chapter 12 Administration and Staffing
- Alan Tait
- Published Date:
- June 1988
The cost of administering such a simple tax has proved to be less than 1 per cent of the tax revenue. We estimated there would be 180,000 [vat] registrations; at present there have been 282,000 and the number is still rising. Over 90,000 businesses previously unknown under the old tax system have registered. Even more amazing was the fact that public approval rating of the tax went up after it came into force….
—Roger Douglas, Minister of Finance of New Zealand, in Toward Prosperity (Auckland: David Bateman), p. 221
Experience with vat shows, once more, that in developing countries tax administration is tax policy.
— Milka Casanegra de Jantscher, “Problems of Administering a Value-Added Tax in Developing Countries,” IMF Working Paper 86/15, December 18, 1986, p. 18
The vat may be complicated to administer but it is not as complex as personal or corporate income taxes. Compared with the administration of a sales tax at the manufacturer or wholesale stage, the vat authorities must deal frequently with many more traders. Compared with a multistage turnover tax or a retail sales tax, although the number of taxpayers might be almost the same,1 the information needed to levy the tax is more complex to provide and more difficult to check. However, three of the supposed advantages of the vat are administrative: First, as each trader’s sale is someone else’s purchase, the invoices required for a credit-invoice vat theoretically allow the authorities to cross-check the accuracy of a trader’s purchases by adding up the supplier’s sales. Second, the data on purchases and sales in the vat returns should provide a valuable input to the calculation of profits for the purpose of income taxation, thus improving the administration of direct taxes. Furthermore, none of the complicated rules about uplifts, discounts, industry agreements, and special regulations under wholesale and manufacturers taxes are needed.
Usually, the introduction of vat has been a major upheaval in a country’s tax system2 and the opportunity has been taken to reform a large part of the tax administration. This chapter is concerned with the details of vat administration and staffing. Administration and staffing may appear to be a mundane preoccupation, but without it all the theory and broad generalities are vitiated. This chapter addresses four questions. First, who should run a vat? Second, how should functions be specified and allocated? Third, how many staff are needed? And, finally, what are the likely costs of administration?
Who Should Run a VAT?
One of the earliest, and most important, decisions that has to be made is to determine which branch of tax administration shall be responsible for the vat. In Latin America, most taxes are organized by functions such as collection, audit, and appeals, and automatically the department administering the internal taxes administers the vat. In other countries, there are four options: the existing sales tax authority can run the new tax, the income tax, or the customs and excise, can take over, or a completely new authority can be established. Since the existing sales taxes being replaced by the vat are likely to be simpler, they are also likely to rely on a less sophisticated administration and on a not so highly trained staff. Moving to a vat is bound to involve extensive retraining and, probably, recruitment and upgrading of staff.
In most cases, where the income tax is efficiently collected, there are advantages in associating the administration of vat with the income tax. As mentioned already, the potential for cross-checking between vat and income tax liabilities is one of the advantages claimed for vat. Such coordination should be much easier if the new tax is under the wing of the income tax organization. Nevertheless, there are sometimes doubts about expanding the scope of the income tax by adding such a large and important tax as vat to what is already the biggest branch of taxation.
The U.K. vat might be thought of as a good example for other countries to emulate. However, the U.K. case was unusual; the Customs and Excise had been responsible for the former purchase tax (a tax charged at the manufacturing and wholesale stages on wholesale values) introduced in wartime, when the income tax organization was understaffed and overburdened and when Customs and Excise was less busy than usual. Because of this accident of timing, the Customs and Excise gained experience in dealing with the wholesale trade in goods, and when the time came to introduce vat, its experience in warehouse control, inventories, verification of customs import values, and factory visits for excise purposes was taken as a deciding factor in allocating responsibility for vat administration. It was recognized that Customs had no experience in taxing services. The elimination of the purchase tax also freed 2,000 customs officers for other work. Perhaps there was also a feeling that the income tax authorities should not be made even more dominant than they were already. In general, few countries have followed this example, and it is not a road many tax advisors would follow.
Sometimes, the tug of war between the customs and excise and the income tax authorities can be such that to avoid antagonizing either, it is deemed more appropriate to allot the vat to an entirely new organization, as has been done in Portugal, where the collection, audit, and legal interpretation under the Serviço de Administraçāo do IVA, the vat branch of the Tax Department, is a separate operational agency. However, there are dangers in this solution: staff may be reluctant to be transferred from income tax or customs and excise to a new sales tax; the personnel of the existing agencies may unreasonably resent the new organization; and there may be problems of cooperation among staff. It should also be borne in mind that, in any case, customs will collect usually a fourth (and sometimes much more) of the vat revenue (on imports). Whatever is decided will require strong and sustained management interest and direction to get the new tax up and running.
Possibly the most complicated relationship between the income tax and vat authorities is in Israel. In that country, the low rate (6.25 percent) is levied on municipalities and nonprofit organizations by the income tax authorities (as the vat in this case is levied on organizations that make no profit, it can be viewed basically as a tax on wages and hence is collected monthly). The standard rate vat (15 percent) is administered by the Customs and Excise except for financial services. The vat on financial services is levied as a direct additive base on wages and profits (profits are defined as they are for the purposes of the income tax) and hence, in this case, the income tax authorities again collect the vat on a monthly basis.
Even if the vat is fully integrated with the income tax there is still the question of whether it is preferable to have a separate organization for vat at headquarters or at regional or district levels, or whether to integrate the vat into the activities at all, or none, of the administrative levels.
There are several basic arguments for having a separate vat oṙganization:
It allows officers to identify closely with a particular function.
There is better specialization; indeed, in developing countries especially, few officers are truly able to be good auditors for vat and good auditors for income taxes at the same time.
It allows for the integration of officers into the vat (for example, customs or excise officers) who might not be readily accepted by the income tax personnel (who frequently consider themselves a more technically and financially sophisticated branch).
The methods of operation of the vat and the income tax are too different to make a combined effort efficient. For instance, audit for vat is simpler and requires a less comprehensive accounting knowledge than an audit for income tax (basically because less analysis and less complex legislation have to be applied to the treatment of capital). This suggests a different and more frequent kind of audit than that for income taxation.
