Globalization and the size of the largest firms : a long-term view on the cases of Germany and the United Kingdom

(1)

Master Thesis in Political Science - track Political Economy: Our Changing Global Economic Order: Growing Risks and Imbalances

Globalization and the Size of the Largest

Firms:

A Long-Term View on the Cases of Germany

and the United Kingdom

Ádám Forgács

11701587

First reader: Dr. L. A. Linsi

Second reader: Dr. E. M. Heemskerk Date: 12. 06. 2018.

(2)

2

Fears about the continuously growing size of corporations arise time after time. It is commonly brought forward by the politically motivated groups of the radical left, anti-globalization movements, or by the ordinary man who afresh experienced the Kafkaesque nature of a multinational corporation, by calling a customer service, situated on the other end of the planet. The commonsensical explanation behind these fears starts to be, single-handendly, globalization. When something becomes such a widespread part of the common sense, social scientist should delve into the topic with even greater appetite for critical scrutiny.

First, let us see, why it is sound to think that globalization makes firms bigger. Approaching from the demand side, corporations need markets to grow. Liberalizing world trade means more, easily-accessible markets. Due to economies of scale and scope, firms get more efficient, thus a firm is just going to be larger if given the opportunity to export goods to more markets. Simple reasoning, larger the market, larger the firm. If we add liberalized capital flows to the equation, we get the same result. More capital can be invested (since the market is larger), on more capital, more returns (not necessarily higher) can be made, there is more capital available at the end of the cycle, thus firms get bigger. It seems, from the viewpoint of common sense, that liberalizing trade and capital flows, essentially economic globalization, unleashes an immense, never-seen extent of corporate size, which can grow to infinity.

But is it reasonable to believe this explanation regarding the relationship of firm size and globalization? The data makes us think twice.

4.1% 5.6% 8.4% 8.4% 7.7% 1907 1917 1927 1937 1947 1957 1967 1977 1987 1997 0% 1% 2% 3% 4% 5% 6% 7% 8% 9% 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 19 07 19 10 19 13 19 16 19 19 19 22 19 25 19 28 19 31 19 34 19 37 19 40 19 43 19 46 19 49 19 52 19 55 19 58 19 61 19 64 19 67 19 70 19 73 19 76 19 79 19 82 19 85 19 88 19 91 19 94 19 97 20 00 Size of the la rge st fir ms W or ld tr ade a s % of G DP

Globalization and firm size in the 20th century (1907-2002, Gospel and Fiedler (2008), Jones (2005), Penn World Tables 8.1, Klasing and Milionis (2014))

World trade as % of world GDP Top 100 firms as % of total labour force Figure 1: Largest firms in developed economies and globalization in the 20th century

(6)

6

On Figure 1, we can see the relative size of the largest firms and globalization. Firm size is operationalized by the number of employees of the top 100 companies in the five most developed economies: United States, United Kingdom, Germany, Japan and France; relative to the active labour force, in each year (Gospel and Fiedler, 2008). Globalization is operationalized by world trade (exports and imports). What we see on Figure 1, is the end of the first phase of globalization (1820-1914) and the whole second phase of globalization (1960-) (Baldwin and Martin, 1999). Surprisingly, in the light of the outlined common-sense explanation, relative firm size was growing in the first, stagnating and then declining in the second phase. What should explain this?

1.2. Existing answers and the research gap

There are three streams of literature, which concern the development of firm size over time. One is influenced by the field of business history, one by standard economics and one by evolutionary theory.

Mostly the studies by business historians focusing on the size of the largest firms in developed economies (Fiedler, 1999, Wardley, 1991, 1999, Jeremy, 1991, Payne, 1967,) are collecting and systematizing data and discovering patterns of firm survival over time. Alfred Chandler, the doyen of the field, published three seminal works on the matter (1962, 1977, 1990). His argument throughout the books is that technological and market developments resulted in the emergence of large, vertically integrated firms. The analysis of his time-span ends before the late 20th_century, thus does not include the decrease in question. There are several studies (Cassis, 1997, 2000, 2008, Fiedler and Gospel, 2008, 2010), which aim to elucidate the long-run dynamics of big business. Both, Cassis and Gospel and Fiedler, take the 20th_{century as their time span. In these} studies, the focus is on the organizational, structural economic and technological developments shaping firm size, neglecting the role of globalization and politics. It can be concluded, that none of the cited studies take account of politics, and all of them fail to explain the recent downturn in the size of the largest firms.

Economics literature concerning firm size over time focuses solely on the distribution of firm size, uses quantitative methods and takes shorter time periods than business historians. There are three currents in the literature, one examines firm size and globalization and yields inconclusive results (Strobbe, 2002, Georgiou, 2013, Biermann, 2016, Colantone et al, 2008 and Nocke and Yeaple, 2006, 2014). Another, line of scholarship examines the relationship between firm size, export activity and productivity, and concludes that firms that export are more productive and tend to be larger (Di Giovanni et al, 2011, Melitz, 2003, Sun and Zhang, 2012, Lee, 2014). A third direction in the academic literature (Lucas, 1978, Poschke, 2014, Aquilina et al 2006, Congregado et al, 2012) investigates the hypothesized positive relationship between firm size and economic

(7)

7

development. The findings of the studies come out as inconclusive. In all of the abovementioned studies the longest time span analysed is between 1972 and 2008 by Congregado et al (2012). Evolutionary economists have the most developed explanations for the puzzle outlined in Figure 1. Especially Langlois’s compelling theory of the vanishing hand (2003, 2007) comes close, which explicitly aims to explain the inverted U-shaped curve of the size of the largest firms in the broad 20th_{century. He bases his theory on the Smithian process of division of labour, which is} determined by the extent of the market. He argues that the Chandlerian era was an anomaly in the long-term development of firm size. The reason being, that in times of economic transformation the capabilities needed for growth are not or expensively available on the market, thus it was necessary to have them in house. The market needed to develop to provide opportunities for specialization. But, there are several shortcomings of Langlois’ work, namely the lack of empirical evidence, the sole focus on the USA and the relative neglect of institutional and political influences, which affect market size. Therefore, I aim to complement, his nevertheless convincing theory of the vanishing hand.

