TRADE POLICY INCENTIVE STRUCTURE IN BANGLADESH 1993

TRADE POLICY INCENTIVE STRUCTURE IN BANGLADESH
1993 - 1996
ISS Working Paper #28
First Draft - November 1994
Revised - August 1996
Kamil Yilmaz
Koç University
Istanbul, Turkey
I. Introduction
One of the poorest countries in the world, Bangladesh poses challenging policy
questions for development economists. Progress in population control, agricultural
production and macroeconomic stabilization over the last decade has not generated
faster economic growth. The past experience taught at least one important lesson.
Whatever the priorities of the Bangladeshi government in the future, industrialization
should be the key objective. It is not at all possible to achieve robust economic
growth without vibrant, competitive industrial sector, currently absent in Bangladesh.
Over the last decade, no significant change in the relative position of domestic
industry took place. Manufacturing value added grew by 3.5 percent per annum
during the period FY 1974/75 - FY 1988/89, while the whole economy grew by 3.8
percent per annum. The contribution of the manufacturing sector to GDP has
stagnated at 10 percent since early 1980s. Although manufacturing employment has
almost tripled between FY 1983/84 and 1989, its share in total employment in 1989
was only 14 percent.
While frequent natural disasters have been an important factor in explaining
fluctuations in industrial and agricultural production, the single most important
element determining the growth prospects of the Bangladeshi economy is the absence
of comprehensive and radical economic policies. Not unlike many developing
countries, trade policy has played its fair role of providing skewed incentives to
Bangladeshi industry. Complex arrangements of trade instruments and regulation
(tariffs, quantitative restrictions, import licences and other non-tariff barriers)
provided domestic industry with excessively high levels of protection against
competition from imports. Not suprisingly, restrictive trade policy measures lead to
inefficnet allocation of resources.
In 1992, the government took serious steps in the right direction and went
ahead with a major overhaul of its trade regime. Since then a big part of importdiscriminating tariffs have been replaced by trade-neutral value-added tax. Rather
complicated tariff regime has been streamlined by reducing the tax dispersion across
goods and sectors, by reducing the maximum rate and the number of statutory rates, as
well as slightly increasing the minimum rate. Quantitative restrictions on imports
have, in the most part, been eliminated.
With the decline in the average rate of nominal protection from 90 percent in
1989 to 25 percent in 1995, Bangladeshi economy has become a fairly open economy.
Yet, the government of Bangladesh continues to exert substantial influence over the
entire economy through industrial, financial and trade policy measures. Furthermore,
aggregate or disaggregate trade policy measures do not provide a complete picture of
incentive structure faced by the industry. The decline in average tariffs and the
dispersion of tariffs do not provide much information about the implications of the
trade regime for various subsectors of the economy.
It is the objective of this paper to quantify the impact of reforms on various
subsectors of the economy using the methodology of effective protection. The paper
first focuses on the incentive structure in effect in the fiscal year 1993/94. Once a
complete picture of the trade incentive structure for the 1993/4 is depicted, the paper
1
moves to the analysis of the ground covered by reforms since then, including the
recent changes introduced in the fiscal year 1995/96.
The analytical framework of the paper is quite straightforward. Effective rate
of protection can be best defined as the percentage at which the value added at
domestic prices exceeds the value added at world prices. Using firm-level data derived
from the ISS survey effective rate of protection is calculated for each manufacturing
subsector.
The paper is organized as follows. In the next section we summarize the
methodology of effective protection. A more detailed exposition of the methodology
and a discussion of the ISS survey data can be found in the Appendix. Section 3
presents ERPs for various manufacturing subsectors in the FY 1993/94 in detail.
Effective protection in the FY 1995/96 is analyzed in Section 4. It is followed with an
international comparison of effective protection to develop directions for future tariff
reforms in Bangladesh. Section 6 concludes the paper.
II. Methodology of Effective Protection
The theory of effective protection was developed in the late sixties and early
seventies when any countries used to follow predominantly import-substitution trade
policies. The methodology emerged as a formalization of every day practices of
national tariff administrations. Under pressure from various industrial groups,
governments had to have a clear idea on the potential impact of industry-specific
protective measures on the rest of the economy. In the presence of intermediate input
use in production, an increase in the domestic price of a particular product brought
about by higher tariff protection would mean increased production costs for sectors
using it as input. The analysis of nominal protection would not be adequate to identify
economy-wide effects as it does not take into account backward and forward linkages
introduced by the presence of intermediate input use. The relevant measure of
protection would be the effective protection to the value added product rather than the
nominal protection to the final product.
Given the availability of data, estimation of effective rates of protection
(ERPs) is straightforward. A detailed description of the ERP methodology is
presented in the appendix. The main thrust of this methodological review lies with the
adaptations I had to make to accommodate for lack of data. In particular, lack of
disaggregated world price data forced me to rely on nominal protection rates in our
ERP estimations. Using ISS survey data on output and inputs valued at domestic
prices I calculated value added at domestic prices. I obtained the value added at
international prices after deflating the value of output and inputs with the appropriate
nominal protection rates.
There are certain problems with using nominal tariff rates to obtain world
prices from domestic prices. If tariff rates were actually binding (that is, if there were
no "water in the tariff") then the use of nominal protection rates (NPRs) to deflate
value added at domestic prices to obtain value added at world prices is no different
from inflating the value added at world prices to obtain value added at domestic
2
prices. However, the two procedures will not necessarily lead to the same result when
tariffs are not binding, which is generally the case.
There are several reasons to expect that high tariff rates are not binding. First,
high tariff rates provide incentives for smuggling with very high returns. Bangladesh
is no exception to this rule. According to a 1990 study by Bangladesh Institute of
Development Studies (BIDS, 1990) the list of goods smuggled into Bangladesh
include hundreds of items ranging from agricultural goods (fruits, cattle, hides and
skins, spices and pulses, poultry, edible oil, etc.) to manufacturing goods (yarn and
textiles, chemicals, cement, cigarettes, machinery and spares, medicine, electronic
goods, etc.). A second reason is related to the low degree of substitutability between
domestic and imported varieties of the same product, which, in turn, is dictated by
differences in quality. One would expect that, in the case of many products, the quality
of imported varieties is higher than the variety produced in Bangladesh. As a result,
the price of imported varieties should be higher than the Bangladeshi variety. In order
to control for quality induced price differences, price of imported variety relative to
domestic variety must be adjusted to reflect quality considerations. However, in the
absence of detailed information there is no way to make the necessary adjustment. By
deflating domestic prices with tariff rates to obtain international prices, I implicitly
assume that the two varieties of the same product are of the same quality.
Consequently, the use of tariff rates inherently leads to bias in ERP estimates. The
direction of the bias, however, is not determinate. Finally, tariffs do not reflect the
impact of other trade policy measures, such as quantitative restrictions and bans,
import licenses and other regulations, which may directly or indirectly cause a wedge
between domestic and international prices. In certain cases, other trade policy
instruments may be more restrictive than tariffs would indicate.
Despite the problems mentioned above, lack of world price data forced us to
use tariffs to obtain measures of effective protection. Since neither the domestic nor
the world price data were available at a disaggregated level I tried to incorporate
certain additional features to our ERP estimations to counter the bias introduced
because of the presence of "water in the tariff".
Table 1 presents measures of nominal and effective protection, profitability
and efficiency in 12 manufacturing sectors for the FY 1993/94 as defined by ISS
survey.1 Table 2, presents the same information for 3- and 4-digit BSIC (Bangladesh
Standard Industrial Classification) sectors.2 ERPs are calculated from the sectoral data
obtained by blowing up the results from the stratified ISS sample survey for each
subsector and size class. In order to show the extend of the coverage for each sector
the table also presents the number of firms in the ISS survey sample and the universe.3
1 We had to deviate from the ISS definition in the case of Engineering sector, as we did not have data
on the number of enterprises in the universe to calculate blow-up factors for 4-digit subsectors of 381.
We define Steel and Engineering sector as 371 + 38.
2 Three-digit subsectors presented in Table 2 are not inclusive of the 4-digit subsectors.
3 The sample coverage ratio is at or above 10 % for the whole manufacturing sector and in 6 out of 14
ISS subsectors. (See Table 1.)
3
Two different effective protection measures are calculated for each subsector.
The first one (ERPds) measures the effective protection on domestic sales, whereas
the second one (ERP) is based on total sales, which includes exports as well as
domestic sales. From the perspective of trade policy analysis the correct measure of
effective protection is ERPds as long as exporters are granted duty exemptions on
imported inputs. ERPs based on total sales are presented in order to show the impact
of duty exemptions and drawbacks in export oriented sectors on effective protection.
There is no monotonic relationship between the two sets of ERPs. Depending on the
reduction in nominal protection on output relative to inputs as a result of duty
exemptions, ERPds can be higher than ERP on total sales. This is actually the case in
the majority of subsectors. The two are equal if no export activity takes place in the
subsector in question.
