Boyer George R. An Economic History of the English Poor Law, 1750-1850
Подождите немного. Документ загружается.
An Economic Model
of
the English
Poor
Law 107
seasonal layoffs were more likely
to
occur
the
larger
the
share
of the
poor rate paid
by
non-labor-hiring taxpayers.
The
value
of
e
was
deter-
mined
by the
distribution
of
property holdings
in the
parish
and was
therefore largely insulated from parish politics. Because parish authori-
ties determined property assessments, however,
the
political makeup
of
the parish could potentially have affected
the
value
of e. If
those
in
power tended
to
underassess their own property, then,
for a
given distri-
bution
of
property,
e
would have been smaller
in
parishes dominated
by
labor-hiring farmers than
in
parishes dominated
by
taxpayers
who did
not hire labor.
The size
of
poor relief benefits
was
determined
by the
parish vestry.
Recall that unemployed workers received income
d
t
,
which consisted
of
private (severance) payments from farmers
to
workers,
b
n
and
public
relief benefits.
It was
clearly
in the
interest
of
labor-hiring farmers
to
substitute public
for
private
relief,
since public relief was partly paid
for
by non-labor-hiring taxpayers. Where labor-hiring farmers dominated
parish vestries, therefore, unemployed workers should have received
only poor
relief.
Conversely, where vestries were controlled
by
taxpay-
ers
who did not
hire labor, poor relief benefits might have been
set
below
the
efficient level
of d
t
(determined from equation
(7)),
forcing
farmers
to
offer private relief payments
to
workers they laid
off
during
slack seasons.
29
Because
the
level
of
public relief benefits
was set by the
parish vestry,
the
size
of the
poor relief subsidy
and,
therefore,
the
number
of
layoffs were partly determined
by the
political makeup
of the
parish.
For any
given values
of e and x
n
the
number
of
layoffs should
have been larger
in
parishes dominated
by
labor-hiring farmers than
in
parishes dominated
by
taxpayers
who did not
hire labor.
The model shows that layoffs should never occur
if
s
= 0.
This result
may appear surprising
at
first,
but it has an
intuitive explanation.
Con-
sider
an
allocation
of
labor hours
in
which there
is
less than full employ-
ment, which yields
an
expected utility
V for
each worker
and a
profit
TT'
for
the
farmer. Suppose this allocation
of
labor hours
is
replaced
by a
new allocation
in
which
all
workers
are
employed,
and
hours
per
worker
are reduced
by
enough
to
equate total labor hours across
the two
alloca-
tions.
If
s
=
0 and the
total man-hours worked
are the
same,
the
farmer's
profits
are the
same
for
both allocations,
and he
should
be
indifferent
29
Private relief benefits typically took
the
form
of
allotments
of
land
or
potato ground
offered
by
farmers
to
their workers
at
zero
or
below-market rent.
108 An Economic History of
the
English Poor Law
between them.
30
However, because workers are risk averse, they prefer
the new allocation of labor, which assures them employment, despite the
fact that the expected income from the two contracts is the same. Thus,
full-employment contracts are always dominant if
s
= 0, that is, if labor-
hiring farmers are the only parish taxpayers (e = 1) or poor relief is not
given to able-bodied workers (g = 0).
One other result obtained from the model should be mentioned. Re-
call that the farmer is assumed to enter into an agreement (implicit
contract) with N workers, where N does not necessarily equal all the
workers in the parish (that is, N ^ 1). The size of N is determined by the
following equation
EX
t
{fMxMx,)*Mx<)
~
c<(x
t
)
+ [1 -
n
t
(x
t
)/N]z
t
(x
t
)}
- y = 0
where y is the Lagrangian multiplier associated with the constraint
TV
<
1. When N < 1, y = 0, and the optimal number of workers under
contract is determined by equating the average output from the marginal
worker to the average cost of hiring him, which is his consumption net of
the amount he is willing to pay not to be laid off times the probability of
layoff.
