The Effects of SME Digital Training and Loan Access on Business Practices and Growth

Last registered on October 07, 2021


Trial Information

General Information

The Effects of SME Digital Training and Loan Access on Business Practices and Growth
Initial registration date
October 04, 2021

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
October 07, 2021, 2:18 PM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.


Primary Investigator

UC San Diego

Other Primary Investigator(s)

PI Affiliation
LMU Munich

Additional Trial Information

On going
Start date
End date
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Studies of the effectiveness of business training for small and medium sized enterprises (SMEs) in lower income countries have shown mixed results, and it has been challenging to reconcile many of these studies because of how much training and opportunities to use the training in practice differs across research settings. In this study we are providing an additional analysis of the effectiveness of training among SMEs, but the training we are analyzing is one that has been delivered to thousands of SMEs across the developing world and is thus, highly standardized and already being heavily invested in. Moreover, we are testing whether the business training is more or less effective when completion is immediately followed by access to a business loan. The study uses data generated from an RCT run by Technoserve Kenya in early 2021 to evaluate their training program among SMEs through Mombasa and Nairobi. Of the just over 14,000 SMEs that applied to participate in the program, a randomly selected one-third was assigned to receive Technoserve’s standard phone-based training program immediately and receive a loan in the size they applied for immediately afterwards, one-third was assigned to receive the training immediately but not receive the loan they applied for until two months after training, and the final one-third was assigned to receive the training followed by the loan they applied for four months following the training of the first two groups. A random sub-set of each group was surveyed and audited for performance before the initial training began, and three and a half months afterwards. Differences in assets, business practices, and performance across groups will be analyzed, and how these differences vary by business type, intensity of COVID-19 cases, and change in knowledge following training will also be analyzed.
External Link(s)

Registration Citation

Ashraf, Anik and Elizabeth Lyons. 2021. "The Effects of SME Digital Training and Loan Access on Business Practices and Growth." AEA RCT Registry. October 07.
Experimental Details


Intervention Start Date
Intervention End Date

Primary Outcomes

Primary Outcomes (end points)
1. Business owner knowledge and practices: Changes in business practices, loan repayment, and change in knowledge to assess whether anything was learned in the training.

2. SME performance: Average daily sales and number of customers, asset counts and types, and growth in employees and inventory.
Primary Outcomes (explanation)
The business practices we are measuring are:

1. Adoption of business budget
2. Maintaining cash savings
3. Having an investment plan and goal
4. Having updated and consistent business records
5. Use of online marketing and sales platforms
6. Use of digital payments
7. Regular engagement with distributor and/or manufacturer
8. Use of formal contracts with suppliers
9. Innovativeness as assessed by enumerator

These measures capture items included in the training modules. However, to avoid testing too many outcomes, we will develop composite measures of practices with items 1-4 measuring financial practices, 5 and 6 measuring digital practices, 7 and 8 measuring supply chain management, and 9 measuring innovativeness.

Measures of what was learned from training will be based on knowledge tests.

Secondary Outcomes

Secondary Outcomes (end points)
We will test whether the effectiveness of training and the extent to which immediate access to a loan changes it varies depending on the size of the business (primarily by comparing dukas to non-dukas), the modules of trainings that the owner/managers completed, the extent of new knowledge the business owners gain through the training, and the type of phone the business owners receive training through (smart phone versus basic phone). In addition, we will also test heterogeneity across owner/managers’ education, and work experience.
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
Over 14,000 business owners applied for a small business loan in response to Technoserve’s call for applications by filling in an application form that asked questions about the business and business owner background. Following the application window, 4,793 applicants were assigned to receive the training followed by the loan immediately (group 1), 4,252 were assigned to receive the training immediately and access to the loan 2 months following training (group 2), and 4,967 were assigned to receive the training followed by loan 4 months following group 1 and 2’s training (group 3).

Following randomization, just under 5,000 businesses dropped out of the program largely because of the training requirement. Of the remaining businesses, 3,272 are in group 1, 3,180 are in group 2, and 3,595 are in group 3.

Our design allows us to test the short-run effects of training, and how they vary by immediate access to financial capital in the form of loans. In particular, comparing changes in performance and business practices between groups 1 and 2 and group 3 before group 3 receives the training will provide an estimate of the effectiveness of business training. Moreover, comparing the performance and practice differences between those in groups 1 and 3 relative to changes between groups 2 and 3 will provide information about whether providing an input that is complementary to business training while the training is still fresh on business owners’ minds enhances the effectiveness of training.

Importantly, Technoserve did not want to deny training or loans to any businesses interested in taking them up, and had a limited window in which they needed to disburse the loans. Thus, the lag between group trainings and loan disbursements only allow us to measure short-run differences in performance and practices. However, whether any short-run effects lead to sustained benefits is also something we will test.
In order to be eligible for the loan, all participants had to complete a financial training module and pass a knowledge test. Participants were given access to other training modules on inventory management, supply chain management, marketing, and alternative revenue streams. Participants were assigned loans in the amounts they applied for.

As the training and knowledge tests were completed via phones (either through WhatsApp or SMS depending on what type of phone business owners possess), we are able to observe which modules they spent time on, and which knowledge tests they completed and how they scored.

A randomly selected subset of SMEs in each group was selected to participate in the baseline survey, and subsequently in the midline survey at the end of June, 2021. This survey and audit assessed business performance, loan status and use, and business expectations. An endline survey will be conducted by the end of year, 2021.
Experimental Design Details
Randomization Method
Technoserve used a clustered randomization approach to assign participant SMEs into experimental arms. First, a cluster was defined as the unique combination of sub-county and geographical neighborhood as identified in the business address of the participant SMEs. Subsequently, 419 clusters were identified from the total pool of 14012 SMEs who showed primary interest in the training. Next, these clusters were randomly assigned into the three experimental arms. Each arm eventually contains about a third of the clusters. Finally, all participant SMEs with businesses in a given cluster were assigned the same treatment status. The randomization into the three groups was performed using Stata's (gen uniform()) command with the seed set at 92187.
Randomization Unit
Small and medium enterprises (SME) subcounty-locations
Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters
419 subcounty location clusters were randomized across groups; SMEs in 361 of these clusters participated in the program.
Sample size: planned number of observations
We have program application data for a total of 14012 SMEs; Technoserve randomized clusters of SME subcounty-locations into treatment groups from this pool. Eventually, 10,047 SMEs in 361 clusters from this pool participated in the loan and training program, and around 550 were surveyed to collect baseline and midline characteristics and outcomes.
Sample size (or number of clusters) by treatment arms
Average number of SMEs per randomized cluster is around 34 in Group 1, 31 in Group 2, and 35 in Group 3. Clusters per treatment group are 140 in group 1, 139 in group 2, 140 in group 3.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)

Institutional Review Boards (IRBs)

IRB Name
IRB Approval Date
IRB Approval Number


Post Trial Information

Study Withdrawal

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information


Is the intervention completed?
Data Collection Complete
Data Publication

Data Publication

Is public data available?

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials