Argentina - Cordoba : public sector assessment - proposals for reform (Vol. 2 of 2) : Annexes

Report No. 15132-AR Argentina Cordoba - Public Sector Assessment: Proposals for Reform (In Two Volumes) Volume II: Annexes May 15, 1996 Country Operations Unit II Country Department I Latin America and the Caribbean Regional Office Document of the World Bank:   CORDOBA PUBLIC SECTOR ASSESSMENT: PROPOSALS FOR REFORM ANNEX CONTENTS 1. Cordoba Public Expenditure Review 2. Education Sector 3. Social Welfare Programs 4. Health 5. Public Sector Reforn: Infrastructure and Public Works 6. Provincial Power Sector and EPEC 7. Public Administration 8. Municipal Finance and Intergovernmental Relations in Cordoba 9. Financial Condition and Prospects of Cordoba's Two Public Banks 10. Evaluation and Prospects of Cordoba's Social Security System 11. Cordoba's Tax Structure   ANNEX 1 CORDOBA PUBLIC E)PENDITURE REVIEW 1. An overview of Cordoba's Public Expenditures Background 1. Cordoba spent 2.5 billion pesos in 1994. Over half went towards the social sectors (education and culture 28%; health 16%; social welfare 7%). Security and justice together accounted for almost 20%; and municipalities for 13% (see Table 1). In terms of an economic classification, personnel accounted for fully half of total expenditures. Transfers, composed largely of transfers to municipalities under coparticipation agreements and subsidies to private education, accounted for another 23%. By contrast, capital expenditures were only 12% of total spending (see Table 2). With the fiscal crisis, new investments have largely been limited to donor- funded projects. 2. Gross provincial budgetary expenditures increased in real terms by about 1,015 million or 102% between 1990-1994. Netting out expenditures which were transferred to the provincial budget lowers the increase to 89%. 1 On a functional basis, expenditures from 1990-94 increased most rapidly in social welfare2 (218%), justice (188%), fiscal administration (13 1%), and health (1 15%) (see Table 1). Interestingly, total public resources devoted to the education sector have not increased rapidly. While provincial educational expenditures increased 75% over this period, part of this increase reflects the transfer of federal education to the provincial budget. Once the federal transfer for servicios nacionales transferidos is netted out, the increase falls to only 48%. Expenditures on infrastructure (development) have hardly increased. On an economic basis, personnel expenditures increased by only 74%, while capital expenditures increased by over 170% (see Table 2). 1 If we net out the federal education transfer (servicios nacionales transferidos) of 111.8 million and the 51.6 million FONAVI transfer from 1994 expenditures (see below), the increase drops to 89%. Water supply (EPOS) was also transferred to the provincial budget as part of DIPAS, the merger of EPOS w-ith DPH (irrigation), starting in 1993. At the time of the transfer, EPOS accounted for 64.4 million pesos. However, what was previously EPOS has substantially shrunk since then. Indeed, total 1994 expenditures for DIPAS as a whole was 57.9 million. The percentage split in recurrent expenditures between EPOS and DPH in 1993 was 77%:23%. Assuming the same percentagc split in 1994 implies that ex-EPOS expenditures were 44.6 million in 1994. Netting this last figure out of total 1994 expenditures would further lower the 1990-1994 expenditure increase to 85%. 2 The growth in provincial expenditures on social welfare is actually 353%. However, thc transfer of FONAVI to the provincial budget must be netted out to obtain the real increase in resources flowing to that sector.  ANNEX I -2 - Where Did the Funds Go? 3. The above percentage changes, however, do not provide a clear indication of where the increase in expenditures went since they are obviously starting from wholly different bases which could already be distorted. A more revealing picture can be obtained from Table 4 which provides a breakdown of the distribution of where the provincial expenditure increase went by ministry and category. In terms of economic classification, the bulk of the 1015 million increase went towards personnel (44%) and transfers (26%). Most of the increase in personnel expenses was on permanent (33%) rather than non-permanent staff(9%). Where the funds unfortunately did not go was towards non-wage operations. Expenditures on bienes de consumo hardly increased during this rapid overall expenditure growth. For example, while overall spending by the Ministries of Education and Health increased by 23 1 and 109 million from 1990-94, spending on bienes de consumo only roes by 0.1 and 2.3 million, respectively. This expenditure imbalance has manifested itself in the current situation of hospitals without medicine and technical schools with no training equipment. 4. In terms of ministerial classification, the bulk went to Economy and Finance (28%) for interest and transfer payments (to municipalities), followed by Education and Culture (23%), which spent it largely on salaries and transfers (subsidies to private education). However, as mentioned earlier, much of the latter was funded from federal sources. Desarollo Social accounted for about 10% of the expenditure increase, which it spent largely on trabajos publicos (e.g., housing programs) and transfers. By contrast, Justice received roughly the same amount (9%), but spent it all on salaries. The above figures indicate where significant increases in inputs to the public sector occurred. In many cases, it is not clear what Cordoba obtained from these increased expenditures by way of outcomes. For example, while real health expenditures Cexcluding saneamiento ambienta!) increased by 72% from 1990-93, the number of patients treated did not change. 2. A Diagnosis of Cordoba's Public Expenditures Guiding Principles: Where Should Cordoba be Spending? 5. In this time of extremely scarce public resources, it is particularly crucial that expenditures motivated on efficiency grounds not substitute for or crowd out private provision; expenditures motivated on equity grounds directly benefit the poorest segments of society; and expenditures be evaluated with regard to the sectoral outcomes that they generate. 6. Public expenditures can be motivated either on grounds of improving efficiency or equity. On efWiciency grounds, govemment intervention should be motivated by some underlying source of market failure, e.g., correcting externalities or providing public goods. The government should not provide goods or services which the private sector would otherwise provide. The presence of market failure, alone, however, is not sufficient to warrant public provision. There may exist other policy instruments (e.g., regulatory, tax, or subsidy instruments) which could improve efficiency short of outright provision. Even when there is a rationale for government financing (e.g., through loans or subsidies) because of credit or insurance market failures, for example,  ANNEX I -3 - there may not be a rationale for public provision of goods and services. This distinction between between the role of the public sector in financing v. provision will be particularly important in the ensuing sectoral analyses. 7. On equity grounds, public expenditures should be motivated by a clear incidence on poverty alleviation. Many expenditures may be motivated on equity grounds, but may largely benefit the reasonably well-off and have a limited impact on poverty alleviation. Lastly, once their rationale has been established, public expenditures should be evaluated based upon their outcomes (e.g., percentage of population served in infrastructure, number of patients treated in health, educational outcomes) rather than on their inputs (e.g., number of staff, hospitals, or schools). Evaluating Cordoba 's Public Expenditures Against These Principles 8. Broadly speaking, Cordoba appears to be spending largely on public rather than private goods. With the notable exception of the provincial banks, there is no large public enterprise sector which receives transfers from the provincial budget, including EPEC. In infrastructure, there has been a significant devolvement of responsibility away from provincial authorities (e.g., private contracting of maintenance in roads; municipal and private participation in water). In education, the high proportion of private school students and the provision of subsidies to private schools translate into a reasonable approximation of a system of private provision with public financing. One possible exception is in health, where much is spent on hospitals, a largely private good. 9. However, the equity impact of public expenditures needs to be improved The ex- Ministry of Social Development has over 100 social welfare programs which lack clear selection criteria of beneficiaries and at best have an uncertain incidence on poverty. The FONAVI and FOVICOR housing programs are heavily subsidized but are not oriented towards the poor. Health expenditures are oriented towards free curative care in hospitals, including expensive al/a complejidad services not primarily used by the poor. 10. Furthermore, Cordoba has been spending very inefficiently. This suggests that substantial expenditure reductions can be achieved wilhout significant loss of outcomes in certain areas. Examples of gross technical inefficiency abound. In health, productivity levels in the public hospitals (as measured by monthly consultations per doctor) were about a quarter of those in private hospitals. In education, the use of substitute teachers (because of absenteeism and other reasons) is about five times higher in public than private schools. II. Expenditures on non-wage operations and maintenance, however, have been inadequate and will need to be increased Personnel expenditures have crowded out non-wage current expenditures, which typically have a far higher impact on outcomes. This expenditure imbalance has led to problems in all the sectors reviewed: hospitals lack medicine with which to treat patients; schools have outdated and inadequate materials; roads, water, and electricity have had inadequate maintenance.  ANINIEX I -4- Health 12. The Ministerio de Salud (MdS, 1994 expenditures 301.9 million) maintains a province- wide, vertically linked network of health care facilities which includes health posts and dispensaries, "neighbourhood" hospitals, sub-regional hospitals, and regional general and specialized hospitals. These facilities provide a wide range of preventive, outpatient, and inpatient curative services free of charge. While sector indicators compare favorably to other provinces and countries, health services are lower quality, inefficient and with a poorer equity incidence than necessary. 