. 5
( 14)


work.110 Not only are men larger, but a greater proportion of their
weight is devoted to muscles. Muscles are about 42 percent of body
weight for men, but only 36 percent for women.111 Men also have larger
hearts and lungs and can use oxygen more ef¬ciently.112 Women™s
bodies have other advantages, such as greater resistance to disease, but
they are inferior when it comes to brute strength.
Differences in muscle strength result in differences in the work that
individuals can do. Gender differences in the performance of exercise
tasks are as well documented as gender differences in muscle strength.
Table 2.4 shows sex differences in the performance of certain tasks
among US Army soldiers.113 Because they have smaller bodies, women
are less disadvantaged in tasks requiring individuals to move their own
bodies, such as running or sit-ups. The gender difference in sit-ups is not
statistically signi¬cant. In running time the gender difference is statis-
tically signi¬cant, but the female mean is less than two standard devi-
ations above the male mean. In lifting, however, the difference in ability
between the sexes is large. Men can lift twice as much as women, and the
gap is more than three times the standard deviation. Thus the largest gap
between the sexes in physical ability occurs in the ability to move
external objects, which is exactly what is required for heavy manual
These differences do not seem to be the result of differences in ¬tness,
since the gaps appear among soldiers, who are required to be ¬t. Table 2.5
compares the gender gap in running time and sit-ups among army

John Nicholson, Men and Women: How Different are They? (Oxford: Oxford University
Press, 1984), p. 28.
John Cooper, Marlene Adrian, and Ruth Glassow, Kinesiology (St. Louis: C. V.
Mosby, 1982), p. 142.
Franz Frohse, Max Brodel, and Leon Schlossberg, Atlas of Anatomy (New York:
Barnes and Noble, 1961), p. 39.
George Brooks and Thomas Fahey, Exercise Physiology (New York: John Wiley, 1984),
p. 642.
James Vogel and Karl Friedl, “Army Data: Body Composition and Physical Capacity,”
in Bernadette Marriot and Judith Grumstrup-Scott, eds., Body Composition and
Physical Performance (Washington, DC: National Academy Press, 1992).
Women™s wages 109
Table 2.4. Differences in physical performance by sex

Male Female Gap/ Gap/
Mean (SD) SDm SDf P(F  mm)
Mean (SD)

All ages
Two-mile run (minutes)* 15.1 (2.0) 17.9 (2.4) 1.4 1.2 0.12
Sit-ups 52 (14) 51 (13) 0.1 0.1 0.47
< 0.001
Maximum lift capacity (kg)* 59.2 (11.8) 29.7 (6.2) 3.5 4.8
By age group
Two-mile run (minutes)
17“20* 14.0 (1.6) 17.2 (2.4) 2.0 1.3 0.09
21“27* 14.4 (1.8) 18.0 (2.0) 2.0 1.8 0.04
28“39* 15.8 (2.0) 18.4 (3.0) 1.3 0.9 0.19
Maximum lift capacity (kg)
< 0.001
17“20* 61.3 (11.7) 30.4 (8.1) 2.6 3.8
< 0.001
21“27* 61.0 (12.0) 29.3 (4.8) 2.6 6.6
< 0.001
28“39* 56.6 (11.0) 30.3 (7.2) 2.4 3.7
40 þ 53.0 (8.3)

* ¼ difference between males and females is signi¬cantly different from zero at the 5% level. The
difference in means divided by the standard deviation is a measure of effect size known as
“Cohen™s d,” and values above 0.8 are considered large. See Jacob Cohen, Statistical Power
Analysis for the Behavioral Sciences, 2nd edn (Hillsdale, NJ: Lawrence Erlbaum Associates,
1988), ch. 2.
Source: Vogel and Friedl, “Army Data,” p. 93. The sample consists of 1126 male and 265
female soldiers.

recruits to the gender gap among soldiers in their twenties.114 Soldiers
are ¬tter than the recruits; running time decreases and sit-up perform-
ance increases as we move from trainees to soldiers. There is some
support for the hypothesis that among the civilian population women are
relatively less ¬t, since women™s relative sit-up performance increases
from 77 to 91 percent of male performance. In running time, however,
the gender gap does not narrow as ¬tness improves. The performance
measures in Table 2.4 control for issues of physical ¬tness because they

Data on army trainees are from Bruce Jones, Matthew Bovee, and Joseph Knapik,
“Associations among Body Composition, Physical Fitness, and Injury in Men and
Women Army Trainees,” in B. Marriott and J. Gumstrup-Scott, eds., Body
Composition and Physical Performance (Washington, DC: National Academy Press,
1992). This study uses 2245 trainees at Fort Jackson, South Carolina, in 1988. Data
on soldiers are from Vogel and Friedl, “Army Data.” This study uses a sample
including both soldiers from Fort Hood, Texas, and students from the Army War
College at Carlisle Barracks, Pennsylvania.
110 Gender, Work and Wages in Industrial Revolution Britain
Table 2.5. Gender gaps in performance for recruits and soldiers

Males Females Ratio

Army trainees
Two-mile run (minutes) 16.4 20.3 1.24
Sit-ups 44.3 33.9 0.77
Soldiers, age 21-27
Two-mile run (minutes) 14.4 18.0 1.25
Sit-ups 57.0 52.0 0.91

Source: Army trainees from Jones, Bovee, and Knapik, “Body Composition,
Physical Fitness, and Injury,” pp. 141“73. Soldiers from Vogel and Friedl, “Army
Data,” pp. 89“103.

are results from a sample of soldiers, who are more uniformly physically
¬t than the general population. Even for this highly ¬t population,
though, differences in lift capacity are large.
Unfortunately we do not have similar direct measures of performance
for tasks done by eighteenth- and nineteenth-century workers. We do,
however, know that much of the work available during this time required
lifting objects or exerting force. Strength was an important component
of productivity in a large percentage of the work available in the ¬rst half
of the nineteenth century. Lane notes that “muscle power was essential
in most trades . . . a wide range of artisans were obliged to lift, carry and
move very heavy weights as a normal part of their work.”115 Francis
Place notes that his father, a baker, was strong enough to carry two sacks
of ¬‚our at the same time.116 While industrialization did reduce the need
for strength, it did not necessarily do so immediately. Humphries doubts
that strength requirements can explain the occupational sorting of
women workers found in the 1851 census because industrialization
decreased the requirements for strength; she claims, “Technical change
in the eighteenth and nineteenth centuries tended to reduce the need for
human muscle power and hence would logically widen female oppor-
tunities.”117 Similarly, Hudson and Lee claim that:
In theory one might expect that technological change would increase the
opportunities for women to work on a greater parity with men. By removing
or lightening tasks requiring great physical strength, more ef¬cient tools and

Lane, Apprenticeship in England, p. 51.
Mary Thale, ed., The Autobiography of Francis Place (Cambridge: Cambridge
University Press, 1972), p. 20.
Humphries, “Most Free from Objection,” p. 934.
Women™s wages 111

mechanization might have changed that aspect of the sexual division of labor
grounded in or at least justi¬ed by the notion of female physical inferiority.118

However, while industrialization eventually eliminated the need for
strength in most industries, it is not true that strength was irrelevant by
1851. Strength continued to be important. In coal-mining, for instance,
machines were introduced for ventilation and draining, but throughout
the nineteenth century the most physically demanding work, hewing the
coal, still relied on shovels, picks, and human muscle.119 Samuel noted
that the new machines of the Industrial Revolution required a great deal
of strength and called it “a cruel caricature to represent machinery as
dispensing with toil.”120
The change in strength requirements was not even monotonic; in at
least a few cases the strength requirement increased substantially before it
decreased. The industrial innovations with the greatest impact on wo-
men™s work were those in spinning. Spinning had been the largest
employer of women, but mechanization reduced total employment, and
eventually women spinners were replaced with men. The mule, a
combination of the earlier jenny and water frame, started as a relatively
small machine that could be worked by women, but it rapidly increased
in size. Soon the mule was so large that it required too much strength for
women to work it. A mule carriage with 336 spindles for spinning coarse
yarn weighed 1400 pounds, and this had to be moved by the spinner
three and a half times per minute.121 Thus, until the mule was fully
mechanized in 1830, mechanization increased the strength requirements
of spinning. A similar increase in the size of a machine occurred in
framework knitting. In 1833 a parliamentary investigator reported:
The labour of working the hand-machines must be very severe; and as fresh
experiments seem to be constantly making on the degree of toil which the human
frame is capable of sustaining, some of the recently constructed machines are
such as none but the most athletic can manage. In 1829 the widest machine
known was a twelve-quarter; that is capable of making a piece of net three yards

