passthrough of exchange rate and tariffs into import prices of india: currency depreciation versus...

Passthrough of Exchange Rate and Tariffs intoImport Prices of India: Currency Depreciationversus Import Liberalization*

Sushanta Mallick and Helena Marques

AbstractThis paper examines the extent of passthrough of exchange rate and tariff changes into import prices usingsectoral panel data (at the two-digit SITC level) for the post-reform period in India (1990–2001). Afterhaving controlled for unobserved effects that might have an impact on the import prices by using sectordummies, we find that on average exchange rate passthrough (ERPT) is a dominant effect compared to tariffrate passthrough (TRPT) in explaining changes in India’s import prices. The sectoral panel results suggestthat the passthrough of exchange rates and tariff rates varies across products. ERPT into import prices issignificant in 12 industries, whereas TRPT is significant only in six industries, with full passthrough. However,ERPT is incomplete only in four industries, but TRPT is incomplete in 36 industries, which means that firmsexporting to India more frequently adopt strategies to maintain their market share against tariffs than againstexchange rate changes.The sectoral differences in passthrough seem to be related to the sector’s share in totalimports and the sector’s effective protection rate. Hence, India’s relatively high levels of protection have animpact on the behavior of foreign exporters.

1. Introduction

In recent years, the growing global external imbalances have motivated renewed inter-est to investigate the link between changes in a country’s exchange rate and the pricesof traded goods—the so-called exchange rate passthrough (ERPT) relationship.1 Thisbecomes even more important in emerging market economies undergoing trade liber-alization and adopting floating exchange rate systems, which seem to have revitalizedthe potential impacts of exchange rate movements on traded goods prices. Given itsimplications for a country’s terms of trade, the evidence on passthrough allows anunderstanding of trade imbalances between developed and emerging market econo-mies. Besides, the degree of ERPT is also critical for the assessment of monetary rules(Devereux et al., 2006), as changes in exchange rate can lead to a rise in import pricesand thus spur overall inflation. Furthermore, the response of local-currency prices ofimported products to changes in exchange rate may not be one-for-one, as has beendebated in the case of many advanced markets.

India’s move to a flexible exchange rate allowed gradual exchange rate depreciationto offset the effects of import liberalization and tariff reduction (Ahluwalia, 2006). This

* Mallick: School of Business and Management, Queen Mary, University of London, Mile End Road, LondonE1 4NS, UK. E-mail: [email protected]. Marques: Manchester Business School, University of Manches-ter, Booth Street West, Manchester M15 6PB, UK. E-mail: [email protected]. We gratefullyacknowledge the constructive comments made by four anonymous referees of this journal on an earlierversion of this paper. We have also benefited from comments and suggestions by participants at EasternEconomic Association annual meetings (New York, February 2007), the European Economics and FinanceSociety Annual Conference (Crete, Greece, May 2006), and the International Conference on MacroeconomicAnalysis and International Finance (University of Crete, May 2006).We are grateful for the useful commentsof Huw Edwards, Subrata Ghatak, David Greenaway, and Kunal Sen. The usual caveat applies.

Review of International Economics, 16(4), 765–782, 2008DOI:10.1111/j.1467-9396.2008.00774.x

© 2008 The AuthorsJournal compilation © 2008 Blackwell Publishing Ltd, 9600 Garsington Road, Oxford, OX4 2DQ, UK and 350 Main St, Malden, MA, 02148, USA

mailto:[email protected]

mailto:[email protected]

suggests that significant currency depreciation was needed to reduce import demandfollowing import liberalization.2 A relatively weak linkage between exchange rate andimport prices seems to be indicating a low level of passthrough, which requires a closerinvestigation of the pricing behavior of foreign exporters. Also, more flexible exchangerate regimes may neutralize the impact of any terms-of-trade shocks on the currentaccount (see Broda, 2004). Although India liberalized its imports by means of lowertariff barriers and elimination of quantitative restrictions, there are still sizable restric-tions, including other nontariff barriers (NTBs) to imports,3 which may diminish theability of the flexible exchange rate regime to neutralize the terms-of-trade shocks.The simultaneous trade liberalization and change of exchange rate regime includedin the reforms of the 1990s thus makes India an interesting case study to investigate therelative contribution of exchange rate depreciation and reduction of trade barriers tothe determination of import prices. Besides, India may serve as an example to otherdeveloping countries that are trying to internationalize their economies and implementliberalizing reforms.

In this paper we are particularly interested in assessing the relative importance ofERPT and tariff rate passthrough (TRPT) into sectoral import prices following thebalance of payments (BOP) crisis in 1991 and the consequent trade liberalization andintroduction of a flexible exchange rate regime in India.4 We further investigatewhether ERPT and TRPT are more pronounced in some sectors, after having con-trolled for a range of unobserved factors captured through sector-specific dummies.5

Existing empirical evidence shows that the deviations from the law of one price6 arelarge and persistent,and the phenomenon of ERPT is country- and even product-specific(Athukorala and Menon, 1994; Bleaney, 1997; Feenstra, 1989; Froot and Klemperer,1989; Gagnon and Knetter, 1995; Goldberg, 1995; Hooper and Mann, 1989; Kim, 1990;Knetter, 1989, 1994; Koch and Rosensweig, 1992; Parsley, 1993; Tange, 1997; Yang, 1997,1998).However,most of the existing studies have looked at the behavior of firms in largehigh-income countries, either US importers, or Japanese and German exporters practic-ing pricing-to-market. Overall, these studies conclude that Japanese and Germanexporters tend to accommodate exchange rate changes, whereas US exporters keepmargins constant and passthrough any exchange rate changes. A second generation ofstudies has dealt with smaller countries:South Korea (Athukorala,1991;Lee,1997;Yangand Hwang, 1994), Australia (Menon, 1992, 1996), Switzerland (Gross and Schmitt,1996), and Ireland (Doyle, 2004).With global integration and trade reforms, incompletepassthrough can also be feasible in emerging markets.7 Recently, Frankel et al. (2005)have examined the passthrough into import prices of eight selected narrowly definedbrand commodities exported by 76 developing countries, reporting a downward trend inERPT. Nevertheless, there is limited evidence in the case of developing countries for abroad spectrum of products. Besides, there are only two studies in the literature thatdiscuss both TRPT and ERPT (Feenstra, 1989; Menon, 1996).

The present paper thus fills a gap in the literature: first, by examining the disaggre-gated sectoral ERPT effect using a panel of two-digit SITC-level products for India asan emerging market economy during 1990–2001; secondly, comparing the relativeimpact of TRPT and ERPT, which becomes crucial to reflect the variation in pricingbehavior across industries and whether such variation would be indicative of any formof nontariff barriers still in place. The main findings can be summarized as follows. Inthe traditional framework of the law of one price and perfect competition, foreignexporters would fully passthrough the exchange rate changes. In the case of India as adeveloping country, we show that there is a significant degree of passthrough of therupee’s movements against a trade-weighted basket of currencies to the local currency

