February 1997                                                           21           2                      N O. 16
0-X   Agriculture
&* Technology Notes A
The Agricultural and Forcstry Systems Division (AGRAF)                                             The WVorld Bank
Promoting Sustainable Agticulture
Improving Policies for Natural Resource Management
Farmers are aware of natural resource degradation, but environmentally unsustainable agricultural practices
remain widespread. Lil<e other differences among farmers, many factors affect the adoption of
sustainable or unsustainable techniques.
While not alwavs a major threat to long-
term  productivity, severe and lasting                                             -
degradationi sometimes occurs because                                                    4 K t,
farmers have insufficient incentives to con- ,I
serve. This Note discusses the degradation i
of agricultural environiments, explains l
somie causes of unsustainable practices,
and suggests policies to encourage more
environmentally sustainable agricultural   l_
techniques. Well-planned sector work and            i.                                             .
policy analysis are vital first steps toward
ensuring that interventions designed to                                                  '1
promote su-stainable agricultuiral develop-    -
ment are both effective and efficient.
Neithier farmers iior agricultural scien-
tists intentionally design agricultural
systems to erode or degrade the resource
base on whlich production depenids. W'hile-
alarmist statistics suggest that natuiral re-
source degradationi mav' threaten wvorld 
food security, it is first necessary to distin-                                                       .     - 
guish real from illusorv dangers to enable   Vegetotion used to stabilize a hillside in Indonesia
policymakers and project managers to
use scarce resources wisel. Second, in    Risks of Resource Degradation        disparity reflects both data scarcity and
order to design effective conservation    The two wvaxs in shich agriculture can    qualty ad the fact that socioeconomic
measures, the reasons for unsustainable   groxv - expansion of cultivated area and    context is oftell neglected in technically-
resource use must be understood. Evi-  iniereases in yiel(d per unit area -  can    based concepts of resource degradatiorn.
dence suggests that where market, policy,   each have positive and negative environ-  Despite these problems. in nmuch of the
and technical conditions are favorable,   mental eflfects. The means of short-termi   world agricultural developmenit appears to
development of intenisive agriculture canl   groxvth clearly atfect lonig-termii viahility of'  be sustainable, and technology anld insti-
be consistent witlh conservation or even   the resource base, but the environmental   tutions continiue to develop and adapt
improve the natural resource base. Where   impact of agricultural growth in different   to the environmiienital stress of growth.
underlying socioeconomic and policy    counitries is the subject of conisiderable    The videly publicized case of natural re-
conditions are unfavorable, introducinsg   debate and iuncertaintv. Estimiiates of   source maniagemiienit (NRM) in Kenya's
improved resource management practices   the extent and conse(quenices of niatural   Machakos District (see box, p. 3 ) illuls-
is likely to fail because only symptoms are   resource degradation vary widely, as   trates potential success (English, Tiffen,
addressed, not root causes.             do their credibility (see box, p. 2 ). This   and Mortinmore 1994).



AGRAF Tech nology Notes
Keys to Successful Innovation              productivity is so low that investments   features of free markets, they often fail to
Evidence from around the world sug-    become profitable. A transition to more    provide adequate incentives for sustainable
gests that farmers are conscious of' the    intensive agriculture occurs -farmers    NRM. When markets fail, prices do not
catuses and consequences of natural    plant trees, build terraces, apply fertilizer    accurately reflect scarcity, making them
resource degra(lation. Farmer surveys    and organic matter, and regulations are    poor indicators of the need to conserve. In
conrsistently show awareness of erosion    imposed on communal forest, pasture, and    many cases market failures arise due to char-
and nutrient loss, deforestation, and range    water resources. In a sustainable and    acteristics of NRM, but they are often
(learadationi. A growing body of literature    intensive agricultural system, land produc-    compounded by government interventions
and  field experience shows that the    tivity mav ultimately be higher than under    that are termed policy failures.
