Introduction to the South African Revenue Service and National Treasury Firm‐Level Panel

• Published date: 01 January 2018
• DOI: http://doi.org/10.1111/saje.12156
• Date: 01 January 2018
• Author: C. Friedrich Kreuser,Duncan Pieterse,Elizabeth Gavin

INTRODUCTION TO THE SOUTH AFRICAN REVENUE

SERVICE AND NATIONAL TREASURY FIRM-LEVEL PANEL

DUNCAN PIETERSE*, ELIZABETH GAVIN

†

AND C.FRIEDRICH KREUSER

‡

Abstract

The South African Revenue Service and National Treasury Firm-Level Panel is an unbalanced panel

data set created by merging several sources of administrative tax data received during 2015. The four

data sources that constitute the panel are: (i) company income tax from registered f‌irms who submit

tax forms; (ii) employee data from employee income tax certif‌icates submitted by employers; (iii)

value-added tax data from registered f‌irms; and (iv) customs records from traders. These data sets con-

stitute a signif‌icant and unique source for the study of f‌irm-level behaviour in post-apartheid South

Africa. We review the key data sources used to construct the panel, highlight some important ques-

tions that arise as a result of panel construction, discuss the biases in the resulting data, compare key

aggregates in the panel to other data sources, and provide a descriptive overview of the tax records.

JEL Classif‌ication: C55, C80

Keywords: administrative data, tax, f‌irm level

1. INTRODUCTION

The South African Revenue Service (SARS) and National Treasury (NT) Firm-Level

Panel (hereinafter the “SARS-NT Panel”), is an unbalanced

1

panel data set created by

merging several sources of South African administrative tax data received during 2015.

2

The four data sources that constitute the SARS-NT Panel are: (i) company income tax

(CIT) data from CIT-registered f‌irms who submit CIT forms; (ii) employee data from

employee income tax certif‌icates submitted by employers (i.e.IRP5andIT3a);(iii)

value-added tax (VAT) data from VAT registered f‌irms; and (iv) customs records from

traders. In this paper, we review the key data sources used to construct the SARS-NT

Panel, highlight some of the important questions that arise as a result of panel construc-

tion, discuss the biases in the data, compare key aggregates in the SARS-NT Panel to

other data sources, and provide a descriptive overview of the tax records.

The forms that tax-paying persons and entities submit to SARS contain various f‌ields,

which are used to construct the variables in the SARS-NT Panel. The CIT records

* Corresponding author: Director, Microeconomic Policy, National Treasury, 240 Madiba

Street, Pretoria, South Africa . E-mail: Dunca n.Pieterse@treasury.gov.za

†

Policy Research, South African Revenue Service, Pretoria, South Africa.

‡

UNU-WIDER Research Collaborator.

1

In an unbalanced panel the number of time periods is not the same for all f‌irms. For example,

this occurs when f‌irms drop out of the panel or new f‌irms join the panel.

2

This could lay the foundation for a broader South African Administrative Database, which

includes other sources of available tax data (such as information from the EMP501 and EMP201

forms and personal income tax) as well as other administrative data. See Barnes et al. (2007) for a

comprehensive review of other sources of administrative data for South Africa.

V

C2017 UNU-WIDER. South African Journal of Economics published by John Wiley & Sons Ltd on behalf of

Economics Society of South Africa.

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which

permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not

used for commercial purposes. doi: 10.1111/saje.12156

6

South African Journal of Economics Vol. 86:S1 January 2018

South African Journal

of Economics

contain f‌irm characteristics, such as the sector in which a f‌irm operates, f‌inancial infor-

mation from their income statements and balance sheets, and tax information. The

employee records from individual IRP5 and IT3a forms contain all employee related

incomes, deductions, taxes and payments made by the f‌irm (such as skills development

levy and unemployment insurance fund (UIF) payments). The VAT records contain

information about input and output VAT paid and payable by the f‌irm as well as VAT

specif‌ic information, such as VAT paid on capital expenditure. The trader and customs

records contain information about the products imported and exported by traders,

including the product codes, value and volume of goods, and origin and destination

information. Together, these data sets constitute a signif‌icant and unique data source for

the study of f‌irm-level behaviour in South Africa.

Using administrative tax records for cutting-edge empirical research is becoming wide-

spread internationally (see Card et al., 2010). First, administrative data offer much larger

sample sizes than traditional survey data sources, and can be harnessed to generate more

compelling research designs. Second, administrative records have an inherent longitudi-

nal structure that enables researchers to follow f‌irms and individuals over time and con-

duct credible public policy evaluation. Third, administrative data provide better

information than is typically available for survey sources, which suffer from high and ris-

ing rates of non-response, attrition, and under-reporting. As a result, using administrative

tax data to conduct policy research has become widespread in such countries as Finland,

Denmark, Sweden, the Netherlands, France, Germany and New Zealand.