As we can see there is a scarce amount of literature which precisely concerns the puzzle outlined in Figure 1. The existing studies in the field of business history cannot find a common ground to explain the long-term dynamics of the size of the largest firms, economics studies are inconclusive, while the evolutionary theory of Langlois needs an empirical backing. In case of the quantitative studies the span analysed is rather short, and never incorporates more than one phase of globalization. Moreover, each of the cited studies examines a time period which falls into the times of expanding globalization. Thus, each finding can be just partial, because, as seen on Figure 1, the unfolding of globalization is a distinctively long-term trend. Studies of business history and Langlois incorporate a longer time span, which smooths out the cycles of globalization. The cited papers by economists examine the distribution of firm size. This approach is superior, since it gives the whole picture, but unfortunately it is a trade-off with the time span. Due to non-existent data a truly long-term take cannot be made on firm size distribution. A solution to bridge this problem is to focus on the largest firms, as business historians do.

However, which both stream of scholarship lack is the explicit consideration of globalization and the underlying international and domestic political factors as the potential explanation for firm size. While business historians partially consider domestic, industrial policy as a potential explanation, they lack the international aspect. Economists do not consider politics at all in the cited studies. As described before the evolutionary school takes account of institutions, but lacks international influences and empirical evidence, which weakens the power of the argument. Additional shared shortcoming between the approaches is the monolithic take on trade. There is

(8)

8

no analysis of trade patterns in a higher degree of disaggregation, be it industry or trade partner level, and they not mention trade policy in any sense.

Based on the above critique of the existing literature, it is possible to conclude, that the puzzle presented in Figure 1 is not yet answered, thus leaving an important gap in our contemporary knowledge of the interplay of firm size and globalization in developed countries. The aim of this thesis is to fill this gap. To do so, the answer is sought for the following overarching research question:

What determined the size of the largest firms in developed economies since the beginning of the 20th century and since the downturn in their size in the 1970s?

1.3. The argument

To answer this question, it is needed to rediscover Adam Smith as a trade theorist. Smith has two important arguments, first that specialisation is an endogenous process to trade (Buchanan and Yoon, 2002 as cited in Schumacher, 2012, Smith, 2012 [1776]), second that the division of labour is limited by the extent of the market. New trade theory is the intellectual descendant of this logic in trade literature. The correspondence of the Smithian approach to new trade theory can be seen in the way how Krugman defines increasing returns to trade due to specialization: “countries were

identical in resources and technology would nonetheless specialize in producing different products, giving rise to trade as consumers sought variety” (Krugman, 2008). This line of thinking, starting

with Smith and continuing with Krugman, explains the recently dominant rich-rich pattern of trade. According to Smith, trading with developed nations broadens the market more, which makes more advanced division of labour possible (Schumacher, 2012, Smith, 2012 [1776]), while Krugman (1980, 2008) explains this with the phenomenon of intra-industry trade and economies of scale. The rich-rich and the underlying intra-industry trade pattern is important, because it started to emerge around the same time as Figure 1 shows stagnation, and the subsequent decline of the relative size of the largest firms (Balassa, 1966, Krugman, 1980, 2008, Baldwin and Martin, 1999). However, the question arises, why would the emergence of intra-industry trade explain the variance in relative size of the largest firms? To answer this question, we first need a leap back in history.

Before the first phase of globalization, and the revolutionary development in transportation and communication in the 19th_{century, even internal markets in a country were highly fragmented.} These were supplied by all-around middle-men merchants, who possessed the most information and made possible the movement of goods between different stages of production (Langlois, 2007). Railways and the telegraph meant unprecedented improvement in business, not just externally, in sales, but also internally, regarding organization. Around the 1900s it became

(9)

9

technically possible to efficiently organize work on a truly large scale, thus bringing to life the large, vertically integrated firms and managerial capitalism, accurately described by Alfred Chandler (1977). One can however ask, that these forces are even more accentuated today, with the presence of commercial air transport, the internet and high-speed trains, then why relative firm size is not growing continuously in developed countries? Here comes to the picture the above-mentioned pattern of intra-industry trade. During the Chandlerian era, up to the 1970s, vertical integration was the most prominent trend in the structure of the largest firms (Langlois, 2007, Chandler, 1977, 1990). It was regarded best to have all stages of production under one roof, to take advantage of the spreading of fixed costs on more products, thus reaching economies of scale and becoming more efficient. Later, horizontal integration started to emerge, based on the proposition that managers can manage everything efficiently, which requires management. This trend followed the same logic, the cost of management was spread over more divisions, thus efficiency is higher, and the company as a whole is more competitive (Langlois, 2007).

Having everything under the same roof was desirable up to the 1970s, but suddenly it became not that desirable afterwards. This, as mentioned, coincides with the emergence of intra-industry and rich-rich trade, which, as Smith and the new trade theory asserts, means a more nuanced specialization in case of developed countries. For instance, Germany makes exceptional premium cars, while France produces less exclusive, but cheaper vehicles. This division makes it rational for these countries to trade in cars, even though that both manufacture them. However, staying with automobiles, not just the end products got more specialized but the needed parts, as well. Thus, French automakers buy German steel pistons and German automakers buy French steering systems. To slice up the value chain in the described manner, countries needed to be on a similar level of technological development and productivity, which explains the rich-rich pattern (Krugman et al, 1995). Such deepening of international value chains makes firms smaller, bringing transactions out from within organizational hierarchies to markets (Coase, 1937, Williamson, 1975). Plus, due to the Smithian division of labour, this was more efficient than having everything under one roof, since countries and companies specialized in given fields of goods and services, making them more productive in their own right.

However, one could ask, why this did not happen before? The answer can be found in the work of another revered classical figure, Friedrich List. List argued in his seminal work, The System of

National Political Economy, that ‘perpetual peace’ is not the result but the condition of commerce

(List, 1841). Around the 1970s, which seems to emerge as a critical juncture in our story, not just the shift in the trend of relative firm size and the emergence of intra-industry trade happened, but institutions such as the EEC, the United Nations and the GATT gained prominence and substantial following. The first and foremost role of these institutions was not the promotion of

(10)

10

international trade, but to introduce List’s perpetual peace, which eventually makes trade possible. These institutional developments at that time lowered political uncertainty and the risk of not having everything under one roof, thus encouraged firms to take transactions to the market, which endogenously reinforced the specialization of countries.

The argument summarized is as follows. The decline in the relative size of the largest firms in developed countries can be attributed to the lowered political uncertainty, through institutionally strengthened peace, in Europe, which made possible expanding international, intra-industry trade and subsequently led to a larger extent of the market. This resulted in an enhanced Smithian division of labour among developed countries and highly internationalized and specialized value chains. These developments made it less risky and more efficient to execute transactions on markets, thus weakening the rationale behind the large vertically and horizontally integrated Chandlerian companies of the previous phases of globalization. Even though the argument builds on Langlois (2003, 2007) there are important differences. In the theory of the vanishing hand the hindering factor to specialization is the lack of development of market capabilities, while I argue that these capabilities were present but institutional and political factors constrained their utilization by firms.