31
All workers in the parish will be under contract if the average
output from the marginal worker is sufficiently high.
If N < 1, the parish contained more able-bodied workers than it
needed. By law the parish was under an obligation to offer poor relief to
all persons unable to provide for themselves. But it was not necessary to
provide all persons with outdoor
relief.
It was in the ratepayers' interest
to force chronically unemployed workers either to perform onerous
tasks or to enter a workhouse in order to obtain
relief.
The purpose of
such tactics was to ensure that the expected utility of redundant workers
was less than V*, and thus to cause them to leave the parish. In this way
the parish vestry was able to maintain the resident work force at approxi-
mately the number of workers needed by labor-hiring farmers.
Up to this point I have assumed that cottage industry did not exist in
the parish. However, although wage rates and employment opportuni-
30
If s > 0, the
farmer's profits
are
larger under
the
original allocation
of
labor time,
because workers
on
layoff
are
supported
in
part
by
non-labor-hiring taxpayers.
31
Burdett
and
Wright (1989b) show that
the
number
of
workers under contract
is
nega-
tively related
to the
value
of a
firm's experience-rating factor, which
is
essentially
the
same
as e in my
model. That
is, the
firm's long-run demand
for
labor
is
affected
by the
marginal
tax
cost
of
layoffs.
In
terms
of my
model,
an
increase
in the
share
of the
poor
rate paid
by
non-labor-hiring taxpayers would increase
the
number
of
workers
the
farmer offered
a
contract,
N.
An Economic Model of
the
English Poor Law 109
ties in cottage industry declined sharply after 1780, women and children
were still employed at straw plaiting, lacemaking, and the like, in parts
of the rural south in 1832.
32
Cottage industry can be incorporated into
the model by including the utility obtained from such employment,
V{H),
in the workers' utility constraint faced by the farmer, as follows
ES
t
V
t
(x
t
) > V* - V(H)
The existence of nonagricultural employment opportunities for women
and children reduced the utility value of the implicit contract farmers
had to offer their workers to keep them from leaving the parish. Agricul-
tural wage rates or relief benefits should have been lower in parishes
with cottage industry than in parishes without cottage industry, other
things equal.
c. Summary of
the
Model's Results
The model developed in this section shows that implicit contracts con-
taining seasonal layoffs and poor relief payments to unemployed work-
ers were cost-minimizing for labor-hiring farmers only under certain
conditions. The major determinants of the form of farmers' profit-
maximizing labor contracts were the extent of seasonal fluctuations in
the demand for labor (a function of crop mix), the share of the poor
rate paid by non-labor-hiring taxpayers, and the political makeup of
the parish. In grain-producing areas, where the demand for labor fluctu-
ated sharply over the crop cycle, farmers minimized labor costs by
laying off unneeded workers during slack seasons and having them
collect poor
relief.
Full-employment contracts were cost-minimizing in
pasture-farming areas, where the demand for labor was spread rela-
tively evenly over the year.
For any given crop mix, the larger the contribution of non-labor-
hiring taxpayers to the poor rate, the lower the cost to farmers of laying
off workers and the larger the number of layoffs. The contribution of
non-labor-hiring taxpayers was determined by the distribution of prop-
erty holdings in the parish and by the size of relief benefits paid to
unemployed workers. The latter was set by the parish vestry and there-
fore was affected by the political makeup of the parish.
County-level data on per capita relief expenditures and information
32
See
Chapter
1,
Section
3.
110
An
Economic History
of
the English
Poor
Law
on regional variations
in the
length
of
labor contracts suggest that,
on
average, labor-hiring farmers responded
in a
cost-minimizing manner
to
the economic environment. Explicit yearlong labor contracts remained
widespread
in
pasture-farming areas throughout
the
Speenhamland
era,
while weekly
(or
even daily) contracts became predominant
in the
south
and east during
the
last
few
decades
of
the eighteenth century (Hasbach
1908:
176-8,
262-3,
329;
Hobsbawm
and
Rude
1968: 40,
43-4).
