Embed Size (px)
Citation preview
This article was downloaded by: [FU Berlin]On: 19 November 2014, At: 06:26Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number:1072954 Registered office: Mortimer House, 37-41 Mortimer Street,London W1T 3JH, UK
Development Southern AfricaPublication details, including instructions forauthors and subscription information:http://www.tandfonline.com/loi/cdsa20
The relative importanceof liquidity and otherconstraints inhibiting thegrowth of small‐scale farmingin KwaZulu‐NatalLouise J Fenwick a & Mike C Lyne ba Research Assistant, Department ofAgricultural Economics , University of Natal ,Pietermaritzburgb Professor, Department of AgriculturalEconomics , University of Natal ,PietermaritzburgPublished online: 27 Feb 2008.
To cite this article: Louise J Fenwick & Mike C Lyne (1999) The relativeimportance of liquidity and other constraints inhibiting the growth of small‐scalefarming in KwaZulu‐Natal, Development Southern Africa, 16:1, 141-155, DOI:10.1080/03768359908440066
To link to this article: http://dx.doi.org/10.1080/03768359908440066
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of allthe information (the “Content”) contained in the publications on ourplatform. However, Taylor & Francis, our agents, and our licensorsmake no representations or warranties whatsoever as to the accuracy,completeness, or suitability for any purpose of the Content. Any opinions
and views expressed in this publication are the opinions and views ofthe authors, and are not the views of or endorsed by Taylor & Francis.The accuracy of the Content should not be relied upon and should beindependently verified with primary sources of information. Taylor andFrancis shall not be liable for any losses, actions, claims, proceedings,demands, costs, expenses, damages, and other liabilities whatsoeveror howsoever caused arising directly or indirectly in connection with, inrelation to or arising out of the use of the Content.
This article may be used for research, teaching, and private studypurposes. Any substantial or systematic reproduction, redistribution,reselling, loan, sub-licensing, systematic supply, or distribution in anyform to anyone is expressly forbidden. Terms & Conditions of accessand use can be found at http://www.tandfonline.com/page/terms-and-conditions
Dow
nloa
ded
by [
FU B
erlin
] at
06:
26 1
9 N
ovem
ber
2014
Development Southern Africa Vol 16, No 1, Autumn 1999
The relative importance of liquidityand other constraints inhibiting thegrowth of small-scale farming inKwaZulu-Natal
Louise J Fenwick & Mike C Lyne1
A logit model is used to examine the extent of liquidity constraints relative to other con-straints inhibiting small-scale farming in KwaZulu-Natal. These other constraints includepoor access to land, labour and information, and high transaction costs. Data for theanalysis were sourced from two rural districts in the former KwaZulu. The results suggestthat liquidity is important, while imperfect land markets, information costs and hightransaction costs are also significant inhibiting factors. Investments in literacy and lan-guage skills, vocational training, and business and financial management skills may im-prove income opportunities for rural people and hence enhance their ability to invest,save and borrow. Better roads, telecommunications and legal institutions are also re-quired to realise the full benefit of investment in extension and credit services.
1. INTRODUCTION
In KwaZulu, small or emerging farmers can be broadly viewed as representinghouseholds for which fanning revenue comprises only a portion of their grossincome. Farm output ranges from near sub-subsistence to the production of amodest marketable surplus. Development and growth of small-scale agricultureare perceived to be constrained primarily by liquidity levels (Strauss Commis-sion, 1996).
Partly as a result of this perception, rural financial services are currently underreview in South Africa with the aim of incorporating previously disadvantagedpeople more fully. International authors, such as Adams (1971), Bottral (1974)and Von Pischke & Adams (1980), highlight improved access to credit as animportant tool for overcoming liquidity constraints. This view progressed fromadvocating the provision of state-subsidised credit to recognising the importanceof savings mobilisation (Adams, 1992).
1Respectively Research Assistant and Professor, Department of Agricultural Economics,University of Natal, Pietermaritzburg. The authors gratefully acknowledge financialsupport from the University of Natal Research Fund, the Centre for Science Develop-ment and the Development Bank of Southern Africa. Opinions expressed in this articleare those of the authors and do not necessarily reflect the views of the sponsors. Recog-nition is also given to Paula Despins, who updated parts of the Mpembeni sample data.
141
Dow
nloa
ded
by [
FU B
erlin
] at
06:
26 1
9 N
ovem
ber
2014
LJ Fenwick & MC Lyne
Proponents of New Growth Theory alternatively argue that physical and humanresources cannot be fully utilised unless incentives are created by the prevailingeconomic and political institutions (Olson, 1996). Fundamental elements of theseinstitutions include property rights, and the legal and regulatory framework (DeGoiter & Swinnen, 1994). While several South African studies have identifiedfactors other than credit that may influence the growth of small-scale farmers(Coetzee, 1995; Lyne, 1992), no attempt has been made to quantify the impact ofextended credit services on homeland agriculture in relation to these other con-straints. Quantifying the relative importance of constraints faced by farmerscould aid in the planning and prioritising of investments aimed at facilitating thedevelopment of small-scale farming (Alwang et al, 1996).
