'The Impact of External Modernization and Internal Modernization for Small-Scale Retailers in Emerging Markets: Evidence from a Field Experiment'

1. “External" Modernization intervention: Firm is partnered with a ‘Modernization Fellow’ (i.e. a top university student majoring in business, economics or related disciplines) who helps them implement external modernization structures in their business over 6 weeks. The Fellow helps the firm through 11 site visits, of 2.5 hours each, that are exclusively dedicated to hands-on implementation of external modernization structures. Fellows are trained and supervised by senior managers who also make 3 site visits (each 2.5 hours).

2. "Internal" Modernization intervention: Exactly the same format and intensity as the first intervention, except the Fellow implements internal modernization structures at the treated business.

2018-01-01

2018-12-01

Modernization levels - external and internal; Sales - revenues and indicators

Store branding; Product-planning

The study consists of three parts: (1) business selection, (2) baseline, randomization and intervention, and (3) collection of followup data. Part one began in fall 2017. Across five residential zones of Mexico City, university students managed by our team approached retail business owners with a marketing pitch to apply for a ‘modernization’ program. Interested retailers signed up via a recruitment survey, which we used to screen candidates. Screened firms were invited to complete a baseline survey. About 1150 firms qualified and were randomized into three equally-sized groups:

• Group 1: offered to get external modernization structures changed.
• Group 2: offered to get internal modernization structures changed.
• Group 3 (control): formed a counterfactual group that did not receive any intervention.

Randomization happened at the firm level immediately prior to the scheduled intervention delivery.

Randomization was stratified by location, baseline profits and sub-sector and happened at the individual firm level immediately prior to the scheduled intervention delivery. During the 6-week intervention period, data was collected on all firms in the treatment group to assess the intervention strength (and comparability across locations). To make the logistics of our RCT as smooth as possible, we ran the three parts of the study (entrepreneur recruitment; baseline, randomization and intervention delivery; follow-up data collection) over four
waves with a manageable sample size of approximately N = 300 in each wave.

We have wrapped up the two marketing interventions for the fourth and final wave of firms in our sample. To measure effects, we have launched our first post-intervention data collection round and expect to complete it by May 2019. This data is being collected 3 months post-intervention for the firms in our sample, wave by wave. In this survey, we measure sales outcomes such as monthly revenues, quantity sold, customer count, basket size per order. Moreover, we measure modernization levels by getting the enumerator to privately score every firm in our sample, upon detailed observation and evidence-gathering, on the 20 customer-facing modernization structures and 20 product-facing using a scale of 0 (no implementation) to 5 (excellent implementation). We also measure the mechanisms of store-level branding (by polling customers on how they rate the store on various branding dimensions) and product-planning (by using independent auditors to rate retailers decisions on own product knowledge and procurement efficiency).

We will conduct another follow-up survey with all firms in the study sample roughly 12 months after the program started, which is summer to fall 2019.

Randomization done by authors on a computer, using statistical programming software (Stata)

Firm

No

N/A

1154 firms

1. N = 385 in External Modernization treatment group
2. N = 385 in Internal Modernization treatment group
3. N = 384 in Control group