Is a 10-sow unit economically sustainable? A profitability assessment of productivity amongst small-holder pig farmers, Mpumalanga, South Africa

The majority of small-holder pig farmers in Mpumalanga had between 1- and 10-sow herds. The main aim of this study is to evaluate the current government agricultural intervention (supply of 10 sows and a boar) in terms of technical and economic feasibilities and ascertain whether the small-scale pig value chain system alleviates poverty. Data were obtained from 220 randomly selected small-holder pig farmers using a semi-structured questionnaire. The results showed that 58% farrowed ≤ 10 piglets/born/sow/litter, 44.2% practiced no weaning method and many fed swill and leftovers alone (41.6%). Pair-wise association revealed that the feeding of commercial feeds had a relationship with pigs in relatively good to very good body condition. Pigs in poor body condition were positively correlated with the feeding of swill alone. The economic models for the 10-sow unit proved that pig farming is unprofitable if the current management and feeding systems that operate in the commercial industry are utilised. However, only through a combination of cooperative systems, benefits of economies of scale, reduction of preweaning mortalities and structured government inputs can pig production be profitable at this scale of production.


Introduction
Animal agriculture in general and pig production in particular are important economic activities globally (Dietze 2011; Mokoele et al. 2015;Roelofse 2013). In South Africa, pig production is distributed in all nine provinces, with higher concentrations in Limpopo, North West, Gauteng and KwaZulu-Natal, partially because of cultural and religious preferences and availability of feedstuffs. Mpumalanga is placed sixth in terms of pig production, contributing approximately 8% of the national pig herd (DAFF 2012;MPG 2013). The province had a relatively high concentration of small-holder pig farmers otherwise known as emerging small-scale pig farmers. Pig farming requires little space, yields a large number of offspring after a shorter gestation period than other small stock and can be combined with other forms of subsistence agriculture where land resources are scarce (DAFF 2012;Makiwane et al. 2012). In addition, it plays a major role in poverty reduction and food security (FAO 2004) and provides a form of investment, emergency cash and meat for home consumption (Drucker & Anderson 2004;Mhlanga 2002). In Mpumalanga, the commonly found breeds of pigs are the Kolbroek, Large White, Landrace and their crosses.
Backyard pig farming and semi-intensive management systems in poorly designed pens are the most common small-holder pig farming practices in the rural and peri-urban areas of Mpumalanga. In general, the farm families rely on family labour and the majority of products are meant for household consumption or converted to cash for the purpose of family maintenance. These contributions are very important for family incomes in Mpumalanga, where poverty rates have ranged from 50. 4% (1996), 59.1% (2004), 50.1% (2008) to 39.4% (2011) and the unemployment rate remains at 29.4% (MPG 2009(MPG , 2013. The Mpumalanga Department of Agriculture, Rural Development, Land and Environmental Affairs (DARDLEA) established the programme called Masibuyele esibayeni, meaning back to the kraal (i.e. returning to the land), with similar programmes in Gauteng and Limpopo amongst others. The programme is aimed at helping the small-holder farmers to upgrade and boost productivity by improving the genetic pool of their livestock. For the pig component, DARDLEA provides farmers with 10 sows and 1 boar with improved genetics for breeding and provides supportive services to such farms. To evaluate the potential success of such a programme, we analysed profit determinants, the economic feasibility and viability of such small-holder projects and suggested options for improvement of the programme.
The majority of small-holder pig farmers in Mpumalanga had between 1-and 10-sow herds. The main aim of this study is to evaluate the current government agricultural intervention (supply of 10 sows and a boar) in terms of technical and economic feasibilities and ascertain whether the small-scale pig value chain system alleviates poverty. Data were obtained from 220 randomly selected small-holder pig farmers using a semi-structured questionnaire. The results showed that 58% farrowed ≤ 10 piglets/born/sow/litter, 44.2% practiced no weaning method and many fed swill and leftovers alone (41.6%). Pair-wise association revealed that the feeding of commercial feeds had a relationship with pigs in relatively good to very good body condition. Pigs in poor body condition were positively correlated with the feeding of swill alone. The economic models for the 10-sow unit proved that pig farming is unprofitable if the current management and feeding systems that operate in the commercial industry are utilised. However, only through a combination of cooperative systems, benefits of economies of scale, reduction of preweaning mortalities and structured government inputs can pig production be profitable at this scale of production.

