Chapter

chapter 13 Control

Author(s):
Alan Tait
Published Date:
June 1988
Share
  • ShareShare
Show Summary Details

It is the … fear induced in ordinary decent trading citizens that they might inadvertently have done something wrong … or done or omitted to do something which could land them up before the Magistrate like a serf before a Tsarist Court… which is at the root of the unpopularity of this tax.

—United Kingdom, Committeeon Enforcement

Powersof the Revenue Departments

(Keith Committee), Report, p. 80

The provision of adequate controls is crucial to a VAT system. Controls are exercised through a number of means, including taxpayer registration, deregistration, invoices, refunds, and audit. These topics are dealt with in separate sections below.

Taxpayer Registration

Purpose of Registration

Registration brings a person within the control of the tax authorities. Usually this means he is required to make periodic returns of purchases and sales and, at the same time, pay the VAT. However, in some countries (for example, Belgium), all traders must register whether or not they will actually be liable for VAT. The compilation and maintenance of a comprehensive and accurate register are essential to the successful introduction and operation of a VAT. Steps toward its compilation must be taken well in advance of the start of the tax. Moreover, VAT information must be sent to all registered persons at various times to inform them of any changes in procedures and their responsibilities to register or not under the law.

The number of taxpayers is not a function of country size as Table 13-1 demonstrates. The highest percentage of taxpayers as a proportion of total population is in New Zealand, a relatively small country; two other smaller countries, Haiti and Niger, have the lowest percentage. The principal influences on the number of taxpayers (see Chapter 12) are the exemption threshold, the treatment of farmers, the treatment of groups, and the coverage of VAT. Low-exemption countries such as Denmark will tend to have more registered taxpayers as will those who tax farmers, as is the case in New Zealand and, again, in Denmark. However, the United Kingdom, even though it taxes farmers, has a lower percentage of VAT taxpayers partly because its exemption threshold is higher. Less sophisticated economies and those where the VAT does not cover the retail stage, for instance, Colombia and Peru, will have fewer taxpayers, as will those where substantial sectors such as services are untaxed. Finally, as described below, a generous willingness to allow groups of companies to register as one taxpayer reduces the number of registrations.

Table 13-1.VAT Taxpayers as a Percentage of Population Ranked by Size
Number of

VAT Taxpayers

(In thousands)
PopulationVAT Taxpayers

as a Percentage

of Population
New Zealand2823,2468.7
Italy4,60156,9458.1
Denmark3855,1017.6
Norway2954,1447.1
Korea1,16820,3575.7
Belgium5619,8535.7
France3,02755,1335.5
Austria4007,5455.3
Sweden4008,3304.8
Chile39911,9903.3
Netherlands46814,4863.2
Panama42,1803.2
Ireland1103,5603.1
Germany, Fed. Rep. of1,85761,0653.0
United Kingdom1,50056,6182.7
Argentina51430,5311.7
Uruguay503,0041.7
Peru1318,6530.7
Colombia15028,4180.5
Ecuador229,3670.2
Niger0.46,3910.01
Haiti0.55,4510.01
Sources: World Bank, The World Bank Atlas 1987 (Washington); “The Rise of VAT: U.S. Holds Out, Hungary to Join,” Financial Times (London), June 12, 1987, p. 20; Milka Casanegra de Jantscher, “Problems of Administering a Value-Added Tax in Developing Countries,” IMF Working Paper WP/86/15 (unpublished; Washington: International Monetary Fund, December 18, 1986); and various country reports.
Sources: World Bank, The World Bank Atlas 1987 (Washington); “The Rise of VAT: U.S. Holds Out, Hungary to Join,” Financial Times (London), June 12, 1987, p. 20; Milka Casanegra de Jantscher, “Problems of Administering a Value-Added Tax in Developing Countries,” IMF Working Paper WP/86/15 (unpublished; Washington: International Monetary Fund, December 18, 1986); and various country reports.

Exemption Limit for Registration

The operation of a sales tax is kept within manageable proportions, either by granting exemption to small enterprises by not requiring those with a low turnover to register, or by ignoring small tax liabilities the collection of which costs more in administrative expense than is worthwhile. A small retail taxpayer under VAT usually buys his goods for resale at tax-paid prices and all that is given up by the government through the exemption is the tax on his margin. As suggested in Chapter 6, small enterprises with annual sales under a fairly generous exemption limit should not be required to register or pay any VAT. The exemption limit should be designed to keep initial VAT registrations to manageable proportions. Through time, with unchanged exemption limits, inflation will increase the fraction of firms having to register.

Preparatory Work on Registration

The appropriate time between the date of registration and the start of the VAT depends on at least three factors. First, whether another form of sales tax already exists in the country; second, whether the coverage of the VAT is to be similar to existing sales taxes or whether it is to embrace a wider range of goods and services; and third, the likely number of VAT taxpayers. For example, for the successful introduction of a VAT entailing a change from a ring-type sales tax at manufacturers’ level to a VAT at the same level, a period of 6 to 12 months may be needed to train inspectors and to instruct businessmen on the operation of the tax and on their responsibilities to the government. However, the number of taxpayers is more or iess unchanged and the tax register should be the same; therefore, the preparatory work on the register is only to ensure it is up to date.

If, however, no form of general sales tax exists, or if the number of taxpayers is likely to be substantially increased (in Spain 400,000 previously unrecorded taxpayers were discovered), a period of at least 18 months or more may be needed. Interestingly enough, in centrally planned economies (for example, in Eastern Europe or some Asian countries) the existing demands on state enterprises and cooperatives by the ministry of finance for detailed returns for both revenue and economic planning purposes mean that the administrative changes needed for a VAT may not be as great as in more market-oriented regimes. Indeed, the greater concentration ratios of centrally planned economies with larger and fewer enterprises, and state-owned activities that include retailing, mean that the registration and accounting for VAT is easier.

To allow for flexibility, the latest date for registration need not be fixed in the law, but provision can be made for its determination by Ministerial Order at whatever might be considered an appropriate time. The intended date should be announced, however, when the draft bill is circulated and, of course, should ensure that all taxpayers are registered before the start of VAT and not, as in one developing country, leave the final date for registration three months into the operation of the new tax. Information about VAT is sent before the final date for registration to all persons already registered for whatever tax is to be replaced and, of course, checked against the register for other taxes (for example, on corporate income). There should be extensive publicity in the press, radio, and television to ensure that all likely taxpayers are aware of the registration requirements.

Unit of Business for Registration

Where a trader operates through two or more places of business, he may find it convenient to register each unit separately or, alternatively, as one entity. However, safeguards must be provided to prevent the provision being used for tax evasion by splitting up a business into two or more separate units to bring each below the limit for registration. To check such avoidance, a provision can state that if the aggregate sales of the various branches exceed a certain amount, they will all be chargeable to VAT on their sales, either as a group or individually.

