Sri Lanka: Selected Issues
Author:
International Monetary Fund
Search for other papers by International Monetary Fund in
Current site
Google Scholar
Close

This Selected Issues paper on Sri Lanka provides background information on economic developments and on selected policy issues facing Sri Lanka. The main economic developments in 1996 and the first quarter of 1997 are discussed. The paper highlights that in 1996, a severe drought, power shortages, and an escalation in the military conflict contributed to a sharp deterioration in the economic situation. With the end of the drought and power shortages, and a rise in investor confidence, macroeconomic conditions in 1997 were more favorable.

Abstract

This Selected Issues paper on Sri Lanka provides background information on economic developments and on selected policy issues facing Sri Lanka. The main economic developments in 1996 and the first quarter of 1997 are discussed. The paper highlights that in 1996, a severe drought, power shortages, and an escalation in the military conflict contributed to a sharp deterioration in the economic situation. With the end of the drought and power shortages, and a rise in investor confidence, macroeconomic conditions in 1997 were more favorable.

V. Exchange Rate and Trade Policy Effects on Trade Performance58

A. Introduction

182. Sri Lanka started a major trade and economic liberalization program in 1977 aimed at increasing its economic growth rate through diversification of its export base. Although a large number of trade and exchange restrictions were removed at the outset, the further reduction of tariffs and removal of export taxes has been spread over the two intervening decades, and this process is still far from complete in the case of tariffs. Section B of this chapter describes the trade liberalization over these 20 years. Section C examines the changes in trade patterns and factors that might have contributed to those trade patterns. Section D estimates the impact of Sri Lankan and foreign economic activity, real exchange rates, and trade liberalization on Sri Lanka’s exports and imports. The priorities for reform are discussed in the concluding section (Section E). The following conclusions emerge.

  • The strong export performance of Sri Lanka’s non-traditional exporters is highly dependent on the liberalization of its import trade regime.

  • Both imports and exports are substantially constrained by Sri Lanka’s tariffs; further liberalization of import tariffs would result in substantial increases in Sri Lanka’s exports.

  • The estimates suggest that an increase in the real exchange rate is associated with a decline not only in Sri Lanka’s real non-traditional exports, but also in its real imports. The latter may reflect the high import content of Sri Lanka’s non-traditional exports.

  • Based on these results, the key policy priorities for the future are: (i) a phased move from the present tariff system of three non-zero bands with a maximum rate of 35 percent, to a system of two non-zero bands with a maximum rate of 15 percent in the next few years; and (ii) the integration of the Board of Investment (BOI) tariff regime with the general tariff regime, which would involve the phased elimination of tariffs in the general regime on products where the BOI has effectively a zero tariff. Such steps would lead to substantial trade liberalization, further efficiency gains, and promote diversification of the still-narrow export base.

B. Trade and Exchange Rate Policy Since 1977

183. Sri Lanka had pursued an aggressive policy of import substitution and state control of trade beginning in the late 1950s. A single attempt at liberalization in 1968 was fully reversed within two years. By 1977, Sri Lanka’s highly restrictive trade policy and exchange regime was characterized by quantitative restrictions on all imports, export licenses for most exports, and multiple exchange rates backed up by extensive exchange controls (IMF, 1976; Papageorgiou et al., 1981).

184. Following elections in July, an ambitious economic reform program was launched in November 1977. The trade elements of the liberalization program involved the elimination of most quantitative restrictions on imports and exports, and the replacement of these measures by tariffs on imports and, in particular, taxes on exports as the main instruments of trade policy. The liberalization was also accompanied by a large devaluation, a liberalization of exchange restrictions, and a shift to a managed float of the exchange rate.

185. The liberalization was substantial, but nevertheless left significant trade barriers and exchange restrictions in place. The Sri Lankan government was unwilling to reduce the remaining tariffs and export taxes for reasons relating to both fiscal revenue they generated and the effects of import liberalization on import-competing producers. Nevertheless, the authorities were concerned that the remaining restrictions would unduly discourage the private investment needed to diversify and expand the export base. Therefore, in tandem with the liberalization of the general trade regime, the Sri Lankan authorities launched various export promotion programs. The most important of these were free trade/export processing zone programs; eventually consolidated as the Board of Investment (BOI) programs. These BOI programs (described below) place participating foreign and domestic exporters effectively outside the customs territory of Sri Lanka in order to insulate them from the anti-export bias effects of the remaining trade restrictions (Papageorgiou et al., 1981; Bruton et al., 1992).

