Tuesday, 27 July 2021

Reconstructing income inequality in a colonial cash crop economy: five social tables for Uganda, 1925–1965

Michiel de Haas (Wageningen University)

Few doubt that colonialism generated new economic cleavages in African societies (Van de Walle 2009). At the same time, we know very little about the extent of such economic inequality in different African colonies and across time. In this article, I measure income inequality in Uganda, located in central-East Africa and colonized by Britain, in five benchmark years between 1925 (mid-colonial period) and 1965 (just after independence). I find that overall income inequality was low compared to other African colonies, but also find that sharp fault lines existed, especially along racial lines.

Based on what we know about Uganda’s pre-colonial and colonial history (Reid 2017), how much inequality should we expect? Uganda’s economy was predicated on smallholder production of export crops – cotton and coffee most importantly (De Haas 2017). In a land-abundant context, this economic structure fostered broad-based access to self-provisioning and cash income, which may have suppressed income inequality. On the other hand, we have reasons to hypothesize that colonial Uganda was unequal.  First, pre-colonial Uganda was far from egalitarian. For example there were sharp inequalities within the centralized Great Lakes Kingdoms, some of which came to make up a large portion of colonial Uganda, as well as between these Kingdoms and other parts of the territory. Second, the agricultural trading and processing sector in colonial Uganda was in the hands of South Asian expatriates, while the higher rungs of the colonial government were filled with Europeans. To what extent were these inequality-suppressing and -inducing factors visible in the country’s income distribution?

Attuning social tables to the context of colonial Africa

To reconstruct Uganda’s colonial inequality landscape, I use the ‘social table approach’, which simplifies the full income distribution of a given population into a smaller number of ‘social classes’. I distinguish eleven such classes for 15 districts each: three mutually exclusive income classes of African wage workers, five mutually exclusive income classes of self-employed earners, two classes of Europeans, and one class of Asians. Social tables have been employed by others in research on colonial Africa as well: Botswana (Bolt and Hillbom 2016), Ghana (Aboagye and Bolt, 2021), Kenya (Bigsten 1987) and Ivory Coast and Senegal (Alfani and Tadei 2019). This literature is synthesized in a recent paper by Hillbom, Bolt, De Haas and Tadei (2021).

My aim is not just to provide in-depth analysis of the Ugandan case, but also to make progress in finetuning the use of social tables in an African colonial context. Most importantly, in the latter respect, I show that it matters a lot what population we use to rank incomes – households or individuals – and how we think of income distribution within households. I demonstrate that households with higher incomes also tended to have more members. This means that such households come out as relatively ‘high income’ when we compare them on an aggregate level, but not when we break down the distribution to their constituent individual members. However, when we allow for intra-household inequality, a household head still benefit from adding more members to the household, as long as he (typically a polygamous man) is able to extract at least some income from additional household members. Thus, in an African, context, where people accumulate ‘wealth in people’ (Guyer and Belinga 1987) and households differ considerably in size, it is important to consider these dynamics explicitly when thinking about and measuring inequality. The results below are robust to using these different assumptions about the ranking population.

How unequal was colonial Uganda?

I answer this question along four lines: in comparison to other African colonies, and in terms of space (regional inequality), race (Africans versus expatriates) and class (income differentiation among African wage earners and the self-employed). I find that in comparative perspective, Uganda was fairly equal and did not see overall inequality rise during the period studied (Figure 1). I attribute this to Uganda’s land abundance and favourable agricultural conditions, which enabled the far majority of households to provision their own food. Moreover, Uganda’s main crop – cotton – was labour intensive and did not favour the rate of unequal accumulation that, for example, livestock did in colonial Botswana (Bolt and Hillbom 2016).

Figure 1: Gini coefficients for Uganda (1925-1965) under different assumptions

Using the Theil coefficient, we can zoom in on how different cleavages contributed to overall inequality (Figure 2). Here, I find that race dominated as the most important fault line in Uganda, especially before the 1950s. The 1930s were most extreme in this respect, as European households earned, on average, some 60 times as much as the average African household, and Asians close to 30 times as much. Regional differences, in the meanwhile were certainly present but contributed only moderately to overall inequality – especially when self-provisioning is taken into account. Inequality among African classes rose in the late-colonial period, as some farmers specialized in coffee and livestock, while the number and average incomes of salaried professionals rose as well.


Figure 2: decomposition of inequality in Uganda along lines of race, class and region.

Note - a similar text also appears on the African Economic History Network homepage


Aboagye, P. Y., & Bolt, J. (2021). Long-Term Trends in Income Inequality: Winners and Losers of Economic Change in Ghana, 1891-1960. Explorations in Economic History, forthcoming.

Alfani, G. & Tadei, F. (2019). Income Inequality in French West Africa: Building Social Tables for Pre-Independence Senegal and Ivory Coast. UB Economics Working Papers, 396.

Bigsten, A. (1987). Income distribution and growth in a dual economy: Kenya, 1914-1976. (Unpublished doctoral dissertation, Gothenburg University, Gothenburg)

Bolt, J., & Hillbom, E. (2016). Long‐term trends in economic inequality: lessons from colonial Botswana, 1921–74. The Economic History Review69(4), 1255-1284.

De Haas, M. (2017). Measuring rural welfare in colonial Africa: did Uganda's smallholders thrive? The Economic History Review70(2), 605-631.

