Contents lists available at SciVerse ScienceDirect

Livestock Science

Livestock Science 155 (2013) 8–16

1871-14http://d

n CorrE-m

vmuche

journal homepage: www.elsevier.com/locate/livsci

Determination of economic weights for breeding traitsin indigenous Nguni cattle under in-situ conservation

O. Tada a, V. Muchenje a,n, J. Madzimure a, K. Dzama b

a Department of Livestock and Pasture Science, University of Fort Hare, P. Bag X1314, Alice 5700, South Africab Department of Animal Sciences, Stellenbosch University, P. Bag X1, Matieland 7602, South Africa

a r t i c l e i n f o

Article history:Received 4 January 2013Received in revised form8 April 2013Accepted 9 April 2013

Keywords:Choice experimentAge at first calvingAggression and mating behaviorBody condition scoreTick infestation

13/$ - see front matter & 2013 Elsevier B.V. Ax.doi.org/10.1016/j.livsci.2013.04.011

esponding author. Tel.: +27 40602 2059; faxail addresses: vmuchenje@ufh.ac.za,nje@hotmail.com (V. Muchenje).

a b s t r a c t

This study was conducted to determine the economic weights of most preferred traits inyoung breeding Nguni bulls and first parity cows. Fifty-four farmers from low-input cattleproduction enterprises participated in the choice experiment. Sixteen individual animalprofiles were formulated from four traits of three levels each using a fractional orthogonaldesign of SPSS 14.0 (2005). 120 pair-wise choices were deduced for each breeding animalclass. A total of 6480 (54�120) observations were obtained for each class of the animals.Data was subjected to multinomial logit (MNL) models using econometric softwareNLOGIT 4.0.1 Version (2007). All computed economic values for bull traits were significant(po0.05). The economic weights of bull traits were poor body condition score(−0.9970.095), good body condition score (0.4570.073), over-conditioned (base level),low tick infestation (0.5770.103), medium tick infestation (0.5870.084), high tickinfestation (base level), high aggression and mating behavior (4.4170.095), averageaggression and mating behavior (2.5370.094), and low aggression and mating behavior(base level). The economic weights of first parity cow traits were poor body conditionscore (−0.0670.055), good body condition score (1.0870.061), over-conditioned (baselevel), low tick infestation (1.5070.059), medium tick infestation (0.8370.067), high tickinfestation (base level), age at first calving of ≤27 months (2.3770.068), age at firstcalving of 27–36 months (1.3070.076), and age at first calving of 436 months (baselevel). Farmers were willing to pay R37,939 (US$4864) for a bull with high aggression andmating behavior score and R17,185 (US$2203) for a first parity cow of less than 27 monthsold. Enterprise ownership and demographics factors of the farmers were significant indetermining economic weights within trait levels. Economic weights were high forreproductive efficiency of the breeding animals followed by the high adaptive character-istics. The choice experiment procedure can be the tool for determining importance ofanimal characteristics under low-input production systems. It is recommended to makeuse of the economic weights of preferential traits in designing selection models.

& 2013 Elsevier B.V. All rights reserved.

1. Introduction

Diversity of an indigenous genetic resource is a keycomponent for a low-input production system to over-come destabilizing factors of uncertainty over future

ll rights reserved.

: +27 86 628 2967.

production environments such as climate change, diseasesand changing market demands (Ruto et al., 2008; Kassieet al., 2010; Zander, 2011). The indigenous Nguni cattlebreed in South Africa is an example of Animal GeneticResources (AnGR) currently under in-situ conservation inthe communal and small-scale farming enterprises of theEastern Cape Province (Muchenje et al., 2008; Tada et al.,2012). The majority of the farmers in these sectors (67%)perceived the low-input in-situ conservation as profitablebecause the indigenous breed possesses traits of economic

O. Tada et al. / Livestock Science 155 (2013) 8–16 9

and socio-cultural importance (Tada et al., 2012). Informeddecisions on sustainable conservation of this geneticresource could be made easier if information on theeconomic value of traits and breeding objectives wereavailable.

The development of breeding objectives has longinvolved quantification of the levels of economical benefitassociated with genetic traits expressed by farmed live-stock (Amer, 2007). The breeding objectives determineoptimal herd size and direction of genetic changes inproduction traits. Thus, they considerably influence theneed of economic weights of production traits in selection.The economic weights of traits for beef cattle are oftendisregarded under low-input production systems probablydue to the difficulty of measuring and valuing them asreported by Roeleveld (1996). The development and appli-cation of adequate tools to economically characterize thetraits was therefore important. A review of potential AnGRvaluation methods by Roosen et al. (2005) highlighted thepotential role of non-market valuation methodologies indeveloping countries. This follows the premise that manyof the benefits derived from the existence of well adaptedindigenous breeds are not transacted in any market (Rutoet al., 2008). An indirect stated preference approach, thechoice experiment (CE) (Louviere et al., 2000) ,can there-fore be used to investigate farmers' preferences over cattletraits in livestock selection markets.

