cupf S- /53 POLIcY RESEARCH WORKING PAPER 1430 Intergovernmental Fiscal Keyissuesof intergovernmental fiscal Relations and Poverty relations arise in all three s 1 t * - @ w s Y aspecrs~~~~aspe of Viet Nam'soert Alleviation in Viet Nam alleviaionstategNsbroad-e - based growth, human Richard M. Bird resource development and Jennie L Litvack safety nets. Spending and M. Govinda Rao revenue decisions need to be. more decentralized to ensure at pro-poor expenditures (such as local infirasmcure, health care, and educationj reflect the preferences. needs. and fisal abilities of different localities. The central government can ensure a minimum social safely net throughout the country by designing intergovernmental tansfers accordingly. The World Bank East Asia and Pacific, Country Departrent I Country Operations Division March 1995 | POLICY 1IF.SFARCII WORKING PAPER 1430 Summary findings A successful poverty alleviation strategy has four distinct Sincc poor provinces are less able to mobilize additional clements: (1) identifying who the poor are, where they local revenues to support services, well-designed are located, and what they do, (2) analyzing why they are intergovernmental transfers are particularly important. poor, (3) developing policies to improve their standards Provinces must play a greater role both in raising of living (usually aimed at accelerating economic growth revenues and in allocating expenditures, with incentives and improving their income-earning opportunities), and built in to ensure that they do so rcsponsibly and (4) supplementing income-improving policies with direct efficiently. Local governments must - if they are to be .safety net' policies to increase the poor's short-term held accountable for their actions - have some consumption entitlements. responsibility for deterrnining local tax rates. This will The precise mixture of 'capacity-improving" allow them to vary rates to collect more revenues to investments and "safety net' policies appropriate for any finance higher levels of public services if they so choose, country will depend on the country's income ievel, the and at the same time allow the central government to extent and nature of its povcrty problem, and many design its transfers in such a way as to ensure that local other factors. The strategy chosen must be implemented fiscal efforts are not discouraged by the receipt of such effectively. Spending and revenue decisions need ro be transfers. more decentralized to ensure that the poverty alleviation Richer provinces wi;l rend to coliect greater revenues. policies adopted reflect the preferences, needs, and fiscal When transfers are needed to finance local spending in abilities of different regions of the country. The nature of poorer areas, they should provide incentivecs for local that decentralization depends on the country. revenue mobilization and allow for some degree of Pro-poor services throughout Viet Nam are equalization. Services deemed of national importance underfunded. This problem is particularly acute in the (for example, a minimum level of educarion, health care, poorer areas. Improvements in the system of and social relief) can be promoted by designing specific- intergovernmental finances could help ensure that each purpose transfers. These scrvices must be identified and level of government, even in the poorer provinces, is varying matching requirements established for different adequatcly funded - and provided with sufficient provinces depending on such factors as their own expenditure and revenue raising autonomy - to support revenue base and the cost of providing services in that local investments and their operation and maintenance. province. This paper - a product of the Country Operations Division, East Asia and Pacific, Country Department I - is part of a larger effort in the department to develop a comprehensive povery alleviation strategy for Viet Nam. This paper was prepared as a background document for the Viet Nam Poverty Assessment and Strategy Report 13442VN, January 1995. Copies of this paper are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Grace Coward, room D8-088, extension 80494 (48 pages). March 1995. TPhe Policy Research Worl;ing Paper SCries dissteinarcs tbc fndngs of work in progress to encowage t*c exchange of ideas about developmnent issues. An objcstive of the series is to get the findings out quickdy, even ;.fthc presentations are less thant fully polished. The papers carry thc namtes of the authors and should be used and cited arcordingly. The findings, inferpretatkzons and conslusior.s are thc autbors' own and should not be attributed to the World Bank, its Executive Board of Directotrs, or any of its memnber countries. Pro-duced by the Policy Rcscarch Dissemination Center Intergovernmental Fiscal Relations and Poverty Alleviation in Viet Nam Bv: Richard NI. Bird Jennie 1. Litvack M. Govinda Rao Richard M. Bird is Professor of Economics and Director of the International Center for Tax Studies. University of Toronto. Jennie 1. Litvack is Country Economist for Viet Nam in the World Bank's East Asia and Pacific Region. M. Govinda Rao is Senior Fellow at National Institute of Public Finance and Policy, India. INTERGOVERNMENTAL FISCAL RELATIONS AND POVERTY ALLEVIATION IN VIET NAM I. Introduction A successful poverty alleviation strategy in any country has four distinct elements. First, who are the poor? Where are they located, and what do rhey do? Second, why are they poor? Only after we know who the poor are, where they live, what they do, and have some idea of the causes of poverty, can the third element in the strategy, the design of a set of specific policies to improve the living standards of the poor, be determined. Experience suggests that policies aimed at accelerating economic growth and enhancing income-earning opportunities for the poor are generally needed to provide a lasting solution to the poverty problem. In addition, however, such policies often need to be supplemented by more direct measures to increase the consumption entitlement of the poor in the short term. The precise mixture of "capacity- improving" and "safety net" policies appropriate for any country will depend upon its income level, the extent and nature of the poverty problem, and many other factors. Finally, whatever policies are finally chosen must of course be implemented effectively tO ensure the benefits actually accrue to the targeted groups. Fiscal decentralization enters into this picture ac several different levels. First, some degree of decentra!liztion of expenditure and revenue decisions may be considered desirable - even essential - to ensure that the policies adopted with respect to poverty alleviation (or anything else) adequately reflect the varying preferences and characteristics of different regions of the country. Although the extent to which this "decentralization theorem" (or "principle of subsidiarity") actually influences policy in any country will of course depend upon how a nation's political institutions are organized, the academic literature demonstrates clearly that decentralized provision of public services is preferable whenever there are significant localized dilferences in preferences.' The extent of actual, desirable or potential decentralization is thus one important factor in determining the appropriate mix of poverty alleviation policies in any country. Secondly, regardless of the extent to which local preferences determine the policies that are adiopted, the efficient and effective implementation of these policies may in many cases depen-d upon the close involvement of local governance institutions. That is, even if the same levei of service is in principle to be provided everywhere, the assumption is that those -overnments that are "closer to the people" - if only because they are responsible to (or for) fewer of them - should as a rule be able to provide services more efficiently and effectively than a remote, centralized authority. From an efficiency point of view, the basic rule of expenditure assignment is thus to assign each function to the lowest level of government consistent with its efficient peifonnatace. So long as there are local variations in tastes or costs, there are potential efficiency gains from carrying out public sector activities in as decentralized a fashion as possible. Fro;_ this perspective, the only services that should be provided centrally are those for which there are no differences in demands in different localities, where there are substantial 'spillovTrs" between jurisdictions that cannot be handled in some other way (by contracting, or bv Capita GDP) > Capila GDP < 450 th. Dongs 300 - 450 th. Dongs 300 th. Dongs 1991 (Actual) 1. Capital expenditure 22.4 10.0 12.2 15.9 2. Current expenditure. of 70.9 47.4 49.0 58.2 which (i) Education 13.4 11.3 10.9 12.2 (ii) Health 8.8 4.6 11.9 6.5 3. Total expenditure 93.3 57.4 61.2 74.1 1992 (Actual) 1. Capital expenditure 34.6 14.9 20.8 25.0 2. Current cxpenditure. of 126.3 73.6 79.7 98.1 which (i) Education 22.0 17.3 17.4 19.4 (ii) Health 1337 7.8 7.8 10.4 3. Total expenditurc 160.9 88.5 100.5 123.1 1993 ( Budget Estimates) 1. Capital cxpenditure 27.0 14.0 19.9 19.9 2 Working capital and 12.4 2.8 2.4 5.6 subsidy to SOES 3. Current expenditurc. of 139.2 78.9 82.5 100.6 which (i) Economic services 40.4 11.1 13.7 21.9 (ii) Education 26.5 21.8 23.3 23.0 (iii) Hcalth 14.7 9.3 4.4 11.3 (iv)Training & scientific 6.2 4.0 9.8 4.8 rescarch (v) Others 51.4 32.7 31.3 39.6 4. Total expenditure 178.6 95.7 104.7 126.6 I The high income provinces are: Ba Ria-Vung Tau. Ho Chi Minh city. Ha Noi. Dong Nai. Kien Giang. Quang Ninh . Khanh Hoa. Hai Phong. Hoa Binh. Vinh Long. Tra Vnh. Long An. An Giang. Can Tho, Minh Hai and Dac Lac 2. The middle income provinces are: Ticn Giang. Yen Bai. Soc Trang. Bac Thai. Tay Ninh. Ben Tre. Song Be. Lam Dong. Dong Thap. Quang Nam. Hai Hung. Vinh Phu, Thanh Hoa, Nam Ha. Thai Binh. Kon Tum. Lai Chau. Phu Yen and Ha Tay. 3. The low income provincres are: Long Son. NinL. Binh. Tuyen Quang. Ha Bac. Ninh Thuan, Gia Lai, Thua Thien Hue. Quang Binh, Binh Dinh, Binh Thuan. Cao Bang. Quang Tri. Nghe An. Quang Ngai. Ha Tinh. Ha Giang, Lao Cai and Son La. Source: Ministry of Finance and State Planning Committee, Government of Viet Nam. 13 Unfortunately, the provincial distribution of central government expendiEure is not available. Although it is possible that such expenditure may be allocated to some extent to offset the disequalizing effects of provincial expenditure, only the latter can be analyzed here. In general. as Table 6 shows, per capita provincial expenditures are clearly higher in more prosperous re-ions. The highest budgeted per capita expenditure was in the North-east of Southland (Ho Chi Minh City) and the lowest was in the North-Central coastal region, where per capita income was also lowest. The correlation is not perfect, however, since the relatively prosperous Mekong river delta region also had very low per capita expenditures, exceeding only the North-central coastal region. A similar pattern emerges for expenditures on different items for all three years shown in Table 7, with per capita expenditure in high income provinces being consistently highest. In aggregate, per capita expenditure in the high income provinces was 26 per cent above average in 1991. 