The Impact of Computer-Generated Credit Scoring on Lending in Colombia
Last registered on February 08, 2017

Pre-Trial

Trial Information
General Information
Title
The Impact of Computer-Generated Credit Scoring on Lending in Colombia
RCT ID
AEARCTR-0001978
Initial registration date
February 08, 2017
Last updated
February 08, 2017 7:48 PM EST
Location(s)
Region
Primary Investigator
Affiliation
MIT
Other Primary Investigator(s)
PI Affiliation
London School of Economics
Additional Trial Information
Status
Completed
Start date
2010-10-01
End date
2011-06-30
Secondary IDs
Abstract
We distinguish the impact of information technology adoption on information processing costs and agency costs by conducting a randomized control trial with a bank that adopts a new credit-scoring tool. The availability of scores significantly increases credit committees' effort and output on difficult-to- evaluate loan applications. Output increases almost as much in a treatment where the committee receives no new information, but anticipates the score becoming available after it evaluates a application, which suggests that scores reduce incentive problems inside the credit committee. We also show that scores improve efficiency by decentralizing decision-making and equalizing marginal returns across loans.
External Link(s)
Registration Citation
Citation
Paravisini, Daniel and Antoinette Schoar. 2017. "The Impact of Computer-Generated Credit Scoring on Lending in Colombia." AEA RCT Registry. February 08. https://www.socialscienceregistry.org/trials/1978/history/13889
Experimental Details
Interventions
Intervention(s)
Intervention Start Date
2010-11-01
Intervention End Date
2011-05-31
Primary Outcomes
Primary Outcomes (end points)
Effects of computer-generated credit scores on small enterprise loans; Impact of information technology adoption on decision-making
Primary Outcomes (explanation)
Secondary Outcomes
Secondary Outcomes (end points)
Secondary Outcomes (explanation)
Experimental Design
Experimental Design
We implement a four-month pilot program with an RCT design in eight urban branches. In each pilot branch we randomly select the treatment applications for which the committee will be able to see the score of the applicant. Randomization occurs in real time when the committee begins to discuss an application. The committee members are informed of the group assignment at the beginning of the discussion.

In the control group, the committee evaluates the application without observing the score. In the first treatment group (T1), the committee receives the score before evaluating the application. This first treatment allows us to measure the overall effect of scores on committee effort, output and productivity. In the second treatment group (T2), the committee evaluates the application and chooses an interim action before receiving the score, receives the score after taking the interim action, and then may revise its choice to take a final action.
Experimental Design Details
Randomization Method
Randomization by computer
Randomization Unit
Individuals
Was the treatment clustered?
No
Experiment Characteristics
Sample size: planned number of clusters
1,421 individuals
Sample size: planned number of observations
1,421 individuals
Sample size (or number of clusters) by treatment arms
335 individuals in the control group
563 individuals in treatment group 1 (T1)
523 individuals in treatment group 2 (T2)
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB
INSTITUTIONAL REVIEW BOARDS (IRBs)
IRB Name
Study has received IRB approval. Details not available.
IRB Approval Date
Details not available
IRB Approval Number
Details not available
Post-Trial
Post Trial Information
Study Withdrawal
Intervention
Is the intervention completed?
Yes
Intervention Completion Date
May 31, 2011, 12:00 AM +00:00
Is data collection complete?
Yes
Data Collection Completion Date
June 30, 2011, 12:00 AM +00:00
Final Sample Size: Number of Clusters (Unit of Randomization)
1,421 individuals
Was attrition correlated with treatment status?
No
Final Sample Size: Total Number of Observations
1,421 individuals
Final Sample Size (or Number of Clusters) by Treatment Arms
335 individuals in the control group, 563 individuals in T1, 523 individuals in T2
Data Publication
Data Publication
Is public data available?
No
Program Files
Program Files
Reports and Papers
Preliminary Reports
Relevant Papers
Abstract
Citation
Paravisini, Daniel, and Antoinette Schoar. "The Incentive Effect of IT: Randomized Evidence from Credit Committees." NBER Working Paper No. 19303, August 2013.