13. The mixing of the role of the government as both provider and financer of health has led to structural problems in the sector. Expenditures have been extremely technically ineff-icient because it is difficult for the public sector to efficiently manage health care facilities. As an example, the provision of services is 2.4 per person per day in the public hospital examined v. 2.4 per person per houir in the private hospital examined. The true productivity is almost certainly worse than the above production numbers indicate because health professionals may be overproviding certain services: the length of stay in the public system is three times as long as in the private sector. Furthermore, expenditures have had a poor equity impact because (i) the poor are segregated into low quality care with no other options; and (ii) in its role as the supplier of last resort, the govemment ends up providing expensive alta complejidad services for free. The latter have low effectiveness in promoting health, are not primarily used by the poor, and should be paid for through insurance mechanisms. The poor equity impact has been exacerbated as a result of the fiscal crisis. The imbalance between wage and non-wage operating expenditures has meant that patients who cannot afford to bring their own drugs for treatment have been turned away because the public hospitals lack drugs. Expenditure Recommendations 14. The following near-term measures could yield savings on the order of 67.4-82.4 million (plus an additional 34 million if all pre-schools were transferred to the municipalities and an additional 7-8 million if subsidies to private schools were tied to students and private class sizes increased in response). (1) Better management of personnel --e.g., eliminating of ghost workers and double salaries, reducing use of substitutes--could yield substantial savings. In 1994, substitutes as a share of permanent positions were 4-5 times higher in public than priate schools. Since sickness and pregnancy rates are likely to be similar, this difference is largely attributable to better private management practices. If the public sector had performed at private sector levels in their use of substitutes, this alone could have yielded savings on the order of 25-40 million. (2) Eliminating one in ten of non-school level positions in the MEC is unlikely to significantly jeopardize outcomes and would yield 4.4 million. (3) Reducing funding for the MIEC's ancillary programs (e.g., scientific-technological, cultural, adult education) by 20% could save about 10 million. (4) Raising school fees at the secondary and higher levels-equivalent to 15% of those paid within the private sector should raise about 5 million. (5) Another 23 million could be saved by reducing school-level personnel without any drastic increases in student-teacher ratios (10 million from eliminating the vice-director position, 13 million from increasing average class size by 10% and reducing the number of maestros de grado).  ANNEX I -5- 15. Several other measures could result in additional budgetary savings but were not included in the above figures. First, tying the subsidies to students rather than school personnel would be less distortionary and provide incentives to raise student/teacher ratios. Such a consolidation of the same number of students into fewer classes could save the MiEC 7.6 million in transfers at the primary level alone. However, this savings figure is somewhat speculative and unlikely to be ralized in the near-term. Second, the MiEC's pilot program of transferring 19 pre-schools to municipalities would save less than I million from the provincial budget. While a transfer of all pre-schools could yield about 34 million, the absorptive and administrative capacities of municipalities encourages caution. 16. Expenditures on non-wage operations should be increased. Increasing expenditures on bienes de consumo and servicios no personiales by 32 million would only bring these expenditures to 5% of 1994 MEC expenditures, as compared to an OECD figure of 36%. In doing so, Cordoba should ensure that it optimizes federal financing potentially available under the Plan Social Educativo, a national compensatory program to provide books, educational materials, and teacher training to poor regions. Using just a per capita measure, the plan should allocate 10.6 million pesos to Cordoba. 17. Improving quality over the longer-term will require sWlitching expenditures towards non- personnel operating expenditures and financing key investments. Increasing student-teacher ratios at the primary level alone could free up substantial resources. If Cordoba achieved student- teacher ratios equivalent to those of Chile (25), slightly below the LAC average of 26, the net savings to the system would be about 70 million in the public sector alone. Strategies to achieve these savings will require a reconsideration of the current model for school staffing.3 Education 18. Education falls under the purview of the Ministry of Education and Culture (MEC; 1994 expenditures 637.1 million). Cordoba has an extensive educational system of pre-primary to higher and secondary schools which includes participation by the Federal, provincial, and municipal governments as well as a very significant subsidized private school sector.4 Cordoba's student-teacher ratios are 17 and 12 at the primary and secondary levels compare well with with national averages (18 and 7) but remain far below in international comparators (26 and 19 in Latin America; 27 and 19 in East Asia; 22 and 15 in the U.K.). While Cordoba's sectoral indicators--in terms of coverage and enrollment rates, national achievement test scores, and primary repetition rates--compare favorably with national figures, there remains significant scope to increase the efficiency and quality of educational services. To illustrate, raising the ratio of teaching to non-teaching staff were raised from 4:1 to 6:1 would save about 15.6 million, of which 10 million could be obtained from the elimination of the vice-director position cited above. Another option is consolidating classes and reducing the number of maestros de grado. Universities are financed by the federal government.  ANNEX 1 -6 - 19. Technical efficiency has declined. Overall cost per student (total spending on education divided by the number of students enrolled) increased at an annual average of 12% since 1990. This outpaced the increase in each of the education sub-sectors (primary 9.7%; secondary -1.0%; higher 8.6%) suggesting that the largest growth occured in areas not directly related to the delivery of education (e.g., cultural activities, ministerial overhead). Increases in real wages probably accounted for some of this increase. While the number of personnel increased by an annual average of 4% between 1993 and 1995, the wage bill rose has risen much faster. Average wages rose 19% in real terms between 1994 and 1995.5 20. The substantial imbalance between personnel and non-wage operating expenses is likely to have eroded the quality of education. Of the Ministry of Education and Culture (MEC) budget, effectively 95% goes to personnel expenses (73% are own personnel expenses and 22% are transfers to private institutions to fin-ance teacher and administrative salaries). Of the scant 3% going towards non-personnel operating expenses, almost all goes to non-personnel services--e.g., transportation and utilities--resulting in virtually no financing of equipment, educational materials, and other inputs. However, most studies indicate that it is precisely these types of expenditures-- rather than personnel and class size--which are the most effective determinants of effective learning.6 By comparison, OECD countries spend 55% on personnel and 36% on non-personnel operating expenditures. 21. Cordoba has in place an educational system which approximates public financing with private provision. About 90% of private schools receive public subsidies and account for over 20% and 40% of primary and secondary students. A total of 150 million pesos was budgeted in 1995 for subsidies to private schools. While the subsidy is currently tied to the number of school personnel rather than students and therefore distortionary, it is worth noting that the subsidy per student in private schools is lower than the cost per student in public schools (463 v. 618 in primary; 1,088-1,292 v. 928-1,367 in secondary). Public provision has a worse fiscal impact than public financing. However, once private school fees are added, the total per student cost is higher in the private than public sector. Social Welfare Programs 22. The ex-Ministry of Social Development (MSD; 1994 expenditures 165 million) accounts for the lion's share of govemment spending in social welfare programs. Almost half of the ministry's budget is financed from transfers from federal programs, e.g., PROSONU (child nutrition) and PROSOCO (social welfare). At the end of 1994, the MSD had over 100 programs largely carried out under the broad categories of general social assistance (about 30% of the budget; transfers to individuals and institutions for a broad range of purposes, e.g., food aid, purchase of medicines, rental assistance) and community and family promotion (20%; support for 5 However, the budget figures may understate the number of people hired, and hence inflate increases in real wages. 6 See, for example, World Bank (1995). Priorities and Strategiesfor Edutcation and Hanushek, E. (1995) "Interpreting Recent Research on Schooling in Developing Countries." WVorld Bank Research Observer, vol. 10, no. 2.  ANNEX 1 -7 - the Provincial Council for the Protection of Minors, CPPM, and assistance programs for the elderly, handicapped, etc.).7 About half of the funds went to institutions and about 44% to individuals. The average individual subsidy was 1,414, which represents about 18% of per capita GPP. The program of individual subsidies was unmonitored in terms of the discretion in choosing recipients and benefit levels. Institutional subsidies (to NGOs and municipalities) were also unmonitored. 23. While many of the MSD programs appear to address legitimate social needs, the equity impact of expenditures is at best uncertain because of (i) the absence of clear criteria for beneficiary selection; (ii) the lack of reliable information on the type and number of beneficiaries and the level and impact of benefits delivered; and (iii) the dispersion of programs. 24. PAICOR (Programa de Asisiencia In?iegral de Cordoba; 55 mil]ion 1994 budget), a provincial school lunch program under the Ministry of Coordination, currently reaches 130,000 children or about 1/3 of the primary school population. Beneficiaries are selected at the school level. Parents decide whether to register their children in the program, and teachers determine which children are enrolled. While PAICOR appears to have a reasonable equity impact based upon geographical targeting (poorer departments have a higher share of students enrolled in PAICOR), it appears to be run rather inefficiently. While 48% of its budget went to food rations, almost the same amount (46%) went to personnel expenses. Expenditure Recommendations 25. The government has already made significant advances in recent months to rationalize its social program expenditures. While staff was cut from 2,414 to 1,932 when the MSD was s-ubsumed within the Ministry of Institutional Affairs as the Secretariat of Social Development (SDS), personnel remains excessive. Reducing the use of substitute teachers in the CPPM and excess administrative personnel should save on the order of 14 million. The danger of misuse and the difficulties in efficiently screening and monitoring the individual subsidy program call for its termination. Individuals with urgent needs can instead be directed to on-going programs and institutions. Eliminating these transfers would save 21 million off 1994 budget levels.8 A bottom- up assessment of PAICOR suggests that the same service could be provided for about 47 million, saving 8 million off 1994 budget levels, with about 12 (rather than about 25) million going to personnel. 7 Housing programs previously accounted for the remaining 50%, but havc been transferred elsewhere and discussed below. 8 It is worth noting that the individual subsidy program was shut down in thc second half of 1995, saving 11 million, at virtually no public outcry.  ANNEX I -8 - Roads 26. The Direccion Provinicial de Vialidad (DPV, 1994 expenditures 64 million) administers, maintains, and improves nearly 60,000 km of provincial road. Of these, the vast majority (82%) are earth roads with less than 100 vehicles per day; 12% are unpaved roads with between 100-500 vehicles per day; and 6% are paved with generally more than 300 vehicles per day. While conditions of Cordoba's paved road system are only slightly inferior to those of other provinces, the situation of the unpaved network is somewhat less satisfactory. 27. Compared to the road directorates of Buenos Aires, Misiones, and Santa Fe, Cordoba ranks best on virtually all performance indicators, e.g., spending per km of provincial road, length of network, staffing efficiency (staff per 1000 km is 7, compared to 54 and I ll for Buenos Aires and Santa Fe). DPV had 402 employees as of October 1995, down 30% from a year earlier, and intends to reduce staffing further to 340. 