. 7
( 14)


distinctiveness, but also that our explanations may be ¬‚awed if we
assume that only economic rules held sway.”66 However, economic rules
do not suggest that the gender division of labor should be the same
everywhere. The gender division of labor should respond to local
demand conditions. It is not a contradiction of the competitive model if
we ¬nd that women did strenuous work in coal mines while in other
areas men did less strenuous tasks. If individuals were completely mobile
then we would expect women to migrate from regions with strength-
intensive jobs, such as mining, to regions with low-strength jobs, such as
cottage industry, but the fact that men and women usually live in family
groups limits the extent of such migration.67 All local labor markets will
contain both men and women, and economic theory predicts that within
the local market women should do the less strenuous of the local jobs.

BPP 1842 (381) XVI, p. 248. 65 BPP 1842 (380) XV, p. 37.
Sharpe, “The Female Labor Market,” p. 178.
Migration did produce some gender sorting in response to relative wages. Women would
earn high wages in the lace-making industry in seventeenth-century Devon, and the sex
ratio of burials fell to 75 males per 100 females. Sharpe, “Literally Spinsters,” pp. 49, 55.
166 Gender, Work and Wages in Industrial Revolution Britain

In fact, we do observe sorting by strength within local labor markets,
and changes in the gender division of labor that respond to local demand
conditions. In textile areas, men wove because it was more strenuous
than spinning. In mining areas, women “hurried” coal because that was
less strenuous than hewing it. The work that women did depended on
local opportunities. In regions with high demand for strength-intensive
jobs, women engaged in heavier work. In the more labor-scarce north,
women more often did “men™s work.” In 1833 Thomas Baker, rector of
Whitburn, Durham, noted: “In the Northern Counties, the Women
engage in Men™s work much more than in the Southern Districts; serving
the masons with mortar, bricks, &c. is not uncommonly done by Women
in the Towns.”68 Women and children were not employed in mining in
northern Staffordshire because they worked in the potteries instead.69
One young woman in West Yorkshire stated that she only worked in the
mines because she could not get work in textiles.70 In Ireland, no women
or girls were employed in mining, only men. The reason seem to be the
excess supply of labor. A parliamentary investigator asked the Irish
employers why they did not employ children and was told, “that as labor
was so abundant and cheap they would not be troubled with Children.”71
Women were less productive than men at agricultural work, so where
other work requiring little or no strength was available, they did less farm
work. In Eden™s 1797 State of the Poor we ¬nd that, in areas where the
pillow-lace trade ¬‚ourished, women did not work in agriculture. Eden
claimed that in Roade, Northamptonshire, “Women here are never
employed in reaping; and it is even very rare to see them milk a cow.”
The reason for this was the relatively high wage available in lace-making:
“lace-workers earn from 6d. to 1s. or 1s.2d. the day; but generally 8d. or
10d. a day.”72 Male agricultural laborers could earn from 1s. to 1s.6d. a
day in the winter, so the best lace-makers could earn as much as a
man.73 Given their lower productivity in agriculture, the women could
not have made as much in farm work as they earned making lace.
Farmers who wanted to hire women had to pay them high wages to
convince them to accept the work. In 1813 in Buckinghamshire female
servants earned £10.10s. per year, approximately the same as male

BPP 1834 (44) XXX, p. 169. 69 BPP 1842 (380) XV, p. 9.
Ann Eggley, 18, stated, “there was nought else for us to do. I have tried to get winding
to do, but could not.” Ibid., p. 75.
Ibid., p. 106. 72 Eden, State of the Poor, vol. II, p. 544.
The wages of male agricultural workers are actually from the neighboring town of
Yardley Gobion, where we ¬nd that the women also make lace and “do very little out of
doors.” Ibid., vol. II, p. 548.
Explaining occupational sorting 167

Region A






0 5
Region B







Figure 3.9 The effect of the lace-making industry on the market
wage ratio

servants, who earned between £10.10s. and £12.12s per year.74 The
result was the one we would expect: no women worked in agriculture,
even in the traditionally female agricultural jobs such as milking cows.
We can use Model B to understand how a high demand for female
labor in cottage industry can raise the wage ratio and reduce female
employment in farming. Figure 3.9 compares the r-pro¬les and equi-
librium wage ratios in two regions. In region A, agriculture, which I have
assumed to have a wage ratio of 0.6, is large, and is the marginal industry
determining the market wage ratio. Other industries with higher or lower
productivity ratios exist, and hire only men or only women. Region B, by

Verdon, Rural Women Workers, p. 50.
168 Gender, Work and Wages in Industrial Revolution Britain

contrast, has relatively high levels of employment in lace-making, an
industry where women are more productive than men. In region B the
demand for lace making is so great that it employs nearly all the women
in the labor force. This makes the industry where men and women are
equally productive the marginal occupation, and male and female wages
are equal. Comparing regions A and B, we can see that the model
predicts different incomes for employment and wages. In region A both
women and men will be employed in agriculture, and the market wage
ratio will be 0.6. In region B, however, no women will be employed
in agriculture because the market wage ratio, 1.0, is higher than the
productivity ratio. Farmers will not be willing to pay high enough wages
to entice women away from lace making, and may complain that they
can™t hire women.
The relationship between opportunities in cottage industry and the
number of women working in agriculture can be shown statistically.
Even in 1841, when the demands for lace and straw-plait were lower
than during the Napoleonic Wars, these industries still affected agri-
cultural employment. In the 1841 census there is a negative relationship
between the use of females in agriculture and the prevalence of either
lace making or straw plaiting. The correlation between the percentage of
occupied women who were in lace-making and the percentage of agri-
cultural laborers who were female was “0.27, and the correlation
between the percentage of occupied women who were in straw plaiting
and the percentage of agricultural laborers who were female was “0.17.75
Census data are not ideal because the censuses undercounted female
agricultural laborers. However, using farm account books I also ¬nd that
regions with more cottage industry were less likely to employ women in
agriculture. Counties with higher levels of female employment in cottage
industry paid women signi¬cantly higher wages and employed signi¬-
cantly fewer women. The impact of cottage industry employment on
agricultural employment was substantial. The movement from a county
with low employment in cottage industry such as Norfolk, to the county
with the greatest employment in cottage industry, Bedfordshire, resulted
in a decline in the employment of females from 10.1 to 2.8 percent of the
agricultural labor force.76 These statistical results support what Eden
told us and what Model B predicts: opportunities for higher wages in
cottage industry drew women away from agricultural employment.

There are forty-two observations. The ¬rst correlation is statistically signi¬cant at the
10 percent level, but the second is not. Data from BPP 1844 (587) XXVII.
These calculations are based on the regressions presented in Burnette, “Wages and
Employment,” p. 680.
Explaining occupational sorting 169

Differences in work patterns might be very local. While women in
general were frequently engaged in agricultural labor in Dorset, a resi-
dent of Whitchurch, Dorset, noted that “The women are much engaged
in buttoning in this village: it is with dif¬culty they can be got out at 4s.
a-week.”77 Local differences in the gender division of labor are usually
interpreted as the result of local custom. This was true in 1843, when
Mr. Vaughn wrote in his report to parliament:
Custom, too, which by perpetuation of other causes becomes a cause in itself, is
not without its authority in determining the degree and manner in which this
kind of labor is applied. In a small tract of country here spoken of, slight dif-
ferences are observable, and steadily adhered to. So about Tunbridge Wells,
women are rarely employed in opening the hills in the hop-grounds. At Maid-
stone and Farnham it is their common occupation.78

However, we should not necessarily accept Mr. Vaughn™s claim that
custom caused the variation in female employment. Tunbridge Wells
was closer to London than Maidstone; if opportunities in London drew
women away from agriculture, farmers in Tunbridge Wells may have
hired men to open the hop-hills because fewer women were available to
do agricultural work. Even if the allocation of labor seems customary to
the farmers, it may still have a basis in economic incentives. In South
Molton, Devon, employment for women in lace making and wool
manufacturing was suf¬cient to make it dif¬cult for the farmer James
Huxtable to hire women: “I can™t get so many women to work for me as
I should like, owing to the lace-making and woolen manufactory in the
neighborhood. Last spring I had to hire a man to weed corn.”79 Though
this farmer thought that weeding ought to be done by women, the labor
market determined who he actually hired. In other words, the work
women did depended on the resources and requirements of the local
labor market rather than ideology.

