THE LONG-RUN EVOLUTION OF INEQUALITY IN INDONESIA, Arief A. Yusuf, Andy Sumner and Irlan A. Rum 1 ABSTRACT

Transcription

1 THE LONG-RUN EVOLUTION OF INEQUALITY IN INDONESIA, 1990 Arief A. Yusuf, Andy Sumner and Irlan A. Rum 1 ABSTRACT This paper makes new, consistent, long-run estimates of the Gini coefficient and a range of other measures of inequality using national household survey, the Susenas for the period The estimates suggest that inequality in Indonesia has risen quite significantly over the two decades. However, the rise in inequality is much more visible in the data for the 2000s, the era of political reform and democratisation. The magnitude of the rise in inequality is quite spectacular: between 2002, the Gini coefficient and the decile dispersion ratio have increased by roughly 20% and 50% respectively and rising inequality is evident across regions, provinces as well as across urban-rural dimensions. The data also point towards an inequality convergence across provinces because the change in inequality has been faster in provinces with a lower initial level of inequality. Possible immediate and underlying causes of changes in inequality in Indonesia are explored. Keywords: Inequality; Indonesia; Gini; Theil JEL code: D63 1. INTRODUCTION The issue of inequality in Indonesia has become more visible in public policy debate recently because the Gini coefficient has risen to its highest point historically reported by the Central Bureau of Statistics (Badan Pusat Statistik, BPS). 2 In light of this, this paper seeks to analyse the long-run evolution of a range of indicators of expenditure inequality between 1990 and. More specifically, to ask whether inequality in Indonesia has been increasing over the long-run period across the range of measures of inequality indicators, and if so, how rapidly, and the extent to which regional, provincial and urban-rural differences shape the overall pattern and the extent to which these components have driven the overall trend. An important further question is related to whether trends in inequality prior to, and since, the Asian Financial Crisis (AFC) period, in the democratisation period, differ and what might explain such differences. In short, what have been the drivers of changes in inequality over the last decade during which Indonesia has experienced a marked political transformation from the Soeharto era, to a more progressive democratisation since? The paper is structured as follows: Section 2 reviews debates on why inequality matters instrumentally. Section 3 discusses existing literature on inequality in Indonesia. Section 4 presents new long-run estimates of various inequality indicators. Section 5 discusses four hypotheses on the potential drivers of changes in inequality in Indonesia over the last decade in particular. Section 6 concludes. 2. DOES RISING INEQUALITY MATTER IF POVERTY IS FALLING? Indonesia, like a number of developing countries, has made good progress in reducing extreme poverty, whether one takes the $1.25 international poverty line or the national 1 Respectively, Center for Economics and Development Studies (CEDS), Universitas Padjadjaran, King s College London and CEDS, Universitas Padjadjaran. Correspondence to: arief.yusuf@fe.unpad.ac.id and andrew.sumner@kcl.ac.uk. We thanks USAID-SEADI project for financial support. Usual disclaimer applies. 2 The most recent official data on the Gini coefficient are available from BPS at 1

2 poverty line (Sumner & Edward, 2013; Suryahadi et al., ; World Bank, ). Indonesia has had stable and moderate levels of expenditure inequality by international standards prior to the Asian Financial Crisis (AFC) in 1997/8. In spite of rapid economic growth which is often associated with rising inequality, Indonesia s Gini coefficient remained around 0.30 up to the early 2000s without any tendency to drastically rise or fall over the period, and as prevailing inequality determines the nature of poverty reduction at any given level of growth (Ravallion, 2006), this stable inequality pattern likely played an important role in reducing the number of poor people, by the national poverty line from 54.2 million people in 1976 (40.1% of total population) to 22.5 million people (11.3% of total population) in (Alisjahbana et al., 2003). Since the AFC, expenditure inequality has risen in Indonesia, as it has done in many other countries in Asia and the poverty reduction rate has slowed. 3 The responsiveness of poverty to economic growth can vary enormously depending on the indicator of poverty taken, the length of time period used in analysis and country specific factors. 4 However, recent projections of the long-run evolution of poverty in Indonesia suggest a potential opportunity cost to the country of rising inequality which is this: current trends on inequality could delay the end of extreme ($1.25) poverty in Indonesia by a decade if such trends were to continue (Sumner & Edward, 2013). That said, addressing rising income inequality in Indonesia is complex, especially given the multiple dimensions of spatial inequality by regions, provinces and urban-rural dimensions. Indonesia faces large disparities in regional development which are particularly evident between Java and non-java and eastern Indonesia. Of course there are normative or instrumental reasons why inequality may be said to matter too (e.g. fairness, meritocracy, and so forth). However, much global literature has taken an instrumentalist approach as to why high or rising inequality can hinder development. For example, Birdsall (2006) argues that income inequality in developing countries matters for at least three instrumental reasons: where markets are underdeveloped, inequality inhibits growth through economic mechanisms; where institutions of government are weak, inequality exacerbates the problem of creating and maintaining accountable government, increasing the probability of economic and social policies that inhibit growth and poverty reduction; and where social institutions are fragile, inequality further discourages the civic and social life that underpins the effective collective decision-making that is necessary to the functioning of healthy societies. There is plenty of empirical evidence to suggest that high or rising inequality has a negative effect on the rate of growth or the length of growth spells (see, for example, Berg & Ostry, 2011; Easterly, 2002). There is some evidence that this may depend on the level of economic development (GDP per capita) and openness (Agénor, 2002; Barro, 2000; Milanovic, 2002). Cornia et al., (2004) is an interesting study of growth; the authors use a dataset of 73 countries to identify critical threshold levels of inequality. They conclude that rising inequality can assist growth but only up to a Gini value of 0.30 and above ADB () noted that in eleven Asian economies, income disparities have widened in the last two decades. For example, from the early 1990s to the late 2000s, the Gini coefficient increased from 0.32 to 0.43 in China, from 0.33 to 0.37 in India, and from 0.29 to 0.41 in Indonesia. 4 Kraay (2004) provided a decomposition analysis of the change in poverty between household surveys due to growth in average income, the poverty elasticity and changes in relative incomes (changes in income distribution) and found, using, data from the 1990s for 41 countries ( long-run growth spells ) that 80% of poverty reduction, FGT-0 (the poverty headcount) was due to changes in average income but that the FGT-1 (the poverty gap) and FGT-2 (headcount x poverty gap) were respectively, 70% and 60% due to changes in average income, suggesting that substantial contributions of changes in distribution do matter and further, the findings based on the 185 country spells, many of which are shorter than 10 years, gave a contribution of growth to poverty reduction of between 43% to 70%, again suggesting that changes in distribution do matter. 2