De Haas, M. (2021). Reconstructing income inequality in a colonial cash crop economy: five social tables for  Uganda, 1925–1965. European Review of Economic History, forthcoming.

Guyer, J. I., & Belinga, S. M. E. (1995). Wealth in people as wealth in knowledge: Accumulation and composition in Equatorial Africa. Journal of African History, 36(1), 91-120.

Hillbom, E., Bolt, J., De Haas, M. & Tadei, F. (2021). Measuring historical inequality in Africa: What can we learn from social tables? African Economic History Network Working Paper, No. 63

Reid, R. J. (2017). A history of modern Uganda. Cambridge: Cambridge University Press.

Van de Walle, N. (2009). The institutional origins of inequality in Sub-Saharan Africa. Annual Review of Political Science12, 307-327.






L’Histoire Immobile? A Reappraisal of French Economic Growth using the Demand-Side Approach, 1280-1850.

Leonardo Ridolfi (University of Siena) and Alessandro Nuvolari (Sant’Anna School of Advanced Studies)

Historians interested in the study of economic performance of preindustrial societies owe a major intellectual debt to Angus Maddison. Even if by means of somewhat speculative methods, Maddison (2001) produced the first comprehensive set of estimates of GDP per capita for a large group of countries which resulted in an intriguing account of the dynamics of economic growth since the end of the Middle Ages (Maddison 2007).

This preliminary global quantitative assessment has prompted historians to later embark in more refined statistical reconstructions of the long-term evolution of output per capita and to develop new ways to integrate these different country-specific estimates over time and across space (Bolt and Van Zanden 2014, Bolt and Van Zanden 2020).

Notwithstanding the riches of source materials dealing with prices and wages, France has been left at the margins of the ongoing efforts of reconstruction of these “second generation” estimates of GDP per capita.

In this paper, we tackle this issue, by providing new estimates of output per capita for France over the period 1280–1850 using the demand-side approach.

The French case is interesting for two main reasons.

The first reason is that an important historiographical tradition, made popular by the French historian Emmanuel Le Roy Ladurie with a famous article first appeared in the Annales in 1974 (see picture below), has depicted France as a paradigmatic case of an inherently stagnant economy dominated by Malthusian checks and other institutional and cultural constraints. Hence, in this perspective, the French case may perhaps provide new materials for assessing the overall plausibility of the Malthusian model as a suitable empirical characterization of the long run evolution of living standards in Europe before the Industrial Revolution.

The second reason relates to the very origins of modern economic growth in pre-industrial Europe and, more specifically, to the analysis of the so-called “Little Divergence”, the process whereby the North Sea Area became the most prosperous and dynamic part of the continent.

Much discussion surrounds the timing and causes of why, between the Middle Ages and the modern period, the frontier of Europe’s economic development moved from the Mediterranean space and particularly Italy -which so far had been the world economic leader- to the countries of the North Sea region (Allen 2001, Broadberry 2020, De Pleijt and Van Zanden 2016). Despite that, there is still a general lack of consensus on the evolution of living standards and the main determinants of this process. In this context, a fresh assessment of French economic performance, a neighbouring country with multiple political, economic, and military interactions with the North Sea area could contribute to shed new light on the drivers of this initial phase of modern economic growth.




Our analysis highlights three main results (see Figure 6 of the article, reproduced below).

First, French output per capita fluctuated with no trend over the long term, with perhaps the exception of a modest “efflorescence” of economic growth during the seventeenth century. Until the seventeenth century, this pattern of growth chimes with Le Roy Ladurie’s (1974, p.688) notion of “motionless history” according to which ‘malgré immenses changements parmi les superstructures celui-ci tel en lui-même se retrouve finalement très proche la veille des famines de la Fronde et de celles de 1693 ou de 1709 de ce il était trois siècles et demi ou quatre siècles auparavant la veille de la famine de 1315 (despite the immense changes of the superstructures, the system finally finds itself very close, on the eve of the famines of the Fronde and those of 1693 or 1709, to what it was three and a half or four centuries before the eve of the famine of 1315).’

In this respect, Italy, the Netherlands and England appear as the main exceptions to this broad European pattern of relative stagnation.

The second result speaks more closely to the debate about the level of comparative development of England and France over the long run which a long historiographic tradition has seen as a useful vantage point for studying the origins of the Industrial Revolution (Crouzet 1985).

Our new series of GDP per capita challenges the notion of an early modern “Little Divergence” between England and France (Allen 2001) and points instead to a later divergence taking place by about the second half of the seventeenth century.

Finally, in our reappraisal of French economic performance, the French growth experience appears to be as an ‘intermediate case’ between the sustained growth pattern of the North Sea region countries and the stagnating or declining trend of the rest of Europe.



Figure 6. French economic performance in international comparative perspective, 1276–1850

Sources: France: our estimates; England: Broadberry et al. (2015); Holland: Van Zanden and Van Leeuwen (2012); Italy (Centre-North): Malanima (2011); Poland: Malinowski and Van Zanden (2017); Portugal: Palma and Reis (2019); Spain: Alvarez-Nogal and Prados de la Escosura (2013); Sweden: 1300–1560, Krantz (2017); 1560–1850, Schön and Krantz (2012).

