land use changes in llanos

8/11/2019 Land Use Changes in Llanos

1/20

ELSEVIER

Agri cuhral Sy ms, Vol. 54, No. 4, pp. 501-520, 1997

0 1997 Published by Elsevier Science Ltd

All rights reserved. Printed in Great Britain

PII: SO308-5 21X(96)0008 8-1 030 521X/97 S17.00+0.00

Lan d Speculation and Intensification at the Frontier: a

Seeming Paradox in the Colombian Savanna

Joyotee Smith,*

Jo& Vicente Cadavid, Alvaro Rinc6nb

& Ralil Veraa

3Zentro Intemacional de Agricultura Tropical (CIAT), Cali, Colombia

%O RPO ICA , Villavicencio, Me ta, Colombia

(Received 24 May 1996; accepted 2 October 1996)

ABSTRACT

Fron t i e r a reas in La t i n A m er i ca hav e been cha rac te r i zed by lan d specu la t i o n ,

abandonmen t and exp l o i t a t i o n . Th i s paper ana l y zes a r on t i er ca t t l e r an ch i ng

a rea in th e savann a w here in t ens t f i ca t i on has occu r r ed in sp i t e o f l and specu -

l a t i o n A w ho l e - f a rm l i v es t o ck p r oduc t i o n mode l i s used t o quan t i f y t he det e r -

m i n a n t s

of

i n t en s t j i ca t i o n . Resu l t s show t ha t l a nd specu l a t i o n has

s imu l t a neousl y i n c r ea sed t he p r o f i t a b i l i t y

of

ca t t l e r an ch i ng , w h i l e s l ow i ng

dow n i n t en s ca t i o n and im ped i ng t he adop t i o n

of

susta inab le prac t ices .

I n t ensi J i ca t i on occu r r ed because t he p rodu c t i on ga ins o f l e r ed by i m p rov ed

past u res o v er nat i v e grasses w ere l a rge enou gh t o ov ercom e t he negat iv e ef l ec t

of

l a nd specu l a t i o n on i n t en s ca t i o n . By con t r a s t , i n t he A mazon , im p r o ved

techno l og ies hav e to ov er come t he h i gh fer t i l i t y o f new ly defo r est ed la nd .

Thus , techno l ogy , la nd specu la t i o n and cha rac t er i s t i cs o f t he resou r ce base

i n t er a c t i n a comp l e x w ay t o det e rm i ne t he speed and na t u r e

of

i n t ens@ca t i on .

0 1997 Pu b l i shed by Elsev i er

Science

L t d

INTRODUCTION

The sustainable management of frontier areas is of increasing importance in

a world characterized by growing populations and increasing pressure on the

natural resource base. Latin America, in particular, has turned increasingly

to its frontier areas to provide food, export earnings and employment.

*To whom correspondence should be addressed to the following: PO Box 6596 JKPW B,

Jakarta 10065 , Indonesia.

501


2/20

502

J. Smith et al.

Examples abound, however, of Latin American frontier occupation being

characterized by massive resource degradation, with minimal compensating

contributions to agricultural production (World Bank, 1993; Aramburu,

1984; Hecht

et al.,

1988). This paper analyzes a frontier cattle ranching

region in the Colombian savanna, where substantial increases in agricultural

productivity have occurred. While intensification has undoubtedly been

accompanied by resource degradation, the compensating increases in pro-

duction and carrying capacity imply that this is a frontier that offers the

possibility of sustainable increases in production. The paper analyzes the

determinants of intensification in the savanna, and draws implications about

the lack of intensification in the Amazon.

The process of intensification in the Latin American savanna is particu-

larly interesting because it is one of the last significant agricultural frontiers

in the world and, therefore, could potentially make a significant contribution

to world food supplies (Borlaug & Dowsdell, 1994; Macedo, 1994). Intensi-

fication in the savanna may also alleviate pressure on the Amazon rain forest

(CIAT, 1991a). On the other hand, it could lead to savanna biodiversity

losses and to siltation and pollution of watercourses and neighboring wet-

lands (Solbrig, 1993; Macedo, 1994; Smith

et al.

in press; Klink

et al.

1993;

Alho

et al.

1988).

LAND USE CHANGE IN THE COLOMBIAN SAVANNA

The study area covers the municipalities of Puerto Lopez and Puerto Gaitin,

an area of 2364 km* in the department of Meta, in the Colombian savanna

(Fig. 1). The region is sparsely populated (population density is 1.25 persons/

km*, 1985 figures; DANE, 1989), has highly acidic soils (pH 4.2-5), with

high aluminium saturation (> 80%) and low nutrient availability. Annual

rainfall is 1700-2500 mm, with a dry season of 3-4 months. The vegetation

consists primarily of native savanna grasses, with gallery forests along the

banks of watercourses. In 1995 the authors surveyed 98 randomly selected

farmers, preceded by surveys in 1979, 1989 and 1992. Bio-physical case

studies were carried out in 1977-1979 and in 1988-1991, and on-station and

on-farm trials have been on-going since the 1970s.

In the 197Os, the region was still a frontier cattle ranching area, i.e. the

natural ecosystem was just beginning to be brought under human interven-

tion. Cattle stocking rates (0.09 animals/ha) were too low to alter the natural

vegetation. The majority of holdings lacked official land titles and the native

savanna grasses were open to communal grazing (Tamayo, 1975). Rivers

were the main form of transport, the dirt tracks running through the area

being virtually impassable during the wet season (Sere & Estrada, 1985).


3/20


4/20

504

J.

m i th et al.

