What does it cost to recycle household waste? A straightforward question with an answer so elusive that it has kept politicians, waste management contractors, local authorities, producers, consultants and academics busy ever since environmental concerns elevated recycling from a commercial operation to a modern society imperative. The elusiveness of the answer has not prevented mandatory recycling targets from being set in European and national legislation. ln the first of a two-part feature, Julia Hummel from Imperial College in London shows how a new model can improve our understanding of the factors that influence household waste recycling costs. ln the next issue of Warmer Bulletin, Julia Hummel explores the answer to the question' What is the optimum amount that should be recycled from household waste?



With mandatory targets on most UK local authorities to double recycling by 2003 (and under the Best Value proposals, treble it by 2005) and the proposed increase in packaging recycling targets, recycling household waste cannot be postponed any longer, whether or not an answer has been found to what it should cost.

Various studies have addressed the issue of cost, but results are inconclusive. The most recent study by Ecotec for Waste Watch reported costs from 85 - 240 per tonne. The studies conclude that the cost of collecting recyclables depends on the quantity and type of recyclables collected, local demographics and the type of system implemented.

Research being carried out at Imperial College and sponsored by the Environment Agency for England and Wales to complement its environmental analysis tool WISARD, might help determine what it should cost to meet current targets, and establish the optimum level of recycling that could be reached. This research has adopted a different approach from earlier studies, analysing standard rather than actual costs. The model incorporates all the variables that influence collection costs for a range of different collection systems (including bring), from commingled collection through to a full kerbside sort, and it calculates a standard cost for the system. Unlike other models, the infrastructure and resources required for collection are assessed by the model using information on the materials targeted and the demographics of the area in which the system is to be implemented.


The model can compare relative costs of different systems and can be used to analyse the sensitivity of the cost- influencing factors (eg householder behaviour and the addition of new materials). These analyses demonstrate that an understanding of what lies behind the performance (measured by the recycling rate) of a particular collection system makes it possible to improve collection efficiency, under some circumstances even without the need for additional resources.


Influence of householders on costs

So you, a local authority, have cherry-picked the 'best' households in your district and provided them with a separate collection system, your collection vehicles are busy all day, the participation rate is high but you are only recycling seven per cent and it is costing you almost 200 per tonne. Now you have to double recovery levels, and no one is going to help pay for it. A daunting challenge but perhaps a better understanding of the factors influencing the performance might make this a little easier than it might appear.

The involvement of householders in any source separation programme is critical to its success. However, the number of householders participating is not the only important factor.

The number of people with access to recycling facilities, their willingness and ability to use the facilities and their recognition of the materials they need to separate out are all critical factors. Finally, recycling will be limited by the selection of materials targeted for collection, and their availability in the waste stream. Table I shows the impact that these factors can have on the amount that is recycled from a given geographical area. By keeping all other variables constant (eg type of collection, materials targeted, housing density), the model has examined the standard costs of achieving similar recycling levels that result from differences in household behaviour for a generic kerbside system. It is important to note that these are standard costs for a generic system. They are illustrative and do not represent a reference against which costs of existing systems should be compared. The results are shown in Table 2.

Returning to the dilemma facing the authority recycling seven per cent from its cherry-picked high participating households, the analysis illustrates that a cost- effective system would be one in which a lower participation was counter-balanced by a higher capture of targeted materials. Perhaps more surprising is that if the high capture could be achieved by only 20 per cent of households across the whole area and not only in the cherry-picked area, the same recycling rate of seven per cent could still be achieved but at a lower cost.

There are many other factors influencing the recycling rate, not least the range of materials targeted, generally market-driven. However, the analysis illustrates that the same recycling rate can be achieved at very different costs. A better understanding of householder behaviour, knowing whether to increase the capture or participation rate, is one way in which progress could be made towards achieving a higher recycling rate and a more cost effective collection.

Building on the examples given already, the data in Table 3 show the costs and related householder behaviour necessary to . achieve a doubling of the recycling rate in each of the three scenarios, assuming that the same materials are still targeted.

The data in Table 3 show that, although it is possible to double the recycling rates in scenarios 2 and 3 without increasing the number of households served, it would require high levels of both participation and capture. However, the system manager might find it is easier to target a smaller number of cherry-picked households (and motivate them to participate).

Due to the lack of information on the extent of promotion and education required to achieve a particular performance, the promotion and education costs are not included in this analysis. Clearly the high levels of participation and capture required in these examples would require extensive promotion and education campaigns to motivate and educate the householders and these costs would need to be included in the final analysis.

The results of the pilot schemes announced in the waste Strategy to encourage householders to participate should provide guidance to local authorities in planning these campaigns.