Notes: the values in 1990 GK$ dollars are those reported by the authors except for Italy (Centre-North), Poland, Spain, and Sweden because the authors did not report annual series in 1990 GK$ dollars. In these cases, we made the conversions using the following benchmarks: Italy (Centre-North), 1,486$ in 1861 (Malanima 2011, p.218); Poland, 946$ in 1870 (2013 Maddison project); Spain, 1,079$ in 1850 (2013 Maddison project), and Sweden 1,076$ in 1850 (2013 Maddison project). All series are 11-year moving averages.



Allen, R. C. (2001). The great divergence in European wages and prices from the Middle Ages to the First World War. Explorations in Economic History 38, pp. 411447.

Álvarez-Nogal, C. and Prados De La Escosura, L. (2013). The rise and fall of Spain (1270–1850). Economic History Review 66, pp. 1–37.

Bolt, J. and Van Zanden, J. L. (2014). The Maddison Project: collaborative research on historical national accounts. Economic History Review 67, pp. 62751.

Bolt, J. and Van Zanden, J. L. (2020). Maddison style estimates of the evolution of the world economy. A new 2020 update. Maddion Project Working Paper n. 15.

Broadberry, S. (2020). The Industrial Revolution and the Great Divergence: Recent Findings from Historical National Accounting. CEPR Discussion Paper n. 15207.

Broadberry, S. Campbell, B. M. Klein, A. Overton, M. and Van Leeuwen, B. (2015). British economic growth, 1270–1870. Cambridge: Cambridge University Press.

Crouzet, F. (1985). De la supériorité de l’Angleterre sur la France: l'économique et l'imaginaire, XVIIe–XXe siècles. Paris: Libr. Académique Perrin.

De Pleijt, A. and van Zanden, J. L. (2016). Accounting for the “Little Divergence”: What drove economic growth in pre-industrial Europe? European Review of Economic History 20, pp. 387-409.

Krantz, O. (2017). Swedish GDP 1300–1560: A Tentative Estimate. Lund Papers in Economic History 152.

Le Roy Ladurie, E. (1974). L’histoire immobile. Annales. Histoire, Sciences Sociales 29, pp. 673–692.

Maddison, A. (2001), The World Economy: a Millennial Perspective. Paris: OECD.

Maddison, A. (2007), Contours of the World Economy, 1–2030 AD. Oxford: Oxford University Press.

Malanima, P. (2011). The long decline of a leading economy: GDP in central and northern Italy, 1300–1913. European Review of Economic History 15, pp. 169219.

Malinowski, M. and Van Zanden, J. L. (2017). Income and its distribution in preindustrial Poland. Cliometrica 11, pp. 375–404.

Palma, N. and Reis, J. (2019). From Convergence to Divergence: Portuguese Economic Growth, 1527–1850. Journal of Economic History 79, pp. 477506.

Schön, L. and Krantz, O. (2012). The Swedish economy in the early modern period: constructing historical national accounts. European Review of Economic History 16, pp. 529–49.

Van Zanden, J. L. and Van Leeuwen, B. (2012). Persistent but not consistent: The growth of national income in Holland 1347–1807. Explorations in economic history 49, pp. 11930.


Friday, 26 March 2021

The Empire Marketing Board and the Failure of ‘Soft’ Trade Policy, 1926-33

By David M. Higgins and Brian Varian

Can a country with a free-trade policy also pursue a preferential trade policy? In the 1920s, Britain tried. Britain's Empire Marketing Board (EMB) was created in 1926 to extend a non-tariff preference to imports from the empire by means a widespread publicity campaign. As we find, the EMB's campaign failed, and we offer several explanations why.

Between 1926 and 1933, the Empire Marketing Board (EMB) sought to influence the purchasing habits of British consumers.  This initiative was ambitious because British consumers had long been accustomed to purchasing produce from non-empire sources, for example, Argentine beef, Chinese tea, as well as Danish butter and Dutch cheese.  The campaign was also unprecedented: public money was expended on a marketing campaign intended to provide a ‘soft’ trade barrier to favour empire producers.

As an instrument of trade policy during the interwar period, the EMB appears unusual in the economic history literature which has been dominated by analyses of ‘hard’ policies involving tariffs.  The creation of the EMB was an attempt by the British government to reciprocate the formal tariff preferences that the Dominions extended to British exports.  The establishment of the EMB was necessary precisely because Britain maintained an essentially free trade policy until 1932.  Indeed, prior to that date, tariff reform, which sought to abolish Britain’s commitment to free trade, was responsible for the Conservative government losing its overall majority in the General Election of 1923.


Evaluating the effectiveness of the EMB

The fundamental objective of the EMB was to ensure that the British empire accounted for a growing proportion of Britain’s produce imports.  Consequently, the Board pursued three major policies: it financed research into the problems affecting empire food production; provided market intelligence to British trade organisations; and launched a major, national, publicity campaign involving ‘empire shopping weeks’ and the commissioning and distribution of iconic posters which appeared on billboards in major cities and towns.  Each policy was interrelated. For example, improvements in the quality and regularity of the supply of empire produce would help make the Board’s publicity more effective (Figure 1).


Figure 1.      Example of an EMB poster

In our study we assess the impact of the EMB’s publicity campaign on Britain’s empire imports.  Posters were a key component of the EMB’s total publicity, accounting, on average, for 35 per cent of the Board’s total publicity expenditure between 1926 and 1932.  We selected posters issued in 1927 that advertised a particular commodity and Dominion, for example, Australian frozen beef.  Our total sample includes advertised and non-advertised commodities originating from other parts of the empire and from non-empire countries.