Some applications of CEs show that such methodolo-gies reveal useful estimates of the values that are placedon the market, non-market, and potential breed attributes(Scarpa et al., 2003; Tano et al., 2003; Ruto et al., 2008).The contribution of preferred traits in the breeding objec-tive is the basis for determination of their economicweights. The need to include economic weights of traitsin a selection model for low-input animal productionenvironments is recognized (Hazell et al., 2007; Zander,2011). This has not yet been implemented for Nguni cattleand many other indigenous breeds in the developingcountries where performance recording systems are mini-mal. The objective of the study was, therefore, to deter-mine the economic weights of most preferred traits ofyoung breeding Nguni animals by farmers in the low-inputproduction enterprises using a CE approach. It washypothesized that the contribution of animal traits to themarket value of the breeding animal from different enter-prise types and farmers of different demographic factorswere the same.

2. Materials and methods

2.1. Study area

The study was conducted in the Eastern Cape Province ofSouth Africa with representative farmers (75%) from com-munal and small-scale Nguni cattle enterprises. The enter-prises were the beneficiaries of Nguni Cattle RestorationProgram that was enacted in 2004 (Raats et al., 2004). TheEastern Cape Province is the second largest Province with anarea of 169,580 km², representing 13.9% of South Africa's totalland mass (Acocks, 1988). The climate varies according to thedistance from the Indian Ocean. The coastal areas enjoy mild

temperate conditions ranging between 14 and 23 1C, whilethe inland areas experience slightly more extreme conditionswith temperatures of 5–35 1C. Inland mountainous areasexperience winter snows and summer rainfalls.

2.2. The rationale of choice experiment and its design

Wurzinger et al. (2006) reported that choice experi-ments are important for identifying selection criteria intraditional production systems where literacy level is lowand recording practices are not in place. The breeding goalis generally described as a linear function of traits to beimproved as described by Hazel (1943); each of these traitsis multiplied by its economic weight (EW) expressing thevalue of a unit change in the trait while keeping the othertraits in the breeding goal constant. Due to the complexityand diversity of the low-input production systems, the lackof good records and good estimates of inputs and outputs,a simplified CE procedure was deemed appropriate toderive EWs of most preferred traits of young breedingNguni cattle. Choice experiment permits the analysis offarmer's preferences in terms of the benefits that theyexpect to attain from different genetically determinedtraits. Hypothetical profiles were described in terms oftrait levels. Traits were identified by farmers during apreliminary survey and these are easily recorded at farmlevel with minimum literacy. The three most importanttraits identified for, and price ranges of the young breedingNguni, i.e. first parity cow and 2–3 year old bull, arepresented in Table 1. When policy makers promote cattlewith desired traits, farmers are likely to conserve the breedand at the same time generate income (Zander, 2011).With four traits of three levels each in both classes ofanimals, there were 64 (43) possible Nguni cattle profilesin a full factorial design. These were reduced to a manage-able size of 16 profiles using a balanced orthogonal i.e.fractional factorial design (SPSS 14.0, 2005). The designensures the identification of the main effects with aminimum number of profile combinations. A choice setwith uncorrelated attributes was then generated. Descrip-tive cards in Xhosa (vernacular) with pictorial illustrationswere used.

2.3. Data collection

Data were collected in the form of an in-person surveyinstrument. Fifty-four respondents representing low-inputNguni cattle conservation enterprises were convenientlysampled. The criteria involved selecting a representativewho is literate and willing to implement cattle recordingsystem. Respondents were first exposed to interactivediscussions on the value of animal records, traits of eco-nomic importance and recording. The demographic data ofthe respondents were also gathered. Age, education level,gender and ownership pattern have been identified as keydemographic parameters affecting selection of animal traitsand their pricing under the South African environment(Madzimure, 2011). After a cheap talk script, respondentswere introduced to the type of choice task required i.e. a fullset of 120 pair-wise choices from 16 individual animalprofiles. The respondents were tasked to hypothetically

Table 1Description of variables used in the choice experiment.

Attribute/Trait Level description

Body condition score (BCS) Poor (BCS 1–3); Good (BCS 4–6) and Over-conditioned (7–9)Tick infestation (TI) Low (visible ticks less than 10); Medium (tick count of 10 to 30) and High (tick count of more than 30)

Attributes specific to the first parity cowAge at first calving (AFC) o27 months; 27–36 months; and 436 monthsAnimal purchase price R4500; R5500 and R7750

(US$577; US$705 and US$994)

Attributes specific to the young breeding bullAggression and mating behavior (AMB) Below average; Average and Above averageAnimal purchase price R5250; R7500 and R9 50

(US$673; US$962 and US$1250)

NB: The trait levels used as reference bases are shown in italics. Animal purchase price indicate the market price of the breeding animal, the attribute levelswere based on the study by Tada et al. (2012) i.e. lower, medium and upper quartiles.