3 inequality impedes GDP growth. 5 In short, Indonesia currently stands immediately before the point at which inequality may hinder economic growth. There is a wealth of empirical evidence linking poverty and inequality: specifically, that the extent of poverty reduction depends on prevailing inequality levels and trends and that a higher level of inequality leads to less poverty reduction at any given level of growth (Chen & Ravallion, 1997; Deininger & Squire, 1998; Fosu, 2009; Hanmer & Naschold, 2001; Kalwij & Verschoor, 2007; Ravallion, 1995, 1997, 2001, 2004, 2005; Son & Kakwani, 2003; Stewart, 2000). For example, Ravallion (2005) argues that inequality is bad for the poor, and shows that elasticities of poverty reduction are related to initial country levels of inequality and the changes in income distribution; arguing that this is why inequality matters for poverty reduction. Kanbur affirms that growth is a plus for poverty reduction, inequality is a minus (Kanbur, 2005, p. 224). In light of the above, this paper makes new, consistent, long-run estimates of inequality in Indonesia by a range of measures using the national household survey, the Susenas for the period 1990 to assess the extent to which the long-run trends in inequality by a range of measures of inequality are consistent with the Gini trends. As is well known, expenditure survey data (as is Indonesia s data from the Susenas) understates income inequality since it ignores savings (for example) while top incomes largely elude surveys (see for details, Leigh & van der Eng, 2010). Those are inherent shortcomings of all such surveys, not just Indonesia s and Cornwell and Anas (2013) note ongoing attempts at BPS to address the issue STUDIES OF INEQUALITY IN INDONESIA 3.1. Existing studies Studies that have focused on long-run inequality typically use the Susenas and compute the Gini coefficient or the Theil index. The findings are as follows: before the 1997/8 crisis, there is general agreement that inequality was relatively low or even declining (Akita et al., 2011), and inequality did not increase drastically as a result of economic growth (van der Eng, 2009), and/or was offset by the gains from growth (Cameron, 2000). However, there are some detractors from this, arguing that inequality was high or increasing pre-crisis (Frankema & Marks, 2009; Leigh & van der Eng, 2010; van Leeuwen & Foldvari, ) and overall, growth has been largely distributionally neutral, although areas such as Java have grown slightly faster than the national average (Hill, 2008; Hill et al., 2008). Edward and Sumner (2013) find that total inequality, measured by the Gini and Theil, fell in the early 1990s although this masked rising urban inequality as it was rural inequality that fell substantially. However, post the 1997/8 crisis, inequality has risen in two waves in particular: one wave visible in the data for and a more recent wave This is consistent with others who have noted inequality has increased post-crisis; mainly intra-group and urbanrural inequality (Akita, 2002; Akita & Miyata, 2008; Skoufias, 2001; Sumarto, 2013; Suryadarma et al., 2005, 2006). A tendency of increasing regional disparity was noted by Sakamoto (2007), who suggests that for 28 years to 2005, there is evidence of increasing regional disparity. 5 Cornia et al., find a distinct non-linear relationship between initial income inequality and economic growth. They argue that too low a level of inequality is bad for growth (free-riding, high supervision costs), but also that too high levels of inequality can have serious negative consequences. 6 One recent study by Nugraha and Lewis (2013) argues, using Susenas data that the different forms of nonmarket income could be taken into account. 3

