AID AND POVERTY REDUCTION: THE INDONESIAN EXPERIENCE TULUS TAMBUNAN Faculty of Economics, University of Trisakti, Jakarta-Indonesia

AID AND POVERTY REDUCTION: THE INDONESIAN EXPERIENCE TULUS TAMBUNAN Faculty of Economics, University of Trisakti, Jakarta-Indonesia INTRODUCTION In the beginning of the new order (NO) regime in 1966, the average Indonesian earned only roughly US$50 a year; about 60% of adult Indonesian could not read or write; and close to 65% of the country s population lived in absolute poverty. Facing this condition, the new order government launched five-year economic development plans, with the first plan started in 1969, and made several crucial economic policies in the 1970s and 1980s, including liberalization in investment, capital account, banking and external trade. During the new order government, industry and agriculture were two priority sectors. To support development of national industry, the government adopted two subsequent strategies. Started first with an import-substitution based industrial development strategy, focusing on labor-intensive industries such as textile and garments, footwear, wood products, and food and beverages, followed latter by assembling industries of automotive, and then gradually shifted to an export promotion strategy by reducing some import tariffs and export restrictions, also focusing on labor-intensive industries. To support agriculture, the government adopted modernization or intensification of agriculture, known as the green revolution, as the main strategy. This strategy which was marked by the adoption of better of agricultural inputs [such as fabricated fertilizers and improved seeds], new technologies [including modern irrigation system], modern marketing, and mechanized processes of production, along with massive public investments in rural areas. The main aim of this strategy was twofold: to boost agricultural productivity and thus to achieve the goal of rice self-sufficiency, and to increase real income per capita in rural areas, and thus to reduce rural poverty and hence national poverty. All these steps had generated a sustained rapid economic growth in the 1980s up to 1997, just before the Asian financial crisis occurred. During that period, the growth strategy accompanied with government s efforts to reduce the poor with special designed measures, which included labour intensive (particularly for unskilled workers) projects (such as construction of village road and technical irrigation); more access to primary education and health care facilities for the poor families through government subsidies; development of backward villages through Inpres Desa Tertinggal (IDT) program under the Presidential Instruction for development of isolated/under-developed villages; and development of micro and small sized enterprises. Together with these pro-poor policies, the rapid and sustained economic growth has not only led the real income per capita to increase, but also the poverty incidence (people living under current official poverty line as percentage of total population) to fall substantially. It is generally argued, however, that without foreign aid in the context of official development assistance (ODA) together with official foreign loans, the NO government could never achieve such results. Indeed, during the 1970s and 1980s, Indonesia was among important recipient developing countries of ODA. The early 1960s saw a much acrimonious exit of foreign donors as the first President Soekarno during the old order (OO) era, asked the donors to go to hell with their aid (Hill, 1996). In sharp contrast, when Soeharto came to power in 1966, as he immediately realized that with the country s current condition in that time he could never lift the economy from its disastrous collapse that marked the end of the OO without big push from foreign capital, and thus he reopened relationship with the West (including World Bank and IMF), and welcomed foreign aid and loans. The donors enjoyed most congenial relationship with the NO government 1

to the extent that Indonesia received aid, including a large trade adjustment loan from the World Bank in 1987 without any conditionally attached (Mosely, et al., 1991). This paper is a contribution to the literature on aid, economic growth and poverty reduction. Despite an extensive empirical literature in this area, existing studies came with uncommon conclusions regarding the effectiveness of aid for economic growth and poverty reduction in developing countries. This paper examines the role of foreign aid or development assistance in helping reducing poverty in Indonesia. RECENT DEVELOPMENT OF OFFICIAL AID Official flows of development finance to developing countries have shown a dramatic shift from loans to grants over the past three years (Figure 1). Foreign aid grants have increased by a cumulative total of US$20 billion during the period, while net official lending has declined by about $52 billion, implying a $32 billion decline in net official flows (aid and lending combined). In 2003, ODA increased by $10 billion to reach $69 billion, following a $6 billion increase in 2002. This represents a nominal increase of 18 percent in 2003, following a 2002 increase of 11%. But in real terms (adjusting for inflation and exchange-rate changes), ODA increased by just 5% in 2003 and 7% in 2002 (Figure 2). Figure 1: Official debt flows and foreign aid grants, 1990 2004 Source: World Bank (2005). From the perspective of the recipient countries, net ODA flows have grown gradually over the past few years. ODA has been quite stable as a share of GDP in recipient countries, averaging just over 1% since 1996, below the high of 2% reached in 1991 (Figure 3). For the poorest recipient countries (excluding those in conflict or post conflict), ODA has averaged just over 2% since 1996, well below the high of 3.7% in 1992. Half of net ODA flows in 2003 were comprised of special-purpose grants, which include technical cooperation, debt forgiveness, emergency and disaster relief, and administrative costs (Table 1). Although special-purpose grants are an essential element of the development process and have budgetary consequences for donor countries, they do not provide additional financial resources to recipient countries to support programs that are needed to achieve the Millennium Development Goals (MDGs). Next, Table 2 shows selected ODA recipient countries in Southeast Asia. As can be seen, the least developed or the most poor countries in the region, such as Cambodia and Laos, have received the highest ODA per capita, US$ 45 and 2

US$ 33 in 2001, respectively. Malaysia received only US$ 1 in 2001, and in absolute terms it also received the least amount. Thailand received US$ 5, Indonesia and the Philippines received US$ 7 each, and Vietnam a higher amount of US$ 18 of ODA per capita. These correlate with their levels of economic development. Cambodia and Laos still depend on foreign aid for 60 to 70% of their investment (gross capital formation). About 18% of the central government s budget of Vietnam was met by foreign aid. ODA to East Asia (and the Pacific) amounted to US$ 4 on a per capita basis. South Asia also received that same amount. This was much lower than ODA per capita to all developing countries. Figure 2: Net Official ODA to Developing Countries, 1990-2003 (US$ billion) 80 70 60 50 40 30 20 10 0 54.3 58.3 62.4 56.1 58.8 58.8 55.6 48.5 52.1 53.2 53.7 52.4 58.3 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 69 Source: World Bank (2005). Figure 3: ODA as a percentage of GDP in recipient countries, 1990 2003 Source: World Bank (2005). 3

