Download presentation

Presentation is loading. Please wait.

Published byJosiah Moxley Modified over 2 years ago

1
Marie Briguglio Kerb-side enthusiasm The determinants of voluntary waste separation effort in Malta

2
INTRODUCTION Imagine: daily kerb-side collection of mixed municipal solid waste is free. Introduce: weekly voluntary kerb-side collection of separated waste at a fee Requiring more space, more time, more money Outcome: National kerb-side recycling scheme takes off… ….at different rates in different localities. 2 Introduction | Literature | Model | Data | Estimation | Conclusion

3
INTRODUCTION 3 Total Collected over 86 weeks Introduction | Literature | Model | Data | Estimation | Conclusion

4
- a case study in behavioural-environmental-economics - co-operation in a social dilemma situation – in the field - empirical analysis of effectiveness of intervention elements …contribution to the Literature INTRODUCTION – motivation 4 - high-stakes issue (externalities, resources, EU obligations) - potential for voluntary approach if enforcement too expensive. - potential to examine if policy for rational egoists misdirected. …contribution to Policy Introduction | Literature | Model | Data | Estimation | Conclusion

5
LITERATURE – Overview 5 Behavioural Environmental Economics e.g. Gsottberg e. 2011 (Behavioural environmental policy) Shogren Taylor 2008 (Behavioural Environmental Economics) Theoretical models for household recycling behaviour e.g. Halvorsen 2008 (Moral motives) Nyborg Rege 2003 (Crowding out/in) Empirical studies on recycling e.g. Sidique et al 2010 (Minnesota) Hage and Söderholm 2008 (Sweden) + OECD 2008 (review) + 30 field studies Introduction | Literature | Model | Data | Estimation | Conclusion

6
LITERATURE - Summary Interventions Pecuniary, Convenience, Communication Socio Economics Education, Gender, Income, Age, Political, + Dwelling/Community Motives Narrow self interest, Moral, Social + Non cognitive Introduction | Literature | Model | Data | Estimation | Conclusion

7
Household derives utility from consumption, leisure, environment, household space, moral well-being and perceived social respect. Conditional upon household characteristics. U i = U i (x i, l c i, e i, a i, s i, v c i, ; d i ) MODEL 7 Recycling increases household’s moral well-being. Positive in government communication. a i = a i (w r i ; GC r ). where δa i /δw r i >0 and (δa i /δw r i )/δGC r >0 Perceived social-respect decreases in distance of the household’s recycling from the norm. Positive in government communication and frequency of kerbside collection s i = s i (|w r i - w rn | (GC r, GF r )). where δs i /δw r i > 0, δw rn /δGC r >0 and δw rn /δGF r >0 Introduction | Literature | Model | Data | Estimation | Conclusion

8
Perceived environmental quality increases with recycling, decreases with waste. Positive in government communication and frequency (and hence visibility) of kerbside collection. e i = e i [w r i, w u i, w rn (GC r, GF r ), w un (GF u )] where δe i /δw r i >0, δw rn /δGC r >0, δw rn /δGF r >0, δw un /δGF u >0 Waste collection comes at a price (as do all other goods and services) m i = GP r.w r i + GP u.w u i + x i Recycling diverts time from leisure. Government communication helps. l c i = l t i - l r i (w r i ; GC r ) where δl r i /δw r i >0 and (δl r i /δw r i )/ δGC r <0 Recycling consumes space. Frequency of waste collection helps v c i = v t i - v r i (w r i ; GF r ) - v u i (w u i ; GF u ) where δv r i /δw r i >0, (δv r i /δw r i )/δGF r <0 Recycling is drawn from waste, a function of consumption w t i (x i ) = w u i + w r i 8 MODEL Introduction | Literature | Model | Data | Estimation | Conclusion