4 3.2. Our methodology We make new estimates of the long-run evolution of expenditure inequality, 1990 using the National Socioeconomic Survey, the Susenas. The Susenas is a series of large-scale multi-purpose socioeconomic surveys initiated in and fielded every year or two since then. Since 1993, the Susenas surveys cover a nationally representative sample typically composed of around 200,000 households. Each survey contains a core questionnaire which consists of a household roster listing gender, age, marital status, and educational attainment of household members, supplemented by modules covering about 60,000 households that are rotated over time to collect additional information such as health care and nutrition, household income and expenditure, and labour force experience (Surbakti, 1995). In terms of long-run patterns of inequality, we make estimates of the following indicators of inequality: the Gini coefficient (of expenditure per person); the decile dispersion ratio of the average consumption of the richest 10% of the population to the average consumption of the poorest 10% and the inter-regional disparity by the Theil index of interprovincial inequality and inter-district inequality. We also decompose inequality by different subgroups of population and by region using the decomposition technique of Haughton and Khandker (2009). This makes it possible to make estimates of the contribution of urban-rural inequality to total inequality and also the contribution of inter-provincial disparity to total inequality. 4. NEW ESTIMATES OF THE LONG-RUN EVOLUTION OF INEQUALITY IN INDONESIA, The Gini coefficient Our estimates suggest the Gini coefficient has risen from 0.33 to 0.41 between 1990 and. As noted, this is the highest ever recorded Gini in Indonesian history. Our estimates are consistent with BPS (see Figure 1) since 2009 but prior to 2009 there are some differences. The reason for this difference is due to the use of a different methodology to calculate the Gini coefficient. According to BPS 7, the Gini coefficient before 2009 was calculated using grouped data, while from 2009 onwards the Gini was calculated using individual data. The use of grouped data (expenditure group) is not as accurate as using individual data from household surveys because it may degrade the underlying data somewhat. We find the increase in the long-run trend is generally similar across urban and rural areas as well as across Java and non-java islands (see Figures 1 to 2). Moreover, in urban areas the Gini was already 0.43 in. In urban areas of Java Island the Gini coefficient reached as high as 0.44 in 2011 albeit slightly falling the following year to In the disaggregated estimates, especially so by regions, the level of current inequality, as measured by the Gini coefficient, seems to be worse than is currently thought in the literature (see earlier discussion). Our estimates suggest that the Gini coefficient was stable, if not slightly decreasing, before the Asian Financial Crisis (AFC). However, after the AFC period the Gini coefficient had a strong upward trend. The Gini coefficient in rural areas is stable and lower than in urban areas. The upward trend in the Gini coefficient is consistently evident in both urban and rural areas, as well as across regions in Indonesia, such as in Java and in the non-java islands. 7 See 4

5 In terms of the proportional increase, the Gini coefficient after the AFC (2001) was 0.34 but rose to 0.41 in or an increase of 0.1 point or 32%. Surprisingly perhaps, the rate of change was faster in rural areas, where in 2001 the Gini coefficient was 0.24, and in it rose to 0.33, a startling increase of almost 40%. This trend is slightly stronger in rural areas of non-java compared to Java. The data for the most recent period, 2006 (see Figures 3) suggest that rising inequality is common across all provinces of Indonesia. The annual percentage change in the Gini coefficient across the last decade (2002 ) ranges from 0.7% to 9.1%. Interestingly, the provincial level analysis captures a pattern of convergence across Indonesian provinces in urban areas (see Figure 3, left panel). Inequality 2002, by the Gini coefficient tends to fall in provinces with initially high inequality and vice versa. The data at district level suggests (Figure 3, right panel) also suggests convergence Income share by group Figures 5 and 6 show the share of national income (proxied by expenditure) of three income groups: the 20% richest households, the 40% middle-income households, and the 40% poorest households. Similar to the Gini coefficient, the data suggest that the income share of the 20% richest households is rather stable from 1990 until the AFC period but increases immediately after the AFC (2001) and continues to do so to. While in 1990 the income share of the 20% richest households was 42.1%, in it was 49.5%. The increase in the income share of the 20% richest households during the last decade has been accompanied by the decline in the income share of the 40% poorest households and the 40% middle-income households. However, the decline of the 40% poorest households is noticeably larger. For example, from 2001 to, the 20% highest income group has gained a 5.4% additional share at the cost of the 40% poorest income group and the 40% middle-income group losing - 4.8% and -2.8% respectively. The widening gap as measured by the income share of the three income groups are common across urban and rural areas, as well as across Java and non-java islands. One particular observation in rural areas deserves more attention. In 1990, the share of the 40% middle-income group was 39.3% while that of the 20% richest group 37.1%, making the former slightly higher than the latter. This remained the same for the following 15 years. However from 2007 onwards, the share of the 20% richest household was consistently above the 40% middle-income household. Figure 5 and 6 also show that from the year 1990 to 2003 the income share of the 40% poorest households increased albeit slowly. From 2004 onwards their income share started declining and that decline grew faster during the most recent years. During this specific period, the income share of the 40% poorest declined by 4%. This means on average that the 40% poorest households in rural areas saw their income share reduced by 0.5% every year The decile dispersion ratio The decile dispersion ratio is the ratio of the mean expenditure of the 10% richest households to that of the 10% poorest households. Figure 7 shows the evolution of the decile dispersion ratio from 1990 to for all Indonesia, urban areas, rural areas, Java Island, and non-java regions. Over the period from 1990 to, the decile dispersion ratio tended to decline 8 The convergence at both district and province level is confirmed by the regression of the change in Gini coefficient (2002-) with its initial level in The coefficient of the initial value is negative and statistically significant at 1% level. 5