Pat Hudson and W. R. Lee, “Women™s Work and the Family Economy in Historical
Perspective,” in Hudson and Lee, eds., Women™s Work and the Family Economy in
Historical Perspective (Manchester: Manchester University Press, 1990), p. 9.
Raphael Samuel, “Mechanization and Hand Labour in Industrializing Britain,” in
Lenard R. Berlanstein, ed., The Industrial Revolution and Work in Nineteenth-Century
Europe (London: Routledge, 1992), p. 30.
Raphael Samuel, “Workshop of the World: Steam Power and Hand Technology in
mid-Victorian Britain,” History Workshop 3 (1977), p. 8.
William Lazonick, “Industrial Relations and Technical Change: The Case of the Self-
Acting Mule,” Cambridge Journal of Economics 3 (1979), p. 235.
112 Gender, Work and Wages in Industrial Revolution Britain

wide. Since that time they have progressively enlarged, and I saw one man at
work on a stupendous hand-machine twenty quarters or ¬ve yards wide.122

An increase in strength requirements can also be seen in agriculture.
During the early nineteenth century, the need for strength in harvesting
increased because of the gradual replacement of the sickle with the
scythe.123 The scythe was much heavier than the sickle, and the mower
who used it bene¬ted from both height and strength. Harvesting
required large amounts of strength until it was mechanized in the late
nineteenth century.124 In the early part of the nineteenth century,
mechanization had by no means freed women from their natural dis-
The fact that women can occasionally be observed doing a task proves
that their productivity was greater than zero, but it does not prove that it
equaled male productivity in the same job. Women rarely plowed, but
on occasion have been observed to do this task. Most cases of women
plowing seem to have occurred in the north, where labor was relatively
scarce. In 1794 Andrew Pringle noted that, in Westmorland, female
servants “drive the harrows, or the plow, when they are drawn by three
or four horses.”125 Judy Gielgud interviewed women from the north who
had plow with horses in the early twentieth century.126 This proves that
women could plow, but it does not prove that they did the job as well
as men. If men were scarce, or if small farm size limited the possibilities
for division of labor according to comparative advantage, women may
have been set to work plowing even though they did the job more slowly
than men.
We must also keep in mind that what matters is not the effort put in by
the worker, but the work output. The market rewards workers for what
they produce, not for how hard they try. Elizabeth Roberts dismisses the
claim that men™s greater strength justi¬ed their higher wages because
“Women expended prodigious amounts of strength and energy in, for
example, the mining industry, in agriculture and in domestic work.”127

BPP 1833 (450) XX, C1, p. 35.
See E. J. T. Collins, “Harvest Technology and Labour Supply in Britain, 1790-1870,”
Economic History Review 22 (1969), pp. 453“73, and Roberts, “Sickles and Scythes.”
The mechanical reaper, which was widely adopted in the US in the 1850s, was not well
adapted to the small ¬elds and rough, wet terrain of Britain. Paul David estimates that
“in 1874 probably more than 53 percent of the British corn-harvest was still being cut
by the sickle and scythe.” Paul David, Technical Choice, Innovation and Economic
Growth (Cambridge: Cambridge University Press, 1975), p. 236.
Andrew Pringle, General View of the Agriculture of the County of Westmoreland
(Edinburgh: Chapman and Co., for the Board of Agriculture, 1794), p. 41.
Gielgud, “Nineteenth Century Farm Women,” pp. 108“9.
Roberts, Women™s Work, p. 14.
Women™s wages 113

She seems to be suggesting that women should be paid as much as men
because they put a great deal of effort into their jobs. However, even the
woman who strains herself to the limit will not be able to produce as
much output as a man. If paid a market wage, the man will earn more
because he produces more.
Contemporary observers were aware of the importance of strength
for productivity. Alfred Austin, who investigated agriculture for a par-
liamentary committee in 1843, concluded: “The strength required for
the work performed by men effectively prevents women from being
employed in it; and the lower rate of wages for which they work
has not had any tendency, therefore, to make them more generally
employed.”128 Interpreting such statements is dif¬cult. Do these
statements re¬‚ect what work was really like, or are they simply re¬‚ec-
tions of ideology that connects male work with strength? It may be
useful to note that such statements about physical strength limiting
productivity came not only from those who accepted women™s lesser
social status, but also from those trying to improve women™s status.
William Thompson, a supporter of socialism and equal political rights
for women, connected this physical disadvantage to lesser productivity
in work. He noted that, “Two circumstances “ permanent inferiority of
strength, and occasional loss of time in gestation and rearing infants “
must eternally render the average exertions of women in the race of the
competition for wealth less successful than those of men.”129 The early
feminist Mary Wollstonecraft admitted that: “In the government of the
physical world it is observable that the female in point of strength is, in
general, inferior to the male . . . A degree of physical superiority cannot,
therefore, be denied.”130 Even those who thought that women were
equal to men in intellect, and should be equal to men politically, agreed
that women were disadvantaged in the world of work by their lesser
Could differences in strength really have created differences in
productivity large enough to account for the wage differences observed?
In certain cases the answer is clearly yes. Even among men, strength
differences caused substantial differences in wages. From a framework
knitter we learn that, “The strong man can earn 18s. or 20s. a week on

BPP 1843 (510) XII, p. 27.
William Thompson, Appeal of One Half of the Human Race, Women, against the Pretensions
of the Other Half, Men, to Retain them in Political and thence in Civil and Domestic Slavery
(New York: Source Book Press, [1825] 1970), p. x. He goes on to say that since women
are disadvantaged in this way, they must have the protection of political rights.
Mary Wollstonecraft, A Vindication of the Rights of Woman (New York: Norton, [1792]
1967), p. 32.
114 Gender, Work and Wages in Industrial Revolution Britain

the wide work, while another poor man sits for 6s.”131 In this case,
strength could increase a man™s wage by a factor of three. In light of such
differences in productivity, the gender wage differences of the Industrial
Revolution no longer seem surprising. Indeed, if in some cases differ-
ences in strength might alter earnings by a factor of three, then an
ef¬cient allocation of jobs was necessary to keep the gender wage gap
from being even larger than it actually was.
Wage premiums for occupations requiring strength have been inter-
preted as ef¬ciency wages. Allen noted that wages in occupations
requiring strength remained high while those in occupations not
requiring strength were eroded. He interpreted this as an ef¬ciency
wage; the excess wages in the “privileged sectors” were necessary to
“secure a work-force that was suf¬ciently nourished.”132 This point
reinforces my claim that individuals who had more strength were more
productive and thus could earn higher wages, but it adds a circularity to
the argument, since according to this argument differences in strength
were partially the result of higher wages. This raises the question of
whether the differences in strength themselves were the result of dis-
criminatory practices that gave men privileged access to food. While I
admit the possibility that high wages in strength-intensive jobs may have
an ef¬ciency wage component, I do not believe that differences in access
to food is suf¬cient to explain the differences between men and women.
The evidence presented above documents large gender differences in
strength among well-fed modern populations. Evidence from army
soldiers and recruits presented in Table 2.5 suggests that differences in
performance between ¬t and un¬t individuals are smaller than differ-
ences between men and women. While ef¬ciency wages may have been
part of the story, productivity differences between men and women were
not simply the result of men™s privileged access to food.
This section has provided clear evidence that men are stronger than
women, and that these strength differences lead to large differences in
certain types of performance such as the maximum amount of weight an
individual can lift. Since Industrial Revolution technology required a
great deal of strength, and wages differed among members of the same
gender according to their strength, it seems reasonable that strength was
an important reason for the differences in productivity documented in
Section II.

