Public Works Employment

PUBLIC WORKS EMPLOYMENT AND RURAL POVERTY ALLEVIATION IN THE PHILIPPINES Country Report Prepared for the International Food Policy Research Institute May 10, 2001

Roehlano M. Briones Assistant Professor Economics Department School of Social Sciences Ateneo de Manila University Katipunan Avenue, 1108 Loyola Heights, Q.C. Tel/fax: 63-2-426-56-61 Trunk: 63-2-426-6001 L5220,5221 e-mail: [email protected]

1 Chapter One. INTRODUCTION 1.1. Overview of the study Alleviation of rural poverty is accorded the highest priority in the country's development agenda. As the capacity to work is the primary asset of the poor, one common strategy to alleviate poverty is to generate employment. A direct way to do this is through public works employment (PWE) schemes. The degree to which poverty can be alleviated through such schemes depends on the accuracy in which benefits are targeted to the poor. In many developing countries, PWE in its various forms is a longstanding and well-recognized strategy for rural poverty alleviation. However, Philippine policy does not accord such a status to PWE schemes. Information on wage employment creation through rural works is also thin: Subbarao, Ahmed, and Teklu (1996), in their study on safety nets, observe that available evidence is “fragmentary”, and that “the experiences of various public works projects are poorly documented. Despite the country’s long experience in labor-based projects, no synthesis yet exists to distill its past experiences (p. 34).” The promotion of PWE is therefore hobbled by inadequate knowledge as to what works and what doesn’t. In view of the IFAD’s desire to promote rural nonfarm employment in the country, this study has been commissioned to partially remedy this inadequacy. Our study undertakes two tasks. First, we review the country’s experience to consolidate existing knowledge about PWE in the Philippines. Second, we shall conduct case studies intended to illustrate “best practice” for PWE. Guidelines for PWE policy and implementation shall then be drawn from the review and casework. 1.2. Poverty and employment trends in the Philippines Table 1.1 provides official figures for urban and rural poverty (see however Box 1). Nearly 40 percent of the country’s population in 1997 was deemed poor. Poverty is unevenly spread throughout the country, with rural poverty consistently higher than urban poverty. Poverty incidence hardly budged over the period 1985-1997, whereas poverty in urban areas has fallen substantially over the same period.1 Rural poverty remains a massive and recalcitrant problem for the country. Poverty trends move closely with trends in aggregate income. Poverty fell perceptibly during the growth episodes of ’85-’88 (average of 4.8% growth) and 1

These figure in Table 1.1 for a developing country with levels of per capita income and other welfare indicators similar to that of the Philippines. Box 1 offers

2 ’95-’96 (average of 5.3% growth), with a greater decline in urban than in rural areas. The period of least poverty reduction (’88-’91) was also a period of stagnation (below 3% annual growth). The employment picture in 1998 seemed bright: only 6.7 % of the labor force was jobless in 1998, with rural areas apparently being far better off than urban areas. However underemployment statistics restore the gloom: persons working less Table 1.1 Urban-rural comparison of poverty and employment, Philippines

Rural

Urban

National

1985

56.4

37.9

49.3

1988

52.3

34.3

45.5

1991

55.1

35.6

45.3

1994

53.1

28.0

40.6

1997

51.2

22.5

37.5

5.4

10.7

6.7

Worked below 40 hours/wk

34.8

22.4

45.3

Worked below 20 hours/wk

11.4

6.8

14.8

Poverty:

Employment (1998) Unemployment rate Underemployment (percentage of the employed)

SOURCE: National Statistical Coordination Board

3 Box 1. Calculation of poverty lines Poverty and household income and expenditure statistics for the Philippines is calculated on the basis of a Family Income and Expenditure Survey, conducted once every three years (results are available only up to 1997). Regional subsistence thresholds are calculated based on a food menu reflecting the average dietary composition of that region. The menu is then valued using prices differentiated by urban and rural areas; urban and rural food lines are therefore separately for each region. Poverty lines (urban and rural by region) are then calculated by adding a nonfood component, using the nonfood expenditure shares of households whose incomes fall within a predetermined band around the subsistence threshold. This method has generated poverty lines consistently and significantly above the international standards, if we are to compare the poverty figures with other indicators of development of the country. Consider the following: Poverty incidence Year Official line

International line

GDP per capita, $US (1998)

HDI rank (1999)

Philippines

1997

37.1

26.9

1050

98

Bangladesh

1996

35.6

-

350

144

India

1994

35.0

47.0

440

138

Cambodia

1997

36.1

-

260

153

SOURCE: World Bank, UNDP. The International Line column reports the poverty incidence using a PPP-adjusted $US 1.00 a day as the poverty threshold. The official line is too high; it is incredible that the Philippines, with its per capita income and HDI rank, would have higher poverty incidence than Bangladesh, India, and Cambodia. In 1997, the national per capita poverty line was 31.63 Php per day, which is $US 2.44 using the World Bank PPP-adjusted exchange rate. The reason for the overestimation is that the food menu is not sufficiently representative of actual preferences from the lower-income groups. Philippine poverty lines also tend to co-move with living standards, biasing interregional and intertemporal comparisons to the conservative side. The reason is that the food composition for calculating the subsistence line itself depends on living standards. Balisacan (1997) corrects for these problem. (Incidentally, he calculates regional poverty lines that are 48% lower on the average than the official lines.)

4

than full-time accounted for 45.3% of the rural employed, more than double the proportion in urban areas. As many as 14% of the rural employed are in fact working on a half-time basis or less. This is consistent with established trends pointing to the intermittent nature of rural occupations, and may well be linked to the low and stagnant productivity in agriculture (Balisacan and Babila, 1994). The geographic distribution of poverty can be described in terms of the country’s 16 administrative regions (Table 1.2). Outside the capital, poverty incidence ranged from a low of 20% to a high of 64% in 1997. The least poor regions are in the northern island group of Luzon, near the capital; the poorest are in the southern island group of Mindanao. The middle island group in the Visayas however has experienced some worsening of poverty from 1994 to 1997.

Table 1.2. Recent estimates of poverty by administrative region, Philippines Region*

Poverty incidence

Population share

1994

1997

1997

National Capital

10.5

9.6

14.1

Cordillera Autonomous

56.4

49.6

1.7

Ilocos

53.6

44.7

5.5

Cagayan Valley

42.1

38.0

3.9

Central Luzon

29.2

20.1

10.3

Southern Luzon

34.9

30.1

13.4

Bicol

60.8

57.5

7.1

Western Visayas

49.9

48.2

8.6

Central Visayas

37.5

39.5

7.2

Eastern Visayas

44.8

49.0

5.1

Western Mindanao

50.6

45.5

4.0

Central Mindanao

54.1

53.2

5.9

Southern Mindanao

45.6

44.6

7.1

Eastern Mindanao

58.7

55.4

3.3

Autonomous Muslim Mindanao

65.3

63.8

2.9

NORTH (LUZON)

CENTRAL (VISAYAS)

SOUTH (MINDANAO)

SOURCE: National Statistics Office.

5 Rural households are usually associated with agricultural employment. Estimates by Balisacan (1997, 2000) identify agriculture to be the greatest contibutor to poverty among the employment sectors (Table 1.3). More importantly, its contribution is much larger than its population share. All the other sectors of employment contribute less to poverty than its respective population share. Table 1.3. Contribution to total poverty by economic sector over time

1985

1994

1997

Population share

Contribution to poverty

Population share

Contribution to poverty

Population share

Contribution to poverty

Agriculture

52.9

75.8

49.8

77.7

40.1

67.8

Manufacturing

9.6

6.2

9.4

3.9

8.1

4.8

Construction

5.9

4.7

7.6

6.5

7.7

7.1

Trade

9.0

4.6

9.3

3.9

8.8

4.7

Transport

6.9

3.3

8.1

4.0

8.0

4.4

Finance

2.2

0.3

1.9

0.2

1.9

0.2

Services

13.6

5.1

13.9

3.8

12.5

4.9

Sector

NOTE: Agriculture – agriculture, fishery, and forestry; Manufacturing – mining and quarrying, electricity, gas, and water, manufacturing; trade – wholesale and retail trade; transport – transportation, storage, and communication; finance – finance, insurance, real estate, and business; services – community, personal, and social services. SOURCE: Balisacan, 1997, 2000

Agriculture is of course not the only source of income and employment of rural households. Data throughout the developing world have revealed the substantial role played by nonagricultural activities in the rural economy. For the Philippines however, little work has been done to characterize the rural nonfarm sector2 on a nationwide basis. The few efforts at obtaining data are limited to scattered household-village surveys. Leones and Feldman (1999) find that nonfarm income accounts for 24% of household income, based on microdata from a Philippine village. Panel data survey for villages in the Philippines' "rice bowl" (Region 3) finds that nonfarm 2

Here "rural nonfarm" is identical to "rural nonagricultural".

6 income as a proportion of total income has climbed from 27% in the 1960s to 51% in the mid-1990s (Estudillo and Otsuka, 1999). Studies of land reform beneficiaries nationwide in 1997 indicate a similar share (Institute of Agrarian and Rurban Development Studies, 1999).These figures suggest that nonfarm incomes are a substantial and growing component of economic activity in the rural areas of the Philippines. These trends suggest a strong potential for boosting nonfarm wage employment as an antipoverty strategy. 1.3. Direct poverty alleviation and wage employment creation It has been rightly pointed out that the eradication of poverty in the Philippines ultimately rests on rapid and sustained overall growth. Nevertheless, direct poverty alleviation remains at the core of development strategy. First, precisely when indirect poverty reduction is sluggish, direct means (that do not weaken nascent growth) are crucial to making any headway at all against poverty. Second, when some of the poor may be left out or even harmed by the growth process, direct programs can be rationalized as "safety nets”. Thus many developing countries have adopted PWE creation as a direct strategy to alleviate poverty. South Asian countries such as India and Bangladesh are well known for a poverty alleviation strategy involving PWE. Other countries that have undertaken PWE programs on a massive scale are China and Indonesia (Esmara, 1987). In Africa meanwhile, Botswana and Zimbabwe have been relatively successful in generating employment through public works (von Braun, Teklu, and Webb, 1992). Public works schemes in Latin America, such as in Bolivia, Chile, and Peru, were used to counter temporary drops in employment due to structural adjustment or macro shortages (Deolalikar, 1995). An important merit of the PWE approach is a built-in targeting feature: if wages are sufficiently low and work is arduous (but not excessively so) then workers who opt to join the program would tend to be poor. This is referred to as selftargeting. Other forms of assistance may offer incentives to both poor and nonpoor (e.g. food price subsidies), making exclusion of the nonpoor difficult. Moreover, labor markets, particularly in developing countries, are subject to numerous market failures. Wages do not adjust to eliminate unemployment and underemployment, for reasons such as wage legislation, costly search, incentive problems, and so forth. Unemployment can also be a short-term phenomenon, arising from environmental disturbance, macroeconomic downturns, and seasonal requirements of agriculture. Such shocks should not be dismissed outright as mere short-term phenomena, as these episodes can push affected households into a "poverty trap" and cause their long-term ruin. In response to these market failures, PWE schemes can stoke labor demand, particularly for the unskilled labor of the poor. Projections made by the government estimate that incorporation of labor-intensive methods in public

7 works could have generated up to 260,000 jobs for 1999; this is nearly 10% of the total number of unemployed (DAP, 1998). 1.4. Types of public works employment projects There are various ways to classify PWE schemes. One classification is based on the assets provided: productive or economic infrastructure directly improves productivity (e.g. irrigation, roads, bridges), while social infrastructure addresses basic needs of households (e.g. potable water systems, school buildings, health clinics). The third category is agroforestry projects (tree planting, upland flood and erosion control). Clay (1986) suggests the following classification based on the type of unemployment problem addressed. Relief works provide temporary employment to address emergencies arising from civil or natural calamity. Income augmenting programs aim to supplement below subsistence incomes of the very poor, and are often seasonal. Long term employment programs provide livelihood for the unemployed. Finally, low cost infrastructure programs incorporate a labor component although are aimed mainly at asset formation. A major type of low cost infrastructure is community works, which are undertaken "in response to demand by, and for the benefit of, specific local groups; and they usually include elements of self-help" (Gaude and Watzlawick, 1992). Public works are often narrowly defined as encompassing only the first three types; e.g. Basu (1996). If this restriction is accepted however, virtually no significant initiatives in recent years can observed for the Philippines. In fact the World Bank (1996) has remarked: “However, as now structured, Philippine programs do not serve as an employer of last resort, address the problem of seasonal unemployment, create assets that benefit the poor (such as schools), nor provide disaster relief." For our purposes then the infrastructure projects with an intentional employment-generation component (reflected in the choice of technology) is classified as a PWE project. The rest of the report is organized as follows: we review country’s policy framework for PWE (Chapter 2), and its experience in PWE schemes (Chapter 3). The case studies are then introduced with a discussion of the research method, a brief background of the cases, and the issues covered (Chapter 4). An in-depth analysis of the case studies follows (Chapter 5). Wage benefits and targeting performance are treated separately (Chapter 6). We conclude with a final assessment of lessons learned, and a set of recommendations based on “best practices” for the implementation of PWE schemes in the Philippines (Chapter 7).

8 Chapter Two INSTITUTIONAL AND POLICY FRAMEWORK FOR PUBLIC WORKS EMPLOYMENT SCHEMES

2.1. Institutional setting Traditional framework Traditionally the national government has been in charge of public works in the country. Executive line agencies (called Departments), work independently or in tandem for public works provision. The agency most directly involved is the Department of Public Works and Highways (DPWH). The Department of Agrarian Reform (DAR) also initiates rural works, though the actual implementation of agrarian reform infrastructure projects is handed over to the DPWH. The Department of Agriculture (DA) meanwhile finances the construction of farm-tomarket roads and postharvest facilities. An agency attached to the DA is the National Irrigation Administration (NIA), which implements irrigation projects, including the establishment of communal irrigation facilities. Meanwhile the Department of Interior and Local Government (DILG), as well as the Department of Labor and Employment (DOLE), collaborate with the local government for rural works provision. Most line agencies have regional offices, which in turn used to coordinate provincial and sub-provincial offices. With no direct involvement in project implementation, but still influential in terms of broad policy thrusts, is the economic planning agency called the National Economic Development Authority (NEDA).

Box 2. Calamity response Special mention must be made of the Department of Social Welfare and Development, which supports disaster relief operations through reconstruction of damaged facilities and houses by members of the affected community. Workers are "paid" with daily rations, including 2 kg. of rice, and canned food (currently the ration value is Php 90.) The assistance lasts from from 5-10 days. However working does not seem to be the condition for receiving rations. That is, this type of relief program does not seem to aim at generating employment.

9 Local government units (LGUs) are stratified into the provincial, municipal, and village (barangay) levels. Barangays may be further divided into sitios or puroks. Leaders at each level (purok leader, barangay captain, municipal mayor, and provincial governor) are elected by popular vote. Local government elections are held every three years. The traditional set-up entailed a highly centralized concentration of authority, not only in public works, but in the entire gamut of social services. Running a huge bureaucracy however proved cumbersome. Major decisions had to await approval from some high authority in the distant (both geographically and politically) regional or central office. Regulations and policies were typically standardized, with no flexibility for local conditions. Finances were first pooled nationally, then disbursed according to a budget approved by Congress, finally reaching the field in trickles, delaying implementation.

Decentralization and rural works The most innovative development in the public institutional framework has been the shift to decentralization (Cabanilla, 1999). Initial moves would be the formation of Regional Development Councils (RDCs), composed of regional heads of line agencies, provincial governors, city mayors, legislative representatives, and, NGO representatives. The RDCs are involved in regional planning, inter-agency coordination, coordination of budgetary and project allocation from the central offices to regional offices and LGUs, and assist LGUs in formulating plans and securing funds. Lower level versions of the RDCs are the Provincial Development Councils (PDCs) and Municipal Development Councils (MDCs). The most radical break with tradition was however initiated by the Local Government Code of 1991, which bestowed autonomy on LGUs. The Code devolves responsibility for road and bridge maintenance, as well as water resources infrastructure. Also devolved are other basic government services, such health care, agricultural extension, and welfare services. The authority of LGUs to collect taxes and fees is widened, while their share in the national government’s internal revenues is increased. The principle behind decentralization is that decision-making and implementation will be improved because of, first, the familiarity of LGU officials with local conditions, and second, because of greater accountability to their constituents. Evidence suggests that the more economically developed countries do tend to have less centralized bureaucracies (Garnier and Majeres, 1992). Local governments are supposed to be better placed to provide local public goods. Alonzo (1999) explains that observing the principle of subsidiarity implies better choice of projects, and a more efficient choice of techniques. Unit costs for

10 construction and maintenance of public works may decline as LGUs favor the use of labor and indigenous materials in locations where these are cheaper. A national government agency, in contrast, may be less sensitive to relative price differences across locations, or may be compelled to adhere to a national design standard. LGUs are reportedly constructing barangay roads at one-third less the cost of the same roads provided by the DPWH. A similar cost savings is heard for classroom construction. Other improvements due to decentralization have been documented over the past decade (Abesamis, 1995). Decentralization though is plagued by its own share of problems. The distribution of devolved personnel across locations appears to have been flawed: for example, in the case of agricultural staff, Southern Tagalog and Southern Mindanao have approximately the same farm size, but the latter’s devolved personnel is only half that of the former. Conflict between devolved and LGU personnel and have also been noted, exacerbated by duplication of functions (Elazegui and de Guzman, 1999). Moreover, whereas successful decentralization requires devolution of control over resources, a mismatch between incremental revenue allotment and devolved functions has been observed (Capuno and Solon, 1996). Evidence suggests inequity in the mismatch: the poorer provinces have a larger fiscal gap compared to the wealthier provinces (Quitazol, 1996). LGUs could possibly raise their own funds, but so far enthusiasm in doing so is weak, owing to the political sensitivity of their fund-raising instruments (such as property taxes). Fiscal problems at the local level is an impediment to development project implementation. A recent evaluation conducted by the national government of ongoing ODA programs identifies insufficient provision of local counterpart funds as a major bottleneck in public works programs. Contributing to this may be the lackadaisical commitment to development projects on the part of some LGUs. The pool of capable individuals from which to choose being shallower; moreover, local elections may be prone to selecting officials not on the basis of merit but on economic influence and clan-name recall. Fears have been raised regarding the increased influence of the local elite and intensification of government corruption under a decentralized regime. Proponents of decentralization counter that the depredations of a corrupt central bureaucracy, in connivance with a national elite, are worse. Decentralization should be viewed as one means to curtail rent-seeking and official corruption, as local governments are more directly observable and accountable to their immediate constituencies. The a priori predictions regarding the relationship between corruption and decentralization are therefore ambiguous. The issue can only be settled empirically. Huther and Shah (1998) finds a negative association between corruption and poverty. Fisman and Gatti (2000) base their study on a pooled cross-country time series regression. Their model

11 controls for the endogeneity associated with the correlation between decentralization and development. It also checks for robustness by using alternative measures of corruption. They conclude persuasively that decentralization is a significant check on corruption. Both the national government and the local government have important roles to play in development of communities. Serafica (1999) notes that there is an asymmetry of information and capabilities between national and local governments; that is, while “central government does not know what to do… the local government does not know how to do it.” (Bird, 1994, p. 28). Local infrastructure should be provided in such a way as to utilize the comparative advantage of each level. Projects must be selected and implemented by local governments, with some supervision and technical assistance from the national government. Unfortunately, few if any success stories of national-local government collaboration have been documented. Perhaps over time LGUs will gradually acquire the capacity to undertake more complex projects (the incrementalist view); for now, the national government may have to be selective as to which provinces and municipalities to collaborate with (the selective approach), and what types of activities and projects to delegate to LGUs (stratification approach). To summarize: gains from decentralization can be further increased with stronger commitment from the national government in terms of financing and operational skills transfer. With this, and gradual maturation of grassroots democracy, the shift to local governance should lead to a more inclusive development process. 2.2. Rural development We now move to the policy framework for rural PWE schemes, which first discussion of rural development strategy. The framework for rural development is set by the Comprehensive Agrarian Reform Program (CARP), implemented since 1988.3 The Program mandates a comprehensive redistribution of land, as well as the provision of support services to agrarian reform beneficiaries. To transform Philippine agriculture, massive interventions in the provision of infrastructure, credit, and technical assistance are deemed essential. The DAR spearheads implementation of the Program, while the rural works component is undertaken by the DPWH. In terms of actual accomplishment though, data on funding CARP infrastructure (coursed through the DPWH) is dispiriting: Over a ten-year period from 1988 to 1998, of the Php 12.165 billion allocation, only Php 2.8 billion was actually disbursed. The bulk of this was done in the first five years, with an average 3

There have been numerous land reform programs in the country. The Marcos land reform immediately preceding CARP covered rice and corn and, in some regions at least (notably Central and Southern Luzon), succeeded in its redistribution objectives.