O. Tada et al. / Livestock Science 155 (2013) 8–1610

buy for breeding one of the two available animal profiles.If neither of the animal profiles was found satisfactory, therespondents could choose the ‘I prefer none’ option. Figs. 1and 2 show the examples of the choice task used. A total of6480 (54�120) observations were obtained from therespondents for each of the two animal breeding categoriesi.e. first parity Nguni cow and young breeding Nguni bull.

3. Statistical analyses

The choice data were analyzed using econometric soft-ware NLOGIT 4.0.1 Version (2007). The multinomial logit(MNL) model is one of the most recognized discrete choicemodels (Train, 2003; Roessler et al., 2008). It assumes thateach individual chooses the alternative that has the high-est perceived utility. Individuals are assumed to evaluatechoice alternatives on the basis of their attribute levels,finally selecting the alternative they subjectively assess toprovide them with highest utility (Roessler et al., 2008).The economic weight of a trait level is represented bylinear utility function. Utility is assumed to either increaseor decrease according to price and animal traits, depend-ing on how the respondent regards animal characteristics.For an individual n choosing alternative j, the indirectutility is assumed to take the following form

Unj ¼ αnj þ γjSn þ β′nxnj þ εnj

The obtained indirect utility may vary between choicej and individuals n (the total number of individuals is n¼1,…, N). Indirect utility is assumed to consist of a determi-nistic part Unj¼αnj+γjSn+β

'nxnj and a stochastic part εnj. The

deterministic component of the utility function consists ofαnj which is the option specific intercept that correspondsto individual n's intrinsic preference for alternative j. Thesocio-economic and demographic characteristics of theindividual, Sn, and the coefficient vector γj correspond tothe systematic preference heterogeneity among the indi-viduals in the sample. Altogether, three animal traitattribute coefficients are estimated and, with the pricecoefficient β′n1, there are four coefficients, (β′n¼[(β′n1,…, β′n).These coefficients are assumed to be generic (i.e. thecoefficients of the explanatory variables do not vary acrossthe options). Hence, an assumption of stable preferenceswas made.

3.1. Part worth values of attributes as estimates of economicvalues

Part worth values reflect the relative importancerespondents put on attributes, or trade-offs they arewilling to make among them. As the cost was includedas an attribute in the CE, it was possible to estimateindirectly the willingness to pay (WTP) or willingness toaccept compensation (WTAC) for all other attributesincluded in the study. The WTP for a certain attribute orattribute level indicates the price (“implicit price”) therespondent was willing to pay for a unit increase in thisattribute or the compensation he/she was willing to acceptfor a decrease in this attribute. The ‘implicit price (W)’ orpart worth is the economic value of a trait level and iscalculated as follows:

W ¼ −1ðβx=βpriceÞ, where βx is the estimate for theattribute x from the MNL model, and βprice is the estimatedprice coefficient.

To compare if parameter estimates of the pooled modelwere different across demographic factors and enterprisetypes, random parameter MNL models were run to obtainutility functions (economic weight estimates) for eachcategory.

4. Results

4.1. Economic values of trait levels in young breeding Ngunibulls

The discrete choice (MNL) model used was good with acoefficient of determination of 39%. All the computedeconomic value estimates were significant (po0.05) witha notable insignificant difference on economic weightsbetween low and medium tick infestation level. Theeconomic weight for poor BCS was negative and subse-quently the farmers were willing to accept compensation(WTAC) of R8494.00 (US$1088.97) for such characteristicanimal. The most weighted trait level was a high AMBwhereby the farmers were willing to pay (WTP) R37,939.00(US$4863.97) for such a bull. The economic weights relativeto the base levels and economic values of all bull trait levelsevaluated are shown in Tables 2 and 3. On accounting forheterogeneity in mean economic weight estimates, poor

Fig. 1. Example of the pair-wise comparison choice set used for first parity Nguni cow.

Fig. 2. Example of the pair-wise comparison choice set used for a 2–3 years old Nguni bull.

O. Tada et al. / Livestock Science 155 (2013) 8–16 11

body condition score (BCS) was not significantly affected bythe demographics factors of the farmers (p40.05) (Table 4).Type of enterprise had no significant influence on economicweights of BCSs and average AMB (p40.05). The low and

medium TI had significant negative economic weights incommunal enterprises. A high AMB was observed to besignificantly associated with positive economic weight incommunal enterprises (po0.05). Highest education level

Table 2Estimates of economic weights and values (in Rands) of trait levels for young breeding Nguni bulls.