BPP 1833 (450) XX, C1, p. 25. Evidence of George Goode. Wages also differed
among women according to strength. The parliamentary investigator Alfred Austin
found that “a woman who is strong and active, a good work-woman, is paid higher
than one of inferior strength.” Women and Children in Agriculture, p. 6.
Allen, Enclosure and the Yeoman, p. 300.
Women™s wages 115

B. Training
While strength was important, it was not the only reason for differences
in productivity. Differences in skill also affected productivity. Women
certainly did not lack skills, but on average they had less training than
men, and they tended to acquire different skills than men. This section
will discuss both formal training, through schooling and apprenticeship,
and informal on-the-job training through work experience. Women
received less schooling than men, and were less likely to be apprenticed,
so they had less formal training. Women were probably as likely as men
to acquire skills through experience, but they acquired different skills
than men, and the value of women™s skills fell during the Industrial
During the Industrial Revolution women received less formal educa-
tion than men. While many girls were taught to read, few learned more
than this. Girls were often kept home from school to help with the
housework, and rarely progressed beyond primary subjects. In 1792
Mary Wollstonecraft argued that the intellectual inferiority of women
was the result of inadequate education rather than natural ability.133
Few women, whatever their ability, had any opportunity to become well
educated. Beyond the elementary level, boys and girls were educated
very differently. Grammar schools, which provided a classical education,
were only open to boys.134 Exceptions were rare. A few girls were
admitted to a Blackburn grammar school in 1830, but in 1833 the
school™s governors ruled that no girls were to be admitted without a
special vote.135 Roach concludes that, “It is probable . . . that it became
more and more uncommon as time went on for girls to receive classical
training.”136 Girls were taught basic literacy in elementary schools, but
additional education was limited to household arts such as sewing and
“accomplishments” meant to attract a husband. The wealthiest girls had
governesses, but they were likely to learn subjects such as music and
French, subjects designed to help them in society rather than the
workplace because these girls were not expected to enter the workforce.
The boarding schools for girls were generally ¬nishing schools that did
not teach academic subjects. Music and embroidery were more common
at girls™ schools than Latin or science.137
Women™s opportunities for higher education were severely limited.
Oxford and Cambridge were closed to women until the twentieth century.

Wollstonecraft, A Vindication of the Rights of Woman.
Davidoff and Hall, Family Fortunes, p. 290.
Roach, A History of Secondary Education, p. 82. 136 Ibid.
Rosemary O™Day, Education and Society, 1500“1800 (New York: Longman, 1982), p. 189.
116 Gender, Work and Wages in Industrial Revolution Britain

Women were not fully admitted to Oxford until 1920.138 Cambridge
admitted women to exams in 1882, but did not allow them full privileges
until 1948.139 A degree from Oxford or Cambridge was a requirement for
entry into the professional elite. The Royal College of Physicians, for
example, admitted only graduates from these two universities.140
Differences in schooling led to marked differences in literacy levels.
Measuring literacy by the ability to sign the marriage register, Scho¬eld
found large differences in the illiteracy rate by gender. In the 1750s, 60
percent of women and 40 percent of men could not sign the marriage
register. In 1840 the rates were lower, but women were still less literate
than men; 50 percent of women and 33 percent of men could not sign
their names.141 Because reading was taught before writing, and many
women stopped attending school before they learned to write, differ-
ences in the ability to read were smaller than differences in the ability to
write. While factory workers are not a random sample of British workers,
an 1816 survey conducted by a parliamentary committee studying
children™s employment shows that, at least among factory workers,
females were less literate. Among these workers females were almost as
likely as males to be able to read; 84 percent of women over age 18 could
read, compared to 89 percent of men. For writing, however, the gap is
much larger; only 32 percent of the women could write, while 70 percent
of the men could.142
While literacy was not required for most jobs, it was required for many
of the best jobs. Literacy improved the chances of upward mobility for
both men and women, and women who wished to escape manual labor
and enter the most skilled occupations would ¬nd literacy valuable.
Mitch divided occupations into four categories: those requiring literacy,
those where literacy was likely to be useful, those where literacy may
have been useful, and those where it was unlikely to be useful. He found

L. Grier, “Women™s Education at Oxford,” in Handbook of the University of Oxford
(Oxford: Oxford University Press, 1969), p. 286. The ¬rst women came to Oxford to
study in 1878, but the road was long. By 1895 examinations were open to women. In
1896 an attempt to allow women to take the BA degree failed.
Christopher Brooke, A History of the University of Cambridge (Cambridge: Cambridge
University Press, 1993), vol. IV.
W. J. Reader, Professional Men: The Rise of the Professional Classes in Nineteenth-Century
England (London: Weidenfeld and Nicolson, 1966), p. 16.
R. S. Scho¬eld, “Dimensions of Illiteracy, 1750“1850,” Explorations in Economic
History 10 (1973), p. 443. Scho¬eld claims that the signature is a good measure of
moderate literacy skill, because “school curricula had been so phased that reading was
taught before writing, and the intermittent nature of school attendance thus ensured
that large numbers of children left school having acquired some reading ability, but
little or no ability to write” (p. 440).
BPP 1816 (397) III.
Women™s wages 117

that women were less likely to be found in the top two categories. In
1841, only 7.4 percent of women workers were in occupations where
literacy was required or likely to be useful, while 27.4 percent of working
men were.143 In a sample of marriage registers, literate brides were more
likely than illiterate brides to be dressmakers, proprietors, or professionals,
and they were less likely to be textile workers or unskilled laborers.
To some extent, women™s lack of schooling was a rational investment
choice. Since they would spend less time in the labor force, girls would
receive lower returns from education than boys. To some extent, how-
ever, women™s low education levels re¬‚ect discrimination in families and
in schools themselves. Families may have made their investment deci-
sions based on gender roles rather than the expected returns of the
investment. Also, many schools discriminated against women by not
admitting them. Even those women who wanted to could not attend
grammar schools or universities.
Apprenticeship was another important method of obtaining job skills,
though it became less important during the Industrial Revolution period.
Girls could be apprenticed and often were, but they were much less likely
than boys to receive this training. Parish apprentices were about 30 per-
cent female.144 Parish apprentices, though, often received little training
because they were apprenticed mainly to save the parish the expense of
maintaining them. A more accurate measure of human capital acquisition
is the number of girls apprenticed by their parents, and these percentages
are much lower. In the sixteenth and seventeenth centuries only 3 to 4
percent of private apprentices were female.145 For the eighteenth century,
estimates range from 3 to 8 percent, but average 5 percent.146 Clearly girls
were not gaining the same human capital as boys through apprenticeship.