766 Sushanta Mallick and Helena Marques

© 2008 The AuthorsJournal compilation © Blackwell Publishing Ltd 2008

https://www.researchgate.net/publication/24066601_Exchange_Rates_and_Prices_of_Australian_Manufactured_Exports?el=1_x_8&enrichId=rgreq-fea82a8b-ba27-4ab2-b5ab-b3450bcc362d&enrichSource=Y292ZXJQYWdlOzQ3ODQzOTk7QVM6MTAzNTczNzMxOTM4MzA1QDE0MDE3MDUzMTAzMTE=

https://www.researchgate.net/publication/222658454_Terms_of_Trade_and_Exchange_Rate_Regimes_in_Developing_Countries?el=1_x_8&enrichId=rgreq-fea82a8b-ba27-4ab2-b5ab-b3450bcc362d&enrichSource=Y292ZXJQYWdlOzQ3ODQzOTk7QVM6MTAzNTczNzMxOTM4MzA1QDE0MDE3MDUzMTAzMTE=

https://www.researchgate.net/publication/24075423_Exchange_Rate_Pass-Through_in_a_Small_Open_Economy_The_Anglo-Irish_Case?el=1_x_8&enrichId=rgreq-fea82a8b-ba27-4ab2-b5ab-b3450bcc362d&enrichSource=Y292ZXJQYWdlOzQ3ODQzOTk7QVM6MTAzNTczNzMxOTM4MzA1QDE0MDE3MDUzMTAzMTE=

https://www.researchgate.net/publication/223126655_Symmetric_Pass-Trough_of_Tariffs_and_Exchange_Rates_Under_Imperfect_Competition_An_Empirical_Test?el=1_x_8&enrichId=rgreq-fea82a8b-ba27-4ab2-b5ab-b3450bcc362d&enrichSource=Y292ZXJQYWdlOzQ3ODQzOTk7QVM6MTAzNTczNzMxOTM4MzA1QDE0MDE3MDUzMTAzMTE=

https://www.researchgate.net/publication/227348327_'Slow_Pass-Through_Around_the_World_A_New_Import_for_Developing_Countries'?el=1_x_8&enrichId=rgreq-fea82a8b-ba27-4ab2-b5ab-b3450bcc362d&enrichSource=Y292ZXJQYWdlOzQ3ODQzOTk7QVM6MTAzNTczNzMxOTM4MzA1QDE0MDE3MDUzMTAzMTE=

prices of the Indian importers. The exchange rate used is a trade-weighted average ofnominal bilateral exchange rates against India’s main trading partners8 that account forthe bulk of transactions (NEER—nominal effective exchange rate). Compared toERPT, TRPT is significant less often (5 against 12 sectors), but incomplete more often(35 against 4 sectors). These results hold after having controlled for unobserved effectsusing industry-specific dummies. In 34 sectors foreign exporters completely passthroughthe exchange rate changes, but in 36 sectors exporters absorb tariff changes to varyingdegrees. Although it might be to the benefit of foreign exporters to refrain from fullypassing through exchange rate shocks to the local currency price of Indian imports, theirreaction is sector-specific.The results suggest that foreign exporters can, to some extent,manipulate the foreign price of their exports, but they react more to tariffs than toexchange rates. In other words, foreign exporters appear to adjust their profit marginsto tariff rates by changing prices in their own currency, and the relative sensitivity of theforeign currency prices is translated into incomplete passthrough.

The remainder of the paper is organized as follows. Section 2 describes a simplemodel of ERPT into import prices in the presence of tariffs, from which an empiricalspecification is derived. Section 3 discusses the data and model specification. Theestimation results are presented in section 4, with alternative explanations for thesectoral passthrough variation in section 5. A summary and discussion of implicationsof the findings are provided in section 6.

2. A Conceptual Framework of Exchange Rate Passthrough with Tariffs

The study of ERPT, defined as the elasticity of import prices induced by a change in theexchange rate, goes back to the 1970s (see, for example, the survey in Goldberg andKnetter, 1997). Empirical studies have provided substantial evidence of incompleteERPT (see Menon, 1995, for an earlier survey). This phenomenon is made possible byimperfect competition and the associated markup pricing: when the exchange ratechanges, exporters change the price in their own currency to stabilize their export pricesin the importer’s currency. This exporter pricing behavior framework is our startingpoint in order to examine ERPT into import prices. In theoretical terms, the phenom-enon can be explained through a markup model (Campa and Goldberg, 2005; Gagnonand Knetter, 1995). This model is based on the definition of the price of exports inforeign currency as the product of marginal cost and a markup. The firm’s profits willequal the difference between its revenue and its cost:

∏ ∑ ∑= +( )( ) − +( )( )⎛⎝

⎞⎠= =

P q e T P C q e T P wix

i ix

i

n

i ix

i

n

1 11 1

, , (1)

where w is an index of input prices, including the imported raw materials, q is thequantity demanded of exports, which can be assumed as a function of the export price(price in exporter’s currency) relative to the price level in the destination market, e isthe exchange rate defined as the domestic currency (e.g. rupee) price of foreigncurrency (e.g. US dollar); T is the unit tariff rate.

Assume that the firm’s external demand changes as the exchange rate changes. Tomaintain competitiveness, the representative exporter may be constrained to keep theprice of its products in its own currency stable despite exchange rate fluctuations. Thismeans that the exporter would maximize its profit function by setting its export priceas a markup over the production cost, where the exchange rate is assumed to determinethe profit markup at a given price elasticity of external demand. Taking the first-orderderivative of equation (1) with respect to Px, the following expression is obtained:

EXCHANGE RATE AND TARIFFS PASSTHROUGH 767


P MCe T P

e T Pi ni

x i ix

i ix

= +( )( )+( )( ) −

⎡⎣⎢

⎤⎦⎥

=ηη

11 1

1, , . . . , , (2)

where h is the absolute value of the price elasticity of demand in the foreign market. Atypical exporting firm sets the price of a good as a constant markup over marginal costs.As external demand increases, the exporting firm is likely to charge a higher markupover its marginal production cost, if products are differentiated under an imperfectlycompetitive market condition. This means that if 0 < hi < 1, the foreign currency exportprice could go up or remain stable via markup adjustment.The extent to which the localcurrency import price will increase is an empirical question and product-specific.Lowering markup, as a function of exchange rate, would result in an increase in thelocal currency import price, indicating incomplete passthrough. Using log-linear differ-entiation, equation (2) can be written as:

d P d MCd

d e T P

d P d eT

Td T

ix i

ix

ix

i

ln lnln

ln

ln ln ln= −

+( )( )+ +

+−

⎛

⎝η

η11

1⎜⎜⎜

⎞

⎠⎟⎟ . (3)

Collecting terms for d Pixln on the left-hand side yields the following testable

equation:

d P d MC d eT

Td Tit

xi i i tln ln ln ln= −( ) − +

+( )11

δ δ , (4)

where

δ η η ηi

i

ix i

i

ixe T P e T P

=+( )( ) − +

+( )( )⎡⎣⎢

⎤⎦⎥

−∂∂

∂∂

lnln

lnln1

11

1

is a function of both the level and the elasticity of hi. The coefficient d is a coefficientof pricing-to-market and influences both ERPT and TRPT. When the exchange rateor tariff rate elasticity of the export price (d) is zero, there is complete passthroughwith both the demand elasticity for exports and the marginal costs constant. Ifneither the export demand elasticity nor the marginal cost of production is constant,the elasticity of the export price will range between 0 and 1. If d = 1, exporters fullyabsorb exchange rate changes and no passthrough to importing currency prices willtake place.

The dependent variable is the price in the exporter’s currency and, assuming mar-ginal costs are independent from the importing markets, it represents the exporter’smarkup. The relationship between foreign currency export prices (Px) and domesticcurrency import prices (Pm) can be written as Pm = ePx.This means the import prices forany country are a transformation of the export prices of that country’s trading partnersusing the exchange rate. Taking logs and differentiating:

d P d e d Pm xln ln ln .= + (5)

Substituting (4) in (5), the equation to be estimated can be written as follows:

d P d e d Titm

i i t i itln ln ln= + −( ) +τ δ β1 , (6)

where ti = (1 - di)d ln MCi is a sector-specific term,9 1 - di is the ERPT coefficient, andbi = -di[T/(1 + T )] is the TRPT coefficient.



