A NEW INDEX OF RESIDENTIAL CONSTRUCTION AND LONG CYCLES IN HOUSE‐BUILDING IN GREAT BRITAJN, 1838–1950
| Date | 01 February 1955 |
| Author | B. Werer |
| DOI | http://doi.org/10.1111/j.1467-9485.1955.tb00721.x |
| Published date | 01 February 1955 |
A
NEW
INDEX
OF
RESIDENTIAL
CONSTRUCTION
AND
LONG
CYCLES
IN
HOUSE-BUILDING
IN
GREAT
BRITAJN,
1838.1950
I
INTRODUCTION
ONE
of the niore iniportant gaps
in
the statistics
of
the nineteenth
cciitury and first quarter
of
the twentieth relates to house-building.
Apart froni the statistics of
one
or
two towns. and the figures obtain-
itblc
from
the
Inhabited House Duty, there is no annual information
Ixariiig directly on the subject throughout the whole
of
the Victorian
pcriod.l
'I'hc
House Duty figures. moreover, cannot be used for years
bdure
1874.
and from
1874
to
1913 they suffer from
a
number of
tlcfccts
which render their reliability problematical. From 1785
10
18-19
we liavc Shannon's brick index2 but
it
is affected by railway
conslruction and represents the whole
of
the building industry, not
house-building
alone.
As
non-residential building fornis
a
large
proportion
of
the
total
and as the various sectors
of
the industry
do
IIO!
necessarily follow the same trend
or
the same fluctuation, Shannon's
inJcx
cariiiot
be taken as
a
reliable indicator
of
house-building. even
Ihuugh
it
may be a better guide
to
the movement of all building
oulpitt combined. The
1850s
and
1860s
are
a
specially obscure period
arid
rlicrc
is
no
adequate link between
1914
and
1924,
the year from
irliicli iiiore comprehensive statistics become available. The only
srJrisrics
;ivailrrble for this period are the returns
of
building plans
passed by
I
number of urban districts and published from
1910
in the
f,rcbour
Grrsc~tlc.
They are given
in
value terms and.
as
the period
WIS
subject
10
violent fluctuations
in
prices, are especially dependent
i)ii
a11
accurate index
of
building
costs
for deflation into ternis
of
v
The deficiency
in
house-building statistics is fortunately. however.
riot
;IS
grciit
as
the absciice
of
published materials would appear
to
suggcsl.
A
good deal
of
information
on
building activity was origin-
ally
collcc~cd
by Local Authorities and sonie
of
it.
at any rate, has
in
'
1.w
IWI-16
we
have.
however.
a
very
useful
series giving
the
number
of
II~IUSC~
crcclcd
in
the
forty
largest
municipal areas in England and
Wales.
The
\tmicc
is
a
p;ciiipIilrt
entilled
The
Natiowal
Conference
on
Hoirsiicy
uflrr
/lr~,
Il'crr
:
f,'t*p[tr/
of
the
Orgawisiiig
Committee
(hlanchcster
1917).
Thc
licurc>
Jiavc
bccn
rrproduccd by Professor
A.
K.
Cairncross in
flume
uiid
'.
I
I.
A. SIi:inrion,
'
Bricks--A
Trade
Index,
1785-1849
',
E'cuwmica
(N.S.).
/.'
;!/ti /III*CS/IIICI~~,
IX7f/-I9Z.Y,
(IY53),
p.
155.
\',)I.
I
N~I.
J,
htigu~t
1934,
pp.
300-18.
A
NEW
INDEX
OF
RESIDENTIAL
CONSTRUCTION.
1838-1950
105
some form
or
other survived the salvage campaigns
of
two World
Wars.
The primary purpose of this paper
is
to present
a
new index
of
house-building activity which has been compiled from precisely such
local data. The earliest information relates to Liverpool and
goes
back
to
1838.
but statistics of a number
of
towns become available
from
the
1850s
and by
1900
the index is based on thirty-four towns.
In
1924
it
can be chained
to
an
index based on the comprehensive
national data.
