Environmental and Resource Economics 13: 473–491, 1999.

473
© 1999 Kluwer Academic Publishers. Printed in the Netherlands.

Household Waste Management in a Swedish

Municipality: Determinants of Waste Disposal,
Recycling and Composting1

THOMAS STERNER and HELEEN BARTELINGS

Department of Economics, University of Göteborg, Vasagatan 1, Gothenburg, Sweden

Accepted 21 April 1998

Abstract. This paper analyzes waste disposal, recycling and composting in a municipality in south-
west Sweden. In 1994, Varberg introduced a weight-based billing system for household waste charg-
ing 1 kr/kg of waste and at the same time recycling centers were set up and a “green shopping”
campaign was launched. This led to a significant reduction in waste collected and increased recycling.
This study had access to actual measured data on waste disposal at the household level for a residen-
tial area called Tvååker, in addition to survey data for the same households. This makes it possible to
carry out a more reliable and more detailed analysis than has been previously possible, particularly
with respect to attitudinal variables. The most important determinants of each individual household’s
waste were composting of kitchen waste, living area, age and attitudes concerning the difficulty of
recycling various materials. Separate sections look at composting behaviour, at willingness to pay for
sound waste management and for the sake of comparison three other municipalities are also studied.
The main finding is that economic incentives, although important, are not the only driving force
behind the observed reduction in municipal waste: Given the proper infrastructure that facilitates
recycling, people are willing to invest more time than can be motivated purely by savings on their
waste management bill.

Key words: waste management, recycling, incentives

JEL classification: D12, H31, Q20

1. Introduction
The management of solid waste from households is important for various reasons.
One of these is that landfill space is becoming a scarce resource in many countries.
More profound is perhaps the concern for ecological damage from hazardous com-
ponents in the waste collected by the municipality will not automatically alleviate
the concern about the spread of hazardous waste into the environment. To the
contrary: thoughtless construction of waste handling tariffs might even have the
effect of encouraging illicit dumping, burning or other improper disposal: see for
instance Fullerton and Kinnaman (1995).
This study focuses on the determinants of total household waste and the effects
of unit price payments combined with better access to recycling possibilities. One
474 THOMAS STERNER AND HELEEN BARTELINGS

of the difficulties encountered by most studies in this field is that they are forced
to make do with self-reported data. In Varberg and a number of other Swedish
municipalities a new system of waste-handling and rates is being tested that is
based on the weight of the garbage. The garbage disposal van weighs the dustbins
which have microchips for identification. The weights are used for billing. In this
study we have given access to the actual individual weights for the residents of
Tvååker, a residential area just outside Varberg, during a year with identification
making it possible to send surveys to the same residents and match the answers
with the actual amounts of waste delivered.

2. Earlier Research
The number of studies analyzing household waste generation is too large to discuss
here. The interested reader can consult Goddard (1995) for an extensive survey.
Most of the studies available rely explicitly or implicitly on utility theory of the
consumer or household production theory to derive a demand function for solid
waste management services. Most empirical studies use models that emphasize
socio-demographic factors. Not too many studies have made much use of disposal
charge since these have until recently not been very common but there are some
exceptions.2
In an early study Wertz (1976) estimated the responsiveness of consumers to
disposal charges for Solid Waste Services (SWS) by comparing two data points
from 1970; the per capita quantity of waste disposed in a city that employed user
prices and the per capita quantity of waste in the rest of the cities in the USA (rep-
resenting non-unit pricing cities). He then calculated a price elasticity of demand
equal to –0.15, that is, unit prices seem to have a substantial negative effect on the
waste discarded. He also collected cross-sectional data for 10 suburbs of Detroit
and estimated a model relating waste collected per capita to annual income per
capita. The estimates of income elasticity were about 0.3.
Jenkins (1993) has developed a model where households maximize utility
which depends positively on the quantity of goods consumed and negatively on
the amount of recycling. Included in the household budget constraint is a disposal
charge for SWS. Other variables that affect the demand for SWS are: the house-
hold income, the prices of goods consumed, the payment for recycled items, the
size and age distribution of the household, weather conditions and the degree of
urbanization. She empirically tested the resulting demand equation for data from
nine communities in the USA. Five of these had a volume-based fee and the other
four had the usual flat fee system.
The model relates the average quantity of household waste discarded per capita
per month in an area to the averages of the variables above. Income, high tem-
perature, percentage of population aged 18 to 49 and population density all had a
positive effect on quantity of waste discarded. The price for SWS, average house-
hold size and the price of recycled paper had negative effects. The price elasticity3
WASTE MANGAGEMENT 475

