Nikki Thompson Recycling and Waste Evidence and recommendations for high quality recycling programs. Canada, along with the rest of the world, is continually growing and expanding. Increased production and urbanization results in increased waste production leading to many negative environmental impacts. When faced with these ever increasing environmental concerns, Canada's focus needs to shift towards promoting sustainable recycling behaviours in order to mitigate these impacts. There is a link between high quality recycling programs and waste reduction, as shown in the multiple regression models discussed in this paper. Expanding on this, we take an in depth look at the impact of various factors on recycling programs. This can be used to build the basis of a successful and effective recycling program. As Canada continues to develop as a global economic player, the country also needs to address the by-products of this growth by developing a high quality recycling program. Such a program would support continued sustainable production and consumption by alleviating the negative side effects associated with the creation of more waste. Increased production and consumption results in increased waste; most commonly waste is disposed through the use of landfills. The impacts of growing landfills vary from minor inconveniences to much more serious issues. On the smaller scale, landfills create odour and noise and are not aesthetically pleasing. Landfills also require the use of large tracks of land. Larger impacts include local pollution, such as the contamination of ground water and soil. Electronic waste, such as televisions or computers, release hazardous substances as they begin to break down in landfills. These toxins are then absorbed into the soil, creating long term environmental impacts (Environment Victoria). The most concerning issue linked to landfills is the production of methane that results from the breakdown of organic material. Methane is a component of natural gas and is also a greenhouse gas. It is created when organic materials, such as paper, food scraps or yard waste begin to decompose (Howard, 2009). While carbon dioxide is commonly associated with the issue of global warming, methane is actually more potent on a per unit basis (Abrahm, 2011). In Canada, emissions from landfills account for 20% of the countries methane emissions (Environment Canada, 2013). Increasing levels of methane exacerbates the problem of global warming. Global warming is the result of greenhouse gasses, including methane and carbon dioxide, being released and collecting in the atmosphere. This traps the sun's heat and results in worldwide changes in climate. Scientific estimates predict that if emissions continue at their current rate, the temperature in some regions will increase by 3 to 9 degrees by the end of the century (NRDC, 2011). Global warming not only affects climate, it also will likely impact human health, food, water, and energy sources. From a health perspective, it is likely that an increase in temperature and increased occurrence of heat waves will lead to an increase in risk of death and illness. Furthermore, we will see increased incidences of diseases being transmitted through food, water, and insects. These health issues affect society disproportionally, increasing the risk to the elderly, children, and/or the poor (NRDC, 2011). On top of health impacts, Canada will see changes in its diverse geographical landscape. As temperatures increase, glaciers in Arctic ecosystems will continue to melt at an ever increasing rate. It is already estimated that in the last century more than fifty percent of B.C's Glacier National Park have melted. Not only are Arctic ecosystems impacted by climate change, but surrounding communities are affected as well. As glaciers decrease, the annual water flows produced in the summer months also diminish. In the last fifty years the summer flows of the Mistaya River have already decreased by thirty nine percent (David Suzuki Foundation). Many communities depend on these water flows each year. Other water supply sources, such as lakes and streams, are also impacted by climate change. Changes in rainfall and evaporation will cause lake levels to fall, and a decreased ability to recharge ground water will lead to smaller steams drying up as well (David Suzuki Foundation). The impacts of climate change are varied and pose a serious issue; recycling programs can be utilized to help limit the contribution of landfill emissions. Recycling programs can address the negative impacts associated with landfills and the resulting methane emissions, and in addition, recycling programs are also linked to a variety of other benefits. Recycling conserves natural resources, reduces energy, and can improve the economy. As we do not have an endless supply of natural resources, recycling can help to conserve these resources by reusing collected materials. On top of this, using recycled materials in the manufacturing process requires much less energy than using raw materials. Studies have indicated that systems using recycled materials in the production process can save up to 94 percent of the energy required when compared with systems that manufacture from raw materials (United States Environmental Protection Agency, 1998). These systems are also characterized by reductions in 10 major categories of air pollutants, 8 major categories of water quality indicators and pollutants, and solid waste (United States Environmental Protection Agency, 1998). Recycling also yields economic benefits. Implementing effective recycling programs and infrastructure will lead to decreased costs associated with landfill operation and waste collection. Expansion of the recycling sector generates new jobs in the areas of collection, brokering, manufacturing, and distribution of recycled and reproduced materials. Studies conducted in North Carolina, where the recycling sector supports around 9,000 state wide jobs, found that for every 100 jobs created in the recycling sector only 13 jobs were lost in the solid waste and raw material collection sectors (United States Environmental Protection Agency, 1998). Municipalities can also generate additional revenue through the sale of recycled materials to manufacturing companies (Leigh). The role of recycling goes beyond reducing waste, it also offers opportunities for economic growth. A major factor determining the level of waste in a province and the effectiveness of recycling programs is the environmental values and social norms of the province. These values are going to impact the disposal decisions of the individual. Such values and norms will differ significantly across provinces. To illustrate this we can compare British Columbia with Alberta. B.C., especially the city of Vancouver, places a high value on the environment and sustainable behaviours, and as such is considered to be a very green province. Currently the city of Vancouver has initiatives in place to become the 'greenest city' by 2020; this will be accomplished by targeting carbon and waste reduction. (City of Vancouver, 2013). As these attitudes have become the expected norm of the province, we assume those who violate this norm are subject to social stigma and are looked down upon by society. Looking at Alberta in comparison, social pressures associated with recycling are not evident. A major difference between B.C and Alberta is the presence of the oil sands. The Alberta oil sands represent the third largest proven crude oil reserve in the world and the largest growing source of greenhouse gas emissions in Canada (Best & Hoberg , 2008). Instead of setting limits on total emissions, the Alberta government aims to reduce emissions per unit of production (Best & Hoberg , 2008), which is expected to reach 3 million barrels of crude oil a day by 2018 (Alberta Energy). If we assume the approach Alberta's government is taking towards the oil sands is representative of the province's attitudes towards sustainability, we see there is not the same social pressures to be 'green' as there are in B.C. If both B.C and Alberta were to implement the same high quality recycling program, we may see an increase in the ratio of recycling to waste in both provinces, however the program is likely to be more effective in Alberta and result in larger gains. If residents of