28. The road system closely approximates public financing with private provision. One of the reasons for Cordoba's efficiency is its high use of private contractors for construction and maintenance. Most maintenance on the unpaved network is carried out by 220 conisorcios camineros managed by the residents of the area served. Expenditure Recommendations 29. Cordoba already has in place a very efficient road management system. In view of the impending transfer of national roads and the beginning deterioration of the quality of its existing network, it would be undesirable to cut resources and staffing further in 1996.9 Indeed, increasing the contribution to maintenance by the consorcios by 3 million to reach the level of previous years would be desirable. 30. The scope for further cost recovery appears limited. The contracts between the DPV and the consorcios already specify a local contribution of 20%. Toll collection is difficult on many provincial roads because of low traffic volumes. Even on the two pilot toll roads, collection costs and VAT consume about 30% and 20% of toll revenues, respectively. Water, Sewerage, and Irrigation 31. The Direccion Provincial de Aguay San7eamienito (DfPAS; 1994 expenditures 57.9 million; net deficit 15.5 million) is responsible for water management, including the operation of the provincial water supply, sewerage, and irrigation systems. While access to piped water supply is above the national average (84% of Cordoba's population in built-up areas in 1991 had water supplied through house connections, compared to a national average of 72%), sewerage 9 As part of the reform of the administration of the national road sector, a net transfer of about 10,000 km of national highways to the provinces is expected by mid-1996. This could increase Cordoba's paved network by more than 10%.  ANNEX I -9- connections are particularly deficient (18% in Cordoba, compared to an already very low national average of 38% and a Latin American average of 53%). Virtually all public sewerage has already been transferred to the municipal level. 32. Cordoba's staffing per 1,000 water customers in 1992 was near the average of comparable Argentine provinces. Despite abysmal cost recovery (a collection rate in water of 60%, compared to over 90% for Buenos Aires, Mendoza, and Santa Fe; and less than 10% cost recovery in irrigation), DIPAS' operating deficit has now been virtually eliminated, apparently at the cost of inadequate maintenance and depreciation. The estimated backlog of rehabilitation and similar works is estimated at 120 million. Expenditure Recommendations 33. A far greater effort needs to be made in improving collections. In principle, DIPAS could collect about 60 million per year more than today, including 28 million for unpaid water bills; 4 million from an estimated 30,000 unregistered water connections; 15-20 million for unregistered private wells; and 3-6 million from EPEC for hydro-generation fees. The theoretical total of accumulated back payments due DIPAS is 139 million, covering unregistered services plus 60 million in accumulated unpaid water bills. While obviously difficult to collect, a more concerted collection effort is certainly called for in this time of fiscal crisis. In irrigation, improving the pricing structure could provide some savings since less than 10% of the cost of providing bulk water is recovered from user charges. 34. It is estimated that a restructuring of DIPAS could reduce personnel costs by about 15-20 million (about 490 redundant staff in water and 230 in irrigation). Several other particular expenditure items bear investigating. VAT payments to the federal government for payments niot received amount to 6 million per year. Furthermore, it is not clear why the provincial bank demands a service fee of 8.2% for receiving water payments (over 3 million per year), especially when it only charges the Housing Institute (IPV) 1.7% for receiving rent payments. Paying the IPV rate would save another 2 million. 35. Of the 368,000 water connections which still remain with DIPAS rather than the municipalities, 327,000 are part of the Cordoba city system which is to be privatized. However, urgent reforms should not be postponed pending privatization since it is likely that a government entity remain responsible for the city water supply until at least 1997. While tender documents for the privatization were issued in 1994, legal objections were raised by the private consortia submitting bids and the matter is currently under review in the courts.  ANNEX I _ _ ~~~~~~- 10- Housing 36. The Instituto Provincial de la Vivienda (IPV; 1994 expenditures 76.3 million) administers several housing programs in Cordoba, which are funded primarily through the federal FONAVI program (51.5 million). The housing programs are heavily subsidized and charge a uniform interest rate of 8.7%, compared to commercial mortgage interest rates of about 18%. 10 Despite their high subsidy component, the housing programs have terrible incidence and primarily benefit the middle class: 79% of the loans go to households with a monthly income of over $560. The outstanding loan portfolio is 721 million. Expenditure Recommendations 37. IPV's record of loan repayment is abysmal and desperately needs improvement. 29% of its clients are behind by more than 45% of billings; another 15% by 25-40%. Only 34% are up- to-date with their payments (behind by less than 10%). The accumulated total of overdue repayments is about 50 million. Increases in loan repayments can help reduce the provincial contribution to the IPV budget, which was 9.2 million (6 million from the budget and 3.2 from FOVIS, the provincial vehicle tax) in 1994. 3. Expenditure Allocations 38. The macroeconomic projections indicate that an expenditure reduction of about 270 million from 1994 levels is required to close the fiscal gap. Where should these cuts fall? From an economic classification standpoint, as a first approximation it appears preferable that the reduction fall on personnel expenses given the extreme imbalances between personnel v. non- personnel expenditures discovered in virtually every sector evaluated. As discussed elsewhere, the preferred scenario for reducing personnel expenditures combines a 13.6% average salary reduction with a 8.6% employment reduction. Since the salary reduction is based on pay scale rather than sector or ministry, where should the employment reduction take place? 10 IPV has an annual construction program of over 60 million. Taking this as the annual "steady state" flow of new loans suggests implicit annual subsidies of 5.6 million.  ANNEX I 39. The mission evaluated the health, education, social welfare, and infrastructure sectors. Together with the PAICOR program under the Ministry of Coordination, the ministries corresponding to these sectors accounted for 54% of 1994 expenditures. For these areas, Table 5 summarizes the near term expenditure recommendations by type and sector. Expenditure "quick fixes" (type I) alone, which include improved cost recovery and elimination of some subsidies (notably individual subsidies under the SDS), can yield substantial savings on the order of 55 million without any reduction in personnel. Relatively conservative estimates were used to arrive at the numbers presented in Table 5. For example, only 10% of the estimated annual uncollected or unbilled amount of 60 million in water is assumed to be collected; housing loan repayments are assumed to increase only by 10%. Furthermore, no recovery of the stock of accumulated back payments in water (139 million) or housing (50 million) is included in these figures. The potential for improved cost recovery and collections is very large and warrants further scrutiny. 40. Ancillary reductions (type II) include reductions in administrative personnel, cutbacks in ancillary programs, and especially reduced use of substitutes (e.g., in education). These would save over 60 million. Reductions among line staff (e.g., in hospitals and schools; type III) could save another 75 million. However, there is also a significant need for increases in non-wage operations in several sectors, which total 54 million. Undertaking these additional expenditures would reduce net savings from 191 to 137 million. 41. Provincial budgetary savings on roughly 91 million from transferring health and education services to municipalities are not included in the above figures. I I The govemment should be extremely wary about trying to resolve the expenditure crisis by relying excessively on tranferring services to the municipalities. While this may ease provincial budget pressures in the near-term, this may only postpone needed efficiency improvements and create the beginnings of an impeding fiscal crisis at the municipal level in the future. 42. An 8.6% employment reduction translates roughly (at average wages) into a 110 million reduction in 1994 personnel expenses. The personnel expenditure reductions in Table 5 alone total approximately 136 million.12 However, from the standpoint of intersectoral allocation, it is completely unwarranted to let the brunt of the adjustment fall only on these sectors, especially there are likely to be significant expenditure savings in the sectors not evaluated. The government estimates that the planned transfer of health facilities to thc municipalities could save around 90 million from the provincial budget. A pilot transfer of prc-schools wNould save less than one million. 12 Type II and type III reductions (136 million) are virtuallv all personnel.  ANNEX I - 12 - 43. Again, the key criterion should be the outcomes foregone from any expenditure reduction. As a first approximation, the ancillary (type II) reductions are likely to have a lesser impact on outcomes since they consist largely of expenditure savings from retrenching overhead and administrative staff and more efficient use of substitute teachers. Limiting the first round of personnel reductions to this type of reduction in the evaluated sectors would imply a target reduction of [49.1] million in the remaining, unevaluated sectors, which represents a reduction of only 4.2% of total 1994 expenditures among these other ministries. If ancillary reductions in these other sectors are insufficient to meet this target, then attention should be again focused on the direct line personnel reductions (type III) cited in the evaluated sectors. 44. While an in-depth analysis of the other sectors is difficult and time-consuming, the distribution of the 1990-1994 provincial expenditure increase presented in Table 4 may be a useful starting point. The table indicates where significant increases in inputs occurred. For example, there were large increases in personnel expenditures at the Ministries of Justice and Govermment. If outputs and outcomes (appropriately defined) did not increase or improve with increased inputs, then it is reasonable to expect that they will not deteriorate significantly with decreased inputs. 45. The above personnel expenditure reductions do not take into account the increased savings from the expenditure quick fixes. However, these additional savings should be allocated towards critical non-wage operating expenditures, rather than used to mitigate personnel reductions which may be desirable for long-term efficiency reasons.  