Table 1: Net bilateral ODA and special purpose grants, 1990 2003 Source:.World Bank (2005). Table 2: ODA Recipient Indicators Selected East Asian Countries, 2001 Source: World Bank (2004b). INDONESIAN DEPENDENCY ON AID Historically, Indonesia has been one of the highly aid dependent countries. On an average foreign economic assistance contributed more than 40% of national development expenditure, peaking at over 90%. Aid has always been a key factor in taking the country s economy out of any macroeconomic shock. For example, during 1987-88 in the face of continued declines in oil revenues, foreign aid peaked at 6.5% of GDP, accounted for close to 40% of government revenue, and financed about 80% of development expenditure. Also following the 1997-98 crisis, ODA has become the major source of external finance. Most of donor countries have linked their assistance to the IMF program and associated structural reform conditions. As a matter of fact, ODA has been the dominant source of external financing. Only since the early 1990s, private capital flows overtook ODA flows (Chowdhury and Sugema, 2003?). 4

It can be taken as a rough and ready indicator of the important of foreign aid in an economy. As shown in Figure 4, foreign aid flows to Indonesia relate very closely to the periods of economic difficulties. Gross aid flows declined steadily to less than 2% of GDP as the economy began to improve in the early 1970s, but since 1977 it increased continuously until 1988 at roughly 6.5% of GDP. 1 7 Figure 4. Foreign Aid as a percentage of GDP in Indonesia, 1971-2001 6 5 4 3 2 1 0 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 Source: Ministry of Finance, R.I. (financial note; various issues). The importance of ODA vis-á-vis private capital flows shows that until late 1980s, ODA was the main source of external financing. Net official flows averaged 4% of GDP until 1989. Only since early 1990s, private capital flows exceeded official flows, but the situation was reversed after the 1997-98 crisis. Figure 5 presents another indicator of foreign aid dependency of Indonesia. As can be seen, the trend of foreign aid as a percentage of government (domestic) expenditure revenue is almost identical to that of aid-gdp ratio shown in Figure 4. After a steady decline since 1988 from almost 39% of government revenue to nearly 12% in 1997, foreign aid increased to almost 28% in 1998, the year when the crisis reached its climax. Figure 6 presents the most important indicator of significance of foreign aid in an economy, i.e. its share in total domestic expenditure. It is stated as the most important indicator especially in positioning the importance of foreign aid in poverty reduction since government expenditure plays the most crucial role in financing both directly and indirectly poverty alleviation projects and programs. As can be seen, foreign aid financed closely to 70% of total development expenditure in 1971, and then dropped to about 22% in 1974. After that up to 1985, it fluctuated between 20% and 30%. In 1988, the contribution of foreign aid to development expenditure reached about 78%, but it continued to decline in the 1990s until the crisis emerged in 1997/98. As shown in Figure 7, the economic crisis and the consequent increased in government debt have turned Indonesia into one of the most aid-dependent countries in the region. In 1990, official flows from all sources was 2.68% of GDP in 1 This figure does not include the received IMF loan, since this latter is not regarded as development assistance, and is provided as a short-to medium-term support for balance of payments. It is not recorded in the state budget, but goes to the Bank of Indonesia (BI) as a supplementary fund to be used when BI s foreign exchange reserves fall short of meeting the balance of payments needs. 5

Indonesia as compared to, for instance, almost 8% in Sri Lanka as the highest in the group. In 1998, the share in Indonesia reached 3.44%, while that of Sri Lanka was 4.05% (Chowdhury and Sugema, 2003?). Figure 5. Foreign Aid as a percentage of Government Revenue in Indonesia, 1971-2001 45 40 35 30 25 20 15 10 5 0 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 Source: see Figure 4. Figure 6. Foreign Aid as a percentage of Government Expenditure in Indonesia, 1971-2001 100 90 80 70 60 50 40 30 20 10 0 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 Source: see Figure 4. 6

Figure 7. Official flows from all sources as a percentage of GDP in Indonesia and other selected Asian Countries, 1990-200 12 10 8 6 4 2 0-2 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Indonesia 2.7 3.1 2.8 1.9 1.3 0.93-0.08 0.51 3.44 2.96 1.63 Thailand 0.6 0.6 0.12 0.8 0.5 0.5 0.41 4.13 1.23 2.1 0.9 Philippines 3.5 3.13 4.02 3.1 1.24-0.13 0.63 0.67 0.7 0.5 0.23 Viet Nam 1.7 2.4 5.2 1.2 4.5 3.1 2.5 3.13 5.6 4.6 4.63 Bangladesh 6.3 6.3 5.7 4.2 5.22 3.3 3.57 2.6 2.9 2.8 2.53 Pakistan 4.4 4.7 3.8 3.5 3.3 2.4 2.71 1.95 2.1 2.1 1.2 Sri Lanka 7.98 9.7 4.8 5.6 4.6 4.75 4.45 3.8 4.1 2.54 2.4 Source: ADB (database, Key Indicators) TRENDS OF ECONOMIC GROWTH AND POVERTY IN INDONESIA Indonesia experienced many years of deteriorating economic performance during the old order period. However, several years after Soeharto took power in 1966 (the beginning of the NO government), the country s economic condition changed dramatically. From the macroeconomic perspective, the Indonesia s economy had performed very well, especially during the 1980s up to mid 1997, just before the 1997 Asian crisis occurred. The real GDP per capita increased significantly with an average growth rate per year of about 7%. By the last quarter of 1997, Indonesia was subject to an economic shock of sizable magnitude, known latter as the Asian crisis, leading to the fall in the country s GDP by 13.4% (Figure 8). Only a few countries in East and South-East Asia achieved high growth rates above 5% before the crisis, namely China, Republic of Korea, Indonesia, Malaysia, Thailand, and Vietnam. Almost all these so-called miracle economies (except China, Malaysia and Vietnam) suffered a slowdown in growth as a result of the crisis. Korea managed much better than the other crisis affected countries to rebound (Table 3). One of most striking aspects of development in this region over the period around 1970-1990 is that many countries achieved remarkably high rates of economic growth accompanied by steep reduction in poverty (Table 4). Based on national poverty lines, poverty declined over this period in Indonesia from 60% to 16.8%; in China from 33% to 10%; in Malaysia from 18% to 6%; in the Republic of Korea from 23% to 8%, and so on. A more comprehensive figure of trend of poverty in Indonesia is presented in Table 5. With the sustained rapid economic growth in the 1970s up to 1997, the percentage of population deemed as poor has declined dramatically. When 7