David Mitch, The Rise of Popular Literacy in Victorian England: The In¬‚uence of Private
Choice and Public Policy (Philadelphia: University of Philadelphia Press, 1992), p. 15.
K. D. M. Snell, Annals of the Laboring Poor: Social Change and Agrarian England, 1660“
1900 (Cambridge: Cambridge University Press, 1985), Table 6.1. Steve Hindle ¬nds
the same sex ratio among pauper apprentices in the 1630s. Steve Hindle, “ ˜Waste™
Children? Pauper Apprenticeship under the Elizabethan Poor Laws, c. 1598“1697,”
in P. Lane, N. Raven, and K. D. M. Snell, eds., Women, Work and Wages in England,
1600“1850 (Woodbridge: Boydell Press, 2004), p. 35.
Michael Roberts ¬nds that 4 percent of apprentices in Bristol 1532“65 were female,
while Sue Wright ¬nds 3 percent for Bristol 1542“53, and 3.6 percent for Salisbury
1603“14. Michael Roberts, “ ˜Words They Are Women, and Deeds They Are Men™:
Images of Work and Gender in Early Modern England,” and Sue Wright,
“˜Churmaids, Huswyfes and Hucksters™: The Employment of Women in Tudor and
Stuart Salisbury,” in Lindsay Charles and Lorna Duf¬n, eds., Women and Work in Pre-
Industrial England (London: Croom Helm, 1985).
Roberts ¬nds the percentage of female apprentices to be 8 percent in Wiltshire 1710“60,
3 percent in Sussex 1710“60, 4 percent in Warwickshire 1710“60, 5 percent in
118 Gender, Work and Wages in Industrial Revolution Britain

Some of women™s lower productivity, then, seems to be the result of
deliberate choices to provide less formal training for girls, both through
education and through apprenticeship. To some extent these choices
were in¬‚uenced by gender roles that assigned girls more domestic tasks.
Girls were more likely than boys to be absent from school. In one public
school, between 1832 and 1834, boys had a 21 percent chance of being
absent, and girls a 27 percent chance.147 This difference was in large
part due to the expectation that girls would help with the housework. A
parliamentary investigator in Ireland found that “the services of females
are more frequently required at home than those of boys, and the con-
sequence is, their attendance at school is more irregular.”148 Gender
roles that assigned domestic work to girls rather than boys thus led them
to receive less education.
Even if parents were in principle willing to provide training for their
daughters, calculation of costs and bene¬ts would have discouraged it.
Apprenticeship was often a substantial investment; Table 2.6 shows
some examples of apprenticeship premiums paid. Parents were more
willing to invest in sons because doing so brought a larger reward. A girl
would be expected to spend less time in the labor market over her life-
time, lowering the potential value of human capital.149 Also, a woman™s
place in the economy was more often determined by the man she
married. Wives of tradesmen often helped their husbands rather than
engaging in independent work. In the guild system, wives had the special
position of being allowed to work in the trade without having been
apprenticed to it. A woman was likely to give up the trade in which she
had been trained when she got married. An example is the wife of James
Hopkinson; she had learned dressmaking and was managing a business
when she became engaged, but gave it up when she got married in order
to help James with his cabinet-making shop.150 If a girl was likely to
abandon her training to work in her husband™s shop, this would reduce
the potential payoff to her human capital and thus discourage investment.
Women™s lower level of human capital was to some extent due to
circular reasoning. Women needed less education because they were less

Bedfordshire 1711“20, and 5 percent in Surrey 1711“31. Roberts, “Words They Are
Women.” See also Snell, Annals of the Labouring Poor, Table 6.4.
Beryl Madoc-Jones, “Patterns of Attendance and Their Social Signi¬cance: Mitcham
National School, 1830“39,” in Phillip McCann, ed., Popular Education and
Socialization in the Nineteenth Century (London: Methuen, 1977), p. 58.
BPP 1840 (43) XXIII, p. 673.
Women™s lower labor force participation may have been partially due to their lower
skill levels and lower wages, but it was also at least partially due to women™s role in
Hopkinson, Victorian Cabinet Maker, pp. 85, 88, 96.
Women™s wages 119
Table 2.6. Examples of apprenticeship premiums

Date Location Occupation Premium Sex Src

1710 London Carpenter £4 F a
1713 London Grocer £180 M b
1715 London Apothecary £50 M c
1720 Birmingham Baker £15 M c
1733 London Mercer £126 M c
1737 Yarmouth Cordwainer £6 M d
1741 Chippenham Saddler £20 M c
1741 Dorchester Joiner £18 M c
1743 Hertford Butcher £20 M d
1750 Salisbury Milliner £40 F c
1759 Westminster Mantua-maker £10 F e
1767 Colchester Milliner £25 F f
1773 Coventry Surgeon £130 M c
1778 London Bookbinder £4 M g
1785 London Surgeon £420 M b
1792 London Bookbinder £15.15s M g
1796 Hertford Grocer £50 M d
1800 London “Wholesale £100 M h
1800 Essex Wholesale draper £500 M i
1819 London Stationer £156 M k
1824 Dublin Cabinet maker £100 M l
1833 Leeds Dressmaker 10s.6d. per year F m
1834 Nottingham Cabinet maker £20 M n
1840 Wool sorter £70 M o
1843 Nottingham Dressmaker £50“60 F p

a. E. B. Jupp and W. W. Pocock, Company of Carpenters (London: Pickering and Chatto,
b. Harold Perkin, The Origins of Modern English Society (London: Routledge and Kegan
Paul, 1969).
c. Lane, Apprenticeship in England, pp. 117, 122, 124, 138, 143.
d. O. Jocelyn Dunlap, English Apprenticeship and Child Labour (New York: Macmillan, 1912).
e. George, London Life in the Eighteenth Century.
f. D™Cruze, “To Acquaint the Ladies,” p. 161.
g. Ellic Howe, A List of London Bookbinders (London: The Bibliographical Society, 1950).
h. The Times, February 16, 1819.
i. Davidoff and Hall, Family Fortunes.
k. The Times, January 14, 1819.
l. BPP 1824 (51) V, p. 456.
m. BPP 1833 (450) XX, C1, p. 73.
n. Hopkinson, Victorian Cabinet Maker.
o. BPP 1841 (296) X, p. 41.
p. BPP 1843 (430) XIV.
120 Gender, Work and Wages in Industrial Revolution Britain

likely to work outside the home, but women were less likely to work
outside the home because their wages were lower, and these lower wages
were partially due to lower levels of education and training. If this trap
was purely due to socially determined gender roles, then we would call it
discriminatory. However, there were biological as well as ideological
forces keeping women in this trap. Women give birth and breast feed
infants, while men don™t. This is one factor, in addition to lower wages,
that makes women more likely than men to ful¬ll the domestic role. As
discussed above, in many occupations women™s lower wages were the
result of their lesser strength. Thus women™s lower levels of formal
training were only partially, and not wholly, due to discrimination.
Formal training, though, was not nearly as important during the
Industrial Revolution period as it is today. While a few of the most
prestigious occupations required formal education, most occupations
did not. Mitch calculates that, in 1841, 5 percent of males worked in
occupations where literacy was required and a further 23 percent worked
in occupations where literacy was likely to be useful. The bulk of the
male labor force, approximately three-fourths of male workers, worked
in occupations where literacy was not useful or only possibly useful.151
Women do seem to be underrepresented in occupations requiring lit-
eracy, but differences in literacy can explain occupational segregation for
only a minority of workers.
Differences in apprenticeship rates are likely to be a relatively minor
reason for occupational segregation because apprenticeship was becoming
increasingly irrelevant during the Industrial Revolution period. Even
before the Industrial Revolution, apprenticeship was not the only way to
acquire skills. In her study of female apprentices in Bristol, Ben-Amos
notes that “training and skills were also acquired by young women in the
town without a formal, recorded apprenticeship.”152 Many individuals
acquired their job skills through practical experience. Apprenticeship was
a legal requirement for entry into a trade under the Elizabethan Statute of
Arti¬cers until this law was repealed in 1814.153 Even before 1814,
though, the law was widely evaded. The Hammonds conclude that, in the
woollen industry, “The practice of enforcing a seven years™ apprenticeship
for weavers and cloth workers had fallen into disuse by 1802.”154 In 1803
only 13 percent of the weavers hired by a Gloucester manufacturer,