In addition to ERPs, I present nominal tariff equivalent (NTE) on output and
tradeable inputs. NTEs are presented along with ERPs to double check the accuracy
of ERP estimations. NTE on sectoral output (NTEo) can be defined as the average
tariff rate on products of the sector weighted by their value of production. Similarly
NTE on tradeable inputs (NTEi) is defined a la Corden over tradeable inputs and
tradeable components of non-traded inputs and depreciation. As can be seen from
equation 5 in the Appendix, NTEs are directly linked with ERPs. One would expect
ERP to be greater (less) than NTEo, when NTEo is greater (less) than NTEi. In other
words, level of effective protection is larger than the nominal protection if output is
protected more heavily than inputs.
Table 1 also presents measures of profitability and efficiency: the financial rate
of return (FRR) and the economic rate of return (ERR). Financial rate of return is a
measure of profitability defined relative to the sum of the value of fixed and working
capital valued at domestic prices. Economic rate of return, on the other hand, is a
measure of profitability of the industry if it were to operate at international prices. In
other words, it measures whether the sector is economically efficient to compete once
the existing protection barriers are removed.
Finally, the tables include total sales valued at domestic prices, employment
and value added share in total output to facilitate the interpration of ERPs as well as to
highlight differences across subsectors. According to our results the manufacturing
sector produced Tk. 294 billion worth of goods in the fiscal year 1992, employing 1.6
million workers. Value added share of the output was 27 percent, low compared to
figures for 1988 and 1989 (36 and 34 %, respectively) published by UNIDO. The
whole manufacturing sector exported 44 percent of its production in FY 1992, a
substantially large share. Total wage bill of the manufacturing sector in 1992 amounts
to Tk. 28.7 billions, approximately 10 % of the total output.
III. Effective Protection in the Manufacturing Sector - FY 1993/94
The level of effective protection on domestic sales of the entire manufacturing
sector is 82 percent. (See Table 1) Although the difference between the nominal
protection on total output and inputs is just 3 percentage points, the effective
4
protection on total output is 34 %. As can be seen from equation 5 in the appendix,
this result is quite normal. First observe that the value added share in the
manufacturing sector is only 27%. Second, in the definition of ERP (equation 5) the
discrepancy between NTEo, on the one hand, and NTEi multiplied by the share of
intermediate inputs times, on the other, is weighted by the inverse of the value added
share.
One of the most questionable results in Table 1 is on the measures of
profitability and efficiency. Financial rate of return (FRR) in the whole manufacturing
sector is 37 %, while economic rate of return (ERR) is 33%. Together these results
indicate that manufacturing production is profitable and the source of profitability
does not disappear once the protection is removed. The implication is rather counter
factual. Together these two ratios depict the picture of a manufacturing sector which
can easily compete against imports and manage to survive once the high protection
walls are removed. It seems as if this is not the sector previous World Bank studies
reported on. They invariably described an inefficient manufacturing sector plagued
with all sorts of problems (including the lack of financial capital and of
infrastructure). However, when one analyzes other columns in Tables 1 and 2, it
becomes apparent that the difference in our sectorwide efficiency and profitability
measures from the earlier studies is partly due to the composition of the ISS survey
sample. One third of total manufacturing output is accounted for by the two most
profitable and efficient subsectors: Readymade Garments, and Tanning and Leather.
To some extent, these two sectors counter balances inefficiencies in other sectors.
Five out of the remaining 11 ISS subsectors (Residual A and B combined) have
negative economic rate of returns. Jute is the least efficient and the sole unprofitable
sector with an ERR of -28 % and FRR of -34%. (See Table 1)
The rest of the analysis moves from the most aggregate definition of
manufacturing subsectors (namely ISS definition) towards 3 and 4 digit BSIC sectoral
definition. Knitwear (ISS 6) is the most protected ISS subsector. (Table 1) ERPds is
not defined for knitwear as its value added at international prices (denominator) is
negative. Surprisingly, the second in the line is ready made garments, RMG, (222 %),
where domestic sales account only 3% of the sectoral sales.4 In four other sectors
effective protection levels are above 100 percent. These sectors are Food Processing
(139 %), Tanning & Leather (127%), Perfume & Toiletries (116%) and Textiles
(105%).
It is interesting to note that two major export sectors (RMG and tanning &
leather) enjoyed easy profits in the domestic market thanks to high tariff rates.
Although the 1992 share of domestic sales were quite low in both sectors, high level
of effective protection has reduced incentives for exports. In addition, the high level of
ERPs show that if the export incentive measures such as duty exemptions were to be
removed, even temporarily, Bangladesh will suffer a significant decline in its exports.
It is therefore highly advised that the government should further reduce high tariffs in
these two sectors along with others.
4
Indeed, Residual A and Residual B covering 378 enterprises (half of which is in bidi production)
were subject to higher protection level.
5
It is not possible to make a similar observation in the case of the other export
sectors. They are the least protected subsectors after pharmaceuticals. ERPds is only
36 % in Jute Textiles (Table 1) and 34 % in Fish and Seafood. (See Table 2) Since
Bangladesh is the most important jute exporter in the world, its jute sector does not
need heavy protection. However, its being the major supplier of jute to the world
market does not require that Bangladeshi jute sector is an efficient one. Indeed, based
on FRR and ERR, it is not difficult to conclude that jute sector was a loss making and
inefficient sector in 1992. Its losses amounted to 34 % of its total capital. The figures
clearly show that jute sector needed restructuring in 1992, a project which has been in
the process of implementation since then. Fish and seafood sector does not need
protection either. Moreover, its potential as an export sector is displayed in its
relatively large profitability and efficiency measures.
In five manufacturing subsectors nominal protection levels were extremely
high so that the value added at world prices was negative, and ERPs not defined.
These sub-sectors are Cigarettes and Tobacco (3141), Bidies (3143) and knitwear
(3223), (369) and (372). (See Table 2) The negative value added in these sectors
(especially in tobacco) play an important role in increasing the ERP estimates for the
whole manufacturing industry. When these subsectors are excluded, effective
protection level for the whole manufacturing sector goes down to 64 percent.
In the rest of this section I discuss measures of effective protection and
efficiency in major manufacturing sectors in detail, starting with the export-oriented
sectors, and moving to textiles, food processing and beverages, chemicals and steel
and engineering, and finally tobacco manufacturing.
i. Export Sectors
Among the major export sectors of Bangladesh are ready made garments
(323), Jute (Jute Spinning and Weaving (3213) and Jute Pressing and Baling (3263)),
Processed Fish and Seafood (3114), and Leather and Leather Products (324). In FY
1992/93, these sectors exported 1,240, 292, 174, and 151 million US dollars worth of
goods respectively. The ISS survey covered a large number of enterprises in all five
sectors.5 Together major export sectors account for 42 percent of total output at
domestic prices. Not surprisingly, their share increases to 50 percent when the output
is valued at world prices. At this stage it is important to note that the survey results
for the export sectors can expected to be closer to reality than other subsectors.
Sample coverage ratios in these subsectors are higher than 10 percent while the same
ratios are much lower in other subsectors.
In the export-oriented sectors only 7% of output is sold domestically.6 Despite
the low share of domestic sales the relevant measure of effective protection is the one
based on domestic sales. Effective protection is very high in RMG sector. The ERP on
domestic sales is 222%. (Table 1) Then comes the leather and tanning with an ERP of
5
The sample coverage ratios are respectively 12, 22, 5, 23 and 15 percent.
This ratio is the lowest in RMG sector, with only 3 %. The share of domestic sales in other sectors
are 20, 24, 8 and 12 percent respectively.
6
6
115%. The other three sectors are well below the average for the whole manufacturing
sector.
The share of value added in total RMG output is 24 percent, which
demonstrates that the minimum value added limit of 30 percent introduced by the
government in 1993 was actually binding in 1992. The largest value added per unit of
output is generated in tanning and leather finishing with a share of 52 percent. High
value added share together with a high economic rate of return (154%) underlies
Bangladesh's comparative advantage in leather and tanning sector. (Table 2) Among
the major export sectors fish and seafood has the lowest value added share with 21%. .
Tanning & Leather sector contributes approximately 7% of the manufacturing
output with only 1.3% share in total manufacturing employment. Output-labor ratio is
1 million Taka per worker, a substantially large amount for this sector, since leather is
known to be a labor intensive sector. (Table 1)
Tanning and leather industry has three subsectors at 4-digit BSIC level:
tanning & leather finishing (3241), leather products excluding footwear (3243) and
leather footwear (3251). (See Table 2) In our sample 95% of the output and 62 % of
the employment in this ISS sector is accounted for by leather & tanning finishing.
High value added share in the subsector (52%) explains the high value added share
for the Tanning & Leather (51%). Similarly the profitability and efficiency measures
for the whole leather & tanning sector (116% and 142 %, respectively) are closely
related to these measures in the leather & tanning finishing subsector (127% and
154%).
Fish & seafood (a subsector of Food Processing (311)), which exports 92 % of
its output, is also highly profitable and efficient, with 41% FRR and 56% ERR.
(Table 2) As a result, the level of effective protection on total output is negligible.
This is reflected in the relative magnitudes of FRR and ERR. ERR is greater than
FRR, which implies that this sector is more profitable (hence efficient) at
international prices.
ii. Textiles
The definition of the Textiles subsector in ISS survey (see Table 1) is different
from the BSIC definition under code 321. Basicly ISS definition includes only cotton
spinning and weaving, classifying other textile subsectors within the Residual A. This
classification is misleading because the survey results show that other textile
subsectors account for half of the total output (compare figures in Table 2 with those
in Table 1). For the purposes of this analysis I take BSIC 321 as the textile industry. In
addition to cotton spinning and weaving (3211), this definition includes silk and
synthetic spinning and weaving (3214), and dyeing & printing (3217) are the major
subsectors of the textile industry.