the crisis occurred in 1997 and reached its climax in 1998, the poverty rate increased to 16.7% and reached its peak level at 23.5% in 1999. The rebound of the country s economy in 2000 has led to a drop again in poverty incidence. In the precrisis period, the poverty rate in rural areas declined faster than that in urban areas. There were at least three causes: (i) agricultural output growth that led employment in the sector and farm income to increase; (ii) employment increased in rural non-farm activities like agro-industries, trade, services and rural transportations as a result of improved rural infrastructure and rural-urban connections; and (iii) many unskilled labor, unabsorbed by the growth in agriculture and rural non-farm activities, migrated to urban areas and worked in labor intensive manufacturing industries such as food and beverages, textile and garments, leather products, electronics and footwear, construction, transportation and services. These were boomed industries and sectors during the NO era. Figure 8: GDP per capita and GDP growth rate in INDONESIA: 1970-2002 Source: Statistical Year Book of Indonesia (SI), published annually by the Indonesian National Agency for Statistics (BPS) Table 3: Average annual GDP p.c. growth in the 1990s in selected Asian countries (%) Countries 1990-1995 1996-2000 China Republic of Korea Cambodia Indonesia Lao People s Democratic Republic Malaysia Philippines Thailand Viet Nam 9.3 6.7 2.0 6.2 3.5 6.7 0.0 7.8 5.7 7.2 3.8 1.6-0.8 3.6 2.4 1.4-0.4 4.8 Source: World Bank (2002). 8

Table 4: Trends in Poverty in Selected Asian Countries, based on National Poverty Lines, 1970-2000 Country Population below the national poverty line (%) Annual change (%) 1970 1980 1990 2000 1980s 1990s China Republic of Korea Indonesia Malaysia Philippines Thailand Viet Nam Bangladesh India Nepal Pakistan Sri Lanka 33.0 23.0 60.0 18.0 61.6 (1971) 26.0.. 71.0 (1973/74) 55.6.. 54.0 (1961) 37.0 (1963) 28.0 10.0 29.0 9.0 59.7 (1985) 17.0 75.0 (1988) 52.3 (1983/84) 48.4 (1978) 41.4 (1984/85) 29.1 (1986/87) 27.3 (1985/86) 10.1 8.2 (1993) 16.8 6.1 (1989) 45.2 (1991) 18.0 58.0 (1993) 49.7 (1991/92) 40.9 (1992).. 26.1 (1990/91) 22.4 (1990/91) Source: Indonesia: BPS; for other countries in the table: ESCAP & UNDP (2003). 4.6 11.6 (1999) 14.6 8.1 (1999) 40.0 14.2 37.0 (1998) 39.8 26.1 (1999/2000) 44.6 (1995/96) 32.6 22.9 (1995/96) -6.4-1.4-4.8-3.6-4.0 0.6-4.5-0.6-1.1.. -2.6-3.6-5.4 6.9 8.8 3.3-1.3-2.1-7.2-2.3-4.5.. 2.9 0.4 Table 5: Percentage of Population Living Under Current Poverty Lines in Indonesia (Headcount Index): 1976-2004 Year Poor People (%) Urban* Rural** National*** 1976 1978 1980 1981 1984 1987 1990 1993 1996 1998 1999 2000 2001 2002 2003 2004 38.8 30.8 29.0 28.1 23.1 20.1 16.8 13.4 9.7 21.9 19.4 14.6 9.8 14.5 13.57 12.6 40.4 33.4 28.4 26.5 21.2 16.1 14.3 13.8 12.3 25.7 26.0 22.4 24.8 21.1 20.23 19.5 40.1 33.3 28.6 26.9 21.6 17.4 15.1 13.7 11.3 16.7 23.5 19.1 18.4 18.2 17.4 16.6 Notes: * = % of urban population; ** = % of rural population; *** = % of total population Source: BPS (SUSENAS) The crisis caused the poverty rate in rural and urban areas to increase in 1998 and 1999. The increase of rural poverty in those two years was partly a result of returning unemployed people from urban areas. During the crisis, many laid-off workers particularly from manufacturing industries and construction (the two sectors that mostly hit by the crisis) were reportedly leaving urban centres to return to their villages where subsistence could at least meet their basic needs. However, in the crisis situation in which poverty in both urban and rural settings was on the increase, many rural originated people who became unemployed stayed in cities and considered self-employment or do any kind of low-paid works in urban informal sector as an option (Amin [1998], Hugo [1998]). So no doubt that during the crisis agriculture together with urban informal sector had played an important role as the last resort for the laid off workers from the formal sector. ECONOMIC GROWTH AND POVERTY REDUCTION In the general development debate, economic growth is viewed as an important, though not sufficient, means of achieving reduction in poverty. Findings from many individual country case and cross-country studies support this proposition. Dollar and Kraay (2002), for example, used data from 80 countries and find that as overall income increases, 9

