Open Economies Review

, Volume 22, Issue 5, pp 955–968

A Measure of Trade Induced Adjustment in Volume and Quality Space

  • Abdul K. M. Azhar
  • Robert James Ross Elliott
RESEARCH ARTICLE

DOI: 10.1007/s11079-010-9186-9

Cite this article as:
Azhar, A.K.M. & Elliott, R.J.R. Open Econ Rev (2011) 22: 955. doi:10.1007/s11079-010-9186-9
  • 117 Downloads

Abstract

In this paper we contribute to the literature on the measurement of international trade flows. Specifically, we combine changes in the volume and quality in matched trade changes to present a simple new index together with a geometric framework that can be used to visualise changes in quality and volume simultaneously. We illustrate the usefulness of our simple extension with data for trade between Malaysia and China between 1994 and 2004.

Keywords

Exports Intra-industry trade Adjustment costs Quality Product differentiation 

JEL classification

F19 

1 Introduction

The literature on the measurement of international trade flows has grown considerably in recent years. Whilst the Balassa (1966) measure of comparative advantage and the Grubel and Lloyd (1975) index of intra-industry trade (IIT) have stood the test of time, there have been some considerable developments in related measures of IIT (the simultaneous import and export of goods from the same industry). The recent literature concentrates on two distinct stands. The first examines the relationship between changes in intra-industry trade flows and the associated adjustment costs (Hamilton and Kniest 1991; Greenaway et al. 1994; Brülhart 1994; Menon and Dixon 1997; Azhar and Elliott 2003). The second examines the extent to which products are differentiated in terms of quality within matched trade flows or to put it another way, differ in quality (Abd-el-Rahman 1991; Greenaway et al. 1994, 1995; Fontagné and Freudenberg 1997; Azhar and Elliott 2006).1

The contribution of this short paper is to fill the remaining gap in the family of static and dynamic trade measures. Specifically, we combine changes in volume and quality in matched trade changes to present a simple new index together with a geometric tool to allow us to visualise both types of change simultaneously.

The background and primary motivation for the first strand looking at the volume based trade adjustment literature is to be able to contribute to the smooth adjustment hypothesis (SAH) debate that hypothesises that trade changes that are intra-industry in nature will experience trade induced adjustment costs that are less severe than the costs associated with inter-industry or net trade changes. The SAH has been invoked by numerous authors including Balassa (1966), Krugman (1981), Cadot et al. (1995) and Egger et al. (2007).2

Assuming constant domestic demand, the SAH in its simplest form argues that if both imports and exports for products within a given industry are matched (either increasing or decreasing) then, whilst some firms in a given industry will win and others will lose, the reallocation of factors between the winners and losers should be achievable at relatively low cost (smooth). This result hinges on the assumption that the factor requirements of firms in the same industry are more similar than the factor requirements of firms across sectors (Elliott et al. 2000). The greater the degree of factor heterogeneity, the longer the economy will produce inside the long-run production possibility frontier. A second interpretation would be costs in the presence of downward sticky wages that fail to adjust to a new equilibrium. The difficulty for applied economists attempting to estimate the economic significance is how to measure adjustment costs. Most proxies of adjustment cost rely on measures of factor reallocation, usually labour and include Greenaway et al. (2002), Brülhart and Elliott (2002), Haynes et al. (2002), Erlat and Erlat (2003), Elliott and Lindley (2006), Brülhart et al. (2006), Cabral and Silva (2006), Cabral et al. (2006) and Ferto (2008).

More recently however, a number of papers have argued that greater consideration needs to be given to the second stand of the measurement literature that considers the role of quality in matched trade changes (Brülhart and Elliott 2002, Greenaway et al. 2002, Ferto 2005, Cabral et al. 2006). For example, Cabral et al. (2006) split intra-industry into horizontal and vertical IIT to examine the factor content of trade and demonstrate that the adjustment implications differ which has important implications for the smooth adjustment hypothesis. Indeed, Cabral et al. (2006) when talking about their results go as far as to say “...it confirms the importance of the distinction between matched trade in horizontally and vertically differentiated products. It seems to be as important as the distinction between inter-and intra-industry trade” (pg. 562). The importance of this distinction is reinforced by Ferto (2005).

