I. Indonesia’s Commodity Boom: Dutch Disease in the Making?1
Indonesia experienced a remarkable export boom in the years preceding the 2008 global crisis, driven mainly by surging commodity exports. While helping to sustain high economic growth, the commodity boom and the accompanying real appreciation has raised questions about the effect on the manufacturing sector (Dutch Disease) and the growing vulnerability to volatile commodity prices. This paper takes an in-depth look into recent trade patterns to assess the extent of such concerns. It is found that (i) there is no strong evidence of Dutch Disease; (ii) weak performance in some sectors, so far, does not appear to be linked to the commodity boom; and (iii) while further reliance on commodities has increased Indonesia’s vulnerability to export price volatility, the terms of trade have actually been rather stable as import and export prices co-move markedly, mitigating such vulnerability.
1. Indonesia went through an impressive period of export growth in the five years preceding the 2008–09 global crisis. Following somewhat stagnant export growth early in this decade, exports accelerated sharply, increasing by about 120 percent from 2003 to 2008 (until the crisis). This remarkable growth is noticeably stronger than the already impressive performance seen in the early 1990’s. After being interrupted by the collapse in trade in late 2008 and early 2009, the export boom seems to have continued despite a still modest recovery in global activity, with exports rebounding to near pre-crisis levels by late 2009 (Figure I.1).
Figure I.1.Exports of Goods
2. Much of this extraordinary performance reflected a booming commodity sector (Figure I.2). Commodity exports grew by 180 percent during this period—notably faster than manufacturing (75 percent). This is noticeably different from the export boom of the early l990’s, which was solely driven by rapidly growing manufacturing exports. Both renewable resource commodities (most noticeably, vegetable oils and rubber) and nonrenewable resource commodities (mainly oil and gas) contributed to the sharp increase in exports, although the former reflected the combination of rapidly increasing volumes and prices, while the latter mainly reflected a sharp increase in prices.23 Manufacturing exports, meanwhile, displayed decent but significantly lower volume growth (Figure I.3).
Figure I.2.Commodity Exports Figure I.3.Export Performance by Main Products, 1990–2010
Sources: CEIC Data Co., Ltd.; and staff estimates.
3. High commodity prices were accompanied by sharply accelerating intra-regional demand. The increase in commodity prices, both renewable and nonrenewable resources, (Figure I.4) was accompanied by sharply accelerating demand from neighboring countries. Most noticeable was the quadrupling of exports to India, the tripling of exports to China and, because of its already high starting level, the doubling of exports to Japan (Figure I.5). As a result, exposure to neighboring emerging market countries increased markedly although the overall exposure to the region remained broadly stable (Box I.1 discusses recent trends and potential vulnerabilities related to the composition of trading partner countries).
Figure I.4.Deflators of Main Export Products, 1990–2010 Figure I.5.Indonesia Exports to Main Destinations
4. But an appreciating real exchange rate and a sharp pickup in imports have raised questions about the effect on other sectors and the increasing vulnerability to external shocks. Driven mainly by marked upward pressures on the rupiah, the real exchange rate appreciated by about 14 percent from late-2003 to mid-2008. The sharp depreciation following Lehman’s collapse reverted most of the previous appreciation, but the subsequent rebound quickly brought the real exchange rate to above pre-crisis levels and 20 percent above the pre-commodity boom levels (Figure I.6). Notwithstanding appreciating pressures, fast growing exports allowed Indonesia to maintain trade (and current account) surpluses for a prolonged period of time. However, after a prolonged period of sluggish import growth, a rapid catch up in 2007–08, mainly reflecting capital and intermediate goods, led to trade deficits for the first time in more than a decade (Figure I.7). Against this backdrop of weakening external balances, the increased reliance on commodity exports could be a source of vulnerability as their prices tend to be highly volatile, potentially exposing the economy to large terms of trade shocks that could rapidly translate into mounting external imbalances. This is of particular concern if the commodity boom comes at the expense of growth in the manufacturing sector (a phenomenon often referred to as Dutch Disease), as the latter sector is unlikely to recover quickly in the event of a turnaround in commodity exports. The next section assesses whether there is evidence of Dutch Disease in Indonesia.