Another consideration is that a group of closely related taxable persons might wish to register as a single unit and to have the business carried on by the various members of the group treated for VAT purposes as if it was carried on by one of them. Such treatment is particularly advantageous under a VAT because invoices would have to be issued only for sales transactions outside the group and VAT liability arises only on such transactions. This procedure creates administrative savings both to the taxpayers concerned and to the authorities. Such arrangements need the approval of the tax authorities, and the members of the group have to undertake to be jointly and severally liable to comply with all the statutory provisions of VAT. In some countries, all group registrations are dealt with in a special control unit or tax district that handles group registration for the whole country; this seems a sensible treatment given the specialized law and accounting issues that can be involved.

Penalty for Nonregistration

Failure to register must be treated as a serious tax offense and a significant penalty for nonregistration must be specified and imposed. The advantages of such a provision are not always appreciated in developing countries. Liability to VAT should not depend on registration; VAT is liable whether or not the trader is registered, as long as his sales exceed whatever threshold has been fixed. A trader who “forgets” to register is still liable to pay VAT. Thus, any tax for a period before registration should be recoverable in full with penalties. The penalty for nonregistration can be a percentage of the VAT unpaid increasing the longer the amount is outstanding. As such calculations can be subject to protracted dispute, the authorities can be given the option to levy a lump sum, plus a continuing accumulating penalty for each day of nonregistration.

Keeping the Register Up to Date

To ensure that the register is up to date, taxable persons should be required to inform the tax authorities of changes in circumstances affecting registration, including death. The local tax offices should review periodically against other information in the office (even the humble telephone book—especially the yellow pages—can be most useful) the lists of registered persons to ensure that new traders are brought on to the tax rolls, and particularly that those whose sales have risen beyond the threshold for treatment as small enterprises are brought into the full VAT.

Table 13-2 shows the result of an interesting exercise in Argentina. The existence of firms was established by checking many sources, including the telephone yellow pages and lists of professions and farmers. These numbers (Column 1) were compared with the actual number of taxpayers (Column 2); that implied the possible percentage (Column 3) of evaders. For example, 90 “restaurants” emerged in the sample, of which only 49 percent appeared as taxpayers (that is, approximately 50 percent evasion) or, of “ladies apparel” businesses, in which only 39 percent showed up as evaders, contrasted with “bookstores,” all of which were registered.

Table 13-2.Argentina: Characteristics of a Sample of Hard-to-Tax Groups, July 1968
Type of Business or

Profession
Total Number

in the Sample
Of which:

Number of

Taxpayers
Of which:

Non tax payers

(In percent)
Jewelry stores674139
Furniture stores755724
Dentists1149814
Physicians14311420
Beauty parlors834645
Representations746019
Restaurants904451
Insurance agents and companies584424
Civil engineers907319
Exporters211529
Importers322716
Funeral homes383313
Real estate78754
Hardware785628
Clothing manufacturers775923
Shoe stores885933
Advisers1086243
Ladies’ apparel643939
Electric appliances513825
Architects1039310
Lawyers19015916
Tourism453337
Bakers815038
Bookstores54540
Dairy shops231439
Liquor stores352227
Night clubs341750
Cattle raising14511223
Source: Adapted from Federico J. Herschel, “Tax Evasion and Its Measurement in Developing Countries,” Public Finance (The Hague), Vol. 33, No. 3 (1978), p. 254.
Source: Adapted from Federico J. Herschel, “Tax Evasion and Its Measurement in Developing Countries,” Public Finance (The Hague), Vol. 33, No. 3 (1978), p. 254.

Persons Who May Elect to Register

It is a good idea to allow a supplier of taxable goods and services, whose sales do not exceed the minimum annual exemption limit, to elect to register if he wishes to do so for trade reasons. Persons who volunteer to register are subject to the same obligations as other registered persons. The procedure for electing to become registered is identical with that for traders who are obliged to register.

A person who is not obliged to register might consider it advisable to become registered (1) if he supplies taxable goods or taxable services to registered persons—because he could then pass on a tax credit to his registered customers or clients; (2) if he exports goods—since as a registered trader, he would have no tax to pay on his export sales and could claim a credit or repayment of any tax invoiced to him on his purchases relating to the goods exported; or (3) if a VAT registration number can make his business presentation (letterhead, invoices) look more professional.

Usually a registered person is chargeable on all taxable sales even if his turnover does not exceed the exemption limit (but not in Belgium—see above).

VAT Registration Numbers

A taxpayer identification number is essential for VAT. Some countries adopt the existing tax identification system for the VAT and some use the introduction of this major new tax as the occasion for an overhaul of the identification system. Indeed, one of the principal contentious issues, that thwarted the 1987 attempt by Japan to introduce a VAT, was the obligation of traders to seek a code number designed to guard against fraudulent tax paid invoices and “other irregularities.”1 Tax identification numbers seem to be provocative; in 1987 the Australian Government was obliged to drop a similar proposal. By contrast, every Swede is given a taxpayer identification number that “the law-abiding Swedes accept without noticeable complaints [so] that even personal checks are embossed with the ID number, regardless of the fact that the number starts out with six digits revealing the birthday of the holder.2 If the income tax authorities are to administer the VAT scheme, the use of income tax code numbers for both the VAT and the income tax can be useful. Some countries use the old tax code with an additional suffix to denote the VAT. An identification numbering system common to both importers and traders is useful for cross-checking; frequently an audit trail can be established for a particular imported shipment and traced through to wholesalers and retailers. This can be particularly useful in smaller developing countries (particularly, island economies). If such codes do not exist, the VAT can be made the occasion for their introduction.

The most important key to an effective computer system is the unique identifier that can distinguish one record from another. In the case of a taxpayer master file, the identifier ensures that each taxpayer’s account contains all the relevant tax and payment data for that taxpayer alone. The best identifier is one that is unique, and permanent. It must be capable of being easily checked for transposition of numbers and other errors. It should be easy to remember as it is used frequently by others. It should not be offensive (as can be the case with three or more letters).

The identification number system shown below meets most of these conditions. Only the first eight digits are permanent and susceptible to computer check. New codes of ten digits or more (Latin American countries typically use eleven-digit codes) can be produced (see Chapter 12 for another example):

12 to 789 to 10
Type of taxpayercode (corporation, individual government agency, etc.)SequentialCheck digitBranch office code

The first eight digits are designed to be permanent numbers, the eighth digit capable of detecting error in the first seven digits using a standard formula, but modified to designate the reciprocal of the remainder as the check digit. The last two digits, which are not subject to error detection, can be used to identify the home office or main office of a taxable firm and each of its branch offices separately. Such a number will be entered by the authorities on the registration form—see Appendix I.

The other extreme is to have no identification number, as at present in Japan. This, of course, makes computerization of returns much more difficult, introduces the possibilities of duplication and error, and greatly complicates audit. Basically, a taxpayer identification number is essential.

Application Forms for Registration for VAT

Application forms should be provided free of charge, and they should be widely available not only in tax offices but also in banks and other public institutions (or perhaps sample forms can be available in such public places). A copy of the trade classification list showing the definitions of special traders (for example, hotels, boarding houses, rental accommodation, and restaurants) and the VAT Guide should be supplied along with this form. See Appendix I for a sample of a VAT registration form.

Checking VAT Forms by Local Tax Offices

The details on the initial VAT application forms must be checked during the educational visits by local tax officers. The particulars given by businessmen on their application forms for registration under VAT are of vital importance, especially the trade classifications allotted to businesses, because these may be used at a later stage to select cases for audit. Where a claim is made for treatment under the small trader scheme, the local office should satisfy itself at registration that the annual turnover is likely to be no more than the limit for the first year.