Equation (6) suggests that a depreciation of the rupee (increase in e) must result ina rise in India’s import prices, of the same magnitude, unless there is a decline in theforeign producer prices via reduction in markup (d). So the ERPT coefficient is goingto depend on the markup parameter (d), and as long as markups vary with exchangerates, passthrough will be incomplete. If d = 0, producer currency pricing (fullpassthrough) takes place; if d = 1, local currency-pricing occurs and the exportersabsorb the exchange rate changes in their own markups (see Campa and Goldberg,2005, for a similar interpretation). Between the two extremes, there is the possibility ofincomplete passthrough. On the other hand, with tariff reduction in the context of tradeliberalization, import prices will decrease by the same percentage if the foreign pro-ducers do not adjust their markups. Note that the TRPT coefficient (b) will be zero witheither free trade (zero tariffs) or full ERPT (d = 0).10 In India’s context, tariffs tend tobe very high and so b will be close to one independent of the degree of ERPT. If foreignexporters were to react to tariff reductions, they would have the opportunity of increas-ing their markups, thus partially offsetting the tariff reduction. They might, however,have an interest in keeping their prices constant to maintain competitiveness. What wemay find in the case of India is an empirical question.11

The passthrough coefficients are crucial estimates to gauge the pricing behavior ofexporters in different products. The extent of ERPT and TRPT depends on the level ofmarkups and product differentiation, which influence the degree of imperfect compe-tition. In other words, product differentiation gives the firm a degree of monopoly, andit is this monopoly power that allows the firm to use the markup approach to pricedetermination. Exchange rates and tariffs influence markups and thus export prices.In turn, industry-specific characteristics—namely, product differentiation, marketshares, and economies of scale—are important determinants of the degree of ERPTand TRPT into import prices of the exporters’ destination market.12

3. Data and Empirical Framework

The unit value indices of imports for a number of sectoral groups are regressed againstthe rupee NEER and the tariff rates so as to investigate the relative contribution ofexchange rate depreciation and tariff reduction into changes in the unit values ofimports. If foreign firms exporting to India price to market when the exchange rate ortariffs change, such pricing behavior will be reflected in the import prices measured inrupees as partial or incomplete passthrough. To test this, we use a sample of 38two-digit-level SITC (standard international trade classification) products in the period1990–2001.The detailed definitions and sources of variables are listed in the Appendix.

It should be noted that, in spite of the extensive reforms undertaken by India duringthe 1990s, the country remains very protectionist in terms of international standards.The simple average of tariff rates within each of the 38 two-digit sectors used in thispaper had declined from 213% in 1990 to 127% in 2001. Although this representssubstantial liberalization, the average tariff level is still very high in India. On the otherhand, in 2001 there was substantial variation across sectors, with tariff rates rangingfrom a maximum of 210% in Beverages and Organic Chemicals to a minimum of 0%in Cereals, Crude Fertilizers, Pharmaceuticals, Metalworking and Electrical Machinery,and Scientific and Photographic Instruments.

The role of each sector’s share in total imports is important in the context of India,as the composition of India’s import structure has been changing since the start of thereforms (Table 1). One of the main characteristics of a developing country is its depen-dency on intermediate goods. However, during the reform period there was a shift from



Petrol into Crude, indicating that India has developed refining capacity.The decrease inimportance of Iron and Steel in total imports also highlights the industrialization effortassociated with the reforms.

The empirical measurement of ERPT has been commonly carried out in a panel dataframework (Feenstra et al., 1996; Gagnon and Knetter, 1995; Goldberg and Knetter,1999;Knetter,1994;Madsen,1998).Referring back to equation (6), import prices dependon tariffs and exchange rates, as well as on sector-specific factors. Hence, the empiricalspecification for India’s imports of sector i in period t can be written as follows:

d P d T d eitm

it i i i t itln ln ln= + + +τ β μ ε , (7)

where d Pitmln is the change in the log of import prices in domestic currency (rupees),

d ln et is the variation in the log of the NEER (an increase indicates depreciation),d ln Ti is the change in the log of the tariff rate, t is the industry-specific dummyvariable, and the error term, e, is assumed to be independently and identically distrib-uted. The degree of passthrough to import prices will be analyzed from India’s point ofview. In the import price equation (7), if b = 0 or m = 0 (b = 1 or m = 1), there is no(complete) passthrough into India’s import prices as the rupee price of imports doesnot change (changes one-to-one) with the tariff rate or the exchange rate. If both b andd are strictly between 0 and 1, then there is incomplete passthrough to import prices.

In order to validate the first-differences theoretical specification in equation (6) forour sample, we test the integration order of each variable in the panel. For this purpose,we use the tests provided by Levin et al. (2002) and Im et al. (2003). All the variablesare nonstationary in levels in Model 2 (see Table 2), as the null hypothesis of a unit rootis not rejected, and thus the variables are used in first differences, with which we obtaina stationary panel.

4. Evidence for Sectoral Passthrough Effects in India

Tables 3 and 4 show the regression results for 1990–2001 using panel estimation withcommon sector coefficients (Table 3) and considering sector-specific coefficients

Table 1. Sectoral Share (%) of Main Import Categories in India’s Total Imports (1990–2001)

Product name 1990 1992 1997 1999 2001

Crude 14.97 17.77 10.74 19.34 25.42Petrol 11.62 11.08 9.91 6.64 2.26Nonmetallic Minerals 9.75 12.34 8.82 11.54 9.80Iron and Steel 5.22 3.80 3.59 1.96 1.65Organic Chemicals 3.56 3.14 4.44 3.16 3.17General Industrial Machinery 3.53 3.02 3.60 2.68 2.61Electrical Machinery 3.37 3.15 2.96 2.79 3.34Specialized Machinery 3.23 2.98 3.77 1.72 1.97Inorganic Chemicals 2.09 3.80 3.02 2.73 2.35Transport Equipment 2.83 1.28 1.60 1.44 1.67

SUM (1) 60 62 52 54 54SUM (2) 81 83 71 69 69

Notes: SUM (1) represents the import share of all 10 sectors in the table; SUM (2) represents the importshare of all 38 sectors used in the regressions. Own calculations using data from TRAINS as described in theAppendix.



https://www.researchgate.net/publication/4853964_Unit_Root_Test_in_Panel_Data_Asymptotic_Finite-Sample_Properties?el=1_x_8&enrichId=rgreq-fea82a8b-ba27-4ab2-b5ab-b3450bcc362d&enrichSource=Y292ZXJQYWdlOzQ3ODQzOTk7QVM6MTAzNTczNzMxOTM4MzA1QDE0MDE3MDUzMTAzMTE=

https://www.researchgate.net/publication/4854199_Testing_For_Unit_Roots_in_Heterogeneous_Panels?el=1_x_8&enrichId=rgreq-fea82a8b-ba27-4ab2-b5ab-b3450bcc362d&enrichSource=Y292ZXJQYWdlOzQ3ODQzOTk7QVM6MTAzNTczNzMxOTM4MzA1QDE0MDE3MDUzMTAzMTE=

Table 2. Panel Unit Root Tests (1990–2001)

H0: I(1)H1: I(0)

Levin, Lin, and Chu (2002) Im, Pesaran, and Shin (2003)

(1) (2) (3) (1) (2) (3)

LevelsLn(imprice) -0.078*** -0.499*** -0.969*** -2.538*** -0.807 0.770Ln(neer) NA NA NA -4.398*** NA NALn(tariff) -0.086*** -0.087 -0.567*** 1.072 6.532 0.951

First differencesDLn(imprice) -1.489*** -1.667*** -1.988*** -10.750*** -6.849*** -5.327***DLn(neer) NA NA NA -28.662*** NA NADLn(tariff) -0.743*** -0.917*** -1.637 -6.068*** -2.764*** 0.526

Notes: These tests were performed using the STATA commands levinlin and ipshin, written by FabianBornhorst and Chris Baum and fully described at http://ideas.repec.org. The command levinlin allows threechoices of deterministics: (1) none; (2) constant; (3) constant and trend. The command ipshin allows anotherthree choices of deterministics: (1) constant; (2) constant with cross-sectionally demeaned variable; (3)constant and trend with cross-sectionally demeaned variable. Nonstationarity is the null.The values given arethe coefficient for levinlin and the W[t-bar] statistic for ipshin. *** represents significance at the 1% level.

