Research in Economic History

Christopher Hanes, Susan Wolcott, Christopher Hanes, Susan Walcott

• Publisher: Emerald Publishing Limited
• Year: 2018
• ISBN:
  9781787565838
• Topic: Économie
• Subtopic: Histoire économique

Volume 34 contains articles on the economic history of Europe, North America and South America and brings new analysis, and newly created datasets to address issues of interest. Two of the papers present newly constructed datasets. In " Prices, Wages and the Cost of Living in Old Republic São Paulo: 1891-1930 ", Ball presents a newly constructed real wage index. São Paulo was the main destination for immigrants to Brazil in this period, but there has never before been sufficient data to analyse why. In " Multiple Core Regions: Regional Inequality in Switzerland, 1860 to 2008 ", Stohr uses the wealth of available Swiss data on agriculture and employment to create GDP measures for subregions in Switzerland. He uses these data to argue that aggregate inequality in Switzerland was low in the initial push to industrialization because there were multiple, similar centers industrializing simultaneously, thus mitigating inequality across regions. Two of the papers gather together existing data so that it can be analysed for the first time in a consistent manner. In " The forgotten half of finance: working-class saving in late nineteenth-century New Jersey ", Bodenhorn uses previously unexplored consumer surveys to characterize the savings behavior of the working class. And in " Heights across the last 2000 years in England ", Galofré-Vilà, Hinde, and Guntupalli gather all existing skeletal data for England for 2000 years to create a consistent longitudinal height series. They compare the series to height series of other regions as well as other measures of well being in England. And finally, in " Monetary Policy and the Copper Price Bust: A Reassessment of the Causes of the 1907 Panic", Rogers and Payne dig into the details of copper prices to discover the link between the Bank of England's contractionary monetary policy and changes in real asset prices. Their findings have important implications for understanding the mechanisms of monetary policy.

MULTIPLE CORE REGIONS: REGIONAL INEQUALITY IN SWITZERLAND, 1860–2008

Christian Stohr

ABSTRACT

This chapter does three things. First, it estimates regional gross domestic product (GDP) for three different geographical levels in Switzerland (97 micro regions, 16 labor market basins, and 3 large regions). Second, it analyzes the evolution of regional inequality relying on a heuristic model inspired by Williamson (1965), which features an initial growth impulse in one or several core regions and subsequent diffusion. Third, it uses index number theory to decompose regional inequality into three different effects: sectoral structure, productivity, and comparative advantage.

The results can be summarized as follows: As a consequence of the existence of multiple core regions, Swiss regional inequality has been comparatively low at higher geographical levels. Spatial diffusion of economic growth occurred across different parts of the country and within different labor market regions. This resulted in a bell-shaped evolution of regional inequality at the micro regional level and convergence at higher geographical levels. In early and in late stages of the development process, productivity differentials were the main drivers of inequality, whereas economic structure was determinant between 1888 and 1941. The poorest regions suffered from comparative disadvantage, that is, they were specialized in the vary sector (agriculture), where their relative productivity was comparatively lowest.

Keywords Regional data; inequality; industrial structure; productivity; comparative advantage; Switzerland

JEL classifications R10; R11; N93; N94; O14; O18

1. INTRODUCTION

Over the last decades, regional inequality has become a widely studied topic in economic history. The European Integration and the spatial turn in economic theory, which materialized in the new economic geography, the new trade theories, and the resurgence of urban economics spurred interest in regional disparities. The acknowledged importance of path dependency and historically grown spatial structures favored historical and long-term perspectives as well as close investigation of regional inequality during the nineteenth-century transport revolution.

Historians have, thus, estimated regional gross domestic product (GDP) and analyzed regional inequality and the spatial dynamics of production in many countries. Research has often focused on countries with large differentials, such as the USA, Italy, Spain, France, and the UK, but little attention has been paid to countries with low levels of inequality. This article provides estimates of regional GDP in Switzerland from 1860 to 2008, which was characterized by low regional disparities. In order to understand the drivers of regional inequality, we need to study not only cases where regional inequality is high but also cases where it is low.

Switzerland is also a complementary case because of its small size. Note that regional inequality is not related to the size of a country (Felsenstein, 2005). Portugal, Finland, and recently Belgium have seen comparatively high levels of regional inequality, while Sweden and Switzerland experienced rather low regional inequality. However, a few particularities of small countries, such as greater openness and the potentially more important incidence of commuting can have a distinct impact on regional inequality.

Another important factor for the Swiss case lies in the federalist political-administrative structure of the country, which distinguishes it clearly from France, the UK, Italy, and Spain. Fiscal decentralization is particularly strong in Switzerland since the collection of direct taxes is in the competency of the cantons and municipalities, while the federal state relies mainly on indirect taxes.