Different interests may be difficult to reconcile. For instance, vat inspectors are typically interested in the current fiscal year, whereas income tax inspectors tend to concentrate on earlier years.
Against this, or in favor of a more complete integration with income tax, are arguments such as:
A more flexible use of manpower (provided the whole staff is highly trained).
A less divided staff.
Better appreciation of mutual problems in vat and income tax, both within the offices and on the part of the traders.
Better internal communications of ideas and solutions.
Joint controls and audits check fraud and help identify undeclared activities.
Probably less duplication of resources.
No separate career paths to create jealousies and elite groups.
Reduced compliance costs to taxpayers, fewer inspections, fewer communications with traders, and more consistent treatment that reduces conflicts.
However, whether or not the functions of the officer running the vat can be combined with the existing income tax system sometimes turns on prosaic but unyielding structural considerations such as the civil service job-grading system. If income tax officials have been given a high job grading, they are, first, unwilling to accept functions for vat that might dilute that grading and, second, those responsible for vat will not work “under” an income tax officer. Such structural rigidities can stymie attempts to combine functions.
Two options (with suggested functions) for the district level are shown in Chart 12-1 and 12-2. In purely logical terms, the combined option (Chart 12-2) looks better. The structure is more cohesive and the duties can be combined more flexibly. The other two options of either shifting the customs and excise function from being primarily concerned with goods crossing international borders to taking on widespread domestic tax functions, or setting up a wholly separate agency, appear less happy.
Chart 12-1.Possible Administrative Structure with VAT Operations as a Separate Organization at District Level
Source: Adapted from the Irish organization chart.
Chart 12-2.Possible Administrative Structure with VAT Integrated with Income Tax at District Level
The circumstances of each country are different and sometimes the decision to whom to give the vat can depend on the personalities of the senior people involved. A young dynamic executive may seem more prepared to try new ideas with a new tax. Perhaps a well-tried experienced hand at the helm is needed to drive through a controversial new tax. The case has to be evaluated for each country, and Table 12-1 shows how diverse the decisions can be. Although the preference for a single department is clear, it is less so for a combined inspectorate. Whatever the decision, it has to be taken early and not left hanging while the planning for the new tax starts off in an administrative limbo. Senior appointees to the vat administration should be participants in the initial planning and committee work and should be seen to have the authoritative backing of the senior revenue and treasury officials (as happened, for example, in New Zealand).
|VAT and Income Tax Administrations|
|One department||One inspectorate|
|Germany, Fed. Rep. of||Yes||Yes|
Central Versus Local Administration: Organization and Functions
There are several arguments in favor of decentralizing many of the administrative functions of vat:
Current management philosophy in many countries favors further decision-making power at the managerial and functional levels that have the information most immediately at hand for prompt action. Traders like to have ready access to local offices to register and to discuss vat problems. The local staff get to know the traders, visit the trading premises, and carry out field audits. Traders should not feel they are dealing with an impersonal centralized machine.
Since the vat is a self-assessment system, voluntary compliance is important, and local personal contact between administrator and trader enhances the likelihood of good compliance.
Decision making is quicker, although it may be uneven—see below.
Coordination between headquarters and traders is probably improved.
Auditing, with local knowledge of trading conditions, is likely to be better.
Regional offices are smaller and staff should be more versatile; if income tax and vat are run together it should make for a more flexible use of staff, a most important consideration in many countries.
Against decentralization we have at least six arguments:
Central control ensures uniformity of decision making and of equal treatment for all traders (but this can be countered by good training and up-to-date manuals).
Officers from a central office are less likely to be swayed and influenced by local conditions and powerful local interests; they will not identify with local power structures.
Corruption is less likely. This may be a crucial consideration in some countries.
Decentralized offices may not have an even work load and, unless planning and execution are good, there can be underused staff; a central staff can be moved to even out local disparities.
A large central operation may allow greater specialization.
There are administrative advantages in having centralized records, accounting, and financial control.
As the points above illustrate, the balance of argument can go either way and each country’s decision reflects its own political, administrative, and social priorities. It does seem desirable, for “customer relations” at the least, to have a substantial local presence, and, for staff morale, a reasonable degree of autonomy should be given to local offices. However, local staff must be capable of handling the responsibility; in some developing countries it has proved better to move gradually in this respect.
Sometimes external technical considerations can dictate the solution. In some cases, communications between the headquarters and districts are so poor that, despite any opposing arguments on technical grounds, the primary discretion on vat decisions has to be left to the district. In another example, although a large central computing capacity existed, telecommunications links were so poor and inadequate that for at least ten years the authorities could only envisage main reliance on local administration. Indeed, the great advances in personal computer capacity and programming have meant that local offices can be given more duties to be carried out using local computing facilities with common software. Data can be transferred to the central authorities on tape or disk by the regular post or courier routes on a daily or weekly basis.
The way in which countries have decided to allocate responsibility for vat illustrates the range and diversity of options. As mentioned above, the United Kingdom is an example of a country that decided to use the Customs and Excise for the vat. Chart 12-3 shows the U.K. organization. Although the computer is centrally housed, some ninety local offices are located in principal towns (and some of these offices have suboffices in other towns).
Chart 12-3.United Kingdom: Organization Chart Showing Customs and Excise Headquarters Offices and Local Offices, and the Main Communication Links Between These and Traders and the VAT Tribunals, 19771
Source: Adapted from U.K. Revenue Commissioners.
1In 1977 mere were 1.25 million vat taxpayers in the United Kingdom. Total staff employed on vat at headquarters was 2,000; local vat offices and suboffices employed more than 7,500 staff.
These offices are generally responsible, under collectors of Customs and Excise, for the local administration of the [vat]. Staff are available … to advise traders about the tax and to deal with their enquiries. In addition, some 35,000 registered traders are visited at their premises each month to ensure that they understand the tax and their responsibilities within the system and that their tax returns are correct and their tax accounts adequate. These routine control visits are an essential part of the system, given its self-assessed basis; but local control is supplemented through use of the central computer system to check the credibility of traders’ returns to draw attention to apparent inconsistencies in patterns of trade.3
Enforcement of vat is a local issue in Belgium. There is one Local Control Office [LCO] for every 2,000 traders and, as a trader, your local office is never very far away. The LCO consists of about 12-15 staff depending on the work load and it carries out all the functions of registration, enquiries, variations, typing, and most of the preliminary work connected with enforcement. . . .
