Journal Issue

Misconceptions About the Value-Added Tax

International Monetary Fund. External Relations Dept.
Published Date:
January 1989
  • ShareShare
Show Summary Details

As a source of revenue, this tax has many attractions. But some aspects of it are not well understood

A value-added tax (VAT) is a sales tax, whose burden falls, in principle, only on final consumers. VAT is collected as a proportion of the value added (sales minus purchases which is equivalent to wages plus profits) at each stage of the production and distribution chain. Each trader’s tax return shows his sales and the VAT on those sales; against this tax liability he can offset the VAT he has paid on his purchases to determine his net VAT liability. In this way collection is divided over the entire cumulation of value added.

The rise of the value-added tax is an unparalleled tax phenomenon. About 46 countries now use the VAT, and each year sees new countries introducing it. VAT as a buoyant revenue source, closely linked to increases in consumption, has become a crucial part of overall revenue for all countries using it. VAT is often seen as the best way to promote neutrality and uniformity of the tax burden and to provide incentives for increased productivity and industrialization. There are still a number of controversial points about VAT; this article, however, focuses on three typical misconceptions: that VAT is expensive to administer; that it is regressive; and that it cannot be evaded.

Is VAT expensive to administer?

A VAT might be expensive to run in at least three ways: (1) it could require a large staff relative to alternative taxes; (2) it might need expensive computers; (3) or, it could be costly for traders to comply with the tax.

Staffing. The number of staff relative to registered taxpayers needed for a VAT is a function of many variables which can make the tax more or less expensive to run. These include:

• The exemption turnover limit. In many countries 60–70 percent of VAT revenue will be collected from 10–15 percent of the registered traders. If many small traders are registered, the staff needed to check the registration, check returns, audit, and chase up small traders will be employed without producing major amounts of revenue. If the exemption limit is set at a high level, as proposed in Japan, the staff can concentrate on the more important and remunerative revenue sources.

• Exempted goods and services. If important lines of goods and services are exempted from the VAT, staffing needs are reduced. If, for example, professional services are excluded from the VAT base, a difficult group to tax is avoided (admittedly at the expense of revenue, neutrality, and equity), and certainly fewer staff are needed. In many countries farmers are the largest group to be left out of the VAT. For example, in New Zealand the difference between including all farmers as registered VAT taxpayers and excluding them was 33 percent of the potential registration base. Again, in Latin America agriculturists are not taxed under the VAT but usually their major inputs (fertilizers, insecticides, seeds) are exempt from VAT.

• The treatment of affiliated companies and group trading. Returns can be simplified and the number of staff needed reduced by allowing subsidiary companies to register as a group.

• Collection by other agencies. The amount of VAT collected by other agencies affects staffing requirements. For instance, an island economy typically imports a large part of the national consumption, whether as final or intermediate goods. The VAT is collected by customs officers at the port of entry. In some countries this can represent over 60 percent of the VAT revenue, and, even in Europe, can be as high as 30 percent and it is probably the most secure portion of the VAT revenue.

Hence, the demand for non-customs VAT staff is likely to be lower.

• The use of staff transferred from work on taxes that have been replaced. The VAT usually replaces some older tax and the staff from the old tax can be retrained and transferred to implement the new tax. This can radically alter the net recruitment of staff depending on the numbers involved and their adaptability to the VAT.

• Frequency of returns. A monthly VAT requires a heavy flow of paper between taxpayers and administration; a quarterly VAT requires less paperwork. A balance must be struck between the promptness of cash flow to the government and the volume of administration. A monthly VAT may well be needed in countries where inflation is high.

