Trading Trash on Tricycles

This trial studies how responsive informal waste collectors are to price subsidies at waste transfer stations in their choice of where to dispose of their collected waste. Treated collectors receive daily payments at disposal sites via mobile money. Collectors daily choices are monitored via daily forms, site visits, and a self-developed smartphone application.

This field experiment builds on ongoing work and several rounds of data collection with households and collectors in Accra, which have allowed us to estimate a structural model of waste collection and disposal in the Greater Accra Metropolitan Area. Counterfactual results on second-best environmental policies suggest that halving transfer stations fees can correct environmental externalities unpriced by the market up to the planner's choice. In this field experiment, we seek to tests the conjectured effectiveness of pricing regulation, either validating our current estimates obtained with data for the current equilibrium, or helping us understand if there may be resistance to environmental policy that creates big shifts from the status quo. We will examine how collectors respond to large relative price changes at transfer stations and whether they face any new barriers to do so.

2025-10-12

2025-10-26

Disposal site choice (categorical), Disposed of at formal transfer station (0/1)

Via daily forms and registry's at disposal sites, we will track the choice of disposal site for each collector in our sample. This is a categorical variable which covers all disposal sites in the Greater Accra Metropolitan Area, based on previous field assessments. Based on this variable, we will construct a dummy variable that indicates if a collector disposed of the waste at one of the operational transfer stations in the city.

Number of customers, revenue, collection price per transaction, disposal fee, reasons for choosing site, recycling revenue, collection locality change.

Via daily forms we will gather data on number of customers, revenue in GHS, disposal fee in GHS. In a smartphone application we developed for the study, collectors can register their collection transactions. Collectors will be incentivised to report their data daily and accurately. This provides us real-time transaction data on collection prices and their location. We will also construct a complementary measure on daily revenues and number of customers too based on the transactions data. Finally, to assess whether dumpsite operators resist losing market share and exert any soft pressure on collectors who may be willing to change their choice of disposal site, we add a question on the reasons behind choosing a disposal site daily. We add "Leaders' advice" and "Dumpsite operators' advice" as potential reasons, amongst fees, waiting time, commuting distance, and others. This variable will allow us to capture if a shift out of the existing equilibrium leads to new forces that may challenge the effectiveness of environmental regulation and are not currently present in the data.

Treatment assignment (timing of price subsidies): The experiment uses a randomised design to assign the timing of payments at disposal sites for the 400 participating waste collectors. 200 collectors randomly assigned to the treatment 1 group receive subsidy payments at disposal sites during the first week of the intervention. The remaining 200 collectors are the control group during the first week. At the end of the week, the collectors assigned to the control group during the first week start receiving the payment for a period of 1 additional week, while the treatment 1 collectors stop receiving the payment during this second week. We track outcomes daily for both groups via different tools: a daily form administered by enumerators, registries upon payments at disposal sites, and a self-developed smartphone application where collectors register their waste collection transactions.

Intervention details: In both weeks, the treatment consists of payments at disposal sites of different quantities. At dumpsites, collectors may receive 10 or 15 GHS daily. At formal waste transfer stations, collectors are eligible to claim a subsidy payment of 100 GHS daily. The intervention thus induce exogenous variation in the relative prices of disposal sites, and more specifically, it makes formal solid waste disposal substantially cheaper in the city. The payment will be conducted via mobile money daily.

We will estimate ITT effects on formal disposal at transfer stations (primary outcome) and on additional outcomes potentially affected by this choice (i.e. number of customers, revenues, collection location, collection prices). We will do this aggregating at treatment period (week 1 and week 2 are the 2 periods), but also by plotting raw shares over intervention days. During baseline we will gather information on disposal site choice during the last 7 days and report daily event studies (using the first day of receiving payments as reference period) for formal disposal as the outcome variable.

We will estimate ITT effects on whether any soft power from leaders or dumpsite operators is a reason for the choice of disposal site. Finally, we will test for heterogeneity of the ITT effect on disposal at transfer station by pre-existing distance between collection area and transfer stations.

We will use assignment to the treatment as an instrument for disposal fees and estimate TOT effects to understand the passthrough of lower disposal fees on number of customers and collection prices.

We will use the assignment to treatment as an instrument for route profits and estimate the taste parameters governing collectors’ route choices in a collection-disposal site discrete choice framework via GMM. This will complement the current structural approach we use to estimate our full model of waste collection and disposal.

Details and rationale for the experimental design: In our experimental design, the treatment assignment determines the timing of payments, and all collectors in the study eventually receive a payment. This is done with one main objective. It incentivises participation in our daily data collection efforts for both treatment and control collectors, reducing the risk of differential attrition across groups. Waste collectors in Accra work in the same disposal sites and have open channels of informal communication and some are members of an association. Having a pure control group would have posed an important risk of attrition or non-participation, as control collectors would have learned about the intervention and potentially refused to participate in the daily data collection efforts. The intervention would have also presented challenges for the collectors' association, whose collaboration is fundamental to ensure both collectors and enumerators can operate without interference.

The rationale for adding a small subsidy to dumpsites is the same. It is done to ensure that the relevant stakeholders allow for the project and data collection efforts to take place. Additionally, it allows us to estimate in a straightforward way the impact of lower disposal fees on collector outcomes beyond site choices. As treated collectors can claim fees subsidies at all sites, independently on their site choice, we can implement an instrumental variables approach to estimate the consequences of lower disposal fees.

Randomisation done in office by a computer after baseline registry is conducted.

Individual (waste collector)

No

400 collectors

400 collectors (daily data collection for 14-15 days - 5600 collector-day observations)

200 collectors control, 200 collectors treatment (disposal fee subsidy) for 1 week. At the end of the week the treatment switches and collectors in initial control group start getting the price subsidy, while treatment collectors stop receiving it. The treatment assignment defines the timing at which collectors receive the payment but all 400 collectors participating in the study receive the price subsidy -- either in week 1 or week 2.