Turkey : export miracle or accounting trick?

Policy, Research, and Extemal Affairs WORKING PAPERS Country Operations l Latin Amerca and the Caribbean Regional Office, Country Department II The World Bank April 1990 WPS 370 Turkey: Export Miracle or Accounting Trick? Ismail Arslan and Sweder van Wijnbergen Was Turkey's recovery from debt crisis an export miracle, as some contend? A consequence of its proximity to the Middle East? Or just an accounting trick - the result of a shift from underinvoicing to overinvoicing? Or a response to sound export incentives and exchange rate policy? The Policy, Research. and Extenal Affairs Cmnplex disuibutes PRE Working Papers to disseminate the findings of work in progress and to alcourage the exchange of ideas among Bank slaffand all others iieeated in development issues. These papes cany the names of the authors. reflect only their views, and should be used and cited accordingly. The findings, interpretatiom and conclusions are the authors' own. They should not be attributed to the World Bank. its Board of Directors, its managemnat, or any of its member countries.  Policy, Research, and Extemal Affairs Contry Operations This paper is a product of the Country Operations Division, Latin America and the Caribbean Regional Office, Country Department II. Copies are available free from the World Bank, 1818 H Street NW, Washington DC 20433. Please contact Margaret Stroude, room 18-163, extension 38831 (15 pages with figures). Alone among major debtor countries, Turkey being overpromoted? What pulled Turkey out substantially lowered its debt-export ratio - by of debt? more than a third between 1980 and 1987. But the driving force behind the Turkish export Using an econometric model, Arslan and miracle - indeed, its very existence - have van Wijnbergen assessed the contributions of been a matter of debate. various factors to the Turkish export miraclc, whose existence they confirm. Some contend that Turkey's export boom had little to do with export incentives or ex- Surprisingly, they leam that import growth change rate policy but was mostly a conse- in the Middle East made a negative contribution quence of Turkey's proximity to the Middle to Turkey's 1980-87 export boom. And al- East. Others claim that export growth reflected though there was a switch from underinvoicing a shift from underinvoicing before 1980 toward to overinvoicing, this accounting trick was overinvoicing afterward - a product of ac- nowhere near enough to explain the export counting tricks in response to changing incen- miracle. tives to be truthful about export receipts. After extraneous factors such as the oil If what happened to Turkey is spillover from boom in the Middle East are accounted for, its proximity to the Middle East, there is little Turkey's export miracle was more than a other countries can learn from the experience. response to explicit export incentives. It was a And if export subsidies were behind Turkey's result of macroeconomic policies and trade export growth, are GATT and free trade maybe reform that allowed a steady real depreciation of the Turkish lira. 'Pe PRE Working Paper Series disseminates the findings of work under way in the Banr s Policy, Research. and Extemal Affairs Complex. An objective of the series is to get these fminngs out quickly, even if presentations are less than fully polished. The findings, interpretations, and conclusions in these papers do not necessarily represent of ficial Bank policy. Prodniced at the PRE Dissemination Center  Turkey: Export Miracle or Accounting Trick? by Ismail Arslan and Sweder van Wijnbergen Table of Contents 1 Introduction I 2 Export Incentives and Export Performance 2 2.1 Export Development in Turkey 2 2.2 Export Incentives 4 3 Structure of the Model and Estimation Results 6 3.1 Model Structure 6 3.2 Econometric Results 8 4 What was Behind the Turkey Export Miracle? 10 4.1 Export Subsidies, the Real Exchange Rate, and 10 External Competitiveness 4.2 The Sources of Turkey's Export Growth 12 5 Conclusions 14 References 14 The first author thanks the Turkey and Pakistan Department in the World Bank and the UNDP for financing a visit to the Bank during which this paper was written. We are indebted to Bela Balassa for extensive comments on an earlier draft.  1 Introduction Turkey's recovery from its debt crisis in 78-80 has ever since made it the paragon of export led growth. In striking contrast to the LAtln American experience after 1982, Turkish growth recovered rapidly, with exports leading the way. Alone among the major debtor countries, Turkey substantially lowered its debt export ratio, by more than a third between 1980 and 1987. Real export growth rates were close to those obtained in for example South Korea in the late sixties and seventies. But the driving force behind the Turkish export miracle, and in fact its very existence, have remained a matter of debate. Some have pointed to Turkey's liberal provision of export incentives. Others have focused on the set of macroeconomic and import liberalization policies that caused Turkey's agressive nominal exchange rate policy to actually produce sustained real depreciation (Anand et alii (1988)). A much more skeptical view has been expressed by Celasun and Rodrik (1987) and Rodrik (1988). They argue that not more than 30% of the increase in exports can be traced to the real depreciation of the TL, and find little empirical support for any effect of the export incentives. They conclude that Turkey's