The econometric results indicate that the EMB’s poster campaign did not affect the empire’s share of British imports for any of the advertised commodities in our sample: Australian frozen beef, Canadian grain, Ceylon tea, Indian rice, Mauritian sugar, New Zealand butter, and New Zealand cheese.  We advance three explanations for the failure of the EMB.


Explanations for failure

First, our re-examination of the EMB’s publicity expenditure indicates that it was small when compared with the advertising expenditures of the official Dominion organisations responsible for exports -- the Australian Dairy Produce Control Board, the New Zealand Dairy Produce Control Board, and the New Zealand Meat Producers Board.  On a standardised basis, the publicity expenditure of the EMB was considerably smaller than that of the Control Boards.  Moreover, the EMB’s expenditure was much less focused: not only did the EMB advertise a diverse range of empire produce, from Cyprus brandy to Malayan pineapples, but much of its advertising did not indicate an association between a Dominion and a commodity or, indeed, feature either a specific Dominion or commodity at all.

Second, it is remarkable that so much effort was devoted to persuading consumers to ‘buy Empire’, when many empire commodities were retailed without an indication of geographical origin.  Landmark legislation was introduced by the Merchandise Marks Act, 1926.  This Act stipulated that producers could make an application for a Marking Order to ensure that specific foodstuffs indicated country of origin.  However, this Act was never applied to cheese, grain, rice, or sugar.  Attempts were made to secure the marking of tea, but this was vociferously rejected by tea distributors and retailers, and famous British tea-blending companies for whom company brands were incomparable marketing assets.  A Marking Order for butter was secured in 1931, but even this Order left considerable leeway to grocers on the precise indications of origin they applied to pats of butter.

Beef (including frozen beef) was subject to a Marking Order from 1933, but by then there had been a pronounced change in the preferences of British consumers toward chilled beef, which was largely supplied by Argentina and the River Plate.  In fact, technological factors meant it was not possible for Australia to supply chilled beef to Britain during the interwar period. By the late 1920s, British consumers’ views of frozen beef were so disapproving that sales of this  commodity were restricted to asylums, other public institutions, and the British army!

The key conclusion to emerge from our study is that soft policies were not an effective substitute for tariffs and other types of quantitative trade restrictions. The introduction of hard barriers to trade had to await changes in the political and economic environment which occurred in the early 1930s.  


David M. Higgins

Brian Varian


Friday, 5 March 2021

YSI Economic History Graduate Webinar, Spring 2021


Dear all,

We are launching a third YSI Economic History Graduate Webinar this Spring. In previous editions we provided a platform for young researchers to present their ongoing work and get feedback from senior scholars. The online format made exchanges from people from different regions and research areas possible, offering early stage researchers an important venue in these times of disconnection. As social distancing remains a reality, so does connecting online to reach out to the community.

For these reasons we want to invite all young scholars working in Economic History to submit a paper for our 2021 Spring series. We welcome all sorts of contributions, regardless of time period or geographic area, as well as qualitative or quantitative approaches. You do not need to be registered with YSI to apply but we encourage all young scholars to join the community.

Send us a full paper version and an updated CV to: eh@youngscholarsinitiative.org

The deadline for papers is March 26th, and we’re planning on starting the sessions in April.

If you are interested in attending the webinar and receive the programme, please register using this form. The seminars will be held on Zoom and last 60 minutes on Tuesdays afternoon (Western Europe time). 

See you online!  

The YSI graduate seminar in Economic history is a joint collaboration between Ester Treccani, Jordi Caum Julio, Maylis Avaro and Xabier Garcia Fuente, with support from the Institute for New Economic Thinking and the European Historical Economics Society. 

Wednesday, 3 March 2021

Optimism or pessimism? A composite view on English living standards during the Industrial Revolution

By Daniel Gallardo Albarrán (Wageningen University, @DanielGalAlb) and Herman de Jong (Groningen University)

blog post based on the article, "Optimism or pessimism? A composite view on English living standards during the Industrial Revolution ", available on EHER here


The consequences of industrialization for the living standards of the mass of the population have been intensively debated ever since the days of William Blake, Karl Marx, and Charles Dickens. Over a period of roughly 100 years after ca. 1750, Great Britain set the basis for a dynamic and self-sustained process of economic development that eventually would improve the lives of millions of people. Although the positive outcomes of this process for human well-being since the 19th century are not disputed, the same does not apply to the years between 1750 and 1850. New methods of production and labor organization brought economic benefits, but they had a deep impact on citizens’ lives and the environment, as illustrated by the picture below.

On one side, a branch of the literature, represented by the so-called optimists, has argued that the benefits of improved methods of production trickled down in the form of substantial real wage increases after the Napoleonic wars (Clark, 2005; Lindert & Williamson, 1983). On the other side, the so-called pessimists have found that the increase in real wages was much less pronounced than what the optimists claim (Allen, 2009; Feinstein, 1998). Also, further supporting the pessimists’ case, health levels stagnated after the 1820s, annual working time reached new heights in the 1830s, and inequality remained at high levels (Allen, 2019; Broadberry, Campbell, Klein, Overton, & van Leeuwen, 2015; Voth, 2001; Wrigley, Davies, Oeppen, & Schofield, 1997).

The lack of consensus on the evolution of living standards during the classical years of the industrial revolution partially stems from the study of a large number of indicators individually. This can be problematic because these variables often exhibit opposite trends, thus having disparate implications for the analysis of well-being. One way to deal with this is by building a  composite index of welfare combining information on a number of key aspects of people’s lives into a single metric.