Trait level Economic weight7s.e p-value Economic values (WTP/WTAC) CI lower limit CI upper limit

Poor body condition score (1–3) −0.9970.095 po0.05 −8494 −10104 −6885Good body condition score (4–6) 0.4570.073 po0.05 3849 2619 5079Over body condition score (7–9) Base level 4645 3225 6065Low tick infestation 0.5770.103 po0.05 4927 3190 6665Medium tick infestation 0.5870.084 po0.05 5001 3586 6415High tick infestation Base level −9928 −11,504 −8352High aggression and mating behavior 4.4170.095 po0.05 37,939 36,342 39,535Average aggression and mating behavior 2.5370.094 po0.05 21,807 20,218 23,397Low aggression and mating behavior Base level −59,746 −61,339 −58,153Price −0.0070.002 po0.05Constant 10.1170.375 po0.05

NB: economic value of trait level used as a base is zero (0). WTP—willingness to pay. WTAC—willingness to accept compensation. CI—confidence interval(95%). US$1.00¼R7.80 (South Africa Reserve Bank, 2011).

Table 3Heterogeneity in mean variables and derived standard deviation of distributions in young breeding Nguni bulls.

Trait level Parameter: variable (base level) Coefficient Standard error Significance

Poor BCS Gender: male (female) −0.07 0.127 NSAge:o50 years (≥50 years) 0.05 0.141 NSEducation: formal (informal) 0.07 0.166 NSOwnership: communal (small-scale) 0.11 0.100 NS

Good BCS Gender: male (female) 0.20 0.105 NSAge:o50 years (≥50 years) −0.50 0.123 nn

Education: formal (informal) −0.06 0.133 NSOwnership: communal (small-scale) −0.04 0.078 NS

Low TI Gender: male (female) 0.83 0.131 nn

Age:o50 years (≥50 years) −0.72 0.143 nn

Education: formal (informal) 0.40 0.170 n

Ownership: communal (small-scale) −0.38 0.102 nn

Medium TI Gender: male (female) 0.70 0.112 nn

Age:o50 years (≥50 years) −0.52 0.128 nn

Education: formal (informal) 0.22 0.141 NSOwnership: communal (small-scale) −0.36 0.083 nn

High AMB Gender: male (female) −1.05 0.138 nn

Age:o50 years (≥50 years) 0.86 0.146 nn

Education: formal (informal) −0.59 0.174 nn

Ownership: communal (small-scale) 0.58 0.106 nn

Average AMB Gender: male (female) −0.63 0.134 nn

Age:o50 years (≥50 years) 0.63 0.149 nn

Education: formal (informal) 0.01 0.176 NSOwnership: communal (small-scale) 0.09 0.106 NS

Derived standard deviations of parameter distributionsPoor BCS NS 0.81 0.115 nn

Good BCS NS 0.16 0.006 NSLow TI NS 0.03 0.166 NSMedium TI NS 0.03 0.200 NSHigh AMB NS 0.07 0.178 NSAverage AMB NS 0.41 0.202 n

NB: heterogeneity co-efficient of variable used as a base is zero (0). NS: not significant at α¼0.05.n Significant at po0.05.nn Significant at po0.01.

O. Tada et al. / Livestock Science 155 (2013) 8–1612

attained by the farmer significantly influenced preference ofTI levels and a high AMB.

4.2. Economic values of trait levels in first parity Nguni cows

The MNL model had a coefficient of determination valueof 24% and produced significant estimates of economic

weights (po0.05) except on poor BCS (Table 4). The baselevels of the three traits used in the model were over-conditioned (BCS 7–9), high tick infestation (TI), and age atfirst calving (AFC) of greater than 36 months. The base levelshad negative economic values. The highest economic weightas indicated by a high utility coefficient was observed onbreeding cows with less than 27 months AFC. On average the

Table 4Estimates of economic weights and values (in Rands) of trait evel sfor first parity breeding Nguni cows.

Trait level Economic weight7s.e p-value Economic value (WTP/WTAC) CI lower limit CI upper limit

Poor body condition score (1–3) −0.0670.055 p40.05 −413 −1194 368Good body condition score (4–6) 1.0870.061 po0.05 7834 6962 8705Over Body Condition score (7–9) Base level −7421 −8247 −6595Low tick infestation 1.5070.059 po0.05 10,859 10,021 11,697Medium tick infestation 0.8370.067 po0.05 6015 5059 6971High tick infestation Base level −16,874 −17,771 −15,977Age at first calving ≤27 months 2.3770.068 po0.05 17,185 16,213 18,156Age at first calving 27–36 months 1.3070.076 po0.05 9454 8372 10,535Age at First calving 436 months Base level −26,638 −27,665 −25,612Price −0.0070.002 po0.05Constant 8.9770.310 po0.05

NB: economic value of trait level used as a base is zero (0). WTP—willingness To Pay. WTAC—willingness to accept compensation. CI—confidence interval(95%). US$1.00¼R7.80 (South Africa Reserve Bank, 2011).

Table 5Heterogeneity in mean variables and derived standard deviation of distributions in first parity Nguni cows.