Figure I.6.Nominal and Real Exchange Rates, 1995–2010 Figure I.7.Trade Balance, 1990–2010 Box I.1.Trading Partner Diversification
The composition of trading partner countries has changed markedly in recent years, shifting towards rapidly growing emerging and developing economies. The overall exposure (export share) to advanced countries fell from 73 percent in 2003 (and an average of 81 percent during the l990’s) to 65 percent in 2008. While this has helped diversify export destinations somewhat and gain market share in rapidly growing economies, partner concentration has remained somewhat high, with 5 countries accounting for more than 50 percent of total exports. A group of 10 countries has retained about 75 percent of export share for more than a decade, although China, India and Malaysia have gained ground at the expense of Japan and the United States. The result has been a sustained downward trend in Indonesia’ Herfindahl index of trading partner concentration. Despite the sharp acceleration in demand from neighboring countries, exposure to the region as a whole has increased only marginally from 60 percent of total exports in 2003 to 62 percent in 2008, as increases in the share of exports to China, India, and Malaysia have been offset by a declining share to Japan.
Herfindahl Index on Export Markets
Shares of Main Export Destinations
This shift has been accompanied by heightened vulnerability to foreign demand shocks, reflecting primarily increased co-movement across trading partners. We construct a measure of foreign demand volatility as a weighted average of trading partners’ domestic demand volatility, with weights given by the trading partner’s share in Indonesia’s exports.1/ This index, depicted below, points to a sharp increase in foreign demand volatility in the last couple of years after a prolonged period of very low volatility (consistent with the phenomenon often referred to as the “Great Moderation”). Heightened volatility of foreign demand, also affecting other countries, reflects mainly a sharp increase in correlations across trading partners rather than their changing shares in Indonesia’s exports (as suggested by the negligible gap between the index with fixed and moving weights). At the same time, covariance with trading partners picked up recently, after a long period of negative or nil correlation, although, again, the changing structure of trading partners has play no significant role.
Volatility of Foreign Demand, 2000–10
Covariance of Indonesia’s GDP and Trading Partners’ Domestic Demand, 2000–10
B. Dutch Disease in the Making?
5. Commodity booms often lead to ‘DutchDisease.’ As extensively documented, first by Corden (1984) and Corden and Neary (1982) and later by an extensive literature,4 commodity booms—resulting from sharp increases in production (e.g., following the discovery of new sources) or in prices—often have pervasive affects on other sectors (see Box I.2). Dutch Disease is normally associated with (i) real appreciation, (ii) a slowdown in manufacturing exports, output and employment, and (iii) an increase in wages.
Box I.2.The Dutch Disease Hypothesis
The Dutch Disease phenomenon is normally associated with two main effects:
A resource movement effect that refers to the reallocation of factors from other sectors of the economy (e.g., manufacturing) to the natural resource booming sector. This effect reflects increased demand in the resource intense sector that tends to attract labor from other sectors of the economy by means of higher wages. That is, if labor is mobile across sectors, higher wages in the export booming sector would cause a movement of labor toward this sector, leading to lower output in other sector (if the economy is operating at full capacity). This process of resource reallocation often leads also to an appreciating real exchange rate. Lower production in the nonbooming sector (including nontradable) results in a loss of production giving rise to excess demand for nontradables, and leading to an increase in the relative price of nontradables (thus, the real exchange rate).
A spending effect that relates to the appreciation of the real exchange rate as a result of increased spending of (at least part of) the booming sector’s extra income. Increased demand of nontradables leads to exchange rate appreciation, as nontradable prices need to adjust upwards to induce higher production in response to higher demand. The magnitude of this effect normally depends on the propensity to consume nontradable goods, which tends to be higher when large parts of the additional income is received by the government, as the latter tends to have a high propensity to consume nontradable goods.
Combining the two effects, the Dutch Disease hypothesis generates three unambiguous predictions:
(i) since the relative price of nontradable goods increases, the real exchange rate appreciates;
(ii) manufacturing output and employment fall due to factor reallocation; (iii) the overall wage level increases (possibly starting with higher wages in the booming sector) in response to higher demand for labor. The combined effect on output and employment in the nontradable sector is ambiguous, as the spending and resource movement effects push in opposite directions.
6. In Indonesia, the commodity boom in recent years has been accompanied by significant real exchange rate appreciation, although there is no evidence of overvaluation. Marked appreciation in recent years followed rapid income and productivity gains—mainly earlier in this decade—and served to revert much of the overshooting experienced during the 1997/98 crisis. As a result, today, the exchange rate is broadly in equilibrium with economic fundamentals, as suggested by the different CGER5 methodologies.