31 per cent in 1992 and about 41 per cent in 1993. Interestingly, however, the distribution of per capita current expenditures on education and health was markedly more even. Moreover, per capita expendiEures in low income provinces were, on the whole, marginally higher than those in middle income provinces. To some extent, this result may reflect the "equalizing" effect of the present system of intergovermnental transfers (see later discussion) as well as the fact that some of the centrally-determined salary scales of provincial employees are hicher in more remote regions, thus raising expenditures in those regions On the other hand, the low levels of expenditures in the middle income provinces, some of which are located in the delta region, may suggest that in part their high concentrations of poverty reflect less their lack of development potential than their lack of the infrastructure needed to realize that potential. Nonetheless, a positive and significant correlation was found between per capita GDP and per capita provincial expenditures in both 1992 and 1993, confirming that on the average relatively more affluent provinces have higher per capita expenditures even though provincial budgets are, as discussed below, largely determined by the central government. 14 Two important inferences emerge from this discussion. First, the amount of resources available at the provincial level is so low in absolute terms that it is unlikely to have more than a very limited impact on poverty. Second, the distribution of resources among the provinces appears to favor those who have, in the sense that the relatively better off provinces on the whole spend more in per capita terms. Although some unit costs of providing some public services may be higher than the national average in more urbanized areas, others will be lower. Cost differentials alone seem unlikely to account for these results. Indeed, in the north mountainous region, where low population density means that the unit costs of providing public services are also likely to be well above average, per capita expenditures were actually slightly below average. Role of Local Governments in Poverty Alleviation What is the role of local govermnents in undertaking expenditure policies targeted to reduce poverty? Although the potential inter-provincial mobility of population limnits the efficacy of sub-central governments in undertaking redistributive policies.'° local authorities should have a comparative advantage in identifying the poor. They may also be more efficient providers both of social and community services such as health and water supply and economic services such as agricultural extension, irrigation and roads, particularly in rural areas. In addition, as noted earlier, in principle an essential function of local governments is to satisfy the varying preferences of people residing in different jurisdictions by providing differing mixes and levels of public services. Local governments may thus play an important role in deciding and implementing expenditure policies, including those intended to alleviate poverty. 10 See, Brown, C.C. and W.E. Oates, "Assistance to the Poor in a Federal System" Jounal of Public Economics, Vol. 32 pp. 307-330, 1987 and Ladd, H., and F.C. Doolittle, 'Which Level of Government Should Assist the Poor", National Tax Journal, Vol. 35, 1982. 15 On the other hand, since most taxes are more efficiently collected by central governments, an essential ingredient of the fiscal structure in most countries is also a system for transferring resources from the central to the local governments. An ideal set of intergovernmental fiscal arrangements, for example, might include the following elements: (i) adequate resources in the hands of local governments as a whole, from a combination of local taxes and central transfers, to enable them to carry out their assigned functions; (ii) a transfer system that ensures each individual local government has sufficient resources to provide essential functions at an acceptable standard, provided local taxes are imposed at reasonable rates; (iii) sufficient tlexibility in setting local taxes and charges so that local govermnments can respond to the preferences. special problems, and resource endowments prevailing in different regions; (iv) but at the same time, sufficient incentive to ensure that local fiscal effort is maintained at reasonable levels and that local budgets are managed efficiently; and (v) where minimum levels of spending on particular programs are considered of national importance, special central transfers financed in whole or in part by specific-purpose transfers. Such an ideal system is seldom found in practice, and certainly not in countries sucb as Viet Nam in which traditionally a unified budget has been employed as a tool to implement central planning. Nonetheless, an essential but often unduly neglected part of the transition towards a more market-oriented economy is precisely a change in the budgetary and intergovernmental system. From this perspective, it may prove useful to keep this "ideal" in mind as a point of reference. III.Intergovernmental Fiscal Arrangements in Viet Nam Expenditure Alocation The assignment of expenditures to different levels of government in Viet Nam is based on a resolution of the Council of Ministers passed in November, 1989, as subsequently 16 amended. Local governments are responsible for ensuring law and order within their jurisdictions. but, as in many other countries, with a few exceptions such as defence, most responsibilities are shared more on the basis of the size of the projects than the type of expenditure. Large investment projects, the benefits of which spill over to a number of provinces, are central responsibilities, while small projects benefiting mainly the residents of a province are undertaken by the provinces. Thus, all major irrigation, flood control and embankment projects and national highways are central responsibilities, but the maintenance and repair of minor irrigation works and roads other than national highways is a local responsibility. Similarly, primary and secondary education is primarily assigned to the provinces, but higher education is central, and major hospitals are managed by the center, while the provinces run hospitals in provincial towns and districts. Provinces are also responsible for the public health centers run at the communal level. Table 8 Share of Provincial Expenditures in Total, 1992 Total Provincial Share of Provincial Expenditure Expenditure Expenditure in (Bn. Dongs) (Bn. Dongs) Total (Per cent) 1. Currcnt Expenditure 18671.0 6697.5 35.9 1. Gencral administrative services 2404.0 1076.7 43.1 2. Economic services 1490.0 1176.3 78.9 3. Social services 6245.0 2872.6 46.0 (i) Education 1495.0 1325.0 88.6 (ii) Health 1136.0 1000.3 88.1 (iii) Others 3614.0 547.2 15.1 4. Other expenditures 8532.0 1612.0 18.9 11. Capital Expenditure 4955.8 1706.6 34.4 111. Total Expenditure 23626. 8404.1 35.5 Source:Ministry of Fmance, Government of Viet Nam. 17 Table 8 shows that in aggregate, provinces account for about a third of total as well as capital expenditures. Their role is especially important with respect to the provision of social services, particularly education and health. Moreover, although most capital expenditures are central, the provinces play a major role in maintaining public investments, as may be seen from the high proportion of current expenditure on economic services shown in the table. In recent years, the central government has introduced a number of special programs, as shown in Table 9, to cope with the perceived inadequacy of provincial spending in areas to which the center wishes to give priority. Such programs are designed by the center but implemented through the local governments. They are thus similar to the specific puTpose (non-matching) transfers seen in other countries, although the expenditures on these programs are shown only in the central budget. Expenditure on all these programs taken together amounted in 1993 only to 1615 bn dongs, or about four per cent of total expenditure, with a third of this figure being accounted for by reforestation. Many of these programs are directed to very specific activities, and the expenditure is spread so thinly across various sectors and provinces that it is unlikely to have any significant impact. In addition to these special programs, the central government also administers pensions to war veterans through the provinces: in 1993, expenditure on this item alone amounted to 2374 bn. dongs, or far more than the total of all the programs shown in Table 9 Although the provinces account for a significant proportion of govermment expenditures in Viet- Nam, their effective role in exercising expenditure decisions is very limited. During the third quarter of the year, after consultations with the State Planning Committee and various line ministries, the