Trait level Parameter: variable (base level) Coefficient Standard error Significance

Poor BCS Gender: male (female) 0.24 0.072 nn

Age:o50 years (≥50 years) 0.16 0.099 NSEducation: formal (informal) 0.26 0.099 nn

Ownership: communal (small-scale) −0.15 0.060 n

Good BCS Gender: male (female) −0.02 0.081 NSAge:o50 years (≥50 years) −0.13 0.114 NSEducation: Formal (informal) 0.11 0.114 NSOwnership: communal (small-scale) 0.05 0.069 NS

Low TI Gender: male (female) 0.05 0.075 NSAge:o50 years (≥50 years) 0.01 0.104 NSEducation: formal (informal) −0.01 0.104 NSOwnership: communal (small-scale) −0.11 0.063 NS

Medium TI Gender: male (female) 0.01 0.089 NSAge:o50 years (≥50 years) −0.34 0.128 nn

Education: formal (informal) −0.54 0.131 nn

Ownership: communal (small-scale) −0.06 0.077 NS

o27 months AFC Gender: male (female) −0.35 0.076 nn

Age:o50 years (≥50 years) −0.11 0.107 NSEducation: formal (informal) −0.29 0.108 nn

Ownership: communal (small-scale) 0.37 0.065 nn

27–36 months AFC Gender: male (female) −0.10 0.088 NSAge:o50 years (≥50 years) 0.30 0.121 n

Education: formal (informal) 0.08 0.121 NSOwnership: communal (small-scale) −0.22 0.075 nn

Derived standard deviations of parameter distributionsPoor BCS NS 0.56 0.106 nn

Good BCS NS 0.01 0.122 NSLow TI NS 0.03 0.157 NSMedium TI NS 0.00 0.161 NSo27 months AFC NS 0.13 0.212 NS27–36 months AFC NS 0.48 0.131 nn

NB: heterogeneity co-efficient of variable used as a base is zero (0). NS: not significant at α¼0.05.n Significant at po0.05.nn Significant at po0.01.

O. Tada et al. / Livestock Science 155 (2013) 8–16 13

farmers were willing to pay (WTP) up to R17,185.00 (US$2203.21) for such a breeding cow. On average farmers werewilling to accept compensation (WTAC) of R26,638.00 (US$3415.13) on breeding cows with an AFC of greater than 36months. Enterprise type had significant effect on the hetero-geneity of economic weights within trait levels as observed

on communal enterprises negatively affecting the poor BCSand AFC of 27–36 months while positively affecting AFC ofless than 27 months (po0.05) (Table 5). The gender of thefarmer significantly influenced the poor BCS and AFC ofless than 27 months (po0.05) where the male farmershad positive and negative economic weights, respectively.

O. Tada et al. / Livestock Science 155 (2013) 8–1614

Age of the farmer had no significant effect on economicweights except for the positive effect on AFC of 27–36months. The highest level of education attained by thefarmer had significant influence on economic weights ofcow traits. Farmers with primary education had negativescores on medium TI and AFC of less than 27 months whilehaving a positive scores on poor BCS (po0.05).

5. Discussion

The pooled models of both breeding bulls and cows hada negative price utility co-efficient, which is good evidencethat farmers did not want expensive breeding animals. Thefarmers were observed to have a limited source of incomeand therefore practice low-input agriculture productionsystem (Tada et al., 2012) that would mean a constrainedbudget on expenses. Furthermore, farmers in the studycame from rural communities whereby expenditure ofbreeding stock is less likely especially for indigenous andnon-descript genotypes which are thought to be climati-cally adapted where natural selection is at play. The fittedmodels were considered good with a co-efficient of deter-mination (R2) of 24% and 39% on cows and bulls, respec-tively. This suggests fewer inconsistencies in the responses(List et al., 2006). The likelihood models are known to havea high co-efficient of variation due to the nature ofrespondents of which differences in the demographicand psychological factors lead to an array of variability inresponses as reported in breed and trait preference studieswith indigenous cattle kept under traditional productionsystems (Kassie et al., 2010; Desta et al., 2011). The valuesobtained may have been influenced by a higher number ofresponses from individual farmers. A total of 120 choicesets from 16 animal profiles per breeding animal class withno missing responses indicated a full fractional orthogonaldesign that can maximize R2 values.

