What does the data really show when it comes to Ontario’s Blue Box Recycling program*?
October 6, 2016
Dr. Calvin Lakhan, York University, Faculty of Environmental Studies, email@example.com, 416-736-2100 ext: 22612
In Ontario, the recovery of total recyclable material (per annum) of Blue Box materials has increased from 779,884 tonnes to 884,504 tonnes between the periods of 2002 and 2014. The costs of managing this system have increased by 91% during this same period. Both packaging producers and municipalities have expressed extreme concern over the inordinate rise in system costs relative to the increase in waste diversion. At this juncture, there remains considerable debate surrounding why material management costs have increased (where material management costs are defined as the costs incurred for collecting, processing and providing administrative support for recycling waste) . Increases in costs have been attributed to decreased revenue from the sale of recyclable material, an increasing trend for producers to switch to “light weight” and more complex packaging, and inefficiencies in municipal waste collection and processing systems. However, is it possible that rising system costs are a result of the province’s decision to emphasize diversion and recycling of the broadest range of materials?
While the 90s and much of the 2000s were characterized by numerous successes, our provincial Blue Box recycling rate has stalled in recent years and actually decreased in the past two. Recycling system costs have increased by more than $100 million dollars since the Waste Diversion Act’s formal inception.
This study attempts to explore some of the potential drivers for increases in recycling system costs over the past decade. York University has already been exploring the economic drivers of recycling behavior for quite a while, but this was our first attempt to quantify the relationship between changes in the packaging mix and increases in recycling system costs (using the historical data set from the Stewardship Ontario Pay in Model).
A brief comment on methodology and definitions
For the purposes of this study, the terms core and non-core Blue Box materials refer to:
Core Material: We define a core material as possessing the following qualities: 1) High recyclability 2) Generated in significant quantities by households 3) Low cost of material management and 4) Accepted by most municipalities for inclusion in the Blue Box program. Using this criteria, the following eleven materials have been classified as core materials: Newsprint, Magazines and Catalogs, Telephone Books, Other Printed Paper, Corrugated Cardboard, Boxboard, PET Bottles, HDPE Bottles, Steel Packaging, Aluminum Packaging & Glass.
Non-Core Materials: There is little available literature regarding what constitutes a “non-core” material. Generally speaking, the characteristics of a “non- core” material include: 1) Low levels of recyclability 2) Poorly developed end markets 3) High cost of material management and 4) Low realized revenues from sale of material. Using these criteria, 7 materials were classified as non-core materials: Gable Top Cartons, Aseptic Containers, Paper Laminates Plastic Film, Plastic Laminates, Polystyrene and Other Plastics.
These definitions are developed by the author, and do not refer to legislative definitions of obligated vs. non obligated materials as per REG 101/94 or the BBPP (A definition of basic and supplementary Blue Box waste as per Reg 101/94 can be found in the Appendix).
Our primary data sources include the Stewardship Ontario Pay in Model (From 2003-2016).
To convert all dollar values from nominal to real values, we use the Statistics Canada CPI index, with 2002 being used as our benchmark year. Denominating values using 2002 values allows us to make meaningful comparisons between time periods. There is an argument to be made that the inflationary index does not adequately reflect how costs specific to recycling have changed, particularly with respect to collection and processing operations. While this is indeed a valid criticism, there is greater utility in using an established inflationary measure, than attempting to construct our own.
While all reasonable attempts were made to maintain the integrity of the data set, the following manipulations were made:
- Aggregate all of the worksheets (From 2003-2016) into one master file that allowed us to examine changes over time (and potential drivers).
- Convert all dollar values into 2002 dollars using the Statistics Canada CPI Index. 2002 is chosen, as it is the benchmark year used by Statistics Canada to track changes in inflation.
From 2003 (used for 2004 and 2005 fees as well) to 2016, the average net cost per tonne has increased from $150 to $287 a tonne. However, this data is expressed in nominal terms.
Figure 1 Below illustrates the inflation adjusted Net Cost Per Tonne for the Blue Box program over time (the 2003-2005 data set was dropped from the analysis, as one data year was used for the 2004 and 2005 fee setting periods).
Figure 1: Inflation Adjusted Net Cost Per Tonne
Using our inflation adjusted data, we observe that net costs per tonne has increased from $135 to $221 a tonne (a ~63% increase). While bad, this is a slightly better story than the 91% increase observed using the nominal data.
However, given that one of the common assumptions surrounding increases in recycling cost are attributed to changes in the packaging mix, the next step was to model a scenario that constructed a “hypothetical Blue Box”. In our hypothetical scenario, we consider a program that does not collect non-core materials (using the definition described above). Instead, the Blue Box is confined to the “core” materials, i.e. Printed Paper, Boxboard/Cardboard, PET/HDPE Bottles, Steel, Aluminum and Glass.
These results are summarized in Figure 2.
Figure 2: Inflation Adjusted Net Cost Per Tonne for our Modeled Scenario.
In our modeled scenario, the inflation adjusted net cost per tonne has increased from $124 to $171 a tonne – a 39% increase over time.
