Pergamon World Development, Vol. 25, No. 10, pp. 1589-1607, 1997

© 1997 Elsevier Science Ltd All rights reserved. Printed in Great Britain

0305-750X/97 $17.00 + 0.00 P I h SO305-750X(97)O0059-4

Growth, Structural Change and Plantation Tree Crops: The Case of Rubber

C O L I N B A R L O W *

Austra l ian Nat iona l University, Canberra, Aus tra l ia

Summary. - - The effects of advancing economic growth on plantations are classed in five stages, starting with conditions in a backward subsistence economy and ending under circumstances where manufacturing is dominant and planting tree crops no longer economic. Changes in relative resource prices and other factors and consequent adjustments of estates and smallholdings are taken into account, doing this in light of international experiences with such crops. The case of natural rubber is scrutinized in depth, comparing economic effects and responses at each stage in chief producing countries. The key elements in plantation adjustments of market conditions, technology, institutional arrangements and government interventions are finally addressed, with policies likely to facilitate appropriate modifications being indicated. © 1997 Elsevier Science Ltd

Key words - - Asia, Africa, growth, rubber, stages model, plantations, interventions

1. INTRODUCTION

This paper considers how traditional labor and land-intensive plantation tree crops I have reacted to general economic development, scrutinizing this from when they were introduced as cash crops in subsistence situations to their probable final demise under conditions of dominant manufacturing. It addresses the case of rubber, examining how estates and smallholdings modified production activities with altering labor, land and capital prices, using new technologies, and taking advantage of other transformations including more integrated markets, better infrastructures and broadening economic opportunities. These opportunities included making local rubber goods, in response to domestic demand and export possibilities.

Plantation tree crops merit special consideration in this regard, being major players on the farming scene. They chiefly involve coconuts, rubber, coffee, oil palm, tea, cocoa and various fruits, which today comprise over 25% of the value of agricultural produce in their main growing region of Southeast Asia (Food and Agriculture Organization, 1995) and a higher share of farm exports. They likewise raise special economic issues compared to annuals, owing to long gestation periods. Their exposure to structur- al change has affected several hundred million persons producing them around the world, and has had great economic and social significance for countries concerned.

The approach of the paper is first to present an analytical framework embodying successive stages of economic growth in relation to plantation crops. The circumstances of rubber production around the world, and differing modes of national adjustment from the late 19th century to the present, are then addressed. Key elements of adjustment are next selected, and are reviewed in various national situations. Broad conclusions are finally presented, and general observations made on processes in- volved.

2. ANALYTICAL FRAMEWORK

Consider for the purpose of this framework an economy with three sectors, agriculture, services and industry including manufacturing. Prior to modem growth and introduction of tree crops there is stage (1) of a backward economy, having a dominant subsistence agriculture based on family farms with miniscule services and industry (Table 1). Land is plentiful and underutilized, while labor engaged in shifting cultivation has a low marginal product. Capital is very scarce. These are typical original

*Thanks are due to colleagues in the plantation and rubber industries of many countries, who willingly supplied guidance and information. Final revision accepted: April 26, 1997.

1589

U~

Tab

le 1

. Ana

lyti

cal f

ram

ewor

k of

pla

ntat

ion

tree

cro

p de

velo

pmen

t

Stag

es

Cha

ract

eris

tics

N

atio

nal

posi

tion

s w

ith

rubb

er a

(1)

Bac

kwar

d ec

onom

y (s

ubsi

sten

ce a

gric

ultu

re,

no p

lant

atio

n cr

ops)

(2)

Ear

ly a

gric

ultu

ral

tran

sfor

mat

ion

(sim

ple

plan

tati

on c

rops

te

chno

logy

)

-- P

redo

min

ant

subs

iste

nce

agri

cult

ure,

min

uscu

le s

ervi

ces

and

indu

stry

. -

-

Lit

tle

trad

e an

d fr

agm

ente

d ru

ral

mar

kets

. -

-

Plen

tifu

l an

d un

deru

tili

zed

land

, la

bor

wit

h lo

w m

argi

nal

prod

uct,

ver

y sc

arce

cap

ital

. --

His

tori

call

y es

tabl

ishe

d sh

ifti

ng c

ulti

vati

on t

echn

olog

y on

fa

mil

y sm

allh

oldi

ngs.

-

-

Min

imal

gov

ernm

ent c

--

C

omm

erci

aliz

ing

agri

cult

ure

the

dom

inan

t se

ctor

, w

ith

esta

tes

and

smal

lhol

ding

s bo

th i

nvol

ved

in p

lant

atio

n cr

op c

ulti

vati

on.

Sm

all

but

grow

ing

serv

ices

and

ind

ustr

y.

--

In

tern

atio

nal

trad

e e a

nd r

ural

mar

ket

deve

lopm

ent

com

men

cing

. --

Lan

d an

d la

bor

beco

min

g sc

arce

and

pri

ces

risi

ng,

capi

tal

beco

min

g m

ore

avai

labl

e.

--

R

apid

ado

ptio

n of

sim

ple

labo

r-in

tens

ive

tree

cro

p te

chno

logi

es,

firs

t on

est

ates

and

the

n on

sm

allh

oldi

ngs.

-

-

Cen

tral

gov

ernm

ent

begi

ns l

evyi

ng t

axes

and

pro

vidi

ng s

mal

l se

rvic

es.

Tha

ilan

d to

192

0 In

done

sia

(out

er i

slan

ds)

to 1

870 b

M

alay

sia

to 1

870 b

In

dia

(Ker

ala)

to

1920

C

hina

(Y

unna

n &

Hai

nan)

to

1960

Sr

i L

anka

to

1870

b N

iger

ia (

sout

hern

are

a) t

o 19

10 d

Vie

tnam

to

1920

P

hili

ppin

es (

Min

dana

o) t

o 19

50

Ivor

y C

oast

to

1950

C

amer

oon

to 1

950

Kam

puch

ea to

192

0

Tha

ilan

d, 1

920-

60

Indo

nesi

a, 1

900-

30 b

Mal

aysi

a, 1

900-

30 b

Indi

a, 1

920-

-40

Chi

na,

mis

sed

out

Sri

Lan

ka,

1900

-193

0 b

Nig

eria

, 19

10-7

0 V

ietn

am,

1920

-40

Phi

lipp

ines

, 19

50-7

0 Iv

ory

Coa

st,

1950

-70

Cam

eroo

n, 1

950-

70

Kam

puch

ea,

1920

-80

©

,<

©

Stag

es

Cha

ract

eris

tics

N

atio

nal

posi

tion

s w

ith

rubb

er a

(3)

Lat

e ag

ricu

ltur

al t

rans

form

atio

n (n

ew p

lant

atio

n cr

ops

tech

nolo

gy)

(4)

Ear

ly a

dvan

ced

econ

omy

(pla

ntat

ion

crop

s be

com

ing

less

pr

ofit

able

)

-- A

gric

ultu

re r

emai

ns o

ne o

f bi

gger

sec

tors

, bu

t se

rvic

es

incr

easi

ng a

nd m

anuf

actu

ring

bas

ed o

n im

port

-sub

stit

utio

n pa

sses

agr

icul

ture

tow

ard

end

of p

erio

d.

Rur

al m

arke

t de

velo

pmen

t pro

gres

sing

, esp

ecia

lly

wit

h go

vern

men

t in

terv

enti

ons.

But

man

y im

perf

ecti

ons

pers

ist.

--

Lan

d an

d la

bor

pric

es r

isin

g; c

apit

al,

man

agem

ent

and

tran

spor

t pr

ices

fal

ling

. -

-

Gen

erat

ion

of n

ew l

and

and

labo

r-sa

ving

but

mor

e ca

pita

l an

d m

anag

emen

t-in

tens

ive

high

-yie

ldin

g tr

ee c

rop

tech

nolo

gies

. A

dopt

ion

firs

t by

est

ates

and

muc

h la

ter

by s

mal

lhol

ding

s.

--

C

entr

al g

over

nmen

t gr

adua

lly

wid

enin

g su

ppor

ting

rol

e, a

nd

foll

owin

g en

d of

col

onia

lism

pro

vidi

ng w

ides

prea

d ru

ral

infr

astr

uctu

res

and

serv

ices

.

