Household participation in waste recycling: Some national survey evidence from Scotland

Abstract

Increasing levels of household waste have raced up national, regional and municipal environmental policy agendas around the world, especially as heavy reliance on landfill and incineration is becoming prohibitively expensive and also a high profile source of local voter dissent or vociferous NIMBY behaviour. In an attempt to reduce reliance on these options, UK local authorities have increasingly been forced to turn to recycling. This paper reports on a broad empirical study of household recycling, utilizing national survey-based evidence (drawn from over 31 000 interviews in Scotland). Hitherto, very few published empirical studies have addressed the household recycling decision using national survey data to furnish baseline evidence. For this purpose logit analysis is undertaken, using data extracted from the Scottish Household Surveys of 2000 and 2001. This work also serves as the basis for further detailed modelling of a selection of illustrative household types. After consideration of particular waste streams and the availability of different recycling schemes, the effects on the probability of recycling participation of age, income level, household composition, car ownership and local authority were analysed. While there are clear income effects, the picture regarding the impact of household composition and some other factors is more complex.

Notes

1 That said, a pay-as-you-throw scheme for household waste disposal is currently being trialled on a small scale in the city of Peterborough, Cambridgeshire, England.

2 In another study estimating the social net benefits of various kerbside recycling schemes, Aadland & Caplan (2003b) suggest, on the basis of their results, that recycling policies should actually be evaluated on a city-by-city basis.

3 In the 1999 SHS the basic recycling participation question was different (‘How often do you use recycling facilities such as bottle banks, paper banks and the like?’). Nine predefined categories were given, including ‘At least once a month’ and 32% gave this response or a more frequent instance. It would appear that fewer people recycled in 1999 but this is likely to be due to the differences in the question. The 2000/2001 surveys would include people who recycled erratically but who just happened to have done so in the preceding month. To maintain consistency, the 1999 dataset was not used.

4 More detailed information on the data, with standard deviations and a split according to whether the household undertook recycling or not, was presented in an earlier version of this paper, available at http://userweb.port.ac.uk/∼snellm/papers/recycle.pdf

5 This question only featured in the 2000 dataset.

6 This information was unfortunately only available for those in paid employment in the week prior to the interview.

7 Although imputed income data were available, it would be inappropriate to use these values since they are based on variables we wish to use to predict propensity to recycle.

8 The time of year might also influence recycling, either through the production of more waste material (e.g. the Christmas period) or through vacations. Monthly dummies were included in the logit analysis (although not reported here) and showed significantly less recycling compared to January interviews when the interview was in March, May and October.

9 The proportions will be slightly different from the combined 2000/2001 Scottish Household Surveys due to the case-wise deletion of missing data for the logit analysis.

10 This only serves a purpose where the household composition is significantly different from the base-line category.

11 Reflecting the negative income*age slope dummy which reduces the slope of the curve (in fact changing it from a positive through to a negative number) as age increases.

12 This classification of children is arbitrary. It could be argued that it is the younger children who make the heaviest demands on adults' time and ‘push out’ the time available for recycling activities. Alternatively, the older children might be more able to influence the household decision-making process (it is not clear in which direction this would act).

13 Some investigation was also undertaken as to whether to distinguish between those dwellings starting on the third and/or fourth floors.

14 By using a cross-tabulation of the LA and rural/urban variables it is also possible to devise a 112 (= 14 × 8) category variable which more precisely categorizes households into geographical regions. This might control more effectively for supply-side variations. In practice, about one-half of these combinations are not sensible or contain just a handful of households, so the number of dummy variables is much less than 112. The results for the non-geographical variables after undertaking this logit estimation were close to those of Table 3.