Data from a randomized controlled trial that evaluated the impact of individual-liability microloans on borrowers in the Philippines. Loans were randomly assigned to applicants depending on credit score. The dataset contains 1113 observations and 51 variables, prepared following the processing in Athey, Fava, Karlan, Osman & Zinman (2025).
Format
A data frame with 1113 rows and 51 variables:
- agg_id
Unique identifier for each individual.
- treat
Treatment indicator (1 = received microloan offer, 0 = control).
- prop_score
Propensity score (probability of treatment assignment).
- fu_survey_year
Year of the follow-up survey (2007 or 2008).
- hhinc_yrly_end
Household income at endline (yearly, Philippine pesos).
- profits_yrly_end
Business profits at endline (yearly, Philippine pesos).
- rev_yrly_end
Business revenue at endline (yearly, Philippine pesos).
- exp_yrly_end
Business expenses at endline (yearly, Philippine pesos).
- hhexp_yrly_end
Household expenses at endline (yearly, Philippine pesos).
- loan_size
Size of the microloan received (Philippine pesos; 0 if no loan taken).
- prinpaid
Principal amount paid back.
- interest_rate
Annual interest rate on the loan (percent).
- repaytime
Repayment time indicator.
- intpaid
Interest paid on the loan.
- penalty
Penalty fees paid.
- paid_total
Total amount paid (principal + interest + penalties).
- net_revenue
Net revenue earned by the lender from the loan.
- css_creditscorefinal
Credit score at application.
- css_yearsataddress
Years at current address at application.
- own_house
Indicator for home ownership (1 = yes).
- sari
Indicator for ownership of a sari-sari store (small convenience store; 1 = yes).
- own_anybus
Indicator for owning any business (1 = yes).
- max_yearsinbusiness
Maximum years in business among owned businesses.
- css_regularworkers
Number of regular workers at application.
- css_traveltime
Travel time to lender branch (minutes).
- css_travelcost
Travel cost to lender branch (Philippine pesos).
- css_stockvalue
Stock value of the business at application (Philippine pesos).
- css_assetvalue
Asset value of the business at application (Philippine pesos).
- css_noofbusiness
Number of businesses at application.
- lower_window
Indicator for being in the lower window of the credit score distribution (1 = yes).
- app_year
Year of the loan application.
- hhsize
Household size.
- age
Age of the applicant (years).
- gender
Gender of the applicant (1 = female, 0 = male).
- education
Education level (categorical).
- married
Marital status (1 = married, 0 = not married).
- savingsamt
Savings amount (Philippine pesos).
- hhinc_yrly_base
Household income at baseline (yearly, Philippine pesos).
- profits_yrly_base
Business profits at baseline (yearly, Philippine pesos).
- rev_yrly_base
Business revenue at baseline (yearly, Philippine pesos).
- exp_yrly_base
Business expenses at baseline (yearly, Philippine pesos).
- hhexp_yrly_base
Household expenses at baseline (yearly, Philippine pesos).
- hhasset_yrly_base
Household assets at baseline (yearly, Philippine pesos).
- revenue_fees
Revenue from fees earned by the lender.
- revenue_interest
Revenue from interest earned by the lender.
- log_hhinc
Log of household income at endline.
- default_value
Default value of the loan (amount unpaid).
- high_school
Indicator for high school completion (1 = yes).
- college
Indicator for college completion (1 = yes).
- bank_profits_level
Bank profits in levels (Philippine pesos).
- bank_profits_pp
Bank profits per peso lent.
Source
Karlan, D. and Zinman, J. (2011). Microcredit in Theory and Practice: Using Randomized Credit Scoring for Impact Evaluation. Science, 332(6035), 1278–1284. doi:10.1126/science.1200138
Athey, S., Fava, B., Karlan, D., Osman, A. and Zinman, J. (2025). Profits and Social Impacts: Complements vs. Tradeoffs for Lenders in Three Countries. Working paper.
Details
The original experiment randomly assigned individual-liability microloans to applicants at a microlender in the Philippines using credit scoring as the randomization device. Applicants near the credit-score cutoff were randomly assigned to treatment (loan approved) or control (loan denied). The data include baseline covariates (demographics, business characteristics, credit score components), treatment assignment, loan repayment outcomes, and follow-up survey outcomes measured 11–22 months after treatment.