on average incomes of the poor increase by exactly the same rate. 2 Data from Indonesia seem to support this notion. As illustrated in Figure 9, the lines of percentage changes in poverty rate (dp) and in real GDP per capita (dy) indicate that during the crisis period the significant decline in per capita real GDP was strongly associated with the substantial increase in poverty rate. In the general development debate, economic development measured by growth in real GDP per capita is viewed as an important, though not sufficient, means of achieving improvements in human well-being or poverty reduction. Dollar and Kraay (2000), for instance, tested the hypothesis growth is good for the poor by investigating the link between the income of the poor (defined as the bottom 20% of the income distribution) and overall income or per capita GDP. The data used consist of income of the poor and mean income for 80 countries over 40 years. The study further examines the poverty growth relationship in cases of poor countries versus rich countries, crisis periods versus normal growth periods, and the recent period compared to earlier times. Their study also introduces other institutions and policies into the analysis and asks whether these influence the extent to which growth benefits the poor. Their basic finding is that as overall income increases, on average incomes of the poor increase by exactly the same rate. As for the impact of policies and institutions, their study found that openness to international trade as well as improvement in rule of law (e.g. property rights) raise incomes of the poor by raising overall per capita GDP. Overall, their findings suggest that growth tends to lift the incomes of the poor proportionately with overall growth (Figure 10). Figure 9: Lines of Annual Percentage Change in Poverty Rate (dp) and Real GDP per capita (dy): 1982-2002 % 60 40 20 0-20 82 84 86 88 90 92 94 96 98 00 02 Source: BPS [SI & SUSENAS] dp dy Similar evidence also provided by Ravallion and Chen (1997). By using data from household surveys for 67 LDCs and transitional economies over 1981-94, they found that in most cases poverty fell with growth in average income and rose with contraction. By regressing the growth of average income for the poorest 20% and the poorest 40% of the population against the growth of GDP per capita, Roemer and Gugerty (1997) found that on average the poor do benefit from economic growth. An increase in the rate of per capita GDP growth translates into a one-for-one increase in average income of the poorest 40%. For the poorest 20%, the elasticity of response is 0.921. Another conclusion from this study is that income distribution changes only very slowly, and that a policy that aims at redistributing income at the expense of economic growth may have very low payoffs in terms of poverty reduction. Also Gallup et al. (1999) have estimated the 2 See also other studies from for example Ravallion and Chen (1997), Roemer and Gugerty (1997), Timmer (1997), and Balisacan and Pernia (2002), Balisacan, et al. (2002). 10

growth elasticity of per capita incomes of the poor to be close to unity, which implies that growth in average income leads to one-for-one increase in incomes of the poor. Figure 10: Economic Growth and Poverty Source: Figure 3 in Dollar and Kraay (2000) By using data on income distribution for 27 LDCs, Timmer (1997) estimates the impact of average per capita income growth on the growth of per capita income of each income quintile. He found that the elasticity of overall growth and the growth in the per capita income of the poorest quintile was only 0.8 (and significantly less than one) and rose steadily to slightly greater than one for the richest quintile. With this result, he argues that the apparent failure of growth to reach the poor in the countries with wide income gaps, while disappointing, should not be taken as a general indictment of economic growth itself. Using provincial data from the Philippines over the 1980s and 1990s, Balisacan and Pernia (2002) find that, on average, the growth elasticity is just above 0.5. This finding indicates that income growth across the country s provinces has not translated into one-for-one to changes in the welfare of the poor. There are still, however, 11

controversial arguments in the literature regarding the role of inequality as a poverty determinant. Empirical studies using different sample and different econometric techniques come with mixed results. For example, a summary of few cross countries studies by Tambunan (2005) shows that some of these studies find a negative impact of inequality or increase in Gini index of income or consumption expenditure (as a general used measurement of inequality) on economic growth; while others have the opposite results. A study by Deolalikar (2002) used provincial data from Thailand suggests that, while economic growth has a strong positive effect on poverty reduction, income inequality has a sharply negative effect. Income inequality reduces the rate of poverty reduction in two ways: (i) increased inequality is association with increased poverty after controlling for economic growth, and (ii) high level of initial inequality reduces future growth rates, thereby impeding the poverty reduction that would have taken place in the presence of rapid economic growth. So, as economic growth and poverty are assumed (or based on empirically evidence) to have a positive correlation, the findings suggest that changes in income inequality, along with economic growth, jointly affect the rate at which poverty is reduced. The impact of economic growth on poverty reduction will be smaller if economic growth is associated with worsening distribution of income. 3 Based on a panel estimator using a larger sample of countries with 10-year periods, Barro (1999) finds that the negative impact of income inequality on GDP growth may depend on a country s wealth level, although the overall effects are weak and the relationship lacks robustness. Deininger and Squire (1996) find that for the 95 growth spells for which data on income shares were available, there was no a systematic link between growth and inequality. However, there was a strong positive relationship between growth and poverty alleviation. In particular, growth benefited the poor in the vast majority (87.5%) of cases, whereas economic decline hurt the poor disproportionately (in five out of seven cases). Use of better data that allow incorporation of panel aspects (using 5-or 10-year averages) suggests, however, that the empirical relationship weakens considerably, and may actually be reversed. On the other hand, Forbes (2000) and Li and Zou (1998), for instance, both find positive effects of income inequality on growth. Following the approach used by Dollar and Kraay [2002] and the others discussed above, the statistical analysis shows that not only the two variables (dp and dy) are negatively correlated (as generally expected) and the regression coefficient is significant from zero at 90% confidence interval, but the growth elasticity is more than unity. The F-statistic is statistically significant with the critical point 0.01, suggesting that dy contributes significantly to the linear prediction of dp based on the observed data. dp = 6.45 2.0dY (2.2)* (-3.997) R² = 0.46 F-statistic = 15.98 * t-values in brackets The regression result suggests that the resulting economic growth is strongly pro-poor, and, as shown in Table 6, Indonesia with Malaysia were two countries in Asia with the greatest total poverty elasticity of growth during the 1990s. This elasticity measures the extent to which economic growth translates into poverty reduction. 4 Another way of 3 Of course, what matters to poverty reduction is not the degree of overall income inequality but the inequality of incomes in the vicinity of the poverty line (Deolalikar, 2002). 4 It is the number of percentage points of change in poverty observed for every 1% change in real per capita income. A negative figure indicates that income grows reduces poverty (or income declines raises poverty). 12