Tha

ilan

d, 1

960-

85

Indo

nesi

a, 1

930-

pres

ent

Mal

aysi

a, 1

930-

70

Indi

a, 1

940-

pres

ent

Chi

na,

1960

-pre

sent

Sr

i L

anka

, 19

30-p

rese

nt

Nig

eria

, 19

70-p

rese

nt

Vie

tnam

, 19

70-p

rese

nt

Phil

ippi

nes,

197

0-pr

esen

t Iv

ory

Coa

st,

1970

-pre

sent

C

amer

oon,

197

0-pr

esen

t K

ampu

chea

, 19

80-p

rese

nt

-- M

anuf

actu

ring

bec

omes

muc

h la

rger

tha

n ag

ricu

ltur

e, b

eing

T

hail

and,

198

5-90

in

crea

sing

ly e

xpor

t-or

ient

ed a

nd i

nclu

ding

dow

nstr

eam

pla

ntat

ion

Mal

aysi

a, 1

970-

85

crop

pro

cess

ing

into

fin

al g

oods

. -

-

Rur

al m

arke

t m

uch

bett

er i

nteg

rate

d an

d co

mpe

titi

ve,

pock

ets

of i

mpe

rfec

tion

per

sist

. --

Res

ourc

e pr

ice

tren

ds o

f (3

) co

ntin

uing

, but

lan

d an

d la

bor p

rice

ris

es

acce

lera

ting

and

rur

al-u

rban

wag

e di

ffer

enti

als

wid

enin

g. C

onse

quen

t la

bor

mig

rati

on t

o to

wns

. -

-

Tre

e cr

op t

echn

olog

y ge

nera

tion

and

ado

ptio

n co

ntin

uing

in

dire

ctio

ns u

nder

(3)

, w

ith

conc

omit

ant w

iden

ing

of a

vail

able

tec

hniq

ues

to a

ssis

t ad

just

men

t to

dif

fere

nt c

ircu

mst

ance

s. S

hift

ing

tow

ards

ge

nera

tion

and

ado

ptio

n (b

y pl

anta

tion

cro

p pr

oces

sors

) of

qua

lity

- im

prov

ing

tech

niqu

es n

eede

d by

pla

ntat

ion

crop

goo

ds s

ubse

ctor

. -

-

Gov

ernm

ent

goin

g on

wit

h pr

ovis

ions

of

rura

l in

fras

truc

ture

s an

d se

rvic

es u

nder

(3)

. Pr

evio

us r

egul

atio

ns i

n tr

ade

bein

g gr

adua

lly

rem

oved

.

C~

0 Z

Z

Z

Z

©

Tab

le 1

--

cont

inue

d ov

erle

af

b~

Tab

le 1

. con

tinu

ed

Stag

es

Cha

ract

eris

tics

a N

atio

nal

posi

tion

s w

ith

rubb

er a

(5)

Lat

e ad

vanc

ed e

cono

my

(pla

ntat

ion

crop

s no

t pr

ofit

able

) --

Man

ufac

turi

ng n

ow p

redo

min

ant,

bei

ng t

wo

or m

ore

tim

es

Tha

ilan

d, 1

990-

pres

ent

bigg

er t

han

agri

cult

ure

and

incl

udin

g si

gnif

ican

t pla

ntat

ion

crop

M

alay

sia,

198

5-pr

esen

t pr

oces

sing

int

o fi

nal

good

s. A

lso

impo

rtat

ion

of n

atur

al r

ubbe

r fr

om

othe

r co

untr

ies

to s

uppl

y th

is g

oods

ind

ustr

y.

--

R

ural

mar

ket

as f

or (

4),

albe

it s

uper

ior

to i

t in

int

egra

tion

and

co

mpe

titi

on.

--

R

esou

rce

pric

e an

d ex

chan

ge r

ate

tren

ds o

f (4

) pe

rsis

ting

. Tra

diti

onal

pl

anta

tion

cro

p pr

oduc

tion

is u

neco

nom

ic e

xcep

t in

rem

ote

sett

ings

, but

ex

isti

ng t

rees

are

sti

ll b

eing

exp

loit

ed a

s a

"sun

set"

cro

p.

Tre

e cr

op t

echn

olog

y ge

nera

tion

and

ado

ptio

n ch

iefl

y co

ncen

trat

ing

on q

uali

ty-i

mpr

ovin

g te

chni

ques

for

goo

ds s

ubse

ctor

. --

Gov

ernm

ent

cont

inui

ng a

s un

der

(4).

Plan

tati

on c

rop

supp

ort

mea

sure

s in

crea

sing

ly b

ecom

e w

elfa

re f

or o

lder

gen

erat

ions

.

©

< ©

Z

awhi

ch c

ame

in a

fter

the

Bac

kwar

d E

cono

my

(sta

ge 1

). bC

ocon

ut,

coff

ee a

nd c

ocoa

fro

m 1

870,

rub

ber

firs

t pl

ante

d in

190

0.

CT

radi

tiona

l go

vern

men

t co

mpr

isin

g li

ttle

loc

al k

ingd

oms

whi

ch m

ay n

onet

hele

ss c

olle

ct t

ax.

"Cen

tral

gov

ernm

ent"

fro

m (

stag

e 2)

oft

en e

ntai

ls a

n in

vadi

ng c

olon

ial

regi

me,

al

thou

gh b

y th

e en

d of

(st

age

3) i

t is

usu

ally

of

"ind

epen

dent

" lo

cal

orig

in.

The

mod

ern

stat

es t

hen

invo

lved

hav

e va

ryin

g de

gree

s of

dev

olut

ion

at l

ocal

lev

el.

dWild

Fun

tum

ia r

ubbe

r ex

ploi

tati

on b

egan

in

the

late

19t

h ce

ntur

y, a

nd a

ccou

nts

for

earl

y N

iger

ian

outp

ut (

Tab

le 2

).

eThi

s fe

atur

e pe

rsis

ts f

or r

emai

ning

sta

ges,

bec

omin

g in

crea

sing

ly i

mpo

rtan

t.

GROWTH, STRUCTURAL CHANGE AND PLANTATION TREE CROPS 1593

circumstances in the humid tropics, which form a necessary growing environment for plantation crops.

Enter now international trade and the opening of markets in stage (2) of economic growth, termed early agricultural transformation. Agriculture re- mains the chief sector, but now has a commercializ- ing orientation while services and industry are growing. This stage frequently coincides with the onset of colonial government and arrival of estates motivated by profits to cultivate plantation crops (Hayami and Kawagoe, 1993, p. 4). The estates need managers and laborers for their institutional struc- ture, while capital from international markets is required to finance these resources. 2 Estate manage- ments and other entrepreneurs now acquire a plantation tree crop technology 3, promoting its cultivation geared to world markets.

The tree crop is first planted on estates, but is soon taken up by smallholders. It has the key trait of being easily grown, involving simple planting of seeds under relatively unskilled land and labor-intensive husbandry. The world market provides what Myint (1954) terms a "vent" for surplus productive capacity, enabling utilization of spare land and (on smallholdings) underemployed labor to produce a valuable export. The widespread adoption of this simple crop with its cash-earning capacity starts transforming the backward domestic economy, encouraging demand for tradeables comprising locally produced foods and imports of other items. Demand for non-tradeables including various public services also rises, and government provides the latter through taxing the new crop. Yet given that tree-growing technology remains unchanged, the crop can only continue extending while land and labor resources remain available. Eventually in the absence of resource-saving techniques, no further advances can be made.

Yet an agricultural sector rarely maintains one tree crop and one local food subsector for long, since once entrepreneurs become active and farmers' receptivity improves, other simple technology-based export-oriented cash crops are adopted. Constraints on land and labor soon appear, and when these are reached competition among individual crops, and between estates and smallholdings intensifies. Prices of previously plentiful land, labor and other factors begin rising, and profitable subsectors expand, exerting "resource pull" effects on others. 4 Factor prices themselves affect profitability, which in turn is dependent on resource-use configurations dictated by underlying technologies. This is the only stage when plantation agriculture leads the economy, for it is later overtaken by other activities. It is termed "getting agriculture moving" by Timmer (1988), who analyzes agricultural transformation from a perspective of extracting resources for broad national development.

Now economic growth enters stage (3) of late agricultural transformation, when industry is gain- ing momentum and manufacturing growing fast. Agriculture starts as the biggest sector, but is overtaken by manufacturing which may indeed be encouraged by government policies supporting import substitution. A vital initiative in this period is demand by producers for new plantation technol- ogies, with resource requirements better matching changed factor prices; these involve increases for land and labor, and decreases for management and capital which are easier to obtain. Such demand entails the induced institutional and technical innovation outlined by Binswanger and Ruttan (1978), motivating governments and other agencies to supply research services and researchers to generate less land and labor-intensive technologies.

The new technologies markedly include varieties of trees yielding greater outputs per hectare of land and person-day, albeit requiring more skill as well as more capital to buy seeds, fertilizers and other chemicals. Once available they are quickly taken up by estates subject to production cycles; 5 this adoption is facilitated by trained management, capacities to tap central information sources and direct access to capital markets.

Meanwhile a critical limit appears on plantation smallholding development and other small farm sectors. For the new technology is not adopted by most farmers, who continue using old methods though with greater efficiency flowing from learn- ing-by-doing. The limit basically springs from incomplete markets, where despite rising competi- tion many imperfections remain. Thus knowledge and handling skills pertinent to adopting new technologies have not yet diffused to the farm level, while credit for plantation investments with long gestation periods is not available from private lenders. The lack of knowledge means smallholders do not match estate operators in demanding new techniques, and such circumstances are worsened by exclusion of rural dwellers from political

• - 6 actwlty. Demand for new technologies to be brought to

smallholders ultimately occurs at this stage, how- ever, commonly springing from governments and international agencies in the modified public philo- sophy following transition from colonialism to independence. There is fresh official concern for broad-based economic development and social wel- fare, with these being pursued through enhanced rural infrastructures and health and education services as well as programs targeting particular subsectors. The latter programs include extension and credit to help planting of higher yielding trees on smallholder plantations, often encouraging wide- spread adoption. Participating farmers further secure "latecomer" advantages, since new technology

1594 WORLD DEVELOPMENT

originally developed for and used by estates now has greater efficiency in terms of higher output-input ratios and wider menus of techniques. The energized government interventions encourage more integrated and competitive rural markets, including better information and lower transport costs.