… when the LCO is satisfied that the return contains valid information they code the essential details on to a special magnetic roll and send these rolls each day to the Computer Centre (CTI). The CTI never gets any information direct from traders nor does it correspond with them. In this way access to the CTI and its sophisticated equipment is restricted to those who know how it works.
The CTI holds a current account for every registered trader and every week it runs the LCO magnetic rolls against it. Payments are all made via one control “Office de Cheque Postale” in Brussels on a prescribed form. Payment information too is coded on a roll, sent to the CTI and this too is run against the current account. Thus the current account shows every movement in the traders position and to assist everyone the CTI sends a quarterly extract of the current account to the trader and to the LCO.4
The LCO has on line direct access to the CTI computer at all times. Regional ctis deal with the collection of vat and fines resulting from assessments by LCO officers.
Korea and Ireland are examples of fully integrated revenue services under common umbrellas, the Office of National Tax Administration (ONTA) in Korea and the office of the Revenue Commission in Ireland. However, the structure in each differs. In Korea, in the regional tax office, there are both direct and indirect tax bureaus, and the indirect tax bureau has separate sections for excise and vat. At the district level, the vat division is in charge of assessment and collection. The central computer prints lists of traders whose data are “nonmatching” (see below), and these are sent to the local officers who visit each trader to inquire the reasons for the discrepancies (which often prove to be figures transcribed wrongly rather than deliberate evasion).
The Irish organization is an interesting example of a homogeneous revenue agency working in a small country. As Chart 12-4 shows, the collection and inspectorate functions are general to all taxes, as are the head office functions of legislation and statistics, general administration, refunds, control audits, investigation, and fraud. At the local level, there are 14 districts (of which one, Dublin, covers about 50 percent of the registered persons).
Source: Adapted from a chart of the Irish Revenue Commissioners.
1Not all staff are assigned full-time to vat; for instance, at the Assistant Secretary level, approximately a quarter of their time would be allocated to vat, and at the Principal Officer level, perhaps less than a third of their time.
2Numbers in parentheses refer to the full-time equivalent staff in each functional category; total revenue staff on vat = 524.
In another small country, the Dominican Republic, vat is administered from headquarters with all audit and enforcement centralized. Registration, vat returns, and collection are made in 29 regional offices (one for each province).5 In the larger Latin American countries, vat is integrated within the tax department, electronic data processing is centralized, and auditing is decentralized.
New Zealand opted to integrate the vat with the income tax at the district level (although in this case there was a peculiar problem in that the split of records for income tax between the two largest district offices did not necessarily form the ideal base for the vat). On the other hand, Morocco, despite using the income tax to administer the vat, decided that administration should be carried out from different offices.
In Brazil, the tax is run at both federal and state levels (federal at the manufacturing stage only, and state at all stages down to retail) and the states enjoy considerable autonomy in the administration of their vat (and may, under the new Constitution currently being debated, have even more, including differential vat rates). Alternatively, in Mexico, although also a federal system, the administrative discretion left to the states on the vat is much less than that in Brazil. Nevertheless, collection and auditing are primarily the responsibility of the states and, as the same states do not have an immediate direct share in the revenue collected, this can give rise to strains between the central control and the states.
In the Federal Republic of Germany, the states collect and administer the vat, and a formula is used to allocate the proportion to be remitted to the Central Government. In this case, the states have a clear interest in collecting all the tax due as this increases their own revenue as well as that of Bonn.
The desirable balance seems to be one where local offices are given administrative decision-making responsibilities that will be improved through local knowledge of local traders. These would include registration and deregistration, delinquencies, field and desk audits, complaints, and physical checks. Collection and financial control, including penalties, however, are probably better organized from a more remote center. Some control audits should also be made from headquarters. If state governments are to share in the revenue, better local administration will be obtained if that revenue share is seen to fluctuate directly with the vat collections.
Once broad agreements have been reached on the general structure and responsibilities of the administration of vat, then the crucial question of staffing must be tackled. Of course, changes in the structure can induce changes in staffing levels, but basically the structure comes first. Thereafter, the variables affecting staffing must be considered, staffing levels in general outlined, and staffing by function allocated; finally, the staffing problems peculiar to the transitional period should be delineated.
Variables Affecting Staffing
The most important factor affecting staffing levels is the number of taxpayers. As the vat usually covers more taxpayers than the taxes replaced, the case is immediately made for an increase in staff. However, it should be noted that it is the number of registered taxpayers that is at issue here. Thus, staffing requirement is a function of the following:
(1) The exemption turnover limit.
(2) The exempted goods and services.
(3) The treatment of agriculture.
(4) The treatment of affiliated companies and group trading.
(5) The collection by other agencies.
(6) The staff transferred from taxes replaced.
(7) The frequency of returns.
(8) The complexity of tax rates;
(9) The standard of administration.
(10) The existing computer systems.
Each of these variables is discussed below.
(1) A high exemption limit can transform the scale of administrative problems. If many smaller traders need not register, the staff can concentrate on the more important and remunerative revenue sources. For instance, in Latin America, small traders typically make an annual return and quarterly or biannual presumptive payments. Of course, other problems increase, including evasion, company splitting, and a reduced tax base, but the saving in staff can be substantial.
(2) Similarly, if important lines of goods and services are omitted from the vat, staffing needs are reduced. If, for example, all professional services are not included in the vat base, an irritating and difficult to tax group is avoided, admittedly at the expense of revenue, neutrality, and equity, but fewer staff are needed.
(3) In many countries, farmers are the largest group to be left out of the vat. In New Zealand, for instance, the difference between including all farmers as registered taxpayers (estimated 180,000 registered traders) and excluding them (120,000 registered traders) was 33 percent of the potential registration base. Ways to avoid requiring farmers to register, for example, the EC flat rate offset (see Chapter 7), not only avoid taxing a difficult sector, they greatly reduce the need for staff. However, in countries with advanced, literate, and numerate farmers, who frequently employ accountants to keep their books, registration and control for vat need not involve a disproportionate number of administrative staff.