12 annual allocation of Php 352.6 million; in the last five years, the average annual allocation was only Php 208.4 million or 30% lower than in the previous period. This decreasing level of commitment is ominous. Financing is hobbled by weak political support from Congress, which includes a representative body elected by according to geographic district (hence the co-opting of numerous elective positions by landed interests). Doubts have been raised as to whether CARP has been properly targeted. Deininger, et.al. 1999) suggest that the DAR's drive to "finish the unfinished business of land reform" has reduced the antipoverty targeting of the Program in rice lands, compared to the previous land reform efforts. However, a nationwide survey of land awardees (IARDS, 2000) has estimated poverty incidence in this group (using official poverty lines) at 62.5%, which is higher than the rural poverty figure, but is consistent with overall poverty in agriculture. Without downplaying valid targeting concerns, agrarian reform nevertheless provides a workable framework for credit and infrastructure programs aimed at alleviating mass poverty in the countryside.

Box 2. CARP and land reform (Briones, 2000) Officially, the CARP covers 83% of the Philippines’ agricultural lands. It follows the traditional approach to land reform. Lands are redistributed with compensation, while restrictions are imposed on land rental arrangements. Persons prioritized to receive land awards are tenants, followed by regular farmworkers, seasonal farmworkers, other farmworkers, tillers of public land, and other cultivators. Land awardees are entitled to no more than 3 ha. of land, and must amortize their lands either to the landowner (under a negotiated settlement) or the national Land Bank. So far, land redistribution is most successful for government-owned and public lands, as well as lands under voluntary sale or transfer. Least successful are private lands under compulsory redistribution, - only 3% of such lands have been transferred to cultivators. This can be readily attributed to landowners’ resistance, particularly aimed against the below market values used by the 2.3. Infrastructure policy land compensation. government in determining

To focus the support services of CARP, the DAR has identified Agrarian Reform Communities (ARCs). An ARC is a barangay or cluster of barangays in which CARP accomplishment is significant and on which support service efforts may be concentrated and packaged. The emphasis is on community-based

13 development, under which local governments and grassroots organizations can be mainstreamed into rural development programs under CARP. This is in line with the participatory approach advocated in development projects, as stated in the country’s Medium Term Development Plan: “the most vulnerable sectors of society shall be economically empowered by expanding their capabilities and opportunities to generate sustainable sources of income, propagating awareness for positive collective action, and increasing their participation in decisionmaking.” 2.3. Infrastructure policy The Plan also sets down the public works policy of the national government. Compared to its fast growing neighbors, the Philippines has mustered a far lower allocation for infrastructure spending. From 1990 to 1992 for example, the country had an average annual allocation of 2.6 percent of GDP for infrastructure, in contrast to 4.0 percent for Thailand, and 5.8 percent for Malaysia (Kohli, 1994). The situation has hardly improved in the rest of the 1990s - the public sector share in construction value added has been mostly on the decline since 1993 (Table 2.1.). The comparison with East Asian frontrunners is even less favorable: 6 – 8 percent for Japan in its years of growth, 8 percent or more for Korea in the 1980s, and similar magnitudes for Taiwan (China), Hong, Kong, and Singapore (Mody, 1998). Table 2.1. Percentage share of construction in income and employment, 1991 – 1997

Year

Construction share in GDP

Construction share in employment

Public sector share in construction

Public sector construction share in GDP

1991

4.9

4.6

39.4

1.9

1992

5.0

4.3

42.7

2.1

1993

5.2

4.6

44.6

2.3

1994

5.5

4.7

40.1

2.2

1995

6.2

4.8

40.5

2.5

1996

5.8

5.7

36.1

2.1

1997

6.4

5.9

38.1

2.4

SOURCE: National Statistics Office.

The Plan also makes explicit the preference for basic infrastructure in rural areas, namely, “farm-to-market and feeder roads, feeder ports, irrigation, rural water supply, and missionary electrification.” For roads, the Medium-Term Plan

14 mandates the construction of an additional 6,000 kilometers of barangay (village) feeder roads, as well as the rehabilitation of an additional 26,800 kilometers, over a 5 year period. The total cost is Php 88 billion ($US 1.96 billion at an exchange rate of $US 45:Php 1). Irrigation meanwhile covers only 43% of the potential area; the Plan requires irrigation of 45% of this potential area (amounting to 766,000 ha.). River drainage and management should also be improved as as present coverage reaches only 60%. Finally, although 87% of the rural population have access to water for household use, the predominant form of access is the rudimentary first level system, which consists of a protected well or developed spring, for a small number of households (around 15) found in remote areas. The Plan evisages expansion of piped-in areas. While the benefit of infrastructure provision is often viewed in terms of the consequent productivity improvement, it can also be pointed out (e.g. Lanjouw, 1999) that infrastructure provision in poor rural areas can do double-duty by providing wage employment. The infrastructure gap may therefore be seen as an opportunity for massive and sustained implementation rural PWE programs. In this vein, the Plan does mention a preference for labor-based methods, where feasible. The emphasis nevertheless remains on the provision of assets, due to the infrastructure gap and a “catch-up” mindset among policymakers. Public works officials note that LBES methods require more time for preparation and construction, compared to equipment-based methods. The former entail a recruitment stage, and deployment of workers whose pace of work is affected by wide variety of factors, such as quality of supervision, degree of motivation, skill level, remuneration, etc. Human work is also subject to physical limitations. In contrast, capital-intensive methods rely on equipment already at-hand, which is specifically engineered for rapid work, and is subject to a smaller set of human human influences. At a certain point therefore a tradeoff between work pace and labor intensity can be observed. Such a tradeoff is not unique to the Philippines: in Cameroon for example, donor agencies and line ministries have usually sought the fastest technical solutions to infrastructure needs, without regard to the employment dimensions (von Braun, Teklu, and Webb, 1991). 2.4. Implementation framework The conventional approach towards public works in general involves contracting with private enterprises. The conventional contractors are subject to equipment regulations (fixed by law) that stipulate the types of machinery the firm is required to own. This is a significant disincentive to adopting labor-intensive technology4 4

What engineers refer to as different types of “technology” (labor-based versus equipmentbased) are, from the viewpoint of the neoclassical production function, merely varieties of capitallabor ratios for the same technology. However we shall adopt the engineers’ usage here.

15 (DAP, 1998). Similar regulations that discourage labor-based methods have been observed in other countries, e.g. Nigeria (von Braun, Teklu, and Webb, 1991). In addition, labor market regulation entails increasing remuneration for labor as firm size increases, favoring substitution of capital for labor. Hence, the conventional approach has been referred to as equipment-based. In the absence of these distortions, it can be argued that adoption of laborintensive methods is more efficient. A rough estimate of the wage at which laborintensive methods just match the costs of equipment-intensive methods is $US 4.00 (Stock and de Veen, 1996). This level is far above above market rates, especially in agricultural areas. (It is in fact higher than the minimum wage set outside the capital.) Labor-based equipment supported technology In contrast to the conventional approach, the official policy on employment generation in public works is to apply "labor-based equipment supported" (LBES) methods.5 These methods pertain to a technology in which labor, supported by light equipment, is used as a cost-effective method of providing or maintaining infrastructure to specific standards. The policy mandate for PWE in the country was made official only in 1988, by an executive order under the Aquino administration. The order required the effective and efficient adoption of labor-based methods throughout the country. Public works agencies are to incorporate such methods within their regular programs on a nationwide scale, particularly for rural-based projects. Later implementing regulations required projects under force account to draw labor from the barangay hosting the project. For rural works undertaken by a private contractor, 50% of the unskilled labor must be drawn from the host barangay. In 1999, an executive order of the Estrada administration recognized the need “to strengthen the implementation of LBES methods through institutionalization of a more focused policy direction and operational framework that will ensure the success of this strategy in alleviating unemployment.” The order mandated the creation of a special inter-agency committee to formulate and oversee a National Program on LBES. These efforts point to the special thrust in the Philippines regarding PWE schemes. Instead of devoting resources to specific employment programs, the government has adopted an integrative approach. Employment generation is to be incorporated in the regular public investment programs for rural areas. In all cases technical soundness is to be maintained. This is far from the “make-work” thrust typically associated with PWE schemes. Again, this may hark back to the 5

Incidentally, this reflects the major advocacy role played by the International Labour Organisation in promoting PWE schemes in the country.

16 “catch-up” mindset which views rural works as primarily as suppliers of muchneeded assets, rather than of jobs. Programming of labor intensity is deliberate, but moderate. Specific standards for a project to be classified as a PWE scheme have not been imposed, at least recently. Rather, project implementers are given suggestive, elaborate guidelines (in the form of manuals) to effectively substitute labor for equipment. The most prominent omission is a requirement on labor content. It is noteworthy that the definition of “labor-based equipment supported” refers only to the value added component of construction and is silent about the materials component. Thus if labor accounts for the bulk of value added, then a project can be “laborbased” even though total project cost is accounted for mostly by materials. Contracting system For the DPWH, the type of small contractors accepted in LBES works are laboronly contractors, typically informally organized. These labor gangs are hired on the basis of pakyaw contracts, which are derived from traditional construction arrangements. The principal to the contract is the public works agency, while the agent is a pakyaw team represented by a leader. A pakyaw contract engages the team to provide labor for the completion of a specific task, within a specific time period and for a stipulated amount. Equipment and tools are all provided by the principal. As the principal recognizes no employer-employee relationship with the pakyaw, legally the project is still undertaken under force account. According to DPWH guidelines, the pakyaw contract begins with formation of pakyaw groups, under the leadership of a project facilitator (from the DPWH) with the involvement of barangay leaders. Workers should belong to the various barangay associations near the project site. Unskilled labor should be drawn from the host barangay, while semi-skilled labor may be recruited from the municipality; skilled labor may be obtained from within the province. In any case qualified workers nearer the barangay are given priority. The teams are composed of groups of twenty workers. As many groups of workers as possible should be formed. The contract price is limited to 50,000 pesos. The contract ordinarily stipulates completion within one month and should conform to normal productivity standards. While payment is normally made once a month, more frequent cases may be arranged; some retention (say 10%) is to be withheld until completion. To promote transparency with communities, invitations to bid, contracts approved, and payments to contractors are all posted in public facilities of the barangay. One obvious advantage of the pakyaw is that it lightens the recruitment and supervision load on government personnel. Another is that it transfers the burden of avoiding minimum wage legislation from a national government agency to the

17 pakyaw leaders, in a rural setting where statutory wage enforcement is usually trivial. We shall be discussing the pakyaw system in greater detail in one of the case studies. Having presented the implementation framework, we may now readily discuss the country’s experience in rural PWE schemes.

Box 3. The task work system An alternative approach to pakyaw is the takay, a task work system in which a worker is paid for completion of a particular task within a specific time period (usually a day). The takay however is highly unpopular with DPWH engineers as it is very time-consuming, both in terms of preparing task assignments, and measuring performance upon completion of tasks.

18 Chapter Three. Public Works Employment Schemes in the Philippines

2.1. Overview PWE schemes for rural areas is said to have started in the early 1970s, following the devastating floods in Central Luzon. Some roads in the calamity-stricken areas were built using LBES methods. Meanwhile, for irrigation facilities, the National Irrigation Administration (NIA) applied labor-based methods in the late 1970s for communal irrigation projects. In the 1980s, foreign donors extended support for pilot PWE schemes, mostly for rural road construction, pathways, and erosion control. The pilot schemes include the following (DAP, 1998): − The Philippine Rural Infrastructure Project constructed 55 km of barangay roads from 1981 to 1995 under World Bank funding. It was implemented by the DPWH with technical assistance from the ILO. − The Rural Roads Project meanwhile constructed 15 km of roads in three provinces in 1985. A detailed study by the ILO showed that labor-based methods offered a technically and financially viable alternative technology to conventional methods for building small scale rural infrastructure. The project was implemented by DILG with funding from the USAID. − The Upland Access Project was larger scale the preceding ones; 300 km of minor roads and trails in 15 provinces were constructed beginning from 1984. The project was likewise implemented by the DILG with USAID funding. The Agency discontinued funding of the project in 1990, citing its relatively high management requirement cost. Unfortunately further details about these projects is unavailable due to loss of project documents6 and movements of project personnel. The foregoing were localized programs. It was only in 1986, under the Aquino administration, that PWE schemes were implemented on a nationwide scale in the Community Employment and Development Program (CEDP). Another major initiative for PWE was the Second Rural Road Improvement Project (SRRIP), a foreign-funded project with an explicit employment generation feature located in southern Philippines. The Local Infrastructure Development Program (LIDP) under the Kabuhayan 2000 (1994-1996) under the Ramos administration included an employment component focused on reforestation, 6

A major problem facing a country program review is the absence of an archiving policy on project documents. On the contrary there is a standing order to destroy files over five years old, which fortunately spares personal copies of government staff, and is itself poorly implemented.

19 land development, and infrastructure. Meanwhile, a Food-for-Work Program was undertaken in the country’s major sugar-producing province from 1986-1987 and from 1989-1995, in response to economic dislocation. The most recent employment generation scheme is the Rural Works Program, initiated by the DOLE, in response to the Asian crisis and the El Niño phenomenon, which ravaged the agricultural sector in 1997. The foregoing are distinct and completed programs with explicit employment objectives. These shall be discussed in detail in the rest of the chapter. Special mention though should be made of the various CARP infrastructure projects, composed of numerous individual programs, which from the late 1980s became a major impetus towards PWE generation. Rural works undertaken a support service of the CARP are administered by the DPWH and officially falls under LBES implementation. Even agrarian reform projects implemented by LGUs were influenced by the employment emphasis. The on-going Agrarian Reform Communities Development Project (ARCDP), begun in 1996 with funding from -the World Bank, identifies as an objective the creation of wage employment through labor-based rural works, under the administration of LGUs. 2.2. The Community Employment and Development Program The CEDP was aimed as a pump-priming measure to counter the employment stagnation from the 1983-1985 economic crisis. Financed by the Philippine government, the CEDP aimed to generate 48.4 million person-days of employment. Over its implementation period from 1986-1987, it attained 76% of its employment target at a cost of 7.65 billion pesos. Project composition, selection, and monitoring The CEDP involved all the major agencies. It emphasized small-scale infrastructure projects such as barangay roads, communal irrigation systems, and schoolbuildings. The DPWH (roads) were allocated 56.6% of the CEDP budget of 9 billion; DILG-LGUs were allotted 18.5%, while the NIA was given 4.4%. Minor program components included agricultural production (DA), reforestation (DENR), and livelihood creation (DSWD). Funds for the project all came from the national government, with no cost-recovery measures imposed. Actual disbursement and person-days generation is shown in Table 3.1. Employment generation was mostly accounted for by the DPWH, which was able to meet more than three-fourths of its employment target, using 78% of its fund support. Relatively weak in fund utilization and even employment generation was the NIA.

