Experiences with and the viability of a recycling pilot project in a Cape Town township

ABSTRACT

This paper documents the experience with an NGO-initiated recycling programme in Hangberg Township, Cape Town. The 11-week trial project achieved 70% sign-up but suffered substantial drop-out. The NGO’s attempt to brand block leaders as ‘eco-heroes’ was unsuccessful and there were concerns about the distribution of project benefits. The project diverted 26% by mass of the estimated solid waste refuse stream of 2.5 kg per person per week and made R1752 from selling 1.965 tons of recycling. Some 970 kg of compostables were diverted but not processed as part of the project. Project accounts revealed a benefit–cost ratio of 0.123 in which the sale of recyclables and landfill costs avoided were the only benefits and wages and consumables the main costs. While a more equitable distribution of benefits could improve the cost–benefit ratio, there are other good reasons for donors to fund projects of this nature.

KEYWORDS:

• Solid waste management
• recycling
• composting
• townships
• environmental health

JEL CLASSIFICATION:

• Q24
• Q52
• Q56

1. Introduction

Recycling has been described as the obvious solution to society’s mounting solid waste problems because it cuts down on landfill use, which has environmental and sanitation benefits and enables reuse which limits extraction and could lower the cost of production (Hornik et al., 1995). In the first world, there is a focus on how to efficiently divert the maximum fraction of household waste at source (Everett & Pierce, 1992; Hornik et al., 1995). In the third world, where solid waste management is usually inadequate, research centres on how to improve working conditions for informal waste pickers (scavengers) (Agarwal et al., 2005; Asim et al., 2012; Tirado-Soto & Zamberlan, 2013) and how to strengthen public landfill management (Henry et al., 2006; Ezeah & Roberts, 2012; Singhirunnusorn et al., 2012). The stylised facts of third-world recycling are that household waste contains a higher compostable fraction than in the first world and that its overall size correlates with income and ‘development’ (environmental awareness; Troschinetz & Mihelcic, 2009). In Gabarone, Botswana, household income affects the amount of plastic and paper packaging generated but not the overall size of the waste stream (Bolane & Ali, 2004). Recovery rates are generally unknown (Troschinetz & Mihelcic, 2009). Business models for private scavenging are not always viable (Asim et al., 2012), and where they could be improved the proposed improvements are not always socio-politically acceptable (Agarwal et al., 2005).

South Africa is an interesting, sparsely documented, case. White South Africa has had good municipal services, including a weekly curbside refuse collection service linked to well-maintained landfills. With the principle of equitable access entrenched in the Local Government Municipal Services Act (Act 32 of 2000), these services were extended to all neighbourhoods, although a lack of vehicle access makes it difficult to run the service in informal shack developments. In such cases ‘conveniently located’ skips are provided, which are emptied when full. Because the service is irregular, unattended black bags are often torn open by scavenging dogs, leaving a stinking mess that attracts flies and rodents (Froestad, 2005) and so many township residents dispose of their waste informally between shacks or on the margins of settlements (Froestad, 2005; Henry et al., 2006). Littering is a secondary problem caused by the generally unsanitary living conditions.

Recycling was not widely practised in South Africa before 1994. It entered the public discourse in 1998 as part of the National Environmental Management Act’s (Act 107 of 1998) three-pronged Zero Waste attack on refuse production. Some municipalities appointed private contractors, many of whom failed owing to low participation rates. Other municipalities collect clear bags alongside black bags, which also undermines commitment to recycling because the practice creates the impression that recyclables will simply end up back in the landfill. There is no penalty for not recycling or a service charge for waste collection. Scavengers work curbsides in rich neighbourhoods and certain landfill sites. Experience indicates that approximately 30% of the residential solid waste stream is commercially recyclable, although proponents of the Zero Waste concept believe the recyclable fraction to be much higher (Matete & Trois, 2008). Hout Bay has had more positive NGO-initiated experience. Between 1997 and 2004 the prize-winning Kronendal Recycling Depot created 14 local jobs to sort up to 75 tons of recycling per month dropped off by the community.

