Fifty Years of Growth in American Consumption, Income, and Wages By Bruce Sacerdote (Darmouth) Abstract: Despite the large increase in U.S. income inequality, consumption for families at the 25th and 50th percentiles of income has grown steadily over the time period 1960-2015. The number of cars per household with below median income has doubled since […]
by Kathryn E. Gary, PhD candidate, Lund University
Women were important workers in the past, but they are still under-studied and their contributions largely absent from big-picture discussions of historical living standards. This is largely because women’s work remains to some extent a black box, but recent research has both challenged assumptions about how women participated in the paid labor market (c.f. Humphries and Sarasua 2012) and provided data about women’s payment for different kinds of labor (c.f. Humphries and Weisdorf 2015). The current work contributes to both these areas, by creating series of men’s and women’s wages in early modern Sweden, and by exploring both the mechanisms behind the gender gap in pay as well as the conditions under which women enter paid labor, with the goal of better understanding work in the past in general.
Primary data come from unskilled workers in the construction industry in Southern Sweden, predominantly from the towns Malmö and Kalmar; these are combined with published data from Stockholm, also from construction workers (Jansson, Andersson Palm, and Söderberg 1991). All data are for individuals paid by the day; relative wages are simply the percentage of men’s wages that women earn.
Figure 1 shows women’s relative wages from 1550 to 1759. Relative wages are high at the beginning of the period, around 80 percent, and increase to levels of parity in the early 17th century, after which they decline substantially, reaching as low as 40 percent during the end of the seventeenth century and into the eighteenth. This is a substantial decline over the period of not much more than a generation.
Some relative wage peaks are related to events that change both the demand for and supply of labor. Kalmar was a border town between Sweden and Denmark; from 1611 to 1613 the two countries fought the Kalmar War. Following these years women’s wages peaked, likely due to necessary rebuilding and a shortage in the supply of men. There is a wage spike in the same city following a fire in 1647 – while the national average weighs down the peak values, the deviations are still clear in the series, and when Kalmar is examined individually women’s relative wages peak as high as 1.33.
Table 1: Women’s work days as a percentage of all workdays in Kalmar, 1614-1710
Women’s ability to earn high wages goes against many of our theories about women’s earning potential – women are expected to earn less than men in physical tasks, because women are not as strong as men, and so are less productive physical laborers (Burnette 2008). Other theories suggest that women face constant wage discrimination (c.f. Bardsley 1999) – but this, too, is confounded by women’s ability to out-earn men, and by the large changes in the relative wage series. Something else is happening.
To understand we must look more closely at the data. In Kalmar workers are almost universally identifiable, allowing for deeper examination of the workforce. Table 1 shows the percentage of paid workdays that were worked by women, compared with the total number of paid work days in five year periods. Comparing the proportional feminization of the workforce with the amount of work, we see that the periods with the greatest amount of work are those in which the workforce is the most feminized – these periods are also those during which women’s relative wages are highest (see figure 1).
In combination with the relationship between total paid workdays and women’s relative wages across the whole country (figure 2), we are faced with a pattern that is familiar from the first and second world wars – when labor demand is high, women enter the labor force in higher numbers and are able to command higher wages. There is less evidence that women were systematically paid less either due to discrimination or because of their lower productivity – instead, women are responsive to economic forces, and especially to demand forces.
It is simple to to extend our sense of what is ‘traditional’ deep into the past, and to apply broad categories of ‘men’s’ and ‘women’s’ work. However, when we are able to suspend our assumptions and dig deeper into the evidence, the data tell a less expected story; women in Sweden worked in physical occupations, alongside men, often for similar wages. They worked especially hard when the need was highest, and women’s wages only fell away from men’s when work became less regular and men and women weren’t employed together.
Accounting for women’s work shifts our understanding of household living standards in the long run, and provides strong evidence for what is intuitively clear: we cannot truly understand the past if we continue to discount the experiences or contribution of half the population.
by Daniel Gallardo Albarrán, appeared on 22nd May 2016
by Chris Minns, Economic History Department, LSE
The Great Depression devastated North American labour markets for a decade, with about a quarter of the work force unemployed at the peak of the crisis. It is well known that the headline figure conceals the extent to which the burden of the Depression was shared unequally. In addition to sharp differences in employment patterns between cities and regions, it was less-skilled workers who saw demand for their labour fall more than those able to access white-collar work. Older men who lost their jobs were particularly vulnerable to falling into the trap of long-term unemployment. There is some evidence to suggest that the Depression may have exacerbated ethnic differences in the labour market, with black men in the United States were affected more heavily than their white counterparts. How was the Depression experienced by the foreign-born population who had settled in large numbers in both Canada and the United States up until the early 1920s? A recent research paper by Kris Inwood, Fraser Summerfield, and myself sought to answer this question using statistical evidence drawn from new digital samples of the Canadian Censuses from 1911 to 1931.