We take this approach and build a new indicator to study workers’ living standards during the early phase of industrialization that combines four dimensions of well-being: income, health, working time, and inequality. We draw on Jones and Klenow (2016) to construct a metric that aggregates the utility flows that an average British worker could expect from them those four aspects of living standards. By using utility theory, this article provides a novel and interesting perspective to the literature, which has mostly relied on other methodologies to construct similar indicators.


Our analysis of the evolution of workers’ well-being during the traditional period of early industrialization (i.e., 1760–1850) presents three main findings. First, unlike earlier composite indices or income per capita, our broad welfare series points to worsening living standards until 1800 (see column II in Table 2 of the article, reproduced below). This is the result of a steep rise in both working time and income inequality after 1760 that is not accounted for by other traditional indicators. Welfare growth rates could have been highly negative if life expectancy had not increased by 5 years between 1760 and 1800.


Second, we find that well-being improved after 1800 when real wages started rising and the negative effect of longer working time and higher inequality reached a plateau. Although well-being grew by almost 0.7 percentage points annually during these years (column III, last row), the resulting average level of welfare by the mid-nineteenth century does not support an optimistic interpretation of the evolution of workers’ living standards. According to our results, welfare was only 22 percent higher in 1850 than in 1760.

Our third main finding is that welfare exhibits much lower growth than our widely-used measures of overall living standards. If we consider GDP per capita, our calculations suggest that national income tends to overestimate welfare growth for the average citizen during the period by 20 percent. On the other hand, if we consider the well-known Human Development Index or the Dasgupta and Weale index, we find that they show a clear improving pattern between benchmarks, whereas our metric shows a much more pessimistic pattern, especially before 1800. 


The last decades of research into the consequences of the industrial revolution have brought a large amount of evidence on a number of economic, demographic, and social aspects of English workers’ lives in the 18th and 19th centuries. An important part of this new evidence is characterized by painting a more complex picture of what earlier generations of scholars initially brought forward and by adding new indicators that revealed opposite movements of well-being for sub-periods.

This article uses an encompassing framework of living standards to put together information about four key aspects of the lives of citizens at that time: material living standards, health, working time, and inequality.  We find that earlier studies drawing on composite indices of well-being are probably too optimistic about trends before 1800, since they do not fully take into account rising annual working time and increasing inequality. By 1850, our calculations show that welfare was 22 percent higher than in 1760 (20 percent less than the improvement in living standards suggested by GDP per capita). Therefore, welfare gains from health and material living standards slightly compensated for the negative effects of increasing levels of working time and inequality.

While encompassing, our indicator does not measure other important aspects of England’s welfare at the time, such as access to knowledge through education, environmental damage, or the social costs of the factory system. Their study in the future may reinforce our view that workers’ lives would not change substantially until the post-1850 period when the productivity benefits of the new forms of production trickled down to the working classes and public health regulation tackled the poor health conditions of the population.


Allen, R. C. (2009). Engels' pause: Technical change, capital accumulation, and inequality in the British industrial revolution. Explorations in Economic History, 46(4), 418-435.

Allen, R. C. (2019). Class structure and inequality during the industrial revolution: lessons from England's social tables, 1688-1867. The Economic History Review, 72(1), 88-125.

Broadberry, S., Campbell, B. M. S., Klein, A., Overton, M., & van Leeuwen, B. (2015). British Economic Growth, 1270-1870. Cambridge: Cambridge University Press.

Clark, G. (2005). The Condition of the Working Class in England, 1209-2004. Journal of Political Economy, 113(6), 1307-1340.

Feinstein, C. H. (1998). Pessimism Perpetuated: Real Wages and the Standard of Living. The Journal of Economic History, 58(3), 625-658.

Jones, C. I., & Klenow, P. J. (2016). Beyond GDP? Welfare across Countries and Time. American Economic Review, 106(9), 2426-2457.

Lindert, P. H., & Williamson, J. G. (1983). English Workers' Living Standards during the Industrial Revolution: A New Look. The Economic History Review, 36(1), 1-25.

Voth, H.-J. (2001). The longest years - new estimates of labor input in England, 1760-1830. The Journal of Economic History, 61(4), 1065-1082.

Wrigley, E. A., Davies, R. S., Oeppen, J. E., & Schofield, R. S. (1997). English Population History from Family Reconstitution 1580-1837. Cambridge: Cambridge University Press.


Monday, 1 March 2021

Changing Places: The Spatial Dispersion of U.S. Manufacturing during the 20th Century

Nicholas Crafts and Alexander Klein

We provide new estimates of changes in the spatial concentration of U.S. manufacturing from 1880 to 2007.  The average level across all industries fell by more than half over the period.  Creative destruction has had a strong spatial component which eroded the manufacturing belt and when compounded by globalization left a legacy of left-behind voters.  Even so, almost all industries can be described as significantly spatially concentrated at all times. 

See the full paper, now on early view at the European Review of Economic History, here

Everybody knows that the geography of industrial production changed dramatically during the 20th century both across and within countries.  There was clearly a strong spatial aspect to the forces of creative destruction.  The ‘left-behind’ victims of these geographic trends have become an important constituency in contemporary politics.