6 moderately prior to the AFC period, falling even further during the AFC period, but increasing rapidly after the AFC period up to. The decile dispersion ratio trend for the last decade suggests a picture of rising inequality that is more significant than that made visible by the Gini coefficient. For example, from 2001 to, the decile ratio for all Indonesia rose by 65% or 0.44 points every year. This is quadruple the 0.13 point a year rise between 1990 and The rising decile dispersion ratio is more prominent for urban areas and in Java. In urban areas the gap between the 10% richest and the 10% poorest from 2001 to widened by around 70%. The gap between the top and bottom income group also grew tremendously in rural areas, albeit slower than in urban areas Inter-regional inequality A final dimension of income disparity in Indonesia that we consider is inter-regional inequality. The Theil index of inter-regional inequality is one of the most common measures of inter-regional inequality. The Theil index was calculated for both inter-provincial inequality and inter-district inequality. The result is shown in Figure 8. The estimates suggest that for each year, inter-regional inequality is always higher for inter-district inequality than inter-province inequality. The trend over the course of 1990 to, however, is rather similar showing general long-run stability with some fluctuations during the middle of the 1990s towards the AFC period. There seems to be a slight declining trend from 1990 to the start of the AFC, particularly for the inter-district inequality. After the AFC period, interprovince inequality seems to decline slightly but this trend levelled off in the last five years. Inter-district equality stabilises earlier (starting in 2008 towards ). The overall observation coming from the Theil index is that there is no tendency to increasing interregional disparity in Indonesia over the last 20 years Contribution of between- and within-group to inequality The discussion so far, on various indicators of inequality, suggests that inequality in Indonesia has risen quite rapidly for the last ten years after a rather stable period prior to the AFC. The Gini coefficient, the income share of various groups and decile dispersion ratio measure inequality across individual households. The later discussion also analysed the interregional dimension of the inequality. There are also some measures of inequality, such as the Generalised Entropy Index that can be decomposed into inequality of individual households within one particular group and inequality within those particular groups. For example, looking at Figures 9 we can say that inequality in Indonesia in is driven primarily (94.7%) by within-province inequality. Inequality between provinces only contributed to 5.3% of the overall inequality. Similar patterns are evident with urban-rural groups. Only a slight proportion (4.3%) is contributed by inequality between urban and rural areas. The largest contribution is inequality within urban and rural areas. 9 We also calculated the mean expenditure per capita ratio between Java and non-java, and between the noneastern region and eastern regions of Indonesia. The ratio for non-eastern to eastern regions is always higher than the ratio of Java to non-java. This is because the eastern region in Indonesia is still lagging behind in economic development. Over the period, however, it can be observed that the gap is more or less stable over the first part of the period but declining during the most recent period. This is more or less consistent with the evolution of the Theil index discussed earlier, but there is a rather stronger decline during the end of the period. In sum, it is clear that the gap between more and less developed regions in Indonesia has declined gradually over the last ten years. 6