In light of these comments, Azhar and Elliott (2008) argue that measures of changes in product quality and volume need to be considered together and marry these two strands of the literature to provide a measure of the changing structure of product quality associated with changes in matched trade. The measure is essentially a dynamic measure of quality in matched trade in the same way that marginal intra-industry trade (MIIT) is a dynamic measure of IIT.

The crux of the quality adjustment argument relates to the assumption of the SAH that factors can more smoothly transfer between firms within the same industry. Whilst this may hold for homogenous products it is less clear for vertically differentiated products (products that differ in quality). Consider an increase in the volume of matched trade for a given industry that also involves an increase in the output from firms producing high quality varieties at the expense of firms producing low quality varieties. In this example it is likely that displaced workers from the low quality, contracting, firms will not have the required levels of skill or attributes to be quickly and painlessly re-employed by the high quality, expanding, firms without a degree of potentially costly retraining (an adjustment cost). Furthermore, it is unlikely that the capital equipment used to previously manufacture low quality varieties could be inexpensively reassigned, relocated or re-engineered to produce high quality varieties of the same product. Similarly, whilst expanding low quality producing firms could easily employ high skilled workers this if likely to involve a degree of resource underemployment which translates into a subsequent loss to the national economy. For capital the problem remains. Even machinery designed to produce high quality products may be incompatible with the production of its low quality counter-part and require a degree of re-engineering (an adjustment cost). The solution proposed by Azhar and Elliott (2008) is a marginal quality (MQ) index that is able to capture changes in quality in matched trade changes.3

The contribution of this paper is to present a simple framework to enable academics and policymakers to visualise changes in both the volume and the quality in matched trade changes simultaneously. In terms of the empirical literature described above such a measure would enter as an additional right hand side variable to explain employment changes of one sort or another. At a more practical level, employed on a month to month basis, such an approach would provide timely information on those sectors that are coming under competitive pressures whether these pressures are volume or quality related. Subsequent government support could include the allocation of funds to facilitate employee retraining or more careful tailoring of strategic industrial or trade policy.

The rest of the paper is organised as follows. Section 2 briefly summarises the existing literature, Section 3 presents our novel extension. Section 4 provides a numerical example and Section 5 concludes.

2 Measuring Changes in Volume and Quality in Matched Trade Changes

As discussed in the introduction, the IIT measurement literature has developed along two distinct lines. The first considers the measurement of changes in the volume in matched trade changes while the second considers quality differences in matched trade. Azhar and Elliott (2008) combine these two strands to measure quality changes in matched trade changes and suggest a three stage approach for empirical investigations. We briefly outline the existing measures of volume and quality changes in matched trade changes that are consistent with the development of our new approach. It is the various properties associated with these measures that make our new index possible.4

Changes in the volume of matched trade changes are measured as;
$$ S = \frac{{\Delta X - \Delta M}}{{2\max \left\{ {{{\left| {\Delta X} \right|}_t},{{\left| {\Delta M} \right|}_t}} \right\}}}{\hbox{for }}t \in N,N = \left\{ {1,2,3,.....n} \right\} $$
(1)
where the S index is scaled between minus one and one and ΔX and ΔM represent changes in exports and imports respectively for the years t to N. S can be subscripted by either H or F to represent the index from the perspective of the home (H) or foreign (F) country. SH is a simple monotonically increasing function of ΔXM that has the important properties of consistency and country specificity (see Azhar and Elliott 2003 for details). The relationship between the Home and Foreign country is given by SF = −SH. A S index of zero means changes in imports and exports are exactly matched. At the extremes, imports and exports move in opposite directions either beneficially or detrimentally for the Home country. Calculating S indices is stage one of the three stage approach and is used to select those industries are to be taken to the second stage. This group will tend to be those that have experienced large increases or decreases in matched trade which the traditional SAH model predicts will have benign adjustment costs.