(4) Returns can be simplified by allowing subsidiary companies to register as a group, through a central parent company as the registered trader.
(5) The amount of vat collected by other agencies affects staffing requirements. An island economy typically imports a large part of national consumption, whether as final or intermediate goods. The vat is collected by customs officers at the point of entry. In some countries, this can represent over 60 percent of the vat revenue and, even in Europe, can be as high as 35 percent and is probably the most secure portion of the vat revenue. Moreover, in the control of vat, the document established at the port of entry is the single most important control element. Manufacturers and wholesalers can be efficiently audited only if an audit trail is started by the import document. By contrast, if the United States were to introduce a vat, the potential handle of collection at the import stage would be a much smaller proportion of consumption than in most countries now using vat, so the likely demands on noncustoms staff would be higher. Similarly, state tax personnel can take over federal vat administration, as has been the case in the Federal Republic of Germany.
(6) Staffing requirements are also affected by using staff released from the taxes abolished. Frequently this requires retraining and the strains between old and new staff in the vat can persist and disrupt staffing for many years. Staff can be transferred only gradually since the taxes to be replaced still have to be administered for several months following the start of the vat (see Chart 12-5, below).
Chart 12-5.An Example of Recruitment and Placing of VAT Staff
Source: See text.
Note: Recruitment numbers are shown in preimplementation functions before the circular boxes: after the circular boxes recruitment is shown with reference to the postimplementation function.
1Date when skeleton head office starts.
2Date when regional offices are set up.
3Date when district offices open.
(7) The volume of paperwork and hence the staff needed to handle it can be significantly altered by the required frequency of vat returns. A balance must be struck between the evenness of cash flow to the government and the volume of administration. An annual vat is conceivable but would create a large interest-free loan to the traders and leave the government with an uneven lumpy revenue, a potentially more difficult administration, and probably increase evasion. A monthly vat is used (especially in Latin America where inflation is high), but the flow of documentation is larger. Indeed, in Chile in the mid-1970s, the authorities tried to collect vatin advance every 15 days, to cope with high inflation. Many countries collect on a two-monthly or quarterly basis (with a provision for monthly returns for established cases where large refunds are due) with returns staggered through the period to ensure an even flow of work and revenue.6 A quarterly return should economize a trifle on staff, leaving more staff for audit by requiring fewer for processing returns.
In some developing countries, such as Argentina, Mexico, and Peru, vat taxpayers have been required to file yearly returns but make provisional monthly payments. However, as one commentator noted, “Experience shows that the yearly return does not seem to add to the effectiveness of vat administration because most administrations lack the resources to match the provisional monthly payments with it.”7
(8) The more complex the vat, the more staff is needed to run it. A single rate vat requires less staff than a multirate tax. Legislators, anxious to exhibit their consciousness of equity, introduce multiple rates and complex exemptions; they ignore the cost to the taxpaying public of the extra staff required to administer the more complicated vat. Even worse, legislators will introduce vat complexities and not create the staff positions to administer the tax properly, thus actually creating more inequities than they “solved” by the new legislation. It is interesting that, broadly speaking, the trends in the newer vats (Indonesia, Korea, Mexico, New Zealand, Spain, Taiwan Province of China, and Turkey) have been toward fewer rates, and those countries already using many rates (Belgium and Italy) have been trying to reduce the number.
(9) The standard of administration is a variable that can profoundly influence staffing. A quality system, sparing no expenses, might allow for visits to traders each year; in a less efficient system, the visits could take place, say, once every 6 years (one system acknowledges its audit frequency as once every 30 years, but this inevitably leads to greater evasion and is indefensible). Thus, it is clear that the level of compliance is an important factor in assessing staffing needs. There is always a trade-off between using staff to ensure that the really large vat taxpayers are audited regularly and properly (and knowing that this would mean less staff is allocated to audit small traders who will, therefore, be evading tax) and needing staff to audit smaller traders (knowing that this could put some major revenues at risk). A typical frequency might entail visits to very large traders twice a year and visits to small traders once every four to five years. Much depends on the standard of accounting and auditing and the actual underpayments found at each type of trader.
(10) Whether to use the existing computer system or to upgrade it is the final variable. Of course, the vat software must be unique to vat but if there is hardware available, then staffing and other administrative costs may be reduced. Ireland, for instance, uses one computer for all taxes, as does Korea (see Chapter 15).
The range of variation shown in Table 12-2 is quite surprising: Italy employs one staff member for every 726 taxpayers, whereas the U.K. figure drops to 149. In one Asian country, the initial staffing ratio was 1:45 and in an African country, 1:31. What does seem clear is that the suggested staffing ratio for a U.S. vat is way out of line. The U.S. Internal Revenue Service (IRS) audits about 2 percent of its income, estate, and gift tax returns and relies on withholding to collect. By contrast, countries using vat audit 10 percent to 15 percent of traders annually, cannot use withholding for vat and, typically, collect a higher proportion of the vat through the (secure) customs import stage than would be true for the United States.8 The various country examples indicate that the IRS may have to budget for a substantially lower ratio of staff to taxpayers than estimated hitherto. Looking at Table 12-2, a more traditional ratio would suggest 1:250 and a consequent fourfold increase in the number of staff needed (from 20,000 to 80,000) and in the expense (from $410 million to, presumably, something well over $ 1 billion). In turn, this emphasizes that to justify such an administrative expense, the U.S. in voice-credit vat rate would not need to be 5 percent but more like 10 percent to 15 percent. Whatever the final outcome, the actual figure should emerge from the careful assessment of staffing at each level.
Staff to Taxpayers
Staffing in General
The functions of the staff at different levels must be clearly specified, then the number of staff should be allocated provisionally, followed by a check on the overall ratio to see how the “bottom-up” approach matches the aggregative approach and estimates outlined above.
Naturally, staff ratios, at different functional levels and overall, would be country-specific, but to illustrate some not implausible figures as we go along, let us presume we have a country with 300,000 registered traders where three levels of administration have been chosen—district, regional, and head office. Broadly, the district offices would be responsible for registration, advice, troubleshooting, and compliance; the regional offices would deal with inspection control and some legal and personnel matters; and the head office would collect and process all the vat revenues and be responsible for general administration, policy, and research. Naturally, these functions can be reshuffled in many ways; for instance, where postal, telecommunications, and banking communications are subject to delays, collection may be pushed down to the district level.
