In November 2004, Sri Lanka reversed its relatively open trade policies and now has one of the world’s most complex and protective import regimes. This has been done by deploying a variety of para-tariffs over and above customs duties, the significance of which appears to have escaped the notice of most of Sri Lanka’s trading partners, including India. Unless rapidly unwound, these new policies will seriously damage the country’s future economic growth. They also subvert Sri Lanka’s preferential trade agreements, and breach its World Trade Organisation commitments.

Garry Pursell (garry.pursell@gmail.com) is at the Australia South Asia Research Centre, Australian National University, Canberra.

I

These changes have been described and analysed for the first time in a recent paper (Pursell and Ahsan 2011). This article summarises some of the principal features of these developments. It concludes by pointing out the serious potential damage of Sri Lanka’s protectionism on its economic growth, and the resulting subversion of its preferential trade agreements (PTAs), especially the agreements with India (India Lanka Free Trade Agreement (ILFTA)), Pakistan (Pakistan Sri Lanka Free Trade Agreement (PSFTA)) and with the south Asian countries as a group (South Asia Free Trade Agreement (SAFTA)). It also comments

Economic & Political Weekly

A Complex Tariff Structure

Sri Lanka probably holds the world record for the complexity of its tariff system. In addition to customs duties (CDs), during 2009 and up to June 2010, the Sri Lankan tariff schedule included nine other import taxes that were imposed or potentially imposable on imports of individual products. By early January 2011 two of these additional import taxes had been removed, but that still left seven other import taxes.

Of the nine taxes during 2009 to June 2010, four (referred to hereafter as “paratariffs”) added to whatever protection is provided to domestic production by customs duties, because they were only applied to imports and there was no domestic equivalent. These were the Port and Airport Development Levy (PAL), the Customs Surcharge (SUR), the cess,1 and the Regional Infrastructure Development Levy (RIDL). The SUR and the RIDL were abolished after June 2010. In addition to the value added tax (VAT), there are three other import taxes for which there are domestic equivalents or approximately equivalent taxes, and which provisionally seem to be roughly neutral in terms of protectiveness. These are the Social Responsibility Levy (SRL), the Nation Building Tax (NBT), and excise duties. Finally, there is a “Special Commodity Levy” (SCL) on imports of a small number (22) of “essential” primary commodities. The SCL imposes specific duties which replace all the other import taxes (including customs duties and VAT) that would otherwise be imposed on imports of these products. The complexity of the system is further increased by the fact that each import tax has its own product coverage. Moreover, the nominal rates of the various import taxes on individual products are frequently increased or reduced, taxes on some products are temporarily waived, and others are exempted. This is particularly the case for the cess, which from the beginning has been used as a selective protective instrument.

COMMENTARY

By 2009, just under a quarter of the total number of tariff lines (half of agricultural lines and a fifth of industrial lines) had a specific component, compared to 1.2% in January 2004. This major policy change was partly implemented by expanding the number of specific CDs, but mainly by the use of specific cesses. Of the total number of tariff lines (2,601) subject to a cess, 32.4% were alternative-specific (i e, the higher of an ad valorem rate or a specific amount per physical unit) and 21.1% were specific-only (i e, a specific amount per physical unit). The use of specific import taxes – strongly discouraged by the WTO – is a classic method for insulating domestic producers from fluctuations in world prices, while at the same time making it very difficult to measure the protectiveness of a country’s tariff system.

Average Protection Levels

Table 1 compares the estimated average total protection rates resulting from CDs and para-tariffs in 2002, early 2004, 2009, and January 2011. During 2009 the total protection rate (TPR) was the sum of the protection provided by CDs and para-tariffs, i e,

TPR = CD + PAL + SUR + Cess + RIDL, where all these import taxes are expressed as percentages of the cif price. In November 2002 and February 2004 the cess and the RIDL did not exist, so for those years TPR=CD+PAL+SUR. In January 2011 the SUR and the RIDL had been abolished, so TPR=CD+PAL+cess. A number of points are apparent from these calculations.