We now turn to static and then dynamic measures of quality in matched trade. The standard approach to measuring quality is to use quantity and price data to calculate unit values (UVs) which are then used to classify quality under the implicit assumption that price is an indicator of quality (Stiglitz 1987).5 Once the degree of matched trade has been calculated, it is split into two components, vertical IIT (VIIT) and horizontal IIT (HIIT), where the degree of the former provides an indicator of the differences in quality between traded goods from the same industry.6

To calculate static quality differences in IIT we take the ratio of unit values to measure the level of VIIT in matched trade for any given year. Hence, the dispersion of product quality in matched trade flows is measured as;
$$ PQV = 1 + \frac{{U{V^X} - U{V^M}}}{{\left( {U{V^X} + U{V^M}} \right)}} $$
(2)
where UVX and UVM are unit values for exports and imports respectively. The PQV index is scaled between zero and two. See Azhar and Elliott (2006) for details. This is stage two of the three stage approach. By construction the PQV index is proportional, symmetric and scaled. Products that share 85% or more of their costs are assumed to be horizontally differentiated with any value above or below this threshold are considered vertically differentiated.7 Thus, from a Home country perspective, IIT is classified as high quality, VIIT (high) if PQV >1.15, low quality (VIIT low) if PQV <0.85 and similar quality (HIIT) if PQV >0.85 and <1.15.
Finally, stage three involves the measurement of changes in product quality in matched trade changes where;
$$ M{Q_t} = \frac{{\Delta U{V^X} - \Delta U{V^M}}}{{2\max \left\{ {{{\left| {\Delta U{V^X}} \right|}_t},\left. {{{\left| {\Delta U{V^M}} \right|}_t}} \right\}} \right.}}{\hbox{for }}t \in N,N = \left\{ {1,2,3,.....n} \right\} $$
(3)
where the MQ index ranges from minus one to one for the years t to N. See Azhar and Elliott (2008) for details. The index is scaled by the maximum of the UV changes for any given time period or range of products. This is the third and final stage of the three stage approach.

Azhar and Elliott (2008) demonstrate that if researchers calculate PQV, S and MQ indices separately and then compare the final two it is possible to arrive at a better understanding of the adjustment implications of changes the volume and quality in matched trade changes. Specifically, the suggestion is to take MQ and S indices for each product and then allocate each pair into one of four groups: (1) positive S and negative MQ; (2) positive S and positive MQ; (3) negative S and negative MQ; (4) negative S and positive MQ. From a Home country perspective, governments are likely to be most concerned when there is a large negative S index and a large negative MQ index which would indicate that exports of a given product have decreased relative to its imports and that the relative quality of the product has fallen although any large negative irrespective of its pairing is worth highlighting.

Whilst the Azhar and Elliott (2008) approach is useful, one weakness is that there is no single measure of volume and quality adjustment or a simple way of presenting the results. In Section 3 we show how the S and MQ indices can be combined into a single index and also how the results can be presented diagrammatically. We believe this fills the final gap in the range of IIT quality adjustment indices.

3 A Quality Adjusted Measure of Trade Induced Adjustment

Our extension takes advantage of an important attribute shared by the S and MQ indices which is that they are both proportional, symmetric and scaled between minus one and one. If these indices are to be combined into a single dynamic quality adjusted measure of trade induced adjustment without weighting the index then the quality and volume induced adjustment costs are assumed to be equal. We discuss the weighting implications later and suggest a method for generating the relevant weights if required.

