making negotiated land reform work: initial experience from colombia, brazil and south africa

Making Negotiated Land Reform Work: Initial

Experience from Colombia, Brazil and South Africa

KLAUS DEININGER *

The World Bank, Washington DC, USA

Summary. Ð The paper describes background, initial experience, and future challengesassociated with a new ``negotiated'' approach to land reform. This approach hasemerged as, following the end of the Cold War and broad macroeconomic adjustment,many countries face a ``second generation'' of reforms to address deep-rooted structuralproblems and provide the basis for sustainable poverty reduction and economic growth.It reviews possible theoretical linksÐthrough credit market or political channelsÐbe-tween asset ownership and economic performance. Program characteristics in eachcountry, as well as lessons for implementation, and implications for monitoring andimpact assessment are discussed. Ó 1999 Elsevier Science Ltd. All rights reserved.

1. INTRODUCTION

Theoretical reasons and empirical evidencesuggest that land reform may provide equityand e�ciency bene®ts. A large body of researchhas demonstrated the existence of a robustlynegative relationship between farm size andproductivity due to the supervision cost asso-ciated with employing hired labor. This impliesthat redistribution of land from wage-operatedlarge farms to family-operated smaller ones canincrease productivity (Binswanger, Deiningerand Feder, 1995). In addition, access to assetsin general and land ownership in particular isassociated with improved access to creditmarkets and can provide bene®ts as an insur-ance substitute to smooth consumption inter-temporally. By enabling the poor to undertakeindivisible productive investments (or by pre-venting them from irreversibly depleting theirasset-base) measures to improve the distribu-tion of assets could lead to higher aggregategrowth, thus improving both equity and e�-ciency (see Bardhan, Bowles and Gintis,forthcoming for references). Aggregate cross-country regressions as well as more micro-levelevidence con®rm the poverty-reducing andgrowth-enhancing impact of a better distribu-tion of productive assets.

This apparent potential notwithstanding,actual experience with land reform has in manyinstances fallen short of expectations. De-spiteÐor becauseÐof this, land reform re-

mains a hotly debated issue in a number ofcountries (e.g., Zimbabwe, Malawi, South Af-rica, Guatemala, El Salvador, Brazil, Colom-bia) some of which are spending considerableamounts of resources for this purpose. Amechanism to provide an e�ciency- and equity-enhancing redistribution of assets that wouldincrease overall investment at a cost that iscomparable to other types of government in-terventions would be very desirable.

This paper describes a new type of negotiatedland reform that relies on voluntary landtransfers based on negotiation between buyersand sellers, where the government's role is re-stricted to establishing the necessary frame-work and making available a land purchase

World Development Vol. 27, No. 4, pp. 651±672, 1999Ó 1999 Elsevier Science Ltd

All rights reserved. Printed in Great Britain0305-750X/99 $ ± see front matter

PII: S0305-750X(99)00023-6

* I would like to thank Hans Binswanger, Juliana

Bottia, Alain de Janvry, Gershon Feder, Gustavo

Gordillo de Anda, Adriana Herrera, John Heath, Nick

Kra�t, Marcos Lins, Michael Lipton, Anibal Llano,

Absalon Machado, Indran Naidoo, Pedro Olinto,

Manuel Rojas, Edson Teo®lo, Dina Umali-Deininger,

Hernando Urbina, Martien van Nieuwkoop, Stefan

Oehrlein, two anonymous reviewers, and seminar par-

ticipants in Helsinki, Santiago, Bogota, the University of

Sussex, and Washington, for valuable insights and

discussions. The opinions expressed in this paper are

those of the author and do not necessarily represent the

views of the World Bank, its Board of Directors, or the

countries they represent. Final revision accepted: 21

September 1998.

651

grant to eligible bene®ciaries. Section 2 dis-cusses the land reform experience in general,thus providing the historical background andconceptual basis for the subsequent argument.Section 3 is devoted to a more detailed de-scription of negotiated land reform in Colom-bia, covering the reasons for choosing anegotiated approach, its principles, and its im-plementation in a number of pilot municipios.Section 4 compares the mechanisms utilized tothose adopted in Brazil and South Africa andbrie¯y highlights some of the implications formonitoring the new approach. Section 5 con-cludes.

2. LAND REFORM: POTENTIAL ANDHISTORICAL EXPERIENCE

While land reform has long been the subjectof policy debate, the motivation for addressingland issues has shifted considerably over time.Initial discussions were largely motivated bypolitical considerations, supported by the no-tion of the inverse farm-size productivity rela-tionship. This notion had received empiricalsupport from cross-sectional studies acrossstates (Kutcher and Scandizzo, 1981), countries(Berry and Cline, 1979), or individual farms(Barraclough, 1970).1 More recent contribu-tions have emphasized the importance of assetownership in situations characterized by in-complete contracting (Bardhan, Bowles andGintis, forthcoming; Ho� and Lyon, 1994),elaborating on models such as Dasgupta andRay (1986, 1987) and Moene (1992). The un-derlying idea is that, in situations characterizedby credit rationing, individuals may not be ableto undertake indivisible investments in humancapital (schooling) or productive assets (wells,bullocks, or perennials with a long gestationperiod) that need to be ®nanced through credit.This idea has been formalized in a number oftheoretical models where lack of collateralkeeps individuals in ``poverty traps,'' unable toundertake highly pro®table indivisible invest-ments (Galor and Zeira, 1993; Eckstein andZilcha, 1994).2

In such a setting, the poor would fail to getout of poverty, not because they are inherentlyless productive or lack the necessary skills, butbecause informational imperfections precludethem from access to credit markets and be-cause, as a consequence, they never get theopportunity to utilize or develop fully theirabilities. If this is true, a one-o� redistribution

to low-wealth groups could strictly dominateother policy instruments (Ho� and Lyon,1994). In particular it will, in the medium tolong term, more e�ective and less costly thancontinuing redistribution of income (e.g.,through social programs) which would be as-sociated with strong disincentive e�ects (Ba-narjee and Newman, 1993; Mookherjee, 1997).

Support for the importance of asset (and in-come) distribution for economic outcomes isprovided by cross-country regressions whichindicate the presence of a signi®cant negativeimpact of the initial asset distribution on sub-sequent economic growth (Birdsall and Lon-donÄo, 1997). This growth-reducing impact isparticularly severe for the poor (Deininger andSquire, 1998). Indeed, overall inequality seemsto have an important impact on societies'ability to e�ectively and quickly respond toexogenous shocks (Rodrik, 1998), the level ofcrime (Fajnzylber, Lederman and Loayza,1998), and the degree to which special interestgroups are able to appropriate rentsÐwithimplications for overall productive e�ciency(Banarjee et al., 1997).

Although inequality may signi®cantly a�ecteconomic performance, historical examples formajor changes of inequality within a countryare rare (Li, Squire and Zou, 1998). Morespeci®cally, the success of land reform wascritically dependent on the form of productioninto which it was introduced. In landlord estateswhere tenants already cultivated the land andall that was required was a reassignment ofproperty rights, land reform was relativelystraightforward and associated with signi®cantproductivity increases and the emergence ofstable systems of production.3 The main reasonis that the organization of production remainedthe same family farm system, and that bene®-ciaries already had the skills and implementsnecessary to cultivate their ®elds. Organiza-tional requirements of conducting such landreforms were minimalÐmaking them compa-rable to the ``stroke of a pen'' reforms familiarfrom the literature on macroeconomic reform.Indeed, since the end of WW II, landlord es-tates in Bolivia, large areas of China, EasternIndia, Ethiopia, Iran, Japan, Korea, and Tai-wan have been transferred to tenants in thecourse of successful land reforms.4

By contrast, land reform in haciendas, i.e.systems where tenants had a small house-plotfor subsistence but worked the majority of thetime on the landlord's home farm, has beenvery di�cult, up to the point where the ``game

652 WORLD DEVELOPMENT

of Latin American Land Reform'' was declaredto be lost (De Janvry and Sadoulet, 1989). Inthe large majority of cases large landownersresponded to the threat of land reform withlarge-scale evictions long before governmentswere able to e�ectively implement laws aimedat tenant protection or land reform. They eitherresumed extensive livestock production andranching orÐaided by signi®cant credit subsi-diesÐstarted highly mechanized self-cultiva-tion (Binswanger, Deininger and Feder, 1995).This reduced tenant welfare, depopulatedfarms, and created further di�culties for re-distributive land reform. The experience withland reform in these environments points to-ward three speci®c di�culties:

First, the transfer from large to small farmersrequires a change in the pattern of production,construction of complementary infrastructure,subdivision of the farm, and settlement of ad-ditional bene®ciaries over and above theworkers who have already been living on thefarm.5 Farms acquired for purposes of landreform have generally not been farmed at fullcapacity, were run down or decapitalized, orhighly mechanized. In all of these cases failureto bring in additional bene®ciaries, to provideresources for simple works (cleaning of pas-tures, fencing, construction of basic infra-structure, etc.) during the startup phase, and toensure the availability of productive assets andtechnical assistance to go with the land haveoften contributed to the failure of reform ef-forts.6

Second, land reform bene®ciaries, even ifthey are workers of the former farm, are rarelyaccustomed to making independent entrepre-neurial decisions, a constraint that is particu-larly important if (as in many of the casesdescribed above) realization of the potentialbene®ts from land reform requires signi®cantmodi®cations in the farm's cropping pattern.Programs that are limited to the mere transferof land, without training and technical assis-tance, have made it di�cult for bene®ciaries toreach quickly an equilibrium characterized byhigh levels of productivity and savings and, tothe degree that bene®ciaries were not able to getaccess to these, may have resulted in a perma-nent decrease in agricultural productivity.

Third, in rural environments with multiplemarket imperfections, providing bene®ciarieswith access to land but not with access tomarkets for output and credit may fail to makethem better o� than before. This will be the caseparticularly if landlords had provided their la-

bor tenants with inputs, credit, or market out-lets before the reform.7 Land markets thereforehave to be viewed in the context of the opera-tion of other factor markets.

These generic land reform di�culties were, inpast attempts to e�ect the redistribution of landto the poor, often exacerbated by implementa-tion-related issues. Instead of aiming to in-crease productivity and reduce poverty, themain goal of many land reforms in the past hasbeen to calm social unrest and allay politicalpressures by peasant organizations.8 Such re-forms had often been initiated in response topolitical pressure (or to divert attention fromother problems) rather than as part of a long-term rural development strategy.9 The resultingreform measures were generally designed adhoc, bore little relation to actual needs on theground, and commitment to them faltered oncesocial emergencies had subsided. Furthermore,attention often focused on the politically vocaland well-connected peasants rather than ruraldwellers with the best ability to make produc-tive use of the land, or the most deserving onpoverty grounds.10

The costs of carrying out land reform wereoften increased by the continued existence ofimplicit and explicit distortions which droveland prices above the capitalized value of ag-ricultural pro®ts and made it attractive for landreform bene®ciaries to sell out to large farmers,thus contributing to reconstruction on hold-ings.11 In addition, instead of aiming to createconditions that would improve the functioningof land rental and sales markets to complementstate-led reform e�orts, governments have of-ten completely outlawed or severely restrictedthe operation of land rental (and to a lesserdegree sales) markets. This has eliminated animportant opportunity for landless individualsto acquire farming experience, made the pro-gress of land reform totally dependent on bu-reaucratic e�orts, and complicated the task oftargeting assistance to the poor. This completereliance on government spawned complex reg-ulations and cumbersome bureaucratic re-quirements to implement land reform, stretchedavailable administrative capacity (Lipton,1974), and resulted in highly centralized pro-cesses of implementation. Government bu-reaucracies at the central levelÐjusti®ed by theneed to provide technical assistance and othersupport services to bene®ciariesÐproved ex-pensive and, unable to utilize information fromthe local level, often also quite ine�ective. (seeTable 1).