186. For five years after the 1977 liberalization, the trade reforms were maintained. There was no increase in the average level of import taxes and export taxes were steadily reduced. Subsequently, between 1982 and 1987, export taxes continued to fall while a series of reversals in the import liberalization took place as the Sri Lankan authorities sought to promote exports at the expense of imports. Increased quantitative restrictions were generally avoided and there was progress in some elements of trade reform—for example, the maximum tariff was reduced from 100 percent to 60 percent in 1985 (WTO, 1995). Nevertheless the import trade regime became more restrictive on balance. Adjustments were made to the tariff schedule in order to restore or increase the effective rates of protection available to individual import competing industries, and a surcharge was imposed on imports to fund export promotion programs. By the late 1980s, the average tariff on a collections basis (i.e., the ratio of import tax revenues to imports) had nearly doubled from its levels of the early 1980s (Weerakoon and White, 1995).

187. Trade liberalization resumed in the 1990s. Maximum tariffs were reduced from 60 percent to 50 percent in 1992, 45 percent in 1993, and most recently to 35 percent in 1995; some tariffs not affected by the reductions in the maximum were moved to lower tariff bands. The average tariff on a collections basis was cut by more than half between 1989 and 1996, both because of the reductions in the statutory tariff rates and because of the continued expansion of the Board of Investment programs. Almost all of the remaining export taxes and most of the remaining import licensing requirements were removed in this period as well. Finally, the import surcharge that funded the export promotion programs remained in place but became largely irrelevant. The surcharge only applied to tariffs at or above 45 percent, and the February 1995 reduction of the maximum tariff to 35 percent eliminated the surcharge on all but a handful of exceptional items remaining outside the system of three non-zero tariff bands.

188. The progression of readily measurable aspects of Sri Lanka’s trade regime over these two decades can be seen in Charts V.1 and V.2. The evolution of export taxes on traditional exports can be seen in Chart V.1. Export taxes applied almost exclusively to the traditional exports (tea, rubber, coconut products, and gems) which accounted for the bulk of Sri Lanka’s exports in 1977. Of the export taxes, those on tea were the most important, accounting for over half of the export tax revenue in all years from 1978 through 1990, by which time all export taxes were well on their way to elimination. The tax on tea was steadily reduced from 46 percent in 1978 to 11 percent by 1986, before finally being eliminated in 1994.59 Export taxes on other traditional exports followed a similar pattern, falling from an average of 41 percent in 1978 to 11 percent in 1986, before being eliminated in 1994.60

CHART V.1
CHART V.1

SRI LANKA: EXPORT TAX RATES, 1977 - 96

(In percent)

Citation: IMF Staff Country Reports 1997, 095; 10.5089/9781451823370.002.A005

Source: Data provided by the Sri Lanka authorities; and staff estimates.
CHART V.2
CHART V.2

SRI LANKA: AVERAGE TARIFFS AND EXPORT TAXES, 1977 - 96

(In percent)

Citation: IMF Staff Country Reports 1997, 095; 10.5089/9781451823370.002.A005

Source: Data provided by the Sri Lanka authorities; and staff estimates.

189. The major trade policy instrument for import protection after 1977 was the tariff schedule and related import taxes. These import tariffs followed a more circuitous route toward liberalization (Chart V.2).61 They were largely unaffected by the 1977 liberalization, at least in the aggregate, and remained in the 8 to 10 percent range through the 1977–82 period. The liberalization of imports was then reversed, with average tariffs on a collections basis rising to 20 percent in 1987 before falling under 10 percent by 1994.

190. Consistent data on the extent of quantitative restrictions and other nontariff barriers are not available for this period. Although quantitative restrictions ceased to be a major feature of Sri Lanka’s trade regime after 1977, they did not disappear. In addition to import licensing for health, security, and environmental reasons, restrictive import licensing on other products, particularly food products, persisted for many years. Several basic food commodities (chilies, onions, and potatoes) remained under restrictive import licensing until mid-1996, and wheat, wheat flour, and maize remain under license (although these last remaining licensing requirements appear to have little or no trade-restricting effect).

191. Much of the liberalization of the import regime reflected in the decline in import tax rates, particularly in the early 1990s, was attributable to the expansion of the BOI free trade zone and export processing zone programs rather than reductions in tariffs in the general trade regime. The BOI programs effectively exempt participating exporters of non-traditional goods from the Sri Lankan trade regime. Enterprises operating from the zones also benefit from tax holidays, greater access to foreign borrowing, greater ability to repatriate profits and capital, and better infrastructure. Nevertheless, the main benefit of BOI status was that these enterprises were effectively outside the customs territory of Sri Lanka. Imports of raw materials, other inputs, or capital goods by BOI enterprises are not subject to tariffs; the only tariffs paid by enterprises in the zone were on their sales “exported” into the local Sri Lankan market.