Study area and data collection
The study was approved by the Ethical Committee of the College of Agriculture and Environmental Sciences, UNISA (CAES) with an Ethical approval number: 2013/CAES/140. A recent document targeting small-holder farmers indicated that at least 5889 small-holder farms exist in Mpumalanga (DAFF 2013; Figure 1). We used this number as the sample frame. The sample size was calculated for frequency using the formula: A total of 361 farms were needed. We continued to recruit small-holder pig farms randomly until we could identify no more farms of interest. A final list of 220 farmers was generated from the list provided by DARDLEA and the additions were made through consultations with farmers, extension officers, animal health technicians and community leaders. All identified farmers were visited and data were collected through the use of a semi-structured pretested questionnaire. Direct observations were evaluated through a checklist, and photographic documentation was obtained, where necessary.
The services of the extension officers and animal health technicians from DARDLEA, who were previously trained on questionnaire administration, were employed. The inclusion criteria were as follows: (1) ownership of ≥ 1 to ≤ 50 pigs and (2) resident within the province and active in the small-holder industry. The English questionnaire was translated and administered using local home languages (Zulu, IsiNdebele, Shangaan and IsiSwati) for the understanding of the study participants.

Statistical analyses and management
All responses were entered into Microsoft Excel 2007 ® spreadsheet and filtered. Data were analysed using Stata v9 (Statacorp., Texas, USA) and hypotheses were tested using appropriate analytical methods. To determine associations, all data were re-entered as 1 = yes and 0 = no and coded correctly for the Stata programme. Using Pearson's Chi-square test, outputs were generated to associate certain variables and preferred methods, including markets, market determinants, treatment methods for sick pigs, feed preference, body conditions of the sows and age at weaning.
To integrate economic analyses, a partial budgeting and return-on-investment (ROI) model was developed in a Microsoft Excel 2007 ® spreadsheet. Outcomes from the data obtained, including details from the field and published materials, were used to develop and validate the model. Economic feasibility and viability of a 10-sow unit was tested for a 3-year farm operation. Details of the inputs and outputs are available in the Appendix. The sensitivity analyses were tested by varying some parameters, including the reduction in feed price, removal of farmer's remuneration, transport cost and reduction or preweaning deaths. Outputs were generated in tables and graphs, and the model is freely available in excel format for the use of small-holder farmers and development partners (see Appendix).

Descriptive statistics
Approximately 41% of the farms surveyed confirmed that the sows farrowed ≥ 11 piglets per litter and more than 58% farrowed ≤ 10/sow/litter. Only about 19% weaned at 1 month (industry standard) and only 11% depended on commercial feed completely (Table 1). The majority of the farmers mixed commercial feed with swill (41.6%) or fed swill and leftovers alone (47%), and about 69% resorted to home medication, allowed the animal to die or sent any sick animal for slaughter (Table 1). Only 27% sold their porkers at less than 6 months, whereas the majority (73%) marketed their pigs above 7 months. Less than 10% of all sows were in adequate body condition (at least a score of 3). The prevailing local and market prices were the main determinant for marketing pigs and incomes arising from the sale were used mainly in the home or to maintain the remaining pigs (Table 1).
Significant association existed between production of a larger number of piglets and feeding of commercial ration (χ 2 = 11.57, p = 0.001). In contrast, the mixing of swill and commercial ration was associated with the production of low (χ 2 = 17.25, p < 0.0001) to medium (χ 2 = 23.11, p < 0.0001) numbers of piglets per litter ( Table 5). The feeding of swill only produced similar significant results (Table 5). Similarly, sows in poor body condition produced low (χ 2 = 6.37, p = 0.01) to medium numbers of piglets per litter (χ 2 = 5.44, p = 0.02) ( Table 5). Only sows in very good body condition were associated with a large number of piglets per litter (χ 2 = 7.77, p = 0.005).