Mitch, The Rise of Popular Literacy, p. 15.
Ilana Krausman Ben-Amos, “Women Apprentices in the Trades and Crafts of Early
Modern Brisol,” Continuity and Change 6 (1991), p. 228.
Peter Kirby, Child Labor in Britain, 1750“1870 (Basingstoke: Palgrave, 2003), p. 95.
J. L. Hammond and B. Hammond, The Skilled Laborers, 1760“1832 (London:
Longmans, Green and Co., 1920), p. 170.
Women™s wages 121

and none of the cloth workers, had served an apprenticeship.155 In the
nineteenth century apprenticeship continued largely through parish
apprenticeship, which was an institution for providing for the poor rather
than for human capital acquisition.
Both men and women acquired valuable skills through work experi-
ence, but, because they did different work, they tended to acquire dif-
ferent skills. One occupation where women acquired signi¬cant skills,
and were paid good wages for their skills, was in dairy management.
Skilled dairywomen could earn more than male agricultural laborers. In
1821 a Gloucestershire bailiff paid £3.14s.3½d. to “Jos. Wilcox for his
wife attending the Dairy 7 weeks.”156 Even assuming that she worked
seven days per week, Mrs. Wilcox earned more than 18d. per day for this
work. At the same date the modal wage for male laborers was 16d. per
day, so Mrs. Wilcox earned a wage that was higher than the majority of
the male laborers.
Women were more likely than men to work at tasks requiring manual
dexterity rather than strength, and they acquired superior skills in
spinning and sewing. Though people sometimes interpreted these skills
as “natural” female skills, they were learned rather than innate. (To
check this, simply try asking a twenty-¬rst-century woman to spin or
sew.) Women learned these skills through practice. Men learned skills
too, but they tended to learn different skills. Skills in hand spinning, for
example, were acquired by women but not men. Sometimes young boys
would spin, but they do not seem to have done this work often enough to
acquire the same skills as women. Men, it seems, could not spin as well
as women. In the early seventeenth century the town of¬cials of Bocking
noted that spinning was the only work available for employing the poor,
and that “spinning work will not yield maintenance to those that want
work, they being for the most part men that have not been exercised in
the art of spinning.”157
When women™s skills were highly valued, women did well. An example
is the boom in the straw-plaiting industry during the Napoleonic Wars;

Estcourt accounts, Gloucestershire Record Of¬ce, D1571 vol. A36, Feb. 17, 1821. In
1821 Mrs. Wilcox was paid £13.16s. for a half year, which is the same weekly wage.
The 1821 payment for seven weeks of work was made when the Estcourts closed their
dairy. There are other examples of high wages given to dairywomen. A large farmer
from Worcestershire claimed that he paid the woman in charge of his dairy “55l. per
year for herself and servant, including maintenance.” In this case, unfortunately, the
wages of the dairywoman and her assistant were combined, but this case con¬rms that
dairywomen could earn relatively high salaries. Women and Children in Agriculture,
Quoted in Sharpe, Adapting to Capitalism, p. 31.
122 Gender, Work and Wages in Industrial Revolution Britain

women™s wages reportedly rose to 21s. per week, about twice the weekly
wage of a male agricultural laborer.158 Women lace-makers in late
seventeenth-century Devon also earned twice as much as male agricul-
tural laborers.159 In most cases where women™s wages rose above male
laborers™ wages, the high wages were short-lived.160 Maxine Berg claims
that high wages in lace-making were more enduring.
Wages in the seventeenth and early eighteenth century were high, higher than
those for wool spinners and much higher than those for local male agricultural
labor. Yet despite the evident prosperity of this occupation for a time, men were
not employed in it, neither did they seek to enter it.161

If the high wages earned by women in lace-making were indeed sus-
tained over a long period of time, this does suggest some skill barrier
preventing men from pursuing this occupation.
Unfortunately, the Industrial Revolution was not good to women.
The machinery of the Industrial Revolution replaced many hand skills
with machines. Sometimes this happened to men such as the cloth
dressers and woolcombers (who resisted strongly). But none of these
occupations could match hand spinning in terms of numbers employed.
Recall that, before the arrival of machinery, spinning was said to employ
all the women in certain areas. What had been a valued skill for most
women suddenly became worthless as the jenny, water frame, and mule
replaced hand spinning. To some extent, then, women™s low wages
re¬‚ected their bad luck in having their most important skills replaced by
machines. Unfortunately for women, the Industrial Revolution led to a
collapse in the value of women™s skills. The fact that women™s wages fell
relative to men™s wages during the Industrial Revolution was probably
due to bad luck; the Industrial Revolution eroded the value of certain
skills, and women happened to be more heavily invested in the hardest-
hit skills than men.
To some extent women™s lower productivity can be explained by
women™s lower levels of human capital. Women received less schooling
than men, and had lower apprenticeship rates. If these women were paid
wages equal to their productivity, then there was no discrimination in

Sharpe, Adapting to Capitalism, p. 57.
Pamela Sharpe, “Literally Spinsters: A New Interpretation of Local Economy and
Demography in Colyton in the Seventeenth and Eighteenth Centuries,” Economic
History Review, 44 (1991), p. 52.
Maxine Berg, “What Difference Did Women™s Work Make,” p. 37.
Ibid., pp. 32“3. Here Berg refers speci¬cally to Sharpe™s article on lace-makers at
Colyton. While Sharpe gives only one wage citation, she does show that lace-making
affected sex ratios and marriage patterns over a long period of time.
Women™s wages 123

the labor market, but the wage gap may still re¬‚ect discrimination if the
differences in education were the result of pre-market discrimination
that limited women™s schooling and apprenticeship. Women also suf-
fered from bad luck; they acquired signi¬cant skills through work
experience, but certain skills such as hand spinning lost their market
value during the Industrial Revolution.

IV. Did women earn customary wages?
As noted above, economic historians do not agree about how to explain
the gender wage gap. Some assume that the wage gap is evidence of
differences in productivity, while others suggest that women were paid
customary wages, which were lower because women were assumed to be
inferior to men. There are now quite a few articles discussing the
question of whether women™s wages were market wages or customary
wages, and many of these conclude that women™s wages were customary.
Pamela Sharpe argues that “there is evidence for the importance of
cultural factors outweighing rational economic decision-making.”162
Penelope Lane claims that the level of women™s wages must have been
set by custom because the wage gap cannot be entirely explained by
differences in productivity.163 Sonya Rose emphasizes the expectation
that women were secondary earners, who did not need to support a
family, and whose wages were only supplementary to the wages of the
men on whom they were dependent.164
Unfortunately the claim that women were paid customary wages is
dif¬cult to evaluate because the claim is not clearly de¬ned. Different
historians seem to mean different things when they use the term. For
some historians, in¬‚exibility is an important characteristic of customary
wages. Sharpe suggests that: “On the whole these women™s wages were
highly inelastic and could remain unchanged across generations
regardless of other factors taking place in the economy. This suggests
that women™s wages had a large customary element.”165 Penelope Lane,
however, suggests that wages may be customary even if they are ¬‚exible:
“Historians are aware of the effect produced on female wage levels by
male labor shortages, or the availability of alternative employment, but a

Pamela Sharpe, “The Female Labor Market in English Agriculture during the
Industrial Revolution: Expansion or Contraction?” Agricultural History Review 47
(1999), pp. 161“81.
Penelope Lane, “A Customary or Market Wage? Women and Work in the East
Midlands, c. 1700“1840,” in P. Lane, N. Raven, and K. D. M. Snell, eds., Women,
Work and Wages in England, 1600“1850 (Woodbridge: Boydell Press, 2004).
Rose, “Gender at Work.” 165 Sharpe, Adapting to Capitalism, pp. 145“6.
124 Gender, Work and Wages in Industrial Revolution Britain

¬‚exibility that breaks with custom is not evidence of a market wage.”166 It
is not clear whether in¬‚exibility is a necessary characteristic of customary
If we want to determine whether women™s wages were customary
wages, we must ¬rst de¬ne what that term means. In this section I
present ¬ve possible de¬nitions for “customary wages” and examine
whether each de¬nition ¬ts the facts. Wages might be customary in the
sense that they were set by the government, or in the sense that they did
not respond to changes in supply and demand. Wages might be cus-
tomary in the sense that they were lower than fair market wages would
have been. Wages might be customary in the sense that, when custom
and market forces con¬‚icted, custom took precedence. Finally, wages
might be customary in the sense that society developed ideologies and
customs designed to justify the wage differences that were created by the
market, and most people thought about wages in these terms. I ¬nd that
the evidence does not support the ¬rst four de¬nitions of customary
wages, and only the ¬fth de¬nition matches the evidence.