B.C. already face very strong social pressures to engage in recycling and sustainable behaviours, these social norms will drive their behaviour. Thus, an increase in the recycling-waste ratio cannot be solely attributed to the recycling program . However, in Alberta, if these social norms and attitudes are absent, the increase in the recycling to waste ratio is driven by the implementation of a sophisticated recycling program. Therefore we find such programs will be more effective in provinces like Alberta than provinces that already place a high valuation on the environment This idea of social norms can also be examined in other provinces as well. We can compare larger provinces with higher GDPs like Ontario and Quebec with smaller provinces like Prince Edward Island and New Brunswick. In such cases, the priorities and values of the provinces may differ when it comes to recycling and the environment. A province with a high GDP is able to focus on environmental sustainability; whereas smaller provinces are more likely to focus their efforts on economic growth. Environmental values and social norms differ across provinces; these values play a critical role in the disposal and recycling decisions of the individual. We will demonstrate the importance of these values in our regression and illustrate how high quality recycling programs work to increase the environmental value and awareness in society. Previous research done in the area of waste and recycling can offer valuable insight into the factors motivating the individual’s disposal decision. Studies done by Biswas, Licata, McKee, Pullig and Daughtridge find that only 27 percent of consumption waste is recycled with the remaining waste being thrown out (Biswas, Licata, McKee, Pullig & Daughtridge, 2000). The authors suggest that it is not only the attitudes toward an action that influence individual behaviour, but also what they believe others think of that behaviour as well (Biswas, Licata, McKee, Pullig & Daughtridge, 2000). In such a case, if you were indifferent towards recycling but knew your peers held very strong positive views about recycling, you will be more inclined to recycle than you would be if you believed your peer group was indifferent. In their paper the authors suggest using promotional content and communication to affect and shape attitudes towards recycling in a positive way. Advertisements about recycling will be most effective when they incorporate and highlight aspects of personal intrinsic rewards and approval of others (Biswas, Licata, McKee, Pullig & Daughtridge, 2000). Further studies conclude similar results, illustrating the importance of intrinsic motivation. Authors Ferrara and Missios find there is a positive correlation between environmental concern and waste prevention (Ferrara & Missios, 2012). They find that individuals who recognize environmental degradation and exhibit a strong level of environmental concern are more likely to adjust not only their recycling behaviours but also purchasing and consumption behaviours. On the other hand, individuals who are only concerned about increasing waste are less likely to adjust their behaviours (Ferrara & Missios, 2012). The findings of these two studies support our claim that high quality recycling programs need to take on an educational aspect which makes consumers aware of the impacts of their disposal decisions. We will create a model that illustrates the link between the quality of recycling programs and waste reduction. Expanding on this, we examine in depth the factors which contribute to increasing the recycling to waste ratio. From these results, we are able to suggest appropriate actions and policies to be undertaken by the government to design a high quality and effective recycling program. Regression 1 We first look to establish a link between high quality recycling programs and decreases in waste. This acts as a foundation from which we can further investigate the impacts of various factors on waste reduction. Our first regression model is as follows, Waste = ß0 + ß1inc + ß2age + ß3quality + ß4govspend + δiyeari + ∂jprovj Waste is our dependent variable and it represents the total amount of non-hazardous waste disposed in public and private waste disposal facilities. Waste is measured in tonnes per capita and our data was obtained through the Statistics Canada survey Waste Management Industry Survey: Businesses and Government Sectors. In order to utilize the data, units were converted from kilograms to tonnes and then the data was divided by population estimates for the provinces to obtain a per capita measure. Data was not available for Newfoundland or Nunavut for any of the years included and data is also missing for both the Yukon and the Northwest Territories for every year expect 2004 and 2006. Income is measured in millions of dollars per capita. To construct this measure, data on the Gross Domestic Product, calculated using the expenditure approach for each province, was obtained through Statistics Canada. This data was then divided by population estimates for each province to obtain income per capita. Data is missing for the Northwest Territories and Nunavut for the years 1996 and 1998. The expectation is that as income per capita rises, waste disposed will decrease and the amount of recycling increases. Age represents the median age of the province. It is the central age of the population in each province, as the median is the value that lies at the midpoint of a frequency distribution. Data was retrieved from Statistics Canada. The expectation is that as age increases, waste will also increase. This is based on the assumption that the younger generation will be more willing to adopt new norms centred around environmental sustainability, whereas the older generation is slower to adopt these changes. Government spending captures the amount governments spend on the operation of recycling facilities. This is measured in thousands of dollars and is obtained through Statistics Canada. Data was not available for Newfoundland, Prince Edward Island, The Northwest Territories, The Yukon, or Nunavut for any of the years and data was missing for 1996 and 1998 in Alberta. We expect that as government spending increases, waste will decrease. As government spending increases, we expect that recycling facilities become more efficient and thus able to handle and process higher levels of materials. The variable of interest in this case is quality, as we aim to establish a relationship between high quality recycling programs and waste reduction. We measure quality as the ratio between recycling and waste. The measures for recycling and waste were obtained from the Statistics Canada report Waste Management Industry Survey: Businesses and Government Sectors. The recycling portion of this ratio represents all residential and non-residential materials prepared for recycling. The residential sources include all non-hazardous recyclable materials either picked up by the municipality or taken by residents to depot or transfer stations. Non-residential sources include solid non-hazardous recyclable materials produced by commercial, institutional, construction, renovation, and demolition sectors. The waste portion represents the total amount of non-hazardous waste disposed in both private and public waste disposal facilities. This also includes all waste exported out of the source province or country for disposal. Both recycling and waste are measured in tonnes per capita. The recycling and waste measures were then divided to create the ratio between the two. Data was not available for Prince Edward Island, the Yukon, Northwest Territories, or Nunavut over the entire time period, and the observation for Newfoundland in 2008 is also missing. We expect that as the quality of programs increases, total waste will decrease. We assume that social norms and values of a province stay constant over time. As beliefs are adopted at a young age and learned through family ties and peer groups, we believe they are slow to change. This then allows us to use panel data and fixed effects model with province and year dummies to capture variation that we are unable to directly account for in the regression. Each year and province is represented in the model with a dummy variable. Running the regression we obtain the following results: Table 1: Regression 