************************************************************
*  "Sources of Rising Student Debt in the U.S.:	
*   College Costs, Wage Inequality, and Delinquency"
*   Written by Heejeong Kim and Jung Hwan Kim
*   E-mail: heejeong.kim@concordia.ca
*   	    junghwankim31@gmail.com
*   last updated: Nov-10-2025
************************************************************
All the results are saved in the folder "Results"
All the Figures are saved in the folder "Figures"

*If not specified, running time is around 1 minute.

Folder Model_Fortran
- Solves quantitative model
- See below for details

Folder Section 4.1.1
- PSID regression: run Main_PSID_regression.do first
- Wage shock estimation: next run Main_estimation.m. (Running time: 6 hours)
- To generate Figures: Run Sectin_4_1_1_Figure.m (Running time: 10 minutes)
- Estimation results are automatically saved in "Results/wageprocess.xlsx"
- Generate Results/wagepremium.txt

Folder NLSY79
- Run nlsy79_code_final.do
- Generate calibration moments and logafqt80.csv | dropout_abil.txt
- Run nlsy79_abilityfi9lter.m to generate skilled_abil_HP_time.txt; unskilled_abil_HP_time.txt

Folder Section 4.2
- Run Figure_7.m to generate Figure7 (Need to run NLSY79 code first)
- Figure7.m also writes ability_10.txt and abilitygrid.txt

Folder NLSY97
- Run nlsy97_code_final.do to first to generate nlsy97_output.csv
- Next, run nlsy97_stat.m

Folder NCES
 - Has tuition information.
 
 
 ****************
 Folder Model_Fortran
 
 Fortran90, compiled in Microsoft Visual Studio. Open MP is used.

System Requirement
	- Stack Reserve Size: 2048000000
	- Stack Commit  Size: 864000000
	- Enable Large Addresses: Support Addresses larger than 2GB
	- Linker - Additional Dependencies: mkl_intel_lp64.lib mkl_sequential.lib mkl_core.lib  legacy_stdio_definitions.lib hdf5_fortran.lib


HDF5 v 1.10.7 is installed (https://www.hdfgroup.org/download-hdf5/)
	
F90 files to compile
	- StdDebtGrowth.f90     : main file for the program
	- steadystate_solve.f90 : subroutine to solve steady state
	- StdDebtGrowth_mod.f90 : subroutines 
	- transition_solve 	: subroutine to solve transition
	- Lib_HK		: folder that contains all the libraries

Running time 
	 The model is solved on a PC with two Intel Xeon CPU E5-2640 2.40GHz 	 processors (40 logical cores) and 64GB RAM.
	 With OpenMP, the running time for the benchmark case including 	 transition is around 8 hours

All the results are saved "../Results_h5"

Figures will be generated by running "Figure_plot_final.m" and figures are saved in the folder "../Figures"


1. Benchmarkcase  
	integer(ik), parameter  :: idefault = 0_ik              ! 1 - solve the model without delinquency option
	integer(ik), parameter  :: iidr = 0_ik                  ! 1 - solve the model with    IDR
	integer(ik), parameter  :: ituition = 0_ik              ! 1 - solve the model without time-varying education costs
	integer(ik), parameter  :: iwp = 0_ik                   ! 1 - solve the model without time-varying wage premia
	integer(ik), parameter  :: ishock = 0_ik                ! 1 - solve the model without time-varying wage shock processes
	integer(ik), parameter  :: iability = 0_ik              ! 1 - solve the model without time-varying ability premia

2. Fixed College Costs (All else is equal to the benchmark case)
	integer(ik), parameter  :: ituition = 1_ik              ! 1 - solve the model without time-varying education costs

3. Fixed ability premia
	integer(ik), parameter  :: iability = 1_ik              ! 1 - solve the model without time-varying ability premia

4. Fixed college wage premia
	integer(ik), parameter  :: iwp = 1_ik                   ! 1 - solve the model without time-varying wage premia

5. Fixed variances of wage shocks
	integer(ik), parameter  :: ishock = 1_ik                ! 1 - solve the model without time-varying wage shock processes

6. No-delinquency economy
	integer(ik), parameter  :: idefault = 1_ik              ! 1 - solve the model without delinquency option

7. IDR economy
	integer(ik), parameter  :: iidr = 1_ik                  ! 1 - solve the model with    IDR

8. The economy with wage garnishment
	real(rk),    parameter  ::  wgar = 0.02_rk               ! wage garnishment in robustness