Unlike the export-oriented RMG sector, for the most part textile industries
(BSIC 321) produce for the domestic market. ISS survey database carried fairly good
quality data for 162 textiles firms (out of 1730 firms) whose total production
amounted to Tk. 28.8 billion, less than 10 % of total manufacturing output. (Table 2)
The sector is primarily domestic market oriented, selling 80 percent of its output in the
7
domestic market. Share of value added in total output is 26 percent, produced by
218,000 strong workforce, which accounts for approximately 13.5% of
manufacturing employment. Effective protection on domestic sales of the textile
industry is 157 %, resulting from a 44% NTE on output and 31% NTE on inputs.
Although the industry is profitable with a 5 % FRR, its operations are inefficient and
the industry may not survive competition under free trade (ERR = -10%).
Since independence in 1971, the textile industry has always been protected
from import competition. Despite the success of the RMG sector in opening up to
international export markets, the rest of the sector continued to benefit from high
protection. There was no significant change in protection levels even after the
privatization of many textile firms in 1982, which had been nationalized after
independence.
Quantitative bans and restrictions as well as tariff rates at or above 100 percent
was the norm in the sector during the eighties. Even after the most serious trade
liberalization effort in FY 1991/92 tariffs and QRs continue to thwart import
competition. In FY 1991/92, the average nominal protection rate (including VAT)
stood at 93 percent. Although trade liberalization brought about a substantial decline
in the tariff rates, it was not sufficient to ensure a modest level of effective protection.
An important feature of protection in Bangladesh is the tariff escalation with
the level of processing. Textiles is not the exception. Nominal protection rates
(inclusive of VATs because initially domestic textile production was exempt from
VAT and 2.5 percent import license fee) on primary textile inputs (cotton, wool, raw
silk, etc.) was less than 75 percent. NPRs on chemicals and dyes ranged from 52 to
86.5 percent. On the other hand, the NPRs on finished textile items, (yarn, thread and
fabrics) ranged between 52 and 190 percent. (See Table av.1 in GATT (1992).)
Effective protection varied substantially among the subsectors of the textile
industry. Export-oriented jute spinning and weaving sector has the lowest ERP
estimate, 36 percent. It is followed by the dyeing and finishing which has an ERP of
91 percent. ERPs in cotton spinning and weaving and silk and synthetic S&W are,
respectively, 105 and 6282 percent. (Table 2)
Finally, although not part of the textile industry the level of effective
protection was very high in knitwear textiles sector (3222), which produces negative
value added at world prices. While there are several other 4-digit textiles subsectors, I
do not discuss them here because the estimates for these subsectors are all based on
observations for only a few enterprises.
The huge discrepancy between ERPs total output of textiles and the RMG
definitely underscores the need for the long overdue restructuring efforts in textiles.
As long as the domestic textile sector is protected from international competition and
major restructuring and investment decisions are postponed, it will never be possible
to establish firm backward linkages between textiles and the RMG. In order to cater
to its international customers RMG sector will always be pressed to use relatively high
quality inputs, which can not be delivered by the existing textile sector. Even lower
protection rates would not suffice. The development of the textiles industry will
require investment in modern weaving, spinning and finishing technology. Unless
8
such long term measures are taken, government will have to continue with its
minimum domestic component requirement measure in the RMG sector. This in
return will harm the international competitiveness of the RMG sector.
iii. Food Processing and Beverages
The food processing sector is primarily domestic market oriented. Actually,
when processed fish and seafood subsector is excluded the output of the rest is
completely sold domestically. According to Bangladeshi statistics, the sector
accounted for 22 percent of total industrial output, second only to textiles (inclusive of
RMG). ISS estimate for the share of the industry in total manufacturing is close to
this figure, 21 %. Furthermore, the industry employs 305,600 workers, 19% of the
total manufacturing workforce.
The level of effective protection in food processing and beverages (311, 312)
industries are close to the average ERP for the whole manufacturing sector. (Table 2)
When combined these sectors enjoy an effective protection of 69%, resulting from an
23% NTE on output and 18% NTE on inputs. The two sectors together are both
profitable and efficient according top our measures of FRR (42%) and ERR (30%).
The ERP on food processing is 45 percent, while it is 139 percent on beverages and
bakery products. The highest ERP in food industry is obtained for dairy products at
121 percent whereas the least protected subsector is edible oils, which has an ERP of
20%.
Rice milling and husking received an effective protection of 69%. (Table 2)
Zero operative tariffs on paddy and husked rice was reflected on an 8% NTE on inputs
compared to 14% NTE on output. The output of the sector was relatively large,
accounting for approximately 6% of manufacturing output and 4% of manufacturing
employment. Moreover, the sector was one of the most efficient 4-digit BSIC
subsectors. The economic rate of return was 40 % in FY92.
Another important production activity under the food processing and
beverages is tea processing. Tea has been one of the traditional export commodities of
Bangladesh and makes up a substantial portion of the total output in the beverages and
bakery sector (BSIC 312). The level of effective protection on domestic sales was 91
% and declines to 64% when exports are also taken into account. (Table 2) This is
extremely high for an export sector, especially when the tariff rate applicable to tea
leaves and processed tea are the same. High protection, it turns out, stems from the
fact that most of the tea processing companies surveyed for ISS are the estates which
produce tea leaves. Duty exemptions on agricultural inputs such as fertilizers is the
real factor behind this result. This is also reflected in the 42% NTE on inputs
compared to 51% NTE on output. Another important point about tea processing is the
extent of intra-industry trade. While Bangladesh produced Tk. 2.2 billion worth of tea
in FY 1991/92, in the first nine months of the fiscal year it imported Tk. 824 million
worth of tea. Tea imports have been on a steep climb from a low level of Tk. 8 million
in FY 1988/89. Therefore, it is not completely incorrect to expect the government to
help tea processors by increasing the effective protection.
iv. Chemicals
9
The development of the chemicals industry in Bangladesh is based on import
substitution. Although, an important intermediate good sector, chemicals industry is
yet to achieve a significant growth. Domestic production of intermediate chemical
products, such as sulphuric acid, hydrochloric acid, caustic soda, and chlorine,
recorded no significant expansion between FY 1988/89 and FY 1991/92. (See
Bangladesh: Monthly Bulletin of Statistics) Similarly, the production of soaps and
detergents, PVC products, matches, cycle tyres and tubes has either declined or
expanded very little.
Despite, import substitution strategy Bangladesh continues to rely heavily on
imported chemical products. The only exception is the soap and detergents where in
1985 imports accounted for only 2 percent of apparent domestic consumption. A
comparison of HS based import data with BSIC based production data shows that
there was no significant change in the share of imports. Import share in
pharmaceuticals was approximately 25 percent in FY 1990/91.
Among the subsectors of the chemicals industry the ISS data cover drugs and
pharmaceuticals, industrial chemicals, soaps and detergents, miscellaneous petroluem
products, rubber products, and plastic products. Altogether chemicals subsectors
produce Tk. 28.9 billion worth of output, accounting for approximately 10% of the
manufacturing output. 83 % of this output is domestically sold. Total employment in
the sector amounts to 92,558. The sector value added is 32% of the total output.
Overall chemicals industry is protected very little with only 11% ERP on domestic
sales. (Table 2)
Not unlike other manufacturing sectors the level of effective protection varies
substantially among the chemicals subsectors. While drugs and pharmaceuticals is
faced with negative protection with an ERP estimate of -5 percent, effective
protection on rubber products, miscellaenous petroluem and coal products, soaps and
detergents are 298, 280 and 155 percent, respectively. (Table 2) Plastic products
production is accorded 138 percent effective protection, whereas the ERP estimate for
industrial chemicals is 97 percent. (Table 2) The low level of industry wide effective
protection is due to the large output share of drugs & pharmaceuticals, which accounts
for 65 % of the output.
Average NPR for inorganic and organic chemicals, chemical elements and
compounds was approximately 58 percent, whereas the average NPR on soaps,
detergents and perfumes was in the range of 85 percent. This may explain excessively
high ERP estimates for soaps and detergents. Again, in line with negative ERP
estimates the average NPR on medical and pharmaceutical products was only 41
percent, lower than the average NPR on the main inputs of the sector, intermediate
chemicals.
v. Steel & Engineering
In our study steel & engineering comprises BSIC subsectors 371, 372, and 38.
In this respect, it is not completely consistent with the initial definition of this sector.
In total S&E produces Tk. 40 million worth of output, of which only 3% is exported
and 24% is the share of value added by the sector. S&E has a 121 thousand strong
10
labor force and appears to be both profitable and slightly efficient according FRR and
ERR.
The sector as a whole faces a very complex and often contradictory incentive
structure. While some local production are highly protected many others receive little
protection. Unlike the textiles industry, I do not observe an escalated tariff structure
in the steel and engineering industry. Iron and steel products are on average assessed
a higher duty than metal products and machinery in whose production they are used as
inputs. This is reflected in the ERP estimates for the industry. While ERP in basic
metal industries (BSIC 38) is 97 percent, it is only 62 percent in metal products and
machinery.