export boom had little to do with the incentive regime or exchange rate policy, but was mostly a consequence of Turkey's proximity to the Middle East. The fact that recorded data show rapid export growth to the OECD countries too is, they argue, a consequence of a shift from underinvoicing before 1980 towards overinvoicing after 1980. This shift in invoicing strategy was, they agree, a response to exchange rate and export incentive policies, as black market premia fell and export subsidies rose. Thus these policies produced, to some extent, a chimera of export growth, a shift from unofficial to official exports, with no real underlying increases in excess of what could simply be explained by the high growth rate of imperts into Iran and Iraq. The importance of this issue goes much beyond the merely academic. There is a persistent belief among some observers (for example Balassa (1978,1985)) that agressive export policies promote growth, a practitioner's view that has recently begun to gather theoretical support (Romer (1989)). Moreover, almost every study of creditworthiness has shown the importance of debt export ratios in creditors' assessment of sovereign risk (McFadden et alii (1988)). But if exports play such an important role, the policies that lead to high export growth are of substantial interest. If all that happened in Turkey was a favorable spillover of its proximity to the Middle East, there is little other countries can learn from the Turkish experience. If on the other hand, active export subsidies were the main determinants of Turkey's export growth, one should wonder about the costs of joining GATT, and whether the static efficiency gains of free trade have not been promoted too much by its adherents (World Bank and IMF, to mention some). The third explanation, real exchange rate depreciation, would put the focus muchl more on the macroeconomic policies and trade reforms that made such a depreciation, where actively pursued through nominal exchange rate policy, sustainable in real terms. Thus the rationale for what we pursue In this paper; an assessment of the relative contributions of all the factors mentioned sofar to Turkey's export growth. Our efefort starts with an assessment of whether there was, in fact, an export miracle once the Middle East is discounted; or was it all a product of accounting tricks in response to changing incentives to be truthful about the true extent of export receipts. In Section 2 we compare Turkey's  2 trade statistics with those reported by its main trading partners to assess whether there was in fact significant growth to countries outside the Middle East. We then present a simple model focusing on the role of export incentives, relative prices and foreign income growth in export determination. This model is estimated econometrically and then used for an assessment of the contributions of the various factors mentioned to the Turkish export miracle, a miracle, by the way, the existence of which we do confirm. 2 ExRort incentives and Export Performance 2.1 Export Developments in Turkey. Since 1980, Turkey's trade strategy has shifted from interventionist import substitution towards a more market-based outward orientation. Export growth over the same period has been impressive. In volume terms growth averaged 24 percent per annum. As overall exports expanded, the structure of exports changed drastically too, both in terms of commodity composition, and markets reached. Industrial exports grew spectacularly during 1980-1987, from $ 1.0 billion in 1980 to $ 8.1 billion in 1987. Industrial exports, which in 1980 represented 36 percent of total Turkish exports, had more than doubled their share by 1987. But agricultural exports remained at the same level in value, with their share in total exports falling from 57 percent to 18.2 percent. The pattern of exports markets has also changed very rapidly, with the Middle Eastern Countries taking an increasing share (cf figure 1). Exports to Oil-exporting Countries (RXO in Fig.l) jumped from $ 0.4 billion ( 13.8 percent of total exports) to $ 2.9 billion ( 36.4 percent of total exports) in 1985. Iran and Iraq have become Turkey's major export markets, mainly because of Turkish exporters easy access to these markets, although supply factors too must have played a role. 1/ However since then exports to oil-exporting countries have stagnated, as falling oil prices forced these countries to curtail their expenditure (cf Fig. 1). However Turkish export growth was much more than a spill-over from the Middle East, triggered by buoyant OPEC expenditure. Turkish exports to other countries, mostly the major OECD Countries, grew at an equally impressive rate of 17.5 percent per year on average (again, in volume terms). Exports to the OECD Countries expanded in both labor-intensive industries (such as textile and apparel) and specialized products (mostly specialized industrial and electrical machinery). Low value added exports consisting mostly of raw materials (such as processed food), diminished in importance. Exports of textiles, leather products, chemicals, machinery, electrical equipment expanded at a higher growth rate than the average of total exports. Textiles and leather products remain the largest exporting industries, while the iron and steel industry has emerged as the second most important export industry. 1 The economic boom in the Middle East started well before 1980.  3 12500- 10000 7500 7foOO -~~~~~~~~~~~~~~~~~~~~~~~~~~~~, ,,/ 50001 <t,^