ANNEX 1 - 13 - LONG-TERM RECOMMENDATIONS 46. Over the long-term in each of the sectors, the government should continue to pursue the guiding principles laid out in the diagnosis of Cordoba's public expenditures: (1) Provide public rather thani private goods and services. In the infrastructure sectors, the government should privatize the Cordoba city water system and EPEC in electricity. (2) Separate out the role of government as provider v. financer of public goods. In roads, this has already been accomplished to a noteworthy degree. In education, a sound basis for this has been established with the provision of public subsidies to private schools. Basing educational subsidies on the number of students rather than school personnel would be an important step towards moving further in this direction. In health, the government should finance health insurance rather than provide free curative care in public hospitals. In health and education, where there is a large private sector co-existing alongside the public sector, separating financing and provision would also help ensure that public provision does not crowd out private provision. Providing financing--be it through health insurance or educational subsidies--would allow patients and students to "vote with their feet" about hospital and school performance. Both public and private facilites would be subject to a market test which should improve efficiency. If the former were inadequate, they would naturally and endogenously shrink. (3) Subsidized, equity-motivated interventions shotuld have a clear benefit onl the poorest segments of society. Targeting, screening and evaluation mechanisms should be established for all social programs, including housing as well as those under the SDS. Health 47. Providing health insurance rather than free care in public hospitals would solve the three structural problems in the sector cited earlier and should yield long-term savings of 61 to 150 million. (1) Low quality: with patients with mobile dollars, the provider has more incentive to be attentive and responsive. (2) Hospital management: the government is relieved of the onerous task of micro-managing hospitals and health centers. The government would instead choose and manage the payment scheme. (3) Alta complejidad: Setting up a system of financing specific activities will facilitate decisions about which services are and are not covered. It will also stop private insurers from pushing the costs from beneficiaries who require these expensive treatments onto the public sector.  ANNEX 1 - 14 - Education 48. Strategies to increase the efficiency and quality of educational services include (I) effective control over wage bill and personnel. Immediate investments in computers and technical assistance to modernize personnel administration should be a high priority. (2) Increasing student teacher ratios to a minimum which still guarantees relative quality. This will require school consolidation, provision for student transportation, and reorienting personnel towards classroom rather than non-classroom positions. A necessary first step is to analyze on a school-by-school basis the unit costs associated with the provision of services. (3) Reorienting sector expenditures to favor educational services at the school-level rather than MEC overhead. This should translate into greater school-level control over resources and increased attention to non-personnel inputs to educational quality. (4) Achievement of greater equity in the system whereby richer students pay more than the poor and greater levels of public subsidies are applied to basic than higher levels (e.g., adult training, fine arts schools). (5) Increasing the quality and relevance of education by modernizing curricula and educational materials. Over the longer-term, increasing student- teacher ratios at the primary level alone to those of Chile (25), slightly below the LAC average of 26, could save 70 million in the public sector alone. Social Welfare 49. Elements of a strategy for social programs include (1) anticipating increased needs for social assistance due to the crisis. Select key programs to protect, such as PAICOR, which provides both basic nutrition and creates incentives to stay in school. (2) Optimizing resources from Federal social programs, e.g., PROSONU and PROSOCO, which will continue to be an important source of financing of core social programs. (3) Establishing targeting, screening, and evaluation mechanisms for all social programs. An initial ranking of cost per beneficiary of the various social programs is necessary to build a priority setting system. This should be complemented with the establishment of selection criteria for screening potential beneficiaries. (4) Review the program methodology and internal costs of PAICOR. (5) Consolidate SDS programs to key groups and areas of action to heighten impacts. (6) Evaluate elements of a social safety net, including the coordination of on-going programs in various sectors, monitoring of the poverty situation, identification of key vulnerable groups, and leveraging of resources for basic elements of the safety net. Roads 50. Cordoba already has in place a very efficient road management system. Increasing maintenance expenditures will be important to arrest the beginning deterioration of its road network.  ANN'EX 1 - 15 - Water,Sewerage, and Irrigation 51. While the privatization of the Cordoba city system should be pursued, urgent reforms should not be postponed pending privatization since it is likely that a government entity remain responsible for the city water supply until at least 1997. It is recommended that a program be devised to register all water customers and collect as much as possible of the outstanding payments. Because of DIPAS' past record in non-collection and because surveys will be required to identify unregistered wells and water connections, it would be most effective to define this program with the assistance of an outside firm. It would also be useful to undertake a management and technical audit of DIPAS to restructure it and improve its efficiency. In irrigation, the pricing structure needs to be re-evaluated since less than 10% of the cost of providing bulk water is recovered from user charges. 52. While sewerage is under municipal purview and off the provincial budget, it is worth noting that an investment of 250-300 million would be required to provide sewerage connections for just half of the population living in Cordoba's built-up areas, which would still be below the Latin American average of 53%. Housing 53. Given the large subsidy component of its loans, IPV should ensure that its housing program directly benefit the low rather than middle-income population. Otherwise, the program should be run on a commercial, unsubsidized basis.13 IPV is now elaborating new standards, and outside expertise on the financial, management, and planning aspects of low-income settlements sJhould be provided. IPV should also significantly improve the abysmal loan repayment record An upgrading of IPV's data processing capabilities to improve its billings and collections should be given high priority in view of the potentially high payback.14 Over the long term, the goal should be to make IPV a self-sustaining agency which does not require any contribution from the provincial budget, especially in view of the regular, annual federal FONAVI transfer that it receives. 13 The terms of the loans it provides are under provincial and not federal purview. 14 IPV has defined a S 100,000 upgrading program which in principlc should be eligible for funding under the World Bank-supported Provincial Development Project.  -16 - ANNEX 2 EDUCATION SECTOR 1. Cordoba has an extensive educational system which includes participation by the Federal, provincial and municipal governments as well as private schools. In general, Cordoba has better coverage and enrollment rates than the average for Argentina, and is approaching its goal of ensuring nine years of universal education. To achieve these ends, spending levels are substantial, including significant support from the provincial budget to private sector schools. The fiscal crisis facing the province in 1995 threatens to undermine sector achievements in terms of quality and coverage unless gains can be made in the efficiency of expenditures and reallocations made toward priority areas. This chapter reviews the situation in the education sector in Cordoba and provides recommendations on short-term strategies for addressing the fiscal crisis as well as suggested elements of a medium-term strategy for the sector. A. Analysis of the Education Sector in Cordoba Leading up to the Crisis Coverage 2. The table below shows the overall distribution of students by level since 1989, excluding students enrolled at universities in Cordoba, which remain under the tutelage of the Federal Government'. Cordoba: Total Enrollment by Level 700000 'i . ... i' .i .; . ....;,t-" '-' . ....... '. " ',''-' 'l.:''. i' '''E ' . . " ; 0'i i.!,'.....' " T . '-' ', ",'. ' i 'l' - '; 600000 500000 : :- :-: -::::: E:5 ::::::::-:: erdr : 7:::X :: : .n 400000- = _1 E3~~~~~~~~~~~~~~~~~~~~~~~ Higher 5300000 _C yecondary z _ _ _ - la ~~~~~~~~~~~~~~Pnmary 200000 - _ llll-l1 Pre-Primary 100000 1989 1990 1991 1992 1993 1994 3. At the pre-primary level, almost half of all four-year-olds and over 90 percent of all five- year-olds attend pre-school. The absolute number of students has grown by about 4 percent per year over the last ten years. As shown in the table below, seventy-two percent of pre-primary students attend provincial public schools, twenty-two percent attend private schools, with the remaining six percent evenly divided between Federal and municipally financed pre-schools. This review does not include information pertaining to thesc universities since they are not financed by the provincial government.  - 17 - ANNEX 2 4. At the primary level, where the bulk of the student population is concentrated, coverage has been universal for over two decades. With a gross enrollment rate of 98 percent and a net enrollment rate of 96 percent, the system is one of the more efficient in terms of student flows for Argentina which has a national average of 115 percent and 90 percent, respectively, and compares favorably with OECD countries. Enrollments have decreased slightly over the last five years, which may be caused by some combination of improvements in the repetition rate, slight contraction of coverage in the more remote areas of the province, or eventually, shifts in the age structure of the population. The projections of the future are for slightly decreasing primary school population. In terms of service providers, the distribution is almost the same as that of pre-primary, with about three-quarters enrolled in provincial public schools. In addition to the standard primary schools, there are about 80 special education schools serving approximately 5,541 students with learning disabilities (2 percent of the primary school population), as well as 7,700 adults enrolled in primary education programs. 5. At the secondary level, gross and net enrollment rates are 75 and 64 percent, respectively, slightly above the national averages of 74 percent and 56 percent. Although far below OECD levels (over 80 percent for net enrollment rates), Cordoba is well above the LAC average for gross secondary enrollment of 48 percent and the average of 54 percent for countries of comparable income levels. The secondary system is diversified into tracks: general (or bachillerato), commercial, technical and agricultural. Forty-four percent of students are enrolled in the commercial track, followed by 29 percent in the technical track, 22 percent in the bachillerato and 4 percent in agricultural studies. In addition to the 160,000 students enrolled in secondary schools, about 13,000 are enrolled in special secondary school courses for adults. 6. Expansion of absolute enrollment has been most significant at the secondary level, although the actual figures are greatly influenced by the transfer of Federally-financed secondary schools to the provincial government starting in 1993. In total, 29,000 public school students and 46,000 private-subsidized students were transferred in 1993, effectively doubling the number of secondary students under the provincial government. With this transfer, over 40 percent of all secondary students in Cordoba are now attending private, subsidized schools, almost twice the national average in 1991.  -18 - ANNEX 2 Cordoba: 1994 Enrollment by Type of School 100% 7.7 ..... 1 o .S...., , i -. T ... , !,, ........ ,#: ,,ti;,......... f :i; ..... . .. _ E :- ........... . ....... ... ........i S . 0%~~~~~~~ . ;..., . -... :. -.f. .. '_''-fV0-;t:.. :0.........?tt : 0- 60% - ^. .0.: _ . -- . i :- $ - _ : t: :: -: -; : - ...:. .... 50% - - -}- _ E00 ;-;; li -t-0 NMunicipal . . _ ,.. E,,..:: _ %:T:i ........... ... .. .................- . :..