In the following sections, the functions and staffing of each level will be reviewed briefly to give an idea of the options and some acceptable staffing ratios (which, of course, will vary widely for different societies and economies).
The District Office
Control.—The usual control functions of the district office would include the following:
Dealing with applications for registration.
Educational visits to new traders.
Responding to inquiries.
Identifying new registrations and making changes in existing registrations.
Cancellation of registrations for traders who go out of business.
Dealing with credit and refund problems.
Low-level “audits” and inspections and follow-up action on the basis of inspection reports from regional and headquarters offices.
The ratios for these functions may be a general figure of about 750—1,000 traders to each staff member, which seems appropriate for a relatively simple vat (a standard rate, a zero rate for exports, a fairly high turnover exemption limit, and, say, a two-month return period).
Filing.—The district office would need, in addition, staff for routing and filing duties, to
Establish and maintain the filing system.
Check out and retrieve returns for audits.
Experience shows that each sorter can be expected to control some 30,000 returns a year,9 or some 4,000-5,000 taxpayers or cases.
Inspectorate.—Probably the most important function of the district office will be the inspectorate. The vat functions better with a high-visibility inspectorate and the trader knowing there is a reasonable risk of an inspector’s visit fairly frequently. The inspectors would be responsible for both routine visits and fraud investigations. Their duty would be to be seen in the field rather than at the desk. They would not be performing full audits—see Chapter 13. The inspectors should have the following functions:
Make educational visits to potential or newly registered traders.
Make routine visits, selected principally by computer, and conduct selective checks.
Help traders to comply with vat regulations.
Ensure that traders’ records are kept in a way that fulfills vat regulations.
Detect and follow up suggested cases of evasion or fraud.
The number of inspectors is crucially dependent on the frequency of their visits and on the number of visits an inspector can accomplish each year. Information is difficult to compare internationally, as administrations differ so much in what they expect officers to do in the field and in the office and, indeed, in the provisions they make for their staff to travel about easily and the time allowed to do so. Some visits may take only an hour, for example, for cross-checking some invoices or for explaining a new ruling, while a fraud investigation might take a week, and some time has to be allowed for office work, updating of records, and so on. However, with these allowances, inspectors should on average manage to make at least one visit every two days. Obviously inspectors will give priority to high-risk or high-yield traders (say, the building trade or retailers) and spend less time on low-yield or low-risk traders, but with approximately 22610 working days in a year, an inspector should conduct at least 113 visits.
So, with 113 visits by inspector each year and a registered trader population of 300,000, the implied inspectorate is shown in Table 12-3. With a four-year cycle and 300,000 traders in our illustrative economy, 664 inspectors would be needed, a ratio of 1:452.
|Average Frequency of Visits|
by Each Inspector
|Ratio of Inspectors to|
|One year (300,000 x 1/113)||2,655||113|
|Two years (300,000/2 x 1/113)||1,327||226|
|Three years (300,000/3 x 1/113)||885||339|
|Four years (300,000/4 x 1/113)||664||452|
|Five years (300,000/5 x 1/113)||531||565|
Assuming an inspector conducts 113 visits a year.
Assuming an inspector conducts 113 visits a year.
Arrears.—The final function of the district office in our example would be to ensure that vat is paid, including vat debts. Staff must visit traders, sympathetically assess the trader’s ability to pay (it may be a genuine temporary liquidity problem), arrange a payment schedule, monitor any arrangement, and be firm in following up delinquencies. In some countries (including European examples), staff do not visit traders to collect arrears as it is considered too dangerous! Instead, the taxpayer is requested to come to the tax office. To some extent, the staff needed depends on the automaticity of the penalties and the commitment of the courts to back up the tax prosecutions (see Chapter 14). However, arrears are also a function of the general perception of the government’s relationship to the private sector. In some countries, traders have frequent experience of government agencies delaying payment of their bills and, in response, traders sometimes withhold taxes due. This can affect the relationship between the government and the entire commercial sector, undermining tax compliance and the authority of the staff. However, this is not a problem peculiar to vat.
If we assume a ratio of staff to registered traders of 1:2,000 for the recovery of arrears, the overall implied staff for our example would be 150. Or, with a district office responsible for 10,000 traders, about 5 staff in that office.
Administration.—Support staff duties include secretarial, training, and management services. Every country has a very different idea about the desirable ratio of support staff to “professionals.” The training function will be at its most demanding at the initial stages of vat, but will continue even after the tax is up and running. Some countries prefer to have training a function of the regional office with training at the district level being more on the job. If district offices become large, then a separate training office can be allocated to that area. Similarly, depending on the size of the district, one or two managers can be added to the administrative staff.
Overall, in a district office with 10,000 registered traders, and with 300,000 registered traders altogether, the staffing needs might be as follows:
The Regional Office
In our example, we consider there should be about one regional office for every 7 to 10 district offices; we will assume four regional offices to service the 30 district offices. The regional office, in this example, would deal with the following:
Monitoring district office performance.
Legal interpretation of vat legislation and cases for revenue recovery.
Personnel recruitment, training, and coordination.
Support services to districts for emergencies.
The day-to-day responsibility for inspection would lie with the district; the regional office would have one or two inspectors to carry out spot checks on the regional offices and to take over extensive and difficult investigations of evasion and fraud.
Each regional office might need a regional attorney or advisor. Alternatively, if communications are good and/or the country is compact, the administration might prefer to keep all the legal services at head office. The regional legal offices would be available for opinions on request from traders, district inspectors, and the regional comptroller. If there are local vat tribunals where appeals are heard, the regional legal advisor would present the government’s cases.
The personnel officer would be responsible for recruitment in accordance with head office guidelines. He might develop and maintain training programs, monitor careers, ensure that annual performance reviews are conducted, and maintain regional records.
Finally, the regional staff might be expected to help district offices if there were special problem cases in the district or to arrange interdistrict coordination should one district be shorthanded or have particular difficulties.
As noted earlier, such regional vat offices would probably be expected to work closely with existing income tax officers and would probably share their accommodations. In our example, we are looking at numbers approximately as follows:
The Head Office
In our example, the head office is responsible for revenue collection and running the main electronic data processing (edp) unit. Other functions would include vat policy and research, inspection, legal services, and management.
Data Processing.—An important duty of data processing staff in vat is to scan returns to ensure that the required data are available for entry and that they are entered correctly. A typical, uncomplicated vat return would need checking for the following:
Trader registration number.
Goods and services sold liable to vat correctly specified.
Goods and services sold correctly exempted.
Goods and services sold correctly zero rated.
Goods and services purchased liable to vat.
Goods and services purchased exempt from vat.
Goods and services purchased zero rated.
Net vat due or net credit and evidence of payment.
A way to approach the problem of estimating the number of staff needed would be to take the total number of registered traders (300,000 in our example), and assume the two monthly return periods, crediting 150,000 returns a month, on a staggered basis. Evidence from some pay-as-you-earn (paye) and vat data suggests that a staff member deals with some 3,000 to 4,000 payments a month, which means that some 38 to 50 staff would be needed. However, even the simple vat entries discussed here suggest that, depending on the precise form of vat chosen, more time will be needed than for paye. So the staff needed might be 50 percent to 80 percent more, say, 57 to 90; for our example, let us assume 75 staff for edp screening (depending on expertise and duties this figure may be toward the high end of the options).
Similarly, data entry, updating the data base, and keeping the master file and the tax deduction ledger base requires about double the number of fields to be keyed, compared with paye. Therefore, using the same ratio as before, the staff needed for edp entry might be about 85 (again, a possibly generous figure).
Policy and Research.—Although policy and research are most important for the launch of the vat, they must be kept running after the initial decisions and problems have been dealt with. Obviously, many of the staff who have considered the issues and have thought and written about them more carefully would be valuable in advising regional and district offices. Nevertheless, governments keep changing the vat structure and the head office must advise ministers on the practicality and cost of changes proposed; it should keep in touch with developments overseas, help prepare legislation and regulations, analyze data, and think ahead about potential improvements in the vat. These functions are often underrated in the staffing of vat, but in fact are crucial to the continued effectiveness and development of the tax.
Inspections, Legal Services, and Management.—The inspection and compliance staff at headquarters are, in a sense, the “keeper of the crown jewels” for the vat. They should be the most highly qualified officers with field and desk experience to provide specialist investigation services, highly trained accounting support to the regions and districts, and have the ability to scrutinize computer selections for possible fraud cases and tax intelligence. Such a staff should certainly include statisticians, accountants, and inspectors.
Legal services would have to be employed to meet the casework load generated by vat. This would depend greatly on the complexity of the tax, the attitude of traders to tax administration in general, and the general litigiousness of the society. The better the staff at the district inspectorate are in explaining and tactfully defining issues at the trader’s level, the less the need would be for the expensive services of legal staff at the head office.
Finally, the head office should provide management and office services. These responsibilities would include the following:
Publicity, including radio and television.
Manuals, both initial and updated.
Design of forms.
Checking on the flow of reports and job sheets.
Checking that recognized office procedures are followed in districts and regions.
Overseeing personnel, training, and staff development.
Of course, the number of staff allocated to these functions would vary greatly, but for our example could be summarized as follows:
|EDP data entry||85|
|Policy and research||4|
|Inspection and compliance||10|
|Management and office workers||25|
Another way to look at the head office figures, if the head office is responsible for raising returns and monitoring and running the vat collections, is to estimate, as the British did with quarterly returns from over a million traders, that the average number of communications received by the central unit each working day would be about 50,000 (an average of three communications per trader per return—some are monthly—and 240 working days a year). Say an average of five minutes per communication and each effective seven-hour working day yielded a staff of 600 to add to the 300 needed for computer operations. In this way, an estimate of the staff needed for the central office alone amounted to about 1,000. However, “it may seem even more astonishing, and it certainly depressed us”11 that the complementary staff for a head office—initial establishment, training, solicitor’s office, the accountant, comptroller general’s office, investigation, clerks, typists, registries, messengers, and cleaners accounted for another 1,000. Another general rule of thumb is that such staff is about a 1:1 ratio to the professional staff. So the U.K. Control Office was estimated to need a staff of about 2,000, which indicated a local staff of about 6,000 and 8,000 overall. With about a million traders, the implied staff/trader ratio was 1:125.
In our example, using the U.K. rule of thumb, with 226 working days, two monthly returns, and three communications per return, the implied staff needed for scanning would be 284. With about 84 staff needed for computer operations, the basic staff for the head office would be 368, and if the 1:1 ratio of head office complementary staff is needed, the head office ends up with 736—which appears a high figure for the example of our relatively small country. Clearly, such calculations are country-specific, but it pays to be aware of the underlying perceptions of the given advice—particularly if you are paying for the consultant.
The total staff in our example is 1,260 at the district level, 44 at regional offices, and 203 at the head office, making 1,507 in all. As the assumed number of traders is 300,000, the overall staff ratio is 1:200, which is above the French and the U.K. figures, about the same as that for Portugal, but well below the figures for Italy or that suggested for the United States (see Table 12-2). However, it should be noted that no figure has been added for the proportion of time spent by customs officials collecting vat. Whatever figures would be added for this in terms of full-time staff equivalents would reduce the staff/taxpayer ratio.
Table 12-4 shows the proposed percentage distribution of staff calculated for the U.S. vat (with a staff to taxpayer ratio of 1:1,000). Compared with the example just considered above, both allocate exactly 35.8 percent to processing returns, computers, and statistical services; however, within this figure, the United States seems to assign an extremely small number to actually run the computers (this may simply reflect a different description of functions). For the rest, the U.S. Internal Revenue Service appears to assign more staff to taxpayer services and appeals, at the expense of the examination services.
|Total staff required||1,507||100.0||20,694||100.0|
|Computer and statistical services||160||10.6||22||0.1|
|Taxpayer services and other costs||131||8.7||2,407||11.7|
|Appeals and counsel||16||1.1||940||4.5|
It must be emphasized that the example discussed above does not represent any country and is not intended as a recommendation; it does, however, highlight the issues to be discussed and gives at least an idea of possible relative levels of staffing and allocation of functions.