De¬nition 1: Wages were set by government regulation
Wages might be fairly described as customary if they were set by law
rather than the free market and legal wages were determined by cus-
tomary ideas about the relative worth of males and females. England did
have a legal apparatus for setting wages until 1813. The Elizabethan
Statute of Arti¬cers empowered local justices to set maximum wages for
various kinds of work; it was passed in 1563 and not repealed until
1813.167 In principle this would mean that wages were set by local
authorities rather than the market, but the evidence suggests that the
laws were not effective. Woodward suggests that wage regulations were
effective in the ¬rst half of the sixteenth century, but largely ineffective in
both the second half of the ¬fteenth century and the second half of the
sixteenth century. He concludes that after 1563 “of¬cial attempts to
control wage rates were largely unsucessful.”168 Certainly the system
was little used in the eighteenth century. Examining the evidence that
compares actual wages to assessed wages, Kelsall concludes that “there
is clearly a tendency for assessed and actual rates to diverge in the

Lane, “A Customary or Market Wage?” p. 118.
The Statute of Arti¬cers was not the ¬rst English law to provide for maximum
wages. Wage regulation began with the Statute of Laborers of 1349. R. H. Tawney,
“The Assessment of Wages in England by the Justices of the Peace,” reprinted in
W. E. Minchinton, ed., Wage Regulation in Pre-Industrial England (Newton Abbot:
David and Charles, 1972), p. 38.
Woodward, “The Determination of Wage Rates,” pp. 26, 28.
Women™s wages 125

eighteenth century.”169 By 1813 the law was so irrelevant that members
of parliament did not even know of its existence. The Hammonds report
that, when the Lancashire cotton weavers appealed to parliament to
have wages set according to the law:
In moving the second reading of the repealing Bill in the House of Lords, Lord
Sidmouth remarked that at the time that recent petitions for regulating wages
had been discussed in the House of Commons it was not known that there were
Acts in existence for regulating the rate of wages “but in the course of the last
year, it had been discovered that there were Acts both in England and Scotland
rendering it imperative on magistrates to ¬x the rate of wages.” Sidmouth as-
sumed “ and rightly assumed “ that it was only necessary to mention the ex-
istence of this legislation to secure its repeal.170

While the law permitting justices to set wages was not repealed until
1813, the law had fallen out of use in the eighteenth century, so wages
were not set by government regulation during the Industrial Revolution

De¬nition 2: Wages did not respond to the forces of supply and demand
Some historians have speci¬cally stated that women™s wages were not set
by supply and demand. For example, Hudson and Lee claim that “The
labor-market was segmented so that excess demand for female labor did
not translate itself into higher female wages.”171 Pamela Sharpe focuses
on the in¬‚exibility of women™s wages: “Whenever the women worked
and whatever they did, the most likely sum they would be paid was 6d.
This . . . must beg the question of the extent to which the payment
is an arbitrary, or customary ¬gure rather than representing a market
value.”172 Scholliers and Schwarz claim that “For most of the eight-
eenth and earlier nineteenth centuries the pay of women in agriculture
was set at 6d. a day over most of England, irrespective of price move-
ments, but also irrespective of a surplus or shortage of female labor.”173
While he shows that male wages did respond to supply and demand,
Woodward claims that women™s wages “were not simply re¬‚ections of

R. Keith Kelsall, “Wage Regulations under the Statute of Arti¬cers,” reprinted in
W. E. Minchinton, ed., Wage Regulation in Pre-Industrial England (Newton Abbot:
David and Charles, [1938] 1972), p. 118.
Hammond and Hammond, The Skilled Labourer, p. 87.
Hudson and Lee, “Women™s Work and the Family Economy,” p. 18.
Sharpe, Adapting to Capitalism, p. 80.
P. Scholliers and L. Schwarz, “The Wage in Europe since the Sixteenth Century,” in
Scholliers and Schwarz, eds., Experiencing Wages: Social and Cultural Aspects of Wage
Forms in Europe since 1500 (New York: Berghahn, 2003), p. 9.
126 Gender, Work and Wages in Industrial Revolution Britain

the supply of and demand for female labor.”174 Paul Johnson identi¬es
customary wages as stable wages, in contrast to market wages, which are
Has the labor market in Britain since early industrialization been characterized
by customary and stable wage differentials, or by ¬‚exible wages that have re¬‚ected
the marginal productivity of the worker and which have readily adjusted to
changing supply and demand conditions?175

Sometimes the claim that wages did not respond to market forces is a
claim about general tendency rather than an absolute rejection of any
responses to market conditions. In her 1996 study of women in Essex,
Pamela Sharpe also claims that wages were not responsive to economic
conditions. She notes that the wages of unskilled domestic servants
“were a matter of custom bearing little relationship to economic
determinants.”176 In a 1999 article, though, she acknowledged that the
market could occasionally affect wages: while women™s wages were
“sticky” at 6d. per day, in a few cases higher wages were paid, “which
shows that the market certainly had some effect.”177 Thus wages might
on occasion respond to market forces, but this was the exceptional case,
and on the whole they were in¬‚exible and unresponsive to the market.
Were women™s wages in¬‚exible, with only a few rare exceptions? If we
look at a large sample of women™s wages over time it is clear that women™s
wages were not ¬xed for generations. Figures 2.5 and 2.6 show the wages
paid to women in winter and summer at a sample of ninety-¬ve farms.
There is a great deal of geographical variation in wages at any one point
in time, and there is also movement in wages over time. Figures 2.5 and
2.6 combine wages from many different farms, but even if we con¬ne
ourselves to wages paid at a particular farm we ¬nd that women™s wages
were ¬‚exible over time. Buckland Abbey in Devon paid all its female
workers the same wage. In 1798 and 1799 they earned 6d. per day year-
round; by 1803 this wage had risen to 7d. in the winter and 8d. in the
summer.178 A farm in Lilistock, Somerset, owned by the Marquis of
Buckingham, also paid uniform wages to its female laborers; this was 10d.
in 1815 and 8d. in 1816 and 1817.179 At a farm in Mangursbury,
Gloucestershire, the female summer wage was 9d. per day in 1823 and

Donald Woodward, “The Determination of Wage Rates in the Early Modern North of
England,” Economic History Review 47 (1994), p. 37.
Johnson, “Age, Gender and the Wage,” p. 229.
Sharpe, Adapting to Capitalism, p. 114.
Sharpe, “The Female Labor Market,” p. 174.
Devonshire Record Of¬ce, Drake 346M/E8“E11.
Rural History Centre, BUC 11/1/11.
Women™s wages 127



Female Winter Wage (d./day)






1720 1740 1760 1780 1800 1820 1840 1860
Figure 2.5 Female winter wages


Female Summer Wage (d./day)







1720 1740 1760 1780 1800 1820 1840 1860
Figure 2.6 Female summer wages
128 Gender, Work and Wages in Industrial Revolution Britain