of demand for SWS is calculated as –0.12 and the welfare loss of not using unit
pricing for an average US community to be about $125 per ton.
Hong et al. (1993) derived a household recycling choice model and a demand
function for SWS and estimated these for a sample of households from the Port-
land, Oregon metropolitan area. Disposal fees are based on a block payment
schedule for pickup of a specific volume at a given time interval. The demand
for waste collection services is assumed to be a function of the incremental fee
associated with contracting an additional bin for waste disposal and the opportunity
cost of sorting waste into recyclables and non-recyclables, here equal to the female
wage rate. The number of persons per household, education level, race and rent
or ownership of their house was also assumed to influence the demand for SWS.
The results indicate a positive but small relation between an increased payment
difference and the demand for the quantity of waste collected. The effect of income
is also positive and significant but the relationship is inelastic. Of the other variables
only the value of time and the education level were found significant.
Miranda et al. (1994) examined the changes in household waste production
behavior upon implementation of unit-pricing systems. They used data collected
from 21 cities through the USA over a 18-month period. The results show that
unit-pricing and recycling-programs can have a dramatic effect on solid waste
flows.
Morris and Holthausen (1994)4 estimated a household production model of
solid waste management and implemented a simplified version using data on
household expenditures and waste flows in Perkasie in two periods. Period one
represents conditions and outcomes of household production choice with a fixed
fee. Period two represents conditions and outcomes when the household pays for
waste management by volume or weight and receives curbside recycling service
at no additional charge. They estimated price elasticities of –0.51 to –0.60 and a
Hicks’ compensation required by the household to offset benefits of a simultaneous
change to unit pricing, curbside recycling and once-a-week collection of $117 per
year.
Reading the above we find that not too many studies include many attitudinal
variables and not too many have access to detailed individual data on actual waste
flows. Outside economics there are however a couple of relevant studies that have
looked specifically at attitude and habit formation: Gamba and Oskamp (1994)
are psychologists and have concentrated on attitudes, showing that concern for the
environment, social pressure and knowledge are important positive determinants
of recycling while personal inconvenience is a strong negative determinant. Ronis
et al. (1989) find empirical evidence that there are two stages of importance in
habit formation with respect to composting and recycling: that of initiation and
that of persistence. This finding is corroborated by Åberg et al. (1996) studying
composting behaviour in a suburb of Gothenburg.
476 THOMAS STERNER AND HELEEN BARTELINGS

Figure 1. Waste disposal in Tvååker.

3. A Survey in Tvååker

In 1994, Varberg introduced a weight-based billing system for household waste

charging 1 kr/kg of waste and at the same time a green shopping campaign was
launched in the stores. The fee was actually 1.25 with VAT and has been raised to
1.31 SEK per kg waste, with a minimum yearly fee of 300 SEK As a complement
to the disposal fee there is a free curbside collection of recycled paper and glass,
and the possibility of leaving other material fractions at a recycling center nearby.
The households themselves have now the responsibility of taking the bin to the
curbside for the solid waste truck to pick up. If there is no bin outside, the truck
just passes by and there is no fee for the households to pay. On average the yearly
fee has been reduced by about 230 SEK, and the average waste per household has
declined by 35%.
A mail questionnaire was sent (in August 1994) to each of nearly 600 house-
holds in Tvååker where all houses are single-family homes. The response rate was
approximately 76% after two reminders. The main attraction of carrying out the
study in Tvååker was that we had access to the actual quantity of household waste
as measured by the weighing and billing system. These data for each household,
between April 1993 and March 1994, were provided by the local authorities. We
could thus in this study assess the effects of attitudes, income and other variables
on actual behaviour as opposed to other questionnaires which are forced to make
do with self-reported behaviour. 5
We did however also include various self-reporting questions concerning recy-
cling and a question on Willingness To Pay for environmentally sound waste
management without any work by the respondent. Responses to the mail question-
WASTE MANGAGEMENT 477

Figure 2. Waste disposal in Varberg.

naire also reflect attitudes concerning recycling, environmental issues and waste
handling in general.
Figure 1 shows the average waste disposed of in Tvååker during the year when
weight-based charges were introduced (smoothed using a three-month moving
average). There is a significant decrease in waste disposed, which is much stronger
if one starts comparing from 19926 when the average was bout 25% higher but this
is masked by seasonal variation.
The data for the whole of Varberg in Figure 2 are more complete and up-to-date.
The effect of weight-based payment is clearly visible after January 1995. When
seasonal effects are removed we found an effective, once and for all, decrease by
29% or 300 tons/month. In the two years after that there is a discernible but very
small (12 tons/month or 1%) increase again; see Ljunge (1997) for more details.
To explain individual behavior we turn to our sample (see Table I) which shows
summary statistics; the variables were coded as 1–4, 0–1 or 1–3 depending on
the number of alternatives. We see that total waste for the year averaged 242 kg
but could actually vary from 6 to 914 kg! Turning to attitudes, 68% thought that
environmental problems were very important (value “1”), 28% thought they were
important (value “2”) and 4% had a “3” for little or no importance, giving a
“mean”7 of 1.36.
On average people claimed that they were composting 75% of their garden
waste and 60% of their kitchen waste. 11% did not do any composting at all. The
majority of the respondents gave as reason for this: lack of space or that they had
not yet bought the necessary equipment, but were planning to do so. The mean
figures for recycling are very high. People claim to be recycling about 90% of their
glass, paper, batteries and even hazardous waste. At least the latter is surprisingly
high and not generally supported by the quantities actually collected.
478 THOMAS STERNER AND HELEEN BARTELINGS

Table I. Characteristics of sample.