looking at the effectiveness of translating economic growth into poverty reduction is to examine the elasticity with constant income distribution. This so called the distribution-neutral growth elasticity is also presented in Table 6, which shows that the elasticity in Indonesia would have been higher had income inequality in the country been held constant. Table 6: Total Poverty Elasticity of Growth in Selected Asian Countries in the 1990s Country Total poverty elasticity, 1990s Distribution-neutral growth elasticity, early 1990s China Republic of Korea Indonesia Malaysia Philippines Thailand Viet Nam Bangladesh India Pakistan Sri Lanka -0.6 1.5-2.0-2.0-1.7-0.6-1.0-0.8-0.9 3.6.. -2.2-5.4-2.4-1.9-1.8-1.9.. -2.3-1.8-1.6-3.8 Source: ESCAP & UNDP (2003) (for total poverty elasticity in Indonesia for 1970-2004, based on own calculation). The strong pro-poor growth in Indonesia was attributed, among many others, to two main factors. First, the growth strategy adopted during the NO era emphasized rapid increases in the demand for unskilled labour (Manning, 1998). Second, the growth strategy accompanied by poverty alleviation measures (as discussed in the introduction section) that at least had protected the poor from becoming poorer as a consequence of a rapid economic transformation that also took place during that period that normally goes along with a rapid economic growth. These poverty alleviation measures combined with the labor-intensive oriented growth strategy may have also influenced positively income distribution in Indonesia. Official statistics show that income distribution (measured by Gini coefficient) has improved somewhat from 1970s up to 1993 and from 1994 to 1995 has deteriorated a little bit. Even, income distribution has improved to a significant degree in rural areas during the period 1965-2004 (Figure 11). Figure 11: Income Distribution (Gini Ratio) in Rural and Urban Areas, 1965-2002 0.4 0.35 0.3 0.25 0.2 0.15 0.1 0.05 0 1965 1970 1976 1980 1986 1990 1995 1998 2002 2003 2004 Urban Rural Source: BPS (SI). DOES AID CONTRIBUTE TO POVERTY REDUCTION? Aid can affect poverty directly and/or indirectly (Figure 12). In the direct way, aid is distributed to the poor either in the form of cash or by financing poverty alleviation projects and programs. In the indirect way, aid affects poverty 13

through two channels. The first channel is through economic growth by financing investment or infrastructure (including irrigation in agriculture), or providing assistance directly to small and medium enterprises. In the second channel, aid is used to support investment in human capital (education and health care facilities) and other public utilities. This paper focuses only on aid-economic growth-poverty links. Before, it was shown that during the NO era until 1997, Indonesia has experienced a high economic growth and a significant decline in poverty, It was also shown, statistically, that economic growth and poverty in this Indonesian case are negatively and significantly correlated. Now the question is, how effective is aid for economic growth? Figure 12: Channels through which Aid affects Poverty Economic growth Aid Education, Health, Public Utilities Poverty reduction Poverty alleviation projects/programsms A fundamental argument for aid, at least on economic grounds, is that it contributes to economic growth in recipient countries. Two recent papers have breathed new life into the empirical question of aid effectiveness. Burnside and Dollar (1997) find that when other determinants of growth are controlled for, especially an indicator of economic policy, aid has no effect. Aid only makes a positive contribution to growth in those countries with high values for the policy indicator; if policy is poor, aid is ineffective. This result is explained by the tendency of recipients, especially if they have poor policies, to divert aid to government consumption spending rather than using it to finance economic growth-promoting investment (Burnside and Dollar, 2000). Using essentially the same data for the same sample, but with different specifications and estimators, Hansen and Tarp (2001) find that aid does have a positive effect on economic growth and this result is not conditional on policy. Boone (1994, 1996) tests that if some aid is allocated to consumption rather than investment, it cannot have a positive impact on growth. He finds that most aid is allocated to consumption, and concludes that aid is therefore ineffective. This finding was criticized by Lensink and Morrissey (1999) who argue that his conclusion is potentially misleading. First, much aid is in practice intended for consumption; as this can permit higher savings, such aid can contribute to growth in principle. Second, measured consumption includes investment in human capital; this may often be an intended use of aid, and by increasing the productivity of human capital aid can contribute to increasing economic growth. This is of course difficult to validate empirically, partly because it is difficult to identify how much of aid is invested in human capital and partly because one would anticipate long lags between the receipt of aid and any increase in labour productivity. Lensink and Morrissey examined the effectiveness of aid controlling for uncertainty of aid inflows, and found that aid uncertainty 14