Table 3. Panel Regression Results for Import Prices (commonsector slopes), 1990–2001

Exchange rate 0.927***(0.094)

Tariff rate 0.001†††(0.013)

Cons. 0.026(0.033)

Sector dummies YesChi-squared test (H0: sector dummies

jointly equal to zero)19.09

Symmetry test 95.11***Homogeneity test 0.56Ramsey RESET test 1.31

N obs. 418N sectors 38Log-likelihood 43.14Wald chi-squared 116.19***

Notes: ***, **, * indicate a coefficient significantly different from zero at,respectively, the 1%, 5%, 10% level. In sectoral passthrough coefficients,†††, ††, † indicate a coefficient significantly different from one at, respec-tively, the 1%, 5%, 10% level. Standard errors are in parentheses. Alikelihood-ratio chi-squared test for panel heteroskedasticity and theWooldridge (2002) panel autocorrelation test were conducted on importsshowing both heteroskedasticity and autocorrelation. These tests arefully described at http://www.stata.com/support/faqs/stat/panel.html. Allestimates were produced using cross-sectional time-series FGLS withheteroskedastic panels and first-order autocorrelation. Symmetry test:chi-squared test where H0: each sector’s slope equal for exchange ratesand tariffs. Homogeneity test: chi-squared test where H0: the sum of eachsector’s exchange rate and tariff coefficients is significantly equal to one.



http://ideas.repec.org

http://www.stata.com/support/faqs/stat/panel.html

Tabl

e4.

Pan

elR

egre

ssio

nR

esul

tsfo

rSe

ctor

alIm

port

Pri

ces

(sec

tor-

spec

ific

slop

es),

1990

–200

1

Exc

hang

era

teTa

riff

rate

Sym

met

ryte

stH

omog

enei

tyte

stE

xcha

nge

rate

Tari

ffra

teSy

mm

etry

test

Hom

ogen

eity

test

Man

ufac

ture

sof

Met

als

1.48

8-0

.027

†0.

840.

09P

harm

aceu

tica

ls1.

827*

-0.0

35††

†2.

95*

0.49

(1.4

94)

(0.5

66)

(1.0

82)

(0.2

31)

Frui

t1.

244

0.18

0*††

†1.

320.

18Sp

ecia

lized

Mac

hine

ry1.

422*

**††

0.05

7†††

23.8

4***

5.17

**(0

.958

)(0

.097

)(0

.219

)(0

.116

)M

anm

ade

Fert

ilize

rs1.

211

-0.3

14††

†2.

680.

02R

ubbe

r1.

380*

**-0

.009

†††

17.6

4***

1.25

(0.8

12)

(0.3

00)

(0.3

31)

(0.0

16)

Met

alw

orki

ngM

achi

nery

1.01

40.

972*

**0.

001.

71M

iner

als

1.34

6***

0.05

7†††

6.34

***

0.66

(0.7

83)

(0.2

31)

(0.4

97)

(0.0

81)

Pow

erM

achi

nery

0.97

60.

057†

††1.

740.

00O

ther

Fibe

rs1.

156*

**-0

.199

**††

†32

.28*

**0.

02(0

.693

)(0

.082

)(0

.242

)(0

.079

)P

ulp

0.93

00.

294†

†0.

150.

02P

hoto

grap

hic

Inst

rum

ents

1.10

6**

0.11

7†††

3.20

*0.

11(1

.545

)(0

.374

)(0

.526

)(0

.210

)C

erea

l0.

871

-0.2

85††

†1.

980.

52A

lum

inum

1.06

2**

-0.3

71††

†8.

79**

*0.

34(0

.581

)(0

.406

)(0

.446

)(0

.246

)P

aper

0.83

60.

135†

††1.

200.

00O

rgan

icC

hem

ical

s1.

039*

*0.

096†

††4.

21**

0.10

(0.6

01)

(0.1

23)

(0.4

25)

(0.1

14)

Non

ferr

ous

Met

als

0.75

50.

404*

*†††

0.42

0.11

Tran

spor

tE

quip

men

t1.

037*

**0.

021†

††4.

95**

0.02

(0.4

84)

(0.1

61)

(0.3

92)

(0.1

59)

Gen

eral

Indu

stri

alM

achi

nery

0.67

1-0

.371

†††

1.37

0.58

Spic

es0.

974*

**-0

.067

†††

7.17

***

0.05

(0.8

21)

(0.3

83)

(0.3

72)

(0.1

37)

Pla

stic

0.67

00.

047†

††0.

660.

12C

rude

0.96

9***

0.14

7†††

5.03

**0.

10(0

.741

)(0

.263

)(0

.338

)(0

.141

)C

oppe

r0.

621

0.16

3†††

0.57

0.14

Iron

and

Stee

l0.

949*

*0.

057†

††4.

72**

0.00

(0.5

74)

(0.1

21)

(0.3

83)

(0.2

43)

Dai

ry0.

556

0.06

0†††

0.94

0.75

Con

stan

t0.

051

(0.4

62)

(0.1

26)

(0.0

44)

Inor

gani

cC

hem

ical

s0.

539

-0.0

52††

†0.

430.

35Se

ctor

Dum

mie

sY

es(0

.816

)(0

.344

)D

yes

0.44

4-0

.144

†††

0.66

1.01

Chi

-sq.

(1)

40.7

8(0

.685

)(0

.190

)



Lea

d0.

335

0.39

30.

000.

06C

hi-s

q.(2

)33

.29

(1.0

53)

(0.3

95)

Scie

ntifi

cIn

stru

men

ts0.

294

0.00

6†††

0.28

1.76

Chi

-sq.

(3)

60.4

0***

(0.5

34)

(0.0

47)

Tin

0.28

5-0

.016

†††

0.23

2.13

Nob

s.41

8(0

.523

)(0

.223

)B

ever

ages

-0.2

44†

-0.0

49*†

††0.

093.

63**

Nse

ctor

s38

(0.6

71)

(0.0

27)

Nic

kel

-0.3

230.

124*

*†††

0.24

1.66

Log

-lik

elih

ood

82.1

1(0

.919

)(0

.061

)Y

arn

-0.4

04-0

.197

†††

0.02

0.85

Wal

dch

i-sq

.27

7.27

***

(1.6

68)

(0.1

27)

Cru

deFe

rtili

zers

-0.4

090.

111†

†0.

151.

11R

amse

yR

ESE

T0.

76(1

.196

)(0

.453

)M

etal

s-0

.507

-0.0

61†

0.05

0.67

(1.8

60)

(0.6

00)

Non

met

allic

Min

eral

s-0

.558

-0.0

01††

†0.

171.

62(1

.219

)(0

.394

)P

etro

l-0

.610

††-0

.092

†††

0.60

5.21

**(0

.704

)(0

.064

)E

lect

rica

lMac

hine

ry-1

.250

††0.

312†

2.05

2.94

*(1

.042

)(0

.385

)

Not

es:

***,

**,

*in

dica

tea

coef

ficie

ntsi

gnifi

cant

lydi

ffer

ent

from

zero

at,

resp

ecti

vely

,th

e1%

,5%

,10

%le

vel.

Inse

ctor

alpa

ssth

roug

hco

effic

ient

s,††

†,††

,†

indi

cate

aco

effic

ient

sign

ifica

ntly

diff

eren

tfr

omon

eat

,re

spec

tive

ly,

the

1%,

5%,

10%

leve

l.St

anda

rder

rors

are

inpa

rent

hese

s.A

likel

ihoo

d-ra

tio

chi-

squa

red

test

for

pane

lhe

tero

sked

asti

city

and

the

Woo

ldri

dge

(200

2)pa

nela

utoc

orre

lati

onte

stw

ere

cond

ucte

don

impo

rts

show

ing

both

hete

rosk

edas

tici

tyan

dau

toco

rrel

atio

n.T

hese

test

sar

efu

llyde

scri

bed

atht

tp://

ww

w.s

tata

.com

/sup

port

/faq

s/st

at/p

anel

.htm

l.A

lles

tim

ates

wer

epr

oduc

edus

ing

cros

s-se

ctio

nal

tim

e-se

ries

FG

LS

wit

hhe

tero

sked

asti

cpa

nels

and

first

-ord

erau

toco

rrel

atio

n.Sy

mm

etry

test

:chi

-sq.

test

whe

reH

0:ea

chse

ctor

’ssl

ope

equa

lfor

exch

ange

rate

san

dta

riff

s.H

omog

enei

tyte

st:c

hi-s

q.te

stw

here

H0:

the

sum

ofea

chse

ctor

’sex

chan

gera

tean

dta

riff

coef

ficie

nts

issi

gnifi

cant

lyeq

ualt

oon

e.C

hi-s

q.(1

):C

hi-s

q.te

stw

here

H0:

sect

ordu

mm

ies

join

tly

equa

lto

zero

.Chi

-sq.