Machakos experience -where people
have developed effective resource man-
agem ent practices  wsthout external
assistance in the Lace of rapid population
growth  is not unique. Induced innovma-         A    E
tiot  thIeor lholds  that over  time,gl at s                                         t
technological innovations and institutional B.t   ta
changes mi-ake the imost of' abundant         adsvr   er.to  s lIIgatetmtd5t  0mlin1etr~o
resour-ces anid econtomiize on scarce re-       oladanalyccrigoth   IEsud,4ecntorigedad
souarces. This theor helps to explain the     wre
cireumrowstances tinder wshich agricultural   othe   frret                                       bye o nefite Interv eon
developmaent will take  place  along          tiy,a   during   th 1a     annuae       a  o       perent. ate  in Mai,
pathIs that deerade or conserve natural      Mes ic,sta m ad       Indolloind degra    dvle        ul        tural syste
resources. It is uisuallv' illuistrated by    lscrnlgra  .t  ~p    noG                        4e~(94etmaeta
com lparing the labor-intensive, land-        by teverd tf the systen    lan neverinttins In the itSesw re                -
saving agriculture of Japan with its histor-  duce crop lds by tox10 perhats   to wat thewp ie there.
icalo v  rich  labor supplv with  the Tn  a        onodiatcosInsathIII,anvethe  orentoeapanrboth nrentl f,i-
land-intensive, labor-saving agriculture Cente th conr   resou           wllo an d  real in preq wnl d           e 21
of the land-abundant United States.           du,adprodu        esv urceidestort    floods   pownutrnrm partilars
Induced innovation also helps to explain   and otheng. A nuberao  espians         ine: eond  congl iesan Poli     ron-
wilespread land degradation in some          lg are  suchiaith  in   a       particl f Chin satherpoorllandestood
places and more sustainable land use in       formesore  dingri cult                      feavy intensin  stlyter
others. In land-abundant tropical areas,
hush  fallow  was traditional. Bult in
land-scarce areas such as the intensive
rice-growing regions of southeast Asia,   the original forest system.                 Inappropriate Interventions
elaborate terraces, irrigation systemis, and  Besides the optimistic Pathi I, Figure I  Tw,o exam-ples demionstrate policy failuire
pitriat inanagemenkt enabled continuous    also illustrates tho other NRM scenarios,   in NRM  and the potential costs from
ivation without land degradation.      In Path  II, investments to improve    failing to recognize the forces that drive
Jije ted innovation suggests that degra-    resources are made onil following degra-    development of agricultural systems and
datioeimnaysometimnes be self-correcting    dation so severe that the system can never   institutions. In the 1980s, forexample, con-
as resoiutre scarcit induces developiaindit   pecoierhis  productivity (exceptperhaps    cern was widespread that deforestation in
of' tinre sustainable land use (Path I in    at uneconomical costs). In Path III, invest-    Nepal increased erosion and runoff, in turn
Figutre 1) (Sclherr and Hazell 1994). Coni-    ments to conser-ve resources are never    increasing the frequency and severity of
si(ler- thie case of land that can lie mnanaged    niade, and productivity dwindles to virtu-   floods downstream, particularly in India
as forest. pasture, or extensive or irsten-    ally nothing. A niumber of quiestions arise:    and Bangladesh. Policymakers responded
sieagricuilture. The first development    WVill fsituire productivitv in a particular    with attempts to control land use. They eii-
stag,e is a bhsis-fallow svstemi in which    svsteni be high or low? Will adequiate    couraged farmers to reforest and to make
thlere are no lanid productivity invest-   investments be madic in agricultuiral    heavv' investmeiits in costly' terrace con-
moents. but fallow periods are long and    research, environmyental axvareness, an-d    struction (beyond those areas already
pressuire is low enough to avoid degrada-    reguilation? And, in places where people   terraced by' farmers on their own). Impact
tion. Productivity losses bgnas grrowing    both recognize natural resource degrada-    was liniited, adoption rates were disap-
polpulation and market pressures lead to    tion and haive the technical know-boxy to    pointing, and poor subsistence farmers
shorter fatllowvs. Degradation continues    stop it. why,b does degradlation continue?  became scapegoats.