Thus, combining the S and MQ indices (Eqs. (1) and (3)) gives us a quality adjusted volume index measured as;
$$ V{Q_H} = S + MQ = \left[ {\frac{{\left( {\Delta X - \Delta M} \right)}}{{2\max \left\{ {{{\left| {\Delta X} \right|}_t},{{\left| {\Delta M} \right|}_t}} \right\}}} + \frac{{\left( {\Delta U{V^X} - \Delta U{V^M}} \right)}}{{2\max \left\{ {{{\left| {\Delta U{V^X}} \right|}_t},{{\left| {\Delta U{V^M}} \right|}_t}} \right\}}}} \right] $$
(4)
i.e.
$$ V{Q_F} = - V{Q_H} $$
(5)
where S and MQ have previously been defined. The subscripts H and F represent the Home and Foreign country respectively. By construction the VQ index is scaled between minus two and two. The scale is self-evident as the index captures total adjustment and reflects the addition of two indices that are each scaled between minus one and one. Another way of thinking about this is that it represents a total adjustment isocline in our geometric presentation. Such an index can be used to analyse any aspect of trade induced adjustment whether it is at the product, industry or country level of multilateral or bilateral trade relationships.
We now construct and present a diagram to complement the VQ index that we call the Quality Adjusted Trade Adjustment Space (QTAS). Such a diagram plots the S index on the horizontal axis and the MQ index on the vertical axis. See Fig. 1. Any change in volume and/or quality in matched trade changes is captured by the QTAS. The diagonal lines represent lines of equal VQ indices. In Fig. 1 these have been drawn at VQ = 1 and VQ = −1 as well as the leading diagonal where VQ = 0.
Fig. 1

Quality adjusted adjustment Space (QTAS

For example, consider a country that that has suffered significant changes in the volume of matched trade changes (a negative S index value). Whilst one might assume that adjustment costs will be fairly high, if a large and negative S index is matched by a large and positive MQ index then the overall adjustment costs may not be as great as first feared as there is a positive quality adjustment. Thus, while the negative S can be thought of, for means of illustration, as a loss of employment, the rise in MQ can likewise be considered as an increase in the overall human capital of those remaining in work. Of course the caveat to this statement is that adjustment to a higher quality equilibrium is not costless. At the extreme, one could argue it is the absolute magnitude of the MQ index that matters with adjustment being as costly for an increase as a decrease in quality. Likewise, if a large positive S index is matched by a fall in quality (negative MQ index) then the positive effects may be less keenly felt. Whilst employment might rise as a result of a high S, the skill requirements will have fallen leading to a country falling into a possible future quality trap.

Clearly, a potential criticism with combining these two indices in this way is the assumption that the costs of adjustment to quality and volume are of equal magnitude, equivalent to saying that the VQ index is like adding apples and oranges. Note that this criticism applies only to the VQ index. The applicability of the QTAS holds. However, given the construction of the index it would be a relatively simple exercise to apply weights to S and MQ when calculating the index. See the appendix for details.

Calculating MQ, S and VQ indices allows comparisons across indices and can be complemented by the use of the QTAS for visual identification. For future econometric tests of the SAH we believe that the VQ index makes a useful addition to the set of independent variables. If the quality adjusted SAH is correct then one would expect to see a hierarchy of significance from inter-industry adjustment (high cost) to quality changes in IIT to HIIT (which should have a low cost). Such an analysis is left for future research.

Finally, Table 1 summarises the family of IIT related measures (both static and dynamic) that have been developed to date showing where the VQ index fits within this wider family of measures. The only empty section is for a combined static quality and volume based measure. Such an index would make little economic or statistical sense.
Table 1

Summary of volume and quality based measures

 

Static measures

Dynamic measures

Volume based

Grubel and Lloyd (1975)

Hamilton and Kniest (1991)

GL index

Greenaway et al. (1994), GHME index

Brülhart (1994), A, B, C indices

Dixon and Menon (1997)

Azhar and Elliott (2003), S index

Quality based

Greenaway et al. (1995)

Azhar and Elliott (2008), MQ index

Fontagné and Freudenberg (1997)

Azhar and Elliott (2006), PQV index

Volume and Quality Based

 

VQ index

4 Empirical Example

To demonstrate the applicability of the VQ index we present an example using data on ten years of trade between China and Malaysia for the period 1994 and 2004. The data are from the Comtrade database for bilateral imports and exports between China and Malaysia measuring both the values (in Dollars) and the quantity (in tonnes). All values were deflated to 1994 prices using a US GDP deflator. Only products measured in tonnes were included in the analysis. China and Malaysian trade was considered as this period marked a transition in China’s openness to trade and its membership of the WTO in 2001. Fears that Malaysia would see a considerable fall in exports to the West in the face of competition from China led to considerable concern among politicians in China’s neighbouring ASEAN countries.

