Distance from core competences and new export survival: Evidence from multi‐product exporters

• Date: 01 November 2019
• Author: Daniel Goya,Andrés Zahler
• Published date: 01 November 2019
• DOI: http://doi.org/10.1111/twec.12835

World Econ. 2019;42:3253–3286. wileyonlinelibrary.com/journal/twec

|

3253

© 2019 John Wiley & Sons Ltd

Received: 26 June 2018

|

Revised: 22 January 2019

|

Accepted: 11 April 2019

DOI: 10.1111/twec.12835

ORIGINAL ARTICLE

Distance from core competences and new export

survival: Evidence from multi‐product exporters

DanielGoya1

|

AndrésZahler2

1School of Business and Economics,Pontificia Universidad Católica de Valparaíso, Valparaíso, Chile

2School of Economics and Business,Diego Portales University, Santiago, Chile

Funding information

This paper was supported by Iniciativa Científica Milenio [Nucleo Milenio NS100017 ‘Intelis Centre’] and Comisión

Nacional de Investigación Científica y Tecnológica [Becas Chile 79090016 to D.G.].

KEYWORDS

core competences, export diversification, export survival, product proximity, survival models

1

|

INTRODUCTION

This paper looks at exporters adding new products to their export baskets in a middle‐income econ-

omy. The aggregate effect of the process of micro‐level diversification appears to be important.

Approximately 60% of the larger exports of developed economies than developing economies can be

explained by differences in the breadth of their export baskets (Hummels & Klenow, 2005).There is

also evidence that export diversification, in terms of the variety of products exported, is associated

with output growth (Funke & Ruhwedel, 2001, 2005) and productivity growth (see, e.g., Addison,

2003 and Feenstra & Kee, 2008). Half of the 500% growth of the non‐mining exports of Chile (our

country of analysis) during the period 1991–2006 is attributable to products added by firms that were

already exporters, and over 95% of the export value in our last year of data corresponds to multi‐prod-

uct firms.1

Motivated by this relevance, we want to dig deeper into how this process happens by analysing

whether firm‐level competences play a role in the survival of new export products. We are motivated

by the idea that similar products require similar competences or capabilities2 on the part of the firm.

This is closely related to the concept of “core competence” used in multi‐product firm models such as

in Eckel and Neary (2010). In their model, successfully producing products “distant” from the core

competence implies greater challenges for the firm, and thus, they have a lower probability of survival

in competitive markets. Following the spirit of Prahalad and Hamel (1990) – who introduced the idea

1 These figures are calculated from Chilean customs data, which will be described in detail below.

2 The concept of capabilities is similar to that of competences and is commonly used in the evolutionary economics literature

stemming from Nelson and Winter (1982).

3254

|

GOYA And ZAHLER

of “core competences” in the management literature – we expand Eckel and Neary's logic, arguing that

firms can have not one but multiple “core competences” or capabilities and that specific combinations

are required to efficiently produce and export particular products.

We argue that a firm's export basket reveals information about the firm's multiple current compe-

tences,3 and we analyse the relationship between the “distance” between a new export product and the

firm's previous export basket, and the new export product's likelihood of survival.

We define a measure of the distance between each new export and the firm's previous export basket using

a measure of the similarity between product pairs defined by Hidalgo, Klinger, Barabási, and Hausmann

(2007), and show that it contains more information than other proxies for distance used in the literature.

For our estimations, we use firm‐level customs data from Chile for the period 1991–2006. We es-

timate linear probability models with fixed‐effects and duration analysis models with random effects.

We find that, ceteris paribus, a new export product that is at a one‐standard‐deviation greater distance

from the firm's previous export basket than another new product is 12% less likely to survive in export

markets for more than one year. This result is consistent with the predictions in Eckel and Neary (2010).

This finding is robust to firm‐level characteristics (size, experience), product‐level characteristics (ini-

tial export size, RCA of the product) and even firm‐time and product‐time fixed effects. We also find

that our distance measure primarily plays a role during the first year of exports of a new product, sug-

gesting the existence of a quick learning process as in Albornoz, Pardo, Corcos, and Ornelas (2012).

To reinforce the interpretation of our results and to explore whether what we are finding is related to

technological (supply) or demand complementarities, we conduct a series of analyses using data from

domestic manufacturing production. This allows us to check whether what we observe in exports also

applies directly to production. Moreover, we build alternative measures of distance between products, a

time‐invariant measure and a measure based on the inputs bought by firms that manufacture each pair

of products, both of which are more strongly linked to production and technology. Although they do not

rule out the possibility that demand complementarities can partly explain our results, these tests tend to

corroborate our hypothesis of the role of technological complementarities on the survival of new ex-

ports.Our findings are closely related to those by Boehm, Dhingra, and Morrow (2016), who find that

similarity in input usage is an important determinant of product adoption across Indian manufacturers.