This article describes and evaluates an NGO’s attempt to introduce a township community, Hangberg in Hout Bay, Cape Town, to the Zero Waste Concept. The NGO is called Thrive Hout Bay and its byline is ‘No potential wasted’. It saw the main potentials in the project as the people of Hangberg and the income to be made from selling recycled materials as well as influencing the community to maintain a clean and safe environment for residents and tourists alike. This project was innovative because it chose the more respectable block leader model over the typical third-world scavenging model and addressed the wet as well as the dry fractions of the solid waste stream.

2. The community

During early colonial days Hout Bay was source of firewood for the Cape Town settlement and in time the valley was cleared for farming. Commercial pine plantations went in on the mountain slopes above the farms and the fishing harbour thrived. Fishing industry workers lived in flats directly above the harbour, at the foot of the Hangberg (Sentinel Peak). When the Harbour became the designated coloured community with the passing of the Group Areas Act (Act 41 of 1950), it had to absorb coloured families displaced from the village and surrounding farmland. The Harbour also attracted African fish industry and forestry workers married to coloured wives to avoid influx control regulations (Horner, 1983). Overcrowding led to the construction of backyard shacks around the Harbour flats that later spilled onto the firebreaks above the settlement. For the backyard dwellers, formal municipal services are available to some degree via their landlords, but further up the mountain conditions are as bad as in the nearby township of Imizamo Yethu.

Imizamo Yethu was established on the site of an abandoned pine plantation in 1991 to serve as the Hout Bay area’s African township. The plan was to provide 417 building sites on 18 hectares of land, with a further 16 hectares reserved as green belt for future infrastructure development (Froestad, 2005). This plan for orderly settlement was quickly overrun by rapid, disorderly immigration following the end of influx control (Oelofse & Dodson, 1997). By 2001, Imizamo Yethu’s population had grown to more than 8000 residents, of whom the majority had limited job prospects (Shortt & Hammett, 2013). Houses are modest shacks packed closely together with little outside space. Water is provided through communal taps and electricity connections are frequently illegal (Harte et al., 2009). Very few roads are vehicle accessible, which combined with the use of paraffin as the main cooking fire, creates a serious fire risk (Froestad, 2005; Harte et al., 2009). Sanitation is almost non-existent in informal housing areas (Froestand, 2005) and the settlement has a negative effect on property prices in its immediate vicinity (Oelofse & Dodson, 1997). Froestad’s (2005) assessment was that, owing to the community’s extreme economic marginality and the patronage offered by its strong civic organisation, it is unlikely that the area’s environmental health problems will be solved cooperatively.

3. Methods

The key metrics for recycling projects are recruitment, drop-out and waste recovery rates. The recruitment rate was defined as the number of participating households divided by the total aimed for in the pilot project. The retention rate was one minus the drop-out rate with the latter defined as the number who abandoned the project part-way through as a proportion of those who started out. The recovery rate was defined as the weight of recyclables and compostables as a proportion of the baseline solid waste stream of the project area.

The basic demographic descriptions for the Hangberg community were found to be rather lacking, as is often the case for rapidly expanding informal neighbourhoods. The most recent community survey reported a total of 12 700 households in the greater Hout Bay and Llandudno area (Statistics South Africa, 2016). The NGO reckoned that about 2700 of this total came from Hangberg, where local law enforcement and social services believed there to be 9000 residents in 2017. This gave a plausible average household size of 3.33 people, an assumption used in the remainder of this analysis.