Evolution of cost with recycling rate

The data in Table 4 show that as the recycling rate increases the costs of the different scenarios become more similar. Using the same generic collection system modelled in the previous examples, but for this analysis keeping all the variables constant except participation and capture rates, it is possible to use the model to explore how these variables influence cost, with increasing recycling rate.

ln the first analysis illustrated in Figure I, all variables are kept constant except the participation rate. ln the second analysis presented in Figure 2 the capture rate is the only variable that is not kept constant. The data in Table 4 show the respective participation and capture rates for both analyses. As in the previous examples it has been assumed that 45 per cent of the waste stream is targeted. Note when looking at the graphs that the scale has a maximum cost in Figure I of 100 and 400 in Figure 2. ln both cases the overall cost and the cost per household increases with the increase in recycling rate, although the trend is much more pronounced in Figure I. Additional infrastructure is required in both cases as more materials are collected; however, in the scenario illustrated in Figure I there is a substantial additional requirement for vehicles that is caused by the additional collection stops that the vehicles have to make as the number of households participating increases.

ln both scenarios, the cost per tonne collected decreases as the quantity of material collected increases. This trend is most marked in the scenario illustrated in Figure 2 in which the increase in recycling rate is due to an increase in the capture rate. This can again be explained by the number of collection stops a vehicle can make during a day (assuming the number of loads per day is not a limiting factor).As there is little difference in the time it takes to collect a full container rather than an empty container the number of stops per day will determine the number of vehicles required. When participation is high, the , collection vehicle must stop at each household, regardless of how much material is set out for collection. Therefore, a vehicle that collects from a round on which only small quantities of materials are made available at each stop (as in Figure 2) will be able to collect less material in a given time than a vehicle that is able to collect larger quantities of material at each stop (as in Figure I). Therefore, a collection vehicle on the high participating low capture round will be able to collect less material per day and consequently will have higher cost per tonne than the vehicle on a lower participating round but in which more material is set out by each householder. The two scenarios illustrated here are extreme cases and in practice it is unlikely for a collection to have an extremely high capture rate coupled with a low participation rate. The results of using the model to evaluate a third scenario in which both the participation and the capture rates are increased simultaneously is presented in Figure 3 together, for the purposes of comparison, with the costs for the first two scenarios.


The trend in costs with increasing recycling rate is the same for this third scenario as for the scenarios in Figures I and 2; however, the costs lie between the costs in Figures I and 2. This illustrates that by modelling the extreme cases it is possible to establish the range in which the costs of a particular system operating in a specific environment should fall.




Clearly, there are many other factors that will influence the cost of achieving different recycling rates, for example, the set-out rate (the number of households who place their materials out for collection, generally slightly lower than the participation rate), the materials targeted, the household density, the type of collection system, the contamination and the extent of kerbside sorting. Also, the changes in cost will actually be stepped, as each additional vehicle is brought into service. These are all variables included in the model and their influence can be evaluated; however, these examples already illustrate the role that modelling can play in explaining the differences in costs between systems and its use in helping to understand the impact that the various cost drivers can have on the total cost. They also serve to illustrate the many factors that system managers must consider when planning how to increase the recycling rate.

Local authorities need to act quickly if they are to achieve their 2003 targets. By a collecting data on some of the key factors influencing the cost and performance of their collection system, municipalities already managing kerbside collections could be in a position to identify how they could meet the targets, for example, through expansion of the service, increasing the range of materials collected or by achieving higher participation, capture rates or both. Combining the operational information in a model such as the one described would further enable them to identify the most cost effective way of meeting the targets in their particular circumstances. Although by modelling standard costs the outputs are inevitably theoretical and do not necessarily represent the price that will eventually paid under a contract for a service, the analyses are valuable in that they provide a level playing field on which to compare the costs of different systems. Providing clarity on the costs of recycling household waste will enable full consideration to be given to the financial as well as the environmental impacts when planning recycling collections as well as in the preparation of legislation. lt should also facilitate collaborative working, as promoted in the Waste Strategy and under BestValue, between the stakeholders concerned with managing household waste, for the successful implementation of the goals and objectives they set out.



Setting optimum targets

The examples presented here illustrate that although the answer to the question What does it cost to recycle household waste? is complex, it need not remain eternally elusive. However, knowledge of the costs and the factors influencing them is useful but only the first step towards cost effective recycling. We need to look further before being able to answer questions such as whether the district specific recycling targets that have been proposed under Best Value provide the most cost effective way of achieving the national recycling targets? Could another basis, for example, setting targets based on housing density result in a more cost effective solution nationally?

The trends in all the examples that have been presented suggest that the most cost effective system is one that achieves the maximum recycling. However, common sense tells us that this cannot be correct and there must be a point after which the costs of collecting each additional tonne of material will increase disproportionately.

The next issue of Warmer Bulletin concludes this extended feature by considering how to establish an optimal level of recycling. Julia Hummel has worked for consultants Enviros and Coopers & Lybrand, and the European Recycling and Recovery Association. She is completing a PhD research thesis on the cost of implementing the EU packaging and packaging waste directive, through the collection of household packaging waste. Julia Hummel can be contacted at:

Imperial College of Science and

Technology, T H Huxley School

Royal School of Mines

Prince Consort Road

London SW7 2BP, UK

Tel: (int+44) (0) 20 7594 6398

Email: j.hummel@ic.ac.uk