The share of value added in electrical and non-electrical machinery output turn
out to be substantially lower than the Indian machinery sectors. The value added
share is respectively 30 and 26 percent in non-electrical and electrical machinery,
while it is 41 percent in the least efficient machinery producing enterprises of India.
(See Aksoy and Ettori (1992)) Although the coverage of data in these sectors are not
very extensive, low value added share in Bangladeshi capital good industries suggest
that they are concentrated on spares rather than machinery production.
While effective protection is excessively high in intermediate product sector
such as Basic metals and Metal products, in machinery and spare parts are subject to
lower protection. Based on data for 18 firms with a total production of Tk. 730
million, ERPs on non-electrical and electrical machinery are respectively 31 and 33
percent, lower than the average NPRs which were 46 and 55 percent in FY 1993/94.
Both sectors are completely domestic market oriented.
Low levels of effective protection on machinery follow from the presence of
numerous special tariff exemptions on capital goods imports. Tariff exemptions effect
53 percent of the 8-digit HS items in non-electrical machinery and 22 percent of the 8digit items in electrical machinery sectors (See GATT 1992). Tariff exemptions for
machinery imports are in place because Bangladesh does not have the technology to
produce many types of machinery. By effectively reducing tariffs on capital good
imports the government aimed at reducing the disincentives for private investment.
The heavy reliance on imported capital goods is evident in the share of imports
to apparent domestic consumption. In 1985, the latest year for which the data is
available, the share of imports in total consumption of these sectors was as low as 30
percent for electrical appliances and houseware, and as high as 100 percent for
electrical industrial machinery and office machinery. It was 48 percent in engines and
turbines, 70 percent in radio and television and 74 percent in agricultural machinery
and equipment. (Table AV.9, GATT 1992) In FY 1991/92, imports of machinery
stood at Tk. 26.2 billion, 27 percent of total imports.
Another S&E subsector is the transport equipment sector. Despite excessively
high levels of CD rates effective protection on transport equipment is 76 percent,
relatively lower from other sectors which are faced with high tariffs. This low ERP
estimate may be a result of the of insufficient coverage of the industry. In our sample
there were only five enterprises with only 20 million Taka worth of production.
11
vi. Tobacco Manufacturing
Tobacco manufacturing is well covered by the data, with a coverage ratio of 9
%. The high protection rate accorded to tobacco manufacturing is distributed equally
to its subsectors. All sub-sectors (Cigarettes and Tobacco, Cigars and Cheroots,
Bidies) are heavily protected with a 354 % NTEo and 65 % NTEi. The average rate
of nominal protection for the whole industry was only 138% in 1992.7 While tariff
rates on raw tobacco ranges between 10 and 30 percent with an average of 22 percent,
tobacco products (cigarettes and bidies) are imparted tariff rates at 300 percent. The
sector is completely domestic market oriented. Value added share in total output
amounts to 20 percent. Since tobacco products are consumer goods and easy to
transport one would expect to observe substantial smuggling of tobacco products from
India. This would definitely reduce the difference between domestic and border prices,
reducing the effective protection below the levels implied by our estimates.
IV. Effective Protection in the Manufacturing Sector - FY 1995/96
There has been a gradual removal of protection in Bangladesh. Average
nominal protection was as high as 89% in 1991. It went down to 50% by 1993 and
further down to 24.8% with the announcement of the new trade regime in FY
1995/96. However, this substantial decline in the nominal protection has not been
reflected in the effective protection. Effective rates of protection for ISS sectos in the
FY 1995/96 are presented in Table 3. They are calculated in the same fashion as the
ERPs for FY 1993/94, using the ISS data for 1992. The only difference between the
two set of ERPs lies in the tariff rates used. In this section, I summarize the results for
ISS sectors only. Effective rates of protection for 3- and 4-digit subsectors are
presented in Table 4.
From FY 1993/94 to FY 1995/96, effective protection in the manufacturing
sector declined from 82 % to 67% (See Table 3). This is consistent with the greater
decline in nominal protection according to manufacturing output compared to
protection on inputs. Nominal Tariff Equivalent on domestic sales of output (NTEods)
(see Equation 13 in Appendix) which used to be 48% in the FY 1993/94 declined to
39% in FY 1995/96, whereas nominal tariff equivalent on inputs used in the
production of domestically sold output (NTE ids) declined from 36% to 32%.
All manufacturing subsectors except Steel & Engineering (S&E) and Food &
Dairy (F&D) accorded a decline in their ERPs. Effective protection in the S&E went
up from 86% to 123%; whereas EP in the F&D went up more dramatically, from 45%
to 123%.
In the case of F&D, the increase in effective protection was a result of an
increase in the nominal protection on output, from 19% in 1993/4 to 26% in 1995/6.
There was no change in the nominal protection on inputs. It stayed put at 15%.
The difference between the NTEo and the average nominal protection rate for tobacco stems from
the fact that the majority of the firms were involved in bidi and cigarette production, both of which were
subject to 350% nominal protection rates.
7
12
Effective protection in S&E was already high in FY 1993/4. With the changes in the
tariff code, effective protection went up, although both NTE on output and inputs
went down. The decline in NTE for inputs was 1% more than the fall in the NTE for
the domestically sold output. Since the share of value added in the total production is
relatively low at 24%, the small change in NTEs resulted in a significant increase in
effective protection.
When we exclude S&E and F&D subsectors from the ISS sample of
manufacturing sector, the reduction in protectionist measures becomes even more
visible: From 93% in FY 1993/94 to 45% in FY 1995/96. This significant change in
the level of effective protection is primarily a consequence of the decline in NTEods
from 59% to 40%. In contrast, NTE on inputs used in the production of domestically
sold output went down from 43% to 38%. (see the last row of both tables)
Changes in the effective protection of readymade garments, tanning and
leather, two leading export sectors, as well as Knitwear and the Residual A category
have been dramatic. ERPs on domestically sold RMG products dropped from 222%
in FY 1993/94 to 48%. In the Residual A category the effective protection rate went
down from 238% to 44%. Similarly, tanning and\ leather experienced a similar decline
from 127% to 41%. In FY 1993/94 tariff protection in Knitwear industry was so high
that value added at world prices was negative. With further tariff reforms effective
protection in this sector went down to 64%, still higher than the average for the whole
manufacturing sector.
V. Cross Country Comparisons
Table 5 presents effective rates of protection (ERPs) for several manufacturing
sectors in Bangladesh, India, Pakistan, Mexico and Malaysia. At this stage a word of
caution about the comparability of ERP estimates accross countries is appropriate
before starting with the cross-country comparison. The methodology of effective
protection may vary across countries in the table for several reasons. Since not all of
the sources discuss their methodology it is not possible for us to get into details of
ERP estimations for each country. Our comparative approach is not undermined by
this problem, because the cross country comparisons are intended to provide a
snapshot of the position of Bangladeshi industry relative to other countries, not
necessarily to define a benchmark tariff structure for Bangladesh.
Table 5. Cross Country Comparisons of Effective Protection (%)
Manufacturing
Sub-sector
Food Processing
Beverages
Tobacco
Textiles
Ready Made Garments
8
Bangladesh
India
Pakistan
(FY 93/94) (1986-89) (1989/90)
45
52
n.a.
139
n.a.
n.a.
VAw<0
n.a.
n.a.
80
94
1008
222
n.a.
n.a.
Synthetic textiles; ERP for Cotton Spinning is 52 %.
13
Mexico
(1991)
n.a.
n.a.
-56
49
8
Malaysia
(1991)
n.a.
-22
-26
15
6
Leather Products
Wood Products
Printing and Publishing
Industrial Chemicals
Rubber Products
Plastic Products
Basic Metals
Metal Products
Fabricated Metal Prod.
Non-Electric Machinery
Electric Machinery
Transport Equipment
115
183
-27
97
298
138
76
139
46
195
84
58
n.a.
n.a.
n.a.
68
n.a.
37
72
72
n.a.
64
42
19
n.a.
n.a.
n.a.
102
23
23
25
19
n.a.
58
-13
-1
-53
9
-57
94
23
9
59
149
54
21
62
-19
n.a.
82
-9
14
163
n.a.
289
30
n.a.
19
12
65
Sources: Mexico: GATT Trade Policy Review, 1993 ; Malaysia: The East Asian Miracle, The World
Bank, 1993; Bangladesh, India and Pakistan: World Bank data and Staff estimates.
Table 5 reveals that overall effective protection enjoyed by Bangladeshi
manufacturing industry is substantially larger than other countries. Industrial
chemicals, printing and publishing, food processing (due to export oriented frozen
sefoods) are among the exceptions. The comparison of protection levels in tobacco is
most interesting leaving no basis to explain the logic behind the substantially high
protection walls in Bangladesh. While Bangladeshi tobacco manufacturing sector
produces negative value added at world price, Tobacco sectors in Mexico and
Malaysia are able to survive under negative effective protection.
Effective rate of protection estimates for textiles industry show the similarity
between the Bangladeshi and Indian textile sectors. Pakistan, Mexico and Malaysia
follow Bangladesh and India in terms of the effective protection accorded to textiles.