7 Looking at the trends in the share of between provinces and between urban-rural inequality, the data suggest that the contribution of inequality between provinces and between urban and rural areas has been consistently declining since as early as Conversely, the contribution of inequality of individual households within provinces and within urban and rural areas is consistently increasing (Figure 9). The explanation for this trend, drawing on the earlier discussion, is simply that rising inequality in Indonesia for the last decade has been common or uniform across geographical locations, whereas the gap between regions is rather stable or slightly declining in more recent years. 5. WHAT HAVE BEEN THE MAIN DRIVERS OF CHANGES IN INEQUALITY IN INDONESIA? 5.1. Immediate causes: The pattern of economic growth in Indonesia If considers all the countries by Gini (figure 10) one finds that over the period of 2001 to 2011, Indonesia moved from low income, low inequality (lower-left quadrant) to low income, high inequality (upper left quadrant). Further, when taking into account that inequality is measured with expenditure and not income in Indonesia, one could say Indonesia may even be higher up in that same quadrant (in figure 10 identified as Indonesia 2011Y). 10 The immediate cause of increasing inequality in Indonesian is the pattern of economic growth. Table 1 shows the growth rate in the real per capita expenditure during 4 distinct periods: the long-term period or the last 22 years of 1990-, the pre-reformasi era (1990- ), the reformasi era of 2000-, and the reformasi era of The last period of may better represent the reformasi era as it was already four years after the Asian Financial Crisis of 1997/8. The growth rate in the annual real per expenditure per capita is calculated for various groups (to avoid determinism by using one poverty line), namely the mean population, the median population, the poor defined by national poverty line, the poor defined by $2 a day poverty line, and another five groups under the 30th percentile. Looking at the long-run growth (1990-) the table shows that while the growth rate in expenditure per person of the mean of population is 4.2% per year, the growth rate of the $2 a day poor and the national poverty line poor grew by 3.5% and 3.2% respectively. In other words, the expenditure per capita of the average population was 33% faster than that of the poor (by national poverty line) whereas it was 19% faster than that of the poor by $2 a day poverty line. Comparing the period of pre-reformasi and reformasi era, it is clear that the propoorness of growth has been considerably worsening. For example, while in the prereformasi era of the growth of the average population was 37% faster than the growth of the poor (national poverty line), during the reformasi era of the expenditure per capita of the average population was 75% higher than the growth rate of the poor. This is also the case in comparison to the poor defined as those living below $2 a day. While during the annual growth rate of the expenditure per person (in real term) has grown by 4.5% a year, the growth rate of the people living below $2 a day grow only by 3.6% a year. This implies that the average population grow 37% faster. However, in the reformasi era, the growth of the average population is faster by 45%. The difference between the two periods is the speed at which the expenditure per person grew. 10 This is based on Indonesia s Gini coefficient in 2011 adjusted for the difference between expenditure-gini and Income-Gini estimated from years when both Gini coefficient was measured). 7

8 Figure 11 illustrate these data in the growth incidence curves. In short, long-run Indonesian economic growth (over the last 20 years, period before the reformasi, or period after the reformasi) has never really been income-equalizing growth. However, during the reformasi era, the growth was a lot more income-inequalizing in tendency and a lot less propoor relative to growth during the period before the reformasi era. 5.2 Underlying drivers of distributional change in Indonesia The causes of recent changes in inequality are complex but we can identify four specific factors that have contributed, some of which would suggest public policy could better manage drivers of rising inequality in Indonesia in future years. 11 One driver of changes in inequality is likely to be that Indonesia is experiencing a commodity boom in coal and palm oil in particular (Burke & Resosudarmo, ) and this is impacting on inequality. Yusuf (2013, forthcoming) uses a CGE model to show that the changes in inequality are due to world prices of mining commodities rather than estate crops; relatedly, Suryahadi et al., () note changing sectoral contributions to growth are associated with slowing poverty reduction, and thus potentially with changes in inequality. The commodity boom hypothesis can be advanced to explain the widening gap between poor and rich groups in rural areas. During more or less the same period, there has been a trend of the increase in the price of commodities, particularly those that are traditionally Indonesian export commodities such as estate crops. Those estate crops are mainly located in rural areas and are owned by landowners in rural areas. The rural richer households will benefit from this commodity boom disproportionately. One way to see the signs of this mechanism at work is to look at the evolution of inequality in rural areas distinguished by Java and the non-java islands because estate crop plantations are mostly located outside Java. Figures 10 and 11 show exactly this. However, the rising share of the 10% richest and the declining share of the 40% poorest are quite similar in both Java and the non-java islands, hence such a hypothesis is not much supported. A second driver of changes in inequality in Indonesia relates to changes in the labour market. Yusuf et al., (2013) propose a hypothesis on labour market rigidity in the formal labour market including inter-related changes in labour market regulation an increase in severance payment, the strengthening of labour unions, rising minimum wages, reduced demand for unskilled labour, and an increase in informality in lower wage employment. Before the East Asian Financial Crisis, the manufacturing sector was the primary source of Indonesian economic growth. The growth of the manufacturing sector s GDP was 11.2% during 1990 (while the average economic growth was 7.9%) and its employment growth was 6% (while the average national employment growth was only 2.3%). Almost a decade after the crisis, the role of the manufacturing sector in generating employment seems to have halted. While its growth for the period of was almost the same as the national average (4.7%), its employment growth was only 0.9% (Yusuf et al., 2013). The employment opportunities in the formal manufacturing sector, historically, have been a haven for people in rural areas to find better paying jobs. When such opportunities are limited, there will be more excess supply of unskilled labour in rural areas. As the labour market in rural areas is flexible, overall rural real wages will be pushed down with the consequence of increasing inequality in rural areas. The official data from BPS suggests that real wages for 11 It has also possible that better data are driving the recent findings. Although high-income groups are not well captured in the data (see Leigh & van der Eng, 2010, p. 173), the BPS has recently sought to make efforts to address this (Cornwell & Anas, 2013, p. 27), potentially meaning that recorded inequality may have risen as a result. 8