There are
at
least
two
distinct types
of
fluctuation which have from
time to
time
bedevilled the course
of
house-building activity:
a
relatively short movement lasting approximately five to seven years,
and
a
movement
of
considerably longer duration and wider amplitude.
The former includes some random fluctuations but
is
on the whole best
regarded as
a
rnanifeslation of the business
cvde.
It is
a
fluctuation
which is
at
times almost completely submerged by the long cycles, and
on occasion it emerges only in variations in the rate of movement.
Even
so,
this fluctuation, and
its
relationship to the trade cycle,
is
far
from being without interest, and it has indeed aroused
a
good deal of
controversy.
It
is. however, a subject which merits treatment on
a
more
extensive scale than can be allowed in this article.
Our interest here
is
in the long fluctuations. These were noted as
early
as
1908
by Dear1e.J and Beveridge described the
'
hyper-cyclical
fluctuations of the building trade" in
1921.
But
in
spite of this early
'
discovery
'.
and in spite
of
the intrinsic interest of the subject to the
student of economic fluctuation and growth, it has not received a great
deal of attention in Great Britain. There is, in fact. no comprehensive
study of the long
'
building cycle'' relating
to
this country.' This is
N.
B.
Dearle,
Problems
of
Unemployment in the London Building
Tradee,
(London, 19qS).
Lord Beveridge.
Some
Aspects
of
Trade Fluctuation
',
Transactions
of
the
Mancheeter Stattlical Society,
Session 1920-21.
The term
'cycle is
employed in this paper without its usual connotation
implying
a
periodically recurrent fluctuation.
'The main works on the subject are: Professor
A.
K.
Cairncross. 'The
Clasgow Building Industry (1870-1914)'.
Review
of
Economic
Studies,
Vol.
11.
No.
1.
October 1934,
pp.
1-17,
and
'Home
aird
Foreign Investttcent,
1870-
19lJ',
(1953), Ch. 2 and elsewhere; and
H.
W.
Robinson,
The Economics
of
Building
(1939),
Chs.
7
and
8.
There have also been some articles by
Dr.
M.
Bowley and Professor
I,
Bowrn but !hey are concerned mainly
yilh
cyclirnl
fluctuations i? 1924-38
M.
Bowley, Fluctuatioos in House-building and the
Trade Cycle
,
Review
6/
Eroiiomic
Studien,
Vol. IV, No.
3.
June 1937, pq.
167-81;
M.
Bowley, 'Some Regional Aspects
of
the Building Boom, 1914-36
,
f2evierr
of
Erotrotttic
Slitdies,
Vol.
v,
No.
3.
June
1938.
pp. 171-186;
1.
Bowen.
Building Output and the Trade Cycle
W.K.
1924-38)
'.
0zfo1.d
Ecotiotnic.
Papers,
NO.
3. February 1940. pp.
110-30,
More work has been done on long
building cycles in the
U.S.A.
and the literature
has
Feen reviewed by
C.
E.
V.
Leser in 'Building Activity and Housing Demand
,
Yorks.
Bull.
of
Econ.
and
Social
Rcscarch,
Vol. 111,
No.
2,
July 1951, pp. 131-49.
106
B.
WEBER
a
task which cannot be undertaken here and no attempt will be made
to
advance any
'
interpretation
'
of the phenomenon. Our task
is
the
less
ambitious one
of
description and statistical definition
of
the major
I1
ucl
uat
ions.
I1
THE
STATISTICS
OF HOUSE-BUILDING
Before presenting the new index it is desirable
to
survey briefly the
niairi
alternative source materials
on
house-building before 1914. These
Jliay
be sunimansed under three heads
:
(i)
Income Tax (Schedule A) and Inhabited House Duty statistics.
(ii)
Materials from the Populalion Census.
(iii)
Approved building plans or houses erected in individual towns.
Only the last
of
these are reliable indices of residential construction,
but
thc only series published are those of
the
plans passed by the Dean
of
Guild Court of Glasgow from 1863. the number of new houses in
tllc
London Metropolitan Police District
from
1871. and the number
of
houses
erected in the forty largest municipal areas in 1901-16.'