MAKING NEGOTIATED LAND REFORM WORK 653

Tab

le1

.C

om

pa

riso

no

fm

ech

anis

ms

toim

ple

men

tm

ark

etass

iste

dla

nd

refo

rmin

Colo

mbia

,B

razi

l,and

South

Afr

ica

Co

lom

bia

Bra

zil

So

uth

Afr

ica

INC

OR

AN

ego

tia

ted

INC

RA

Neg

oti

ate

dP

ilo

ts

Lan

dse

lect

ion

Sel

ecti

on

by

INC

OR

Ad

efa

cto

rb

ase

do

np

oli

tica

lp

ress

ure

;C

ost

of

$1

80

00

±2

20

00

per

fam

ily.

By

ben

e®ci

ari

es;

wit

hm

ean

sto

incr

ease

tran

s-p

are

ncy

an

dp

rovid

ete

chn

ica

la

ssis

tan

ce.

Pu

rch

ase

or

exp

rop

riati

on

;aver

age

cost

of

$11

600

per

fam

ily;

main

lyle

gali

zati

on

of

occ

up

ied

lan

ds.

Neg

oti

ate

db

yco

mm

un

ity;

wil

lin

gse

ller

(in

clu

din

gb

an

ks)

-w

illi

ng

bu

yer

;ex

pec

ted

cost

$3

000

per

ben

e®ci

ary

.

Co

mm

un

ity

init

iati

ve.

Lan

d®

na

nci

ng

70

%o

fla

nd

va

lue

(up

to$

22

00

0)

as

gra

nt

(20

%ca

sh;

50

%b

on

ds)

;3

0%

thro

ugh

Ca

jaA

gra

ria

cred

it(l

eng

thy

del

ay

s).

Co

mm

erci

al

ban

kad

-m

inis

ters

gra

nt

reso

urc

esa

nd

pro

vid

esa

dd

itio

nal

cred

itfo

rla

nd

pu

rch

ase

an

dw

ork

ing

cap

ital.

TD

As

for

un

imp

roved

lan

dan

dca

shfo

rim

pro

vem

ents

an

dcr

op

s;b

ene®

ciari

esin

theo

ryex

pec

ted

top

ay

back

;n

ot

en-

forc

ed.

Lo

an

toap

pro

ved

ben

e®ci

ari

esfr

om

aco

mm

erci

al

ban

k(c

on

-si

der

ab

lesu

bsi

dy

elem

ent)

.

Max.

gra

nt

of

R15

000

for

lan

dp

urc

hase

;se

para

tep

lan

nin

ggra

nt.

Ben

e®ci

ary

se-

lect

ion

Po

int

sch

eme

for

soci

al

nee

da

nd

ag

ricu

ltu

ral

ex-

per

ien

ce;

inp

ract

ice

ad

-h

oc

sele

ctio

nb

ase

do

nin

div

idu

al

farm

s.

Co

mp

reh

ensi

ve

regis

tra-

tio

n;

an

dp

re-s

elec

tio

nb

ase

do

nso

cia

lcr

iter

ia.

Fin

al

sele

ctio

nb

ase

do

np

rod

uct

ive

pro

ject

s.

Th

rou

gh

INC

RA

base

do

nex

am

inati

on

of

agri

c.k

no

wle

dge;

inp

ract

ice

alm

ost

all

are

regu

-la

rize

dsq

uatt

ers.

Sel

f-se

lect

ion

of

ben

e®ci

ari

es;

clea

ran

ceo

fp

rice

an

dti

tle

by

Sta

teL

an

dIn

stit

ute

;d

ecen

tral-

ized

ap

pro

val.

Occ

up

ied

lan

ds

inel

igib

le.

Sel

f-se

lect

ion

of

ben

e®ci

a-

ries

sub

ject

tom

axim

um

inco

me

crit

erio

n(<

R1

500

per

mo

nth

).

Farm

ing

pro

ject

de®

nit

ion

Per

ceiv

edto

be

nec

es-

sary

on

lyto

ob

tain

ing

ab

an

klo

an

.

Key

issu

efo

rse

lect

ion

;d

i�er

ent

form

so

fte

chn

i-ca

la

ssis

tan

ceavail

ab

le.

Fa

rmm

od

els

avail

ab

leat

mu

nic

ipa

lle

vel

.

No

spec

i®c

arr

an

gem

ents

.U

pto

8%

of

pro

ject

va

lue

avail

ab

lefo

rte

chn

ical

ass

is-

tan

cein

pro

ject

pre

para

tio

nan

dim

ple

men

tati

on

.F

arm

mo

del

sel

ab

ora

ted

at

state

level

.

Pro

vin

cial

pla

nis

ap

re-

con

dit

ion

bu

tfe

wsp

eci®

cgu

idel

ines

pro

vid

edan

dn

ofa

rmm

od

els

are

elab

ora

ted

.

Oth

er®

na

nci

ng

Cre

dit

for

lan

da

nd

wo

rkin

gca

pit

al

isth

eb

igb

ott

len

eck

cau

sin

gim

ple

men

tati

on

del

ay

s.

Ind

epen

den

t®

nan

cial

in-

stit

uti

on

sto

pro

vid

ein

-te

gra

ted

cred

itfo

rth

ew

ho

lep

roje

ct.

Cre

dit

of

up

to$1150

(aver

age

$610)

for

foo

dan

dh

ou

sin

gan

d$7

500

(aver

age

$4

500)

for

wo

rkin

gca

pit

al.

70%

sub

sid

yel

emen

t;m

inim

al

cost

reco

ver

y.

Acc

ess

toP

RO

CE

RA

cred

itli

ke

oth

erla

nd

refo

rmb

ene-

®ci

ari

es.

Res

po

nsi

bil

ity

of

ben

e®ci

a-

ries

.

O�

-fa

rmin

ves

t-m

ents

Co

mp

lex

set

of

inte

rin

-st

itu

tio

na

lco

ord

ina

tio

nw

ith

litt

lere

sult

su

pto

no

w.

Iden

ti®

eda

nd

cost

edin

mu

nic

ipa

lla

nd

refo

rmp

lan

.

Pro

vid

edb

yIN

CR

A($

3200

in1994,

no

wu

pto

$8000);

alm

ost

all

for

road

s.

Gra

nt

of

$4

000

per

ben

e®-

ciary

;d

isb

urs

edd

irec

tly

toco

mm

un

ity.

Ser

vic

esan

dto

som

eex

ten

tin

frast

ruct

ure

are

pro

vin

-ci

al

resp

on

sib

ilit

y;

coo

rdi-

nati

on

wit

hth

ece

nte

ris

stil

lw

eak

.


3. NEGOTIATED LAND REFORM INCOLOMBIA

(a) Background

(i) Land reform before 1994

In Colombia, land reform has been a long-standing concern to correct an extremely ineq-uitable distribution of land, to increase theproductivity and environmental sustainabilityof agricultural production, and to reducewidespread rural violence. Maldistribution ofland in rural areas, while dating back to theencomiendas given out following the Spanishconquest, has been reinforced and exacerbatedin more recent times by a number of policyrelated factors.12 These include:

Ð Tax incentives for agriculture that impliedthat rich individuals acquired land in order too�set taxes on nonagricultural enterprises.Ð Legal impediments to the smooth func-tioning of the land rental and sales markets.Share tenancy was either directly outlawedor, when this was lifted, discouraged by thefact that tenants would receive propertyrights to whatever land improvements theyhad made, making it in principle impossibleto terminate their leases.Ð Credit and interest rate subsidies plus dis-proportionate protection of the livestock sub-sector provided incentives for agriculturalcultivation with very low labor intensity(World Bank, 1996).Ð The use of land to launder money that hadbeen acquired by drug lords.

These factors have profound implications forfactor use, employment generation, and welfarein rural areas. First, while small farmers wereoften driven o� their traditional lands to ekeout a living in marginal and environmentallyfragile areas, much of the best agricultural land(75% of the land suitable for crop production)was devoted to extensive livestock grazing ornot farmed at all due to violence (Heath andBinswanger, 1996). This suggests that there areindeed large tracts of unutilized or underuti-lized land which could be subjected to landreform in order to increase agricultural pro-ductivityÐa notion in line with available em-pirical evidence.13

Second, economic growth has been labor-saving. Since the 1950s, the rate of growth ofrural employment has been signi®cantly lower

than aggregate economic growth, which issurprising even by the standard of other LatinAmerican countries (Mision Social, 1990). Thisappears to have increased peasants' inclinationto support, or at least live with exceptionallyhigh levels of rural violence that increasinglyconstitute a drag on the whole economy (esti-mates in the Colombian press put the lossesassociated with rural violence at about 15% ofGDP). The government sees the reduction ofrural violence as an important goal of land re-form.

Third, structural adjustment made the lack ofadaptability in the large farm sector particularlyblatant. Elimination of credit subsidies caughtlarge mechanized farms that cultivated mainlytraditional crops with minimal labor inputs in adebt-trap that made them unable to adjust tothe new environment and take advantage of theopportunities for exports of nontraditional andmore labor-intensive crops. Unable to respondto the loss of agricultural protection in a pro-ductive way, the large farm sector resorted tolarge-scale lobbying. Establishment of a dy-namic small farm sector would, it was hoped,enable Colombia to capitalize on its agro-eco-logical diversity and signi®cantly increase itsexports of traditional and nontraditional crops.

None of these concerns are new. The mal-distribution of productive resources, especiallyland, was identi®ed as one of the root causes ofeconomic stagnation by a World Bank missionin the 1950s. In 1961, the government estab-lished the National Land Reform Institute(Instituto Nacional Colombiano de ReformaAgraria or INCORA), to bring about a moreequitable distribution of assets in the ruraleconomy. But even though considerableamounts of resources were spent on land re-form (INCORA's average annual budget in thelate 1980s was about US $140 million), mostwas spent on a large bureaucracy14 and almost35 years of operations had produced little vis-ible e�ect on the ground. INCORA appeared tobe more e�ective in regularizing spontaneoussettlement on the frontier than in convertingthe landless into successful agricultural entre-preneurs in areas that were previously culti-vated by large owners. Even where land reformdid distribute land, lack of capital forced manybene®ciaries to abandon full-time agricultureand rent out part or all of their land, often tothe old landlord. In the aggregate, between the1960s and 1990, the Gini coe�cient of the op-erational land distribution fell by only threepercentage points, from 0.87 to 0.84.