- f.... 7. i 40% -0 ; Q t;t _ 0; 00ttt0t0t i; r Federal 20% U 1|1 1 l Provinecial 70% Pre- Primary Secondary Higher/No Primary n-Univ 7. Higher education financed at the provincial level covers about 9,300 students distributed among disciplines as follows: over 60 percent are enrolled in teacher training courses, with the remainder in social sciences, technical training, and other areas such as nursing and police academies. Average annual growth rate of enrollment over the last six years has been approximately seven percent. Educational Quality and EfEMciency 8. Educational Achievement. Overall, Cordoba ranks fairly highly in academic achievement compared with other provinces, reflecting in part the better socio-economic standing of the population. Based on results of the 1994 national academic achievement tests which cover language and math at the primary and secondary levels, Cordoba ranked fifth and sixth among provinces in primary level language and math, and eighth and ninth in secondary level language and math. The table below presents the 1994 test scores. Of note, Cordoba is slightly above the national average score at the primary level and slightly below at the secondary level. Within the province, private schools sampled scored well above the Cordoba and national averages. Public schools scored below these averages, with the wedge between public and private schools highest at the primary level. Primary schools in urban areas outperformed only slightly their rural counterparts. Nonetheless, these results disguise a wide disparity between individual schools, as shown in the gap between the highest and lowest scores for each subject.  -19 - ANNEX 2 Cordoba - National Achievement Test Scores 1994 100.00% 90.00% 80.00% - 70.00% i Natl Avg 60.00% : Cordoba Avg PrimaryC- Primary Secondary Secondary Language Math - - Math Language 9. Internal Effciency. The efficiency of the system in terms of student flows is quite good at the primary level. Estimates of repetition rates suggest that overall rates have dropped by half over the last ten years, from about 8 percent in the early 1980's to 4 percent at present, and well below the LAC average of 14 percent. The significant fall in the repetition rate for first graders from 18 to 10 percent may be due in large part to the expansion of access to pre-school over the last ten years, particularly in the rural areas where the percentage of children entering first grade who had attended pre-school rose from 19 percent in 1982 to 74 percent in 1994. Most students who enter first grade eventually complete the primary cycle, with the low wvedge between the net and gross enrollment rates reflecting the gains in reducing repetition. 10. The system is less efficient at the secondary level, a situation which is a major concern for educational policy in Argentina in general. While estimates vary, about 50 percent of students who enter secondary school drop-out before competing the five year cycle, with the highest rate of drop-outs occurring in the first years of secondary school Private schools tend to have better results, but still exceed forty percent. Public technical secondary schools have the highest drop- out rate at about 60 percent for the 1987/91 cohort. Repetition rates range from about 15 percent to one percent depending on the type of school and year in the cycle, again with the highest rates in the initial years of the cycle. This situation is cause for concern on two fronts. First, as secondary school is often the last stage before entering the labor market, labor market skill levels and individuals' earnings potential are thwarted. As an economic efficiency issue, this situation causes a "lumping" effect in the first years of the secondary cycle, followed by thinning class size in the upper years, which does not lend itself to optimizing on infrastructure and personnel. 11. Student/Teacher Ratios. As a measure of an efficient use of its teaching force, student/teacher ratios are quite low in Argentina in general, and while Cordoba is slightly "better" than national averages, the configuration of the system is relatively high cost. Annex I provides comparative data on student teacher ratios2. To summarize, Cordoba's primary level student/teacher ratio of 17 is slightly lower than the national average. The average for LAG and East Asia are 26 and 27, while Cordoda's ratio is nearer to the US rate of 16. Within the primary 2 Please note that student teachcr ratios are not equivalent to class size since there arc personnel classified as teaching staff who do not perform classroom functions (directors, vice-directors, librarians, etc.).  - 20 - ANNEX2 subsector, the highest student/teacher ratios are found in municipal schools due to the concentration of these schools in densely-populated marginal urban neighborhoods, while the private sector has the lowest ratios. As the private schools are almost exclusively concentrated in urban areas, the low student-teacher ratios may account for some of the differences in quality and test scores between the public and private primary schools. 12. Student/teacher ratios fall at the secondary and higher education levels, to 12 and 9, respectively'. These low averages disguise large variations between the overcrowded initial years of secondary and the upper levels, though disaggregated estimates are not available. Again, Cordoba is better than the national average of 7 at the secondary level, but well below international comparators which range from 15-16 for Chile, the US and UK to 19 for the LAC and East Asia averages. Evolution of Spending on Education 13. Overall Levels. Provincial government spending on education is significant. The 1995 budget of the education sector as a whole (including cultural activities) totaled 807 million pesos, or about 30 percent of the provincial budget. The share of education in the provincial budget has hovered between 28-33 percent over the last ten years. This is equivalent to 3.4 percent of the Gross Provincial Product, above the average for Argentina where education spending by the provinces amounts to 2.5 percent of GDP and total education spending 3.12 percent of GDP in 1992. In comparison, educational expenditures as a share of GNP are about 4 percent for LAC and 5.2 percent in OECD countries. Within the education sector, 40 percent is spent on primary education, 30 percent on secondary, 27 percent on activities not specifically linked to a level, 2 percent on cultural activities and only 1 percent on higher education. The low amount on higher education is a function of the Federal responsibility for financing universities, with the provinces retaining relatively lower-cost non-university higher education such as teacher training programs. 14. Distribution of MEC Budget. Within this envelope, the Ministry of Education and Culture (MEC) accounts for 89 percent, or 715.6 million pesos. Within the MEC budget, personnel expenses account for 73 percent, transfers to private institutions 22 percent (to finance teachers and administrative salaries); non-wage operating expenses account for three percent and investment less than two percent.4 With effectively 95 percent of the budget going to personnel expenses in one form or another, Cordoba runs the risk of an erosion in the quality of the educational services provided. Moreover, of the scant three percent of non-personnel operating expenses, almost all goes to non-personnel services including transportation, utilities, and communications, leaving almost no financing of equipment, educational materials and other inputs. In comparison, OECD countries spend about 9 percent of the education budget on investment, 55 percent on personnel and 36 percent on non-personnel operating expenditures. Although overall spending levels are significant in Cordoba, the pattern of spending, with Estimates of student/teacher ratios at the secondary level are less precise due to the large number of part-time teachers and the payment by 'horas catedras'. 4 This understates total investment in education since many of the education infrastructure works are carried out by the Ministry of Public Works, particulary for foreign-financed projects. This report does not cover the school construction and rehabilitation carried out under the Minsitry of Pubic Works, including: (i) the US$ 12.3 million slated for school rehabilitation to be financed by the IDB/IRD through the Provincial Reform Loan; the majority of which appears oriented to the larger schools in the capital; (ii) new construction, rehabilitation and expansion fianaiced with provincial funds, and (iii) works carried out through Federal-level programs such as the Plan Educativo Social. In 1995, the MPW allocated 2.5 million pesos to primary education, 3.2 million to secondary, and 0.2 million to other education/culture related physical investments.  - 21 - ANNEX 2 personnel crowding out all other expenditures, not only affects quality of education, it tends to erode the productivity of the personnel employed. 15. How have educational expenditures evolved over time? Annex 2 presents the real growth in the education budget over the last six years. Since 1990, average annual growth of expenditures in real terms has been about 14 percent. The highest rates have been for "unspecified" expenditures, which given the budget nomenclature, are difficult to identify. Cultural expenditures have also grown dramatically, albeit off a fairly low base. Real spending on primary education has seen the lowest rate of increase at 6.6 percent. The 22.2 percent average annual increase in secondary education reflects, in part, higher enrollment growth rates and the absorption of the Federal schools within the provincial budget. 16. To achieve a more accurate picture of the evolution of spending on educational services, the table below traces the increase in real cost per student per level over the same years in order to net out the effect of expanding enrollments. Cost per student has increased at an annual average of 8.6 percent since 1990. As presented in Annex 1, the student/teacher ratio at the primary level has fallen by about 10 percent since 1990, driving the increases in per student cost at this level. At the secondary level, per student costs have fallen slightly since 1990, with the changes during the intervening years largely a statistical artifact of absorbing budgets and enrollments from the Federal transferred schools. In line with the findings in Annex 3, overall spending on education (total spending on education divided by the number of students enrolled) at an average of 12 percent has grown faster than any of the education subsectors. In other words, the largest growth in real spending has occurred outside of the services directly related to the delivery of education. These include cultural activities, scientific and technological programs, MEC overhead, and the like. 17. What has driven the increasing unit costs of the system? There are a number of possible factors involved: (i) the enrollment base is shifting increasingly to secondary education, which has higher unit costs, (ii) the MEC has expanded its range of actions, taking on new programs and activities not directly related to enrollment levels; (iii) the hiring of personnel has outrun the real needs of the system; and (iv) real wages have risen. The relative weight of each factor is difficult to assess due to the lack of detailed information and the poor quality of statistics that do exist, particularly on personnel and wages. . s.s . ...... . .. . ..... ..... . (in constant 1991 pesos) Yrly. Ave. 1989 1990 1991 1992 1993 1994 1990 - 1994 Primary 287 397 427 492 535 551 Secondary 777 1155 1115 1211 643 1109 Higher 578 691 764 912 980 927 Total 439 621 674 893 922 919 Annual growth 1990 1991 1992 1993 1994 Primary 38.5% 7.5% 15.1% 8.9% 3.0% 9.7% Secondary 48.5% -3.4% 8.6% -46.9% 72.4% -1.0% Higher 19.6% 10.6% 19.3% 7.4% -5.3% 8.6% Total 41.4% 8.5% 32.5% 3.3% -0.4% 12.0%  - 22 - ANNEX 2 18. To attempt some level of analysis regarding (iii) and (iv), Annex 3 presents the personnel structure of the MEC over the last three years. The bulk of all employees are teachers, followed by general service personnel (mainly school level staff such as janitors, guardians etc.) and administrative positions. The number of personnel, not including teaching hours, increased by an annual average of 4 percent between 1993 and 1995. The wage bill has risen much higher. Between 1994 and 1995, average wages rose 19 percent in real terms. Annex 3 also presents a disaggregated look at the growth of average wages by area of education and MEC administration. The wide variation in the figures reveals the largest increases in areas outside of the programs related to delivery of educational services. Although the evidence does suggest significant increases in real wages, the quality of the data limits the strength of conclusions that can be drawn. In some cases, the budget figures may understate the number of people hired, and hence inflate increases in real wages. 