This study uses a logit model to identify and prioritise liquidity and other con-straints facing small farmers, with data drawn from household surveys conductedin two districts of the former KwaZulu during 1995/6. The following sectionexamines possible constraints facing small-scale farmers in more detail, whileSection 3 briefly describes the study areas. Section 4 presents the logit model andits results. The article concludes with comments in Section 5.
2. FARMING CONSTRAINTSPossible constraints to the growth of small farmers include
• tenure insecurity and the consequent absence of efficient land markets (Thom-son & Lyne, 1993)
• liquidity, including low cash income, and access to credit and savings (Chris-tensen, 1993; Udry, 1995)
• a shortage of quality labour (Nattrass & May, 1986)• information costs (Delgado, 1996)• high transaction costs. Transaction costs affect access to markets for inputs,
products and contractor services and vary with gender, education, off-farm mi-grancy, length of residency in the area and proximity to formal markets (Low,1986: 124; Lyne, 1992; Coetzee, 1995).
2.1 Tenure insecurity and the consequent absence of efficient land marketsHouseholds in the former KwaZulu do not possess title deeds to their land(Thomson & Lyne, 1993). A tribal authority assigns use rights to the householdhead in accordance with traditional laws and customs. Typically, the householdhas exclusive use rights to cultivated land and communal use rights to grazingland. Customary law precludes the legal operation of a sale market for agricul-tural land in the former homeland.
Nieuwoudt (1990) reported that 80 per cent of arable allotments in KwaZulu aresmaller than 2 ha. When compared with wage employment for skilled workers,farms of this size cannot produce attractive incomes even under optimal techno-logical conditions. Returns are low on very small farms owing to the presence of
142
Dow
nloa
ded
by [
FU B
erlin
] at
06:
26 1
9 N
ovem
ber
2014
The relative importance of liquidity and other constraints inhibiting growth
fixed costs, such as the indivisible costs of family labour, management and in-formation input. There are also numerous fixed transaction costs. Renting wouldallow farm consolidation and increase potential earnings from farming for thelessee, while simultaneously creating a higher income stream for the lessor. Lyne& Ortmann (1992) show that access to more land (rental of land) is one of themost important characteristics of households that did not default on loans inKwaZulu. Unfortunately, there is little evidence of an active rental market forfarmland in KwaZulu (Thomson, 1996: 17).
Arable land lies idle over large parts of KwaZulu, while some households arerecorded as farming intensively. Thomson (1996: 23) attributes the underutilisa-tion of arable land to an imperfect land market. Imperfections may also occur inmarkets for contractors, credit, inputs and products. These market failures maynot, however, be responsible for the underutilisation of land. If the rental marketfor land was active, land controlled by households that were without access tofunds, labour or contractors would transfer to those households that were notconstrained. Alternatively, in the presence of a land rental market, the householdcould use interlinked contracts to overcome imperfections in other markets.
Prerequisites for a rental market in land are tenure security, and low transactioncosts relative to rental income. Transaction costs are increased by risk andhouseholds perceive renting to be risky as they believe they could lose their landpermanently (Lyne & Thomson, 1997). As a result, potential lessors prefer not tolease out land, as land offers security against illness, unemployment and old age.Thomson (1996: 83-6) facilitated a series of'adaptive strategies' involving smallincremental changes to customary institutions to encourage land rental transac-tions in parts of the former KwaZulu. The initial strategies were aimed at reduc-ing risk: tribal authorities granted permission for the promotion of a rental mar-ket for arable land; administrative and dispute procedures were established andclarified; and local chiefs called community meetings to voice their approval ofrental transactions and explain procedures. The local radio station, Radio Zulu,was used extensively to promote the process, advertise workshops and commu-nicate their outcomes. As a result of these measures, the number of householdswilling to lease'out land (potential lessors) increased significantly.
Unfortunately, these changes did little to encourage potential lessees. Thomson(1996: 30) found that households' exclusive rights to arable land were not guar-anteed against cattle invasion. Participants in the workshops identified poorlyenforced grazing rules and the consequent need to protect cultivated lands as themost important constraint to crop production. This constraint was ranked aboveimperfect credit markets and unreliable contractor services. In short, insecure landtenure caused by cattle invasion discouraged farmers from hiring additional land.