3.1. Recording stakeholder experiences

Key informant interviews with project facilitators, staff and participating households allowed us to triangulate between competing views of the pilot project’s success. The site was visited by the two field workers on five occasions during March and April, both while the project was still running and after it ended. The first two visits were to the NGO’s offices and dealt mainly with the project’s aims, budget and implementation. Some of these details were clarified during follow-up discussions. During the last three visits the field workers ventured into the Hangberg community, first accompanied by the supervisor and later their own, when it became evident that the supervisor’s introductions led to a degree of social desirability bias. While interviews were mainly conducted as informal conversations, often on the street, most interviews were taped and additional notes were taken. Questions dealt with participants’ experience of the project, its perceived fairness, practicability and recruitment. Respondents’ familiarity with recycling and their assessments of its benefits and practices were also probed. Interviews took about 10–20 minutes, a period probably too brief to achieve full disclosure of concerns and expectations (Table 1).

Table 1. Key informant interviews conducted.

Group	Dates of interviews	Number of interviews per group
NGO	15 and 22 March 2017 11 April 2017	3
Supervisor	11 and 27 April 2017	2
Volunteer	18 April 2017	1
Block leaders	11, 18 and 27 April 2017	6
Households	11, 18 and 27 April 2017	7

The main limitation of the use of a convenience sample is that it produces anecdotal evidence which does not allow for a systematic examination of community attitudes to recycling, and sheds little light on for example the participation rates than might be expected should the city decide to roll out a large-scale recycling project.

3.2. The cost–benefit analysis

A standard cost–benefit framework was adopted to evaluate the project’s financial viability. For ease of comparison with other municipal budgets, pilot figures were converted into monthly figures per thousand participating households. Since this project required no capital expenditure, no discounting was needed. Only the market value of recyclables and landfill space avoided were counted explicitly as project benefits, with non-valued benefits listed for future analysis. This analysis improves on the cost–benefit analysis presented in Matete & Trois (2008) by explicitly valuing the actual sales of recyclables at their real prices and by accounting for project expenditures.

The prices and quantities of recyclables sold were obtained from project accounts. Since the pilot project did not run for long enough to allow for a full composting cycle, the accounts show a compositing fee and no revenue. If mainstreamed, this expense will convert into an income stream which was valued at a price of R30 per compostable ton, based on the performance of commercial composting operations. The quantities of the compostable fraction were estimated by NGO volunteers that are experienced gardeners, and are likely to be a conservative estimate. Some refinement is needed here.

Landfill space avoided applied a standard rate of R1.70 per kg to the total weight of recyclables and compostables removed from the waste stream. This rate divided the monthly operations and maintenance cost at the local landfill (R1.5 million) by its monthly weight accepted (880 tons). Project costs were obtained from project accounts. Although the project was funded by donations, the accounts reflect actual expenditures regardless of the source of funding. For the latter, some assumptions had to be made regarding the size of the baseline solid waste stream of participating households. The NGO reported that the 140 households that signed up for the pilot produced approximately 22 wheelie bin containers of solid waste per week, and that recycling reduced this number to about nine per week. No weighing took place at baseline or during the project. Full bins were estimated to weigh 100 kg each. When converted to per capita daily waste production, this assumption gave a figure 74% higher than an earlier estimate of 2.5 kg of solid waste per person per week and 90% higher than a published estimate for Botswana (Bolane & Ali, 2004). Assuming a lower weight of just 60 kg per full bin pulled the three figures into line, and placed the baseline estimated at 9.4 kg per household per week.

To convert the pilot project’s financial data into a budget for the entire community, the pilot’s figures were divided by 125.4 and multiplied by 1890 households, in other words 70% of the estimated 2700 households in the neighbourhood. The size and composition of the waste stream were taken as is from the pilot, at the actual prices. The only exception was compostables, which were priced at R30 per ton of compostable material, or roughly R120 per ton of compost. This low price was to allow for any unforeseen expenses associated with renting land on which composting could take place as well as operational costs. While it was felt that the pilot project provided generous support and supervision to project staff, this cost item was retained to provide for additional community training, which was felt to be lacking.