Reregistration and Deregistration

It is often not appreciated that net changes in the number of registered traders from year to year do not reflect the whole picture of the work of keeping the register up to date. Thus, in the United Kingdom, with approximately 1.2 million VAT taxpayers, as the Commissioners of Customs and Excise pointed out:

In the 12 months up to September 1977 there were 170,000 new registrations and 158,000 cancellations, and in the following 12 months there were 153,000 new registrations and 164,000 cancellations. In other words, the Department has been required each year to deal with over 300,000 movements in the VAT Register, while the total of registered traders increased by 12,000 in the first year and fell by 11,000 in the second.

Each movement in the Register gives rise to a substantial amount of work. Accurate information must be recorded about every new registration; this is essential to facilitate the smooth operation of the computer system, to ensure that any repayments of tax are made to the person properly entitled to receive them and to safeguard the position should it be found in the future that the registered person has failed to comply with his legal obligations to render returns, etc. Many registrations are followed by advisory visits to the trader’s premises. Before a registered person can be removed finally from the VAT Register care has to be taken to ensure that any outstanding liability to pay tax has been identified and discharged.3

There are really no short cuts to keeping the register up to date. However, it is recommended that where an existing legal entity wishes to reregister as a new legal entity it should be allowed to do so with the minimum of formality. Unincorporated associations should be registered in the name of the association (and act in the names of representative members, that is, treasurer, vice chairman, etc., where otherwise a change of member would require a reregistration). Further descriptions of work on registrations will be found later in this chapter.

Invoices

The invoice is the crucial control document of the usual VAT. It establishes the tax liability of the supplier and the entitlement of the purchaser to a deduction for the VAT charged. Invoices must be carefully completed and kept as records. The law and regulations of each country lay down the form of invoice, the time limit for its issue, changes in invoices, and the use of credit notes. In the EC, each member state may determine criteria for admissibility of invoices.

The invoice must normally show the following:

  • Name and address of person issuing the invoice.

  • VAT registration number.

  • Serial number of the invoice.

  • Date of issue of the invoice.

  • Date of supply of the goods or services (if different from the date of invoice).

  • A description of the goods or services.

  • Amount charged, excluding VAT.

  • Rate (including zero rate) and amount of VAT at each rate.

  • Name and address of customer.

To prevent a possible black market developing in the use of tax invoices (which are a claim on the government, tantamount to printing money and the classic way to evade VAT—see Chapter 14), the requirement to give the name and address and reference of both seller and buyer, as well as the other statutory details should be strictly enforced. In the EC, most countries do not require a VAT registration number on the invoice, on the ground that if the other particulars are tampered with it is unlikely that the number will be correct. Where the buyer is not a taxable person it may seem preferable to issue an invoice quoting the consideration inclusive of VAT. On the other hand, legislators think this procedure conceals the tax from the public and the requirement should be to show the VAT separately on all sales, including retail sales. The EC practice is to quote prices VAT inclusive for retail sales only. Of course, retail invoices can be of a more summary nature.

As long as the required particulars are included, the invoice can take any form the company wishes. Some authorities standardize the invoices for VAT, such as in Taiwan Province of China, but given the variety of business and accounting practices, it imposes a heavy burden on many traders to use only one “official” invoice, and this at a time (the introduction of the VAT) when the authorities are seeking the greatest cooperation from the business community.

As a minimum, one copy of the invoice is kept by the supplier and the other by the purchaser. Each trader retains his copy of the invoice at his office to be inspected, if required, by a VAT auditor. In Chile, three invoices are required, one each for seller and purchaser, and the third to be kept by the seller in a separate file for inspection by the VAT authorities. Korea adopted a more rigorous system whereby each general taxpayer (that is, those who are not small traders) has to complete four invoices for each sale at the time of the supply of goods and services. One copy is kept by the person or business making the sale, another copy is sent by that person to the district tax office, the third copy is kept by the purchaser, and the fourth copy is sent by the purchaser to his district tax office. The two copies of the invoice held by the appropriate tax offices are then sent to the computer data processing unit which carries out a cross-check of sales against purchases. Tax invoices issued or received were to be kept for five years. Tax invoices in which the value of a transaction exceeded W 300,000 (US$236) were computerized for audit. This was a most ambitious attempt to use the underlying theoretical structure of VAT to crosscheck purchase and sale one against the other.

Korea is the only country to have adopted this system, and Table 13-3 shows the progress in the first two years of the VAT. At first, nonmatching data was 7.2 percent of all general returns. The “learning curve” was steep in the first year as the nonmatched data fell from 7.2 percent to 4.4 percent, 3.4 percent, and after two years to 3 percent. Similarly, the number of errors detected fell from 6.4 percent to 1.8 percent. Of course, as both percentages fell, it might become more debatable as to whether it was worthwhile committing the administrative resources to completing the matching process. Indeed, the open question is whether any tax administration other than the Korean could succeed in such an effort and, even if they could, whether the necessary staff and computer resources might be better employed on other controls.

Table 13-3.Korea: VAT Nonmatching and Error Data, 1977–79
1977 (Second Half)1978 (First Half)1978 (Second Half)1979 (First Half)
TotalOutputInputTotalOutputInputTotalOutputInputTotalOutputInput
(In thousands of won)
Total52,24534,12818,11755,49035,71319,77756,64935,30421,34551,52332,33819,185
Nonmatching3,7611,5662,1952,4288971,5311,9096991,210
Error3,3632,2011,1642,2231,5087151,4301,002429954705249
(In percent)
Ratio of non matching data to total7.24.612.14.42.57.73.42.05.83.01.45.7
Ratio of error to total6.46.56.44.04.23.62.52.82.01.82.21.3
Source: Korea, Office of National Tax Administration.
Source: Korea, Office of National Tax Administration.

The Retail Invoice: Cash Registers and Other Schemes

Korea is unusual for another aspect of control associated with invoices. Usually, in other countries, traders may issue invoices for all business transactions but are not obliged to do so for transactions with nontaxable companies and individuals (except for registered persons entitled to repayment—for example, farmers). In Korea the authorities recognized that the final retail sale was the weakest link in the chain and decided to try to oblige retailers to issue invoices. Article 32 of the VAT law required traders to install cash registers that would issue receipts. The Government would advance 80 percent of the cost of buying the cash register (repayable without interest over 18 months) and the trader was entitled to deduct from his VAT bill an amount equivalent to 0.5 percent of his taxable sales; he kept the tally roll in fulfillment of his obligations under VAT law. However, the most valuable benefit was conferred by Article 83 whereby the retail customers could obtain “compensation” from the Government of 1 percent of the amount of their purchases covered by the cash register dockets. To further encourage such customers to ask for and keep their “invoices,” the Government ran lotteries based on consecutive numbers on the tally rolls. This elaborate scheme sacrificed a significant element of the VAT revenue and required increased administration for, perhaps, a rather uncertain extra yield. However, it is an interesting example of a developing country trying to ensure that the weakest link in the VAT chain, the retailer, complied with the law.