(ii) The new Law and its implementation

Not unrelated to the loss of INCORA's tra-ditional source of ®nanceÐa share of duties onagricultural imports that was eliminated withagricultural trade liberalizationÐa law waspassed in 1994 that would allow for a moredecentralized and demand-driven process. De-spite favorable preconditions, however, and thegovernment's expressed determination to dis-tribute one million hectares within four years,the land reform program had a disappointinglyslow start.15 Before describing the new imple-mentation arrangements that emerged to im-prove on this, it is worthwhile to consider themain reasons for this lackluster performanceand the measures taken to address this issue.

To overcome the ``fundamental ®nancingproblem of the poor'' (Binswanger and Elgin,1988;Carter and Mesbah, 1993), i.e. the factthat fully mortgage-based ®nancing of landpurchases by the poor is infeasible, the Co-lombian Land Reform Law provides for a landpurchase grant. The grants amounts to 70% ofthe negotiated land purchase price, up to amaximum that was based on historical landreform allocations.16 The grant, however, wasrestricted to the purchases of land and couldnot be used to undertake complementary in-vestments. This created incentives for collusionbetween sellers and buyers to overstate landprices, divide the surplus between them, and letthe government foot the bill.17 The resultingincentive structure was strongly biased in favorof the transfer of developed agricultural landclose to infrastructure and already well en-dowed with the necessary complementary in-vestment. This tended to reduce land reform toa mere redistribution of existing assets ratherthan the creation of new ones, by targetingunderutilized lands and helping bene®ciaries toundertake signi®cant investments. To deal withthese issues, it was clari®ed that the goal ofmarket assisted land reform is the establish-ment of viable productive projects (proyectosproductivos), rather than the mere transfer ofland. A mechanism was devised to facilitate useof grant funds to ®nance nonland investments,thus overcoming the bias inherent in previouslegal provisions.18

A second issue was that, to create viable ag-ricultural enterprises, rather than a ``ruralproletariat,'' a target income from full-timeagriculture (equivalent to a minimum farm sizeof about 15 hectares) was legally required. This

neglected the potential of the poorÐespeciallythose in proximity to urban areasÐto deriveincome from a variety of sources and togetherwith a prohibition of rental and sale, left littleroom either for the exit of unsuccessful bene-®ciaries or the gradual expansion of the hold-ings of successful ones through rental orpurchase of additional land. It also demon-strated little awareness of the requirements, interms of human capital, other assets, and ex-perience with ®nancial and marketing institu-tions, associated with operating a 15-hectarefarm.19 As a result, the law was in danger ofconcentrating large amounts of subsidies on awell-connected ``agrarian bourgeoisie'' whileleaving the majority of potential bene®ciariesuncovered.20 To overcome this shortcoming,the target income was reduced by one-thirdand, rather than being based on general aver-ages, is to be assessed based on a project-spe-ci®c plan elaborated by bene®ciaries thatincluded income from nonagricultural sources.

Finally, even though the law provides for anexemplary and elaborate institutional struc-ture21 to facilitate an encompassing process ofreform, the fact that there was little incentivefor local leaders to actually establish the nec-essary structures implied that it was di�cult tomake the model operational and ensure e�ec-tive bene®ciary participation. Lack of dissemi-nation of the law prevented a truly democraticprocess at the local level. Continued subsidi-zation of INCORA drove out private serviceproviders22 and ``success'' continued to be de-®ned in terms of transferring land and ex-hausting budgets rather than in establishingviable rural enterprises. To change this, a shiftof responsibility for approval from INCORA'sheadquarters to regional o�ces was accompa-nied by transferring resources directly to localcommunities and by clarifying that, amongothers, existence and functioning of a munici-pal council was a precondition for municipiosto become eligible for land reform funds.

(b) Implementation

While the ®rst two years of program imple-mentation highlighted critical shortcomings,they provided little insight in how to actuallyimplement such a program. In this section weuse the experience from ®ve pilot municipios,selected to re¯ect the heterogeneity of thecountry,23 to illustrate four key elements forimplementation namely


Ð making land reform a program ``owned''by the local government, thus achieving bet-ter integration of land reform with existingmunicipal development priorities or invest-ments and at the same time greatly reducingtransaction costsÐ requiring the elaboration of productiveprojects that provide the basis for a moreprogrammatic approach to bene®ciary train-ing, negotiation of land prices, and an eco-nomic evaluation of the expected bene®tsand costs of land reformÐ establishing a decentralized and ``handson'' program of bene®ciary training thatwould act as a means of pre-selecting bene®-ciaries (based on their willingness to partici-pate), help them to overcome their structuraldi�culties, ensure greater ``ownership'' andultimately economic sustainability of projectsÐ insisting on a transparent and public pro-cess of project approval and linking the need-ed mechanisms of accountability directly to aprocess of monitoring and evaluation thatlinks to the municipal land reform plan, aimsto detect quickly deviations from targets andthe reasons for them, and forms the basis foran in-depth impact assessment of the land re-form process at a later stage.

The expectation is that the steps taken by localgovernments to improve infrastructure andfunctioning of other factor markets, togetherwith land-reform speci®c measures (describedin detail below), would improve potential ben-e®ciaries' capacity to negotiate and make pro-ductive use of land and at the same time reducethe gap between the net present value of agri-cultural pro®ts land prices. This, in turn, wouldreduce the size of the land purchase grant re-quired per person, thus making it possible touse a given amount of grant money to attend toa larger number of bene®ciaries.

(i) The municipal land reform plan

A key document in the pilot municipios hasbeen a municipal land reform plan. This plan,elaborated in a decentralized fashion, containsinformation about demand and supply of landfor land reform purposes, and a characteriza-tion of the institutional environment and re-sponsibilities for land reform. Following asystematic procedure to establish a municipalplan is expected to have three main bene®ts.

The ®rst is to identify potential demand forthis type of land reform. This includes steps

such as raising awareness among the bene®-ciary population and help target the most nee-dy, to establish a transparent process that canground land reform ®rmly within the context ofother local development initiatives, to identifythe potential demand for land reform, and todevelop realistic expectations about the extentto which land reform can contribute to thesolution of existing problems.

A second bene®t is to identify potential supplyand to generate the basis for reasonably com-petitive land markets by ensuring that supply ofland (at reasonable prices and in areas suitablefor small farmer cultivation) exceeds demand.

A ®nal issue is the establishmentÐat the lo-cal levelÐof the institutional infrastructureneeded for e�ective implementation of landreform. To ensure sustainability of land reformprojects, it has proven to be critical to identifynongovernment organizations (NGOs) who areable to provide continuing technical assistance,and ®nancial institutions who are in a positionto extend and e�ectively supervise credit to landreform bene®ciaries. These elements, togetherwith information on the contributions expectedfrom di�erent participants (bene®ciaries, cen-tral government, local institutions), makes itmuch easier for local authorities to elaborate acoordinated program of land reform that is inline with the speci®c needs and opportunities,including the ®scal capacity, of the municipio.

(ii) Identi®cation of potential bene®ciaries

Under the process followed before initiationof the pilots, selection of bene®ciaries was oftenarbitrary and ad hoc. Despite the regulations ofthe new law, INCORA continued to selectbene®ciaries on a case by case basis once agiven farm had been put up for sale and centralapproval for the release of the necessary landpurchase funds had been obtained. In thesecases, to be able to disburse funds quickly, salesof farms were often quite secretive, despite theexistence (on paper) of a needs-based quali®-cation system. The selection committees thatwere established included workers of the exist-ing farm who were generally careful not toadmit too many contenders from outside.24

To ensure participation beyond the mem-bership of well-established campesino organi-zations and a transparent and more competitivemarket for land, this approach has, in the pilotmunicipios, been replaced by a procedurethat aims to create the basis for land transac-tions through a more competitive market. To


identify potential demand, a systematic infor-mation campaign to disseminate the law, withsubsequent inscription of potential land reformbene®ciaries (aspirantes) in a registry to bemaintained by INCORA, is conductedthroughout the municipio. A questionnaireprovides basic information on bene®ciaries'educational level, their agricultural experience(if any), their income sources, and their accessto other types of government services such aseducation or health. Based on this, a pre-quali®cation, essentially a means test based onassets, is conducted.

Experience indicates that enabling potentialbene®ciaries to register at public o�ces andpolice stations in outlying villages has consid-erably broadened the outreach of the program.Contrary to the procedures followed earlier byINCORA, the information supplied is checkedfor consistency, resulting in the elimination of alarge number of non-quali®ed applications. Thenames of rejected and accepted aspirantes (withreasons for rejection) are posted publicly. Thepublicity of the selection process seems to haveincreased accountability, and facilitated a betterunderstanding of the scope and limitations ofland reform by local authorities and potentialbene®ciaries. In all of the pilot municipios, al-ternative programs have been initiated to(temporarily or permanently) take care of thespeci®c needs of groups who will not be able tobene®t from land reform in the immediate fu-ture. These programs include chicken hatcheriesand other microenterprises for female house-hold heads, construction of rural roads underseasonal food for work schemes, as well as re-forestation of environmentally fragile zones.

In addition to generating awareness for theprogram, its potential target group, and thecharacteristics of demand, the process of ben-e®ciary selection also provides a basis for localauthorities to integrate land reform into abroader program of capacity building and so-cial assistance at the municipal level. This couldcontribute to the resolution of at least someaspects of the potential con¯ict between thedual objectives of equity and e�ciency which isto some extent unavoidable if land reform is tomake a long-term sustainable contribution topoverty reduction.

(iii) Creating the basis for a functioning landmarket

The availability of large amounts of unuti-lized or underutilized land in large holdings

implies that, in a reasonably ¯uid land marketthere would be plenty of supply to enable po-tential bene®ciaries to choose the most suitablelands and negotiate to obtain a competitiveprice. In practice, however, land markets havefound to be thin, highly segmented, character-ized by high transaction costs, and often pu-shed into informality (FAO, 1994). Creditmarket imperfections, lack of market informa-tion by potential sellers, and the non-existenceof farm models suited to the speci®c needs andfactor endowments of small agricultural pro-ducers, have prevented such an outcome andcontributed to the fact that bene®ciaries underthe old-style reform program often acquiredmarginal lands at highly exaggerated priceswithout being able to make productive use ofit.25

In the pilot municipios, a procedure similar tothe identi®cation of demand for land is alsobeing followed on the supply side. The ®rst stepis to determine ecologically suitable zones and,based on cadastral information, establish aninventory of the land according to size classi-®cation that could be used to identify targetareas for agrarian reform. Areas where landreform would result in environmental hazards,where soil fertility is insu�cient, or where theexisting ownership structure is already charac-terized by small to medium-sized holdings, arethus eliminated a priori. This gives bene®ciariesa better idea of where to focus their e�orts,helps to set realistic goals, and puts into per-spective the potential contribution of land re-form for solving the social problems of a givenmunicipio. It increases not only awareness ofthe scope for land reform as compared to otheroptions aiming at overall development of themunicipio (and the degree to which the successof land reform will depend on such comple-mentary measures), but also forces local gov-ernments to think about potential levers (fromland taxes to land price information systemsand training to increase the productive capacityof potential bene®ciaries) that they can utilizeto achieve an process of ``integrated land mar-ket development''.26 Experience from the pilotsindicates that land that had traditionally beeno�ered to INCORA for land reform was oftenof marginal quality and hardly suitable for landreform while some of the best land continued tolie idle or underutilized. Speci®c measures toresolve this issue have included:

Ð Increasing sellers' awareness of the scopeand potential for alternative forms, such as


land rental that could temporarily or perma-nently provide potential bene®ciaries with ac-cess to land increases e�ciency and may serveas a springboard for the landless to acquirethe information and agricultural experiencenecessary to put together a productive pro-ject.Ð Encouraging more e�ective collection ofexisting municipal land taxesÐa strategy inline with the central government's desire toincrease the revenue base of local govern-ments and to gradually reduce the need forcentral transfers.Ð More e�ective and systematic dissemina-tion of information not only among potentialbuyers but also any sellers of land, speci®callyinformation on the mechanisms of market-as-sisted land reform and the modalities of pay-ment under this program.