4. Results

4.1. Solid waste stream at baseline

The size of Hangberg’s residential solid waste stream was only captured informally at baseline. Project staff reported that 22 wheelie bins of 1 m³ each were collected from the project area every week, which at an average content weight of 60 kg each implies 2.83 kg of waste per capita per week, or 400 g per person per day. The NGO works on an assumption of 2.5 kg per person per week, and comparable estimates vary from 200 g per person per day (Matete & Trois, 2008) to 330 g per person per day (Bolane & Ali, 2004). The composition of the waste stream was not assessed at baseline. Troschinetz & Mihelcic (2009) reported that in the developing world the compostable fraction of household solid waste was twice as high as in the USA or Europe but that this estimate varies depending on income and diet. In Gabarone, 68% by weight of the waste stream is recyclable (Bolane & Ali, 2004).

4.2. Project implementation

The Hangberg Zero Waste Pilot Project ran for 11 weeks, from January to April 2017. It happened entirely at the NGO’s initiative, whose staff and volunteers were directly involved in implementation. In this sense, it is an example of top-down social engineering. The project was funded by the City of Cape Town via the organisational structures of the Hout Bay Partnership, a local coalition of civil society and business that is committed to finding solutions to the valley’s many social problems. For implementation, the NGO partnered with the Bay Harbour Market, a weekend trading space for community stallholders. The Market, which operates from an empty fish packing plant in the Harbour, is negatively affected by littering in its immediate environment. The Market contributed a holding facility for recyclables and a link to the person appointed as site supervisor.

Project design involved the use of 10 block leaders, whose presence have been shown to positively affect household participation in curbside recycling if block leaders have sufficient status and are well liked by their neighbours (Everett & Pierce, 1992; Derksen & Gartrell, 1993). In this case block leaders were overseen by a part-time site supervisor who lives in the Hangberg community and otherwise does the recycling for the Bay Harbour Market. The site supervisor was coached by NGO staff and volunteers throughout but could recruit block leaders freely from amongst her own circle. It is not clear what her selection criteria were, although it was shown elsewhere that low-income communities do steer public works jobs towards the most vulnerable members of their group (Nattrass et al., 2015). Four training workshops were held with the block leaders and site supervisor by the NGO. Formal contracts had to be signed by the block leaders and the site supervisor continued to meet with them regularly once a week for the first month and every other week during the second and third months. Once households were signed up, a further workshop followed that brought them together with the block leaders, the site supervisor and the NGO’s representatives. On this occasion, leaflets and equipment were distributed and the project’s procedure was explained.

To confer more status to the block leaders than is usually given to refuse workers, the NGO branded them as ‘eco-heroes’. Street sweeping is normally performed by workers of the Extended Public Works Programme (EPWP) in Hout Bay. This programme gained momentum in the context of Working for Water, an effort to clear invasive alien vegetation from fynbos catchments using labour-intensive methods (Hosking & Du Preez, 1999), and has since expanded to include other aspects of environmental clean-up and maintenance, such as Working on Fire and the Coast. Productivity in this sector is low and jobs come with the stigma of poverty. However, Froestad (2005) argued that in the nearby township Imizamo Yethu extensive littering could be interpreted as the community’s attempt to capture more EPWP jobs into their area. The site supervisor certainly promoted the project as a money-making venture, although it is not clear if she also said how the project’s benefits would be distributed.

The project’s exact procedures were not clarified during training. We found at least one block leader who indicated that she watched and imitated the behaviour of other eco-heroes because she was unsure of how to proceed.

As far as we could tell, the block leaders’ main role was to provide the inspiration for community recycling by setting a good example. They were expected to visit each of their client households once a week to advise on practical matters and to check on progress, for a total of 10 hours per week. Their contracts stipulated that they had to record their own working hours and each household’s contribution of recyclables, but as far as we could tell, households had to bring their clear bags to satellite recycling stations where they were gathered for transport down to the market. Satellite stations were available within 50 metres of households’ front doors. This was not the most practical arrangement, as other commitments prevented the eco-heroes from always being on hand to receive and document each household’s contribution to the project. To avoid this problem, certain eco-heroes checked on correct sorting during their weekly visits and took in the households’ contribution at the same time. They also picked up litter on the way there and back to set an example to others. These two practices were soon taken for granted and became the downfall of the project. It also came to our attention that the main holding facility had to be kept locked to prevent theft and was too small to accommodate sufficient material to make it worthwhile for the recycling company to collect the project’s materials.