192. The BOI programs are no longer restricted to free trade zones; similar policies for BOI exporters were subsequently added to cover exporters of non-traditional exports regardless of whether they were located within the physical free trade zones. Additional BOI programs have been added over the years that provided BOI status to exporters using advanced technology, investors in large projects, “exporters” of services such as tourism, and a wide variety of other Sri Lankan investment priorities.

193. By the early 1990s, BOI imports and exports were only slightly smaller than exports and imports in the general trade regime. Although the predecessors of the BOI program had been in place since 1978, it became a central part of the trade regime only after 1990. Between 1990 and 1993, the share of BOI enterprise exports in total exports rose from 22 percent to more than 50 percent (Chart V.3). Imports by BOI enterprises increased by smaller, but still substantial, amounts from 17 percent of total imports in 1991 (the first year for which data are available) to 25 percent in 1996. These figures probably understate the importance of export promotion programs for the import regime; a variety of “BOI-like” policies for non-BOI exporters (i.e., reduction or elimination of tariffs on imported inputs and capital goods) were introduced or expanded in the 1990s. Because of these programs, important sectors of the economy had become dependent on BOI trade policies. An example of the importance of the BOI regime to a major export industry is provided by garment manufacturing, which accounts for half of Sri Lanka’s exports. This industry is heavily dependent on both imported fabric and the BOI trade policies; because of the heavy weight of the BOI program, the average tariff for fabric on a collections basis is only 1 percent in spite of a tariff rate of 35 percent in the general tariff regime.

CHART V.3
CHART V.3

SRI LANKA: BOI TRADE SHARES, 1978–96

(In percent)

Citation: IMF Staff Country Reports 1997, 095; 10.5089/9781451823370.002.A005

Source: Data provided by the Sri Lanka authorities.

C. Trade Patterns, 1977–97

194. The effects of the shift in 1977 from a strong import substitution bias stance toward what became a relatively strong outward orientation were quickly evident. Imports grew rapidly in real value and more than doubled their GDP share between 1977 and 1978, reflecting both the liberalization of the trade regime and the upward revaluation of trade value due to the exchange rate depreciation. Import growth continued after 1978, peaking at more than 50 percent of GDP in 1980 (triple the import-to-GDP ratio of 1977), fueled by both the import liberalization and a large public investment program funded by foreign grants and concessional loans. Exports rose sharply in 1978 relative to GDP reflecting both the depreciation and an unusually good year for traditional exports (Charts V.4 and V.5).62 However, the export-to-GDP ratio remained steady until 1981 and then trended downward until 1986 along with the real appreciation of the exchange rate and the increase in import protection. By 1986, Sri Lanka’s export to GDP ratio had fallen to a level of 19 percent of GDP, only slightly above the 18 percent level of 1977.

CHART V.4
CHART V.4

SRI LANKA: REAL EXPORTS, 1977–96

(In millions of 1990 rupees)

Citation: IMF Staff Country Reports 1997, 095; 10.5089/9781451823370.002.A005

Source: Data provided by the Sri Lanka authorities.
CHART V.5
CHART V.5

SRI LANKA: TRADE TO GDP RATIOS, 1977–96

(In percent)

Citation: IMF Staff Country Reports 1997, 095; 10.5089/9781451823370.002.A005

Source: Data provided by the Sri Lanka authorities.

195. The composition of Sri Lanka’s exports casts a different light on the poor performance of overall exports in the first decade after liberalization. The traditional exports that accounted for 77 percent of Sri Lanka’s exports in 1977 consisted of four product groups: tea, rubber, coconut, and gems. All four of these product groups are commodities for which Sri Lanka is largely a price taker and three of the four are tree crops where production is largely driven by planting decisions taken years earlier and land and climate constraints (for the fourth, gems, production is driven in large part by geology). With this undiversified and inelastic export base as a starting point, the initial positive effects of the 1977 liberalization on Sri Lanka’s export base were manifested in the changing composition of exports during the first decade after liberalization. Non-traditional exports rose from 23 percent to 56 percent of exports (or from 4 percent to 11 percent of GDP) between 1977 and 1986, offsetting much of the real decline in traditional export earnings. Almost all of this increase occurred between 1977 and 1982, when non-traditional exports reached to 49 percent of total exports, lending support to the idea that the increase in import protection subsequent to 1982 had a pronounced negative effect on exports. In a related development, the BOI enterprises had their share of exports grow from 1 percent in 1979 to 15 percent by 1986. In contrast to non-traditional exports as a whole, the BOI share rose as rapidly between 1982 and 1986 as before, perhaps suggesting that exporters responded to the increase in protection on imported inputs by moving into the BOI program to escape the deleterious effects of the trade regime on their competitiveness.