Thomson (1996: 100-5) addressed this problem by negotiating and publicising aspecific date after which all livestock would be removed from arable lands forthe duration of the growing and harvesting season. A system of fines and com-
143
Dow
nloa
ded
by [
FU B
erlin
] at
06:
26 1
9 N
ovem
ber
2014
LJFenwick & MC Lyne
pensation was introduced to penalise offenders. Rental transactions flourishedafter these interventions, with improvements in both equity and farming effi-ciency. The picture that emerges is that growth of small-scale farmers in Kwa-Zulu-Natal may be constrained more by an imperfect rental market for land thanby low levels of savings or poor access to credit.
2.2 Financial restrictionsEmpirical studies suggest that small farmers in South Africa are constrained bylow and irregular income which reduces their ability to save, borrow and investin agriculture (Lyne & Ortmann, 1992: 20-2). Short-term solutions to this prob-lem include using savings, selling assets or accessing credit.
Savings in rural areas may be used to smooth out income fluctuations over. time.Savings can, however, also act as an insurance substitute for rural households.Money saved specifically as an insurance substitute, especially in informal in-stitutions such as burial societies, is not fungible and therefore cannot be used toimprove cash flow. Another option for smoothing out cash flow is the sale ofassets such as jewellery and livestock. According to Low (1986: 111-5), smallfarm households in southern Africa also use cattle and smaller livestock as aform of liquidity to smooth out income fluctuations.
Accessing formal production credit to smooth out income fluctuations is difficultowing to collateral and debt-servicing problems. When applying for credit, col-lateral is usually required because information about borrowers is costly andlargely unavailable to lenders. Agricultural land had no value as collateral in theformer KwaZulu. Legal developments, such as private land rights, create a foun-dation for using land as collateral but these rights must be secure if they are tominimise transaction costs (Johnson, quoted by Thomson & Lyne, 1993). Privateland rights alone, however, do not bestow collateral value on land. Land hasvalue as collateral only if there is a sale market for it or a secondary market forlong-term leases. Title deeds do not give land collateral value unless they assistmarket transfers.
In some countries, lenders have emphasised collateral substitutes, eg savings.The Grameen Bank of Bangladesh's successful programme supplying loans tosmall-scale entrepreneurs used savings as collateral (Pitt & Khandker, 1995).However, it cannot just be assumed that Grameen Bank strategies can be easily
• transplanted to rural South Africa as they were based on a system of small, cohe-sive groups of woman engaged in non-farm activities that facilitated regularsavings deposits. In contrast, small-scale farmers in South Africa are widely dis-persed and farm income irregular. Nevertheless, formal savings may still be animportant collateral substitute as they do not require regular cash commitmentsand so may be suitable for rural households.
An important aspect of obtaining a loan is the ability to repay the principal, withinterest, usually in regular payments. The debt-servicing capacity, or liquidity, of
144
Dow
nloa
ded
by [
FU B
erlin
] at
06:
26 1
9 N
ovem
ber
2014
The relative importance of liquidity and other constraints inhibiting growth
the applicant therefore plays a vital role in determining the success of a creditapplication, as the lender seeks to minimise the risk of loan default. Formalcredit institutions often demand a predetermined, constant stream of repayments(Devereux et al, 1989) which may be difficult for a small-scale farmer to meet.Lugemwa & Darroch (1995) found that off-farm income is a significant determi-nant of a small-scale farmer's ability to secure and repay seasonal loans. Conse-quently, small farmers without off-farm income may find it difficult to borrow inthe formal sector. Extending credit services to include more small-scale farmersmay not therefore effectively address their liquidity problems.
2.3 Labour
Mincer (1963: 71) noted that 'opportunity costs (ie earning powers in the marketand marginal productivity [earning powers] in alternative pursuits) differ amongfamily members'. Viewed against Nieuwoudt's (1990) statement that potentialprofits are low on small farms, it is clear that the more skilled and mobile mem-bers of households have a competitive advantage in off-farm employment. Nat-trass & May (1986) have shown that labour migration from rural areas in SouthAfrica is age, gender and skill specific. The vast majority of migrant workers areyoung and relatively well-educated men, and this loss of quality of labour hasadverse implications for productivity on small-scale farms (Low, 1986: 52).
2.4 High transaction and information costs
Transaction costs reflect the degree of development and maturity of the financialsystem, the available transport and communication facilities, and the efficiencyof the legal system (Coetzee, 1995). Transaction costs can be divided into exante and ex post costs. Drafting and negotiating an agreement are examples of exante costs. The problem of moral hazard in contracts leads to ex post costs,which include contract enforcement and risk. Risk is increased when propertyrights are not secure and when there is uncertainty about legal enforcement ofcontracts (Lyne & Thomson, 1997). While market forces can induce efficiencyin the private sector, efficient markets require public sector investment in bothphysical and legal infrastructure (Timmer, 1992; Sahn & Sarris, 1994).