4. Responses to changes in technology and demand
The models presented above can also accurately predict how the gender
division of labor changes in response to new technology. The most
dramatic change was in spinning. In the eighteenth century, spinners
were generally women, and weavers were generally men. Weaving fre-
quently requires some strength (although the strength requirement
varies with the type of cloth) and hand spinning requires no strength, so
this allocation was consistent with comparative advantage. Spinning
eventually became a male occupation, but only when spinning began to

77 78 79
Women and Children in Agriculture, p. 7. Ibid., p. 133. Ibid., p. 105.
170 Gender, Work and Wages in Industrial Revolution Britain

require strength. The switch from female to male labor did not follow
the invention of new machines, or even the movement from the home to
the factory, but occurred only when strength requirements increased
enough that men were more productive than women. The ¬rst machine
to spin multiple threads was the spinning jenny. This machine was used
in the home, and was designed for use by young women. Arkwright™s
water frame was the ¬rst spinning machine to use a centralized power
source, and moved spinning into the factory. Spinners were still women.
The mule, invented in 1779, was a combination of the jenny and the
water frame and produced a better yarn. Women worked the earlier
mules, and only left the occupation when mules grew so large that
strength became an important factor in a spinner™s productivity. 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.80 The invention of the self-actor eventually removed this
strength requirement by using the power of the steam engine to move
the carriage, but by this time the male spinners had formed a union, and
explicit exclusion prevented women from returning to spinning.81 Even
when women were still working as spinners, the new machines increased
productivity and reduced the total number of spinners needed, so many
women found themselves without employment. Many of these women
moved into handloom weaving.82 Handloom weaving, an occupation
that had been almost completely male, soon had nearly as many female
workers as male workers (see Table 1.3). Thus the gender allocation of
labor between spinning and weaving was completely reversed in
response to a change in spinning technology. The ease of this transition
suggests an allocation based on the market rather than on ideology.
The entry of women into handloom weaving occurred not only in
Britain, but also in France and the United States. For France, Gay
Gullickson ¬nds that, while weavers had been uniformly male in the
eighteenth century, by 1850 women dominated the trade. By 1851 the
canton of Auffay had three times as many female weavers as male weavers,
and women dominated the weaving trade in other regions as well.83 In
New England women moved into weaving, which had been a male
occupation, in the late eighteenth century, though here the circumstances

Lazonick, “Industrial Relations and Technical Change,” p. 235. A mule of 336 spindles
was not particularly large. Beginning in the 1820s employers began to use “doubled”
mules of 500 to 600 spindles. A mule could have as many as 1000 spindles. See
Freifeld, “Technological Change and the ˜Self-Acting™ Mule,” pp. 335“6.
See Chapter 5. 82 See Collier, The Family Economy of the Working Classes, p. 3.
Gullickson, Spinners and Weavers of Auffay, p. 109, and Gullickson, “Love and Power,”
p. 218.
Explaining occupational sorting 171

of work were different for women. While the male weavers had worked as
artisans, the women worked part-time from their homes.84 The models
discussed above clearly show that a biological difference in strength,
combined with a change in technology, could have caused the observed
change. In this case the disappearance of an occupation requiring no
strength pushed women into an occupation requiring some strength, and
thus increased the gender wage gap.85 Over the course of half a century
the handloom weaving trade was transformed from a male trade to a
female trade, not as a result of changing ideology, but as a predictable
response to the substitution of mule-spinning for hand spinning.
I conclude that the division of labor between the sexes in much of the
labor market can be explained by differences in strength. Women had
less strength than men and thus a comparative advantage in tasks
requiring less strength. When technology changed the strength
requirements of different tasks, the gender division of labor changed to
re¬‚ect the new comparative advantage. Comparative advantage also
explains local variations in employment patterns. Though employers
may have had expectations about what sex should be hired for certain
jobs, they responded to market forces. The result was a division of labor
between the sexes based on comparative advantage. The only case that
seems to be an exception to this rule is laundry, which required strength
but was done by women.
Usually the dictates of market ef¬ciency matched well with socially
determined gender roles. When the two did con¬‚ict, comparative
advantage determined the division of labor more often than ideology.
While ideology seems to have prevented men from doing laundry, there
are many more cases where women did “men™s work” because men were
scarce. In agriculture farmers had de¬nite ideas of what work ought to be
done by women, but the labor market determined actual employment. In
1833 a farmer from Cornwall noted that “When married, [ women] leave
off work. Farmers in this neighborhood are obliged to employ men for
what women and children should do.”86

II. Occupational sorting and child care
Another important biological difference between men and women is the
fact that women bear children and men do not. During the Industrial

Ulrich, “Wheels, Looms and the Gender Division of Labor,” pp. 12“13.
For evidence that the gender wage gap increased between 1750 and 1850, see Burnette,
“Wages and Employment,” p. 680.
BPP 1834 (44) XXX, Newlyn East, Cornwall, p. 95.
172 Gender, Work and Wages in Industrial Revolution Britain

Revolution, fertility was high, and women spent a great deal of time
bearing and breast feeding children. As with strength, reproductive
duties led women to be concentrated in certain occupations. Work done
in the home was more convenient for women with young children to
care for, so women workers were concentrated in cottage industries.
Brenner and Ramas suggest that women™s role in child-bearing was an
important determinant of the gender division of labor because of “the
incompatibility of child care and work outside the home.”87 This claim,
though, just leads us to the question of why women were responsible
for child care. Sally Alexander points to social expectations that linked
women to child-bearing:
Women™s vulnerability as wage-workers stemmed from their child-bearing capacity
upon which “natural” foundation the sexual division of labor within the family
was based . . . a wife™s responsibility for the well-being of her husband and children
always came before her work in social production, and in a patriarchal culture,
this was seen to follow naturally from her role in biological reproduction.88

Rose claims that “Although women do have special needs in pregnancy,
childbirth, and nursing, nothing about these biological processes per se
would cause women to work at low-paying jobs in general or homework
in particular.”89 While it is true that there is no necessary relationship,
these biological processes did in¬‚uence decisions by changing relative
prices, and made homework more attractive to mothers. Women™s
monopoly on pregnancy, birth, and breast feeding, combined with their
comparative advantage in child care because of their lesser strength,
resulted in women usually being responsible for child care. Given this
responsibility, substantial child care costs encouraged women to con-
centrate more heavily in work done in the home, such as work in various
cottage industries.

A. Why mothers took responsibility for child care
It is not obvious that mothers must provide child care. Fathers are
capable of most child-care tasks (with a couple of important exceptions),
but mothers provided virtually all the child care during the Industrial
Revolution period. This section will examine why mothers, and not
fathers, were responsible for child care. There were many different
possible criteria for choosing which parent would provide the child care,
and all suggested the same answer “ the mother.

Brenner and Ramas, “Rethinking Women™s Oppression,” p. 51.
Alexander, Becoming a Woman, p. 21. 89 Rose, Limited Livelihoods, p. 98.
Explaining occupational sorting 173
Table 3.3. Age-speci¬c marital fertility (births per 1000 woman-years)

Time period 20“24 25“29 30“34 35“39 40“44 Source

1750“1799 411 338 283 234 118 a
1800“1850 427 361 318 261 162 b

Sources: a. E. A. Wrigley and R. S. Scho¬eld, The Population History of England,
1541“1871 (Cambridge: Cambridge University Press, 1981), p. 254.
b. Michael W. Flinn, The European Demographic System, 1500“1820 (Baltimore:
Johns Hopkins University Press, 1981), p. 104.