Meanwhile downstream manufacturing of planta- tion crops is progressively established, beginning with domestically consumed items and benefiting from accessible raw materials. These manufacturing expansions pull resources and notably labor from other sectors. They likewise involve spending effects and enhanced demand for services, with conse- quently rising prices of non-traded goods again lifting resource charges and appreciating exchange rates. Such appreciation may also be caused directly by import substitution, acting to reduce export prices of tree crop outputs.

Stage (4), early advanced economy (Table 1), begins when manufacturing becomes much larger than agriculture and moves to an export orientation. This reinforces adverse consequences for tree crops just outlined, with increases in land and labor costs now being accompanied by widening rural-urban wage differentials. The latter often trigger urban migration by younger people in particular, posing long-term adjustment difficulties for family-based farms. Increasing exports from manufacturing pro- voke further appreciations in exchange rates.

Manufacturing growth has positive outcomes for plantations, however, notably by strengthening demand linkages through expanded downstream processing. Its spending effects allow heightened government provisions of rural infrastructures and services, all of which stimulate market integration. These provisions may be further increased owing to greater popular demand accompanying better com- munications. The generation of new techniques moves firmly toward less land and labor-intensive and more management and capital-intensive innova- tions, being also induced by manufacturers to improve processing and quality methods for fabri- cating rubber goods.

New production innovations are gradually adopted, assisting moves to less costly resources. There may also be shifts to less labor-intensive tree or annual crops, usually at the end of the cycle. While such changes are again most evident on estates, they are now easier to make on smallhold- ings whose operators have become more sophisti- cated following investments in human capital. The freshly developed processing systems facilitate manufacturers' utilization of plantation products as components of goods chiefly for export.

Finally, economic development reaches stage (5) of a late advanced economy, where manufactur- ing is dominant and becoming far larger than agriculture (Table 1). Now resource pull, spending

and exchange rate effects are greatly accentuated, further undermining the economic viability of plantations. 7 With the exception of remote districts, wages and land values are well above levels justifying further investment in traditional tree crops, while many rural laborers are ageing and unable to cope following outmigration of younger cohorts. Spending on food by urban dwellers is increasingly on meat, fruits and specialist vegeta- bles, reflecting high income elasticities of demand for such products. Both estate operators and better educated smallholders find producing the latter items more profitable, using management and capital-intensive and labor-saving techniques; these activities replace traditional plantations, although old trees are still exploited as long as they earn sufficient returns over direct costs, s Estate owners further invest offshore, using their accumulated technology and management expertise in places still having low labor and land costs.

Manufacturing by the advanced economy stage includes sizeable downstream plantation crop pro- cessing, exploiting economic advantages of utilizing high-quality domestically produced intermediate goods consequent on adopting new techniques. Consumption of these goods has further positive linkage effects, although lower quality needs are often sourced from other countries at earlier stages of development. Technology generation and adoption in plantation items now concentrate on intermediate goods processing and final goods manufacturing, with a focus on meeting demand on international markets.

Yet although rural markets are much more integrated, pockets of smallholders persist who do not adjust and earn declining incomes in burgeoning general economies. This occurs especially in coun- tries occupying large and scattered land masses, with associated high communication and transportation costs. Often pockets may be encouraged by regula- tion, including legal restrictions on land sales to those who could operate larger, more economic units. Such restrictions paradoxically remain when most official thrusts are toward deregulaton. Govern- ments with spending power tend to view small- holders in these pockets as objects of welfare: older individuals are helped with income transfers, while younger people are trained to take up remunerative opportunities elsewhere.

The analytical framework just portrayed rests on observed economic behavior in plantation crop countries at differing stages of economic growth. It broadly portrays evolving reality and cannot precisely fit given national tree crop sectors. Its stages are more clearcut than practical eventuality, where some phenomena begin earlier or persist later than suggested. The framework clearly embraces a complex picture, with positive and negative influ-

GROWTH, STRUCTURAL CHANGE AND PLANTATION TREE CROPS

Table 2. World natural rubber production

1595

Thailand Indonesia Malaysia India China Sri Lanka Vietnam a

Production COO0 t) --1910 - - 3 (3) b 6 (6) - - - - 2 (2) --1930 4 (l) 245 (29) 467 (56) 9 (1) - - 77 (9) I I (l) --1950 114 (6) 707 (37) 761 (40) 16 (1) - - l l6 (6) 92 (5) --1970 287 (9) 815 (26) 1,269 (40) 90 (3) 46 (1) 159 (5) 28 (l) --1990 1,271 (25) 1,262 (25) 1,291 (25) 324 (6) 264 (5) 113 (2) 103 (2) --1995 1,786 (31) 1,420 (24) 1,085 (19) 500 (9) 360 (2) 103 (2) 95 (2) Area, 1990 ('000 ha) 1,844 [95] c 3,155 [83] 1,837 [81] 451 [83] 603 [na] d 199 [33] 250 [na] High-yielding trees, 52 17 95 92 100 f 75 15

1995 (%)~ Yield, 1990 ---(kg/mature ha) 847 677 800 1,057 na 772 na ---(kg/planted ha) 689 400 599 718 438 568 352 Consumption, 1995 150 133 327 516 732 36 30 ('000 t)

Nigeria g Ivory Coast Philippines Cameroon Kampuchea Others h World

Production ('000 t) --1910 14 (14) b g - - g - - 70 (65) 98 --1930 5 (l) g - - g a 20 (2) 838 --1950 56 (3) g 1 (-) g a 27 (1) 1,890 --1970 65 (2) 11 (-) 20 (1) 12 (-) 3 (-) 172 (5) 3,140 --1990 152 (3) 69 (1) 61 (1) 38 (1) 35 (1) 136 (3) 5,120 --1995 93 (2) 77 (1) 60 (1) 55 (1) 44 (???) 142 (2) 5,820 Area, 1990 ('000 ha) 247 [81] c 68 [29] 88[75] ~ 41 [5] ~ 19 [na] 283 [na] 9,085 ~ High-yielding trees, 10 90 30 70 l0 na na

1995 (%)e Yield, 1990 ---(kg/mature ha) na 1,712 841 na na na na --(kg/planted ha) 615 1,015 693 927 1,842 481 564 Consumption, 1995 19 8 j 36 J 10 na 5,920 ('000 t)

Sources: Barlow et al. (1994); International Rubber Study Group (1946-96). aKampuchean and "other" Southeast Asian production included in Vietnam up to 1970. bFigures in parentheses along production lines are shares of each country's NR production in total world production. ~Figures in brackets along this line are per cents of smallholdings in total planted area of early 1990s. dProbably about 30%, with the balance being under state farms. eEstimated by author, using best available information. fBut the smallholder area in particular is badly managed. gAll African production included under "Nigeria" up to 1990. hMainly Brazil and Guatemala up to 1970. Subsequently including Myanmar, Liberia, Zaire and several other small producers. 'Percentage of holdings less than 5 ha. JAil African consumption included under Ivory Coast.

ences at each stage needing to be scrutinized carefully as economic adjustments are explored. The paper now turns to doing the latter for rubber.

3. N A T U R A L R U B B E R AS A P L A N T A T I O N CASE

(a) Overal l pa t terns

Table 2 shows that a lmost three-quarters o f natural rubber in the mid-1990s was produced by

Thailand, Indonesia, and Malaysia, with almost half the balance coming f rom India and China. The history o f planting rubber goes back to early this century, and Indonesia and Malaysia originally began producing it in the 1900s. 9 Thailand entered the sector 20 -30 years later, however . India first grew the crop in the 1920s, b u t only launched it substantially in the 1960s when China too was beginning its cultivation. Other smaller producing countries including Sri Lanka, Vietnam, and several Wes t African nations started hal f a century or more ago, and Brazil was the original native source o f the

1596 WORLD DEVELOPMENT

rubber tree. The Phi l ippines became a small producer from the 1950s.

Malaysia 's production declined from the late 1980s, responding to very high land and labor costs accompanying its economic advance. But Thai output was still increasing in the 1990s, reflecting a huge program of replanting with high-yielding trees from the mid-1960s. Indonesia 's crop was rising steadily due to continuing expansion in planted land area. Indian and Chinese supplies were advancing quickly, following active planting with improved trees. Turnout from other small producing nations was growing, albeit subject to climatic and other constraints explored below.

Table 2 denotes that most planted rubber is on smallholdings, and this notably applies to the three

big producers and India. All countries nonetheless have estate plantings, while the Ivory Coast and Cameroon as well as China and Vietnam with their state farms have most rubber areas in this class. The institutional arrangements of production influence development significantly, as suggested in the framework.