196. The role of real exchange rates in determining exports is difficult to disentangle from other factors influencing trade patterns in the 1977–86 period (Chart V.6). The liberalization of quantitative restrictions in 1977 should have resulted in a real depreciation even absent other factors, but the exchange rate may have been overvalued for other reasons as well. For subsequent years, the increase in capital inflows after the liberalization (in the form official development assistance) may have contributed to the real appreciation observed in the 1978–82 period. Thereafter, in the 1982–86 period, the real effective exchange rate appreciates by some measures and remains constant by others. The real appreciation may have been associated with the increases in import tariffs over that period, with policy changes unrelated to trade policy, or non-policy factors.

CHART V.6
CHART V.6

SRI LANKA: REAL EXCHANGE RATE MEASURES, 1977–96

(Index 1990=100)

Citation: IMF Staff Country Reports 1997, 095; 10.5089/9781451823370.002.A005

Source: Data provided by the Sri Lanka authorities; and staff estimates.

197. The measures shown in Chart V.6 include the conventional real effective exchange rate index based on a consumer price index (CPI). However, because the Sri Lankan CPI is based on a dated basket that has an overly heavy weight on food items, two alternative measures were also examined: the conventional index with the Sri Lankan CPI replaced by the GDP deflator, and the conventional index with the Sri Lankan CPI replaced by a Sri Lankan nominal minimum wage index (in all cases partner countries prices were measured by their consumer prices). For the 1977–86 period, the there is some distinction, but they generally move together.

198. From 1987 onward, both imports and exports resumed a steady upward trend, both in absolute terms and relative to GDP. Imports rose from 30 percent of GDP in 1986 to approximately 40 percent of GDP over the 1993–96 period. Exports rose from 19 percent of GDP in 1986 to 29 percent of GDP in 1996. Real exchange rate movements do not seem to have been an important factor in the 1987–96 period; in contrast to the first decade after liberalization, real exchange rate showed little variation by any measure. Toward the end of this period (1993–96), growth in all measures of openness slows down; the ratio of both imports and exports to GDP flattens out in 1993, as does the share of BOI exports in total exports. One possible explanation for this slowdown in trade growth is the reduced impulse to trade growth from trade liberalization. Tariff reductions slowed in this period, and the reduction of export taxes could proceed little further once the main export taxes had been eliminated in 1993.

D. The Impact of Economic Activity, Real Exchange Rates, and Trade Policy on Sri Lanka’s Trade

199. Statistical estimation of the influence of Sri Lankan and foreign income, real exchange rates, and trade policies on Sri Lanka’s trade patterns is constrained by a number of data problems. First, the heavily reliance on non-price controls in the trade and exchange regime through 1977 make data collected before 1978 difficult to interpret. Second, many key variables, including GDP and GDP deflators for Sri Lanka and some of its trade partners are available only on an annual basis. The combination of annual data and 19 years of usable observations severely constrains the choice of estimation methods. Finally, there are few indices of import or export prices that can be used reliably to capture the price movements of even traditional exports; these lacunae in data availability foreclose the use of estimation of a system of separate import and export demand and supply equations along the lines suggested by Goldstein and Khan (1985) and instead suggest their fall back recommendation of a reduced form equation when demand and supply side variables are combined.

200. The estimates below attempt to measure the effects of the various income, exchange rate, and trade policy influences on the real imports and exports of Sri Lanka. Because preliminary analysis could not reject that the any of variables other than export taxes were integrated of order one (apart from export tax rates), an error correction model approach was used to estimate these relationships. Unit root test rejected higher order integration for all dependent and independent variable in the reported estimates. The estimates were calculated both with and without policy variables. First, the impact of foreign and domestic real income on Sri Lanka’s import and export flows are estimated following the three-stage procedure used for the Asia-Pacific Economic Cooperation countries by Bayoumi (1996), to which the results are compared. These estimates make use of the method developed by Engle and Yoo and employed by Bayoumi, which is parsimonious in its data requirements and has corrects the non-normality of the statistical distributions for the long-run relationship in the Engle-Granger two-stage procedure.63 Next, policy variables are introduced into the estimation to determine the additional influence of tariffs and export taxes. For imports, this involves re-estimating the import demand equation with additional policy variables. For exports, which are divided into traditional and non-traditional exports, policy variables influencing both export demand and supply are included in reduced form regressions.