Exccpt
for the latter the coverage is
too
limited
for
the figures
to
serve
;IS
indications
of
national building activity. However, it
is
from this
kind
of
inforniation that our index for larger areas has been constructed
and
the
scries
for
Glasgow
and London have been incorporated into
thc
iiew
index.
Frorii
the Inhabited House Duty
it
is possible to obtain statistics
of
the
annual increases
in
the housing stock from 1874-5
to
1913-4.'
These
have
generally been niade use of for the study
of
house-building
fluctuations
but they suffer from a number of defects which detract
considerably
from
their value.
'The
figurcs
are
net
quantities and a good deal of new house-
Iwildiiig
which is offset by demolitions or conversions is
lost
in the
.
totals.
If
this magnitude could be established,
or
if it could be
:issunled
that
it
formed
a
fairly constant proportion of the
total
A.
K.
Cairncross.
Home
aid
Foreign
Itrvestment,
1870-1913.
p
16.
155;
J.
C.
Spcnsley. Urban Housing Problems.'
Journal
of
the
f;oyal
.S/u/ix/icul
SocielU.
Vol.
LXXXI,
Part
11,
March 1918, p. 210..
A
series
of
a
JilTsrrnt kind, based on fees received
by
Building Surveyors in London
lor itibpscting ncw buildings un,der
the
provisions
of
the London Building
Acls,
is
given by
E.
W.
Cooney. Capital Exports, and Investment in Building
iii
Ilrilain and the U.S.A., 1856-1914
',
Econonrica
(N.S.),
Vol.
XVI,
November
1Y-i').
pp.
347-s4.
*
The
Jala
arc given in Lord
Slamp,
Rrifivh
Zticotties
atid
P.roperlv,
1916.
p.
141.
Siinilar figures. but relating only
to
houses with an annual value of
f?O
and over. are available from 1851-2. The proportion
of
the
latter
to
all
.~.l~\dling-liousrs
in 1874-5. was, however, only 12.7 per cent. Stanip has
also
an
*
wwve grncriil discussion
of
Inhabited House Duty statistics.
A
NEW
INDEX
OF
RESIDENTIAL
CONSTRUCTION,
1838-1950
107
housing stock
or
of
new house-building, this difficulty could be over-
come, Unfortunately. there is very little information on demolitions,
etc. and what estimates there are indicate considerable variations over
time which are not ostensibly related
to
fluctuations in new building
or
the stock of
houses.
There
is
the further difficulty of the re-assessment years: every
three years in England and Wales before
1888-9.
then every
five
years
(except
in
1910-1,
when the interval was seven years). and every five
years in the Metropolis.' In these years increases in rent were given
effect
to
while reductions were allowed
for
annually. This introduced
a
downward bias into the figures
of
annual value in inter-assessment
years, while the numerous appeals which followed new assessments
and often led to subsequent revision introduced
an
opposite bias in
years
of
assessment.
This
dificulty does not affect the similar statistics
given for the
number
of houses, but these (no less than the value
figures) were affected by the continuous transfers and re-transfers of
buildings between different categories.
The House Duty statistics may therefore be used as a measure of
the
net
change
in
the annual number of
all
premises. but they must be
treated with the greatest caution when used
as
an index
for
house-
building activity.
The Census of Population gives the number of inhabited and
uninhabited houses on dates
on
which the Census was taken from
1801,
and the number
of
houses
in
course of erection from 181
1.
From
these data it is possible to obtain the
net
increases
in
the housing stock
over decennial periods. The Census information is somewhat impaired
by
changes in the definition
of
a
house, and it is likely that the
enumeration of uninhabited houses was incomplete. The statistics
give. however, a valuable-if rough-indication of building movements
over
a
long historical period and they have a major advantage in being
available
for
local areas. The numbers of houses
in
course of erection
provide
a
useful check
on
other data in Census years.
THE
NEW
INDEX
Statistics of building plans came into existence
in
connection with
the responsibilities of Local Authorities for enforcing the local building
bylaws. Builders wishing
to
erect houses were required
to
submit
'In
Scotland
the
system
of
valuation
amounted
in
cffccl
lo
annual re-
assessments.
