The DNA of the domestic value added (DVA) in exports: Firm‐level analysis of DVA in exports

• DOI: http://doi.org/10.1111/twec.12800
• Published date: 01 September 2019
• Author: Nataša Vrh
• Date: 01 September 2019

2566

|

wileyonlinelibrary.com/journal/twec World Econ. 2019;42:2566–2601.

© 2019 John Wiley & Sons Ltd

Received: 30 September 2018

|

Revised: 8 February 2019

|

Accepted: 3 April 2019

DOI: 10.1111/twec.12800

ORIGINAL ARTICLE

The DNA of the domestic value added (DVA) in

exports: Firm‐level analysis of DVA in exports

NatašaVrh

Faculty of Economics,University of Ljubljana, Ljubljana, Slovenia

Funding information

Javna Agencija za Raziskovalno Dejavnost RS, Grant/Award Number: 36385

1

|

INTRODUCTION

The emergence of global value chains has greatly transformed the ways multinational companies

(MNCs) are organised (Lanz & Miroudot, 2011), with international trade in tasks overtaking trade in

goods. By geographically spreading different stages of production, the value added across the value

chain has become unevenly distributed. Thus, a firm's position in the value chain plays an important

role in its ability to create, improve and retain value and considerably affects its competitiveness

(Pavlínek & Ženka, 2010). Firms broadening their boundaries and having a stronger focus on their

core competencies, outsourcing and offshoring have grown in importance. Offshoring was initially

mainly concerned with an MNC relocating its manufacturing stages of production to affiliates or local

suppliers in countries with low labour costs. More recently, offshore activities have evolved to also

include upstream tasks previously performed solely in the parent firm's home country, like R&D or

the design of advanced technology (Antràs & Yeaple, 2014; Nieto & Rodrίguez, 2011).

Firms participating in global value chains (GVCs) interact with each other through a network of affiliates

(via foreign direct investment; FDI) or have contractual/arm's‐length trade connections1 to ensure they are

supplied with inputs (thus playing the role of “value chain organisers”) or as suppliers themselves. Firms that

create their “own” GVC perform (depending on their strategic orientation) one or more key production pro-

cess stages/functions together with their affiliates (“offshoring”), while the other stages required to achieve

the final product are contracted to arm's‐length suppliers at home or abroad (outsourcing; Stare, 2016). The

first approach associates GVCs with FDI flows and is characterised by a relationship whereby the parent

provides its subsidiaries with inputs or as part of a relationship among subsidiaries with the same parent. In

this case, the trade in intermediate goods entails intra‐firm transactions with production stages located

around the world (Amador & Cabral, 2014). In the second approach, firms engage in GVCs through

1 Besides intra‐firm trade (via FDI), firms can participate in GVCs through a contractual relationship with an MNC (in non‐equity

modes—NEMs) or through arm's‐length transactions. The GVC literature mentions three main NEM types: captive, modular and

relational. It is estimated that around 16% of total global trade that involves an MNC is linked to NEM‐related trade. Arm's‐

length and intra‐firm trade each separately represent 6.3% of total global trade that involves an MNC (UNCTAD, 2013).

|

2567

VRH

arm's‐length/contractual transactions or inter‐firm trade with unrelated parties where firms export inputs to

international buyers or import intermediates in the role of domestic final producers (Taglioni & Winkler,

2016).

However, it is only the United States that collects detailed trade statistics that distinguish between intra‐

and inter‐firm trade (Antràs, 2016). In 2009, US intra‐firm trade accounted for 48% imports and about

30% exports of goods, while US arm's‐length trade between unrelated parties represented 50% of imports

and almost 70% of exports of goods. Such statistics are rare for other countries and only available for ex-

ports by foreign affiliates (Lanz & Miroudot, 2011). It is therefore difficult to estimate the average share

of intra‐ and inter‐firm trade at the global level due to the large variations across countries and industries.

The central aim of this paper is to analyse the process of firms upgrading GVCs in terms of domes-

tic value added in exports (DVA) relative to forms of GVC participation and selected firm characteris-

tics. My analysis concentrates on exporting firms which, in theory (Helpman, Melitz, & Yeaple, 2004;

Melitz, 2003), are relatively more productive than firms that exclusively supply domestic markets.

Only the most productive exporting firms engage in FDI (i.e., foreign‐owned firms exhibit higher

productivity than locally owned firms). However, based on existing theory and empirical findings

for Austrian firms, Pfaffermayr and Bellak (2002) investigate possible reasons for one type of firm's

superior performance vis‐à‐vis another. They notice that in certain cases, domestic‐owned firms may

even outperform foreign ones so long as the firm has a multinational character.

This paper examines possible evidence in support of the claim that in terms of DVA in exports,

locally owned exporting firms outperform more productive foreign‐owned firms that generally supply

more demanding and developed markets. The analysis builds on two main research questions: (a) How

does DVA in exports vary depending on the way a firm participates in GVCs (through either a network

of affiliates or inter‐firm trade) by distinguishing firms that are themselves suppliers or have their

own suppliers located abroad? and (b) Which firm characteristics affect a firm's “success in DVA” as

measured by the ratio of DVA in exports. In order to control for firms' heterogeneity, this paper adapts

the recently established methodology for measuring value added in exports using firm‐level data de-

veloped by Kee and Tang (2016).

