

Warnken, P. 1975. The agricultural development of Nicaragua. Columbia International Series 1 (July). Special Report 169, Agricultural Experiment Station. University of Missouri, USA.

Weeks, J. 1985. The economies of Central America. New York. Holmes and Meier.

Williams, R. 1986. Export agriculture and the crisis in Central America. Chapel Hill. USA. University of North Carolina Press.

World Bank. 1987. Informe de misión sobre la situación medioambiental en Honduras. Washington, DC. World Bank. (mimeo, restricted)