This paper contributes to the literature in two ways. First, it provides evidence of a new determinant

of the survival of new trade flows at the firm level: the “distance” between the new export and the

firm's previous export basket. The longer the distance of a new export is, the farther away from the

firm's core competences, and thus, the lower the probability of survival, ceteris paribus. This finding

suggests that firms should move “slowly” when they diversify their export baskets. Second, our re-

sults connect the theoretical literature on multi‐product firms to the literature focusing on firm‐level

capabilities (stemming from Nelson & Winter, 1982 and Prahalad & Hamel, 1990) and on country‐

level capabilities and relatedness between products (e.g. Hidalgo et al., 2007; Neffke, Henning, &

Boschma, 2011). Our results suggest that firm‐level capabilities are important determinants of product

survival and that measures based on the “proximity” concept advanced by Hidalgo et al. (2007) are

good proxies for the theoretical “distance” from the core competence found in multi‐product firm

models such as Eckel and Neary (2010).

Hidalgo et al. (2007) show that countries gradually change their specialisation patterns, expanding

first to products that are at a short distance from their current specialisation patterns. This paper adds

a micro, within‐firm element, showing that firms are also more likely to be successful in introducing

new export products when they move gradually.

3 By doing this, we assume that production competences are the same as export competences. For a discussion on this, see

Section 6.

|

3255

GOYA And ZAHLER

The paper is structured as follows: Section 2 reviews the related theoretical and empirical litera-

ture. Section 3 describes the data we use and discusses the definition of the distance variable and how

we define a “new” export. We also present some preliminary evidence showing that the measure

proposed is consistent with theoretical models. Section 5 presents the main result, namely, that the risk

of failure is higher for new exports that are more distant from the firm's previous export basket. In

Section 6, we discuss potential alternative explanations of our findings and provide analyses to try to

disentangle them from our hypothesis, with the help of alternative distance measures and the use of

domestic manufacturing data.4 Section 7 presents a number of additional robustness tests for the main

result. Finally, Section 8 concludes.

2

|

PREVIOUS RESEARCH

Our paper builds upon three strands of literature, which we analyse below: first, on the theoretical

literature on multi‐product firms; second, on the papers that define measures of distance between

products; and, finally, on the trade literature on export survival. Below, we explain the relation of our

work with each.

2.1

|

Multi‐product firms

There are several approaches to modelling the production and export behaviour of multi‐product

firms, two of which are particularly relevant for our question. Bernard, Redding and Schott (2011)

developed a model in which productivity in a given variety is determined by two parameters: a firm‐

level measure of ability and a firm‐product specific expertise, which is uncorrelated across products.5

Eckel and Neary (2010) used a different approach that assumes that each firm has a core competence.

In this model, there is a continuum of product varieties. A given core competence locates a firm on

that continuum. Products at greater distance from the firm's core competence are produced less effi-

ciently (at a higher marginal cost).6 The products at a greater risk of being abandoned after a negative

shock to a firm are those that are further away from its core (Eckel & Neary, 2010) or those for which

the firm has the lowest product‐specific expertise (Bernard et al., 2011).

To understand whether survival depends on relationships between products, each model contrib-

utes one important point but excludes another. Bernard et al.'s (2011) model is interesting in that it

implicitly recognises that firms require different capabilities to produce different goods. However,

expertise is orthogonal across goods, leaving no room for heterogeneous technological relationships

between products due to, for example, common expertise and ability or the capabilities required to

produce and export a certain pair of goods. Eckel and Neary's (2010) model defines a distance be-

tween different varieties. However, since all of the varieties are placed along a single continuum, the

possibility of multiple underlying capabilities is ruled out. Prahalad and Hamel (1990) refer to “core

competences,” in plural, and argue that firms use different combinations of core competences for dif-

ferent goods. An apt example that they provide is Canon: this firm's core competences are precision

4 We base our analysis on export data and not on manufacturing, because: (a) our interest lies in understanding the micro‐level

dynamics of export diversification processes; (b) with customs data we can explore all sectors of the economy; and (c)

manufacturing accounts for only approximately 10% of Chilean GDP and of its exports.

5 These can also be interpreted as firm productivity or consumer taste parameters.

6 Mayer et al. (2014) develop a model that employs a similar idea of “core competence” but uses monopolistic instead of

oligopolistic competition.