201. Import and export elasticities for real income and real exchange rate movements are presented in Table V.1. Imports and exports are deflated by the Sri Lankan GDP deflator, foreign real income in the export demand equation is an index of real GDP in partner countries constructed using the same weights as in the real effective exchange rate; and the real effective exchange rate is based on consumer prices in partner countries, but uses the GDP deflators a measure of Sri Lankan prices.64 All variables are annual data for the period 1978–96 and all except the trade tax ratios are converted to natural logarithms for estimation.

Table V.1.

Import and Export Demand

article image
One and two asterisks denote statistical significance at five and one percent levels respectively.

202. The estimates for income and real exchange rate elasticities in the table above show typical patterns for such regressions. The income elasticities are of the expected sign in both the short and long run and are either statistically significant at the one percent level or just short of significance at the five percent level. This is consistent with the results presented by Bayoumi for the APEC countries; all of the long-run income elasticities in that study were positive and most were statistically significant at the 5 percent level.65

203. The real exchange rate elasticities present a more mixed picture. The coefficients on the real exchange rate in the export demand equation are of the expected sign but fall short of the usual statistical significance thresholds. More surprising is the coefficient on the real exchange rate in the long-run import demand equation, which is both negative and significant, a result not obtained for any of the countries examined by Bayoumi. This unusual pattern may reflect in part the heavy derived demand for imports by Sri Lankan exporters; under this interpretation, an increase in the real exchange rate suppresses demand for exports, and therefore imported inputs for those exports, and this negative impact is greater than the increase in imports for domestic final consumption. Although, even under this interpretation, one would expect the coefficient on the real exchange rate in the import equation to be smaller in absolute value than the corresponding term in the export equation.

204. The introduction of trade policy variables starts with the addition of a term for the import tariff to the import demand equation.66 Estimates from this equation are presented in Table V.2. The coefficient on the import tariff term itself is highly significant and negative both in the long-run and short-run equations. This term indicates that a one percentage point increase in the tariff level results in more than a three percent drop in real imports. Introduction of the tariff variable results in little change in coefficients other than the long-run real exchange rate coefficient and the error correction term. The coefficient on the real exchange rate continues to be negative and significant, but it by drops almost two thirds in absolute value (although it retains its high level of statistical significance due to the improved fit), while the coefficient on the error correction term more than triples in absolute value, suggesting a very rapid short-run convergence toward the long-run equilibrium.

Table V.2.

Real Import Demand

article image
One and two asterisks denote statistical significance at five and one percent levels respectively.

205. Rather than affecting the demand for Sri Lanka’s exports, trade policy variables are more relevant for Sri Lanka’s export supply. First, because Sri Lanka is largely a price taker in the markets for its traditional exports, the incidence of export taxes on those traditional exports is likely to fall primarily on export suppliers rather than foreign buyers. Second, the Sri Lanka government followed distinctly different trade policies for traditional and non-traditional exports, taxing the former heavily throughout much of the sample period and subsidizing the latter. Third, the influence of trade taxes on other import-competing or export industries may influence exports through the export supply relationship (e.g., through labor demand in those industries) rather than the export demand relationship. For these reasons, the reduced forms for traditional and non-traditional exports are estimated separately.

206. As noted above, one approach to estimating the impact of trade policies on exports would be to estimate a system of export supply and demand equations along the lines suggested in Goldstein and Khan (1985) in which the trade policy variables could be assigned to either the demand or supply equation as appropriate. Such a system could be represented by the pair of equations below.67 Estimation of this system would require data on prices for non-traditional exports (the only common variable to the two equations apart from the level of real exports itself). However, in the Sri Lankan case, good data on price levels for non-traditional exports are not available; existing export price and volume indices tend to capture changes in the composition of exports in the price index rather than export volume, and this problem has become more severe with the diversification of exports.68

X n t d = α 0 + α 1 * Y f + α 2 * R + α 3 * P n t
X n t s = β 0 + β 1 * P n t + β 2 * τ e + β 3 * τ i

207. As an alternative to estimating separate demand and supply equations, following McDermott (1997), the reduced form equations can be estimated to capture both demand-and supply-side influences. This is done by solving the export supply equation for prices and substituting it into the export demand equation. These reduced form estimates complicate the interpretation of the coefficients; instead of being straightforward demand and supply elasticities, they now combine the parameters of the supply and demand equations into the equation shown below. This general form is used for single-equation, reduced-form estimates for both traditional and non-traditional exports. This approach precludes the estimation of price elasticities, but such estimates would likely be biased toward zero because of the measurement problems associated with those price indices (Goldstein and Khan, 1986).