The long run move of manufacturing employment out of the manufacturing belt which is reported in Table 1.  Whereas 87.2 per cent of manufacturing employment in the U. S. economy was in the manufacturing belt in 1880 by 1940 this had fallen to 73.6 per cent and in 2007 to 42.9 per cent.  Within this, the East North Central (mid-west) region has a different chronology with a rising share from 1880 to 1947 and then a steady decline during subsequent decades.

Measuring Spatial Concentration

It is important to control for differences in the size distribution of plants when measuring spatial concentration and also to take account of the geographical position of regions through allowing for ‘neighbourhood effects’.  The spatially weighted version of the Ellison and Glaeser index has these desirable features and provides a better measure of spatial concentration than traditional indices such as Hoover’s localization coefficient but has not previously been used by economic historians.  In our new paper (Crafts and Klein, forthcoming) we present estimates of this index for manufacturing industries in the United States for selected years between 1880 and 2007.  We find as follows.

First, there was a big decline over the long run in the average spatial concentration index.  This occurred in two phases - gradual prior to 1940 and rapid after 1940.  The mean across all industries was 0.223 in 1880 which fell to 0.183 in 1940 and 0.096 in 1997 (Figure 1).  Greater spatial dispersion was characteristic of the vast majority of manufacturing industries by the second half of the 20th century.

Second, the measured decline in spatial concentration before 1940 confirms the views of economic geographers writing around this time who stressed the de-centralization of economic activity but has been overlooked in more recent literature.

Third, nevertheless almost all industries are spatially concentrated in the sense that the index score is always positive and significantly different from zero.  Indeed, the vast majority of scores throughout the period are above 0.05, the level which is conventionally described as ‘highly concentrated’ and indicative of the existence of significant local cost advantages.  This was still true at the end of the period in 2007 when the mean was 0.098.

Decline of the Manufacturing Belt

The context for long-run changes in the location of manufacturing has strong similarities with the stylized core-periphery model associated with Paul Krugman which places transport costs centre stage.  The model envisages a move from very high to intermediate to very low transport costs driving a move from dispersed to spatially concentrated then back to dispersed locations for manufacturing.  In the spatially concentrated (manufacturing belt) phase the core benefits from economies of scale and proximity to markets and suppliers which raises productivity but also tends to raise wages; subsequently, however, in the context of much lower transport costs, the wage gap becomes too high and moves to the periphery promote a convergence of wage rates.

The costs of moving manufactured goods declined by over 90 per cent in real terms between 1890 and 2000 from 18.5 cents per ton-mile to 2.3 cents (at 2001 prices).  In fact, much of this decrease occurred by 1967 when the cost was only 5.6 cents (at 2001 prices) while by 1891 the railroad revolution had already cut transport costs to about 10 per cent of the 1820s’ level and the manufacturing belt had been born.   We calculate that the ratio of the average wage in manufacturing in East North Central and Mid-Atlantic states relative to East and West South Central states rose from 1.22 in 1890 to 1.52 in 1940 before falling to 1.15 in 1987.

An excellent example of this process is Motor Vehicles and Equipment (SIC 371) where overall geographic concentration fell in the second half of the 20th century but where significant localization persisted in a new configuration.  The index for SIC 371 was 0.191 in 1940, 0.120 in 1958, 0.106 in 1977 and 0.094 in 1997.  This is reflected in maps 1 to 4 which show an evolving pattern of spatial concentration over time such that by 1997 the move away from the 1940 situation of a dominant position for Michigan and an east-west corridor in the southern Great Lakes region has been superseded by one in which Michigan is still a major centre but clusters within ‘Auto Alley’ extend as far south as Alabama.

Blue to Red 

Clearly, industrial geography changed greatly during the 20th century.  Spatial adjustment can be seen as an integral part of the creative destruction which was instrumental in promoting this change.  The relative decline of traditional manufacturing areas was driven by domestic cost factors and not simply attributable to globalization.

It might be argued that with a relatively flexible economy and mobile society the United States has coped with these pressures quite well.  Even so, some workers have been left behind and addressing their grievances has become a big political issue.  While cities like Boston have regenerated with knowledge-intensive business services and a highly educated workforce others such as Detroit have not been so well-placed.  Of the six states which were red in 2016 but had been blue in 2012, four (Michigan, Pennsylvania, Ohio and Wisconsin) were in the manufacturing belt – their electoral college votes won it for Trump.



Crafts, N. and Klein, A., “Spatial Concentration of Manufacturing Industries in the United States: Re-Examination of Long-Run Trends”, European Review of Economic History, forthcoming.


Thursday, 4 February 2021

Intergenerational mobility: what about the daughters?

by Vincent Delabastita*^ and Erik Buyst*

*Department of Economics, KU Leuven
^Research Foundation - Flanders (FWO)

blog post based on the article, "Intergenerational mobility of sons and daughters: evidence from nineteenth-century West Flanders", now available on EHER early view here

Research of the intergenerational transmission of socio-economic attainment has long had a restrictive focus on the relationship between fathers and sons. Recently, a growing strand of literature has taken up the challenge to overcome the omission of women. In historical research, however, this is challenging because (1) cultural tradition often prescribes that women change their name upon marriage, making intergenerational tracking of female life courses much more challenging (2) the socio-economic attainment of women on historical labor markets is often poorly documented. The first challenge can be solved by either constructing indirect links based on naming practices (Olivetti & Paserman, 2015; Olivetti, Paserman, & Salisbury, 2018), or by looking at areas for which it is possible to construct direct intergenerational links. This research project adds to a range of recent papers that adopts the latter approach, by studying the case of 19th-century West Flanders (Craig, Eriksson, & Niemesh, 2019; Dribe, Eriksson, & Scalone, 2019).