I focus on these questions by examining Slovenian exporting companies, which are (compared to

MNCs from the region) smaller but relatively more engaged in global trade and have functionally and

geographically dispersed affiliates (Jaklič & Svetličič, 2008). The GVC participation index2 for

Slovenia (58.7) is above the average for developed (48.0) and developing (48.6) economies and one of

the highest among CEE countries according to WTO data for 2011 (WTO, 2017). The recent AMNE

database shows that exports by MNCs (foreign‐affiliate and domestic MNCs) account for more than

70% of Slovenian gross exports (Cadestin, De Backer, Desnoyers‐James, Miroudot, Ye, & Rigo, 2018).

Decisions by firm regarding their participation form are analysed at the firm–country–product

level by considering foreign direct flows and the stability of trade flows. To define inter‐firm rela-

tionships, I rely on detailed transaction‐level data for the population of Slovenian firms at the level of

6‐digit product groups (HS6) and country of destination for the period 2002–14. The detailed export

data by firms, years, destination country (country of dispatch) and products on the HS6 level so ob-

tained allow me to measure firms’ export‐sales stability (the stability of a firm's supply from abroad)

where stable sales indicate a firm is a permanent supplier (or has a permanent supplier from abroad)

based on product‐destination stability. In addition, I matched these data with data on FDI flows and

firms’ balance sheets.

2 The participation index consists of two components: share of foreign inputs (backward GVC participation) and domestically

produced inputs used in third countries’ exports (forward GVC participation) in gross exports (WTO, 2018).

2568

|

VRH

The primary task of many global manufacturing firms in developed economies today is actually not

manufacturing but to provide product design, marketing, logistics, etc., since the goal of their offshoring

activities (through either FDI or at arm's length) is to transfer other stages and functions (usually pro-

duction‐related) to others, while focusing on their core competencies (Milberg & Winkler, 2010). While

this seems the dominant motivation for developed‐country outsourcing, others, such as outsourcing

skill‐intensive phases of production or proprietary inputs, are also likely and may ultimately be impact-

ing the sign and significance of the coefficients. Based on this international trade phenomenon, I expect

firms with foreign ownership and unaffiliated firms identifiable as independent permanent suppliers

for foreign firms to have a lower DVA value, while firms that themselves have foreign affiliates and

unaffiliated firms with permanent suppliers from abroad will capture higher DVA in exports.

This paper makes two contributions to the literature. First, departing from the most common ap-

proach in the literature of using industry‐level data from I‐O tables (Amador & Cabral, 2014), this

paper develops a method for estimating DVA in exports with firm‐level data by adapting the approach

taken by Kee and Tang (2016). The use of firm‐level data allows smaller firms not captured in I‐O

tables (i.e., firm size sample used to construct I‐O tables may be consisted only from large firms) to

be included and avoids possible aggregation biases (i.e., establishing a common classification while

constructing the WIOD tables necessitates the aggregation and disaggregation of national supply and

use tables; trade flow data in WIOD are based on the UN COMTRADE database which includes

trade statistics by commodities and partner countries on an annual basis, etc.). A micro‐level approach

considers firm heterogeneity and controls for firm characteristics to estimate a firm's DVA in exports.

Second, since the literature does not provide guidance on how to define firms participating in GVCs

via a contractual relationship (between unaffiliated firms) and there are insufficient data to allow a bet-

ter definition of such trade linkages, I proxy for the possible underlying relationship by exploring the

stability of firm trade flows. The availability of highly detailed import and export flow information by

products at the 6‐digit level and origin/destination country helps identify unrelated firms participating

in GVCs via a contractual relationship (on both import and export sides) by using a correlation method

between trade and their lagged values (month/month (year−1)) by firm, year, 6‐digit product and coun-

try of destination/country of dispatch to define trade flow stability. This enables me to distinguish two

modes of firms’ organisational structure in GVCs (through either FDI or a contractual relationship).

The rest of the paper is structured as follows. Section 2 presents the related empirical literature,

Section 3 provides details of the DVA estimation at the firm level, Section 4 describes the data used

and empirical framework, and Section 5 outlines the results while Section 6 concludes.

2

|

RELATED RESEARCH LITERATURE

In the area of the ways firms upgrade3 in GVCs, the micro‐level empirical evidence on the factors

influencing how firms achieve such GVC upgrades remains underexplored. Existing evidence is still

mostly based on case studies or company interviews and is frequently limited to certain sectors (e.g.,

Aspers, 2010; Azmeh & Nadvi, 2014; Isaksen & Kalsaas, 2009; Jürgens & Krzywdzinski, 2009;

Kadarusman & Nadvi, 2013; Pickles, Smith, & Begg, 2006; Tokatli, 2007; among many others).

3 Humphrey and Schmitz (2002) identify four different ways of upgrading: (a) process upgrading refers to more efficient and

organised production or implementation of new technologies; (b) product upgrading relates to the production of more

complex products; (c) functional upgrading relates to an increase in the production's overall skill content; and (d) inter‐sec-

toral upgrading is related to the move to new productive activities (industries).