9
MODEL - Formal Introduction | Literature | Model | Data | Estimation | Conclusion 1. U i = U i (x i, l c i, e i, a i, s i, v c i, ; d i ) 2. a i = a i (w r i ; GC r ). where δa i /δw r i >0 and (δa i /δw r i )/δGC r >0 3. s i = s i (|w r i - w rn | (GC r, GF r )). where δs i /δw r i > 0, δw rn /δGC r >0 and δw rn /δGF r >0 4. e i = e i [w r i, w u i, w rn (GC r, GF r ), w un (GF u )] where δe i /δw r i >0, δw rn /δGC r >0, δw rn /δGF r >0, δw un /δGF u >0 5. U i = U i (x i, l c i, v c i, w r i, w u i, w rn (GC r, GF r ), w un (GF u ); d i, GC r ) 6. m i = GP r.w r i + GP u.w u i + x i 7. l c i = l t i - l r i (w r i ; GC r ) where δl r i /δw r i >0 and (δl r i /δw r i )/ δGC r <0 8. v c i = v t i - v r i (w r i ; GF r ) - v u i (w u i ; GF u ) where δv r i /δw r i >0 and (δv r i /δw r i )/δGF r <0 9. w t i (x i ) = w u i + w r i 9

10
10 Solving (Lagrangian s.t. conditions), yields typical predictions about extent of recycling by the optimising household and allows us to make hypothesis about effect of interventions and constraints on voluntary recycling + Implications of intervention if optimisation conditions are relaxed MODEL Introduction | Literature | Model | Data | Estimation | Conclusion

11
MODEL - Estimation Expressing recycling as a linear function of the exogenous variables provides the basis for the empirical estimation Y it = α + β 1 G it + β 2 X it + β 3 C it + u it Y it kilograms of separated waste per capita G it vector of interventions (convenience, price, communication) X it vector of constraints (space, time, income) C it captures a number of relevant controls including SES u it represents the error term. i indexes locality of observation; t indexes time units (week) 11 Introduction | Literature | Model | Data | Estimation | Conclusion

12
The a-priori expectation is that Yit (recycling) increases with Government communication, frequency of collection, lower fees Lower opportunity cost of time, space, price Stronger norm/moral prefs e.g. efficacy belief, homogeneity Habit over time Model - Hypotheses 12 Introduction | Literature | Model | Data | Estimation | Conclusion

13
DATA - Context 13 MALTA 316 km 2 total area 68 locations 410,000 population 2 inhabited islands EU member state V. high voter turnout, bi-party system Lowest r in the EU Introduction | Literature | Model | Data | Estimation | Conclusion

14
DATA - Context 14 THE SCHEME Recycling “Tuesdays” 3 streams of waste In a grey bag At the kerbside At 0.08 euro per bag 86 weeks March 2008 Introduction | Literature | Model | Data | Estimation | Conclusion

15
15 DATA - Context Introduction | Literature | Model | Data | Estimation | Conclusion

16
16 DATA - Context Introduction | Literature | Model | Data | Estimation | Conclusion

17
DATA - Sources Data available on: Recycling waste volumes by locality, week Intervention design elements by locality, week Constraints by locality Socio economic characteristics by locality Controls by locality, week 17 FROM: Malta Environment and Planning Authority 68 Local Councils National Statistics Office Department of Information Malta Tourism Authority The Diocese of Malta interalia Introduction | Literature | Model | Data | Estimation | Conclusion

18
Data – Variables The dependent variable “Yit” Tonnes of separated waste in 68 localities/population (recpc)PANEL 18 The explanatory variables “Git” GC dummy variable extent of promotional effort (recpr) (recprXvote) CROSS GP dummy variable weeks of free bags (freebag) TIME GP dummy variable for period with tax on bags (mswtax) TIME GF variables for frequency of collection (recfreq), (mswfreq) PANEL GF dummy variable for missed collections (holiday) PANEL The constraints “Xit” TIME for leisure (oldpc) (teredupc) CROSS SPACE (densitybuilt) CROSS INCOME (spapc) CROSS The moral/social preferences “C” SOCIAL – (diversity) diversity index (tourists, singles, social cases)PANEL MORAL - (votepnpc) CROSS + Control variable TIME week number in the scheme (week) TIME Introduction | Literature | Model | Data | Estimation | Conclusion