The twin pillars on which any system of VAT must rest are, first, the issue of tax invoices for all transactions between taxable persons and, second, the maintenance of an adequate system of bookkeeping. In some developing countries, the use of cash registers for retail sales and their rolls as VAT checks has become a major preoccupation of the authorities.4 (For example, Mexico is about to embark—in 1988—on a scheme to get traders to use one kind of cash register.) However, requiring the use of cash registers for retail sales is no guarantee that significant “back-door” unregistered sales are not also being carried on from the same premises. Only by auditing a trader’s books by the tax officers and the occasional unannounced check on stock-in-trade can such transactions be brought to light.

The use of a cash register is only one element in a proper system of bookkeeping and to equate it to the fulfillment of the bookkeeping obligations under VAT would be tantamount to giving a trader a license to evade tax. In Korea there was a strong belief that the extensive use of cash registers at the retail level could be of great value in introducing retailers to obligations of running a modern tax system like VAT and this may be true for other developing countries. But the scheme as planned was costly in terms of money and administrative overheads. If a government allows small retailers to write off the cost of buying a new cash register in a year, that should be enough compensation. No EC country gives a tax reduction to traders for using cash registers or to customers for the submission of cash register dockets. Governments get more revenue from using the VAT staff to audit the larger cases where the discovery of fraud often results in substantial payments of tax.

The Turkish authorities tried a different device to encourage households to ask for receipts. Wage earners were allowed to offset a proportion of the VAT paid on their retail purchases against their income tax liability. This was meant to offset some of the regressivity of VAT but, at the same time, households could only claim the credit for VAT paid on production of invoices showing the purchase and this, it is claimed, stimulates households into demanding invoices for their purchases. Of course, it also opens the door to further evasion through false invoices, and it erodes the income tax base and creates uncertainty about the yield of both VAT and income tax.

The most prevalent type of evasion under any sales tax that includes the retail stage is the underdeclaration of sales to final consumers, and it is understandable that this should cause concern. There is, however, much less ground for worry under a VAT than under a retail sales tax, because under VAT reliable figures of purchases can be established, and, by applying computer tests, the cases where the figures of sales appear to be underreported can be picked out and investigated.

Payments and Refunds

The VAT is basically a self-assessed tax. As such, registered traders are expected to submit their payment with their completed assessment forms. Indeed, in some countries such as Indonesia no VAT return is accepted if it is not accompanied by payment, unless, of course, a refund is claimed. This eliminates the need to match returns and payments later. Returns and payments can be processed both centrally and regionally. In some countries, including Denmark, the government GIRO (post office banking system) system is used; in others (for example, Argentina, Bolivia, Brazil, Chile, Colombia, Peru, and the Philippines) the commercial banks collect taxes on behalf of the authorities, and are willing to do so, providing they get four to ten days’ use of the money. Using banks to collect VAT can be particularly efficient where local collectors may not be fully reliable and where the postal system is suspect. Authorized agent banks can be required to sort out returns into three groups: those where the amount paid is the same as that shown in the return; a second group where the amount paid is different to that in the return; and a group where credit exceeds tax liability. Banks stamp all copies of the returns showing the date of receipt and validate receipt of the VAT. The tax return serves as evidence of the tax paid.

In Argentina, Spain, and Hungary, the authorities supply a booklet with slips containing data printed in magnetic characters relating to the registered taxpayer. This slip is used by the trader to identify his tax return. This procedure is claimed to prevent taxpayer errors in transcription and shortens the administrative time to record the taxpayer’s information.

Returns are frequently filed monthly, as is the practice in Brazil, Chile, Costa Rica, Ecuador, Honduras, Mexico, Nicaragua, Panama, and Uruguay, but many countries use a two-month or quarterly system. Argentina and Mexico are among the few countries that have adopted annual filing (with provisional monthly payments), but, as mentioned in Chapter 12, this has not proved particularly useful.

One of the features of the VAT is the possibility of generating balances in favor of the taxpayer through the credit mechanism. Indeed, in one new European VAT, monitoring refunds absorbed almost 70 percent of the initial control effort. If the balances are clearly of a temporary nature, the problem is easy to solve; they will be offset by debit balances in a relatively short period of time. There are, however, situations in which the credit balances not only are created but once created keep growing. A VAT is intended to be a tax on consumption and not a tax on business; in principle, all excess of input tax over output tax should be refunded at once. Failure to do so alters the nature of VAT by making it, in part, a tax on business. The tax law must provide a solution for these cases. These fall into five main categories, which are described below.

Balances Resulting from Export Operations

Since exporters are normally allowed to recover the tax paid on their purchases and there is no tax on their export sales, it is easy to see that they will tend to accumulate credit balances. Application of the destination principle, therefore, requires that taxes paid by exporters be refunded. Ideally, any excess should be refunded promptly. Alternatively, it has been suggested that exporters could be allowed to use their balances against other tax liabilities, for example, income tax, as in the Dominican Republic. Of course, liabilities should always be enough to absorb the balances without undue delays. However, it is not a good principle to mix the liability and collection of two quite separate taxes. This can be especially true in small developing countries where the refunds claimed can be concentrated in very few companies (for example, in Panama, one third of the taxpayers claim credit and pay no tax and of these taxpayers just two account for most of the tax credit).5

Sometimes exporters are allowed to claim their VAT on imports in anticipation of the exports. For instance, under the Chilean “second method,”6 the exporter makes a sworn return specifying by when the export will be made and estimating the amount. Once approved, the refund is made within five days. The exporter must provide a notarized guarantee represented by a bill of exchange or promissary note. If the exporter cannot demonstrate that the export has been carried out within the period promised, he must reimburse the amount, adjusted for inflation, plus 1 percent a month interest.

Balances Resulting from the Use of Multiple Rates

This situation can occur when the products sold are subject to a rate that is lower than that applicable to inputs used in the manufacturing process. In practice, this would be unusual as rate schedules are set up to avoid it happening. The solution should be the same as in the preceding case, that is, credit balances should be refunded. If a single rate is not feasible for political or other reasons, as the EC suggests, the lower rate should be set at a level that is high enough ordinarily to allow complete absorption of the credit originated by tax paid on inputs.

Balances Resulting from Zero-Rated Goods

This case is similar to the preceding one and it arises when taxable inputs are used to produce zero-rated goods. But, as discussed in Chapter 12, administrative considerations may make exemption preferable to zero rating; if so, the taxpayer will not be able to recover the tax paid on purchases and will have to treat it as an additional cost element.

Balances Due to Purchases of Equipment

Large purchases of fixed assets by taxpayers may also give rise to credit balances and their volume may be such that it would take a long time to accumulate the debit balances to allow complete offsetting. During all this time, the tax paid on the purchase of the asset becomes a financial burden for the taxpayer that, if severe, may strain his liquidity. This may be particularly important in the case of new firms that are confronted by substantial investment in equipment. It follows that, as for balances from exports, quick refunds are the best solution. It should be noted that this requires most careful administrative checking and inspection. Also in this case, another possibility would be to use the balances to pay other taxes. In Chile, if credit for fixed assets is accumulated for six consecutive tax periods, it can be offset against other tax liabilities, including import duties. Again, new firms frequently make little profit and hence incur no liability to corporate taxation and, in any case, two tax structures should not be mingled in this way. Such credits in Argentina must be spread over three years and can be claimed only annually.