1. Biological reasons
Certain child-care tasks can be done only by the mother. The father
cannot give birth to or breast feed the child. These tasks were frequently
required of a woman because fertility was high during the Industrial
Revolution. Table 3.3 shows marital fertility for the late eighteenth and
early nineteenth centuries. For every 1000 married women between the
ages of 20 and 24, there were over 400 births in a year. A young wife could
expect to give birth every two and a half years. Because of this high
fertility, women spent a large portion of their time with infants. A married
woman in her child-bearing years could be pregnant or nursing almost
continually. Davidoff and Hall give the example of Mary Brightwen, the
wife of an Essex merchant, to illustrate the possibilities. Between the ages
of 26 and 45 Mary Brightwen bore ten children, so if she nursed each
child for one year, she was either pregnant or breast feeding for 85 percent
of the time during these years.90 Davies described the wives of laborers as
“mere nurses for ten or twelve years after marriage, being always either
with child, or having a child at the breast; consequently incapable of doing
much other work besides the necessary business of their families, such as
baking, washing, and the like.”91 Being a mother in an era of high fertility
meant many years of attention to nursing infants, and there was relatively
little scope for shifting this work to others. The period of nursing could be
shortened, but alternative foods for babies were much inferior. Given the
low quality of alternative foods for infants, a father was a poor substitute
for the mother when the child was an infant, so during the ¬rst months of
a child™s life the child-care duties fell to the mother.

2. Comparative advantage in the labor market
Since child care required less strength than most market work, women
had a comparative advantage in this type of work. Because strength was

Davidoff and Hall, Family Fortunes, pp. 337“8.
Davies, The Case of Laborers in Husbandry, p. 14.
174 Gender, Work and Wages in Industrial Revolution Britain

valued in the market, the father could typically earn a higher wage in the
market than the mother. The mother was at least as productive as the
father in providing child care, and more productive in the feeding of
infants, so the family could maximize its income by sending the father to
work in the market and assigning child-care duties to the mother. Thus,
the fact that women were assigned the responsibility of caring for chil-
dren can be seen as another case of sorting by comparative advantage.

3. Gender roles
While the allocation of work within the family may have been based on
comparative advantage, we must not discount the power of gender roles.
The expectation that a married woman would remain at home was
strong. Many women stayed at home because their husbands wished it
rather than because they chose to. In 1833 Jane Falp, a worker in a
Scottish ¬‚ax mill, gave evidence that she “has been married seven
months; her husband is going to take her away to take care of his house,
not to do anything else: she would rather stay and work at the mill.”92
(Note, however, that at the time of the interview she was working in the
mill.) Even apart from the work done in the home, having a wife who did
not work became a status good for the husband.93 Social expectations
were probably strong enough to prevent fathers from taking on the child
care in all but the most desperate circumstances.
Cases where comparative advantage and gender roles con¬‚icted
would have been relatively rare. All three forces acted in one direction “
to keep the mother in charge of the children. In general the question of
which force determined the allocation of child-care duties is unidenti-
¬ed; the result cannot really be ascribed to either comparative advantage
or gender roles, since both had a part, and both suggested the same
outcome. Only if market incentives changed can we identify whether
economic incentives or gender roles determined work patterns. Heidi
Hartmann argues that the allocation of housework in modern families is
determined by gender roles. The evidence she presents is that the allo-
cation of housework to women did not change when women™s market
work increased.94 If comparative advantage rather than gender ideology
assigned women the household tasks, then this assignment should have
changed when patterns of market work changed.

BPP 1833 (450) XX, A2, p. 4.
This was Thorstein Veblen™s argument in The Theory of the Leisure Class, ¬rst published
in 1899.
Heidi Hartmann, “The Family as the Locus of Gender, Class, and Political Struggle:
The Example of Housework,” Signs: Journal of Women in Culture and Society 6 (1981),
pp. 336“94.
Explaining occupational sorting 175

During the Industrial Revolution, comparative advantage sometimes
overruled gender roles, though only in extreme circumstances. In the
few cases where men did provide the child care, they did so only because
they could not ¬nd work and their wives could. In Lancashire women
had good opportunities for work, and many supported an unemployed
husband. As mentioned earlier, Engels tells a story of a man unable to
¬nd employment who takes on the home production tasks:
[a] working-man, being on tramp, came to St. Helens, in Lancashire, and there
looked up an old friend. He found him in a miserable, damp cellar, scarcely
furnished; and when my poor friend went in, there sat Jack near the ¬re, and what
did he, think you? why he sat and mended his wife™s stockings with the bodkin; as
soon as he saw his old friend at the door-post, he tried to hide them. But Joe, that
is my friend™s name, had seen it, and said: “Jack, what the devil art thou doing?
Where is the missus? Why, is that thy work?” and poor Jack was ashamed and
said: “No, I know that this is not my work, but my poor missus is i™ th™ factory; she
has to leave at half-past ¬ve and works till eight at night, and then she is so
knocked up that she cannot do aught when she gets home, so I have to do
everything for her what I can, for I have no work, nor had any for more nor three
years . . . There is work enough for women folks and childer hereabouts, but none for
men; thou mayest sooner ¬nd a hundred pound in the road than work for men . . . when
I got married I had work plenty . . . and Mary need not go out to work. I could
work for the two of us; but now the world is upside down. Mary has to work and I
have to stop at home, mind the childer, sweep and wash, bake and mend” . . .
And then Jack began to cry again, and he wished he had never married.95

Clearly the only reason that this father took care of the house and
children was that he had no work opportunities. The passage suggests
that the situation was ideologically unacceptable. We are told that “the
world is upside down.” Engels claims that in Manchester “many hun-
dred such men could be cited, condemned to domestic occupation.”96
While market forces were obeyed in this case, a smaller difference
in market opportunities may not have been enough to overcome the
dictates of gender roles. It took a very large difference in market prod-
uctivities to alter the gender allocation of housework. Gender roles could
be overruled, but only in extreme circumstances.

B. Why women worked in cottage industry
One reason that women were so heavily concentrated in cottage industry
was the fact that such work was more convenient for women, given their

Frederick Engels, Condition of the Working Class in England in 1844 (London: George
Allen and Unwin, 1926), pp. 145“6.
Ibid., p. 144.
176 Gender, Work and Wages in Industrial Revolution Britain

domestic duties. Whether the allocation of housework was based on
comparative advantage or gender roles, the sorting of women into cot-
tage industries was an ef¬cient market response to that allocation. Of
course, a preference for work done in the home only reinforced the
allocation of women into cottage industry, since these occupations also
required little strength, and thus would have been chosen by women for
that reason even without the convenient location. However, location in
the home does seem to have been one of the forces encouraging women
to choose cottage industry.
To see the advantage of work done in the home, imagine being a
mother who is deciding whether to work, and if so where. You will
choose the best option available to you, given your constraints (which
are many). Your husband (if present) works in the market; as we have
seen in the previous section, both ideology and comparative advantage
suggest that the father will specialize in market work. If you choose to
work outside the home, you must pay for child care, which is expensive.
Although young children were occasionally left home alone, this was
done only rarely because the dangers were real. The Times reports this
incident in 1819:
A shocking accident occurred at Llandidno, near Conway, on Tuesday night,
during the absence of a miner and his wife, who had gone to attend a methodist
meeting, and locked the house door, leaving two children within; the house by
some means took ¬re, and was, together with the unfortunate children, con-
sumed to ashes; the eldest only four years old!97

Mary Hunt, a laborer from Studley, Wiltshire, seems to have worried a
great deal about the safety of her children when she went out to work:

I have always left my children to themselves, and, God be praised! nothing has
ever happened to them, though I have thought it dangerous. I have many a time
come home, and have thought it a mercy to ¬nd nothing has happened to them.
It would be much better if mothers could be at home, but they must work. Bad
accidents often happen.98

Occasionally children who were left alone died. A farm bailiff from
Wiltshire noted that “I know of two or three cases of deaths from burning
of children, since I have been in the neighborhood.”99 Evidently children
were left alone often enough to result in occasional deaths, but the dan-
gers were great enough that some women did not consider it an option.
Mrs. Smart from Calne, Wiltshire, reported: “Sometimes I have had my