• The complexity of VAT rates. The more complex the VAT the larger the staff needed to run it. A single-rate VAT requires less staff than a multi-rate VAT. Legislators, anxious to display their consciousness of equity, often introduce multiple rates and complex exemptions; they ignore the cost to the taxpaying public of the extra staff required to administer the more complicated tax. Even worse, legislators will introduce VAT complexities and not create staff positions to administer the more complicated tax properly, thus creating more inequities than they “solved” by the new legislation. It is interesting that, broadly speaking, the trends in the newer VATs (Indonesia, Republic of Korea, Mexico, New Zealand, Spain, Taiwan Province of China, and Turkey) have been toward fewer VAT rates, and those countries already using many rates (Belgium and Italy) have been trying to reduce the number.

For a comprehensive treatment of the subject, see Value-Added Tax: International Practice and Problems by Alan A. Tait, International Monetary Fund, 1988, 450 pp., $29.50. Available from Publication Services, IMF, Washington, DC 20431 USA.

• The standard of administration. A typical frequency of audit might entail visits to very large traders twice a year and visits to small traders once every four to five years. One way to economize on staff is to examine accounts less frequently; but this puts more revenue at risk (one system acknowledges its audit frequency as once every 30 years).

Typically, VAT will require more staff per unit of revenue than excise taxes, but less than income taxation. The ratio of VAT staff to registered traders has been as low as 1:31 (in an African country) and as high as a suggested ratio of 1:1000 for the United States. A more traditional ratio suggests about 1:250 but clearly there is a tradeoff between the security of revenue and economy on staffing. The cost of administering a VAT is usually under 2 percent of revenue and sometimes (e.g., New Zealand) is under 1 percent.

Computerization. Frequently people think that because the VAT is usually associated with a computerized system this necessarily implies expense. However, in recent years, the efficiency, capabilities, and capacities of computer hardware have steadily increased, while hardware costs have declined. Maintenance has also become easier. Depending on the size of the system, countries may not need to employ expensive mainframe computers with high maintenance costs. On the other hand, salaries and other costs of computer technicians have increased to a point where it is not unusual for the software, and personnel costs, to exceed or equal the hardware costs. In addition, experienced computer technicians are scarce in many developing countries and learning basic skills by education, special training, or on-the-job application usually requires several years. These skills must be updated frequently as techniques and computer systems’ capabilities are constantly changing. Moreover, in a small administration, the loss of a handful of the top computer center people can cause a major dislocation in the system’s operations.

Two recent developments have reduced these problems. First is the development and mass marketing of personal computers, which are powerful, versatile, reliable, and economical, have interchangeable components, and can even operate on battery power. They are also less sensitive to climate and are easy to use, enabling them to be operated in almost any environment. They can be programmed with fairly simple common languages.

The second breakthrough is the enormous variety of software packages tailored for the personal computer. They are convenient for both the VAT administration and the traders who must comply with VAT. A general software package, such as an accounts receivable program, can be modified to fit specific tax administration requirements for a VAT system.

In the last few years, the costs of a total computer system for a small country have fallen considerably. Depending on the complexity of the VAT system and the number of revenue offices that will issue, receive, or process VAT documents, the total computer system, including all hardware components and software packages, can be acquired for well under $100,000.

Compliance. Most businesses do not actually compute what it costs to comply with government requirements. They just know it costs money and they resent it. This is true of VAT, income tax, or social security contributions. Even before the VAT was introduced in New Zealand, the most widespread concern was the perception that it would impose heavy costs on those having to account for the tax. In the United Kingdom in 1985, VAT was considered the most serious officially-created burden on business—twice as serious as the next tax problem (pay-as-you-earn).

It is true that businessmen must allow for the cost of wages and salaries, the appropriate overhead costs of employees, computer hardware, and any other equipment and supplies to fulfill their obligations under VAT. However, it is argued that there are offsetting benefits. Smaller firms may keep better records than they would have kept without VAT. Also traders usually collect more VAT on sales than they pay on purchases; they will, therefore, enjoy a cash flow benefit that will vary according to their relative sales and purchases, their commercial credit periods, and the proportion of annual VAT payments held on average throughout the year. For example, a large retailer selling mainly for cash but enjoying, say, a three-month credit from suppliers, would do well. Others, selling on credit and getting little suppliers’ credit (typical of smaller firms) would not gain this advantage. These advantages are undoubtedly real but, in most cases, are difficult to measure in surveys. At least one survey shows that compliance costs for traders are halved when account is taken of the offsetting benefits.