We provide the total import and export values as well as the S index, PQV, MQ and VQ indices. We consider China to be the Home country in this analysis (although it would be simple to reverse this assumption). This means we can analyse the results from both a Malaysian and Chinese perspective. Clearly, high positive VQ index values are good for China (both S and MQ are positive) and a concern to Malaysia. We concentrate on a small selection of manufactured products within SITC 8 (Miscellaneous Manufactures). The products were selected after an initial trawl through the data to pick only those 5-digit products with significant levels of matched trade (with an S index score of greater than plus or minus 0.4).

From Table 2 observe that for China, 15 out of 23 products in the SITC 8 category with significant matched trade volumes, have suffered some potential adjustment costs both quality and volume related. The largest negative VQ index is for SITC 89594 (Typewriter ribbons) where imports into China grew much faster than exports to Malaysia leading to a negative S index of −0.32. At the same time there was a considerable fall in the relative quality of Chinese varieties of this product with an MQ index of −0.94. The overall VQ index is −1.26.
Table 2

China-Malaysia trade in SITC 8 (miscellaneous manufactures) 1994–2004

SITC

Product

Import 1994

Exports 1994

Imports 2004

Exports 2004

PQV 1994

PQV 2004

S index

MQ index

VQ index

82131

FURNITURE, N.E.S., OF METAL, USED IN OFFICES

8181

26589

368943

498170

0.51

0.44

0.12

0.30

0.42

82139

FURNITURE, N.E.S., OF METAL, NOT OFFICES

23455

8487

1094637

4158148

0.58

0.22

0.37

−0.57

−0.20

84699

CLOTHING ACCESSORIES, N.E.S.

47127

238113

18371

213583

1.55

0.46

0.07

−0.92

−0.84

84822

RUBBER GLOVES

15926

57186

9754734

5548757

0.80

1.02

−0.23

0.02

−0.21

87145

MICROSCOPES, N.E.S.

6254

148669

350646

294896

0.03

0.02

−0.29

−0.50

−0.79

87325

SPEEDOMETERS AND TACHOMETERS; STROBOSCOPES

1143

31800

13676

44864

0.28

0.00

0.02

−0.59

−0.57

87443

SPECTROMETERS, SPECTROPHOTOMETERS

15199

2100

63762

28480

0.43

0.06

−0.23

−0.55

−0.78

87451

BALANCES OF A SENSITIVITY OF 5 CG OR BETTER.

17716

21546

316317

141206

0.15

0.95

−0.30

0.48

0.18

87452

INSTRUMENTS, APPARATUS OR MODELS

29787

198089

149882

573981

0.10

0.88

0.34

0.50

0.84

87455

HYDROMETERS AND SIMILAR FLOATING INSTRUMENTS, THERMOMETERS, PYROMETERS, BAROMETERS.

22783

53839

224293

252291

0.00

0.43

−0.01

0.50

0.49

87465

AUTOMATIC REGULATING OR CONTROLLING INSTRUMENTS AND APPARATUS, N.E.S.

546424

113714

1859079

580956

0.01

0.28

−0.32

0.50

0.18

87478

INSTRUMENTS AND APPARATUS FOR MEASURING OR CHECKING ELECTRICAL QUANTITIES, N.E.S.

