Его полный текст - http://www-wds.worldbank.org/servlet/WDSContentServer/WDSP/IB/2005/03/17/000012009_20050317113145/Rendered/PDF/289230RU.pdf
Документ большой, под двести страниц весьма технического текста, весит 17 мегабайт.
Мне показался интересным кусок, оценивающий перераспределительный эффект ряда социальных программ. Это страницы 97-99.
Маленькое предисловие. В общем плане государственные расходы на социальные цели обычно регрессивны, то есть перераспределяют от бедных к богатым, хотя и под лозунгом "помощи бедным". Конечно, в разных странах - своя специфика. По-видимому, в развивающихся бедных странах бюджетное перераспределение идет от бедных и беднейших в пользу самых богатых; в развитых богатых странах - от а) небольшой прослойки самых богатых и б) от подавляющей части среднего класса в пользу в) небольшой прослойки самых бедных и г) самой зажиточной части среднего класса (это моя гипотеза, могу и ошибаться).
Но это - если говорить о бюджетном перераспределении в целом. Специальные программы помощи бедным обычно имеют все-таки более целевой характер и в меньшей степени "перехватываются" средними и богатыми слоями населения.
Так вот, по оценке Банка, в России регрессивный характер имеют не только общие социальные программы, но даже и программы с максимальной эффективностью.
Я не имею никакого отношения к подготовке этого доклада, у меня нет личного мнения насчет его аккуратности и точности. Тем не менее, мне кажется полезным сохранить этот кусок вместе с тремя картинками-таблицами (еще две картинки не ставлю, они менее интересны). Возможно, доклад будет переведен на русский, но когда - не знаю.
Хотя выпущен доклад в феврале, подготовлен он на самом деле был, кажется, прошлым летом, а самые последние данные, использованные в нем - за 2002 и 2003 годы.
Да, чтобы понятно было. HUS - это "housing and utility services", то есть, по-нашему, ЖКХ.
----------------------------------------
Inefficient Budget Allocation within Non-contributory Social Assistance Programs
8.15. The budget allocation among various non-contributory social assistance programs marginalizes the programs targeted to the poor at the expense of an inefficient system of privileges. While the amount of public resources redistributed by the system is very large compared to other countries at a similar level of economic development, the share of resources explicitly targeted toward the poor or vulnerable strata of the population is very small. Only 7 percent of the total social assistance spending is explicitly targeted to the poor (23).
(23) The allocation for all targeted social assistance programs represented 0.4 percent of GDP, equivalent to 10 percent of total public spending on non-contributory social programs (3.9 percent of GDP) or 7 percent of the overall (fiscal and quasi-fiscal) spending on non-contributory social programs (5.9 percent of GDP).
-- From a static perspective, the existing programs do not channel sufficient resources toward the most needy because of (i) an over-emphasis on regressive subsidies; (ii) a too generous Minimum Subsistence Level for the minority of targeted programs, diluting the allocation of scarce resources by addressing an overly large group, thereby eroding the adequacy of the allocation; and (iii) the use of targeting methods that have a mediocre performance.
-- From a dynamic perspective, the current program mix focuses too much on coping with poverty or vulnerability, and pays too little attention to prevention. For example, (i) social work and community care programs are underdeveloped, (ii) social workers spend too much time on verifying program eligibility (income) and too little on managing the cases of their clients, and (iii) the policy towards orphans relied until recently on costly institutionalization, ignoring preventive services or family-based care.
Box 8.2: Reforming the System of Subsidies for Privileged Citizens: Why and How
Privileges - categorically-targeted subsidies for a wide range of services and goods - are the most important non-contributory transfer in terms of spending. In 2003, only two types of privileges, for housing and utility services and transport, channeled more than six times the resources allocated for transfers targeted to the poor. These privileges are govemed by a complex system of federal, regional and local laws and regulations. It was estimated that there were 156 types of privileges granted to 236 types of beneficiaries only at the federal level (Ovcharova, 2001). A review of the social protection system in three regions (Komi, Nizhy-Novgorod and Moscow oblast) revealed that regional and local govemments complement the system, either by granting privileges to new categories or by granting more generous discounts to those legislated federally.
Such a program archipelago is overly complex, hindering basic functions such as adequate budgeting or monitoring. According to the World Bank (2002)) the system of privileges covers, de jure, 70 percent of the population, but the government can honor only part of its obligations. In 2000, only 36 percent of the households received their benefits. Thus, privileges continue to operate as an unhnded mandate, with adverse consequences for budgetary planning (without knowing with precision the number of beneficiaries and the level of the benefit, it is hard to estimate the volume of the subsidy required to cover this mandate), as well as for the financial position of the service providers.
Overall, about 45 percent of the population benefits directly or indirectly from at least one type of privilege. Irrespective of their type, the largest share of occupational privileges accrues to the richest strata of the population (Figure 8.1). The share of the households from the poorest quintile (poorest 20 percent) in the overall utilization of such subsidized services (in the number of services, not the total amount of the subsidy) varies between 7 percent and 14 percent. In contrast, the richest 20 percent of the population captures between 22 percent and 33 percent of the number of services. The capture of the subsidy by the richest quintile is particularly acute for a few categories with smaller coverage, such as telephone services, spas and holidays, and tax exemption for house repair. For merit-related privileges, the distributional pattem is similar, but the criterion of pro-poorness is not relevant in their case.