Variable Mean Std dev Min Max

Dependent variables
Total waste (kg) 242.4 165.9 6 914
WTP (SEK) 155 292 0 2000

Questions concerning attitudes

Importance of env. problems 1.36 0.55 1 3
Importance of nature 1.59 0.59 1 3
Importance of waste problem 1.72 0.68 1
Changing buying habits 2.20 0.96 1 3
Change in attitude toward waste problem 0.52 0.50 0 1
Practical difficulties recycling glass 0.02 0.13 0 1
Practical difficulties recycling paper 0.02 0.12 0 1
Practical difficulties recycling batteries 0.10 0.31 0 1
Practical difficulties for hazardous waste 0.18 0.39 0 1
Importance of diff. waste fees 2.07 0.87 1 4

Questions concerning actual behaviour

Recycling of paper (%) 89.0 22.7 0 100
Recycling of glass (%) 94.7 15.0 0 100
Recycling of batteries (%) 95.1 15.8 0 100
Recycling hazardous waste (%) 88.8 23.0 0 100
Composting garden waste (%) 74.6 33.8 0 100
Composting kitchen waste (%) 60.4 36.8 0 100

Personal variables
Gender (1 for female) 0.23 0.42 0 1
Marital status (1 for married) 0.88 0.32 0 1
Education (1 for higher education) 0.37 0.48 0 1
Age 51.1 14.1 24 91
Number of children 1.17 1.24 0 5
Number of people staying at home 0.68 0.77 0 5

Economical variables
Living area (sq m) 137 39 50 500
Garden area (sq m) 1649 5279 90 94000
Income (thousand SEK/yr) 281 98 22 720

Variables related to waste handling

Time spent handling waste (min/week) 29.3 33.8 0 420
Distance to recycling center (m) 898 1013 0 10000
WASTE MANGAGEMENT 479

The results for self-reported recycling and composting might be biased if the
respondents for instance want to give the impression that they are behaving in tune
with certain social norms. On the other hand a reduction in waste collection by 35%
does indeed suggest a considerable amount of recycling and composting, rather in
line with the figures reported.
The average time spent on handling waste within the household was slightly
below 30 minutes per week. The average distance to the recycling center was
under 1 km. The average living space and garden area were 137 m2 and 1.649 m2
respectively. The respondents thought the practical problems of recycling were
rather small, perhaps with the exception of hazardous waste. For paper, glass and
batteries, more than 50% considered it very easy to recycle.
The average age was about 51 years, which is quite high. This partly explains
the low level of formal education; less than 40% had college level or above. The
average level of income was about 280.000 Skr. per year. There were on average
3.1 members in the household (and 1.2 children), and of these, on average “0.7
persons” stayed at home for at least part of the day.
About 26% of the respondents claimed that the differentiated fees were very
important in their decision regarding waste handling, recycling and composting,
and another 41% considered them important. Only 6% said that the new system of
fees did not matter at all. The Willingness To Pay question will be discussed in a
separate section below.

4. Estimating the Demand for Waste Disposal

The demand for waste disposal can be deduced from an explicit utility maximizing
framework or from a household production framework. This is however fairly stan-
dard and will not be done formally here; the reader is referred to the references cited
earlier, in particular Jenkins (1993). Suffice it to point out that the demand must
depend on factors which affect the relative costs of various disposal options. “Ordi-
nary” prices do not generally play such a significant part here, but the generalized
“cost” in terms of time for composting, recycling etc. are the main focus of interest.
Both the composition of waste produced in the family and the relative availability
of time for this type of task depend crucially on the age and gender distribution
of the family as well as on their income and other socio-geographical variables.
Unfortunately some of these, such as living area, may have several opposing effects
on waste disposal and thus the net effect to be expected is not necessarily clear-cut.8
Various other assets and beliefs, knowledge and opinions within the family may
play an important role as well as the ubiquitous formation of habits. The equation
to be estimated can be stated as

D = F (X1 , . . . , Xn ) + e (1)

where D is the demand for waste disposal services measured as the amount of
waste left for the solid waste truck to pick up, and X1 , . . . , Xn are the variables
480 THOMAS STERNER AND HELEEN BARTELINGS

used to explain the amount of waste disposed. These include variables reflecting a
household’s demographic, physical and economic situation, the ease or difficulty
of recycling and finally attitudes and habits. The price of waste disposal would
naturally have been interesting but unfortunately the data we have do not cover
the period before the new weight-based billing system was introduced. It is thus
impossible to analyze the price effect directly.
Three different functional forms were used: Linear, exponential and loglinear.9
All three equations can be estimated using ordinary least squares when the appro-
priate assumptions about the normality of e are made. Table II shows the results
of the regressions. The variables which were highly insignificant (t-value lower
than 0.50) were omitted from the analysis. Only the variables income and dis-
tance were kept in the model irrespectively of their t-value. As economists we
have a special reason to be interested in the effects of income and prices and in
this case the distance to the recycling center was the closest proxy we could find
for a generalised “cost” of solid waste reduction. Considering it is cross-sectional
data, there is an acceptable degree of explanation provided by the independent
variables: They accounted for 35% to 45% of the variance observed. Most of the
categorical variables, like for example attitude toward the waste problem, were
used as dummy-variables which take only two values: 1 for very much concern
about the waste problem against 0 for slight or no concern.
Starting with personal and economic characteristics, the most significant is age:
Older people seem to be generating significantly less waste, perhaps reflecting
a more frugal lifestyle. This coincides with Fullerton and Kinnaman (1994) and
Jenkins (1993). Living area is positive and highly significant which might be sur-
prising since more space should make it easier to compost and recycle. However, it
seems the respondents all have sufficient space in this respect and it could be that
the larger the house the more consumption and thus the larger is the production
of waste. In principle this should not happen in a regression including income.
However, the variable “income” may not actually be a sufficiently good reflection
of consumption possibilities since the reporting of income is far from perfect.
This suspicion is reinforced by the negative and highly insignificant coefficient
for “income” itself. It is worth noting that this area has many old-age pensioners
(remember the average age was 52!) for whom official “income” may not be a good
description of their consumption possibilities.
The variable people staying at home is negative (and significant in the exponen-
tial model). If more people are at home at least during part of the day, they will
compost more and produce less waste. The variable time has a small positive value
and is significant in the log-linear model. If recycling takes more time, people
are less willing to recycle and produce more waste. The coefficients of the log-
linear model give us elasticities directly for the different variables. All the variables
are inelastic. This is especially interesting for the variable household members,
although this variable is not significant. It shows that the amount of waste per
person declines as the family size increases. This coincides with the results of
WASTE MANGAGEMENT 481