19. The lions' share of personnel expenditures go to teachers' salaries, the regulations for which are established by the Estatuto Docente. The teachers' salary scales establish a base salary upon which a number of additional payments are factored in. These include: education level, seniority pay, hierarchy, overtime, family allowances, payments for training and educational materials, premiums for hardship posts, and productivity bonuses. On average, the base salary accounts for only about 40 percent of the gross salary. On top of this, the MEC must contribute an additional 26.5 percent in employer payments related to pensions and insurance. While teachers were exempt from the wage cuts imposed by the Emergency Law, they did receive a 20 percent temporary reduction in the salary amount linked to seniority. The complexity of the salary system makes analysis of real wages difficult. It appears that automatic raises for seniority, as well as the expansion of 'pseudo-wage' benefits above the base wage, have driven increasing costs to a greater degree than adjustments in the base wage levels. The Role of the Private Sector 20. Expansion of private sector coverage has been based largely on public sector financing. The Ministry of Education provides financing for personnel to the non-profit private sector schools, of which about 80 percent are administered by the Catholic Church. The subsidy system functions based on a sliding scale depending on the level of school fees charged. Although the statutes limit the amount of the school fee allowable per level of subsidy, in practice many private schools find ways of levying additional charges to supplement school resources while maximizing the subsidy from the Government. Teachers are paid the same and the same standards for number of teachers and administrative positions per size of school apply. A separate MEC department supervises the private system at all levels. 21. A total of 150 million pesos was budgeted in 1995 for subsidies to private schools, over half of which was allocated to the secondary level. Only about ten percent of the private sector functions without any subsidies from the state. Of those schools that receive subsidies, 93 percent receive the maximum amount of subsidy allowable (100 percent of the value of teachers' salaries). As presented in the table below, the per student subsidy to private schools is less than the per student cost of public education. This is due to the fact that the subsidy only covers certain positions within a given school, and not the entire complement of school personnel. If an estimate of private school fees is added, the total per student cost of public schools is higher at all levels than that of public schools. This higher cost is associated with higher quality of services, both in  - 23 - ANNEX 2 terms of the range of educational programs available, lower class size and better school management (lower use of substitute teachers, less days lost due to strikes, etc.). Cordoba - Estimate of Differential Subsidy Levels in Public and Private Sectors (in 1994 current pesos) Estimated Amount of School Fees Total Cost Per Total Cost Per Public Subsidy to Charged in Student in Student in Public Level . Private Schools Private Schools* Private Schools Schools Primary 463 230 693 618 Secondary 1,088 - 1,292 360 1,448 - 1,652 928 - 1,367 Nti Transferred 1,026 1,314 Higher 1,332 487 1,819 1,504 Special. - 2,507 329 2,836 2,708 * These figures are rough estimates based on the distribution of schools within ranges of school fees. In addition, they tend to underestimate actual private contributions through various charges not labeled as general school fees. For instance, in one school visited, additional charges amounted to 65 percent over the general fee level. B. Policy Options for Savings in the Education Sector in the Context of the Fiscal Crisis 22. There is a general consensus that the overall level of spending on education is sufficient to deliver educational services of a reasonable quality to the over 600,000 students enrolled in the system. The 1995 fiscal crisis has exacerbated the already apparent trend towards sacrifices in quality. Over time, the wage bill has crowded out almost all other expenditures on education. In the face of fiscal contraction, the Government will be hard-pressed to find ways of preserving its overall policy goals of guaranteeing access and a minimum of quality for at least nine years of basic schooling. This section explores the following options available to the Government to increase efficiency of expenditures while protecting overall sector goals: (i) improved personnel management, (ii) consolidation of Ministry structures and functions, (iii) revisions of subsidies to private schools, (iv) identification of other sources of financing, and (v) revisions to wage policies. Improved Personnel Management 23. A significant portion of the wage bill is lost to 'unproductive' uses, in part due to the lack of effective control systems. Examples include: (a) teachers that have accumulated excess teaching slots (either 'cargos' or 'horas caledras'); (b) outdated personnel information, leading to either 'ghost teachers' or unclaimed payments returned to the Ministry of Economy but recorded as educational disbursements; (c) heavy use of substitute teachers, generated by, in certain cases, excessive granting of various types of special arrangements (licencias, comisiones, lareas pasivas, jubilacion provisional, etc.); and (d) recourse to non-permanent 'contracted' staff. At the time of this review, information on (a) and (b) was not sufficient to permit an estimate of the financial savings available from the cleaning of these phenomena from the payroll.  - 24 - ANNEX 2 24. Regarding substitutes ('suplentes'), data indicate significant potential for efficiency gains. In most instances (although not always), hiring substitutes results in two salaries being paid for delivering the same service. The table below compares the rate of substitute positions per permanent positions for public versus private schools in 1994 and 1995. In all cases, private schools used substitute positions less than public schools. This difference is largely attributed to better management practices in the private sector, since other factors like sickness and pregnancy rates should be similar. With the onset of the fiscal crisis in 1995, the use of substitute positions decreased notably in both the public and private sectors as the Ministry of Education sought greater control over the granting of these positions. Substitutes as a Share of Permanent Positions Teachers 1994 1995 Pre-Primary Private 10.0% 2.7% Public 37.6% 23.4% Primary Private 5.7% 0.4% Public 28.6% 17.7% Secondary Private 3.6% 3.3% Public 15.7% 7.5% Directors, primary Private 5.5% 3.6% Public 20.3% 15.9% Special Ed. Teachers Private 8.2% 1.7% Public 15.4% 6.6% Total positions Private 6.8% 2.7% (cargos) Public 21.4% 13.9% Total teaching hours Private 3.8% 3.5% (horas catedras) Public 15.6% 7.5% 25. To estimate the potential savings from better management of substitute positions, 1994 is taken as a 'normal' year and the substitute rate within the private sector as a reasonable efficiency target. Using these parameters, if the public sector had performed at private sector levels in 1994, a maximum potential savings would have been about 44 million pesos, based on the calculations presented in Annex 4. Since not all substitute positions represent payment of double salaries for the same task, the actual savings would have been less. For instance, in reviewing a sample of almost three thousand positions which would have generated substitutes, including 'inactive duty', sick leave, maternity leave, promotions, relocations, provisional retirement, temporary reassignments, and the like, 45 percent would have generated two salaries, 10 percent would not have, and the remaining 45 percent is unclear. Therefore, a more accurate estimate of potential savings is in the 25 - 40 million pesos range, or about 4 - 7 percent of the total MEC 1994 budget. 26. Regarding contracted staff, as of the end of 1994, there were approximately 120 such positions within the MEC, which represents a relatively small share of total contracted staff as well as total personnel of the MEC. The overall government policy of not renewing contracts would result in about 850,000 pesos in savings for the MEC. However, it is worth noting that many of the contracted staff fill important functions, such as computer operators, the elimination of which would have an adverse effect on MEC functions. Consolidation of Ministry Structures and Functions  - 25 - ANNEX 2 27. At the beginning of 1995, the Ministry of Education was comprised of six secretariats and subsecretariats and about fifteen separate departments. The new administration has moved to consolidate the MEC structure, reducing the number of subsecretaries and departments. While this move has provided greater focus and control over MEC activities, without a corresponding elimination of redundant positions arising out of consolidation, the financial savings appear to have been minimal. 28. Is the Ministry of Education overstaffed? This question can be broken down into two: Are there too many people for the functions proscribed? And, are there too many low-priority activities that have generated additional personnel? Regarding the first question, there are clear indications of overstaffing at a number of levels. In terms of teaching staff, the low student/teacher ratios point to a more structural issue of average class size, staffing patterns and practices within schools, and the physical configuration of schools, particularly in low-density rural areas. To give a comparison of the overall size of the teaching force, about 4.4 percent of the labor force in Cordoba are teachers, excluding the university level, according to the 1991 National Census of Teachers. The OECD average for teaching staff as a percentage of the labor force is 3.1 percent including higher education and 2.6 percent counting only primary and secondary levels. 29. Because personnel structures differ significantly between countries, it is not possible to characterize the relative distribution of teaching to non-teaching staff. In Cordoba, 82 percent of the MEC staff are in 'teaching' positions (including school directors and other technical school level staff), 11 percent are in general services, and the remaining 7 percent fill the technical, administrative and managerial functions of the MEC. While general services appear high, in fact looking at the primary level alone, the ratio of general service staff (janitors, guardians, etc.) per public school is about 1.4 and at the secondary level 1.7 per school. There may be some overstaffing in the various school-level non-classroom "teaching" personnel (directors, vice- directors, section chiefs, etc.) which net of general services averages about 2.2 staff per primary school and 5 staff per secondary school. At the central MEC level, there has been an increase in technical level staff, typical of such ministries where the career path is out of the classroom into school management positions and then into MEC technical functions. It is not possible to say With any accuracy how many MEC staff are either redundant, underemployed or engaged in low priority tasks. However, for the sake of argument, assuming just one in ten of the non-school level positions could be eliminated without significantly jeopardizing the M1EC's ability to function (a fairly safe assumption), this would be the equivalent of 230 positions, or an annual budget savings to the MEC of 4.4 million pesos. 