By combining different sources of administrative tax data, the SARS-NT Panel covers

various aspects of f‌irm behaviour. In the construction of the panel, all variables that could

be used to identify f‌irms were removed to create an anonymised data set where the conf‌i-

dentially of f‌irms and individuals is ensured.

The remainder of this paper is organised as follows: Section 2 explores the key ele-

ments of the SARS-NT Panel including which tax records are included, how a f‌irm is a

def‌ined, as well as how the SARS-NT Panel is constructed; Section 3 contains a discus-

sion of data characteristics including biases in the data; Section 4 compares key variables

in the data to other data sources such as the Quarterly Employment and Financial Statis-

tics released by Statistics South Africa; and Section 5 brief‌ly draws out some key descrip-

tive statistics to convey the richness of the data.

2. KEY ELEMENTS OF THE SARS-NT PANEL

2.1 The Tax Records That Constitute the SARS-NT Panel

CIT data form the backbone of the SARS-NT Panel. CIT forms are completed by all

businesses resident

3

for tax purposes in South Africa and include information in respect

of taxable income.

4

CIT records are drawn from the IT14 and ITR14 forms. The ITR14

3

A business is resident in South Africa if it is incorporated or effectively managed in South Africa.

Non-resident companies which operate through a branch or which have a permanent establish-

ment within the Republic are subject to tax on all income from a source within South Africa.

4

Not-for-prof‌it organizations that receive preferential tax treatment are an exception provided

they meet the requirements set out in the Income TaxAct. Once approved by SARS, the organiza-

tion is registered as a public benef‌it organization (PBO), allocated a unique PBO reference num-

ber, and annually submits a separate IT12EI income tax return.

7South African Journal of Economics Vol. 86:S1 January 2018

V

C2017 UNU-WIDER. South African Journal of Economics published by John Wiley & Sons Ltd on behalf of

Economics Society of South Africa.

form replaced the IT14 form on 4 May 2013. In its current form, the SARS-NT Panel

is available for the tax years from 2008 to 2014. Since all tax returns submitted by a tax-

registered entity must be completed electronically or at a SARS branch within 12 months

of its f‌inancial year end, data for the 2014 tax year are still somewhat incomplete. The

CIT data are augmented with variables from the following databases: (i) customs records;

(ii) VAT vendors; and (iii) employee income tax certif‌icates.

(i) Customs Records The customs records contain information on the import and export

activities of trading entities. Entities that appear in the trader database include the follow-

ing: local importers and exporters; foreign importers and exporters; rebate users;

5

“drawback” manufacturers;

6

participants in the Automotive Production and Develop-

ment Programme (APDP);

7

and f‌irms who operate in an Industrial Development Zone.

8

Customs records are available at the transaction level for entities identif‌ied with a cus-

toms reference number. Each transaction entry includes information on the Rand

amount imported or exported, the destination or origin, the quantity and unit of the

product, as well as detailed product codes using the Harmonised System (HS) of the

World Customs Organisation. Customs reference numbers are linked to a CIT reference

number using a link table provided by SARS. These f‌irms are then matched to the CIT

database as detailed in Section 2.3. Customs records were retained without any changes,

which mean that duplicate entries (to the extent that they exist) were not identif‌ied or

dealt with. While customs records are available to researchers at the transaction level, the

SARS-NT Panel includes aggregated information for each customs reference number

level for the tax year in question. The SARS-NT Panel includes the total Rand value of

imports and exports, number of import origins and export destinations, and number of

unique products imported or exported.

(ii) VAT Data VAT is an indirect tax on the consumption of goods and services in the

economy and is charged at each stage of the production and distribution process. VAT is

presently levied at the standard rate of 14% on the supply of most goods and services

and on the importation of goods. Some goods and services are subject to VAT at zero

rate or are exempt from VAT. Although any person that carries on a business may register

for VAT, VATregistration is only mandatory if a company’s taxable supplies are in excess

of R1 million in any consecutive twelve-month period. These businesses become vendors

that act as the agent for government in collecting VAT. VAT-registered entities submit

5

Individuals and entities who are allowed to import goods without paying duties under certain

conditions, e.g. for processing, manufacturing and packaging with the intention of exporting

them.

6

All local persons who import goods into South Africa that are used in the manufacture of goods

which are subsequently exported should register if they wish to apply for a drawback (refund) for

exporting goods on which duties were already paid.

7

APDP is a production incentive programme for the motor industry whereby components are

imported under rebate item 317.03 for the manufacture of light motor vehicles and 317.06 for the

manufacture of automotive components. Anyone who imports components for use in terms of

such provisions should register for the APDP.

8

Available at: http://www.sars.gov.za/ClientSegments/Customs-Excise/Processing/Pre-assess-

ment/Registration/Pages/default.aspx.

8 South African Journal of Economics Vol. 86:S1 January 2018

V

C2017 UNU-WIDER. South African Journal of Economics published by John Wiley & Sons Ltd on behalf of

Economics Society of South Africa.