Table II. Regression results explaining the determinants of waste disposal demand.

Variable Equation (1) Equation (2) Equation (3)

Linear model Exponential model log linear model
Estimate t-value Estimate t-value Estimate t-value

Intercept 143.6 1.05 5.29 7.29 7.420 2.67

attitudinal variables
Importance of the waste problem –0.13 –0.89
Composting –113.2 –1.93 –0.48 –1.62
Change in buying habits
Change attitude toward waste problem 0.15 1.54
Difficulties recycling paper and glass 199.3a 2.37 1.22a 2.70
Difficulties recycling batteries –66.0 –1.48 –0.417 –1.70
Difficulties recycling hazardous waste 61.0 1.61 0.375 1.77
Importance different fees –15.6 –0.73 –0.16 –1.46

Behavioral variables
Recycling paper –0.002 –0.87 –0.218 –0.91
Recycling glass –0.004 –0.75 0.435 0.90
Recycling batteries –0.343 –0.90
Recycling hazardous waste 0.003 1.14
Composting of garden waste 0.560 1.19 0.003 1.14 0.192 1.19
Composting of kitchen waste –1.13a –3.26 a
–0.005 –2.78 a
–0.265 –2.91

Personal characteristics
Gender (fem = 1) –29.2 –1.21 –0.187 –1.55
Marital status 39.3 0.91 0.257 1.20
Education –36.1 –1.25 –0.110 –0.77
Age –3.11a –3.07 –0.019a –3.75 –1.17a –4.16
Household members 12.2 1.28 0.048 0.93 0.094 0.55
People staying at home –8.840 –0.68 –0.145b –2.03

Economical characteristics
Living area 1.22a 4.21 0.005a 3.38 0.460b 2.07
Garden area 0.000 –0.65
Income –9.780 –0.79 –0.060 –1.00 –0.070 –0.50

Variables related to waste management

Time spent on waste management 0.340 1.25 0.002 1.61 0.239a 3.22
Distance to recycling center 2.004 0.22 –0.064 –1.00 –0.015 –0.51

R2 0.39 0.45 0.35

R2 -adj 0.34 0.37 0.31

a significant at 99% level; b significant at 95% level.

482 THOMAS STERNER AND HELEEN BARTELINGS

various other authors such as Hong et al. (1993), Gamba and Oskamp (1994) and
Jenkins (1993).
Composting of kitchen waste has a strong and significant negative influence
on the production of waste. The more people compost, the more potential mixed
waste for conventional disposal is diverted and therefore it is logical that these
coefficients are negative. It may seem surprising that the coefficient for composting
of garden waste is positive (albeit small and insignificant) but this is most probably
due to the fact that this variable is correlated with the composting of kitchen waste.
Furthermore garden waste is probably not to any greater extent thrown into the
garbage anyway – it may just be left on the ground.
The variables recycling paper, glass, batteries and hazardous waste are found
to be very insignificant. This is surprising because one would normally expect
these variables to have a significant effect on the amount of waste produced. Note
that paper is the main component with respect to weight and that its coefficient is
negative. The insignificance may be due to the fact that these variables all have a
very high mean value. People are claiming that they recycle almost all their paper,
glass, batteries and hazardous waste. It is possible that the correlation between
reported and actual behaviour is causing us problems here.
Turning to the attitudinal questions we find that the expression of “Practical
difficulties with recycling of paper and glass” has a positive and significant effect
on the waste generated as expected. The more difficult the individual’s perception
of recycling, the less likely he or she is actually to recycle. When it comes to
people’s perceptions of the difficulties with recycling of hazardous waste we get
the same, albeit, not significant result. Surprisingly enough we get the opposite
(although again not significant) result for difficulties with batteries which seems
illogical. We should note however that these three attitudinal variables are very
strongly correlated and if we look at the total effect of all three it is still a strong
positive effect. Batteries are normally a very minor share (by weight) of the average
household waste and hence the effect of battery recycling is presumably less than
that of the other types of recycling.
We also find that a positive attitude to composting has an additional, separate,
impact leading to smaller quantities of household waste – over and above the effect
of the variables for actual levels of composting. This impact is not significant at the
95% level but at the 90% level and appears to be the variable that best picks up the
residual “attitudinal” variation since the variables concerning “the importance of
the waste problem” and whether people thought they should change their buying
habits had practically no influence at all.

5. The Determinants of Composing

One of the factors which exerts a large physical effect on the demand for solid waste
management is the degree to which people actually do compost. The compostable
WASTE MANGAGEMENT 483
Table III. Regression results for waste composteda .