32884

49139

15943282

5941357

0.00

0.01

−0.32

−0.50

−0.82

88114

PARTS AND ACCESSORIES FOR PHOTOGRAPHIC (OTHER THAN CINEMATOGRAPHIC) CAMERAS

99548

17067

639517

1757079

0.43

0.22

0.35

−0.80

−0.45

88415

SPECTACLE LENSES OF GLASS

163791

190139

2795

32394

0.77

0.15

0.01

−0.64

−0.63

89281

PAPER AND PAPERBOARD LABELS OF ALL KINDS

280524

10891

1438558

304885

0.47

0.34

−0.37

−0.41

−0.78

89311

SACKS AND BAGS (INCLUDING CONES) OF PLASTICS

233058

83762

2854044

1809818

1.03

0.59

−0.17

−0.68

−0.85

89319

ARTICLES FOR THE CONVEYANCE OR PACKING OF GOODS, N.E.S., OF PLASTICS; STOPPERS, LIDS

1220511

135780

10517878

6938614

0.79

0.44

−0.13

−0.91

−1.04

89395

FITTINGS FOR FURNITURE, COACHWORK OF PLASTICS

11706

2513

168466

192745

0.30

0.36

0.09

0.50

0.59

89399

ARTICLES OF PLASTICS, N.E.S.

1054715

3138952

7018762

21967253

0.79

0.34

0.34

−0.56

−0.22

89594

TYPEWRITER OR SIMILAR RIBBONS, INKED OR OTHERWISE PREPARED, ON SPOOLS ETC.

9405

6510

516928

185156

1.30

0.96

−0.32

−0.94

−1.26

89851

PREPARED UNRECORDED MEDIA FOR SOUND RECORDING OR SIMILAR RECORDING

1072

830132

27219896

6358236

1.54

0.75

−0.40

−0.47

−0.87

89983

PRESS-FASTENERS, SNAP-FASTENERS AND PRESS-STUDS AND PARTS THEREFOR; BUTTONS

4151

154951

41431

313574

0.50

0.49

0.38

−0.36

0.02

89985

SLIDE FASTENERS

637

132656

100315

280754

0.34

0.83

0.16

0.27

0.42

COMTRADE and authors own calculations. All values are in $US in 1994 prices. The selected products are those with significant levels of matched trade

In contrast, the best performing product from a Chinese perspective is SITC 87452 (Instruments apparatus or models) with a VQ index of 0.84 derived from a positive S index of 0.34 and a positive MQ index of 0.50. It is worth noting that of the 23 products, the PQV indices are almost all below one which means that China consistently exports the low quality variety of the product to Malaysia. It is interesting to note that China’s relative quality has not been increasing despite anecdotal evidence that China has been experiencing rapid increases in the quality of its products for export (Rodrik 2006). Our results, for this admittedly small sub-set of products, suggest that Malaysia has been increasing its quality of products more rapidly than China. Such a pattern is however still consistent with China increasing the absolute quality of its goods.

Finally, we take the 23 products from Table 2 and present them in a QTAS diagram in Fig. 2. There is a relatively even spread of products across all four quadrants. From here it is straightforward to assess the changes in volume and quality visually and to make potentially bold statements. Although we caution against any knee jerk reactions due to the lack of theoretical grounding of this analysis, one could propose that the four products in the top right hand quadrant would be of most concern for the Malaysian government as they appear to have experienced deterioration in both volume and quality. Conversely, the Chinese government could examine the eight products in the bottom left quadrant to see how trade in these products has changed with Malaysia and then to extend this to look at other countries to see if this trend is widespread.
Fig. 2

QTAS and VQ indices for SITC 8 (Miscellaneous Manufactures) 1994–2004

5 Conclusions

The relationship between trade and quality is being increasingly studied by trade economists. The contribution of this short paper is to provide a modest but potentially useful and simple to execute extension to the MQ methodology outlined in Azhar and Elliott (2008). Our approach is to combine the MQ index of changes in quality in matched trade changes with the more traditional volume based measure of IIT represented by the S index (Azhar and Elliott 2003). This paper provides both a new method for visualising changes in volume and quality in one diagram but also provides a new summary index.