The Times, February 6, 1819. 98 Women and Children in Agriculture, p. 68.
Evidence of Mr. Henry Phelps, agent of the Marquis of Lansdowne, ibid., p. 63.
Explaining occupational sorting 177

mother, and sometimes my sister, to take care of the children, or I could
not have gone out.”100 If leaving your children alone was too dangerous,
another form of child care must be found.
Sometimes mothers would take their infants to work when working
outside the home. A 1739 poem by Mary Collier, “The Woman™s
Labor,” suggests that women workers carried their babies into the ¬elds:

Our tender Babes into the Field we bear,
And wrap them in our Cloaths to keep them warm,
While round about we gather up the Corn;
When Night comes on, unto our Home we go,
Our Corn we carry, and our Infant too.101

An 1835 Poor Law report found that in Sussex, “the custom of the
mother of a family carrying her infant with her in its cradle into the ¬eld,
rather than lose the opportunity of adding her earnings to the general
stock, though partially practiced before, is becoming very much more
general now.”102 This option, however, would not have been available
for all mothers.
Sometimes relatives were available to provide child care. Grand-
mothers sometimes took care of children. Elizabeth Leadbeater, who
worked for a Birmingham brass founder, worked while she was nursing
and had her mother look after the infant.103 More commonly, though,
older siblings provided the child care. “Older siblings” generally meant
children of 9 or 10 years old, and included boys as well as girls.104 In a
family from Presteign, Wales, containing children aged 9, 7, 5, 3, and 1,
we ¬nd that “The oldest children nurse the youngest.”105 When asked
what income a laborer™s wife and children could earn, some respondents
to the 1833 “Rural Queries” assumed that the eldest child would take
care of the others, leaving the mother free to work. The returns from
Bengeworth, Worcester, report that “if the Mother goes to ¬eld work,
the eldest Child had need to stay at home, to tend the younger branches
of the Family.”106 Ewhurst, Surrey, reported that “If the Mother were
employed, the elder Children at home would probably be required to
attend to the younger Children.”107 This type of child care was not free.

Ibid., p. 65. 101 Collier, “The Woman™s Labor,” p. 10.
BPP 1835 vol. XXIX, p. 221, quoted in Pinchbeck, Women Workers, p. 85.
BPP 1843 (431) XIV, p. 710.
Mrs. Britton of Calne, Wiltshire, left her children in the care of her eldest boy. Women
and Children in Agriculture, p. 66.
Eden, State of the Poor, vol. III, p. 904. 106 BPP 1834 (44) XXX, p. 593.
Ibid., p. 488.
178 Gender, Work and Wages in Industrial Revolution Britain

Although siblings had no direct cost, they did have an opportunity cost
equal to either the market wage, or the value of the education given up.
Children under 10 were employed during the Industrial Revolution, not
only in factories, but also in many other industries. If nothing else, a
child could be hired out to look after a neighbor™s children. Only if
market work was unavailable for that child would the cost be zero.
When relatives were not available, a mother could hire child care, but
such care was expensive. As today, paid child care was available both
inside and outside the home. If a daughter was not available to look after
the younger children, someone else™s daughter, or a neighbor, could be
hired. In 1843 Charlotte Clark, who worked at a paper factory in Kent,
claimed that “The married women who have families, if some of the
children be not old enough to look after the younger, hire someone for
the purpose.”108 Elizabeth Leadbeater of Birmingham claimed that “It is
a common custom for infants to be fed by the hand whilst their mothers
are at the shop; they are under the charge of either young girls, 7, 8, or
9 years old, or put out to some neighbors.”109 A Scottish mother sent
her children to the home of another woman for 2s.6d. per week: “I used
to take them to her house at 4 o™clock in the morning, out of their own
beds, to put them into hers.”110
The closest counterpart of today™s day-care center was the dame
school. Women often took in a number of children and formed a small
school. The quality of the instruction was not high, and many considered
these schools to be, in fact, child-care arrangements. Critics of dame
schools who focus on how little children learned there are ignoring the
other important function of these schools. Dame schools thrived because
they provided day care, along with some minimal instruction. In 1840 an
observer of Spital¬elds noted, “In this neighborhood, where the women
as well as the men are employed in the manufacture of silk, many children
are sent to small schools, not for instruction, but to be taken care of whilst
their mothers are at work.”111 Most likely, the dame schools served both
purposes: child care and some very basic education. While they were
located in homes rather than in specialized buildings, they seem to have
been much like day-care centers. In areas where lace-making or straw-
plaiting thrived, young children were often sent to “schools” where they
learned the trade. At one straw-plaiting school in Hertfordshire:
Children commence learning the trade about seven years old: parents pay 3d.
a-week for each child, and for this they are taught the trade and taught to

108 109
BPP 1843 (431) XIV, p. 20. Ibid., p. 710.
110 111
BPP 1844 (592) XVI, p. 6. BPP 1840 (43) XXIII, p. 261.
Explaining occupational sorting 179

read. The mistress employs about from 15 to 20 at work in a room; the
parents get the pro¬ts of the children™s labor.112

At these schools there was very little instruction; some time was devoted
to teaching the children to read, but they spent most of their time
working. The standard rate of 3d. per week seems to have been paid for
supervision of the children rather than for the instruction.
Mothers might use a combination of different types of child care.
Elizabeth Wells, who worked in a Leicester worsted factory, had ¬ve
children, ages 10, 8, 6, 2, and four months. The eldest, a daughter,
stayed home to tend the house and care for the infant. The second child
worked, and the 6-year-old and the 2-year-old were sent to “an infant
school.”113 In response to a question about shorter hours, she
responded, “I should like to work less, though I made less; that is,
because I have so many little children at home.”114 Mary Wright, an
“over-looker” in the rag-cutting room of a Buckinghamshire paper fac-
tory, had ¬ve children. The eldest worked in the rag-cutting room with
her, the youngest was cared for at home, and the middle three were sent
to a school; “for taking care of an infant she pays 1s.6d. a-week, and 3d.
a-week for the three others. They go to a school, where they are taken
care of and taught to read.”115
The cost of hired child care was substantial. Davies quotes the price at
1s. a week, which was about a quarter of a woman™s weekly earnings in
agriculture.116 In 1843 parliamentary investigator Alfred Austin reports,
“Where a girl is hired to take care of children, she is paid about 9d. a
week, and has her food besides, which is a serious deduction from the
wages of the woman at work.”117 Agricultural wages in the area were
8d. per day, so even without the cost of food, the cost of child care was
about one-¬fth of a woman™s wage. One Scottish woman earned 7s. per
week in a coal mine and paid 2s.6d., or 36 percent of her income, for the
care of her children.118 In 1843 Mary Wright, an “over-looker” at a
Buckinghamshire paper factory, paid even more for child care; she told
parliamentary investigators that “for taking care of an infant she pays
1s.6d. a-week, and 3d. a-week for three others.”119 She earned 10s.6d.
per week, so her total child-care payments were 21 percent of her

BPP 1843 (431) XIV, p. 64. 113 BPP 1833 (450) XX, C1 p. 33. 114 Ibid., p. 369.
BPP 1843 (431) XIV, p. 46.
Davies, The Case of Laborers in Husbandry, p. 14. Agricultural wages for this time
period are found in Eden, State of the Poor.
Women and Children in Agriculture, p. 26. 118 BPP 1844 (592) XVI, p. 6.
BPP 1843 (431) XIV, p. 46.
180 Gender, Work and Wages in Industrial Revolution Britain

wage.120 Engels put the cost of child care at 1s. or 18d. a week.121
Factory workers often made 7s. a week, so again these women may have
paid around one-¬fth of their earnings for child care. Some estimates of
child care costs as a portion of the mother™s earnings were even higher.
The overseer of Wisbech, Cambridgeshire, suggests a higher fraction; he
reports, “The earnings of the Wife we consider comparatively small, in
cases where she has a large family to attend to; if she has one or two
children, she has to pay half, or perhaps more of her earnings for a
person to take care of them.”122 A woman hired for child care had to be
paid the going wage; as Ellspee Thomson, who worked in a Scottish coal
mine, claimed, “neighbors, if they keep the children, they require as
much as women sometimes earn.”123 Child care in dame schools, which
took older children, was less expensive. In 1840 the wife of a Gloucester
weaver earned 2s. a week from running a school; she had twelve students
and charged each 2d. a week.124 In 1843 the lace-making schools of the
midlands generally charged 3d. per week.125
If you were an Industrial Revolution mother deciding whether to work
outside the home, you would have considered the high costs of child
care. Working would have produced a wage, but net earnings from such
work would be substantially reduced by the cost of child care. Against
this lower wage must be set the value of time spent at home. More
ef¬cient household management could save money, and in the end you
might have decided that working outside the home would not increase
the family™s welfare. Mrs. Sumbler of Calne, Wiltshire, told a parlia-
mentary investigator that she thought the net bene¬t of agricultural work
was very small:
I do not think a great deal is got by a mother of a family going out to work;
perhaps she has to hire a girl to look after the children, and there is a great waste
of victuals and spoiling of things; and then working in the ¬elds makes people eat
so much more. I know it was so with me always. I often say there is not
fourpence got in the year by my working out.126