Table 3 presents selected economic parameters for main natural rubber producing countries over the last two decades, indicating wide divergences in GNPs per head, rates of growth, shares of agriculture and manufacturing, movements in exchange rates and levels of agricultural wages. At one extreme Malaysia following sustained high growth had the highest GNP per head of $4,010 in 1995, and could with its 33% manufacturing share dwarfing the 14%

Table 3. Economic circumstances of natural rubber producing countries a, 1970s-90s

Thailand Indonesia Malaysia India China Sri Lanka

Current GNP/head ($), 270 (2,713) 120 (969) 600 (4,010) 130 (331) na (506) 230 (600) 1973 (1993) b'c GDP Growth (%/year), 7.8 (8.7) 6.8 (6.0) 7.6 (6.6) 4.3 (4.9) 6.6 (10.3) 5.8 (4.2) 1973-83 (1984-93) GDP Shares (%) --1973, Agriculture 23 (20) 33 (6) 29 (16) 42 (15) 44 (24) 33 (17) (Manufacturing) --1993, Agriculture 11 (30) 17 (25) 14 (33) 29 (19) 21 (44) 23 (19) (Manufacturing) Exchange Rate Change d --1973 as % 1983 (1983 90 (91) 45 (44) 105 (90) 76 (33) 76 (33) 27 (49) as % 1993) Current Agric. Wage ($ 1.08 (5.50) 0.50 (1.70) 2.70 (8.50) low f (1.70) low f (2.30) per day) 1973 (1996) e

Vietnam Nigeria Ivory Coast Philippines Cameroon World

Current GNP/head ($), na (219) 310 (300) 350 (630) 250 (1,105) 220 (820) 1,130 (4,420) 1973 (1993) b'c GDP Growth (%/year), na (ha) 1.8 (4.1) 4.0 (-0.3) 5.2 (1.1) 9.2 (-1.8) 2.9 (3.1) 1973-83 (1984-93) GDP Shares (%) --1973, Agriculture na (na) 44 (5) 48 (na) 27 (29) 41 (5) na (na) (Manufacturing) --1993, Agriculture 27 (23) 33 (9) 37 (na) 21 (26) 25 (12) na (ha) (Manufacturing) Exchange Rate Change d --1973 as % 1983 (1983 na (neg) 91 (3) 58 (135) 61 (41) 68 (134) - - (-) as % 1993) Current Agric. Wage ($ na (low) f na (low) f low e (2.80) low s (2.30) low f (2.00) na (ha) per day), 1973 (1996) e

Source: World Bank (1996). aExcluding Kampuchea, for which no data is available. bFigures for Thailand, Indonesia, Malaysia, India, China, Vietnam and the Philippines refer to 1995. Cltems in parentheses match with subsequent figures in parentheses in each case. dRate of local currency against US$ in first year as a per cent of rate in second year. Percentages below 100 imply depreciation, and above 100 appreciation, of currency concerned. eEstimated on basis of available information. Rates include value of perquisites and apply to major rubber-growing areas of countries. fBelow $1.00 per day.

GROWTH, STRUCTURAL CHANGE AND PLANTATION TREE CROPS 1597

of agriculture be classed as having reached stage (5) of a late advanced economy. At the other extreme Nigeria and Vietnam with their GNPs in 1993 of $300 and $219 were still poor countries which had only just entered stage (3) of late agricultural transformation. These and other economic variations between countries are reflected in divergent prices of resources including labor.

Rubber, like many tropical tree crops, can only be grown economically in a humid belt of up to 1,000 kilometers around the equator. Conditions become less suitable for rubber with outward moves in this belt, since dry seasons become longer. Thus Thai- land has a three month dry spell in its north east compared to one month in its southern growing areas, while Yunnan which is north of Thailand additionally experiences cold nights and even periodic frosts. Java is good for rubber in the west, but runs into severe dryness constraints toward the east. Yields are lower and production more costly in such outlying places, making rubber less competitive with alternatives.

(b) Transitions between stages

The natural rubber countries of Tables 2 and 3 together covered all five stages of the analytical framework, from the late 19th century to the present. The years when countries entered and left stages are noted in Table 1. Detailed backgrounds and references on natural rubber in national situations are given by Barlow et al. (1994).

The move from stage (1) of a backward economy to stage (2) of early agricultural transformation occurred with rubber in the late 19th and early 20th century in Indonesia, Malaysia, and Sri Lanka. j° This was a crucial phase in global economic development, marked also by the opening of Suez and growing trade from the Western industrial revolution. It conformed with the framework in being aided by extensions of colonialism, offering political stability and regularized contact with external commerce. It was greatly facilitated by introduction of estate enterprises motivated to gain access to and adopt the new rubber-growing technology. While these enterprises secured devel- opment capital in Europe and the United States, colonial regimes backed them in obtaining land, importing labor and constructing roads and other infrastructures required to support large-scale rubber cultivation (Allen and Donnithorne, 1954).

The simple technology assisted the move to rubber, following the stage (2) prescription in this respect (Table 1). Although it was introduced to estates by professional managers, these were really not needed for the technology per se being chiefly hired to control large workforces. It basically

comprised seeds growing into robust trees with scant maintenance and no fertilizer, and was thus easily handled by smallholders possessing little technical knowledge, plentiful land, family labor and some access to new infrastructures. These techniques were in fact observed on estates by smallholders, who adapted them for their small farms.

Similar advances spurred by colonialism and estates occurred over the next two decades in Nigeria and Kerala (Table 1). But transformation in Indo-China in the 1920s and the Ivory Coast and Cameroon in the 1950s hardly proceeded beyond estates, for rubber planting on smallholdings was officially discouraged. Small plantations were slug- gish in Thailand beginning in the 1920s, since there was no colonial power and no estate subsector; they only started spreading after official encouragement in the 1960s (Table 2). Early transformation was not reached in the Philippines until the 1950s, when government encouraged first estate and then small- holding expansions as initiatives for the southern island of Mindanao. Stage (2) was completely missed by China, which from 1960 used state farms to move directly from a backward economy into stage (3) of late agricultural transformation.

The first entrants to stage (3) were Indonesia, Malaysia and Sri Lanka. These having started plantations earliest began after the 1930s depression to face substantially rising prices of land and labor in regions concerned. In Indonesia the generation of land and labor-saving but more management and capital-intensive high yield technologies marking this stage was primarily by private estates, under- taking this cooperatively in response to their needs. In the different colonial philosophies of Malaysia and Sri Lanka most research was carried out by government agencies, albeit benefiting through knowledge transfers from path-breaking Indonesian workJ 1 Yet while new techniques were mostly adopted by all three countries' estates as they completed crop cycles, the limit on smallholder adoption suggested in the framework was evident everywhere. This almost universally prevented transitions to higher yielding varieties (Bauer, 1948).

Yet after two decades and the end of colonialism in the 1950s, the depressed circumstances of rubber smallholders in Malaysia and Sri Lanka changed radically. Newly independent governments greatly enhanced rural infrastructures and services, also using credit and extension to encourage adoption by smallholders of higher-yielding trees; the latter had been improved in the interim, although modifica- tions to suit small farm conditions were frequently needed.

In Indonesia, on the other hand, little altered until the late 1960s due to political instability, absence of economic growth and consequent government in- ability to undertake much development. But then

1598 WORLD DEVELOPMENT

official efforts were made to stimulate rubber and other tree crop improvement during the long period of stability and economic growth from the 1970s to the present. Although progress was made with rural infrastructure and services, efforts targeted at rubber smallholders were constrained by limited resources 12 and bureaucracies resistant to change. Most small- holders in Indonesia today accordingly still use the original simple technology of the 1900s, albeit with greater efficiency learned through practice. This lack of progress is reflected in persistent low yields (Table 2).

The then small Indian natural rubber sector entered stage (3) in 1940 (Table 1), when govern- ment help in crop improvement was spurred by the war and need to produce rubber for the Anglo- American-Russian alliance. Subsequently after in- dependence, government support to the largely smatlholder rubber sector was motivated by joint desires for rural development and self sufficiency for a goods manufacturing sector serving the huge domestic market. Policies resembling those in Malaysia and Sri Lanka were supplemented by high output prices induced through tariffs and quotas on natural rubber imports, which stimulated spurts in output from the 1970s (Table 2). Motivations like those of India attended China's efforts to promote natural rubber when it entered stage (3) from the 1960s. The main vehicle of Chinese improvement was state farms, although independent smallholders also began to participate from the late 1970s under the production responsibility system. The big local demand for natural rubber in China and India is illustrated in Table 2, showing that by 1995 they were third and fourth only to the United States and Japan as global consumers of the commodity. Their rapid consumption increases likewise meant they had to supplement home production with imports.

Yet the most impressive movement in stage (3) took place in Thailand, beginning in the 1960s and building on minor areas of rubber smallholdings. The government greatly improved infrastructures and services, 13 and in institutional spillovers from Malaysia and Sri Lanka established targeted exten- sion and credit arrangements to assist in rubber planting. These actions were very effective, trans- forming the smallholder-based Thai rubber sector from minor player to global leader in less than three decades (Table 2).