Economic models
Using partial budgeting and ROI models, the 10-sow unit pig farm continues to utilise more cash (outflow) than the receipts   that came into the farm account. Feed (using commercial ration) accounted for at least 75% of the annual cash outflow for any 1 year ( Figure 2a; Table 1-A1, Table 2-A1, Table 3-A1).
With a 50% reduction in feed price through supplementation with swill and leftovers from the home, the model became economically viable towards the end of the third year of operation ( Figure 2b). However, a 100% reduction in feed price through complete replacement with swill would make the farm model break even at the beginning of the second year of operation with subsequent profits (Figure 2c). With complete removal of remuneration for the farmer over a 3-year project cycle, the farm was still economically unsustainable ( Figure 2d) and similar results were obtained with a 60% reduction in transport cost ( Figure 2e) and improving the farm productivity through a 25% reduction in preweaning mortalities (Figure 2f).

Discussion
In Mpumalanga, there are at least 15 auction facilities that are randomly dispersed. Many of the small-holder farmers marketed their pigs at auctions or within the communities. Because strong association existed between auctions and prevailing market price, it can be inferred that high pig populations at auctions are indications that the prevailing market prices are good. As such, market price is a driver for moving pigs to auctions. Such pigs often evade ante-and postmortem inspections and may inadvertently spread infectious diseases. It becomes necessary to identify each auction within Mpumalanga and know the farms and road networks that support them so as to plan and apply intervention strategies where and when necessary, for example, in the case of a rapidly spreading animal disease or for surveillance purposes. Secondly, market price had some degree of influence in moving pigs to the abattoirs but only a minority (8.3%) preferred this option for marketing. Similar results have been reported from Limpopo, where farmers travelled a long distance to obtain higher prices primarily at auctions and at abattoirs (Mokoele et al. 2014).
Furthermore, because of the association that exists between local price and local markets (Table 2), we inferred that the local market is highly influenced by prevailing local price, with minor influence from the age and sex of the pig. Primarily, sick, unthrifty pigs, mature boars and latematuring pigs are slaughtered locally and according to our results, only the local slaughter slabs were associated with the slaughter of sick animals. Comparatively, prevailing market prices are higher than the local price range and better quality products are often sent to the auctions (51%) and commercial abattoirs (8.3%). It is known that pork from mature boars has 'boar taint', a pheromonal smell (androstenone and skatole), and generally attracts much lower prices compared with other pork. In addition, slowgrowing, late-maturing pigs are characteristic of most small-holder pig producers, an indication that they will be presented for slaughter at later age. It is therefore not surprising that local market and age of pigs are associated.
Although a positive association exists between auctions and local slaughter slabs and home medication and ethnoveterinary usages, similar qualitative evaluations have revealed that small-holder pig farmers who sell at auction and locally tend to medicate the pigs themselves with ethno-veterinary preparations and also to use, sometimes incorrectly, long-acting oxytetracycline and ivermectin (Fasina et al. 2012;Mokoele et al. 2014). Although DARDLEA has public veterinarians and animal health technicians in all the districts and municipalities within the province who provide free services to small-holder farmers, it appeared that small-holder pig farmers lacked information about veterinary services in the province, or intentionally refused to seek veterinary assistance; these farmers hardly consult veterinarians and para-veterinary professionals (Table 2). Reasons for this disconnect must be established and a drive to gain small-holder farmers' trust and to make service accessible and affordable must be implemented.
Local slaughter for household consumption and sale within the community was prevalent in our study group (64%). This poses a high risk to animal and public health because preslaughter pig inspections are neglected, as highlighted above. Whereas local slaughter for community sale is There is an insignificant degree of association between auctions and sex of animals and the reason for this observation is evident. Top quality breeding boars are priced beyond the reach of small-holder farmers and good commercial boars are equally expensive. As such, the small-holder farms often settle for lower quality boars sourced from auctions. This practice exposes the small-holder farms to risks of infectious http://www.ojvr.org Open Access animal diseases and the genetic value of the boars is doubtful. It will be important for agricultural authorities to revise their strategies, evaluate these gaps and devise means of addressing them. A good interim measure may be the creation of district and municipal pig-breeding centres to multiply quality genetics and distribute them to small-holder farmers at minimum costs. Such an intervention would carry additional benefits of employment generation and provision of training in animal production, animal health and biosecurity.
