I. Export Performance1
A. Introduction and Summary
1. Exports have played a prominent role in the rapid industrialization and sustained high economic growth of Asian countries over the past several decades. The growth in exports, including the substantial rise in intra-regional trade, has helped to sustain the growth of world trade, which over recent years has been well above the growth of world output. The changing composition of Asia’s exports, especially toward manufacturing goods, has been aided by the reduction of distortions in domestic markets and in trade barriers. Within the region, shifts in the pattern of production have reflected differences in the stages of development and industrialization, with labor-intensive and lower value-added industries moving toward the lower wage economies. Capital flows both from outside and within the region have reflected these shifts in the distribution of production.
2. Indonesia has followed broadly the regional pattern, although the share of manufactures in total exports lags behind that of several other ASEAN countries. Export growth and diversification started late mainly because of the inward-looking development strategy until the early 1980s. Foreign investment was closely controlled, limiting access to much needed capital and technology. Rising oil revenues from the two oil shocks of the 1970s allowed this strategy, but, once oil prices started to decline, more export-oriented polices were adopted. A policy of trade deregulation was initiated and there were reduced restrictions on foreign investment. As a result, manufacturing exports increased from 4 percent of GDP in 1980 to 13 percent in 1996, led initially by footwear and garments and subsequently by diversification into electronics and other nontraditional goods.
3. An important objective of this paper is to explain through econometric analysis the contribution of different factors to this growth and assess the prospects for its maintenance in the future. The paper is organized as follows: Section B reviews developments in Indonesian export growth since 1970. Section C develops a model of export demand and export supply, provides empirical estimates from the model and the implications of the results. Section D addresses medium-term prospects for export growth on the basis of the estimated model and the possible impact of exchange rate changes.
4. The main results from the study are as follows:
- Indonesian exports are influenced by world demand for imports, but supply factors are the dominant source of the expansion in manufacturing exports, accounting for 88 percent of the observed rise from 1980 to 1994.
- Investment rates in the export sector have increased sharply over the last fifteen years and contributed 45 percent of the sustained expansion of manufactured exports.
- Trade liberalization measures accounted for 40 percent of the expansion of manufactured exports over the same period.
- The rise in export prices relative to domestic prices of 7 percent over the last fifteen years, has induced an increase in the supply of manufactured exports, that accounts for 3 percent of the total expansion during this time.
- The medium-term outlook for manufactured export growth is highly favorable if access to foreign capital and markets is maintained, and trade liberalization policies are continued. Based on medium-term projections for prices, world import demand, and investment from the World Economic Outlook, exports can be expected to grow at rates well above 10 percent over the next few years due to abundant supply of labor and natural resources.
- Resisting nominal exchange rate appreciation associated with large capital inflow risks, inducing real exchange rate adjustments through higher domestic inflation, with potentially detrimental effects on exports, investment, and growth.
B. Export Developments, 1970-96
5. In the last quarter of a century, Indonesia’s manufactured exports have risen from less than 2 percent of GDP in the early 1970s to 13 percent of GDP in 1996 (Chart 1). By contrast, the share of agricultural and mineral exports fell from nearly 6 percent to 2.5 percent. The share of oil exports in GDP rose dramatically following the two oil shocks of the 1970s; it subsequently fell from a high of 20 percent in 1980 to 6 percent. Part of this decline reflects falling oil prices, but some of the decline is also due to the increased demands for oil from the domestic economy. As a result of these trends, the share of manufactured goods in total exports increased from only 12 percent in 1970 and 14 percent in 1980 to 58 percent in 1996 (Chart 2). While this share is still below the average for the Asian region of about 75 percent, the difference is much less pronounced than in 1970 when the regional share of manufactured exports was 42 percent of total exports.
Chart I.1Indonesia: The Changing Shares Of Exports, 1970-96
Sources: UN International Trade Statistics and Fund staff estimates.
Chart I.2Indonesia: Export Shares, 1996
Source: Data provided by the Indonesian authorities.
6. Exports of manufactures were led initially by plywood, garments, and footwear. Competitiveness in these sectors—which make up approximately 50 percent of manufactured exports and almost 30 percent of total exports—is especially sensitive to large increases in minimum wages, competition from other countries in the region, and slow-moving bureaucratic procedures, which have limited the growth of these exports in recent years (Chart 3). On the other hand, dynamism in attracting high rates of foreign investment into other nontraditional types of manufactured exports, such as electrical appliances, has sustained rapid export growth for manufactured goods over the past decade. Moreover, Indonesia has been relatively less affected by the regional export slowdown than other Asian countries.