1.4. Research design

To investigate the puzzle and whether the argument holds, a purposeful research design utilizing comparative case studies, descriptive and inferential statistics is proposed. The study aims to fill the gap described above, by investigating solely the largest firms1_{, by focusing on qualitative} factors, as well, and by analysing the question on the long-run.

The time span chosen is between 1907 and 2017. The starting point is determined by the availability of data, since there is no systematic information on the size of the largest firms in developed countries before 1907. The selected 110 years makes it possible to incorporate two upward (1907-1914 and 1945-2017) and one downward (1914-1945) trending periods of globalization, as seen on Figure 1. This can help us to see the whole picture and not just focus on one monolithic period, as it is done in the previous studies.

The selection of cases is based on two factors, on theory and data availability. In the initial puzzle presented in Figure 1, the five most developed large countries are shown, the US, UK, Germany,

1_{Due to the long-run nature of globalization the size largest firms are chosen to be explained. The trade-off} is the following. Firm size distribution is better, since it incorporates the whole spectrum of firms in an economy, but data on it is scarce, and on the long run it is not existent. Thus, largest firm serve as an alternative take on the issue, since precise accounts are available going back to 1907. So, for the sake of the long-run perspective, largest firms are examined. The long perspective is needed, because the unfolding of globalization is a very long-term process (See Figure 1).

(11)

11

Japan and France. This is the initial pile for the selection. To make the result of the thesis stronger, the varieties of capitalism approach (Hall and Soskice, 2001) is used as a theoretical guide. One liberal market economy (LME) and one coordinated market economy (CME) should be chosen for in-depth analysis. Among the listed countries the US and the UK can be characterised as LMEs, while Germany and Japan as CMEs, France falling in between. Data for the US is scarce before the Great Depression due to legislative reasons (Wardley, 1999). On the other hand, the data on the largest firms in the United Kingdom is well documented by business historians throughout the 20th_{century (Jeremy, 1991, Cassis 1997, Fiedler and Gospel 2008, 2010). Based on this the United} Kingdom is selected as a liberal market economy. Regarding, CMEs, Germany is much better documented than Japan, thanks to the work of Fiedler (1999), thus it is chosen as a case for the coordinated market economy. Also, both countries being members of the United Nations, the GATT/WTO and, for most of the period, the European Union, makes them more similar, thus more suitable for comparison.

The phenomenon this thesis aims to explain, is the trend in the size of the largest firms between 1907 and 2014/17, and if apparent the potential differences in the respective patterns of the United Kingdom and Germany. It is operationalized as the number of people employed by the top 100 firms divided by the active labour force of the year of observation. However, absolute data is presented, as well. Data for the largest firms is obtained for years 1907, 1938, 1957, 1973, 1995 (Fiedler, 1999, 2002) and 2014 (or more recent) for Germany and 1907, 1935, 1955, 1972, 1992 (Fiedler and Gospel, 2010, 2012, Jeremy, 1991) and 2017 for the United Kingdom. The most recent data, 2014 and 2017, is collected by the author.2_{Data on active labour force is retrieved} from Mitchell (1998), the Federal Statistics Office of Germany, the Bank of England and the Federal Reserve Bank of St. Louis.

Due to the complex nature of the question three sets of explanatory factors are proposed for analysis. These are technological and structural economic development, trade patterns and industrial policy.

Technological and structural economic development is one. Technological development

means two different things, productivity and the advances in transport and communication technologies. Productivity is measured by total factor productivity (Bergeaud et al, 2016). Transport costs and communication cost indices are taken from OECD (2007). General economic development is measured by the per capita GDP, while structural development is operationalized as the industrial composition of an economy, by the sector shares of GDP by added value (Booth, 2001, World Bank, 2018, Fouquin, M., Hugot, J., (2016)).

(12)

12

Trade policy and patterns are operationalized quantitatively in one way as the total

international trade as percentage of GDP per country. The data is retrieved from the Maddison, Penn World Tables and World Bank databases and from the study by Klasing and Milionis (2014) and Fouquin and Hugot (2016). The depth and stability of trade relations is operationalized by the number of preferential and regional trade agreements held by the given country (WTO, 2018). Intra-industry trade is the second facet. Building on the idea of Krugman (2008) it is operationalized by a rough measure, the percentage share of manufacture exports in merchandise export (World Bank, 2018, United Nations, 1962). The analysis of more granular, industry level trade data exceeds the limits of this thesis. The third perspective on globalization are tariff levels, which are retrieved from Bairoch (1993) and the World Bank database. Tariffs serve as a quantitative de jure proxy for trade policy, while trade patterns are seen as the consequence of this policy. Qualitatively trade policy of the respective countries is analysed. It is operationalized as the persisting attitude towards openness or protectionism and international institutions in the political agenda of governments through the time period. Also, the countries’ membership in international organizations, institutions and agreements, which could contribute to the Listian perpetual peace, are reviewed. These are analysed through accounts of economic and general history. Generally, Bairoch (1993) and for separately Germany, Plumpe (2016), Weiss et al (1988) and Borchardt (1991), and for the UK Booth (2001), Johnson et al (2008) and various news source are used.

Industrial policy is important since the government can favour large, ‘national champions’ as in

the post-war France or the clusters of small, networked companies as in the Silicon Valley in the US. These policies directly influence firm size, so it is essential to take them into consideration. Under industrial policy answers are sought to the following questions: what was the government’s view on firm size and how that changed over time? How anti-trust regulation developed? Here the aim is to build on the work of Cassis (1997, 2000, 2008), Fiedler and Gospel (2008, 2010), Fiedler (1999), Wardley (1991, 1999), and the respective economic and business history of Germany and the United Kingdom (sources cited above).

A comparative analysis of descriptive statistics will be executed in section 3 to establish the general plausibility of the argument in the light of long-term, overarching quantitative evidence. Findings on the size of the largest firms and all of the above-mentioned set of quantitative explanatory factors are presented here, as well.

The qualitative part is divided to a macro and a micro level analysis. In the macro level investigation, each country’s trade and industrial policies will be analyzed in each time period (Pre-1907, 1907-1935/38, 1935/8-1955/57, 1955/57-1972/73, 1972/73-1992/95, 1992/95-2014/2017) to shed light on underlying economic and political processes behind the numbers.