I
ox
B.
WEBER
building plans
for
approval
to
the City Surveyor
or
Engineer who
sanctioned the plans
if
they confomied with the requirements
of
the
b)
-laws. Often the oRicer responsible also kept a check
on
the number
of
houses
actually erected, and
for
some towns these statistics alone
hu\e
bccn
preserved. Whenever available these are the more accurate
indicators, but usually they do not extend
as
far
back as the building
l'hcre are a nuniber
of
shortcomings in these statistics which must
he
brielly noted. First, there
is
the problem
of
urban boundary changes.
As
towns enlarged their administrative areas, the building series applied
correspondingly
to
larger districts and some discontinuities were
inevitably introduced.
No
adjustment was possible, but in most cases
ilie
volume of building in the added area
must
have been small in
relation to building
in
the town as
a
whole. Inspection of the building
series
in
years
in
which sonie
of
the larger extensions occurred indicates.
morcover.
no
obvious discontinuities.
It
is therefore reasonable
to
,tx,iiiiic
tliat
comparability is not, on that count, seriously impaired.'"
pl;tns.
Thc
definition
of
a
'
house
'
may have changed
from
time to time
and thcrc
may
have becn variations in definition between different
I-ocal
Authorities. Prior
to
1914 these changes were probably of
a
iiiiiior
character. The prevailing tendency in England and Wales
before
1914
appears
lo
have been to
follow
the Census of Population
aii~l
10
rlcfinc
n
'
house
'
as
'all
the space within the external and
party
walls of
a
building
',
and in the Census
for
England and
Walcs
Illis delinition
was
adhercd to throughout the period
from
1851 to 191
1.
Froiii
I021
the Census housing returns were based on
a
unit described
Ily
il~c
tcriii
'structurally scparate dwelling'.
It
was defined
in
the
1031
Cciisus
as
'
any rooti1
or
suite
of
rooms intended
or
used
for
Iiabitatioii having separate access to the street
or
to
a
common landing
or
staircase
'
(Hozisi?zg
Report
and
Tables)."
This change in
Jcliiiitiun
inay
have
sonie
eflect on the comparability
of
the
post-1918
ligurcs
with
prc-war
ones.
Inasmuch
as
a
given housing stock would
iiidudc
;I
largcr
nuniber of housing units under the 1921 definition
1h;tit
it
iwul~l
under
thc
dcfinition prevailing before the war.
it
is likcly
io
cx;Iggcr;Ite
the
level
of
house-building
in
the inter-war pcriod
in
rcl;itiuii
to
curlier
years.
'"C,
U.
Long:.
who
examined
the
effect
of
more than
firty
boundary
c\tctibions
in iiiliie than a dozen American cities found no evidence that
the
onit
i)riiiity
of
building permit data was seriously alTected
by
boundary changrs
(L.
I).
I.ong,
Jr.,
h'uddittg
c'yclcs
urtd
the
'I'hcory
of
Inocslncettt,
19.10,
p.
97).
I'
A
siiiiilar dclinition applicd
lo
Scotland froin
1881,
but for
1851-71
the
L~~~~IIII~~II~
was
Ihs
smie
3s
that used in England and Walcs
up
to
and including
l')ll.
A
NEW
INDEX
OF
RESIDENTIAL
CONSTRUCTION,
1838-1950
109
Other shortcomings adhere more specifically to
'
plan
'
statistics.
There is a time-lag before plans materialise into actual houses, and
some plans are never executed as builders subsequently decide not
to build. Impending changes in building by-laws-tending towards
tighter regulations-led at times
to
a rush
to
take out a large number
of plans in the hope that they might be executed under the more
liberal old regulations. and many
of
these plans were never imple-
mented.
In
other cases plans were allowed
to
lapse, only to be renewed
at
a
later stage.
A
time limit was generally set. sometimes
as
short
as
one year. during which actual building operations had to begin,
or
the approved plans would expire and have to be filed again
if
the
building was
to
go up.
For
a number
of
reasons therefore, building
plan statistics exaggerate the amount of actual building and
D
statistical
adjustment
to
allow
for
this tendency is required.