Having estimated the net bene¬t of working at only 4d., it is not surprising
to hear her say that “Sometimes the children have prevented my going

Assuming she paid 3d. per week for each of the middle three children, she paid 2s.6d.
per week. I make this assumption because the going rate for schools was 3d. per week.
Engels, Condition of the Working Class in England, p. 143.
BPP 1834 (44) XXX, p. 76. Malcolmson reports a case where a woman earning 11s.
paid 4s.8d. for child care (42 percent of her wage). Malcolmson, English Laundresses,
p. 36.
BPP 1842 (381) XVI, p. 450. 124 BPP 1840 (220) XXIV, p. 419.
BPP 1843 (431) XIV, pp. 46, 64, 71, 72.
Women and Children in Agriculture, pp. 67“8.
Explaining occupational sorting 181
Table 3.4. Women™s wages in cottage industry compared to wages in other industries, 1833
(earnings per week)

County industry Agriculture Factory work Src

Bedfordshire Lace 1s.6d.“3s.
Berkshire 4s.“5s.
Buckinghamshire Lace 2s. 4s.“5s.
Straw 2s.6d.
Derbyshire Embroidery 2s.“7s. 5s. Cotton 6s.“7s. a
Silk 6s.7d. b
Lace 8s.3d. b
Devonshire 3s.“4s. Lace 5s.4d. b
Essex Straw 1s.“5s. 5s.“6s.
Gloucester Stockings 2s.“4s. 4s. 6s.“7s. a
Wool 5s.6d. b
Herefordshire Gloves 3s.“4s. 3s.“4s.
Hertfordshire Straw 1s.6d.“3s.6d. 4s.“6s.
Lancashire Handloom 4s. 5s.“6s. Cotton 8s.7d. b
Leicestershire Lace 3s. 4s.“5s. Worsted 7s.8d. c
Northamptonshire Lace 1s.6d.“3s.6d. 3s.“4s.
Nottinghamshire Lace-running 2s.“4s. 4s.“6s. Lace 8s. c
Oxfordshire Lace 2s.“3s. 3s.“4s.
Somerset 4s. Silk 2s.6d.“4s. a
Westmoreland Knitting 1s. 6s.“9s.
Wiltshire Lace 2s.“3s. 4s.
Worcestershire Gloves 2s. 4s.“5s.
Yorkshire, West Handloom 2s.6d.“3s. 5s. Wool 5s.“6s. a
Riding weaving
Wool 7s.2d. b
Flax 6s.4d. b
Flax 6s.4d. c

Sources: All wages for cottage industry and agriculture are from the “Rural Queries,” BPP
1834 (44) XXX. Factory wages are from various sources, as indicated in the column
labeled “Src.”
a. BPP 1834 (44) XXX.
b. BPP 1834 (167) XIX, pp. 291“5. Average wages of workers age 21 and older.
c. BPP 1833 (450) XX. Average wages of workers age 21 and older.

out.” If work outside the home did not provide much net bene¬t, work
that could be done in the home may have been an attractive option.
The high cost of child care helps to explain why women worked in
cottage industry, even though the wages were lower than in other
occupations. Table 3.4 compares women™s wages in cottage industry to
women™s wages in agriculture and factory work, and shows that wages in
182 Gender, Work and Wages in Industrial Revolution Britain

cottage industry were lower than wages in either agriculture or factory
work. In Devonshire, women working in lace factories averaged 5s.4d.
per week, but those making lace by hand at home made only 6d. per day,
or 3s. per week if they worked six days. At the same time that women
were earning 3s. per week in cottage industry in Leicestershire, women
were earning more than twice that sum in worsted factories. Maria
Hook, 16, earned 7s.6d. per week, and Elizabeth Wells, 35, earned
10s.6d. per week as worsted spinners.127
In spite of the low wages, women exhibited a preference for work done
in the home. Brenner and Ramas claim that “Many jobs that are ˜wo-
men™s work,™ such as charring and dressmaking, were taken up because
they could more easily be combined with family responsibilities than
factory work.”128 When framework knitters worked in their homes,
women most often seamed stockings, but they would also work on the
knitting frames. When knitting frames moved from the home into
workshops in the mid-nineteenth century, women became even more
concentrated in the seaming of stockings, since this work remained in
the home. Osterud found that, in both the hosiery industry and the boot
and shoe industry of Leicester, women worked in their homes, while
men went out to work in workshops.129
Why did women continue to work in cottage industry if they could
earn twice as much in the factories? Pinchbeck observed the preference
that women showed for cottage industry “in spite of the great and
obvious disadvantages,” but she interpreted this as a dislike of factory
discipline, claiming that “domestic workers regarded discipline and
regularity with horror.”130 This suggests that women had very strong
preferences for avoiding factories. While the dislike of factory discipline
is probably part of the story, I do not think it is the whole story. Women
avoided factory discipline, not only because they disliked it, but also
because it made child care much more dif¬cult. Of course, many factors
caused women to choose cottage industry. Cottage industries such as
lace making and straw plaiting were particularly suited to women™s
skills; they required no strength, and they took advantage of a type of
skill that women in particular developed, dexterity. Factory employment
and cottage industry were available in speci¬c locations, and women
often did not have both available where they lived. In addition to these
reasons, though, I suggest that women chose to work in cottage industry,

BPP 1833 (450) pp. XX, pp. 358, 369.
Brenner and Ramas, “Rethinking Women™s Oppression,” p. 58.
Osterud, “Gender Divisions and the Organization of Work.”
Pinchbeck, Women Workers, p. 237.
Explaining occupational sorting 183

in spite of the low wages they received there, because child-care costs
were low.
One factor attracting women to cottage industry was the ¬‚exibility of
the work, and the opportunity to work part-time. A portion of the wage
gap between factory work and cottage industry likely results from a
difference in the number of hours worked. Unfortunately we do not
know how many hours per day, or per week, women worked in cottage
industry. In 1840 women handloom weavers who worked at home
averaged only 4s. per week, compared to 5s.5d. earned by women
working on the same machines in weaving shops. Parliamentary inves-
tigator James Mitchell attributed this difference to the fact that women
working at home spent some of their time in household production
tasks: “The lower average of the wages of the women working at home
is, in a considerable degree, attributable to the circumstance that many
of them are married women, and their time is partly occupied by their
domestic duties.”131 The fact that hours of work in cottage industry
were ¬‚exible, and thus work was more easily combined with household
responsibilities, may have been one of the things that attracted women to
these industries, even if the hourly wages were lower. Even today women
seem to be willing to accept a lower wage for the privilege of working
Another factor attracting women to cottage industry, in spite of the
low wages, was the location of work. Since the work was done in the
home, the mother could be physically present with the children. Part
of child care is simply being present to prevent accidents, the fear of
which prevented most women from leaving their children home alone.
Of course, this is not all there is to child care. As noted in the previous
paragraph, women working at home seem to have worked fewer hours
because of their household responsibilities. One method women used to
increase the hours they could devote to work was to sedate their chil-
dren. Mothers sometimes used laudanum, or other opiate mixtures such
as Godfrey™s cordial, to quiet their babies and give them time to work.
Sarah Johnson of Nottingham claimed that she:
Knows it is quite a common custom for mothers to give Godfrey™s and the
Anodyne cordial to their infants, “it is quite too common.” It is given to infants
at the breast; it is not given because the child is ill, but “to compose it to rest, to