Chart I.3Indonesia: Manufacturing Exports By Major Commodities, 1981-96
Sources: Data provided by the Indonesian authorities and Fund staff estimates.
Tariffs and regulatory issues
7. In the mid 1980s the Indonesian government began to liberalize its trade and investment policies and this process has been continued through successive deregulation packages. Fane and Condon (1996) estimate that the real effective rates of protection for manufacturing, including oil refining, fell from 27 percent in 1987 to 11 percent in 1995; for manufacturing, excluding oil refining, the fall was from 59 percent to 10 percent; and for agriculture from 9 percent to 4 percent.
8. Land resources for establishing free trade zones and industrial estates have been set aside by the government to attract foreign investors. These zones have relaxed exporting and importing regulations and reduced tariffs. Prior to the deregulation of joint ventures in 1994, it was only within a free trade zone that foreign investors were allowed to make investments without forming a joint venture. These export zones along with other parts of Indonesia have attracted investment and trade from Singapore and Malaysia.
9. However, there remain restrictions, which generate inefficiencies and add to production costs. These include export bans; export taxes, including on forestry products, palm oil, logs, and rattan; and export levies. There are a number of domestic features that restrict competition and inhibit export growth, including cartels in the cement, paper, and fertilizer industries; entry/exit controls in the plywood industry and in retail trade; exclusive licensing on clove marketing and wheat flour milling; and a dominant public sector in steel, fertilizer, and refined oil products.
C. Model of Export Performance
10. In most empirical work on export determination, export volumes are typically related to changes in real activity in foreign markets and to changes in relative prices. While such an approach has proven successful in many empirical studies of other countries or regions, it is not entirely appropriate for modeling export performance in Indonesia. In particular, the rapid increase in the value of exports and the increase in the share of manufactured exports over the last quarter of a century reflect dramatic changes in the underlying structure of production. This suggests that Indonesia’s exports have been both demand determined and supply driven.
Demand for exports
11. The theoretical foundation underpinning the approach adopted in this paper is based on the imperfect substitute model discussed in Goldstein and Khan (1985). This model is based on the premise that exported goods are imperfect substitutes for goods produced in the home market of foreign consumers. The demand for exports function can be derived from a foreign consumer’s utility maximization problem.2 Using the optimization approach with an infinitely lived representative foreign consumer, it can be shown that foreign demand for exports is completely determined by foreign activity (typically captured by some measure of permanent income) and by relative prices.
12. The empirical specification of the steady state relationship implied by the model of foreign consumer’s behavior is given by
where xt is exports of the domestic economy (px/p*)t is the relative price of exports to foreign prices, wt is a measure of world activity, and єt is a stochastic error term.
13. The results of estimating (1) for the period 1972-94 are reported in Table 1. They underscore the importance of world import demand for Indonesia’s exports of manufactures. The elasticity of manufactured exports with respect to world import demand is slightly in excess of unity. This implies that a 10 percent increase in world import demand will increase the demand for Indonesia’s manufactured exports by slightly more than 10 percent. The relatively strong impact of external demand suggests that, in common with many other developing countries, Indonesia’s exports of manufactured goods to developed country markets are strongly affected by cyclical fluctuations in the world economy.
|Coefficients||Case 1||Case 2|
14. The impact of changes in relative prices on the demand for manufactured exports is less clear: the coefficient on relative prices, although correctly signed, is insignificant at conventional levels. These results suggest that the rapid growth in Indonesia’s manufacturing exports has not been accompanied by a trend decline in their prices vis-á-vis foreign competitors. This is, however, consistent with findings of a number of other studies in this area.3
Supply of exports
15. Given the substantial changes in the structure of Indonesian exports over the last decade it is important to differentiate between the major categories of exports. This study therefore focuses separately on manufactured exports, agricultural and mineral exports, and oil and gas exports. The export supply functions for each of these categories can be considered to be solutions to the underlying profit maximizing behavior of exporting firms. Under a constant returns to scale production technology, output will depend on prices and factor inputs into the production process. In addition to labor and capital inputs, we also allow for trade distortions. By impeding efficiency and raising production costs, these distortions effectively imply lower output.