20

Table 3.1. Indicators of agency performance in the CEDP

Agency

Share in total disbursement

Percent of fund support

Share in total generated employment

Percent of employment target

DPWH

67.6

78.0

64.3

76.9

4.8

66.9

2.1

52.1

DILG-LGUs

10.0

56.5

6.9

60.9

Others

17.6

57.2

26.7

100.0

NIA

SOURCE: NEDA

Rural works projects under the CEDP were required to allocate at least 30% of project cost to labor. Each project was given a one-month completion deadline. Projects should also be consistent with municipal, provincial, regional, and national development plans. Approved proposals should be found implementable, efficient, and technically ready. In selecting specific projects, the CEDP adopted a bottom-up approach. Identification began at the barangay level, which is then taken up by the Municipal Development Council. The proposal is passed onto the higher Development Councils (municipal and regional), as well as District and Regional Offices of the DPWH. Final evaluation and approval was centralized at NEDA. The specific design of the project (embodied in the Program of Work) is supposed to involve LGUs. Implementation of works was undertaken through either by force account or through private contractors. Under the former, labor was obtained through pakyaw contracts. A distinctive feature of the CEDP was the involvement of NGOs in monitoring. Participation of NGOs was elicited by the Regional Development Councils, who monitor and inspect the projects. A special monitoring committee consisting of NGO representatives submits a status report (distinct from the implementer’s status report) on particular works. Implementation problems encountered The CEDP was the first nationwide attempt at employment generation through public works. At the same time, initial plans proved to be too ambitious: it was first intended for completion by the end of 1986, but agencies took time to issue

21 implementing guidelines. The administrative demands on the DPWH was particular heavy, as its procedures and practices were built on the conventional approach using equipment-based methods. Even upon issuance of guidelines, many engineers resorted to the conventional approach due to lack of familiarity with labor-based methods (DAP, 1998). Another reason for delay was erratic funding and disbursement. Moreover, contract procedures were cumbersome; in some cases, the absence of official guidelines led to confusion over bidding requirements. This caused late approvals, failure of contractors, and delays in payments. It was in response to these criticisms that the DPWH formalized guidelines for the pakyaw contract (see p. 17) to ensure adoption of LBES technology. CEDP funds were not entirely used for direct hiring of workers, whether for public works or for self-employment. Some of the funds were reportedly used by line agencies for personnel expenses; these included the DILG, DA, DOTC, and the even the Department of Justice. The CEDP fund allocation tended to be arbitrary and spread out across the line agencies, without thorough examination as to whether each agency could be actively involved in meeting CEDP objectives. The choice of projects was also prey to local politics. The foregoing selection process was not strictly followed as many fund allocations were realigned under pressure of local politicians. The bottom-up approach was therefore frustrated in some instances. NGO participation in monitoring was supposed to check for some of these implementation problems. The NGOs exposed bottlenecks, and even irregularities involving government personnel. At least 107 cases of alleged anomalies have been reported by NGOs and private individuals to the NEDA and DPWH. Complaints cover rigged bidding, the use of dummy contractors, payroll padding, fraudulent deliveries, materials overpricing, and use of substandard materials. However, the NGOs could have been more influential had they more actively participated in the CEDP. On the contrary, many NGOs initially invited to the CEDP dropped out due to lack of motivation, busy schedules, confusion over weakly coordinated monitoring activities, and inadequate logistic support from the government. Labor intensity was supposed to have been enforced by the 30% labor content requirement. The DPWH in fact had initially set the requirement to 40%. Subsequent experience in the field compelled adjustment. Even this was often not satisfied. Water supply and schoolbuilding projects could not comply due to the nature of the work and the need for relatively expensive construction materials. Moreover, some private contractors did not follow the labor content rule due to unavailability of manpower, peace and order problems, and time pressure. Monitoring and enforcement of the labor content rule was also impeded by confusion and unfamiliarity with implementation guidelines of the CEDP. In the

22 opinion of one DPWH official, not even a tenth of total CEDP expenditures were for labor, so serious were the departures from labor-intensive methods.7 After implementation, no provisions for maintenance were undertaken specifically for CEDP projects. Rather, the completed works were simply treated within the pool of infrastructure to be maintained by the national government (which had not yet devolved rural works at the time.) Communities were not organized to provide maintenance and assure sustainability. In general, community involvement in the project was limited to the provision of labor. Wages and targeting Remuneration for labor was on a cash-for-work basis. There was a modest attempt at food-for-work using US rations, but this was immediately terminated due to problems with distribution, storage and handling, and unacceptably poor quality of the rations. Average daily wage for the CEDP was 85.30 pesos (around $US 4.00), according to Mangahas (1986). This was higher than the prevailing agricultural wage. Information about CEDP workers is available from the Worker’s Profile Survey covering the period of July to August 1987. Interviews were conducted by implementing agencies, headed by the NEDA, through interviews of workers. We discuss this survey in some detail, as it is the only data set available on workers hired by a large-scale PWE scheme for the Philippines. The survey covers 11,086 respondents. It found the following:

7

•

A large minority of workers (36%) hired under the CEDP were females. In one region (Central Luzon) the proportion of males was only 27.3%.

•

Most of the workers (40.7%) were young adults (21-30 years old.)

•

Persons with only primary schooling accounted for 37.5% of CEDP workers. Persons with up to high school attainment accounted for another 35.8%. As many as 26.6% had reached college, with 15.1% being college graduates.

•

The CEDP intended to use local labor. Indeed, 78% of workers were residents of the barangay in which a project is located.

•

The CEDP aimed to increase employment. Before CEDP, 54.4% were indeed unemployed.

In terms of number of projects, a majority did comply with the labor cost requirement, as there were a large number of small projects in which the labor content requirement was realistic.

23

•

Of those employed before the CEDP, 86% were earning below the legal minimum wage.

•

In terms of poverty, before the CEDP, 82% of the workers hired were below the poverty line, while 54% of CEDP workers were below the food threshold. Compare this with national poverty incidence of 40% in 1985. The CEDP therefore appears to be well-targeted.

The regional profile of workers status before CEDP is shown in Table 2.3. Table 3.2 Employment and income characteristics of CEDP workers by region, in percent

Region

Emloyed before CEDP

Below the subsistence threshold

Below the poverty threshold

Philippines

45.6

53.7

81.9

Metro Manila

64.9

39.7

75.8

Ilocos

47.1

62.3

87.5

Cagayan

36.3

49.5

76.2

Central Luzon

41.1

40.5

83.0

Southern Tagalog

40.0

46.3

35.5

Bicol

44.0

55.9

84.2

Western Visayas

38.8

64.0

88.2

Central Visayas

37.7

69.1

83.1

Eastern Visayas

44.8

73.9

84.2

Western Mindanao

49.6

49.8

68.7

Northern Mindanao

32.8

58.5

85.4

Southern Mindanao

42.7

53.0

78.1

Central Mindanao

41.6

67.3

90.0

Source: NEDA

However, there are very strong reasons to be skeptical of the foregoing figures. The purpose of the survey was to evaluate whether program direction had been properly followed; the survey does not intend to measure the impact of the CEDP on incomes and livelihoods of the workers. Though the survey provides some interesting information about wage benefits, in the end we remain ignorant as to

24 the targeting performance of the CEDP. Box 4 highlights some observations regarding interpretation of the survey.

Box 4. Observations on the CEDP Worker Profile Survey •

The survey covers all types of workers of the CEDP. DPWH project workers accounted for 41.5%, while DSWD project workers accounted for 35.7%. The survey report does not disaggregate workers by agency, hence we have no idea whether the foregoing figures apply to rural works laborers. It almost certainly does not in the case of male-female ratios, as the DSWD livelihood component probably reached out to females.

•

The survey question on state of employment prior to the CEDP provides no time frame (hours worked in last two weeks, last month, last six months, etc.). In the case of rural workers with intermittent work (farmers, farm workers, etc.) this could be very important.

•

Moreover, there is no distinction between the unemployed and nonparticipants in the labor force; given the large number of female homemakers in the survey, it is likely that most of them would indicate that they were unemployed.

•

Questions on monthly income pertained to workers; no effort was made to measure household incomes, let alone per capita household incomes.

•

There was only one question in the survey form eliciting monthly income. The entry for this question consisted of ticking off one of several income brackets. This procedure is well-known to result in a serious downward bias in the measurement of income.

2.3. The Local Infrastructure Development Program (Kabuhayan 2000) As the CEDP ceased in 1987, the government attempted to adopt and institutionalize its elements in regular projects of the line agencies (Orbeta and Sanchez, 1996). One such attempt was the Local Infrastructure Development Program (LIDP) under the Kabuhayan 2000. The Kabuhayan 2000 is an umbrella interagency program involving various employment-generating thrusts, with LIPD being the PWE scheme.

25 Overall, the Kabuhayan targets the creation of 2 million jobs. Projects under the Kabuhayan were targeted to priority areas, which were the poorest 19 provinces and the fifth and sixth class municipalities. Within provinces, identified projects would employ fishers, small farmers, landless workers, upland subsistence farmers, indigenous peoples, and underemployed urban dwellers. The LIDP sought to identify projects within the regular infrastructure programs of the DPWH and LGUs in which LBES methods were deemed feasible. True to the integrative approach of the government towards PWE schemes, no new funds were infused into PWE schemes, nor was there any transfer or pooling of funds across departments or agencies. The specific targets under the LIDP were as follows: •

The DPWH would enroll 14% of its projects under LBES methods. Its own rules and procedures (particularly the pakyaw system) would be followed in meeting this target. An additional guideline is that 35% of project cost is to consist of labor. An estimated 36.5 million person-days is to be generated at a total cost of 9.5 billion.

•

The LGUs would enroll road maintenance projects under LBES. The maintenance approach to be adopted by LGUs is the length man system in which individual workers assigned to maintain particular portions of a road project through a set of pakyaw contracts. The targets are itemized in Table 3.3.

Table 3.3 Targets for LGU road maintenance under the LIDP

Target road

Length (km)

Estimated cost (‘000 pesos)

Jobs generated

30% of provincial roads

8,679

624,888

26,037*

30% of city roads

1,195

107,550

3,585*

100% of municipal roads

12,874

463,464

25,748**

100% of barangay roads

85,940

2,062,560

85,940***

Totals

3,258,462

*

3 jobs per km. per year. 2 jobs per km. per year ***- 1 job per km. per year -

** -

SOURCE: DILG

141,310

26

Strategies to meet these targets include the introduction and transfer of LBES technology to all LGUs, as well as national government assistance in the form of training, supervision, technical, and financial support. These will serve as a transition mechanism to full decentralization of LBES methods. In contrast with the CEDP, we may note a strong emphasis on local governments in the Kabuhayan 2000. This is a direct consequence of decentralization reforms of the early 1990s. In fact, the DILG was selected as the lead agency precisely due to its specialization in LGU concerns, as well as the fact that its offices down to the municipal level remained under the national government. Nevertheless, no matching grant scheme was implemented vis-à-vis the local governments, as funding was sourced entirely from within the regular budgets of the agencies and units themselves. The Kabuhayan mentions the necessity of involving NGOs; in practice however, NGOs were not included in any aspect of the LIDP, but were confined to the agroforestry component of the Kabuhayan. Participation of communities is limited to supplying labor LBES works. As the LIDP monitoring system was limited to employment generation, figures generated for the project are meager. The LGUs generated 417,329 jobs, nearly triple the target. The DPWH however generated only 9.3% of its target of persondays. Based on interviews of officials, the DPWH part was more properly implemented compared to the CEDP, as the guidelines were already in place and personnel were already becoming familiarized with LBES methods. The projects enrolled under the LIDP were CARP projects. The DPWH was selective in identifying the enrolled projects, making sure that the 35% labor content rule was followed. However, employment targets were not met partly because of slippage from expected CARP funding (see p. 11). The DPWH identified a few other remaining problems with the LBES approach. Policy inconsistencies at the donor and national level continue to undermine employment-generation. Funding agencies sometimes restrict the number of projects to be implemented on force account, thus favoring the contractor system that is biased towards equipment-based methods. The civil service commission has also tried to impose strict requirements regarding the hiring of common laborers for rural works. National government agencies also seem to disagree about the appropriate policies towards contracting. The revenue agency has attempted to tax pakyaw transactions, making them less attractive to contractors. (DAP, 1998). These problems have cropped up persistently since the time of the CEDP, the Kabuhayan, and the other infrastructure programs, but no progress has been made as of now in resolving these policy inconsistencies.

27 Finally, it is not clear whether roads are still being maintained under the length man system (or any system for that matter); after the lapse of the program, monitoring is no longer undertaken. The DPWH components were also turned over to LGUs upon completion, as is customary; however the state of the infrastructure after the turnover was not monitored. 2.3. The Second Rural Roads Improvement Project The SRRIP was a project on rural roads which began in 1986 and was completed in 1995. It covered 23 provinces in Visayas and Mindanao. The project was based mostly on feasibility studies conducted in 1982 under the First Rural Roads Improvement Project. The two major components were the rehabilitation and improvement of provincial roads There were two major components. The first was the rehabilitation and improvement of provincial roads, implemented by the DILG and provincial governments. The second was the improvement of barangay roads, implemented by the DPWH. The latter component was located in agrarian reform settlement areas and was conducted using LBES methods. The second component lasted only until 1993 and generated around 1 million person-days of employment in the course of constructing 261.8 km. of dirt roads. Before implementation, roads in two of the project areas were impassable during the rainy season, with vehicle access possible for only 60 to 125 days per year. Some of the “roads” were not much more than trails accessible only with packhorses and buffalo carts. Thus, the poor and mostly Muslim communities in the remote settlement areas were denied easy access to trade and interaction with the lowlands. Once completed though, the impacts of the SRRIP could be easily discerned. The settlement areas have become accessible all year round; travel time to the nearest major center was cut by 30 to 70 percent, and transport costs were cut by 60 to 80 percent. All the barangays enjoyed regular public transport service after the project. In constructing the roads, the DPWH implemented labor-based methods mostly with pakyaw contracts, based on the usual DPWH guidelines. In determining the task size for each pakyaw contract, a benchmark of 70.40 pesos/day (US$ 3.40) was used (based on the prevailing statutory agricultural wage) throughout the duration of the project. The actual amount received by workers is however not known, as this is purely an internal arrangement between the pakyaw contractor and the workers. Given the estimate of agricultural wages to be around 50-56 pesos at the start of the project (Subbarao, Ahmed, and Teklu, 1996), both contractor and workers could have shared in the 20-30% margin.

28 On average, over half of construction cost was spent for labor (Table 3.4). According to the SRRIP implementers, as the projects consisted of dirt roads, material costs were brought down by obtaining them on site. There were no problems in generating a relatively high labor content for the road works. Variations in materials cost roads were due to varying proximity to sediment areas (riverbeds) which provided road materials. Where no suitable “as dug” materials were available, small rock crushers were used upon boulders collected by workers. Thus environmental factors largely explained variations in labor days per kilometer. The implementation of LBES methods realized cost savings: expenditures per kilometer were 12 % lower than projected. Additional benefits include the training of workers in concreting, masonry, and even the repair and maintenance of vehicles and light equipment. Table 3.4 Selected statistics for the SRRIP Barangay Roads Component 1987-1993

Year

Km. of roads constructed

Labor days per km.

Labor cost ratio

1987

26.4

4,229

54.1

1988

55.4

3,719

54.4

1989

66.4

4,050

54.3

1990

63.3

4,230

54.8

1991

5.1

4,269

55.0

1992

29.3

4,438

55.0

1993

4.3

5,329

55.0

Total

250.1

4,114

54.6

SOURCE: Subbarao, Ahmed, and Teklu (1996).

Local governments were involved to the extent of identifying projects, and providing inputs to the Program of Works. Actual construction work was handled entirely by the DPWH. No pattern of matching grants was required. Communities were involved to the extent of assisting in labor recruitment for pakyaw group; no cost-sharing arrangement was reached with communities. Implementers though noted that the labor-based scheme was widely approved among the villagers due to its employment benefits. Upon completion, the roads were turned over to barangay officials for maintenance. It is though that community involvement in maintenance was

29 encouraged by taking a labor-based approach. There is however no information as to whether communities did go about conducting maintenance works, nor about the current status of these roads. 2.3. The Food-for-Work Program The FFW Program was implemented in the province of Negros Occidental (Eastern Visayas Region), the major sugarcane province of the country. Over 70% of the province’s agricultural area is devoted to the crop. In 1986, sugarcane output of the province had fallen by 40% from its peak in 1976. As an estimated 300,000 workers faced unemployment; alarms were being raised on the prospect of mass starvation.

Description of the scheme That year a food-for-work scheme was implemented in the province. Officially known as “Special Employment Program for Displaced Rural Workers,” it was supported by food aid from the United Nations World Food Program (WFP). The special program lasted only one year under the leadership of the DOLE. The food–for-work scheme was continued and expanded in the “Assistance to the Rehabilitation of Displaced Sugar Workers and Rural Development in Negros Occidental”, from 1989 to 1995. The lead agency was identified as the DILG, which was to work in collaboration with the Negros Occidental LGUs. The components of the FFW Program were agricultural assistance to CARP beneficiaries, agro-forestry, self-employment promotion, and rural works. The following is devoted to the rural works component of the expanded phase. Planning for the infrastructure components took note of the fact that only 6.2% of the province's cultivated area was irrigated, compared to the potential area of 19%; furthermore, that the share of community works in the province in proportion to the total stock were lower than its population share (around 3%). The project selection process is typically initiated at the barangay level. Each barangay had an “implementation group” which was made part of the organization structure of the projects. Members of farmer associations and cooperatives identified plans and programs and discussed these with barangay and LGU officials. Upon identification, the barangay sends its project proposal to the municipal office. This is then endorsed by the mayor to the DILG provincial office, which then conducts an evaluation; approved proposals are then sent to the an interagency task for final review and approval. Construction of the infrastructure was supposed to be labor-intensive. Each projects lasts around 5-6 months, and was scheduled to avoid planting and harvesting seasons. Community works were undertaken by provincial LGUs, while irrigation projects were administered by the NIA. The former targeted

30 790,000 person-days of employment, while the latter targeted 500,000 persondays. Payment was partly in the form of rice, which was obtained from the World Food Program. The rice was shipped from Australia and replaced with local rice at a fixed conversion rate (1 ton foreign : 1.3 ton local) from the government-owned grains trader. The Philippine government meanwhile shouldered the cash component of wages as well as the capital outlay for the rural works. Actual distribution started in 1991 and lasted until 1995. Releases of rice to the province occurred monthly, while implementers disbursed cash and wages weekly. The daily ration for all beneficiaries was 2 kg. of rice, for which Php 10.00 was deducted from the daily wage of Php 100.00. Accomplishments The WFP provided a total of 8,269.2 mt of rice, costing US$ 3.15 million. Total government expenditure for the FFW project (all components) was US$ 22.9 million (using a conversion of US$ 1:Php 22, the average exchange rate from 1989-1995.) Of this amount, 23% consisted of capital outlay, while 31.5% was contributed by the LGUs. The infrastructure works employed 14,257 workers and 402,035 person-days of employment, for an average of 28.2 days per person. Meanwhile irrigation works accounted for another 101,492 person-days. The infrastructure and irrigation works reached 51% and 20% of their employment targets, respectively. Irrigation eventually covered 6647 ha, exceeding the target. In all, there were 3,693 farmers benefited from the new irrigation facilities (98% of target). Targets and accomplishment for the other rural works items are listed in Table 3.5. Problems encountered The project completion report observed that the majority of the target outputs in infrastructure could not be realized due to slow disbursement of the national government, and slow liquidation of advances of the provincial government. Bidding procedures even at the local level, particularly for construction materials, were also cited as a problem. Late delivery of rice to the project site by the government trader was often observed, with the lag dragging on for months in some instances. This slowed down project implementation. Such delays have been encountered in other foodfor-work programs; in Bangladesh, for one, workers were often paid cash wages due to the late delivery of wheat (Hossain and Akash, 1993).

