Evaluating the Success of Trade Liberalization in Bangladesh
by John Williamson, Peterson Institute for International Economics
Paper for the meeting on "Bangladesh Trade Liberalization: Its Pace and Impacts"
September 16, 1999
© Peterson Institute for International Economics
This paper was written while Mr. Williamson was the Chief Economist for the South Asia Region at the World Bank.
Economic theory suggests that the primary impact of trade liberalization will be on the overall level of trade, with a roughly parallel increase in exports and imports as a percentage of GDP. This is likely to require a modest depreciation of the real exchange rate in order to ensure that exports increase as much as imports, and leave the balance of trade unchanged. The export increase is likely to be concentrated in a relatively narrow range of products, products which use intensively the country's abundant factor of production (unskilled labour in the case of Bangladesh), while the import growth is expected to be much more diversified and more capital intensive. Some import-substituting industries are likely to find themselves squeezed, but others will respond to the competitive challenge by modernization. The net effect will be to increase the growth rate, through various channels, including concentration on industries in which the country has a comparative advantage, possibly increasing returns to scale in some of the export industries, the exposure of import-competing industries to competition from imports, improved technology when the price of imported capital goods declines and the country can afford to import more, and the chance to use imported intermediate goods when these are more suitable than domestic inputs.
I believe that most foreign economists who have looked at Bangladesh's trade performance in the last 8 years have concluded that its trade liberalization since 1990 has been a textbook case of success. Exports have increased at double-digit rates, and imports have increased in parallel, leaving the trade balance largely unchanged in dollar terms. These exports have been heavily concentrated in the garment industry, which is an industry well-suited to Bangladesh's comparative advantage in view of its heavy use of abundant unskilled labour. GDP growth has accelerated. What more could anyone ask for, apart from more of the same? (Well, I suppose one could have asked for much more progress in establishing an enabling business environment, reforming the financial sector and state enterprises, and improving infrastructure, measures that would have been well worthwhile in themselves and would also have enabled Bangladesh to draw even greater benefits from trade liberalization.)
What has surprised many of us is that it seems that public opinion in Bangladesh, and even some of its leading economists, do not see things the same way. They have complained of a flood of imports having a devastating effect on the import-competing industries, and reducing the breadth of the country's industrial base. They complain of the trade deficit with India. They complain that the garment industry uses imported textiles, so that the value-added is much less than the gross value of exports. There are even complaints that smuggled imports from India are taking markets away from local industry, with an apparent belief that smuggling is stimulated by trade liberalization (a notion so counter to all existing theory that any economist who builds a convincing model demonstrating how it could happen would surely deserve the Nobel prize).
The study we are discussing today, Bangladesh Trade Liberalization: Its Pace and Impacts, provides the best basis we have yet for evaluating these charges. It documents the parallel increase in exports and imports, and shows that the share of exports in GDP has more than doubled in the 1990s (Table A.9), compared to a 50% increase in the 1980s. It has a pioneering and most illuminating analysis of the microeconomic impact on the import-competing industries, and concludes that these experienced much more dramatic productivity gains than the export industries, but that their fortunes were far more varied than those of the export industries. These findings seem plausible. Export industries had to be efficient in order to be capable of exporting when there was a 66% bias against exporting, so they had little scope for productivity improvement after liberalization. Import-competing industries, in contrast, had an easy life under the ancien regime. After liberalization, some succeeded in modernizing under the stimulus of competition, but some were out-competed by imports, as we know must happen if exports and imports are to expand in parallel. (Only in Lake Wobegon are all the children above average.)
In case this does not reassure those who worry about deindustrialization in Bangladesh, let us look at the statistics on growth of the manufacturing sector. Manufacturing output grew at 3.0% per year in the 1980s and at 6.6% per year in the 1990s. Even if one excludes garments, growth was about 4.4% in the 1990s, i.e. faster than in the pre-liberalization years.1 Admittedly the growth of manufacturing value-added, at 4.7% per year, was faster than the 3% growth of gross production in the 1980s, and if one guestimates2 that garments also contributed 2.2% to the 6.8% growth of manufacturing value-added in the 1990s, then the growth of value-added in the rest of manufacturing showed no acceleration in the 1990s. But neither did it decline, as the pessimists seem to believe. The accelerated growth of garments was a net addition to manufacturing growth, not something that displaced the rest of manufacturing.
It is of course true that the net exports of the garment industry are substantially less than the gross exports, and it is worth knowing that in net terms jute is still Bangladesh's biggest export industry. That fact surely makes the case for an enquiry into the future of the jute industry, which I recently heard Rehman Sobhan plead for. But it does not constitute an indictment of the garment industry, which remains an invaluable national asset. The test of its value to the nation is a comparison of the value added with the value of the resources employed to produce it, and the fact that the industry generates good profits and many jobs implies that the industry passes the test.
The charge that trade liberalization is responsible for the smuggling from India is, as I have already implied, absurd. That said, the report is right to point out that smuggling is also encouraged by the failure to collect VAT properly in Bangladesh. If retailers were charged VAT on the value of their sales, and entitled to a rebate on the value of their purchases if and only if those goods could be shown to have already had tax paid on them, the distorted incentive to smuggle so as to avoid VAT would be eliminated. The worry about a bilateral trade deficit with India is of course misplaced: there is nothing wrong with triangular trade, which is often welfare-enhancing. But it may be acting as surrogate for a perfectly legitimate complaint, which is that Bangladesh's big neigbour India retains a very restrictive import regime. That regime hurts India as well as Bangladesh; its existence provides no rationale for Bangladesh to hurt itself as well as India by reciprocating.
Nothing in the above challenges my tentative conclusion in the second paragraph that Bangladesh would be well advised to seek more of the same: more trade liberalization, leading to more trade, leading to faster growth. It would of course be better still if those reforms were supplemented by more reforms in the sectors where Bangladesh has lagged, like the financial sector and privatization and the civil service and infrastructure. But experience has shown that the country can benefit from trade liberalization even without these other reforms. Just think what it might accomplish if it had these other reforms as well!
1. Excluding jute, the increase would be about 0.25% per year faster in both the 1980s and the 1990s.
2. A guestimate is necessary because there are no statistics on the growth of value-added in garments.