The effects of business loans and technology support on the growth of U.S. online entrepreneurs

Last registered on January 09, 2023
View Trial History

Pre-Trial

Trial Information

General Information

Title
The effects of business loans and technology support on the growth of U.S. online entrepreneurs
RCT ID
AEARCTR-0010695
Initial registration date
January 05, 2023

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
January 09, 2023, 5:32 PM EST

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

Locations

Country
United States of America
Region
All of US

Primary Investigator

Name
Mihai Alexandru Codreanu
Affiliation
Stanford University
Contact Primary Investigator

Other Primary Investigator(s)

PI Name
Nick Bloom
PI Affiliation
Stanford University
Contact Investigator

Additional Trial Information

Status
In development
Start date
2023-01-30
End date
2025-12-31
Keywords
Finance & Microfinance, Firms & Productivity
Additional Keywords
JEL code(s)
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
This study is focused on the relationship between borrowing constraints, access to cutting-edge technology and information about cutting-edge technology on the performance of U.S. online businesses. With the help of two large U.S. technology companies we will be able to randomize access to loans and free cloud computing credits (as well as information about the potential use of technology) to otherwise identical (generally small, but fast growing) firms, to see if they will have a causal impact on firm development. We will check the performance of these companies on a wide variety of metrics (sales, growth, input choices, etc.), heterogeneity in the treatment effects, as well as possible channels through which the effect may take place using both administrative data provided by our partners and data collected independently by the researchers (webscraping, surveying, etc.). We will survey the firms in our sample at baseline (before the program is implemented), short-term (around 6 months after), and longer term (around 18 months after) to check the persistency of the effects. We posit that at least the capital loans intervention will have a sizable short and long-term effect on firm performance, while the cloud computing experiments will induce firms to use more cloud computing for their day-to-day activities.
External Link(s)

Registration Citation

Citation
Bloom, Nick and Mihai Alexandru Codreanu. 2023. "The effects of business loans and technology support on the growth of U.S. online entrepreneurs." AEA RCT Registry. January 09. https://doi.org/10.1257/rct.10695-1.0
Sponsors & Partners

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

Request Information
Experimental Details

Interventions

Intervention(s)
There are three types of interventions that we will run: 1) randomly hold-out a number of firms from receiving a business loan offer (at advantageous conditions) from one of our partners, 2) randomly offer free cloud computing credits to a number of firms (both to firms presented with the loan offer, and not presented) from one of our partners, 3) randomly offer a simple information treatment about the benefits of cloud computing technology (both to firms presented with the loan offer, and not presented)
Intervention Start Date
2023-01-30
Intervention End Date
2025-02-03

Primary Outcomes

Primary Outcomes (end points)
Understand whether and by how much access to capital and cloud computing technology (and information about it) benefits small U.S. firms and enables them to accelerate their revenue growth. We are mainly interested in observing merchants’ longer term (e.g. around 15-18 months) payment volumes since the experiment. However, we are also interested in shorter-term measures (e.g. around 6 months) since the experiment
Primary Outcomes (explanation)
The main outcome of interest will be the monthly sales volume as provided by our partner. We will run robustness checks on this figure using self-reported total monthly sales volumes across all platforms, as well as other webscraped data and third party data on firm performance.

Secondary Outcomes

Secondary Outcomes (end points)
All secondary outcomes will also measures both short-term (e.g. around 6 months), and longer-term (e.g. around 15-18 months) after the experiment.
Secondary Outcomes (explanation)
We will also track measures of firm productivity, growth rates, input choices, survival (having transaction in a 6 month period), as well as cloud computing/credit usage and expectations/confidence about the firms' future prospects.

Experimental Design

Experimental Design
The first intervention (capital loans) will be given to a larger set of firms. Out of those firms, some will be cross-randomized into the cloud computing experiments providing credits and information. We will run this experiment design on a sample of firms at the beginning of each month for 9 months in 2023.

For the first capital loans intervention, we randomly assign around 75% of the firms to treatment (i.e. receiving the loan offer) and 25% of the firms to control (not receiving the loan offer). While taking our baseline survey (that happens before the loan offers are presented to the customers), the treatment and control firms will be presented with a 50% probability with free credits from one of our partners, and with 50% with an advert of the usefulness of cloud computing and services that our partner provides. Then, we will extend the capital loan offers to those 75% firms initially selected into the capital treatment, and track their performance both using our partners' administrative data, third party data, online data we will collect, as well as and survey the firms after around 6 months and 15-18 months since the experiment. Answering surveys will be rewarded with some additional financial incentives.
Experimental Design Details
Not available
Randomization Method
Randomization will be done based on last digits/letters of the firm account number, that are as good as random
Randomization Unit
Firm
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
Around 63,000 firms for the capital loan treatment, around 9,000 firms for the cloud computing treatments.
Sample size: planned number of observations
Around 63,000 firms for the capital loan treatment, around 9,000 firms for the cloud computing treatments.
Sample size (or number of clusters) by treatment arms
Around 40,500 firms receive the capital loan offer and no cloud computing offer (+no initial survey).
Around 12,500 firms don't receive the capital loan offer and no cloud computing offer (+no initial survey).

Around 1,700 firms receive the capital loan offer and the cloud computing credits and the cloud computing information pack (+receive initial survey).
Around 1,700 firms receive the capital loan offer and the cloud computing credits but not the cloud computing information pack (+receive initial survey).
Around 1,700 firms receive the capital loan offer and the cloud computing information pack but not the the cloud computing credits (+receive initial survey).
Around 1,700 firms receive the capital loan offer but not the cloud computing credits or the cloud computing information pack (+receive initial survey).

Around 550 firms don't receive the capital loan offer but receive the cloud computing credits and the cloud computing information pack (+receive initial survey).
Around 550 firms don't receive the capital loan offer or the cloud computing information pack but receive the cloud computing credits (+receive initial survey).
Around 550 firms don't receive the capital loan offer or the cloud computing credits but receive the cloud computing information pack (+receive initial survey).
Around 550 firms don't receive the capital loan offer, the cloud computing credits or the cloud computing information pack (+receive initial survey).
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
Supporting Documents and Materials

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

Request Information
IRB

Institutional Review Boards (IRBs)

Analysis Plan