Transitional Staff Problems
Recruitment and Transfer
Whether the vat replaces existing taxes or is introduced as a new tax, there are important transitional issues for staffing. The vat staff cannot be created with a complete hierarchy overnight. Senior staff for all levels must be brought in from elsewhere in the service.
Unfortunately, real life does not allow such a simple smooth changeover. The taxes to be replaced must be monitored up to and past the day on which the new tax is introduced. No matter how desirable it may be to detach staff from the old tax for training and thinking about the new tax, the machinery of government has to be maintained so that only a limited number of trainees can be released at any one time. Moreover, personnel have to continue working on the old tax, overlapping the staff of the vat; collections must be made, delinquents chased, audits conducted, and penalties collected. For a time the two systems will operate side by side and this requires a “hump” of staff during the changeover. This is no easy matter to arrange and is certain to cause strain. The responsibilities and timing of job changes need to be worked out in detail well ahead, so that everyone is fully aware of what is to happen and unpleasant surprises are minimized.
It is wasteful to recruit permanent staff solely for the temporary “hump,” when careful anticipation of retirements or re-employment of recently retired staff for the temporary overlap can cut the long-run costs of staffing the vat. Chart 12-5 gives one example of a possible schema for the recruitment and placing of staff. Duties are divided into preimplementation and postimplementation functions and the transfers are shown so that the cumulative recruitment problem is presented, literally, as the bottom line. Again, this is only an example, and each country will create its own method of analyzing the problem, but the types of problems are similar and can be represented in similar ways.
The assumptions behind Chart 12-5 are that 21 months are allowed to plan and introduce vat. Legislation for vat will be introduced during this period; the exact timing depends on political and legislative situations unique to each country (Hungary put the legislation through only three months before the vat was introduced). Obviously, administrators cannot train and organize for the vat until they have a clear mandate from parliament. In our example, legislation might be introduced at the end of the first year. The first six months involve an initial design team of about two dozen persons and the first major recruitments are the 70 trainee supervisors, 10 for the head office eventually and 60 to form the core groups of six at the ten district offices. Another early and important recruitment is that of the 10 edp staff to develop the computerization of the vat at head office. Indeed, half of this group is recruited in the second month of planning and is involved in all the early discussions. Further main edp and data processing recruitment is left to later in the program.
In the example shown in Chart 12-5, only 280 (250 for inspection and compliance and 30 for administration) persons are transferred from the taxes being replaced; naturally, this figure could be much larger, depending on the number employed in the taxes replaced. However, experience shows that sometimes it is difficult to get people to leave the existing tax administrations and, even when they do move, frequently their skills and motivation are not necessarily suited to the vat.
In this example, the monthly recruitment reaches a peak nine months before the vat is introduced, with the addition of edp personnel and a large batch of trainee supervisors and field and data registration officers, all bound for regional offices for control and inspection functions. The total cumulative number recruited for vat, 1,507, is the same as that used in the earlier discussion on relative functional responsibilities; the totals under each function at headquarters and at regional and district offices is, similarly, the same as that described earlier.
Finally, three dates are assumed to be important. A skeleton headquarters is set up after 6 months (that is, 15 months before the vat is started). The small regional offices are opened 2 months later and, a year before the “vat date,” the district offices are started with 6 inspection control “core” trained officers each, and 10 more officers to be trained by the core officers. After another 3 months, a further 380 officers are added to the districts (30 for inspection and 8 for control in each office) to be trained by those already trained. There is another influx 3 months later, and the final additions occur just before and after the vat starts, as staff is transferred from the taxes replaced.
Plans for Staffing
The planning for vat can take many forms, depending on the political, administrative, and legal relationships and responsibilities. The first thinking about vat might best be conducted in a small cohesive committee with an outside (academic?) chairman, government lawyers, and administration economists (such as the format in Portugal). Such a committee would consider the feasibility and broad implications of vat for revenue, prices, savings, equity, and so on and as such is not shown in Chart 12-5. The committee’s report would be a precursor to the decisions to adopt vat after which the timetable outlined in Chart 12-5 could start.
If it is desirable to push ahead more quickly and involve from the outset those who will conduct the vat, then the outline organization for planning the vat could take the form shown in Chart 12-6. This embodies some of the structure used in the United Kingdom where the planning (and the present vat administration) is split into two broad categories—”vat liabilities” and “vat machinery” (not the most elegant usage). In Chart 12-6, it is suggested that “vat: Goods and services” would deal with the liabilities of different goods and services to tax and another principal branch “vat: Administration” would be subdivided into problems connected with “Procedures” and others connected with “Traders.” Of course, few distinctions are watertight, and those dealing with, say, “retailers” would also have to know about, for instance, the decisions being made in “food, agriculture, hotels, and secondhand goods.” Indeed, this whole division might be too complex for a country where the number of traders was below, say, 300,000 and some consolidation might be needed. An alternative organizational structure is given in Chart 12-7, which draws on the experience of New Zealand.
Chart 12-6.A Possible Outline Organization for Planning VAT (I)
Source: See text.
Chart 12-7.A Possible Outline Organization for Planning VAT (II)
Source: See text.
In Chart 12-6, a third major responsibility is shown for the economic effects of vat (this responsibility is subsumed under the main subheading “Technical” in the New Zealand example—Chart 12-7). Ministers are keenly aware of the potential political ammunition provided by the introduction of vat and those responsible will be asked for initial assessments (and continuous updates) of the likely effects on prices, household expenditures, investments, and exports. As some of the earliest work on vat will involve the calculation of alternative tax bases and possible combinations of tax rates, this unit should be created at the beginning of the planning stage (and, indeed, may include the continuation of the original appraisal committee referred to above), but would be run down and abolished, or phased into research, once vat was introduced and working efficiently.
It may be desirable to create a small legal team to advise the vat planners on the numerous issues involving the law in nearly all aspects of early decisions on vat (this is shown in Chart 12-5, with two persons recruited in the early stages of thinking about vat and eventually assigned to vat headquarters).