(2):

Chi

-sq.

test

whe

reH

0:eq

uals

ecto

rsl

opes

for

exch

ange

rate

s.C

hi-s

q.(3

):C

hi-s

q.te

stw

here

H0:

equa

lsec

tor

slop

esfo

rta

riff

s.












(Table 4). The dependent variable in both cases is the annual change in import pricesfor 38 two-digit sectors. The TRPT coefficient is not significantly different from zero,which suggests that on average there was no TRPT into import prices. That is, althoughtariffs were reduced, this liberalization was matched by foreign exporters increasingprices in their own currency, and consequently rupee import prices did not benefit fromthe liberalization. The ERPT coefficient is significantly different from zero, but wecannot reject that rupee prices changed one-to-one with the exchange rate, so thatforeign exporter prices may have been kept constant, suggesting complete ERPT. Thismeans that Indian import prices responded to changes in the rupee NEER, but not tochanges in the tariff rates, because foreign exporters were more responsive to liberal-ization than to depreciation.

We also test for symmetry and homogeneity of the exchange rate and tariff ratecoefficients. The nulls of these tests are, respectively, the two coefficients being equal,and the sum of the two coefficients being one. The symmetry test confirms that onaverage import prices react differently to exchange rates and tariff rates. However, thenull of the homogeneity test cannot be rejected, so that on average there could be fulljoint passthrough of exchange rates and tariffs into import prices. This is due to ERPTbeing very close to one and TRPT being very close to zero.

We now want to look at sectoral differences in ERPT and TRPT. As expected, apositive passthrough coefficient in Table 4 implies that the rupee price of importsincreases with exchange-rate depreciation and decreases with tariff reduction. If thepositive passthrough coefficient is lower than one, the rupee price change is less thanproportional to the exchange-rate depreciation or to the tariff reduction, so the price inforeign currency must have changed reflecting foreign firms “pricing to market” adjust-ing their markup according to the destination market conditions. Also incompletepassthrough may depend on the local market conditions, contingent on the extent ofcompetition an importing firm faces in the local market, its market share, and the extentto which the product is differentiated from similar products. For example, a foreign firmthat is attempting to increase its share of the Indian market may passthrough much of thechange in the exchange rate when the rupee is appreciating, but the same firm will resistwhen the rupee is depreciating in an effort to maintain its market share. In a developingcountry where the exchange rate depreciates more often than it appreciates, one wouldexpect a constant or declining foreign currency price that might explain why in someleading import sectors there is incomplete passthrough.If there are import restrictions orhigh tariffs, then the markup adjustment due to change in exchange rate may not help inthis regard. Thus it is important to look at ERPT and TRPT simultaneously.

The ERPT coefficient is significantly positive in 12 sectors, out of which only inSpecialized Machinery the coefficient is significantly different from one (1.42%).Hence in this sector foreign exporters increase their prices by 42% of the exchange ratedepreciation, forcing an increase in the rupee price that is 42% above the exchange ratedepreciation.This result can be understood in light of India’s dependence on machineryto bolster rapid growth in the 1990s. As a result, foreign exporters have a high degreeof market power. In the other 11 sectors showing significant ERPT, we cannot reject fullpassthrough; that is, rupee import prices change one-to-one with the exchange rate.Along the lines of Dixit (1989), this result suggests that there is active entry and exit offoreign firms in the 11 sectors where we cannot reject full passthrough, whereas foreignfirms neither enter nor exit in the remaining sectors where we cannot reject nopassthrough.

The TRPT coefficient is significant in six sectors, out of which only in MetalworkingMachinery it is not significantly different from one, meaning that in this sector there is



full passthrough of tariff reductions, with the import prices being reduced proportion-ally. In three other sectors—Nickel, Fruits, and Nonferrous Metals—rupee importprices decrease by, respectively, 12%, 18%, and 40% of the tariff cut. Hence the rupeeprice of imports has partly reacted to trade liberalization in these sectors, meaning thatforeign exporters increased their foreign currency price by, respectively, 88%, 82%, and60% of the tariff cut.The consequence is that the impact on the price paid by consumersis not as large as it otherwise could be and trade liberalization benefits the exporters byallowing them to increase their prices, still benefiting from some reduction in the rupeeprice. In this case, the foreign exporters benefit relatively more than Indian consumers.There are also two sectors—Beverages and Other Fibers—with negative and significantTRPT. In these sectors, rupee import prices actually increase by, respectively, 5% and20% of the tariff cut. This is because foreign exporters increase their foreign currencyprices by the same proportion of the tariff cut, which could be due to an inelasticdemand in these two sectors helping the foreign exporters to exploit the tariff cuts toincrease their prices. Nontariff barriers may contribute to the incomplete ERPT.13

The results reported in Table 4 show that the sectoral slope coefficients significantlydiffer for TRPT but not for ERPT.We also repeat the symmetry and homogeneity testsfor each sector in Table 4. The symmetry test fails in 12 sectors which have significantERPT, but no TRPT.14 Only four sectors fail the homogeneity test: Beverages, Petrol,Electrical, and Specialized Machinery. These are the four sectors for which we couldreject full ERPT and full TRPT. The homogeneity condition is satisfied when at leastone of the coefficients is statistically equal to one and the condition fails only when bothcoefficients are statistically different from one.

The regression results in Tables 3 and 4 implicitly assume that India’s imports areinvoiced in a variety of currencies. However, it seems reasonable to presume that a largeproportion of imports is invoiced in US dollars (USD). In the absence of detailedcustoms data on invoice currencies, we can take the extreme assumption that all ofIndia’s imports are invoiced in USD. In order to check the robustness of the NEERregression results to the use of the USD as the only invoice currency, we replicated theregressions of Tables 3 and 4, replacing the NEER with the bilateral rupee/USDexchange rate.15 The only qualitative change occurring in Table 3 is that we reject that theERPT coefficient for the USD is equal to one, whilst we could not reject that hypothesisfor the NEER. With respect to Table 4, only 11 sectors out of 38 are affected by eithergaining or losing significance (nine with respect to ERPT and two with respect toTRPT).16 We then replicated the regressions of Table 3 removing those 11 sectors fromthe sample and verified that the previous results for the USD remain qualitatively thesame, possibly because the sectoral changes cancel out one another. As a consequence,we can conclude that, in general, the higher the proportion of India’s imports invoiced inUSD, the more likely it is that ERPT is incomplete. However, as not all transactions areinvoiced in USD, the NEER results stand as being more general than the USD results.

5. Searching for Explanations in the Sectoral Passthrough Variation

The literature has documented that the degree of ERPT is in general low (see Engel,2002). This observation is in line with our Table 4 results, where rupee import pricesreact to exchange rate and tariff changes in, respectively, 12 and 6 sectors out of 38, andremain unchanged in the rest of the sectors. However there is still disagreement in theliterature regarding the causes of low passthrough. Some authors argue that the expla-nation is microeconomic, based on structural features of international trade, suchas pricing-to-market by imperfectly competitive firms (Corsetti and Dedola, 2005),



domestic content in the distribution of traded goods (Burstein et al., 2003), the import-ance of nontraded goods in consumption, or the role of substitution between goods inresponse to exchange rate changes (Burstein et al., 2005). It has also been shown thatthe degree of passthrough increases with the exporter’s share in the destination market(Feenstra et al., 1996) and with the extent of entry and exit of foreign firms in anindustry (Dixit, 1989). Other authors argue that the failure of passthrough is mostly amacroeconomic phenomenon related to the slow adjustment of goods prices at theconsumer level (Engel, 2002) or motivated by macroeconomic policy reforms, forexample, monetary policy (Taylor, 2000). Campa and Goldberg (2002) provide evi-dence for OECD countries that both macro- and microeconomic factors are importantin the evolution of ERPT estimates over time, but in the end favor a microeconomicexplanation based on the changing composition of import goods.