Occupational privileges are an expensive way to complement the wages and pensions of the benefit holders. They are highly regressive and should be reformed. These types of benefits are contrary to the social protection principles in a market economy; the system represents an inequitable use of scarce resources, as it does not explicitly benefit the poor and vulnerable. The fact that most benefits are captured by well-off beneficiaries is only one of the factors that determine the highly regressive outcome of the occupational privileges. The other factor, which is equally important, is the regressivity of the benefit (i.e., the fact that the amount of subsidy captured by the, richer households is many times larger than for the poor households). The regressivity of the benefit is embodied in the design of the system. As privileges are subsidies for the consumption of services with a high income elasticity of demand, the better off households will tend to consume both a larger quantity of services and better quality (hence more expensive) services, if available.
A first step to reform the current system of privileges is to ensure an equitable access within the each type of privileged category, be it occupational or merit-based. According to this principle, all privileged persons would be entitled to an equal amount of subsidy. For example, all war veterans should receive the same per capita compensation for transport, based on average utilization rates and tariffs. One way to enforce such equitable access is to migrate from the current system of open-ended subsidies to quota-based subsidies (where each privileged citizen will receive a voucher redeemable up to its face value at the service provider) or by replacing them with flat cash benefits. Information on the utilization of these services and the respective tariffs can be obtained from a sample of providers (supply-side information), from a household survey which collected such information (demand-side information), or from both (allowing to cross-check the two sources).
We illustrate such reform scenario in the case of transport-related privileges, using data on the intake of the subsidy from the NOBUS survey in 2003. From the survey, we found that 55.3 million persons benefited, directly or indirectly from urban, commuter or long-distance transport services, free of charge or with discount (assumed at 50 percent). The total amount of subsidy is 3,155 million Rubles per quarter. The distribution of the beneficiaries is not very regressive: about 17 percent of the users belong to the poorest 20 percent of the population. Yet the distribution of the subsidy is very regressive: the beneficiaries from the poorest quintile received about 8% of the total subsidy, while the richest 20% captured 30 percent.
In Table 8.7, we estimate the average benefit per capita for a recipient household from each quintile. For fiscal considerations, we take the average consumption of the second quintile (where the official poverty line is located) as the value to monetize the benefit, or 146 Ruble/month. This value goes from 123 Rubles in the poorest quintile to 303 Rubles for the richest quintile. In the last two lines, we simulate the distribution of this subsidy across quintiles. This scenario is income neutral for the second quintile, raises the subsidy received by the poorest quintile by l9 percent, and reduces the value of the subsidies in the richer quintiles. Overall, it can reduce the cost of this program by over one quarter, freeing up substantial resources to increase the budget of other, poverty-targeted, benefits. Given that HUS cost is almost twice that of transport benefit, potential budget savings from their rationalization
would be even bigger.
Mediocre Targeting Performance
8.16. The two programs that have the largest share of poor among their beneficiaries are the child allowance program and the decentralized social assistance program (Table 8.8). These programs include about 30 percent and 28 percent, respectively, of their beneficiaries from the poorest quintile. As expected, programs not targeted to the poor have lower targeting performance. The programs that include the smaller share of poor recipients among their beneficiaries are transport privileges (only 13 percent), followed by HUS privileges (17 percent), and the targeted HUS allowance (20 percent). There is substantial leakage for all programs, irrespective of the chosen poverty line - official or alternative. About half of the beneficiaries of the targeted social assistance programs come from the richest 60 percent. This share is even higher for privileges, where 62 percent of the beneficiaries are from the richest 60 percent of the population.
8.17. Almost all programs transfer higher benefits to richer households: the incidence of the benefit is regressive. The only exception is the child allowance program. This is immediately visible in Figure 8.2, where the average benefit of a rich household (from the richest quintile) is compared with the average benefit of a household from the poorest quintile. Among the pro-poor programs, only the child allowance program transfers relatively higher amounts to households from the poorest quintile (compared to the richest). This outcome is due to the benefit formula, which provides higher amounts for very vulnerable beneficiaries (the benefit is doubled for a one-parent household). In the case of the other two targeted programs (the HUS allowance and decentralized social assistance) the average amount obtained by a household from the richest quintile is, respectively, 80 percent and 100 percent higher compared to a household from the poorest quintile. The subsidies for privileged citizens for housing and utility services for transport are the most regressive.
8.18. The targeting performance of the “targeted social assistance” programs in the Russian Federation is substantially lower compared to other “good practice” programs in countries that use either a means test, or a proxy means test. Figure 8.3 compares one dimension of the targeting performance of means-tested programs - the share of funds captured by the poorest quintile of the population - across programs from the United States, Latin America, Central Europe, the former Soviet Union and the Russian Federation. According to this indicator, all targeted programs implemented in the Russian Federation rank well below their comparators. The program with the best targeting performance in the Russian Federation, the child allowance program, channels only 30 percent of funding to the beneficiaries from the poorest quintile, while most comparator programs succeed in transferring between 40 percent and 60 percent of funds to this population group.