Information costs fall with declining transaction costs. Taiwan, a success story inrural financial markets, reduced transaction costs by investing in infrastructure(Adams et al, 1993). For example, telephones reduce the travelling and waitingtime of credit applicants and assist the banker in validating the creditworthinessof the applicant more easily. This improves the level of competition, therebyreducing financial charges in rural credit markets (Hoff & Stiglitz, 1990).Fitschen & Klitgaardt (1996) found a strong relationship between per capita in-come and infrastructure among tribal wards in KwaZulu-Natal, with the poorestwards being those farthest removed from markets and main roads. Similarly,Alwang et al (1996) highlight the link between rural poverty and isolation frominfrastructure.
145
Dow
nloa
ded
by [
FU B
erlin
] at
06:
26 1
9 N
ovem
ber
2014
LJ Fenwick & MC Lyne
Extension services supply information about improved agricultural technology tofarmers in order to increase their productivity. Studies, for example by Alwang etal (1996), have illustrated the value of disseminating knowledge via extensionsupport. These services will not, however, produce a significant response withoutthe correct institutional setting. Sachs (1996) concludes that government fundsshould be concentrated on investment in true public goods such as roads, tele-communications and postal services, as well as establishing a dependable judicialsystem to support security of property rights and uphold commercial contracts.
While transaction costs may be relatively low in informal markets, this does notimply that they are low in absolute terms. Transactions in informal markets tendto be highly personalised, with unique terms negotiated by both parties. Berry(1993) frequently refers to the influence of gender and social standing on theoutcome of disputes, with women facing greater uncertainty. In contrast, trans-actions in formal markets are fairly impersonal. Each applicant is assessed on thesame economic criteria: collateral, debt repayment capacity and transactioncosts. A contract is drawn up and is legally binding on both parties, and disputesare handled by an impartial third party.
Transaction costs also vary between members of the same rural household. Amale who works in an urban area will most probably have a relatively high levelof education (Low, 1986: 124), a steady wage, and contact with formal institu-tions such banks and shops. As a result, he will face lower transaction costs inboth formal and informal markets. This contrasts with a married woman whorelies on irregular wage remittances and seasonal crop income. Delgado (1996)describes the constraints faced by female farmers in Africa, which include inse-cure land rights, limited access to common property resources, lack of equipmentand appropriate technology, and lower education. Collaborating evidence fromparts of South Africa show that on average, adult women in rural KwaZulu haveonly 2,9 years of formal education (Lyne, 1992) and live 40 km from sources offormal credit (Coetzee, 1995). It follows that people who live and farm in ruralareas (mainly women) may encounter relatively high transaction costs, even inlocal markets. Consequently, when engaging in credit transactions they tend toborrow from friends and family in order to minimise costs.
3. STUDY AREAS
Data for the study were gathered from 150 small farm households in two com-munal areas of KwaZulu-Natal. Seventy-five households were sampled in eachof two study areas during 1995/6. The first study area - Mpembeni ward - is aremote settlement located in the Hlabisa magisterial district, west of theHluhluwe game reserve. The topography is broken and rainfall unreliable. Cropsare grown largely for household consumption, with some local sales of maizeand vegetables in particular. Households are often far removed from their allot-ted arable lands. The second area - Mkhwanazi ward - is situated in the OngoyeII magisterial district near Port Durnford, just 15 km from the Empangeni-
146
Dow
nloa
ded
by [
FU B
erlin
] at
06:
26 1
9 N
ovem
ber
2014
The relative importance of liquidity and other constraints inhibiting growth
Richards Bay industrial complex. The topography is flat and rainfall high but thesoil is generally poor, with sugar cane being the principal cash crop.
Farm sizes reported by respondents averaged just 2 ha, with less than one-half ofthe respondents cultivating all of their arable land. Uncultivated lands lie idle - asymptom of an imperfect rental market for cropland. Relatively few samplehouseholds participated in rental transactions. Whereas 15 per cent hired land,less than 10 per cent admitted to renting land out. These observations lend sup-port to Thomson's (1996: 83-6) argument that households may be reluctant tolease out land for fear of losing it permanently or unwilling to hire additionalland that they are unable to police. More than one-third of respondents sufferedcrop damage.
Crop incomes are low: mean annual cash incomes of R323 and R971 were esti-mated for Mpembeni and Mkhwanazi respectively. By contrast, annual cash re-ceived from off-farm activities (including non-farm enterprises) averagedR16 013 in Mpembeni and R15 825 in Mkhwanazi. On average, households inMpembeni owned six cattle and fifteen small livestock, while Mkhwanazihouseholds owned three cattle and eight small livestock. Although many house-holds did not cultivate all their land, 68 per cent held savings with commercialbanks, averaging Rl 658 in Mpembeni and Rl 284 in Mkhwanazi. This suggeststhat rural households do participate in formal financial markets, albeit at a lowlevel, and that emerging farmers are constrained by factors other than internal orexternal credit rationing.