Ministry of Finance issues guidelines for preparing the budget. These guidelines (i) set out broad economic policy directions on the basis of projected macroeconomic and balance of payments situation in the ensuing year, (ii) specify the policy parameters to be followed by all spending departmnents including the local governments, and (iii) lay down the detailed procedures for preparing the budget and prescribe the 'norms' for 18 Table 9 Expenditure on Special Central Program in Provinces (Bn. dongs) 1993 1994 (Actuals) (Budgeted Estimates) I. Educadon 216 430 ti) Popularising primary education and spread of literacy 30 40 (ii) Education for mountain regions 60 s0 (iii) Separation or primary schools from secondary schools 50 50 (iv) Improvement in class rooms 100 (v) Elimination or 3rd shift in schools 70 100 (vi) Books and Newspapers for mountain pupils 8 15 (vii) Impruvemcnt in srcondary schools 6 25 (viii) Computer education for schools 20 2. Training Programs 10 105 ti) lmprtvement in teaching material and courses 10 15 (ii) Computer fior schools 20 (iii) Upgrading basic materials 20 3. Medicine and health programs 132 305 (i) UNICEF water supply programs for mountain region 1 5 (ii) Anti-Malaria program 52 70 (iii) Anti-GoiLre program 30 59 (iv) Immunization program 15 25 (v) Anti-HIV program 10 40 (vi) Upgrading material bases 25 70 (vii) Establishing health care centers in non-existent areas 26 4. Population and family planning program 93 225 5. Cultural and information programs 30 ti) Information and cultural activities for the communes 10 (ii) Filmmaking. renovation and developing 10 (iii) Cultural and historical heritage reproduction 10 6. Child care 2 30 7. Job programs. 300 250 8. Support for Khmer, Cham and other minorities 10 30 9. Administration of.boundary mapping 30 42 10. Pricc suppon policy for specified commodities 170 250 11. Project 327: Reforostation 584 610 12. Anti-social vice programs 60 70) (i) Anti-opium plantation S0 S0 fii) Abolition of prostitution 10 20 Total 1615 2377 19 making projections of revenues and expenditures. Based on these guidelines, the tax department prepares the forecast of revenues, and the State Planning Committee in consultation with the Ministry of Finance prepares the investment budget based on the proposals received from the provinces and line ministries which, in turn, are finalized after feasibility studies conducted by them. On the basis of the detailed norms specified for each item of expenditre, the provinces and line ministries then make projections of current expenditure. The Ministry of Finance puts together the estimates of revenue from domestic and foreign sources and attempts to balance the investment and current expenditures with the projected revenues. After two or three rounds of negotiations with the spending ministries and provincial governments, the estimates are finalized and submitted to the Prime Minister, who places it before the economic and budget committee of National Assembly for review. Each year, in early November, the estimates are submitted to the spending committee of the National Assembly which after a scrutiny places them in the National Assembly for approval. The budget thus approved is communicated to all spending departments and provinces and the latter in turn informn the districts. The provinces employ the districts primarily as their spending agencies. They are given a share in public enterprise taxes (profits tax, turnover tax and depreciation from State enterprises). The provinces directly spend most of the capital expenditures, and the responsibility of districts is mainly to execute small investment projects and to maintain the assets within their jurisdiction. Similarly, communes are allowed to retain 10 per cent of revenue from agricultural taxes and in addition receive a small grant, mainly for holding meetings. Most development works are directly undertaken by the - rovinces, and minor works are executed by the communes, often with voluntary contributions (mostly in kind) by the community itself. Provincial budgets are thus passed by the National Assembly, and revenues are basically assigned to meet the budgeted expenditures. If, during the course of the year, the revenue collections assigned to the provinces exceed the targets, the excess revenmes can either 20 be put in the financial reserve fund or spent on capital expenditures, major repairs or to enhance outlays on education, health and social welfare. If expenditures exceed the budgeted amounts as a result of a policy change (e.g., wage revision), an additional grant may be given by the center. Unforeseen increases in expenditure may also be met from withdrawals from a contingency fund established by each province, which must be recouped during the year. Provinces may thus spend more than the budgeted amounts only when their actual revenue collections exceed the targets (budget estimate). Moreover, current expenditures under various heads are determined on the basis of the norms detailed in the budget circular, and provinces cannot alter the allocations between different expenditure heads. Most provinces thus have only limited flexibility in taking expenditure decisions. In practice, however, as discussed later, the more prosperous and fast growing provinces are more likely to exceed the revenue targets in absolute terms and hence to be able to carry out more investmnents.Although the provinces are required to follow the detailed guidelines and norms set out in the budget circular in making their current expenditure projections, the expenditure budget tfiat is fir,ally passed in the National Assembly is invariably only a fraction of what the provinces project. Moreover, although the norms are, as discussed below, relevant in determining the relative shares of the provinces. they do not appear to be based on realistic estimates of the cost of providing reasonable standards of services or satisfactorily maintaining the economic infrastructure. Similarly, the provinces submit numerous investment proposals to the State Planning Committee after detailed project evaluation, but in the final analysis, the projects included in the investment budgets are chosen on the basis of considerations that are by no means clear and are only a small proportion of what is requested. Concentration of Tax Powers Before 1990, the responsibility for collecting most taxes except external trade taxes was vested with the provincial governments. The "unified budget" concept introduced in 21 October 1990 centralized the tax powers. At present, policies relating to tax base and tax rates in respect of all major taxes are decided at the Central level. The tax collection machinery, however, is responsible to both the Ministry of Finance and to the Provincial Peoples' Committees. The salary of the tax officials is paid by the central tax department; at the same time, however, these officials are accountable to the Provincial Peoples' Committees, particularly for fulfilling the revenue targets. Tax collections are deposited with the central treasury. The revenues from various taxes are then retumed to the provinces in part or full, on the basis of their expenditure requirements as determined by the budgetary process describee, earlier. Unstable System of Revenue Assignment The assignment of revenues to the provinces is done on a year to year basis, depending on the allocation of expenditures to them. Revenues and expenditures at the center and in each of the provinces are balanced by making a number of adjustments. To begin with, the entire revenue from certain sources are assigned to the provinces. Proceeds from other taxes are shared with the provinces, with the ratio received by each province being determined by the gap between budgeted expenditures and 100 per cent assigned revenues. If the proceeds from the shared taxes are not adequate to fill the gap, a cash grant equivalent to the remaining gap is riven. Balance is thus brought about through adjustments in expenditures, in assigned revenues, in shared revenues, and in grants. In addition, if, during the course of the year, the government changes policies and prices which affect revenues and expenditures, the Ministry of Finance together with the Provincial People's Committees recalculates the budget item and provides additional grants if necessary. There is thus no fixity or stability about the revenue assigmnent, as may be seen in Table 10. The revenues from external trade taxes, the special consumption tax, depreciation and profit taxes of state-owned enterprises vertically integrated with the center, and from major 22 minerals including crude oil are retained by the center. In 1993, the center retained the revenue from personal income tax as well, but in 1994 this source was assigned to the provinces except for Ba Ria-Vung Tau (where there is oil). The provinces are assigned the revenues from land and housing tax, slaughter tax, license fee, registration fee, depreciatL n and tax on local enterprises, revenue from lotteries, and other minor fees and taxes. In 1993, agricultural tax was shared with the provinces, but in 1994 it was entirely assigned to the provinces. Similarly, in 1994, 50 per cent of profits tax collected in 8 provinces will be retained by the center, though until 1993, all the provinces retained the entire amount."L The turnover tax is the major balancing item which is shared wiLh the provinces in both years though, until 1993, all the provinces received the entire amount. As already mentioned, the share of turnover tax forms the balancing (ap-filling) factor.'2 In 1994, 14 provinces will receive less than 100 per cent of turnover tax and the remaining 39 provinces will get cash grants in addition.