while the returnis to inivestinig in land im-  Answers to these questions depend        Recent analysis, however, sugg ests that
provements are n-ot worth the costs. At    heavril' on economnic forces and govermiient    geologic-al and climatic forces wvell beyond
sonme point, land becoimes scarce and its    policies. Despite the many attractive    the control of humans are the primary
page 2



Febnrary 1997
to describe the livestock population that    . aim  to determine  the  fun(lamental
Reclaiming the resource                  an area can sustainably support on a           scarcity relationships that are driving
base in Machakos                     constant basis. In sub-Saharan Africa, how-   resource use and transformations, and
ever, herders follow a strategy adapted to    . distinguish  them   from   degradation
In the 1930s, the hillyv lanlds
of Macbakeos were markQed by              both spatial and temporal rainfall variation.  caused by the various types of market
erosion, def;restation, low pmdodc-       Stocking rates increase in high rainfall      and policv failures.
de,  '  .            years to inake up for mortalitv and poor       For both sector analysis and project
tivityv and pasture degradation,
v  and pa vture    gra,n.             returns in low rainfall vears, and migration    appraisal, a range of analytic techniques,
People were poor and population
densopy was  considered greater           follows weather. In low  rainfall years,   includinig natural resource accounting
ity ws c              d gatr  -          rangelands mav appear overgrazed and    and other approaches to valuation, can
than the land could support. By
degraded bv temperate standards, mask-    help to identifv specific forms of market
19W, however, incomes were
higher, the resource base had re-        ing their ability to recover with higher    and policv failures and prescribe interven-
covered, and the real value of            rainfall. This opportunistic, disequilibrium  tions and reforms to promote better NRM.
agricultural output per capita had        strategy is opposite to that of the stable    The most important policy constraints
agseutural 3(tpuet peven caitahad         carrying capacity practiced in temperate    affecting the adoption of sustainable agri-
selatib n hadp quintuplev though          zones. Development efforts to settle the    cultural practices are insecure property
population had quintupled and             pastoralists and impose stocking rates    rights, externalities, failure of credit mar-
agrkulture had spread to even 
agricture    hIand sp oreadito evenfailed completely. Analysts trained in tem-    kets, and price, trade, and tax policies,
more marginal lands. According to
the study's authors, good roads          perate zones failed to understand the logic    detailed below.
offered access to a large Nairobi       of the indigenous svstem and proclaimed        Insecure property rights. Market
offered access to a large Nairobi
it unsustainable. Indeed. productivity mea-    failures  affecting  natural resources
thusincreasingthe rclturns to ag-         surements recorded in a single low rainfall   often relate to insecure, unclearly spec-
tiusitureasIngd   returns tof frm         vearwould indicate alackofsustainability.    ified, unenforceable, or nonexclusive
riculture. In addiltion, off-farm
income Xprovided resources for  Onlv writh continued observation can the    property  rights. Insecure  property
income D10Vicled resources for                                                   
land Limprovements such as ter-  svstem be understood (Behnke. Scoones    rights prevent resource  users from
and Kervin 1993: Biot et al. 1995).          planning use over time because future
races, planted trees and hedges,                                                       access is uncertain. A squatter fearing
and organic matter applications,          Operational Implications                     eviction has little incentive to invest
and local institutions developed         While investments are needed to rehab-    in maintaining or improving the land
as rules emerged for managing            ilitate degraded areas, well-structtured    because anv benefits from  develop-
communan  oremource  (English,            economic and policy analysis of natural    ment efforts would be lost along with
Tiffen, and Mortimore 1994).        resource use in agriculture is critical.   eviction. Similarly, a herder may over-
Country- and area-specific sector work    exploit grazing lands because he will
should:                                      gain all of the benefits while some
culprits of downstream land degradationi.    -precede or accompany preparation for    or most of the costs are absorbed bv
Floods in Bangladesh are highl,y correlated   a planned project,                          society as a whlole.