The average sample household has nine family members, of whom almost halfare children. Education levels are low, with de facto heads completing just fouryears of formal schooling in Mpembeni and five years in Mkhwanazi. The aver-age household in the Mpembeni ward is located more than 83 km from formalmarkets and has 1,2 migrant workers. Conversely, the average household in theMkhwanazi ward is fewer than 25 km from formal markets and has 1,5 migrantworkers. Transaction costs in formal markets are therefore expected to be lowerin Mkhwanazi.
4. EMPIRICAL ANALYSIS OF FARM CONSTRAINTS
The empirical model hypothesised that growth of small-scale farming is con-strained largely by factors other than liquidity, including imperfect land markets,inadequate family labour (quality and quantity), poor information and high trans-action costs. Under these conditions, improving farmers' access to credit mayhave little impact on farm production.
A proxy (dependent) variable measuring the intention to farm was developedusing expenditure on seasonal farm inputs (including the opportunity cost of owndraught power) to the value of R800 or more, the ownership of farm assets, theadoption of sugar cane or timber enterprises, and crop sales greater than R300.The inclusion values for expenditure and crop sales were chosen following
147
Dow
nloa
ded
by [
FU B
erlin
] at
06:
26 1
9 N
ovem
ber
2014
LJ Fenwick & MC Lyne
Jiisenbeis' (1977) method of observing a natural break in the frequency distribu-tion. Households displaying one or more of these characteristics were classifiedas 'farmers' and scored a value of 1 on the dependent variable. Conversely,households that reporced no expenditure on farm inputs or that did not cultivateany of their arable land were classified as 'non-farmers' and scored 0 on the de-pendent variable. Minor adjustments were made to the groups identified in theMpembeni ward after the initial classification had been checked by a researcherfamiliar with the area (Despins, 1998).
Owing to the discrete nature of the dependent variable and the problems this pre-sents when applying Ordinary Least Squares regression, a logit function wasestimated using the maximum likelihood technique. The independent variables(and their expected signs) included in the model are summarised in Table 1.
Farm income was excluded from the liquidity variable because it was used as acriterion for distinguishing farmers from non-farmers (ie, it was used to definethe dependent variable). Remittances, pensions and microenterprise income are,in any case, far more important and usually more regular sources of liquidity.
Family labour {Family lab) was expressed as the ratio of household labourequivalents available for farm work relative to household size. The numerator(family labour equivalents) was squared to capture the effect of complex coop-eration in farm production. Complex cooperation describes the increasing effi-ciency in production as more family labour becomes available (Low, 1986: 29).The denominator controls differences in household size.
Tcost is an index of transaction costs computed using exogenous variables. Eachvariable was standardised to give it equal weighting in the index. Householdswith high index scores face low transaction costs. To ensure that this conditionheld true, a constant of 100 was added to each standardised variable to eliminatenegative values.
Following the theory outlined in Section 2.4, a male de facto household headscores 1 on the gender dummy variable because he usually faces lower transac-tion costs than a female who scores 0. Similarly, household heads with moreyears of formal education are expected to face lower transaction costs. The trans-action cost index considers only the interaction effects of these variables, as mi-gration is both gender and skill specific. Interactively, a well-educated malefaces lower transaction costs than an educated female. A longer length of resi-dency in the area suggests higher social standing and lower transaction costs ininformal markets. Likewise, households with migrant workers should face lowertransaction costs in formal markets. A log transformation was applied to thenumber of migrant workers present in a family because transaction costs are ex-pected to fall marginally with each additional migrant. This necessitated addingthe value of 1 to the number of migrant workers in each household before the logtransformation to eliminate problems of a zero log transformation.
148
Dow
nloa
ded
by [
FU B
erlin
] at
06:
26 1
9 N
ovem
ber
2014
The relative importance of liquidity and other constraints inhibiting growth
Table 1: Independent variables included in the logit model
Variable Expected sign Description of variable
Size
Rent
Interlink
Liquidity
Fsavings
Isavings
Family lab
Visits
Tcost
Dependency
Size of allocated arable land (ha)
Dummy scoring 1 for households that hire additional land, and
0 otherwise
Dummy scoring 1 for households with interlinked contracts,
and 0 otherwise
A liquidity indicator calculated as the sum of annual off-farm
and non-farm income plus the monetary value of livestock
owned (rand)
Formal savings level (rand)
Dummy scoring 1 for households participating in informal
savings groups, and 0 otherwise
A ratio indicating the amount of family labour available per
household member. It is calculated as the square of household
labour equivalents1 (calculated as adults - wage employees +
0,5 [children + pensioners]) divided by total household size
Number of visits by an extension officer
Transactions cost index calculated as the standardised values2
of each variable in the index, ie {[(gender' of de facto head) *
(education of de facto head)] + (length of residency in the
area) + (log [1 + number of migrant workers])} / {(district
dummy4) * (car ownership5)}
Ratio of infants and school children to wage earners and pen-
sioners
Notes:
'The household labour equivalent is squared to account for complex cooperation effects.2A constant of 100 was also added to each standardised variable to eliminate negative values.