(13)

13

Here the inquiry will focus on the mechanisms of the argument outlined above and the propositions formulated section 2.2. In the next part, the micro level answers to the macro level processes are going to be investigated in detail, through the historical analysis of two firms in each case. From UK firms GKN, a manufacturer, and Prudential, an insurer, and from Germany Siemens, an electrical engineering giant, and Deutsche Bank, a financial services firm are presented. The history of these firms is analysed with particular attention to their relation to industrial policy, international politics and globalization (restructuring efforts to be competitive). The qualitative part is followed by a brief analysis of inferential statistics to obtain more rigorous evidence on the hypothesized relationship between firm size, politics and the extent of the market. Three models are presented, from which two is a low-N time-series (1955/57-2014/2017) analysis of the data of the largest firms in the two examined countries. The third model is a cross-sectional analysis of firm size data, from 31 developed countries, from the year 2014.

The ultimate goal of the research design is to establish the theoretical framework outlined in the next section with a three-level analysis. First, generally with descriptive and inferential quantitative evidence, second untangling the numbers through a macro-level qualitative analysis and finally find firm-level evidence of the main processes. Overall, a combination of the above described explanatory factors is laid out, which eventually answer the overarching research question of the thesis.

1.5. Structure

In section 2 a detailed literature review follows, which situates the thesis in the broader academic context and also sets out a more nuanced theoretical framework as five propositions are formulated. Section 3 elaborates on the descriptive statistical aspects of the research, namely data and empirics. Considerations on measurement of quantitative variables and their operationalization is laid out. Also, the methods and principles based on which the author collected the most recent data on the largest firms in Germany and the UK is elaborated. The case for comparability of the existing data and the newly collected data by the author is established. Later in the section the findings regarding the quantitative explanatory factors are presented in the form of descriptive statistics, mostly visualized in graphs and tables, and the individual and comparative analysis of the findings is conducted in detail. In Section 4 and 5 the qualitative sources are analysed regarding Germany and the UK, respectively. It is done in a chronological, systematic manner. Analysis of firm level evidence follows the macro level inquiry in each respective chapter. Section 6 consists the brief analysis of inferential statistics. Section 7 pulls the

(14)

14

strings together and concludes what can be learnt from the analysis outlined in this thesis. This part also elaborates on possible implications for policy and future research.

(15)

15

2. Literature review and theoretical framework 2.1. Existing literature

The most renowned scholar of long-term dynamics of big business is Alfred Chandler. He wrote three seminal pieces (1962, 1977, 1990) on the topic. In Strategy and Structure (1962) there are two main arguments he formulates. First that structure follows strategy. Second, he asserts that the technological and market developments, near the end of the 19th_{century, created a condition,} which resulted in the large, vertically integrated, multi-divisional firm and managerial capitalism. The technological and market developments are the advances in transport and communication technology and growing extent of the market. Visible Hand (1977) is another, more refined but similar, testament on the side of managerial capitalism. Chandler builds on data on a long time span and argues that Adam Smith’s invisible hand is transforming into the visible hand of the managerial class. In the above-mentioned books Chandler draws solely on evidence from large, US enterprises. In Scale and Scope (1990) he elaborates a comparative design, utilizing data from Germany and UK, as well. His main argument, regarding the development of industrial capitalism, is still that the managerial hierarchy made firms more efficient. Plus, that the emergence of this hierarchy is based on the mentioned technological and market developments of the late 19th_and early 20th_{century. As, Figure 1, and other data to be presented in this thesis shows, large,} vertically integrated companies are on the decline, but communications and transport technology is developing, and markets are expanding on a breakneck speed. Based on this, Chandler can be only partially right, and some important determinants must be missing from his, nevertheless, stellar analysis.

Gospel and Fiedler’s 2008 study serves as the basis of this thesis’s’ puzzle. They examine a list of global top 100 firms between 1907 and 2002. Global in this case means that the largest companies from Germany, Japan, US, UK and France, which make a combined list of the top 100 firms, by employee number3_{. By this they trace the largest firms in the most developed economies} throughout increases and decreases in globalization in the 20th_{century. Thus, we can say that} these findings can bear more potential for generalization than Chandler’s, due to the wider comparative perspective. Gospel and Fiedler lay out three sets of explanatory factors which could determine firm size: technology and markets, firm strategy and state influence. Among the other factors they put the biggest emphasis on the development of new general-purpose technologies, which, in their explanation, made possible the management of larger organizations. They also label the last period (1972-2002) as the “re-making of the corporate economy” (Gospel and Fiedler, 2008: 17), since the industrial composition changed significantly, switching from

(16)

16

manufacturing to services. Overall, they conclude that technological and market developments had the largest effect on firm size over the course of the 20th_{century, while they dismiss the state’s} role and consider corporate strategy as a residual. However, they give no consideration to factors of international politics.

Cassis’s (2000) is a comparative study of the American and European big business. He outlines several important differences between them. First that US companies were much larger in terms of assets, second that family ownership was very accentuated in Europe compared to the US and third, that full or partial state ownership was more present on the Old Continent. He also concludes that the home of big business in Europe was Germany and the United Kingdom, which corresponds to the cases selected in this thesis. By examining the different countries, he proposes that relationship between industry and finance4_{had little effect on firm size. He also argues that} the European experience with big business has highly Chandlerian characteristics, but also acknowledges that now we are living in a post-Chandlerian era, when deeper specialization is the trend. His overall conclusion is that “national differences had little impact on business performance,

but they did affect business practices” (Cassis, 2000: 410). Cassis’s important book, Big Business: The European Experience (1997), is more preoccupied with the task to identify big business in

Germany, UK and France, he elaborates little on the potential causes of its emergence. However, he identifies several factors, which according to him explain the birth of large firms in the late 19th century. These are technological and market developments, improved communications, improved investment finance, government regulation and the motivation of entrepreneurs. Unfortunately, he does not elaborate on these factors’ relative importance.

Traú (2003) examines a cross-country (France, Germany, Italy, US, UK, Japan) data of the largest manufacturing firms between 1960 and 2000. He argues that the most prominent development in industrial economies, in the last three decades of his time-span, is the emergence of small firms, which on a longer-run shows a V-shaped curve (opposite of the firm size curve on Figure 15_{). Traú} identifies several causes for this pattern in the existing literature: emerging need for product differentiation brought by rising income levels, emergence of flexible production technologies, outbreak of job actions and the oil shocks of the 1970s. However, he argues that two important determinants are omitted, the strengthening of competition brought by the increased importance of external demand (essentially international trade) and rising uncertainty (meaning enhanced financial liberalization and space for speculation) (Traú, 2003: 3). He summarizes the main line of his reasoning the following way: “the need to manage a growing amount of information and

4_{In the UK capital markets had a more short-term orientation, while in Germany the universal banks} focused on the long-term.