The
absence
of
yearlong contracts
in
grain-producing areas suggests that
farmers were indeed laying
off
laborers during certain seasons.
There
is
also
a
strong positive correlation,
at the
county level
of
aggregation, between
per
capita poor relief expenditures
and the
impor-
tance
of
grain production.
33
Figure 3.1 presents county-level data
on per
capita relief expenditures
in
1831. There
is a
clear regional aspect
to the
level
of per
capita expenditures, which were significantly higher
in the
southeast than
in the
west
or
north. County-level data
on the
share
of
farm land recorded
as
arable
and on the
ratio
of
arable
to
pasture land
in
1836
are
given
in
Figure 3.2.
The
land-use estimates were constructed
by
Kain (1986) from
the
tithe surveys conducted under the Tithe Commuta-
tion
Act of
1836.
34
The
data show
the
importance
of
arable farming
in
the southeast
and
pasture farming
in the
west.
The positive relationship between
per
capita relief expenditures
and
arable agriculture
can be
seen
by
combining
the
data from Figures
3.1
and
3.2. In the 12
counties with more than
50% of
their farm land
recorded
as
arable,
per
capita relief expenditures averaged 14.9s.
per
annum
in 1831. In the 12
counties with 35-50%
of
their farm land
arable,
per
capita relief expenditures averaged 11.4s.
per
annum.
Fi-
nally,
in the
11
counties with less than
35%
of
their farm land arable,
per
capita relief expenditures averaged 7.5s.
per
annum.
In
1821,
per
capita
relief expenditures averaged 16.1s., 13.6s.,
and 9.2s. in
counties with
more than 50%, 35-50%,
and
less than 35%
of
their farm land arable.
I
also grouped counties according
to the
ratio
of
arable
to
pasture land.
In
the
12
counties where
the
ratio
of
arable
to
pasture land was greater than
1.8, per
capita relief expenditures averaged 15.2s.
in 1831. In the 11
counties where
the
ratio
was
between
0.8 and 1.8, per
capita relief
33
This correlation was first pointed out by Blaug
(1963:
171), who noted a "striking
coincidence . . . between the spread of Speenhamland and the production of wheat."
34
Kain was able to construct land-use estimates for only 35 English counties (including
Monmouth) from the 1836 tithe surveys. Land-use estimates are not available for Mid-
dlesex, Northampton, Leicester, Nottingham, Westmorland, and Cumberland.
An Economic Model of the
English Poor
Law
111
^ expenditure > 15 s.
J
10s.
£ expenditure < 15s.
[JJ]
expenditure <10s.
Figure
3.1.
County-level data on per capita relief expenditures, 1831-2.
(Source:
Poor
relief expenditure data from Blaug
[1963:
178-9].)
expenditures averaged 10.6s. per annum. In the 12 counties where the
ratio was less than 0.8, per capita relief expenditures averaged 8.2s. per
annum.
35
The relief expenditure data used in these calculations cover all forms of parish poor
relief,
not just outdoor relief to able-bodied laborers. However, there is no reason to
believe that the differential in per capita relief expenditures between "corn and grazing
counties" would decline if the calculations were performed using only data on relief
expenditures for able-bodied laborers.
Arable as a percentage of
total county area
r
Ratio of arable to pasture
•
No data
No data
Figure 3.2. County-level data on crop mix, 1836. (Source: Kain [1986: 462].)
An Economic Model of
the English
Poor Law
5. The Effect of Migrant Labor on the Rural Labor Market
113
The model developed in Section 4 assumed that farmers only employed
workers who resided in the parish during the entire year. But profit-
maximizing farmers might have preferred to reduce the resident labor
force and hire migrant labor at harvest time rather than to maintain a
resident labor force large enough to meet peak-season labor require-
ments. The attractiveness of the former policy depended on the elastic-
ity of the supply curve of migrant labor and on the seasonal distribution
of labor requirements in agriculture.