16. A supply equation for each category of exports is specified in terms of shares of total output, since the relative shares of each sector have changed significantly over the long term. Each export supply equation is specified as a function of relative prices, factor inputs, and trade distortions, as follows:
where sit is the share of the ith sector in GDP (i=1 is manufactured exports, i=2 is agricultural exports, and i=3 is oil exports). The pjt, terms represent the respective export prices relative to prices of nontraded goods, kt is the capital-labor ratio, τt is a measure of trade distortions, and єit is stochastic error term for sector i.
17. In this paper, we proxy the extent of trade distortions as the ratio of import tariff revenue to total imports plus the ratio of export tax revenue to total exports. A broader, and more accurate measure of trade distortions would include other impediments to trade, such as nontariff barriers. However, systematic data on nontariff barriers to imports and export bans is not available. A major drawback of measuring trade distortions by tariff and tax revenues is that this variable also depends on cyclical fluctuations. For example, import tariff revenues will rise when imports increase, such as during periods when economic growth is above potential, even through there may have been no change in tariff rates. To account for such distortions, we use a smoothed series for τt, in the estimation of equation (2). 4
18. The results of estimating (2) for the period 1970-94 are reported in Table 2.5 They demonstrate that the capital-labor ratio plays an important role in the production of manufactured exports. The coefficient estimate implies that when this ratio rises, prices rise by 10 percent, the share of manufactured exports in GDP increases by 0.3 percentage points. This result reflects the process of industrialization in Indonesia. The adoption of foreign technologies, embodied in imported capital has increased the share of manufacturing in the economy, and thereby the share of manufacturing exports in total exports. Over the last fifteen years the capital-labor ratio has been growing at an average annual rate of nearly 10 percent. As a result, investment increased from 13 percent of GDP in 1970 to 32 percent in 1994. This increase in investment has resulted in a sustained increase in the share of manufacturing exports since 1980 of 4.5 percent of GDP, or nearly half of the increase.
|Oil and Gas|
|β own price||0.044||0.039||0.110|
|β cross price||-0.095||-0.095|
|σ X 10-2||1.64||0.95||0.961.44|
19. Manufactured exports have also been boosted by trade liberalization. The measure of trade distortions used in this paper has fallen dramatically (around 3 fold) since 1970 (Chart 4). This implies that over the sample period the reduction in trade distortions has increased the share of manufacturing exports by 4 percent of GDP, despite the very rapid growth in GDP itself.
20. The supply of manufactured exports is not strongly influenced by the price of manufacturing goods. Over the past fifteen years there has been an increase in relative prices of nearly 7 percent, which has accounted for an increase in the share of manufactured exports of 0.3 percent of GDP.
Chart I.4Indonesia: Trade Distortions, 1970-96
Sources: IMF, Government Finance Statistics, International Financial Statistics; and Fund staff estimates.
21. Since 1980, the share of manufactured exports has increased by 10 percent of GDP and supply factors have contributed 88 percent of this increase. The remaining 1.1 percent of the increased share in GDP can be attributed to increased demand from the rest of the world.
Reduced form model
22. The estimates of the demand and supply equations can be combined in a reduced form model to determine aggregate export performance. The advantage of this approach is that it allows us to conduct simulations of the impact of world economic conditions and policy changes on future export performance.
23. The reduced form estimates are derived by obtaining the equilibrium values from the supply and demand equations. For example, inverting the manufacturing supply equation and substituting this in the manufacturing demand equation yields
where Δ In xt is the growth of manufactured exports, ep is the domestic price level in US dollar terms, e is the US dollar-rupiah exchange rate, and εp, εk, and ετ are the elasticities with respect to relative prices, capital and trade distortions respectively.6 Substituting the estimated parameters implies:
24. These (derived) estimates are broadly similar to results obtained by other researchers for countries in the region. Ito et al(1996) run similar regressions of export growth as a function of the real effective exchange rate and a real activity variable for the APEC economies. For Indonesia, their estimate of the coefficient on the real effective exchange rate is -0.32 and 1.27 for the coefficient on the real activity variable. The sum of the coefficients of all the activity variables in (4) is 1.39 (assuming a decrease in trade distortions). The impact of the real exchange rate is estimated to be broadly the same in both studies; a 1 percent appreciation in the real exchange rate typically is associated with a 0.32 percent decline in manufactured export growth, compared with the current estimate of 0.28 percent.