There is justi®ed concern that a program thatprovides targeted support to land purchasesmay contribute to an increase in land prices,and therefore bene®t former landowners ratherthan the poor farmers who receive the land. Todeal with this, it has been decided that, to beeligible for land reform bene®ts, municipiosneed to provide evidence (using actual sale in-scriptions/o�ers by landowners) that existingland supply is at least three times the amount ofland to be transacted under the land reformprogram.

(iv) Establishing the Institutional basis

Experience suggests that, in the absence oftechnical support during the startup phase andwithout access to markets for ®nance and out-puts, the sustainability of newly initiated landreform settlements will be limited. The munic-ipal plan thus contains a list of quali®ed pro-viders of technical assistance from whichpotential bene®ciaries can choose one to use thepart of the land purchase grant that is ear-marked for technical assistance. In addition, itaims to identify ®nancial institutions thatwould be willing to lend to land reform bene-®ciaries. The rationale for this is simple: It islikely to be futile to initiate a large process ofland reform in a municipio where neither creditnor product markets are accessible to potentialbene®ciaries or where capacity to providetechnical assistance is grossly inadequate.While identifying ®nancial intermediaries whowould in principle be willing to lend to landreform bene®ciaries does not imply that every

application will automatically be approved, itcan help both parties to be more clear about therules of the game from the beginning and, inaddition, signi®cantly reduce the search cost tobe incurred by individual bene®ciaries.

(v) Formulation of productive projects

To help bene®ciaries assess the requirements,opportunities, and risks they will face as inde-pendent farmers more realistically, elaborationof model farm projects has proved to be criti-cal. While these models are necessarily abstractand therefore not directly applicable to thecircumstances of speci®c bene®ciaries or farms,they enable bene®ciaries and technical assis-tance providers to be more speci®c about keyfactors such as marketing channels, input sup-plies, working capital requirements, etc. thatneed to be made more concrete during thesubsequent process.

Under the approach followed by INCORA,where agricultural productivity received little ifany consideration, bene®ciaries generally elab-orated their ``productive projects'' after gettingaccess to the land, with little systematic guid-ance and no discretion in the use of the tech-nical assistance funds which were administeredby INCORA (often with ``bene®ciaries'' re-ceiving no bene®ts at all). Without a clear un-derstanding of the economic potential of thefarms to be established, expected returns, andalternative options (within or outside the landreform program), bene®ciaries' ability to en-gage in substantive bargaining was greatly di-minished. It was only natural that INCORAtook the lead in ``negotiating'' with the land-lord; a type of negotiation which ordinarilyamounted to mere formality and generally re-sulted in acceptance of the price set by an``independent'' assessor who was contracted bythe landlord and paid in proportion to the as-sessed farm value.

The pilot experience has indicated that, un-less bene®ciaries have a clear idea of productiveopportunities consistent with their abilities be-fore they formulate productive projects thatform the basis for ``shopping'' for land, it isvery di�cult to break this deadlock. To thisend, agricultural professionals that are con-tracted through bene®ciary representatives helpestablish crop budgets for a range of optionsactually practiced in the municipio and conducttraining courses and meetings to disseminatethem. Aggregation of these into farm plans in-volving more intensive land use and sustainable


income to bene®ciaries provides the basis forthe formulation of ``productive projects'' byindividual bene®ciaries. Only after potentialbene®ciaries have understood the model bud-gets, including their requirements and eco-nomic implications, do they proceed to a pre-selection of farms they might like to visit.

Thus, instead of being regarded as a tediousnecessity to obtaining access to complementarycredit (as under the old program), farm plansunder the pilot program acquired signi®cantimportance in a number of aspects. First, froma substantive point of view the farm modelsthat have been established thus far do consti-tute a break with the past in that the main goalis to provide full employment of the family'slabor force throughout the year. Second, farmplans are characterized by a focus on high valuecrops rather than traditional bulk commodities,greater diversi®cation of crop production, andan important livestock component. Third, allplans include a signi®cant ``garden plot,'' set-ting aside about one hectare for domestic con-sumption needs (including chickens, one pig,and a cow) and intensive cultivation of vege-tables or fruits, the surplus of which is to besold in the market. In the process of shiftingemphasis towards these goals, the importanceof land area has declined signi®cantlyÐinmany cases the land area is between 30% and50% of what had been the standard underearlier land reform programs. In addition tothis, farm plans also serve as a ®rst step towardthe identi®cation and prioritization of invest-ment needs, and to provide a justi®cation forguiding the allocation of public funds to themost productive use.

(vi) Bene®ciary training and project approval

Negotiated land reform requires bene®ciariesto take considerable initiative and performtasks such as group formation, selection of aviable farm model, adaptation of this generalmodel to the conditions of a speci®c farm,identi®cation of the productive value of at leasta number of farms available for sale, negotia-tion of a purchase price with the farm owner,arrangement for a credit to ®nance the land andcapital requirements that are not covered by thepurchase grant, formulation of a strategy toestablish needed on-farm infrastructure, andeventually cope with the challenges and risksassociated with sustaining an economically vi-able farm enterprise. Given their limited en-dowments and experience, potential

bene®ciaries are generally unable to go throughthe steps required in a ``negotiated'' type ofland reform without assistance. In fact, whilemost of the bene®ciaries pre-selected in the pi-lot had some kind of agricultural experience,almost one quarter was illiterate and 70% had®ve years or less of formal education. Whilemany were in a great rush to receive land, theirability to negotiate or manage resources wasclearly limited. Furthermore, even thoughmany bene®ciaries came in pre-existing groups,these groups were often based on coincidencemore than on similarity of interest. Their ca-pacity to resolve internal con¯icts or to devisee�ective strategies to achieve common goalswas low or non-existent. Problems that willinevitably arise in jointly establishing and sus-taining an agricultural enterprise would prob-ably have led to the paralysis or breakup ofmany of these groups.27

To remedy this, and thus increase the scopefor land reform to lead to productivity-en-hancing outcomes, an in-depth training pro-gram for pre-selected aspirantes wasdeveloped.28 This program, which is ®nancedfrom INCORA's administrative budget, aimsto cover not only abstract principles but toenable bene®ciaries to formulate a viable farmplan. The ``theory'' part includes topics thatrange from group dynamics and negotiation toeconomic analysis, farm management, andbudgeting. Simultaneously, or as soon as ben-e®ciaries have tentatively formed groups anddecided for certain crop combinations, this istranslated into practice in the context of visitsto farms that have been o�ered for sale, cal-culation of the potential of these farms togenerate revenue, the implications for the pricethat can be paid, the needed startup invest-ments to allow productive use of the farm, andthe way in which these can be e�ected inpractice by the bene®ciary group.

Contrary to widespread fears, lack of localcapacity has not been a problem in developingthese training programs. Local universities,NGOs, farmers' organizations, and govern-ment institutions (including INCORA) are en-thusiastic to utilize synergies in providing suchsupport in the expectation that they will followthe projects at least through the establishmentphase. While the costs of this component arenot negligible (about US $1,800 per bene®cia-ry), this is not only less than one-third of whatwas spent by INCORA under the old processbut can be more than justi®ed in terms of theoutcomes achieved in negotiations. Prices paid


for comparable farms under the pilot wereabout 40% lower than what had been paid inthe previous year under INCORA's program.29

Furthermore, while under the INCORA pro-cess some bene®ciaries cancelled technical as-sistance contracts that had existed under theprevious owner, pilot bene®ciaries are keenlyaware of the importance of continuing techni-cal support and provisions to pay for this ser-vice out of farm revenues are made in all of therespective farm budgets.

A review of unsuccessful land reform caseshad indicated two main reasons for failure. Onewas the absence of a fully funded plan to un-dertake the investments needed to convert thelarge farm into an enterprise suitable for smallfarmer cultivation, and the lack of funds tocarry bene®ciaries through to the ®rst harvest.To deal with this, the training phase is utilizedto develop, together with bene®ciaries, a de-tailed plan (with monitoring indicators) for thenon-farm investments needed to convert thefarm into a smallholder enterprise. The abilityto obtain partial funding for the startup activ-ities undertaken during this phase has provento be critical for the project's success.

The second problem was related to lack ofaccess to credit and output markets. Under thepilot, agreements have been reached with anumber of cooperative banks already active inrural areas to lend to land reform bene®ciariesand thus compete with the government-ownedbank that has traditionally provided ®nancing

to land reform bene®ciaries. The preferred ar-rangement bears similarity to contract farmingwhereby the bank works closely with the pro-viders of technical assistance (ensuring that thefarm business established by bene®ciarieswould indeed generate the desired revenues)and help farmers market produce. This enablesthem to supervise the use of the credit moreclosely, to ensure that enterprises are indeeddeveloping their productive potential, and todeduct loan repayments at the source, ratherthan relying on unrealistic expectations offoreclosure. Problems of interinstitutional co-ordination (essentially the inability of the in-stitutions who have traditionally administeredthese funds to work with nongovernmental in-stitutions) have prevented broader extension ofthis model. But, bene®ciaries from all the pilotmunicipios are unequivocal in their preferencefor dealing with a predictable private sectorinstitution rather than with an unpredictablebureaucracy that is directly or indirectly de-pendent on government.30

In line with the principle that responsibilityhas to rest at the local level, all of the pilotmunicipios decide about the approval (andfunding) of speci®c productive projects inpublic sessions of the municipal councilÐgen-erally with record attendance. In these sessionsbene®ciaries have to present and defend theirproject, thus not only indicating that they un-derstood the critical issues, but also providingan example to guide other candidates for the

Table 2. Key variables for land reform planning, monitoring, and impact assessment

Municipal land reform plan Monitoring Impact assessment

Bene®ciaries Bene®ciary identi®cation Grant per bene®ciary/employment. Increase in incomeBene®ciary pro®le(capacity; welfare)

Group formation. Additionalemployment generated.

Consumption smoothing(assets).

Training requirements Targeting e�ciency.Improvements in access to land.

Credit market accessSocial services.Reduction of violence

Projects Demand and supply of land Characteristics of farms transferred. Agricultural productivityChar's of productive projects Implementation of projects. Environmental sustaina-

bilityComplementary investmentsneeded

Repayment performance(planned and actual).

Reconcentration of land?