Balances Resulting from Temporary Conditions

There are a number of situations that may give rise to temporary credit balances, for example, substantial increase of inventories or seasonally slumping sales. By their very nature, these credit balances will tend to disappear quite quickly. Any credit should be settled within the normal tax period. Some countries, however, allow the taxpayer to offset such credits only against future debit balances (that is, only export credits are paid at once). Others (for example, Belgium) allow credit to accumulate for three months before they are paid. Spain allows the credit to be claimed at the end of the year.

In discussing the treatment of credit balances, mention should be made of the so-called buffer rule. It originated in France (règle de butoir) and has been adopted, with changes, by some Latin American countries, such as Argentina and Uruguay (regla del tope). Basically, the French approach meant (it was abandoned in 1972) that the tax paid on purchases could be credited against the tax due on sales only up to the latter’s amount and that any excess over the tax on sales was not creditable. The Argentine law went further and established that the creditable balance was limited to the amount that would result from applying to the taxpayer’s purchases the rate that would be applicable to his sales.

This is an undesirable feature that should be avoided. It could distort business organization, leading firms, for instance, to merge to use credits. Ideally, the total tax burden on any particular product should be equal to the rate applicable at the point of final sale to the consumer, and the only way of achieving this result is to do away with these artificial ceilings on credit balances, whenever possible. These difficulties are another argument to minimize the use of exemptions and to avoid multiple rates.

The EC’s Sixth Directive leaves the decision on the means of dealing with excess deductions to the discretion of the country: “Member States may either make a refund or carry the excess forward to the following period according to conditions which they shall determine.”7 In practice, the treatments vary but most countries repay any excess credit within a month.

The Sixth Directive also permits member countries to ignore cases where the excess for a given taxable period is insignificant. With the increasing use of computers, most countries find it more satisfactory for the tax administration and for taxable persons to settle the account for each taxable period as it arises rather than to carry a balance forward from one taxable period to the next. In general, European countries provide for repayment of VAT within about three weeks from the end of the taxable period; if a firm is in a continuous refund position, returns are accepted more frequently than usual and refunds made promptly. In all countries abnormal claims are subjected to prior verification.

However, in many developing countries, the refund is perceived as the major potential weakness of the VAT and the main avenue of evasion; therefore, the authorities are reluctant to be too prompt on refunds, preferring to have time to inspect, in detail, large refund claims. This is a prudent requirement in any VAT system.

Audit

Relationship to Income Tax Audit

Efforts should be made to coordinate VAT audit with the internal and field audits carried out for the income tax. The relationship and exchange of information between the VAT and the income tax is discussed in Chapter 12, but it is assumed that the legislation allows a close relationship. Audit for VAT is usually undertaken independently of income tax audits because the work is less comprehensive and can be done, in some instances, by less technically experienced staff. Some countries, including the Netherlands, consider VAT audit to be as comprehensive as the corporate income tax audit. In these cases, the size and nature of the firm determines the staff to be responsible for the audit. In any case, recent income tax audit reports should be available for examination by the VAT auditor beforehand. Equally, VAT audit reports should provide a useful input for income tax audits.

The universal method for the control of VAT and the prevention of tax evasion is the auditing and spot checking of records, coupled with a system of adequate penalties for detected cases of fraud. Only by an effective audit program can instances of tax evasion be brought to light. Therefore, the central tax administration must have one arm for the direction and control of the audit programs and another for the investigation of cases of suspected fraud. The central audit management division and the investigation division may well deal with income tax and VAT as a joint operation.

Often staff to administer VAT are transferred from other departments or tax offices (for example, staff with experience in manufacturers sales taxes, excise, and wholesale taxes) that have relied more on physical control than on financial appraisal. Before the VAT is introduced a substantial amount of time has to be allowed to train the new staff in VAT administration, including simple financial audit. The training for VAT audit can be used as a staging post on the way to fulfilling the revenue department’s qualifications to be a full income tax auditor.

Physical Checks

There is still a place under VAT for physical checks on warehousing and movements of goods. In the opinion of many “hands on” VAT administrators, physical checks are crucial to effective VAT control. Physical checks establish undeniable facts; paper and records can conceal facts. A VAT fundamentally must tax a fact, the delivery or supply of a good or a service; it should not tax just an invoice. Others claim that physical checks are out of place in a financial accounts based tax. This debate can become quite fractious. There is merit in both views. Most VAT is collected and audited on documentary evidence, but the physical check and the threat of physical checks continue to be important. Truck drivers can be required to carry consecutively numbered manifests corresponding to an invoice giving particulars of what is carried and the names and addresses, with registered number, of the buyer and seller. In Europe, some of the more blatant (and successful) schemes for VAT evasion have involved illegal movement of goods across international boundaries, and physical checks are useful in this context. These physical checks on merchandise and of the movement of goods ought to be considered as additional tools of enforcement. They are usually carried out by staff operating under the immediate control of the district and regional offices and can be especially useful in developing countries in checking on suspected VAT fraud.

Definition of Audit

There must be a common understanding of the definition of an “audit.” A VAT control visit does not necessarily constitute an audit (see Chapter 12). Control visits are meant to ensure, primarily, that the statutory provisions for VAT are being complied with; that, in cases picked at random, the records of purchases and sales are correctly completed from original documents; that monthly returns for selected months correspond with the totals in the records of purchases and sales; and that the totals under the different columns for the month are correct and the records accurately reflect the business operations. When the control officer detects what he thinks may be evidence of fraud, he merely notes the facts. He does not attempt an investigation but makes a report so that a full audit can be made.

An audit for VAT purposes differs from an audit by a chartered accountant. A VAT audit is a selective review of the taxpayer’s books and records, including year-end statements, balance sheets, and profit and loss accounts, to ensure that the major areas of purchases, sales, and stocks are substantially correct, not that the taxpayer has filed a perfect return. An error can be found pertaining to almost every return, but frequently its significance is not really material, particularly when viewed in terms of auditing and processing costs. The usual audit is limited to the significant issues of the current year. If notable changes to the current returns are made, the audit should be linked also to those same issues in the two preceding years, plus any other issues in those years that are apparent on the face of the return. Only in unusual cases should the audit be extended beyond the two preceding years, even though there may be, for instance, a five-year statute.

The purpose of the VAT audit is to find out, according to the law and regulations, the actual VAT liability of the taxpayer. This generates both revenue and voluntary compliance. Voluntary compliance is generally enhanced more by increasing the number of taxpayers audited than by auditing fewer taxpayers with large tax potential. Perversely, revenue is enhanced by concentrating on those taxpayers where a large amount of revenue is at risk. These opposing concerns will always pull audit teams in opposite directions.

The audit report should also include a recommendation as to whether frequent audits should continue. The taxpayer’s books and records should be evaluated, and if found deficient, the taxpayer should be advised that underreporting coupled with inadequate books and records are prima facie evidence of an intent to evade taxes and, therefore, subject to severe civil or criminal penalties or both. In such a case, returns should be earmarked for potential audit, and if inadequate records persist, recommendations may be made for severe penalties or prosecution or both.