'Gender is a dummy variable scoring 1 for males and 0 for females.4A district dummy scoring 1 for the Mpembeni ward and 0 for the Mkhwanazi ward.5Car ownership is a dummy,variable scoring 1 for non-owners and 0 for owners.
The denominator of Tcost reflects interaction between the distance householdsmust travel to participate in formal markets, as captured by the district dummy,and ownership of a car. The impact of distance on transaction costs is most rele-vant when a household does not own a motor vehicle. Households that owned aserviceable vehicle scored a 0 on the dummy variable whereas those that did notscored 1.
Figure 1 shows that Tcost is a continuous variable and that average transactioncosts are higher in the Mpembeni ward (1,01) than in the Mkhwanazi ward(1,03).
149
Dow
nloa
ded
by [
FU B
erlin
] at
06:
26 1
9 N
ovem
ber
2014
LJ Fenwick & MC Lyne
Figure 1: The distribution of Tcost (n = 86, mean = 1,026)
Table 2 presents the mean value of each independent variable computed for non-farmers (dependent variable = 0) and farmers (dependent variable = 1). The re-sults of independent t-tests across the means are also tabulated. As expected,households with higher levels of farm income and expenditure had larger farmsor were able to hire additional land.
Somewhat surprisingly, non-farmers are much less liquid than farmers. Formalsavings, which may be regarded as collateral or a (minor) source of liquidity, var-ied significantly across the groups. Of course, formal savings held by rural house-holds are not always accessible to those family .members engaged in farming.The incidence of informal savings was similar in each group, possibly illustratingthe need for both groups to employ informal savings as an insurance substitute.
The comparison of group means also shows farmers to be better endowed withfamily labour, even though non-farmers have similar dependency ratios. As ex-pected, farmers had more contact with extension staff and faced much lowertransaction costs than non-farmers. The district dummy shows that householdssampled in the Mpembeni ward accounted for around 40 per cent of both farmersand non-farmers. The implication is that Hlabisa is under-represented in the sub-sample and that differences observed between farmers and non-farmers are notdependent on location.
150
Dow
nloa
ded
by [
FU B
erlin
] at
06:
26 1
9 N
ovem
ber
2014
The relative importance of liquidity and other constraints inhibiting growth
Table 2: Comparison of group means
Variable
Size
Rent
Interlink
Liquidity
Fsavings
Isavings
Family lab
Visits
Tcost
District
Dependency
Non-farmersValid obser-
vations
47
47
44
43
40
47
47
47
42
47
47
= 0
Mean
1,25.
0,09
0,07
11947
1 307
0,36
1,90
0,55
1,02
0,36
1,69
Farmers =Valid obser-
vations
51
51
45
46
45
51
51
51
44
51
51
1
Mean
3,09
0,22
0,07
36 678
2 429
0,31
2,50
1,29
1,03
0,41
1,81
t
3,80
1,81
0,03
3,82
1,00
0,50
1,99
2,32
1,62
0,41
0,48
sigt
0,000*
0,074*
0,974
0,000*
0,323*
0,620
0,049*
0,023*
0,005*
0,616
0,632
Note: * T-value recorded as greater than or equal to 1.
Variables with differences in group means of less than 1 on the t-test were ex-cluded from the logit model on the basis that such variables would not contributesignificantly to the analysis. Results of the logit analysis are presented in Table3. The contribution made by each independent variable was evaluated using thelikelihood ratio test (Menard, 1995).
The estimated model predicts 85 per cent of the sample cases correctly and itschi-squared statistic is significant at the 1 per cent level of probability. The nullhypothesis that the independent variables are not related to the dependent vari-able can therefore be rejected. RL
2, at 56 per cent, indicates that the proportionalreduction of the absolute value of the log-likelihood was greater than half.Goodman and Kruskal's tau-p statistic (Menard, 1995), which indicates the re-duction of classification error, is 0,71 and is significant at the 1 per cent level ofprobability. The model therefore reduces classification error by almost three-quarters.
The signs of the estimated coefficients are in agreement with a priori expecta-tions. The standardised coefficients rank the constraints in order of severity, al-though the loading estimated for Tcost may be understated as the transaction costindex is understandably positively correlated with liquidity. From a policy per-spective, small farmers are constrained by low incomes and savings, inadequateinformation, small farm sizes - compounded by insecure land tenure and risksassociated with rental transactions - and high transaction costs.