is negatively related to growth, and this result is robust. Investment appeared to be the principal determinant of growth and, when included with investment, foreign aid does not have a robust effect on growth. Their results suggest that aid, if one controls for uncertainty, has a robust effect on economic growth via the level of investment. This holds both for the entire group of LDCs. Based on these results, they conclude that predictable aid inflows are important for investment, and thus for economic growth. Gomanee et al. (2002) identify investment as the most significant transmission mechanism in aid-growth links.. With the use of residual generated regressors, they achieve a measure of the total effect of aid on growth, accounting for the effect via investment. Pooled panel results for a sample of 25 sub Saharan African countries over the period 1970-1997 point to a significant positive effect of foreign aid on growth, ceteris paribus. On average, each one percentage point increase in the aid/gnp ratio contributes one-quarter of one percentage point to the growth rate. With these results, they argue that Africa s poor growth record should not therefore be attributed to aid ineffectiveness. Of course, the simple relationship between aid and economic or per capita income growth in recipient countries is weak Within LDCs, some countries get a great deal of aid but grow slowly, while others also get a lot and have high economic growth. At least there are two problems in analyzing the effect of aid on growth in a simple way. First, there many factors (beyond aid) that affect growth, so these factors must be considered. Second, in some cases aid may be deliberately given to countries that are growing poorly. Consider a country hit by a natural disaster that destroys the rice crop and thus reducing growth. This calamity may induce a temporary increase in aid. Or as in the case of Indonesia, when the country hit by the financial crisis in 1997/98 which resulted in a negative economic growth in 1998, aid flow increased into the country. Thus, while the simple relationship might appear to be a negative association between aid and growth, it would be a mistake to interpret this as evidence that aid reduces growth. A recent study addressed this problem and still found no relationship between aid and growth (Boone 1994). Even when a range of institutional, political, and policy variables are added to the equation the result is still the same. Burnside and Dollar (1997) tested the proposition that aid has no effect on economic growth when incentives are weak. They found that for countries with poor management aid has no effect at all, or even economic growth is negative. Taking the good-management group and dividing them into high-aid and low-aid groups produces striking results. The good-management, low-aid group grew at 2.2% per capita, but the good-management, high-aid group grew almost twice as fast-at 3.7% per capita. Including other variables (institutional, political, and policy), their study shows that aid has a large, positive effect on growth in good-management countries. Removing middle-income countries, which receive little aid, from the sample makes the effect of aid even stronger. With good management, an additional 1% of GDP in aid increases growth by 0.5 percentage points. The view that aid has a greater impact on growth if it is targeted to countries with better policies and institutions is also supported by Dollar and Levin (2004) who developed two indexes to measure the selectivity of aid. A policy selectivity index measures the elasticity of aid with respect to the quality of recipients policies and institutions (controlling for recipients per capita income and population). A poverty selectivity index measures the elasticity of aid with respect to the recipients per capita income level (controlling for their policy and institutional quality and population). An overall index is computed as an average of the two, i.e. average of the policy elasticity and the (negative of) the poverty elasticity, to measure how selective donors are in terms of focusing assistance on low-income countries with better policies and institutions. The results show that in 2002, a 100% increase in the quality of recipients policies 15

was on average associated with 176% more aid. The relationship is much stronger for multilateral assistance (elasticity of 2.57) than for bilateral assistance (elasticity of 0.63). However, policy elasticity varies considerably among bilateral donors. Overall, their study suggests that the impact of aid on economic growth depends on the interaction between aid and policy. With bad policies, aid has little effect on growth; if anything, the relationship may be negative. With good policies, the effect of aid is positive and is stronger if aid is given in sufficiently large amounts (Figure 13). The above findings are also supported by a World Bank s study, which shows that policies have a critical influence on the effectiveness of aid, for the same reasons that they effect economic growth (World Bank 1998). The study uses the dataset developed in Burnside and Dollar (1997) for its regression analysis. The dependent variable is growth rate of per capita GNP, averaged over four-year periods, beginning with 1970-73 and ending with 1990-93. There are six four-year periods and 56 LDCs in the sample (though there are some missing observations where data were not available. First, the result shows there is no relationship between aid and growth. However, if aid is interacted with the management index, than it indicates that aid has a positive effect on growth in a good policy environment. Now the question is, did aid contribute to high economic growth during the NO era? It is most likely that the answer is yes. Lensink and Morrisey (1999) said that the impact of aid on economic growth is determined by the stability of aid flows, and not by the level or size of aid per se, and the Indonesian case shows that the ODA flows to Indonesia have been very steady and displayed very little volatility, when episodes of external shocks are taken into account, and based on this, Chowdhury and Sugema (2003) conclude that the stability of ODA flows allowed the government to plan its development expenditure and executes it. Hill (1996) also agreed with that and he has summarized the contribution of aid in Indonesia in the following words: The stability of foreign aid flows, in contrast to volatility of private flows,.has been a recurring of the New Order.. The stability of official flows underlines a crucial contribution of foreign aid In a close relationship with donors, aid flows are more consistent, they provide a basis on which governments may play longer-term investment projects and they enable nations to endure difficult economic periods and to enact policy reforms less painfully than would be the case in the absence (p.79-80). Figure 13. The relationship between the quantity of aid, the quality of policies, and GDP growth Source: Dollar and Levin 2004. THE IMPORTANCE OF AGRICULTURAL GROWTH FOR POVERTY REDUCTION 16