In future research we would like to take this new measure to the data to estimate the impact of S and MQ individually and interacted together via the VQ index on labour adjustment costs following the empirical methodologies of Greenaway et al. (2002; Brülhart and Elliott (2002); Brülhart et al. 2006; Cabral and Silva 2006; Ferto 2005). We believe that this extension provides a useful addition to the existing measures of trade induced adjustment that accounts for changes in quality that we believe to be an under researched area but potentially important element in the IIT debate.

For policy makers, a complete family of measures are now available and presented in Table 1. When these approaches are combined in a systematic way policy makers would be able to access a complete picture of changes in trade patterns. Given the large percentage of trade that is intra-industry, the ability to track a countries changes in quality and value will be of increasing importance.

Footnotes
1

Theoretically, Falvey (1981), Falvey and Kierzkowski (1985) and Flam and Helpman (1987) show that, even without increasing returns to scale, large numbers of firms will produce varieties of different quality. More recently, an empirical literature has developed to quantify the importance of product variety (see e.g. Hummels and Klenow 2005; Schott 2004; Broda and Weinstein 2006).

 
2

The precise meaning of the SAH remains subject to differing interpretation although an excellent clarification of the three key components of the SAH, which are trade as an exogenous variable, adjustment costs and IIT, can be found in Brulhart et al. (2006). There is insufficient space to go into detail although simply put, “trade induced” adjustment refers to changes in the domestic market that can be traced back to changes in prices or volume relative to the rest of the world. See also Davidson and Matusz (2001, 2004) and Bacchetta and Jansen (2003) for recent assessments of trade-induced adjustment costs.

 
3

Eicher (1996) discusses the relationship between human capital and technological change with higher technological innovation (improvements in quality) being associated with more skilled labour (skill-biased labour demand) and an increase in the relative wage of skilled to unskilled labour.

 
4

Azhar et al. (2008) follow the three stage approach and investigate the robustness and sensitivity of the existing approaches to measuring VIIT and HIIT (Greenaway et al. 1999; Fontagné and Freudenberg 1997; Azhar and Elliott 2006) using data on the nature of trade flows between China and its East Asian neighbours.

 
5

The premise for using UVs is that goods of a higher quality should demand a higher price (Stiglitz 1987). Thus, price can be considered an imperfect indicator of quality. Greenaway et al. (1994) and Aiginger (1997) provide discussion. Such comparisons however ignore the fact that that prices may also differ due to comparative advantage (Hallak and Schott 2005). Schott (2004) demonstrates that exporter skill and capital abundance are positively related to the within-product variation in export unit values. Unit values as a surrogate for price indices has also been criticised for being liable to measurement bias due to poor data quality and the mix of heterogenous items within product groupings (Silver 2007).

 
6

Rodrik (2006) and Azhar et al. (2008) use UVs to compare the quality of Chinese exports with those from other East Asian countries while Schott (2006) compares Chinese export UVs within different product categories to the prices received by other US trading partners. See Hummels and Klenow (2005) for a general study looking at the relationship between the variety and quality of a nation’s exports and Funke and Ralf (2001) for an East Asian study of export variety and export performance.

 
7

The choice of the 85% cut-off point is arbitrary but the choice is no different to the 15% or 25% cut-off used in previous measures. See Azhar and Elliott (2006) for a longer debate on this issue.

 

Acknowledgements

A.K.M. Azhar is grateful to the Malaysian Ministry of Science, Technology and Innovation for funding grant number 54322. The usual disclaimer applies.

Copyright information

© Springer Science+Business Media, LLC 2010

Authors and Affiliations

  • Abdul K. M. Azhar
    • 1
  • Robert James Ross Elliott
    • 2
  1. 1.GSMUniversiti Putra MalaysiaSerdangMalaysia
  2. 2.Department of EconomicsUniversity of BirminghamBirminghamUK

Personalised recommendations