TABLE

Changes in Land U se 1979-1 995, Municipalities of Puerto L opez and Puerto Gait&r,

Colombian Savanna: Survey Data

1979a 1995b

% farm in improved pasture

2 20

Animals/ha

0.09 0.28

Fertilizer users (% farmers)

3

53

Average farm size (ha)

5162 1857

Estimated beef production (kg/ha)

16 50

Estimated carrying capacity (000 heads)d

959

1726

YWvey data (111 farmers): Sere & Estrada (1985).

bUnpublished survey data (98 farmers).

cBase d on degree of pasture adoption and production data from on-farm trials.

dPotential herd size in well-drained savanna (4.8 m ha) based on degree of pasture adoption

and stocking rate (Seri & Estrada, 1985).

the dominant enterprise, with only 15% of farmers having small plots of

annual or perennial crops.

Although improved pastures occupy only a small part of the area, they

appear to have resulted in significant increases in land productivity and car-

rying capacity. Production data from on-farm sources show that beef pro-

duction is around 12 kg/ha on native savanna (Kleinheisterkamp & Habich,

1985), and this increases more than &fold to around 200 kg/ha on

improved grass pastures under farmers management (CIAT, 1994).

Weighting this data by the reported extent of improved pasture implies that

beef production/ha has more than tripled between 1979 and 1995. Data from

the intensive case studies carried out in the late 1970s (Sere & Estrada, 1985)

indicate also that with 20% of the area converted to improved pasture, and

the current stocking rate on improved pasture of about 1 AU/ha (AU is

Animal Units) (Hoyos et al., 1992), the potential herd size has increased by

80% (Table 1). Consistent with this, the cattle inventory of sample farmers

increased at an annual rate of 2.9% between 1979 and 1989 (Cadavid et al.

1991), which is considerably higher than the increase in the national inven-

tory of 0.93% over the same period (Lorente, 1994).

EXPLAINING LAND USE CHANGE: HYPOTHESES

Data from the Colombian savanna indicate that intensification appears to be

closely related to land values. The smallest farms, with the highest propor-

tion of area in improved pasture and with the highest stocking rates are

located whe re land values are highest (Table 2). This is an area which has


5/20

Lund speculat i on and int enszj i cat i on at t he ront i er

505

TABLE 2

Characteristics of Farms in Puerto Lo pez and Puerto Gaitan (Colombian Savanna), by

Infrastructure and Topograp hy: Survey Data, 1995

Al l arms I nfrast ructur e

M oderat e Poor

Plains Undulating Plains Undulating

Number of sample farms 98 I1 26 29 32

Average land price ( /ha) 289 543 385 293 137

% farm in improved p asture* 20

50 35b

19=

loc

Animals/ha* 0.28

0.6=

0.3b 0.2b

Farm size (ha)* 1857 418

1998=~~

2989

*Differences in letters indicate significant differences (5% level), Duncans Multiple Range Test.

electricity supplies, is connected to the capital of Meta (Villavicencio) with a

paved road and has a flat topography which permits mechanized sowing of

improved pasture. Plains with poor infrastructure and lower land values,

have a significantly lower degree of land-use change. Intensification is even

lower in remote areas with an undulating topography.

It is widely accepted that improved infrastructure increases land values

and induces intensification (Binswanger & McIntire, 1987; Smith

et al.,

1994;

Katzman, 1974). In the study area about 45% of the road between Villavi-

cencio and Carimagua is now paved and the road between Villavicencio and

the nations capital, Santafe de Bogota, is being significantly improved.

Further improvements are expected as a result of major oil discoveries near

the region, which have more than doubled Colombias oil reserves. Thus, the

role of infrastructure in the Colombian savanna has been very different from

its role in the Amazon, where the frontier was open, i.e. penetration roads

made new land available for human intervention, thus exerting a downward

pressure on land values (World Bank, 1993; Kaimovitz, 1996). In contrast, in

the Colombian savanna improvements in infrastructure increased the viabi-

lity of economic activity on already accessible land.

Recently a number of authors have hypothesized that land speculation is

an impediment to intensification, particularly in frontier areas (World Bank,

1993; Brandao & de Rezende, 1992; Kaimowitz, 1996; Smith et al. in press).

Where credit and risk markets are incomplete, or macroeconomic instability

exists, land values exceed the present value of capitalized agricultural profits.

In frontier areas when infrastructure improves and land prices increase

rapidly, this speculative component gains importance relative to productive

use. Available capital is, therefore, likely to be used more for increasing land

holdings than for intensification (Lowenberg-De Boer & Boehije, 1986).

In the study area the extremely high degree of land price appreciation makes

it likely that the role of land as a speculative financial asset affects the


6/20

506

J. Smith et al

intensification decision. Macroeconomic developments have favored the

attractiveness of land as a financial asset. During a severe structural adjust-

ment program in the latter part of the 1980s few investment opportunities,

other than land, could keep up with inflation which increased from 20% in

1980 to 32% in 1990 (IMF, 1995). In the 1990s demand for land was

increased by high levels of capital inflows (averaging 1.7 billion during

1991-1993: E.I.U., 1995), due to foreign investment in the oil sector and to

repatriation of profits from the illegal drug trade. Although the recent

political crisis and crackdown on the drug trade has led to currency

depreciation and an uncertain outlook for land prices, it is likely, in the

medium term, that land will continue to be regarded as a trusted financial

asset. This is borne out by the reported objectives of ranchers. The majority

of farmers (52%) claim that they hold land both because of its productive

capacity and because of expected increases in land values. Twenty eight

per cent of farmers are interested primarily in the lands productive value,

while the rest (20%) claim they are holding land primarily as a financial

asset.