1824, and rose to 10d. in 1825. The winter wage rose at the same time
from 6d. to 7d., but fell back to 6d. in 1826.180 At the Oakes farm in
Derbyshire, the female non-harvest wage remained at 8d. between 1837
and 1846, but in 1847 rose to 10d.181 These are just a few of the indi-
vidual farms at which the female daily wage changes over time.
To further examine wage persistence I made use of the fact that about
half of the farms in my sample have wage observations from more than
one year. Table 2.7 shows the number of wage pairs in my data set
categorized by the number of years between the observations. There
were 131 cases where female summer wages were available for two
consecutive years at the same farm. In 102 of these cases the two wages
were the same, and in twenty-nine cases the two wages were different.
I conclude that the probability of the female wage changing over the
course of one year is 22 percent, and that the probability of wage
persistence is 78 percent. Figure 2.7 graphs the probability of wage per-
sistence from Table 2.7 against the number of years between wage
observations. Wage persistence declines as the number of years
increases. After one year most farms, 78 percent, are still paying the
same wage. After ¬fteen years half of the farms are paying the same
wage, and half of the farms have changed the wage. After three decades
the probability a farm was still paying the same wage is only 10 percent.
Thus there were some farms where the female wage remained ¬xed for
decades, but these farms were not in the majority.
Women™s agricultural wages not only changed, but did so in a way
that responded to supply and demand conditions. Regions that had high
demand for female labor in alternative occupations should have higher
wages in agriculture, since the presence of other work decreased the
supply of women available to agriculture. In my sample of farm accounts
I ¬nd that counties with more cottage industry had higher female wages.
Moving from an average county to Bedfordshire, where cottage industry
was most prevalent, increased female wages in agriculture by 20 percent.
The appearance of textile factories in the north-west also increased
female agricultural wages. The north-west region, including (for my
purposes) Lancashire, Cheshire, and the West Riding of Yorkshire, had
lower female wages than the south-east in the second half of the eight-
eenth century, but in the ¬rst half of the nineteenth century wages were
about 50 percent higher than in the south-east.182 Clearly a high
demand for female labor in competing industries increased the wages
earned by females in agriculture.

Rural History Centre, GLO 1/2/1. 181 Shef¬eld Archives, OD1518, OD1531.
Burnette, “Wage and Employment,” p. 677.
Women™s wages 129
Table 2.7. Wage persistence, female summer wages

Number of years
between Number of pairs
observations Number of pairs with equal wages Persisting percent

1 131 102 78
2 98 71 72
3 88 65 74
4 66 51 77
5 62 47 76
6 50 39 78
7 48 34 71
8 44 29 66
9 37 26 70
10 39 27 69
11 30 20 67
12 25 14 56
13 23 14 61
14 23 14 61
15 22 11 50
16 19 9 47
17 19 9 47
18 18 9 50
19 18 8 44
20 16 8 50
21 17 7 41
22 13 2 15
23 13 4 31
24 13 3 23
25 13 3 23
26 13 3 23
27 11 2 18
28 13 2 15
29 11 1 9
30 10 1 10

If the claim that women™s wages were unresponsive to the market was
an absolute one, then it could be disproven with one example. If the
claim is not absolute, and admits some exceptions, it is harder to dis-
prove. I have shown that women™s wages were not ¬xed over the period
1740 to 1850, and that they responded to regional opportunities for
women™s employment. I have also measured the extent of wage per-
sistence, and have shown that, while most farms paid women the same
wage after one decade, most farms had changed the female wage by the
130 Gender, Work and Wages in Industrial Revolution Britain


Probability of Wage Persistence








0 5 10 15 20 25 30
Year Ahead
Figure 2.7 Wage persistence

end of three decades. I hope this is enough to convince the reader that
instances of wages responding to market forces were not rare.

De¬nition 3: Wages were customary in the sense that they were lower than fair
market wages
Wages, even if ¬‚exible, might be said to be customary wages if they were
lower than market wages would have been. Sometimes the claim that
wages were customary boils down to a claim that women were paid too
low a wage. Penelope Lane, for example, acknowledges that female wages
responded to market conditions, but notes that “¬‚exibility that breaks with
custom is not evidence of a market wage.”183 Lane believes that women™s
wages were discriminatory because, even after correcting for hours
worked, “the gap that remains cannot be accounted for by differences in
productivity.”184 Lane speci¬cally points to the case of weeding, where she
argues that “it could be argued that women were actually more ef¬cient at
this task, since weeding was so closely associated with them in any century,
and therefore they should have received higher wages than men.”185

Lane, “A Customary or Market Wage?” p. 118. 184 Ibid., p. 112.
Ibid., pp. 111“12. Note, however, that the association of women with weeding implies
that women had the comparative advantage in this task, but does not imply that women
were absolutely more productive than men in this task.
Women™s wages 131

Here the claim is not that women™s wages were in¬‚exible, but that they
were too low relative to women™s productivity.
Supporters of customary wages simply do not believe that women
were half as productive as men. If it is true that women™s wages were
lower than their marginal product, then wage discrimination in the sense
de¬ned by Gary Becker existed.186 Evaluation of this claim requires an
evaluation of the relative productivity of male and female workers.
Evidence presented in Section II suggests that women were less pro-
ductive than men, and that the gaps were large enough to explain the
wage gaps. Women were 85 percent as productive as men in picking
cotton, but in agricultural more generally only about 60 percent as
productive. In manufacturing, women were about half as productive as
men in the nineteenth century. Today women are still not as productive
as men in manufacturing, though their relative productivity is higher,
at 83 to 84 percent of male productivity. Econometric estimates of
productivity suggest that the productivity differences were large enough
to explain the wage gap.

De¬nition 4: Wages were customary in the sense that they were determined by
custom even when custom con¬‚icted with market forces
Perhaps wages only seem to follow the dictates of the market because
custom and market did not often con¬‚ict. Michael Roberts speaks of the
“lack of incompatibility” between custom and market.187 Certainly it is
true that custom and market usually prescribed the same thing. Custom
led people to expect that women should be paid less, and, as we have
seen, productivity differences also suggested that women™s wages should
be lower than men™s. If custom and market pushed in the same direc-
tions, perhaps evidence of ¬‚exible wages does not disprove the import-
ance of custom in setting wages. But if there is any meaning to the
debate about whether wages were market or customary, then there must
have been some occasions when market and custom suggested different
outcomes, so that it is still meaningful to ask which was the main
determinant of wages. Which was the locomotive, and which the
caboose? Pamela Sharpe argues that custom was the determining force
for wages: “there is evidence for the importance of cultural factors
outweighing rational economic decision-making.”188 The evidence
I ¬nd, though, supports the opposite conclusion.

Gary Becker, The Economics of Discrimination (Chicago: University of Chicago Press,
Roberts, “Sickles and Scythes Revisited,” p. 89.
Sharpe, “The Female Labor Market,” p. 178.
132 Gender, Work and Wages in Industrial Revolution Britain

To ¬nd evidence of whether custom or market was pulling the train,
we need to ¬nd evidence of cases where custom and market con¬‚icted,
and see which prevailed. One such area of con¬‚ict is cases where the
market pushed women™s wages above men™s wages. It is generally agreed
that society expected women™s wages to be lower than men™s, so if the
market suggested women™s wages should be higher than men™s, there
was a con¬‚ict between the prescriptions of custom and the market.
When market forces increased the demand for occupations where
women had skills, we occasionally do see women earning wages higher
than men earned. An example is in the lace and straw industries during
the Napoleonic Wars. War with France cut off imports, increasing the
demand for domestically supplied lace and straw. In response to this
demand, women™s wages rose. In some cases women in these industries
earned more than male agricultural laborers (who couldn™t switch to
making lace because they didn™t have the skills). In the straw-plaiting
industry, women™s wages reportedly rose to 21s. per week, about twice
the weekly wage of a male agricultural laborer.189 High demand for
females in cottage industry meant that female agricultural servants could
earn as much as some men in Buckinghamshire in 1813.190 Unfortu-
nately these high wages were short-lived, and the end of the wars
brought lower wages. Similarly, the invention of the spinning jenny for a
brief time allowed a woman using this machine to make more than a
man weaving cloth.191 This situation was short-lived, though, because
the spinning mule soon made the jenny obsolete.
Deborah Valenze explains the relatively low wages earned by female
spinners in the eighteenth century as the result of the assumption that
the spinners were supported by men:
Stigmatized by its association with women™s work, spinning never earned wages
commensurate with the demand for thread . . . Purchasers of spun thread,
whether middleman or manufacturer, assumed that spinners came from
households where male wages provided the primary means of support; thus they
deliberately set wages for spinning low, often in complete disregard of other
factors involved in the trade.192