The proposal for the wet fraction was for households to gather their vegetable peelings, food scrapings, used tea bags and so on in a 5 litre ice-cream tub with tight-fitting lid. When full, these tubs could be emptied into larger buckets with lids at the satellite sorting stations, from where compostables would be periodically removed, presumably by NGO staff. The amount of material involved did not justify the elaborate procedure and before long households were expected to bring their compostables to the central sorting station instead. It is not clear how the change in arrangement affected the size of the recovered compostable waste stream.

The eco-heroes generally did not complete their timesheets because it soon become apparent that neither the site supervisor nor the NGO representatives would monitor the block leaders particularly closely. The site supervisor probably took on too many of the day-to-day tasks of the project and paid too little attention to getting the best out of her co-workers. With nobody recording working hours or the individual quantity of sorted recycles, there was room for free-riding and principal–agent problems. Iyer & Kashyap (2007) prescribed quite harsh monetary penalties as the appropriate response to such negligence. Given how small the stipends are, and how difficult and expensive the monitoring of these individuals’ efforts would be, one should be careful not to incur more costs than the action would generate in terms of additional labour productivity.

4.3. Project financial performance

The project aimed to recruit 200 households, or 7% of the estimated number of residents in the community. It signed up 140 households of whom approximately 40 dropped out after 6 weeks. The project diverted 1.965 tons of paper and cardboard, plastic, glass and metal as well as 970 kg of compostable materials over the project’s 11 week duration. Assuming an average household size of 3.33 individuals and a weekly production of 2.5 kg of solid waste per person, this amounted to 26% of the estimated solid waste stream being diverted from landfills. This is 50% higher than the rate of recovery in the slums of Delhi (Agarwal et al., 2005) and similar to the accepted recycling rate of 30% for South Africa (Matete & Trois, 2008). By weight, the diverted portion of the waste stream consisted of 33% compostables, 36% paper and cardboard, 14% plastics, 12% glass and 4% tins and cold drink cans. It was mentioned by a respondent that some of the cardboard was scavenged from businesses but the origin and extent of this behaviour were not documented.

The sale of recyclables is one of two benefits counted in the calculation laid out in Table 2. Prices varied from R0.30 per kg for glass to R1.40 per kg for tins, with a weighted average price of R0.88 per kg for the dry waste fraction. According to the NGO, these are premium prices for the area and would include pick-up if the project was big enough to make it worth the buyer’s while. The project’s total income from recycling was R1752 for the 11 week period. Since a composting cycle could not be completed within this time frame, the wet fraction incurred a composting fee of R3000 for the 0.97 tons of compostable materials presented for recycling. When operating at scale, compost might become a source of revenue. This possibility was valued here at a conservative R30 per ton of compostable material.

Table 2. Project financial performance and scalability.

Budget items	Quantity (kg)	Price (R/kg)	Amount (R)	Per 1000 households targeted (R/month)
Income from recycling
Plastic	420	1.07	450	1 370
Tins	125	1.40	175	533
Paper and cardboard	1 070	0.93	995	3 030
Glass	350	0.30	105	320
Compostables	970	0	0	320
Landfill costs avoided	2 935	1.70	4 990	15 192
Total benefits			6 742	20 764
Consumables			15 563	47 386
Composting fee			3 000	0
Wages eco-heroes			21 600	65 768
Wages site supervisor			9 600	28 926
Management fee			5 000	15 000
Total costs			54 763	157 080
Shortfall			49 021	136 316
Benefit–cost ratio			0.123	0.132