(17)

17

simultaneously to become more efficient leads the firm, given the amount of managerial resources it has at its disposal, to an overall reduction in its size.” (Traú, 2003: 4). Traú also asserts that this

trend came to its end with the 20th_{century, which according to the data presented in this thesis,} is not the case, since the reduction in size continues today. Plus, it is inarguable that financialization might lead to higher uncertainty in a given sense, but the overall uncertainty, especially political uncertainty, is substantially lowered in recent times. Therefore, the foundation of Traú’s argument is debatable.

Langlois (2003, 2007) notices the inverted U-shaped curve of the size of the largest firms throughout the 20th_{century. To explain it, he puts forward the vanishing hand theory. I build} greatly on Langlois’ argument in this thesis. The main line of his reasoning is that as income, the extent of the market and population grows, technology advances, the “the Smithian process of the

division of labour always trends to lead to finer specialization of function and increased coordination through markets” (Langlois, 2003: 2). However, he argues that in the light of this the

Chandlerian era was rather the distinction than the rule, since capabilities for further growth were too expensive on the market thus needed to be developed and kept in-house. This being the constraint for specialization and the incentive for vertical integration. But it is not entirely clear what is the evidence for the high price/non-existence of these capabilities. Additionally, Langlois argues, that a firm is an information processing unit, which’s main goal is to buffer uncertainty6_. Later he adds, that the conditions determining the degree of uncertainty in the firm’s environment are technology, organization and institutions. Under institutions Langlois means business institutions like the financial system, the limited liability company, industrial standards and property rights. His theory explains a great deal of variation in firm size, however there is a major neglect of international politics and its institutions. The extent of the market is largely determined by bi- and multilateral trade agreements and uncertainty, which needs to be buffered, can be substantially lowered thanks to institutions such as the European Union or the United Nations. Also, the level of development of these institutions determines the price of capabilities needed for growth on the market (which were too expensive or non-existent in Langlois’ view in the Chandlerian era). So, this addition seems necessary to Langlois’ line of reasoning. Plus, his neglect of empirical evidence also cries for deeper analysis, to uncover whether this theoretical construct can be taken seriously.

It can be seen, that there is no consensus in explaining the long-run dynamics of firms. Chandler’s (1962, 1977, 1990) explanations are partial and builds solely on advances in technology and organization. Gospel and Fiedler (2008) is no exception. Cassis (1997, 2000) does not put a great

6_{He takes an evolutionary view on the problem, and not approaches it from the viewpoint of transaction} cost economics.

(18)

18

emphasis on the emergence and dynamics of big business, rather he defines and identifies it in the European context. Traú (2003) and Langlois (2003, 2007) give a more refined set of explanatory factors, however their arguments are contradictory. Traú asserts that uncertainty increased, and firms are unable to handle it, thus they rather become smaller, while Langlois argues that there is less uncertainty to be “buffered” thus firms do not need to be large and vertically integrated.

According to the existing literature on the long-run dynamics of large firms, we see that advances in technology, growing extent of the market and the level of uncertainty in the environment are the main determinants of firm size. The other trend, which is obvious from the literature is the increasing specialization among firms, which consequently leads to the reduction in their size. Thus, a rational next step in theory building is to connect increasing specialization to the three explanatory factors, technology, extent of the market and uncertainty. This is going to be done in the next part of this section. Also, it is important to note that all authors examined these dynamics in developed economies, so the following theory deals with them as well. 7

2.2. Theoretical framework

The main mechanism behind the long-run dynamics of firms is the inverse relationship between specialization and concentration. In the recent decades, in developed countries, it worked the way that specialization increased, while the size of the largest firms (concentration) decreased. The existing literature supports this view. Baldwin et al (2001) empirically investigates it in the case of Canadian manufacturing firms and comes to the conclusion, that indeed specialization leads to smaller firms. Jacobides (2005) lays out a theoretical model built on standardization of information and simplified coordination to explain the occurrence of vertical disintegration. Rossi-Hansberg (2005) argues that continuously lowering transport costs lead to higher specialization, which eventually leads to lower concentration. This theoretical proposition is tested in Aiginger and Rossi-Hansberg (2006) and the authors conclude that the relationship holds empirically. Additionally, Liu and Yang (2000) formulate the theory of the irrelevance of the size of the firm, which states that inter-firm specialization leads to decreasing firm size. This theory is empirically tested and supported by Lam and Liu (2004) using data from Hong-Kong on its 11 main industries, between 1982 and 1999.

As this relationship is established, the question arises, what determines specialization? Adam Smith (2012 [1776]) argues that “the division of labour is limited by the extent of the market”. This assertion was later theoretically supported by Stigler (1951). Under this vague expression, extent of the market, we understand not just the mere size, but the depth and the stability of market

(19)

19

relations, as well.8_{Acknowledging this we need to answer the question: what determines the} extent of the market? Stigler (1951) argues that transport and communications costs are one aspect. However, even in the extreme case that these costs are zero, the extent of the market could be limited by politics. If there is autarchy the market can maximally be the domestic one. In a case closer to reality, if a country is protectionist and does not follow an open trade policy the size of the market will be limited by an arbitrary, political decision. Thus, we can conclude that there is an indirect connection between trade policy and specialization, the transmission mechanism being the extent of the market. This theoretical proposition is widely supported by empirical studies. The cited Baldwin et al (2001) article finds an emerging trend of specialization among Canadian manufacturing firms after the free trade agreement between Canada and the US was ratified. Kapri (2016) found that the reduction of tariffs in China against Korean goods resulted in productivity improvement among Korean exporting firms. Sun and Zhang (2012) find empirically, on a dataset of Chinese manufacturing firms between 1998 and 2007 that exporting firms are more productive. Lee (2014) finds a similar pattern by examining 753 Malaysian firms in 2002 and 2006, namely that exporters have a higher productivity, than non-exporters, but he adds that it becomes less important among larger firms.9_{Bresnahan et al (2016), building on} evidence from developing Africa, argues that trade liberalization indeed leads to productivity growth.

According to the above outlined theoretical argument and cited empirical literature, trade policy and tariff levels are regarded as an important determinant of the extent of the market, therefore for specialization and firm size, as well. Two interrelated propositions are formulated regarding this explanatory factor:

1. Liberal trade policy in a developed economy extends the market, thus deeper specialization is possible, consequently the relative size of largest firms tends to be smaller in the economy. 2. Protectionist trade policy in a developed economy decreases the extent of the market, thus lowers the possible degree of specialization, consequently relative size of largest firms tends to be bigger in the economy.