31 Table 3.5. Targets and accomplishments of the rural works component, FFW Program Item

Taget

Accomplishment

roads

258 km.

bridges

56 m.

324 km rehabilitation 18.8 km. concreting 468 m.

SWIMs* artesian wells

250 ha. coverage 167

137

school buildings home economics buildings community centers health stations

145 units 25 units 140 64

34 units 17 units 128 units 28 units

warehouses/ solar dryers

30

fishing port

1

toilets potable water systems.

47 units 2

riprapping/erosion protection

*

40 units 4 3.01 km.

Small Water Impounding Systems for irrigation

SOURCE: DILG

Upon completion, maintenance was to be undertaken by the concerned LGUs; the irrigation facilities though were to be maintained by the users associations. There is however no information available on the status of the infrastructure. Nevertheless, the provincial LGU has signified its intention to continue similar schemes. The program itself has won over the provincial government. It continues to advocate continuation of the program approaches and procedures, despite its termination in 1996. In 1996 it earmarked some funds to purchase rice in support of a similar scheme, making use of the rules, procedures, and standards of the FFW program. Labor policies and targeting The rural works component made no use of labor content requirements, nor did they follow the same guidelines adopted in other agencies such as the DPWH. The NIA component of the rural works is guided by its internal procedures. (Further details on NIA procedures are provided in Chapter 5.) As for the local government component, implementation was by force account. The LGUs

32 directly undertook supervision of workers and management of payroll (including distribution of food). However, the procedures in hiring labor and ensuring adoption of labor-based technology for the LGUs are not clear. Perhaps laborintensity was not pushed to the optimum, hence the observed slippage from employment generation targets. Wage payments were generous compared to market wages. In the first place, the food subsidy to rice (under the Php 5.00/kg pricing of rice) is around 44% subsidy on rice, as the market price is roughly Php 9.00 per kilo. By 1994, the wage paid (including food) was around 104 to 108 pesos, compared to the market wage of less than 90 pesos. According to project officials, many workers are willing to work for as low as 60 pesos. A baseline study (David,1993) shows that there were no significant differences in socioeconomic characteristics between participating and nonparticipating households. These suggest that the program has suffered considerable leakage (Subbarao, Ahmed, and Teklu, 1996). The rural works component did have a high proportion of women participants relative to other PWE schemes in the country. Of the total number of female participants in all the FFW schemes (including the microenterprise and farm development projects), 20% participated in the various infrastructure activities, while close to 50% participated in road rehabilitation. Often they would serve as effective alternates to their husbands who worked in sugarcane plantations. 3.4. The Rural Works Program In 1997, the Asian financial crisis struck simultaneously with the El Niño-induced drought, wreaking havoc on the Philippine economy and the agricultural sector. The RWP consists of local government projects aimed at assisting displaced workers in the aftermath of these crises. Torres (2000) is the basis of the following description. Features The RWP is a cost-sharing scheme between national and local governments. The DOLE shoulders 60% of the total wage bill, while the LGU absorbs other project expenses. The scheme was limited to LGUs with a Provincial Employment Services Office (PESO), an office dedicated to providing employment assistance, employment referral, vocational guidance, and other support to workers in the province. Workers were selected based on the recommendations of the PESO. Project areas in turn were selected based on the size of the displaced worker in the area. The "displaced worker" category includes "workers who suffered from diminution or loss of income as a result of their being displaced from work and/or

33 less working hours available." Prior to the project approval, the LGU must demonstrate its capacity to shoulder its cost share. In order to ensure the attainment of program objectives, regular monitoring is undertaken on a monthly basis, involving field staff of the regional DOLE offices. In addition to monthly progress reports, at a project’s end, the LGU proponent is required to submit a project completion report that attest to the sustainability of the project. The project was deemed successful, mainly due to the close partnership between DOLE and LGU proponents. The projects selected were mostly brief, small-scale efforts (Table 2.6). A total of 14,921 workers have been given work, with an average of 16 person-days each. Economic infrastructure accounted for around 35% of project cost (shouldered by DOLE), while social infrastructure took up 28.3%; the remainder consisted of agroforestry and other projects. Most of the economic infrastructure (76%) was in the form of farm-to-market roads. Table 2.6. Status of the Rural Works Program, August 2000

Region

DOLE expenditure, (Php millions)

Average no. of workers per project

Average duration of employment

Mandated minimum wage, 1999 (Php/day)

Philippines

24.1

43

16

-

CAR

2.6

38

12

147

NCR

0.1

55

1

198

Ilocos

1.5

29

45

166

Cagayan

1.1

49

15

160

Central Luzon

0.9

39

10

175

Southern Tagalog

2.1

26

27

188

Bicol

1.0

31

2

163

Western Visayas

2.0

44

4

146

Central Visayas

1.8

46

27

160

Eastern Visayas

0.9

57

4

153

Western Mindanao

2.4

38

54

142

Northern Mindanao

2.5

60

9

149

Southern Mindanao

1.4

60

28

145

Central Mindanao

1.2

32

32

141

ARRM

2.6

47

10

141

SOURCE: DOLE.

34

Problems identified Several problems have been noted. One is environmental: in some areas (particularly in Mindanao), frequent heavy rains delay implementation. Second, the works seemed to have not been scheduled to take into account agricultural seasonality, as the works reportedly interfere at times activities such as planting, cultivation, and harvesting. Third, the DOLE expects LGUs to implement projects by force account; in fact, LGUs at times resort to private contractors which, according to the DOLE’s, adversely affects the placement of displaced workers. Finally, the DOLE observes that all the workers employed in the RWP were males. To accommodate female workers, the RWP should begin to prioritize projects that do not involve physically intense work, such as beautification drives, eco-tourism, reforestation, tree planting, and similar projects. Labor intensity, wages, and targeting We have reviewed some of the other data for the project and have come up with our own set of observations. The labor cost ratio of a typical RWP project averages 45%. Aggregating over all projects however, the total wage bill is only 34% of total cost. Average labor cost as a ratio of project cost varies from 14% (a farm to market road) to 100% (canal drainage and desilting). The road projects taken up in the RWP tend to have the lowest labor intensity among the projects undertaken; in light of the large share of rural roads in RWP project, this seems to have impeded its job-generating objectives. However rather than allocating resources away from roads, perhaps the type of roads, the technology adopted for road construction, and the use of on-site materials should be carefully evaluated to see of LGUs are optimizing labor intensity. The DPWH labor-based units have extensive experience in this regard, but no linkages, support, or training with the DPWH have been attempted in the RWP. Wages are set at the legal minimum for the region. This alone makes the targeting of the poor in under the approach adopted by the DOLE is highly suspect. Statutory wages are set far above farm and informal sector wages and are certain to attract many of the nonpoor to participate in RWP projects. Another reason Second, the focus on “displaced” does not automatically lead to targeting of the poor; most displaced workers do not come from the ranks of the poorest, and are not necessarily heading there for good in the absence of government intervention. Third, the reliance on PESO referrals may limit selection of workers to better-educated, well-informed individuals who are not poor; the PESO may also have referred workers with skills in construction works, which may not be the least well-off. This is in fact noted by Torres (2000), who observed that the PESO recommendees included masons, carpenters, farm laborers, and unemployed persons, most of whom indicated familiarity with infrastructure work.

35

3.5. General assessment Subbarao, Ahmed, and Teklu (1996) find PWE schemes in the Philippines to be in general poorly guided by clear objectives. That is, If the objective is to stabilize seasonal consumption, public works should be timed to counter seasonal fluctuations. If the programs are intended to offer insurance, employment should be guaranteed on demand, as in India’s Maharashtra Employment Guarantee Scheme. If programs are targeted to the poor, employment should be targeted to the poor, employment should be restricted with poverty targeting indicators, means-tests, or self-targeting measures such as low wages. None of these measures are explicitly embodied in current Philippine programs despite the close association between poverty and seasonal underemployment in rural areas (p.35).

In addition, they point to the poor attention accorded to maintenance strategies: The long term employment and income creation effects of public works programs is largely contingent on the sustainability of assets. To prevent asset loss, a mainenance strategy needs to be designed in the project formulation stage. (p.35).

The quotations refer to the issues of seasonality, employment guarantee, targeting, and sustainability. We conclude our review of Philippine experience by way of discussing these issues. PWE and seasonality. Rural-based PWE schemes in the Philippines are often scheduled so as to avoid competing with agricultural activities. This is true of DPWH programs (which make the timing consideration explicit in their labor recruitment guidelines), NIA programs (see Chapter 5), and some LGUs (e.g the Food-for-Work Program). Competition with seasonal activities was cited only for the RWP. While this decision is administratively convenient, it also inadvertently achieves seasonal targeting, at least in part. Hence strident criticisms on the absence of seasonal targeting have failed to take into account this basic design feature in PWE schemes. Employment guarantee. A large-scale employment guarantee scheme would run counter to the standard Philippine approach to PWE, which is integrative. That is, its major PWE thrust is to incorporate labor-intensive methods in smallscale rural works. The last time a national make-work program was attempted (the CEDP), significant deviation from employment objectives was observed. This is attributed to a desire to preserve asset quality, combined with inadequate operational familiarity and commitment to labor-based methods. Future incarnations (i.e. the LIDP) achieved employment objectives by deliberate selection of the appropriate projects with inherently high labor intensity.

36 Current schemes (i.e. labor-based CARP projects) are clearly integrative, with employment generation strictly a secondary consideration in project design. Even this moderate approach is fraught with problems due to policy inconsistency, which need to be immediately resolved at the national government level. Whether employment guarantee programs are appropriate will depend on the cost-effectiveness of alleviating poverty and providing social protection through a make-work scheme, compared to other instruments (e.g. food stamps, microcredit, education, etc.). This in turn would partly depend on targeting performance of PWE. Targeting. While poverty alleviation is a general intention for all PWE schemes, there has been no project with specific measures towards targeting of the poor. Rather, target groups are identified, and it is assumed that hiring from these groups is equivalent to assisting the poor. The problem is that these groups are usually too broadly defined. The DPWH policy for rural works is to hire unskilled local residents, with the assumption the poor are thereby assisted with wage employment. This boils down to geographic targeting of the poor; it is not clear however, whether the provinces and villages were appropriately targeted in terms of contribution to poverty. Likewise, LGUs implementing rural works are assumed to assist their constituents who are needy of employment, but this criterion is prone to political (rather than welfare) considerations. Self-targeting, an consideration in PWE schemes, is missing in all the programs reviewed. The basis for remunerating labor is the statutory wage, which is always higher than prevailing agricultural wage. A reassessment of this policy should be conducted, with a view of exempting public works schemes with clear employment objectives from wage legislation. Another important targeting indicator is gender, which is ignored in nearly all PWE schemes reviewed. With construction being a male-dominated industry, and in the absence of deliberate hiring policies directed to women, this should not be surprising. As the FFW demonstrates however, meaningful participation of women, perhaps in lighter tasks, can (and should) be accommodated in such schemes. Sustainability and replicability. Asset maintenance is essential for ensuring long-term productivity improvement from public works; a labor-intensive approach to maintenance also ensures a long term employment effect. The absence of a maintenance strategy is another serious gap in PWE schemes. For most assets created by PWE schemes, no funds for maintenance are designated (DAP, 1998). The above suggestion by Subbarao, Ahmed, and Teklu is welltaken. A widespread and sustained strategy of employment-generating public works requires implementation down to the level of LGUs. National government officials

37 admit that LBES methods have encountered stiff resistance from some local executives, mostly due to misconceptions regarding the inability of such methods to match costs, quality, and duration of conventional methods, or even perhaps due to lobbying by construction firms. In general however, local governments may in fact furnish a strong constituency for rural works aimed at promoting local employment, given the popularity of such schemes at the grassroots (as observed for the SRRIP). A final element in maintenance and sustainability that is typically ignored is the participation of communities. In most projects they may have been involved in project identification; in the pakyaw system they were responsible for providing labor and other inputs to public works; however there has been little progress in incorporating communities in actual design, implementation, and maintenance. These two themes – decentralization and community participation – shall be the focus of our case studies.

38 Chapter Four THE CASE STUDIES: INTRODUCTION The review of experience provided some broad generalizations that are useful in making policy and program recommendations. However to obtain insights on best practices, more substantive primary data gathering would be needed. This is the purpose of the case studies. This chapter introduces these studies by presenting the methodology, providing a brief backgroun, and discussing the issues framing the analysis. 4.1. Methodology The selected projects should be exemplars of “best practice”, so as to inform the formulation of guidelines for project implementation. That is, we have chosen the relatively more successful projects that adopt approaches that are innovative in the Philippine context. Government officials helped identify the cases as “models of the LBES approach to public works. Funding for the projects all originate from foreign assistance. For one project, funding was provided by the IFAD, to allow comparisons between IFAD and nonIFAD projects. To compare the different modes of administering a project, we have chosen one project administered by an LGU, and two administered by the national government. We have also factored in CARP in the selection of the projects (two of the projects are CARP-related and one is non-CARP) due to the strategic role of CARP in rural development policy. The projects all fall under the community works classification. Officials are unanimous in attributing project success to the active participation of the beneficiary communities. The element of community participation is novel from the viewpoint of traditional rural works policy. The first case studies an intensive-participation project implemented by a national government agency. The second case (also under a national government agency) studies a participatory approach implemented by pakyaw contract. The third case studies how a community works project is implemented by an LGU. The labor-only contract and decentralization have been identified in the preceding review as major considerations in the conduct of PWE schemes in the Philippines. In addition to project documents, we gather information directly through personal and group interviews of government officials, field staff, community leaders, and project beneficiaries. Structured interviews of household heads were undertaken in the villages hosting the selected projects. Households are classified as working if at least one member was employed in the project. We interviewed heads of 20 working and 20 non-working households. The respondents were

39 selected randomly based on a roster of workers (for the working households) or of villagers (for the non-working households.) 4.2. Background of the selected projects Based on the foregoing selection criteria, the following projects have been chosen for the case studies. In particular, these projects have been cited by government officials as illustrative of best practice within their agencies (the DPWH, the DAR, and NIA), owing to excellent community performance, high asset quality, and successful employment generation. The Tumbaga Access Road Project (ARP). The Tumbaga ARP is a CARPrelated administered by the national government under the Agrarian Reform Infrastructure Support Project (with funds from the Japanese Bank for International Cooperation). It exemplifies the labor-only contract or pakyaw, the mechanism adopted by the public works agency of the country for implementing PWE schemes. The community recognizes the road as well-built, and has maintained it consistently. The Ciabu Water Supply Project (WSP). The Ciabu WSP is a local government project exemplifying the implementation of PWE under a decentralized regime. Funding originated from the World Bank through the ARCDP. The municipal government in the area constructed a potable water system for the remote upland village of Ciabu. The project made use of local labor, and upon completion was operated entirely by the beneficiaries organized in a water users association. The system remains well maintained. The Palanas Communal Irrigation Project (CIP). The Palanas CIP involved the construction of a small (50-ha.) irrigation facility in the rice lands of eastern Philippines. The Palanas CIP was undertaken by the NIA, a national government agency, under funding from the IFAD. It exemplifies intensive participation of beneficiaries in project implementation, beginning from the proposal and design stage, to actual construction and maintenance. The reason why participation is characterized as “intensive” is the unusually large share of the beneficiaries in project cost: the farm community managed to shoulder a 30% share, compared to the 10% share ordinarily shouldered by beneficiaires in similar projects. 4.3. Issues Provision refers to project resources, services, and assets. This concern covers mechanisms to ensure appropriate quantity and quality of materials and equipment. Total and itemized expenditures of the project are to be measured. Timeliness of fund disbursement has consistently turned up as an issue in Philippine projects; David and Innocencio (2000) compute the average unused funds ratio to be 20% in the late 1990s for the agricultural and natural resources agencies of the government.

40 The actual production of public works is closely related to that of provision as labor-based projects are typically undertaken by force account. Here we look at the role of government units and that of private sector entities in the works, including pakyaw groups. Furthermore, the role of government units can be observed at both national and local levels. We are interested in understanding determinants of the relative effectiveness of each level of governance for public works provision, particularly with the firm policy commitment in the Philippines towards decentralization. Community associations play a prominent role in "participatory projects". In Philippine nomenclature, such associations are called People’s Organizations (POs) to distinguish them from Nongovernmental Organizations (NGOs) which are private volunteer associations. Presumably when beneficiaries organized into POs participate in all aspects of the project (from inception to maintenance), delays are shorter, project cost is lower, cost recovery is greater, and the asset is better maintained (Alonzo, 1999). Evidence from South Africa suggests that community participation in decision-making increases the labor intensity of projects that provide community buildings, roads or sewers; it also reduces the cost of transfers to poor individuals (Hodinott, 2001). We shall check whether these findings hold true for the case studies. Employment practices relate to the recruitment, supervision, and remuneration of labor. Relevant concerns would then be: procedures for ensuring that the project uses labor-intensive methods, monitoring of quality of work, basis for remuneration (time rate, task rate, or piece rate) as well as form of remuneration. Sustainability refers to persistence of project benefits over time. We shall examine whether provision was made for maintenance of the asset constructed in the couse of the rural works, and whether maintenance was designed with employment generation in mind. Sustainability of the project approach is related to replicability. That is, the usefulness of identifying best practices for any one project depends on the degree to which these practices demonstrate innovative approaches that can be tried elsewhere. Thus in the case studies we shall identify idiosyncrasies that may hinder replicability. We include rough measures of the benefits from the asset constructed by selected projects. Where applicable, we shall also note any skill transfer and worker training attributed to the project. The major focus of the case studies however is wage benefits. How important was the project in terms of individual household welfare? The Philippine RWP for example averaged only 16 days of employment per worker. Similarly, in the major PWE program of Thailand, the average number of work-days for each participating household was only 12 days; the contribution of the program to the income of participating households averaged only 4.6% (Krongkaew, 1987). On the other hand, the longer the duration of employment, the fewer the workers that can be hired (holding labor cost constant). This tradeoff should be examined in greater detail.