In the modeled scenario, the projected rate of increase (with respect to system costs) is much lower. Things are suddenly starting to look interesting.
Now let’s take it one step further and compare the recycling rates of our current and hypothetical system.
Figure 3 is the Blue Box recycling rate for our current system.
Figure 3: Blue Box System Recycling Rates (Baseline)
Figure 4 is the recycling rate of our “hypothetical” Blue Box that is comprised only of core materials.
Figure 4: Blue Box Recycling Rates in our Modeled Scenario
Suddenly our “lagging Blue Box” doesn’t look so bad – we are actually recycling more than 80% of the material that public readily recognizes as being part of the Blue Box.
Figures 5 and 6 offer direct comparisons between our current (inflation adjusted) system and our modeled scenario.
Figure 5: Comparison of Baseline and Modeled Scenarios
Figure 6: System Impact of Including Non-Core Materials
Figure 6 is perhaps the most telling, as it illustrates the challenges of attempting to manage non-core materials. Their presence within the recycling system results in significant cost increases, while contributed negligibly to overall diversion rates. When viewed relative to the modeled scenario, the inclusion of non-core materials is adversely impacting both costs and recycling rate performance for the system as a whole.
To specifically isolate the statistical relationship between the delta in recycling system costs and changes in the types of materials being recovered over time (how much of the variation in recycling system costs is explained by changes in the mix of materials being recovered), a regression model was developed that included all historical data entries. (For a full description surrounding the statistical methodology, please refer to the appendices, or contact the author)
Table 1 below summarizes the results of our regression (materials have been grouped into larger container categories for simplicity).
While we will concede that there are probably exogenous drivers of system cost that are being omitted from this regression, the explanatory power of our model is sufficiently high to make informed statements. Based on Table 1, changes in the cost of Printed Paper, Composite Packaging and Other Plastics are the most significant drivers of changes in system cost over time. Printed paper is an interesting result, in that it is actually considered one of the “core” materials, and has a low net cost for recycling relative to other Blue Box materials. This, in part, is explained by printed paper making up a significant portion of all Blue Box materials recovered. Even small changes in the cost of managing printed paper are likely to affect the overall cost of managing the program. Other Plastics is also responsible for a significant portion of the overall increase in system costs, lending credence to the municipal position that these materials are difficult to manage within the existing Blue Box system.
While we are reluctant to offer any definitive guidance regarding what materials actually belong in the Blue Box program, it is time to start asking difficult questions regarding what we want the future of Ontario’s recycling system to look like. Particularly in light of the Waste-Free Ontario legislation, there is now an opportunity to revisit what materials are included in the Blue Box program, and perhaps add clarity to what should constitute an obligated vs a non obligated material. This study is not meant to provide any answers, but merely elucidate the economic challenges facing recycling stakeholders. The hope is to start a conversation about what changes need to be made moving forward. Ontario has achieved some amazing things with the Blue Box program and continues to be a global leader in residential recycling. However, the province needs to be able to glean from previous experiences, and be adaptable, flexible and willing to ask tough questions in order to come to rational and evidenced based decisions.
For the statistically inclined, we test to see whether a random or fixed effects regression should be used in place of a pooled OLS analysis. The testing reveals that the null hypothesis is rejected, as the variance across entities is greater than zero. To determine whether a fixed or random effects model should to be used, a Hausman test was conducted to see whether the models unique errors (ui) were correlated with the regressors. The results show that cross-sectional variance components are zero, suggesting that a random effects regressive model is the best available choice given the characteristics of the dataset.
Definitions as per Reg 101/94:
PART I BASIC BLUE BOX WASTE
- Aluminum food or beverage cans (including cans made primarily of aluminum).
- Glass bottles and jars for food or beverages.
- Polyethylene terephthalate bottles for food or beverages (including bottles made primarily of polyethylene terephthalate).
- Steel food or beverage cans (including cans made primarily of steel).
PART II SUPPLEMENTARY BLUE BOX WASTE
- Aluminum foil (including items made from aluminum foil).
- Boxboard and paperboard.
- Cardboard (corrugated).
- Expanded polystyrene food or beverage containers and packing materials.
- Fine paper.
- Paper cups and plates.
- Plastic film being,
- linear low density or low density polyethylene grocery bags or bags used for food or beverages, and
- linear low density or low density polyethylene used for wrapping products.
- Rigid plastic containers being,
- high density polyethylene bottles used for food, beverages, toiletries or household cleaners (including bottles made primarily of high density polyethylene), and
- polystyrene containers used for food or beverages (including containers made primarily of polystyrene).
- Telephone directories.
- Textiles (not including fibreglass or carpet).
- Polycoat paperboard containers, being containers made primarily of paperboard and coated with low density polyethylene or aluminum, and used for food or beverages.
*This blog is provided as an opinion for consideration and discussion only. The views, opinions, and findings expressed are those of the author and do not necessarily reflect those of the CIF, its staff members, affiliates and/or governing bodies. The CIF makes no representations as to the accuracy, completeness, correctness, suitability, or validity of this information and will not be liable for any errors or omissions in the information or any losses, injuries, or damages howsoever arising from its display or use.