Market structures in small open economies: Evidence from Denmark

Abstract

This paper provides some stylized facts about market structure in Denmark, a country exhibiting high rates of exports and imports as is common in small developed economies. Utilizing disaggregated data at the firm-product level for manufacturing industries, we highlight the widespread presence of industries that are neither purely oligopolistic or monopolistically competitive; rather, they contain a few domestic leaders with numerous firms having insignificant domestic market shares. We also document that, relative to the latter type of firms, leaders have greater labor productivity, are more capital intensive, and pay higher wages; additionally, they are more likely to export and import, although they exhibit a greater domestic intensity relative to exporters with negligible domestic market shares. Finally, through a model of leaders and followers, we investigate how leaders can benefit from acting strategically against small firms and quantify its potential impact on industry outcomes through a numerical exercise.

Introduction

One common feature of small countries that distinguishes them from large economies is their heavy dependence on international trade (Benito, Larimo, Narula, Pedersen, 2002, Van Den Bulcke, Verbeke, 2001, Hannerz, Gingrich, 2017). Regarding exports, this is a consequence of the restricted size of their home market, which makes firms commonly sell abroad to increase their production scale and, hence, operate efficiently. As for imports, their limited variety of resources and market size create hurdles to supply the whole diversity of goods that consumers demand. Thus, these countries end up importing more intensively and exposing local firms to tougher competition.

Given these distinctive features of a small open economy, what kind of market structures arise? In this paper, we provide some evidence on the matter by identifying stylized facts for Denmark. Our findings highlight the ubiquity of industries with coexistence of domestic leaders and numerous negligible firms. For these industries, we identify empirical features of their firms and present a model that highlights the incentives of leaders to behave strategically against the negligible firms.

In Section 2, we identify patterns regarding the market structure of Danish manufacturing industries. Our findings are based on highly disaggregated data regarding Danish manufacturing, which primarily come from two sources: Danish Prodcom statistics and international transactions collected by Danish customs. These datasets are part of the country’s official statistics and cover almost all the international transactions and more than 90% of total production by industry.

Two features make these datasets particularly suitable for the analysis of market structure. First, the information is presented at the firm-product level and disaggregated at the 8-digit product level, making it possible to allocate each firm-product to an appropriately defined industry. In this way, we avoid issues arising with typical balance-sheet data declared for tax purposes, where a firm’s total revenue is usually allocated to the specific industry that generates the greatest bulk of it.

Second, the information on international transactions is also presented at the firm-product level and encompasses imports by both manufacturing and non-manufacturing firms. This becomes particularly relevant for small economies like Denmark, where import competition is pronounced, since the importance of a domestic firm within an industry can only be measured once that imports are accounted for.

By obtaining the domestic market share of each domestic firm in terms of expenditures (i.e., domestic sales plus imports), our findings are as follows. First, around half of the industries have numerous Danish firms with negligible domestic market shares operating. Throughout the paper, we refer to these firms as domestic non-leaders (DNLs). Remarkably, when we measure the importance of these industries by total manufacturing expenditure, we find that they determine the majority of it. Specifically, more than 80% of total expenditure is covered by them.

In addition, we show that DNLs are not the only type of firms operating in these industries. Rather, they commonly coexist with a handful of firms that accrue great domestic market share, which we refer to as domestic leaders (DLs). Specifically, among industries with numerous DNLs, more than 85% of them have at least one DL and they represent more than 80% of the total manufacturing expenditure.

Delving into properties of these industries, we analyze how DLs compare with DNLs across several features. The results point out that DLs have greater labor productivity (i.e., revenue per employee), are more capital intensive, and pay higher wages. Additionally, they are more likely to engage in exporting and importing activities, although a different picture emerges when we measure their degree of internationalization through domestic intensity (i.e., a firm’s domestic sales relative to the sales at home plus exports): conditional on exporting, DLs tend to display greater domestic intensity than DNLs. The result becomes relevant in light of the mixed empirical evidence on this matter² and the debate on whether a firm’s position at its home market predicts a specific type of export behavior (for classical studies regarding this, see Mascarenhas, 1986, Bonaccorsi, 1992, Porter, 1998).