There are three reasons we were particularly interested in this topic. First, while some social historians have argued that immigrants suffered greater exposure to labour market discrimination when jobs were rationed in the 1930s, there is surprisingly little published evidence to confirm or contradict this contention. Second, by focusing on the earnings of immigrants over a twenty year period, we wanted to see whether the experience of the Depression had implications for the long-run labour market adjustment of immigrants relative to native-born Canadians. Third, Canada offers an excellent laboratory in which to conduct this research. The Depression experience in Canada was comparable to the United States in terms of unemployment trends in the early 1930s, and the country was a leading destination for European immigrants from the late 19th century. A unique feature of early 20th century Canada was that Census questionnaires asked respondents to report their earnings beginning in 1901. This means that our measure of attainment can reflect changes in pay within occupations and the effects of spell of unemployment on total earnings.
Our analysis of Census earnings yields a striking pattern: immigrants experienced “reverse assimilation” in Canadian labour markets, with the gap in pay between immigrants and the native-born growing between 1921 and 1931. Figures 1 and 2 show predicted earnings for immigrants relative to otherwise identical native-born Canadians for young, recent migrants (Figure 1), and older migrants with longer tenure in Canada (Figure 2). Free migrants had unrestricted access to Canada, and came mainly from the United States and the United Kingdom and Ireland. Preferred and non-preferred migrants hailed mostly from Continental Europe. The figures show that the relative decline of immigrant earnings was strongest among older men, even among those who had lived in Canada for decades when the Depression hit. We also find that the effects are focused almost entirely among immigrants from continental Europe who were not native English speakers, with American and British migrants experiencing no reversal in relative earnings.
Figure 1: Predicted relative immigrant earnings, born 1886, arriving 1911
Figure 2: Predicted relative immigrant earnings, born 1871, arriving 1896
The evidence of reverse assimilation is not a statistical artefact due to selective return migration of European migrants, or selective outmigration of Canadian residents to the United States; international migration flows were much lower in the early 1930s than the late 1920s, despite the government encouraging the return of indigent migrants to their home countries. Nor are the problems of migrants accounted for by a skills mismatch created by the differential shocks of the late 1920s and early 1930s, with immigrants having the misfortune of being concentrated in the jobs that were hit hardest. One factor that does account for a large share in the earnings gap is unemployment; immigrants in the 1931 Census were more likely to have lost time out of work than their native-born counterparts. This suggests that one way in which ethnicity mattered in the Depression was that those who were most obviously foreign were the first to lose their jobs and the last to be rehired. But unemployment does not fully account for reverse assimilation, as non-English speaking immigrants from continental Europe experienced a significant decline in weekly earnings between 1921 and 1931, relative to their native-born counterparts.
Our findings point to a disturbing conclusion: apparently well-integrated immigrants were more vulnerable to the adverse effects of a sustained recession than native-born workers with similar skills. Whether this pattern has been repeated in immigrant receiving economies during the recent crisis is an important question for future research.
Since Phelps Brown Hopkins published ‘Seven centuries’ in the mid 1950s economic historians and cliometricians have used ‘day wages’ – day rates for masons, carpenters and bricklayers taken from building accounts – to estimate the earnings of workers of the past. Whilst recent work has shown that these rates were not what the masons, carpenters and bricklayers actually received  many historians have been working on the means of earnings of other groups. A wage formation conference at the Institute of Historical Research on 16 September aimed to bring the notion that wages are more multifarious than day rates to the fore. The programme brought research on lead and coal miners, hostmen, keelmen, laundresses, sailors, bankers, spinners, agricultural labourers and clergy to debate, and the features that all these groups had in common in their pay before 1900 was an observation that all who attended shared.
Kicking off the day in opening remarks, Leigh Shaw-Taylor put the conclusions that authors such as Greg Clark, and Robert Allen and others have drawn from long run compilations of builder’s day rates within a theoretical context of structural change, pointing out that the role of real wages and average wages has been confused by cliometricians, and reminding us that Malthus predicted shifts in the wages of the poor, not of the average worker.