The utility co-efficient, which translate to economicweights, of poor BCS, high TI and low AMB in breedingbulls, were negative. This gives an indication that the traitlevels were not desirable to the farmers and can beselected against (Roessler et al., 2008; Moyo and Masika,2009). The highest economic weight observed was on highAMB level (4.4170.095), which is a signal that farmer'sdecision on buying a breeding bull is strongly influencedby the AMB. This is consistent with the preliminary surveyresults obtained when farmers were ranking traits ofeconomic importance in these rural enterprises as wellas studies by Desta et al. (2011) on indigenous Sheko cattlein Ethiopia. Breeding bulls are known to have a majorinfluence on the enterprise as they leave more progenythan breeding cows. The farmers realised the importanceof this trait and were willing to pay up to R37,939 (US$4864) for a high performing bull despite the fact that theyare resource-limited. The economic value of this trait levelwas consistent with auction prices of proven pedigreebulls of the same Nguni breed across the country. Theprices of the breeding bulls ranged from R30,000 toR55,000 (US$3846–US$7051) during the auctions heldunder the auspices of the Nguni Cattle Breed Society inthe year 2011, the period this research data was gathered(Nguni Cattle Breeders Society, 2011).

The second highest economic weight was observed onbulls of average AMB (2.5370.094), this indicated thatfarmers compromise for an average performing bull beforeconsidering the low TI, medium TI and good BCS. The lowAMB level attracted the worst negative economic value of–R59,746 (–US$7660). This signifies a complete displea-sure by the farmers for a breeding bull with such unpro-ductive characteristics. The low and medium TI had similareconomic values (Table 2) as these can be thought to becontrollable at minimum cost unlike the high TI whichattracted a negative economic value of up to −–R9928 (US$1273). A positive economic value for high BCS and anegative economic value on poor BCS is evidence thatfarmers prefer to buy an over-conditioned than an under-conditioned bull. An over-conditioned bull can maintainacceptable BCS with minimum costs whilst conditioning apoor bull is associated with costs from farmers whoalready have limited sources of income (Tada et al., 2012).

Enterprise ownership pattern and demographic factorsof the farmers did not significantly influence the prefer-ence of an over-conditioned bull. This suggests that farm-ers would not buy poor-conditioned breeding bulls.Younger (o50 years) and old farmers (≥50 years) hadcontrasting decisions on buying a good-conditioned, lowand medium TI, and high and average AMB levels. Theyoung farmers can be the class of farmers that prefer over-conditioned breeding bulls as management practicestargeting maintenance of body condition may fail andsettle for the ideal condition. These results on low andmedium TI can suggest some levels of inconsistencies fromthe responses of this category. Inconsistencies have beenreported in choice experiments (CE) and drastically reduceR2 values under field conditions mainly because of less-effective cheap talk scripts and less time spent in evaluat-ing the choice sets (List et al., 2006).

The young farmers significantly preferred buying highand average AMB bulls more than old farmers which cansuggest the correlation of sexual behavior of humans asrelated to animals. Contrasting views across genderpatterns were observed as male farmers indicated lesseremphasis of high and average AMB as well as the low andmedium TI in breeding bulls. Primary and secondaryeducated farmers were consistent on the responses forlow TI. This can be attributed to the experience andknowledge of animal husbandry acquired formally and/orinformally through the activities of the Eastern CapeDepartment of Rural Development and Agrarian Reform(ECDRDAR). Kassie et al. (2010) and Wollny (2003) postu-lated a likelihood of different needs, perceptions andpreferences by which village communities in Ethiopiaand Africa in general make decisions for mating or saleof animals. Primary educated farmers did not prefer highAMB bulls which can be a revelation that aggressionnature of the bulls can be misunderstood yet it offers thenaturally fit and naturally selected bull a chance to leavemore progeny.

Communal and small-scale enterprises had contrastingpreferences on low to medium TI levels. The small-scaleenterprises were located in Land Redistribution and Agricul-ture Development (LRAD) farms which may not be theprimary target by the ECDRDAR due to logistical issues as

O. Tada et al. / Livestock Science 155 (2013) 8–16 15

the LRAD farms were relatively a newmethod of empoweringemerging farmers. The work by the ECDRDAR on helping therural communities with dipping infrastructure and chemicals/acaricides (Tada et al., 2012; Moyo and Masika, 2009) couldhave made the communal enterprises think that low andmedium TI are not that crucial compared to responses fromsmall-scale enterprises. The enterprise ownership patternswere not consistent on the need of a high AMB bulls. Thecommunal enterprises were most vocal because the servicesof the Nguni bulls were dearly needed as noted in theprevious studies on breeding concerns bulling rates and herdstructure (Tada et al., 2012).

The respondents in the study had cattle enterprisesthat were yet to payback the nucleus herd animals to theprogram. Ideally, animals that quickly reproduce are mostfavored. This could be one of the main reasons whyhighest economic weights of trait levels in buying a firstparity Nguni cow was observed in AFC of less than 27months. A balanced birth sex ratio postulated throughMendelian principles for the enterprise animals acrossownership patterns indicates the need to have highlyreproductive animals to meet program objectives as wellas generating potential breeding bulls for performancetesting and or slaughter steer production system. Thefarmers realised that Nguni cattle within the enterpriseshad never been selected and therefore were not uniform inperformance. By considering top-notch breeding cowsusing AFC with a reported medium to high heritabilitytrait value of 27–37% in beef cattle (Bourmann and Wilson,2010; Gergovska and Yordanova, 2011) genetic progresswas fore-seen. This translated to a high economic value ofup to R17,185 (US$2203). This price was consistent withauction prices of the breed society (Nguni Cattle BreedersSociety, 2011).