Examples
data(microcredit)
head(microcredit)
#> agg_id treat prop_score fu_survey_year hhinc_yrly_end profits_yrly_end
#> 1 100519 1 0.8434874 2007 15750 15750
#> 2 101793 1 0.5465839 2008 66990 66990
#> 3 102790 1 0.8434874 2007 9450 9450
#> 4 103223 1 0.8434874 2007 29025 29025
#> 5 104470 1 0.8434874 2007 3750 3750
#> 6 105906 1 0.8434874 2007 0 0
#> rev_yrly_end exp_yrly_end hhexp_yrly_end loan_size prinpaid interest_rate
#> 1 21750 29475 3000.0 0 NA NA
#> 2 78750 11760 4500.0 0 NA NA
#> 3 11250 1800 11250.0 625 10000 30
#> 4 30000 975 10500.0 0 NA NA
#> 5 7500 9000 5812.5 0 NA NA
#> 6 0 0 11250.0 625 10000 30
#> repaytime intpaid penalty paid_total net_revenue css_creditscorefinal
#> 1 1 NA NA NA 0.00000 48
#> 2 1 NA NA NA 0.00000 42
#> 3 1 758.29 0 10758.29 47.39313 60
#> 4 1 NA NA NA 0.00000 52
#> 5 1 NA NA NA 0.00000 54
#> 6 1 758.29 0 10758.29 47.39313 56
#> css_yearsataddress own_house sari own_anybus max_yearsinbusiness
#> 1 44 1 1 1 5
#> 2 4 1 0 1 4
#> 3 15 1 1 1 15
#> 4 39 1 1 1 10
#> 5 39 1 1 1 4
#> 6 15 1 0 1 4
#> css_regularworkers css_traveltime css_travelcost css_stockvalue
#> 1 0 10 30 25000
#> 2 0 5 12 150000
#> 3 0 20 24 17000
#> 4 0 10 0 13800
#> 5 0 10 24 15000
#> 6 4 20 0 5000
#> css_assetvalue css_noofbusiness lower_window app_year hhsize age gender
#> 1 140800 3 0 2006 1 44 1
#> 2 0 1 1 2006 1 46 0
#> 3 24000 2 0 2006 5 36 1
#> 4 11000 2 0 2006 4 39 0
#> 5 25000 1 0 2006 3 39 1
#> 6 15000 3 0 2006 6 38 1
#> education married savingsamt hhinc_yrly_base profits_yrly_base rev_yrly_base
#> 1 1 0 0 21443.750 128662.50 224018.8
#> 2 2 1 0 57031.250 342187.50 342187.5
#> 3 2 1 0 10950.000 65700.00 123187.5
#> 4 1 1 0 21907.604 131445.62 250549.7
#> 5 1 1 0 5703.125 34218.75 48362.5
#> 6 1 1 0 23572.917 141437.50 234284.4
#> exp_yrly_base hhexp_yrly_base hhasset_yrly_base revenue_fees revenue_interest
#> 1 95356.25 3390.00 1562.5 0.0000000 0.00000
#> 2 0.00 2325.00 1000.0 0.0000000 0.00000
#> 3 57487.50 2137.50 14012.5 0.6465497 46.74658
#> 4 119104.06 1050.00 2125.0 0.0000000 0.00000
#> 5 14143.75 2756.25 2875.0 0.0000000 0.00000
#> 6 92846.88 8421.00 7875.0 0.6465497 46.74658
#> log_hhinc default_value high_school college bank_profits_level
#> 1 9.664659 0 1 1 0.00000
#> 2 11.112314 0 1 0 0.00000
#> 3 9.153876 0 1 0 -16.93826
#> 4 10.275947 0 1 1 0.00000
#> 5 8.229778 0 1 1 0.00000
#> 6 0.000000 0 1 1 -16.93826
#> bank_profits_pp
#> 1 0.00000
#> 2 0.00000
#> 3 1.06774
#> 4 0.00000
#> 5 0.00000
#> 6 1.06774
summary(microcredit$treat)
#> Min. 1st Qu. Median Mean 3rd Qu. Max.
#> 0.0000 1.0000 1.0000 0.8005 1.0000 1.0000