X = γ0 + γ1*Yf + γ2*R + γ3e + γ4i

Where γ = δ*(α0 + α301), γ1 = δ*α1 γ2 = δ*α2 γ3 = δ*α321 γ4 = δ*α331 δ = β1/(β1 - α3)

208. The estimates for the reduced form equation for non-traditional exports are shown in Table V.3. Several trade policy variables were introduced into this equation, but the only one with significant explanatory power was the average import tariff, which was consistently significant and negative. Foreign real income continues to be highly significant and there is little effect on the value of the real exchange rate coefficient from the addition of the import tariff variable. This negative effect on exports from the import tariff might be the result of general equilibrium effects such as the anti-export bias introduced to the economy by protection. Also, in the Sri Lankan case, it may reflect the direct effects on competitiveness of import tariffs on imported inputs.

Table V.3.

Non-Traditional Exports - Reduced Form

article image
One and two asterisks denote statistical significance at five and one percent levels respectively.

209. Moving to traditional exports, Table V.4 show the reduced form equation for traditional exports estimating the effects of real income and real exchange rate variables as well as effects of export taxes. Specifications with other policy variables, supply variables (e.g., agricultural wages, fertilizer use, and rainfall in tea plantation districts), and tea alone as a dependent variable were also estimated, but both real income and real exchange rates tended to have either insignificant effects or perverse signs in these alternative long-run estimates. This is not surprising when the composition of traditional exports is taken into account. As noted above, the supply of each of these commodity groups is highly inelastic, and prices of plantation crops are heavily affected by weather in other producing countries, a factor not captured in this estimation. The specification shown in Table V.4 is the “best” only in the sense of not having perverse or implausible coefficients on real income and the real exchange rate.

Table V.4.

Traditional Exports - Reduced Form

article image
One and two asterisks denote statistical significance at five and one percent levels respectively.

210. Surprisingly, export taxes have a significant and positive short-run relationship to traditional exports. The straightforward interpretation of this relationship, that export taxes encourage the production of traditional exports, is implausible at best. A more credible explanation of the positive relationship between export taxes and export value is suggested by various issues of the Central Bank of Sri Lanka’s Review of the Economy and Annual Report in which it is made apparent that export taxes are set after it becomes clear what world prices are for traditional exports, and that these taxes are set in response to those prices. For example, when it became clear whether the price of tea or rubber will be particularly good (or bad), export taxes were adjusted upward (or downward) to capture part of the windfall (or cushion the drop).

E. Conclusions and Priorities for Reform

211. The evidence of the last twenty years points to a close positive linkage between Sri Lanka’s trade liberalization and improved export performance. Even apart from any general equilibrium factors not captured in the estimates above, reductions in import tariffs yield increases in exports that are comparable in scale to the increases in import growth induced by tariff reductions. This liberalization has had two major sources: the removal of non-tariff restrictions and the reduction of tariffs and export taxes in the general trade regime, and the expansion of coverage of the BOI trade provisions among Sri Lanka’s exporters. Both aspect of the trade liberalization to date need to be taken into account in planning the further liberalization of Sri Lanka’s trade.

212. Although export taxes and quantitative restrictions have been largely eliminated, the general trade regime still retains high tariffs and high rates of effective protection. Absent the BOI program, average tariff rates would be on the order of 20 percent and the structure of these tariffs has been consciously arranged to maintain high levels of effective protection for import competing industries. For consumers, these tariffs substantially increase the costs of imported and import-competing items. For producers outside the BOI program, these high tariffs reduce competitiveness through the higher cost of imported inputs and through the lack of competitive pressure from foreign producers. For producers inside the BOI program, the lack of competitiveness creates disincentives to use Sri Lanka inputs in production and the high tariffs discourage sales into the domestic market. The reduction of each of these distortions through a substantial reduction in the average and maximum tariffs rates would yield substantial benefits to the Sri Lankan economy. To reduce adjustment costs and allow time for replacement of the fiscal revenue, such a reduction in rates could proceed with the staged reduction of the current maximum tariff of 35 percent down to 15 percent in the next few years with a commensurate reduction in the average tariff. At the same time, to further simplify the tariff system and reduce dispersion, the number of non-zero bands could be reduced from three to two.