41 The transfer benefits need to be stated in terms of incremental income change. Studies of PWE in Bangladesh and India estimated displaced earnings to cost around 50% to 60% of wages received; Datt and Ravallion (1994) though found that incremental benefits are actually larger due to the overestimation of displaced wages. Targeting meanwhile examines the way wage benefits were directed towards alleviating poverty. Large-scale PWE programs in India and Bangladesh have been found to be well targeted (Ravallion, 1991). In Chile and Bolivia, 71% - 77% of participants were from the bottom 40% of households (Grosh, 1991). In rural Botswana and Kenya, the poor were found to be overrepresented in PWE schemes (Teklu and Arifa, 1999). For the Philippines, quantification of inclusion of the poor in PWE schemes is prevented by flawed or missing data. On a microlevel our case studies will try to generate data on targeting performance. Specific measures taken by the case projects to ensure selection of the poor shall be identified, in addition to the outreach of the project to the poorest in the vicinity. Features of the works that repel, crowd out, or draw the poor should also be identified. Location of the project site, arduousness of the work, timeliness of wage payments, are some determinants of acceptability to the poor. The wage rate is a critical feature in self-targeting; in the Philippines wages in PWE schemes have been set near the legislated minimum, thus encouraging the nonpoor to take the available jobs. Targeting moreover means not only selection of poor workers, but also whether among workers, greater wage benefits have paid out to poor individuals. The cases are discussed in-depth in Chapter 5, on the basis of the issues enumerated in the foreoing. The targeting dimension of the cases however requires a detailed analysis, which is relegated to Chapter 6.

42 Chapter Five. THE CASE STUDIES: IN-DEPTH ANALYSIS

5.1. The Tumbaga Access Road Project The rationale for selecting the Tumbaga ARP is its implementation of the laboronly contract (pakyaw). The novelty of this arrangement lies in its intermediate status between traditional modes of production, i.e. the traditional straight contract, and the strict force account (entailing direct management of workers). Barangay Tumbaga 1 is an ARC located in Southern Luzon, around 140 km southwest of the national capital. It is a small village with 201 ha. of rice land, 121 of which are irrigated. Over 150 ha of rice land have been redistributed to 118 agrarian reform beneficiaries, giving the CARP a 97% accomplishment rate in the village. Provision and the role of government units The Tumbaga ARP consists of two gravel roads that connect a wide swathe of rice fields to a national road. Each road is 750 meters long, , with two short bridges, accompanying culverts, and ripraps. Both roads were completed within three months starting from September 1998. The specification and general layout of the road project was undertaken by the DAR, while road construction was undertaken by the DPWH. Project cost was Php 5.8 million. Financing of the foreign assisted component was generally prompt; however, the GOP counterpart, covering 9% of the project cost, was severely delayed - more than a year after the project, the GOP remained in arrears with suppliers. Overall, labor accounts for only 14.5% of the total cost. The project is nevertheless regarded as labor-based, as equipment rental contributes only 5% of the cost. Materials accounted for most of the project expenses (73%). The embankment, bridge work, and culverts probably explain the high cost of the road (running to Php 3.9 million per kilometer).8 Project documents report generating 5,131 mandays of labor (87% of which was unskilled). The project completion reports attest to the high quality of the road and to its usefulness to the community. This was also the conclusion of the assessment reports made by foreign consultants. The study also found no evidence or even suspicion of corruption.

8

The rule of thumb is that the typical gravel-type barangay road should cost Php 1.8 million/km.

43

Table 5.1 Cost breakdown of the Tumbaga Access Road Project, by road

Item

Road 1 cost

Road 2 cost

Total

% of total

45,079

45,292

90,371

embankment

184,381

157,536

341,917

stone masonry

76,659.9

85,876

162,535.9

other activities

53,376.3

60,680

114,056.3

Sub-total

359,496.2

349,384

708,880.2

12.2

78,200

52,600

130,800

2.3

Equipment rental

152,110

131,300

283,410

4.9

Hand tools

10,000

10,000

20,000

0.3

2,158,145.5

2,069,630

4,227,775.5

Supervision

102,900

97,440

200,340

3.4

Administrative costs

141,488

133,980

275,486

4.7

3,002,339.7

2,809,542

5,811,881.7

100

Unskilled labor clearing and excavation

Skilled labor (all activities)

Materials

TOTAL

72.7

NOTE: The 1998 nominal exchange rate was Php 40.9:US$1. SOURCE: DPWH

Production and employment practices In the Tumbaga ARP, there were four pakyaw contracts under different team leaders. The pakyaw groups were selected by the District Office of the DPWH. Two pakyaw leaders were from the municipality and the other two were from the

44 provincial capital. No barangay residents were taken as pakyaw leaders. The DPWH attributed this to the absence of skilled workers in the vicinity. Separate contracts were drawn up for construction materials and for equipment rental. According to the DPWH, the multiplicity of contracts makes raises transaction costs and makes monitoring more difficult. On the plus side, the subdivision of one project into different contracts increases transparency of the transaction, hence reducing the likelihood of anomaly. Supervision of the entire works was undertaken by the DPWH Engineer and project-in-charge. Labor supervision was mostly undertaken by the pakyaw leaders. The pakyaw leaders brought with them their own gangs of skilled workers, none of whom were from the area. At the time of this study, these pakyaw leaders and their gangs were unavailable for interview, as they had moved to other areas. Meanwhile for tasks done in conjunction with equipment operations, supervision of workers was also undertaken by the construction companies who leased out their equipment for the project. While pakyaw leaders received lumpsum tranches of their payment, workers from the community were paid cash wages on a weekly basis. Payments for pakyaw contracts are however made only monthly. It turns out that the construction and material supply companies served as sources of liquidity for funding wage payments. The construction and material supply companies were thus highly visible in terms of task supervision and wage disbursement, hence most of the villagers had the mistaken notion that these were straight contractors. The financing role of the construction firms is of course not a formal arrangement. Rather, the pakyaw leaders, the construction firm, and the field staff reached this arrangement informally under the exigencies of project operations. Pakyaw leaders, who were just workers themselves, could not finance wage payments, hence capital-endowed entities, such as material suppliers or equipment lessors, would bridge the gap. The latters’ interest in providing this service is to allow the project to proceed thus providing them profit from leasing equipment and providing materials. (One naturally suspects that their profit calculations had already incorporated the costs of extending this financial service). DPWH officials noted that similar financing had to be informally arranged in the implementation of some other LBES projects. We hesitate to characterize this de facto arrangement as "best practice". If cash problems afflict the pakyaw contractor, then possibly some other avenue of financing can be explicitly opened. This may range from creation of a special liquid account specifically for pakyaw contracts, to brokering arrangements with banks or other financial intermediaries. The provincial LGU and municipal LGUs involvement in the project is limited to endorsing village resolutions requesting national government assistance for

45 previously identified projects. Occasionally the municipal LGUs sent inspectors to examine the projects in progress. Overall though, the LGUs remained largely a spectator to the proceedings. Community associations The Tumbaga ARC is well known throughout the country for the successful implementation of CARP projects in cooperation with dynamic community associations. These associations are: the Irrigators’ Association (IA), a farmers’ cooperative, and a womens’ association. The IA is the users association of the communal irrigation facility and is composed entirely of farmers (i.e. cultivators; absentee landowners are not included). The latter two organizations are newly formed; the women's association started only in 1993, and is composed mainly of wives of IA and coop members. The main community association involved in the Tumbaga ARP was the IA. The IA is the oldest community association in the area and in general initiates proposals for needed projects in the community. Consultation of the community was mainly undertaken by the DAR and was limited to general design and implementation features of the access road. The village leaders (both in the barangay government and in the IA) helped identify the road layout, with access of IA members as well as easement of rightof-way as the primary considerations. In negotiating for right-of-way, village leaders had to exert moral suasion on the absentee landowners. 9 Selection of workers was directly the responsibility of the pakyaw leaders, usually upon recommendation of the IA members or the purok leader. All unskilled labor was hired from within the village; a few skilled carpenters and masons were taken in as well. The DPWH did not organize residents (whether skilled or unskilled) into pakyaw groups, contrary to pakyaw contract guidelines, as it was apparently pressed for time in completing the project. Barangay residents complained that some of the equipment inputs could easily have been substituted by their labor, thus increasing labor income. (In fact, one respondent claimed that the photographs taken as evidence that residents were working on the project were sham, as they were asked to artificially pose for these shots. This however is an extreme opinion not shared by other workers.) The other public works in Tumbaga were NIA projects under a cost-sharing scheme. The DPWH though has no cost-recovery scheme in any its projects, 9

Opposition of landowners was muted by the fact that under CARP, rental payments can only take the form of fixed lease, and that the current occupants were endowed with security of tenure. Donation of small boundary portions of their lands could hardly have worsened their position.

46 and intended to implement none for the Tumbaga ARP. Interestingly, the the IA did attempt to collect cost-sharing contributions from workers (Php 10-20 daily each worker), to fund the cost-sharing requirements in the other public works. The justification was that the projects came as a package to benefit the IA members. Such charges were however hardly collected as the DPWH did not recognize the equity scheme, and was barely aware that there was one. Moreover, as observed by the IA members themselves, their justification was weakened by the fact that the road project funds were outright grants, not loans, unlike the NIA projects. Sustainability The barangay, being the recipient of the road, is immediately responsible for these maintenance activities. Field staff observed that only minimal maintenance effort is required, such as grass clearing, at least for at least the next ten years, because of the investment in road quality and durability. From the time the road was completed, monthly clearing was undertaken, as part of the barangay's routine general cleaning drive. This work is usually done by women in the community, and recently was subsumed as a regular activity of the womens association. The project illustrates a well-replicated approach of force account implementation under the pakyaw system. Communities have typically welcomed this system and the notion of labor-based public works; Tumbaga is no exception. The one specific feature of the road project that cannot be replicated widely is the large investment in asset quality (unusual for a gravel road). For example had on-site materials been used, it is likely that durability would have suffered, and maintenance requirement would be greater. It is unclear whether the community associations would be capable of sustained maintenance works under this scenario. Asset benefits The access road had been in operation for over a year at the time the interviews were conducted. No benefit-cost analysis of the project had been done on the either the feasibility or the evaluation stage. We undertake the following rought analysis: The first source of savings would be the savings in paddy rice transport cost due to the road. The charge for transporting from the field to the national road is now Php 5 per sack, whereas before, the cost of hiring a laborer to haul a sack was Php 10-20. This amounts to around 2.5% savings in the cost of rice farming. (On the other hand, income to off-farm workers may have fallen.) However, because of low yields and a small coverage area, the benefit from transport cost savings is quite modest: a reasonable estimate would be Php 225,000 per year. This would be the lower bound benefit from the access road. From this source alone the rate of return of the project would only be around

47 1.1%. However the access roads were beneficial to the purok residents, who previously they had to cross the paddy fields in order to reach the national road. Moreover there are dynamic gains in terms of new activities in the sitio made possible by road access. As the task is too difficult, we make no effort to quantifying these important sources of benefit.

5.2. The Ciabu Water Supply Project The next case study is the Ciabu Water Supply Project (WSP), a case study in decentralized implementation of a PWE project. Barangay Ciabu is another agrarian reform community in the town of Baybay, Eastern Philippines. Baybay (1995 pop. 85,306) is a coastal town whose economy is based on farming and fishery. It is well known throughout the country for its mountain spring water, as well as the success of the water distribution program of the municipal LGU. In 1994, the municipality received the Gawad Pook (local government award) for successfully constructing a water system for all of the central districts in its 69 barangays. Nevertheless, there remain remote sitios where only few faucets serve numerous residents, who sometimes have to resort to rudimentary wells. Among the underserved communities were three citios in Ciabu. The village is composed mostly of coconut farmers, although small amounts of other crops (rice, roots, and bananas) are grown. The Ciabu central citio is around 3.7 kilometers from the national highway. The village is accessible only through a hilly gravel road, considerably upgraded by a project also of the ARCDP. Given the record of accomplishment of the municipality and the ARC status of Ciabu, the DAR decided to fund a pilot communal water system project in the barangay. The implementation of the project was however completely undertaken by the municipal LGU. The first phase was the Spring Development Project, to identify a water source and construct water tanks, a reservoir, and main distribution pipes. This phase lasted from 1994 to 1995 using national government funds coursed through DAR. The second phase involves the construction of a distribution network and is the subject of the case study. Provision and the role of the local government The Ciabu WSP began in 1997 and was finished in just two months. It involved the construction of distribution pipes, and communal faucets or hydrants. The local government was wholly responsible for implementation of the Ciabu WSP. The national government confined itself only with channeling project finance, and occasional monitoring of work in progress. The mayor himself is a young old engineer on his first term who is currently oversees other ARC projects within the municipality.

48 The ARCDP follows a matching grant policy, where the LGU was required to shoulder a local counterpart amounting to 30%. Payment of the cost-share could be in-cash or in-kind. Direct project cost was estimated at Php 246,518 (Table 3.2). Total estimated project cost (including contingency) is Php 345,00. The project was well within the administrative and financial capability of the municipal LGU, precisely because of its small scale. Total project expenditure is not however unusually small compared to the typical labor-based LGU project; in the LGU-implemented RWP, 59% of projects had a total cost less than or equal to that of the project. For large-scale community projects, however competence of municipal LGUs will have to be scrutinized on a case-to-case basis. Based on project estimates (using a wage of Php 100/day), the project is expected to generate only 437 person-days. Because of its small scale, no special equipment is necessary, as the work requires only digging pipelines, burying pipes, and masonry work on the communal faucets. Only 3 skilled workers are called for (foreman, plumber, mason), to be hired for a total of 73 person-days. Materials accounted for 64% of direct project cost. The labor cost ratio meanwhile is 19%. Incidentally, the labor cost ratio of LGU-implemented water supply projects in the RWP averages 25%. The completion report noted a major drawback, in that the budget fell short of providing quality pipes appropriate for the pressure conditions of the reservoir. Project expenses hit the mandated ceiling then in place for this type of ARCDP project. Because of this experience of the Ciabu pilot project, the cost cap was removed for subsequent projects undertaken by the LGU for the other ARCs in the municipality.

Table 5.2. Breakdown of costs, Ciabu WSP, in Php Item

ARCDP

LGU

Total

% of total cost

Indirect cost

-

49,372

49,372

16.8

Direct cost materials labor

190,040 17,081

39,394

190,040 56,475

64.2 19.0

Total cost

207,121

88,766

295,887

100

SOURCE: Municipality of Baybay

49 No problems in financing were encountered. The only delay experienced was the lengthy processing of the LGU's project proposal. The waiting period was one year. Once the program was approved, funding from the ARCDP proceeded smoothly. The municipal office was also prompt in providing funds and procuring materials. There was however one incident in which the local politician attempted to block funding for the ARCDP in Ciabu; this lobby was successfully thwarted (see Box 5). Employment policies and the role of the community As early as 1995 (three years before commencement of the works), the municipal LGU was already consulting with barangay officials and community residents regarding the needs of the barangay. Thus the potable water system was conceptualized; during project construction, the municipal engineer was assisted by the barangay captain in locating the water hydrants (who in turn consulted residents informally). However, the final project design (as codified in the program of work) was drawn up by the engineer with few (if any) specific inputs from the community. Decisions about the ratio of labor cost in total cost, material procurement, etc. were made by the municipal engineer. He explained that specific details could easily be planned at his level without need for further consultations. (The barangay captain did have the option of rejecting poor quality materials. This option however never had to be exercised this case). The municipal LGU was subject to a cost-sharing scheme with the ARCDP. Likewise, it engaged the barangay government in a cost-sharing scheme, with the community counterpart consisting solely of labor. Labor provision was mobilized through a traditional community effort locally known as pintakasi, which involves community work by groups, typically on a rotating basis. On any given day, a group of 20 to 40 villagers was at work under the supervision of a municipal engineer. In the earlier spring development phase, the labor contribution format had already been implemented with great success. No payment for wages was made, which was acceptable to the community as the amount of work required was modest. Each worker contributed only a few days of work. Their only "incentive" was free lunch and snacks during the workday, prepared by the women of the community. Interestingly, such meals served in community self-help efforts have been referred to as a type of "food-for-work" scheme (Tajgman and de Veen, 1998). The water supply phase however required greater labor outlay; the payment of wages was therefore warranted. The daily wage was disbursed on-site by the municipal project officials in the form of cash during Saturdays. The barangay and municipal governments set a one-day weekly contribution (usually taken on Saturdays). According to the workers, the importance of community ties and a sense of solidarity motivated them in part to donate the extra day.

50

For each week during the duration of construction, one citio would be assigned to provide labor on a rotating basis. The sitio leaders recruited residents within their jurisdiction and set work assignments for individuals. The project engineer though had some leeway in keeping some productive laborers for a longer interval. No problems in availability of manpower were reported. The farmer's cooperative in Ciabu, in addition to the barangay government, was active in preparation and social mobilization for the project. It should be understood that in this case the elected village leaders could hardly be distinguished from the community association leaders. The Ciabu farmer's cooperative was founded by the current barangay captain; while the WSP was being undertaken, he was also then serving as the cooperative chairman. Moreover, some of the barangay councilors and officials also served as leaders in the cooperative. Given the bias of the DAR towards extending assistance towards organized farmer groups in ARCs, it was natural that the barangay officials would also carry their identity as cooperative leaders in transacting with both national and municipal governments. Maintenance Upon completion, the water system was turned over to users, consisting of 220 households. The cooperative permitted water users to access the system using pipes directly extending into their homes. There are 60 piped-in users. The rest of the users gain access through communal faucets (20 in all). Initially the system was managed by the same farmer’s cooperative mentioned earlier. However, with a change in leadership in the cooperative, many water users raised complaints against the new management. Inasmuch as many users were not members of the coop, the barangay officials convened a general assembly of water users, where they decided to constitute a new user association, the Ciabu Potable Water System. The association membership fee is a Php 50, while the monthly fee is Php 10 for households with piped-in water, and Php 3 for households using communal faucets. Some members of the farmer cooperative though object to the new arrangement and refuse to pay their user fees. The vast majority of villagers though have ceased in their complaints about the water system management. The Ciabu Potable Water System is performing its maintenance tasks well in terms of collecting water fees, repairing damages, and regulating the use of the communal faucets. These regulations include prohibitions against bathing and laundry washing. In the absence of a piped-in connection, households undertaking such activities must resort to a nearby river In the medium term, facility maintenance is assured by through the activity of the water association. (However see the discussion in Box 3, on threats to the

51 project.) The reason given by field staff and barangay leaders is the fact that, under cost-sharing scheme, the villagers were compelled to contribute towards project provision. The end-users, having invested their time in construction, would therefore not waste their earlier efforts by allowing the system to deteriorate.

