NEW UPDATE: Completed trials may now upload and register supplementary documents (e.g. null results reports, populated pre-analysis plans, or post-trial results reports) in the Post Trial section under Reports, Papers, & Other Materials.
Interfirm Relationships and Business Performance
Initial registration date
January 09, 2017
January 09, 2017 11:12 AM EST
University of Michigan
Other Primary Investigator(s)
Central European University and CEPR
Additional Trial Information
We organized business associations for the owner-managers of randomly selected young Chinese firms to study the effect of business networks on firm performance. We randomized 2,800 firms into small groups whose managers held monthly meetings for one year, and into a "nomeetings" control group. We find that: (1) The meetings increased firm revenue by 8.1 percent, and also significantly increased profit, factors, inputs, the number of partners, borrowing, and a management score; (2) These effects persisted one year after the conclusion of the meetings; and (3) Firms randomized to have better peers exhibited higher growth. We exploit additional interventions to document concrete channels. (4) Managers shared exogenous business-relevant information, particularly when they were not competitors, showing that the meetings facilitated learning from peers. (5) Managers created more business partnerships in the regular than in other one-time meetings, showing that the meetings improved supplier-client matching. (6) Firms whose managers discussed management, partners, or finance improved more in the associated domain, suggesting that the content of conversations shaped the nature of gains.
The main intervention consisted of organizing "business associations," or firm meetings. The managers in each meeting group were expected to meet once a month, every month, for one year. Researchers organized the first meetings, and offered the managers in each group print material containing business-relevant information. They also gave the same material to control firms. To provide incentives for participation, researchers offered managers who answered the surveys and attended at least 10 out of the 12 monthly meetings a certificate.
There were also three additional interventions. First, to help measure peer effects, researchers created variation in the composition of groups by size and sector. They created four kinds of groups: small firms in the same sector; large firms in the same sector; mixed size firms in the same sector; and mixed size and mixed sector. Researcheres randomized treated firms into these groups in each region. Second, to measure information diffusion, researchers gave information about two financial products to randomly chosen managers. The first product was a funding opportunity for the firm, the second a savings opportunity for the manager. They distributed information about each product via phone calls and text messages to 0%, 50% or 80% of the managers in each meeting group. Researchers also distributed the information to 40% of control firms to ensure that the same share of treatment and control firms have the information. Finally, to learn about the role of meeting frequency, researchers organized one time cross-group meetings. Each cross-group met once, and in the midline survey we asked managers to play hypothetical trust games (with large payoffs) with a randomly selected regular group member as well as with a randomly selected cross-group member.
Intervention Start Date
Intervention End Date
Primary Outcomes (end points)
firm characteristics (profits, sales, etc.), management, employee well-being, firm networks, firm performance, product innovation, employee satisfaction, peer effects
Primary Outcomes (explanation)
Secondary Outcomes (end points)
Secondary Outcomes (explanation)
Researchers organized "business associations" for firms in the main treatment group where meetings were held between employees and managers. There were three additional randomized interventions within the main treatment arm, and a portion of the treatment managers also played a hypothetical trust game. The interventions were designed to test the effects of business networks on firm performance.
Experimental Design Details
Stratified randomization by firm size and industry
Was the treatment clustered?
Sample size: planned number of clusters
Sample size: planned number of observations
Sample size (or number of clusters) by treatment arms
1,480 firms with owner-manager meetings
1,320 control group firms
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
INSTITUTIONAL REVIEW BOARDS (IRBs)
Post Trial Information
Is the intervention completed?
Intervention Completion Date
August 31, 2014, 12:00 AM +00:00
Is data collection complete?
Data Collection Completion Date
August 31, 2015, 12:00 AM +00:00
Final Sample Size: Number of Clusters (Unit of Randomization)
Was attrition correlated with treatment status?
Final Sample Size: Total Number of Observations
Final Sample Size (or Number of Clusters) by Treatment Arms
Reports, Papers & Other Materials
Interfirm Relationships and Business Performance - Working Paper, December 2016.
Cai, Jing, and Adam Szeidl. "Interfirm Relationships and Business Performance." Working Paper, December 2016.
REPORTS & OTHER MATERIALS