7. Evidence of manufacturing exports being affected is mixed, with significant heterogeneity within the group. While manufacturing exports have been markedly outpaced by commodity exports in recent years, its growth has still been robust at an aggregate level (Figure I.3). Some traditional industries (mainly textiles, wood manufactures and paper products) have performed poorly, while others (e.g., chemicals and machinery and apparatus) have showed remarkable growth. Still, most of the sectors that witnessed sluggish growth in recent years seem to have been on that path long before the commodity boom manifested itself (Figure I.8).
Figure I.8.Manufacturing Exports by Main Products, 1990–2009
8. GDP data confirms that sectoral performance has been uneven, and weak output does not appear to be linked to the recent commodity boom. A long-term view of sectoral output reveals that sectors that have been sluggish in recent years have displayed weak performance long before the commodity boom, suggesting that real appreciation may not have been the main factor behind these sectoral weaknesses (Figure I.9). This is particularly clear for textiles, wood manufactures, and iron and steel industries, which have shown sluggish growth since 2000. At the same time, sectors closely linked to commodities (e.g., food, fertilizers, chemical rubber, cement) have witnessed decent, although slowing, economic performance in recent years. And other capital-intensive industries (e.g., automotive and machinery) have actually been booming in recent years—like in the period preceding the commodity boom—arguably suggesting that long standing labor market frictions may have been a constraint on growth in some labor-intensive industries.6
Figure I.9.Real GDP by Sector, 1993–2003
9. Only recently, wages pressures have started to appear in the manufacturing sector.7 Consistent with the Dutch Disease hypothesis, wages in the commodity sectors (particularly in mining) have grown rapidly and outpaced those in other sectors in recent years. Employment in these sectors (except in agriculture, forestry and fishery) has also grown fast, and faster than in the manufacturing sector. Still, until 2008, wage pressures on the manufacturing sector had not materialized, partly reflecting one-off reforms that helped to lower wages in the sector.8 In fact, wages in this sector decreased by about 15 percent in real terms during 2003–08, along with somewhat smaller decreases in service sectors, while real wages in the mining sector grew by 19 percent (Figure I.10). More recently, however, wage pressures have appeared in the service and manufacturing sectors, with the later showing real wage increases of about 13 percent during 2009 alone.
Figure I.10.Employment and Wage by Sector, 2003–08
10. Finally, econometric analysis also suggests there is little evidence of Dutch Disease. A standard vector autoregressive (VAR) model is estimated to gauge the effect of commodity price shocks on output of key manufacturing sectors. The model is estimated for each sector with monthly data for the period 1993–2008 (until Lehman’s crisis). Sectoral output is measured by the corresponding industrial production index. To control for possible correlation between commodity and sectoral manufacturing prices, the sectoral export price deflator is also included in the model. Finally, an index of total imports in trading partners—weighted by their share in Indonesia’s exports—is introduced to control for external demand shocks. Results (Figure I.11) suggest that only the textile sector may have been affected by commodity price shocks (as indicated by the negative and statistically significant impulse response and the 35 percent of the variance explained by commodity prices). Other sectors do not display any statistically significant evidence of Dutch Disease. In fact, if any, the effect of higher commodity prices seems to be positive, reflecting some correlation of commodity and industry specific export prices, as well as likely inter-sectoral complementarities (e.g., machinery and chemical industries are closely linked to production in the commodity sector), although the role of commodity prices in explaining the variance in output is limited.
Figure I.11.VAR--Impulse Responses and Variance Decomposition
C. Is Increased Reliance on Commodity Exports a Source of Concern?
11. Although there is no clear evidence of Dutch Disease, greater reliance on commodities raises questions about increasing vulnerability to terms of trade shocks (Figure I.12). Despite increasing diversification within manufacturing exports, and to some extent within the commodity group as well, overall diversification has remained broadly constant for more than a decade—as indicated by the overall Herfindahl index of product concentration9—reverting the trend seen in the 1990’s and early 2000. This reflected the fact that greater within-group diversification has been offset by increased reliance on commodities, which have a lower degree of diversification (higher Herfindahl index) than the manufacturing sector (Figure I.13).