30. As of 1995, there were over fifty program areas being carried out under the Ministry of Education. These programs range from the pre-primary, primary, secondary, and higher educational services, to myriad smaller-scale endeavors such as support to scientific-technological development, special education services, activities around the arts.and culture, library development, pilot computer education programs, in-service teacher training, in-school health promotion, non-formal education, labor market training and so forth. Of the total MEC budget of 715 million pesos for 1995, over ninety percent, or 649 million pesos, was allocated to direct educational services and supporting programs at the pre-school, primary, secondary and higher levels including transfers to private schools, adult and physical education. Of the remaining 66 million pesos, 31 million pesos was earmarked for activities directly under the Minister's Office  - 26 - ANNEX 2 including the financial and administrative operations of the Ministry (of which 10 million pesos is for capital investments and 12 million pesos for non-personnel services), about 17 million pesos was allocated to cultural and artistic activities, 11 million pesos to activities related to promoting science and technology, and about 7 million pesos to planning, curricular and educational administration functions. 31. There has been little evaluation of the impact and effectiveness of many of the ancillary programs carried out by the MEC. Annual budgets and work programs are developed primarily based on historical activities. In addition, since 1993, the MEC has inherited a number of on- going programs from the Federal level, including a doubling of the expenditures on adult education programs. While not possible within the time constraints of this review, an assessment of these programs will be necessary within the fiscal constraints facing the Government in the upcoming year. Some of these programs are not the highest priority and/or impact and could be scaled-back or eliminated (permanently or temporarily) in the face of budget constraints. 32. Possible candidates include parts of the 11 million pesos for scientific-technological programs, some of which finances subsidies to promote innovative projects carried out by the private sector, the operation of an Advanced Technology Center and several non-formal training courses, seminars, scholarships and various events. The 16.3 million pesos in the 1995 budget for adult education also warrants review. About 31,000 adults participate in formal and non-formal courses in primary and secondary education as well as labor market training. Unit costs per student rose by over 30 percent between 1994 and 1995. Student teacher ratios are generally low, for instance on average about 14 students per teacher for adult primary education, particularly in the programs transferred from the Federal level. And the impact on labor market outcomes of the recipients of the non-formal labor training programs is not clear. In addition, while there is no clear methodology for assessing the merits of public investments in the cultural area (maintenance of historical patrimony, support to symphonies, ballet, drama and other events, promotion of Cordoba's culture), the amounts are not insignificant at 17 million pesos for 1995. Diversification of funding sources for cultural activities, including greater municipal support, is worth exploring. Overall, the potential savings depends on the will of the MEC to cut programs. If programs were reduced by 20 percent (not including capital investments), the overall savings would be about 10 million pesos or about two percent of the MEC budget. Subsidies to Private Schools 33. In the 1995 budget, subsidies to private schools totaled about 150 million pesos. Exploring potential budget savings from a reduction in subsidies to the private sector is tricky. At first glance, it would appear attractive to scale back these subsidies since there are immediate savings which could be realized, to be compensated by a greater share of contribution from the private sector. For instance, if the province selected positions which were no longer eligible for financing or reworked the sliding scale to elicit greater private contributions, it is assumed that school fees would have to be raised or certain types of services now financed by fees reduced (i.e., language, religious or computer courses). The danger lies in the potential response to increased school fees, particularly given the current economic crisis. Since the elasticity of demand for private schooling is not known, a rise in school fees may cause parents to switch their children to public schools, where the unit costs to the public sector are even higher, thereby exacerbating the budget situation (or reducing public school quality through overcrowding).  - 27 - ANNEX 2 34. Is there a way to rework the subsidy to make it more efficient and better targeted? The present sliding scale in which almost all schools receive 100 percent subsidy serves no targeting purpose other than to suppress the incentive for raising school fees in the private sector (at least the official fees). In addition, the subsidy level is tied to the number of personnel of the school and not directly to the number of students. For instance, a private school with average class size of 15 gets the same absolute level of subsidy as one with class size of 25. There is no incentive to raise student/teacher ratio, and private schools can exclude students to maintain smaller class size. 35. One option to adjust the incentives is for the subsidy to be conceptually attached to the student and not the teacher. A minimum class size would be established where the per student subsidy would be sufficient to cover the teachers' salary (plus a coefficient for other personnel). If the subsidy were attached to the student, a smaller class would require a greater level of private subsidy (i.e., parents that wanted smaller class size would have to pay for it). To trace the possible effects of such a reworking of the subsidy, the MEC would have to determine a target student/teacher ratio. The current weighted average of pre-school and primary level student teacher ratios in the private sector is 19 (see Annex 1). The public sector is slightly less at 18, while the municipal and Federal student/teacher ratios are significantly higher. To illustrate potential savings, we will assume a target student/teacher ratio of 23 for the pre-school and primary levels together, a reasonable level given its almost exclusively urban presence. 36. Annex 5 includes the calculations of impacts depending on three different scenarios, under each of which the primary level public subsidy per student would fall from 463 pesos to 382 pesos. Under the first scenario, the private sector would consolidate the same number of total students into fewer classes, resulting in a net of 7.6 million pesos in MEC transfers. Under the second scenario, the private schools choose to retain the previous student teacher ratios, and raise fees to cover the 7.6 million pesos shortfall, equivalent to an increase of about 20 percent over current fee levels. Under the third scenario, the private schools keep the same number of classes and teachers but attract more students to come up with the target student/teacher ratio. Enrollment in private primary schools would increase by 21 percent given the same budget transfer. Since there is full enrollment in the primary sector at present, students would shift from public to private. The theoretical savings to the public sector is equivalent to 12.3 million pesos, assuming that there would be an equivalent reduction in the public sector teaching force.' 37. What could be expected in practice is a differentiation in the private sector market, with some schools opting to raise fees and others increasing class size. Moreover, there are many private schools that would feel no effects, particularly those private schools which tend to be quite crowded already. Anecdotal evidence would suggest these tend to serve more populous low- income neighborhoods. Therefore the effects of such a policy would likely be progressive, with a greater contribution extracted from the private sector willing to pay for the premium of lower class size. In addition, since the subsidy would be made in a lump-sum per enrollment levels and not via teachers' salaries, private schools would be free to use the resources as they prefer, while the MEC would cease to incur the administrative costs of running the personnel system of the private schools. Such a system would require three regulatory aspects: (i) accurate monitoring of student enrollments to avoid exaggeration, (ii) deregulation of the amounts private schools are It should be ntoed that these calculations apply only to the pre-school and primary levels. Potential savings arc much greater at the secondary level due the higher unit costs and absolute levels of support to the private sector. However, the calculation of per student subsidies would require a cleaning of the data resulting from the Federal transfers and a more detailed treatment of horas catedras which is beyond the scope of this report.  - 28 - ANNEX 2 allowed to charge (delinking or rescaling of subsidy with fee levels), (iii) definition of the 'free- standing' private sector to preclude its support; and (iv) regulations to avoid poorer students being pushed out the private system due to escalating school fees; for instance a minimum percentage of 'scholarship' students in each school, to be financed with internal cross-subsidies. Diversify Sources of Sector Financing 38. A further option would be to seek out and optimize other sources of sector financing. Three are worth exploring: (i) raise school fees, (ii) decentralize schools to the municipalities, and (iii) optimize resources coming from the Federal level. Regarding school fees, at present education is free at all levels in Cordoba. This is not to say that there is not significant private spending even within the public system. As the state has retreated from financing anything but the salaries of personnel within schools, purchases of educational materials, school supplies and the like must be generated by the families of students, as well as the individual expenses related to purchase of books, transport to school, uniforms etc. Nonetheless, it is difficult to justify no school fees, particularly at the upper levels of education, in a province with the per capita income levels of Cordoba. Any potential revenue generated from school fees should remain as part of the school budget to address critical needs defined by the schools themselves. Since the MEC currently spends virtually nothing on school-level materials, there would be no direct budget savings from these measures. However, applied as such, minimum school fees may be the only way in the short run to attempt to reverse the decline in educational quality. How much could be raised? If school fees equivalent to 15 percent of those paid within the private sector were applied to secondary and higher education (48 and 73 per student), the amount collected would be approximately 4.5 million pesos and 0.5 million pesos respectively, or roughly 16,000 pesos per secondary school and 10,000 pesos per higher education institute, which could be applied to supplementing school libraries, computer programs, lab equipment, etc. 39. Decentralization of education services offers the immediate attraction of passing expenditures along to municipal governments. Much in the way the Federal government transferred secondary education to the provinces, the current discussion centers on the feasibility of transferring pre-school and, eventually, primary education to the municipal level. The provincial government is exploring a pilot for 1996 of transferring 19 stand-alone pre-schools. Without the actual enrollment and staffing figures, an accurate estimate of budget savings is not possible. Using global figures for cost per student and average school size (645 x 70 x 19), savings would be under I million pesos'. An eventual transfer of all pre-schools would save, using the same methodology, about 34 million pesos. However, there are a number of issues involved in such a transfer, the two most important of which are the equity implications of variations in absorptive capacity of municipalities and the administrative difficulties inherent in pre-schools that are physically integrated with primary schools (the majority). 