In this section, we compare the importance of a number of factors in explaining thesectoral differences found in Table 4. The factors considered are: (i) each sector’s sharein total imports; (ii) the effective rate of protection; (iii) the importance of nontariffbarriers; and (iv) the import penetration rate. The first measure was directly calculatedfrom the TRAINS database (see the Appendix), whilst the other three are taken fromDas (2003). Here we follow Das (2003) in distinguishing two phases in India’s tradeliberalization: the first (1991–95) starts with the 1990–91 reforms; the second (1996–2001) includes the EXIM policies that aim at simplifying procedures and rationalizingtariff rates. In general, the second phase slowed down the tariff reduction, especially inthe most sensitive sectors.

Table 5 shows the results of a cross-sectional regression of the exchange rate andtariff passthrough coefficients on the four measures indicated above for 1990–95 and1996–2001. In the linear model (1), the degree of ERPT increases with the effectiveprotection rate in 1996–2001; that is, import prices react more to exchange rate changesin sectors with higher effective protection rates.The degree of TRPT decreases with thesector’s share in total imports; that is, import prices react less to changes in tariff ratesin sectors with higher share in total imports. The significance of the sector’s share intotal imports confirms the important role of import shares in determining TRPT. Theimport penetration ratio has no effect on the degree of passthrough, which could beinterpreted as a consequence of the analysis in Dixit (1989), according to which asufficiently small increase in import penetration17 does not have any effect on entry andexit of firms, consequently having no effect on passthrough. However, as Dixit (1989)pointed out, nonlinearities may be important in this context. To investigate this possi-bility, Table 5 also presents the results of a quadratic specification (2). Indeed, theimpact of import penetration appears as a second-order effect in the case of ERPT in1996–2001. Moreover, in the case of TRPT the Ramsey RESET test shows that second-order effects are required for a correct specification. The lack of significance in othercases may be due to data limitations or to small sample problems.

6. Conclusions

This paper makes a contribution to the literature by empirically testing the degreeof passthrough of exchange rates and tariffs in the context of an emerging marketeconomy undergoing deep structural change. India started extensive reforms in thebeginning of the 1990s, comprising both exchange rate depreciation and tariff reduc-tion. The paper examines the responsiveness of Indian import prices to exchange ratechanges and tariff variations in the post-reform period, modeling TRPT alongsideERPT, both theoretically and empirically.



Tabl

e5.

Cro

ss-S

ectio

nal

Reg

ress

ion

Res

ults

for

Pas

sthr

ough

Coe

ffici

ents

,199

0–20

01

Exc

hang

era

teco

effic

ient

sTa

riff

rate

coef

ficie

nts

1990

–95

1996

–200

019

90–9

519

96–2

000

(1)

(2)

(1)

(2)

(1)

(2)

(1)

(2)

Sect

orsh

are

into

tali

mpo

rts

-0.0

480.

099

-0.0

370.

177

-0.0

25*

-0.0

46-0

.040

*-0

.052

(0.0

55)

(0.2

07)

(0.0

71)

(0.2

03)

(0.0

14)

(0.0

62)

(0.0

24)

(0.0

78)

Sect

orsh

are

into

tali

mpo

rts

(squ

ared

)-0

.016

-0.0

330.

001

-0.0

00(0

.017

)(0

.028

)(0

.005

)(0

.010

)E

ffec

tive

prot

ecti

onra

te0.

005

0.03

40.

032*

0.18

7-0

.003

-0.0

30-0

.009

-0.1

15(0

.008

)(0

.052

)(0

.018

)(0

.231

)(0

.003

)(0

.022

)(0

.007

)(0

.089

)E

ffec

tive

prot

ecti

onra

te(s

quar

ed)

-0.0

00-0

.002

0.00

00.

001

(0.0

00)

(0.0

03)

(0.0

00)

(0.0

01)

Non

tari

ffba

rrie

rs-0

.005

-0.0

09-0

.010

0.02

2-0

.001

0.00

3-0

.002

-0.0

00(0

.006

)(0

.019

)(0

.007

)(0

.022

)(0

.002

)(0

.005

)(0

.002

)(0

.009

)N

onta

riff

barr

iers

(squ

ared

)0.

000

-0.0

01-0

.000

-0.0

00(0

.000

)(0

.000

)(0

.000

)(0

.000

)Im

port

pene

trat

ion

rati

o-1

.154

2.59

9-0

.469

-5.5

890.

409

1.16

50.

349

1.03

8(1

.263

)(6

.597

)(0

.981

)(3

.474

)(0

.360

)(1

.783

)(0

.243

)(0

.919

)Im

port

pene

trat

ion

rati

o(s

quar

ed)

-7.0

2910

.559

*-0

.935

-1.2

77(1

1.26

6)(5

.551

)(3

.021

)(1

.794

)C

ons.

0.66

8-0

.965

-0.2

23-3

.040

0.24

61.

175

0.36

62.

302

(0.8

97)

(2.3

31)

(0.8

73)

(4.7

93)

(0.3

53)

(0.8

96)

(0.3

07)

(1.7

39)

Nob

s.29

2929

2929

2929

29R

-squ

ared

0.13

40.

223

0.21

70.

396

0.23

20.

344

0.21

80.

342

Ram

sey

RE

SET

test

1.26

1.81

0.83

1.45

2.12

*1.

094.

52**

*1.

82F

-tes

t(H

0:co

effic

ient

sjo

intl

yeq

ualt

oze

ro)

1.41

9.91

***

2.97

**12

.94*

**2.

16*

2.54

**1.

151.

30

Not

es:

***,

**,*

indi

cate

aco

effic

ient

sign

ifica

ntly

diff

eren

tfro

mze

roat

,res

pect

ivel

y,th

e1%

,5%

,10%

leve

l.O

LS

robu

stst

anda

rder

rors

are

inpa

rent

hese

s.(1

)is

alin

ear

mod

el;(

2)is

aqu

adra

tic

mod

el.F

orm

ore

deta

ilson

how

the

expl

anat

ory

vari

able

sar

eca

lcul

ated

and

thei

rva

lues

see

Das

(200

3).



Using data for a panel of 38 two-digit SITC sectors over the period from 1990 to2001, the passthrough of changes in both the NEER of the rupee and the tariff ratesinto import prices is often found to be incomplete or imperfect in the reform period.This finding suggests that the pricing behavior of foreign exporters varies across indus-tries, with ERPT being complete in 32% (12 out of 38) of sectors, and zero orincomplete in the remaining 68%. The results also indicate that the foreign exportersabsorb at least part of the exchange rate and tariff changes in, respectively, 4 and 35industries.The share of each sector in total imports and the effective protection rate arefound to contribute to sectoral differences in the degree of passthrough, although in anonlinear way.

In India’s context, tariffs are still high despite more than a decade of liberalization.However, the paper’s results show that TRPT is not complete, although zeropassthrough cannot be ruled out. Hence, tariff rates are not high enough to verifyempirically the theoretical presumption that, under high tariff rates,TRPT will be closeto one independent of the degree of ERPT. Our results hint that, at least in somesectors, tariff liberalization has benefited less the Indian consumers and more theforeign exporters, who react to tariff reductions by increasing their markups, thuspartially or even totally offsetting the tariff reduction. On the contrary, exchange ratedepreciation has not benefited foreign exporters, since either the rupee import pricechanges one-to-one with the exchange rate, or does not react to the depreciation,implying that foreign exporters totally absorbed the price increase. It is thus interestingthat, in India’s import markets, trade liberalization leads to an increase in foreignexporters’ markups and the benefit is not passed on to the consumers, but exchangerate depreciation causes a decrease in foreign exporters’ markups because the foreignproducer seems to absorb a part of the increase in the cost of imports.