Agriculture remains central to the Indonesian economy for two main reasons. First, from the employment perspective, until early 1990s agriculture absorbed more than 50% of total labor force in the country. Although its employment share started to fall under 50% since 1993, agriculture is still the biggest employment-generating sector in Indonesia. This pattern of change in employment distribution by sector is also observable in other important agricultural based economies in Southeast Asia such as China, Vietnam and Thailand, where other sectors particularly industry manufacturing, construction and services become increasingly important for employment generation (Figure 14). Figure 14. Shares of total employed workers in agriculture in selected Southeast Asian countries (%): 1985-2003 80 70 60 50 40 30 20 10 0 1985 1990 1995 2000 2003 Indonesia China Thailand Vietnam Source: BPS (SI) and ADB database Second, the National Social and Economic Survey (SUSENAS) data on distribution of poor families by main occupations/income source show that the vast majority of poor families in Indonesia are in agricultural work, predominantly on farms (Table 7), and almost 70% of the poor in rural areas work in agriculture (Table 8). Mason and Baptist (1996) have identified the marginal farmers and agricultural laborers with the lowest income among all agricultural household groups as containing the majority of the poor in rural areas. The National Agricultural Census [NAC] in 2003 (recent data) show that there are 25.437 million land-using farmers, 13.663 million or almost 57% of which are marginal farmers with less than 0.5 ha of land under their control. In 1993 the number of land-using farm households was 20.518 millions or grown by 1.8% per year, whereas the number of marginal farmers was 10.804 million or increased by 2.6% per year during the 1993-2003 period. In Java, where the majority of total population as well as poverty are concentrated, the marginal farmers increased by 2.4% per year [Table 9]. Table 7: Distribution of poor families by main occupation/income source, 1996-2002 (number of heads of families in 1000 persons) Unemployed/others Agriculture Industry Services 1996 % 1998 % 1999 % 2000 % 2001 % 2002 % 994 2.88 941 1.90 4.063 8.47 3.560 9.26 2.349 6.33 3.072 8.61 21739 63.01 28063 56.67 25.997 54.19 20.109 51.73 23.375 62.99 20.605 57.75 2208 6.40 3679 7.43 6.069 12.65 5.380 13.84 4.401 11.86 4.471 12.53 9560 27.71 16837 34.00 11.840 24.68 9.784 25.17 6.984 18.82 7.571 21.22 Total 34.500 Source: BPS (SUSENAS) 100.0 49.520 100.0 47.969 100.0 38.833 100.0 37.109 100.0 35.719 100.0 Table 8: Distribution of poor families by sector and area, 2002 (%) Sector Urban Rural 17

Agriculture Forestry Fishery Mining Industry Electricity Construction Trade Transportation Finance Services Others Source: BPS (SUSENAS). 31.11 0.23 1.48 1.25 12.17 0.10 9.67 14.06 8.94 0.69 8.14 0.04 69.09 1.34 2.23 0.49 4.98 0.02 3.63 5.00 2.73 0.08 2.40 0.06 Table 9: Number and growth rate of land-using farms and marginal farmers in Indonesia: 1993-2003 Category 1993 2003 Java Outside Java Indonesia Java Outside Java Indonesia Number (million) - Marginal farmers - Land-using farms - Agricultural households 8.067 (69.8)* 11.564 11.671 2.737 (30.6) 8.954 9.116 10.894 (52.7) 20.518 20.787 9.989 (74.9) 13.336 13.964 3.674 (33.9) 10.841 11.472 13.663 (56.5) 24.176 25.437 Growth rate per year, 1993-2003 (%) - Marginal farmers - Land-using farms - Agricultural households Note: * = % of land-using farms Source: BPS (NAC) 2.4 1.5 2.0 3.4 2.1 2.6 2.6 1.8 2.2 One interesting fact from Table 8 is that even for urban poor, agriculture is very important as their main source of income. Semi-subsistence urban farming is widely believed to make an important contribution to the livelihoods of the urban poor in many developing countries. 5 Urban agriculture can encompass aquaculture in tanks, ponds, rivers and coastal bays; livestock (particularly micro-livestock) raised in backyards, along roadsides, in poultry sheds and piggeries; orchards, street trees, and backyard trees; and vegetable and other crop production on roof tops, in backyards, in vacant tracts of land on industrial estates, along canals, on the grounds of institutions, on roadsides and in many peri-urban and urban farms (Gordon, et al., 2000) There are two main channels through which the performance of agriculture affects poverty, namely output (or productivity) growth and wage increases in the sector. With respect to the first channel, there have been several major sources of economic growth in Indonesia since the end of 1960s, including rapid output growth in agriculture. The dominant contribution of agricultural growth, however, ended by the late 1980s and manufacturing industry took off rapidly (Timmer, 2004). This was also the period when workers from agriculture (rural areas) began to move to the manufacturing sector (urban areas). As illustrated in Figure 15, in 1986 the GDP share of agriculture started to decline and it continued until 1997; while that of industry continued to increase. In 1998, there was some improvement in agriculture s GDP share, mainly because output in the sector declined at a much smaller degree as compared to other sectors also hit severely by the crisis. On the other hand, the industry s GDP share dropped very slightly in 1997 and went up a little bit in 1998, despite the fact that its output fell by more than 10 per cent during the crisis period. This happened because its output contribution to GDP is already much larger than that of agriculture and some other sectors. 5 See, e.g. UNDP (1996), Sanyal (1985), and Freeman (1991). 18

28 Figure 15: Lines of development of GDP share of agriculture and industry (%) 24 20 16 12 8 82 84 86 88 90 92 94 96 98 00 02 Agriculture's GDP share Industry's GDP share Source: BPS (SI). SHOULD AGRICULTURE GET MORE AID? Based on OECD report in 1998, most aid goes to projects-and most project aid goes to social, economic, and administrative infrastructure; while agriculture received only about 12% (Figure 16). Further, as can be seen from Table 10, sectoral distribution of aid in Indonesia changed over time. In fiscal year (FY) 1990/1993, the five largest aid recipients were mining, transportation, agriculture, education and national defense sectors. In FY 1996/97, just before the crisis the main recipients were mining, transport, irrigation, telecommunications and education sectors. Probably because public investment in mining, transport and telecommunications often requires imported capital goods, thus it is not surprising that much of project aid was allocated to those sectors. After the crisis, most aid is allocated to education, transportation, regional development, irrigation and social welfare. For Indonesia as a large agrarian economy with the fact that the majority of poverty is in this sector, it is thus important that agriculture together with infrastructure, health and education should get the highest priority in the allocation of aid. Researchers have also tried to address sectoral fungibility, that is, does higher foreign assistance for a particular sector (for example, agriculture) increase spending for that sector? There are two approaches to answer that question: by comparing spending over time within a country or comparing spending across countries. Whatever the method, the answer is the same: it varies. In some countries and sectors aid appears to be completely fungible across sectors, while in other countries and sectors the money seems to "stick." Pack and Pack (1990) show that for every dollar increase in donor financing for Indonesian agriculture, expenditures in that sector increase by 92 cents. In contrast, a cross-sectional study of 14 countries finds that agricultural expenditures decrease by five cents for every $1 in aid given to agriculture (Figure 17). Many argue that government commitment to particular sectors is more important than targeting aid (e.g. World Bank, 1998). 19