(plus 100 per cent of the tax). As noted earlier, in an ideal system, a mix of general purpose and specific purpose transfers would likely be required. General purpose transfers would be designed to ensure that fiscally disadvantaged provinces would be able to provide a minimum "national sEandard" package of public services at a reasonable tax rate. Such grants should generally be structured to take into account both the "needs' of different provinces and their "capacity" to meet their needs out of their own resources. If properiy designed, such nequalization" grants can also ensure that local fiscal efforts are maintained at reasonable levels.'3 The eight rich provinces receiving 50 per cent of the share are, Ho Chi Minh City, Ha Noi, Ba Ria-Vung Tau, Long An, Kien Giang, Dong Nai, Song Be and Quang Ninh. 12 Thus, a province's share of turnover tux is given by (ErR,)/T, if (E,-RR) > 0 where E,, Rt and Ti respectively represent expenditures, assigned revenues and revenue from tumover tax of the i* province. 3 For an example of such a grant scheme, see Bird and Wallich, Financing Local Govermnent in Hungary, WPS, 19 869, 1992. 23 Table 10 Reveniic Assignment in Vietnam Central Revenues Revenucs Assigned to the Provinics Revenues Shared between Center and Provinces 1993 1994 1993 1994 1993 1994 1. Extemal trade taxes 1. ExternaJ Trade taxes 1. Land and housing tax 1. Agricultura! tax !. Six provinccs of hlc 1. Turnover aiix: 14 Mvickong River Delta provinces reccived less 2. Special consumption 2. Special consumption 2. Slaughter tax 2. Land and housing tax rcccivcd sharcs of than 100% of the siharc taxes taxes . agricultural taxes and the remaining 39 3. Licensc fcc/taxes 3. Slaughter tax ranging from 60-8 provinces received the 3. Profits and 3. Profits and percent. The remaining entire collections depreciation of public depreciation of central 4. Registration taxes 4. License fee/taxes provinces received the enterprises public enterprises entire collections 2. 45 provinces can 5. Depreciatloon and tax 5. Registration fees receive the entire pTofit 4. Crude oil revenues 4. Crude oil revenues on capital use and profits 2. Tumover tax tax. In 8 provinces, of provincial enterprises 6. Depreciation and tax however, only 50% of 5. Personal income tax on capital use and proftls the revenue from the 6. Profit tax of provincial enterprises tax is assigned 7. Taxes on lottery 7. Personal income tax (except in Ba Ria- Vung 8. Small scale border Tau) import/export8. Taxes on lottery 9. Transportation fees 9. Transportation fees 10. Revenues from forestry and 10. Revenues from reforestation fees forrstry and rcrorestation fees 11. Other minor fees and taxes 24 In principle, it may be argued there is little rationale for making such general grants unconditional.'4 Not only does the central government have a legitimate interest in what is done with its grants, but the nation as a whole also has a legitimate concern in ensuring that services provided by local governments such as education and health are available throughout the country at minimum standards. There is therefore a case for at least limited conditionality, for example, by requiring that grant funds should be spent on e.g. education or health or by requiring local governments receiving such grants to provide services of at least a specified quality and level. Ot course, such legal requirements are inevitably to some extent only pro forna. The fungibility of money and the ability of local governments to alter other expenditures and taxes mean that requiring a grant to be spent on a particular activity does not necessarily imply that total (centrally-funded plus locally-funded) expenditure on the activity has gone up proportionally. Indeed, in most cases it will not. Nonetheless, there is considerable evidence that conditionality is likely to increase the amount spent on the designated services to at least some extent. Specific purpose transfers, of course, are those which are specifically targeted to poverty alleviation or other programs. Such transfers have been found to be more effective when the recipient local governments - the provinces, in this case - are required to make matching contributions. Particularly when there remain significant disparities between localities even after the distribution of general purpose grants, as is likely to be true in Viet Nam in view of the paucity of resources available, higher subsidies (lower matching rates) may be appropriate for low-income provinces with larger concentrations of poverty.'5 14 The position taken here is set out in Bird, op. cit. For an apparently contrary view, although one aimed more at large intermediate governrments in federal states, see A. Shah, "Perspectives on the Design of Intergovernmental Relations,' World Bank WPS 726, 1991. 5 For a detailed discussion of these issues, see, Rao, M.G. and A. Das-Gupta, "Intergovernmental Transfers as an Instrument to Alleviate Poverty", Economic Development Institute, The World Bank. 1993. 25 The present system of transfers in Viet Nam falls far short of these precepts. Indeed, the concept of transfer itself is unclear. At one extreme, as the provinces do not have any independent power to raise revenues, all their expenditure may be treated as transfers. At another. all revenue collections made by a province may be considered as its own; in such a case, the difference between expenditures and revenue collections may be construed as transfers. A third option is to consider the transfers as the balancing component - the share of turnover taxes and the cash grants received by provinces. Finally, even the shares of tumover taxes may be considered as taxes assigned to the provinces, and only the cash grants considered as transfers. Expenditure projections Whatever concept of transfers is adopted, the critical element in determining their amount is, as mentioned earlier, the preparation of the expenditure budget at the provincial level. The process by which investment projects is selected is both complicated and non-transparent, but considerations such as the need to finance continuing projects, the relative backwardness of the provinces, the priority of the projects as perceived by the central government, and the relative priorities of the different provinces appear to influence investment allocations. In addition. as noted above, the provinces may increase investment outlays if they raise more than the targeted (budgeted) revenues in respect of taxes assigned to them. The process of determining current expenditures is equally complex. The budget circular sent to the provinces details the norms to be employed for projecting different items of expenditure. These "standardized" norms are applied to expenditure items such as education, culture and information, training (higher education), sports, health, and administration. In the case of maintenance expenditures on infrastructures like roads, bridges, embankments and irrigation works, standardized norms (on a per capita basis) cannot be applied, so expenditures are projected on the basis of specified norms for each item. 26 The nonns are developed for each expenditure item on the basis of the discussion between the Ministry of Finance, State Planning Committee and the relevant central Ministry. In most cases, population is the major factor taken into account but different weights are assigned to population in different areas. The budget circular for 1994, for example, distinguishes between five different categories: (i) Cities, (ii) Plains, (iii) Midland and Coastal areas. (iv) Low Mountains and Remote areas, and (v) High Mountains and Islands. In principle, this distinction is supposed to take account of differences in the needs as well as varying unit costs of providing these services. The standardized norms for 1994 for various items of expenditure are shown in Table 11. In the case of education, for example, the normative per capita expenditures vary from D 26,700 to D 42,800. Since provinces may contain more than one type of area, each province is required to project its education expenditures on the basis of its composition of population in terms of the different categories. If properly designed, a system of establishing expenditure "norms" may constitute a useful component of a general purpose transfer program, when combined with a measure of local capacity to meet costs out of own resources. Such norms might, for instance, take into account differences in the cost of providing a "standard" level of services in different regions. Such costs might vary with population density, the terrain, and other factors. In Viet Nam, however, although the norms apparently attempt to take some account of cost differences, it is not clear that the central government has adequate cost information on which to base the norms. Moreover, the chosen indicators of 'need' could be improved. In. the case of education, for example, costs are specified on per capita basis and not on the basis of population in the school- going age group or the number of students. If the age composition of population or the student-population ratio differs between provinces, this norm may not properly represent the need factor. Nor is any distinction made 27 between primary and secondary education. In the case of higher education and training, however, the norm is specified in terms of number of students. Health expenditures too are determined on a per capita basis. Administrative expenditures are determined on the basis of number of employees, although all provincial recruitment and the salary scales have to be approved by the Central government. Another problem is that resource constraints mean that the norms are set too low to provide reasonable levels of service. In the case of education, for example, on average the expenditure per student in 1991-92 was only D 106,800 and expenditure per person of school- going aoe was just only D 57,600 (about U.S. $6). Since teachers' salaries itself accounted for about 80 per cent of this amount, little was left for books, furniture and educational aids. Moreover, alEhough the norm system in effect allocates expenditure resources among the provinces, it does not seem to do so in a very satisfactory way. In Soc Trang, one of the poor provinces. for instance, per student expenditure of D 19,600 was actually the lowest. Indeed, all spending on education was less than 2 per cent of GDP, or less than half the average ratio in the group of developing countries with less than $ 300 per capita income.'6 The situation is no different in respect of economic services. In Thanh Hoa province. where cyclones are a recurring phenomenon, the sanctioned budget for strengthening and maintaining 937 kms of embankments is generally only about 10 per cent of what is requested by the province - which itself is only a fraction of what is actually needed for its proper maintenance. Even after mobilizing conmnunity help by way of voluntary labor (through embankment protection committee), the province is thus able to select only the very weakest segments of embanlknent every year for strengthening. The story appears to be the same in other sectors and in other provinces as well. 16 The estimate for the developing countries is for 1987, and is taken from the Statistical Year Book, 1989, UNESCO. 