Appendix: Data Sources and Definitions

The unit value indices of imports for a number of sectoral groups, and the rupee NEER(nominal effective exchange rate) were compiled from the Handbook of Statistics onthe Indian Economy 2002–03, Reserve Bank of India, over the period 1990–91 to2001–02. Financial year (annual average) data are used in this paper. Import valueindices for the two-digit products are calculated by multiplying the quantity index withunit value index, and with base year values in local currency for the respective product,the sectoral value indices are converted to local currency units and the product sharesare then derived.

The NEER is calculated as a weighted geometric average of the bilateral nominalexchange rates of the Indian rupee in terms of foreign currencies. Here it measures theappreciation/depreciation of the rupee against the weighted basket of 36 currencieswhose countries are the main trading partners or competitors of India. The formula is:

NEER ei INRw

i

i= ( )=

∏ , ,1

36

where ei is the exchange rate of the rupee against the currency of the trading partner i,i.e. rupee per currency i (in index form); wi is the 36-country bilateral trade weightsattached to currency/country i in the index.

Data on imports relate to cost, insurance, and freight (c.i.f.) values. All the dataare annual. Data on tariffs were taken from the TRAINS database at http://wits.worldbank.org. The rate used is the weighted average of all the tariff lines within eachtwo-digit category, as provided by TRAINS.



http://wits

Table A1. The Codes and Definition of the Two-Digit SITC (Rev. 2) Sectors

Code DescriptionSITC Rev. 2

code

DAIRY Food and food articles: dairy products 02CEREAL Food and food articles: cereals and cereal preparations 04FRUIT Food and food articles: fruits and nuts 057SPICES Food and food articles: spices 075BEV Beverages and tobacco: beverages 11RUBBER Crude materials, inedible, except fuels: crude rubber incl. synthetic and

reclaimed23

PAPER Crude materials, inedible, except fuels: pulp and waste paper 25OTH FIB Crude materials, inedible, except fuels: textile fibers and waste excl.

cotton26-263

CRUFERT Crude materials, inedible, except fuels: crude fertilizers 272MINERALS Crude materials, inedible, except fuels: minerals (excl. coal, petroleum,

crude fertilizers, sulfur and precious stones)27-272

METALS Crude materials, inedible, except fuel: ores and concentrates of basemetals n.e.s.

287

NF METALS Crude materials, inedible, except fuels: nonferrous base metals, wasteand scrap

288

CRUDE Mineral fuels, lubricants, etc.: petroleum crude 33PETROL Mineral fuels, lubricants, etc.: petroleum products 33ORGCHEM Chemicals and related products: organic chemicals 51INORGCHEM Chemicals and related products: inorganic chemicals 52DYES Chemicals and related products: dyeing, tanning and coloring materials 53PHARM Chemicals and related products: medicinal and pharmaceutical products 54MANFERT Chemicals and related products: fertilizers, manufactured 56PLASTIC Chemicals and related products: artificial resin and plastic material and

cellulose ester57

PAPER Manufactured goods classified chiefly by material: paper, paperboardand articles thereof

64

YARN Manufactured goods classified chiefly by material: textile yarn 651NONMETMIN Manufactured goods classified chiefly by material: nonmetallic mineral

manufactures n.e.s.66

IRONSTEEL Manufactured goods classified chiefly by material: iron and steel 67COPPER Manufactured goods classified chiefly by material: copper 682NICKEL Manufactured goods classified chiefly by material: nickel 683ALUM Manufactured goods classified chiefly by material: aluminum 684LEAD Manufactured goods classified chiefly by material: lead 685TIN Manufactured goods classified chiefly by material: tin 687MANMET Manufactured goods classified chiefly by material: manufactures of

metals69

POWERMACH Machinery and transport equipment: power generating machineryand equipment

71

SPECMACH Machinery and transport equipment: machinery specialized forparticular industries

72

METWORKMACH Machinery and transport equipment: metalworking machinery 73GENINDMACH Machinery and transport equipment: general industrial machinery

and equipment74

ELMACH Machinery and transport equipment: electrical machinery 77TRANSEQ Machinery and transport equipment: transport equipment 79SCINSTR Miscellaneous manufactured articles: professional, scientific and

controlling instruments and apparatus n.e.s.87

PHOTOINSTR Miscellaneous manufactured articles: photographic apparatus, etc. 88

Note: A full description of the SITC codes can be found at http://www.census.gov/foreign-trade/reference/codes/sitc/sitc.txt.



http://www.census.gov/foreign-trade/reference

References

Ahluwalia, M. S., “Economic Reforms in India since 1991: Has Gradualism Worked?” Journal ofEconomic Perspectives 16 (2002):67–88.

———, “India’s Experience with Globalisation,” Australian Economic Review 39 (2006):1–13.Athukorala, P., “Exchange Rate Pass-Through: The Case of Korean Exports of Manufactures,”

Economics Letters 35 (1991):79–84.Athukorala, P. and J. Menon,“Pricing-to-Market Behaviour and Exchange Rate Pass-Through in

Japanese Exports,” Economic Journal 104 (1994):271–81.Bleaney, M. F.,“Invoicing–Currency Effects in the Pricing of Japanese Exports of Manufactures,”

Canadian Journal of Economics 30 (1997):968–74.Broda, C., “Terms of Trade and Exchange Rate Regimes in Developing Countries,” Journal of

International Economics 63 (2004):31–58.Burstein, A., M. Eichenbaum, and S. Rebelo,“Large Devaluations and the Real Exchange Rate,”

Journal of Political Economy 113 (2005):742–84.Burstein, A. T., J. C. Neves, and S. Rebelo, “Distribution Costs and Real Exchange Rate Dynam-

ics,” Journal of Monetary Economics 50 (2003):1189–214.Campa, J. M. and L. S. Goldberg, “Exchange Rate Pass-Through into Import Prices: A Macro or

Micro Phenomenon?” NBER working paper 8934 (2002).———, “Exchange Rate Pass-Through into Import Prices,” Review of Economics and Statistics

87 (2005):679–90.Corsetti, G. and L. Dedola, “A Macroeconomic Model of International Price Discrimination,”

Journal of International Economics 67 (2005):129–55.Das, D. K., “Quantifying Trade Barriers: Has Protection Declined Substantially in Indian Manu-

facturing?” ICRIER working paper 105, New Delhi (2003).Devereux, M. B., P. R. Lane, and J. Xu, “Exchange Rates and Monetary Policy in Emerging

Market Economies,” Economic Journal 116 (2006):478–506.Dixit, A.,“Hysteresis, Import Penetration, and Exchange Rate Pass-Through,” Quarterly Journal

of Economics 104 (1989):205–28.Doyle, E., “Exchange Rate Pass-Through in a Small Open Economy: The Anglo-Irish Case,”

Applied Economics 36 (2004):443–55.Engel, C., “The Responsiveness of Consumer Prices to Exchange Rates: A Synthesis of Some

New Open Economy Macro Models,” The Manchester School 70 (Supplement) (2002):1–15.Feenstra, R., “Symmetric Pass-Through of Tariffs and Exchange Rates under Imperfect Compe-

tition: An Empirical Test,” Journal of International Economics 27 (1989):25–45.Feenstra, R., J. Gagnon, and M. Knetter, “Market Share and Exchange Rate Pass-Through in

World Automobile Trade,” Journal of International Economics 40 (1996):187–207.Frankel, J., D. Parsley, and S. Wei, “Slow Pass-Through around the World: A New Import for

Developing Countries?” NBER working paper 11199 (2005).Froot, K. and P. Klemperer, “Exchange Rate Pass-Through when Market Share Matters,”

American Economic Review 79 (1989):637–54.Gagnon, J. and M. Knetter, “Markup Adjustment and Exchange Rate Fluctuations: Evidence

from Panel Data on Automobile Exports,” Journal of International Money and Finance 14(1995):289–310.

Goldberg, P., “Product Differentiation and Oligopoly in International Markets: The Case of theUS Automobile Industry,” Econometrica 63 (1995):891–951.