Technological advances also interact with incentives to intensify (Smith et

al. 1994). If a new technology results in a major increase in land productiv-

ity, farmers may obtain a higher total income by changing to the new tech-

nology on their current land holding, rather than extensifying, i.e.

augmenting their land holding with the old technology. The lower the land

price the larger will be the extra land that can be acquired if the new tech-

nology is not adopted. Thus, only a very major technological advance indu-

ces intensification when land prices are low. Conversely, the greater the

technological advance, the lower the land price at which intensification

occurs. In the savanna intensification occurred in a land-abundant area

because improved grass pastures are able to offer spectacular increases in

land productivity over the traditional technology. This was made possible by

the inherent nature of the native forage, characterized by low levels of dry

matter yield (which enforce low stocking rates), and poor digestibility which

limits liveweight gains per animal (Fisher et al. 1992). In contrast, in the

Amazon if technologies are to be adopted, they have to significantly surpass

the high returns obtainable with traditional practices in the first few years

after deforestation (World Bank, 1993). Thus, the poor production potential

of native vegetation in the savanna made it easier for technology to contri-

bute to intensification.

Summing up, therefore, the descriptive analysis leads to the hypotheses

that the most significant factors driving the process of intensification have

been technology, land price appreciation, and the nature of the natural

resource base. These factors are incorporated in the quantitative analysis

given below.


7/20

Land speculation and intens cation at the frontier

507

QUANTIFYING THE DETERMINANTS OF INTENSIFICATION

Simulation analysis is used to quantify the determinants of intensification.

An intensification strategy is specified as one in which a given piece of land in

native savanna is purchased at the beginning of the planning period, pro-

gressively converted to improved pasture and sold at the end of the planning

period. The returns to an intensification strategy are the discounted returns

to beef production under increasing amounts of improved pasture, plus

discounted capital gains on the land. Returns to beef production are net of

the extra expenditure required for improved pasture. These include the cost

of establishing, fencing and periodically renovating improved pasture,

constructing infrastructure such as corrals, drinking troughs and wind mills

for pumping water, purchasing additional cattle and inputs such as fertilizer,

animal health, administrative and technical services. Capital gains on land

include the land price premium commanded by the proportion of land in

improved pasture at the end of the planning period. Under an extensification

strategy, the farm consists entirely of native savanna (or a negligible pro-

portion of improved pasture) throughout the planning period, and the

expenditure foregone by non-adoption of improved pasture is used for

acquiring additional land in native savanna, which is also managed with

traditional practices. The returns to an extensification strategy are the dis-

counted returns to traditional beef production on the entire farm (including

the extra land acquired by foregoing intensification), plus the discounted

capital gains from selling the entire farm at the end of the planning period.

The above formulation shows that the more productive the new technol-

ogy relative to traditional practices, the greater will be the degree of intensi-

fication, with land price, land price appreciation and the discount rate held

constant. High land prices and low discount rates have conflicting effects on

the intensification decision. On the one hand, they encourage intensification

by lowering the amount of land that can be acquired by foregoing intensifi-

cation, and by reducing the cost of investing in new technology. On the other

hand, they encourage extensification by increasing the value of capital gains

and lowering the cost of acquiring new land. The higher the rate of land price

appreciation, the greater are the above incentives for extensification. Thus, a

high rate of land price appreciation dilutes the incentives for intensification.

The net effect on the intensification decision is an empirical question, which

is quantified in the simulation analysis.

A whole-farm livestock production model is used to simulate annual

changes in beef production over the planning period, resulting from cumu-

lative changes over time in the establishment of improved pasture. The

structure of the livestock production model is based on the Gittinger (1982)

herd projection model, which has been adapted to Colombian conditions by


8/20

508 J. Smith

et al.

the inclusion of modules of representative production systems in the savanna

(breeding and breeding plus fattening), with stable herd size and composition

calibrated to technical coefficients drawn from on-farm trials (Botero, 1985,

1989~; Ramirez & Botero, 1989; Rivas, 1993). For the analysis presented in

this paper, the model was adapted to farmers objectives by the addition of a

land purchase and sale component, and was recalibrated to the latest tech-

nical coefficients. Starting with a typical breeding herd grazed on native

savanna in the 1970s the model simulates changes in herd size and composi-

tion based on typical reproduction, mortality and culling rates for each age

and sex category, derived from the intensive farm monitoring study carried

out in the 1970s (Kleinheisterkamp & Habich, 1985). As the area in

improved pasture increases, certain key technical coefficients are changed in

accordance with trial results, synthesized in Botero (1989b). With 50% in

improved pasture, for example, the calving rate increases from 46 to 70%

(Kleinheisterkamp & Habich, 1985), and the calf mortality rate falls from 12

to 7% (Botero, 1989b). The breeding enterprise also progressively includes a

fattening component starting with 8% of the herd, when 20% of the area is

planted to improved pasture. Typical stocking rates on improved pasture

and native savanna are achieved via cattle purchases and sales as required.

Data on year-by-year Live Weight Grain (LWG) per animal on native

savanna and improved pasture is obtained from a long-term trial carried out

over 16 years and summarized by Lascano & Euclides (1994).

We begin by quantifying the nature of the relationship between intensifi-

cation and its determinants. Taking the case of a rancher who purchases

500 ha in native savanna in time period 0, we simulate the returns to inten-

sification and extensification over a time horizon of 15 years. The discount

rate is taken as 5% (equivalent to the real interest rate in the early 198Os),

and improved pasture is assumed to sell for 2.5 times the price of native

savanna. Land price is varied from lO/ha to 200/ha, and the rate of land

price appreciation from 0% p.a. to 15% p.a. in real terms.

