Avoiding the Resource Curse Lessons from Indonesia.

• Author: Hill, Hal

1. Introduction

   A central puzzle in development economics is what has come to be known as the "resource curse", the paradox that natural resource abundance appears to be causally associated with poorer economic performance, at least among developing countries. A subset of this analysis refers to the "Dutch Disease", with similarly negative connotations. In their influential paper, Sachs and Warner (2001, and an earlier version circulated as a 1995 NBER paper) establish the proposition empirically, and also provide an analytical framework for understanding the negative correlation between resource abundance and economic performance. Following their empirical approach, Figure 1 shows the relationship between the exports of natural resources as a percentage of GDP in 1970 and real per capita GDP growth in 1970-89 for eighty-eight developing countries. That is, to motivate the study of the relationship, natural resource abundance is defined for a given "base year" and compared with subsequent growth outcomes. The data are for all countries that report at least some natural resource exports. The slope of the regression line is -0.0364, that is, a weakly negative relationship.

   However, there are exceptions to this generalization. Sachs and Warner draw attention to such two countries--Malaysia and Mauritius. (1) Notice from Figure 1 that Indonesia is also an exception, with a similar growth rate. This in fact motivates our paper, to examine and explain how and why Indonesia largely avoided the resource curse. (2)

   Extending this analysis, since the nature and size of resource booms vary over time and across countries, Figure 2 shows the relationship for 151 countries for the period 2000-19, a similar but larger set of countries, allowing also for the existence of some new states, notably in Central Asia. As for Figure 1, the base year, 2000, was also a period when commodity prices were relatively subdued. For the second period, the slope of the line is very slightly positive (0.0085). However, if three extreme outliers--Turkmenistan, Tajikistan and Papua New Guinea--are removed, there is again a negative relationship, with a slightly steeper slope, -0.0124. (3)

   The major conclusion is again that Indonesia is a moderately resource-rich economy with above-average economic performance. Relative to GDP, natural resource exports in 2000 were in the 10-20 per cent range, similar to neighbouring Malaysia and Thailand, higher than Brazil but considerably lower than Nigeria and the Middle Eastern petro states. Indonesia's economic growth over the period was relatively strong, slightly above 4 per cent per annum, and higher than the developing country average. This reinforces our working hypothesis, that Indonesia appears to have managed its resource abundance relatively well.

   Three additional points need to be emphasized at the outset. First, as much as possible, the analysis needs to be comparative. In particular, as reference points we select three additional resource-rich, middleincome, populous (at least in two cases), tropical developing countries which share some common features with Indonesia: Nigeria, which in some respects is the closest comparator country to Indonesia, at least during the initial stages of the resource boom; Brazil, the largest resource-rich Latin American economy; and Malaysia, the other sizeable resource-rich Southeast Asian economy. Second, Indonesia has actually had two resource booms in the last fifty years--an oil and gas boom of the 1970s, when the country was under a centralized, authoritarian government, and the proceeds of the boom accrued mainly to the central government; and a more recent boom, in approximately 2004-11, when Indonesia had democratic and decentralized governance, and the booms were principally in coal, palm oil, and gas. We briefly draw out the implications of these two very different episodes. (4)

   The third point to emphasize is that resource abundance by its very nature involves both booms, during periods of high prices and sudden resource discoveries, and "busts", when prices fall or resources deplete. In fact, it is the bust episodes that require the most careful examination. High economic growth in boom periods is easily achieved, often further facilitated by capital inflow surges that accompany the booms. It is the bust periods that test the quality of economic management--how efficiently the proceeds of the boom (and the additional borrowed funds) were invested, and how effective is the macroeconomic adjustment to the downturn. As Garnaut (2015) observes, "The test of the 'resource curse' is how a country responds to the end of a boom." As we will see, Indonesia's principal strength in managing its resource abundance has been the quality of its macroeconomic management during hard times, and (especially during the first boom) its ability to "recycle" a substantial proportion of the proceeds of the boom into productive investment, notwithstanding the sometimes-egregious corruption that also occurred.

   Our organization is as follows. We commence with a review of the resource curse literature in the next section. The third section examines the various dimensions of the resource booms with special reference to Indonesia and our three comparator countries. In the subsequent section, we investigate outcomes in the four countries, including aggregate rates of economic growth and a range of variables that are commonly associated with the resource curse. The final section sums up.

2. The Resource Curse: An Overview

   We briefly summarize the voluminous literature on the relationship between natural resources and economic performance, referring in particular to three main strands--taxonomies and measurement issues, the Dutch Disease, and the resource curse. The latter is the most relevant for our study, especially as it pertains to developing countries.

   The IMF classifies fifty-one countries as "resource-rich". These are defined as countries which derive at least 20 per cent of merchandise exports or 20 per cent of fiscal revenue from non-renewable natural resources. Twenty-nine of these countries are low- and lower-middle-income. Common characteristics of these twenty-nine countries include extreme (or at least very high) dependence on resource wealth for fiscal revenues, export sales, or both; low saving rates; poor growth performance; and highly volatile resource revenues.

   Booms (and busts) refer to sudden and substantial, often unexpected, increases (decreases) in foreign currency receipts that accrue to a country. They are most commonly discussed in the context of natural resources, through either price or quantity (new discoveries) effects, but they can in principle originate from a wide range of sources. Examples include remittances, aid flows, and even short-term capital flows (the latter is especially an issue for emerging market economies in times of global macroeconomic volatility).

   Boom and bust episodes vary greatly in their intensity and duration, and important policy implications follow from this fact. Countries with very large reserves of a particular commodity will likely experience several episodes depending on international price variability. In these cases, a policy framework needs to be established (for example, a sovereign wealth fund, SWF) that insulates the economy from these episodes. Sudden but short-lived price spikes (or capital flows) will have different effects. The windfall gains will typically be higher in cases where domestic production costs of the booming commodity are low. Booms based on agricultural commodities may have less immediate effects depending on whether new plantings and land acquisitions are required, or there are higher marginal production costs as compared to the classic boom cases of low-cost, easily accessible oil fields.

   The domestic distributional implications also vary greatly. At one extreme is a single commodity and owner of that commodity. A common example is oil, the revenue of which accrues principally to the state through ownership rights and/or taxation arrangements. This approximated the Indonesian case in the 1970s oil boom. At the other extreme are decentralized, private-sector beneficiaries, as is typically the case with agricultural booms. In this case, the principal beneficiaries are the owners of the factors of production, mainly land and perhaps specialized market and technology know-how. The state, and therefore the rest of the nation, are beneficiaries primarily through the increase in state tax revenues, whether general or via specific revenue measures. The owners of the sector-specific factors of production and the workers employed as a result of the boom are also beneficiaries, but these spillovers tend to be local. This latter case approximates conditions in Indonesia's second resource boom, at least for palm oil and some of the coal sector. The political economy implications of these various types of booms are important, and will be addressed subsequently in the Indonesian case.

   Analytically, the issue of booming sectors initially attracted attention through the Dutch Disease literature (see, for example, Corden and Neary 1982, Corden 1984). The increased foreign currency earnings result in a real exchange rate (RER) appreciation, through either higher inflation than major trading partners in the presence of a fixed nominal rate, an appreciating nominal exchange rate, or a combination of the two. The effect is to squeeze the profitability of the non-booming tradable sectors, which then must contract as a share of output and employment. An appreciating RER also increases the profitability of non-tradables, mainly services, and therefore a three-sector model is developed: nontradables (including services) and two tradables, the booming sector and the rest.

   Two effects are at work--the resource-movement effect, as labour and other resources are pulled towards the booming sector; and the...