BPP 1840 (43) XXIII, p. 157.
Controlling for observable differences, British women working part-time earned
6 percent less than similar women working full-time. Heather Joshi and Pierella Paci,
Unequal Pay for Women and Men: Evidence from the British Birth Cohort Studies
(Cambridge, Mass.: MIT Press, 1998), pp. 93“4.
184 Gender, Work and Wages in Industrial Revolution Britain

sleep it,” so that the mother may get to work. “Has seen an infant lay asleep on
its mother™s lap whilst at the lace-frame for six or eight hours at a time.” This has
been from the effects of the cordial.133

This method of coping with child care, though harmful to the children,
had low immediate costs for the mother. The cost of laudanum was only
about 2d. per week. Mary Colton, a lace worker from Nottingham,
described her use of the drug:
Was con¬ned of an illegitimate child in November, 1839. When the child was a
week old she gave it a half teaspoonful of Godfrey™s twice a-day. She could not
afford to pay for the nursing of the child, and so gave it Godfrey™s to keep it
quiet, that she might not be interrupted at the lace piece; she gradually increased
the quantity by a drop or two at a time until it reached a teaspoonful; when the
infant was four months old it was so “wankle” and thin that folks persuaded her
to give it laudanum to bring it on, as it did other children. A halfpenny worth,
which was about a teaspoonful and three-quarters, was given in two days;
continued to give her this quantity since February, 1840, until this last past
(1841), and then reduced the quantity. She now buys a halfpenny worth of
laudanum and a halfpenny worth of Godfrey™s mixed, which lasts her three days.

While aware that this method of child care was not good for her child,
she used the laudanum because it was the only way that she could work
the thirteen-hour days necessary to earn a living:
If it had not been for her having to sit so close to work she would never have
given the child Godfrey™s. She has tried to break it off many times but cannot, for
if she did, she should not have anything to eat.134

Whatever its effects on health, laudanum was a cheap form of child care,
and allowed the mother to work without paying the high cost of hiring
someone else to take care of the child. Cottage industry, combined with
opiates, allowed mothers to earn a (somewhat lower) market wage
without paying the relatively high costs to hire child care.
The fact that women earned low wages in domestic industry does not
necessarily mean they were prevented from entering other occupations.
If the ability to work at home was an advantage to a woman, then the
supply curve for work in the home would lie below the supply curve for
work outside the home. If the demand was the same for both types of

BPP 1843 (430) XIV, p. 613. This mother was probably not an anomaly. Robert W.
Fogel, “Nutrition and the Decline in Mortality since 1700: Some Preliminary
Findings,” in S. Engerman and R. Gallman, eds., Long-Term Factors in American
Economic Growth (Chicago: University of Chicago Press, 1986), p. 507, claims that,
“the administration of opiates to infants also appears to have been widespread for
some stretches of time.”
BPP 1843 (431) XIV, p. 630.
Explaining occupational sorting 185

labor, then work done in the home would pay a lower wage. Thus the
low wages earned in cottage industry do not necessarily indicate a seg-
mented labor market. One of the many reasons that women were con-
centrated in cottage industry was the fact that this work could be more
easily combined with child care than could work outside the home.
Given their child-care duties, many women found that even at lower
wages they could do better working in domestic industry, since they did
not have to hire child care.

In this chapter I have presented two reasons why women and men
would be sorted into different occupations even in a competitive mar-
ket “ strength and child-care duties “ thus showing that differences in
occupations by sex do not necessarily indicate the presence of labor
market discrimination against women. I have also shown that sorting by
comparative advantage reduced, rather than increased, the gender wage
gap. The following chapter presents statistical evidence that, in the more
competitive portions of the labor market such as agriculture, occupa-
tional sorting responded to labor market incentives. Readers less
comfortable with statistics can skip Chapter 4 without losing the train of
the argument.
4 Testing for occupational barriers in

The last chapter presented two models of occupational sorting by
strength and argued that, in the more competitive portions of the labor
market, the division of labor between the sexes can be explained by
comparative advantage. This chapter will present statistical tests for
occupational barriers in agriculture. Both discrimination and free mar-
kets would imply that men and women worked different jobs. This
chapter will attempt to distinguish between those two theories by testing
whether the labor market was segmented or integrated. If customary
barriers kept women in certain types of work, then changes in the supply
of or demand for their labor would not in¬‚uence the wage or employ-
ment of male workers. However, if the division of labor was the result of
the sorting models presented in the last chapter, then men and women
should be substitutable, and changes in the wages of one sex should
affect the labor market opportunities for the other sex.
Having established that sex differences in wages and occupations are
not proof of occupational segregation constraints, I will now provide
evidence that the agricultural labor market did not have discriminatory
occupational constraints. The results in this chapter will show that men
and women were hired in an integrated labor market; employers were
willing to substitute men and women workers in response to wage
changes. As we shall see in later chapters, these results do not apply to
the entire labor market. The evidence provided here is for agriculture
only. However, the agricultural sector was still large, employing more
than a quarter of employed males and somewhere between 4 and 33
percent of employed females in 1841 (see Table 1.2). While I do not
have the same data for other industries, I believe the results presented
here for agriculture would apply to some other sectors where the labor
market was competitive. However, the results would not apply to all
sectors of the labor market. There were parts of the labor market where
barriers to competition were erected, and men and women could not
move freely into any occupation they wished. The portions of the labor
market with occupational barriers are examined in Chapters 5 and 6.

Testing for occupational barriers in agriculture 187

The tests in this chapter provide a useful complement to the discus-
sion of wages in Chapter 2. Two very different ways that labor market
discrimination can occur are wage discrimination and occupational
crowding. Wage discrimination occurs if an individual is paid a wage less
than his or her marginal productivity. Occupational crowding occurs if
women are prevented from moving from low-paying occupations to
high-paying occupations, where they could be more productive, by
discriminatory occupational barriers.1 These forms of discrimination
operate independently, and showing that women™s wages were fair
market wages requires showing that neither form of discrimination
prevailed. In Chapter 2 I argued that women were paid market wages in
the sense that their wages matched their productivity, that is, that there
was no wage discrimination. However, even if women were paid wages
equal to their marginal product they may still have suffered from dis-
crimination in the form of occupational crowding. In the crowding
model women are paid wages equal to their marginal product, but,
because of the diminishing marginal product of labor, con¬ning women
to a small number of jobs results in “overcrowding” and reduces their
marginal product. This chapter tests for discrimination in the form of
occupational crowding. A gender division of labor is not suf¬cient evi-
dence of crowding because, as Chapter 3 argued, a free market that
sorted individuals into occupations ef¬ciently would also have produced
occupational differences. This chapter will attempt to distinguish
between the two possible causes of gender differences in occupations
(occupational barriers and ef¬cient sorting) by examining whether men
and women were substitutable. If there were rigid employment con-
straints, that would imply that men and women were not substitutable.
If men and women were substitutes, this suggests a division of labor
based on the models presented in Chapter 3, rather than a division of
labor based on rigidly de¬ned gender roles.
Historians of the British Industrial Revolution rely on con¬‚icting and
largely untested claims about whether men and women were substitut-
able. For example, Lindert and Williamson justify excluding women™s
wages from their wage index by claiming that women were substitutes
for unskilled men, and that their wages must therefore have followed the
same path.2 Other historians, however, note the extensive differences

The wage discrimination model was developed by Becker, and the occupational
crowding model by Bergmann. Becker, The Economics of Discrimination; Bergmann,
“The Effect on White Incomes of Discrimination” and “Occupational Segregation.” See
Appendix 4.1 for a mathematical description of these models.
Peter Lindert and Jeffrey Williamson, “English Workers™ Living Standards during the
Industrial Revolution: A New Look,” Economic History Review 36 (1983), p. 17.
188 Gender, Work and Wages in Industrial Revolution Britain

between men and women in wages and occupations and conclude that
the labor market was segmented by gender. According to this thesis,
women were con¬ned to occupations considered “women™s work.”
Deborah Valenze claims generally that women were “squeezed out of
mainstream industrial production and con¬ned to low-paid, exploitative
occupations.”3 For agriculture speci¬cally, Duncan Bythell claims that
among farm servants “there was a clear distinction between the work
done by the two sexes, with girls generally con¬ned to work in the dairy
and poultry-yard, to weeding in the ¬elds, and to household tasks”
[italics added].4 In the last chapter we saw that the allocation of labor
between the sexes was consistent with comparative advantage, and thus
with a non-discriminatory labor market. It remains to be shown, how-
ever, whether the unskilled labor market was characterized by occupa-
tional segregation constraints. This chapter presents statistical evidence
that in the unskilled labor market women were not con¬ned to a few
occupations because of their sex.

