Access to the Nation Treasury Secure Data Facility (NT-SDF) is granted under strict supervision based on the submission and acceptance of a research proposal. All research needs to be conducted physically at the secure data facility in Pretoria. Researchers are provided with the de-identified data, computer equipment and software under NT-SDF staff supervision. Researchers must sign confidentiality agreements before any access to data is granted. Researchers need to submit research proposals which are evaluated on the basis of technical merit, policy relevance, capacity building and the research team’s capability. Please see the ‘Request Access to Data’ tab under the Data tab for more information.

None of the data at the NT-SDF can be accessed remotely. All researchers need to be physically present at the NT-SDF in Pretoria to access and work with the data. However, we provide a number of ways to facilitate research remotely.

1. The metadata and dummy data allow you to get to know the data before you arrive at the NT-SDF. The metadata provides descriptions of each dataset, while the dummy data allows you to write your code before testing it in the lab. Please ensure you are sufficiently familiar with the data before you arrive. We encourage all researchers to have do files or other code ready to be tested before their first visit.
2. We encourage international researchers to partner with local research assistants (RAs). Local RAs can be present even when you cannot, and your partnership provides vital capacity building for budding economists and data scientists.
3. We have limited availability within the lab to run code for you. However, we only run code which has already been tested and finalized in the lab and only requires small amendments.

Within the NT-SDF you can access Stata, R, Python, and QGIS.

We primarily work in Stata and are currently working with versions 17 and 18. Your Stata code will need to be compatible with these versions to run.

We encourage work in Stata and R and discourage work in Python. This is because the lab has limited RAM and cannot facilitate many users working in Python.

Due to high demand, booking at the NT-SDF is regulated. The rules below allow us to keep the NT-SDF running. However, if you find your project needs a concession you are welcome to request one.

• Maximum bookings policy: We allow a maximum of 40 bookings per project (each booking is one working day)
• One-project, one-desk policy: Only one person from a project may book in the lab at any one time. However, researchers may work on multiple projects concurrently.
• Project timeline policy: Each project has one year to make their bookings from the date of acceptance of the proposal. PhD projects are an exception and get three years automatically.

Bookings can only be made once a proposal has been accepted.

Demand for the NT-SDF is high. If you book at the lab and do not show up you will be penalized. However, if there is a reasonable reason why you cannot make your booking, or you cancel with due notice, we will not penalize you. The penalties within the lab are as follows:

Not showing up and not cancelling your booking:

- First offence: Your next two bookings will be cancelled

- Second offence: Your bookings within the next two calendar months will be cancelled

Showing up more than 2 hours late for your booking:

- First offence: Your next booking will be cancelled

- Second offence: Your next two bookings will be cancelled

- Third offence: Your bookings within the next two calendar months will be cancelled

Cancelling your booking with short notice (within 24 hours):

- First offence: No penalty

- Second offence: Your next two bookings will be cancelled

- Third offence: Your bookings within the next two calendar months will be cancelled

No raw data will be released from the NT-SDF under any circumstances. All analysis needs to be performed at the secure data facility. Results-based output requests (e.g. in the forms of tables and figures) will be checked by the Data Team before release. Please see the Safe Output Rules document.

There is no funding available for independent projects. On the acceptance of a proposal by the NT-SDF researchers must make their own way to Pretoria. However, the SA-TIED programme publishes requests for proposalson their website for projects in which they are interested.

The National Treasury and UNU-WIDER also provide partial funding for masters and PhD students to write their theses using tax and other National Treasury data. These opportunitiesare advertised yearly.

Be sure to look through the metadata prior to writing your proposal to ensure that your research question is compatible with the data available. Once your proposal is accepted you will be put in touch with the Data Team who will send you an NDA to sign and add you to the booking system. At this stage you should familiarize yourself with the dummy data so that you can begin writing your code and estimating how much time you will need in the lab.

Once you have returned the NDA, written your preliminary code, and booked your desk you are all set to arrive at the National Treasury. Visitors must bring a form of ID (ID card, passport, driver’s license) to present at National Treasury security. Thereafter, the Data Team will provide you with an access card. The access card MUST be returned every Friday. Researchers who do not return access cards will be fined R150.

There is no parking available at the National Treasury. Researchers can park at the Hallmark parking (230 Johannes Ramokhoase Street) if they inform the NT-SDF of their need for parking prior to arrival. Parking for visitors is on level P3, behind the blue gate.

Alternatively, if using the Gautrain from Johannesburg, get off at the Pretoria station and take the Gautrain P3 bus to stop number 8, National Treasury on Thabo Sehume street.