The persistence of debt of government examines the existing range to debt types people owe to government, the types of debt and the length of time the people have had debt to government.
The initial analysis focused on how often people owe debt to more than one government agency, and patterns of debt owed across three agencies: the Ministry of Social Development (MSD), the Ministry of Justice (MoJ), and Inland Revenue (IR). Following this initial work, conversations with the Department of the Prime Minister and Cabinet and with the cross-agency working group on debt identified the persistence of debt and its impact as an area needing further investigation.
The report examines debt persistence, by considering the 713,000 New Zealand residents who owed debt to MSD, MoJ, or IR in 2020 and examining debt records back nine years to 2012. The research also provides further insights on individuals and families who experience prolonged or persistent debt.
This code should be read and used in alongside the accompanying report: The-persistence-of-debt-to-government-research-report.
It is necessary to have an IDI project if you wish to run the code. Visit the Stats NZ website for more information about this. This analysis has been developed for the IDI_Clean_20211020 refresh of the IDI. As changes in database structure can occur between refreshes, the initial preparation of the input information may require updating to run the code in other refreshes.
In addition, MoJ debt information was ad hoc loaded to this refresh. To use this information in another refresh, you must first link it to the 20211020 and then build links between this refresh and your preferred refresh.
The R code makes use of several publicly available R packages. Stats NZ who maintain the IDI have already installed in the IDI the all the key packages that this analysis depends on. Should the version of these packages be important, this analysis was conducted using odbc version 1.2.3, DBI version 1.1.0, dplyr version 1.0.0, and dbplyr version 1.4.4.
If applying this code to another environment other than the IDI, several features of the environment are required: First, R and some database manager (such as SQL Server) must be installed. Second, these must be configured such that R can pass commands to, and retrieve results from, the database. Once the environment is configured correctly, then adjustments to the code in response to the new environment can be considered.
This repository contains all the core code to assemble the data and run the analysis.
-
Definitions: All the definitions of the inputs are found here. This folder contains subfolders for population definitions and measure definitions. These definitions can be used independently of this project.
-
Analysis: This folder contains all the R scripts for executing each stage of the analysis. It also contains the R code for the 2019 - Student Loan Forecast model
-
Output: This folder contains the R script for constructing the resulting aggregated output table.
Note that this project uses the Dataset Assembly Tool to simplify dataset preparation prior to analysis. Researchers interested in the tool itself are advised to see its separate GitHub project. Documentation for the Dataset Assembly Tool can be found on our website. Interested readers are advised to view both the primer and the training presentation.
The GitHub page for the assembly tool is here.
Prior to running the project be sure to review the associated report and documentation.
- Setup SQL inputs, population and measures.
- Make any required changes to the SQL definition scripts.
- Run all scripts inside the SQL folder (except Equivalised hhld income.sql)
- Test the performance of the assembly tool.
- Enter the connection details at the top of dbplyr_helper_functions.R
- Run automated_tests.R to confirm all tests run correctly
- Assemble the data for analysis.
- Set run_assembly.R for the individual assembly and run the script
- Set run_assembly.R for the household assembly and run the script
- Run Equivalised hhld income.sql
- Clean the prepared table and create summary output.
- Run each tidy_variables R script (household and individual)
- Run each summary R script (household and individual)
- Run both median scripts
Social Wellbeing Agency (2022). Understanding debt to government 3. Source code. https://github.com/nz-social-wellbeing-agency/understanding_debt_and_debtors_to_government_3
If you have any questions email [email protected]