The commercial compost price in the Western Cape is determined by the price of chicken (broiler) litter, which currently sells for R265 per cubic metre (weighing approximately 600 kg), or R442 per ton. Compost yield is a function of the materials used, ideally a 50:50 mixture of household compostables and pony bedding in the case of Hout Bay (see Martin & Gershuny, 1992). Even if land for the compost yard is made available free of charge by the municipality and a minimum disturbance approach is taken, there will be some labour and capital requirement that could easily wipe out a composting venture’s profits. For example, a handling fee of just R200 per ton, only 20% of what the pilot was charged, leaves a profit of just R42 per ton of compost, which at most will translate into a price of R20 per ton of compostables if pony bedding is donated in sufficient quantities to make the composting venture viable.

The other financial benefit considered was landfill costs avoided, at a rate of R170 per ton × 2.935 tons of material diverted. The landfill benefit only considers operations and maintenance costs; Hornik et al. (1995) estimated that landfill costs could increase threefold if land acquisition and environmental costs such as emissions and pollution were priced in. No attempt was made to value the environmental health or socio-psychological benefits of living in a clean environment. With the landfill costs avoided representing 74% of total benefit and sales amounting to just R1.27 per household per week, the business outlook for township recycling as a private venture is bleak, which explains why waste scavenging is limited to richer residential neighbourhoods and dump sites. Unfortunately, it also means that the site supervisor rather oversold recycling as a money-making venture for township households, which will make it harder to convince the rest of the Hangberg community to be converted to recycling in the future.

Supervisory and block leader wages accounted for a total of R31 200 or 57% of total project expenditures. Eco-heroes received a stipend of R600 per month, which at the EPWP wage rate of R3200 per month represents 4.125 days of work per month or a 10 hour day per week. The impression exists that their duties could be performed much more quickly, although it is always possible to spend more time motivating reluctant recyclers. The site supervisor drew a stipend of R9600 over a 6 month period, which represents 10.89 days of work per month at the EPWP rate, which comes to 2–3 days of work per week. Were the site supervisor to work only during the time slot allocated to the eco-heroes, her wage would be R400 per day, a trivial amount for the responsibility she held.

Consumables including protective clothing, recycling bags and composting containers, signage and a small amount of training and catering supplies accounted for R15 563 or 28% of the pilot project’s operational budget. The R3000 composting fee increases this percentage to 34% of the project’s recurring budget. The R5000 management fee charged by the NGO for facilitation, training and management support came to 9% of the operating budget.

These costs and benefits resulted in a project shortfall of just over R49 000, or R35.54 per participating household per week. The benefit–cost ratio was therefore 0.123, suggesting that this venture should not be supported using public money, although it is recognised that some of the problems described below did affect participation and hence the amount gained from the sale of recyclables. However, experience elsewhere and the relative magnitudes of costs and benefits involved in this project suggest that a shortfall will persist even under ideal recycling conditions (Martin & Gershuny, 1992; Everett & Pierce, 1992). Since the tourism value of fynbos landscapes has been set at R2268–4570 per hectare in constant 2010 prices (Turpie et al., 2003; Conradie & Garcia, 2013), there might be donors wanting to invest in projects that will reduce landfill growth while others might be interested in improving environmental health.

There would be only marginal changes to the benefit–cost ratio if recycling was to become a community-wide practice. It is unclear if there is a demand for compostable material locally, or how much it would cost for the municipality to set up a dedicated plant in the area, but presumably there would be savings on making compost rather than paying an outside contractor to do it. A larger scale programme might have some economies of scale in management that could bring about higher labour productivity, but at the same time a more comprehensive effort could meet with more community resistance, so that there would not necessarily be savings on street sweeping or the ability to divert resources (trucks, workers) from refuse collection to recycling. For this reason, the final column in Table 2 merely assumed costs and benefits to be scaled proportionally according to the number of households involved.