D. Simulations and Policy Implications
25. In this section we address some of the policy implications of the empirical results and the medium-term prospects for export performance. In particular, we examine the impact of changes in the real exchange rate on external competitiveness, including through the effect of nominal exchange rate changes on domestic prices. The potentially adverse effects of an increase in the real exchange rate are also compared with gains in competitiveness from trade liberalization (through increases in efficiency and reductions in production costs).
26. Using the reduced form model (4) in combination with World Economic Outlook projections for world import demand, foreign prices, and capital stock formation (based on investment projections), we derive medium-term projections for growth of manufactured exports. Supplementing these projections with World Economic Outlook projections for nonmanufactured export growth provides the medium-term outlook for total export growth, shown in Table 3. An important feature of the outlook is that manufactured exports are expected to continue to grow faster than total exports. Aggregate growth of exports is expected to remain buoyant, mainly owing to the abundant supply of labor, access to foreign capital, and large natural resources. Continuing increases in the capital-labor ratio are an important ingredient in the model’s out-of-sample projections. For example, growth in global demand for imports of Indonesian goods is expected to be 6.4 percentage points in 1997, leading to a volume growth in manufactured exports of 4.6 percentage points. If the percentage change in the capital-labor ratio is in the order of 10 percent, then a further 1.5 percentage points in manufactured export growth can be expected. There is only a small change in the relative price of exportables in 1997 leading to negligible change in manufactured exports. Assuming no change in the level of trade tariffs, the overall growth of manufactured exports in volume terms is 6.1 percent. An increase in export prices of 4.4 percent implies that export values are likely to grow by 10.8 percent.
Impact of exchange rate changes
27. The impact of exchange rate changes on competitiveness can be examined in two stages. First, the effect of a nominal exchange rate appreciation assuming domestic and foreign prices do not change: the estimates from equation (4) suggest that a 10 percent nominal appreciation would—through its impact on the real exchange rate—lead to a decline in the growth rate of manufactured exports by about 3 percent. This, however, represents only the short-run impact effect. A higher nominal exchange rate would reduce domestic prices (or alternatively, reduce inflationary pressure), and thereby offset some of the real exchange rate rise due to the increase in nominal exchange rates. Thus, the long-run impact of exchange rate changes depends importantly on the extent of exchange rate pass-through on domestic prices.
28. To measure both the speed with which exchange rate changes affect domestic prices, and also the magnitude of exchange rate pass-through to domestic prices, we estimate a relatively simple inflation forecasting equation. This equation (based on an error correction model) yields the following results:
and where p is the domestic consumer price index, e is the nominal rupiah-US dollar exchange rate, and p* is the U.S. consumer price index, and m is the narrow money supply. The augmented Dickey-Fuller test for a unit root in the disequilibrium term zt was -3.92, compared to a critical value of-3.80 at the 5 percent level, implying cointegration between domestic prices, foreign prices, and exchange rates.
29. The speed of adjustment to any disequilibrium in relative purchasing power is fairly rapid, with approximately 65 percent of any shock to the long run equilibrium dissipated within one year. The main source of the adjustment is not identified because this is a reduced form model of inflation, as opposed to a structural model. In some periods, it may be the exchange rate that adjusts while at other times it may be domestic inflation. However, the impact of external influences on domestic inflation is strong and fairly rapid.
3. Impact of trade liberalization
30. To reinforce the results of the econometric model a simple counter-factual exercise is conducted. In this exercise we suppose that all the trade tariffs that were in place during the period 1991 to 1994 had been removed. The model’s predicted path of the share of manufactured exports in GDP under this zero tariff assumption is then compared with the model’s predicted path with the actual trade distortions—import tariffs and export taxes—that existed at the time.7 In line with the earlier results, the simulation results indicate that the response with trade distortions removed increases the share of manufactured exports by 3 percent of GDP, relative to the response under the baseline scenario, with a 90 percent confidence band of plus or minus 1 percent of GDP. The implications are that past increases in the openness of the economy have been associated with improved manufacturing export performance, that more trade liberalization would have resulted in even better performance, and that future liberalization will lead to further improvements in export performance.
31. Indonesia’s past export performance has been very favorable. In principle, exports should continue to grow due to abundant supply of labor, access to foreign capital, and natural resources. However, Indonesia faces a challenge in order to maintain the momentum of non-oil export performance, especially given the diminishing role of oil exports. Policy action should be taken to ensure that higher labor costs are offset by reduced bureaucratic procedures and better quality infrastructure. Further trade liberalization is also necessary, especially owing to increasing competition from lower cost producers. Heavily protected local producers in some sectors are poorly positioned for freer trade and these industries make Indonesia vulnerable to retaliatory action by trading partners and use valuable resources that could be better used elsewhere. Active participation in multilateral trade agreements, including regional initiatives, will assist the trade liberalization process.