213. Sri Lanka has benefitted from the BOI program through a much more rapid expansion and diversification of its export base than would have been possible with its general trade regime alone. Nevertheless, this form of export expansion and diversification has its limitations; the BOI program has attracted industries that are dependent on imported inputs and able to relocate quickly in response to changing labor costs or new opportunities elsewhere. The phase-out of the BOI program would eliminate such distortions and reintegrate the BOI enterprises into the Sri Lankan economy. This would provide substantial benefits to the Sri Lankan economy, but only if it could be done in a manner that did not increase protection on inputs used by the BOI enterprises. Such a reintegration of the BOI would proceed through lowering tariffs in the general trade regime to the (zero) levels of the BOI regime for inputs where the BOI program has resulted in a de facto zero tariff rate for most of the export industry. This form of reintegration would do much to maintain competitiveness in existing export industries as well as improve competitiveness and increase economic efficiency in the economy more generally. It would also allow Sri Lankan consumers and non-BOI producers to benefit directly from the availability of access to the products used or produced by BOI enterprises at world market prices.

214. The reforms suggested above would lead to substantial trade liberalization, further efficiency gains, and promote diversification of the still-narrow export base. In particular, the lowering of tariff barriers in the general trade regime could increase consumer welfare and export competitiveness outside the BOI sector while the reintegration of the BOI sector into the general economy would contribute to both export diversification and increase local value added in exports.

References

  • Bayoumi, Tamim, 1996, “International Trade and Real Exchange Rates” in Takatoshi Ito, Peter Isard, Steven Symansky, and Tamim Bayoumi, 1996 Exchange Rate Movements and Their Impact on Trade and Investment in the APEC Region, Occasional Paper No. 145. (Washington, International Monetary Fund), December 1996.

    • Search Google Scholar
    • Export Citation
  • Board of Investment of Sri Lanka, Sri Lanka: Investment Policy and Incentives, Colombo, 1996.

  • Bruton, Henry J. and Associates, The Political Economy of Poverty, Equity, and Growth: Sri Lanka and Malaysia. (Oxford, Oxford University Press), 1992.

    • Search Google Scholar
    • Export Citation
  • Central Bank of Sri Lanka, Annual Report - various issues (Colombo).

  • Central Bank of Sri Lanka Review of the Economy - various issues (Colombo).

  • Cuthbertson, Keith, Steven G. Hall, and Mark P. Taylor, Applied Econometric Techniques (Ann Arbor: University of Michigan Press), 1992.

  • Engle, R. F. and C. W. J Granger, Cointegration and Error Correction: Representation, Estimation, and Testing. Econometrica, 55, pp. 25176, 1987.

    • Search Google Scholar
    • Export Citation
  • Goldstein, Morris and Mohsin S. Khan, 1985, “Income and Price Elasticities in Trade,” in Handbook of International Economics, Vol. 1, ed. By Ronald W. Jones and Peter B. Kenan (Amsterdam, North-Holland), 1985.

    • Search Google Scholar
    • Export Citation
  • Hendry, David F. and Jurgen A. Doornik, Empirical Econometric Modelling Using PcGive 9.0 for Windows (London: International Thomson Business Press), 1996.

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 27th Annual Report on Exchange Restrictions (Washington: International Monetary Fund), 1996.

  • McDermott, John.Export Performance” in International Monetary Fund, Indonesia - Selected Issues (SM/97/159), 1997

  • Papageorgiou, Dimitri, Michael Michaely, and Armeane M. Choksi, eds., Liberalizing Foreign Trade, 7 vols., (Cambridge, Massachusetts: Basil Blackwell, Inc.), 1991.

    • Search Google Scholar
    • Export Citation
  • Weerakoon, Dushni and Howard White, How Open Has the Sri Lankan Economy Become? Trends in Trade and Trade Taxes: 1977–93. Working Paper 205, Institute of Social Studies The Hague 1995.

    • Search Google Scholar
    • Export Citation
  • World Trade Organization, Trade Policy Review of Sri Lanka (Geneva), 1996.

58

Prepared by Thomas Dorsey.

59

There are omissions in the measurement of both export and import taxes due to problems with data availability. For export taxes, the data do not include a variety of mostly small, off-budget export taxes referred to as royalties and “cesses.” These range from trivial fees for conservation (the equivalent of US$0.18 per live elephant without tusks and US$8.50 per elephant with tusks) to much more substantial taxes on some minor exports (a 200 percent tax on the export of unshelled cashews). Some of these taxes remain in force.