In order
to
render comparable the data relating to houses actually
erected and the data of plans
it
has been assumed that it took six
months
to
build a house and that plans approved
in
a given year were
implemented in the year ending six months later. It was further
assumed that
10
per
cent. of the annual building plans failed
to
be
executed, this percentage being
an
estimate formed from inspection
of
the statistics of plans and completed houses for a few towns for which
both are available over
a
period of years. Finally. all series given in
financial years were converted to calendar years.
It
was possible
to
assemble series of statistics for thirty-four towns,
twenty-six of these being in the form of 'houses erected' and the
remainder relating
to
the number of houses on approved building plans.
Before
1900
the sample gradually decreases to six in
1856
and for the
period from
1838
to
1855
use
is
made only
of
a series for Liverpool.
The
problem of combining an increasing number
of
series up to
1900
into
a
general index was solved by varying appropriately the number
of towns included
in
the base period
(1900
-
09).
Thus the houses
built
in
the towns included
in,
say,
1856
were added up and the
sums
were expressed as relatives of the average number of houses erected
in
the same towns in the base period.
As
further towns were added,
the average for
1900
-
09
was correspondingly raised by
the
average
amount
of
building
in
the added towns.
In
this way house-building
in
any given year was compared
with
house-building
in
the sanie area
in
the base years. Even
so.
however. the narrowing coverage before
1900
cannot but impair the accuracy
of
the index
for
earlier years.
London prcsciitcd
a
special problcni. Building activity
in
the
capital has at times diverged considerably from that clscwhere and
whcn London is includcd
in
the gcncral index by thc method describcd
110
B.
WEBER
il
tends, on account of its size,
to
dominate the index. The difficulty
was
overcome by
first
constructing an index based on thirty-three
towns. excluding London, and estimating on the basis
of
this index
thc actual number
of
houses erected in Great Britain apart from
London. The base for the number
of
houses built was the official
statistics
of
all houses erected in Great Britain in
1924-37
less
the
Iiouscs
built
in London.Ia The number
of
houses erected
in
the capital
WIS
then added
on
and the result was converted to an index based on
1900
-
09
=
100.
In this way London was given
its
appropriate weight
in
he
gcncral index.
The index
is
based on urban areas only but,
for
reasons given in
Section
VI.
it
is
possible
to
make
it
broadly representative of the
movenient
of
total
national house-building output by a correction
applied to the index in the period 1891
-
1913. As
for
the urban areas,
;ill
towns with
a
population
of
more than
200.000
in 1911 are included,
;IS
well
as
a
number
of
small and medium towns.
Most
of
the major
geographical
areas. such
as
Lancashire. Yorkshire. Scotland. Wales,
rlie
Midlands and the South, are in some way represented, although
sonic
are under-represented while others are perhaps overweighted.
This
is an unavoidable shortcoming since the towns included were
dclerniincd by the existence
of
data and not
by
some statistical
forinula.
If
additional scries can be secured-and
a
good deal may
still
hc
unearthed
from
old
Local Authority reports and files-the index
I!
ill
hc
reviscd and rendered more completely representative
of
the
country as a whole.
Although the final index is
not
based on the total volume
of
house-
building
undertaken. and in spite
of
the
defects discussed,
it
may
riom
helcss
be taken
as
a broadly reliable indicator
of
residential
huiltling fluctuations in Great Britain.
If
not all-inclusive, it does
rcflcct
house-building activity
in
a substantial sector of the economy
and
thc dcfects attached
to
the data prove in practice
to
be more
npparcnt
than real.
IV
THE
COURSE
OF
HOUSE-BUILDING
'I'lircc
final indiccs arc set out
in
the Appendix and shown in
I-.ig.
I.
There
is
an
index
for
Liverpool from
1838,
an index
for
L~wtlori
from
1856.
and
an index
for
the thirty-four towns from the
1937
is
the last year
in
the pre-war period
for
which
the
statistics
of
I
ondori
arc
available.
A
NEW
INDEX
OF
RESIDENTIAL
CONSTRUCTION,
1838-1950
111
I12
B.