1 Variable Parameter Estimate Standard Error Income 6.25 ** 0.56 Age 0.01 2.14 Government Spending 0.00007 0.00046 Quality -0.31 *** 0.07 There were 104 observations in the data and 47 were deleted due to incomplete information. The residual standard error on 38 degrees of freedom is 0.046 and the adjusted R2 is 0.92. Coefficients are significant are the following levels: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1. The age and government variables were both statistically insignificant. Income was statistically significant, however when interpreting the results, we see it will take an increase of one million dollars to lead to 6.25 tonne decrease in waste. When looking at the units we see an increase in income actually has a rather small effect. Our quality measure is significant, indicating that an increase in quality will result in a decrease in waste disposed. Sensitivity analysis indicates that these results are still significant when differing the data by factors of -10%, -5%, +5%, and +10%, which demonstrates the robustness of the variable. Clearly, we can see there is a link between high quality recycling programs and waste reduction. Endogeneity Problem While we are able to demonstrate there is a link between high quality recycling programs and waste reduction, our previous model is subject to endogeneity problems. We are faced with two issues: an omitted variable, and a systematic relationship between the dependent variable, waste, and the independent variable, quality. A major factor influencing the individual’s disposal decision is both their individual values and the social norms of the communities. These social norms and values are a driving motivator of the individual’s actions. These values are not represented in our previous regression. As mentioned earlier, environmental values can vary from province to province for a variety of reasons. Not only do social norms affect the disposal decision of the individual, they can also influence the overall effectiveness of recycling programs. In our previous regression, it was assumed that attitudes of a province are constant over time and thus we were able to utilize panel data and the fixed effects model to address this issue. However, if we want to analyze the impacts of social norms on recycling behaviours, we need to capture the environmental values or “greenness” of a province through the use of an instrument. The second issue with our previous model is the link between the dependent variable and the quality measure. Quality is represented as the ratio between recycling and waste. As such, we see waste appear in both our variables and on both sides of our regression equation. This means if waste were to exogenously increase, we will see a change both in the dependent variable and in our independent quality variable. This leads to a larger endogeneity issue as these two variables are connected in this unintentional way. The regression results found by our original model show there is a link between high quality recycling programs and waste reduction. In order to address the endogeneity problems encountered in the original model and look more in depth at the various factors influencing recycling behaviours, we introduce a new model. Our new regression model takes the quality measure, the ratio of recycling to waste, and makes this the dependent variable. We also introduce an instrument to capture the environmental attitudes of each province. Such a model allows us to look more closely at what specifically impacts the ratio of recycling to waste. By looking at how different variables effect the recycling to waste ratio, we can determine the critical components of a high quality recycling program. We begin by establishing the links between consumption, waste, and recycling, and then introduce our new regression model. Model As we experience continued growth and expansion worldwide, we also see increased demand and consumption. This increase in consumption leads to a decision at the individual level on how to dispose of these units of consumption. The individual is now faced with a tradeoff between recycling or simply throwing these items in the trash. This decision is highly influenced by the both the individual’s personal preferences and the social norms of society. By modelling the relationship between consumption, recycling, and waste, we can illustrate the links between these behavioural choices and the environmental values held by the individual. This relationship can be modelled as follows: C = w/(1-ß) + r/ß + E C represents consumption, w represents waste, and r represents recycling. E is an error term which captures products in which there is not an option to recycle and also durable goods that are long lasting and not consumed and recycled in one period. ß is a preference parameter capturing the individual’s preferences towards recycling and the environment. This can be thought of as the social norms of society towards recycling previously discussed. This function symbolizes the consumer’s tradeoff and preferences towards recycling or simply disposing a unit of consumption. We can solve this model for the ratio of recycling and waste: r/w = ßC/w – ß/(1-ß) – ßE/w From here we can see that when ß increases, r/w increases. If the individual’s preferences towards the environment and recycling suddenly increase, they will be inclined to recycle more out of their consumption, everything else remaining constant. Changes in C will have an ambiguous effect depending on the ß parameter. If the individual naturally places a high value on the environment, and consequently has a high ß, then when the total amount consumed increases, they will recycle a higher proportion of this increased consumption. On the other hand, if the individual has a low ß and places a low value on recycling, an increase in consumption results in increased waste, thus lowering the ratio of recycling to waste. Clearly ß plays an important role in influencing the individual’s behaviours. We extrapolate the ideas evident in this model and apply them to our regression model in order to further examine the impacts of these variables, in particular ß, on the ratio of recycling and waste. Regression 2 The regression model used is as follows, R/W = ß0 + ß1inc + ß2age + ß3consumption + ß4ageorg + ß5gov + ß6log(access) + ß7log(green) + δiyeari + ∂jprovj The dependent variable in this model is the ratio of recycling and waste, the quality variable in the previous regression. The recycling component represents both residential and non-residential materials prepared for recycling. The waste component contains the total amount of public and private waste disposed. Recycling was divided by waste to create the ratio between the two components. In contrast to before, where we established a causal relationship between the quality of a recycling program and waste reduction, we are now interested in examining the effects of various factors linked to people’s recycling behaviour on the ratio of recycling to waste. Income is the same measure used in our original regression and represents income per capita measured in millions of dollars. We expect that as income rises, the ratio between recycling and waste will decrease. Median age is the same data used in the previous regression. This represents the central age of the province. We expect as age increases, this ratio will decrease. This is based on the belief that the younger generation is more environmentally conscious and has been exposed to the social norms surrounding recycling at a younger age. The older generation, on the other hand, will be slower to change previously learned behaviours and adopt new beliefs. Consumption represents consumer spending on goods. This data is collected from the expenditure based Gross Domestic Product, and includes all spending on durable, semi-durable, and non-durable goods. The data is collected through Statistics Canada. This is measured in millions of dollars and is chained to 2007 dollars. Data is missing in the years 1996 and 1998 for the Northwest Territories and Nunavut. Consumption is believed to have an ambiguous effect on the ratio of recycling and waste. Relating back to the previous model linking composition, waste, and recycling, if there is a high value placed on the environmentally friendly actions, or a high ß, an increase in