Viewed in long-term perspective, Indonesia's record of sustained poverty reduction is remarkable. Over the two decades ending in 1996, the proportion of the population with expenditure below the government's poverty line declined from almost two-thirds to less than one-fifth. Poverty incidence declined dramatically in both rural and urban areas and in all provinces. The principal driver was economic growth (Figure 3.1). Since 1996 the rate of poverty reduction has been more moderate but there were just two periods during which poverty incidence did not fall (Figure 3.2). The first was the Asian financial crisis of 1997–98. A massive economic contraction occurred during which GDP declined by 13 per cent in a single year. It is hardly surprising that poverty incidence temporarily increased (by 5.7 per cent). The second period was 2005–2006. Explaining this episode is more difficult, because poverty incidence rose (from 16 per cent in 2005 to 17.8 per cent in 2006) despite GDP growth exceeding 5 per cent.

Viewed in regional perspective, the experiences of Indonesia and its neighbours share some key features. First, other rapidly growing economies, such as Thailand and Malaysia, also achieved dramatic long-term reductions in poverty incidence. Even the Philippines, while recording lower growth rates than the other three, nevertheless experienced significant long-term reductions in poverty incidence. Second, in all four countries poverty incidence rose during the recessions accompanying the Asian financial crisis. Third, Thailand and the Philippines, like Indonesia, experienced a mysterious increase in poverty incidence around 2006, of a similar magnitude to Indonesia's. In all three countries, poverty incidence rose despite continued GDP growth above 5 per cent. This chapter looks at this puzzling increase, focusing on Indonesia.

In the next section, I review some of the empirical and analytical issues arising from the measurement of poverty incidence and comparisons over time and across countries. I then examine the long-term data on changes in poverty incidence in the above four Southeast Asian countries. The analysis then focuses on the 2006 increase in measured poverty incidence in Indonesia. I show that a projection of changes in poverty incidence based on the econometrically estimated relationship between changes in poverty incidence and the rate and sectoral composition of economic growth predicts a significant reduction in poverty incidence between 2005 and 2006, rather than the increase that actually occurred.