First, as measured by unweighted average TPRs, protection went up slightly between late 2002 and early 2004, but then increased by two-thirds or more between 2004 and 2011. The average TPR for agricultural tariff lines increased from 28.1% to 46.8%, for industrial tariff lines from 10.7% to 19.7%, and for all tariff lines from 13.4% to 23.7%. These average protection rates are very high by world standards and have clearly reversed Sri Lanka’s relatively open trade policies.

Second, nearly all of the dramatic increase in the average protectiveness of the

Table 1: Estimated Unweighted Averages of Protection from Tariffs and Para-tariffs

January 2004

January 2011 ucts e g, chicken meat from 30% to 70.2%.

Customs Para- Total Customs Para- Total Customs Para- Total Customs Para- Total

Duties tariffs Protective Duties tariffs Protective Duties tariffs Protective Duties tariffs Protective

Rate Rate Rate Rate rate for toilet soap went up from 30% to

Agriculture 21.1 5.2 26.3 24.6 3.5 28.1 24.6 25.0 49.6 25.4 21.4 46.8 64.7%, bicycles from 20% to 43.4%, and

Industry 7.6 2.5 10.1 8.8 1.9 10.7 10.0 14.1 24.1 9.1 10.6 19.7 motor car tyres from 30% to 48.2%. Even

All tariff lines 9.6 2.9 12.5 11.3 2.1 13.4 12.1 15.7 27.9 11.5 12.2 23.7 though throughout the period Sri Lanka Sources: See Pursell and Ahsan (2011), Table 8.

continued to have a large, highly competi-Table 2: Protection Rates for Select Agricultural and Industrial Products (1999 and in 2009) tive export-oriented garment industry, the

HS Code CD=TPR CD PT TPR

Agricultural products 1999 to 53% in 2009. Likewise, the TPR on 0201.10 Beef carcasses 30 28 42.2 70.2 textile fabrics which had been cut to zero

0207.11 Chicken meat 30 28 42.2 70.2

0402.10	Skimmed milk powder	30	28	5.0	33.0 Finally, the most protective of the para
0702.00	Tomatoes	35	28	31.2	59.2
0704 to 0712	Most vegetables	35	28	42.2	70.2 tariffs in 2009 was the cess, followed by
0804	Mangoes, guavas, pineapples, etc	35	28	42.2	70.2 the PAL, SUR, and RIDL. As measured by
0805	Oranges, mandarins, grapefruit, etc	35	28	31.2	59.2 average protection rates over all tariff
0901.21	Coffee (roasted)	35	28	20.2	48.2	lines, the protectiveness of the cess (8.8%)
1005.90	Maize	35	15	34.8	49.8	was not much less than the protectiveness
1007.00	Sorghum	35	28	57.7	85.7	of CDs (12.1%). In addition, as are CDs, the
1601.00 2009 3208.10.10	Meat sausages 30 Orange and other fruit juices (brix value <20) 35 Industrial products Paints (polyester based) 20	28 28 28	42.2 47.7 42.2	70.2 75.770.2	cess is used selectively by setting high cess rates on finished products that the government wishes to protect and either
3401.11.10	Toilet soaps (retail packs)	30	28	36.7	64.7	exempting or setting low cess rates on
3924.10	Plastic tableware	30	28	42.2	70.2	intermediate inputs.
4011.10	Motor car tyres	30	28	20.2	48.2
HS 52,54 & 55	Textile fabrics	0	0	38.0	38.0	Distribution
HS 61 & 62 8703.21.91 8703.22.50	Clothing Cars <1000 cc Cars 1000cc-1500 cc	10 30 30	15 28 28	38.0 33.2 35.2	53.0 61.2 63.2	During 2009 there was an apparently simple structure of CDs with just five normal
8712.00.90	Bicycles	20	28	15.4	43.4	rates (0%, 2.5%, 6%, 15% and 28%) with a

in 1997 had increased to 38% by 2009.