Variable Composting kitchen

waste
Estimate t-value

Intercept 12.51 0.56

Attitude
Importance of the waste problem 6.60 0.99
Change attitude concerning waste –4.90 –1.15
Importance different fees 8.80 1.86
Composting of garden waste 0.46b 6.19

Personal variables
Gender (fem = 1) –4.15 –0.81
Education 2.83 0.45
Age 0.21 0.96
Household members 2.45 1.14
People staying at home 2.07 0.75

Economic variables
Living area –0.11 –1.77
Garden area –0.002 –0.99
Income 2.751 1.01

Other
Time –0.13c –2.27

R2 0.23
R2 -adj 0.18
a The equation for waste composted was estimated using a simple linear form and
OLS; b significant at 99% level; c significant at 95% level.

fraction of household wastes can lie in the range from about 30–50%. We therefore
turn now to an explicit analysis of the factors determining composting behaviour.
The dependent variable here was the percentage composted kitchen waste. The
results re shown in Table III. The explanatory power of the regression is somewhat
lower, perhaps reflecting the fact that the data in this case are not actually measured
but self-reported estimations. It could be that the answers to these questions are
biased because people want to give an estimation that is socially acceptable or
because they simply do not know the exact percentage of the amount of waste that
they are composting.
484 THOMAS STERNER AND HELEEN BARTELINGS

The most important variable explaining the composting of household or kitchen

waste appears to be the quantity of garden waste composted. This is very natural
not only from the viewpoint of habit formation but for purely chemical or tech-
nical reasons: it is actually a lot easier to compost both types of waste together.
Composts consisting only of kitchen waste tend to have an excess of compounds
rich in nitrogen but lacking in carbon. This can create a messy and smelly compost
but is easily rectified by adding garden compost material such as grass, leaves or
twigs which are rich in carbon and also provide a physical structure increasing the
penetration of air and acting as a barrier that keeps flies etc. away! People who have
access to garden waste are thus not only more likely to start composting kitchen
waste since they have the habits and equipment – they are furthermore much more
likely to be successful and thus to persist in their composting in the long run!
The variables reflecting personal and economical characteristics were generally
supportive of what might be thought of as natural but were not generally significant.
This applies to income, number of people in household, age and education. It also
concurs with findings from earlier studies by for instance Hong et al. (1993) and
Fullerton and Kinnaman (1994) (although the latter did find a negative influence of
education). Another rather natural finding is that the number of people staying at
home had a positive effect on composting.
Other variables of interest explaining composting behaviour include some atti-
tudinal variables such as the importance of fee structures for waste management
and the importance of the waste problem in general. These were however not sig-
nificant. The variable “time” is an answer to the question whether the respondent
considers that recycling and other waste management takes a lot of time. Naturally
those who think it takes a long time (be it because they are unaccustomed to it or
for other practical reasons), tend to be more reticent about doing it and thus the
negative coefficient is rather natural.

6. Willingness to Pay for Environmentally Sound Waste Management

So far we have seen a considerable interest in waste management and related issues
among our respondents: not only by their attitudes but more directly through their
actual behaviour; most of our respondents have shown that they are interested in
a sensible waste management and in fact more interested than would be motivated
purely by savings on their bills. However, some people are invariably less interested
or less capable of actually getting themselves involved than others. We wanted to
study this group and to see if they, too, had an environmental interest. We therefore
asked the following question:
How much more are you willing to pay in yearly fees so that another organi-
zation (such as the county council) would be responsible for taking care of the
waste and recycling problem?
The purpose of this question was to ascertain how much people would be willing to
pay – in cash – for an environmentally sound waste management without any effort
WASTE MANGAGEMENT 485
Table IV. Regression results for Willingness To Pay.

Variable Willingness To Pay

Estimate t-value

Intercept 604.584 1.78

Importance of the waste problem 74.089 0.91
Change in buying habits –147.868 –1.23
Composting –55.315 –1.01
Change attitude toward waste problem –71.194 –1.32
Difficulties recycling hazardous waste 98.321 1.29
Importance different fees 56.002 0.88
Gender (fem = 1) 128.272a 1.81
Martial status 103.439 0.72
Education –153.467a –1.73
Recycling paper 2.651 2.29
Recycling glass –1.542 –0.65
Recycling batteries –2.567 –1.30
Composting of kitchen waste –0.835 –0.88
Garden area –0.037 –1.21
Age –5.415b –2.42
Household members 23.608 0.69
Time –0.245 –0.37
Distance to recycling center 39.024 1.09
Income 4.153 0.12

R2 0.256

a significant at 90% level; b significant at 95% level.