In contrast with the textiles sector in all three countries, for which the estimates are
available, RMG sector appears to have relatively low rate of effective protection (3%).
However, when one looks at domestic sales, only Bangladeshi RMG sector appears to
be heavily protected (222%). Malaysia and Mexico are among the major exporters of
RMG in international markets. I would expect ERPs to be small for India and Pakistan
as well. Because of their competitive-edge in apparel trade none of the countries need
to raise protection walls against import competition.
In general, the ranking of ERPs in textiles and RMG sectors, as well as the
comparison of ERPs between the two sectors are not surprising. Thanks to duty
exemptions and special bonded warehouse system for exporters, Bangladesh has a low
rate of effective protection in RMG sector compared to textiles even though it has
higher nominal rates in RMG sector on average. Since textiles products are the main
intermediate inputs for the RMG sector, the small difference between the average
tariff rates of the two sectors explains the relatively low rates of effective protection
for the domestic sales of the RMG sector.
In this respect, the case of Malaysia may be very instructive for Bangladeshi
policy makers. Malaysia had adopted the export-oriented development strategy earlier
than many other developing countries. In addition, Malaysia invested heavily in
sectors which it did not initially have comparative advantage. Textiles is one these
sectors as compared to labor intensive RMG sector. Today, while the Malaysian
14
textiles can easily compete in international markets, it is difficult to make a similar
statement for Bangladesh, India and Pakistan which have catered to domestic market
only.
ERP estimates for leather and footwear sectors show a substantially high
effective protection for footwear sector whereas the ERP for the leather sector is lower
than its average tariff rate. Because of the data limitations it is not possible to
compare Bangladeshi leather and footwear industries with their counterparts in other
countries. The only exception is Mexico. The comparison with Mexico’s leather
sector, which appears to have been hurt by high protection accorded to its inputs, leads
one to conclude that effective protection in Bangladeshi leather sector was relatively
high as of FY 1993/94.
Bangladesh has one of the most protected steel and engineering sector among
the countries for which ERP measures are presented in Table 5. ERPs in metal
products, non-electrical and electrical machinery are higher than the ERPs in other
countries. The comparison of Bangladeshi iron and steel sector with its counterparts in
India and Pakistan indicates, however, that it is imperative for Bangladesh to reduce
nominal protection rates in this sector which is continued to take place in the last two
budgets.
Levels of effective protection in basic metals, fabricated metal products and
transport equipment sectors are commensurate with other countries although they
provide substantially levels of effective protection. ERPs for these sectors are indeed
lower than at least one another country. With the exception of Bangladeshi nonelectrical machinery, in all countries effective protection on machinery and spares are
among the lowest in the manufacturing industry. This is a good indication of the
heavy reliance on capital good imports not only by Bangadesh but other countries as
well.
The variation in ERPs in chemical subsectors is observed in other countries as
well. While ERPs in Bangladeshi chemical subsectors vary between 97 and 298
percent, they are as high as 163 percent in Malaysian rubber sector. Bangladeshi
chemical industry appears to be excessively protected compared to other countries. In
all countries printing and publishing receives negative effective protection. This may
be due to high tariff rates on paper products, which is not included in the table because
of dependence.
In all sectors but textiles Bangladeshi tariff regime in was more protectionist
than the Indian regime that had been in effect between 1986 and 1989. Similarly
effective protection levels in Pakistan in FY 1989/90 were all below the
corresponding measures in Bangladesh, with the exception Industrial chemicals.
Although, important significant steps have been taken towards tariff reforms in
Bangladesh since FY1992, the analysis of nominal rates show that these efforts were
not sufficient to bring about the much needed chamge in the incentive structure.
Bangladeshi government needs to realign its tariffs to insure that the minimum
effective protection levels which do not completely alter the incentive structure under
free trade.
15
VI. Conclusions
This paper presented the analysis of effective protection in Bangladeshi
manufacturing sector. ERP estimates are based on the ISS survey data which was
conducted during the FY 1992/93. The analysis is a big step forward in drawing the
incentive structure faced by the Bangladeshi manufacturing in FY 1993/94 and FY
1995/96.
Although I have taken all steps to make sure that the analysis reflects the true
incentive structure in Bangladesh, there is still room for improvement in the
estimation of ERPs. The most important task in the future is to gather information on
comparable world prices. Since gathering price data for hundreds of items will be a
very time consuming and costly process, an alternative route can be taken. A subset
of manufacturing subsectors which currently play a central role in Bangladesh or are
expected to do so in the future can be chosen. The collection of price data for a
smaller subset of goods will be much easier and improve the results substantially.
Another criteria to identify the subsectors should be the quality of ISS data for those
subsectors. An important first step before starting to gather price data is the
identification and coding of products and inputs. Unless the names of products and
inputs are identified there is no way to be confident about the results.
Using nominal protection rates for two years along with the ISS data, the paper
showed that there has been partial success in tariff liberalization efforts. Bangladeshi
government should continue with tariff reforms in the future, especially in the Steel
and Engineering, and Food and Dairy sectors. In addition, the emphasis of future
changes in the tariff structure should focus on reducing the tariff dispersion, so as to
reduce the effective protection from its current 67% .
Another important policy implication of the results is the protection received
by domestic sales of export oriented sectors. Although enterprises in export sectors
receive substantial incentives to export, their domestic markets are continued to be
protected heavily from import competition. This policy continues to distort the
incentive structure towards domestic market and hence reduces the potential for
improved efficiency in export sectors. This was especially true for RMG, and tanning
& leather, which used to receive an effective protection of 222 and 127 percent,
respectively, in FY 1993/94. Thanks to futher tariff changes in subsequent years,
trade-distortions in the export sectors has been reduced substantially.
Another result of the analysis lends support to claims surfaced in the
Bangladeshi press about negative effective protection received by allopathic
medicines (drugs & pharmaceuticals). The results support the view that domestic drug
companies are not protected from foreign competition for the FY 1993/94. The
situation has somewhat changed by FY 1995/96, when the ERP was 8%. Although
the result of the paper is in line with charges surfaced in the press, it is not correct to
jump to the same conclusion. Rather, it makes sense for the government to insure the
availability of essential drugs in the domestic market. Since domestic companies may
not be able to produce some of these essential drugs, it is understandable to observe
no protection or even negative protection.
16
Finally, on the basis of cross-country comparisons of effective protection it is
argued that the government needs to realign its tariff structure further to reduce
disincentives for an export oriented growth. It is too early to decide whether the recent
changes in tariff structure has contributed to the realization of this objective. One
certain point, however, is that other countries and expecially India has moved swiftly
in recent years in opening up domestic markets to import competition. This has
certainly increased the pressure on Bangladesh to intensify its efforts towards a less
protectionist trade regime.
17
APPENDIX
Methodology of Effective Protection and Bangladeshi Data
In this appendix I present methodological background of the analysis of
effective protection (EP) in Bangladesh. The appendix starts with a simple,
generalized framework to discuss the theoretical underpinnings of the EP analysis,
where only traded inputs are considered. Following this introduction, various
complicating factors such as non-traded goods and depreciation are incorporated into
the framework and Corden and Balassa approaches in handling the non-tradeables are
discussed. After a theoretical review of the EP analysis, I discuss data requirements
for the analysis and the availability of data, especially on economy-wide variables,
such as the price of non-tradeable inputs, the use of non-tradeable inputs in the
production of non-tradeable goods, and the valuation of fixed assets and depreciation.
Finally, modifications in the analysis due to lack of data are summarized.
I. A Simple Framework for the Effective Protection Analysis
In a market economy prices play a key signalling role, providing producers
with information about market conditions. Through their impact on relative prices,
tariffs and other indirect taxes alter the incentives faced by the producers. The
effective rate of protection, in essence, measures the degree of the change in the
incentive structure, taking into account backward and forward linkages accross the
sectors. Effective protection analysis falls short of studying the effect of all forms of
government interventions, including subsidies and tax holidays, etc., on producers'
incentives. The same methology can be extented to the analysis of effective assistance
received by the industry.
Effective rate of protection in activity i is defined as the difference between the
value added at domestic prices (VAd) and the value added at world prices (VAw)
relative to the value added at world prices:
d
ERPi =
w
VAi - VAi
(1)
w
VAi
This definition of ERP is vacuous as it stands. To see the link between the
ERP and the tariff regime, one needs to define VAd and VAw incorporating the
nominal tariff rates on inputs and the output, which effectively determines the
difference between the VAd and VAw. Value added per unit of output i at domestic
prices and world prices are respectively defined as
N
VAi = Pi [(1 + ti ) - ∑ (1 + t j ) a ji ]
d
j =1
N
(2)
VAi = Pi [1 - ∑ a ji ]
w
j =1
18
where aji is the share of tradeable input j in cost of production of one unit of output i
under free trade, ti and tj are the nominal tariff rates on output i and input j, and Pi is
the unit output price under the free trade.
There are three underlying assumptions implicit in the definition of value
added in equation (2) which carry on to the analysis of effective protection. First, the
country in question is assumed to be a small country with no effect on world prices.