Type of Audit

There are two main types of audits: field audit and office audit.8 A field audit is conducted at the taxpayer’s place of business. It frequently involves third party contacts with suppliers or other government entities to obtain information to validate the accuracy of the books and records, or to reconstruct net income. A field audit can require, in addition to a thorough knowledge of the tax law, the application of accounting principles, and auditing knowledge and techniques. A field audit can include an audit of associates such as parent and subsidiary corporations.

An office audit for VAT does not require an extensive knowledge of accounting or auditing principles. It is generally limited to the checking of returns for arithmetical correctness and general conformity with legal requirements. It can include the verification of exemptions, claims for credit, and checking for self consumption, etc. Audits can also be described according to the function. For instance, in Denmark an “accounts analytic audit” is an introductory survey of the facts; a “system analytic audit” is for large companies and identifies weak and strong areas; and a “transactions audit” is more time consuming and investigates in depth areas of weakness.

Audit Management Activities

Audit management activities, which management should design, organize, monitor, evaluate, adopt, correct, and approve, include the following:

  • Search for unregistered taxpayers.

  • Identifying the information needed by the taxpayers about VAT.

  • Establishment of criteria to select audit cases.

  • Selection of cases.

  • Collection of information about trader practices.

  • Preparation of VAT audits.

  • Performance of office and field audits.

  • Reporting on audits.

  • Discussion and decisions about the outcome of audits.

  • Processing of results for use in other audits and for management decision making.

  • The quantity and quality of audit work.

  • Training needs for audit.

  • Staffing requirements.

The management responsibility for each of these activities can be explained by considering, for example, the first activity mentioned above of a “search for unregistered taxpayers” in terms of definition of problem, the criteria needed, the options, and the possible solutions.

The problem in this case is that there are persons and companies that according to the law are taxpayers but are not yet registered. The criteria involve testing the possible solutions for their cost effectiveness, that is, the cost of the extra audit against the possible revenue obtained.

Possible operations could include the following:

  • Creating a publicity campaign using radio, television, the press, and videotapes.

  • Mounting an information campaign through locally organized meetings to inform potential new taxpayers.

  • Seeking the collaboration of organizations (trade, professional, journals) to inform members of their obligations to register and possible penalties for not doing so.

  • Using readily available information sources (for example, telephone yellow pages, business guides, local newsletters, publications of professional organizations, newspaper advertisements, and local tax officials’ observations).

  • Making checks on unregulated retailers to ascertain the sources of their stocks.

  • Cross-checking with other tax sources, especially customs clearances in economies that rely heavily on imported raw materials and finished goods; checking excise returns, property tax registrations, and even business truck registrations.

Probably the first three operations have already been tried, so that attention would be concentrated on the last three. Having selected the thrust of the activity, management must assess the resources to be used and then organize the activity. Although it sounds obvious, experience in many countries suggests any organization of work should be kept as simple and cheap as possible. Paperwork should be reduced to the minimum commensurate with what is needed to monitor the activities.

Monitoring should try to establish effectiveness without fuss. A big difficulty, experienced in many businesses (even in international organizations), is that people tend to report only positive things to their superior (or things they think their superior wants to hear). To anticipate this, the VAT manager needs to devise simple statistics on a weekly or monthly basis to keep himself informed of what is going on and, of course, to make it clear that fake reporting will not be tolerated. At the same time, the VAT manager should ensure that all VAT officials involved are kept informed of the overall situation so that the information flows both up and down. Some suggestions are as follows:

  • Regional directors conduct a sample check in their regions.

  • Headquarters conducts some direct regional sample checks.

  • Results from different regions are compared and large differences must be explained.

  • All regional results are reported to every other region.

Once the activity is under way and monitored, it needs to be evaluated and, if necessary, corrected and improved. However, the core of audit work is actual office and field audit, and reporting on audits, as mentioned above.

Outline Procedures for Audit of VAT at the District Level

In general, most tax administrations stress the uneasy relationship between the tax officer with obligations to secure revenue and traders with businesses to run. Traders cannot be expected to welcome VAT control visits; equally, the demands of most revenue services on traders are infrequent and reasonable: “Officers will often have to pit their wits against the keen business minds of traders who are, as it were, playing on their own ground. They should accept this as a challenge and strive to maintain the Department’s reputation for courtesy, commonsense, shrewdness and fairness.9

Preliminary Inquiries

Before visiting a taxpayer, the auditor should:

  • Check that filing of periodic returns and payments are up to date.

  • Examine the annual summaries of monthly (or quarterly) sales and compare them with the returns for income tax.

  • Check the most recent income tax audit report available before making the audit.

  • Examine the file for any peculiarities or uncompleted matters from previous visits.

General Approach

At the beginning of the audit the auditor reviews the business activities in a general way with the proprietor or with a responsible officer of the company to gain an overall picture of operations—the main sources of supply, whether goods were paid for in cash or by check, the number of employees, amounts withdrawn for private use, how cash receipts were dealt with, and so on. This should help to pinpoint the areas that might require special attention.

The First Vat Audit

The duties of a VAT auditor (always remembering the overriding requirement to be fair and courteous) on the first audit to a taxpayer could be summarized as:

  • Ensuring that the taxpayer is in no doubt about his legal obligations and the procedures he is expected to follow.

  • Analyzing and recording in the file the bookkeeping procedures and records. He should be prepared to educate the taxpayer in the best way to carry out his obligations under VAT.

  • Arranging for payment of tax where underpayments have been discovered and to ensure the correct operation of the system by the taxpayer in future.

  • Considering any impact on income tax that comes to light.

The training of the junior auditors should be directed toward exercising sound judgment and discretion in handling any features which seem to point to irregularities. In particular, they should not seek to investigate fraud but should obtain enough information to enable their office to decide whether the case is one calling for full investigation.

While VAT provides certain built-in checks in that the output items of taxable sellers constitute the input items of their taxable customers, if the cross-checking of such items does not bring any discrepancy to light, it does not mean that everything is satisfactory, as items might be omitted by both the seller and the buyer.

The nature of the taxpayer’s business can suggest possibilities for special scrutiny. Confidence in checking is acquired by junior auditors, as part of on-the-job training, by examining in the first instance 20 or 30 cases that are believed to be operating satisfactory bookkeeping systems. While examining the accounts in those cases, the junior auditors familiarize themselves with any features which might be new to them.

Subsequent Audits

At subsequent audits, the following additional procedures should receive attention:

  • A particular month is selected and the figures of purchases and sales on the monthly returns are checked against the original documents.

  • The totals for purchases and sales are compared with the bank statements.

  • Whatever records that are kept for goods withdrawn for private use are examined, and inquiries made to verify their reliability. If there are no records, agreed estimates are made.

  • A search is made among the invoices of purchases for any items that might be destined for private use. This search might be extended to invoices outside the particular month selected. The purpose of the search is to ensure that a deduction was not claimed for tax on such expenditure and that it has been excluded from the business “expenses” for income tax purposes. Any unusual items appearing in the record of purchases are investigated to see to what use they were put.