28 Table 11 Norms for Determining Current Expenditures for Standardized Items in Provinces 1994 Expenditure Itcms Unit Cides Plains Midland Low Moun- High Moun- and Coas- ta nous and tains and tal Areas Remote Areas Islands 1. Norms Based tn Catcgorization of Regions 1. Education Dongs per capita 37.500 26.700 J9.400 33.500 42.800 2. Culture and Dongs per capita 3.750 2.500 2,940 3.350 4.Z80 information 3. Sports Dongs per capia 1.540 1.250 1.150 700 960 4. Health: Dongs per capita i) Prcventive 4.750 3.340 3.700 4.170 5.340 ii) Curative 10.520 7.520 8.270 9.400 12.030 5. Administration Mn dongs/per (i) Provincial lev-el employte in pay mil 9-5 6.7 8.2 13.5 - (ii) District levcl 7.5 6.7 5.5 11.0 12.0 11. Norms Uniform for all Regions Norms for Science and Pailgogy Pedagogy Agri.. Economics Culture and Teehnical Education Technotogy - Genral - Higher Formsty. and Lcgal Arts - Mn dongs per Fisheries Affairs sLudent and Health .ti Cnilkcc and 3.8 4.0 2.7 3.9 3.3 4.2 Universi1t (iil Technical (3 yrs 2.4 2.5 - 2.1 1.9 3.0 after gencral eduatkion) (iii) Vocational 2.9 2.4 - 2.9 Source: Ministsy of Finance. Govcrnmnmt or Vic Nanm. Revenue projections The expenditure budgets of provinces are determined by the revenue projection for domestic resources (plus planned domestic borrowing and external assistance). Analysis of 29 revenue projections, however, reveals that the tax department consistently tends to underestimate revenues. The tax department tends to set targets which can easily be reached. Underestimation is in the interest of provinces as well, for, once a tax is assigned to a province for the year, collections in excess of projections provide some flexibility in the use of funds. On the other hand, the State Planning Comrnittee would like to have high revenue targets so as to have a larger invesiment budget- Often, the projections of tax department and State Planning Committee differ widely, requiring the Prime Minister to take the final decision. Even after such negotiations, however, as Table 12 shows, for the last three years, actual collections have exceeded targets in every single province. Although a larger share of some of the taxes collected in higher income provinces goes to the center and though in relative terms the increase in richer provinces was actually lower than in poor provinces, nonetheless in per capita terms, the richer provinces still retained higher excess collections and consequently had higher per capita expenditures. Estimated expenditures in per capita terms exceeded the budgeted amount by D 51,670 in high income provinces. by D 26,800 in middle income provinces, and by D 31,030 in the low income provinces. The actual distribution of public expenditures among provinces thus differed from the approved budget. General purpose transfers Tihe system of intergovernmental relationships and the design of intergovernmental transfers in Viet Nam can clearly be improved in a number of respects: First, local authorities do not have much autonomy and flexibility in deciding the level and composition of their expenditures. With virtually no independent revenue-raisig powers and with both revenue assignment and expenditures being decided at the central level, local governments are unable to increase outlays on agricultural extension, irrigation, and roads. 30 They are thus unable to exert much influence on either the pattern of economic growth or, indirectly, on poverty alleviation. Second, the impression gathered from field visits to various provinces is that very few of the investment projects forwarded after feasibility studies are included in the plan. Moreover. the repair and maintenance expenditures approved are only a fraction of those requested (and needed). The criteria determining which projects are selected are obscure. Table 12 Per Capita Revenues and Expenditures - Targets and Estimates Per Capit Rcevcnues - 199 Per Capiu Expendicurcs 1993 1 TarScEs Latest Increasc in Percen- Targets Latest Increase in Percen- (Th. Estimates Estitnaes tacg (Th. Estinate Estimates tage dongs) (h over Increase Dongs) (Tb. ovcr Inasc Dongs) Targets Dongs) Targets Cmb (Th- Dones) Dongs) High inctimn 656.43 758-10 101.67 15.5 178.53 230.2 51.67 28.9 provinces Middle income 95.43 111.19 15.76 16.5 95.65 122-45 26.80 28.0 prov mnce% L,w Inictmnic 54.53 69.92 15.39 282 104.73 135.76 31.03 29.6 provinces All Ponvinces 27S.60 323-89 45.29 16.3 126.54 162.99 36.45 288 Source: 1. State Planning Committee 2.Ministry of Finance. Government of Viet Nam Third, as noted above, the norms used for determining expenditure allocations between different provinces could be improved by obtaining better information on unit costs and by targeting them more specifically (e.g. on students, rather than on population). 31 Fourth. the lack} of stability in tax assignments and tax sharing arrangements creates botlh uncertaintles and disincentives for provincial govemments. Uncertainty in revenues renders planning for the medium term difficult. In addition, larger revenue collections in a province in one year set higher targets for the next year and appear in at least some cases to result in a lower provincial share of turnover taxes. In Ha Noi, for example, revenue collections in 1993 were highler than the previous year by over 50 per cent and exceeded the budget estimates by 15 pur cent. Much the same happened in 1992. Consequently, the share of turnover tax accruing to the cint was reduced from 70 per cent in 1992 to 36 per cent in 1993 and was further reduced to just 6.3 per cent in 1994. Fifth, the lack of transparency in the method of allocating investment expenditures and the complicated method of determining recurring expenditures make the interprovincial distribution of social and economic infrastructures a matter less of objective analysis than of negotiation and bargaining. In the final analysis, the allocation of expenditures basically responds to bureaucratic and political judgments at the central level, but the relationship between political leaders at the central and provincial levels also appears to play an important role in expenditure allocation. The impact of these and other factors on the provision of public services to poor people. particularly in the poorer parts of the country, is on the whole adverse. On one side, the predominant role of population in determining the expenditure norms obviously tends to equalize expenditures; on the other, population is not necessarily an accurate way to measure 'need.' Similarly, one reason higher weights are assigned to sparsely populated and mountainous regions is presumably to take their higher unit costs of providing services into account, but no one knows exactly how these weights are derived, or if they accurately reflect cost differentials. Given the general tendency to underestimate revenues, affluent and fast- growing provinces can generate larger per capita surpluses over targeted revenues, thus accentuating inequalities in social and economic infrastructures. Given the backlog in 32 infrastructure in poor provinces, even when equal per capita expenditures are allocated for investment. these provinces will find it difficult to 'catch up' with their more prosperous counterparts. To the extent the development potential of these regions is equal, this is clearly detrimental to economic growth. Even if poorer regions are poor in part because of a deficient resource base. so that in the long run it might be advisable for people in search of a higher income to move to where they can be more productive, it is generally in the national interest to ensure that educational and health standards meet at least minimal national standards. The allocation of expenditures among different provinces does not succeed in offsetting the revenue and cost disabilities of the pooTer areas. On the other hand, the central allocation of expenditures does result in expenditures being more equitably distributed than if the expenditures are incurred according to revenue derivation. The correlation analysis reported in Table 13 confirms that provinces with higher per capita GDP had higher per capita expenditures. Moreover, per capita expenditures were higher in more densely populated provinces, which perhaps suggests that the hi-her unit costs in mountainous and remote areas are not adequately reflected in the determination of aggregate expenditures. "7 The expenditure pattern has no si-nificant association with infant mortality rate, another 'need' variable. Similarly, per capita actual expenditure in 1992 is not significantly correlated with the illiteracy rate. but the budoeted expenditures in 1993 has a significant correlation. However, both actual and budcgeted expenditures are significantly negatively correlated with agricultural productivity. The picture is similar for current and capital expenditures. Both are higher in more developed provinces. Per capita capital and current expenditures are significantly related to per capita industrial output, but per capita agricultural output and productivity are negatively associated. Capital expenditures in provinces with higher infant mortality rates and illiteracy were higher in 1993 though there was no significant positive correlation in 1992. Current expenditures, however, showed no significant correlation with these need variables. 