19
Estimation - specification Recpc it = B 0 + recpr i B 1 + freebag t B 2 + mswprice t B 3 + recfreq it B 4 + mswfreq it B 5 + holiday it B 6 + densitybuilt i B 7 + spapc i B 8 + teredupc i B 9 + oldpc i B 10 + votepnpc i B 11 + diversity i B 12 + week t B 13 + u it That is, recycling per capita is a function of government interventions (price, frequency, communication), household constraints (space, income, time), moral preferences, normative effects and time. 19 Introduction | Literature | Model | Data | Estimation | Conclusion

20
GOZO COUNCILS LEFT OUT VariableNMeanSDMinMax recpc44690.3680.30202.980198 freebag44690.1620.36901 mswprice44690.500 01 recfreq44691.0300.17012 mswfreq44695.7841.25707 holiday44690.0350.18301 recpr44690.5770.49401 densitybuilt44697.3852.8062.37019.031 spapc44690.0710.0380.0110.188 teredupc44690.0790.0410.0220.187 oldpc44690.2200.0720.0700.378 votepnpc44690.4360.1510.1980.831 diverse44690.1072.386-3.0576.701 week446943.51724.819186 Introduction | Literature | Model | Data | Estimation | Conclusion Estimation – statistics

21
21 VariablesRE Estimation CoefficientSE Freebag0.01270.0088 Mswprice-0.07670.0101*** Recfreq0.1830.0286*** Mswfreq0.03210.0104*** Holiday-0.1980.0134*** Recpr0.1210.0353*** densitybuilt-0.01940.0063*** Spapc-2.0751.235* teredupc-2.6111.057** Oldpc0.3170.474 votepnpc0.8210.191*** diverse-0.01870.00951** week0.002210.00024*** Constant-0.0580.118 Notes: *** p<0.01, ** p<0.05, * p<0.1 N: 4469; Councils 52; r2_o 0.187; r2_b 0.233; chi2 383.3; sigma_u 0.113; sigma_e 0.156; rho 0.343 Introduction | Literature | Model | Data | Estimation | Conclusion Estimation – Random Effects

22
CoefficientSE freebag0.01280.00881 mswprice-0.07680.0101*** recfreq0.1790.0286*** mswfreq0.03150.0102*** holiday-0.1980.0134*** recpr-0.1390.0968 densitybuilt-0.01890.00599*** spapc-1.4221.194 teredupc-2.1431.016** oldpc0.1980.452 votepnpc0.5310.208** diverse-0.02460.00926*** week0.002210.000242*** recprXvote0.6050.212*** Constant0.01250.115 DISCUSSION : INTERACTING PR WITH VOTE Introduction | Literature | Model | Data | Estimation | Conclusion DISCUSSION : INTERACTING PR WITH VOTE

23
23 Estimation - tests Random Effects? Breusch and Pagan LM test Hausmann Test Multicollinearity? Pair-wise reveal some high correlations. Endogeneity? Price: Freebag, time dummy Promotion: Recpr, cross section dummy. Frequency: Recfreq and mswfreq not much variation in time. Introduction | Literature | Model | Data | Estimation | Conclusion Robust SE?

24
CONCLUSION - limitations 1. Regional data: Inferences about households/individuals? More households/individuals or more recycling? 2. Contamination of data: (regions, streams, sources for recpc) and reporting lag. 3. Reliance on time/cross section dummies for identification, omitted variables? 24 Introduction | Context | Literature | Model | Data | Estimation | Conclusion

25
CONCLUSION - strengths 1. Employs rich spatial and temporal panel data set with low level of aggregation, based on actual recycling, no influence on data collection processes, no distortion between reported and actual. 2. Documents recycling uptake w/o of pecuniary incentive and confirms role of intervention design other than price. 3. Finds importance of pro-government sentiment – for further research. 25 Introduction | Context | Literature | Model | Data | Estimation | Conclusion

26
Thank You 26 marie.briguglio@stir.ac.uk

Similar presentations

© 2017 SlidePlayer.com Inc.

All rights reserved.

Ads by Google