Excepting Kampuchea, all other countries detailed in the tables but not yet mentioned moved to late agricultural transformation with rubber in the 1970s (Table 1), but apart from the Philippines shifts to new techniques were chiefly confined to estates. In Vietnam as in China efforts were channelled through state farms, albeit with small moves to encourage smallholders from the 1980s. In Nigeria even stage (3) transformations on estates were muted, since

workers were drawn from all rural activities by massive urban projects financed by oil boom revenues of government. Although the Ivory Coast and Cameroon tried from the 1980s to include smallholders in "nucleus estates" giving technical help and other services to surrounding farmers, these had limited impact since managers tried to maximize benefits to core operations. It is interesting that nucleus estates were employed in Indonesia with similar disappointing outcomes. In the Philippines, however, smallholdings benefited from estate spread effects as well as direct government help. All countries coming to stage (3) at later times gained through technical spillovers from earlier entrants, notably from larger rubber nations sustaining economies in research. Yet local adaptation of techniques was also usually needed, and its absence in Indonesia and West Africa acted to further brake adoption in these places.

The postulated economic growth effects on resource prices were universally evident during early and late agricultural transformation, although some- times mitigated by market imperfections and remo- teness from development centers; the latter especially occurred in countries where government did not attempt to improve market integration and enhance competition. Rural wages were placed under further upward pressures owing to import-substitut- ing policies during late transformation, and this notably took place where rubber was grown near industrializing centers; it was true, for example, of peninsular Malaysia 14 in the 1960s and Hainan- Guangdong in the 1990s. Although import substitu- tion tended to appreciate exchange rates, most countries except Thailand, Malaysia and some West African nations acted from the 1970s to strongly depreciate their rates against the US dollar (Table 3); this was not only the unit in which rubber was (and still is) largely traded, but also a main item in the basket determining most national currencies. These reactions to freer world financial markets followed earlier overvaluations, acting to usefully raise domestic prices of rubber and other export items.

Given the new techniques, estates in stage (3) moved strongly in altered price configurations to plant rubber, albeit sometimes shifting to less labor- intensive crops such as oil palm, cocoa and fruit trees. Smallholders getting effective official help mainly established rubber, which for them was the most suitable crop. J5 These investments caused large rubber output rises from the 1950s to the 1990s in Thailand, Malaysia, India, China and other countries growing improved trees (Table 2).

(i) Later transitions Desires to invest in rubber planting diminished

with the entry to the early advanced economy, stage (4) beginning with Malaysia in 1970 and Thailand in

GROWTH, STRUCTURAL CHANGE AND PLANTATION TREE CROPS 1599

1985 (Table 1). Rubber workers in Malaysia now had many new opportunities, including jobs in urban areas; these were actively taken up by younger people, causing dramatic shifts of population to towns from the mid-1970s. The changes enhanced rural wages, although this trend was countered through mounting and often "illegal" entry of migrant laborers from Indonesia and elsewhere. Modes of production from existing trees shifted toward labor-saving methods, drawing on techniques now generated by researchers. Similar trends occurred in Thailand, but alterations were less widespread owing to distances between the Bangkok manufacturing centers and outlying districts. Labor shortages were again counteracted by migrants from depressed national areas and Myanmar, while targeted improvement schemes now allowed produ- cers to interplant rubber with fruit trees, thus hedging against profitability declines. The Thai government further tried to stimulate rubber planting in the northeast of the country; that proved difficult, however, owing both to low profit expectations in a less suitable climate and the existence of superior alternatives.

Shifts to export-oriented manufacturing occurred in both countries in stage (4), further enhancing the wage pressures outlined. They were spurred by new emphases on economic deregulation, leading to even more open economies. The new emphases on manufacturing followed the framework in including rubber goods sectors favoring natural-rubber inten- sive products; these especially comprised heavy-duty tires and "latex" items such as surgical gloves and condoms, being supported by novel techniques from research institutes. Manufacturing growth was helped through joint ventures providing markets and expertise, and Malaysia now became a global specialist in the products concerned. The Malaysian and Thai developments from the 1980s caused slight currency appreciations, accordingly penalizing tradi- tional plantation and other export crop producers.

Finally, Malaysia entered stage (5) of a late advanced economy in 1985, and Thailand did so in 1990 (Table 1). In Malaysia especially, further rubber planting was not economic with very high wages (Table 3) and land prices accompanying the dispersal across the peninsula of industrial centers, but estates continued to exploit old rubber trees employing mainly foreign migrants. At the ends of crop cycles they partly replanted with less labor- intensive crops, utilizing other land for housing or industry. Many Malaysian rubber estate companies used their technology and management in invest- ments overseas, especially in Indonesia and the Philippines with their far lower labor and land costs. Smallholders too went on tapping rubber with hired migrants, although their higher price elasticities of supply meant outputs declined greatly during the low

output prices of the early 1990s. ~6 The family ageing factor suggested in the framework also operated, and taken with land transfer restrictions meant old smallholding rubber parcels were often abandoned rather than sold, amalgamated and planted to new more crops. Government support for older Malaysian rubber workers was increasingly seen as welfare, albeit accompanied by reorganization in block development schemes. 17

In Thailand rubber planting slowed markedly, while fruit and other trees were increasingly favored. But active exploitation of existing trees continued with help from migrant laborers, and total national rubber production seemed likely to go on rising until the year 2000 owing to areas extensively replanted up to the late 1980s. Progressive declines in production were forecast thereafter, however, paralleling exten- sions of manufacturing throughout the country.

The growth of export-oriented rubber goods manufacturing continued apace in stage (5), and Malaysia with its specialization had become the fifth biggest global consumer of natural rubber by the mid 1990s (Table 2). It had through its manufacturing technology generation become the source of new processing techniques licensed to other ventures around the world. As suggested in the framework, it secured lower quality rubber from other less developed sectors, notably Vietnam and Nigeria. Thailand too considerably raised its rubber con- sumption, which was expected to expand greatly through demand for locally manufactured tires from its expanding domestic auto industry.

(ii) Resource price effects The impacts on rubber plantations of changing

resource prices in stages (3), (4) and (5) are shown in Table 4, which is based on studies of national production sectors and presents estimated costs of production per kg using improved techniques. The wage of $2 per day and land price of $250 per hectare under late agricultural transformation are relevant opportunity costs in countries such as Indonesia and the Philippines, while for Malaysia the figures are $10 and $12,000 under late advanced economy. Table 4 takes into account of achievement during the transition of improved labor efficiency, lowered transport and forwarding charges, and substitutions of material inputs such as fertilizers and yield stimulants for labor.

Only producers in late agricultural transformation consistently earn profits over grand total costs at the border prices of 75-150 cts per kg obtaining in 1990-95. Producers in early advanced economy secure small profits at the high end of the price range, while those in late advanced economy always incur losses. Circumstances look better, however, when variable production costs alone are considered in an appropriate short-term outlook, for then even

1600 WORLD DEVELOPMENT

producers in late advanced economy get some profits at higher price levels. These comparisons illustrate growers' loss of comparative advantage as labor and land prices rise, stressing that new investments in rubber will not be profitable beyond stage (3); at the same time, however, growers from this stage have new more remunerative opportunities, comprising not only industrial jobs but fruit trees and vegetables in Thailand and oil palm in Malaysia. Such trends have crucial implications for development policy.

Responses of Malaysian rubber estates and smallholdings to changing factor prices over 1922- 95 are presented in Table 5. This details altering

shares of land, labor, management and capital as well as costs of these items and other information at intervals corresponding to stages (2) to (5) of the model. Average quoted market prices of land and labor are also given, although opportunity costs of smallholder resources were often lower than these. The comparisons are not controlled, for surveys underlying the data were conducted for varying purposes. Yet they are thought to portray evolving circumstances quite well.

Table 5 denotes large shifts in factor shares, with land on estates climbing from minor roles in initial development and production phases in 1922 to

Table 4. Estimated revenues and costs from smallholder natural rubber production in advancing economies

Late agricultural Early advanced economy Late advanced economy transformation

Labor (S/day) 2 6 10 Land ($.ha) 250 6,000 12,000 Revenue --Yield (kg/ha) 900 900 900 --Border price (cts/kg, fob) a 75-150 75-150 75-150

Costs (cts/kg) --Field Production b Management and labor c 27.2 49.0 68.1 Materials d 6.0 7.5 9.0 Transport e'f 3.0 2.5 2.0

36.2 59.0 79.1 --Processing g Management and labor 2.2 6.5 10.9 Materials 4.0 5.0 6.0

6.2 11.5 16.9 --Forwarding Domestic e 4.5 3.0 1.5 External h 12.0 11.0 10.0

16.5 14.0 11.5 Total Variable 58.9 84.5 107.5 ~ a p i t a l Planting' 8.9 26.8 44.6 Processin~ 0.8 0.8 0.8

9.7 27.6 45.4 --Land k 1.4 33.3 66.7 Grand Total 70.0 145.4 219.6

aFor Technically Specified Rubber (TSR) as the major internationally marketed grade. blncluding initial processing to low-level intermediate goods (sheet or slab). CTaking task sizes of 300 trees per tapper (late agricultural transformation), 500 trees (early advanced economy) and 600 trees (late advanced economy), with common densities of 350 tapped trees per ha and tapping days of 105 per year. dlncluding costs of equipment. eTo point of high-level intermediate processing. This cost is usually incurred by traders. fAssuming progressive falls in domestic transportation costs with better roads in advancing economies. glnto high-level intermediate product. hFrom national boundaries to rubber goods factories in foreign industrial centers. Assuming some decline in external forwarding costs with better port facilities in advancing economies. ~Principal repayment and 5% real interest charges over 20 years on gross investments in bringing plantings to maturity of $1,000/ha (late agricultural transformation with $2/day labor cost), $3,000/ha (early advanced economy with $6/day labor cost) and $5,000/ha (late advanced economy with $ I 0/day labor cost). JPrincipal repayment and 5% real interest charges over 20 years on investments in all cases of $1.2 million for a 60t/day TSR factory with 75% capacity utilization. k5% real interest charges on land costs shown.