Training for vat depends on the tax it is replacing and the officers to be trained. If the income tax department is responsible for vat and uses its own officers, the training may be an extension of its existing courses, simplified to the extent that vat audits do not need the full galaxy of accounting tools used in income taxation. If, however, the vat employs many officers previously used to administer taxes based on physical controls, such as customs and excise, then they will have to be retrained to administer a tax based principally on financial returns and controls.
Whatever the scale of retraining, the broad methods are similar. A small group of officers is selected to be trained, say, 12 to 18 months before the vat is to be introduced. This “core” group may have to be taught for 4 to 5 months, perhaps helping draft manuals, often with the aid of counterparts from other countries. After their training, each of them becomes a teacher to train similar groups on, say, three-to six-week courses. In one developing country, a typical course designed for training local (district) junior officers engaged on the straightforward administration of vat involved approximately three weeks: for tests (three days); theory and practice of vat (two days); vat law (one day); bookkeeping and accounting for vat (four days); “control” visits (four days); reports on control visits (one day); and fraud, law of evidence, and psychological aspects (one day). Depending on the size of recruitment and training, the “teachers” can continue training successive groups of officers, and their newly trained personnel can themselves return to provincial centers to teach yet more staff. Whatever system is adopted, it has to be articulated well in advance, so that the cadre of trained personnel will be available at the locations on the dates needed to implement the vat (see possible sequences outlined in Chart 12-5).
|A. Manpower, Control, Costs, and Revenue|
|(Numbers in thousands)|
|Directly employed on activity||10||…||10||…||10||1.9|
|Directly employed on activity and domestic administration||13||12||12||11||11||–7.3|
|Registered traders at March 31||1,286||1,327||1,338||1,380||1,398||5.4|
|(In millions of pounds sterling)|
|Underdeclarations of tax discovered on visits||61||84||146||169||290||375.4|
|Overdeclarations of tax discovered on visits||4||4||10||5||24||585.7|
|Cost of administration (including domestic administration)||99||105||130||145||155||48.0|
|(In pounds sterling)|
|Average underdeclarations discovered per visit||152||228||409||472||866||469.7|
|B. Cost of Administration as Percentage of Revenue|
|vat and customs and excise||1.59||1.35||1.36||1.32||1.26||–21.0|
|Inland Revenue duties||2.00||1.89||1.98||1.72||1.73||–13.5|
|Inland Revenue and National Insurance||1.41||1.31||1.40||1.23||1.24||–12.1|
|Ratio of registered traders to manpower directly employed (vat only)||128||139||141||147||149||16.4|
Revenues increased in 1979/80 because of the increase in the vat rate to 15 percent (from 8 percent and 12.5 percent).
Revenues increased in 1979/80 because of the increase in the vat rate to 15 percent (from 8 percent and 12.5 percent).
Costs of Administration
It is notoriously dangerous to relate costs of tax administration to the revenue collected. First, it is often difficult to designate precisely which officers administer which taxes; often the responsibility between taxes is blurred for one tax helps another (after all, one point in favor of vat is that it is supposed to help the income tax). Second, changes in rates of duty and tax may need more staff, yet greatly “improve” their efficiency ratios. Conversely, multiple rates of vat might both raise revenue and require more staff, thus not improving the ratio at all. Nevertheless, as long as we are not wedded to the significance of these ratios to several places of decimals, their broad movement over time and their relation to other similar tax ratios may be informative.
Table 12-5 shows how the costs of administering the U.K. vat have fallen over the period 1978/79-1982/83 and amounted to only about 1.08 percent of revenue in 1982/83. Of course, as the table also shows, excises were the cheapest taxes to administer (under 0.5 percent of revenue collected). Indirect taxes were less expensive, in toto, to administer than personal and corporate income tax (above 1.72 percent). It should be noted that if national insurance is included as part of direct taxation, the administration costs as a percentage of revenue fall to a figure much the same as that for indirect taxation.
A fair comment seems to be that in 1982/83 there were about 150 registered traders per person employed in administering the U.K. vat (up from 128 five years earlier). However, as noted earlier, this ratio can be as high as over 700 (as in Italy) or even 1,000 (as proposed in the United States), but at the expense of the number of control visits. As Table 12-5 shows, each control visit in the United Kingdom on average uncovered underdeclarations of £866 (US$ 1,561), so that the potential revenue costs of not employing sufficient staff can be high.
Thus, evidence indicates that a vat introduced at low rates would need almost the same number of staff, yield less revenue, and be less revenue effective as a vat at much higher rates. Given the complexity of the changeover, the recommendation has to be that the costs of the vat can only be justified if it is a major revenue source which, in the best of circumstances, generates sufficient resources to allow the government leeway to make discretionary changes in other taxes (a point debated in the U.S. context). The buoyant revenue from a broad-based single rate vat also can be used to compensate low-income households through transfer payments for equity costs incurred through the introduction of a vat (see Chapters 1 and 11).
This may look surprising, but under a retail sales tax, particularly in developing countries, many manufacturers and wholesalers make occasional sales to final consumers and therefore have to be registered. The U.S. authorities estimated that there would be only 10 percent fewer taxpayers under a retail sales tax than under a vat.
Not always; some Central American countries have adopted a form of manufacturing vat almost by accident, through extending a credit mechanism for tax on inputs, as was the case in Costa Rica.
United Kingdom, Commissioners of H.M. Customs and Excise (1978, p. 7).
Meyer and Due (1988, p. 14).
Some advisors do not favor staggered returns; they argue that returns should be filed to ensure the best possible compliance and maintain that to concentrate on, say, only four submission dates a year improves compliance. This may be true, but there are offsetting inefficiencies involved in bunching returns. The flow of work for accountants, advisors, and tax officials is better if returns are staggered—and the effect on compliance is not likely to be great.
Casanegra de Jantscher (1986, p. 14).
Casanegra de Jantscher (1987, pp. 302-305).
“Returns” does not necessarily mean one piece of correspondence; in the U.K. context each “return” might involve three actual letters to the trader.
Assumes a 5-day week, four weeks of holiday, and 14 public days or long-service days; but note, a European country assumed 240 working days a year.
Johnstone (1975, p. 47).