I. Cross-price elasticities
The ¬rst test of occupational constraints I use is the cross-price elasticity
between male and female farm servants.5 This elasticity measures the
extent to which the employment of male servants responds to changes in
the price of female servants, or the extent to which the employment of
female servants responds to changes in the price of male servants. If men
and women were substitutes, then employers would respond to an
increase in the price of men by substituting women for men “ hiring
fewer men and more women. Employers should also respond to an
increase in the price of women by hiring more men and fewer women.
Both situations would produce a positive cross-price elasticity (the
employment of one group increases when the price of the other group
increases). Alternatively, if tasks were strictly assigned by gender, then
an employer facing a higher price for male workers might economize on
the use of men, but would not hire women in their place, and would not
be expected to hire more women. If men and women were not substi-
tutable, we would expect to ¬nd no relationship between the employ-
ment of one group and the price of the other group, and the cross-price

Valenze, The First Industrial Woman, p. 4.
Duncan Bythell, “Women in the Work Force,” in Patrick O™Brien and Roland Quinault,
eds., The Industrial Revolution in British Society (Cambridge: Cambridge University Press,
1993), p. 39.
Appendix 4.2 describes this test mathematically.
Testing for occupational barriers in agriculture 189

elasticity would be zero. I ¬nd positive cross-price elasticities, indicating
that male and female workers were substitutes. This suggests that the
division of labor was not ¬xed by gender constraints, but responded to
changes in prices. If, in addition, British women were paid their mar-
ginal product, then the agricultural labor market in Industrial Revolu-
tion Britain was not discriminatory.
While many studies of the modern labor market have examined
substitutability between different types of labor, these studies have not
generally been used to examine occupational constraints. Most studies
have used the elasticity of complementarity, which measures how wages
respond to exogenous changes in employment. The elasticity of com-
plementarity provides an explanation of how changes in the demo-
graphic structure of the workforce have changed wages. By calculating
the elasticity of complementarity, Freeman found that young men and
older men were not good substitutes.6 He thus explained a change in the
male age“earning pro¬le during the 1970s as a response to the increase
in the number of younger workers. Grant and Hamermesh found that
youths and white women were substitutes; the more women in a city™s
labor force, the lower the wages of youths.7 They suggest that the growth
of female participation has hurt the earnings of youths. Borjas examined
substitutability between black, whites, and Hispanics by looking at how
wages responded to changes in the percentage of each group in the labor
force of the city.8 He interpreted the lack of substitutability between
blacks and Hispanics as showing that the immigration of Hispanics has
not hurt blacks. While the elasticity of complementarity has been most
popular because of the types of questions researchers were asking, a few
studies have used the elasticity of substitution, which examines the
response of employment to an exogenous change in wages.9 Welch and
Cunningham, for example, used the elasticity of substitution to examine

Richard Freeman, “The Effect of Demographic Factors on the Age“Earnings Pro¬les,”
Journal of Human Resources 14 (1979), pp. 289“318.
James Grant and Daniel Hamermesh, “Labour Market Competition among Youths,
White Women and Others,” Review of Economics and Statistics 63 (1981), pp. 354“60.
George Borjas, “The Substitutability of Black, Hispanic, and White Labor,” Economic
Inquiry 21 (1983), pp. 93“106.
Whether we expect wages to respond to employment or employment to respond to
wages depends on the level of analysis. Wages are set by supply and demand in the labor
market as a whole, but individual ¬rms, if the are small enough that they do not have
monopsony power, are price-takers and must pay the going wage. Studies using the
elasticity of complementarity examine the response of wages in a city, state, or country to
changes in labor supply. Below I will examine wages and employment at the level of an
individual farm. These farms are price-takers and thus the wage is exogenous, and the
number of workers hired at the farm responds to exogenous changes in the wage.
190 Gender, Work and Wages in Industrial Revolution Britain

the effects of the minimum wage on youth employment.10 One study used
the elasticity of substitution to study labor market segmentation. Merrilees
calculated the elasticity of substitution for four labor market groups: male
and female youths and adults.11 He found no substitutability and inter-
preted this as evidence of labor market segmentation. Like Merrilees,
I will examine the response of employment to an exogenous change in
wages and will interpret the result as a test of labor market segmentation.
In this section I will use data on English farm servants from 1768 to
1770 to test whether men and women were substitutes. Section II will
show that the conclusions in this section are valid throughout the
Industrial Revolution. While I test only agriculture, this industry was an
important one and provides an appropriate test of the claims I have
made. In keeping with the models I have outlined, agricultural work
required strength, and the labor market was competitive.

A. Data
In the late 1760s, Arthur Young set out to tour England. He believed
that scienti¬c study would improve English farming, and to that end he
collected data about the farms he visited. The data for this chapter come
from two of his books, both of which were based on extensive travels in
England. The ¬rst, A Six Months™ Tour through the North of England
(1770), is based on travels in 1768, and the second, A Farmer™s Tour
through the East of England (1771), is based on travels in 1770. At this
point in his career, Young was well known enough that he could get
landlords to cooperate with his inquiries, but he was not yet so famous
that he spent most of his touring time being entertained by gentlemen, as
he did during his 1776 Irish tour.12 Both A Six Months™ Tour and A
Farmer™s Tour present information for the towns Young visited and end
with his conclusions from the data he collected. Young was very careful
collecting his data, and while his conclusions have not held up under
further examination (his own data have been used to contradict his
conclusions), his data have fared much better. Allen used Young™s data
on particular farms to examine how employment varied with farm size

Finis Welch and James Cunningham, “Effects of Minimum Wages on the Level and
Age Composition of Youth Employment,” Review of Economics and Statistics 60 (1978),
pp. 140“5.
William Merrilees, “Labor Market Segmentation in Canada: An Econometric
Approach,” Canadian Journal of Economics 15 (1982), pp. 458“73.
Robert C. Allen and Cormac O™ Grada, “On the Road Again with Arthur Young:
English, Irish and French Agriculture during the Industrial Revolution,” Journal of
Economic History 48 (1988), pp. 100, 106.
Testing for occupational barriers in agriculture 191
Table 4.1. Descriptive statistics: Arthur Young™s data

Variable Mean SD Min. Max. N

Farm-level data
Acres 324.39 620.88 35 6000 222
Arable acres 153.76 261.99 0 2000 222
Percent arable 0.51 0.28 0 1 222
Rent 167.54 197.93 21 1500 220
Men 1.89 2.10 0 17 222
Boys 1.29 1.10 0 6 222
Women 1.44 1.04 0 6 222
Laborers 3.56 7.54 0 80 222
Milk cows 11.29 9.45 0 60 222
Other cattle 20.73 30.82 0 280 222
Horses 8.61 10.04 0 100 222
Sheep 241.63 651.48 0 8000 222

Town-level data
Men™s wage 8.53 1.47 6.0 12.0 68
Women™s wage 3.55 0.68 2.5 5.0 68
Boys™ wage 3.28 1.23 1.0 6.0 67
Laborers™ wage 19.15 3.07 13.0 26.2 67
Alternative wage 3.78 3.56 0.0 12.0 20
Spinning employment 0.71 0.45 0.0 1.0 52
High-wage employment 0.10 0.30 0.0 1.0 52

Source: Young, Northern Tour and Eastern Tour.