In other words, the elasticity of supply of importables (both the output and the inputs)
are infinite. This assumption assures that imposition of the tariff on various products
have no effect on the world price of those products. Secondly, both domestic and
foreign markets are assumed to be perfectly competitive, such that unit cost of
production is equal to the market price of the output (Pi). Finally, it is assumed that
the production relationship between the output and the intermediate inputs can be
approximated by a Leontief-type fixed coefficient technology. This way the share of
each input in cost of production (aji) is fixed and independent from the changes in
relative prices. It amounts to assuming away the substitution posssibilities among the
inputs. Note, however, that this assumption does not extend to the primary factors of
production, such as labor, capital and land, which produce the value-added.
After substituting VAd and VAw from equation (2) in equation (1) and
rearranging the resulting expression, ERP can be written in terms of nominal tariff
rates and input coefficients:
N
ERPi =
ti - ∑ t j a ji
j =1
N
(3)
1 - ∑ a ji
j =1
In the absence of non-traded inputs ERP on any particular good depends on (i)
the nominal rate of protection on the output, (ii) the nominal rate protection on
tradeable inputs, and (iii) the share of each tradeable input in cost of production.
Since there may be more than one tradeable input used in the production process and
custom duties on inputs may vary considerably, whether any particular product
receives positive protection or not cannot be gauged from a simple analysis of the
nominal rates.
It is, on the other hand, possible to calculate the weighted average tariff rate on
inputs using the fixed input coefficients:
N
∑t a
j
ti =
ji
j =1
N
∑a
(4)
ji
j =1
Substituting the weighted average duty on inputs in equation (3) it is possible to
simplify the algebra of ERP so that:
19
N
ti - ti
∑a
ji
j =1
ERPi =
(5)
N
1 -
∑a
ji
j =1
Given that the value added at world prices is positive, one can easily determine
the sign of ERP on the basis of eqaution 5:
ERPi is positive, nill, or negative, when the nominal tariff rate on output is
greater than, equal to or less than the tariff rate on inputs. ERPi >,=,< 0, when
ti/tj >,=,< Σ aji.
Equation 5 can also be used to show the the relationship between the nominal
and effective rates of protection:
The effective rate of protection for good i is greater than, equal to or less than
the nominal tariff rate, when the nominal tariff rate itself is greater than, equal
to or less than the (weighted) average tariff on inputs to produce one (Taka
worth of or unit of) output. ( ERPi >,=,< ti when ti >,=,< ti)
Another possibility in the analysis of effective protection is to observe a
negative value added at world prices. In algebraic terms this means the denominator
of equation 5, 1 - Σ aji < 0, is negative. As a result equation 5 can not be used to
calculate the ERP, as it will obtain a negative value when the value added at domestic
prices is larger than the value added at world prices. However, one does not need
equation 5 to conclude that the ERP for this particular product is infinite given that the
value added at domestic prices is positive. Furthermore, one can easily conclude that
the production of the good in question is not economically viable and should be
ceased at once.
II. The Estimation of ERPs:
Until now I confined the analysis of effective protection into a simple
framework where it is easy to understand the relationship between various
components of the analysis. As already mentioned above, however, this simple
framework is not instrumental in the estimation of ERPs for several reasons. In the
rest of this chapter I will introduce all dimensions of the methodology of the ERP
estimation incorporating non-traded inputs, various forms of indirect taxes and
depreciation of capital into the framework.
While keeping the definition of ERP in Equation 1 unchanged, the definition
of the value added is refined in order to capture the details of the production process.
a) Non-traded Inputs:
In developing the concept and the measure of effective protection above I
assumed, for simplicity, that only traded inputs are used in the production of the final
good which is subject to protection. In reality non-traded inputs may also be used in
production. When non-traded inputs are incorporated in the analysis, however, the
definition of ERP in equations 2 through 5 does not carry through.
20
One of the conditions underlying the simple framework of the EP analysis will
be violated in the presence of non-traded inputs: supply of non-traded inputs are not
infinitely elastic. This is true irrespective of the small country assumption. When this
is the case the imposition of a tariff on the output will lead to an increase in the
demand for and the price of non-traded inputs which are used intensively in the
production process. As a result the supply of non-traded inputs will increase. This
effectively means that in the presence of non-traded inputs the protection on the
product is extended to activities producing non-traded inputs as well.
What complicates the analysis further is the use of traded inputs in the
production of non-traded goods which are themselves used as inputs. Treating nontraded inputs in the same way as value added, which is equivalent to assuming there is
no possibility of using traded goods in the production of non-traded goods, will
introduce a bias in ERP estimates.
For example, petroleum being a traded input is used in the production of
electricity, a non-traded input, which is used in the textile industry. For simplicity let
us assume that there are no tariffs with the exception of a 50 percent tariff on
petroluem imports, increasing the cost of electricity production and its domestic price
above the level which would have prevailed under the free trade. Higher domestic
electricity price will push the cost of production of textiles above the prevailing world
price. If I do not take higher domestic electricity price into account I will obtain an
ERP estimate of 0, indicating a neutral trade policy. When the higher domestic price
of electricity is incorporated in ERP estimations, however, ERP will certainly turn out
to be negative.
An alternative approach is the Balassa method which assumes that non-traded
inputs are no different from traded inputs, and should be excluded from the value
added. The Balassa method will also produce biased ERP estimates, because it
ignores the portion of non-traded inputs which are produced using other non-traded
inputs. Going back to our example in the previous paragraph, this is nothing but to
assume that the cost of producing electricity will increase by 50 percent due to a 50
percent tariff on fuel imports. However, this is not the case as long as fuel costs
account for less than 100 percent of the cost of electricity production. Despite its
methodological deficiencies, the Balassa method is widely used in the estimation of
ERPs, because detailed data on non-traded inputs and their traded components are
very difficult to obtain.
A more plausible approach to the estimation of ERPs in the presence of nontraded inputs is the Corden method. The Corden method breaks down the non-traded
inputs used into its traded goods and the value added components. The traded good
component of the non-traded input is treated in the same way as other traded inputs in
the original activity. The value added component of the non-traded input, on the other
hand, is added to the value added produced in the original traded activity. This way
any protection which filters back into the production activity of non-traded inputs will
be accounted for.
b) Depreciation:
21
Another factor complicating the analysis of effective protection is the
treatment of depreciation. The capital stock is not homogeneous and composed of
various capital goods, such as machinery and equipments, buildings and structures,
etc., some of which are non-traded goods. Non-homogeneity of the capital stock
forces us to choose one of the following methods to account for depreciation in the
estimation of ERPs.
One approach is to consider the depreciation adjustments as part of the return
to capital stock because of the loss of the value of capital stock used in the production,
in which case depreciation is treated as part of the value added. There are two
problems with this approach. The first is conceptual. Depreciation is the amount of
resources which is put aside to be used in replacing that part of the capital stock lost
during the production process and not a return to capital. Secondly, by including
depreciation payments in the value added one implicitly assumes that capital goods
are non-traded. While the installed capital goods can be considered as non-traded,
new capital goods which are in theory to be purchased to keep the capital stock intact
can be a tradeable or non-tradeable good.
An alternative approach, which is also used in the TIRS-I report, is to treat
depreciation as part of traded inputs and subtract the depreciation payments from the
value added. This approach is no better than the first one, for it ignores the fact that
buildings and structures are non-traded and the replacement of these capital goods will
increase the demand for construction services as well as the demand for traded inputs
to construction.
A more desirable method, which is adapted below, is to distinguish between
traded and non-traded components of the capital stock and account for the
depreciation of the seperate components appropriately in the estimation of ERP. This
approach to depreciation is also consistent with the Corden method. Depreciation of
the non-traded components of non-traded capital goods will be treated as part of value
added, whereas the depreciation which corresponds to traded goods used in the
production of non-traded capital stock, such as cement used in buildings and
structures, will be treated as tradeables. Depreciation of traded capital goods will be
treated as part of traded inputs.
The problem with this approach is the availability of detailed data to calculate
the traded and non-traded components of the capital stock. However, as will be
discussed below, it was possiblw to gather detailed information to estimate ERPs.
Using the Corden method in the case of non-traded goods and the quasi-Corden
approach to depreciation, the value added at domestic prices and value added at world
prices are redefined:9
9
In the rest of this appendix we drop the product subscripts i and j.
22
d d
VAd = Od-Id-NT -D -T
T T
w w
VAw = Ow-Iw-NT -D
T T
(2')
where superscripts w and d denote the values evaluated at world and domestic prices,
respectively. Furthemore, O is the value of total production, I is the value of traded
inputs, NTT is the value of traded inputs to non-traded goods which are themselves
used as inputs in the activity, DT is the amount of depreciation of the tradeable
component of the capital stock such as machinery and equipment. Finally, T denotes
net domestic indirect taxes and is defined as the value added tax plus the
supplementary duty on the product of the firm minus duty drawbacks for exporters.
The value of output at domestic prices (Od) is inclusive of value added tax and the
supplementary duty on output.
III. Valuation of Goods at Domestic and World Prices:
a) Tradeables:
While the domestic price of any tradeable good is equivalent to its ex-factory
price, there are two alternative ways to obtain the value of tradeables at world prices.