In the first presented case of the day Jane Humphries and Ben Schneider (Oxford) overturned the notion, common in recent historiography, that spinners were well paid and part of a high wage economy in England in the 18th century; rather they showed only the most productive spinners in England earned what Arthur Young described, moreover many spinners were employed by parishes at low piece rates under the poor laws. Amy Ridgway (Exeter) presented the only data from agriculture at the conference. Using the records of Kingston Lacy in Dorset she showed that the number of day labourers hired on a casual basis increased throughout the late 18th century and early 19th century, contrary to the established literature. Kathryn Gary (Lund) presented a new wage series for unskilled men in Sweden in the long run. She showed definitively that the wages unskilled men were not enough to support a family.
Four papers presented at the workshop dealt with the earnings of miners or those engaged in the coal industry. Andy Burn (Durham) showed that the keelmen of Newcastle-on-Tyne in the late 17th and early 18th century had pay that consisted of variable elements. Part was for hauling, another part for loading, and the rates varied according to location and season. Although the men were relatively well-paid when they were at work, the seasonality of the trade challenged living standards, and created a public order problem for the authorities. Tim Barmby (Newcastle) has been researching the Allendale lead miners. There men and mine owners bargained a price per fathom to be mined. To bargain effectively they needed to be able to predict, or have better information about the seams and geology that they were mining. Barmby shows that wage bargains were a means by which the mine owners extracted information from the more knowledgable miners. Unsurprisingly, the system produced unequal gains, with the best teams repeatedly winning the bargains. Guy Solomon (Exeter), who has fully quantitatively analysed Peter Kirby’s 2010 data shows that piece rates in coal mining in Northumberland brought about large variations in wage amongst workers doing the same job. Matthew Pawelski (Lancaster) showed how a Derbyshire free miner of the mid 18th century, John Naylor, used his own rights to common mining land to earn a large amount to take him out of a period of significant indebtedness. The case shows that as well as having his own resources, Naylor took local work with other employers when he could, and highlights the multifarious nature of earning for men of this class, and the role of book credit in such small enterprise.
Richard Blakemore (Reading) has spent the last three years looking at how sailors were paid. He debunked the common myth that sailors were an early modern global proletariat paid poorly wages. Instead he shows that Sailors earnings were, again, highly variable – many mariners made money from trading goods between ports. The form in which sailors were paid varied according to risk. Blakemore showed that the bargaining systems between shipowners and mariners benefited both parties at different times. Laundresses – a vital group never properly examined before – are the subject of Kathryne Crossley’s (Oxford) research. Drawing on the records of Oxford Colleges she shows that their status, and the means by which they were paid shifted over the 17th and 18th centuries. In the earlier period they operated as enterprising sole traders, in the 19th century they were integrated into the discipline of college staff. Anne Murphy (Hertfordshire) brought some badly needed research into white collar workers. Bank of England clerks had much in common with sailors – and laundresses – it turns out. The basic salary that the clerks received was at the very lowest end of white-collar earnings in in London. Variation and extra income were earned by the clerks through gratuities, frequently for favours for clients, and trading illegitimately as brokers. Judy Stephenson (Oxford) gave a review approach, centred around the question of trying to work out how representative day wages used in macroeconomics series really are of earners in London across the long eighteenth century. Early research, funded by Cambridge Humanities Grant, indicates that few London workers were paid by the day before 1800. Wouter Marchand (Utrecht) demonstrated that the pay of clergy in early modern Friesland was dependent on the quality of land that church lands produced income from. The clergy are one of those groups that economists love to refer to as sacrificing wages for status. Marchand shows that their wages were not determined by custom. The best paid clergy were in merged or combined parishes on fertile soil.
The commonalities between the cases presented at the workshop was remarkable. These kept coffee breaks and lunch and dinner abuzz with debate, conversation and connections. The most marked was the observation of varying levels of income due to the effects of piece rates, bargaining and variable pay structures. Variation in earnings of people doing the same jobs was a consistent theme throughout the cases presented. Moreover, nearly all the cases showed only small part of income came from basic pay, and auxiliary rates, gratuities, alternate employment and bargains, were used to meet the problems of information asymmetry, seasonality or uncertainty. This was directly related to the materiality of some of the occupations. It was also noted that the agency or bargaining power of workers in a number of sectors was a determinant of their income. A final comment was that that ‘custom’, which dominates a great deal of historical literature, was not mentioned all day as as a determining variable in any of the cases presented.
The conference reinforced the idea held by many participants that wages in the early modern period and nineteenth century were a more complex issue than the use of real wages in long run studies have suggested, but it also showed that the topic of wage formation is ripe for further research. The full proceedings and papers will be published at a later date.
Judy Stephenson. Judy.Stephenson@wadh.ox.ac.uk
 Stephenson, EcHR, forthcoming.