Is VAT regressive?

A popular argument against VAT is that it is regressive. The evidence shows that a broad-based VAT is likely to be regressive and that making it progressive by not taxing say, basic necessities (e.g., as in Ireland and the United Kingdom) is undesirable on administrative and revenue grounds. Moreover, such use of zero-rating does not ensure that the underprivileged groups who are supposed to be helped actually are helped. As an Irish report points out, the addition to tax revenues from applying the VAT to expenditures not taxed on grounds of regressivity far exceeds the need to compensate the poor for the diminution of their household consumption due to VAT.

Given that exemptions, zero rating, and multiple rates are inefficient ways to reduce the potential regressivity of VAT, can low-income households be compensated for the VAT they pay? New Zealand used a wage supplement and changes in social security to ensure compensation to low-income families following the VAT introduction. The crucial problem remains as to whether or not compensation can reach those with small rural incomes, the poor in the urban informal sector, and the old, especially where the social security system is rudimentary or non-existent. These problems are especially important in developing countries where such groups are likely to be large and hence more difficult to reach and compensate.

However, this itself does not rule out the use of VAT. After all, it is not the regressivity of one particular tax that matters but the overall progressivity or regressivity of the entire budgetary tax and expenditure system. For instance, although the VAT may be regressive, its regressivity could be offset by efficiently run progressive direct taxes on income and wealth. This might be an appropriate argument where income taxes are efficiently administered and where the incomes and wealth of the better-off households are taxed effectively; unfortunately, in many developing countries such progressive taxes, even though apparently draconian in legislation, may be paper tigers in application. Therefore, the presumed correction to VAT’s regressivity through progressive direct taxation may not take place.

Even this is not the end of the matter. After all, if revenue is not raised through VAT it must be raised by some other tax. If the alternative tax is even more regressive than VAT, then the use of VAT may prove to be less regressive than the alternative. This is frequently the case. If VAT revenue allows less to be collected from other forms of highly discriminatory sales taxes (e.g., on tobacco, alcohol, sugar, soft drinks) which are typically consumed by low-income households, the VAT by taxing all consumption may prove to be less regressive than the highly discriminatory taxes.

So, therefore, the question is not simply whether VAT is regressive; by itself it certainly is. The more interesting question is whether its regressivity can be ameliorated by transfers, whether the tax can be designed to be less regressive, and whether the whole system is more or less regressive with or without a VAT.

Can VAT be evaded?

Most commentators theorize that the VAT contains a self-checking mechanism. As each trader’s purchases are some other trader’s sales it is theoretically possible to use the invoices showing the credit to be allowed for VAT paid as a check on the sales of other traders. Korea has tried formally to cross-match sales and purchase invoices.

However, like most other taxes, VAT can be evaded. To minimize evasion, it is essential to have a strict enforcement system with known and applied penalties culminating in criminal prosecutions in rare cases.

The most common forms of VAT evasion include:

• Traders liable to VAT but not registering. As in any tax system registration is a crucial matter. An unregistered potential taxpayer is a lost taxpayer. It is not always recognized what a substantial administrative task is involved in keeping the VAT register up to date. For instance, in the United Kingdom in one year, about 170,000 new taxpayers were registered and some 158,000 did not register; this churning of the register is a major problem and creates many opportunities for traders not to register.

• Exaggerated refund claims. Another way to evade VAT is to inflate the claims for deduction of VAT paid at earlier stages. The simplest method is to fabricate fake invoices for purchases never made. Clearly, there is limited scope for this as too many fake invoices would squeeze the taxable value to the point where the authorities would become suspicious. Further, the more fake invoices there are the greater is the risk that an audit will identify a fraudulent invoice. A more complicated, but more remunerative (and riskier), device is to claim big refunds when a new business is started. Here, the authorities would expect VAT deductions to exceed VAT liability since there would be large purchases of capital equipment and raw materials. Most authorities pay refunds quickly to avoid squeezing business liquidity; this provides the opportunity for the firm to close the business and disappear before the time for an audit. For this reason, among others, some countries will not make large refunds—especially to newly established traders—without throughly first checking the authenticity of the claim.