CD=Customs duty, PT=Para-tariffs, TPR=Total Protection Rate. There were no para-tariffs on these products in 1999. low-to-moderate unweighted average CD

june 18, 2011 vol xlvi no 25

COMMENTARY

rate over all tariff lines of just 12.1%. But generalisations about the protective structure without allowing for the protective para-tariffs are meaningless. After allowing for them the protective structure was complex in the extreme, with 75 different total protective rates ranging from 0% to more than 90%. In particular,

(i) The TPRs on almost half (46.1%) of agricultural tariff lines exceeded 50%, clearly breaching Sri Lanka’s Uruguay Round commitment which bound nearly all agricultural tariffs at 50%.

(ii) The cess on its own is a major contributor to the wide dispersion of TPRs. High cess rates are frequently fixed when CDs are also high. For example, in 2009 the TPR of 431 tariff lines (6.6% of the total) was 70.2%, which is the sum of CD (28%), SUR (4.2%), PAL (5%) and cess (33%). Conversely, there is generally no cess on products with low or zero CDs (mainly inputs used by local industries). Cesses at varying rates below the top two levels seem to be fixed on a case to case basis in order to provide what are apparently considered to be appropriate import tax rates on imports under Sri Lanka’s PTAs, especially the agreements with India (ILFTA) and Pakistan (PSFTA).

(iii) As discussed below, the cess was introduced to provide protection against, and revenue from, preferential imports on which CDs were going to zero under these agreements. In this way it is a second, flexible and discretionary customs duty which is used by the government to tax imports from preferential sources in whatever way it considers appropriate, while at the same time providing extra protection against most favoured nation (MFN) imports.

Para-Tariffs and Sri Lanka’s PTAs

The nationalist government that came to power in April 2004 explicitly rejected Sri Lanka’s previously open trade policies, and the intensification of the civil war in the following years increased revenue needs. To meet these needs it was politically convenient to raise import taxes rather than domestic taxes. In this regard the scheduled reductions in CDs agreed under Sri Lanka’s PTAs were seen as undercutting both objectives, by reducing protection for domestic industries, and reducing revenue from import taxes.

The way out was to expand the use of India under ILFTA would have been 3.9%, para-tariffs, especially the cess, on the 61% below the average MFN TPR. But inargument that the PTA agreements applied cluding the para-tariffs, the average TPR only to CDs. From the government’s view-on industrial tariff lines under ILFTA is point, the use of the cess and the other 17.3%, just 28% below the average MFN three para-tariffs therefore killed four TPR. The de facto abrogation or partial abbirds with one stone (or more accurately rogation by Sri Lanka of this and its other with four!). First, para-tariffs can be set at PTAs from the para-tariffs is apparent whatever levels are desired to block or from Table 3. limit imports from MFN sources. Second, they can be set at any desired level to con-Consequences trol imports from PTA countries, especial-For a number of related reasons Sri Lanka’s ly imports from India and Pakistan. current protectionist import tax structure Third, since customs duty reductions un-has serious costs for economic welfare and der the PTAs have proceeded as originally growth, both in the recent past, at present, agreed upon, the government can argue and in the foreseeable future, unless that some margin of preference for these present policies are rapidly unwound. products is being maintained (even if the Some of these concerns are discussed by para-tariffs are on their own prohibi-Pursell and Ahsan (2011): tive!). Fourth, the para-tariffs are a sepa-(1) The country’s openness to internationrate and important source of revenue al trade as measured by the trade/gross from imports which reduces the need for domestic product ratio has drastically revenue from unpopular domestic indirect declined since 2000, with a particularly or direct taxes. sharp drop between 2008 and 2009.

Table 3 compares some statistics on the (2) This declining share of trade in the average protectiveness of Sri Lanka’s PTAs economy is consistent with the move during 2009, after allowing

Table 3: Unweighted Average Protection Rates under Sri Lanka’s for the para-tariffs. In this Preferential Trade Agreements