Cost by component Targeting of underutilized lands

Institutions Institutional capacity: Local andcentral government (tech. assist.legal framework private sector(banks, input suppliers, marketing)NGOs (training, evaluation))

E�ectiveness in dissemination andcapacity building

Strengthened localgovernment.

E�ciency of land transfer process. Fiscal sustainabilityPrivate sector/NGO participation. Degree of decentraliza-

tion


land purchase grant and thus setting the stagefor a transformation of the image of land re-form in more general terms. In addition togenerating positive feedback loopsÐbene®cia-ries who were selected in last year's INCORAland reform projects have already demandedaccess to similar technical assistanceÐthis alsoestablishes the basis for community-basedmonitoring and social control to ensure thatbene®ciaries' performance does actually live upto expectations.

(vii) Monitoring and evaluation

It is well known that decentralization withoutadequate mechanisms of accountability maynot have the desired consequences. In additionto helping focus on project quality rather thanmerely physical quantities (e.g., land trans-ferred) as an outcome indicator, a system thatmonitors successive stages of land reform im-plementation would help to quickly identifyand rectify unforeseen deviations from theprogram's overall objectives, in addition toassessing its long-term impacts.

Use of a grant-based mechanism that relieson market transactions to redistribute produc-tive assets is an innovative approach and thestrong reliance on decentralized mechanisms ofimplementation generates a tremendous op-portunity to learn from innovative practicesthat are developed in some communities.Careful monitoring is therefore essential to as-sess the degree to which the program attains itsoverall goals and to identify means for im-proving on implementation. This is critical toprovide greater responsiveness to operationaldi�culties than has been available in practices.Table 2 relates key components of monitoringand impact evaluation to the issues discussed inthe municipal land reform plan.

To provide answers to these questions, it isnecessary to consider which instruments arebest suited to reach particular target groups,maximize the net bene®ts of land reform (orminimize the cost of its execution), and arecompatible with a rapid attainment of thegovernment's quantitative goals and how theyperform in terms of bene®ciaries' adherence tothe project plans they have established (and theassociated economic bene®ts). Of course, themost important question is what the direct andindirect impacts of such an approach on agri-cultural productivity and poverty reduction areand how it compares with other instruments at

the government's disposal, such as constructionof infrastructure or support to education. Thiswould facilitate a judgement regarding when,and under what conditions, land reform canconstitute a sustainable and worthwhile use ofpublic funds.

Monitoring aims to ascertain whether theprogram actually achieves its objectives andwhether it is doing so at low cost. To this end, itdescribes what is happening on the ground,compares it to original objectives, and uses theinformation to update original parameters. Theinformation provided can also be used to assesswhether the mechanisms utilized do reach targetpopulations and areas (i.e. poor rural dwellersand underutilized lands); provide an ex-anteestimate of the expected bene®ts of land reform,both in terms of productivity and poverty re-duction; and provide information on whetherimplementation is progressing as expected.

Impact assessment, in turn, is concerned withthe ultimate impact, both direct and indirect, ofthe program on household well-being, agricul-tural productivity, environmental sustainabili-ty, and institutional strengthening. To this enda nationwide LSMS-type household survey,and a farm survey, will be used as a controlgroup, to be compared to a panel of bene®-ciaries and lands that will be established underthe program.

4. COMPARISON WITH OTHERCOUNTRIES

In addition to Colombia, Brazil and SouthAfrica have, under di�erent conditions, recentlyinitiated programs of negotiated land reform.With an institutional background very similarto that of Colombia (presence of land reformlegislation and a central land reform institutedating from the early 1960s), negotiated landreform in Brazil has been driven by individualstates' initiative. The purpose of the Brazilianinterventions is to establish cheaper, more agilepolicy alternatives to centralized land reform inan environment where the issue of land reformis high up on the political agenda and potentialbene®ciaries have an idea of what to do withthe land.

By contrast, negotiated land reform in SouthAfrica has been adopted in the context of thenational reconstruction program, in an envi-ronment in which productive small-scale agri-culture was eradicated almost a century ago.This implies that, in its present form, the pro-


gram bears many similarities to an urban set-tlement project and greater e�ort is required toestablish the decentralized infrastructure nec-essary to implement land reform, to providecomplementary services such as marketing andtechnical assistance, and to increase bene®cia-ries' entrepreneurial capacity.

(a) Brazil

With a land distribution among the mostunequal in the world, Brazil's situation is sim-ilar to Colombia's in a number of respects.There is a very large and vocal political demandfor land reform; a recent FAO study estimatesthe number of families who are potential can-didates for land reform at 2.5 million. A landreform institute (INCRA) was established in1969, distributing 10 million hectares to200,000 families and colonizing about 14 mil-lion hectares for about 75,000 bene®ciaryfamilies since then. Land reform has recentlyacquired considerable political importance; afederal Minister for Agrarian Reform was ap-pointed in 1996 and the land reform budgettripled from US $0.4 billion in 1994 to $1.3billion in 1995 with a further increase to $2.6billion proposed in 1997.

The large majority of these federal funds willbe spent according to the old process which hasbeen costly (about $30,000 per bene®ciary) andbureaucratically cumbersome. State govern-ments in the Northeast have been movingahead and have set up a decentralized marketassisted pilot scheme.31 This scheme (supportedby a $90 million external loan) aims to speed upthe land reform process, reduce costs, andprovide the basis for a model that could even-tually be adopted nation-wide.32

The main di�erences from the old mechanismare the following (compare Table 1):

Ð Rather than relying on a lengthy processof expropriation, land is selected by commu-nity-groups on a willing seller-willing buyerbasis. This is expected to reduce the price ofland from currently US $11,000 to $3,000,mainly by avoiding the need to pay for expen-sive land improvements that are of little usefor small scale agriculture. Funds for the landpurchase are extended to bene®ciaries as apro forma credit that has to be repaid uponemancipacion, an o�cial declaration that thefarmer is now able to farm independently,that is combined with the transfer of the ti-tle.33

Ð Instead of compensating landlords withhighly discounted government bonds, theyare paid cash. This provides a strong incen-tive for landownersÐincluding many bankswhich hold title to large tracts of land as acollateral for non-performing loansÐto sellland to land reform bene®ciaries.Ð Government's role is reduced to providingassurance that there are no problems with theland titles, and ensuring that the price negoti-ated between community groups and land-lords is within acceptable boundaries.Projects are approved at the state level.Ð Technical assistance is provided on a strict-ly demand driven basis; bene®ciaries can usepart of the community grant made availableunder a World Bank loan to contract privateproviders; CONTAG (Federation of RuralWorkers) participates in the state councilsand assists with information disseminationand land purchase negotiations.Ð The only commonality between pilots andthe nationwide land reform process is the factthat bene®ciaries under the new process haveaccess to a subsidized loan under a specialprogram (PROCERA) for land reform bene-®ciaries.

While the broad principles are similar tothose in Colombia, the process is considerablymore ¯exible and agile. The main points ofdi�erence are the following:

Ð Since grant ®nancing is provided for com-plementary and community-level infrastruc-ture rather than land itself, bene®ciarieswho expect to repay their land purchase loanhave an incentive to bid down the price forland as much as possible. This creates an in-centive to focus on lands that are currentlyunderutilized, thus reducing the expectedpurchase price and directing land reform toareas where the social gains from the inter-vention are maximized.Ð The process of bene®ciary selection is lessformal and bureaucratic than in Colombia,relying on an infrastructure of existing com-munity association to conduct informationcampaigns. Community control facilitatesgreater ¯exibility in project execution but alsocreates a danger that, without mechanismsfor supervision and ex-post accountability,landowners select former workers as bene®-ciaries and a minority of politically vocaland well-represented monopolizes a largepart of program bene®ts.


Ð As long as bene®ciary organizations existthat can assist farmers in preparing projectsand in the initial stages of establishment,the solution chosenÐwhere technical assis-tance can be ®nanced, on a voluntary basis,through the community grantÐis appropri-ate. Evaluation of the initial projects shouldexamine whether this prevents bene®ciariesfrom shifting to higher value crops and a shiftin production patterns.Ð Given the high subsidy element (about70%) in the working capital credit providedto bene®ciaries, availability of governmentfunds for this credit may easily become abinding constraint for program implementa-tion. In addition, the longer-term economicviability of land reform bene®ciaries will de-pend critically on their ability to gain accessto alternative sources of working capitaland marketing channels. This, together withthe question to what degree bene®ciaries doactually expect to repay even working capitalcredits, could become one of the critical is-sues for the long-term success of land reform.

Given the political importance of land re-form and the limited knowledge of both themost appropriate mechanisms to implementthis reform as well as the magnitude of theproductivity and poverty impact, the govern-ment has established the Central Institute forAgrarian Studies to encourage discussion be-tween all parts of civil society, academics, andpoliticians on land reform issues; carry out athorough and careful monitoring and evalua-tion of the land reform process either directlyor through local institutions; make the datacollected in this process available to nationaland international researchers, thus acting as ahub in a broader network of countries andresearchers interested in negotiated landreform.

(b) South Africa

While South Africa shares with Brazil andColombia a highly unequal land distribution,policies that neglected the land rights of thenonwhite population have systematically exac-erbated these inequalities. The Native LandsAct of 1912 prohibited the establishment ofnew farming operations, sharecropping, or cashrentals by blacks outside of the reserves, whichmade up only 7.7% of the country's area. Insidethe reserves an arti®cial form of ``traditional''tenure with maximum holding sizes and re-

strictions on land transactions was imposed.Subsequent policies of ``black spot removal''transferred the large majority of black farmerswho had legitimately owned land outside thereserves into the homelands where tenure re-strictions, high population density, and lack ofcapital and market access made commercialagriculture virtually impossible. Labor lawsthat discriminated against blacks in favor ofwhite workers and generous capital subsidiescontributed to successive evictions of largeparts of the black population from white farms,where they had been employed as labor tenantsand farm workers (Binswanger and Deininger,1993).