consumption leads to an increase in recycling. This is due to the fact that more of what the consumer purchases and consumes is being recycled as opposed to thrown in the waste. Conversely, if the ß or the environmental values of the consumer is low, we expect an increase in consumption to lead to a decrease in the recycling-waste ratio. The consumer is purchasing and consuming more, however, due to the low value placed on the environment, these additional units of consumption are being thrown out. The effect of consumption is then linked with the environmental values of the population. Age of organization captures how long recycling and sustainability driven councils and organizations have been present in each province. This measure is developed by calculating the age of these organizations from when they were first implemented. All provinces with the exception of Nunavut, Newfoundland and the Yukon, have recycling and sustainability organizations. An increase in the age of the organization is expected to increase the recycling-waste ratio. An older organization is more established and experienced within its community, and likely more influential. Government represents whether the environmental organization is government based or nonprofit and non-political. This measure is represented by a dummy variable indicating whether the recycling organizations in each province are government run or non-profit. The dummy is 1 if the organization is government based and 0 if it is a non-profit, non-political organization. Organizations in Prince Edward Island, New Brunswick, Quebec, Manitoba, and Northwest Territories are all government organizations. On the other hand Nova Scotia, Ontario, Saskatchewan, Alberta, and British Columbia's organizations are non-profit or non-political. Newfoundland, Nunavut, and the Yukon do not have any of these organizations. This captures the role for government in implementing recycling organizations in the province. The expectation is that government organizations will increase the ratio due to the fact they likely are more prominent and have a wider access to funding and resources. Access represents the percentage of all households that had access to recycling programs. Data was available for 1994 and 2006. A linear approximation connecting the data points was used to estimate the remaining years. The log of this data is used in the regression. Data is missing for Nunavut, Northwest Territories, and the Yukon. The expectation is that an increase in access will lead to an increase in the ratio of recycling to waste. Increasing the access to recycling programs makes the act of recycling easier and more attractive to consumers. Increasing access to recycling programs leads to increased efficiency for both households and the institutions that collect and handle recycled materials. This should result in an overall decrease in waste and increase in recycling. The green measure is an instrument used to capture the norms and values towards recycling in the province. The instrument used is percentage of households that purchase environmentally friendly or “green” cleaning products. This data is split into five different categories; always, often, sometimes, rarely, and never purchase these products. The green measure was constructed by adding together the percentage of households that indicated they always or often purchased green cleaning products. The log of this data is then used in the regression. Data was not available for the Yukon, Nunavut, or the Northwest Territories. An increase in the values of society, shifting towards a more environmentally friendly position, will result in an increase in the recycling-waste ratio. If the population becomes more environmentally conscious, both their disposal and consumption behaviours change. As the population places an increased importance on being green, they will make more of an effort to recycle as opposed to disposing items in the garbage and landfills. Additionally, as the environmental values of the population change, we believe consumption behaviours will change as well. We believe the population will begin opting for products they know to be recyclable. Both these effects will lead to an increase in recycling and a decrease in waste. δiyeari represents the year dummy of the model. Each year from 1996 to 2010 is represented in the regression. For example, if looking at δ11996, δ is equal to 1 if the year is 1996 and 0 if otherwise. If looking at δ21998, δ is equal to 1 if the year is 1998 and 0 otherwise. This is done for every year in the model. Additionally, the province dummy, ∂iprovi , represents each province in the model. If we have ∂1BC, ∂ is equal to 1 given the province is BC and 0 for every other province. Similarly, ∂2Manitoba, is where ∂ is 1 if the province is Manitoba and 0 otherwise. There is a province dummy for each province represented in the model. Below we present a summary of our data to illustrate the range and variation between provinces. Our government variable has been excluded as it is a dummy variable. Table 2: Summary Statistics Variable Mean Max Min Median 1st Quintile 3rd Quintile Observations (minus NA) Income 0.04 0.12 0.02 0.04 0.03 0.05 4 Age 36.09 43.30 21.80 37.10 34.98 38.92 NA Consumption 0.006 0.014 0.007 0.010 0.009 0.012 4 Age Organization 22 40 6 23.5 11 27 24 Access 4.4 4.6 3.5 4.5 4.4 4.5 24 Green 3.3 3.6 2.6 3.3 3.1 3.5 24 We can take a closer look the variables, age of organization and green. There is large variation in the age of recycling organizations in provinces. British Columbia's recycling organization, the Recycling Council of BC, was established in 1974, making it the maximum value at 40 years. The minimum comes from New Brunswick, where the organization, Recycle NB, was established in 2008. Our green variable uses the percentage of households which responded that they always or often purchase environmentally friendly cleaning products to capture the environmental values of a province. The log of this percentage was then taken. British Columbia has the highest value with 37% and the lowest value comes from Newfoundland at 14%. Results When running the regression we obtain the following results: Table 3: Regression 2 Variable Parameter Estimate Standard Error Income 19.9 ** 6.6 Age 0.11 *** 0.02 Consumption -210.86 * 78.72 Age Organization 0.05 ** 0.01 Government 0.08 0.09 Access 0.68 ** 0.23 Green 3.35 ** 1.02 All variables are found to be significant with the exception of government; this indicates that it is important to have environmental and recycling based organizations in the province, however it is not important who runs these organizations. All variables match our predictions with the exception being age. This could be attributed to the fact that the older generation may have both the time and the resources to devote to environmentally conscious actions, whereas the younger generation may still be establishing themselves, and thus, may not view recycling as a high priority. While consumption appears to have quite a large effect, when examining the units in which it is measured, we see that it will take a one million dollar increase in consumption to induce a decrease in the recycling to waste ratio by 210.86. When taking the units into account, the effect of consumption is quite small. Although the result for income is much smaller, the same issue of units of measurement takes place. A one million dollar increase in income results in a 19.9 increase in the ratio between recycling and waste. The variable age of the organization indicates that older organizations are more effective; increasing the age of the organization by one year, and holding all else constant, results in a 0.05 increase in the ratio of recycling to waste. Long standing organizations show a commitment to sustainability and recycling initiatives. These results suggest established organizations whose primary goals focus on environmental issues can have a positive impact on recycling and waste decisions of the province. The coefficients on the two variables of interest, access and green, are both expected and significant. These results indicate that increasing the number of households with access to recycling programs by 1% will lead to an increase in the ratio of recycling to waste by 0.68, holding all else constant. Similarly, increasing the environmental