17 This result may also be because the population density in the relatively affluent provinces is simply higher. But the results of multivariate analysis shown later also support the above observation. 33 Table 13 Correlation Coefricients of Per Capita Expenditures with Capacity and Need Variables Current Expenditure Capital Expenditure Total Expenditurc 5 I992 1993 I 7 1993 1993 A. Variable Representing "Capacity' 1. Per capita GDP (1989 0.574* 0.338* 0.3660 0.282** 0.543* 0.327* prices) 2. Per capita aigricultural .0.424* -0.354* -0.351* -0.388* -04426* -0.368 productmin 3. Per capita industrial 0.479* 0.243** 0.279"* 0.207 0.445* 0.236** production eicr capita 4. Per capita production 0.493* 0.179 0.216 0.136 0.368* 0.170 (Industry + Agzriculturc) B. Variable Representing 'Need' 5. Population deLnsity 0.324* 0.349* 0.132 0.239** 0.2R3*" 0.323* 6. Road len-thisq. knm. 0.050 0.122 -0.008 0.10 0.035 0.092* 7. Agricultural -0.228** -0.282** - 0.277* -0.350' -0.257** -0.305** prtductivily S. Infant Mortality Rate 0.093 0.144 0.053 0.244'' 0.086 0.174 9. Illiteracy Rate 0.160 0.214 0.140 0.365' 0.163 0.259** 1992 - Actuals 1993 - Budget Estimates Significant at I per cent level Significant at 5 per cent level Source: Staff Estimates based on information provided by the Ministry of Finance and State Planning Committee, Government of Viet Nam. 34 Table 14 Inker-Provincilu Variationis in l'er Cnpita Expenditures: Regressioti ResUIts === ~~~~~~~~~~1. Linealr Equatlionn II1. Lozg-L.inc;ur Eqiuatio)ns Independent Variables Total Eap endi- Titatl ExpIendi- Currcnt Expendi- Current ExpeLndi. Total Uxpendi- Trn:ti IRxpesidi- Currelit Current Expenditure ture (1992) lure (1993) ture (1992) [Lire (1993) lure (19921 lure (1993) lExpetidittirc (1992) (1993) Constant 78.47450 ' 85.6801 " .27.2925 .59.7011 " ' 0.7706 1.3153 0.3465 0.2444 (1.9627) (1.9460) (.0.9904) (.1.7958) (0.7260) (1.2016) (0.3078) (0.2169) Per capita GDP 0.0883' 0.0490' 0.0824' 0.0533' 0.6512' 0.49X7' 0.6501' 0.5775' (5.1232) (2.9620) (7.4827) (4.5983) (6.6394) (5.1592) (6.9738) (6.4965) Population density 0.0525' 0.0759' 0.0149 0.0211 *0.0756 .0.0209 *0.1169' *0.1084"' (2.9370) (3.8386) (1.3349) (1.5712) (-1.2083) (.0.3175) (-2.4647) (-2.2957) Agricultural .20.6833 ' -25.4634" . . -0.3553' -0.4748". productliviy (-2.5513) (-2.8317) (-2.1285) (-2.6999) Illiteracy rate 1.7383' 2.3134". 0.2446* 6.3051" . (2.3005) (2.7598) (2.2682) (2.6933) Infant mortality rate . 0.3415 0.4450 0.0746 0.0902 (0.7023) (0.8242) (0.4243) (0.4869) Length of roads per sq. -9.0623 -6.8225' -27.87650 .27.42814'* 0.0410 0.0611 -0.0153 0.0575 km. area - (-0.4154) (-0.2821) (-2.1605) (-1.7522) (0.7706) (1.0974) (-0.3252) (-1.1855) Enrolment ratio . .17.3764 -0.1764 . -0.1870 0.0200 (-1.4597) (-0.0122) (-1.5992) (0.1708) Teacher - student ratio - . -0.3371 -0.0309 . . 0.1611 0.2303 (-0.5685) (-0.0430) (0.7177) (0.9924) Hospital beds per 1000 16.6056' 19.5164' 0.2939' 0.2463" persons (4.0406) (3.9126) (2.6172) (2.1310) Hospital staff per 1000 . . 39.1342' 47.06760 . . . persons (4.5253) (4.4741) 2 0.4918 0.4452 0.6567 0.5332 0.4935 0.4122 0.5408 0.5089 1 P 9.3860 7.9539 15.2124 9.4841 9.4438 7.0776 11.2058 8.6987 Significant at I per cent level, " Significant at S Fer cent level. ** Significant at 10 per cent level I'lie regression results for interprovincial differences in per capita current, capital and total expenditures presented in Table 14 generally confimi thcsc observations. In all the equations. the regression coefficient of per capita GDP is positive and significant for both 1992 aiid 1993. Similarly, population density shows significant diseconomies of scale though the rate of increasc in expenditures was lower as the rate of population density increases (log-linear equation). These regressions do suggest, however, that there is some attempt to spend more in provinces with lower agricultural productivity and higher illiteracy rate. But per capita expenditures were not significantly related to other need variables such as infant mortality and road length per sq. knm. of area. Unfortunately, data on provincial expenditures at the required level of disaggregation are available only for education and health. Similar correlation and regression analysis was carried out for per capita education and health expenditures with various -capacity' and need' variables. The results (not presented in detail here) are broadly similar to those already discussed, although some interesting observations emerged. For example, although the budgeted expenditures on education and health in 1993 in different provinces were not si-nificantly correlated with per capita GDP, actual expenditures in 1992 were significantly hiiher in provinces having higher per capita GDP. Expenditure on education was higher in provinces with a higher illiteracy rate, but there was no significant correlation with population density. Interestingly, education expenditure was negatively and significantly correlated with enrolment ratios. It is not clear how to interpret this result. If the low enrolment ratio is due to inadequate facilities for education, presumably higher expenditures in low enrolment provinces would increase enrolment. But this scenario seems unlikely in Viet Nam, where facilities (as evidenced by the ratio of teachers to school going population) are fairly evenly spread across the country. In this instance, the negative correlation may perhaps indicate that more is spent on education in provinces where the effectiveness of such expenditure may be lower. 36 In the case of health expenditures, significant pusitive correlationi with population density suggests, as mentioned above, that there may be inadequate recognition of cost disabilitics in sparsely populated provinces. Although per capita health expenditure has no correlation witlh such 'need' variables as infan,ut mortality rates, it is, unsurprisingly influcnccd sienificantlv bv the number of health workers and hospital beds. One way to analyze inter-provincial redistribuLion is by examiiining the diffcrence btLween the revenues collected in a province and provincial expenditures. The elasticities of per capita revenues an-id per capita expenditures in a log-linear equation arc as follows: I . Log (R) = -2.8141 + 1.2362 log (y) (-3.4369) (8.9660) -2 R 0.604 F = 80.389 N = 53 2. Log (E) = 2.3487 + 0.4044 log (y) (3.6976) (3.7812) -2 R = 0.204 F = 14.297 N = 53 Where R = Per capita Revenue collections in the provinces in 1992. E = Per capita expenditures in the provinces in 1992 y = Per capita GDP in the provinces in 1992 [Figures in the parenthesis denote 't' values of the regression coefficients.] These regressions, as illustrated in Figure 1, bring out two important conclusions. First, the income elasticity of expenditures in 1992 (0.404) was substantially lower than that of revenues (1.236). That is, expenditures are much iore equally distributed than revenues. Second, in spite of this equalization, per capita expend'-:ire was significantly higher in provinces with higher per capita GDP. On average, when the pzr capita GDP was higher by one per cent, expenditure was higher by 0.40 per cent. 37 Figure. .1: PROVINCIAL REVENUE AND EXPENDITURE PER CAPITA, 1992 10 - 2 - - l: ll 7 5 5.5 6 46.7 .8 85 a.~~~~~~~ uJ 5 5.5 6 6.5 7 7.5 8 8.5 GOP per Capita (Log Scale) In addition, there may of course be significant tax and expenditure spillovers across provinces. In 1992, for example, total expenditure of the central and provincial governments together accounted for almost 25 per cent of GDP. Of this, a little less than one- half was contributed by revenues from State enterprises and turnover and profits taxes on non- State enterprises: and about 4 per cent of GDP or 16 per cent of total expenditures was financed by fiscal deficit. Taxes on enterprises (profits tax, depreciation and turnover tax) are basically collected predominantly from the two major cities (Ha Noi and Ho Chi Minh) by virtue of the location of head offices of major enterprises and the concentration of production and trading activities in these cities. In reality, however, these taxes tend to be passed on to consumers throughout the country. Similar spillovers will also occur, probably to a lesser extent, on the expenditure side. Although the inter-regional incidence of various revenues and expenditures is too complex a question to explore here, it seems reasonable to suggest that if these factors were taken into account the net transfers received by the poorer provinces would be generally lower than indicated by the gross transfer concept used in the preceding analysis. 38 Specific Purpose Transfers The basic transfer system thus probably does something, but not much, to alleviate the effects of differential provincial capacities to provide services related to poverty alleviation. As noted earlier. the central government has also introduced a number of special programs in the education, health and social welfare sectors which are implemented by the local governments. However, apart from the pension payments to war veterans, none of the programs seems large enough to make any significant impact on poverty. One scheme that is specifically targeted to alleviate poverty is a subsidized loan pro-ram introduced by the Treasury and implemented through the Ministry of Labor to create employment for the unemployed. The program was introduced in mid-1992, and budgeted expenditure is about D 250 bn in 1994. The loans are given to individuals (D 50 bn) and to small enterprises and workshops (D 200 bn). The scheme covers Ha Noi and Ho Chi Minh cities and mountainous low density provinces. Loans for self-employment at subsidized interest rates are given after scrutinising applications submitted to the Provincial Peoples' Committees. The process is similar for the small enterprise loans, which are not directly targeted to the poor, but are given to generate employment. Loans carry a monthly interest rate of 0.5 per cent for loans up to one year, 0.4 per cent for loans up to 2 years and 0.3 per cent for loan up to 3 years. This employment program could be useful but at present it suffers frtom a number of shortcomings. First, it is only a small pilot program in