Effectiveness of Tax Control Framework – single audit via statistical sampling

Effectiveness of Tax Control Framework – single audit via statistical sampling

26 November 2013

Authors: Hein Kloosterman, Ferry Geertman and Robbert Hoogeveen

In order to quickly gain insight into the level of tax risks for all tax types (calculation of the potential assessment), the single audit method of tax authorities can be used. This method provides a quick insight into the performance of the Tax Control Framework since the results can be used to evaluate whether or not these are within the risk tolerance limits of the company.

An overview of this method is presented below.

The term ‘Single Audit’ originates from the government accountancy. Initially, it referred to an audit simultaneously carried out for multiple supervisors/regulators and for whom the controlled has an accountability to report to each of the supervisors/ regulators.

Because of such a single audit much unnecessary duplication of work can be avoided. After all, separately monitoring different type of reports that all originates from the same (financial) accounting system, leads to extensive audit burden. A single audit reduces this burden.

For the purpose of the Horizontal Monitoring, the Dutch Tax Authority also uses the term ‘Single Audit’. This encompasses simultaneous monitoring of all relevant different tax types, such as:

  • Corporate Income Tax
  • Value Added Tax
  • Wage taxExcise
  • Customs duties
  • Etcetera

The Tax Authority previously used the term ‘integrated control’ for the same concept. Although the Tax Authority is only a single party, the single audit can still, as with other single audits, limit the audit burden.

In order to ensure sufficient depth, each ‘single audit’ must be performed by specialists. The various fiscal (financial) statements all require control and therefore the use of different tax expertise is essential. Most of these ‘single audits’ will have to be monitored by a multidisciplinary team and in addition to audit expertise, knowledge on different types of tax must incorporated.

What makes the fiscal ‘single audit’ special is that the accuracy with which the Tax Authority assesses the collection of tax returns of a taxpayer is a priori expressed in terms of tax money. That is, the financial statements to be assessed are often expressed in monetary units of the base amounts of the different taxes.

A few examples:

  • Corporate Income Tax – the General Ledger, and thus the CIT return, only includes revenues, expenditures and balance sheet accounts but lacks the applicable CIT tax rate.
  • Value Added Tax – the expenditure (investments/costs) will be posted in the General Ledger net of applicable to input tax and the VAT wis posted on a separate VAT balance sheet account
  • Wage tax: The payroll posting in the General Ledger is based gross amounts while tax audit samples are generally based on the net amount of wages.


A ‘single audit’ does not imply that the (monetary units of the) defined transactions must be seen as one indivisible population. The indivisibility can however be assumed in the case of sample drawing in which risk analysis is not taken into account.

It should be evident that certain subpopulations are created as a result of more complex interpretation and are therefore prone to a higher error probability. Other subpopulations have a lowered risk profile (e.g. due to regular, adequate internal monitoring). This needs to be taken into account with respect to the amount of work dedicated and if possible, the sample could be reduced in size.

In practice this means that the following steps should be taken:

  1. Defining the scope of the single audit
  2. Defining the required data from the systems
  3. Defining the sample size based on the parameters of the Tax Authority
  4. Drawing the non-reduced sample
  5. Determining the degree of sample size reduction based on the performed risk analysis per type of tax. Although the Tax Authority makes use of standard tables, this reduction is highly subjective.
  6. Drawing the reduced sample. This concerns drawing a sample from the non-reduced sample (step 4)
  7. Evaluating the sample; in case a reduced sample is used, the items that are not reviewed are assumed to be without errors.
  8. Evaluating the results of the sample


It is obviously always allowed to use a non-reduced sample (skip step 5 and 6) when performing an internal sample check into the effectiveness of one’s own Tax Control Framework.

In the event that the Tax Authority has performed a tax audit based on the single audit method and has imposed substantial tax assessments, it is not only important to check whether the position of the Tax Authority is tenable, but also to evaluate the process. That is, it must be determined whether this process can withstand scientific scrutiny.

In the following paragraphs the methodology of reducing the sample will be outlined. The figure below depicts the population for which from every subpopulation a proportional sample of money is drawn. The starting point is that the tolerable rate in money for each of the subpopulations is equal to that of the entire population. A subpopulation of the total collection of expenditure as such could for instance be net wages, or expenditure for which the input tax on VAT returns is deducted. Furthermore, one can choose for example every class of transactions that have similar risks.

  • First Arrow – sample size: sample size proportionally distributed over subpopulations sample size per subpopulation after incorporating “prior information”
  • Second Arrow – incomplete transactions


The dotted lined rectangle depicts the possibility that – due to whatever reason – not all transactions are included in the to be evaluated overall population. The figure above displays the situation in which the risk is equal across all subpopulations.

The figure below depicts different sample sizes per subpopulation. This results from a different amount of ‘prior information’ for each subpopulation.

  • Arrow: Sample size per subpopulation after incorporating “prior information”


Monitoring different types of tax simultaneously can lead to a different amount of prior information per type and per subpopulation. It could then happen that a sample of a very small size (very low risk) can be drawn for the Value Added Tax, while a sample of medium size (no specific risk) is possible for Wage Tax. For Corporate Income Tax a non-reduced sample might be appropriate. By means of a data analysis tool, such as ACL, subsamples can then be drawn, where a part of the will be examined for all three tax types, a part for two of these and an additional part for only one type of tax.

Literature (in Dutch):

  • Belastingdienst (2013) ControleAanpakBelastingdienst; document cab_dv4221z1fd.pdf via de site van de Belastingdienst.
  • Kerpershoek, L. (2010), Single information, single audit: de impact op stakeholders, PWC Spotlight, Jaargang 17- 2010 uitgave 4, pp 38 - 41
  • Kloosterman, H.H.W. (ed)(2003) Optimalisatie van Controlebeslissingen; project Platform Versterking Vaktechniek Belastingdienst

 

Back

Take aways 

Anticipate what users would want

We combine technical knowledge with industry understanding and knowhow of technologically advanced tools and methodologies available in the market or developed by ourselves.

What do we like to achieve

  • Focus on tax processes that could be improved
    • Manual process: same data requests are made by different stakeholders
  • As Is assessment
  • Anticipate future changes and the data needed
    • What are tax trends?
    • What is happening locally and what should be considered across jurisdictions where you operate?
    • Anticipate new stakeholders and their data needs or requests (internal and external)
  • Define scope and actions for short, mid and long term
  • Write business case for change
  • Realize sponsorship for implementation

‘As is’ assessment, actions and business case

  • What tax data is requested and by whom?
  • What tax process can be improved and what can be automated?
    • CIT, VAT, tax data warehouse
  • What is the Return on Investment?
    • Hard saving: process improvement
    • Meeting (new) tax requirement
  • What systems are in use: SAP, Oracle, etc
    • By which entities?
  • How many end-use computing tools (e.g. excel spreadsheet) do we have?
  • How do we avoid an ad-hoc solution?
    • Understand the bigger picture
    • Real problem and not the symptom

Risk and reward

Technology-related tax risk: understand and address the potential harms and benefits of (new) technology.

Technology tools & systems integration

Ascertaining proper IT support for ensuring efficient, timely and reliable reporting.

Change and project management

VAT should be considered in every aspect of the process, from concept through completion and beyond. Managing by design — looking at any process or transaction from end to end and factoring in all the requirements and controls essential to designing and optimizing a compliant VAT process.

Effective communication and teaming

We speak the language of the business and IT and no translation is needed.