valuation of households by 1% leads to an increase in the ratio of recycling and waste by 3.35, holding all else constant. These results indicate these two factors play a crucial role in influencing the individual’s recycling and waste behaviours. There were 104 observations in the data, with 40 deleted due to incomplete information. This was likely generated by provinces where the data was unavailable for all years, such as the Northwest Territories or the Yukon. The residual standard error on 45 degrees of freedom is 0.08554 and the 2 adjusted R is 0.7541. Coefficients are significant are the following levels: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1. By completing a sensitivity analysis on the two main variables of interest, we test the robustness of the variables and their sensitivity to change. The data for both variables has been differed by a factor of -10%, -5%, +5%, and +10%. Table 4: Sensitivity Analysis, Green Variable Parameter Coefficient Standard Error -10% 3.35 ** 1.02 -5% 3.35 ** 1.02 Green 3.35 ** 1.02 +5% 3.35 ** 1.02 +10% 3.35 ** 1.02 Table 5: Sensitivity Analysis, Access Variable Parameter Coefficient Standard Error -10% 0.68 ** 0.23 -5% 0.68 ** 0.23 Access 0.68 ** 0.23 +5% 0.67 ** 0.25 +10% 0.59 ** 0.26 Differing the data for both the access and green variables by various percentages has a very small effect on the results, and in the case of the green variable, zero effect. This shows the results obtained are significant and the variables are robust. Robustness of green variable The choice of green variable is an important factor in the above regression. This variable works to capture the underlying values of society, values which motivate recycling behaviours. The green variable used in our regression was constructed from a Statistics Canada Survey in which Canadian households were asked how often they purchased environmentally friendly or green cleaning products. The five possible responses included; always, often, sometimes, rarely, and never. Our green variable was constructed by adding the percentage of households that responded with always and often.Another potential variable that captures these social norms is looked at to determine the robustness of the green variable used. This variable uses the same Statistics Canada data on green cleaning products. The alternative instrument is constructed by summing the percentage of households who responded they often and sometimes purchase environmentally friendly or 'green' cleaning products. Regression1 lists the previous results found using our green variable whereas Regression2 indicates the parameter coefficients obtained from running the same regression with the alternative green measure. Table 6: Robustness of green instrument Variable Regression1 (Green1) Regression2 (Green2) Income 19.9 ** 19.9 ** Age 0.11 *** 0.11 *** Consumption -210.86 * -210.86 * Age Organization 0.05 ** 0.34 ** Government 0.08 5.5 ** Access 0.68 ** 0.67 ** Green1 3.35 ** Green2 60.07 ** In Regression2 we see that both government and age of environmental organization have larger impacts on the ratio of recycling and waste. If the population does not already have a high environmental valuation then recycling and government programs will be much more effective in influencing individual behaviour. Furthermore, we see the parameters on the green variables themselves are quite different. In the original regression, the parameter coefficient is 3.34 as opposed to 60.07 in the second regression. Again, if society already has strong established positive recycling and environmental norms, any shifts in preferences that increase the 'greenness' of the population will have less of an effect than a population who are more neutral. Our alternative variable produces fairly consistent results, indicating the green measure used in our regression to represent society’s preferences is both robust and appropriate. The alternative variable does produce much stronger results. This is likely because it covers a larger percentage of the population. There is a loss of variation in the data though; and as such, we are unable to properly isolate the true impacts and benefits of our variables. This variation will also be important when conducting further counter factual analysis. Counter Factual Analysis We can use the results generated in the model to illustrate the potential gains from improving and influencing the environmental values of the province. When looking at our green variable, which is an instrument that captures the environmental attitudes or 'greenness' of a province, we can compare a province with a high value, like B.C., to a province with a lower valuation, like New Brunswick. We discussed earlier why values may differ between two provinces; in this case B.C is not only known for being very environmentally conscious, it is also a larger province in terms of GDP and population. These factors allow B.C. to make the environment a much higher priority, leading to strong environmental values. The same is likely not true for New Brunswick, who will devote its efforts to building and expanding their economy, potentially leading to a lesser focus on environmental issues. When we look at our green variable, the percentage of households that indicated they purchased environmentally friendly cleaning products always or often, the percentage in B.C. is 37% compared with 22% in New Brunswick. Through the parameter coefficients found in our regression, we estimate the recycling to waste ratio for New Brunswick in 2010 given the available data. The steps are illustrated in the appendix. We find the estimated recycling to waste ratio for New Brunswick is 0.68. To find the benefit of increasing the 'greenness' of a province we first assume that our regression model is correct and also that all other factors do not change. We then estimate the recycling to waste ratio in the same way as before, but this time assume New Brunswick has as the same environmental valuation as B.C. In this case, we find the estimated recycling to waste ratio for New Brunswick is 2.43. By keeping all else constant and assigning New Brunswick the same level of environmental consciousness as a high valuation province, there is a 1.75 increase in the recycling-waste ratio. In such a case, New Brunswick is going from producing more waste than recycling, as the recycling to waste ratio is less than one, to producing more than two times the amount recycling to waste. We can see increasing the environmental valuation and 'greenness' of a province can lead to gains in terms of amount recycled and disposed. We can graph the relationship between the ratio of recycling to waste and our green variable. Ignoring any outliers we can see that as the 'greenness' of a province increases, the ratio of recycling to waste is increasing as well. This provides further evidence that there are benefits to targeting the environmental attitudes of a province in an effort to increase the ratio of recycling to waste. We can conduct the same experiment when looking into the age of environmental organizations in a province. In this case, we will compare B.C. with Quebec. Both provinces are fairly similar in terms of income per capita, and as such they will have access to relatively similar resources. B.C.'s major recycling organization is the Recycling Council of BC, which is 40 years old. In Quebec, the major organization is the crown corporation Recyc-Quebec, which is 24 years old. Again we start by estimating the recycling to waste ratio for Quebec in the year 2010 using the data available and the year of their recycling organization. The estimated ratio of recycling to waste is 0.2. Next, we calculate the same ratio by assuming the age of Quebec's recycling program is the same as B.C.’s and hold all other factors at their original level. We find that the estimated recycling to waste ratio is now 1; this is an increase in the recycling-waste ratio by 0.8. By appointing Quebec a higher organization age, they move from producing more waste than recycling, to an even amount of recycling and waste. This provides evidence that recycling and environmental organizations should be implemented earlier as opposed to later. A long standing recycling organization will be recognizable and credible to the communities, which contributes to increasing the recycling-waste ratio. We can