While the Native Lands Act was repealed in1993, the momentous task of a comprehensivereversal of these policies and their consequenceswas left to the government that entered powerfollowing the 1994 elections. In attempting todo so, this government had to contend not onlywith the extremely unequal land distribution(the average amount held per person was 1.3hectares by blacks compared to 1,570 hectaresby whites) but also the lack of any local gov-ernment structure, widespread absence of ad-ministrative capacity, a highly indebted largefarm sector, and fear that redistribution wouldwreak havoc with agricultural productivity andjeopardize national food security. The govern-ment decided to adopt a land reform policythat would redress the injustices of apartheid,foster national reconciliation and stability, un-derpin economic growth, improve householdwelfare, and alleviate poverty (Government ofSouth Africa, 1996). The three central compo-nents of this policy are as follows:

Restitution: Legal processes have been put inplace to compensate (in cash or kind) individ-uals who had been victims of forced removalsafter 1913. All restitution cases are dealt withthrough the Land Claims Court and Commis-sion, established in 1994 to which claims haveto be submitted within 3 years (i.e. by the endof 1997). Even if the legal process can be com-pleted in a speedy manner, the inability of thevast majority of the population to furnishwritten evidence makes this option feasiblefor only a small part of the population.Land tenure reform: This component seeks toimprove tenure security of all South Africansby recognizing individual as well as commu-nal ownership rights to land, giving peoplethe right to make decisions about their owntenure system, adjudicating disputes, reform-


ing tenancy laws, and attempting to end dis-crimination against women in land allocationand holding. It is intended to create the ad-ministrative infrastructure that will providehitherto disadvantaged groups with accessto land under a wide array of arrangementsthat are in line with agro-ecological endow-ments and community characteristics. It ishoped that this will provide the regulatoryenvironment for a land rental market thatcould transfer land to more productive users,redressing some of the ine�ciencies of theapartheid system.Redistribution: As the main component of thegovernment's land reform policy, redistribu-tion aims to complement the market by pro-viding land for productive and residentialpurposes to a large number of rural blackswho were dispossessed during apartheid andwho are interested in obtaining land. It aimsto do so by providing a one-time Settle-ment/Land Acquisition Grant of R15,000(US $3,300), the amount of which is equiva-lent to the National Housing Subsidy to eligi-ble bene®ciaries, de®ned as anyone with amonthly salary below R 1,500. The choiceof negotiated land reform rather than expro-priation (which, as in Colombia, can still beused as an instrument of last resort) wasbased on the need to maintain public con®-dence in the land market, and more generallyto a�rm the government's respect for indi-vidual property rights. It also re¯ects the rec-ognition that in other countries expropriationhas failed to provide rapid access to land fora large number of people and instead degen-erated into lengthy political maneuveringand rent-seeking. The number of potentialland reform bene®ciaries is considerable; esti-mates indicate that there are about 200,000labor tenants and one million farm workers,and as many as 7±8 million blacks in the re-serves.

The fact that the large majority of the ruralpopulation has never seen a successful andproductive small farm, and that many of theland reform bene®ciaries themselves seem tobelieve that e�cient agricultural production ispossible only on large farms, gives capacity-building particular importance. The South Af-rican government is well aware of these issuesand has set up a number of pilots in di�erentprovinces to accumulate experience and im-prove the execution of land reform. Theseprojects are now gathering momentum and

have provided a number of valuable lessons,many of which are currently being integratedinto government policies:

Ð The almost complete fungibility of theland purchase grant represents an importantadvantage over the Brazilian and Colombianmodels. It prevents individuals without com-parative advantage in farming from becom-ing land reform bene®ciaries just to securethe government subsidy, while at the sametime eliminating the possibility that land re-form will in¯ate land prices. The absence ofany e�ective bene®ciary contribution has,however, resulted in the formation of un-wieldy organizational structures (the average``Community Trust'' established under theprogram comprises about 200 households)which are geared more toward acquisitionof land than e�ective operation of an agricul-tural enterprise.Ð While it is the government's goal to exe-cute land reform with maximum local partic-ipation, the lack of an institutional structurehas up to now made e�ective decentralizationand bene®ciary participation di�cult. Thishas resulted in huge demands being placedon DLA sta� and at the same time consider-ably slowed down the process (the projectedtime for DLA to process a project is about14 months, and often more time is required).While progress has been made in addressingthis issue, real involvement of local playersis likely to require some devolution of deci-sion-making authority (at present each landdesignation has to be signed by the Ministerhimself) and a shift toward ex post control.This, of course, is contingent on appropriateinformation systems being in place.Ð Bene®ciaries' access to complementaryservices and infrastructure has been problem-aticÐa situation that was exacerbated be-cause responsibility for these services restsoutside DLA, the agency responsible for landreform execution.34 These shortcomings canbe addressed by empowering emerging localgovernments to play a stronger role in coor-dinating the di�erent institutions involved atthe local level. In addition, more systemati-cally incorporating the private sector (banksand other ®nancial institutions, as well as cur-rent land owners) in the formulation andevaluation of farm plans, the provision oftechnical assistance, as well as marketingand input supply, could e�ectively addressthe current problems of ``aftercare.''


Ð While there is strong emphasis on elabora-tion of land use and business plans by bene®-ciaries, many of these plans have beenelaborated by entities who do not have re-sponsibility for the long-term economic suc-cess of the project and do not containarrangements for technical assistance in theproject's establishment phase. The experiencefrom land reform in Latin America illustratesthat, in the absence of clear plan that is avail-able before accessing the land, there is a dan-ger that bene®ciaries will eat up whatevercapital stock was transferred with the farmand subsequently revert to survival farming.This neglect of the productivity aspect of landreform, often exacerbated by unwieldy groupsizes and lack of coordination among di�er-ent institutions, could seriously threaten thepoverty-reducing potential of this interven-tion, in addition to having negative environ-mental consequences (Cross et al., 1996).35

5. CONCLUSION

To demonstrate that negotiated land reformcan provide a solution to the problem of assetmaldistribution observed in many developingcountries, two key questions need to be answ-ered. First, one needs to show that the new app-roach is more e�ective and less costly than earlierland reform attempts. Second, it will be neces-sary to establish that investment of scarce gov-ernment resources in land reform is warranted.

While a ®nal judgement on whether negoti-ated land reform can rise to the challenges thatadministrative land reform has failed to solvewill have to await availability of appropriatedata,36 the experience from the three countriesprovides valuable lessons that can guide at-tempts to implement programs of land reform.It suggests that (a) land reform through nego-tiation can only succeed if measures are takento make the market for land sales and rentalmore transparent and ¯uid; (b) productiveprojects are a core element of market-assistedland reform that is designed to establish eco-nomically viable and productive projects at asocially-justi®able cost rather than to transferassets; (c) the only way to achieve e�ective co-ordination of the various entities involved inthis process is through demand-driven and de-centralized implementation; and (d) the long-run success of land reform is likely to dependcritically on getting the private sector involvedin implementation, and the ability to utilize the

land purchase grant to ``crowd in'' privatemoney.

Making land markets more transparent and¯uid: While there is broad agreement thatselection of land and bene®ciaries should bedemand-driven and e�ected at the local level,the problem is that the conditions for local landmarkets to function, such as information onland prices as well as bene®ciaries' ability toassess the value of a piece of land or the po-tential productive returns of its more intensivecultivation, are often limited. Three mutuallyreinforcing strategies to deal with this constraintare (i) the provision of technical assistance atthe community level, including assessment ofthe adequacy of the land price at the point oftransaction; (ii) co-®nancing of the land pur-chase through a private ®nancial intermediarywhich, because it shares in the risk of default,will have an incentive to assess the economicfeasibility of the proposed farming project; (iii)a ``market information system'' to provideprices for plots transacted in the marketÐbothwith and without use of a land purchase grant.

Negotiated land reform is a complement,rather than a substitute for other forms ofgaining access to land, especially land rental. Inthis sense, land rental should be understood asa means for bene®ciaries to accumulate expe-rience and start-up capital thus reducing thesize of the land purchase grant required under amodel of negotiated land reform. This has toreplace an understanding whereby renting outmight cause owners to lose their land or rentingin may cause tenants to become ineligible for aland purchase grant. Up to now none of themodels considered here have contemplated indepth the potential bene®ts of interregionalmigration, and ways to encourage such migra-tion to reduce the cost of a land reform pro-gram and at the same time enhance its impacton productivity.

Focusing on productive projects: Productiveprojects are likely to be the key of market-as-sisted land reform because the income generatedby such a project it is the only objective criterionto put an upper bound on the price that can beexpected to be paid; because it provides a nec-essary basis for ®nancial intermediaries toevaluate and eventually support such projects;and because it requires bene®ciaries to famil-iarize themselves with the realities they are likelyto confront as independent farmers, and in theprocess provides them with greater clarity ontheir own aspirations as well as the potentialÐand the limitationsÐof land reform to con-


tribute to the attainment of these goals. Incontrast to previous plans that were imposedfrom the top with little awareness of local con-straints, these plans aim to create the basis for amore transparent market at the local level andare the starting point rather than a substitute formore active bene®ciary involvement.

Decentralizing implementation: Experiencewith centralized land reform has revealed it as aslowÐand costlyÐalternative to decentralizednegotiation. In Colombia, the whole pilot, fromthe ®rst dissemination e�orts to project ap-proval by a technical committee and local ®-nancial intermediaries, was completed withinseven months37Ðwith landlords being the mosteager party to see it advance. Prices paid underthe decentralized approach were about 40%below the cost of land that had been paid ear-lier and bene®ciaries see a major bene®t fromtraining and detailed project planning in theirability to negotiate independently with land-owners and the fact that they have a clear ideaof how to proceed once they receive the land.Similar ®gures are available from Brazil wherethe presence of community-organizations al-lowed even more speedy implementation. Evenin South Africa, where the absence of localinstitutions at the start of the program made arelatively centralized mode of implementationnecessary, emergence of local governments hasalready facilitated signi®cant steps towardgreater decentralization and advances in pro-gram implementation.

Maximizing private sector involvement: Twolessons have become evident about the ®nancingof land reform. First, restricting a land reformgrant to a speci®c part of the land reformpackage to the exclusion of others (as wasoriginally the case in Colombia) is likely to becounterproductive; a ¯at grant that can be usedfor all types of expenditures (as in South Africa)

is clearly preferable. Second, without e�ectivebene®ciary contribution and assured access to®nancial markets as a part of the land reformpackage either bene®ciary self-selection of thesustainability of project operations will becompromised. There seems to be scope for ex-ploring possibilities of combining this withmechanisms aimed at sustainable savings gen-eration. Given the high transaction costs ofproviding credit in rural areas and the increasedneed for monitoring when most clients lackprevious exposure to credit, it is important thatthis issue be given su�cient attention. This reliesheavily on the formulation of economically vi-able and technically feasible productive projects.Getting participantsÐgovernments and NGOsas well as potential bene®ciariesÐto realisticallyassess the potential as well as the dangers in-herent in negotiated land reform is critical.

A focus on ``integrated land market devel-opment'' that aims to develop land marketsjointly with markets for other factors may o�erpotential not only in situations such as the onesdescribedherewhere landisdistributedveryineq-uitably. It may also help countries (e.g., Nica-ragua, Honduras, El Salvador) where large landreforms has not had the desired productivityimpact to realize the productive potential of thereform sector and provide a model for ``fair''dispute resolution in situations (e.g., Uganda)where overlapping claims and long-standingdisputes over land ownership have severely af-fected the productivity of land use. If negotiatedland reform achieves to direct e�orts towardpolicies that make markets work better for thepoor, and at the same time helps bene®ciariesimprove their human capital endowments,change from passive objects into subjects of theprocess, and convert a one-time subsidy into apermanent improvement of their livelihood, itwill have more than achieved its purpose.

NOTES

1. One of the ®rst studies to provide both a theoretical

model and an empirical investigation of the potential

impact of land reform indicates that, based on district

level data from India, land reform can indeed have a

positive impact on wages and employment (Rosenzweig,

1978).

2. Though still accumulating, empirical evidence does

support some of these conclusions (Jalan and Ravallion,

1997;Fafchamps and Pender, 1997).