Goldberg, P. and M. Knetter, “Goods Prices and Exchange Rates: What Have We Learned?”Journal of Economic Literature 35 (1997):1243–72.

———, “Measuring the Intensity of Competition in Export Markets,” Journal of InternationalEconomics 47 (1999):27–60.

Gross, D. and N. Schmitt, “Exchange Rate Pass-Through and Rivalry in the Swiss AutomobileMarket,” Weltwirtschaftliches Archiv 123 (1996):278–303.

Hooper, P. and C. Mann, “Exchange Rate Pass-Through in the 1980s: The Case of US Imports ofManufactures,” Brookings Papers on Economic Activity 1 (1989):297–329.



Im, K. S., M. H. Pesaran, and Y. Shin, “Testing for Unit Roots in Heterogeneous Panels,” Journalof Econometrics 115 (2003):53–74.

Joshi, V., “India and the Impossible Trinity,” The World Economy 26 (2003):555–83.Kim, Y., “Exchange Rates and Import Prices in the US: A Varying-Parameter Estimation of

Exchange Rate Pass-Through,” Journal of Business and Economic Statistics 8 (1990):305–15.Knetter, M., “Price Discrimination by US and German Exporters,” American Economic Review

79 (1989):198–210.———, “Is Export Price Adjustment Asymmetric? Evaluating the Market Share and Marketing

Bottlenecks Hypotheses,” Journal of International Money and Finance 13 (1994):55–70.Koch, P. and J. Rosensweig,“The Dollar and the US Terms of Trade,” Journal of Macroeconomics

14 (1992):467–86.Lee, J., “The Response of Exchange Rate Pass-Through to Market Concentration in a Small

Economy: The Evidence from Korea,” Review of Economics and Statistics 79 (1997):142–45.Levin, A., C.-F. Lin, and C.-S. J. Chu, “Unit Root Tests in Panel Data: Asymptotic and Finite

Sample Properties,” Journal of Econometrics 108 (2002):1–24.Madsen, J., “Errors-in-Variables, Supply Side Effects, and Price Elasticities in Foreign Trade,”

Weltwirtschaftliches Archiv 134 (1998):612–37.Mallick, S. K. and H. Marques, “Sectoral Exchange Rate Pass-Through: Testing the Impact of

Policy Reforms in India,” Scottish Journal of Political Economy 53 (2006):280–303.Mattoo, A. and R. Stern, India and the WTO, Oxford: Oxford University Press (2003).Menon, J., “Exchange Rates and Prices of Australian Manufactured Exports,” Weltwirtschaft-

liches Archiv 128 (1992):695–710.———, “Exchange Rate Pass-Through,” Journal of Economic Surveys 9 (1995):197–231.———, “The Degree and Determinants of Exchange Rate Pass-Through: Market Structure,

Non-Tariff Barriers and Multinational Corporations,” Economic Journal 106 (1996):434–44.Panagariya, A., “India’s Trade Reform,” in A. Panagariya (ed.), The India Policy Forum 2004,

Vol. 1, Washington, DC: Brookings Institution, and New Delhi: NCAER (2005).Parsley, D., “Exchange Rate Pass-Through: Evidence from Aggregate Japanese Exports,” South-

ern Economic Journal 60 (1993):454–62.Steel, D. and A. King, “Exchange Rate Pass-Through: The Role of Regime Changes,” Interna-

tional Review of Applied Economics 18 (2004):301–22.Tange, T., “Exchange Rates and Export Prices of Japanese Manufacturing,” Journal of Policy

Modelling 19 (1997):195–206.Taylor, J. B.,“Low Inflation, Pass-Through, and the Pricing Power of Firms,” European Economic

Review 44 (2000):1389–408.Taylor, M. P., “Purchasing Power Parity,” Review of International Economics 11 (2003):436–52.Wooldridge, J., Econometric Analysis of Cross Section and Panel Data, Cambridge, MA: MIT

Press (2002).Yang, J., “Exchange Rate Pass-Through in US Manufacturing Industries,” Review of Economics

and Statistics 79 (1997):95–105.———, “Pricing-to-Market in US Imports and Exports: A Time Series and Cross-Sectional

Study,” Quarterly Review of Economics and Finance 38 (1998):843–61.Yang, Y. and M. Hwang, “Price Behaviour in Korean Manufacturing,” Review of Economics and

Statistics 76 (1994):461–70.

Notes

1. ERPT is the percentage change in local currency import prices resulting from a 1% change inthe exchange rate between the exporting and importing countries.2. Between 1990–91 and 2001–02, the rupee depreciated at an average annual rate of over 8%,whereas the local currency import prices increased at an annual average rate of 6.8% during thesame period.3. The highest tariff rate was brought down from 150% in 1991–92 to 30.8% in 2002–03, whilstthe average import-weighted tariff was reduced from 72.5% in 1991–92 to 29% in 2002–03



(Ahluwalia, 2002). However, this average hides important sectoral differences, with imports suchas textiles and footwear still subject to tariffs higher than 40% (Mattoo and Stern, 2003).4. These reforms included devaluation of the rupee vis-à-vis the US dollar, subsequently leadingto a managed-float regime. For a detailed discussion of the 1990s trade policy reforms, seeAhluwalia (2002) and Panagariya (2005). Also see Joshi (2003), particularly for a discussion ofthe management of India’s BOP in the 1990s.5. Similarly to ERPT, TRPT is the percentage change in local currency import prices resultingfrom a 1% change in the tariff rate between the exporting and importing countries.6. See Taylor (2003) for a review on purchasing power parity with reference to the literature ofthe last two decades.7. See Mallick and Marques (2006), which is the only previous study on Indian ERPT at thesectoral level (SITC one-digit). This study compares ERPT in India’s export and import pricesbefore and after the introduction of the 1991 reform package.8. A total of 36-country bilateral weights of Indian rupee have been used in the index (indexbase: 1985 = 100). For full details, see www.rbi.org.in.9. The term ti becomes a sector-specific term, as it is assumed that marginal costs are constantover time.10. On the other hand, one would expect import liberalization (tariff cuts in a large number ofsectors) to substantially raise the import demand causing depreciation of the Indian currency.Therefore, a part of ERPT may be due to tariff cuts. In other words, instead of passing on theadvantage of lower tariffs to the Indian consumers the foreign exporters absorb a part of theincrease in the import cost caused by currency depreciation that results from trade liberalization.11. From equation (4), when h is larger (or smaller) than unity and d i > 1, the sign of thenumerator of the d function is positive (or negative). If h is smaller (or larger) than unity and0 < d i < 1, the sign of the numerator of the d function is negative (or positive).Thus the exact signof exchange rate coefficient is indeed an empirical question.12. It is also possible that the ERPT could depend on forward contracts. But due to the fact thatIndia did not have a well-developed forward FX market until the late 1990s, it is hard toinvestigate the influence of forward rates for the time span used in this paper (1990–2001).13. If foreign exporters are able to extract rents as a result of a quota, a depreciation of theimporter’s currency should only have the effect of reducing the rent component of the good’sprice and not increase the price itself, so long as the quota remains binding (Steel and King, 2004).14. Except Other Fibers, this has a significantly positive ERPT but a significantly negative tariffpassthrough. These two coefficients do significantly differ.15. The full results are available from the authors on request.16. In detail, the ERPT coefficient in Iron and Steel, Organic Chemicals and Pharmaceuticals isno longer significantly different from zero, whilst it becomes significantly different from zero inCereal and Metalworking Machinery; it is no longer significantly different from one in Special-ized Machinery, whilst it becomes significantly different from one in Scientific Instruments andTin. In turn, the TRPT coefficient in Fruit is no longer significantly different from zero, whilst itbecomes significantly different from zero in Petrol.17. In our sample the unweighted average import penetration ratio across the 38 two-digitsectors increases from 0.216 in 1991–95 to 0.249 in 1996–2001.



http://www.rbi.org.in

passthrough of exchange rate and tariffs into import prices of india: currency depreciation versus...

Documents