year, CDs on tariff lines subject to preferences (70.5% of		CD	Averages Para- tariffs	TPR	TPR:Standard Deviation
the total number of lines)	Agriculture
under ILFTA had come down	MFN	24.6	25.0	49.5	34.2
to zero, and CDs on tariff	GSTP	24.5	25.0	49.5	34.2
lines subject to preferences	APTA	24.5	25.0	49.4	34.2
(81.6% of the total) under PSFTA had been reduced by 60%. A number of major points are apparent from	SAFTA (India and Pakistan) SAFTA LDCs (e g, Bangladesh) PSFTA (Pakistan) ILFTA (India) Industry	23.4 21.2 15.8 22.2	24.8 24.5 23.7 24.6	48.3 45.6 39.5 46.9	34.5 35.8 34.7 35.6
these comparisons.	MFN	9.9	14.1	24.1	19.9
(i) Under every	PTA, the	GSTP	9.9	14.1	24.0	19.8
para-tariffs more than double	APTA	9.7	14.1	23.8	19.7
the average protection rate	SAFTA (India and Pakistan)	8.6	13.9	22.5	19.1
available from CDs alone.	SAFTA LDCs (e g, Bangladesh)	5.7	13.6	19.2	18.4
(ii) After allowing for the para-tariffs, even the PTAs with the most generous pref-	PSFTA (Pakistan) ILFTA (India) All tariff lines MFN	5.9 3.9 12.1	13.6 13.3 15.7	19.5 17.3 27.9	18.2 18.2 24.4
erences provide very high	GSTP	12.1	15.7	27.8	24.3
average TPRs which are not	APTA	11.9	15.7	27.6	24.3

far below average MFN TPRs. SAFTA (India and Pakistan) 10.8 15.6 26.3 23.9

(iii) The para-tariffs drasti-SAFTA LDCs (e g, Bangladesh) 8.0 15.2 23.2 23.8

PSFTA (Pakistan) 7.4 15.1 22.5 22.7

cally reduce the original per-

ILFTA (India) 6.7 15.0 21.7 24.1

centage preference margins

CD=Customs Duty; TPR=Total Protection Rate. agreed under the PTAs. For MFN=Most Favoured Nation; GSTP=Global System of Trade Preferences; APTA=Asia Pacific Trade Agreement; SAFTA=South Asia Free Trade Agreement; LDCs=Less

example, during 2009 the

Developed Countries; PSFTA=Pakistan Sri Lanka Free Trade Agreement; ILFTA=India average TPR on industrial (Sri) Lanka Free Trade Agreement. Source: See Pursell and Ahsan (2011), text and Table 11. The average protection rates tariff lines for imports from are over all tariff lines including tariff lines without preferences.

EPW

june 18, 2011 vol xlvi no 25

COMMENTARY

towards protectionist trade policies, especially after 2004. These policies pull resources out of export industries, appreciate the real exchange rate (or slow the devaluation rate), and slow export growth.

(3) There is a serious danger that capital (including foreign capital), labour and land will be disproportionately pulled into highly protected import substitution manufacturing and agricultural industries with low or negative economic rates of return.

(4) The revival of discretionary, ad hoc protectionism in Sri Lanka brings with it opportunities for substantial economic rents for favoured firms, especially if they have or develop market power behind high tariff and para-tariff barriers to import competition. A special danger is domestic or foreign direct investment in industries in which the final product is protected by CDs and para-tariffs (say against imports from China) while inputs can be imported over zero or low tariffs under Sri Lanka’s PTAs (say from India under ILFTA).

(5) Despite recent simplifications, the import tax and protective system is nontransparent, complex, constantly changing, and unpredictable. This combination is likely to deter long-term business commitments, both in production and trade. Related to this, the system appears to be highly discretionary (by whom and how is it decided to fix a Cess on a given product, and at what rate?). In such systems, financial success disproportionately goes to people and groups with lobbying skills and access to bureaucrats, politicians and other knowledgeable insiders.

(6) The new trade policies have sharply increased the government’s public finance reliance on trade taxes, reversing many years of successful efforts to instead use domestic and trade-neutral taxes. This increase was especially apparent following the largescale use of para-tariffs after 2004.

Implications for the WTO

As far as I know, none of Sri Lanka’s PTA partners have so far formally objected to its use of para-tariffs, even though they are yet another setback to many years of efforts to free up regional trade, especially trade between the south Asian countries. One possible reason is that starting on a large scale well before Sri Lanka, Bangladesh since 1997 has routinely used para-tariffs as flexible, discretionary and non-transparent instruments for providing extra protection to selected industries.2 A second reason is probably that from India’s perspective the two affected export markets – Bangladesh and Sri Lanka – are too small to warrant much negotiating effort. This is also very likely the attitude of China and Korea in the APTA.