40. A brief comment on optimization of Federal financing available to the province. The Plan Social Educativo, a national compensatory program to provide books, educational materials and teacher training to primary schools in poor regions, allocated 126 million pesos in 1995 and 97 million pesos in 1995. The Plan's budget for 1996 is anticipated at around 122.7 million pesos. If the primary motivation for this transfer is budget savings in 1996, it is worth comapring the effects of transfer of some of the cultural programs to the municipal level: program 350 - support to regional orchestras, choirs, thcater groups and museums (3 million pesos); program 354 - cultural and artistic promotion ( 11.9 million pesos), prooram 355 - preservation of cultural patrimony (0.3 million pesos).  - 29 - ANNEX 2 This review was not able to locate data on how much of the overall program was being spent in Cordoba. Using just a per capita measure, the Plan would allocate 8.7 percent of its resources to Cordoba, or 10.6 million pesos in 1996, equivalent to about S5 percent of the investment budget of the MiEC in 1995. Given the significance of the amounts, Cordoba should optimize (as well as maximize) to the extent possible the resources from this program. About 167 primary schools have received Plan Social support in Cordoba, representing approximately 12 percent of all public primary schools. Forty percent of these have been in the capital area, where 10 percent of the public primary schools and 34 percent of the public primary students reside. Targeting could be improved to ensure that the poorest students and regions benefit from this support Summary of Short-Term Options 41. Barring the discovery a significant number of double salaries, ghost teachers and other anomalies, the range of savings available to the MEC in the short-term are relatively limited. In terms of "quick fixes" which would have no significant impact on sector objectives and would not require personnel retrenchments of permanent positions (A, B, C and G in the table below), potential savings represent between 26.7 and 41.7 million pesos, or from 4 - 7 percent of the MEC's 1995 budget. Adding measures which imply a reduction in permanent positions and the scaling back of non-priority programs would probably result in about 14 million pesos in additional savings (D and E below) or about 2 percent of the 1995 MEC budget ' To summarize, applying all measures including quick fixes, better personnel management, retrenchment of redundant staff and non-priority programs, and pilot decentralization, as difficult as these measures would be, would only reduce the MEC's budget by 10 percent at most, and not all of this would represent net savings to the provincial government due to severance and retirement payments. If further cuts are necessary, reductions in wages, decreasing the number of school- level personnel, or deeper cuts in lower-priority programs, and eventually investments, would be the only remaining options. 7 If all infrastructure investments were curtailed for 1996, an additional 10 mullion pesos or 1.4 perccnt of thc 1995 budget could be saved, although such a move risks quality and, to a lesser degree, access to educational services  -30- ANNEX 2 Summary of Strategies in the Short-Term Range of 1996 Savings Savings? Observations A. Clean-up payroll - eliminate double salaries, ????? Yes ghosts B. Better control over Yes use of substitute teachers 25 - 40 million C. Eliminate contracted Yes Need to preserve certain key personnel 0.8 million functions D. Eliminate low priority Possible Little net savings to government, need programs (cut by 20%) 10 million change in employment policy E. Reduce overhead Possible No net savings to government, need personnel by 10% 4.4 million change in employment policy F. Revise basis for ???? Possible Need change in legal framework private school subsidies G. Municipalize pre- Yes Equity/ absorptive capacity schools (19) 0.9 million considerations D. Medium-Term Strategies for the Education Sector 42. The short-term measures will help to increase the efficiency of the education system. However, certain trade-offs in terms of a medium-term strategy to increase quality are inherent in some of the options. For instance, an eventual reduction in teachers wages may, over the medium-term, create a disincentive to attracting qualified candidates and exacerbate the tendency of teachers to maximize teaching hours over schools (the so-called 'taxi teachers'). Moreover, in atight budget situation, it will be extremely difficult to undertake expenditure switching towards non-personnel operating expenditures or the financing of key investments which will improve quality over the longer-term. Therefore, the short-term options available to the government of Cordoba should be compatible with the medium-term strategies necessary to meet sector objectives. Over the medium-term, strategies to increase the efficiency and quality of educational services include: 43. Increase student/teacher ratios to a minimum wvhich still guarantees relative quality. The structure of the educational system in Cordoba, and indeed in Argentina as a whole, is very high cost in terms of the relation between service providers (i.e., teachers and other sector personnel) and the number of people they serve. The most important element of a strategy to make Cordoba's education system more efficient over the medium-term will be to increase the overall student-teacher ratios. This will require school consolidation and provision for student transportation in certain areas, as well as rethinking the personnel structure at the school level to favor teaching positions in the classroom over non-classroom positions. A necessary first step will be to analyze on a school-by-school basis the unit costs associated with the provision of services. A simple cross-reference of enrollment per school with the monthly wage bill associated with that school would permit the MEC to detect which schools are the most expensive on a per student basis. Local solutions can be taken to bring these schools in line with sector targets. In some cases, the cause may be too few students, such as in isolated rural schools, in others it may be too many teachers due to a lack of effective personnel control. Potential solutions will differ  - @'1 - ANNEX 2 from each school. Annex 6 presents some initial estimates of the economic impacts of raising student teacher/ratios looking only at the primary education system. In the base case scenario, if Cordoba achieved student/teacher ratios equivalent to those of Chile (25), slightly below the LAC average of 26, the net savings to the system would be about 70 million pesos in the public sector alone, with an additional 10.5 million pesos in savings if this rate were achieved in the private sector as well. The strategies to achieve these savings will require micro-level planning and a reconsideration of the current model of school staffing.' 44. Effective control over wvage bill and personnel. To achieve this end greater transparency in the personnel structure and wage calculations will be required along with the modernization of personnel administration. Immediate investments in computers and technical assistance to modernize personnel administration should be a high priority for the Ministry in 1996. 45. Reorient sector expenditures tojfavor educational services at the school-level and less to protect the interests of the 'educational establishment'. MEC overhead should be made as efficient as possible, including an orientation to programs which support improved quality of the learning environment. This should translate into greater school-level control over resources and increased attention to the non-personnel inputs to educational quality. This also implies protecting essential investments in physical rehabilitation and upgrading of schools. The MEC should seek a medium-term target of 10 percent for non-personnel operational expenditures. This would allow for the purchase of essential inputs and the equipping of classrooms, especially at the secondary level, with modern technologies and teaching materials. 46. Achievement of greater equity in the system, whereby richer students pay more for educational services than do the poor and greater levels of public subsidies are applied to basic education over the higher levels (including adult labor market training, fine arts schools, etc.). In a context of budget constraints unlike any Cordoba has faced in the past, these cross-subsidies ivill be necessary to ensure that access and quality do not erode. 47. Increasing the quality and relevance of education, especially at the secondary level, including modernizing curriculum content and educational materials in line with a rapidly changing global economy. Implementation of the new structure and curriculum of secondary education will be an important element in increasing the internal efficiency of the system while boosting the quality of Cordoba's labor market in the coming years. To illustrate, if the ratio of teaching to non-teaching staff in schools were raised from the current 4:1 to 6:1, the savings would amount to about 15.6 million pesos, eliminating the position of vice-director alone would save about 10 million pesos. Other options include increasing classroom size. If classrooms were consolidated to achieve a 10 percent increase in average class size, 13 million pesos would be saved in reducing the numbcr of macstros dc grado (not ilcludinig similar reductions in other associated teachers).  -32 - ANNEX 2 Pre-school 191 I I - Urban 20A , . i -Rural 18 . Primary 171 181 26! 251 161 221 27 - Urban - Public 18 I - Urban Private 19 Ii - Rural Public 1 3 I I i - Rural Private 14 I i Secondary 12 71 191 18 16 15j 19 -Fed. Trans '94 8 i | I i Higher 91 7 1 I Special Ed (Primary) 3j i _ _ - Public 31 I i_i - Private 41; I II _ _ _ _ _I _ _ _ i I i _ _ _ _ Disaggregated Pre-School and Primary Levels since 1989 1 1 Pre-School 1989: 19901 1991 1992 19931 National 231 211 19 191 231 Municipal _ 291 271 3 331 291 Provincial Public 201 201 21 201 201 Private 21 221 211 201 171 Total | 211 211 211j 201 191 Primary i i i I i National _ 21! 201 171 201 201 Municipal i 27 251 271 27 271 Provincial Public 1s, 181 181 17 i 17 Private 23: 201 201 201 19 Total i 201 191 181 181 171  tMEbt1otitk i>{t.;Z4&Z1989 1990 1991 1992 1993 1994 1995 3,453,616 4,931,223 6,164,190 8,075,774 10,490 996 11,219,921 14,300,435 Priwy:Et$4m. 128,286,934 1 77,286,546 190,785,333 218,961,775 235,352,669 245,332,943 242,586,285 47,398,927 74,378,891 79,274,081 92,219,501 101,585,470 178,362,259 181,619,871 3,887,401 5,429,725 6,264,273 7,689,656 8,446,012 8,637,219 8,805,791 43,253,756 __60,196,096 72,067,586 146,321,054 203,173,981 121,577,623 163,254,427 'g'~"' 2 " ||| 226,280,635 322,222,481 354,555,463 473,267,761 559,049,128 565,129,966 61015665808 Annual 1990 1991 1992 1993 1994 1995 ,growth 42.8% 25.0% 31.0% 29.9% 6.9% 27.5% 38.2% 7.6% 14.8% 7.5% 4.2% -1.1% ____________ 56.9% 6.6% 16.3% 10.2% 75.6% 1.8% ____________ 39.7% 15.4% 22.8% 9.8% 2.3% 2.0% 39.2% 19.7% 103.0% 38.9% .40.2% 34.3%- 42.4% 10.0% 33.5% 18.1% 1.1% 8.0%- Average Average ______ Annual 89-95 Annual 90-95_ 27.2% 24.1% _ ____________ 1. % 6.6% _ _ _ _ _ _ _ _ _ 27.9% - 22.1% 15.3% 10.4% _ _ _ _ _ _ _ _ _ _ 32.5% 31.1 %__ _ _ _ __ _ _ _ _ _ _ _ _ 18.9% 14.2% ,_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ x  C~~u$rdJME~~~~~ ~~ 340 535 5,778,431 10,801 533 7,865,474 177 14,112307 383 22633 211,057,673 9,325 22677 258855,777 14,15 10,916 17.1% iesrg''.th1 __ 387 7510 _ 85,286,163 1 137 _1_35 _ _ 79 _5 0 0 14,049 1 3 3 j5 18.3% 389 7-60 9, _6 8 1,7 86 1i27'~3 9 760~ 1 1 398, 91 1-4999 14,343 _126% 390 282 2,387,012 8,465 282 3;057,948 10,844 10,370 22.5% 359 5 29,922 5,984 5 53,690 10,738 10,269 71.6% . .dii?IhhoiL ? = 376 76 1,125,160 14,805 76 1,161,524 15,283 14,615 -1.31% fr?erdj?~ . -. ..> .......378 199 2,159,293 10,851 199 2,196,552 11,038 10,556 2.7% .:re~MEC =nft~ '~ ____ 1 5_ 41 7,405,481 13,689 _ 539 10,407,324 19,309 18,465 34.9% Su>Cu>'.tzr>t.