or work on their behalf. The idea was that there might be some people who have
an abstract “interest” or feeling that they want to be “environmentally correct” but
who know little about the environment and about waste and may not even want to
know or get involved.
Obviously there is a considerable risk of misunderstanding associated with a
question such as this one. The people who are concerned with waste management
and environmental questions will generally realize that the most sensible solution
to waste management policies is source separation and necessarily does involve
the individual in sorting, recycling and composting. In fact the whole information
concerning waste management in the last few years has emphasized precisely these
aspects. Many who are interested and motivated could be expected to reject this
question as illegitimate and either answer 0 or not answer at all.
This is, in fact, exactly what happened! Only 57% of the respondents answered
this question. Of this 57% approximately 60% gave a WTP of zero. It is quite
486 THOMAS STERNER AND HELEEN BARTELINGS

understandable that all these respondents rejected as unreasonable the suggestion

that they should pay someone else to sort their waste. Thus the fact that many
respondents gave a WTP of zero is not because they do not care about waste col-
lection, but because they feel it is better to take care of he recycling and composting
themselves. We, however, wanted to reach the others – those who do not compost or
recycle but may have a less articulate and consistent analysis of the waste problem
but still some vague desire to be environmental. For the 23% of the respondents
who did give a positive reply this may still be an indication that some of them have
a willingness to pay specifically with cash (420 SEK on average) rather than time
and effort in order to contribute to a sound waste management.
Bearing this in mind the regression does seem to give some reasonable infor-
mation; see Table IV. Most of the variables had expected signs although only three
were significant even at the 90% significance level: gender, education and age.
The variable gender has a positive sign, which implies that women are willing to
pay more than men. Education and age both have negative signs. People with less
education seem to be willing to pay more. In Table III we found that people with
less education produce more waste so it might be logical that they would be willing
to pay more for waste collection. Also more education might make it obvious that
the suggested waste management is rather unreasonable. Younger people were also
found to produce more waste and be more willing to pay with money than time to
take care of it.
It is also worth nothing that there is a negative (although not significant) rela-
tionship between willingness to pay and time actually spent on waste management.
This would seem to strengthen the hypothesis that the payment is being seen as an
alternative to physical involvement and not as an abstract measure of interest.
Another measure of Willingness To Pay is precisely this estimated time spent on
waste management. On average the respondents reported that they spent just under
half an hour per week or roughly 25 hours per year. The average income in this area
would correspond very roughly to an after tax income of 100 SEK/hr. This suggests
that the average “willingness to actually get involved” in waste management 25
hours per year is much more valuable (corresponding to roughly 2500 SEK) than
the average 400 SEK/yr that the respondents to the Willingness To Pay question
reported. The question of course arises which value is most reasonable. Perhaps the
answer is in some sense between the two. On the one hand people’s valuation of
their free time spent on waste management may very well, at the margin, be lower
than the market wage. They may derive some satisfaction from this activity. On the
other hand the WTP question had a number of weaknesses already mentioned and
probably the 25 hours does give a better indication of interest for this issue even if
it is hard to translate the time spent exactly into monetary terms.
WASTE MANGAGEMENT 487

7. Waste Management in Eda, Mark and Åmål

A special study of attitudes and habits concerning recycling and waste was carried
out, by way of comparison, in three communities using a questionnaire based on
the one used in Tvååker.10 The three municipalities were chosen to give as much
variation as possible with respect to fee structures and changes in fee structure
during the period studied (1994). Eda in Värmland has in 1993 introduced a weight
based fee system much like the one used in Tvååker. They started with a fee of
1 kr/kg of unsorted household waste which has then been raised to 1.25 and now
1.40 kr/kg. Mark in Västergötland introduced during 1994–1995 a waste sorting
scheme with a differentiation in fee depending on the frequency of collection and
Åmål had no change in fee or waste collection system during the period at all. Their
system was similar to the one introduced in Mark.
For 1994 Eda reports a drastic reduction in waste collection from a level of
553 kg/household to 284. This is a reduction by 48% but it should be noted
that the figure for 1993 was estimated (from average number of bags times esti-
mated average weight) while the figure for 1994 was actually weighed which
could imply some bias. Mark had a reduction by about 12% from 600 to
526 kg/household. In Åmål the amount of waste was constant between 1990–1993
at about 628 kg/household but decreased during 1994, by 11% (and without any
apparent reason at least as far as the fee structure goes) to 559 kg/household in
1994.
In all three municipalities there has been an increase in recycling of glass,
paper and various other types of material an waste during 1994. This was particu-
larly noticeable in Eda where sorted waste of various kinds (including household
machines, metal, electronic gear etc.) can be left at recycling stations free of charge.
The figure for Eda is now exceptionally low and while part of this may be
explained by composting and recycling (see below) one might have suspected some
illegal dumping. This has not been observed by the local authorities (Rydström
1996). The waste disposal figures for Eda have actually continued to decline and
by 1996 they have fallen by another 18% compared to 1994! If this decrease can
be attributed to a rise in tariff from 1 to 1.25 kr/kg then we can calculate a crude
price elasticity which turns out quite high at around –0.7! Eda is now down to
1100 tons (233 kgs/household or 118 kgs/cap) which is very low indeed. It should
however be pointed out that the low level is to some extent explained by the fact
that this is a very rural community with a low population density of 11 persons/sq
km compared to 28 and 36 for Åmål and Mark respectively. The low population
density may make composting (and other local disposal?) easier.
According to our enquiry the decrease in waste can be explained by large
increases in composting, changes in buying habits and increased recycling. Eda
has the highest reported percentages of recycling for paper (85%) and glass (95%).
A number of regressions were run explaining the amount of reported recy-
cling; see Table V. Only for glass was there a significant difference between the
488 THOMAS STERNER AND HELEEN BARTELINGS

Table V. The determinants of recycling

Glass Paper Refundables Batteries Hazardous Household Textiles

waste machines

Constant 58.42 29.14 65.78 67.12 47.95 32.94 11.92

(8.55) (2.94) (16.02) (11.58) (4.52) (3.71) (1.36)