However, the now well-documented unconstrained misuse of para-tariffs by Bangladesh and Sri Lanka does illustrate the importance of dealing effectively with this longstanding issue at the WTO. If a large-scale misuse on the Bangladesh and Sri Lankan pattern were to spread to many countries, tariff bindings and therefore the entire General Agreement on Tariffs and Trade (GATT)/WTO process would be meaningless.

Para-tariffs (known as “Other Duties and Charges” (ODCs) in WTO terminology) were recognised as a problem from the beginning of the GATT in 1947. Article II [1(b)] of the GATT says that when a product is subject to a “concession” (meaning a binding setting a maximum customs duty), it will also be exempt from “all other duties or charges...in excess of those imposed on the date of this agreement”.

Sri Lanka has bound only about a quarter of its non-agricultural tariff lines, so there is no effective WTO constraint on the level either of CDs or para-tariffs of the other three quarters. However a careful look would probably show that many applied industrial TPRs exceed their bound levels. This is unambiguously the case in agriculture, where nearly all of Sri Lanka’s tariff lines are bound at 50%. During 2009 to June 2010 almost half (46%) of applied agricultural TPRs exceeded 50%, and about 40% were within a range of 70% to 80%.

In the light of this it is (to say the least) surprising that the executive summary of the recent WTO Trade Policy Review report on Sri Lanka (2010), opens with the statement, “Sri Lanka has a fairly open and transparent trade regime, characterised by ...in general, relatively low tariffs”, and downplays the impact of “import surcharges and other charges” because “they have been taken mostly for revenue purposes” (p vii).

This statement misrepresents and blurs the implications of detailed information on Sri Lanka’s ODCs in the body of the same report. As Pursell and Ahsan (2011) show, the Sri Lankan trade regime is neither open nor transparent, after allowing for para-tariffs the protection available is very high, and the principal reason for the large-scale use of para-tariffs is not revenue generation, but to provide a protection-increasing instrument which the government hopes will pass under the radar of Sri Lanka’s PTA partners and the WTO. It is disappointing that this tactic seems to have succeeded with respect to what is intended to be one of the few sources in the world of objective and disinterested information on national trade policies.

Notes

1 Despite its official title the objectives and role of the cess are to provide extra protection to local import substitution industries and to raise revenue. It has nothing to do with promoting or subsidising exports.

2 Bangladesh’s para-tariffs are discussed in Daly et al (2001) and World Bank (2004), Vol II. I understand that they were still being extensively used during 2010 (personal communication with Zaidi Sattar, Policy Research Institute of Bangladesh).

References

Athukorala, P C (2007): “Outward-Oriented Policy Reforms and Industrialisation in Sri Lanka”, Economic Papers, 26(4): 372-91.

Athukorala, P C and S Rajapatirana (2000): Liberalisation and Industrial Transformation: Sri Lanka in International Perspective (New Delhi: Oxford University Press).

Daly, M, M A Khan and M Oshikawa (2001): “Tariff and Non-Tariff Barriers to Trade and Economic Development in Bangladesh”, Journal of World Trade 35(2): 253-73.

Pursell, Garry and F M Ziaul Ahsan (2011): “Sri Lanka’s Trade Policies: Back to Protectionism”, Australia South Asia Research Centre, Working Paper 2011/03, 21 January, available at http://www. crawford.anu.edu.au /acde/asarc/pdf/papers/2011/WP2011_03.pdf.

World Bank (2004): Trade Policies in South Asia: An Overview, Report No 29949, in three volumes, Washington DC, September.

World Trade Organisation (2010): Trade Policy Review Report on Sri Lanka, 30 November .

Permission for Reproduction of Articles Published in EPW No article published in EPW or part thereof should be reproduced in any form without prior permission of the author(s). A soft/hard copy of the author(s)’s approval should be sent to EPW. In cases where the email address of the author has not been published along with the articles, EPW can be contacted for help.

june 18, 2011 vol xlvi no 25