two cities and a few provinces - and, as mentioned earlier, the incidence of poverty in the two major cities is relatively low. in terms of the amount of money spent, D-250 bn is less than 0.5 per cent of the total budget. Second, although the loans to enterprises are ostensibly given to generate employment, the recipients may be quite affluent, and there is no mechanism in place to ensure that these enterprises in fact generate the number of jobs as proposed, once the loan is received. Third, the maxinum loan given to an individual is D-500,000 or about $ 45 (in Ho Chi Minh.city, it is D-600,000). Even at a 25 per cent rate of return, the additional income generated will only be 125,000 or about 39 U.S. S12 per year: any impact on the living conditions of the poor is going to be marginal. Moreover. the fact that the loan is given for a maximum of three years makes it more difficult for poor persons to benefit from the scheme. Finally, while obviously repayment capacity should be taken into account to some extent in any loan program, there is no explicit attempt to allow for the intensity of poverty in determining who benefits from this program. IV. Reorganizing Intergovernmental Fiscal Arrangements for Poverty Alleviation Intergovernmental fiscal arrangements in Viet Nam need to be reoriented in order to improve the consumption standards of the poor, particularly those living in the poorer areas of the country. Poverty alleviation through economic growth reouires that the supply of social and economic infrastructure responds adequately to the increasing demands of the expanding market economy. Appropriate fiscal decentralization can help in this task. Concern is sometimes expressed about the possible effects of decentralizing public sector activities owing to the poor quality of local government administration in many developing countries. To a considerable extent, however. each country gets the local government it wants. In Viet Nam local government officials, like those anywhere, respond to the incentives with which they are faced. If those incentives discourage initiative and reward inefficiency, then it should come as no surprise to find inefficient local governments. The answer to this problem, if it is one, is obviously to alter the incentive structure to make it possible and attractive for honest, well-trained people to make a career in local government. Similarly, -'ie answer to local governments that make "wrong" decisions is to provide an incentive structure that leads them, in their own interests, to make the ''riclht" decisions, that is, decisions that are both economically efficient and politically acceptable. The fiscal arrangements needed to achieve this result will promote accountability and provide incentives for the cost-effective provision of public services. To the extent local governments are best suited to providing the services needed to accelerate economic growth and reduce poverty, they should therefore have a greater role both in raising revenues and in allocating expenditure, subject to the provision of adequate incentives to ensure that they do so in a responsible and efficient way. 40 One aspect of greater fiscal autonomy is the power to raise revenues. At present, the provinces can increase their spending only when the actual collection of revenues assigned to them exceeds the budgeted targets. As the tax department is not entirely within their jurisdiction. provinces thus have little flexibility in raising revenues. Moreover, the provinces do not have the right to determine the rates of any taxes. If provinces are to be able, within centrally-detennined limits, more freely to determine their level and composition of spending, they will need more independent revenue-raising authority. Of course, the power to levy taxes whose bases are mobile across different provinces should remain with the center, and local governments should be restrained from imposing taxes that will basically be exported to other areas. The basic principles of local revenue assignment are: (i) Local own-source revenues should ideally be sufficient to enable at least the richest local governments to finance from their own resources all local services primarily benefitting local residents(as opposed to those with si_nificant externalities such as education); and (ii) to the extent possible, local revenues should be collected from local residents only, preferably in relation to the perceived benefits they receive from local services. When transfers are needed to finance local expenditures, they should provide incentives for local revenue mobilization and, as a rule, allow for some degree of equalization The critical point is that local governments must, if they are to be held accountable for their acElons. have some responsibility for determining tax rates. This would both allow them to vary rates to collect larger revenues to finance higher levels of public services if they so choose and at the same time allow the central govermnent to design its transfers in such a way as to ensure that local fiscal efforts were not discouraged by the receipt of such funds. Vesting the provinces with at least some independent revenue-raising authority is essential to establishing a linkage between revenue raising and expenditure decisions at the margin. Of course, if local goverrunments are vested with the power to levy important taxes, inevitably richer provinces will collect larger revenues. To offset this disequalizing effect, as just noted, transfers can be designed to provide sufficient revenues to local governments to carry out a "standard" (minimum) set of local expenditure, as set for example by a revised (and more 41 accurate) set of norms, provided they levy taxes at average national rates. If they lower taxes below this level. they are not rewarded by higher transfers: the result is thus that local residents are penalized. If. on the other hand, they raise taxes above the required level, then total expenditures may be increased. The precise design and implementation of such a transfer system would require further study, but it should certainly be feasible in Viet Nam, if desired. Sensible local government decisions require a greater degree of stability and certainty in financial arrangements than now prevails: the new budget law would improve matters in this respect. A more scientific method of forecasting revenues and expenditures would also be an improvement, although this too is a matter for further study. As in the case of transfers, simpler and more transparent arrangements in determining investment allocations are also needed. Another area where reform seems needed is in regard to the special programs initiated by the center. Most of them are so small they seem unlikely to have any significant impact on anything. It would seem more sensible to focus on a few programs that might have an impact on poverty reduction. Providing drinking water, family planning, and adult literacy, for instance, are programs that would appear useful from this perspective. Programs to combat widely prevalent diseases (like malaria and goitre) may also of course be essential. Given the nation-wide externalities from spending on human capital formation, provincial initiative in these areas may be encouraged by specific purpose matching grants. There may also be a place for such grants, at higher rates of subsidization, in mountainous regions and for ethnic minorities. If the provinces are given the right to levy and collect some taxes, introduction of specific purpose transfers with matching resource requirements can result in greater provincial participation and help to augment more resources for their special programs. The correct matching rate is, in this approach, set by the size of the spillovers. Basically, a matching grant program designed to encourage the optimal provision of public services should therefore vary primarily with the nature of the activity, that is, depending on the level of associated externalities. Since, however, no country has achieved fIll equalization of local fiscal capacities, a uniform matching level offering the same "price" to all local goverrnents will yield non- 42 uniform responses in rich and poor localiries. Even if revenue bases are fully equalized, need or cost differentials may require an equalization element in matching grant formulas. For example, per capita grants for roads in sparsely populated and mountainous regions should as a rule be larger simply because the per capita cost of achieving any particular standard of road service will obviously be higher. In mountainous reaions and drought prone areas and in areas where there is large scale unemployment of unskilled labor, perhaps a rural works program of creating wage employment can be introduced to create infrastructures like rural roads, creating arnd deepening of ponds, lakes and canals, repairing and strengthening of embankments and building school and health centers. Such a program miaht, for example, generate a fixed number of person days of employment during the lean season. One advantage of such a program is its self-selection process: targeting the program to benefit the poor is easy as only the poor will seek such employment. In evolvina improved intergovernmental fiscal arrangements, it is obviously useful to build on the existing system to the extent possible by minimizing its undesirable effects and imparting the necessary, autonomy, incentives and accountability to various levels of government. The key elements in this process would seem to include the following: (i) devolution of some independent revenue-raising authority to local governments, in particular, the right to set the rates of some local taxes; (ii) creation of a more adequate method of resolving vertical and horizontal imbalance through revenue sharing arrangements and general purpose grants; (iii) identification of activities for which specific purpose transfers should be made and determination of the matching requirements that should be established for different provinces; 43 (iv) provision of adequate stability in intergovernmental fiscal arrangements; (v) and perhaps, although this aspect has not been discussed here, the assignment to the provinces of some limited borrowing powers for investment purposes. A system along these lines should prove much more responsive to the changing needs of an emerging market eccnomy while providing social and economic infrastructure in an efficient manner. An Organic Budget Law has been under preparation in Viet Nam for the past two years now. It has undergone several revisions; the most recent version, September 1994 does contain many of the suggestions made here. However, further revisions are expected in early 1995. On paper, Viet Nam currently has a highly centralized public sector; in practice, however, the way the system works- is to produce unplanned decentralization which basically places the poorer provinces in a relatively weaker position. More formal decentralization in a coordinated and coherent framework along the lines sketched above would be a substantial irnprovement and could prove a substantial step towards the implementation of more effective policies for poverty alleviation. 