The case of extensification is represented by a farm containing only native

savanna. Given that adoption of improved pasture occurred very slowly in

the early 1980s the intensification strategy is represented by a farm that

converts 10% of the area to improved pasture by the end of the planning

period. Figure 2 shows how the returns to intensification (INT) and extensi-

fication (EXT) change with land price and the rate of land price apprecia-

tion. The numerical suffixes after INT and EXT indicate the assumed rate of

land price appreciation. Results show that net present values (NPV) of stra-

tegies based only on native savanna are negative for all land values unless

land appreciates in real terms by at least 10% p.a. This is consistent with

results from the early 1980s (SerC & Estrada, 1985). INT15 and EXT15,

representing land appreciation rates of 15% p.a. (which is the situation


9/20

Land speculation and intensification at the frontier

509

prevailing during the last 15 years) intersect at a threshold land price (T15)

just below lOO/ha. This indicates that intensification is unlikely unless land

prices are at least lOO/ha. Figure 2 also shows that with land appreciation

held constant at 15%, the incentive for intensification (i.e. the positive

difference between INTl5 and EXT15), increases with the price of land. This

is consistent with the positive relation between pasture adoption and land

price shown in Table 2. In 1995 average land prices were above the threshold

level for intensification in all parts of the study area. Therefore, it is not

surprising that 93% of farmers had at least a part of their farms in improved

pasture, and that the proportion of improved pasture averaged 10% or

greater in all parts of. the study area (Table 2). Figure 2 also shows that if

improved pastures had been about 30% less productive, causing INT15 to

shift downwards to INT15, the threshold land price for intensification

would have risen to T15 (just under 200/ha). Thus, the dramatically higher

500

400

300

200

100

0

- 100

W (S 000)

INT15

EXTI 0

111 EXT5

111 EXTO

Fig. 2. Effect

50

100 150

Land Prices ( /ha)

200

Farm size: 500 ha, Time horizon: 15 years,

INT=lO in improved pasture

EXT=O in improved pasture,

Discount rate= 5 ,

Live weight gain data from: Lascano and Euclides (lQQ4),

Suffixes for INT and EXT refer to land price appreciation.

of techno logy on intensification: thresh old land prices; simulation results,

Colombian savanna.


10/20

510 J. Smith et al.

productivity of improved pastures over native savanna appears to have

induced earlier intensification.

Figure 3 analyzes the impact of land price appreciation on intensification.

Given the negative returns to native savanna, the strategy of extensification

is represented in the rest of the paper as a slow rate of adoption of improved

pasture, resulting in 10% of the farm being converted to planted pastures by

the end of the planning period. Intensification is represented as a more rapid

adoption, resulting in 50% of the farm being converted by the end of the

planning period. As before, the foregone expenditure required for intensifi-

cation is used to acquire additional land in the extensification strategy. Fig-

ure 3 shows that when land prices appreciate at 15%, the threshold

intensification price (T15) is > 150/ha. If the expectation of land apprecia-

tion drops to 5%, intensification occurs at T5, i.e. at land prices > 50. Thus,

a high rate of land price appreciation slows down the process of intensifica-

tion. This occurs because when land prices appreciate rapidly, capital gains

dominate income gains from intensification. Therefore, farmers with rela-

tively cheap land are better off purchasing more land in order to capture

capital gains, rather than investing in intensification. At higher land prices

less land can be acquired by foregoing intensification and, therefore, the

, o o o

NW SOW

I

800

600

NT15

EXTlS

INT10

EXTlO

INTS

INTO

EXT5

-200

5

200

l and Pri ces / ha)

Farm size: 500 ha

Time horizon: 15 years

INT=CQ in improved pash~re

EXT=lO in improved pas-hare

Discount rate= 5

Live weigM gain data from Lascano and Euclidea (1994)

Suffixes for INT and EXT refer to land price appreciation

Fig.

3. Effect of land price apprec iation on intensification: thresh old land prices; simulation

results. Colombian savanna.


11/20

Land speculat i on and int ensiJi cat i on at t he ronti er

511

opportunity for capturing capital gains on additional land declines. The

higher the rate of land price appreciation, the slower the rate at which this

phenomenon occurs. This quantifies the negative effect of land speculation

on intensification, and shows that the high rate of land price appreciation in

the study area explains why intensification has progressed relatively slowly.

Intensification can be expected to progress slowly as long as expectations

about land appreciation remain high.

Figure 4 holds land appreciation constant at 15% and varies the discount

rate. At a higher discount rate of lo%, the intensification threshold (T* 15) is

around 75/ha, which is lower than at a discount rate of 5% (T15). Both the

INT and EXT curves shift downwards at higher discount rates, but the

downward shift in EXT is greater than that for INT, particularly at low land

prices. Thus, at a high rate of land price appreciation, a low discount rate

slows down intensification, because the lower cost of acquiring new land

dominates the lower cost of investing in new technology.

Figure 4 indicates how sensitive NPVs are to discount rates. In fact, Fig. 5

shows that with interest rates at 15% in real terms, even 50% in improved

pasture is barely viable, if land appreciation drops below 10% (INT**5 and

INT**O). Thus, in a climate of high interest rates, speculative land price

1

800

600

o-

l u

Ipv (5 Cal

INT15

EXTIS

IN715

EXT.15

15

200

Land Prices ( /ha)

Farm size: 500 ha

Time horizon: 15 years

INT=50 in improved pasture

EXT=lO in improved pasture

Disw~nt rate= 5 for IM15 and EXT15

Disuxmni rate = 10 for IW15 and EXP15

Land price appmcktion = 15 p.a.

Live wigM gain data from Lascano and Eudides (1994)

Fig. 4. Effect of discount rate on intensification: threshold land prices; simulation results,

Colombian savanna.


12/20

512

J. Smith et al.

INT-15

50 100 150 200

Land Prices ( /ha)

Farm size: 5M) ha

Time horiron: 15 years

INT=50 in improved pasture.