Table 10. Sectoral Distribution of Project Aid (%) in Indonesia, 1990/91-2002 Ssector 1990/91 1995/96 1996/97 1998/99 1999/2000 2000 2001 2002 Industry Agriculture Irrigation Manpower Trade, finance & cooperative Transportation, meteorology, geophysics Mining & energy Telecommunication, post & tourism Regional development & transmigration Environment & spatial planning Education, culture, youth, & sport Population & family welfare Social welfare, health & women Housing & settlement Religion Science & technology Law State apparatus Politics, international relation, information National defense 4.34 15.62 5.27 0.47 3.82 14.08 18.83 2.30 3.20 1.45 11.80 0.53 0.20 6.09 0.01 2.89 0.00 0.03 0.36 8.71 2.75 3.72 9.93 0.32 3.23 17.65 25.71 7.89 4.45 2.04 6.39 0.41 1.63 5.00 0.37 1.79 0.07 1.03 0.57 5.04 2.41 3.79 8.68 0.22 1.77 20.48 25.47 7.64 2.31 2.36 7.36 0.40 2.21 5.87 0.61 1.60 0.11 1.25 0.35 5.10 2.03 4.33 8.78 0.62 2.63 19.59 20.82 4.75 4.74 1.72 9.03 1.19 4.06 6.18 0.70 2.13 0.04 1.45 1.01 4.21 1.30 4.41 6.87 0.26 0.98 19.32 19.45 2.78 11.80 1.18 11.88 1.17 6.26 5.02 1.05 1.11. 0.03 1.58 0.15 3.41 0.37 10.34 8.57 0.00 1.01 15.75 8.36 4.17 6.94 2.77 17.32 1.29 9.42 3.10 0.03 1.99 0.00 1.26 0.00 7.31 6.26 7.50 8.45 0.38 0.51 14.15 8.16 4.61 11.96 2.35 18.28 0.16 7.77 0.95 0.02 0.92 0.00 1.66 0.15 5.75 6.46 6.11 8.31 0.09 0.01 19.95 11.53 6.07 11.35 1.57 12.35 0.26 4.42 1.14 0.00 0.69 0.16 1.63 0.21 7.68 Total 100.00 100.0 Source: Ministry of Finance, R.I. (Financial Notes; various issues). 100.00 100.00 100.00 100.00 100.00 100.00 20

Figure 16 Distribution of Aid, by (A)Type 3 6 26 65 Food aid Program assistance Other Project (B) by Sector 7 12 Industri & other production Agriculture 40 Economic infrastructure Social & administration infrastructure 41 Source: OECD (1998). Figure 17. A Dollar's Worth of Aid to Agriculture and Spending on Agriculture Source: from Figure 3.7 in Pack and Pack (1990) 21

IS FOREIGN AID EFFECTIVE IN REDUCING POVERTY IN INDONESIA? This question is not easy to answer, due to the fact, as already explained before, that there are many channels through which aid affect poverty, directly as well as indirectly. However, two things that all may agree is that high economic growth accompanied by many poverty alleviation programs played an important role in poverty reduction during the NO era on the one hand, and, on the other hand, although there has been no systematic study of aid effectiveness in Indonesia, it is widely believed that foreign assistance played a significant role in Indonesia s success in terms of sustaining rapid economic growth during that period. Starting as an IDA (International Development Association) 6 recipient in the late 1960s, Indonesia graduated to an IBRD client and bulk of its foreign financing since the early 1990s until the 1997/98 crisis was from commercial sources. So, indirectly, it can be argued that during the NO regime, foreign aid has contributed to poverty reduction in Indonesia. Now, after the crisis, with donor help, the government established a special agency for poverty reduction in 2001 attached to the vice-president s Office. A reorganized poverty reduction committee, chaired by the Coordinating Minister for Social Welfare (MENKOKESRA) is responsible for designing and implementing poverty reduction program, However, the cost-effectiveness of foreign aid in poverty reduction still needs to be measured, especially regarding the following question: through what ways aid is the most effective in reducing poverty: either indirectly (e.g. through financing development of infrastructure in rural areas or human resource development) or directly to the poor (e.g. financing specific poverty alleviation projects or programs)? References Amin, A.T.M. Nurul, 1998, The Urban Informal Sector in Indonesia: through the economic recession and recovery, ARTEP Working Papers, No.2, New Delhi: ILO. Arnade, Carlos, 1998, Using a Programming Approach to Measure International Agricultural Efficiency and Productivity, Journal of Agricultural Economics, Vol.49: 67-84 Balisacan, Arsenio M. and Ernesto M. Pernia, 2002, Probing Beneath Cross-National Averages: Poverty, Inequality, and Growth in the Philippines, ERD Working Paper Series No.7, Manila: Asian Development Bank. Balisacan, Arsenio M., Ernesto M. Pernia and Abuzar Asra, 2002, Revisiting Growth and Poverty Reduction in Indonesia: What do Subnational Data Show?, ERD Working Paper No.25, October, Manila: Asian Development Bank. Boone, Peter (1994) "The Impact of Foreign Aid on Savings and Growth," London School of Economics. Boone, Peter, (1996), Politics and the effectiveness of foreign aid, European Economic Review, 40. Burnside, Craig, and David Dollar (1997), "Aid, Policies and Growth." Policy Research Working Paper 1777, World Bank, Development Research Group, Washington, D.C. 6 It is the World Bank facility for highly concessional loans for low income/poor countries. 22