While trade policy is the tool, which extends or curtails the market, trade patterns are the consequence. If the market increases, thanks to an open trade policy, larger specialization follows, however only in the case if there is a possibility to buy goods from other specialized firms. So, if

8_{Note that it is important to add these two dimensions, since depth and stability make possible long-lasting} supplier relationships, which eventually result in more specialization. If solely the size of the market increases then firm size should in fact grow, however if stability and depth is introduced specialization should deepen and decrease firm size.

9_{Here productivity is used as a proxy for specialization. Weinhold and Rauch (1997) in a cross-country} study of 39 economies that specialization is positively correlated with productivity.

(20)

20

specialization follows trade liberalization, the pattern of intra-industry trade emerges between similar countries (Balassa et al, 1987). This line of reasoning was first empirically brought forward by Bela Balassa (1966), while investigating the effects of the Common Market on the trade patterns between industrialized countries. The field culminated in the work of Krugman (1980, 1995, 2008) with establishment of new trade theory. Feenstra (1998) concluded that while there is an integration in world trade, there is a disintegration in production, which is due to the increasing sliced up nature of global value chains. Also, it is underpinned by Wolff (1999), who examined trade patterns of 14 OECD countries between 1970 and 1993. He found that the economies in question, with growing trade as share of GDP, instead of convergence, retained their respective specializations. Additionally, Hummels et al (2001) find that vertical specialization account for more than 30% growth in between 1970 and 1990 in the OECD countries. Based on the described theoretical and empirical background the following proposition is formulated:

3. The importance of intra-industry trade in a country’s trade pattern is positively related to the decrease of relative size of the largest firms in a developed economy.

However, trade patterns and policy are not independent of international politics, since as Friedrich List (1841) argues ‘perpetual peace’ is the condition and not the result of commerce. In connection with specialization, it can be argued that institutions and agreements ensuring peace on the international level, lower the risk of conflict, thus lower the long-term uncertainty the firm has to deal with. So, for firms, if there is lower uncertainty it makes it sense to buy from other more specialized companies, which may be located abroad, than to make it in a less specialized way themselves. This increases competitiveness and may reduce costs, as well. Thus, membership in international institutions and agreements indirectly affects size of the largest firms in the economy through decreasing political uncertainty. The proposition regarding this is:

4. Membership in international institutions, organizations and agreements10 is positively correlated to the decrease of relative size of the largest firms in a developed economy.

Another important factor influencing the extent of the market as elaborated by Stigler (1951) and most of the literature on the long-run dynamics of large firms11_{is the lowering of transport and} communication costs, or phrased differently, advances in transport and communication technology. Rossi-Hansberg et al (2005, 2006) argues that lowering transport costs leads to higher specialization. On the other hand, Chandler (1962, 1977, 1990) argues that the development in these technologies resulted in the large, vertically integrated, Chandlerian firms.

10_{Institutions such as the European Union, United Nations, World Trade Organization, International} Monetary Fund, multilateral, regional and preferential trade agreements.

(21)

21

This line of reasoning is not contested by Fiedler and Gospel (2008) and neither by Cassis (1997, 2000). However, while the development in these technologies have been continuous in the last century, the size pattern of the largest firms shows a different shape. Since there seem to be no direct correlation between the two, the question arises, whether this is indeed an important determinant of firm size? In this thesis, relatively advanced communication and transport technology, thus lower costs, are seen as a necessary, but not sufficient condition and not as a determinant. In other words, if the development of these technologies reach a given threshold, which allows firms larger than the pre-Chandlerian ones12_{, then not transport and} communication technology will determine firm size. According to the historical pattern, if the threshold is met the organization of markets can either trend towards larger or smaller firms. The next factor, which is independent of the abovementioned ones, is industrial policy. Different states tend to favour different firm sizes. The national champion industrial policy (Falck et al, 2011) was prevalent in given European countries, like France, in the last century (Hancke, 2002). This can be understood as a form of favouring larger, national firms against small- and middle-sized companies. This policy not necessarily affects specialization, but directly the size of the largest firms, which are helped to grow through government subsidies and policies. The other take on this matter is the favouring of smaller companies or clusters of them. A stellar example of this is the decades of government subsidies to Silicon Valley in the United States (Mazzucato, 2011). As these policies directly affect size, the theoretical proposition is the following:

5. A developed country with an industrial policy favouring national champions is positively correlated with the relative size of the largest firms in the economy.

2.3. Summary

In summary, the driver of the theoretical framework is the mechanism that specialization leads to a decrease in the size of the largest firms in developed economies. This builds on the Smithian assertion that the “the division of labour is limited by the extent of the market”, trade policy and membership in international organizations, institutions and agreements are taken into account as factors influencing the extent of the market. Industrial policy is considered independently, since it is in itself can affect the size of firms in an economy. In accordance with previous studies

12_{Langlois (2007) argues that markets in the antebellum period were fragmented, connected through} generalist, middle-man merchants. In this period firm size was generally very low compared to even the late 19th century. The first major boom in transportation and communication technologies (railroad and telegraph), thus, had a substantial effect on the organization of markets and firms, as well. This first boom was the one, where the advancement in technologies and costs reached the threshold condition to allow Chandlerian firms.

(22)

22

advances in communication and transport technology are incorporated, as well, but only as a necessary condition of the presence of large firms and not as a direct determinant of firm size. In the following section a quantitative analysis is laid out to strengthen the overall plausibility of the framework.

(23)

23

3. Descriptive statistics

3.1. Methodological considerations 3.1.1. Size of the largest firms

The most important empirical basis of this thesis is the data on the 100 largest firms in Germany and the United Kingdom. The backbone of this dataset is the work done by Gospel and Fiedler (2008, 2010) who collected the data for the years 1907, 1938, 1957, 1973 and 1995 for Germany and 1907, 1935, 1955, 1972 and 1992 for the UK. The dataset is complemented with recent findings for the year 2014 for Germany and 2017 for the UK. Gospel and Fiedler mostly focused on domestic employment in the largest firms. However, they did collect data regarding the global workforce, but these findings are partial in 1992/95, 1972/73 and extremely scarce in the earlier years13_{. Overall, the methodology for the most recent entries corresponds to that of Gospel and} Fiedler (2008, 2010).

To obtain the 2014 figures for Germany, Die Welt’s database of top 500 firms was used. Unfortunately, this was ranked by revenue, however also consisted of the number of global employees. First, the data of the top 200 firms was obtained, then ranked based on global employees. Then, the number of domestic employees of each company was researched in annual reports, the business press and company websites. After this step the companies were ranked once again, now based on the domestic employees, to correspond to the existing data. As the last step, these findings were cross-checked with that of Gospel and Fiedler from the year 1995, whether there some serious omissions.