• Underreported sales. Understating sales is the most usual way to evade VAT. The lower the value of sales, the less VAT owed. However, if an invoice is issued, the purchaser’s claim on his VAT deduction would not be matched in an audit against the VAT paid. This danger, of course, depends on the likelihood of an audit. Again, any trader who underreports on a major scale, out of line with the average for other traders in similar categories, may be picked up by the computer as a case for audit.

The VAT at risk through this type of evasion is only the tax on the particular stage of production or distribution. Earlier activities in the production chain will have declared their sales and purchases and paid the VAT appropriate to their value added. Indeed, if the invoice for the non-reported sale is suppressed (other than at the retail stage), the VAT liability will catch up at the next stage as the purchaser will be unable to claim his VAT deduction.

Retail sales of services are particularly prone to evasion through underreporting. Any trade where taxable inputs are small and value added is high is open to evasion, particularly by small operators. A householder may be offered a lower price for gardening, plumbing, decorating, carpentry, or building, and the sale is not reported for tax purposes. The trader’s purchases of inputs are so small that he is prepared to pay the value-added tax on his raw materials, not be registered and not be liable for VAT. Such cash sales are difficult to check and such services, particularly in developing countries, should probably not form the basis for VAT liability. This type of evasion has been tackled in various countries. In Belgium, for example, service retailers such as garages, restaurants, hotels, and builders are required to issue numbered receipts giving their name, address, and VAT registration number.

• Multiple rates and incorrect descriptions. Multiple rates of VAT are the bane of VAT administration. Tax forms become more complicated as VAT at the appropriate rate has to be applied to both inputs and outputs. The chances for genuine error are enhanced and the opportunity for deliberate misclassification is widened. Clearly, traders might shift high VAT goods to standard or lower rates; equally, the refunds can be inflated by allocating an undue proportion of sales to exempt or low rate categories. Such deliberate misclassification is easiest in a small retail shop where the owner can persistently misallocate goods to a lower VAT on the cash register. In larger stores, a salesman in exactly the same way can favor friends, defrauding the revenue. Of course such evasion is not unique to VAT and can be applied in the same way in a retail sales tax. This consideration as far as VAT is concerned emphasizes the need to limit the number of tax rates and exemptions to as few as possible.

• False export claims. All VAT is rebated on exports. Therefore, a completely false export sale can generate an immediate payment by government—very much like having your hand in the government purse. The fear of false export claims is a major worry for VAT administrators, especially in developing countries. However, export invoices are usually associated with customs and shipping documents and it requires great effort to prepare all the paperwork required. Nevertheless, many countries are reluctant to grant large immediate VAT repayments to exporters until the administration has had the opportunity to examine the documents in detail.

There are many other forms of evasion which have not been mentioned, for instance, unrecorded cash purchases, credit claims for invoices from unregistered suppliers, credit notes on purchases including VAT not shown on returns, credit claims for taxable supplies used in exempt activities, credit claims for purchases that are not creditable, omission of self-deliveries, bogus traders, and barter arrangements.

It is clear that, like all taxes, VAT is open to evasion. However, experience has shown that VAT can work efficiently with evasion accounting for only 2–4 percent of revenue forgone. On the other hand, there are countries where estimates of VAT evasion have been as high as 40 percent.

This brief discussion of three aspects of VAT emphasizes that this vastly productive tax is not without its disadvantages. All taxes have their unpleasant sides. However, VAT, properly administered, on a broad base, with few rates, is often much more efficient, fair, and effective in raising resources for the government than other types ot taxes.

Other Resources Citing This Publication