The realization of the importance of tick control undercommunal low-input production systems was shown bythe second ranking of this low TI level (Table 4). Tickinfestation is regarded as a measure of disease resistanceand therefore adaptability of the animals, therefore theeconomic value of the trait level observed was positive atR10,859 (US$1398). Cows with an undesirable high TI hadhigh negative economic values which the resource-limitedfarmers were not likely to accept as the animal would be aliability to the enterprise. Farmers had the notion ofchoosing animals which would offer benefits at the leastpossible cost. This notion was supported by a negativeeconomic value observed on over-conditioned and poor-conditioned cows that are known to have problems withregard to cow productivity (Veerkamp et al., 2001; Destaet al., 2011).

Male and female farmers concurred on the negative effectsto the enterprise of cows with a poor BCS with males having abigger voice. With regard to the highly weighted trait level,o27 months AFC, male farmers contributed less on valuingthis characteristic cow as well as farmers who attendedprimary education. A stronger response was observed onbreeding cows of poor BCS with farmers who attended onlyprimary education than secondary education. On the mediumTI and o27 months AFC, farmers who attended secondaryeducation level were more influential in valuing the breedingcow. Communal enterprises were negatively inconsistent with

small-scale enterprises on poor BCS and 27–36 month AFC,which could be attributed to the notion that communalenterprises thought it was a normal case for animals to havea poor condition and calve after 3 years (Nguni Cattle BreedersSociety, 2011). The small-scale enterprises had less positiveinfluence on choosing o27 months AFC, which can be due toa higher influence on 27–36 month AFC.

6. Conclusion and recommendation

The utility co-efficient of the trait levels translated tothe economic weight while the implicit price or part worthvalue translated to the economic value of the trait levels.Farmers indicated high economic weight on reproductiveefficiency of the breeding animals followed by the adap-tive characteristics. A high and medium AMB resulted inthe highest economic value for Nguni breeding bulls whilea low AMB and a high TI were discouraged by the farmers.Enterprise ownership patterns and demographic factorshad significant influence in making choice of Ngunibreeding bulls. The most treasured trait levels in Ngunibreeding cows were the reproductively efficient cows witha good condition score and highly adaptive to the localticks. To realize high incomes, farmers are urged to keepanimal performance records and sell their breeding stockat competitive prices especially at formal auctions therebysustaining an enabling policy for sustainable managementof indigenous cattle genetic resources. It is recommendedto make use of the economic weights of preferential traitsin designing selection models and mating strategies.

Conflict of interest statement

All authors declare that there are no actual or potentialconflicts of interest between the authors and other peopleor organizations that could inappropriately bias their work.

Acknowledgments

The authors are grateful to the farmers under the UFHNguni Cattle Program in the Eastern Cape Province forcooperation during the study period. The project wasfunded by Adam Fleming through the Nguni ProjectOperations (P329) of the University of Fort Hare.

References

Acocks, J.P.H., 1988. Veld types of South Africa, 3rd Edition. Memoirs of theBotanical Survey of South Africa, 57. Government Printer, Pretoria1–146.

Amer, P.R., 2007. Approaches to formulating breeding objectives. In:Proceedings of the 8th World Congress on Genetics Applied toLivestock Production. August 13–18, 2006. Belo Horizonte, MG, Brasil.

Bourmann, M.J., Wilson, D.E., 2010. Calving day and age at first calving inAngus heifers. J. Anim. Sci. 88, 1947–1956.

Desta, T.T., Ayalew, W., Hegde, B.P., 2011. Breed and trait preferences ofSheko cattle keepers in south-western Ethiopia. Trop. Anim. HealthProd. 43 (4), 851–856.

Gergovska, Z.H., Yordanova, L., 2011. Effect of age at first calving on theevaluation of breeding potential of dairy cattle and its correlation totest day productivity. Agric. Sci. Tech. 3 (1), 3–7.

Hazel, L.N., 1943. The genetic basis for constructing selection indexes.Genet 28, 476–490.

O. Tada et al. / Livestock Science 155 (2013) 8–1616

Hazell, P., Poulton, C., Wiggins, S., Dorward, A., 2007. The future of smallfarms for poverty reduction and growth. IFPRI Discussion Paper 42.Available from: ⟨http://www.ifpri.org/2020/dp/vp42.asp⟩.

Kassie, G.T., Abdulai, A., Wollny, C., 2010. Implicit prices of indigenous bulltraits in crop-livestock mixed production systems of Ethiopia. Afric.Dev. Rev. 22 (4), 482–494.