WEBEK
s;inic
Thc latter is chained in 1924-8 with an index
of
the
toial
iiuiiibcr
of
houses erected
in
Great Britain.
It
is continued
to
19.50.
in
order
to
include the trough in house-building occasioned
by
the
war. and
it
can easily be brought up to date. For some purposes'
it
is uscful to know
also
the actual number
of
houses built
as
distinct
from
irdiccs.
and
for
Great Britain these
too
are
given
in the Appendix.
Ihc
index
of
thirty-four towns exhibits some minor fluctuations
iii
\tic
1850s
and
1860s. the most important
of
which appears
to
have
IWI~
tlic
lluctuation
with
its
peak
centred on 1863. The trcnd
is
almost
horimiital
for
tlic first ten years
or
so of
the index. but rises sharply
from
thr:
trough
of
the
niinor cycle
in
1865-6. There
is
a
minor peak
in
1872, followcd by
a
trough
in
1873 and then
by
a
sharp rise
to
the
pcak
of
the first
long
cycle
in
1876. From 1877
to
1879 house-building
sluiiips
heavily. Thereafter building remains
in
the
doldrums for some
sixtceii
ycars, but there are some minor ups and downs and there
is
a
gentle
upward trend from the trough year 1886.
In
the mid-1890s the
iiidrx
riscs sharply to a niajor double peak in 1898 and 1903. After
I')O3
the
do\vnswing
sets
in.
Activity
falls
almost persistently
in
the
Ilc'xi
1lccaJc.
with
only
slight reversals
in
1906 and 1913.
to
a
levcl
lov,cr
iliiln
at
any time since the 1850s and early 1860s. The outbreak
(If
war
iii
I914
further accentuates the building slump and reduces
htwsc-building to almost nil by 1918.
1
Iic
upsving of
a
third building cycle is located
in
the inter-war
pcritxl. Thcrc
is
ii
niinor
peak
in
1922
and
a
trough
in
1923. This
is
follmvcd
by a much grcater boom centred on 1927.
a
break with
a
trcriigli
in
1030,
and
a
final all-time
peak
in
1936. This movement is
liii.illy
rcvcrscd by
thc
Sccond World War. wEhouse-building came
11)
;I
si;iiidstill.
The
turning-points
of
the niajor fluctuations
in
Great Britain, as
\wll
;is
llie
turning-points
in
London and Liverpool. arc given
in
-l-;ll~lc
1.
'flic
gciicral
outline
of
the fluctuations is the
sanic
for London
as
ftlr
the
thirty-four towns
but
there are some divcrgcnccs
in
timing and
;IniplituJc
;ind
there is
a
further cycle-though not
on
a
par
with any
of
tlic
subscqiicrit fluctuations-with
a
pcak
in
1868.
Jn
contr~tst
to
I
lic
gcricral index, house-building
in
London increased
in
the later
I
XYk.
llic
early
1800s
and
up
to
tlic
boorn
of
1868. There followed
a
ii\i:ird
iiiw.*c!iici\l
to
I871 when the index
of
thirty-four towns
.\!IOU
Y
;I
rise.
llitrc
is
a
minor
pc;ik
in
London
in
I873 and
in
he
tlc.xt
ft)tir
ycars
the
two
indiccs niovc
in
agrccnicnl. They diverge again
;iftcr
I
%7(1.
wlicii
house-l>uiltlin_r was beginning
to
slump
in
thc gcncral
I.'
'I
IIC
ititlcx
will
For
conveniencc be referred
to
as
thc
'
index
of
thirfy-four
towti>
*
In
spik
of
ils
narrowing coverage
beforc IYW.
GREAT
BRITAIN
LONDON
Peaks
Troughs
Peaks
Troughs
-
......
......
......
......
......
......
1868 1873
I876 1886 1881 1890
1898 1918 1899
app.
1918
1936
1944
1934
......
LIVERPOOL
Peaks
Trough8
----
1845 1850
I864 1873
1878 1895
1903 1918
1927 1942
114
B.
WEBER
liiially reached rock-bottom in
1895.