  • The accounting system is examined generally, with particular reference to the calculation of the figures for taxable sales and any changes in the system as recorded in the file being noted.

  • Special attention is paid to the treatment of payments and receipts in cash.

  • Records of payments are checked against receipts issued as well as invoices received.

  • The most recent inventory is examined, particularly items related to the records of purchases and sales.

  • In a few selected cases the last two inventories can be examined and a reconciliation sought with the figures returned for purchase and sales in the intervening period.

  • Checks are made for claims for tax credits based on invoices in the names of other firms (that is, multiple claims associated typically with groups of companies).

  • Checks are made that sales to tourists are genuinely exported. See Chapter 16 for a further discussion of the items officials may look for from the trader’s point of view.

In the beginning of the U.K. VAT, traders were to be visited, on average, once every three years. As trader experience with compliance grew (and there was some pressure to reduce staff), this frequency fell to one in four. (By comparison, income tax audits can be as infrequent as one every ten years.) Audits for the U.S. state retail sales taxes are strikingly infrequent, from below 1 percent of accounts audited annually, up to a maximum of 8.8 percent (in Mississippi).10 In one Latin American country, the probability of audit is at most once every hundred years, but in Chile it is once every seven. Clearly, the greater the likelihood of a visit, the better compliance should be. It is also true that in all countries, commissioners of tax can show that the employment of an additional inspector will easily pay for his salary through the increased tax collected. Yet, no legislature will authorize increases in numbers of tax inspectors up to the level where the marginal inspector’s salary just equals his additional collections. Any decision on the inspectorate is a compromise. An average of about once every four years would seem about as infrequent as is compatible with keeping a grip on the tax. Of course, some taxpayers might be visited very infrequently and others much more so.

Number of Audits

Auditing capacities of field offices are determined by calculating the number of returns that can be audited, using estimates of the number of returns each field and office auditor can reasonably be expected to complete during one year. There should be an overselection of 25 percent to 30 percent of audit capacity. With proper pre-audit planning, strong supervision, and effective use of the review staff as a training vehicle, it is not unreasonable to expect an office auditor to complete an average of 10 to 20 cases a month, and for a field auditor (exclusive of auditors dealing with controlled cases) to complete an average of 4 to 6 cases a month.

Table 13-4 shows a typical presentation of data on central visits. Part of the assessment of the required number of visits is clearly the effectiveness based on the net receipts and the staff effort involved.

Table 13-4.United Kingdom: VAT Control Visits, Year Ended March 31
198019811982
Number of control visits (in thousands)368357358
Number of underdeclarations (in thousands)144142133
Value of underdeclarations (in millions of pounds)84146160
Average value of each underdeclaration (in pounds)5831,0281,203
Number of overdeclarations (in thousands)13118
Value of overdeclarations (in millions of pounds)4105
Average value of each overdeclaration (in pounds)308909625
Approximate staff effort in control visits (in thousands of pounds)4,2004,1004,100
Net receipts (in thousands of pounds)8,18910,96711,856
Net receipts per visit (in pounds)1,9492,6742,892
Source: Adapted from United Kingdom, Committee on Enforcement Powers of the Revenue Departments, Report, Vol. 1, Cmnd. 8822 (London: H. M. Stationery Office, 1983), p. 74, Table 2.
Source: Adapted from United Kingdom, Committee on Enforcement Powers of the Revenue Departments, Report, Vol. 1, Cmnd. 8822 (London: H. M. Stationery Office, 1983), p. 74, Table 2.

In a tax department operating in an environment where there is a widespread belief that most tax returns are underreported, careful planning is needed, otherwise there may be an overselection of returns for audit, with the result that the inventory of work in process will increase to unmanageable proportions and make an orderly approach to an annual audit work plan impossible. The first step in a monthly work plan is to devise a system to control the work load and to accumulate data to refine the audit plan. The results of the audit and the characteristics of the return before and after audit must also be checked. Computers can provide control for each return selected for audit, and for accumulating the data on audit results.

The frequency of audit is a contentious issue. The judgment is always a delicate balance between the threat of audit to check the temptation to evade and the cost. Some countries are prepared to envisage a high frequency of audit (for example, Indonesia) but usually, where the VAT system is fairly well established, audits of 15 percent to 20 percent of registered traders a year are sufficient. In the Federal Republic of Germany, small traders may be audited by the tax authorities only once in every 8 to 10 years.

However, it is not only the crude number of audits that is the most useful measure of need. The form of audit is important. Often too much time is spent on each audit visit; in some countries, five days on average has been noted, whereas this should be reduced to one to two days with an average of one and a quarter days. Again, for each trader, the checks appropriate to his business should be set out and applied selectively on each visit. Also the number of individual inquiries can be reduced by ensuring that the trader’s file contains cross-references from audits of other firms as they are received, which are dealt with in bulk on the next inspection visit.

An allocation of about 200 cases a year should be reasonable for an experienced junior auditor, who should be eligible for advancement to senior auditor after fours years of satisfactory service in that grade. Usually, where serious irregularities are discovered, or where fraud is suspected, the case, in coordination with the income tax auditor, should be reported for attention by the regional office.

In Panama, a “preventive audit” has proved effective in detecting and preventing violations. This is a spot check control over inventories designed for the time of the year when sales activities increase, as during Easter, Carnival, and Christmas; sales by the month are taken and combined with a review of inventories of seller and purchaser companies.

Refunds of VAT on exports must be closely monitored. Auditing of the larger manufacturers is carried out by senior auditors under the control of the regional offices. Audit visits in such cases is undertaken yearly, if possible. Manufacturers not sufficiently large for audit of that nature are dealt with by the local tax office, where the more straightforward cases are handled by junior auditors and the more complex cases by senior auditors.

Audit Selection Criteria

There are several sources from which the audit work load is derived. Returns will be drawn for audit from these sources by either selecting returns through a special process, or selecting returns through a routine process, based on developed or established criteria.

Special Process

This category could include the following:

  • Controlled cases as defined.

  • Sectoral audits made to establish guidelines to be used for reference purposes during audits.

  • Cases that must be audited based on policy decisions. These should be kept to a minimum. A most common case of this type is one including a large refund. Others might be repeated claims for refunds or “nil” returns. Also included would be traders who repeatedly fail to file returns or those whose category of trade suggests a higher rate of VAT than that returned. Small builders and restaurants fall into this category.

  • Cases on which information is received from external sources, often a dissatisfied competitor. These should not be automatically audited; they should be reviewed, and only those that appear to contain substantial error should be audited.

  • Cases on which information was developed through the audit of another taxpayer. These should also be screened to select only those involving a substantial omission.

  • Returns audited as a result of a project to determine compliance within a specific profession or industry, such as advertising and contractors.

The investigation to assess compliance can be more general. In Colombia in 1983, the Government organized a special audit of some 7,000 commercial traders. Those audited produced revenue more than 5 percent higher than the comparator group not audited. No penalties were levied, but advice on compliance, and a warning, was given. Subsequently, one fifth of those audited were visited again to see whether the recommendations were adopted. Four fifths of the sample had adopted at least some of the recommendations and the remaining 20 percent were subject to a “punitive audit” and the full application of penalties.11

An interesting special method was found useful in Argentina when government purchases were used as a starting point for audit. It was found that up to 50 percent of some traders’ sales to the Government had not been reported and, of course, this resulted in substantial assessments for underpayments.