Box 5. Communities against internal and external threats External threat to the works. The following illustrates the role of communities in resisting arbitrary moves of local politicians. The ARCDP projects in Ciabu were earlier threatened by the machinations of the congressional representative, who requested the national authorities to realign the projects to benefit her favored municipalities. In response, LGUs at the village and municipal levels sent petitions to the agrarian reform secretary to thwart the representative's request. The realignment in request was turned down. In numerous other cases however the representatives are not as frustrated in their abuse of power (as in some instances reported in the CEDP). They can only be countervailed by communities who remain united and vocal in their stand. Internal threat to maintenance. Communities of course are not themselves free from activities of its members Some time after the completion of the construction phase management problems arose. After a rash of complaints from the community regarding the lack of transparency in the use of the water fees, the move to constitute a Ciabu Potable Water System was approved in a Project benefits and targeting general assembly. The decision remains an acrimonious affair within the village. Interestingly, the farmer cooperative management who were against the Potable Water Supply management were hoping to credit the monetary value of the water system in their balance sheet in order to secure a large loan from a Manila-based NGO. Some POs are obviously not above naked displays of opportunism. The dispute reached a high point when the cooperative filed a case with the municipal council, where the matter remains unsettled. It is expected though that the authorities would rule in favor of the larger majority of water users represented in the current water system management.

An alternative explanation would be the selection effect: under an equity scheme, only communities with a long term interest in and capability for sustaining project benefits would accept a cost-sharing arrangement. From the viewpoint of economic rationality, this alternative explanation would be the more likely one, as

52 earlier efforts are after all “sunk cost”, which would therefore not be a factor in their decisions to maintain or not maintain the water system. In the long term a serious problem will inevitably confront the Ciabu water users. The rate of Php 10 per month for piped-in water users is too low, given their consumption levels as well as depreciation of the facility. A major breakdown of the water system may require extensive rehabilitation, which may once more require external assistance. To make the system completely self-reliant within the reasonable life-span of the facility, higher rates need to be collected, especially from the piped-in users. The most sustainable arrangement would be to meter the water usage and water fees the piped-in users should charged metered rates. Asset benefits The asset provided was of satisfactory quality, according to Ciabu residents. Currently the system shows signs only of ordinary wear on the pipes and faucets. At the source however, upon ocular inspection one can observe that some pipes used have indeed been substandard PVC pipes. According to the project engineer this was a result of the abovementioned cost cap. Among the households whose water source shifted from spring-wells to the communal faucet, the daily time savings was 31.5 minutes. This however conceals many other benefits. Previously, large household reserves of water were needed, which was very inconvenient. The presence of a water system has greater benefits in periods of drought, where water flow may weaken to a trickle, and queues could be long and time-consuming. For example, in a nearby village households suffered from endless queing at a well during the El Niño drought . The Ciabu citios were then savoring the first few months of their new water system and avoided this problem entirely.

5.3. The Palanas Communal Irrigation Project The third case study treats an IFAD project in irrigation that highlights an intensive participatory approach of farmer-beneficiaries. This approach is anchored on an innovative cost-sharing scheme. The NIA participatory approach: a backgrounder The participatory approach has been a standard format followed by the NIA since the 1980s. Communities are organized into Irrigator's Associations (IAs), which are formal association of cultivators who operate a communal irrigation facility. The principle behind the IA is the necessity of cooperation among water users. Irrigation facilities are common pool resources: whereas persons outside the coverage area may be easily barred from accessing water, abusers within the

53 area cannot be as easily excluded. For example, one type of problem is the the “tail-end syndrome”: in the absence of an enforced rotation system, downstream users typically get the least water supply (Cruz and Cruz, 1984). The participatory approach consists of the active involvement of an IA in all phases of a project (planning, construction, and maintenance) as well as in shouldering project cost. The NIA’s approach was developed over years of experience in community-based irrigation works. The IFAD has had a lengthy experience funding the institutional aspects of communal irrigation facilities in the Philippines. There are admittedly cases in which the approach was merely prescriptive; “It is not uncommon, for example, to find communal irrigation systems constructucted by the NIA (the National Irrigation Administration) which, at turnover time, are rejected by the community (p. 219)” (Alonzo, 1999). Nevertheless instances of successful compliance with participation guidelines can also be found. In any case, few other government agencies conducting public works schemes have as thoroughly ingrained the participatory approach as their norm for provision. The NIA requires organization of the IA prior to construction of facilities, with at membership including at least 80% of farmers in the coverage area. The inclusion of the IA in the planning and construction phase might be seen as facilitating the flow of information between end-users and implementers, as well as checking agency problems (i.e. incompetence or negligence of the NIA in construction). Supervision over project construction is still exercised by NIA staff, composed of a supervising engineer and an Irrigation Development Officer or IDO. While the supervising engineer oversees the technical side of the irrigation works, the IDO focuses on community organizing, and plays the key role in coordinating and even initiating the participation of communities in the irrigation system. IA participation in irrigation works is anchored on cost sharing. The cost share of the user association is called equity, which (as in the Ciabu WSP) can take the form of labor contribution. There are two kinds of equity schemes. In the 10:90 scheme, the IA shoulders 10% of project cost by the time of completion, while the IA amortizes the remaining 90% for a period not exceeding 50 years. Meanwhile in the 30:70 scheme, the IA shoulders 30% of project cost by the time of completion, while 70% is shouldered by the NIA. The NIA offers these two schemes as a menu of options – itself an innovative feature in public works programs. In practice, the 30:70 option is seldom chosen. The reason is that this scheme would require large contributions from the IA members during the construction phase, as well as a high degree of cooperation between them. Few communities are optimistic enough to take this option.

54

Box 6. Communal irrigation in the Philippines An irrigation facility with a coverage area of 1,000 ha. or less is classified as a communal system. Rivers with small catchment areas can been controlled with simple brush dams (which need to be rehabilitated after a significant flood). For centuries, local communities have built such dams to irrigate up to a few hundred hectares, with little assistance from any central authority. (NIA, 2000). Most notable among these traditional associations are the zanjeras, which survive in northern Philippines and is based on apportioning each farmer a parcel both at the upstream and downstream portions of the water source (Cruz and Cruz, 1984). These and other traditional associations are prototypes of the modern IA.

Provision and financing The Palanas CIP consists of an extensive upgrade of an existing facility. The previous system consisted of a small dam and earthen canals. This was constructed in 1975, the year the Palanas IA was organized. The IA was able to complete amortization of the facility, which was well-maintained until 1991. That year the dam was washed out by massive flooding brought about by a massive typhoon. Much later (June 1999) the Palanas CIP began construction of a larger dam better suited to local weather conditions. The main canal was also widened and cemented. The works were completed on January 2000. During the datagathering period communal irrigation system was only partly operational; nevertheless the beneficiaries interviewed consistently attested to the sufficiency of construction materials and adequacy of workmanship of the current system. The computed direct project cost is Php 4.18 million, whose breakdown is shown in Table 5.3. Of the direct project cost in Phase 2 (Php 3.37 million), one-third is in the form of labor. (The indirect cost is the amount charged by the NIA for administrative overhead. It excludes the salaries of the on-site staff for the project duration.) Financing of the project is generally adequate. No delays were reported with respect to securing approval and release of funding from the funding agency. For this particular project, NIA was also timely in disbursing funds. This contrasts with findings of other studies that highlights the slow fund utilization characteristic of NIA operations (e.g. David and Innocencio, 2000).

55

Table 5.3. Programmed and actual expenditures for the Palanas CIP, in Php (as of 31 December 1999)

Item

Programmed

Direct cost

% of direct cost

Actual

% of direct cost

3,371,518

3,173,877

428,571

428,571

Total cost

3,800,089

3,602,448

Equity

1,011,455

30.5

1,011,519

32.0

Labor

589,663

17.5

424,745

13.4

Materials

437,010

13.0

586,773

18.4

Indirect cost

NOTE: The 1999 nominal exchange rate was Php 39.1: $US 1. SOURCE: NIA Region 8 Office.

Community participation The project IDO noted that organizational work was limited to reconstituting the previous association, facilitating processing of papers (especially the IA’s rights to the water source), and coordinating with the NIA. The IDO observed that the members were highly motivated in the irrigation works, met often to make group decisions, and collectively enforced these decisions. No major organizational or managerial problems were encountered. In the design stage, the Palanas IA insisted upon a strong and durable facility, even at some added cost (given their experience with typhoons). One of the major changes they insisted upon was a canal width that was wider than that specified in the original NIA plan. The project plan and schedule, which is detailed in a program of work, was made upon consultation with the IA. The IA officers agreed that the program of work was generally followed throughout the course of the construction, with delays attributed only to occasional typhoons. To arrive at the program estimates, the NIA project staff and the IA met in a series of equity programming sessions. The IA was able within technical limits to design the largest possible labor and material contribution that could be mustered to meet the 30% requirement. This type of discretion is absent from the cost-sharing arrangements in the previous projects in Ciabu and Tumbaga. It is consistent with evidence that greater labor intensity accompanies community works characterized by intensive participation (Hoddinott, et.al. 2001).

56 Table 5.4 breaks down the equity by type of work. Based on the equity programming, the IA will freely contribute around 53% of labor cost; donated materials make up the remainder of the equity. What happened in fact was a shortfall in labor contribution, but a significant overshooting in materials contribution. The latter is mostly composed of sand, gravel, and boulder. Materials were costed at their market value, i.e. Php 526.41/m3 for sand, Php 507.14/m3 for gravel, and Php 824.00/m3 for boulder. During the data-gathering period, the IA was about to complete its equity requirement. Table 5.4. Breakdown of equity contribution for the Palanas CIP, in Php

Item of work

Direct cost

Labor equity

Materials equity

sand

gravel

boulder

lumber

Diversion works

1,792,498

304,470

34,526

58,012

168,253

10,576

Canal system

1,258,689

285,196

111,696

45,993

0

7,954

0

0

0

Others Total

320,331 3,371,518

0 589,666

0 146,222

104,005

168,253

18,530

SOURCE: NIA Region 8 Office.

The involvement of the IA extends to such details as project finances, and material procurement, all according to NIA's guidelines on beneficiary participation. A process called "cost reconciliation" permits the IA (through its financial committee) to audit project financial records. Another committee joins the NIA canvassing and procurement personnel. The extensive participation of the IA is understandable given their interest in obtaining quality material, and properly accounting for the equity contribution. Moreover, these steps ensure that transactions are above board, avoiding the rumors of corruption tainting many public works projects. No delays were reported because of the involvement of the IA in these procedures. There were some problems associated with adoption of the intensive participation format. First, the remarkable equity contribution of the farmers relies greatly on accounting sleight-of-hand. In the case of materials, the farmerbeneficiaries did not actually finance any material purchases out-of-pocket; instead, these materials were dredged from the river bottom close to the facility's water source. Procurement and hauling of these materials relied mostly on the

57 labor provided by the IA members, hence the crediting of these items to the beneficiaries' equity. However, the valuation of these materials was not based on dredging and hauling (actual) cost, but rather was imputed a value based on the market price. This certainly overestimates the cost of materials. Second, it turn out that the "community" which participated in this project is only a small portion of the total number of farmers covered by the Palanas communal irrigation system. Out of the 72 water-users expected to benefit from the facility, only 23 are members of the IA (of whom 5 were recruited after the project started). This is contrary to the implementing guidelines on IA membership. (At the field level such deviations are allowed as long as the IA voices an intention to expand membership.) This smallness is viewed by the NIA as a weakness in project implementation. Both the IA members and the NIA staff agree that the narrow membership base is due to the selectivity of the leadership. According to the IA leaders, the IA has excluded farmers who may be a source of dissent within the IA. Most particularly, farmers who are “beholden” to their landlords, and might therefore introduce the influence of the elite within their group, are barred from membership. Moreover, the members are well screened so as to recruit only the reliable and cooperative individuals. The small size of the IA is itself conducive to cohesion, thus avoiding transaction costs of coordinating large groups, as well as the vulnerability of large groups to the free rider problem. Labor policies Responsibility for providing construction labor, both skilled and unskilled, was vested on the IA. Work slots were assigned to each individual IA members, who must fill it personally or recommend a substitute. Prioritization in recruiting non-IA members, as outlined in the standard procedures of the NIA, is as follows: the member's relatives, residing in the coverage area, followed by other residents in the same area, and finally, residents in the barangay. The NIA for its part hired only two skilled personnel (a machine operator and foreman) and no unskilled workers. Wages are paid in cash twice a month on a 15-30 schedule by the NIA. There is no evidence of underpayment. Despite the overall timeliness in project funding, there were occasional delays (up to one week) in the payment of the wages, which dissatisfied some of the workers. Role of government units According to NIA officials, its participatory approach is made possible because projects are implemented on force account. Except for the IA, there is no other private entity involved in the project. (NIA staff mentioned that irrigation works under straight contract typically deviate from their participatory guidelines. Such

58 contracts normally occur when funds originate from congressional representatives.) The local government had no direct hand in the project. The municipal and barangay governments did however make two important contributions: first, in providing equipment support; and second, in resolving right-of-way problems. LGU support was greatly facilitated by the fact that the IA president is also the barangay captain and heads the association of barangay captains in the municipality. The IA president was therefore able to coordinate easily with the town mayor and other local officials. With regard to equipment support, the mayor lent municipal dump (rent-free) upon request of the barangay captain. She also assisted the IA in borrowing (rent-free) bulldozers and trucks from a nearby LPG factory. This equipment support was not trivial: while NIA equipment rental ran to Php 66,370, private equipment costs in terms of fuel and parts alone amounted to Php 52,544. As for right-of-way, objections from landowners initially threatened to halt the project.10 Serious opposition was however forestalled through efforts of moral suasion, first by the barangay captain, and by the town mayor. Sustainability and replicability The IA is in charge of managing and maintaining the facility. Maintenance funds would be generated by charging a user's fee, which may be collected even from non-members.11 Prospects for maintenance of the facility appear favorable, barring severe calamities, based on the cooperativeness within the current IA, and the organization's track record in maintaining the previous facility. Furthermore, the burden of paying water fees is smaller as the cost of amortization has been eliminated. This raises the question of free-riding by users who did not make equity contributions upon users who did make equity contributions. First of all, the wage payments were satisfactory enough to the members (see Chapter 6); second, the contributors reason argue that free-riding is irrelevant; in the end, what mattered to them was that they personally stand to benefit from the facility even accounting for the cost of their contributions.

10

Interstingly, the reason for their opposition lay in the land reform history of the area. In the 1970s, land reform covered only rice and corn lands. Landowners were then rumored to be planning conversion of the rice land into sugar land, as well as eviction of the tenants. To forestall this a group of tenants requested an irrigation facility under the government irrigation program, thus establishing that rice was the major crop in the area. Henceforth the irrigation facility was a source of animosity between landlords and some of the tenants. 11

The right to charge for use is given by the government's Water Board

59 The characteristics of this group are quite idiosyncratic, which raises some doubts about the sustainability and replicability of the intensive participation format. Their situation contrasts with that of other areas in which no IA exists, and the IDO must undertake protracted community organizing efforts. In other areas in which IAs exist, farmers typically settle for the easier 10:70 scheme; moreover, frequently members are apathetic, or whether active, factious. Region 8 staff have voiced some disappointment with the overall performance of IAs in the region, with cultural or social "values" as the culprit behind the belligerency within some groups. It is indeed unrealistic to expect this approach to be disseminated widely in community irrigation works. The point however is not to replicate this approach widely, but to adopt it systematically in locations cohesive and capable groups exist or can potentially be organized. The merit of a equity scheme menu is very clear, as the menu allows such groups to self-identify in an incentive compatible manner. Benefits from the asset We rely on the NIA feasibility study to estimate ex ante benefits from the irrigation facility. The estimated total project cost is Php 5.24 million (higher than actual because of additional overhead items and contingencies). Project benefit is expected to take the form of increased rice yield. The current rice yield at 2.4 mt/ha; with the project, the yield should increase by 92% to 4.6 mt/ha, a 92% increase. This corresponds to an increase a 112% increase in net farm income per hectare, reaching a total of Php 11,616. The IRR is estimated at 13.9%. Our own estimate of the NPV is Php 0.96 million using a discount rate of 15%. The assumptions on project impact made in the feasibility study seem to be wellfounded (e.g. our own estimate of farm yield in 1999 is 2.9 mt/ha.) It is likely moreover that variability of yield from season to season (and from year to year) would decline as a consequence of the facility. For the poor farmer such stabilization of income would be invaluable.

60 Chapter Six CASE STUDIES: WAGE BENEFITS AND TARGETING

So far little has been said about wage benefits and none at all about targeting in the case studies. The analysis of these issues in this chapter should make clear the relevance of the studies to poverty alleviation. 6.1. The Tumbaga ARP The average wage paid for the Tumbaga ARP is Php 143 daily, as determined by the pakyaw leaders. This is approximately 5% below the farm wage of Php 150 prevailing in the area. The prevailing farm wage is the highest among the project sites, and may reflect the relative progressiveness or high cost of living in the Southern Luzon region, just 120 km. from the national capital. The margin between actual and market wage may be expected to lead to some degree of self-targeting. Intertemporal targeting could have also been promoted by the timing of the works during the months of September to December season, are mostly lean months for rice farming (except September, which is harvest season). Table 6.1 presents a profile of workers and nonworkers. The two groups are demographically similar, except that the control group is on average less educated than the worker group. Farm size is greater in the nonworker group. Though there were about the same number of landless workers among in both groups, duration of off-farm employment is greater among the workers. Worker income is slightly lower than nonworker income. Farm income share is lower for workers (consistent with their greater reliance on off-farm income and employment). Even for the nonworker group, farm income accounts for only a quarter of total income; rather, nearly half of income arises from nonfarm sources. Most of the farmers in the sample do not report any nonfarm or off-farm occupations. Of persons who were engaged in off-farm employment, nonfarm work accounted for 100 days of employment in both worker and nonworker category over the previous year. The rest of off-farm employment is largely taken up in hunus, an informal agrarian contract in which the landless worker supplies labor to a farmer in exchange for a share of the harvest. The type of work provided for hunus is mostly light except during harvest season. The table suggests that workers tend to earn relatively less of their income from farming and more from off-farm work compared to nonworkers.

