https://doi.org/10.5061/dryad.bzkh189h4

In this repository, we provide all the data and necessary information for replication of our paper titled "Of the first five US states with food waste bans, Massachusetts alone has reduced landfilled waste". We include all the raw data and software we used to produce all tables and figures in this paper. Additionally, for easy replication, we include some outputs generated by our code, such as power analysis results. These are available in the "Data from Code" section.

Raw data

1. Waste data: includes all the data we use for our analysis (power2_impexp.csv)
   • Note: In this file, "disposal" refers to state-generated MSW for disposal. That is, we have (i) used only the MSW fraction of the total waste and (ii) excluded the imported waste and included waste exports. (For more details, please see sections A.1.2 and A.1.3 of the SM). When county_name is NA, it implies that the relevant state reports at the state level.
   • There are 3 additional datasets for waste disposal: one that has Boulder's disposal (boulder_waste.csv), and two files that have information for Seattle's disposal (seattle_disposal.csv) and composting (seattle_composting.csv)
2. Waste characterizations data: (wcs_2.csv)
3. Food waste generators lists: (food_generators_MA.csv & food_generators_VT.csv). The names and addresses of these businesses have been redacted. 
4. Food waste processors lists: (composting_infrastructure_all_states_gov.csv, composting_capacity_all_states.csv, food_processors_list_MA.csv, food_processors_list_VT.csv). When a field is NA, it implies that the information is not available.
5. Other data:
   • Population of US counties: needed to create the per capita waste disposal (population.csv, boulder_population.csv, population_seattle.csv)
   • US Cities: coordinates of US cities---needed to estimate the minimum distance between generators and processors (uscities.csv, town_coordinates_VT.csv)

Code

Primary analysis

1. placebo_all.RMD: calculation of the placebo confidence intervals ("power") (right panel of Fig. 2). This script additionally includes:
   • the statistics of the placebo distributions for state-level bans (Table S6) (see section "State") --> The results of the placebo tests are saved in files power_state_plac_1.csv through power_state_plac_10.csv
   • the calculation of the placebo confidence intervals for city-level specifications (right panel of Fig. S6) (see section "Placebo Confidence Intervals-County") --> The results of the placebo tests are saved in files plac_for_histogram4.csv  & plac_for_histogram6.csv
   • the calculation of the placebo confidence intervals for the anticipatory robustness check (right panel of Fig. S8) (see section "Passage") --> The results of the placebo tests are saved in the file: power_state_pass_data.csv
2. xy.R: main results for the point estimate of the ATT of the bans + the building of the SC for the treated states (left panel of Fig. 2). This script also includes:
   • the calculation of the point estimate of the ATT + the SC for city-level bans (left panel of Fig. S6)
   • a section that saves all necessary txt files
   • A Monte Carlo analysis for the city-level bans. The equivalent for state-level bans is found in script mcmc.R. (In this script, we repeat the calculation of the ATT and the associated p-values many times to ensure robustness to the randomness of the algorithm. The associated files with the results of these experiments are saved in mc1--10.csv, plseed refers to the input placebo seed from the creation of the placebo outcomes, and mc_for_cities.csv).

Data section

1. data_section: producing all the data-related figures and tables (Fig. 1, Fig. S2, Fig. S3, Table S4)

2. WCS: calculations regarding the expected effect of the bans (Table 1, Fig. S1)

   This script also includes an analysis of the WCSs of treated and non-treated states in our sample (Fig. S4)

Mechanism

infrastructure_gov.R: includes analyses related to the final section of the paper, the potential drivers of MA's ban's success (Fig. 3, Fig. 4, Fig. S8).

Robustness checks

1. year_placebo: produces results for California-specific provision (Fig. S7)
2. composting: all analyses regarding bans' effects on composting (Fig. S9)
3. ghg: all analyses regarding emissions (Fig. S5, Table S8). For this analysis, we find that the results are not robust across different specifications that handle mid-year ban-implementation (i)Baseline spec: that re-centers the time-series based on the month that the ban was implemented. This analysis in the ghg.R script is called spec1 and produces null effects for all states; (ii) Alternative specification 2: we move the ban to the beginning of the year, regardless of the month of implementation. This analysis in the ghg.R script is called spec2 and produces null effects for all states, apart from MA which produces a significant -25.7% decrease in emissions.
4. xy_passage: all analyses regarding bans's anticipation effects (left panel of Fig. S8)
5. diff_in_diff: all analyses for the difference in differences specification (Table S7)
6. placebo_all_alt2.R & xy_alt2.R: we use an alternative specification to handle the fact that the bans of VT, MA, and CA are implemented in July, Oct, and Apr, respectively. We shift it to the beginning of the calendar year, regardless of the ban-month implementation (placebo_all_alt2.R to re-estimate the placebo confidence intervals, and xy_alt2.R to estimate ATTs).
7. plac_sf.RMD: the calculation of the placebo confidence intervals for San Francisco (see section "Municipal-San Francisco") --> The results of the placebo tests are saved in the file: plac_sf.csv

Intermediate data

These files include intermediate data that have been produced at midpoints in the code. They are needed to replicate parts of the code without running every script (e.g., run xy.R without running placebo_all.RMD)