Consider first the demand for migrant labor. The ability of farmers to
profit from an elastic supply of migrant labor depended on their farms'
crop mix, which determined the seasonal distribution of their labor re-
quirements. In the pastoral-farming west and north, the demand for
labor was fairly steady across seasons and there was little need for
migrant labor. The grain-producing south and east experienced large
seasonal fluctuations in demand for labor and therefore would appear to
have been amenable to migrant labor. Within the south and east, how-
ever, the demand for migrant labor varied across parishes, according to
the precise crop rotation in use. Figure 3.3 shows the monthly labor
requirements of two typical 500-acre farms, as calculated by C. Peter
Timmer (1969). The "old" farm is cropped according to the traditional
three-course rotation, with "120 acres wheat, 120 acres barley, 120 acres
fallow, and 140 acres in permanent grass" (Timmer 1969: 386). The
"new" farm is cropped according to the four-course "Norfolk" system, in
Old Farm
Figure 3.3. Monthly labor requirements of two 500-acre farms. (Source: Timmer [1969:
394].)
114 An Economic History of the
English Poor
Law
which the 260 acres of fallow and permanent grass are replaced by 120
acres of turnips, 120 acres of clover, and 20 acres of permanent grass.
The transition from the three-course system of cultivation to the four-
course system began on the well-drained, thin, and infertile soils of "the
home counties, East Anglia, and much of Southern England" during the
late seventeenth century (Chambers and Mingay 1966: 59). The number
of southern and eastern farms practicing the Norfolk system substan-
tially increased during the eighteenth century, although the traditional
three-course technique continued to be used in areas that received large
amounts of rainfall or had soils unamenable to the new rotation.
36
The adoption of the Norfolk system had an important effect on the
seasonal distribution of labor requirements. As seen in Figure 3.3, the
cultivation of turnips on the new farm led to two new peak periods of
labor demand, in March and June. The increased demand for labor
during the early spring made the new farm much less amenable to the
use of seasonal migrant labor than was the old farm. A farm practicing
the Norfolk system required as many laborers in March as it did in
August, but migrant labor generally was available only during the sum-
mer. Thus, a farmer whose crop mix corresponded to that of the new
farm would require a resident labor force as large as his harvest labor
force.
Farmers using the Norfolk system had the option of using migrant
labor instead of resident labor at harvest time, but any reduction in
resident laborers' harvest earnings had to be made up either in poor
relief or in higher wages during other seasons, in order to maintain
workers' expected utility at its reservation level. So long as farmers
could not use migrant labor to reduce the number of workers under
contract, it was cheaper to use resident labor rather than migrant labor
at harvest time. According to E. J. T. Collins (1976: 56)
As a general rule outsiders were not taken on until all local labour was fully
employed and custom decreed that first refusal of casual work was given to the
wives and dependents of permanent workers. Farmers were anyhow mindful of
the connexion between earnings in summer and poor relief in winter, and, thus
cautioned, were loath to interfere in what were commonly regarded as the
"rights of labor."
36
According to Chambers and Mingay (1966: 58), "the difficulty of growing roots and the
new legumes and grasses on ... the wet and cold clays provided a serious and persistent
obstacle to agricultural progress in the midland clay triangle and other districts of similar
soils.
... Of necessity, two crops and a fallow remained the basic rotation on them . . .
until cheap under-drainage came in towards the middle of the nineteenth century."
An Economic Model of
the English Poor
Law 115
The situation was very different for the farmer whose crop mix corre-
sponded to that of the old farm. The only major peak period of labor
demand on the old farm occurred at harvest time, when migrant labor
was available. Given an elastic supply of migrant labor, a farmer practic-
ing the traditional three-course rotation could minimize costs by having
a resident labor force large enough to meet labor requirements on the
farm for all months except August and September, and hiring migrant
labor during the peak season. The substitution of migrant labor for
resident labor reduced costs because the farmer did not have to support
migrant workers during slack seasons.