61 Table 6.1 Characteristics of workers and nonworkers in the Tumbaga ARP

Characteristic

Average

Worker Age of household head (years)

Nonworker

47.1

48.2

Schooling of household head (years)

6.6

4.9

Household size

4.8

4.6

Number of landless

12

13

Number of landowners

3

4

Farm size (ha)

1.1

2.0

Days off-farm work per year

Annual income (Php)

98

68

40,604

43,025

Sources (%) Farming

14.9

26.9

Off-farm employment

41.1

35.4

Project employment

14.5

-

Other nonfarm employment

26.9

33.3

2.6

4.4

Other nonfarm sources SOURCE: Author's data.

The duration of project employment averages only two weeks. The reason is the rapid turnover of workers, which led to an implicit rotation of work. Once reason for the rapid turnover is that there were four different pakyaw leaders and teams; this implied discontinuity in worker selection. The rotation system reflected both the desire of the pakyaw leaders to appear accommodating to all the members of the community, as well as rationing of work. Table 6.2 examines the distribution of household per capita incomes within worker and nonworker category. We focus on per capita figures as this is a better measure of the living standard of the household. For the worker group, both withand without-project incomes of the worker group are shown.

62 Table 6.2 Comparison of average per capita income and project employment in the Tumbaga ARP

Control group

Worker group without project

with project

Days of project work

Per capita income bottom 10

4,200

2,991

3,574

15

bottom 15

6,558

5,083

5,729

25

whole group

9,126

9,258

9,926

22

-

0.12

0.13

1.0

international line

7

9

8

official line

12

15

15

correlation coefficient with duration of project work number of the poor

SOURCE: Author's data.

To calculate the without-project income, we first note that project wage earnings accounts only for a small proportion of the incomes of workers. Seven of the workers worked for two weeks or less, while eight worked between three weeks and one month. For the eight, we imputed foregone earnings of the few workers who insisted that some employment opportunities were indeed missed due to the project. The adjustment results in without-project per capita income of workers only 7.2% lower than the with-project income. On a per capita basis, the without-project incomes of workers are surprisingly higher than the incomes of nonworkers. On first glance this indicates imprecise targeting. However, without-project per capita incomes of workers are much lower than those of nonworkers if the top five per capita incomes are removed. This is confirmed in the poverty headcount: 7 workers are classified as poor by an international poverty line (see Box 1), while 9 of the workers would be classified as poor on a without-project basis. The disparity in the poverty headcount is larger going by official poverty lines. The top five without-project per capita incomes of workers turn out to be mostly persons with high nonfarm incomes, who were hired because of their carpentry skills. To a limited extent project employment to the unskilled seems to have been welltargeted. This could be explained by the self-targeting effect of the below-market

63 wage. However, one should not discount the possibility that recruitment recommendations by IA members to pakyaw leaders were directed to underemployed persons in the community. The project could be said to have reduced the poverty headcount by just one. This need not be held strongly against the project if the wage benefits were biased towards the poorer workers. However, the correlation coefficient between without-project income of workers and days of work in the project is positive though insignificant. The same characterizes the correlation coefficient of withproject income of workers with days of work. Apparently, the rotation system adopted in the project essentially randomized the selection of the workers, removing any bias towards the poorest. With rotation, the profile of workers would begin to reflect the profile of residents in the village, many of whom are not poor. The nonpoor still suffer from underemployment, and therefore might desire a stint in the project despite its lower-than-market wage as long as its schedule is flexible. The rotation is not entirely disadvantageous; while the "leakage" problem in targeting may have been worsened, the "undercoverage" problem may have been reduced. The poorest village residents, particularly the landless, did manage to join the project. Their work stints however were all too brief.

6.2. The Ciabu WSP The modal wage among the sample workers of the Ciabu WSP is Php 75. This is lower than the planned daily wage of Php 83 (adjusted for equity). There were however seven skilled workers were paid between Php 90 and Php 150 daily wages. These persons pushed up the sample average wage to Php 103. These wage benefits arrived at a most auspicious time, when the El Nino drought was then wreaking havoc on the countryside. The worker-nonworker profile is shown in Table 6.3. Workers and nonworkers are demographically similar, except for average age (which is much lower for the workers). Several young persons managed to enter the Ciabu WSP, including some non-household heads. Tenurial status is similar across categories, except for smaller average farm size for the workers. The farm sizes in Ciabu are apparently much greater than those of other case study sites, but the quality of these lands is much poorer, because of the hilly terrain. The annual incomes for Ciabu are low as a whole compared to other case studies. Several facts should be noted: the year is 1997, hence the purchasing power is somewhat lower than in the other projects which came later. There is also the phenomenon of the El Niño, which reduced coconut and rice yields.

64 Table 6.3 Characteristics of workers and nonworkers in the Ciabu WSP

Characteristic

Average

Worker

Nonworker

38.4

45.1

Schooling of household head (years)

6.5

6.3

Household size

6.3

5.0

Number of landless

4

4

Number of landowners

9

8

Farm size (ha)

2.1

3.5

Age of household head (years)

Days off-farm work per year

Annual income (Php)

81

73

39,307

19,850

Sources (%) Farming

45.1

53.1

Off-farm employment

26.4

17.1

Project employment

7.7

-

Other nonfarm employment Other nonfarm sources

15.0

21.0

5.9

8.8

SOURCE: Author's data

Nevertheless, it is safe to conclude that the Ciabu residents are poorer than those of the other project sites, consistent with the fact that Ciabu is a marginalized upland village devoted to coconut growing. Farming is the biggest income source for either category. For workers off-farm employment is the next biggest source of income, while for nonworkers it is nonfarm employment. Project employment has only a small contribution to worker income. The very low nonworker incomes suggest that the poorest members of the community have not been covered by the project. The tabulation of per capita incomes is shown in Table 6.4. Using a method similar to that in the Tumbaga case study, wages and employment for the counterfactual without-project situation are estimated. The incremental

65 contribution of project wages to income is only 3.2%, implying that nearly 60% of project wage earnings simply replaced alternative earnings. This is mainly because the skilled workers earned a greater part of the project’s wage benefits, and claim that they could have found employment elsewhere. Table 6.4 Comparison of per capita incomes, Ciabu WSP

Control group

Worker group without project

with project

Days of work on the project

Per capita income bottom 10

2,285

2,196

2,367

13

bottom 15

2,772

3,262

3,420

12

whole group

4,310

6,768

6,986

14

-

0.49

0.50

International line

14

9

9

Official line

18

17

17

0.87

1.19

1.16

correlation coefficient with duration of work

Number of poor

Coefficient of variation

SOURCE: Author's data

Average per capita incomes of the workers (with or without the project) are much higher than those of nonworkers. Most of the nonworkers are poor on the basis of the international line; this same line classifies only a minority of workers as poor. (The official line fails to distinguish significantly between the poverty headcount of the two categories.) The superiority of workers’ incomes originates from the top half of sample workers. The bottom half of households are quite similar (in fact for these households the workers are slightly worse off.) The difference becomes increasingly pronounced as the higher income households are added. The coefficient of variation of worker income is large relative to that of the control group, confirming the relatively larger dispersion of per capita incomes in the former category.

66 Furthermore, within the worker category, the distribution of wage benefits does not favor the poorer individuals; there is in fact a mild tendency for project employment to rise with rising household income. This is seen in the fewer days worked by the bottom half of the workers, and more clearly, the positive value of the correlation coefficient for with-project or without-project incomes. As in the Tumbaga ARP case, the mistargeting arises from accommodating skilled workers, and rotating of work assignments in compliance with the pintakasi tradition. Of the three workers with the highest per capita incomes, one is a record-keeper and a high school graduate, coming from a family with one overseas worker remitting foreign earnings, while two are skilled carpenters. (When these workers are excluded, the correlation coefficient of per capita income without-project is -0.15). Furthermore, excluding the poorest worker (who worked continuously on the project for two months), the poor workers contributed only 7 days of labor on average. While work rotation did allow many of the poor to be included in the project, it also ensured that wage benefits could not be adjusted to provide greater assistance to the most needy.

6.3. The Palanas CIP For the Palanas CIP the wage was set at Php 200, which is the lowest salary in the government pay scale but way above the minimum wage. Effectively though the wage is 100, as 50% is collected as equity. This is obviously another case of overvaluation (just as with the materials component of equity; see Chapter 5.3). We estimate the daily farm wage in the locality to range from Php 60 (slack) to Php 75 (peak). Thus, the effective wage (equity excluded) remains significantly higher than the alternative wage. Targeting therefore might be a problem unless there is effectual means to ration out the nonpoor. Based on the standard NIA procedures, the program of work was scheduled to avoid conflict with the peak agricultural season around November to December, and June to July. In reality though delays in the project led to conflict with farm occupations. The farmers who worked actively on the project claimed that their role in farming over the project year was confined to overseeing the farm work of his wife and children, and occasional fieldwork on Sundays.12 A comparison of workers and nonworkers (Table 6.5) shows noticeably different characteristics between the two categories. Workers are younger than nonworkers and are somewhat better schooled. Only a few workers are landless, while none in the control group are landless.

12

The project in Palanas (as in the other case studies) adopted a six-day work week..

67 Table 6.5 Characteristics of workers and nonworkers in the Palanas CIP projects

Characteristic

Average

Worker Age of household head (years)

Nonworker

47.1

56.5

Schooling of household head (years)

6.6

5.9

Household size

4.8

4.2

Number of landless

3

0

Number of landowners

6

6

Farm size (ha)

0.8

1.1

Days off-farm work per year

Annual income (Php)

24

52,258

59

37,622

Sources (%) Farming

26.6

55.2

Off-farm employment

6.3

3.2

Project employment

44.8

Other nonfarm employment

21.0

30.5

0.9

14.2

Other nonfarm sources

-

SOURCE: Author's data.

Income of workers is much higher than that of nonworkers. The biggest chunk of earnings of the workers came from the project, in contrast to the earlier case studies in which project work hardly made a dent in worker incomes. Meanwhile the bulk of earnings of the control group came from farming. Off-farm work frequently takes the form of angkon, a local version of the hunus. As for nonfarm employment, the project had only two workers with nonfarm occupations. The first had a non-regular job which was not affected by the irrigation works, and the second suffered some income loss due to his participation in the works (owing to the moral suasion of his brother, who happens to be the IA president). Other worker households with large nonfarm incomes actually earned them through household members not working in the irrigation project.

68

Per capita income figures are summarized in Table 6.6. Calculation of withoutproject incomes requires some care, given the substantial amount of time that individual workers have devoted to the project. We impute the number of days employed and wage as follows: for the landless worker: for the farmer:

average off-farm wage*150 days average off-farm wage*20 days

Table 6.6. Comparison of per capita incomes for the Palanas CIP

Control group

Worker group without project

with project

Days of employment

Per capita income bottom 10

4,018

2,211

5,688

177

bottom 15

5,757

3,905

7,893

167

whole group

9,615

9,340

13,690

155

-

-0.44

-0.04

1.00

International line

7

10

5

Official line

13

14

13

correlation coefficient with duration of work

number of poor

SOURCE: Author's data.

Angkon work accounts for the remainder of the man-year of the landless worker. The imputation of 150 days for landless workers is based on their own estimates, while for the farmers the off-farm work estimate is based on the off-farm employment of the control group. On the aggregate, the increase in per capita income due to the project was 47%. The ratio of incremental contribution to project income is 84%, meaning only 16% of project wage earnings replaced incomes that would have been earned in the absence of the project. Averaging incremental incomes over workers, the proportion shoots up to 142%. That is, per capita incomes of workers increased

69 by nearly two-and-a-half times on average. This follows from the prolonged employment duration, the relatively high wage offered by the project, and the limited off-farm and nonfarm opportunities in the area. Without-project incomes of workers are much lower than those of the control group. The disparity is larger for the bottom half of workers. The are only 6 poor persons in the worker category compared to 13 without the project, representing a poverty incidence reduction from 65% to 30% (at least during for the duration of the works). Meanwhile only 35% of the control group are poor - higher than the without-project poverty incidence of the worker group, but lower than the withproject poverty incidence. The project appears to have been well-targeted to the poor as well as to the poorer workers. It is noteworthy that targeting performed well despite the high wage and the absence of a deliberate policy to hire the poor. Hence the most important factor in the targeting performance of the project must have been the characteristics and hiring decisions of the IA members. If the IA member or his son worked, he tended to be poor; if another, then the designate tended to be poor, e.g. a landless relative. Targeting happened to coincide with the fact that the project had to be located in a very poor area with highly fragmented landholdings and limited nonfarm opportunities. Within the worker group, there seems to have been some exercise of targeting, i.e. the poorer workers get to work more days in the project. The correlation coefficients confirm this pattern: the without-incomes of the workers are negatively correlated with days of project employment; the with-project incomes however are very weakly correlated with the employment duration, suggesting that the project contributed to equalizing worker incomes. The correlation coefficient between days of project work and share of nonfarm income in total income is –0.75. Clearly, the better-off workers were less willing to devote more man-days to the project. The targeting of wage benefits towards the poorer workers was not seen in the other two cases as these implemented a rotation system. 6.4. Overall evaluation of targeting Earlier we had stated that no explicit antipoverty targeting methods had been adopted in the projects. However, specific practices in each project inadvertently determined targeting performance. the Palanas CIP, the fact that the IA members and their relatives were among the poorest in the community led to a superior targeting of wage benefits. Targeting was weaker in the Ciabu WSP, and more so in the Tumbaga ARP, first due to adoption of rotating work assignments, and secondly due to the premium placed on hiring skilled labor. However the recruitment practice is not the only location-specific variable; other variables such as income, demographic characteristics, tenure, etc. vary with

70 location. To bolster our contention we must be able to control for these other variables. We do this through ordinary least analysis. The left hand side variable is the days of project employment. Correlates on the right hand side are worker characteristics, household per capita income, and the location dummies (The Palanas dummy appears only as an interaction term with household per capita income, making it a “slope shifter”; the Ciabu and Tumbaga dummies are represented as “intercept shifters”.) We take the log transformation of the variables (except for the dummy variables) to capture nonlinearities in the correlations. The terms "correlation" and "correlate" are chosen carefully; we are not attempting here to construct a causal model, particularly because worker characteristics and incomes are related, and the sample collection was more purposive than random. Hence, the coefficients should be interpreted as indicators of correlation; the t-values likewise should be interpreted in terms of correlations, rather than statistical values for testing hypotheses. The results of the OLS correlation are reported in Table 6.7. Table 6.7 Results of correlation analysis Pooled workers

Pooled workers (Palanas dropped)

Coefficient

t-value

Coefficient

t-value

Distance from project (km)

-0.512

-1.427

0.029

0.051

Age (years)

-0.281

-0.623

-0.348

-0.677

Schooling (years)

0.129

0.451

0.041

0.124

Household size

-0.369

-1.283

0.112

0.318

Size of landholding (ha)

0.296

0.940

0.062

0.175

Is a landowner (landowner = 1)

-0.416

-1.318

-0.511

-1.329

Income (w/o project, in Php)

-0.020

-0.150

0.056

0.405

Palanas income (interaction term)

-0.455

-2.173

---

---

Tumbaga sample

-5.928

-3.090

0.286

0.634

Ciabu sample

-6.526

-3.632

--

Constant

10.550

4.379

3.122

Worker characteristic

adj-R

2

0.62 SOURCE: Author's calculations

-0.06

0.212

71 Chapter Seven. FINAL EVALUATION: LESSONS LEARNED AND BEST PRACTICES

In this concluding chapter we synthesize and consolidate lessons learned based on the review of PWE experience and specific observations from the case studies. Some recommendations for project design have been incorporated. 7.1. On employment generation. The Philippines has pursued an integrative approach to PWE schemes: instead of providing make-work schemes that happen to generate productive assets, it has tried to incorporate labor-intensive methods in its general infrastructure program. The country does not adopt as a matter of policy explicit quantitative indicators of labor intensity to set apart PWE schemes from other infrastructure projects. While laudable in intent, such an approach has blurred the distinctiveness of PWE schemes. The CEDP experience met instances of tradeoffs with technical adequacy. However other experiences (i.e. the LIDP) have shown that labor content rules are not inconsistent with technical adequacy as long as the implementer is selective about the types and locations of the projects to be enrolled under PWE objectives. Despite this, hard and fast labor content rules have not been adopted since the LIDP. Instead, policy embraces LBES technology; in practice though the vagueness of the LBES rubric has permitted numerous departures from optimum labor content in public works. We recommend the imposition of strict guidelines on labor content for priority PWE schemes, in order to highlight the distinctiveness of such schemes. A benchmark figure of 50% project cost paid to labor is reasonable. To preserve asset quality, the enrolled priority projects must be chosen carefully for technical soundness and suitability for labor-based construction. Box 7 presents a checklist of the types of works that may be suitable for LBES technology. Meanwhile, other PWE schemes can follow the usual integrative approach advocated in government policy. 7.2. Remuneration systems PWE schemes currently hire on a cash-for-work basis. Earlier projects under a food-for-work format were bogged down by the high cost of food distribution. As for basis of remuneration, the experience with piece rate payment and taskbased payment has generally been negative from the implementers’ viewpoint. Rural works engineers find the daily metering of output (as well as setting individual assignments under the task-based payment) too time-consuming.

72

Box 5. Projects suitable for labor-based methods Item

Typical labor content

Potential labor content

Low cost housing

25-35

30-40

Social buildings

20-30

25-35

Water reticulation

5-15

25-35

Stormwater

5-15

40-50

Sanitation

5-15

25-35

Secondary roads

5-15

30-70

Dams

10-20

50-70

Railways

5-15

20-30

Electrication

10-15

35-45

Irrigation

15-25

30-70

Forestry

25-35

35-45

SOURCE: DPWH

We recommend this policy be continued, i.e. payment should typically take the form of cash, except for special food-for-work schemes to support specific emergency or relief operations within a particular time frame. Payment should also take the form of daily or weekly wages. This runs counter to some authors’ advocacy of piece or task rating (e.g. Ravallion, 1996). In terms of Philippine experience such remuneration schemes are too difficult to administer. 7.3. The pakyaw system. From the viewpoint of the public provider, intermediate to daily wage payment and task-based payment is the pakyaw contract. The pakyaw system works because the provider saves on supervision effort, while the agent faces incentives to produce specific outputs within a reasonable time frame. Unfortunately DPWH expertise and experience with the pakyaw system is not widely disseminated, particularly among LGUs. Within the national government, policy inconsistencies related to hiring policies, wage setting, and taxation of pakyaw contractors need to be resolved.