Previous 0.21 0.42 0.30 0.22 0.54 0.56 0.91

Experience (6.79) (8.66) (9.87) (6.94) (10.33) (11.04) (19.67)

Information 1.91 9.03 1.30 –0.41 –1.09 –1.10 5.20

on waste problems (0.71) (2.27) (1.17) (–0.19) (–0.21) (–0.26) (1.17)

Change in buying –0.88 7.60 0.42 –1.77 6.03 5.83 –2.46

behavior (–0.35) (2.06) (0.41) (0.86) (1.33) (1.51) (–0.61)

Nr of persons 1.00 3.59 0.07 –0.11 1.06 1.29 –4.43

in household (0.78) (1.94) (0.13) (–0.11) (0.47) (0.66) (–2.44)

Ease of recycling 9.67 11.17 0.26 6.55 0.66 –1.40 –0.29

(2.32) (1.68) (0.13) (2.13) (0.14) (–0.36) (–0.07)

Average age of 0.01 –0.23 0.08 0.06 –0.23 –0.01 0.04

adults in family (0.06) (–1.44) (1.81) (0.78) (–1.20) (–0.06) (0.23)

Attitude about 2.76 4.34 –0.77 3.70 –1.44 3.74 9.08

importance of waste (0.97) (1.02) (–0.65) (1.64) (–0.29) (0.87) (2.11)

Dummy 1, 6.81 0.75 –0.71 0.86 4.57 1.00 –2.03

Eda (2.27) (0.17) (–0.57) (0.36) (0.86) (0.21) (–0.39)

Dummy 2, 8.50 4.87 0.26 1.08 8.88 4.11 1.66

Mark (2.56) (0.97) (0.19) (0.41) (1.49) (0.83) (0.32)

R2 0.26 0.37 0.35 0.26 0.48 0.51 0.83

t-statistics in brackets.

municipalities (Eda and Mark and higher percentages than Åmål). Other signifi-
cant variables were the degree to which the respondent had been accustomed to
recycling in the past (this variable naturally picks up a large share of the relevant
individual characteristics), information and number of persons in the household.
The latter was positive for recycling of newspaper and various other types of bulky
waste but negative for textiles – presumably because families with many children
“recycle” clothes internally and thus have less waste textiles per family! Another
variable that had different effects on different types of waste was age: older people
were better at recycling refundables but not as good as younger people at newspa-
per and hazardous waste which is not entirely surprising considering differences in
economy, information and lifestyle. The last variable which had some explanatory
WASTE MANGAGEMENT 489

power was how difficult respondents judged the recycling to be, i.e. the distance
etc. to recycling centers. This was particularly important for glass, batteries and
newspapers.

8. Conclusions
The experience in Tvååker shows a significant reduction in waste collected and
increased recycling after a weight-based billing system was introduced. The most
important determinants were composting of kitchen waste, living area, age and
the perceived difficulties with recycling of various materials. The effect of age on
the amount of waste produced is quite strong. The elasticity of age is –0.9. Other
slightly less significant determinants were the number of people staying at home
during some part of the day and the time it took to recycle.
It is clear that most people are willing to spend a considerable amount of time on
recycling and composting and while economic incentives are an important element
in encouraging this process they are not its only determinant. The amount of effort
and time invested are greater than would be purely motivated by savings on their
waste management bill.
The surveys from Eda, Mark and Åmål broadly speaking confirm these results.
There is an increased interest across the country in recycling and composting but
it does obviously help a lot if the local municipality provides correct economic
incentives and good physical infrastructure that facilitates environmentally sound
waste management. Price sensitivity to tariffs appears to be a significant factor.
One of the contributions of this paper is that it has used actual (measured)
household data on waste delivered, together with survey data on attitudes and
other variables. An important result is that perceptions about the difficulties related
to recycling and composting are important determinants of the effort people are
prepared to make.

Notes
1. Thanks are also due to Håkan Wahlberg who collected the data and participated in the writing
of a preliminary report on Tvååker; Anette Hällerdahl and Marita Fagerling who carried out a
first version of the analysis described in Section 7. We would also lie to thank one anonymous
referee and Elbert Dijkgraaf for very useful comments at the EAERE meeting in Tilburg and
finally we thank the Swedish Waste Research Council for financial support. We would like
to emphasize that the paper studies waste reduction through increased recycling but does not
necessarily recommend it. Recent work, see for instance Bruvoll (1998), suggests that increased
recycling may, if wrongly designed, be bad rather than good for the environment.
2. See also Derksen and Gartell (1993), Gamba and Oskamp (1994), Fullerton and Kinnaman
(1994), Menell (1990), Miedema (1983), Neal and Schubel (1994), Pearce and Kerry (1992)
and US EPA (1990).
490 THOMAS STERNER AND HELEEN BARTELINGS

3. Two other studies are worth mention here: McFarland et al. (1972) reports –0.46 and Skumatz
(1990) –0.14. Both used revenues for the waste collection firms as proxies for the disposal price
variable.
4. See also Richardson (1978), Duggal et al. (1991) or Lackman (1976).
5. Unfortunately the figures are all for the period after the weight-based billing system is intro-
duced (billing is the reason why the trash is weighed in the first place). We can thus not, at the
household level, compare with conditions prior to the reform and therefore we cannot calculate
price elasticities.
6. Report (RVF 94:2). For 1992 we only have an aggregate level of waste collection for comparison.
7. Clearly this type of “mean” is not legitimate and we do not use them in the regressions, rather
we use dummies for more or less interest.
8. More space may facilitate recycling but may also reflect a tendency to have more objects around!
9. Variables which can take on the value zero (such as dummies for gender) were always used
in linear form. Tests between the functional forms were inconclusive and we felt it was most
instructive to keep all three for the sake of comparison.
10. See Hällerdahl and Fagerling (1995). 600 questionnaires were sent to 200 house-owners
(individual homes only) in each of the three municipalities. The reply frequency was 61%.