GROWTH, STRUCTURAL CHANGE AND PLANTATION TREE CROPS 1601

Table 5. Mean factor prices and factor use in initial development and field production of natural rubber, Malaysian estates and smallholdings, 1922-95.

Malaysian estates Malaysian smallholdings Factor Prices 1922 1963 1978 1995 1964 1995

Land (S/ha) a 84 735 6,477 12,000 735 b Labor ($ per day) 0.39 1.47 3,82 8.50 ~ 1.47 b

Initial development d (%) (%) (%) (%) (%) Land e 3 (25) e'f 19 (220) 49 (1943) 47 (3600) na Labor 61 (472) 42 (482) 25 (959) 20 11558) na Management 23 (180) 15 (170) 6 (240) 6 c492) na Capital g 13 (98) 24 (273) 20 (800) 27 i2,050) na Total 100 (775) 100 (1145) 100 (3942) 100 (7700)

(days/ (days/ (days/ (days/ (days/ ha) ha) ha) ha) ha)

Person inputs 1203 362 251 183 na Field production 1964

(%) (%) (%) (%) (%) Land b 4 (4) 11 (37) 33 (324) 45 (600) 10 Labor 56 (59) 58 (208) 42 (409) } 47 (624) } 74 Management 23 (24) 10 (37) 10 (93)} } } Capital d 17 (18) 21 (76) 16 (155) 8 (107) 16 Total 100 (105) 100 (358) (981) 100 (1331) 100

(days/ (days/ (days/ (days/ (days/ ha) ha) ha) ha) ha)

Person inputs 151 141 107 59 191 (Yield,kg/ha 300 1,308 1,448 1,500 1,173

(38) (281)

(58) (377)

1 2 , 0 0 0 b 8.50 b

(%) 43 (3600) } 3- (2444) } 27 (2256) I00 (8300) (days/ ha) 235

(%) 46 (600) } 47 (6131

7 (90) 100 ( 1,303) (days/ ha) 72 900)

Sources: Large sample surveys of estate companies and smallholdings by agencies and the author in countries concerned. Detailed sources given by Barlow et al. (1994). ~Land prices are averages of ranges quoted in surveys for years shown. bActual opportunity costs of smallholding land and labor were frequently lower. CWage for local Malaysian labor. Migrant labor was often cheaper. dlnitial development by clearing, planting and maintenance to maturity (and tapping commencement) over a six-year gestation period. ¢Land costs are estimated as 5% interest on current values over the six-year gestation period. fFigures in parentheses are nominal costs in $ per hectare (at concurrent $US exchange rates with ringgits in which costs originally quoted). gAll purchased inputs other than labor and management, and including working capital.

almost 50% in 1995. Labor, in contrast, declined in share, where its substitution by capital inputs such as heavy land-clearing machines much reduced its part in initial development; thus the huge absolute labor inputs in this phase in 1922 had fallen to less than one-sixth by 1995. These and other changes such as the drop over recent decades in capital inputs in field production on estates and smallholdings also re- flected increased technical efficiency through learn- ing-by-doing. It is also interesting to note that large estate yields increased up to the 1960s, with little advance thereafter; this trend suggests diminishing returns in yield improvement, resembling most crops involved in the "Green Revolution" in Asia and elsewhere.

(c) Key elements in adjustment

The four elements of market conditions, technol-

ogies, institutional arrangements and government interventions are seen as crucially influencing rubber sector adjustments in the frame of the model just explored. They are now reviewed in turn.

(i) Market conditions The differing condit ions facing estates and

smallholdings growing rubber and other plantation crops are well portrayed by what Myint (1992, p. 26) terms "organizational dualism." This entails the coexistence of parallel economies, one which is "fully organized" economy with free f lowing linkages, and one which is "underdeveloped." The output, capital and labor markets of the latter are incomplete, with high transport, information and other transaction costs. They are also segmented, with differing prices facing participants in each portion. The underdeveloped economy likewise has a poorly funded bureaucracy, resulting in high charges

1602 WORLD DEVELOPMENT

for inferior services. The parallel economies are still closely connected, however, and far from the entire separation postulated by Boeke (1966).

Estates fall in the fully organized portion of this paradigm, where their large size provides economies of scale in accessing markets for outputs, capital, information and other inputs as well as in supplying infrastructures and services; these economies mean they are not subject to markets and facilities of surrounding rural areas. Smallholdings fall in the underdeveloped portion, and do not possess such economies. These differing market conditions cru- cially affect adjustments through technology adop- tion and other means.

Yet incomplete markets did not bar smallholders from adopting the early simple and almost exclu- sively land and labor-using tree crop growing technologies, given they became available through the estates. The slight necessary information flowed easily to individual farmers, with acceptance as usual starting with progressive individuals and spreading to others in a bandwagon effect. Little skill was entailed in planting or subsequent harvesting, and this together with high anticipated profitability explained the rapid diffusion of early smallholding plantations, is Smallholders in most countries had good access to output markets through ubiquitous local traders, who in Southeast Asia were largely Chinese. These traders also helped in distributing seed from estates.

But when it came to new skill and capital- intensive high-yielding technologies rapidly taken up by estates I9 in stage (3) of late agricultural transformation, acceptance by smallholders in all countries was virtually nonexistent. Farmers' infor- mation about complex methods and related activities was deficient, while long-term credit for sizeable purchased inputs was not available from trader- lenders uneasy about apparently risky ventures with deferred returns. Incentives were further reduced by high transportation charges, which raised costs of purchased items. As mentioned above, however, incomplete markets often meant prices of land and labor remained depressed, enabling old techniques to stay profitable; this was a positive short-term feature, but further delayed adoption of new technologies.

While such barriers to adoption were overcome by targeted government interventions in Malaysia, India, Thailand and other countries, they were ultimately circumvented even in contexts without official assistance. This occurred because technology spread effects eventually took place from estates and other centers to smallholdings, albeit facilitated by public infrastructures and services reducing transport costs and improving information. The process was encoraged by enhanced linkages and competition attending overall economic growth.

In Indonesia in the 1980s and 1990s, for example,

ex-estate workers and their children began 50--60 years after improved rubber and other plantation crops were introduced on estates to plant these on lands acquired around estate borders. They also set up private nurseries selling higher yielding stock to the huge majority of smallholders still using 1900s technology (Barlow, 1997). 20 These moves were facilitated by government road and education programs, and also encouraged by rapid national economic expansion (Table 3). Yet these delayed and usually poorly executed Indonesian adjustments were still small in effect, highlighting advantages of targeted interventions in other countries. It should nonetheless be recalled that Indonesia in the late 1990s is still only in stage (3) of the transformation (Table 1), and that by stages (4) and (5) of advanced economy its rural markets should be more integrated and competitive. Provided there are no constraints such as those on land in Malaysia, adjustments including conversions to newer crops will probably occur without much official assistance.

(ii) Technologies The role of this key element in adjustment is

explored through considering its use of resources. Viewing plantation technologies as isoquants utiliz- ing labor and capital, robust trees introduced in stage (2) of early agricultural transformation are suitably represented in Figure 1 as YoY0; this isoquant enables wide possible substitution between re- sources, permitting estates with cheap capital to operate at El and smallholdings to work at St. Such circumstances explain how different production systems used the same early technology.

The new technology introduced during stage (3) (Table 1) may be appropriately portrayed as Y1YI, allowing far less substitution between factors and

• E o

Y2" \\

\ ~ 1 Y ~ - - - Y 4 ~ ,~--.+_. ~ -,x~,, .So

"" .... Y3 ~ " ~ " YO

Labour

Figure 1. Capital and labor use by plantation technologies.

GROWTH, STRUCTURAL CHANGE AND PLANTATION TREE CROPS 1603

needing proportionately more capital; this denotes why adoption was so hard for smallholders. Yet subsequent production research continuing into stage (4) of early advanced economy generated a menu of technologies like Y2Y2 or Y3Y3, making possible choices of techniques more suitable for given smallholding (and estate) conditions. This later research thrust greatly facilitated adjustment.