to downstream rain and less to rainf'all in
the bills, suggesting  that rainfall in                                       _
Bangladesh, not Nepal, triggers flooding.             Figure I. Induced innovation in natural resource management
More detailed studies demonstrated that                           Deonence   L              I  re           l
on naturally,               rehabilitation
deforestation is not as rampant as in pre-                        occurring                   and transition  Dependente on human-
vious (lecades, and that forest cover is                    R    resources   l Resource      I to intensive  l managed resources
actually increasing in many areas (Gill  ~~Total supply of  NR    (e.g., natural   degradation  Imanagement  (e.g., agroforestry, forest
actuallv iiicreasinor in manyv areas (Gill    services and        woodlands,    (e.g., deforestation)  (e.g., tree planting    plantations, managed
1995, Biot et al. 1995). The Himalaya are      products from a    management  I             I on farms, regulation I reserves)
the voungest, steepest, and most tnstable     given natural       of natural                  of communal
ountain range on ear,th, and as the recipi-   resource            vegetation)                 woodlands)
ent of intense monsoon rainfall, landslides                NR
and soil erosion result inore from geologi-    f (quantity,
cal factors than from poor NRM.                 quality,
A second example is inappropriate trans-       productivity
fer of rangelanid science from temperate                     l u
zones to the semi-arid tropical rangelands                                   I
of sub-Saharan Africa. In temperate envi-                                                             X * .      L         _ lll
ronments where rainifall patterns are                                        I               II
relatively stable, ranigeland scieintists de-                 to             ti             t2              t3              t4
s   1 1   r~~~~~~~~~~~~~~~~~~- Soucsh- -ad H-11 1994            Time, Population density,Market demand    >
veloped the concept of carrying capacity                                                        _
page 3



AGRAF Technology Notes
lPolicy analysis should first assess the    consequences of externalities are useful,      Price, trade, and tax policies. Some-
magnituide and sources of property right    especially to alert decisionmakers, but are    times governments control prices of
insecurity, ideintify the resource use man-    not an essential precondition. Moving    agricultural or forest products or otherwise
if estations  of insecurity,  and  assess    directlytotheidentificationofregulations,    restrict trade. These policies may affect
spontaneous and government responses.    information campaigns, or other techni-    NRM even if they are designed and imple-
Policy reforms and investments could    cal interventions is often possible. Where    mented for other purposes. For example,
include provision of i ndivicdu al or grouIp    specific inputs such as agrochemicals are    restrictions against cutting trees on private
titles, legal recognition of traditional claims    involved. taxes, fees, and other market in-    land or on transporting timber -  intro-
and conflict resolution methods, and per-    terventions can be low-cost xvavs to    duced in India, Thailand, and elsewhere
lhaps insurance sclieiies and cost-sbaring    promote reduced applications. Sometimes    to halt deforestation -  reduce the profit-
schelmies to reduce the costs of conserva-    a combinationj of policv interven1tion1s    ability of growiing trees on farm  lauid
tion and fanrin development to target    is most effective. Introduction of new           without addressing the root causes of de-
beneficiaries. Experience has show,n that    technologies such as integrated pest man-    forestation on government forest lands
it is crucial to recognize that 'modern' pri-    agement-even though thev may involve    (Chambers et al. 1989). In Brazil, exempt-
vate tenure svstemis are not uiliiversallv                         )     I
vate tenure systems arenotuniversally increased public expenditures -  mav be    ing agricultural income from  taxation
superior to traditional or gyroup ovniership,
supe.ior to traditional or group ownership,    the best way to deal with some externali-    encouraged people to clear the forest
'sp)ecially in marginal environments,        ties. Transaction costs are almost alwavs    for agricultural production. Pesticide
Externalities. Spillovers that occur         ..                 .......                       t
Externalities. Spi.lovers that .occur    high when dealing with externalities, and    subsidies in several countries encouraged
when costs or benefits from the economic
an important role for economic analvsis is   their overuse, contributing to high disease
activity of one person accrue to some-
one else or to society at large withot l to ensure that poliev interventions are ef-    rates and even death in some intensive
onte else or to societv at large without 
compenstion 'arcalle  extalit. . fective and efficient and do not introduce    agricultural areas.
comnpensation are called externalities.