3. While it would be desirable to have more studies

examining the productivity impact of such reforms in

quantitative terms, there is agreement in the literature

that such reforms have been associated with signi®cant

increases in output and/or productivity (Callison, 1983;

Koo, 1968; King, 1977; Lieten, 1996; Besley and

Burgess, 1998; Dorner and Thiesenhusen, 1990; Otsuka,

1991; Banarjee and Ghatak, 1996).

4. Conceptually, one would expect productivity gains

to be proportional to the improvement in work and


investment incentives associated with the post-reform

regime. In cases where security of tenure had already

been high before the reform, where cash-rent (rather

than share rent) contracts had prevailed, and where

landlords had provided tenants with access to markets

for credit, inputs, and outputs, one would expect static

e�ciency gains from land reforms to be modest and the

bulk of reform bene®ts to come through enhanced

investment incentives and credit access associated with

land ownership.

5. In many cases where large farms were distributed

exclusively to the resident labor force (e.g., in Peru,

Nicaragua, Honduras, and Cuba), the resulting units

were too large to be able to reap the bene®ts from

utilization of family labor. As a result, production on the

reformed farms was often characterized by pervasive

labor problems and productivity increases failed to

materialize.

6. The only exception is where already well-established

plantations were redistributed to the former workers, a

case that is very costly and characterized by relatively

low social bene®ts (Adriano, Hayami and Quisumbing,

1990).

7. In Ireland, in the early 20th century, a large scale

``negotiated'' land reform transferred of 9.3 million acres

(about half of the agricultural land available) from

landowners to tenants but had a very limited impact on

productivity. One reason is that land reformÐwhich was

enacted on top of earlier legislationÐdid little to alter

the structure of production or increase investment

incentives. Another issue, probably more important, is

that land reform led to a worsening of access to credit, by

limiting the ability of new landowners to mortgage land

while at the same time cutting o� informal credit they

had earlier obtained from the landlord (Guinnane and

Miller, 1997). Severely restricted access to credit,

together with insecure property rights have also led to

widespread selling of land by former land reform

bene®ciaries in NicaraguaÐoften at prices well below

the productive value of the land (Jonakin, 1996). Lack of

access other markets has been identi®ed as a key

obstacle to the initiation and success of independent

small farming operations in Eastern Europe (Brooks and

Lerman, 1994).

8. This would be consistent with an interpretation of

land reform as a piecemeal strategy by the rich to avoid

the imminent threat of revoltÐwith backtracking as

soon as the threat weakens as modeled by Horowitz,

1993.

9. Even where there was a genuine commitment to

breaking the power of landed elites, agrarian reforms

were generally designed by urban intellectuals with little

idea of the realities of agricultural production and a

prejudice against the ability of small-scale cultivators to

manage on their own farms, let alone increase produc-

tivity (Barraclough, 1970).

10. The importance of political factors is also evident

from a number of Eastern European countries where

political constraints generally led to a relatively ``ine�-

cient'' way of implementing land reformÐthrough

physical restitution of plots rather than compensation

of former owners through fungible cash payments

(Swinnen, 1997).

11. Despite attempts to limit bene®ciary desertion

through imposition of legal restrictions and severe

punishment in case of contravention, there is consider-

able anecdotal evidence on land sales by reform bene-

®ciaries in Nicaragua, Colombia, and El Salvador. In a

recent census of Brazilian land reform settlements, only

about 60% of land reform bene®ciaries were actually

found tilling their land.

12. It is well known that market imperfections com-

monly encountered in rural areas of developing coun-

tries can lead to concentration of land in the hands of

larger producers (e.g., Carter and Mesbah, 1993). But

credit market imperfections aloneÐwithout policy in-

terventionsÐcannot explain the heavy underutilization

of land in Colombia where 75% of potential cropland is

currently under pastureÐit would presumably be more

pro®table for large landowners to rent out to shareten-

ants rather than to use land for extensive cattle ranching

or to leave it completely fallow.

13. Balcazar (1990) summarizes the existing literature

in three points, namely (a) small farms are cultivated

more intensively than large farms, as measured by value

of output per unit area; (b) between 1973±76 and 1988

average physical yields on small farms increased by

about 82% (this seems to be an unweighed average

across di�erent types of crops) whereas those on large

farms have remained stagnant; (c) regional as well as

commodity-speci®c studies ®nd no systematic relation-

ship between farm size and adoption of new technology

or improvements in productivity.

14. In the early 1990s the administrative costs of

transferring land were very high, amounting to about

50% of the total land reform budget or about $15,000

per bene®ciary.

15. This failure to proceed more swiftly with imple-

mentation of the negotiated model of land reform was


not due to resource constraints but rather institutional

rigidities and resistance. In fact, resources available for

market assisted land reform were accumulated until the

end of the year and then disbursed in a rushed process

that bore little relationship to the established procedure,

with the idea that it would be better to spend the

resources imperfectly than to lose them.

16. For historical reasons, the grant was set to be 70%

of the purchase price of the land. The remaining 30% of

the land purchase price, plus any additional start-up

investment, had to be obtained from other sourcesÐ

either the farm household's own resources or a regular

loan from a ®nancial institution at market rates, i.e.

without any speci®c subsidy for land reform bene®cia-

ries. It was hoped that by sharing the risk, ®nancial

intermediaries would provide additional assurance of the

economic viability of land reform enterprises. While the

high level of subsidy received strong political support

from large landowners, experience thus far indicates that

bene®ciaries would be better served by a more modest

grant that can be used for either land purchase or

investments related to the startup of the new enterprise,

i.e. that the 70% is too high.

17. Landlords have in many instances overstated the

price of land, and-by covering the complete land value

with the 70% grant obtained a subsidy element of 100%.

Consequently, in 1996 the price of land acquired

through ``direct intervention'' by INCORA (under a

residual budget) was lower than the price of land

acquired by bene®ciaries through ``negotiated'' land

reform in the open market, leading to widespread

dissatisfaction and calls for the return to the interven-

tionist paradigm.

18. Given that this regulation creates another layer of

bureaucracy, transaction costs, and uncertainty about

approval (or scope for politically motivated interven-

tions), it would have been more desirable to change the

underlying legal framework than develop ad-hoc solu-

tions.

19. The concentration of bene®ts may be expedient

from a political point of viewÐespecially if the bene®ts

can be appropriated by the about 10±15% of the

peasantry that is politically well organized and that

closely collaborates with INCORA. But comparison

with European standards puts the farm size issue in

perspective. In 1990 (the latest year for which these

statistics are available), average farm size in the EU

(including former East Germany) was, with 14.8 hect-

ares, slightly below the Colombian ``minimum size'' of

15 hectares. Average farm sizes in Greece and Italy are

only 4 and 5.6 hectares, respectively, and even in the

Netherlands and Germany, the average size of well±

established farms was only 16 hectares (Eurostat, 1995).

20. The categorical prohibition of rental of land

reform land (included in the 1994 law) is unlikely to be

enforceable, and may even be counterproductive in that

it would discourage successful reform bene®ciaries from

intensi®cation on part of their land and renting out the

other part while preventing unsuccessful reform bene®-

ciaries from exiting.

21. The institutional structure is exemplary from a

conceptual point of view, comprising (a) decentralized

decision-making characterized by maximum local par-

ticipation expressed through the pre-eminent role of the

local councils; (b) private sector involvement in bringing

together potential buyers, provision of complementary

credit, and technical assistance to continue during the

®rst two years of production on the land received; and

the (c) limitation of INCORA to a regulatory role

preventing misuse of funds, ensuring that regulatory

requirements are met, and coordinating the di�erent

government agencies involved.

22. Private real estate agents, who were supposed to

act as information brokers in making potential buyers

and sellers aware of market demands, found it di�cult

to compete with comparable INCORA services which

were o�ered free of charge. In addition, INCORA's

continued involvement in the decision-making processÐ

including its ability to manipulate political leversÐput

the institution in a position to virtually guarantee a

``successful'' outcome.

23. These municipios are San Benito Abad in Sucre,

Riveria in Huila, Fuente de Oro in Meta, Montelibano

in Cordoba, and Puertowilches in Santander. While

continued guerrilla activity throughout Colombia un-

doubtedly a�ects the scope for implementing land

reform, acceptance of and support for the process at

the local level has virtually eliminated guerrilla-related

disruptions.

24. While one would expect that the transfer of large

and relatively extensively cultivated farms would provide

an opportunity for accommodating additional workers,

the opposite often happened in practice, due to the

unwillingness of existing workers to reduce their share

and the associated incentives to exaggerate the amount

of land needed to establish a productive unit. This

tendency to expel laborers is well known from the theory

of cooperatives.

25. As highlighted earlier, even under the new land

reform law, rationing was largely by nonprice meansÐ


including outright corruptionÐand the main challenge

for landowners willing to sell was to obtain clearance

from the regional and central INCORA o�ces regarding

the availability of funds to purchase their land which

was seen as the precondition for proceeding with

negotiations.

26. In principle the decentralized implementation

mechanism, if coupled with competitive allocation of

funds across municipos, should eliminate the scope for

spending scarce resources on municipios where land

reform is not economically feasible and outmigration is

the most cost-e�ective way to satisfy demands for land.

Due to the limited extent of the pilot, no such cases have,

as of yet, been encountered.

27. Indeed one of the surprising insights from the pilot

was that virtually all of the groups that had initially

existed were disbanded and replaced by new ones that

were based more on commonality of interest (e.g. in

speci®c production systems) or complementarity in

experience.

28. The number of pre-selected aspirants is about

double the number of those that could be accommodat-

ed as bene®ciaries with given budgets to allow for

attrition and to ensure that potential bene®ciaries do

indeed compete in trying to put together the most viable

farm projects.

29. These price di�erences are not due to quality

factors. In one case, a farm neighboring (and almost

identical) to one that had been acquired in 1996 by

INCORA (supposedly under the market-assisted pro-

gram) was acquired under the pilot with a price saving of

about 40%Ðmoney which bene®ciaries used to pay for

complementary investments and working capital. In

other cases, discounts of similar magnitude were ob-

tained on the ``o�cial valuations'' that had been

commissioned by the landowners.

30. Indeed, obtaining ®nancing for the projects elab-

orated by bene®ciaries in the pilot municipios developed

into one of the main obstacles for smooth implementa-

tion. Even though all projects had been thoroughly

evaluated and approved at the local level (including

visits by bank representatives), it took the government

owned Caja Agraria between four and ®ve months to

obtain central approval from Bogota. This severely

damaged the goodwill of landlords who wanted to sell

and jeopardized bene®ciaries' ability to proceed with

their plans in a timely fashion and was often combined

with dictating of changes to the projects which were

incompatible with bene®ciaries' preferences or needs.. In

view of this, all the parties (including bene®ciaries)

involved in the pilot agree that the pilot experience can

be successfully transformed into a broader national

program only if the monopoly of Caja Agraria is broken

and private sector institutions ®nance land reform

projects. This would, of course, imply that they assume

some of the associated risks.