The best possible data for this purpose is the CIF value of the the tradeable good. A
comparison of domestic price and the CIF value provides us with an accurate measure
of the actual level of protection. However, firms may choose not to provide
information on CIF values. In that case, one can use the nominal tariff rate and other
indirect taxes discriminating against imports to deduce the world price of the tradeable
good from its domestic price:
Ow = Ocif, when the CIF price is available;
Ow = Od - Tariffs - net indirect taxes, when the CIF price is not available.
Using the CIF value is preferred to the use of nominal tariff rates because
nominal tariff rates announced by the government are not always binding. When the
trade policy is very restrictive it is possible to observe an increase in the smuggling of
goods whose domestic production is heavily protected. Smuggling will have a direct
impact on the effectiveness of the trade policy in realizing the government's
objectives. The domestic price of a tradeable good will effectively be lower if it is
possible to import this good through illegal means. A second reason for preferring the
use of CIF values is the extensive use of quantitative restrictions as a means of
protection. When the quantitative restriction on imports is binding the domestic price
of the tradeable good will be higher than the world price plus the tariffs. Therefore, in
the presence of smuggling and/or binding quantitative restrictions the world price
estimate obtained by subtracting tariffs and indirect domestic taxes from the domestic
price will be biased.
While it is desirable to obtain the CIF value for each product, it is also
important to make use of the distinction between the CIF price and the world price
23
obtained from using the NPR in analyzing the impact of smuggling on the
effectiveness of the government policies. One can calculate the ERPs using both
definitions of the world price. The ERP corresponding to the CIF price shows us the
actual level of effective protection. In contrast, the ERP based on the world price
calculated indirectly using the nominal tariff rates is a measure of the level of effective
protection which was intended by the government.
b) Non-tradeables:
The domestic price of non-tradeables is obtained from the ISS survey database.
However, it is impossible to collect the border prices of non-tradeables from the firms
surveyed. This information gap is filled with data from a study by O. Shahabuddin
and K. Mustahidur Rahman, with the title "Estimation of Economic Prices of Selected
Commodities for Use in FAP Planning Studies" (April 1992). Below I present the
conversion factors (the ratio of border to domestic prices) for major non-traded goods
which are calculated on the basis of data from this study:
i) Building and Structures: 0.637
ii) Selling and Distribution Costs: 0.893
It is an unweighted average of the ratios for transport and trade, which are 0.79
and 1.019, respectively.
iii) Water, Electricity, Telephone and other utilities: 0.877
Since these items are part of non-industrial costs, they are the expenses in
headquarters rather than in the factory. Therefore, it is acceptable to use the
ratio for the commercial electricity use category rather than one of industrial
use categories. The distinction between the industrial and non-industrial use
of electricity is important because of the existence of subsidy for the industrial
use of electricity, which is not enjoyed by other commercial users.
iv) Electricity used in production :
It is divided into three categories so that it is possible to apply different ratios
to different user. This way it is possible to distinguish the price of electricity
used by a large chemical complex (high voltage) from the price of electricity
used by a smaller firm.
Small industry: 0.971
Medium voltage: 1.010
High voltage : 1.136
v) Repair and Maintenance:
If the firm provides information on the division of repair and maintenance
expenditures across the subcategories (Machinery and Equipment, and
Buildings and Structures) then I use conversion factors for the corresponding
type of capital. If this information is not provided repair and maintenance
charges are classified in subcategories on the basis of the share of each type of
24
capital in total capital stock, and then the conversion factors are applied to
each category seperately.
vi) Rent Payments:
To account for rent payments as part of non-industrial costs, I use the data on
rent payments for leased property provided in Section 4 of the ISS survey. A
similar question on rent payments was also asked in section 8 of the survey.
However, the latter is different from the former in that it includes the
opportunity cost of land, building and others as well as the actual rents paid by
the firm for leased property. Since the effective rate of protection is calculated
on the basis of market prices, rather than the shadow prices, I use the rent
payment figures provided in section 4.
IV. Non-traded Inputs In the Production of Non-traded Goods:
In our framework only three categories of non-traded inputs are assumed to be
produced by a combination of non-traded and traded inputs. These categories are
water, electricity, telephone and others (EL); selling and distribution activities (S&D),
Repair and Maintenance (R&M); and building and structures (B&S) which is part of
the capital stock. Other non-traded inputs such as interest and other bank charges;
advertising, electricity and other utility charges for non-industrial use; rents; and other
facilitation expenses are assumed to be produced by using non-traded inputs only.
Share of traded and non-traded inputs used in the production of non-traded
inputs are calculated using information from various sources, including 1981/82
Input-Output Table and Shahabuddin and Rahman (1992).
a) Electricity:
Table A1, compiled from Shahabuddin and Rahman (1992), identifies three main cost
items in the production of electricity. Assuming that the unit operation costs and
maintenance services are only non-tradeable inputs, the share of traded inputs in
electricity production is estimated to be 90.3 percent.10
Table A1
Cost of Electricity per Kwh (1985 border prices, Taka)
Annualized Capital Cost
1.32
Fuel Costs
0.28
Unit Operation and Maintenance Cost
0.17
Total Costs
1.77
10
Alternatively, we use data from 1981/82 I-O Table to calculate the share of traded inputs in the
production of electricity. Including value added, the share of non-traded goods amounts to a relatively
high value of 29.9 percent, compared to 9.7 percent estimated on the basis of data from the FAP
Report. We decided to use 9.7 percent in our ERP estimations because it is based on more recent
information.
25
0.903
Share of Traded Goods In Total Costs (αEL)
b) Selling and Distribution:
Since it was not possible to obtain more recent information on selling and
distribution activity I use data from 1981/82 Bangladeshi Input-Output Table,
published in 1990 by Planning Commission. The share of traded inputs in S&D
activity are obtained using the input and value added data for trade, and transport
sectors of the I-O table. While the share of traded inputs is quite low in the trade
sector (2.8 percent), it is balanced with a relatively high share in the transport sector
(27.4 percent). When these two sectors are combined to form the Selling and
Distribution activity, the share of traded inputs in total production costs (αSD)
amounts to 18.3 percent.
Table A2
Share of Traded Inputs in Selling & Distribution Activity
Trade
Transport
S&D
Non-Traded (mil.Tk)
15790
21484
37272
Output (mil. Tk)
16040
29598
45638
Traded Component
0.028
0.274
0.183
The share of traded inputs in the total of non-traded goods is calculated by
adding up the traded input used in the production Electricity (αEL) and selling and
distribution activities (αSD):
NTT = 0.903*EL + 0.817*SD
c) Building and Structures (B&S):
The share of traded inputs in Building and Structures is estimated using the
data from 1981/82 I-O Table for the Other Construction sector, rather than the Urban
and the Rural House Building sectors. According to the data 80 percent of the output
of Other Construction sector is accounted for by traded inputs. Assuming that B&S is
produced mainly by this sector,
αBS = 0.80.
V. Valuation of Fixed Assets and Depreciation
In the measurement of depreciation, I use standard depreciation factors for
three components of the capital stock:
Land (L): zero percent
Building and Structures (B&S): 5 %
Machinery and equipment (inc. transport eq.) (M&E): 12.5 %
26
Before applying the rate of depreciation on each type of capital I calculated the
traded input component of building and structures by multiplying it with the traded
component coefficient calculated in the previous section. Since land depreciation is
zero, there is no need to consider its traded component. Similarly, I do not consider
the components of M&E because overall it is a tradeable capital good. As a result, the
traded component of depreciation is
DT = 0.05*0.80*BS + 0.125*ME
The most difficult step in estimating depreciation is the valuation of fixed
capital assets. There are several factors which affect the valuation of fixed assets
through time. As long as the capital is used in the production process it depreciates.
However, at an aggregate level one can only determine some approximate rate of
depreciation as it is very difficult to obtain data even at a micro level. Secondly, the
price of the installed capital increases through time along with the price of new capital
goods. However, there is no way to determine the exact amount of increase in the
value of capital as the technological progress and change in quality may in the
extreme lead to obseleteness of machinery and equipment.
In order to obtain an accurate data on the value of capital assets, ISS
manufacturing survey included questions on the current purchase price as well as the
book value of the capital assets. However, during the data editing process it is
realized that the respondents' guestimates for the current purchase price do not seem to
be very realistic. Consequently, I decided to calculate the current value of the asset
based on historical cost accounting method. The historical cost of the capital asset is
calculated by adjusting the initial cost of the capital asset upwards by the amount of
inflation, and downwards by the amount of depreciation. As a result of these
adjustments, the net percentage change in the value of the asset is equal to the rate of
inflation minus the rate of depreciation.
In order to account for the heterogeneity of capital I used different price
deflators for each group of capital asset. For land I use the price index for the urban
industrial land when available, otherwise I use the general price index. In the case of
Buildings and Structures I use the cost of construction index, and in the case
machinery and equipment I used the price index for industrial products.
VI. Adapting the Effective Protection Analysis to ISS Survey Data
Sections I through III presented the underlying methodology of effective
protection analysis without any consideration of the data availability. In the rest of the
appendix, I present adjustments to the methodology to circumvent the problems
imposed by the data. In doing so, I go deeper into a detailed presentation of the
computation of the variables which are directly or indirectly used to estmate the net
value added at domestic and/or at world prices, as given in equation (2') .