The common approach to overcoming the second challenge with respect to the definition of women’s socio-economic attainment is to take the husband’s or father’s occupational status as a proxy. We argue that this stance leads to a problematic neglect of the female experience in labor markets in the past, as working women were definitely ubiquitous in European history (for example, see Humphries & Sarasúa, 2012). An examination of women’s social mobility solely based on their marital mobility or the attainment of her father/husband shows only part of the picture. Therefore, our paper takes a different approach by examining parental influence on women’s own occupational decisions on the labor market.

19th-century West Flanders presents itself as a suitable case study, given that its economic structure was characterized by its export-oriented rural linen industry and typically relied on women in the role of flax spinners. This makes that marriage certificates from the West-Flemish civil register system are an excellent opportunity to overcome both challenges with respect to the study of female intergenerational mobility: not only were brides identified by their maiden names, their occupational activity at marriage was also commonly recorded (see Figure 1). Furthermore, the economic history of West Flanders presents us with interesting variation. In the middle of the 19th century, its once-flourishing rural linen industry collapsed under the pressure of mechanized competition in neighboring regions. This dramatic demise hit women disproportionally hard, as flax spinning was a common source of income among West-Flemish women. Towards the end of the 19th century, this period of economic turmoil was followed by a gradual process of industrialization.

Figure 1: Registration of an economic activity of the bride; 1830-1900

Results and discussion

Building on the intra- and intergenerational linkage of more than a million digitized civil birth and marriage certificates, we were able to construct a comprehensive sample of 40,703 parent-child pairs. We find evidence of a gender gap in occupational mobility, with sons being more attached to their socio-economic roots. Throughout the period under observation, however, there were only modest mobility gains for daughters compared to sons, leading to a gender convergence in mobility. Moreover, the risk of ending up in an unskilled occupation became progressively bigger for West-Flemish women as the rural industry was replaced by mechanized industries in neighboring areas. Overall, this presents a gloomy picture for 19th-century daughters, as they missed out on the possibilities offered by industrialization in terms of intergenerational mobility and socio-economic status.

In the background of the demise and resurgence of West Flanders’ industry, we point to two causal factors underlying these differential trends in socio-economic attainment. First, hand spinning – a typically female activity - was mechanized much more rapidly, leading to a starker decrease in the payoff of investing in daughters’ human capital and to higher levels of female mobility. In contrast, traditional linen weaving remained competitive against mechanized production for much longer, so the deindustrialization process went more smoothly for men. A second explanation for the overall lower mobility in the post-crisis period as well as the observed gender differential in mobility can be found in the gradual emergence of migration. We find that selection effects due to geographic mobility played a more important role in the determination of male intergenerational mobility, suggesting that migration was a more effective way to achieve social mobility for male workers.

From an international perspective, our results largely align with recent developments in the literature. Our estimates for father-son mobility are consistent with the idea that intergenerational mobility was significantly higher across the Atlantic Ocean (see Pérez 2019). Importantly, we present first evidence that a similar case can be made for daughters. Expanding our empirical framework to marital mobility, in which we take the traditional approach of imputing female social status by their husband’s attainment, a direct comparison with recent work on the US reveals that mobility in the US was markedly larger not only for sons, but also daughters (Craig, Eriksson, & Niemesh, 2019). Strikingly, a similar pattern is also found for American women, as daughters enjoyed less benefits in terms of mobility growth throughout the 19th century.

Read more about Vincent Delabastita's research at his website here; you can follow him on twitter here

Read more about Erik Buyst's research at his website here


Craig, J., Eriksson, K., & Niemesh, G. T. (2019). Marriage and the intergenerational mobility of women: Evidence from marriage certificates 1850-1910 (Tech. Rep.). Department of Economics, UC Davis. (Mimeo)

Dribe, M., Eriksson, B., & Scalone, F. (2019). Migration, marriage, and social mobility: Women in Sweden 1880-1900. Explorations in Economic History, 71, 93 - 111.

Humphries, J., & Sarasúa, C. (2012). Off the record: Reconstructing women’s labor force participation in the European past. Feminist Economics, 18(4), 39–67.

Olivetti, C., & Paserman, M. D. (2015). In the name of the son (and the daughter): Intergenerational mobility in the United States, 1850–1940. American Economic Review, 105(8), 2695–2724.

Olivetti, C., Paserman, M. D., & Salisbury, L. (2018). Three-generation mobility in the United States, 1850–1940: The role of maternal and paternal grandparents. Explorations in Economic History, 70, 73–90.

Pérez, S. (2019). Intergenerational occupational mobility across three continents. The Journal of Economic History, 79(2), 383416.

Tuesday, 2 February 2021

Comparing Income and Wealth Inequality in Pre-Industrial Economies: the case of Castile (Spain), c. 1750


by Esteban Nicolini and Fernando Ramos-Palencia (@framospalencia)

blog post based on the article, "Comparing income and wealth inequality in pre-industrial economies: the case of Castile (Spain) in the eighteenth century", available on EHER early view here

Our knowledge of the evolution of economic inequality within countries in pre-industrial Europe has expanded considerably in the last years. The two most important dimensions of economic inequality in the literature are related with income (a flow) and wealth (a set of assets); although many researchers implicitly assume that these two variables are very good substitutes of each other, there is no study on the relationship between these two variables for pre-industrial Europe. 