5. CONCLUSION
The results of the logit analysis confirm earlier findings that small farmers in
151
Dow
nloa
ded
by [
FU B
erlin
] at
06:
26 1
9 N
ovem
ber
2014
LJ Fenwick & MC Lyne
Table 3: Results of the logit model
Standardised
coefficients
-2 Log LR
Significance
Classification rates
% correctly classified
Number of cases
RL2
Model chi-squared
Goodman and Kruskal
Size
0,807
6,1815
0,0090
's tau-p
Explanatory variables
Rent
0,687
2,660
0,1029
Li-
quidity
2,339
14,070
0,0002
Non-
Family
savings
0,372
1,641
0,2002
farmers
86
Family
lab
0,842
4,512
0,0337
70
0,54
44,62***
0,71***
Visits
0,851
6,771
0,0093
Farmers
83
Total
cost
0,441
1,133
0,2871
Note: * Significant at the 1 per cent level of probability.
KwaZulu-Natal are severely constrained by low levels of liquidity which restrictinvestment in farm inputs, including hired labour. It is plausible that institutionalcredit might help alleviate liquidity problems, but successful participation in fi-nancial markets is unlikely while rural incomes are low. As rural householdsearn most of their income from off-farm sources, investment in literacy and lan-guage skills, vocational training, and business and financial management skills isexpected to improve their income opportunities and hence their ability to invest,save and borrow.
The results also identified farm size as a constraint to farming. Households withlarger arable land allotments are more likely to farm intensively. Promoting anefficient rental market for land would allow land-poor farmers to increase thescale of their operations, raising the incomes of both farmers and potential les-sors. In addition, households endowed with land but little else could securecapital, labour and contractor services through interlinked contracts. Other stud-ies in KwaZulu have shown that an efficient rental market will require more ex-clusive and assured rights to arable land, as well as a judicial system that can berelied upon to uphold contracts. Adaptive strategies, involving small incrementalchanges to customary institutions, appear to be an appropriate way of achievingthis goal.
Poor access to information (visits) and high transaction costs are further impor-tant constraints to small farmers. The implication is that government could alle-viate binding constraints by investing in real public goods, such as physical andinstitutional infrastructure. This includes developing roads, telecommunications,electricity and postal services in rural areas, as well as taking steps to ensure
152
Dow
nloa
ded
by [
FU B
erlin
] at
06:
26 1
9 N
ovem
ber
2014
The relative importance of liquidity and other constraints inhibiting growth
consistency between rulings in tribal and national courts, particularly those re-lating to commercial contracts involving women married under customary law.
Of course, local researchers should not discount the importance of financial mar-kets in developing regions. On the contrary, demand for credit is likely to growas more farmers emerge, increasing the value of information about successfulborrowers, lenders and financial technologies.
REFERENCES
ADAMS, DW, 1971. Agricultural credit in Latin America: a critical review ofexternal funding policy. American Journal of Agricultural Economics, 53(2):163-72.ADAMS, DW, 1992. Building durable rural financial markets in Africa. AfricanReview of Money, Finance and Banking, 1: 5-14.ADAMS, DW, CHEN, HY & LAMBERTE, MB, 1993. Differences in uses ofrural financial markets in Taiwan and the Philippines. World Development,21(4): 555-63.ALWANG, J, SIEGEL, PB, & JORGENSEN, SJ, 1996. Seeking guidelines forpoverty reduction in rural Zambia. World Development, 24(11): 1711-23.BERRY, S, 1993. No condition is permanent: the social dynamics of agrarianchange in sub-Saharan Africa. Wisconsin: University of Wisconsin Press.BOTTRAL, AF, 1976. Financing small farmers: a range of strategies. In Hunter,G, Bunting, AH & Bottral, AF (Eds), Policy and practice in rural development.London: Croom Helm.CHRISTENSEN, G, 1993. The limits to informal financial intermediation.World Development, 21(5): 721-31.COETZEE, G, 1995. Credit. In Singini, R & Van Rooyen, J (Eds), Servingsmall-scale farmers: an evaluation of DBSA 's farmer support programmes.Midrand: Development Bank of Southern Africa.DE GORTER, H & SWINNEN, J, 1994. The political economy and institutionaldeterminants of public policy in agriculture. In Peters, GH & Hedley, DD (Eds),Agricultural competitiveness: market forces and policy choice. Proceedings ofthe 22nd International Conference of Agricultural Economists, Dartmouth. Ha-rare, Zimbabwe, 22-9 August.DELGADO, CL, 1996. Bringing previously disadvantaged rural people into theeconomic mainstream: the role of smallholder agricultural production in sub-Saharan Africa. Paper prepared for the Symposium on Meeting the Challenge ofOvercoming Food Insecurity in Southern Africa. International Food Policy Re-search Institute. Washington DC, 24-5 June.DESPINS, P, 1998. Personal communication. Visiting scholar, Department ofAgricultural Economics, University of Natal, Pietermaritzburg.DEVEREUX, S, PARES, H & BEST, J, 1989. Credit and savings for develop-ment. Oxford: Oxfam.EISENBEIS, AR, 1977. Pitfalls in the application of discriminant analysis inbusiness, finance and economics. Journal of Finance, 32(3): 875-900.