Figure I.12.Product Concentration, 1990–2010 Figure I.13.Product Concentration, 1990–2010
12. Export price volatility has increased significantly, reflecting increased volatility in international prices as well as increased concentration in price-volatile products. We construct an index of export price volatility that tracks the variance of main export prices over time, weighting them by their share in total exports (Figure I.14).10 As expected, the index shows that volatility has increased significantly in the last couple of years, even before the sharp and generalized fall in commodity prices associated with the collapse in trade (after Lehman). This increase is explained both by heightened variance of underlying prices as well as increased correlation among them. As suggested by the difference between the index with fixed weights and the index with moving weights, increased reliance on commodity exports has significantly contributed to Indonesia’s export price volatility, precisely because commodities display higher volatility than manufacturing products (even after detrending and seasonally adjusting them).
Figure I.14.Volatility of Export Price, 1991–2010
13. However, Indonesia’s terms of trade have displayed limited volatility. Being also a significant importer of raw materials, Indonesia’s import prices have also fluctuated sharply in recent years, and closely co-moved with export prices (Figure I.15). As a result, terms of trade have actually remained quite stable over the commodity boom period, as well as during the commodity price bust of the 2008 global crisis. This suggests that, despite significant exposure to commodities, Indonesia’s external balances are likely to remain quite resilient to external price shocks.
Figure I.15.Terms of Trade, 1990–2010
14. The unprecedented export boom preceding the 2008/09 global crisis, and the associated marked real appreciation and import boom, have raised questions about the impact of the commodity boom on other sectors of the economy, as a turnaround on commodity prices (and demand) could quickly lead to mounting external imbalances. An in-depth look at recent trade patterns, sectoral output and labor markets, however, does not point to an obvious case of Dutch Disease at this point, although there is evidence of pressures building in some sectors. While there is evidence of some stagnating manufacturing sectors, this does not seem to relate to the recent commodity boom, suggesting that other structural factors (e.g., infrastructure bottlenecks and labor market frictions) may have played a role. Still, increasing reliance on commodity exports could make the economy vulnerable to external price shocks, although the high correlation between export and import prices (reflecting a high commodity content in imports) go a long way in mitigating price shocks. Addressing structural problems will be key to foster growth in the manufacturing sector and diversifying the economy away from commodities, while still exploiting its comparative advantage.
CeruttiE. and M.Mansilla2008“Bolivia: The Hydrocarbons Boom and the Risk of Dutch Disease,”IMF Working Paper No. 08/154 (Washington: International Monetary Fund).
CordenW. Max1984“Booming Sector and Dutch Disease Economics: Survey and Consolidation,”Oxford Economic PapersVol. 36 (November) pp. 359–80.
CordenW. Wax and J.Peter Neary1982“Booming Sector and De-Industrialisation in a Small Open Economy,”The Economic JournalVol. 92 (December) pp. 825–48.
IsmailK.2010“The Structural Manifestation of the ‘Dutch Disease’: The Case of Oil Exporting Countries,”IMF Working Paper No. 10/103 (Washington: International Monetary Fund).
OomesN. and K.Kalcheva2008“Diagnosing Dutch Disease: Does Russia Have the Symptoms?,”IMF Working Paper No. 07/102 (Washington: International Monetary Fund).
Prepared by Gustavo Adler, with research assistance from Agnes Isnawangsih.
Commodities labeled as renewable resources refer mainly to agriculture, animal, fishery and forestry related activities, while nonrenewable resource commodities refer mainly to mining activities.
Volume and deflators are estimated by staff, based on value and volume data for export and import groups corresponding to SITC 2 digit level disaggregation, as aggregate official statistics do not weight subgroups by their economic value.
Consultative Group on Exchange Rates.
In particular, high severance payments have been a mayor constraint on labor-intensive industries (see Thacker, 2005, “Labor Market Policies and Job Creation.”
Serious data limitations on wages and employment prevented more in-depth analysis of sectoral trends.
Minimum wage setting was decentralized to the provinces, giving rise to inter-province wage competition in recent year.
The Herfindahl index of product concentration is constructed from SITC 2-digit disaggregation groups.
Based on the construction of the overall export deflator, the vulnerability index is computed as, where P is the export deflator, Pi is product i’s price deflator, wi is product i’s share in total exports and τR is the variance of P. As before, weights are fixed or allowed to move over time depending on the desired information. Variances and co-variances are measured on seasonally-adjusted and (HP-filter) detrended series of export deflators.