However, she also presents evidence that high demand for female labor
could overcome such concerns and push women™s wages above men™s.
When the invention of the ¬‚ying shuttle caused a shortage of thread,
increasing its price, Valenze notes that, “Decrying the fact that at times

Verdon, Rural Women Workers, p. 143. See also Sharpe, Adapting to Capitalism, p. 57.
Female annual wages were £10.10s., and male annual wages were £10.10s. to
£12.12s. Verdon, Rural Women Workers, p. 50.
Valenze, The First Industrial Woman, p. 79. 192 Ibid., p. 72.
Women™s wages 133

women obtained more for their work than weavers, contemporaries
called attention to the seeming injustice of it all.”193 If it is true that
spinning wages were set by the expectation that women were supported
by male wages, then how can we explain the fact that for a time women™s
wages for spinning were higher than male wages? The high prices paid to
spinners con¬‚icted with custom, leading to complaints from contem-
poraries, but the high prices were paid anyway.
Similar con¬‚icts arose in seventeenth-century Devon, where women
lace-makers could earn 7s. per week, while male agricultural laborers
earned only 6d. to 8d. per day.194 In this case the high female wages had
important social consequences. The sex ratio of burials fell to around 75
males per 100 females as male workers migrated out of the region and
female workers migrated in. Age at marriage was high for women, higher
even than for men, and few widows remarried, leaving large numbers of
women living independently.195 High female wages did not ¬t with
cultural expectations, and caused social changes that were probably
unsettling to some, but the market prevailed. In this case, where market
valuations con¬‚icted with cultural expectations, market wages were
paid. Thus, while women™s wages were justi¬ed by appeals to religion
and women™s roles in the family, these justi¬cations do not seem to have
prevented women from earning high wages when the demand for their
services was high.

De¬nition 5: Wages are set by market forces but justi¬ed by custom
It is possible that wages were governed by market forces, but explained
and justi¬ed by appeals to custom and gender ideology. I do believe that
women™s wages were customary in this sense. Most people do not under-
stand how economic forces set prices, and tell other stories about how
prices are determined. When gasoline prices rise, conspiracy theories
emerge.196 Increases in gasoline prices prompt Congressional investi-
gations. In a 2004 poll, 77 percent of Californians thought that high
gasoline prices were due to the greed of the oil companies, and only
14 percent thought that the high prices were due to “legitimate changes

Ibid., p. 79. 194 Sharpe, “Literally Spinsters,” p. 52. 195 Ibid., pp. 49, 55.
In 2000, Congress asked the Federal Trade Commission to investigate whether high
gasoline prices were caused by illegal price-¬xing. See the “Interim Report of the
Federal Trade Commission, Midwest Gasoline Price Investigation,” July 28, 2000,
www.ftc.gov/os/2000/gasprice.htm. In May 2007, MoveOn.org asked for support of a
bill against price-gouging, noting that “As consumers suffer, the oil industry continues
to reap the windfall “ breaking pro¬t records on an almost quarterly basis. It™s
outrageous! . . . Hearings start today on H.R. 1252, a House bill that would make gas
price-gouging a federal crime, punishable by 10 years in prison.”
134 Gender, Work and Wages in Industrial Revolution Britain

in market conditions.”197 Similarly, individuals who did not understand
market forces sought alternative explanations of women™s lower wages.
Sonya Rose suggests that wages must have been set by custom because,
when asked, employers could only justify women™s wages by appealing to
what women usually earn.
Industrialists evidently paid women a customary wage rate based on their gen-
der. A study of women™s work and wages in Birmingham in 1906 reported that
“employers can usually give no other reasons for the actual wage than the fact
that such and such a ¬gure is what women usually get in Birmingham.”198

However, wages may be set by the market even if employers do not
understand how the market works. In fact, the competitive models
suggest that employers should be price takers, paying the going rate for
labor. One of the strengths of the market system is that it works even if
the individual decision-makers have very little information. I may not
understand why the price of milk is $2.94 per gallon, but that does not
prevent it from being a market price. Buyers of milk do not need to know
why the current price is $2.94 “ that™s one of the greatest strengths of the
capitalist system. The fact that these employers could not explain how
wages were determined does not contradict the claim that they were
market wages.
If people do not understand how market forces work, they can be
expected to appeal to other factors to explain the levels of wages. In
nineteenth-century Britain women™s wages were explained in terms of
custom, or in terms of their domestic role. For example, James Mitchell,
reporting on the wages of factory workers in 1834, noted the fact that
women™s wages were lower than men™s wages, and commented:
Some persons feel much regret at seeing the wages of females so low, in some
cases full grown women averaging under 6s., and comparatively few more than
8s., but perhaps such persons are wrong; and nature effects her own purposes
more wisely and more effectually than could be done by the wisest of men.
The low price of female labor makes it the most pro¬table as well as the most
agreeable occupation for a female to superintend her own domestic establish-
ment, and her low wages do not tempt her to abandon the care of her own
children. Nature thereby provides that her designs shall not be disappointed.199

Mitchell clearly was seeking some explanation of what struck him as
gross inequality. He did not have access to economic explanations of the
wage gap, and he justi¬ed women™s lower wages as a means to encourage
them to pursue domestic duties. However, that does not mean he was

Field Poll #2117, http://¬eld.com/¬eldpollonline/subscribers/.
Rose, “Gender Antagonism,” p. 197. 199 BPP 1834 (167) XIX, p. 39.
Women™s wages 135

correct in his assessment. Even if custom or ideology was invoked to
justify the wage gap, that does not mean that custom or ideology was
necessarily the cause of the wage gap. While I would agree that women™s
wages were customary in the sense that they were justi¬ed by appeals to
ideology and understood by contemporaries in these terms, I do not
believe that ideology really determined the level of women™s wages.
I do not believe that women™s wages were in¬‚exible, unresponsive to
supply and demand, or set lower than women™s productivity. I do believe
that contemporaries understood women™s lower wages through the lens
of gender ideology rather than through the lens of economic models.
However, individuals do not always understand the true causes of
phenomena they observe. If the Greeks, observing lightning, explained it
as a thunderbolt thrown by Zeus, that was part of their culture, but it
doesn™t mean that the lightning really did come from Zeus. Similarly,
nineteenth-century Britain could interpret women™s lower wages as the
result of their inferiority or dependency on men, without those being the
real causes of the low wages. I conclude that both the level of wages and
changes in wages over time were the result of market forces, thus that
women™s wages should be called market wages, even if contemporaries
did not understand them that way. Customary explanations for women™s
wage were part of the culture, but they did not cause the gender wage gap.
3 Explaining occupational sorting

Two circumstances “ permanent inferiority of strength, and occasional
loss of time in gestation and rearing of infants “ must eternally render the
average exertions of women in the race of the competition for wealth less
William Thompson, An Appeal, 18251
successful than those of men.

The previous two chapters have documented large gender differences in
wages and occupations. These differences are well known and easy to
document, but explaining why these differences occurred is a more
dif¬cult task. Most historians attribute occupational sorting by gender to
some form of discrimination. However, this conclusion is too hasty if we
have not ¬rst explored whether a non-discriminatory labor market would
produce the observed results. This chapter will present some models of
market-based occupational sorting, and will argue that in the most
competitive parts of the labor market the division of labor between the
sexes matched would have been produced by the market.
In the last chapter we saw that sex differences in wages could be
explained by differences in productivity. Of course, discrimination could
still be the cause of women™s low wages if gender discrimination con-
¬ned women to a limited number of occupations, where their prod-
uctivity was low. One way to explain the observed occupational sorting
and wage differences is the crowding hypothesis formulated by Barbara
Bergmann.2 This theory states that because women are prevented from
entering many occupations, the few occupations which they are permitted
to enter become overcrowded, lowering the marginal productivity of
workers in these occupations and thus also their wages. This model
implies that the occupational sorting is based on gender discrimination,
on barriers that prevented women from entering the male occupations.


. 5
( 14)