60

The breakdown of export taxes for products other than tea is not available, although it is known that rubber, rubber products, coconuts, and coconut products accounted for almost all of the remaining export taxes. The calculations presented here assume all of the non-tea export taxes were applied to exports of other traditional exports.

61

The data referred to here capture most taxes on imports. However, other tax policy instruments, primarily discriminatory valuation of imported goods for domestic turnover and other indirect taxes, have been employed to tax imports in Sri Lanka over the 1977–97 period. There is an artificial markup of 25 percent in the valuation of imports (but not domestic goods) for domestic indirect taxes that is not captured in the data used here; this also remains in force. Data are not available on the additional tax revenue from these policy instruments.

62

Tea, which accounted for over 50 percent of export value in 1978, reached a real dollar price level in that year exceeded in only one of the eighteen years between 1979 and 1996.

63

The estimation method is that suggested by Engle and Yoo (1989) as described in Cuthbertson, Hall, and Taylor (1992). It begins with the standard two-step procedure suggested by Engle and Granger (1987) in which the parameter estimates for a long-run relationship (embodying the cointegrating vector) are used to estimate a second-stage dynamic relationship using an error-correction (or equilibrium compensating) mechanism. The Engle-Yoo procedure adds a third stage in which the residuals from the second stage regression are regressed on the independent variables of the original long-run regression with those variables pre-multiplied by the coefficient on the error correction term from the second stage regression with the sign reversed. The coefficients from this third stage regression are added to those from the first stage regression and the standard errors from the third stage regression, now cleared of the influences of unit roots and trends, are used in place of the standard errors from the first stage regressions. This procedures gives estimates of coefficients and standard errors with normal statistical distributions for which the usual test of statistical significance are valid.

64

The heavy weight on basic food commodities in the Sri Lanka consumer price index suggests that it might be less representative of Sri Lanka prices than the GDP deflator. Estimates of all of the equations presented in this chapter were also prepared using the real effective exchange rate index based on Sri Lankan consumer prices. These estimates, however, tended to have a poorer fit, lower real exchange rate coefficients, and higher standard errors than the GDP deflator-based estimates presented here.

65

The significance levels for t-statistics of individual coefficients are calculated for two-tail probability levels. Equation test statistics in this and other tables of this chapter are (from left to right) the R2 adjusted for degrees of freedom, the Durbin-Watson statistic, the equation F statistic for a zero slope, and the marginal probabilities of the x2 test for normality and the RESET test. For the last two test statistics, significant values suggest non-normality and mis-specification respectively. Test statistics for the long-run relationship are those from the first stage of the estimation. All test are calculated using the default settings of PcGive 9.0 for Windows (Hendry and Doornik, 1996) except for R2 where the unadjusted levels reported by PcGive are adjusted for degrees of freedom in the statistics shown here.

66

Other authors have used different combinations of trade policy variables. For example, McDermott (1997) used the sum of the ratio of import tariff revenues to total imports and the ratio of export tax revenues to total exports as an aggregate measure of trade distortions. Preliminary estimates of the import and export equations were conducted using this composite ratio, its two components separately, and the ratio of export taxes to traditional exports both separately and in combination.

67

The independent variables in the export demand equation for non-traditional exports are foreign real income (Yf), the real effective exchange rate (R), and the price of non-traditional exports (Pnt) respectively. The independent variables in the export supply equation for non-traditional exports are the price of non-traditional exports, average import tariffs (τi), and average export taxes on traditional exports (τe).

68

For example, volume indices for garments, Sri Lanka’s largest export industry, are based on piece counts that do not distinguish between high and low value exports in spite of a known shift in the composition of garment exports toward higher value products.

  • Collapse
  • Expand
Sri Lanka: Selected Issues
Author:
International Monetary Fund
  • CHART V.1

    SRI LANKA: EXPORT TAX RATES, 1977 - 96

    (In percent)

  • CHART V.2

    SRI LANKA: AVERAGE TARIFFS AND EXPORT TAXES, 1977 - 96

    (In percent)

  • CHART V.3

    SRI LANKA: BOI TRADE SHARES, 1978–96

    (In percent)

  • CHART V.4

    SRI LANKA: REAL EXPORTS, 1977–96

    (In millions of 1990 rupees)

  • CHART V.5

    SRI LANKA: TRADE TO GDP RATIOS, 1977–96

    (In percent)

  • CHART V.6

    SRI LANKA: REAL EXCHANGE RATE MEASURES, 1977–96

    (Index 1990=100)