construct a graph illustrating the relationship between the ratio of recycling to waste and the age of a recycling organization. If we again ignore the outliers we see a positive relationship between these two variables. This shows that as the age of a recycling organization increases, the ratio of recycling to waste will increase; therefore, it is beneficial to implement such organizations as early as possible. Impacts on organic waste A natural extension of this model is to examine not only the effect of these variables on the recycling to waste ratio, but also the compost to waste ratio. One of the major impacts associated with increasing levels of landfill waste is the increased production of methane. Methane is released during the breakdown of organic materials, adding to greenhouse gasses in the atmosphere. It is useful then to determine whether these variables have a similar effect when concentrating solely on composting. We run the previous regression using the same independent variables; however, the dependent variable is the ratio of compost to waste measured. Compost data was obtained from the same report used to collect recycling and waste data, Waste Management Industry Survey: Businesses and Government Sectors, and represents organic materials. Data is missing for Newfoundland, Prince Edward Island, Yukon, Northwest Territories, and Nunavut in all years. Data is also missing in 1996, 1998, 2000, and 2006 for New Brunswick, in 1998 and 2008 for Manitoba, and in 1998, 2002, 2004, and 2010 in Saskatchewan. Running the regression with this as the dependent variable gives the following results. Table 7: Regression 3 Variable Parameter Estimate Standard Error Income 7.92 * 3.63 Age 0.08 *** 0.02 Consumption -61.63 42.89 Age Organization 0.03 ** 0.01 Government -0.13 * 0.06 Access 0.34 * 0.14 Green 2.15 *** 0.58 All variables are significant with the exception of consumption. Again we see positive and expected signs on our two main variables of interest, access and green, indicating programs that influence these variables will result in an increase in the compost to waste ratio. An interesting result of this regression is the negative government coefficient. This could potentially indicate that, when it comes to composting, individuals are motivated by intrinsic factors opposed to external factors. As such, they may be more likely to engage in composting behaviours independent of government programs. There were 104 observations with 50 deleted due to incomplete information. The residual standard error is 0.04 on 35 degrees of freedom and the adjusted R2 is 0.81. Coefficients are significant are the following levels: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1. We can compare the results obtained by using the ratio of recycling to waste as the dependent variable in the regression with using the ratio of compost to waste. Table 8: Comparison of recycling and compost regressions Variable Regression -recycling to waste Regression -compost to waste Income 19.9 ** 7.92 * Age 0.11 *** 0.08 *** Consumption -210.86 * -61.63 Age Organization 0.05 ** 0.03 ** Government 0.08 -0.13 * Access 0.68 ** 0.34 * Green 3.35 ** 2.15 *** These results are clearly aligned with the results obtained in our original regression with the ratio between recycling and waste as the dependent variable. Given these results, we can incorporate similar strategies when addressing both the issues of recycling and composting. Existing recycling programs should be expanded to include a composting component. The city of Vancouver has adopted such a strategy by implementing green bin programs that collect and encourage composting and reduces the amount of organic materials which ends up in landfills (City of Vancouver, 2013). A high quality recycling program will then incorporate aspects that target not only traditional recyclable materials, such as cardboard and plastic, but also organic and compostable materials. Policy Implications These results indicate there are ways in which the ratio between recycling and waste can be influenced, as such, the amount of recycling can be increased and the amount of waste decreased. We see that the age of the organization can influence the ratio of recycled materials to waste. Older organizations have a bigger impact on this ratio as long standing organizations are rooted and established within the community. Accordingly, older organizations demonstrate a commitment by the province to continue to implement sustainable practices. If the province doesn’t already have an organization committed to environmental awareness, one should be established. . In addition, if a province has multiple, smaller committees, these should be consolidated into one organization. These environmental organizations can work to increase the ratio of recycling to waste by providing their communities with all the necessary resources and information needed to effectively recycle. In order to increase the ratio of recycling to waste and promote sustainable behaviours, the government should focus on targeting the quality of recycling programs and their accessibility and ease. This corresponds to the green and access variables from our regression. As shown by the regression results, a focus on these two aspects will lead to an increase in the recycling to waste ratio. The government should be looking to implement high quality recycling programs, a key aspect of which is an educational component. This will target and influence the community’s environmental values and social norms, the ß in our model. By including an educational component to a recycling program, the government could increase the awareness of the impacts of recycling and waste. Providing not only the tools to recycle but also the knowledge as to why it is important will increase the overall effectiveness of the program. Over time, this will shift the underlying norms and values of the province, moving towards a more environmentally conscious society. Above influencing the amount of materials recycled, a high quality recycling program based on education also has the potential to influence society’s consumption choices as well. When an educational component is introduced, individuals are made aware of the impacts of their consumption and disposal decisions. As such, we potentially will see a shift towards the purchase of environmentally friendly products, either in the form of items made from recycled, as opposed to raw materials, or items the consumer knows to be recyclable. The second component of a high quality recycling program works to increase the ease and accessibility of the program. By making the act of recycling as easy as possible for the individual, there will be an increase in the amount of people willing to participate. The study by Ferrara and Missios found that the more accessible recycling services are, the more likely individuals are to engage in environmentally conscious behaviours. They show that curbside recycling programs in which the recycled materials are picked up from households are more effective then drop off programs where the individual is responsible for delivering the recycled materials (Ferrara & Missios, 2012).This can be accomplished by increasing the current scope of recycling programs and accepting a larger variety of items. For example, Vancouver currently has initiatives to accept organic food waste, which also targets the compost to waste ratio. The city should also provide households with all the necessary materials required, such as bags and bins, and increase the pick-up times for recycled materials. Furthermore, a focus should be placed on making recycling accessible in public areas. Installing recycling bins alongside trash cans in public areas, like parks and on the streets, gives the individual options for disposing of items when outside their homes that previously were not present. These strategies can be mimicked at the business and institutional level as well. Providing households and businesses with the education and tools to efficiently recycle will increase the effectiveness of recycling programs and increase the recycling to waste ratio. When individuals not only understand the impacts of their disposal decisions but also have all the resources to properly and easily recycle, the overall effectiveness of recycling programs increases. Moreover, not only does