The manufacturing export demand equation is estimated using annual data for the period 1970-94. The variable definitions and sources are given below.
- px is the price of manufactured exports in U.S. dollar terms (assumed to be equal to the price of manufactured goods) (source: World Bank, World Tables). The period average rupiah/dollar exchange rate is used to convert the price index from local currency to U.S. dollars (source: International Financial Statistics).
- x is the volume of manufactured exports. It is defined as the share of manufactured exports in total exports (source: United Nations, Yearbook of International Trade Statistics) multiplied by the value of total exports (source: International Financial Statistics) and deflated by the index of manufactured export prices, px.
- w is a measure of world import demand in U.S. dollar terms. It is constructed on the basis of trade-weighted imports of trading partner countries (source: International Financial Statistics and World Economic Outlook).
- p* is the price (unit value) of manufactured exports from the rest of the world in U.S. dollar terms (source: World Economic Outlook).
The data used to estimate the parameters of the supply equations are annual and cover the period 1970-94. In addition to the variables used to estimate the demand equation, we use the following:
- sit is the share of commodity i in GDP, given by the share of commodity i in total exports (source: United Nations, Yearbook of International Trade Statistics) multiplied by the ratio of total exports to GDP (source: International Financial Statistics).
- p is the consumer price index (source: International Financial Statistics).
- k is the ratio of capital per worker. The capital stock is constructed from investment data, (source: International Financial Statistics) assuming a depreciation rate of 5 percent. The number of workers is proxied by the size of the population (source: International Financial Statistics).
- τ is a measure of trade distortions. It is defined as the sum of government revenue from import tariffs as a proportion of total imports, and revenue obtained from export taxes as a proportion of total exports (source: Government Finance Statistics). This series is then smoothed using the Hodrick-Prescott filter with the smoothing parameter determined by the method of generalized cross-validation described in McDermott (1997).
AndrewsDonald W. K. (1991) “Heteroskedasticity and Autocorrelation Consistent Covariance Matrix Estimation” Econometrica 59817–858.
DiewertE. (1974) “Applications of Duality Theory” inM.Intriligator and D.Kendrickeds.Frontiers of Quantitative Economics vol. 2 (Amsterdam: North Holland).
DiewertE. and CatherineJ. Morrison (1986) “Export Supply and Import Demand Functions: A Production Theory Approach” NBER Working Paper #2011.
FaneGeorge and TimothyCondon (1996) “Trade Reform in Indonesia, 1987-95” Bulletin of Indonesian Economic Studies 3233–54.
FreedmanD.A. and S.C.Peters (1984) “Bootstrapping an Econometric Model: Some Empirical Results” Journal of Business and Economic Statistics 2150–158.
GiorgianniLorenzo and GianMaria Milesi-Ferretti (1997) “Determinants of Korean Trade Flows and their Geographical Destination” IMF Working Paper WP/97/54.
GoldsteinMorris and MohsinS. Khan (1976) “The Supply and Demand for Exports: A Simultaneous Approach” Review of Economics and Statistics pp 257–286.
GoldsteinMorris and MohsinS. Khan (1985) “Income and Price Effects in Foreign Trade” inR.W.Jones and P.B.Kenen (eds.) Handbook of International Economics vol. IIElsevier Science pp 1041–1105.
ItoTakatoshiPeterIsardStevenSymansky and TamimBayoumi (1996) Exchange Rate Movements and Their Impact on Trade and Investment in the APEC Region IMF Occasional Paper 145 (Washington D.C.).
McDermottC. John (1997) “An Automatic Method of Choosing the Smoothing Parameter in the HP Filter” (Unpublished: Washington: International Monetary Fund).
ParkJoon Y. (1992) “Canonical Cointegrating Regressions” Econometrica 60119–143.
PhillipsPeterC.B. and SamOuliaris (1990) “Asymptotic Properties of Residual Based Tests for Cointegration” Econometrica 58165–193.
ReinhartCarmen M. (1995) “Devaluation, Relative Prices, and International Trade” IMF Staff Papers 42 pp. 290–312.