In general, incomes are composed by the returns to physical assets (capital or land), to financial assets, to human capital and to raw labor; on the other hand, for a given person or household, the stream of incomes influence savings that accumulate in future wealth. The relative importance of the different kinds of assets in total wealth and in the generation of income, changed substantially with economic growth. In modern societies, the agricultural sector plays a relatively minor role in aggregate production, income inequality is only weakly linked to the distribution of land property and a large bulk of income inequality is related with labor incomes and retribution to human capital (Shorrocks 1982). However, in traditional pre-industrial economies, most of the population worked in the primary sector, land and labor were the most important productive factors and land property was a major source of income, power and status. In these economies, where average human capital was relatively low, most of economic inequality is expected to be explained by land distribution; even though labor retributions can be an important share of the total value of production, if labor is evenly distributed across individuals, its contribution to total income inequality would be small. So far, many scholars have relied on the methodological assumption that, in preindustrial economies, inequality of assets like land or real estate could be considered a reasonable proxy of income inequality because the different subsets of wealth would correlate very well with each other and all of them would correlate very well with income (for instance, Alfani 2015; Lindert 2014; Alfani and Ammannati 2017).

We have scrutinized the validity of this assumption in our paper. In highly urbanized commercial junctures, trading capital is probably important as well as Soltow and Van Zanden (1998) assume for Amsterdam in the eighteenth century and even in less economically advanced societies, labor income differences can be important: Nicolini and Ramos-Palencia (2016) have suggested that labor incomes contribute up to 65% of income inequality in urban areas of Old Castile in the 18th century and Álvarez and Ramos-Palencia (2018) have stressed the importance of human capital to explain income inequality in the same region and period.

Our article presents a new data set to analyze economic inequality in Spain based on information, circa 1750, from Palencia, Madrid, Guadalajara and Granada. This data set has some unique characteristics. First, it combines information from two different sources: probate inventories, which contain detailed descriptions of household wealth; and the Ensenada Cadastre, a mid-century government census that contains information about household income, the contribution of each income source (for instance land or labor) to total income and other characteristics like household head’s occupation and ability to sign. Second, the data set enables us to link the households from the set of inventories with their corresponding records in the Cadastre; this connection makes it possible to analyze the relationship between the income of a household when the Cadastre was produced and the wealth of that household some years later, when its head passed away. This data set opens the possibility to link the distributions of income and wealth so that we can propose hypotheses via which their differences can be better understood and the possible shortcomings of using one as a proxy for the other.

We find that the income assigned by the EC and the wealth registered in the PIs are closely associated suggesting that, even though income inequality seems to be consistently less than wealth inequality, both variables capture very well a unique dimension of economic inequality in pre-industrial economies and that a given household’s location in one distribution depends strongly on its location in the other. Given that many times, data scarcity forces researchers to use the distribution of wealth, real estate or other assets to approximate the distribution of income (Alfani 2017, Lindert 2014), the confirmation that household’s wealth can be a very good predictor of income is extremely valuable from a methodological point of view.

Using an econometric specification in which both income and wealth are stated in terms of their logarithms the elasticity of income with respect to wealth varies between 0.4 and 0.6 (depending on the specification). These values imply that a 10% increase in the wealth of a household is associated with its income being from 4% to 6% higher. Elasticity that is less than 1 is consistent with general observations –confirmed with our data- that wealth inequality is greater than income inequality.

The parameters associated with our Secondary and Tertiary dummy variables are positive and the former is statistically significant in all the specifications. This result suggests that, for a given level of wealth, households with a head who works in one of those sectors, particularly in the secondary sector, tend to have more income than households with a head who works in the primary sector. Variables proxying human capital also have a sizable impact on the income distribution: for a given level of wealth those heads of households with skills (mainly captured by the kind of occupation but also by the ability to sign) have significantly larger incomes than those without skills highlighting the importance of the human capital in the determination of labor incomes (Álvarez and Ramos-Palencia 2018). 

For instance, using the parameters obtained in the regression in levels (see specification C in the table above), we can compare the income predicted by our equation for a head of a household without any wealth or human capital and working in the agricultural sector (394 reales) with the income predicted for a similar household but with some human capital; if we add literacy to this head of the household, income would increase 90% (up to 749 reales) and if we predict the income with a high-skill occupation income would increase 224% (up to 1278 reales).

These examples show that the way in which the wealth and income distributions are related is more complex than the one suggested by a pure traditional and agricultural society in which land and real estate are the only productive assets generating social differentiation. Our results suggest the relationship between income and wealth can be affected in some non -trivial ways if the whole society or some households experience shocks like mortality picks or migration (voluntary or forced) that change the nature and strength of that correlation. This multidimensional nature of the income inequality is not necessarily surprising and the roles of different kind of assets and human capital in the income distribution have been already emphasized for urban sophisticated economies in 17th and 18th centuries (Soltow and Van Zanden 1998). However, the confirmation of this pattern in a relatively backward and traditional economy (Alvarez-Nogal and Prados de la Escosura 2013) would suggest that in Modern Europe, structural change, urbanization and sophistication of labor markets would generate complex changes in the income distribution and in the relationship between overall income, income sources and wealth that can be overlooked if we focus only on one dimension of economic inequality.