Data Governance for Tax Administrations E-Book

Inter-American Center of Tax Administrations – CIAT

• Herausgeber: Centro Interamericano de Administraciones Tributarias
• Kategorie: Fachliteratur
• Sprache: Englisch

Seitenzahl: 188

Veröffentlichungsjahr: 2024

Beschreibung

The manual outlines various tools for data management, such as data catalogues, documentation portals, and information lifecycle management. Its goal is to provide tax administrations with a comparative perspective to self-evaluate their Data Governance processes based on best practices. This also helps tax administrations to develop their own roadmap for improving their Data Governance. It is a comprehensive guide that explains how Data Governance functions within the context of tax administrations. 

Leseprobe

Data governance for tax administrations. A practical guide

© 2022, All rights reserved

Inter-American Center of Tax Administrations - CIAT

ISBN: 978-9962-722-27-4

Cover design

Illustrates the butterfly effect that data errors can cause in tax administrations, with increasing consequences expanding in various data domains. The cover includes an adapted version of TwoLorenzOrbits by XaosBits CC BY 2.5. It is the trajectory of a Lorenz System.

This publication is funded by GIZ, - the Deutsche Gesellschaft für Internationale Zusammenarbeit GmbH, within a joint effort with CIAT.

ABOUT THE AUTHORS

Andrés Duque

Mr. Duque is an Ecuadorian national. He is a Systems Engineer with a master's degrees in Systems Management and Business Intelligence, as well as Data Management and Enterprise Architecture certifications. For 15 years he has participated and led data management and data governance projects in the public and private sectors.

He owns Datalegio, a data management consulting firm and is president of DAMA Ecuador Chapter. As an independent consultant he has supported the Inter-American Center of Tax Administrations - CIAT.

Wolney Martins

Mr. Martins is a Brazilian national. He holds a degree in engineering with a postgraduate degree in Computer Networks. For 20 years, he worked at SERPRO, the IT company of the Ministry of Finance of Brazil, where he held positions as an analyst, department head, director (operations, technology, and systems development) and president.

Since 2014, he is an independent IT consultant, working in public finance and tax administration systems, crisis management, IT planning and innovation, working for CIAT in several initiatives.

Antonio Seco

Mr. Seco is a Brazilian national from Cabo Frio-RJ. He holds a degree in Electrical Engineering, with a Master of Science degree in ICT Management and a postgraduate in Tax Administration. He was a SERPRO official in the Ministry of Finance of Brazil, where he participated in the implementation of SIAFI and other public finance systems, also having worked as consultant and IT component leader in Public Finance modernization projects in Latin America, the Caribbean and Africa.

He is an Independent Senior Consultant of the Inter-American Center of Tax Administrations (CIAT) and the Inter-American Development Bank (IADB).

Raul Zambrano

Mr. Zambrano is an Ecuadorian national. He is a Systems Engineer with a master's degree in public policy management and serves at the CIAT organization as Director for Technical Assistance and Technology.

He has directed international multidisciplinary teams in modernization projects carried out by CIAT in six countries and has participated in the improvement of processes and in the development and evaluation of information systems as a consultant in over 20 tax and customs administrations at national and subnational levels in Latin-America, the Caribbean and Africa.

TABLE OF CONTENTS

1.Setting the Landscape

1.1.Data, Information, Knowledge

1.2.The DIKW Model

1.3.The Growing Importance of Data Governance in Tax Administrations

1.4.Data Management vs. Data Governance

1.4.1.Data Management

1.4.2.Data Governance in Data Management

1.4.3.What is data governance all about?