Although it was established that the feeding of commercial rations alone is significantly associated only with poor to fairly good body condition, it was also established that such feeds are often rationed and pigs were underfed because of cost. However, results showed that well-conditioned animals tend to produce larger numbers of piglets per litter and early weaning was positively associated with good condition across the panel of feeding. The combination of good quality commercial ration supplemented with risk-free cooked swill may offer alternative feeding strategies in small-scale pig production.
Using partial budgeting and ROI, the 10-sow unit pig farm is economically unsustainable with commercial feed.   It appeared that the main driver of profitability in smallholder pig farms is the feed cost. Whereas the profitability of a pig production unit increases with an increase in the number of live-born piglets per litter (Kyriazakis & Whittemore 2006), our results only partially agreed with this assertion. Even when the production efficiency increased and preweaning mortality was reduced by 25%, a 10-sow pig production unit was still not able to break even in this analysis.
Feed is a pig farm input that is indisputably of utmost importance (Kyriazakis & Whittemore 2006); our analyses have confirmed this, as it was the most important determinant of profitability in small-holder farms. Although using commercially compounded ration by farmers should yield better quality products and improve reproduction and overall production, at this scale of production it was not financially feasible and viable. Previous studies have highlighted some of the reasons for the relatively high feed cost and suggested reasons why small-holder pig farmers rely on swill as an alternative form of feed (Phengsavanh et al. 2010;Roelofse 2013). We have similarly concluded that because of the infeasibility of feeding commercial ration, small-holder farmers will continue to feed swill and alternative feed sources for the unforeseeable future.
We are aware that it is potentially possible for small-holder pig farmers to make profit (Lapar & Staal 2010;Petrus et al. 2011;Phengsavanh et al. 2011) and have demonstrated that in this economic analysis. However, the ≤ 10-sow unit, small-holder farmers in Mpumalanga were able to achieve profitability mainly through the use of swill as a major source of feed. This practice is common to most small-holder pig farms elsewhere in South Africa (Gcumisa 2013; Roelofse 2013). The feeding of swill comes with potential risk of spread of diseases to pigs (e.g. salmonellosis, campylobacteriosis, African swine fever, classical swine fever, porcine reproductive and respiratory syndrome and foot and mouth disease) and possible transmission of zoonotic diseases from pigs to humans (Beltrán-Alcrudo et al. 2008;DAFF 2005;Haynes 2001). In addition, pig products originating from swill feeding may not reach the quality required by South African pig abattoirs.
In view of the above risks and knowing that it is more realistic for small-holder farmers to make a profit with units of between 50 and 100 sows (Roelofse 2013), we suggest that the agricultural authorities should assist farmers in the development of community self-help groups and farmers' cooperatives. Such calls have been made previously (Munyai 2012) and the successes associated with such organisations by small-holder pig farmers have been documented in Namibia (Petrus et al. 2011), Vietnam (Lapar & Staal 2010 and Lao People's Democratic Republic (Phengsavanh et al. 2011). Such cooperative organisations have the advantages of bulk purchase of feed with benefits of economies of scale and discounts (Costales et al. 2007;Lapar & Staal 2010), reduced transport tariffs through bulk transport and better negotiating power.
In addition, because an improvement in efficiency and reduction of preweaning loss by about 25% will improve profitability, it is necessary to implement the measures required to achieve these objectives. The government may also consider tax rebates on animal feed products that are directed to small-holder pig farmers. Only through the combination of the above measures and interventions will small-holder farmers with ≤ 10-sow units be able to break even and use pig production as a means of poverty alleviation.

Conclusion
Small-holder pig production and health management will continue to be relevant in an emerging economy like South Africa. However, for government agricultural interventions to provide the desired benefits, empirical evaluations for technical and economic feasibilities must be carried out. Although a 10-sow unit is technically feasible, in Mpumalanga and elsewhere in South Africa, the current input systems negate the benefits that should come with such programmes. Proposed models and revisions as suggested above may facilitate government interventions and make pig production more attractive to small-holder farmers.  Feb.

May
June July Aug.

Total
Opening bank balance   Feb.

May
June July Aug.

Total
Opening bank balance