4.4. Community perceptions of the project

The consensus feeling amongst beneficiary households was that recycling is a good thing, because it takes hardly any time or effort and leads to a cleaner environment, which is desirable. These positive comments were contradicted by the high level (30%) of drop-out experienced during the project. The eco-heroes were equally positive, although they revealed a certain degree of confusion about their roles and responsibilities despite signing a contract that described these. The site supervisor was more ambivalent. She recognised that being paid compromised her and the block leaders’ acceptability to the rest of the community. We are not clear, however, whether she would be willing to perform these functions as a volunteer. While she seems to be wanting to do so, the amount of time she would be able to devote to the task is likely to be constrained by her own family’s needs.

The interviews revealed two complaints, both to do with how project benefits were distributed. One women told us very clearly that she was not prepared to ‘work for’ the site supervisor, but rather wanted to recycle by herself for her own account. This person was clearly unaware of how little profit there was in recycling and was envious of the stipends drawn by project staff. Other people articulated the perceived unfairness of the way in which project benefits were distributed as the eco-heroes not doing enough to justify their stipends. Complaints included that streets were still full of litter and that they did not even bother to collect curbside recycling. Although some eco-heroes were fully aware of these accusations, they fought back, telling us that these functions were not part of what they were contracted to do. It is evident from this account that the eco-hero branding failed, as residents failed to distinguish between the role of the eco-heroes and the functions performed by their usual EPWP street sweepers. The eco-heroes’ failure to comply with community expectations was considered an important reason for the high drop-out rate, amongst both households and the eco-heroes themselves.

Another interesting comment made by one of the eco-heroes was that there were not enough community workshops held to motivate and educate participating households and that this was the main reason for the high level of drop-out. Clearly this individual failed to understand that her main role was education and motivation and the community workshops were only meant to initiate this process. This view might be linked back to standing, as this person clearly felt unable to provide the leadership usually associated with the position of block leader. The real question is if it is reasonable to expect the simple act of branding to confer sufficient leadership status on an unemployed community member to allow them to influence their neighbours’ habits. If this cannot be done, the whole model collapses to that of public works programme, in which case Froestad’s (2005) poor prognosis regarding non-cooperation will probably hold.

A final revealing observation was that some eco-heroes were willing to assert that some of their fellow workers were lazy in private, but not in public, when it was generally claimed that everything was ‘fine’. The site supervisor sometimes recounted different versions of events than our field workers were given, which raises a broader question of reliability. For example, four of the eco-heroes claimed that they would continue their tasks even without the stipend, but according to the site supervisor only two of the four actually did this. The inability to really understand motivations and actions could compromise project planning during the next stage of implementation.

5. Conclusions

The Hangberg pilot project was a brave attempt to mobilise community cooperation to advance the Zero Waste vision embodied in the National Environmental Management Act (Act 107 of 1998). The project demonstrated that township recycling is possible and can incorporate the wet fraction of the household solid waste stream. It confirmed that the main financial benefit of recycling was landfill costs avoided (Everett & Pierce, 1992; Martin & Gershuny, 1992; Hornik et al., 1995). Since fewer landfills are a public good, this kind of initiative ought to continue with external funding. One of the most interesting lessons learnt was that, owing to the marginality of many of the livelihoods in this community, there is a high level of distrust and envy amongst residents, which means that exceptional care should be taken to make explicit project benefits and their distribution. The NGO tested an interesting blend of the typical first-world self-help recycling and third-world scavenge-for-a-livelihood models, which presented implementers with some difficulty. It seems as if the community might prefer the more familiar public works service delivery model, although this mode of operation is likely to undermine the importance of the recycling message. More community building will be required before the community’s preferences could be investigated and it remains to be worked out how the new recycling system and the public refuse collection will best interact.

Acknowledgements

Input from Gwynn Banning of Thrive Head Office (and two other anonymous reviewers) is gratefully acknowledged.

Disclosure statement

No potential conflict of interest was reported by the authors.