9 agricultural workers have been declining for the past few years suggesting a possibility of constraints for people to migrate to cities and find formal employment (Manning and Pratomo, 2013). A third driver of changes in inequality is rice prices. One particular inequality spike is that which occurred in There are various possible reasons why inequality rose sharply from 2003 to The domestic price of rice increased by almost 20% during that period after being very stable for a long period. This may have reduced the real expenditure of the poor. A fourth driver of changes in inequality is social expenditures and notably the cash transfers of During the inequality spike of world oil prices rose by 70% during the same period, leading to an increase of fuel subsidies from 6.5% of the total government budget (APBN) in 2003 to almost 16% in This meant the fiscal space for additional government social spending was curtailed and the benefit of the fuel subsidy disproportionately benefited the non-poor. In October 2005, the government made a large adjustment to fuel prices, increasing retail fuel prices for gasoline, kerosene and diesel. The price of gasoline increased by 87.5%, diesel by 104.7%, and surprisingly kerosene by 185.7% (Yusuf & Resosudarmo, 2008). The Indonesian government also distributed cash transfers of Rp 18 trillion to the poor and the near poor to mitigate the inflationary effect of the price increase. This policy package likely contributed to the decline of the Gini coefficient from March 2005 to March However, the Gini coefficient started rising steadily again thereafter. 6. CONCLUSION The objective of this paper has been to analyse the long-term evolution of various indicators of inequality. The paper has made estimates of various measures of inequality, namely the Gini coefficient, the income share of various groups, the decile dispersion ratio, the Theil index of inter-regional inequality, the income gap between certain regions, as well as the decomposition of inequality into between- and within-group inequality. The estimates suggest that inequality in Indonesia has been rising quite significantly. However, the rise in inequality is predominantly visible in the last decade after the Asian Financial Crisis or the era of political reform and democratisation. Before that period, inequality was relatively stable and at the moderate level. The magnitude of more recent rising inequality is rather startling. In the ten years up to, the Gini coefficient and the decile dispersion ratio have increased by roughly 20% and 50% respectively. Furthermore, rising inequality is more or less common across regional dimensions, urban-rural dimensions and Java and non-java regions. We found, however, no strong evidence of increasing interregional inequality in Indonesia over the period. We propose a set of hypotheses, not necessarily mutually exclusive, that may play a significant role in explaining rising inequality in Indonesia. First, the recent commodity boom may play a role in urban-rural inequality, but this is not clear. Second, labour market rigidity in the formal labour market may be significant and the evidence for this is clearer. Finally, rice prices and cash transfer programmes, have likely contributed to changes in inequality because of their relative impacts on the standards of living of the poorer parts of the population versus better off groups, all of which would suggest public policy could be developed to better manage rising inequality. Sustaining economic growth and poverty reduction rates may even be increasingly dependent on this in the years ahead. 9

10 REFERENCES ADB (Asian Development Bank) (). Asian Development Outlook : Confronting Rising Inequality in Asia, Manila: ADB. Agénor, P. (2002). Does Globalisation Hurt the Poor? Mimeograph. Washington, DC: World Bank. Akita, T. (2002). Regional Income Inequality in Indonesia and the Initial Impact of the Economic Crisis. Bulletin of Indonesian Economic Studies 38(2), Akita, T., & Miyata, S. (2008). Urbanization, Educational Expansion, and Expenditure Inequality in Indonesia in, 1999, and Journal of the Asia Pacific Economy 13(2). Akita, T., Kurniawan, P. A., & Miyata, S. (2011). Structural Changes and Regional Income Inequality in Indonesia: A Bidimensional Decomposition Analysis. Asian Economic Journal 25, Alisjahbana, A., Yusuf, A., Chotib, Yasin, M., & Soeprobo, T. (2003). Understanding the Determinants and Consequences of Income Inequality in Indonesia. Bangkok, East Asian Development Network (EADN). Bank Working Paper Number Washington, DC: World Bank. Barro, R. (2000). Inequality and Growth in a Panel of Countries, Journal of Economic Growth, 5:5 32. Berg, A. and Ostry, J. (2011). Inequality and Unsustainable Growth: Two Sides of the Same Coin? IMF Staff Discussion Note. IMF: Washington, DC Birdsall, N. (2007). Income Distribution: Effects on Growth and Development. CGD Working Paper 188. Washington DC: CGD. Chen, S. and Ravallion, M. (1997). What Can New Survey Data Tell Us about Recent Changes in Distribution and Poverty, World Bank Economic Review, 11(2): Cornia et al., (2004). Income Distribution Changes and their Impact in the Post Second World War Period. In Cornia, G.A. (Ed.) (2004). Inequality, growth, and poverty in an era of liberalization and globalization. Oxford: Oxford University Press. Cornwell, K. and Anas, T. (2013). Survey of Recent Developments. Bulletin of Indonesian Economic Studies. 49(1): Deininger, K. and Squire, L. (1998). New Ways of Looking at Old Issues: Inequality and Growth, Journal of Development Economics, 57(2): Dollar, David, and Aart Kraay. (2002). Growth is Good for the Poor. Journal of Economic Growth Easterly, W. (2002). Inequality Does Cause Underdevelopment. Centre for Global Development Working Paper Number 1. Washington, DC: Centre for Global Development. Fosu, A. (2011). Growth, Inequality, and Poverty Reduction in Developing Countries: Recent Global Evidence. WIDER Working Paper No. 2011/01. WIDER: Helsinki. Frankema, E., & Marks D. (2009). Was it really Growth with Equity under Soeharto? A Theil Analysis of Indonesian Income Inequality, , Economics and Finance in Indonesia 57(1), Hanmer, L. and Naschold, F. (2001). Attaining the International Development Targets: will Growth be Enough?, Development Policy Review, 18: Hill, H. (2008). Globalization, Inequality, and Local-level Dynamics: Indonesia and the Philippines. Asian Economic Policy Review 3, Hill, H., Resosudarmo, B. P., & Vidyattama, Y. (2008). Indonesia's Changing Economic Geography. Bulletin of Indonesian Economic Studies 44(3),