Based on the ubiquity of these types of industries, in Section 3 we investigate the implications that such a market structure might entail. In particular, the asymmetry of these firms in terms of size and position at home can lead DLs to make strategic moves against DNLs to gain a better position domestically (see Kwoka and White 2001 and D’Aveni 2002 for several examples of such behavior).

To explore this matter, we present an open-economy model where DLs are represented as in monopolistic competition and embed a set of non-negligible heterogeneous firms to capture the presence of DLs. Moreover, we suppose that DLs engage in strategic moves and make investment decisions to gain a better position locally. To isolate the strategic motive to invest by these DLs, we utilize the two-stages approach by Fudenberg and Tirole (1984). Thus, we consider a scenario where DLs decide on domestic investments prior to both entry of DNLs and the market stage. Then, we compare the outcomes of this scenario with a non-strategic benchmark where domestic investments are not observed by rival firms.

The results of the model indicate that DLs act more aggressively through overinvesting. The goal is to deter entry of DNLs and capture domestic sales that, otherwise, would go to DNLs. A corollary of this is that, even when we assume competition à la Bertrand, DLs never accommodate entry. The result clearly contrasts with what occurs in models with one incumbent and one potential entrant à la Fudenberg and Tirole (1984). This occurs because, in our model, DNLs are governed by free-entry rules. Thus, as in Etro, 2006, Etro, 2008, accommodating strategies are always unprofitable since they end up inducing additional entry and, hence, turning futile the attempt of softening competition.

To assess the potential impact of this type of behavior, in Section 4 we calibrate the model and perform some numerical exercises. We find that, on average, each DL could increase its domestic revenue by around 36%. In addition, the domestic market share of each would become greater by around 2.5 percentage points and its domestic intensity higher by around 5 percentage points. Nonetheless, the results exhibit great heterogeneity across and within industries.

Our paper is related to different strands of the literature. First, it touches upon studies that characterize markets in open economies. In particular, it is related to studies that have identified a coexistence of large and small firms. At the country level, this has been documented in Axtell (2001) for the USA and Fujiwara et al. (2004) for several European countries; at the industry level, a similar pattern has been shown by Bronnenberg et al. (2009) and Hottman et al. (2016) for various industries in the USA and Gaubert and Itskhoki (2018) for French manufacturing.

We contribute to the literature in several respects. First, the evidence that we provide covers all the industries belonging to manufacturing. Thus, in contrast to studies that analyze some specific industries, we can measure their relative importance with respect to the whole manufacturing sector. Specifically, this allows us to conclude that industries with leaders and negligible firms coexisting cover the great bulk of total manufacturing expenditure. Second, our results are based on highly disaggregated information at the firm-product level and including imports. Thus, we improve upon results utilizing balance-sheet data, which allocates a firm’s total revenue to the main activity of the firm without splitting it into the different industries from which this is generated. Finally, we focus on the case of Denmark, which is a small economy exhibiting high rates of exports and imports.³

Furthermore, our paper is related to studies that search for stylized facts of firms through the use of microdata. Regarding this, there is a vast literature initiated by Bernard and Jensen (1995) on the differences between exporters and firms that exclusively serve domestic markets. See in particular Eaton et al. (2004) for France, Mayer and Ottaviano (2008) for several European countries, and Bernard et al. (2012) for the USA. Furthermore, there is a growing literature on the so-called “superstar firms”, including Freund and Pierola (2015), Autor et al. (2017), Gutiérrez and Philippon (2017), and De Loecker and Eeckhout (2018). While these papers define large firms according to a firm’s exports or total revenue, we deal with large firms defined by their domestic market shares and, hence, their position at the home market.