To obtain the 2017 figures for the UK, the FTSE100 was used for the 100 largest publicly listed companies and The Sunday Times Top Track 100 2017 league table for the 100 largest private companies. The problem was similar to the German case, only the number of the global employees was in these lists. So, first these two hundred companies were ranked by their global employees then searched for the domestic numbers in annual reports, the business press and company websites. Once, the domestic data were available the firms were ranked again. As the last step, these findings were cross-checked with that of Gospel and Fiedler from the year 1992, whether there are some serious omissions.

These data are then aggregated in each case to get the absolute number of domestic employees in the 100 largest firms in each year. To control for population growth a relative figure is calculated, as well by dividing the absolute domestic employment with the active labour force (Mitchell, 1998, St. Louis Fed, 2018, Destatis, 2018).

13_{For Germany there are 3 entries of global employee number for 1907, 1938 and 1 for 1957. For the UK} there are 3 entries for 1907, 7 for 1935 and 20 for 1955.

(24)

24

The reason that size is measured in the number of domestic employees is straightforward, in case of a monetary measure (revenue, assets, value) various valuation exercises would have been needed, which would raise the possibility of measurement errors. Additionally, the existing data on the size of the largest firms were in this format, for the same reason described above, as well. Plus, for the long-term global employment data is scarce. While the shortcomings are clear, like the overweighting of companies, which have higher workforce needs it is still the best available measure for long-term analysis.

3.1.2. Globalization

In this thesis globalization is broadly measured as international trade. The reason for not using any of the available globalization indices, like KOF (Gygli et al, 2018), CSGR (Lockwood et al, 2005), MGI (Figge et al, 2014) is their short time-span. KOF covers the longest time period, which is between 1970 and 2018.

Based on van Bergelijk and Mensink (1997) and Samimi et al (2011) the best available figure to measure globalization are aggregated trade statistics. The advantage of trade against foreign investment on the long run is the fact that trade was always a base of taxation, thus it’s precise measurement was in interest of any government, so the figures available are more trustworthy (Bergelijk and Mensink, 1997). Plus, aggregation filters out the potential misclassification of bilateral data. The all-around representation of trade data is a goal in this study, thus were it is possible, imports and exports are presented, as well.

However, this is a de facto measure, since the main proposition of this thesis is that politics influences firm size, a de jure one is needed, as well. Building on Samimi et al (2011) and the availability of long-term data, the weighted average tariff rate on manufactured products is chosen. To asses the depth and the stability of market relations the number of preferential and regional trade agreements held by the given country is included, as well.

(25)

25

3.2. Top 100 firms in Germany and the UK

In this part, the empirical findings are presented concerning the size of the largest firms in the German and the British economy between 1907 and 2017. Based on Figure 2, the fact that the size of largest firms is on the decline is supported empirically. What is more striking is that this trend is apparent not just relative to the active labour force, but in absolute terms, as well. The patterns over the course of the 20th and the early 21st century are similar in both countries and aspects, relative and absolute. There is a rise between 1907 and 1972/73 and a subsequent decline after that, till 2014/2017. An interesting difference is the higher relative firm size of the UK throughout the whole examined period and the higher absolute values in 1955/57 and 1972/73.

Based on Figure 2, there are two questions to which answers are sought in the following sections. First, what are the determinants of the relative size of the largest firms throughout the examined period? Second, what accounts for cross-country differences in the size of the largest firms, as the UK figures are consistently higher than that of Germany?

However, since Figure 2 includes only the domestic employees of the 100 largest firms the question naturally arises, what if the missing workers are relocated overseas and firm size have not declined? 0% 5% 10% 15% 20% 25% 30% 0 1,000,000 2,000,000 3,000,000 4,000,000 5,000,000 6,000,000 7,000,000

1907 1938 and 1935 1957 and 1955 1973 and 1972 1995 and 1992 2014 and 2017

A bso lut e si ze

Germany and the United Kingdom, domestic employment in the top 100 firms over the 20th century, absolute and relative to active labour force (1907-2017, Gospel and

Fiedler, 2008, 2010 and own research)

GER absolute employees UK absolute employees

GER relative employees UK relative employees

(26)

26

Figure 3: UK - Global and domestic workforce Figure 4: Germany - Global and domestic workforce

Figure 3 and 4 14_{gives us an answer, in case of the UK there is an overall decline, while German} firms indeed reorganized their production through foreign direct investment. There can be two potential reasons for this difference. It is obvious that in the German economy industry is more important (Figure 6, section 3.3.2., 30.5% versus 19.2% in 2016) than in the British. First, it can be argued that the low-value added, but critical production processes in industrial firms are relocated to countries with more favourable cost conditions, but due to their criticality they are held inside the boundaries of the firm. While in services there are lack of such processes, which could be relocated, since the majority of personnel is working in customer service. Thus, the largest UK companies cannot leverage this to extent of their German counterparts, for the simple reason that there are fewer industrial firms in the top 100. Second, another factor contributing, can be the superior German productivity (Figure 9, section 3.3.3.). Its growth was steeper after the second World War than in the UK, thus it makes sense, that German firms enter markets (not with the purpose to relocate production, but primarily to compete with the local firms) in countries with lower productivity, bring their technology and processes and operate there. For the first Volkswagen, Daimler, Siemens and Bosch, while for the second Lidl, Aldi and Deutsche Telekom are the most prominent examples in the German context. However, it is important to notice, that the uptick in German global employment starts around the same time, 1970s, which is labelled as a critical juncture regarding the matter of inquiry.

However, it is worth emphasizing that the data on global employees in rather scarce before 2014/17 and is not complete in any sense. Thus, these findings should be handled with caution and just aim to present general trends. A more thorough inquiry is needed into the overseas

14_{Here global employment is relative to the same active labour force figure in the given country as domestic} employment, in order to give it the same basis for presentation and easier comparability.

0% 5% 10% 15% 20% 25% 30% 1907 1938 1957 1973 1995 2014

Germany top 100 firms domestic and combined (global + domestic) employee

number relative to active labour force (1907-2014, Fiedler and Gospel, 2008,

2010 and own research)

Relative to active labour force (domestic workforce)

Relative to active labour force (global+domestic workforce) 0% 10% 20% 30% 1907 1935 1955 1972 1992 2017

United Kingdom top 100 firms domestic and combined (global + domestic) employee number relative to active labour force (1907-2017, Fiedler and Gospel, 2008, 2010 and own research)

Relative to active labour force (domestic workforce)

Relative to active labour force (global+domestic workforce)

(27)

27

employment numbers of the largest companies in the 20th_{century to give a solid ground for the} above conclusion. However, the scope of this thesis does not let to re-examine the past findings regarding global employment at the largest firms. Therefore, in the following I build on the domestic figures.