73 We recommend that the pakyaw system be widely disseminated through collaborative PWE schemes implemented by the DPWH and LGUs. The national government should immediately establish a uniform policy exempting pakyaw contracts from tax and pakyaw workers from the usual labor market regulations. Further advantages from the pakyaw system can be realized if greater emphasis were placed on preparatory organization of communities. Organizational work could also solidify the pakyaw groups from loose associations to active coalitions of workers. While DPWH guidelines provide for such preliminaries, in practice time pressure and absence of resources (especially personnel) make prior community organizing unrealistic. This attitude is to be contrasted with the NIA, which provides a specialist community organizer for each irrigation project. In our case study, the DPWH implementer left community organizing to the DAR. But the DAR’s organizing work extended only to community associations not directly involved with road maintenance (e.g. farmer cooperative, IA).

Therefore, special guidelines should be drawn up directing additional personnel and generous time allocation (say, 1 to 3 months) for community organizing prior to pakyaw contracts. To ease the administrative burden, the special guidelines could be applied only to priority PWE schemes identified in 7.1. Finally, if a project is interagency (as for CARP infrastructure projects), community organization should be elevated to the status of an interagency activity as well. 7.4. Timeliness of wages Adequate preparation should be made to ensure frequent and timely payment of wages, which is critical to workers’ welfare in PWE schemes. However, the inflexible schedule of pakyaw payment, as well as the generic viscosity of cash flow afflicting government operations, makes this requirement a problem. The pakyaw contractor has no capacity to fill in the liquidity gap. Implementers should anticipate financing problems and make preparations to ensure that at least the wage payments do not suffer from the general liquidity malaise. In the case of the Tumbaga ARP, a material supplier bridged the financing gap. Such improvisation cannot be expected to yield consistently adequate results in rural PWE schemes. We recommend instituting a special revolving fund large enough to finance three months worth of pakyaw contracts. The possibility of creating a new contracting system for small and medium companies owning light equipment should be carefully studied, so as to provide an alternative to cash-strapped pakyaw contractors. Guidelines for formulating such a contracting system are enumerated in Box 6.

74

Box 6. Guidelines for an alternative contracting system (Edmonds and de Veen, 1996) There are several requisites for this widespread adoption: • A training period is needed to pass on the organizational and management skills • Contract procedures should be streamlined, that is: o design specifications should be more flexible, to allow application of labor-based methods utilizing on-site construction materials. o contract documents need to be simple and straightforward. o government agencies need to decentralize approval and monitoring of contracts o large contracts can be broken down into smaller contracts of manageable size. • General constraints on credit need to be removed or mitigated. Government agency may have to back up contractors in their relations to banks. • It is essential to ensure an adequate cash flow in paying contractors. One system may be to introduce government certificates that can serve as overdraft facilities. Contracts must include penalties for late payment.

7.5. Local governments Our case project had the ingredients that are supposed to make decentralized implementation work: an LGU sensitive to the basic needs of its local communities, with the means to obtain funds, and the know-how to conduct community works. We note however our case study pertains to a small-scale project which placed few administrative demands on the municipal LGU. However the national government appears technically more capable of mplementing the larger-scale projects. As local governments are now vested the major responsibility for rural infrastructure, undersupply of larger works within localities has become a problem. Part of the reason though why capacities of the LGUs are stunted is weak financing from the national government. Successful cases such as the one recounted here could help generate financial and other support from the national to the local units. At the village level, elected barangay officials play a crucial role in implementing development projects. Often these officials are simultaneously leaders of the major village associations. We saw the importance of village officials not only in the case of the LGU project, but in our other cases as well; they were part of

75 project preparation and design, essential in negotiating access rights, as well as part-organizers of labor recruitment efforts. For the communal irrigation facility, not only was the village head active in the works, he also obtained support from the municipal LGU in terms of equipment support, as well as ensuring acceptability of the facility to landowners. Doubts may be raised as to how reliable these leaders are in representing the wishes of the villagers. However, the question can be raised at any level of leadership or of government; the question is at which level leaders are most sensitive to the wishes of their constituency with regard to local rural works. We recommend orienting project administration towards local governments. Admittedly, at the outset there will be some difficulties finding LGUs with proven capacity and experience, and receiving their appropriate share in the national government revenue. One way to deal with this difficulty is to engage national agencies in collaborative ventures with LGUs, at least initially and for larger projects with village-specific benefits. If a project is small enough and not technically demanding, then LGUs may implement it in its entirety. At the village level, barangay officials should take a lead role in mobilizing and coordinating with villagers. 7.6. Cost sharing or equity Participation of communities from inception, planning, design, to construction and maintenance of rural works, is integral to project success. IFAD involvement in Philippine irrigation projects is characterized by a laudable emphasis on community organizing and participation. The concrete mechanism for generating and shaping this participation is however often overlooked. The self-help format, involving equity contribution from communities, is one such mechanism (Alonzo, 1999). Cost sharing should be motivated by more than just stretching a tight government budget. It should serve as an incentive mechanism for galvanizing communities to supply inputs in which they have comparative advantage, namely information, labor, and monitoring effort. Collective effort should be seen as more than a by-product of communal ethos, or of traditions of mutual help (e.g. pintakasi), but rather of incentive design. Of course, the prerequisite to all these is that the community must clearly understand that equity contributions are compulsory for recipients of public assistance. One often hears the argument that cost sharing encourages communities to take responsibility for infrastructure maintenance. A better explanation would be that cost sharing is also a screening device: only those communities ready to ensure the enjoyment of long term benefits from an asset would be willing to accept an equity scheme. As a corollary, we may reason out that the accuracy of selfselection is improved by a wider menu of equity schemes available to communities. This is precisely what was observed for the case of the Palanas IA.

76

While we have applied the foregoing arguments about cost-sharing to communities, practically the same reasoning applies to local governments. This justifies the requirement of matching grants for specific development projects. In turn, LGUs can extend equity schemes to beneficiary villages. We recommend a policy of equity schemes and matching grants to reinforce participatory and decentralized approaches to development projects. The policy may extend a menu of options to be voluntary selected by the barangay government, the community association, or the LGU. Further study is needed regarding the optimum structure of these cost sharing arrangements. 7.7. Labor contribution and self-targeting In providing equity contributions, communities should be given the option of contributing labor (rather than cash), a factor they possess in relative abundance. Effectively, this entails paying workers a lower wage compared to a no-equity arrangement. Thus, in our case studies, quoted wages were set at or near the statutory minimum; however, effective wages were set nearer to farm work wages because of the equity contribution. This is an important tactic for achieving self-targeting wage rates, particularly if there are political and regulatory problems in evading statutory wages. Ideally workers under this arrangement are paid approximately their reservation wage, because of prevailing underemployment. Incidentally, the fact that the poor “contributors” are adequately compensated at the margin renders moot the problem of nonworkers “free riding” on the equity “contributions” of the workers. We recommend the equity scheme allow contributions in the form of labor. The equity labor contribution effectively becomes a wage reduction device that can also serve to self-target the poor. 7.8. Location and targeting Geographic targeting is widely practiced in Philippine PWE schemes. Despite the broad intention of alleviating poverty, selection of workers is merely “targeted” to the unskilled members of the community. Thus targeting of the poor depends in a significant way on prior selection of a poor village. We recommend a rigorous procedure of geographic targeting be adopted in identifying the beneficiary communities for rural PWE. That is, the poorer areas may be selected in succeeding order: poor provinces, then poor municipalities of the poor provinces, and poor villages of the poor municipalities. Funds may not be sufficient to cover all such villages; hence, those which are ARCs should be given priority. 7.9. Rotating work and targeting

77

In two of the case studies, work assignments were rationed to numerous villages explicitly to encourage community participation. Thus many of the poor workers’ stints were cut short, whereas some of nonpoor were able to join the works during their idle days. While coverage of the poor may have increased commensurate with an increased number of workers, it remains likely that wage payments were less efficiently allocated towards alleviating poverty. We recommend the adoption of guidelines discouraging communities in self-help schemes from adopting rotating work assignments. Instead, persons paid belowmarket wages should be allowed to work for the duration of a project. 7.10. Maintenance That assets characterized as common pool resources be managed and maintained by users’ associations is hardly debatable. However other assets benefit more than a narrow group of users, e.g. feeder roads, flood control structures, etc. Thus a user association may not be held solely responsible for upkeep; some joint maintenance between “frequent users” and the government, particularly the LGU, may be required. However the absence of maintenance provision is endemic to public works projects. The general tendency is to approach asset provision in a piecemeal manner; the effort of producing an asset differs from the effort of maintaining it. Unfortunately this may lead to wastage as there are economies of scope in coordinating both projects into one sustained effort. We recommend PWE project to adopt a clear procedure that appropriates funds, sets up a monitoring system, and identifies lines of responsibility, with respect to maintenance. The procedure should be adopted simultaneously with the program of work for constructing the asset. Hopefully, such moves would help eliminate an implementation mentality that is too project-oriented but not welfare-oriented. 7.11. Final remarks In the Philippines rural works is only weakly linked with employment generation directed at poverty alleviation. This link can be strengthened considerably. This report finds that production of rural infrastructure and other community works can be feasibly undertaken with a high employment content oriented to the poor. Funding agencies can aid in strengthening these links, by devoting funds for rural works programs that have a high employment content, and are directed towards hiring the poor, without compromising asset quality. The findings and conclusions of this report can perhaps serve as a preliminary basis for formulating projects along these lines.

78 REFERENCES This Report has relied on various mimeos, handouts, project reports from both local and foreign sources, not separately listed here. Abesamis, T., ed. (1995). Technical Cooperation for the Management of Change. National Economic Development Authority and U.S. Agency for International Development: Manila. Alonzo, R. (1999). "Local governance and poverty alleviation." Causes of Poverty: Myths, Facts, and Policies. A. Balisacan and S. Fujisaki, eds. Quezon City: University of the Philippines Press. Balisacan, A. (1997). “Getting the story right: Growth, redistribution, and poverty alleviation in the Philippines.” Philippine Review of Economics and Business 34:1-37. Balisacan, A. (1999). "Poverty Profile in the Philippines: An Update and Reexamination in the Wake of the Asian Crisis." University of the Philippines School of Economics. Processed. Balisacan, A., and A. Babila (1994). “Employment structure and urban growth.” Spatial development, land use, and urban-rural growth linkages in the Philippines. A. Balisacan and F. Medalla, eds. Pasig, Metro Manila: National Economic Development Authority. Bird, R. (1994). “Decentralizing infrastructure: For good or for ill?” Policy Research Working Paper 1258. Washington D.C.: World Bank. Briones, R. (2000). "Property Rights Reform In Philippine Agriculture: Framework For Analysis And Review Of Recent Experience". Philippine Institute for Development Studies Discussion Paper Series 2000-29. Cabanilla, L. (1999). “Achieving food security: some critical points to consider.” Food Security in the Philippines. L. Cabanilla and M. Paulagui, eds. Quezon City, Philippines: University of the Philippines Center for Integrative and Development Studies. Capuno, J. and O. Solon (1996). "The impact of devolution on local health expenditures: Anecdotes and some estimates from the Philippines." Philippine Review of Economics and Business 33:238-318. Clay, E. (1986). "Rural public works and food-for-work: A survey." World Development 14:1237-1252.

79 Cruz, W., and M. Cruz (1984). “Nonseparable externality, collective action, and irrigation institutions.” Philippine Economic Journal 23(2-3): Datt, G. and M. Ravallion (1994). “Transfer benefits from public works employment.” Economic Journal 104:1346-69. David, F. (1993). “The socioeconomic condition of households in the Food-forWork Project target communities in Negros Occidental: a baseline study.” Iloilo City: Central Philippine University. Processed. David, C. and A. Innocencio (2000). "Key Indicators for Public Expenditure in Agriculture, National Resources, and the Environment." Philippine Institute for Development Studies. Processed. Deininger, K., F. Lara, Jr., M. Maerten, and A. Quisumbing. (1999). "Agrarian reform in the Philippines: Past impact and future challenges." Processed. Deolalikar, A. (1995). "Special employment programs and poverty alleviation." Asian Development Review 13:50-73. Development Academy of the Philippines (1998). A Study on the Potential Expansion of Labour-based, Equipment-Supported Infrastructure Programs in the Philippines. International Labour Organisation AsiaPacific and the Development Academy of the Philippines:Manila. Edmonds, G., and J. de Veen (1992). "A labour-based approach to roads and rural transport in developing countries." International Labor Review 131:95-123. Elazegui, D. and P. de Guzman (1999). “Devolution in the agriculture sector.” Food Security in the Philippines. L. Cabanilla and M. Paulagui, eds. Quezon City, Philippines: University of the Philippines Center for Integrative and Development Studies. Esmara, H. (1987). "Creating employment through the labour-intensive public works programmes: The Indonesian experience." Philippine Review of Economics and Business 24:214-236. Estudillo, J. and K. Otsuka. (1999). “Green revolution, human capital, and off-far employment: changing sources of income among farm households in Central Luzon, 1966-1994.” Economic Development and Cultural Change 47: 497-523. Freedman, D. (1990). “Special employment programs in developed and developing countries.” International Labour Review 129:165-184.

80 Fisman, R. and R. Gatti (2000). “Decentralization and corruption: evidence across countries.” Annual Bank Conference on Development Economics, Washington D.C., April 18-20. Garnier, P. and J. Majeres (1992). "Fighting poverty by promoting employment and socio-econmic rights at the grassroots level." International Labor Review 131:63-75. Gaude, S. and H. Watzlawick (1992). “Employment generation and poverty alleviation through labor-intensive public works in least developed countries.” International Labor Review 131:3-18. Grosh, M. (1994). Administering Targeted Employment Programs in Latin America: From Platitudes to Practice. World Bank: Washington, D.C. Hoddinott, J., M. Adato, T. Besley, and L. Haddad (2001). “Participation and poverty reduction: issues, theory, and new evidence from South Africa.” Discussion Paper No. 98. Food and Consumption Nutrition Division, International Food Policy Research Institute: Washington D.C. Hossain, M. and M. Akash (1993). "Public Rural Works for Relief and Development: A Review of Bangladesh Experience." International Food Policy Research Institute Working Papers on Food Subsidies No. 7. Huther, Jeff and Anwar Shah (1998). “Applying a Simple Measure of Good Governance to the Debate on Fiscal Decentralization.” World Bank Policy Research Working Papers, no. 1894. Imschoot, M. (1992). "Water as a source of employment." International Labor Review 131:125-137. Institute for Agrarian and Rurban Studies (2000). The Agrarian Reform Beneficiaries Project Monitoring and Evaluation Survey: Phase 3. Processed. International Labour Organisation (1983). Target Wages, Actual Wages and Small Contractors: The Pakyaw System in the Philippines. Vol. 1. Main Report. International Labour Organisation: Manila. Lanjouw, P. (1999). "Rural nonagricultural employment and poverty in Ecuador." Economic Development and Cultural Change 48(1):91-122. Leones, J. and S. Feldman (1998). "Nonfarm activity and rural household income: Evidence from Philippine microdata." Economic Development and Cultural Change 46:789-806.

81 Kohli, H. (1994). Infrastructure Development in East Asia and Pacific. World Bank: Washington, D.C. Krongkaew, M. (1987) "The economic and social impact of Thailand's rural job creation programs." Philippine Review of Economics and Business 24:237-272. Mangahas, M. (1986). "Rural Employment Creation in the Philippines and Thailand." Rural Employment Creation in Asia and the Pacific: Papers and Proceedings of the ADB-ILO Regional Workshop on Rural Employment Creation, 24-28 November. ILO: Manila. Mody, A. (1998). "Introduction." Infrastructure Strategies in East Asia: The Untold Story. A. Mody, ed. Economic Development Institute of the World Bank: Washington, D.C. National Economic Development Authority (1998). Medium Term Philippine Development Plan: 1998 – 2004. National Economic Development Authority: Manila. Orbeta, A. and M. Sanchez (1997). Micro Interventions for Poverty Alleviation in the Philippines. New Delhi: South Asia Multidisciplinary Advisory Team (SAAT), International Labour Organization. Quitazol, J. (1996). "Local government fiscal behavior: A look at the pre- and post-devolution scenarios." Developing an Action Plan for the Financing of Local Government Projects with Social and Environmental Objectives. R. P. Alonzo et. al., eds. Submitted to the National Economic Development Authority. Ravallion, M. (1991). “Reaching the poor through public employment: Arguments, evidence, lessons from South Asia.” World Bank Research Observer 6: 153-75. Serafica, R. (1999). “Beyond 2000: an assessment of infrastructure policies.” Discussion Paper No. 98-07. Makati, Philippines: Philippine Institute for Development Studies. Stock, E. and J. de Veen (1996). "Expanding Labour-based Methods for Road Works in Africa." Technical Paper No. 347. World Bank: Washington D.C. Stock, E. (1996). The Problems Facing Labour-based Road Programs and What To Do About Them: Evidence from Ghana. World Bank: Washington, D.C..

82 Subbarao, K., A. Ahmed, and T. Teklu (1996). “Selected social safety net programs in the Philippines: targeting, cost-effectiveness, and options for reform.” World Bank Discussion Paper No. 317. Tajgman, D. and J. de Veen (1998). Employment-Intensive Infrastructure Programs: Labour Policies and Practices. International Labour Office: Geneva. Teklu, T., S. Arifa (1999) "Who participates in labor-intensive public works in Sub-Saharan Africa? Evidence from Rural Botswana and Kenya." World Development 27(2):431-38. Torres, J. (2000). “The Rural Works Employment Program.” Mimeo. Department of Labor and Employment. Von Bran, J., T. Teklu, and P. Webb (1992) “Labor-intensive public works for food security in Africa: Past experience and future potential.” International Labor Review 131: 19-33. Von Bran, J., T. Teklu, and P. Webb (1992) “Labor-intensive public works for food security: Experience in Africa." International Food Policy Research Institute Working Papers on Food Subsidies No. 6. World Bank (1996). A Strategy to Fight Poverty: Philippines. Country Operations Division, East Asia and Pacific Region.