the overall ratio of recycling to waste increase, but the environmental values and social norms of the province begin to shift as well. By focusing on these two aspects, quality and accessibility, the government has the tools to implement very successful recycling programs in their provinces. Limitations & Extensions Our model and regression illustrate the link between high quality, accessible recycling programs and increases in the ratio of recycled materials to waste. However, as with any model, there are limitations encountered. One major barrier faced in running this regression was the source and quality of data. All data used in the regressions was obtained from Statistics Canada surveys. While this does provide a consistent source of data, there is the potential for reporting bias stemming from the actual reporting agency. Using multiple and diverse sources of data would potentially limit this type of bias, however at this time such data was unavailable. Furthermore, the number of observations was very limited, producing a small data set. On top of an initially modest data set, we also encountered years and provinces with missing observations. This not only impacts the reliability of the results, it also restricts our ability to conduct additional analysis or include supplementary variables of interest. For example, the interaction term illustrating the impact of green values, or ß, of a province on consumption was excluded from the model since a lack of data produced unreliable results. The results found in our regressions are statistically significant; however, these issues, rooted in the data, are potential causes of error and limitation. The issue of recycling and environment is becoming increasingly important and is receiving a higher level of attention. Therefore, it is an area that can benefit from more in depth research. Our model provides interesting results and strategies for increasing the ratio of recycling to waste within each province. This model can also be used not only to examine Canada on a provincial level, but can also be extended to look at specific municipalities within each province. Each province is very diverse with its own unique identify, and the same is true for the municipalities within these provinces. For example, if we were to look at the province of B.C we see that metro areas like the Lower Mainland and Vancouver are very different than Northern B.C. These regions differ not only in aspects like population and geography, but also in their values and social norms. Strategies to reduce waste in metro Vancouver may not be successful in a northern town like Prince George. Using the model to determine the impacts of our variables on distinct municipalities can help develop specific, specialized recycling programs to be implemented in different cities versus a general province wide program. It is also useful to look into the costs of implementing such programs. Extensions of this model should incorporate the government budget constraints and determine the level of funding needed to implement the high quality programs proposed. Furthermore, this would provide an interesting comparison when contrasting provinces with large budgets and resources to those with less funding available. Finally, when approaching the issue of recycling and environmental sustainability, it is important to note that this is not a concern limited to Canada. The impacts of recycling and waste are experienced on a global level, as illustrated through the effects of global climate change. Therefore, a relevant extension would be to apply this model to various countries. Just as we see variation from municipality to municipality and province to province, there are also large differences between countries. The diversity in population, geography, political atmosphere, and social norms will likely lead to different results from our model. In particular, if we use the model to compare developed and developing countries, we will see different variables having larger impacts. For example, income per capita had a very minor effect in our regression. If we assume that a developed country, like Canada, operates using a Kuznet's Curve type model in terms of their environmental views and behaviours, we would expect income to have a low impact. Under a Kuznet's Curve assumption, as a country develops, their focus and priority is on industrialization and not on sustainability. This translates into increased waste and environmental degradation. As the country develops, and a higher average income is reached, we expect to see a shift towards more sustainable behaviours and a decrease in waste produced. Thus, the expectation is income per capita will have a much larger impact on the recycling to waste ratio of a developing country than a developed one. Expanding our model to look at different countries can help to formulate strategies to increase environmental awareness and program effectiveness in developing countries. As Canada and the rest of the world’s population continues to grow and expand, both production and consumption will increase. This translates into an increase in the waste by-product produced. While landfill waste is associated with smaller impacts like increased odour, noise, and loss of aesthetics, there are also much more serious impacts. Landfill waste can potentially lead to climate change through increased methane levels resulting from the breakdown of organic materials in landfills. Recycling provides not only the opportunity to reduce and mitigate the negative impacts associated with increased waste levels, but also has its own inherently positive impacts. Using recycled materials in the production process conserves natural resources and energy required for production. Furthermore, expansion of the recycling sector creates additional jobs and revenue for the municipality through the sale of recycled materials. A multitude of factors influence the provinces’ disposal decisions, and can play an important role in increasing the ratio of recycling to waste. Our model identifies key variables that are crucial when designing recycling programs with the overall goal of increasing recycling and reducing waste. Both access to recycling programs and the green value and norms of society are linked with increases in the recycling to waste ratio. Governments can utilize these findings when designing and implementing recycling programs. High quality recycling programs should improve the accessibility and ease of recycling by increasing the scope of programs and variety of items accepted. Furthermore, a high quality program will have an educational aspect which makes consumers aware of the impacts of their disposal decisions. This educational component influences not only disposal behaviour but also potentially consumption behaviour. By implementing high quality recycling programs, governments will be able to successfully and effectively limit the amount of waste and negative effects produced while providing benefits to society and the environment. References: Abhijit Biswas, Jane W. Licata, Daryl McKee, Chris Pullig and Christopher Daughtridge. Journal of Public Policy & Marketing , Vol. 19, No. 1, Privacy and Ethical Issues in Database/Interactive Marketing and Public Policy (Spring, 2000), pp. 93-105 Abrahm , L. (2011, Jan 25). Climate benefits of natural gas may be overstated. Retrieved from http://www.propublica.org/article/natural-gas-and-coal-pollution-gap-in-doubt Alberta Energy. (n.d.). Oil sands. Retrieved from http://www.energy.alberta.ca/OurBusiness/oilsands.asp Best, J., & Hoberg , G. (2008, Jun 18). Alberta’s oil sands: Key issues and impacts. Retrieved from http://mapleleafweb.com/features/alberta-s-oil-sands-key-issues-and-impacts Biswas, A., Licata, J. W., McKee, D., Pullig, C., & Daughtridge, C. (2000). The recycling cycle: An empirical examination of consumer waste recycling and recycling shopping behaviors. Journal of Public Policy & Marketing, 19(1), 93-105. Retrieved from http://www.jstor.org.proxy.lib.sfu.ca/stable/30000490 City of Vancouver. (2013, Nov 21). Zero Waste. Retrieved from http://vancouver.ca/greenvancouver/zero-waste.aspx Clean Nova Scotia. (n.d.). About us. Retrieved from http://clean.ns.ca/about-us/ David Suzuki Foundation. (n.d.). Water impacts. Retrieved from http://davidsuzuki.org/issues/climatechange/science/impacts/water-impacts/ Environment Victoria. (n.d.). The problem with landfill. Retrieved from http://environmentvictoria.org.au/content/problem-landfill. Ferrara I, Missios P. Land Economics, Vol. 88, No. 4, A Cross-Country Study of Household Waste Prevention and Recycling: Assessing the Effectiveness of Policy Instruments. (November 2012), pp. 710-744. Green Manitoba. (n.d.). Background. Retrieved from http://greenmanitoba.ca/home/ Howard, A. (2009, 8 6).Landfills and the environmental effects. Retrieved from http://livelifegreen.com/landfills- and-th-environmental-effects/ I Care NWT (n.d.). About us. Retrieved from http://icarenwt.ca/about-us Island Waste Management Corporation. (2014). Retrieved from http://www.iwmc.pe.ca/ Leigh, E. (n.d.). Effects of recycling on humans. Retrieved from http://homeguides.sfgate.com/effects- recycling-humans-79735.html Levinson , A. (n.d.). Environmental kuznets curve. Retrieved from http://www9.georgetown.edu/faculty/aml6/pdfs&zips/PalgraveEKC.pdf NRDC. (2011, 11 8). An introduction to climate change. Retrieved from http://www.nrdc.org/globalwarming/climatebasics.asp R-bloggers. (2012, April 5). Melt. Retrieved from http://www.r-bloggers.com/melt/ Recyc-Quebec. (n.d.). Frequently asked questions. Retrieved from http://www.recycquebec.gouv.qc.ca/client/fr/rubriques/questions.asp?idCat=15 Recycle NB. (n.d.). About us. Retrieved from http://www.recyclenb.com/en/about_us/ Recycling Council of Alberta. (n.d.). About the RCA. Retrieved from https://www.recycle.ab.ca/ Recycling Council of Ontario. (n.d.). History. Retrieved from https://www.rco.on.ca/history Recycling Council of British Columbia. (n.d.). What is RCBC and what do we do?. Retrieved from http://www.rcbc.ca/ Saskatchewan Waste Reduction Council. (2011). SWRC - the first 10 years. Retrieved from http://www.saskwastereduction.ca/about/ United States Environmental Protection Agency. (1998, Jan).Puzzled about recycling’s value? look beyond the bin. Retrieved from http://www.epa.gov/osw/conserve/downloads/benefits.pdf Data References Statistics Canada. (1999, Oct ). Waste management industry survey: Business and government sectors 1996. Retrieved from http://www.statcan.gc.ca/pub/16f0023x/16f0023x1996001- eng.pdf Statistics Canada. (2000, Oct 12). Waste management industry survey: Business and government sectors 1998. Retrieved from http://www.statcan.gc.ca/pub/16f0023x/16f0023x1998001- eng.pdf Statistics Canada. (2003, Mar 14). Waste management industry survey: Business and government sectors 2000. Retrieved from http://www.statcan.gc.ca/pub/16f0023x/16f0023x2000001- eng.pdf Statistics Canada. (2004, Sep 24). Waste management industry survey: Business and government sectors 2002. Retrieved from http://www.statcan.gc.ca/pub/16f0023x/16f0023x2002001-eng.pdf Statistics Canada. (2007, Feb 5). Waste management industry survey: Business and government Sectors 2004. Retrieved from http://www.statcan.gc.ca/pub/16f0023x/16f0023x200400 eng.pdf Statistics Canada. (2008, Jun 23). Waste management industry survey: Business and government sectors 2006. Retrieved from http://www.statcan.gc.ca/pub/16f0023x/16f0023x200600 eng.pdf Statistics Canada. (2010, Dec 22). Waste management industry survey: Business and government sectors 2008. Retrieved from http://www.statcan.gc.ca/pub/16f0023x/16f0023x2010001- eng.pdf Statistics Canada. (2012, Dec 19). Households that had access to and used recycling programs in Canada, by province, 1994 and 2007. Retrieved from http://www.statcan.gc.ca/pub/16-001m/2010013/t001-eng.htm Statistics Canada. (2012, Sept). Households and the environment. Retrieved from http://www.statcan.gc.ca/pub/11-526-x/11-526-x2011001-eng.pdf Statistics Canada. (2013, Aug 21). Waste management industry survey: Business and government sectors 2010. Retrieved from http://www.statcan.gc.ca/pub/16f0023x/16f0023x2013001- eng.pdf Statistics Canada. (2013, Nov 8). Gross domestic product, expenditure-based, by province and t erritory. Retrieved from http://www.statcan.gc.ca/tables-tableaux/sum- som/l01/cst01/econ15eng.htm Statistics Canada. (2013, Nov 25). Population by year, by province and territory (number). Retrieved from http://www.statcan.gc.ca/tables-tableaux/sum-som/l01/cst01/demo02a eng.htm Statistics Canada. (2013, Nov 25). Estimates of population, by age group and sex for july 1, Canada, provinces and territories. Retrieved from http://www5.statcan.gc.ca/cansim/a26? lang=eng&retrLang=eng&id=0510001&paSer=&pattern=&stByVal=1&p1=1&p2=1&tabMode=dataTable&csid= Statistics Canada. (2014, Feb 28). Gross domestic product, expenditure-based. Table 380-0064 . Retrieved from http://www5.statcan.gc.ca/cansim/a26 Appendix Regressions Regressions with intercept, year, and province factors included. Regression 1 Variable Parameter Estimate Standard Error Intercept 0.4 0.56 Income 6.25 ** 0.56 Age 0.01 2.14 Government Spending 0.00007 0.00046 Quality -0.31 *** 0.07 British Columbia -0.2 ** 0.06 Manitoba -0.09 0.06 New Brunswick -0.25 ** 0.08 Nova Scotia -0.34 *** 0.07 Ontario -0.18 ** 0.05 Quebec -0.05 0.07 Saskatchewan -0.1 * 0.04 1998 0.006 0.03 2000 -0.06 0.04 2002 -0.05 0.05 2004 -0.06 0.07 2006 -0.07 0.08 2008 -0.1 0.1 2010 -0.1 0.1 Regression 2 Variable Parameter Estimate Standard Error Intercept -18.42 *** 4.05 Income 19.9 ** 6.6 Age 0.11 *** 0.02 Consumption -210.86 * 78.72 Age Organization 0.05 ** 0.01 Government 0.08 0.09 Access 0.68 ** 0.23 Green 3.35 ** 1.02 British Columbia -1.26** 0.42 Manitoba 1.9 ** 0.57 New Brunswick 2.14 ** 0.67 Nova Scotia 1.11 *** 0.28 Ontario NA NA Quebec NA NA Saskatchewan NA NA 1998 0.072 0.06 2000 -0.04 0.08 2002 -0.05 0.11 2004 -0.16 0.14 2006 -0.14 0.17 2008 -0.18 0.2 2010 -0.22 0.22 Regression 3: Compost to waste ratio Variable Parameter Estimate Standard Error Intercept -11.98 *** 2.35 Income 7.92 * 3.63 Age 0.08 *** 0.02 Consumption -61.63 42.89 Age Organization 0.03 ** 0.01 Government -0.13 * 0.06 Access 0.34 * 0.14 Green 2.15 *** 0.58 British Columbia -0.92 *** 0.24 Manitoba 1.33 *** 0.33 New Brunswick 1.49 *** 0.38 Nova Scotia 0.61 *** 0.16 Ontario NA NA Quebec NA NA Saskatchewan NA NA 1998 -0.07 * 0.03 2000 -0.12 ** 0.04 2002 -0.17 ** 0.06 2004 -0.22 ** 0.07 2006 -0.26 ** 0.09 2008 -0.28 * 0.11 2010 -0.31 ** 0.11 Regression 4: Robustness Testing, Green alternative Variable Parameter Estimate Standard Error Intercept -255.6 ** 75.85 Income 19.9 ** 6.6 Age 0.11 *** 0.03 Consumption -210.86 * 78.72 Age Organization 0.34 ** 0.11 Government 5.5 ** 1.63 Access 0.67 ** 0.23 Green 60.07 ** 18.24 British Columbia -2.41 ** 0.77 Manitoba 5.5 ** 1.67 New Brunswick 3.9 ** 1.2 Nova Scotia -0.64 * 0.29 Ontario NA NA Quebec NA NA Saskatchewan NA NA 1998 0.07 0.06 2000 -0.04 0.08 2002 -0.05 0.11 2004 -0.16 0.14 2006 -0.14 0.17 2008 -0.18 0.2 2010 -0.21 0.21 Counter Factual Analysis Calculation 1. Green variable Model with estimated parameter coefficients. R/What = -18.42 + 19.9(inc) + 0.11(age) -210.87(consumption) + 0.08(gov) +0.05(ageorg) +0.68(access) +3.35(green) -0.21(year) + 2.14(prov) i. New Brunswick: Original data GreenNew Brunswick = 22% Log(22) = 3.09 R/What = -18.42 + 19.9(0.04) + 0.11(42.7) -210.87(0.01) + 0.08(1) +0.05(6) +0.68(4.48) + 3.35(3.09) -0.21 + 2.14 = -18.42+ 0.8 + 4.7 -2.11 + 0.08 + 0.3 + 3.05 + 10.35 – 0.21 + 2.14 = 0.68 ii. New Brunswick: New green data Replace New Brunswick's green data with British Columbia's green data GreenBritish Columba = 37% Log(37) = 3.61 R/What = -18.42 + 19.9(0.04) + 0.11(42.7) -210.87(0.01) + 0.08(1) +0.05(6) +0.68(4.48) + 3.35(3.61) -0.21(year) + 2.14(prov) = -18.42 + 0.8 + 4.7 -2.11 + 0.08 + 0.3 + 3.05 + 12.1 – 0.21 + 2.14 = 2.43 2. Age of Organization Model with estimated parameter coefficients R/What = -18.42 + 19.9(inc) + 0.11(age) -210.87(consumption) + 0.08(gov) +0.05(ageorg) +0.68(accesslong) +3.35(greenlog) -0.21(year) i. Quebec: Original data Age of OrganizationQuebec = 24 R/What = -18.42 + 19.9(0.04) + 0.11(41.2) -210.87(0.012) + 0.08(1) +0.05(24) +0.68(4.61) +3.35(3.47) -0.21 = -18.42 + 0.8 + 4.53 – 2.53 + 0.08 + 1.2 + 3.13 + 11.62 – 0.21 = 0.2 ii. Quebec: New age of organization Replace age of Quebec's recycling organization with age of British Columbia's organization. Age of OrganizationBritish Columbia = 40 R/What = -18.42 + 19.9(0.04) + 0.11(41.2) - 210.87(0.012) + 0.08(1) + 0.05(40) + 0.68(4.61) + 3.35(3.47) - 0.21(year) = -18.42 + 0.8 + 4.53 – 2.53 + 0.08 + 2 + 3.13 + 11.62 – 0.21 =1