3.3. Technological change and structural development 3.3.1. Transport and communication costs

The set of explanatory factors most commonly cited in the existing literature as the main determinant of the size of the largest firms (Chandler, 1962, 1977, 1990, Gospel and Fiedler, 2008, 2010, Cassis, 1997, 2000) are technological and structural development. In this section I show that in the light of the new evidence, presented in section 3.2., this set of factors does not hold high explanatory value on the long-term.

Figure 5: Transport and communication costs

On Figure 5 the development of trade and communication costs between 1930 and 2005 (OECD, 2007) 15_{can be seen plotted together with the relative size of the largest firms in the UK and} Germany. It can be seen on Figure 5 that there is no serious correlation between the explanatory factors and the size of the largest firms. It corresponds to the theoretical framework laid out in this thesis, that certain level of development in transport and communication is needed for large firms to exist, thus they act as a necessary condition. After this condition is met the size is not primarily determined by transport and communication costs. Up until the 1970s they seem to be inversely correlated, but after the 1970s until today, there is a sudden shift in the trend of the size of the largest firms, while there’s no break in the trend of transport and communication costs.

15_{In case of transport and communication costs this is an index which concerns the world and where 1930} = 100. 0% 5% 10% 15% 20% 25% 30% 0 20 40 60 80 100 1930 1934 1938 1942 1946 1950 1954 1958 1962 1966 1970 1974 1978 1982 1986 1990 1994 1998 2002 2006 2010 2014 T ra nspo rt & co m m . c ost s

Transport and communication costs and the relative size of the largest firms in the UK

and Germany (1930-2017, OECD) UK - Nr. Of domestic

employees (top 100) relative to active labour force Germany - Nr. Of domestic employees (top 100) relative to active labour force International calling costs (relative to 1930) (%) Passenger air transport cost (relative to 1930) (%) Sea freight cost (relative to 1930) (%)

(28)

28

It is important to note that Chandler worked out his main thesis, which sees these developments as the determinants of the emergence of the large vertically integrated firm, in the 1960s and 1970s. Thus, based on Figure 5, his thesis is not wrong, however lost its relevance in the last 40 years, as the sole explanatory factor for the size of the largest firms. This reinforces the view that there are other major determinants in play on the long-run.

Moreover, there are studies which argue that recent advances in information technology, thus communication cost reductions, will eventually make firms smaller. Malone et al (1987) builds a theoretical argument that lowering coordination costs by IT will lead to more transactions executed through markets, rather than hierarchies. Hitt (1999) empirically examines this question and finds that the use of more information technology leads to decreased vertical integration.

To summarize, while transport and communication cost decreases, before the advent of information technology, resulted in larger firms and more vertical integration (Chandler, 1962, 1977, 1990, Cassis, 1997, 2000, Gospel and Fiedler, 2008, 20108), from the 1970s developments in IT might lead to smaller firms, lower degree of integration and more transactions on the market (Malone, 1987, Hitt, 1999). So, something else must be the general determinant of the size of the largest firms.

3.3.2. Structural change

Figure 6: UK and Germany - Structure of output and size of the largest firms (1900-2017)

Structural change, first from agriculture to industry then to services, is another widely used explanatory factor for the size of largest firms, Chandler (1962, 1977, 1990) and Cassis (1997,

0% 5% 10% 15% 20% 25% 30% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% Sha re o f GD P

Germany - Structure of output and the size of the largest firms (1900-2016, source: Booth, 2001, World

Bank) Domestic employees relative to active labour force Agriculture Industry Services 0% 5% 10% 15% 20% 25% 30% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% Sha re o f GD P

UK - Structure of output and the size of the largest firms (1900-2017, source:

(29)

29

2000) dub it as market development, while Gospel and Fiedler (2008, 2010) as the ‘re-making of the corporate economy’. On Figure 6 the sector shares of GDP are plotted together with the relative size of the largest firms. The patterns are similar in both countries; however, the sector share of services was higher throughout the time-period in the United Kingdom. Additionally, in Germany agriculture was much more accentuated between 1900 and 1973, going from 29% to 3%, while in the UK going from 6% to 3% of GDP. This corresponds to Booth (2001) who argues that the shift of agricultural workforce to higher productivity sectors, industry and services, happened during the early-to-mid 1800s in the UK, while in the late 1800s and early 1900s in Germany. It is uncontestable that there is a correlation between the decline of agriculture, the expansion of industry and the increase in the relative size of the largest firms. But, as seen on Figure 6, in case of the UK this growth in firm size continued until the 1970s, while the categorical shift in sector trends, towards services, occurred in the 1950s. In case of Germany, the picture is similar, but less accentuated. The growth of industry flattened from the 1930s onwards, while the services sector was intensely growing since the beginning of the century. Additionally, relative firm size was growing together with the share of services up till the 1970s, however, after that services continued to grow, while firm size went on to decline. Based on this piece of evidence structural shift might have played a role in the development of firm size, however it seems less determinative than argued by the existing literature.

3.3.3. Prosperity, productivity and capital intensity

Figure 7: Germany and UK - capital intensity and the size of the largest firms (1900-2016)

Long-term trends and cross-country differences might be also explained by capital intensity in a given country (Kumar et al, 1999). The size of the largest firms and capital intensity is plotted together on Figure 7. The long-run pattern is similar in both countries in both measures, thus the very trend in the shifts of the largest firms inside an economy might not be explained by capital intensity. Additionally, capital intensity is a questionable determinant of cross-country

0% 10% 20% 30% 0 50 100 150 200 250 1900 1908 1916 1924 1932 1940 1948 1956 1964 1972 1980 1988 1996 2004 2012 Ca pi ta l i nt en si ty in dex

Germany - capital intensity and the size of the largest firms (1900-2015, Bergeaud et

al, 2016)

Capital intensity

Domestic employees relative to active labour force

0% 5% 10% 15% 20% 25% 30% 0 50 100 150 200 250 1900 1908 1916 1924 1932 1940 1948 1956 1964 1972 1980 1988 1996 2004 2012 Ca pi ta l i nt en si ty in dex

UK - Capital intensity and the size of the largest firms (1900-2016, Bergeaud et al,

2016)

Capital intensity

Globalization and the size of the largest firms : a long-term view on the cases of Germany and the United Kingdom