11 Kanbur. (2005). Growth, Inequality and Poverty: Some Hard Questions. NY: Cornell University. Kalwij, A. and Verschoor, A. (2007). Not by Growth Alone: The Role of the Distribution of Income in Regional Diversity in Poverty Reduction, European Economic Review, 51(4): Kanbur, R. (1998). Income Distribution and Development, in A.B. Atkinson & F. Bourguignon, eds., 'Handbook of Income Distribution', North Holland. Kraay, A. (2004). When is Growth Pro-Poor? Cross-Country Evidence. World Bank Working Paper. Leigh, A., & van der Eng, P. (2010). Inequality in Indonesia: What Can We Learn from Top Incomes? Journal of Public Economics 93(1-2), Milanovic, B. (2002). Can We Discern the Effect of Globalisation on Income Distribution? World Bank Policy Research Working Paper. Washington, DC: World Bank. Nugraha, K. and Lewis, P. (2013). Towards a Better Measure of Income Inequality in Indonesia. Bulletin of Indonesian Economic Studies. 49(1): Ravallion, M. (1995). Growth and Poverty: Evidence for Developing Countries in the 1980s. Economic Letters, 48: Ravallion, M. (1997). Can High-Inequality Developing Countries Escape Absolute Poverty?, Economic Letters, 56:51 7. Ravallion, M. (1998). Does Aggregation Hide the Harmful Effects of Inequality on Growth?, Economics Letters, 61:73 7. Ravallion, M. (2001). Growth, Inequality and Poverty: Looking Behind the Averages, World Development, 29(11): Ravallion, M. (2004). Measuring Pro-Poor Growth: A Primer. World Bank Working Paper Washington, DC: World Bank. Ravallion, M. (2005). Looking Beyond Averages in the Trade and Poverty Debate, World Bank Working Paper Number Washington, DC: World Bank. Ravallion, M. (2005). "Inequality is Bad for the Poor". World Bank Policy Research, Working Paper Sakamoto, H. (2007). The Dynamics of Inter-Provincial Income Distribution in Indonesia. The International Centre for the Study of East Asian Development, Kitakyushu, Working Paper, 2007, 25. Skoufias, E. (2001). Changes in Regional Inequality and Social Welfare in Indonesia from to Journal of International Development 13, Son, H. and Kakwani, N. (2003). Poverty Reduction: Do Initial Conditions Matter? Mimeograph. Washington, DC: World Bank. Stewart, F. (2000). Income Distribution and Development. Queen Elizabeth House (QEH) Working Paper. Oxford University, Oxford: QEH. Sumner, A. and Edward, P. (2013). From Low Income, High Poverty to High-Income, No Poverty? An Optimistic View of the Long-Run Evolution of Poverty in Indonesia by International Poverty Lines, Working Paper in Economics and Development Studies. Department of Economics, Padjadjaran University, Indonesia Suryadarma, D., Artha, R. P., Suryahadi, A., & Sumarto, S. (2005). A Reassessment of Inequality and Its Role in Poverty Reduction in Indonesia. SMERU Research Institute Working Paper. Jakarta: Social Monitoring and Early Response Unit. Suryadarma, D., Widyanti, W., Suryahadi, A., & Sumarto, S. (2006). From Access to Income: Regional and Ethnic Inequality in Indonesia. SMERU Research Institute Working Paper. Jakarta: Social Monitoring and Early Response Unit. Suryahadi, A., Hadiwidjaja, G., & Sumatro, S. (). Economic Growth and Poverty Reduction in Indonesia Before and After the Asian Financial Crisis. SMERU 11