Live weight gain data from Lascano and Euclides (lgg4)

Suffixes for INT and EXl refer to land price appreciation

Fig. 5. Effec t of land price appreciation on viability of cattle ranching (discount rate = 15% ):

simulation resu lts, Colom bian savanna.

increases have made a key contribution to the viability of cattle ranching.

Moreover, as shown earlier, increases in land prices have attracted new,

presumably more capitalized owners, for whom intensification may have

been easier. Thus, land speculation has both aided and hindered intensifica-

tion. Currently in Colombia interest rates are around the 20% level in real

terms. Moreover, political instability has clouded land price prospects and

increased insecurity in cattle ranching areas. These developments seriously

threaten the viability of cattle ranching in the savanna, at least while the

current situation continues.

RESOURCE DEGRADATION, INTENSIFICATION AND LAND

SPECULATION

Intensification in the savanna appears to have had both positive and negative

effects on the environment. On the positive side, Fisher et al. (1994) show

that improved pastures sequester substantially more carbon than native

savanna grasses. On the negative side, farmers and scientists report consid-

erable resource degradation, some aspects of which appear to be systemati-

cally related to intensification. In the sample as a whole a little over 60% of


13/20

Land speculat i on and int ens@cat i on at t he ront i er 513

farmers report degradation of native savanna, which results in changes in

plant communities and threatens biodiversity (Rippstein & Girard, 1994).

Table 3 compares reported degradation between areas of moderate and poor

infrastructure in the plains. Reported degradation of native savanna appears

to be significantly lower in areas of moderately good infrastructure where

50% of the area is in improved pasture compared with areas where improved

pastures occupy only 19%. Native savanna degrades because the common

practice for coaxing higher productivity out of native grasses is frequent

burning and continuous grazing of young regrowth after burning. The lower

incidence of native savanna degradation in intensified areas implies that

when farms increase the area in improved pasture, the pressure to produce

more from native savanna declines. To the extent that land speculation slows

down intensification, it prolongs the pressure on native savanna. Also,

improved pasture is not intentionally burned. Thus, intensification reduces

the greenhouse gases emitted by burning native savanna. On the other hand,

intensification may ultimately lead to only small pockets of native savanna

remaining. Thus, it is important to study the spatial variability in plant

communities in native savanna in order to identify suitable protected areas

as intensification proceeds.

Other widely reported ecological changes include degradation of gallery

forests reported by 60% of farmers, and declines in bird, animal and fish

species each reported by around half the farmers. Gallery forests appear to

be under greater threat in the more intensified areas (Table 3). Ninety per

cent of farms that report degradation of gallery forests relate this to the

extraction of species for fencing paddocks of improved pasture. Twenty-

seven per cent of farmers had also cut down gallery forests in order to plant

improved pastures. Thus the adoption of improved pastures appears to

threaten gallery forests. Thirty five per cent of farmers report that logs are

extracted for sale from gallery forests. This is likely to be more prevalent in

TABLE 3

Repo rted Resou rce Degrad ation in the Plains of the Colombian Savanna: Municipalities of

Puerto Lo pez and Puerto Gaitin: Survey Data, 1995

Infrastructure

M oderat e

Poor

% area in improved pasture

Reported degradation

Native savanna

Gallery forest

Improved pasture

Animals (species/supply)

Birds (species/supply)

Fish (species/supply)

50

37

89

67

89

78

63

% farmers

19

83

52

52

66

59

57


14/20

514 J. mith et al.

intensified areas because infrastructure is better and, therefore, the cost of

getting logs to market is lower. Table 3 shows that while fish species and

supplies have declined uniformly in all areas, birds and animals are under

greater threat in areas of better infrastructure. Farmers relate this to uncon-

trolled hunting and fishing and degradation of gallery forests, both of which

appear to be more prevalent where infrastructure is more developed.

A little over 60% of farmers report degradation of improved pastures,

mainly loss of ground cover, which could result in wind and water induced

soil loss. Pasture degradation also results in significant declines in animal

production. Data from trials on two farms under farmers management show

that beef production/ha declined by one-third within 3 years on one farm,

and declined to a negative value in the other farm (CIAT, 1994). An inten-

sive monitoring of pasture management practices indicates that overgrazing

may partially account for the degree of pasture degradation. Improved pas-

tures are grazed for about 67% of the year, with grazing pressures particu-

larly high during the dry season, when instantaneous stocking rates are

about 0.8-1.6 AU/ha (Hoyos

et al.

1992). Unlike in the Amazon rainforest

however, degraded pastures are not abandoned. Forty-four to sixty per cent

of paddocks are renovated at least every 4 years. Sixty-six per cent of farmers

renovate by harrowing, with or without the application of small quantities of

P fertilizer, while 5% use fertilizer alone. Seventy-eight per cent of paddocks

were reported to have been attacked by spittle bug, which is commonly

controlled through overgrazing followed by mechanical renovation (CIAT,

1991b). Thus, overgrazing, nutrient deficiencies (particularly N) and pest

problems appear to contribute to pasture degradation. Rotation with annual

crops has been proposed as a means of renovating degraded pastures, the

cost of renovation and fertilizer apphcation being paid out of crop sale pro-

ceeds (Vera

et al.

1992). However, only 2% of the pasture area is renovated

with this technology.

Technologies for preventing pasture degradation include mixed grass-

legume pastures which maintain soil N levels (Thomas, 1992), increase soil

carbon in sand fractions (Guggenberger

et al.

1995) and earthworm popu-

lations (Decaens

et al.

1994). Adoption, however, remains disappointing.

Only 2% of the pasture area contains legumes and 88% of farmers who

plant legumes report that legumes fail to persist under farmers grazing

pressures.