It is useful to distinguish embodied and disembo- died technologies, where the former are "built into" long-term production inputs such as trees or proces- sing machines. Disembodied technologies are not so tied, however, comprising shorter term items such as fertilizers, herbicides, yield stimulants and modified factory routines; they may indeed be applied with embodied technologies at particular times, enabling factor substitution as well as yield and quality adjustments. Applying rubber yield stimulants with the embodied technology of improved trees repre- sented by Y2Y2, for example, extends the combined disembodied-embodied isoquant along the dotted line through Y~ in the direction of capital (Figure 1); this permits reduced labor use to secure a given output. Yield stimulants are still being profitably employed on remaining estate rubber areas in the advanced economy stage in Malaysia.

A major economic aspect of technology use is learning-by-doing, which was measured with Indo- nesian rubber smallholders. Thus farm records over 70 years show sizeable rises in technical efficiency, giving yield advances of 30% associated with large reductions in labor use (Barlow et al., 1994). If YoYo in Figure 1 is now taken as a frontier production function indicating best possible performance, these efficiency improvements involved a move like that from So to S1 for smallholdings and Eo to E1 for estates. Improved allocative efficiency would also have occurred, but was not quantified. While bigger learning gains were certainly secured with new technologies, these were hard to assess owing to the wide range of other inputs also involved. Such gains have been connected with many government-sponsored programs, constituting a vital feature of adjustment.

Most transferred technologies adopted by lateco- mers performed indifferently in new environments, with outcomes represented by the inferior position of isoquant Y4Y4 in Figure 1. Modifying research was needed to move isoquants to higher yielding sites such as Y1Y1, Y2Y2, and Y3Y3, usually concerning both embodied and disembodied technologies. Hence Chinese researchers cleverly improved rubber trees imported from other regions, and also devised cultivation methods to help cope with periodically windy, cold and dry conditions in Yunnan. They consequently made it possible to secure yields as high as those in more favored places (Huang and Pan, 1992). Similar adjustments were made elsewhere, being crucial to the viability of sectors concerned.

(iii) Institutional arrangements The different organizational arrangements of

estates and smallholdings have manifestly influenced economic performances of plantations at all stages of adjustment. These arrangements closely relate to divergent socioeconomic contexts and the dualism previously addressed.

Estates sustained marked scale economies in accessing inputs and outputs, as well as in initial development and output processing and handling. They secured cheap capital and information, ad- justed quickly to new techniques, and had low costs per unit in most operations. They likewise obtained vertical integration economies through membership of conglomerates with central selling organizations, giving good access to information about technolo- gies and buyers' requirements. They were insulated from incomplete markets facing smallholders, while at least during colonialism they often gained from political influences on governments. The latter frequently assisted early land acquisitions, making available areas which were comparatively level, fertile and close to points of export. Estate arrange- ments also had disadvantages, however, especially when their vertical integration under nationalized structures following independence incurred diseco- nomies owing to actions of bureaucracies and politicians.

Smallholdings had none of these estate merits, although their family basis conferred low interper- sonal monitoring charges. They benefited too through multiproduct orientation, where given plan- tation crops rarely occupied more than half the total value of their outputs; this hedged them against downward price fluctuations, helping explain char- acteristically high price elasticities of supply. More- over, smallholdings had to be viewed together with their network of buying traders, who usually offered competitive output markets embodying scale and vertical integration economies. These output markets still had relatively high transport costs, however, while input markets for information, capital, new technologies and associated inputs were all incom- plete as described.

A pertinent and fascinating question is why smallholding plantations have survived and in fact expanded beside estates with their economic super- iority in most spheres. One explanation lies in low opportunity costs of resources, where especially up to stage (4) family members often worked for returns under market wages. It is also because land has frequently had little actual value owing to legal and other restrictions on transfers beyond narrowly defined groups. Remoteness has likewise protected immediate economic viability, reinforcing market imperfections and preserving low land and labor costs. Smallholdings have further been helped by popular dislike of estates, which owing to associa-

1604 WORLD DEVELOPMENT

tions first with colonialism and later with modern elites have often been limited to areas alienated during early development or previously unoccupied lands.

Smallholdings in some countries have been greatly bolstered by the targeted government inter- ventions described above. These by overcoming incomplete markets and enabling adoption of better trees have allowed farmers to earn good incomes and later autonomously adjust to newer opportunities. Finally, the public control of many estate sectors following independence has reduced their economic performances, accordingly benefiting smallholdings. Although this control included subsidies and poli- tical support, it visibly undermined economic opera- tion and occasioned moves to reprivatization in the 1990s.

(iv) Government interventions These have crucially affected adjustment of

plantations, being usefully divided between micro and macro measures. The former comprise roads, other infrastructure and general education and health services on the one hand, and targeted programs including credit, extension and research on the other; all these components except the credit 21 are collectively consumed public goods, profitable for society as a whole but either not supplied or only produced at suboptimal levels by private firms. Their furnishing by government has aided smallholdings in overcoming incomplete markets and defective bu- reaucracies, with this occurring in many national contexts. Sometimes, however, targeted programs in particular have been ineffective owing to scarce funds, poor planning or persisting administrative

2 z deficiencies. This indeed has been true in Indone- sia, although infrastructure and service provisions in that country have been generally appropriate.

It is relevant to query whether private sector agents could have undertaken some adjustments more efficiently. Such agents largely failed in all smallholder input markets in stage (3) of late agricultural transformation, although sluggish pri- vate adaptations commenced after half a century in Indonesia. The author has indeed calculated highly positive rates of return to official Malaysian and Thai rubber improvement schemes 23, also indicating big gains compared to situations in countries with minor or non-existent public targeted programs. Govern- ment microinterventions accordingly seem to have been hugely helpful to adjustment, although con- tracting private agencies for certain tasks might have raised effectiveness in some instances.

Macro interventions including exchange rate, trade and price administration measures have also substantially influenced economic performance, but often negatively. Hence overvalued exchange rates disadvantaged rubber and traditional sectors in all

countries except Malaysia and Thailand in the 1950s and 1960s. The import substitution policies often associated with stage (3) also put upward pressures on rates and resource prices, with few apparent gains. The Indian and Chinese output price supports and input subsidies greatly boosted rubber outputs, but led to expensive crops of doubtful economic merit. The cold region rubber production of Yunnan, for example, is unlikely to be profitable at world price levels. Although Indian and Chinese policies were politically justified by self-sufficiency, they may now be abandoned under globally reducing protection combined with availability of cheaper imported rubbers; smallholders and others will then face painful adjustments, whose costs may well offset benefits received to date.

Within this scenario the far more open Thai and Malaysian macro regimes stand out as much assisting adjustments in rubber and other sectors. These regimes helped stimulate continuing high GDP growth in recent decades (Table 2), and despite some inconsistencies favored local prices reflecting global levels and impelling pertinent economic change. The strong export orientation of manufactur- ing sectors not only helped promote overall econom- ic adjustment, but also facilitated downstream rubber processing realizing comparative advantages in fabricating certain rubber goods.

4. CONCLUSIONS

The five-stage analytical framework of this paper is thought to assist appreciation of how economic growth affects plantation growers, and how appro- priate adjustments to change might be achieved. The scrutiny of natural rubber highlights the nature of processes involved, as well as policies likely to assist producers in effective responses. The discussion also indicates key elements critically influencing adjust- ments.

Market conditions including seriously incomplete input trade is seen as a large constraint on smallholdings needing to be addressed through interventions, while the further element of technol- ogy entails the generation and adoption of suitable cost-reducing and quality-improving techniques. The element of institutional arrangements of estates and smallholdings manifestly influences economic per- formance, where the size and vertical linkages of estates confer big advantages. But smallholdings, despite difficulties with input markets and technol- ogies, do have some merits, including low monitor- ing charges and access to competitive output markets with economies in handling their crops.

The final key element in adjustment of govern- ment interventions has major policy implications, warranting special concluding emphasis. Micro-level

GROWTH, STRUCTURAL CHANGE AND PLANTATION TREE CROPS 1605

help in overcoming difficulties o f smallholding input markets seems very necessary and economical ly advantageous, one aspect being the provision o f general infrastructures and services and the other the mount ing o f targeted programs for particular crops. Further, micro support is critical in generat ing technologies for estates and smallholdings, but this is the only need for the former which otherwise seem able to accomodate change without external help.

G o v e r n m e n t macro in te rven t ions have of ten obst ructed adjus tment , however , and the paper denotes gains from open trading policies compared to p ro tec t ive approaches . Thai and Malays ian

policies emerge as superior in this respect, with economic pricing of inputs and outputs having been helpful in stimulating posi t ive changes. It is no coinc idence that these countries have the most developed economies o f all tree crop producing countries.

Except ing Malaysia and Thailand, plantations in all countries have still to traverse all five stages of the analytical f ramework, and most have only reached late agricultural transformation. The fore- going analyses indicate relevant adjustment paths for what in most instances remain highly significant sectors of national economies .