and found that the results closely match the results from a different data
set, which suggests that we can have con¬dence in Young™s data.13
Brunt has recently defended Young against critics, and concluded that
he was a careful researcher, and that “Although the Young data-set is
not perfect, its quality, quantity, and scope make it far superior to any
other data source available on English agriculture in the late eighteenth
and early nineteenth centuries.”14
For a number of locations, Young presented statistical descriptions
of a few farms in the area, which he labeled “particulars of farms.”
Table 4.1 provides descriptive statistics of some of the variables he
provides. Farm employees are listed in four categories. “Men,” “boys,”
and “maids” are servants who were hired annually and received room
and board from the farmer, while “laborers” received daily wages and
lived on their own. Servants were typically single, and married workers

Allen, Enclosure and the Yeoman.
Liam Brunt, “Rehabilitating Arthur Young,” Economic History Review 56 (2003),
pp. 294“5.
192 Gender, Work and Wages in Industrial Revolution Britain
Table 4.2. Distribution of farm size

Arthur Young sample Allen™s tax sample

% > 30
Size (acres) N % %

0“30 0 0 58.4
30“59 26 11.7 9.9 23.8
60“99 37 16.7 8.9 21.5
100“199 62 27.9 12.6 30.3
200“299 31 14.0 5.4 12.9
300“399 19 8.6 2.4 5.7
400“499 16 7.2 1.1 2.5
500“1000 18 8.1 1.2 2.8
1000 þ 13 5.9 0.2 0.5

The second column of Allen™s sample reports the percentage distribution among
farms greater than 30 acres.
Source: Young, Northern Tour and Eastern Tour, and Allen, Enclosure and the
Yeoman, p. 82. Allen™s sample is from tax assessments, c. 1790.

were typically laborers. My study will focus on the employment of ser-
vants because the employment of day-laborers is not well represented in
this data set. The number of “laborers” given by Young probably counts
only the male laborers who worked regularly through the year and thus
ignores female laborers, as well as male laborers whose work was more
casual and intermittent.
Although not a random sample, the farms described by Young were
broadly representative of English agriculture. While advertised as
“Northern” and “Eastern” tours, Young™s books actually cover most of
England, and Brunt ¬nds that Young did not over-sample arable
farms.15 The farms Young recorded are larger than average, though.
The average size of farms in his sample is 324 acres, while the average
farm size in a sample taken by Robert Allen from 1790 tax assessments is
119 acres.16 Table 4.2 compares the distribution of farm size in Young™s
farm sample with the actual distribution of farm size in 1790. The
smallest farm in my sample is 35 acres, so the smallest farms are missed.
The last column of the table shows that Young™s sample is much closer
to the actual distribution of farms greater than 30 acres, though he still
over-sampled larger farms. While Young™s sample omits the smallest
farms, that is not a serious problem for this study, since the smallest farms
would not have hired servants. Young did report on one 30-acre farm, but
this was eliminated from the sample because it employed no servants.

15 16
Ibid., pp. 290“1. Allen, Enclosure and the Yeoman, p. 81.
Testing for occupational barriers in agriculture 193

Young included a number of farms small enough to hire only one
servant, the smallest farms useful in a study of employment.
For each town, Young gave wages that indicate the prevailing wage in
the local labor market. I have constructed a two-tiered data set that
matches the wages for a given location with data from one or more farms
in that area. In order to arrive at one wage for each employment category,
I have averaged some wages together. Averaging is necessary because in
most cases Young recorded annual money wages for ¬ve different cat-
egories of servants: “¬rst man,” “second man,” “dairy maid,” “other
maid,” and “boy.”17 The “¬rst man” received a higher wage than the
“second man,” presumably because he had more responsibility and/or
more skill. Similarly, the wages of “dairy maids” are higher than those of
“other maids.” The wage variables used in the regressions result from
averaging together the higher and the lower wage for each sex.

B. Results
I estimate the cross-price effect in both directions, using both women™s
employment and men™s employment as dependent variables. I estimate
each equation by both OLS and Tobit, since a number of farms have
zero employment levels for either men or women (about 15 percent of
farms for each). For women, I also test whether wages were exogenous.
The explanatory variables include the wages of men and women ser-
vants, measures of farm size, and controls for the type of farming. Farm
size is measured either in acres or in rental value, and I will present
estimates using both measures of farm size. The variable “Rent”
measures the total rent paid for the farm, not rent per acre, and is a
measure of farm size that weights each acre by its value. Eighteenth-
century farmers favored rent as a measure of farm size; they usually
described farms by their rent payment rather than acres, and the poor-
rate taxes were assessed as a percentage of the farm™s rent.18
In order to control for across-task substitution and show that farmers
were willing to substitute male and female workers rather than just
adjusting the farm™s output and thus the tasks required in response to
changes in wages, I also include variables indicating the farm™s labor
needs.19 These variables are the percentage of the land that is arable and
the number of milk cows, other cattle, horses, and sheep on the farm.
Cows were used for dairying, an occupation in which women had the

In some cases, though, only one wage is given for men or for maids.
This preference may have been due to the fact that rent was more easily observable than
acres. A survey was required to measure the acreage of a farm.
See Appendix 4.2 for a discussion of within-task versus across-task substitution.
194 Gender, Work and Wages in Industrial Revolution Britain
Table 4.3. OLS and Tobit estimations: speci¬cation one

OLS Tobit

Women Men Women Men

Constant 0.5727 “0.9913 0.5209 “2.2139**
(0.4342) (0.8308) (0.4933) (0.9592)
Acres 0.0158 0.1272** 0.0128 0.1422**
(in hundreds) (0.0278) (0.0531) (0.0312) (0.0587)
Acres squared “0.0016** “0.0056** “0.0016** “0.0064**
(in 10,000s) (0.0005) (0.0010) (0.0006) (0.0011)
Percent arable “0.0622 0.4740 “0.0819 0.9875*
(0.2347) (0.4491) (0.2689) (0.5216)
Milk cows 0.0273** 0.0106 0.0294** 0.0176
(0.0065) (0.0125) (0.0074) (0.0139)
Other cattle 0.0025 “0.0075 0.0030 “0.0025
(0.0031) (0.0060) (0.0036) (0.0067)
Horses 0.0551** 0.1813** 0.0569** 0.1897**
(0.0122) (0.0233) (0.0138) (0.0258)
Sheep 0.0347 0.0641 0.0374 0.0655
(in hundreds) (0.0225) (0.0431) (0.0253) (0.0476)
Women™s wage “0.2509** 0.1424 “0.2987** 0.1723
(0.0876) (0.1676) (0.0999) (0.1896)
Men™s wage 0.1034** 0.0363 0.1178** 0.0809
(0.0392) (0.0751) (0.0446) (0.0853)
N 222 222 222 222
R 0.455 0.510
Log-likelihood “281.94 “385.23

Standard errors in parentheses.
* ¼ signi¬cantly different from zero at the 10% level
** ¼ signi¬cantly different from zero at the 5% level
Normalized coef¬cients are presented for the Tobit estimations.
Source: Young, Northern Tour and Eastern Tour.

comparative advantage, so this variable should increase the number of
women hired. Arable agriculture, on the other hand, required more
strength and was therefore better suited to male workers. Horses may be
associated with plowing, which was a male task. It is less clear how cattle
or sheep should affect the relative demand for men and women, so I do
not expect any particular effect.
The results of OLS and Tobit estimations are presented in Tables 4.3
and 4.4. The ¬rst speci¬cation, in Table 4.3, includes acres and acres
squared, allowing a non-linear relationship between employment and
farm size. The second speci¬cation, in Table 4.4, divides the dependent
variable and each farm stock variable by rent. Both speci¬cations give
Testing for occupational barriers in agriculture 195
Table 4.4. OLS and Tobit estimations: speci¬cation two

OLS Tobit

Women/rent Men/rent Women/rent Men/rent

Constant 0.5096 0.2140 0.4595 “0.1630
(0.4771) (0.5750) (0.5434) (0.6654)
Rent “0.0825** “0.0494 “0.0776** “0.0279
(in hundreds) (0.0330) (0.0398) (0.0374) (0.0457)


. 7
( 14)