31. The steps involved in the process of expropriation,

which is applicable to land that is utilized to less than

80%, are as follows: First, there is a visit by an INCRA

mission to assess the value of land and improvements,

followed by expropriation through presidential decree

and con®rmation of expropriation through a federal

court (emissao de posse) in a process that takes about a

year. Once this is accomplished, landowners are com-

pensated with Titulos da Reforma Agraria (with a real

interest rate of 6%; bearing a discount of 25±40% in the

market); anecdotal evidence of excessive compensation

abounds. The necessary infrastructure investment is then

included in the subsequent year's INCRA budget (it

takes 1±2 years for the infrastructure to be established)

and INCRA announces that the land is available,

selecting bene®ciaries based on agricultural skills, al-

though in practice all cases are limited to regularization

of existing squatter settlements. Once selected, bene®-

ciaries are eligible for credit from PROCERA (maxi-

mum US$7,500; average US$4,500), a special program

for land reform bene®ciaries administered through

INCRA (with an e�ective subsidy of about 70%).

32. The emphasis on land reform by the federal

government and the initiation of alternative approaches

by state governors (essentially in the hope of being able

to demonstrate that they are able to deal with the

problem more e�ectively) is certainly not unrelated to

the political strength of the landless movement (Mo-

vimiento sem Terra, MST). Whether the MST will

modify its approach in favor of the old type of land

reform projects will, among others, also depend on the

degree to which the new approach will be able to live up

to expectations.

33. Not surprisingly, the number of land reform

properties that have been emancipated is minuscule.

34. Based on a case study of one of the provincial land

reform pilot in Kwazulu Natal, Cross et al. (1996) report

four main shortcomings, namely (a) a complete lack of

interinstitutional coordination (e.g. concerning access to

water resources); (b) underestimation of the time and

energy required for legal issues such as adjudication

between con¯icting land claims; (c) political tensions

between ANC and IFP (at the national and provincial

level, respectively) which prevent the e�ective delivery of

services other than land to the bene®ciaries, and (d) a


tendency to interfere by local chiefs who perceive land

reform mainly as a means to enhance their own power

which depends on the number of people they are able to

accommodate rather than the productive success of the

programs that are established.

35. ``NGOs and pilot structures are.... beginning to be

afraid that they would not be able to do more than

provide land redistribution bene®ciaries with the mini-

mum of land and secure tenure and that this alone would

amount to just dumping themÐequivalent to the apart-

heid practice of dumping removals victims in the middle

of the veld with no shelter and no way to make a living''

(Cross et al., 1996 p. 166).

36. Data for such an exercise will hopefully be supplied

by the monitoring systems that are established in all of

the countries described.

37. Following approval of productive projects at the

local level, centralized institutions (INCORA and Caja

Agraria) engaged in endless soul-searching and tried to

reject projects because they were ``too far from infra-

structure'' (even though construction of a bridges or

rehabilitation of roads was part of the farm development

plan) or otherwise not suited for smallholder cultivation.

Even though in the end all of the projects were

approved, this did much to undermine the credibility

of the process with sellers.

REFERENCES

Adriano, L. S., Hayami, Y. and Quisumbing, M. A. R.(1990) Toward an Alternative Land Reform Para-digm, A Philippine Perspective Ateneo de ManilaUniversity Press, Manila.

Balcazar, A. (1990) Tana_o de ®nca, dinamica tecnolo-gia y rendimientos agricolas, Coyuntura Agropecu-aria 7(3), 107±125.

Banarjee, A. V. and Ghatak, M. (1996) Empower-ment and e�ciency: the economics of tenancyreform, Massachusetts Institute of Technology,MA, USA.

Banarjee, A. V. and Newman, A. F. (1993) Occupationalchoice and the process of development, Journal ofPolitical Economy 101(2), 274±298, The University ofChicago, Chicago, USA.

Banarjee, A. V., Mookherjee, D., Munshi, K. and Ray,Debraj (1997) Inequality, control rights and rentseeking: A theoretical and empirical analysis of sugarcooperatives in Maharashtra. Bostan University,Boston, MA, USA.

Bardhan, P., Bowles, S. and Gintis, H. forthcomingWealth Inequality, Wealth Constraints and Econom-ic Performance. In: Anthony Atkinson and FrancßoisBourguignon (Eds.), Handbook on Income Distribu-tion, forthcoming 1998.

Barraclough, S. L. (1970) Agricultural Policy and LandReform. Journal of Political Economy pp. 906±947.

Berry, R. A. and Cline, W. R. (1979) Agrarian Structureand Productivity in Developing Countries, ILOGeneva.

Besley, T. and Burgess, R. (1998) Land Reform, PovertyReduction, and Growth: Evidence from India, LSEmimeo.

Binswanger, H. P. and Elgin, M. (1988) What are theProspects for Land Reform? in: A. Maunder and A.Valdes, (Eds.) Agriculture and Governments in anInterdependent World. Proceedings of the TwentiethInternational Conference of Agricultural Econo-mists, Buenos Aires.

Binswanger, H. P. and Deininger, K. (1993) SouthAfrican Land Policy: The Legacy of History andCurrent Options, World Development 21(9), 1451±1475.

Binswanger, H. P., Deininger, K. and Feder, G. (1995)Power, Distortions, Revolt and Reform in Agricul-tural Land Relations. In: J. Behrman and T. N.Srinivasan (Eds.), Handbook of Development Eco-nomics, Volume III, Elsevier, Amsterdam.

Birdsall, N. and Londo~no, J. L. (1997) Asset Inequalitymatters: an assessment of the World Bank's ap-proach to poverty reduction. In: American EconomicReview 87(2), 32±37.

Brooks, K. and Lerman, Zvi (1994) Land reform andfarm restructuring in Russia. World Bank DiscussionPaper 233, World Bank, Washington, DC.

Callison, C. S. (1983) Land to the Tiller in the MekongDelta, Economic, Social and Political of LandReform in Four Villages of South Vietnam, Univer-sity Press of America, Lanham, MD.

Carter, M. R. and Mesbah, D. (1993) Can land marketreform mitigate the exclusionary aspects of rapidagro-export growth? World Development 21(7),1085±1100.

Cross, C., Mngadi, T., Sibanda, S. and Jama, V. (1996)Making a living under land reform: Weighing up thechances in KwaZulu-Natal, Chapter six. In: Lipton,M, Ellis, F and Lipton, M. (eds), Land, Labour andLivelihoods in Rural South Africa: Volume II: Kwa-Zulu-Natal and Northern Province. Indicator, Uni-versity of Natal, Durban, South Africa.

Dasgupta, P. and Ray, D. (1986) Inequality as adeterminant of malnutrition and unemployment:Theory, Economic Journal 96, 1011±1034.

Dasgupta, P. and Ray, D. (1987) Inequality as adeterminant of malnutrition and unemployment:Policy, Economic Journal 97, 177±188.

De Janvry, A. and Sadoulet, E. (1989) A study inresistance to institutional change: the lost game ofLatin American land reform, World Development 17,1397±1407.

Deininger, K. and Squire, L. (1998) New Ways ofLooking at Old Issues: Inequality and Growth.Journal of Development Economics 57, 249±287.

Dorner, P. and Thiesenhusen, W. C. (1990) Selected landreforms in East and Southeast Asia: their origins andimpacts, Asian Paci®c Economic Literature 4, 69±95.


Eckstein, Z. and Zilcha, I. (1994) The e�ects ofcompulsory schooling on growth income distributionand welfare. Journal of Public Economics, no. 54, pp.339±359.

Eurostat (1995) Daten fur die landwirtschaft der EU.O�ce des Publications O�cielles des CommunautesEuropeen, Luzembourg.

Fafchamps, M. and Pender, J. (1997) Precautionarysaving, credit constraints, and irreversible invest-ment: Theory and evidence from semi-arid India.

Fajnzylber, P., Lederman, D. and Loayza, N. (1998)What causes violent crime? mimeo, World Bank,Washington DC.

FAO (1994) El mercado de tierras y la formacion depropietarios en Colombia: Estudios de casos. FAO,Rome.

Galor, O. and Zeira, J. (1993) Income distribution andmacroeconomics. Review of Economic Studies 60, 35±52.

Government of South Africa. (1996) Our Land. GreenPaper on South African Land Policy. Head O�ce,Department of Land A�airs.

Guinnane, T. W. and Miller, Ronald I. (1997) Thelimits to land reform: The land acts in Ireland, 1870±1909. Economic Development and Cultural Change45(3), 591±612.

Heath, J. and Binswanger, H. P. (1996) Natural resourceDegradation E�ects of Poverty and PopulationGrowth are Largely Policy-Induced. Environmentand Development Economics, Volume 1, Part 1.Cambridge University Press, MA.

Ho�, K. and Lyon, A. B. (1994) Non-Leaky Buckets:Optimal Redistributive Taxation and Agency Costs.NBER Working Paper Series, no. 4652, NationalBureau of Economic Research, Cambridge, MA.

Horowitz, A. W. (1993) Time paths of land reform: Atheooretical model of reform dynamics. The Amer-ican Economic Review 83(4), 1003±1010.

Jalan, J. and Ravallion, M. (1997) Are the poor less wellinsured? Evidence on vulnerability to income risk inrural China, The World Bank.

Jonakin, J. (1996) The impact of structural adjustmentand property rights con¯icts on Nicaragua agrarianreform bene®ciaries. World Development 24(7), 1179±1191.

King, R. (1977) Land reform: a world survey. Bell,London.

Koo, A. Y. C. (1968) Land reform and economicdevelopment: a case study of Taiwan. Praeger, NewYork.

Kutcher, G. P. and Scandizzo, P. L. (1981) Theagricultural economy of Northeast Brazil. WorldBank, Washington DC .

Li, H., Squire, L. Zou, H. F. (1998) Explaininginternational and intertemporal variations in incomeinequality. Economic Journal 108, 26±43.

Lieten, G. K. (1996) Land reforms at center stage: Theevidence on West Bengal, Development and Change27, 111±130.

Lipton, M. (1974) Towards a theory of land reform, In:D. Lehman, ed., Peasants, Landlords and Govern-ments: Agrarian Reform in the Third World, Holmes& Meyer, New York.

Mision Social (1990) Ministerio de Agricultura, Depart-amento Nacional de Planeacion.. El DesarrolloAgropecuario en Colombia, Tomo 1. Bogot�a, D. E.,Rep�ublica de Colombia.

Moene, K. O. (1992) Poverty and Landownership,American Economic Review 82, 52±64.

Mookherjee, D. (1997) Informational rents and propertyrights in land, In: J. Roemer, ed., Property Rights,Incentives & Welfare, Macmillan, 1997.

Otsuka, K. (1991) Determinants and consequences ofland reform implementation in the Philippines,Journal of Development Economics 35, 339-55.

Rodrik, D. (1998) Rural wages, labor supply and landreform: A theoretical and empirical analysis. Amer-ican Economic Review 67, 847±861.

Rosenzweig, M. R. (1978) Rural wages, labor supplyand land reform: A theoretical and empirical anal-ysis, American Economic Review 67, 847±861.

Swinnen, J. F. M. (1997). Does compensation fordisruptions stimulate reforms? The case for agr-arian reform in Central and Eastern Europe. Euro-pean Review of Agricultural Economics 24(2), 249±266.

World Bank (1996) Review of Colombia's agriculturaland rural development strategy. Washington DC.


making negotiated land reform work: initial experience from colombia, brazil and south africa

Documents