The relative profitability of pure grass, and grass-legume pastures is com-

pared in Table 4. As before we simulate the case of a rancher who purchases

500 ha in native savanna, progressively converts a part of it to improved

pasture, and disposes of the farm at the end of the planning period of 15

years. Land price appreciation is assumed to be 15% p.a. in real terms, and

improved pasture (whether pure grass or grass-legume) sells for 2.5 times the


15/20

Land speculat i on and int ensi Ji cati on at t he ront i er

515

price of native savanna. The first two rows of Table 4 present results based

on data from on-farm trials under farmers management. In the first trial

LWG was lower on the grass-legume pasture than in the pure grass pasture.

In the second LWG was 30% higher on grass-legumes. In both trials

legumes effectively disappeared after 3 years and had to be resown, while the

grass persisted. Results show that the internal rate of return (IRR) for grass-

legumes was at best only marginally higher than for pure grass. Thus, there is

clearly a need to improve the performance of currently available legumes

under farmers management. In the third row of Table 4 a controlled trial

with best practices is analyzed, in which the legume persisted for 16 years.

Here the LWG on grass-legumes was 34% higher than on grasses, and no

expenditure on replanting legumes was required. Even in this trial, however,

IRR was only 8% higher than for grass pastures. The fourth row simulates a

further 30% increase in LWG for grass-legume pastures compared with the

data in the third row, giving a total increase of 74% over pure grass pastures.

Under these conditions a considerable increase of 20% in the IRR is

achieved. In the last row LWG remains at the level of the third row, but the

rate of land price appreciation declines to 5%. The results show that in this

situation the inclusion of legumes can increase IRR by 23%. Thus when land

appreciates rapidly, the productivity gains of improved technology are over-

whelmed by the capital gains from land sales, which dominate the IRR. Thus

technologies have to offer spectacular increases in productivity before they

are adopted. Recent data indicate that a newly released legume,

Arachis

pintoi offers significant gains in productivity, while persisting under farmers

management (Lascano, 1994). However, adequate data are not yet available

to quantify whether it is sufficiently productive to make a significant impact

on profitability, under high rates of land price appreciation. Thus land

TABLE 4

Internal Rate of Return: Pure Grass Versus Grass-Legum e Pastures: Simulation Results,

Colombian Savannaa

Grass Grass-legume

On-farm trial (undulating topograp hy, fragile soil)b*c

On-farm trial (sandy soil)b,c

Well-managed long-term trial (well-drained plain)bgd

Well-managed long-term trial (well-drained plain)

(+ 30% production for grass-legume)b,d

Well-managed long-term trial (land price appreciation 5% )4

19.5 18.3

17.6 18.9

19.4 21.0

19.4 23.3

11.8

14.5

aFarm size 500 ha, time horizon 15 years, 50 % in improved pasture.

bLand price appreciation 15% .

Live weight gain data from CIAT (1994).

dLive weight gain data from Lascano & Euclides (1 994).


16/20

516

J. Smith et

al.

speculation may impede the adoption of technologies which have the poten-

tial to prevent resource degradation.

TOWARDS A DEVELOPMENT PATH FOR FRONTIER AREAS

The above results indicate that technology, land speculation and biophysical

conditions have all been key determinants of the process of intensification in

the savanna. These results are used to take a first step towards formulating a

tentative development path for frontier cattle ranching areas in the savanna

(Thomas et

al.

1996). In emerging frontier areas when land price is stagnant

at very low levels, extensive production systems based on native savanna

grasses are likely to prevail. With very low stocking rates and minimal

infrastructure resource degradation is unlikely. This is similar to the situa-

tion prevailing in the Colombian savanna in the 197Os, and is depicted as

phase I in Fig. 6. As infrastructure improves land prices may increase very

rapidly if conditions in the larger economy encourage speculative land

T

Phase

I

Extensive Ranching

Minim al resource

degradation

Phase II

Slow increase

in

intensification:

gradual adoption of improved

grass pastures.

Adoption of legumes unl ikely.

Degradation of improved

pasture, gallery forest.

T

e

Phase Ill

Rapid

ncrease in

intensifcatlon.

Legumes? Crops?

Ley-farming?

Native savanna,gallery ore.4

threatened w ith extinction.

Fig. 6.

Stages of intensification in frontier cattle ranchin g areas in the savan na.


17/20

Land speculat i on and int ensifi cati on at t he ront i er

517

purchases. In this situation capital gains on land investment dominate prof-

itability and overwhelm technological advances which offer moderate pro-

duction increases. Thus, only technologies which lead to spectacular

increases in productivity are likely to be able to induce intensification. Thus

the speculative motive slows down the pace of intensification and technology

adoption. On the other hand, it may support the productive use of land by

maintaining the viability of cattle ranching under unfavorable economic

conditions. This scenario is similar to the current situation in the Colombian

savanna (phase II in Fig. 6) in which the dramatic production increases

obtained with grass pastures has stimulated their adoption, and consequent

intensification. Sustainable technologies, such as grass-legume pastures have

not, however, been adopted, resulting in widespread pasture degradation. In

this situation farmers demand highly productive grass pastures capable of

withstanding severe grazing pressures, or technologies which recuperate

degraded grass pastures. Our results also imply that once the frontier

matures and land prices plateau at high levels, as in established cattle

ranching areas in the Brazilian savanna today (phase III), farmers may be

more open to technological advances even if the production gains are not

dramatic. Sustainable technologies such as ley farming systems and grass-

legume pastures may be better targeted at this phase. However, native

savanna and gallery forests may be threatened with extinction in this phase.

Carefully designated protected areas may be required, or market-based

institutional mechanisms which internalize the environmental services of

these resources (Smith et al . in press).