37
Farmers
7
use of migrant labor was determined by its availability as
well as by crop mix. There were two broad sources of seasonal migrant
labor: workers from nonagricultural (often urban) occupations who par-
ticipated in the harvest in nearby agricultural areas; and agricultural
workers from other regions where the harvest occurred at different
times.
The first source of migrant labor declined in importance during
the early nineteenth century as a result of "the decay of rural indus-
try ... and new work rhythms [of factory employment] which pre-
cluded the allocation of labour-time between field and workshop" (Col-
lins 1976: 40). The Kent hops
fields
continued to attract London workers
at harvest time, but much of the grain-producing southeast was too far
from London, and other cities, to attract harvest labor.
The second, and more important, source of migrant labor was the
agricultural sector of the " 'Celtic fringe' - the Scottish Highlands, the
Welsh hill country, and above all, western Ireland" (Collins 1976: 45).
There is evidence of Scottish and Irish harvesters in southern England in
the late eighteenth century. Their numbers declined during the French
wars (1793-1815) because of agricultural prosperity in Scotland and
Ireland, then increased sharply after the Irish famine of 1822 (Redford
1964:
142). Large-scale Irish migration into the southeast began in the
late 1820s (Kerr 1942-3: 373).
The "vast influx" of Irish labor was largely a result of agricultural
depression in Ireland and of the introduction of cheap steamship service
between Ireland and Britain. The pattern of migration
was
determined by
37
This conclusion assumes that resident workers made redundant by migrant labor left the
parish. If they remained, the parish was obliged to pay them poor
relief,
which increased
the cost of using migrant labor. However, the parish could make the conditions for
obtaining relief onerous enough to convince redundant workers to leave the parish, by
forcing them to enter a workhouse or to perform manual labor.
116 An Economic History of the
English
Poor Law
the seasonal labor requirements of both origin and destination. Laborers
left Ireland after planting their potato crops and "returned in time for the
potato-digging," so that they were in England from May through August
(Redford 1964: 147). In England they "took two or even three harvests a
season [by] exploiting the different harvest timings of upland and low-
land, north and south, clay land and light land" (Collins 1969a: 77).
What effect did the availability of seasonal migrant labor have on the
form of farmers' cost-minimizing implicit labor contracts? The model
developed in Section 4 shows that farmers should lay off workers during
season t if the marginal product of labor at full employment (n
t
= N) was
sufficiently small. Migrant labor affects the form of the cost-minimizing
contract by reducing the number of workers the farmer offers an implicit
contract, N. A decline in
TV
raises the marginal product of labor at full
employment in each season (because f
n
< 0), and therefore reduces the
number of layoffs. Under certain conditions the availability of migrant
labor will reduce
TV
by enough to eliminate seasonal layoffs altogether,
and therefore change the form of the cost-minimizing labor contract.
This will occur if the demand for labor is relatively constant across all
seasons except harvest.
The data in Figure 3.3 can be used to estimate the effect of migrant
labor on farmers' implicit contracts in parishes that practiced the three-
course system of
agriculture.
An elastic supply of migrant labor enabled a
farmer to reduce the number of workers under contract by up to
40%.
38
The reduction in N reduced the number of months
in
which farmers would
have laid off workers from perhaps 10 to between
1
and 5.
One must be careful, however, not to overstate the role played by
migrant labor. The actual decline in the number of workers under con-
tract was probably smaller than the data in Figure 3.3 suggest, because
there was added uncertainty (for farmers) associated with the use of
migrant labor. In particular, a farmer could not be sure that migrant
labor would be available when it was needed. To reduce uncertainty,
farmers entered into implicit contracts with migrant workers to ensure
their return. A contemporary wrote that Irish harvesters "mostly resort
38
Timmer's (1969: 393) calculations show that a 500-acre farm cropped under the three-
course system required approximately 50 workers in August and September, but no
more than 30 in any other month. If migrant labor was available in August and Septem-
ber, the farmer could have reduced the number of laborers under contract by 20, a
decline of 40%.