Lawyers for Consumer Protection: Evidence from a Field Experiment in Uganda

The study will consist of two treatment arms and one control group. The first treatment will consist in offering the services of a law student within CTDR-U to resolve the mobile money issue. The lawyers and law students at CTDR-U will implement an innovative legal strategy, based on several steps, starting with sending a formal letter, and then escalating the matter to the regulator (the Uganda Communications Commission, UCC). The second treatment arm will consist in creating a web page where the different letter formats are posted, as well as the contact information of the complaints department for the different institutions. With this treatment, clients would be empowered to complain on their own.

2022-04-08

2024-02-03

The primary outcomes are the proportion of disputes resolved by CTDR-U, trust in mobile money operators and use of mobile money (including mobile banking, mobile loans). We will collect a household survey for each participant containing questions on the resolution of disputes according to the respondent.

We will use an ancova analysis, controlling for the baseline value of the dependent variable. If the baseline value is missing, it will be replaced by the value zero and a dummy variable “Baseline_Missing” will be defined as 1, 0 otherwise. This dummy will be included in the model.

The research plan includes subgroup analyses by:

• Socio-economic background of respondent (lower income respondents may have less access to formal legal institutions and beenefit more from the intervention)

• Method of recruiting (as explained below, we will explore three different ways of recruiting respondents; the existing literature has different predictions if the lawyer went to the client or the client came to the lawyer)

• Type of dispute (funds sent but not delivered correctly, or utility bills not paid, or fraudulent agent; some disputes may be easier to resolve than others)

• Financial amount of the dispute (smaller disputes may see a quicker resolution on the part of the mobile money operators)

The secondary outcomes will be the outcomes identified in the existing literature on mobile money as the potential benefits of this new technology:

• Risk sharing (meaning an increase in consumption (or at least no decrease) if the household experiences a negative shock, which can mean: death, job loss, health shock, loss of valuable, agricultural losses)

• Greater access to (digital) finance: saving, credit

• Therefore more investment

• Less poverty, more expenditures

Our survey includes sections on all these aspects. Of course, we will look at these outcomes only if the primary outcomes are significant, i.e., if the treatments work at resolving disputes.

Multiple hypothesis testing will be done with the Sharpened False Discovery Rate (FDR) adjusted q-values (Anderson, 2008).

The unit of randomization is the individual. There will be three groups of equal size:

• Treatment 1: the legal team within CTDR-U handles the case

• Treatment 2: the respondent will get access to the relevant forms and documentation and will pursue the legal strategy on its own

• A control group of individuals continuing business-as-usual

The target population consists of those who use mobile money technology to complete transactions, such as paying utility or rent bills. We will restrict our sample to Kampala, the capital. In order to construct a representative sample of mobile money users, we will follow three different strategies. First, we will contact the Village Health Team (VHT) and Chairman Local Council (LC1) leaders. VHTs are volunteer community health workers who deliver predominantly health education and preventive services in communities. They have a good knowledge of the communities they serve. They will act as a referral for households in their communities which have encountered a mobile money issue. Chairman Local Council (LC1) also need to be informed of any research going on in their village, they will act as a source of referral after getting their consent. We will survey these households. The fieldworkers will determine eligibility of the cases in the field, using a list of eligible cases determined by CTDR-U. We will ask participants whether they wish to receive legal aid services.

We will start with treatment 1 only in a first stage, because we need to debvelop the material to be delivered in treatment 2 based on the protocol we develop and refine in treatment 1.

We will do paired matching randomization (with Mahalanobis matching). We will match on the baseline outcomes of interest (proportion of disputes resolved, trust in mobile money operators and use of mobile money), geographic indicators (village), and basic demographics (gender, age, income). This paired matching randomization will allow us to address the potential partial compliance that may occur in the treatment group. People might drop out out of the intervention if they deem their disputes too small financially, or when they have to travel to CTDR-U's offices to sign a consent form for legal services. If this drop out happens, we will drop the entire pair to maintain balance between the treatment and control groups.

Second, we will ask the regulator called UCC (Uganda Communications Commission) to refer customers who raise these types of issues to the CTRD-U. UCC has a built-in mechanism to register complaints from customers.

Third, we will deploy a promotional campaign on social media and SMS platform (with geo-restricted posts), with sponsored ads.

In the first method of recruiting, CTDR-U will go towards the client, whereas in the second and third method, the client will come towards CTDR-U. There might be a different success rate for different methods of recruiting. We will first do a pilot to test out the best method of recruiting. If we end up using the three methods, we will stratify by the method of recruiting to make sure there is good balance withing each method.

Not available

The randomization is done in office by a computer.

The unit of randomization is the individual.

No

The planned number of villages is 33.

The planned number of observations is 1,000 individuals.

333 individuals will receive treatment 1, 333 treatment 2 and 333 will be in the control group.