Frontier areas in Latin America have been rife with land speculation,

abandonment and exploitation. Intensification has been an elusive com-

modity. Our analysis indicates that intensification can occur even in frontier

areas. Intensification occurred in the savanna because improved pastures

offered spectacular production increases over native grasses, the gains being

large enough to overcome the negative effect of land speculation on intensi-

fication. The implication is that biophysical conditions may be one factor

contributing to the lack of intensification in the Amazon. Improved tech-

nologies in the Amazon have to out-perform the high native fertility of newly

deforested land. Also, in the Amazon the supply of new land has been pro-

gressively augmented by the construction of penetration roads, while in the

Colombian savanna infrastructure has increased the economic viability of

already accessible land. Thus, technology, land speculation and biophysical

characteristics of the resource base interact in a complex way to determine

the nature and speed of the intensification process. These crucial elements

should be taken into consideration in the development of sustainable land

use strategies and in the design of medium- and long-term research strate-

gies, particularly in frontier areas.


18/20

5 8

J. Sm i t h

et al.

ACKNOWLEDGEMENTS

The authors are grateful to Jos6 I. Sanz, Peter Kerridge, Carlos Lascano,

Luis R. Sanint, Richard J. Thomas and Idupulapati M. Rao for helpful

comments.

REFERENCES

Alho, C. J. R., Lather, T. E., Jr and G oncalves, H. C. (1988). Environmental

degradation in the Pantan al ecosystem. B ioscience 38 , 164-171 .

Aram buru, C. E. (1984). Expansion of the agrarian and demo graphic frontier in the

Peruvian Selva. In

F r on t i e r Expans i o n i n A mazon i a ,

eds M. Schmink & C. H.

Wood. University of Florida Press, Gainesville, pp . 153-179.

Binswanger, H. P. and McIntire, J. (1987). Behavioral and material determinants of

production relations in land abund ant tropical agriculture.

Econom i c Deve l op -

m en t a nd Cu l t u r a l Change 34, 73-99 .

Borlaug, N. E. and Dowsdell, C. R. (1994). Feeding a human population that

increasingly crowds a fragile planet. Keynote Lecture. Supplement to Transac-

tions, 15th World C ongress of Soil Science, Acapulco, Mexico.

Botero, B. R. (1985). Sistema de production para &a extensiva de ganado de came

en 10s Llanos Orientales de Colom bia. MSc thesis, University of Costa Rica,

San Jose, Costa Rica.

Botero, B. R. (1989a). Proyeccion de hatos en ganado bovino, CIAT , Cali, Colom bia.

Botero, B. R. (1989b). Manejo de explotaciones ganad eras en las saban as bien dre-

nadas de 10s Llanos Orientales de Colombia, CIAT, Cali, Colombia.

Brandao, A. S. P. & de Rezende, G. C. (1992). Credit subsidies, inflation and the

land market in Brazil: a theoretical and empirical analysis, World Bank/

AGRAP, Washington, DC.

Cadavid, J. V., SerC, C., Botero, B. R. and Rivas, L. (1991). Adoption de pastos en

10s mun icipios de Puerto Lopez y Puerto Gaitan en la altillanura oriental de

Colombia. CIAT, C ali, Colombia.

Cad avid, J. V. (1995). Com portamiento y limitantes de la adoption de pastos y de

cultivos asociados en 10s Llanos Orientales de Colom bia. MSc thesis, Universi-

dad de1 Valle, Cali, Co lombia.

CIAT (1991a). (Centro International de Agricultura Tropical),

C IA T i n t h e I 9 9O s

and Beyond : A St r a t eg i c P la n ,

CIAT, Cali, Colombia.

CIAT (199lb).(Centro International de Agricultura T ropical), Annua l Report,

Tropical Pastures Program 1987-1 991. Vol. 2. CIAT, Cali, Colom bia.

CIAT (1994). (Centro International de Agricultura Tropical), Savan nas Program,

Biennial Report 1992-1993, Working Document No. 134, CIAT, Cali, Colom-

bia.

DA NE , (1989). Colombia E stadistica, M unicipal, Bo gota, II.

Decaen s, T., Lavelle, P., JimCn ez, J. J., Escobar, G. and Rippstein, G. (1994).

Impac t of land man agem ent on soil macrofauna in the Oriental Llanos of

Colombia.

Eur . J. So i l B i o l . 30 ,

157-168.

E.I.U. (1995). (The Economist Intelligence Unit), Country Profile, Colom bia, 1994--

1995.


19/20

Land specu l a t i o n nd n t en st l i c a t i o n t t he r on t i er 519

Fisher, M . J., Lascano, C. E., Vera, R. R. and Ripp stein, G. (1992). Integrating the

native savann a resource with improved pastures, In

Pas t u res fo r t he T rop i ca l

L ow l a n d s: C I A T s Con t r i b u t i o n ,


Fisher, M. J., Rao, I. M ., Ayarza, M . A., Lascano, C. E., Sanz, J. I., Thomas, R. J.

and Vera, R. R. (1994). Carbon storage by introduced deep-rooted grasses in

the South American savannas.

N at ur e 371,236238.

Gittinger, J. P. (1982).

Econom i c A na l y s i s o f A g r i cu l t u r a l P r o j ect s .

Johns Hopkins,

University Press, Baltimore, MD.

Guggenbe rger, G., Zech, W. and Thomas, R. J. (1995). Lignin and carbohydrate

alteration in partial-size separates of an Oxisol under tropical pastures following

native savanna.

So i l B i o l . B iochem . 27,

1629-1638.

Hecht, S., Norgaa rd, R. B. and Possio, G . (1988). The economics of cattle ranching

in eastern Am azonia.

I n t erc i enci a 13, 233-240.

Hoyos, P., Vera, R . R., Lascano, C. and France, M. A. (1992). Manejo de1 pastoreo

por productores de la altillanura plana de 10s Llanos Orientales de Colom bia. In

Red I n t er na t i o na l de Eva l ua c i d n de Pas t o s T r op i ca l es , R I EPT .