1.4.4.Data Lifecycle

1.5.Data Attributes

1.5.1.Common Business Vocabulary

1.5.2.Master and Reference Data

1.5.3.Metadata

1.5.4.Operational and Analytical Data

1.5.5.Structured and Unstructured Data

1.5.6.Security and Privacy

1.5.7.Data Classification

1.5.8.Data Retention

1.5.9.Data Lineage

1.5.10.Data Masking

1.5.11.Cloud Systems, Data, and Sovereignty

1.5.12.Data Domain

1.5.13.Data Quality

2.Data Governance at-a-glance

2.1.Data Governance Frameworks

2.2.Data Governance Policies

2.3.Data Governance Processes

2.4.Data Governance Roles

2.5.Data Governance Committees and Councils

2.6.Data Governance Roles and the IT Department

2.6.1.About the organizational titles of a data governance structure

2.7.Data Literacy

3.Data Governance for Tax Administrations: Strategic Perspectives

3.1.Data Strategy

3.2.Metrics to Monitor and Measure the Impact of Data Strategy

3.3.Mapping Technical Capabilities to Processes and Analytics

3.4.Mapping Organizational and Program Capabilities to the Data Strategy

3.5.Change Management

3.6.Final Comments

4.Data Governance for Tax Administration: Modeling Proposal

4.1.Data governance principles and policies

4.2.Data Governance Capabilities

4.2.1.Data Governance Strategy Management

4.2.2.Data Governance Operation Management

4.2.3.Data governance support management

4.3.Data Governance Organization

4.3.1.Basic Data Governance Implementation and Evolution in Small Economies

4.4.Organizational Structure Roles and Responsibilities

4.5.Light Data Governance Model

4.6.Data Stewardship

4.7.Data Quality Dimensions

5.Data Governance for Tax Administration: Maturity Assessment

5.1.Maturity Models

5.1.1.Lack of Precision in the Description of Maturity Models

5.1.2.Maturity Model: It is not about "how to do" data governance

5.1.3.ISORA and TADAT

5.1.4.Using an Existing Maturity Model

5.2.Data Governance Maturity Models

5.2.1.Why use DAMA-DMBoK2?

5.2.2.The Importance of Measuring

5.2.3.How to Measure

5.2.4.DAMA-DMBoK2 in a Nutshell

5.2.5.Short Description of the Stanford Data Governance Maturity Model

5.2.6.Data Governance Matters

5.2.7.Data Governance and COBIT

6.Data Governance Tools

6.1.Glossary of Terms

6.2.Data Catalog

6.3.Data Lineage

6.4.Document Management and Collaboration Portals

6.5.Other beneficial tools

6.6.References: Market Research

6.6.1.Gartner Magic Quadrant

6.6.2.The Forrester Wave™

7.Roadmap for the Implementation of data Governance in a tax Administration

7.1.First Activities

7.2.How to Implement Data Governance?

7.3.Why deploy data governance?

7.4.Why not implement data governance or implement it only partially?

7.5.Initial Studies

7.6.Pay Attention to Change Management and Communication

7.7.Roles and Responsibilities

7.8.Address a maximum of four knowledge areas at a time

7.9.Framework

7.10.Maturity assessment

7.11.Progressive Implementation of Data Governance

7.12.Final Comments

8.Data Governance Guides

8.1.Data Strategy Definition Guide

8.1.1.Data Strategy Route

8.1.2.Data Strategy Execution

8.2.Data Management Principles and Policies Definition Guide

8.2.1.Principles

8.2.2.Policies

8.3.Data quality dimensions definition guide

8.4.Data Management Maturity Assessment Guide (data governance focus)

8.4.1.Stanford Assessment

8.4.2.Other assessment tools - examples

8.5.Data Governance Roles Designation Guide

8.5.1.Assignment of Roles

8.5.2.Use of RACI Matrix

8.6.Data Governance Stakeholder Identification Guide

8.6.1.Identify Stakeholders

8.6.2.Analyze and Map Stakeholders

8.7.Practical Implementation Guide

List of tables and figures

References

1.SETTING THE LANDSCAPE

Information and knowledge are keys for organizations to fulfill their objectives.

The DAMA association1 emphasizes that organizations with reliable, high-quality data about their users, products, services, and operations can make better decisions than those without. The absence of these properties will result in a waste of opportunities and deficient performance (DAMA-DMBoK2, 2017). This assertion is valid with greater emphasis for tax administrations, where data and its products are fundamental to accomplishing its mission.

1.1.Data, Information, Knowledge

A still current and passionate discussion in information sciences and knowledge management is the differentiation among data, information, knowledge, and (sometimes) wisdom.

Models available often present these concepts as a hierarchy, in which mastery of the lower level provides the opportunity to scale to the next level. This structured ascension is not a point of agreement among scholars, but it can be a starting point to understanding the concepts and establishing more precise communication among different users.

A theoretical model helps in understanding the transformations and relationships among these concepts.

1.2.The DIKW Model

Among the available models, one of the most visible, but not without controversies, is the so-called DIKW (Data, Information, Knowledge, Wisdom), presented in the form of a pyramid (Figure 1-1). One of the high points of the controversies is the inclusion and definition of the last attribute, “wisdom”2.

Figure 1-1 The DIKW model.