153
Dow
nloa
ded
by [
FU B
erlin
] at
06:
26 1
9 N
ovem
ber
2014
LJ Fenwick & MC Lyne
FITSCHEN, A & KLITGAARDT, R, 1996. Exploring income variations acrosstribal wards in KwaZulu-Natal. Unpublished paper. Department of Economics.Durban: University of Natal.HOFF, K & STIGLITZ, JE, 1990. Introduction: imperfect information and ruralcredit markets - puzzles and policy perspectives. World Bank Economic Review,4(3): 235-50.LOW, A, 1986. Agricultural development in Southern Africa: farm-householdeconomics and the food crisis. Cape Town: David Philip.LUGEMWA, WH & DARROCH, MAG, 1995. Discriminant analysis of sea-sonal agricultural loan repayment by small-scale farmers in Transkei. Agrekon,34(4): 231-7.LYNE, MC, 1992. Constraints to emerging farmers in the homeland areas: re-sults of a sample survey. Unpublished report. Department of Agricultural Eco-nomics. Pietermaritzburg: University of Natal.LYNE, MC & ORTMANN, GF, 1992. Evaluation of the KwaZulu Farmer Sup-port Programme: second interim report. Unpublished report. Department of Ag-ricultural Economics. Pietermaritzburg: University of Natal.LYNE, MC & THOMSON, DN, 1997. Creating opportunities for farmers incommunal areas: adapting institutions to promote an efficient rental market inarable land. Unpublished report. Department of Agricultural Economics. Pieter-maritzburg: University of Natal.MENARD, S, 1995. Applied logistical regression analysis. London: Sage Uni-versity.MINCER, J, 1963. Market prices, opportunity costs and income effects. InChrist, CF (Ed.), Measurement in economic studies in mathematical economicsin memory of Yehuda Grunfeld. Stanford, USA: Stanford University Press.NATTRASS, J & MAY, J, 1986. Migration and dependency: sources and levelsof income in KwaZulu. Development Southern Africa, 3(4): 583-99.NIEUWOUDT, WL, 1990. Efficiency of land use. Agrekon, 29(4): 210-5.OLSON, M, 1996. Big bills left on the sidewalk: why some nations are rich andothers poor. Journal of Economic Perspectives, 10(2): 3-24.PITT, MM & KHANDKER, SR, 1995. Household and intrahousehold impactsof the Grameen Bank and similar targeted credit programs in Bangladesh. Paperprepared for presentation at the World Bank/Bangladesh Institute of Develop-ment Studies (BIDS) sponsored 'Workshop on Credit Programs for the Poor,'Dhaka, Bangladesh, 19-22 March.
SACHS, J, 1996. Growth in Africa: it can be done. The Economist, 29 June, 23-4.SAHN, DE & SARRIS, A, 1994. The evolution of states, markets and civil in-stitutions in rural Africa. Journal of Modem African Studies, 32(2): 279-303.STRAUSS COMMISSION, 1996. Interim report of the commission of inquiryinto the provision of rural financial services. South Africa, March.THOMSON, DN, 1996. A study of land rental markets and institutions in com-munal areas of rural KwaZulu-Natal. Unpublished PhD thesis, Department ofAgricultural Economics. Pietermaritzburg: University of Natal.
154
Dow
nloa
ded
by [
FU B
erlin
] at
06:
26 1
9 N
ovem
ber
2014
The relative importance of liquidity and other constraints inhibiting growth
THOMSON, DN, & LYNE, MC, 1993. Constraints to land rental in KwaZulu:analysing the transaction costs. Oxford Agrarian Studies, 21(2): 143-50.TIMMER, CP, 1992. Agriculture and economic development revisited. Agricul-tural Systems, 40: 21-58.UDRY, C, 1995. Risk and saving in Northern Nigeria. American Economic Re-view, 85(5): 1287-300.VON PISCHKE, JD & ADAMS, DW, 1980. Fungibility and the design andevaluation of agricultural credit projects. American Journal of Agricultural Eco-nomics, 62(4): 719-26.
Submitted March 1998; accepted for publication June 1998.
155
Dow
nloa
ded
by [
FU B
erlin
] at
06:
26 1
9 N
ovem
ber
2014