Routine Process

The balance of the work load should be selected through a regular classification procedure.

Regular classification consists of selecting returns in one of two ways, by audit ratios or by computer selection. The audit plan should calculate the number of returns to be audited from each audit class. From this, a ratio of returns to be audited to total returns filed by class can be determined. Generally, a higher ratio of returns should be selected from the higher income and higher turnover groups. The ratio should be used to determine the number of returns to be reviewed to select those for audit. All returns are, of course, not reviewed. A balanced plan might assume the following:

  • For large returns, a selection is made of 30 percent to 40 percent of the returns reviewed.

  • For medium returns, a selection is made of 10 percent to 20 percent of returns reviewed.

  • For small returns, a selection is made of 1 percent to 5 percent of returns reviewed.

  • A small percentage, say, 1 percent, to be selected purely at random.

A balance must be maintained between returns selected through the special classification and returns selected through regular classification. In some countries, the number of returns that can be selected through special classification—basically informants—exceed the entire annual work load. To actually select them would congest the audit plan and result in a deterioration of compliance. Theoretically, the ideal balance seems to be for some 40 percent to be selected through special classification and the rest through regular classification, although conditions vary greatly among countries. Based on the above, a projection of the number of returns that should be reviewed can be made.

Strict monitoring is required since it is difficult to refrain from reviewing an excessive number of returns, especially when there is a belief that most taxpayers are underreporting. Over a period of time serious underreporting will, however, be identified through the random qualities of the process.

The auditors assigned to the task of reviewing the returns should be experienced in auditing and well versed in the law. They should recognize the importance of the duty, be able to read, analyze, and interpret financial statements, and adapt to the confining task of sitting at a desk, reviewing return after return. A good classifier should be able to review some 75 to 100 of the larger returns a day. It should be noted that the classifier is not reviewing the return for the purpose of making an assessment; he is exercising judgment in determining returns offering the best potential for audit.

A checksheet should be developed that lists the reasons (by line item) why the return was selected for audit. The checksheet should accompany the return through the audit process, and, at the end of the audit, should be used to evaluate the effectiveness of the classification program and the effectiveness of the individual classifier. Check-sheets can also be used to generate statistics on noncompliance. For a representative sample of a particular class of taxpayer, the ratio of unaudited tax liability to the correct tax after audit gives the level of compliance. Staff can concentrate on areas where compliance is low.

The classifiers should be given training on their special responsibilities. They should learn the process of developing an audit plan and the process of analysis required to determine whether a taxpayer should be selected for audit, as well as the criteria used in selecting a taxpayer for examination. These criteria can best be developed by conveying a group of experienced auditors to adapt criteria to the circumstances of each country.

Computer Selection

Within each tax district a certain number of tax officers are available for field audit work. To direct those field auditors into the most productive inquiries a computer-based case selection system can be invaluable.

Computer selection for sales tax audit has been used for many years.12 Interestingly enough, computer selection for the relatively simple retail sales taxes of the United States has not been an unqualified success. A major problem has been the growth of multiproduct stores so that the norms are more difficult to establish and, even if established, their validity does not hold for long. However, more sophisticated analysis has produced estimated dollar returns from audits or a ranking of firms on the basis of potential audit productivity.13 Nevertheless, experienced auditors still express strong reservations about the usefulness of computer-generated selections, especially in developing countries.

To help select those taxpayers who appear to warrant auditing more than others, the information already stored within the computer on each trader’s tax and trade performance over a period of 12 months should be compared with performance of his competitors in the same class of business. Those taxpayers who are showing much lower-than-average gross profit margins should be the subject of early audit.

The output and input tax and the sales purchases for a current year should be calculated and used to calculate for each taxpayer: (1) the markup (output minus input as a percent of inputs) and (2) annual sales growth (current sales less previous year sales as a percent of the previous year sales). It should be borne in mind, however, that opening and closing stocks may fluctuate and therefore affect the apparent markup.

Each taxpayer’s figures should be calculated individually and should be used to calculate the average markup and sales growth for the business classification as a whole. The comparative performance data should be calculated separately for payment and refund taxpayers.

There should also be a separate performance ratio calculated to assess the comparative growth of sales of exempt goods and services over successive tax periods. The calculation should be total exempt sales as a percent of total sales.

Each general taxpayer should be assigned an indicator within the computer relating to performance over a two-year period. The indicator values should take into account the following:

  • Markup over the period compared with other taxpayers in the same business classification.

  • Returns performance—if any returns have been submitted late or have had to be corrected.

  • Tax performance—whether or not the tax has been paid promptly.

The range of values might go from “good” (one) to “bad” (four) in four stages. It would be set by reference to statistics extracted from the computer record or by reference to the local tax officer or by both.

In any selection for audit, the potential revenue risk should be assessed. For taxpayers with a high tax base the risk is generally greater than one with a low tax base if other indicators are the same. Accordingly, selections for audit should be biased toward the larger taxpayers. Naturally, the computer selected returns for audit still need a senior auditor to scan them to make the final decision about which traders should be audited.

The more traditional tax administrators who view computer selection with a jaundiced eye maintain that plain random selection generates the same results with less effort and less administrative cost. A problem with the so-called random selection is that too often it can degenerate into personal selection and, in bad examples, into bias and personal vendetta. Some rules have to be made and kept whether or not they are computer generated.

Audit Manuals

The preparation of an audit manual is a complex job and is particular to each country although, of course, many of the procedures (as indicated above) are common. What needs to be emphasized is not only the importance of creating a detailed manual but of insisting that officers use it and continuously review and update it.

In many countries, it is sad to report, tax manuals are written and used when a tax is introduced or when it is radically overhauled. Within a few years, officers refer to them infrequently, often ignore the recommended procedures, and indeed, sometimes cannot even find a copy of the manual in their office or in the offices of the staff they are supervising. Manuals are meant to be used and their procedures followed. To the extent they are not, it usually indicates the erosion of the tax base.

It follows that one way to keep attention focused on manual procedures is to have statistical records on the procedures the manuals recommend and to issue regular updates to the manuals that, on inspection, are expected to be included in each officer’s copy.

“Japan: Sales Tax,” World Tax Report (1987, p. 14).

Mutén (1988, pp. 2—3).

United Kingdom, Commissioners of H.M. Customs and Excise (1978, p. 16).

See the list of problems with cash registers in Panama in Porras (1979).

Porras (1979).

See “Chile,” in International Bureau of Fiscal Documentation (1986, pp. 51—52).

European Community, Sixth Council Directive, Article 18(4).

Some would classify audits as (1) examinations, that is, the administrative check of returns at the district office; (2) verifications, which basically are field audits; and (3) in-depth audits, usually where a suspicion of fraud exists.

United Kingdom, Committee on Enforcement Powers of the Revenue Departments (1983, p. 686).

Due (1983b. Table 9.5, p. 232).

Perry and Orozco de Triana (1986, p. 43).

See an interesting account by Due (1985).

Due (1985, p. 238).

    Other Resources Citing This Publication