12 Research Institute Working Paper. Jakarta: Social Monitoring and Early Response Unit. Suryahadi, A., Suryadarma, D., & Sumarto, S. (2006). Economic Growth and Poverty Reduction in Indonesia: The Effects of Location and Sectoral Components of Growth. SMERU Research Institute Working Paper. Jakarta: Social Monitoring and Early Response Unit. Suryahadi, Asep, Gracia Hadiwidjaja, and Sudarno Sumarto. "Economic Growth and Poverty Reduction in Indonesia before and after the Asian Financial Crisis." Bulletin of Indonesian Economic Studies 48.2 (): Van der Eng, P. (2009). Growth and Inequality: The Case of Indonesia, MPRA Paper Munich: Munich Personal RePEc Archive. Van Leeuwen, B., & Foldvari, P. (). The Development of Inequality and Poverty in Indonesia, CGEH Working Paper 26. Utrecht University: Centre for Global Economic History. World Bank (a). PovcalNet, (accessed 1 April ). Yusuf A.A, A. Komarulzaman & M. Purnagunawan & B. Resosudarmo (2013). "Growth, Poverty and Labor Market Rigidity in Indonesia: A General Equilibrium Investigation," Working Papers in Economics and Development Studies (WoPEDS) , Department of Economics, Padjadjaran University, revised Jan Yusuf, A.A., Resosudarmo, B.P. (2008). Mitigating Distributional Impact of Fuel Pricing Reform: Indonesian Experience. ASEAN Economic Bulletin 25,

13 TABLE Table 1. Indonesia s rate of pro-poor growth, for various time periods Growth rate in mean Growth rate at median Mean percentile growth rate Rate of pro-poor growth (nat. pov Line) Rate of pro-poor growth ($2/day) Rate of pro-poor growth 10th percentile th percentile th percentile th percentile th percentile Note: Numbers in italics are ratio of the growth rate in mean to the corresponding growth rate FIGURES Indonesia Urban Rural Indonesia (BPS) Figure 1. Gini coefficient by urban and rural areas of Indonesia 13

14 0.45 Java 0.45 Non-Java All Java Urban Rural All Non-Java Urban Rural Figure 2. Gini coefficient in Java and Non-Java Annual change (%) Provinces Gini coefficient at 2002 Districts Gini coefficient at 2002 Figure 3. Inequality convergence across Indonesian provinces and districts (2002-) 14

15 % Richest 40% Middle 40% Poorest Figure 5. Expenditure share by income groups, all Indonesia (%) Urban Rural % Richest 40% Middle 20% Richest 40% Middle 40% Poorest 40% Poorest Figure 6. Expenditure share by income groups in urban and rural areas (%) 15

16 All Indonesia Urban Rural Java Non-Java Figure 7. Ratio of mean expenditure of 10% richest to 10% poorest households Theil Index (Province) Theil Index (districts) Figure 8. Theil Index of inter-regional inequality 16

17 Percent Within Provinces Within Urban Rural Figure 9. Share of inequality (generalised entropy) explained by within-region inequality SIERRA LEONE LESOTHO C. AFRICAN R. HAITI HONDURAS BOTSWANA SOUTH AFRICA NAMIBIA CONGO, REP. Gini Index (%) GUATEMALA PARAGUAY BOLIVIA THAILAND HONG KONG BRAZIL PANAMA CHILE ZAMBIA GAMBIA, THE PNG SWAZILAND COSTA RICA SRI LANKA MEXICO MADAGASCAR ECUADOR CHINA SINGAPORE RWANDA EL SALVADOR DOMINICAN R. MOZAMBIQUE INDONESIA (2011Y) PERU ARGENTINA MALAYSIA BULGARIA URUGUAY CAMEROON PHILIPPINES UNITED STATES UGANDA IRAN NIGERIA GUYANA GEORGIA MACAO BURUNDI KENYA COTE D'IVOIRE KOREA, REP. SENEGAL RUSSIA INDONESIA (2011) MALI TURKMENISTAN UNITED KINGDOM BURKINA GUINEAGHANA FASO JORDAN TUNISIA TURKEY MALAWI MAURITANIA NICARAGUA MAURITIUS VANUATU ISRAEL PORTUGAL TANZANIA CAMBODIA YEMEN, MOLDOVA MOROCCO JAPAN BENIN UZBEKISTAN LAO INDIA PDR REP. MONGOLIA BOSNIA NEW ZEALAND ALGERIA LATVIA LITHUANIA EGYPT ALBANIA NIGER AZERBAIJAN POLAND IRELAND BANGLADESH NEPAL INDONESIA (1991) ROMANIA TAJIKISTAN GREECE KYRGYZ REP. FRANCE TIMOR-LESTE CROATIA SPAIN ITALYCANADA ARMENIA ESTONIA PAKISTAN CZECH EUROPEAN REPUBLIC NETHERLANDS ETHIOPIA AUSTRALIA UNION SWITZERLAND KAZAKHSTAN CYPRUS INDONESIA (2001) UKRAINE SERBIA BELGIUM ICELAND BELARUS MALTA SLOVAK REPUBLIC GERMANY FINLAND AUSTRIA LUXEMBOURG MONTENEGRO HUNGARY DENMARK NORWAY SLOVENIA SWEDEN Log of GDP per capita (2005 PPP US$) Figure 10. Indonesian Gini coefficient in comparative perspective 17

18 Median spline Percentiles Figure 11. Indonesia s Growth Incidence Curves, vs