List, J.A., Sinha, P., Taylor, M.H., 2006. Using choice experiments to value non-market goods and services: evidence from field experiments. Adv. Econ.Anal. Policy 6 (2). (Article 2)⟨http://www.fieldexperiments.com/uploads/UsingChoiceExperimentstoValueNon-MarketGoodsandServices.pdf⟩.

Louviere, J.J., Hensher, D.A., Swait, J.D., 2000. Stated Choice Methods:Analysis and Applications. Cambridge University Press, Cambridge.

Madzimure, J., 2011. Climate Change Adaptation and Economic Valuationof Local Pig Genetic Resources in Communal Production Systems ofSouth Africa (PhD thesis). University of Fort Hare, Alice, South Africa.

Moyo, B., Masika, P.J., 2009. Tick control methods used by resource-limitedfarmers and the effect of ticks on cattle in rural areas of the EasternCape Province, South Africa. Trop. Anim. Health Prod. 41 (4), 517–523.

Muchenje, V., Dzama, K., Chimonyo, M., Raats, J.G., Strydom, P.E., 2008. Meatquality of Nguni, Bonsmara and Aberdeen Angus steers raised on naturalpasture in the Eastern Cape, South Africa. Meat Sci. 79, 20–28.

Nguni Cattle Breeders Society, 2011. In: Nguni 2011 25 Years – Breed fromthe past for the future. Nguni Cattle Breeders Society, Bloemfontein,South Africa 35.

NLOGIT, 2007. NLOGIT 4.0.1 Student Version, User Guide. EconometricSoftware, Catle Hill, Australia.

Raats, J.G., Magadlela, A.M., Fraser G.C.G., Hugo, A., 2004. Re-introducingNguni Nucleus Herds in 100 Communal Villages of the Eastern CapeProvince. A proposed co-operative project between the University ofFort Hare, Agricultural and Development Research Institute (ARDRI)and the Eastern Cape Department of Agriculture and the KelloggFoundation.

Roeleveld, A.C.W., 1996. The diagnostic phase in research on livestocksystems. In: Roeleveld, A.C.W., van den Broek, A. (Eds.), FocusingLivestock Systems Research, Royal Tropical Institute, Amsterdam, TheNetherlands, pp. 14–28.

Roessler, R., Drucker, A., Scarpa, R., Markemann, A., Lemke, U., Thuy, L., 2008.Using choice experiments to assess smallholder farmers' preferences forpig breeding traits in different production systems in North-WestVietnam. Ecol. Econ. 66 (1), 184–192.

Roosen, J., Fadlaoui, A., Bertaglia, M., 2005. Economic evaluation for conserva-tion of farm animal genetic resources. J. Anim. Breed Genet. 122, 217–228.

Ruto, E., Garrod, G., Scarpa, R., 2008. Valuing animal genetic resources: achoice modelling application to indigenous cattle in Kenya. Agric. Econ.38, 89–98.

Scarpa, R., Drucker, A.G., Anderson, S., Ferraes-Ehuan, N., Gómez, V.,Risopatrón, C.R., Rubio-Leonel, O., 2003. Valuing genetic resources inpeasant economies: the case of ‘hairless’ creole pigs in Yucatan. Ecol.Econ. 45 (3), 427–443.

Statistical Package for Social Scientists, 2005. In: SPSS 14.0 for WindowsEvaluation VersionCopyright© SPSS Inc., USA.

Tada, O., Muchenje, V., Dzama, K., 2012. Monetary Value of Nguni cattleand socio-economic profiles of farmers in the low-input communalproduction system of Eastern Cape Province, South Africa. Afric. J.Bus. Manage. 6 (45), 11304–11311.

Tano, K., Kamuanga, M., Faminow, M.D., Swallow, B., 2003. Using conjointanalysis to estimate farmer's preferences for cattle traits in WestAfrica. Ecol. Econ. 45, 393–407.

Train, K.E., 2003. Discrete Choice Methods with Simulation. CambridgeUniversity Press, Cambridge, UK.

Veerkamp, R.F., Koenen, E.P.C., De Jong, G., 2001. Genetic correlationsamong body condition score, yield, and fertility in first-parity cowsestimated by random regression models. J. Dairy Sci. 84, 2327–2335.

Wollny, C.B.A., 2003. The need to conserve farm animal genetic resources inAfrica: should policy makers be concerned? Ecol. Econ. 45 (3), 341–351.

Wurzinger, M., Ndumu, D., Baumung, R., Drucker, A., Okeyo, A.M.,Semambo, D.K., Byamumgu, N., Sölkner, J., 2006. Comparison ofproduction systems and selection criteria of Ankole cattle by breedersin Burundi, Rwanda, Tanzania and Uganda. Trop. Anim. Health Prod. 38,571–581.

Zander, K.K., 2011. Attitudes of livestock keepers to breeding strategies—threats and opportunities for on-farm conservation of the Boranacattle breed. J. Agric Sci. 3 (2), 3–12.