Source: Prepared by the authors

The implicit assumption of this model is that tax administrations can use data to create information; information can be used to develop knowledge, and knowledge can be used to create wisdom.

The following definitions and associations to different types of information systems can be performed on this model:

Table 1-1 DIKW Model - elements definitions and information systems associations.

Element

Definition (Ackoff, 1989)

Association (Rowley, 2007)

Data

Symbols

Transaction Processing Systems

Information

Data processed to be useful; provides answers to who, what, where, and when questions

Management Information Systems

Knowledge

Application of data and information; answers how questions

Decision Support Systems

Wisdom

Evaluated understanding

Expert Systems

Source: Prepared by the authors

1.3.The Growing Importance of Data Governance in Tax Administrations

Tax administrations are related to the automated processing of data from the beginning. After all, they were (along with the census bureau) the first users of the so-called “data processing machines” in government.

Tax returns and the provision of ancillary information in digital format by taxpayers and auxiliary institutions (especially financial institutions) have been part of the life of tax administrations and taxpayers in the recent past.

In those times, the data was structured with a minimal data management schema, consisting fundamentally of a data dictionary3. IT4 personnel had control of the processes of extracting, transforming, and loading the data. The data needed to be cleaned5, mostly manually.

Data management was the responsibility of the IT area, with occasional advice from the business areas. Thus, organizations merged data management with IT management.

Nowadays, data availability has increased dramatically in quantity and formats, as well as the dependence of tax administrations on its treatment. As established in (Collosa, 2021), this is mainly due to:

The significant expansion of computer processing and storage capacity associated with the reducing their costs.

The increasing availability of communications networks and broadband Internet.

The development of effective models to capture, store and process massive data and advanced cognitive algorithms.

The emergence of new data sources and formats e.g., sensors, GPS

6

, OCR

7

cameras for truck plates, RFID

8

chips and antennas, social networks, etc. (Arias & Zambrano, 2020) , including electronic invoices (Barreix & Zambrano, 2018) and tax information exchange between countries.

A few years ago, the importance of using data in the work of organizations was mentioned with a quote from the famous “total quality guru” W. E. Deming “without data, you’re just another person with an opinion” (ETF-Europa, 2018). Currently, KPMG analysts have rephrased this quote: “without trust in your data, you’re just another person that consumes data” (KPMG, 2021).

Tax administrations are strongly linked to this reality.

Over the past several years, tax administrations worldwide have started to undergo digital transformation, collecting data from non-traditional sources and formats, and accumulating them in their databases. Tax administrations can rely heavily on data and algorithms for their internal processes and provide more and better services to the taxpayers and other stakeholders, so tax administrations can count on data accuracy, completeness, and availability.

The following numbers illustrate these aspects as presented by the OECD

From 2014 to 2019, average e-filing rates have increased significantly between 13 and 18%.

Over 80% of payments (by value and numbers) are made electronically.

Close to 50% of tax administrations pre-fill PIT (Personal Income Tax) returns with specific deductibles expenses.

New data sources allow pre-filling to move to VAT (Value-Added Tax) and CIT (Corporate Income Tax) returns.

A growing number of tax administrations use virtual assistants to respond to taxpayers enquires and support self-service.

Use artificial intelligence in services supporting taxpayers and tax officials.

Percentage of tax administrations that allow taxpayers to register online up from 70% (2015) to 97% (2019).

With the increasing availability of data, compliance work focus can change to prevention.

At the same time, society demands more responsibility from the entities that obtain and consume data from citizens and companies, establishing a series of data protection laws and regulations.

In this context, a modern data governance landscape must be set up to ensure data confidentiality, availability, quality, and integrity and reinforce the legal protection instruments (as data protection regulations) and compliance rules.

In other words, data governance must ensure that data are consistent and trustworthy and don’t get misused, so as in the transactional operations up to enable the effective use of data analytics helping to optimize operations and drive business decision-making.

This data governance landscape includes all hierarchical levels of a tax administration, intending to define policies, standards, processes, and participating in data governance committees.

1.4.Data Management vs. Data Governance

Data is an essential asset within tax administrations. Data can give tax administrations different benefits through its use and exploitation, as well as through its correct administration.

To generate value, tax administrations require data. It needs to be managed consciously; for this, the organization must put a set of fundamental practices in place to allow it to manage data like any other business asset.

1.4.1.Data Management

According to DAMA (DAMA-DMBoK2, 2017), Data Management is defined as the development, execution, and supervision of plans, policies, programs, and practices that deliver, control, protect, and enhance the value of data and information assets throughout their lifecycles.