NOTES

1. A "plantation tree crop" is understood to be one cultivated systematically in a plantation, as opposed to growing naturally in "native groves." Plantations can be established on family smallholdings of a few hectares or commercial estates with hired managers and workforces. The latter were privately owned in the early historical period, but are now often "public estates" run under government auspices. In China and Vietnam public estates are called "state farms."

2. Local farmers engaged in shifting cultivation rarely wanted work on estates in these beginning circumstances (Bauer, 1948), meaning laborers had to be imported from outside. This attitude of farmers sprang from objective functions emphasizing "subsistence affluence" and dislik- ing the regimented disease-ridden estate employment.

3. A technology may be defined as the body of knowledge concerning the production of a particular commodity from the appropriate resources. It is represen- table by a production function (Figure 1). Unique combinations of resources within the technology are referred to as "techniques."

4. Land prices also rise with increasing populations, while the latter tend to reduce wages of labor.

5. Most tree crops have gestation or "immature" periods of at least 3-5 years from initial planting to first harvesting. Certain crops have longer periods than this, and smallholder rubber in Indonesia and Nigeria commonly takes 10 years to come to tapping. The "mature" period during which crop is harvested is 15-25 years, making a total crop cycle of up to 35 years.

7. The spending and exchange rate effects may also be of "booming sector" origin, springing from exploitation of natural resources including petroleum. Such developments do not use much labor or land, and do not greatly modify prices of those factors.

8. It is interesting to note that output prices of traditional tree crops secured by countries at this stage may be affected by changes in elasticity of demand. Thus Lewis (1970) postulates a growing inelasticity, and consequent increases in output prices, as more cultivators of tropical commodities exit the market and hence reduce total global production. Those countries coming later to stage (5) may with such rising prices be able to offset greater costs, allowing production of commodities to continue longer. This depends on whether substitutes for these commodities are available, however.

9. Names of countries have frequently been altered over the years, but in this paper are referred to in 1990s terminology. Thailand was called Siam until 1950, while Indonesia was known as the Dutch East Indies until 1945. Myanmar was named Burma until the 1970s. Again, some earlier "Indonesian" developments were actually initiated by Dutch colonials, and the same applies to events in ex- British and French colonies.

10. No changes were abrupt, and other plantation crops in the three countries were actually introduced over half a century from the mid-1800s. Rubber was brought in during 1890-1910. A similar comment applies to later transitions defined in Table 1, although a decade rather than 50 years is relevant as leeway in these instances. The specified transition years refer to when large alterations first became evident.

6. Thus especially during the colonial period character- istically accompanying early development of plantation crops, smallholder sectors receive short shrift from colonial regimes (see, for example, Murray, 1980, on French lndochina). In contrast, estate sectors are strongly supported during this time, especially in allocating land. Sometimes lack of smallholder consideration is carried forward after independence, particularly when new regimes have no democratic structure.

11. Indonesian research into improved rubber actually began in the 1900s in the government botanic gardens at Bogor. The first breakthrough into high-yielding clones was made in 1915. But under the "laissez-faire" philosophy of the Dutch colonial government, most subsesquent research was undertaken by private estate enterprises.

12. Average real cumulative investment per smallholder rubber farmer in Indonesia is about one-fifth of levels in

1606 WORLD DEVELOPMENT

Malaysia, Sri Lanka, and Thailand, according to estimates made by the author.

13. These efforts were also connected with attempts to eliminate widespread communist insurgency in southern Thailand.

14. Malaysia is divided into the two main segments of the much more industrialized Peninsular Malaysia north of Singapore and East Malaysia which is part of the island of Borneo.

15. The distribution between producers and consumers of benefits from these improvements depended on relative price elasticities of supply and demand for rubber. Under the usual inelastic rubber supply and elastic demand, perhaps half the benefits accrued to consumers.

16. These were the "last straw" for many farmers, who were thereby encouraged to permanently leave their lands.

17. These "mini-estate" and other schemes were man- aged by the rural development bureaucracy, and while technically efficient were quite uneconomic. They entailed heavy subsidies and employment of migrant workers to produce rubber, proceeds from which were paid to original smallholder owners who were now shareholders. They resembled state-sponsored agricultural schemes in Korea, Taiwan and other countries well into the advanced economy stage, being further justified by support from politically powerful groups.

18. Acceptance of early simple tree crop technologies was eased by the fact that they fitted concurrent swidden systems, whereby land was cleared and planted with subsistence items. The seeds or seedlings of the technol- ogies were planted alongside subsistence items, involving minimal extra costs together with anticipations of sub- stantial extra returns.

19. The generation of technologies themselves usually required government intervention even with estates, and was justified by the public goods argument. The only exception was in the Dutch East Indies, where estates were constrained by government to mount two big cooperative research centers.

20. Unfortunately the quality of high-yielding materials was poor, and far below that of trees distributed through government-operated schemes in other countries. Small-

holders with their inferior information did not recognize this, however.

21. Government's provision of credit is basically justified on "infant industry" grounds that see the establishment of subsequently autonomous enterprises with important spread effect to other farmers.

22. The institutional arrangements of targeted govern- ment programs for smallholdings have likewise been critical, and may be seen in a global context as broadly occupying two categories. One is "dispersed," in the sense of arrangements making improved trees, credit and training available to all individual farmers wishing to take part. It is well illustrated by rubber extension in Thailand, which ingeniously and effectively meets farmers' planting needs. The other category is "focused," concentrating on selected cultivator groups which receive great financial support and management assistance. The latter is typified by "nucleus estate" projects operated by commercial enterprises, as well as comprehensive "land developments" administered directly by govemment; the latter developments include the work of the Federal Land Development Authority in Malaysia. Focused schemes attempt to capture economic advantages of scale and integration, and also frequently include welfare purposes in settling landless people. While nucleus programs have largely failed on both economic and social criteria, land developments have often been a success.International experiences with tree crop small- holders and other small farmers indicate that dispersed strategies are more economically effective than focused approaches, not least because they distribute benefits more widely (Tomich et al., 1995); immediate returns to intervention are higher, while significant spread effects occur in surrounding communities. But a mix of the categories is usually implemented owing to welfare and political considerations.

23. Calculations were made separately for research and planting support programs, taking into account flows of costs and benefits over 30-40 years. While all costs were accounted for, only "immediate" benefits to producers in terms of estimated values of higher yields were considered; substantial secondary gains flowing from farmer training and improved market integration were excluded. Returns from both Thai and Malaysian research programs over the 30-year period 1960--90 were both well over 40%, while the Thai planting support secured over 30%. The Malaysian planting support program earned some 25% up to 1975.

GROWTH, STRUCTURAL CHANGE AND PLANTATION TREE CROPS 1607

REFERENCES

Allen, G. C. and Donnithorne, A. (1954) Western Enterprise in Indonesia and Malaya. A Study in Economic Development. George Allen and Unwin, Ltd., London.

Barlow, C. (1997) The market for tree crop technology: A Sumatran case. Journal of Agricultural Economics 48(2), 193-210.

Barlow, C., Jayasuriya, S. K. and Tan, C. S. (1994) The World Rubber Industry. Routledge, London and New York.

Bauer, P. T. (1948) The Rubber Industry. A Study in Competition and Monopoly. Longman, Green and Co., Ltd., London.

Binswanger, H. P. and Ruttan, V. W. (1978) Induced Innovation, Technology, Institutions and Development. The Johns Hopkins University Press, Baltimore.

Boeke, J. H. (1966) Dualistic economics. In Indonesian Economics: The Concept of Dualism in Theory and Policy, ed. W. F. Wertheim, PP. 165-192. W. van Hoeve, The Hague.

Food and Agriculture Organization (1995) Production Yearbook. FAO, Rome.

Hayami, Y. and Kawagoe, T. (1993) The Agrarian Origins of Commerce and Industry. A Study of Peasant Market- ing in Indonesia. St Martin's Press, New York.

Huang, Z. D. and Pan, Y. Q. (1992) Rubber cultivation under climatic stresses in China. In Natural Rubber: Biology, Cultivation and Technology, eds. M. R. Sethuraj and N. M. Mathew. Elsevier, Amsterdam.

Lewis, W. A. (Ed.) (1970) Tropical Development, 1880- 1913. Studies in Economic Progress. George Allen and Unwin, London.

Murray, M. J. (1980) The Development of Capitalism in Colonial Indochina (1870-1940). University of Califor- nia Press, Berkeley.

Myint, H. (1954) An interpretation of economic back- wardness. Oxford Economic Papers 6, 132-162.

Myint, H. (1992) Organizational dualism and economic development. Asian Development Review, pp. 24-42.

Timmer, C. P. (1988) The agricultural transformation. In Handbook of Development Economics, eds. H. Chenery and T. N. Srinivasan, Vol. 1, pp. 276-331. North Holland, Amsterdam.

Tomich, T. P., Kilby, P. and Johnston, B. F. (1995) Transforming Agrarian Economies. Opportunities Seized, Opportunities Missed. Cornell University Press, Ithaca.

World Bank (1996) World Tables, 1996. Johns Hopkins University Press, Baltimore.