References
Bruvoll, A. (1998), ‘The Costs of Alternative Policies for Paper and Plastic waste’, Rapporter 98/2,
Statistisk Sentralbyrå, Oslo.
Derksen, L. and J. Gartrell (1993), ‘The Social Context of Recycling’, American Sociological Review
58, June.
Duggal V. G., C. Saltzman and M. L. Williams (1991), ‘Recycling: An Economic Analysis’, Eastern
Economic Journal 17(3), July–Sept.
‘The Effects of Weight- or Volume-based Pricing on Solid Waste Management’. EPA/530–SW–90–
047. Sept. 1990.
Fullerton, D. and T. C. Kinnaman (1994), ‘Household Demand for Garbage and Recycling Collection
With the Start of a Price per Bag’, NBER WP 4670.
Fullerton, D. and T. C. Kinnaman (1995), ‘Garbage, Recycling and Illicit Burning or Dumping’,
Journal of Environmental Economics and Management 29(1).
Gamba, R. J. and S. Oskamp (1994), ‘Factors Influencing Community Residents’ Participation in
Commingled Curbside Recycling Programs’, Environment and Behaviour 26(5).
Goddard, H. G. (1995), ‘The Benefits of Costs of Alternative Solid Waste Management Policies’,
Resources, Conservation and Recycling 13, 183–213.
Hong, S., R. M. Adams and H. A. Love (1993), ‘An Economic Analysis of Household Recycling
of Solid Wastes: The Case of Portland, Oregon’, Journal of Environmental Economics and
Management 25, 136–146.
Hällerdahl, A. and M. Fagerling (1995), Kompostering och Återvinningsvanor. Project paper for the
B Sc Degree 5:7, Unit for Environmental Economics, University of Göteborg.
Jenkins, R. R. (1993), The Economics of Solid Waste Reduction. The Impact of User Fees. Edward
Elgar.
Lackman, C. L. (1976), ‘Consumption of Solid Waste’, Journal of Economic Theory 13, 478–483.
Ljunge, M. (1997), Economic Instruments in Local environmental Policy. Project paper for the M Sc
Degree: 5, Unit for Environmental Economics, University of Göteborg.
McFarland J. L. et al. (1972), ‘Comprehensive Studies of Solid Waste Management’, Report to the
US EPA. Berkeley, May: The University of California.
Miedema, A. K. (1983), ‘Fundamental Economic Comparison of Solid Waste Policy Options’,
Resources and Energy 5, 21–43.
WASTE MANGAGEMENT 491

Miranda, M. L., J. W. Everett, D. Blume and B. A. Roy (1994), ‘Market-Based Incentives and
Residential Municipal Solid Waste’, Journal of Policy Analysis and Management 13(4), 681–698.
Morris, G. E. and D. Holthausen (1994), ‘The Economics of Household Solid Waste Generation and
Disposal’, Journal of Environmental Economics and Management, 215–234.
Neal, H. A. and J. R. Schubel (1994), Solid Waste Management and the Environment: The mounting
Garbage and Trash Crisis, Englewood Cliffs: Prentice Hall Inc.
Pearce, D. and K. T. Kerry (1992), Market-Based Approaches to Solid Waste Management, CSERGE.
Working Paper WM 92–102.
Richardson, R. A. (1978), ‘Economic Analysis of the Composition of Household Solid Wastes’,
Journal of Environmental Economics and Management (5), 103–111.
Ronis, D. L., J. F. Yates and J. P. Kirscht (1989), ‘Attitudes, Decisions and Habits as Determinants
of Repeated Behaviour’, in A. R. Pratkanis, S. J. Becker and A. J. Greenwald, eds., Attitude
Structure and Function. Hillsdale: NJ Erlbaum, 213–239.
RVF. Rapport (1994:2), Tvååker-projektet. Försök med vägning av avfall i kärl.
RVF. Rapport (1991:16), Differentierade taxor. Förstudie.
Rydström (1996), Telephone Interview with Mates Rydström, Miljö och hälskoskyddsinspektör, Eda
Kommun, 951127 and 961227.
Skumatz, L. (1990), ‘Volume-Based Rates in Solid Waste: Seattle’s Experience’, Report for the
Seattle Solid Waste Utility, mimeo.
US Environmental Protection Agency (1990), ‘Charging Household for Waste Collection and
Disposal: The Effects of Weight- or Volume-based Pricing on Solid Waste Management’,
EPA/530–SW–90–047. Sept.
Wertz, K. L. (1967), ‘Economic Factors Influencing Households’ Production of Refuse’, Journal of
Environmental Economics and Management April, 263–272.
Åberg, H., S. Dahlman, H. Shanahan and R Säljö (1996), ‘Towards Sound Environmental Behaviour:
Exploring Household Participation in Waste Management’, Journal of Consumer Policy 19, 45–
67.