Moreover, the available data in Switzerland cover a very fine-grained geographical level, which makes it possible to study the evolution and the determinants of regional inequality at different geographical levels and allow for a better understanding of the phenomenon.

1.1. Estimating Regional GDP for Switzerland

Currently, a couple of datasets on cantonal national accounts are available for Switzerland. Fischer, Altermatt, and Resegatti (1980) have developed the first estimates of net cantonal income from 1950 onwards and the federal statistical office continued these estimates until 2005 (Bundesamt für Statistik, 1995). However, at the regional level, the difference between net national income (NNI) and GDP is substantial because of the large commuter flows between certain regions. Hence, these data are ill suited to analyze the spatial dynamics of production processes. Since 2008 the federal statistical office has, therefore, adopted an alternative approach to cantonal national accounting based on the GDP concept.

For the period 1890–1960, Ritzmann-Blickenstorfer and David (2012) carried out a regional decomposition of Swiss GDP based on national value added by industry and regional employment statistics. However, these estimates are problematic for two reasons. First, they rely on employment data that correspond to the national concept rather than the domestic concept. As a consequence, they are closer to gross national product (GNP) than to GDP. Second, the authors corrected their estimates with a so-called location factor (Standortfaktor), which was constructed from different proxy variables (population density, road and rail kilometers, size of agricultural exploitations, industrial employment rate, and service employment rate). Unfortunately, the authors do not provide any information on the methodology used to construct this location factor nor on the weight attributed to the different proxy variables. The inclusion of such a location factor in the estimation represents a problem if the data shall be used for spatial analyses. Such studies generally try to test and measure the effect of population density or market access on regional productivity. If these variables are already used in the data estimation, the analysis will not measure these effects but merely uncover the assumptions made for the data construction.

All the currently available regional datasets are limited to the cantonal level. However, cantons are of very unequal size, which makes them hardly comparable. The smallest canton fits more than 180 times into the largest one. Moreover, cantons are not functional areas but administrative regions. For this reason the federal statistical office has constructed alternative territorial subdivisions (e.g., micro regions and labor market basins), which are better suited for the spatial analysis of economic and social phenomena. But no national accounting data are available for these regions so far.

The present article takes advantage of very fine-grained data on agricultural inputs, employment, and commuting. Thereby it overcomes the shortcomings of the aforementioned datasets and manages to provide GDP estimates for three different territorial subdivisions that are well suited for spatial analyses (micro regions, labor market basins, and large regions).

Most of the international studies rely on a method proposed by Geary and Stark (2002). This method distributes national value added of each sector over different regions according to regional employment shares and uses regional wages as a proxy for productivity differentials (Badia-Miró, Guilera, & Lains, 2012; Buyst, 2009; Crafts, 2005; Enflo, 2014; Enflo, Henning, & Schön, 2010; Felice, 2011; Martínez-Galarraga, Rosés, & Tirado, 2013). Other studies have adopted a more eclectic method making use, where possible, of direct estimates of regional value added or output (Combes, Lafourcade, Thisse, & Toutain, 2011; Schulze, 2000, 2007). This chapter also applies an eclectic approach, which provides a solution to the problem of lacking regional wage data in Switzerland and, at the same time, allows me to take advantage of very detailed regional data available on agriculture (land use and livestock censuses) as well as manufacturing and services (fine-grained employment data and value-added estimates). The advantage of this method is that it rests on more detailed data than the Geary-Stark method and that instead of relying on a proxy it uses data that are directly related to production and productivity.

1.2. Explaining Regional Inequality in Switzerland

In a seminal paper, Williamson (1965) argued that in most countries regional inequality followed a bell-shaped evolution along the development trajectory. He provided evidence for this pattern from both longitudinal data and cross-section analysis. His finding has been confirmed in recent research on regional inequality in the USA (Kim, 1998), Britain (Crafts, 2005), Spain (Martínez-Galarraga et al., 2013), Italy (Felice, 2011), and Portugal (Badia-Miró et al., 2012). But other studies have found evidence for continuous convergence, for example, in Sweden (Enflo & Rosés, 2015), in France (Combes et al., 2011), and in Finland (Enflo, 2014).

Williamson’s explanation of the bell-shaped evolution of regional inequality rests on a model of diffusion, according to which an initial growth impulse first affects only one core region and then diffuses to other regions (Williamson, 1965). By focusing only on large geographical entities, such as the North and the South of the USA, Williamson’s diffusion model neglects different forms of diffusion and inequality. Hence, it cannot explain, why in some cases regional inequality caused a very strong and durable rise, while in others convergence started much faster. To distinguish different forms of diffusion and inequality I extend Williamson’s model to ...