27
Starting with the order of importance I first present the definition of the value of total output at
domestic prices (Od):11
O d = ∑ Oi +
d
i
∑O
i
d
+ Gd + H d
i
(6)
where Oid is the product of the sales price of good i (Pid) and its output (Qi), both of
which are taken from section 9 of the survey. Whenever the sales price of the product
is not provided by the respondent I use the sales value (Sjd), which is not necessarily
equal to Ojd. In equation (9) The products for which the sales price is not provided
are denoted by a subscript j. The value of total output also includes the value of work
in progress (Gd) and the value of services rendered to others (Hd).
The second variable in the estimation of ERPs is the value of output at world
prices (Ow), which can not directly be obtained from the survey. I compute Ow using
the data on the value of output at domestic prices, import taxes and indirect domestic
taxes. If a firm reports its indirect domestic tax payments (VAT and supplementary
duty (SD)) then I use this data and the customs duty rate to obtain the value of the
output at world prices. Unfortunately, the data on indirect tax payments, which I
denote by VAT+SD, is scarce and to alleviate this problem I calculate Ow using the
VAT and SD rates obatined from the Operative Tariff Schedule:
Oiw =
Oid
(1 + vi ) + 0. 025 if VAT + SD = 0
(1 + ci + si )
(7)
Oid
= [O −
(VAT + SD)] / [1 + ci + 0. 025], otherwise.
∑ Oid
d
i
i
where c, v, and s are respectively customs duty, VAT and supplementry duty rates.
The additional import tax 0.025 corresponds to the import license fee. I also need to
include Advance Income Tax (AIT) as an import tax. When the data on total VAT
and SD payments by the enterprise are available, VAT and SD payments
corresponding to each product is obtained by multiplying the total VAT and SD
payments with the value share of each product. The value of output at world prices is
then equal to the net value of output (net of indirect domestic taxes) divided by
(1+ci+0.025).
The value of total output at world prices is obtained by adding all the
components of the output as in the case of
Ow =
∑ Oi +
w
i
∑ Sj +
w
j
Gd
Hd
+
1+ t
1+ t
(8)
11
It is important to remember that we calculate the ERPs at a sectoral level rather than for each
product.
28
The value of work in progress (G) and the services rendered by the enterprise (H) at
world prices is calculated by deflating their respective values at domestic prices by an
average tax rate ( τ ) which incorporates the domestic indirect tax rate as well as the
import tax rate.12
As it is already pointed out above, indirect domestic tax payments are not
reported by the majority of the enterprises, in which case I need to compute the tax
payments using the available data. One complicating factor, however, is that the base
of indirect domestic taxes is the ex-factory price which is not available, either. Below
I first compute the value of the output at ex-factory prices (Ox) using the definition of
the value at domestic prices (Od):13
d
d
. x
Oi
Oi
d
vi
d
Oi = [
+ (
)(
+ NTT )] / (1+ v i )
d ) (I
1 + ci
1 + vi ∑ Oi
i
x
i
VATi + SDi = si O + vi (1 + si )[Oix −
I d + NT d
]
1 + vi
(9)
(10)
I obtain a measure of the domestic indirect tax payments (VAT+SD) on a product
basis by multiplying the value of the product at ex-factory prices with the
corresponding VAT and SD rates
VII. Nominal Protection on Packing Materials and Fuel:
i) Packing Materials:
Although the ISS survey was conducted at an enterprise level it was not
always possible to get a detailed description of each input as well as
each product. This difficulty was most evident in the case of packing
materials, which was a seperate input item for which the information
was sought. However, the survey questionarie did not ask for the
descripton of the packing materials for each sector. Since I donot have
detailed information on the nature of packing materials I are forced to
use an average tariff rate for packing materials irrespective of the
sector they are used as an input. This rate is calculated as an average of
operative tariff rates of all HS products which can be classified as
packing material. While the use of an average tariff rate would
introduce a source on noise in our ERP calculations I intend to address
the sensitivity of our results on this assumption.
ii) Energy and Fuel:
12 One plus the average tax rate (tau bar) is equal to the ratio of the value of output at domestic and
world prices.
13
The value of output at domestic prices is defined as the value of output at ex-factory prices plus the
indirect domestic taxes: Oid = Oiw + VATi + SDi
29
In Bangladesh, petroluem, oil and lubricants (POL) imports are subject
to specific import duty which can not directly be used in ERP analysis.
In our analysis, I used an ad valorem duty rate for POL products which
is calculated as the ratio of customs duty payments on POL imports and
POL imports. Calculated as such this duty rate is an ex post, rather that
an ex ante, measure of nominal protection.
VIII. Domestic Resource Cost
As it is discussed above the effective rate of protection is a measure of the
extent that resource flows are distorted in the economy as a result of protection.
Despite its usefulness to provide us with the picture of the incentive structure faced by
the manufacturing industry, ERP calculations do not take into account other sources of
distortions prevalent in the economy. Its adequecy is especially questionable when
there are distortions in the product and factor markets which are not necessarily
created by the specific trade policy. For example, the analysis of effective protection
doesnot allow one to consider factor market imperfections resulting from minimum
wage legislation or repressed financial markets. Similarly, oligopolistic market
structures lead to divergence of market prices from the social marginal cost of the
product.
The fact that effective protection analysis does not correct for the presence of
market distortions renders ERP useful only as a measure of the net protection
accorded to domestic value added. However, the analysis of trade and industrial
policy measures also requires us to understand the "technical" efficiency and hence the
comparative advantage of the domestic industry, which is in part determined by longrun policy measures. This is where the domestic resource cost (DRC) coefficient
plays an important role.
DRC is defined as the opportunity cost of domestic resources which are used
to save (through the production of import substitutes) and/or earn (through exports)
one unit of foreign exchange. In other words, given the existing technology, DRC
measures how worthwile it is to use domestic factors of production and non-tradeable
goods in order to add further value to raw materials or semi-manufactured inputs to
earn or save foreign exchange. It can be written as the ratio of the value added at
shadow prices and the value added at world prices:
DRC =
VA s
VA w
(11)
At first glance the definition of DRC does not differ much from that of ERP.
The differences arise from the use of shadow prices rather than the market prices, and
the direct measurement of domestic resources which are used to produce the value
added rather than obtaining it as the residual of output after all tradeable inputs are
paid for. The use of shadow prices ensures that, unlike ERPs, the estimation of DRCs
excludes any economic surplus earned as a result of domestic market distortions
(tariffs, indirect taxes, subsidies) .
30
The DRC concept was initially developed along the lines of evaluating social
profitability of alternative investments and can be seen as an extension of the rate of
return criterion to the foreign exchange. However, unlike the use of rate of return
criterion it should not be used as a means to rank alternative investment projects in
different industries. Rather, it should be seen as a static efficiency criterion which, on
the basis of available information, can also be used to evaluate the social profitability
of replacement investment.
While ERPs are ex post, positive measures of protection accorded to value
added, DRCs can be used both as an ex ante, normative measure to evaluate the social
profitability of future investments as well as an ex post, positive measure of the
technical efficiency of the industry. (Bruno (1968))
s
DRC =
s
ws L + r s K F +K W
Es
VA w
m
E
(12)
where ws and rs are the shadow wage and interest rates, L is the employment level,
KsF is the opportunity cost of fixed capital at shadow prices and KsW is the working
capital at shadow prices. Finally Em and Es denote the market and shadow exchange
rates. The value added at world prices (VAw) in the denominator is defined in
equation (2'). The tilde denotes that variable inputs and output are adjusted to reflect
the full capacity utilization. The working capital is defined as the sum of the average
stock of inputs and output throughout the year.14
IX. Other indicators of Assistance and Efficiency
Since our ERP estimations are based on the tariff rates rather than the
comparison of domestic and world prices I calcualte measures of nominal protection
on output and inputs.
These are the net nominal tariff equivalent on inputs and on output and are measured
as
1. Net nominal tariff equivalent on inputs:
NTEI =
( I D + NT D + D D ) - ( I W + NT W + D W )
*100
I W + NT W + DW
(13)
2. Net nominal tariff equivalent on output
NTEO =
(O D - T ) - OW
*100
OW
(14)
The average inventory stock of each input and product is equal to one half of the sum of
inventoruies at the debinning and end of the year.
14
31
In addition to DRC and ERP I include two other measures of profitability and
efficiency, respectively: the financial rate of return and the economic rate of return.
The financial rate of return is based on domestic prices and measures the profitability
of the firm/industry given the existing protection regime. The economic rate of return,
on the other hand, measures returns to capital under world prices.
3. Financial Rate of Return
ρ=
(O d - I d - NT d - D d - T ) - Ld
* 100
d
d
KF + KW
(15)
The bracketed term in the numerator is the value added defined al a Balassa and
denoted as VAdB. Financial rate of return the profitability of an enterprise under the
existing trade and industrial policy.
4. Economic Rate of Return
ρE =
(O w - I w - NT w - D w ) - Lw
*100
w
w
KF + KW
(16)
Unlike financial rate of return, ERR measures the profitability of an activity when it is
valued at world prices rather than distorted domestic prices. It is, therefore, an
efficiency measure and very important for evaluating the performance of an industry.
32