44 Annex I The Draft Budget Lawv A new State Budget Law regulating the budgetary process is expected to be considered by the National Assembly in June, 1995. Several drafts of this proposed law have been considered during the past 2 years. The latest draft (dated September 1, 1994) would affect some important aspects of the present system while leaving others unchanged. (i) Instead of the entire national assembly, its standing committee would decide the current and capital expenditures at the Central and provincial levels. It will also decide on the volume of Central transfers to the provinces (article 18). (ii) The draft budget law sets out the revenues and expenditures of different budget levels. The expenditure responsibilities of the central government (article 30) include state administration, defence, price support, non-productive economic activities of the central ministries and state agencies, non-productive training, nationally established and managed aspects of health care, training, education, culture. sports and scientific research. Investment expenditure on basic infrastructure, key economic projects, state-owned enterprises and joint ventures is also allocated to the center. (iii) Provincial responsibilities (article 31) include expenditure on provincially managed state activities, economic operations, education, health care, cultural, social activities, scientific research and sports managed by the locality, security and social order in the localities and investment expenditures on basic infrastructure (excluding the projects administered by the center). (iv) The revenues of the central budget (clause 28) will consist of external trade taxes, special consumption taxes, profits and income tax on foreign enterprises, 45 individual income tax, revenues from activities like oil and gas, resources tax and mineral revenues, profits and retained incomes from state-owned enterprises and joint ventures, proceeds from the sale of assets, and fees and non-tax revenues specified by the goverm-nent. The turnover (value added) tax will be shared betwveen the center and provinces, with the shares will be determined by the central novernment for a number of years. In addition, the center may borrow from domestic and foreign sources. (v) Provincial revenues (article 29) would consist of house and land taxes, a portion of agricultural land use taxes, natural resources tax (except those assigned to the Center), provincial share of turnover tax, taxes on land use transfer, license tax, slaughter tax outside the communes, state economic activities specified by the government, fees and non-tax revenues, grants from the central budget, and direct foreign aid as permitted by the central government. (vi) Transfers from higher to lower level governments (article 33) will be made to support current expenditure obligations, which are to be determined by factors such as population, natural conditions, and the level of economic development in each regions. The transfers will 'take into special account the need to create favorable conditions for economic and social development in deep and remote areas, former revolutionary bases, areas of minority peoples and disadvantaged areas.' Transfers to support capital expenditures will be made according to the list of investment projects already reviewed and approved. Some of these proposals seem good from the perspective of the previous discussion. The proposal to assign specified revenue sources to the provinces and to have stable tax sharing arrangements for the turnover tax for a few years for example, will bring about some degree of stability, reduce the disincentives prevailing at present, and help in better planning of the provision of public services at local levels. It is not clear from the draft law if provinces would be able to set the rate of their assigned taxes or if this would still be done by 46 the center. The September draft law refers only to the provincial level of local government and removes specific reference (which appeared in earlier drafts) to the commune budget. Since many important pro-poor expenditure items are delivered at the commune level, this lack of clarity could result in their underfunding. Also, the tax department will continue to be under the dual charge of the Ministry of Finance and the Provincial People's Cormnittees. Investment projects will continue to be decided in a non-transparent manner. This draft law has undergone several revisions but has not yet been submitted to the National Assembly mainly because of the sensitive issue of central-local relations. The September, 1994 draft version contained several important components which would increase decentralization of both expenditures and revenues; make the system of intergovernmental transfers more transparent; and increase the stability in the local budgetary process. This draft law is expected to be revised once again in early 1995. 47 REFERENCES Bird. R.M. "Financing Local Services: Patterns, Problems, and Possibilities," Paper prepared for Global Report on Human Settlements, April 1994. Bird, R.M. "Threading the Fiscal Labyrinth: Some Issues in Fiscal Decentralization," National Tax Journal, September 1993. Bird. Richard and Christine Wallich, "Financing Local Govermment in Hungary," (Washington: World Bank Working Paper No. WPS 869, March 1992). Brown, C.C. and W.E. Oats, "Assistance to the Poor in a Federal System" Journal of Public Economics, vol.32 pp.307-330, 1987. Tadd. H-I. and F.C. Doolittle, "Which Level of Government Should Assist the Poor", National Tax Journal, Vol. 35, 1982. Oakland. William, "Fiscal Equalization: An Empty Box?" National Tax Journal, March 1994. Oates, Wallace E. Fiscal Federalism (New York: Harcourt Brace Jovanovich, Inc. 1972.) Rao, M.G. and A. Das-Gupta, "Intergovernmental Transfers as an Instrument to Alleviate Poverty", Economic Development Institute, The World Bank, 1993. Shaw, Anwar, "Perspectives on the Design of Intergovernmental Relations," (Washington: World Bank Working paper No. WPS 726, July, 1991). World Bank, Transition to the Market, (Washington: Report no. 11902-VN, September, 1993). 48 Policy Research Working Paper Series Contact Title Author Date for paper WPS1411 Income Inequality. Welfare. and Nanak Kakwani January 1995 G. Evans Poverty: An Illustration Using 85783 Ukrainian Data WPS1412 Foreign Technology Imports and Xiaoming Zhang January 1995 C. Jones Economic Growth in Developing Heng-fu Zou 37754 Countries WPS1413 Endogenous Distortions in Product Martin Rama January 1995 S. Fallon and Labor Markets Guido Tabellini 38009 WPS1414 The World Bank and Legal Technical The World Bank January 1995 K. Mathernova Assistance: Initial Lessons Legal Department 82782 WPS1415 China's GDP in U.S. Dollars Based Ren Ruoen January 1995 E. O'Rielly-Campbell on Purchasing Power Parity Chen Kai 33707 WPS1416 Informa! Regulation of Industrial Sheoli Pargal February 1995 E. Schaper Pollution in Developing Countries: David Wheeler 33457 Evidence from Indonesia WPS1417 Uncertainty and Global Warming: An Andrea Baranzini February 1995 C. Dell Option-Pricing Approach to Policy Marc Chesney 85148 Jacques Morisset WPS1418 The Impact of Labor Market Lyn Squire February 1995 G. Bayard Regulations Sethaput Suthiwart- 37460 Narueput WPS1419 Industry Structure and Regulation Martin C. Stewart-Smnth February 1995 N. James 82758 WPS1420 Legislative Frameworks Used to William T. Onorato February 1995 W. Onorato Foster Petroleum Development 81611 WPS1421 Distribution of Income and the Income Zeliko Bogetic February 1995 F. Smith Tax Burden in Bulgaria Fareed M. A. Hassan 36072 WPS1422 Efficiency in Bulgaria's Schools: Zeljko Bogetic February 1995 F. Smith A Nonparametric Study Sajal Chattophadyay 36072 WPS1423 The Role of Commercial Banks in Millard Long February 1995 R. Gamer Enterprise Restructuring in Central Izabela Rutkowska 37670 and Eastem Europe WPS1424 Terms-of-Trade Shocks and Optimal Luis Serven February 1995 E. Khine Investment: Another Look at the 37471 Laursen-Metzler Effect Policy Research Working Paper Series Contact Title Author Date for paper WPS1425 On the Intersectoral Migration of Donald Larson February 1995 J. Jacobson Agricultural Labor Yair Mundlak 33710 WPS1426 Russian Unemployment: Its Simon Commander February 1995 V. Reid Magnitude. Characteristics, and Ruslan Yemisov 35195 Regional Dimensions WPS1427 Corporate Govemance and Equity Stijn Claessens February 1995 F. Hatab Prices: Evidence from the Czech 35835 and Slovak Republics WPS1 428 Short-Term Supply Response to a Bruno Boccara February 1995 M. Pfeiffenberger Devaluation: A Model's Implications Fabien Nsengiyumva 34963 for Primary Commodity-Exporting Developing Countries WPS1429 The World Trade Organization's Bemard M. Hoekman March 1995 F. Hatab Agreement on Government Petros C. Mavroidis 38535 Procurement: Expanding Disciplines, Declining Membership? WPS1430 Intergovernmental Fiscal Relations Richard M. Bird March 1935 G. Coward and Poverty Alleviation ir. Viet Nam Jennie 1. Litvack 80494 M. Govinda Rao