Figure 1-2 The DAMA-DMBoK2 Data Management Framework (The DAMA Wheel).

Source: (DAMA-DMBoK2, 2017)

Organizations develop data management practices through different disciplines that cover all activities around the data lifecycles, e.g., Data Governance, Data Architecture, Data Quality, Business Intelligence, etc.

DAMA-DMBoK2 defines 11 disciplines for data management, with data governance at the center, as shown in Figure 1-2.

1.4.2.Data Governance in Data Management

As tax administrations face different challenges of information systems implementations, be it to support analytical capabilities, transactional, or business processes, it is recognized that data assets deserve to be managed correctly.

Traditionally, IT departments in organizations have been responsible for promoting data projects. Now, IT departments cannot operationalize these projects in isolation or without the commitment of the whole institution.

To manage data correctly, it is essential to have roles and responsibilities that allow accountability for the problems that data usually present and their inherent definitions. Here is where data governance intervenes as a framework that allows organizations to establish a system of rights and obligations for decision-making throughout the entire data lifecycle.

Data management requires a structure that controls and guarantees the correct administration of data, and that is why the implementation of data governance programs is gaining greater importance.

DAMA-DMBoK2 defines data governance as “the exercise of authority and control (planning, monitoring, and enforcement) over the management of data assets” (DAMA-DMBoK2, 2017). On the other hand, Ladley (Ladley, 2020) mentions that the purpose of data governance is to ensure that the data is managed properly, according to policies and best practices.

As we can see from DAMA-DMBoK2 Management Framework (Figure 1-2), data governance is at the center of all DAMA-DMBoK2 disciplines because it is crucial to control all kinds of data projects through centered guidance.

Data governance provides the best tools to manage data correctly, e.g., principles, policies, functions, processes, procedures, etc.

1.4.3.What is data governance all about?

Data governance is a key component of data management. Tableau (Tableau Software, 2020) proposes that data governance helps answer questions like:

Who has ownership of the data?

Who can access what data?

What are security measures are in place to protect data and privacy?

How much of our data is compliant with new regulations?

Which data sources are approved to use?

Governance models and practices won’t be the same across every organization, even among tax administrations, but these models are crucial pieces of the process. As also mentioned in the paper referenced above, the following stand out:

Data quality is a pillar of data management. It doesn’t matter how robust your governance program is if you don’t have quality data. Having data that is accurate, complete, and reliable is a cornerstone of any data-driven organization.

Data security and compliance is defining and labeling data by their levels of risk and then creating secure access points, keeping a balance between user interaction and safety, considering access levels that can go at the functional, object, or even field level (Martins, Nieto, Seco, & Zambrano, 2020).

Data stewardship helps monitor how teams use data, and stewards lead by example to ensure data access, security, and quality, defining clear interactions and responsibilities of different data stakeholders.

Data transparency matters because every piece of the process and the procedures you put in place should work within a model of clarity.

Analysts and business users should quickly find out where their data comes from and know if there are any special considerations.

1.4.4.Data Lifecycle

The data lifecycle is the sequence of stages a particular data unit goes through, from its initial generation or capture to its eventual archival or deletion at the end of its useful life (Wigmore, 2017).

Figure 1-3 The data lifecycle key activities.

Source: Prepared by the authors based on (DAMA-DMBoK2, 2017)

The data governance practices must cover all data lifecycle, as it is shown in Figure 1-3.

1.5.Data Attributes

Attributes are specification or characteristic that helps define a data entity. In data management, some attributes refer to the processing characteristics of the data and its lifecycle, use and structure, security requirements, quality parameters, and compliance needs.

The following topics present summaries of several essential data attributes for their management.

Specific chapters of this document will take up these attributes.

1.5.1.Common Business Vocabulary

A typical business vocabulary is a set of commonly defined data names and definitions documented in a business glossary, for example, within a data catalog or independently.

Its purpose is to ensure that data is consistently named and commonly understood, especially when it is shared.

A specialized software may supports creating and maintaining a business glossary with a common business vocabulary of common data names, definitions, and attributes for data entities. This is critical to promoting the proper common understanding and use of tax terms.

Most countries already have some formalization of tax terms, but often in scattered or incomplete documents. These documents can be good sources for everyday business vocabulary.