Thex, imaytv he beneficial or- hiarmiiiful,    additional undesirable distortions.          Removal of distortionary policies invari-
Failure of credit markets. A common        ably involves a political process that pits
agricultural chemicals, siltation of wsater-    characteristic of natural resource invest-   the privileged elite against broader social
w7 av's from upland erosion, and! overfishing7  ments is their long gestation period.    welfare. Few guidelines are available, but
are primne examples. The g7reenhlouse ef-    C^rowing trees and conserving soil, for ex-    well-stnictured analysis, wide dissemina-
feet of largre-scale tropical deforestationi    ample, offer only distant payoffs. In    tion of results, and linking investments
amnd redlueed profitability of grazing on    addition, many natural resource invest-   to reforms and policy dialogue with a
open access pastures are others. WNithout    ments, such as terraces or irrigation wells,    wide range of stakeholders, can all be
government intervenTtion or other coop-    have high initial costs. \Vhen costs are high    part of the process to improve policy en-
erative action. those who impose harmful    or returns are not immediate, credit may    vironments in support of sustainable
extemalities have little incentive to change  be necessary for investors, but conventional    agriculture. Experience has clearly show, n
thteir wavs,                                 banking services are often inaccessible to    that favorable economic conditions and
A wide range of market-based and regu-    natural resource managers. Sometimes    good policies and institutions need to be
latorv  (often  termed  'command  and        credit market failures are compounded by    in place before farmers adopt sustainable
control') instrulTents can be used to    insecure property rights -creditors may    agricultural practices. These lessons are
redluce thie negative effects of externali-    hesitate to lend to people without reliable    the cornerstone of any effort to promote
ties. Valuation  exercises to express the    collateral such as land titles.              sustainable agriculture.
References                                                          Gill, GC. 1995. "Major Natural Resource Managenment Concerns in Sooth
Beihuke. H.H., 1. Scoones. andl C. Kenen. 1993. Range Ecology at Disequilib-  Asia." Prepared for the Conference on a 2020 Vision for Agricuilture.
rit HzV N,t' Models of Natur-^al Variability and Pastoral Adaptation in  Washington, D.C.: IFPRI.
.Ajricas Saarannas. London: Overseas Developmnent Institute.     Oldemaii, L.R., R.T.A. Hakkeling, and W.G. Sombroek. 1991. "World Map
Biot, Y.. P Blaikie, C. Jacl;son. anld B. Palmer-Jones. 1995. 'Rethinkinlg  ofthe Status of'Human-Induced Soil Degradation: An Explanatorn Note."
Research ori LaLnd Degradationi in Developing Countries." 5Vorldl Bank  Revised ed. XVageningen and Naiirobi: International Soil Reference and
DiSclSsi(n Paper No. 289. WXashington, D.C.: W,Vorld Bank.          Information Centre and United Nations Environment Programme.
Chalmbeicrs, Robert, NC.. Saxena, aisd Tislsaar Shah. 1989. lo the Hands ofthe  Pierce, F., R. Dowdy, 'W Larson, and W Graharn. 1984. "Soil Pro(luctivity in
Poor' Wu'itr andI Trees. New Delhii: Oxford and IBH.                the Corr Belt: An Assessment of Erosion's Long-Term Effect "Jooirnal
Eiielislm, J.. M. Tiflen. andN M. Mortirnore. 1994. "Land Resolurce Manage-  of Soil and Water Consecration 39(2):131-136.
mn1C:it in Machakos District, Kenya, 1930-1990." b'orld Bank Environment  ScherT, S. and E Hazell. 1994. "Sustainable Agnciltural Developmcent Strategies
Paper No. 5. \Washington. D .C .: orld Bank.                        in Fragile Lands." EPTD Discuission Paper lo. 1. NWashington, D.C.: IFPRI.
Agriculture Technology Notes provide up-to-date infornation on selected agricultural topics to World Bank professionals. This Note was
prepared byJohn Kerr (IFPRI) and Willian Magrath (YWorld Bank). To obtain copies of this or other Notes, orto enquire about the series,
please call (202) 473-9406. Notes: in this series indlude Sustainable Agricultre, Reform of Seed Reglatioms, Information Technology and
Rural Devlpment, SimpleAnimal Health Techniques, Cmpetitive Research Grant Systems, Provision ofMineral Supplewnts, Integrated
Pest Management, Privatiztion ofAnimnaHeaith Services, Wonrd Fisheries, Natoa Seed Systens, Participatory Rural Assessment, Integrated
Soil Managemnentfor the Tr, Biotechnology, and Livstock Production an  sustallnale Resource Use. Edited by Seth Beckernnn,
Business and Technical Communnkations, Pittr, PA. Lyut by Tony Condello, Condello Design, Pittsburgh, PA
page 4