Document0 de

Trabajo No. 117, CIAT, Cali, Colombia.

IMF. (1995). International Financial Statistics.

Kaimo witz, D. (1996). Livestock and deforestation in Central Am erica in the 1980s

and 1990s: a policy perspective, CIFOR , Bogor, Ind onesia.

Katzm an, M. T. (1974). The von Thuenen paradigm , the industrial-urban hypoth-

esis, and the spatial structure of agriculture.

Am. J . Ag r i cu l t u r a l Econ . 56 ,

683-696.

Kleinheisterkamp, I. and Habich, G. (1985). Estudio biologic0 y tecnico. In

Sis te -

m as de Prod uccid n Pecua r i a Ex t ens i v a : Bra s i l , Co l om b ia , Venezuela .

eds R. R.

Vera & C. Sere, CIAT, Cali, Colombia.

Klink, C. A., Mo reira, A. G. and Solbrig, 0. T. (1993). Ecological impact of agri-

cultural development in the Brazilian cerrados. In

The W or l d s Sav annas : Eco -

nom i c D r i v i n g Fo r ces , Eco l og i ca l Cons t r a i n t s and Po l i c y O p t i o n s o r Sus t a i n ab l e

Land Use .

eds M . D. Young and 0. T. Solbrig, The Parthenon Publishing

Group Ltd, Carnforth.

Lascano, C. E. (1994). Nutritive value and animal production of forage Arachis. In

B i o l o g y and A g r onomy o f Fo r age A r a ch i s .

eds P. C. Kerridge & B. Hardy,


Lascano, C. E. and E uclides, V. P. B. (1994). Nutrition quality and animal produc-

tion on Brachiaria pastures. In

Pr oceed i ng s o f W o r kshop on B i o l o g y , Ag r onom y

and Im p r o vemen t o f B r a ch i a r i a .

October 3-7, CIAT, Cali, Colombia.

Lorente, L. (1994). Estrategias de desarrollo ganadero.

Coyu n t u r a Co l omb i a n a

11,

111-182.

Lowenberg-De Boer, J. and Boehije, M. (1986). The impact of farmland price

changes on farm size and financial structure.

A m , J. Ag r i cu l t u r a l Econ . 68 ,

839-848.

Macedo, J. (1994). State of the art of research on man agem ent of acid soils in the

Brazilian cerrados. EM BRA PA/CP AC, Brasilia.

Ram irez, A. & Botero, B. R. (1989). Program a en LOTU S l-2-3 para el calculo de1

hato estable. CIAT, Cali, Colombia.

Rippstein, G. and Girard, C. M. (1994). Utilization de datos radiometricos espacial

y terrestre para el estudio e inventario de la vegetation de 10s Llanos Orientales

de Colombia.

Pa stur as Tr op i ca les 16, 2-9 .


20/20

520 J. Smith et al.

Rivas, R . (1993). Ana lisis economico en investigation en fincas: Mo delos microeco-

nomicos. In

Z nv es t i gacid n con Pas tu r as en F incas .

Docum ent0 de Trabajo No.

124. CIAT, Cali, Colombia.

SerC, C. & Estrada, R. D. (1985). Marco socioeconomico de la ganade ria en 10s

Llanos Orientales de Colom bia. In

Sis t em as de Pro duccio n Pecua r i a Ex tensi v a :

B ra s i l , Co l om b ia , V enezuela .

eds R. R. Vera & C. Sere, CIAT, Cali, Colombia.

Sm ith, J., Barau, A. D., Goldman, A. and M areck, J. H. (1994). The role of tech-

nology in agricultural intensification: the evolution of maize production in the

Northern Guinea sav anna of Nigeria.

Econom i c D eve l opmen t and Cu l t u r a l

Change 42,537-554.

Sm ith, J., Winograd, M., Gallopin, G. & Pachico, D. (1996). Dynam ics of the agri-

cultural frontier in the Am azon and Savannas of Brazil: analyzing the impact of

policy and technology.

Env i r o nm en t a l M odel l i n g and A ssessm en t

(in press).

Solbrig, 0. T. (1993). Ecological constraints to savanna land use. In

The W o r l d s

Savannas : Econom i c D r i v i n g Fo r ces , Eco l ogi ca l Cons t r a i n t s and Po l i c y O p t i o n s

fo r Sus ta i nab l e La nd U se.

eds M. D. Young & 0. T. Solbrig, The Parthenon

Publishing Group Ltd, Carnforth.

Tamayo, H. (1975). Los llanos orientales de Colom bia; estudio descriptive, Instituto

de Estudios Colombianos, Bogota, Colombia.

Thom as, R. J. (1992). The role of the legume in the nitrogen cycle of productive and

sustainable pastures.

G ra ss an d Forages Sci . 4 7, 133-142.

Thom as, R. J., Sm ith, J., Sanz, J. I. and Vera, R . R. (1996). Approaches to the

sustainable man agement of the Tropical Lowlands in Latin Am erica. In

Ag r i -

cu l t u r a l Resea r ch and Sus ta i nab l e M anagem en t o f N a t u ra l Resou r ces ,

ed. H.-J.

A. Preuss. U niversity of Giessern.

Vera, R. R., Thom as, R., Sanint, L . and Sanz, J. I. (1992). Development of sus-

tainable ley-farming systems for the acid soil savann as of tropical Am erica.

An .

Acad . Bras . C i .

64(l), 105-125.

Wo rld Bank, (1993). Government and economy on the Am azon frontier. Environ-

ment Division, Technical Department, Latin Am erican and Caribbean Region,

Report No. 12644 LAC, The World Bank, Washington DC.

land use changes in llanos

Documents