1.5.2.Master and Reference Data

According to (DAMA-Dictionary, 2009), Master Data is “the data that provides the context for business activity data in the form of common and abstract concepts that relate to the activity. It includes the details (definitions and identifiers) of internal and external objects involved in business transactions, such as customers, products, employees, vendors, and controlled domains (code values)”.

Another definition by the consultant company Gartner Group for Master Data is the consistent and uniform set of identifiers and extended attributes that describe the core entities of the enterprise, including customers, prospects, citizens, suppliers, sites, hierarchies, and chart of accounts.

Transaction processing applications and analytical systems need Master Data, so they must be application agnostic.

An example of a Master Data, a subset of the suggested elements for taxpayer identification (Falkenbach, González, Redondo, & Zambrano, 2020), is shown below.

Table 1-2 Master Data (Taxpayer identification)

Source: Prepared by the authors

Reference Data is any data used to characterize or classify other data or to relate data to information external to an organization. The most basic Reference Data consists of codes and descriptions, but some Reference Data can be more complex and incorporate mappings and hierarchies (DAMA-DMBoK2, 2017).

Reference Data has characteristics that distinguish them from Master Data: they are less volatile; data sets are generally less complex and smaller; they have fewer columns and rows. The management focus differs between Master and Reference Data.

Among the types of Reference Data, we mention Internal Reference Data (created to support internal processes and applications), Industry Reference Data (created and maintained by industry associations or government bodies), and Computational Reference Data (which differs from other types because of the frequency with which it changes).

Reference data could be presented and used in many ways, using a code-value strategy or fixed labels (Zambrano, 2010). A basic Reference Data example is shown below.

Table 1-3 Reference data (list)

Code Value

Description

AR

Argentina

BR

Brazil

PY

Paraguay

Source: Prepared by the authors

1.5.3.Metadata

The Gartner Glossary defines metadata as “information that describes various facets of an information asset to improve its usability throughout its lifecycle” (Gartner Inc., 2012). The DAMA, in (DAMA-DMBoK2, 2017), adds other features: “metadata” includes information about technical and business processes, data rules and constraints, and logical and physical data structures. It describes the data itself (e.g., databases, data elements, data models), the concepts the data represents (e.g., business processes, application systems, software code, technology infrastructure), and the connections (relationships) between data and concepts.

Necessary for structured data, metadata is perhaps most important for unstructured data (see the basics of structured and unstructured data later in this chapter). New practices are emerging for treating unstructured data in data lakes, for example, a minimum set of metadata attributes of ingested objects is collected as part of the ingestion process, such as name, format, source, version, and date received, producing a catalog.

There is also a requirement for a metadata lineage to provide an audit trail to know where the data originated and how it has been transformed in this way to the point of use. It may also trace who or what is maintaining data, including when and where it occurs.

Metadata turns information into an asset, and accurate metadata can help prolong the lifetime of existing data by assisting users in finding new ways to apply it.

Many IT tools are available to deal with metadata, as we will see later in this document.

1.5.4.Operational and Analytical Data

The world of data used to be divided between the applications and processes creating and updating data (operational) and the solutions and processes analyzing data (analytical). The two are structurally different and provide different types of insight.

Operational data is produced by the day-to-day operations of a tax administration, such as changes in the tax registry, tax payments, taxpayers’ appeals, etc. Operational data are produced mainly by the OLTP9 systems, supporting high-volume, low-latency access. These systems create, read, update, or delete one piece of data at a time.

Analytical data is used to support business decisions, instead of recording the data from actual operational business processes. Examples include grouping taxpayers by income or amount of tax due over time. Every organization will have different questions to answer and other decisions, so analytical data is definitely not one-size-fits-all. Analytical data is best stored in a system designed for heavy aggregation, data mining, and ad hoc queries, called an OLAP10 system or a Data Warehouse (Simpson, 2016).

The core of the analytical data is the institution’s operational data.

Figure 1-4 Operational and analytical data.

Source: Prepared by the authors based on (Simpson, 2016)

Operational databases contain transactional data, while analytical databases are designed for efficient analysis, as can be seen in Figure 1-4.

1.5.5.Structured and Unstructured Data

According to Talend (Talend Company, 2020) structured data is data that has been predefined and formatted to a set structure before being placed in data storage, which is often referred to as schema-on-write11. The best example of structured data is the relational database: the data has been formatted into precisely defined fields, such as tax identification numbers or addresses, to be easily queried with programming languages like SQL.

The same source establishes that unstructured data is stored in its native format and not processed until it is used, known as schema-on-read12