1. Expected effects
   • State-level expected effects for each phase: (disposal_effect_size2.csv)
   • (Robustness check) City-level expected effects (municipal_effect.csv)
   • (Robustness check) Effects of the bans on composting (composting_effect.csv)
2. Time series of actual and synthetic disposal
   • Summarized results of the ATT estimates for state-level and city-level bans (bt_with_power_data.csv)
   • (Robustness check) Detailed results of the time series when using methane emissions as the main outcome (sc_data_ghg_spec1.csv, sc_data_ghg_spec2.csv)
   • (Robustness check) Detailed results of the California provision placebo distribution (year_placebo.csv)
   • (Robustness check) Summarized results of the estimates of the ATT when using composting as the main outcome for state-level (tr_res_state_multiple_composting.csv) & for the city-level bans (tr_res_municipal_multiple_composting.csv)
3. Placebo distribution results
   • Detailed results of the state-level placebo distribution (power_state_plac_1--10.csv) for 10 different input seeds
   • (Robustness check) Detailed results of the county-level placebo distribution (plac_for_histogram4.csv, plac_for_histogram6.csv
   • (Robustness check) Summarized results of the county-level placebo distribution (power_county.csv)
   • (Robustness check) Detailed results of the SF placebo distribution (plac_sf.csv)
   • (Robustness check) Summarized results of the SF placebo distribution (sf_power.csv)
   • (Robustness check) Detailed results of the state-level placebo distribution when assuming the passage date as the implementation date (power_state_pass_data.csv)
   • (Robustness check) Summarized results of the state-level placebo distribution when assuming the passage date as the implementation date (power_state_p.csv)
   • i. (Robustness check) Detailed results of the state-level placebo distribution when using methane emissions as the main outcome (power_gas_res_spec1--2.csv & sc_data_ghg_spec1--2.csv)
   • (Robustness check) Detailed results of the California provision placebo distribution (year_placebo.csv)
   • (Robustness check) Summarized results of the state-level placebo distribution when using composting as our outcome (power_state_composting.csv)
   • (Robustness check) Detailed results of the county-level placebo distribution when using composting as our outcome (plac_for_histogram4_composting.csv, plac_for_histogram6_composting.csv)
   • (Robustness check) Summarized results of the county-level placebo distribution when using composting as our outcome (power_county_composting.csv)
   • (Robustness check) ATTs and p-values for multiple different seeds for the five treated states (mc1--10.csv) (plseed is the input placebo seed from a)
   • (Robustness check) ATTs and placebos for a different specification that handles mid-year ban-implementation (power_state_plac_alt2.csv).

Definition of variables

Raw data

1. power2_impexp.csv:
   - year: the calendar year for which waste is reported
   - state_id: the state for which waste is reported
   - county_name: the county for which waste is reported (NA if the state reports waste at the state level)
   - type: the disposition method of waste (can be disposal, composting, food_donation, etc.)
   - tons: the amount of waste in tons
2. wcs_2.csv:
   - Activity: the disposition method of waste (only Disposal)
   - state_id: the state for which waste is reported
   - City: the jurisdiction where the WCS was conducted (if statewide, then it applies to the whole state)
   - year: the year when the WCS was conducted
   - generator_category: the category of a generator generating the reporting waste
   - material: the category of the material (see SM for additional details)
   - tons: the amount of waste reported (NA if the WCS is in %)
   - percent: the percent of waste reported (some states only report the fraction in their WCSs) (NA if the WCS is in tons)
   - need_to_find: binary variable. =1 if the state does not report the breakdown between commercial and non-commercial waste (NA otherwise)
3. food_generators_MA.csv
   - DEP_Code: the unique code assigned to the generator by MassDEP
   - Name: the name of the generator (has been redacted)
   - Street_Add: the address of the generator (has been redacted)
   - Town_City: the town the generator is located in
   - Type: the type of generator (F: Food and Beverage Manufacturers/Processors, W: Wholesale Distributors, IH: Institutions-Healthcare Facilities, IS: Institutions -Independent Schools, IC: Institutions -Colleges/Universities, IP: Institutions -Correctional Facilities, C: Resorts and Conference Facilities, G: Supermarkets and Grocery Stores, R: Restaurants)
   - ZIP_Code: the zip code of the generator
   - tons: the estimated annual generation in tons
   - Lat: the latitude of the generator
   - Long: the longitude of the generator
   - X_Coord: The Massachusetts State Plane coordinates (not used)
   - Y_Coord: The Massachusetts State Plane coordinates (not used)
   - Status: The Status field denotes whether the potential food waste generator was definitively matched (M) or the highest ranking tie of one or more locations (T).
   - Score: the longitude of the composting facility
4. food_generators_VT.csv
   - X: The Vermont State Plane coordinates (not used)
   - Y: The Vermont State Plane coordinates (not used)
   - ID: the unique ID of the generator
   - FSG_Name: the name of the generator (redacted)
   - TonsPerWeek: the estimated generation of food waste in tons per week
   - TYPE1: the classification of the generator by the Vermont state agency
   - TYPE2: the classification of the generator by the Vermont state agency
   - TYPE3: the classification of the generator by the Vermont state agency
   - Address: the address of the generator (redacted)
   - Town: the town the generator is located in
   - County: the county the generator is located in
   - ZIP: the ZIP code of the generator
5. composting_infrastructure_all_states_gov.csv
   - state_name: the name of the state in which the composting facility is located
   - composting_facility: the name of the relevant composting facility (has been redacted)
   - lat: the latitude of the composting facility
   - long: the longitude of the composting facility
   - date_type: the source of the data (only gov meaning some government agency)
   - capacity: the annual capacity of the facility in tons
6. composting_capacity_all_states.csv
   - state_id: the state for which the capacity is reported
   - capacity: the capacity of the state's composting network
   - year: the year for which we report the capacity