Only
247
houses were completed
in
that
year.
‘The
pattern of house-building activity
in
the
1880s
and
early
1890s
diverges therefore from that
of
most other towns
in
that
it
follows
a
downward trend and reaches its trough rather later than
elsewhere.
The
upswing from
1895
to
1899
was both steeper and shorter
tli;1li
in
iiiost
other
towns. The temporary reversa1 in
1900
and the
sccoiid
peak
ol
the cycle are clearly marked, but there is a further
divergence
from
the more general experience after
1903,
when building
in
Liverpool remained high for some years longer while house-building
elsewhere declined.
After
the minor peak in
1906
Liverpool too
su
trcrcd depression. although there were
slight
reversals
in
activity
in
1909
and
1914:
In the inter-war period cyclical peaks are shown in
lU22.
1927
:rid
1934.
and troughs in
1924
and
1930.
The highest level
of
activity was reached
in
the boom year
1927.
almost ten
years
before
the
peak
in
the various available national indices.
The
fluctuation
in
Liverpool
in
the
1840s
is
of
special interest, partly
bcciluse
it
is the only available statistical evidence bearing directly on
housc-huilding
in
this period and partly because it suggests
a
further
long
cycle comparable
to
those with peaks
in
1864,
1878
and
1903.
It
is.
however. doubtful
to
what extent the movement in Liverpool may
he
ciikcit
as
representative for the country
as
a
whole.
It
is true that
Shaiinon’s figurcs for England and
Wales
show an output of bricks
in
1x40-9
greater
than
in
the corresponding years in the
1830s
by
3.1
hilltoits or ncarly
25
per cent.
But
there are
no
comparable figures
for
ihc
1850s
and
it
is
not
clear how much of the output
of
bricks
in
ihc
1840s
was
absorbed by the rapidly increasing railway construction
ii
1i0
by
orhcr
lion-residential
uses.
There were also at least two special local forces operating in
Liverpool
to
stimulate house-building that existed elsewhere only
to
;I
itiiich
lcsscr degree
(in
uric case)
or
not
at all
(in
the other). The
tirst
was
the
inipctus givcn to housing demand by large-scale Irish
iiiiniigration into the city; the other was the agitation for, and final
iiriplcwcnlation
of,
the
first Building By-Law
in
1846.
Faced with the
prospccl
of
being tied by building rcgulations, the Liverpool speculative
builJcrs anticipated the By-Law and rushed to provide
as
many
clwcllings
of
the
old
type as possible
in
the period immediately pre-
ceding
its i~itroduction.~‘
I’crhaps
ihc
best
tcst is
to
take the Census net increases
in
housing
51ocL
in
L:iigland and
Wnlcs
in
1841-51
and the adjoining decadcs and
c‘ciiiipilrc
thcm with housc-building
in
the city. The result indicates
a
I‘
RII\;II
C’ciniiiiiasitrn
on
the
Housing
of
the
W~irking-Classes.
Vol.
ti.
1’1
11’
IHRS.
A
NEW
INDEX
OF
RESIDENTIAL
CONSTRUCTION.
1838-1950 115
level of house-building
in
the 1840s considerably below both the 1830s
and the 1850s.
TABLE
11
NET
INCREASES
IN
HOUSING
STOCK
IN
ENGLAND
AND
WALES,
1831
-61
1831
-41
ia/i
-61
1851
-
61
5
15,733
314,341
492,666
Some allowance must be made for the greater volume
of
denlolition
which accompanied railway construction
in
the 1840s
;
and the Census
estimates of the increase
in
housing stock may be subject to consider-
able error since, prior to 1851, the interpretation of the term
'
house
'
was
left to the discretion of the enumerator. Even
so,
it is difficult
to
see how the volume
of
house-building
in
the 1840s could have
approached the level of the 1830s or 1850s.
When one comes to look at the other individual town indices. the
most striking fact to emerge is the general participation of all towns
in
the rhythm of the long cycles. although there were very considerable
variations
in
amplitude and timing and although not every town
participated
in
each cycle.
The cycle in the 1860s. the peak of which is strongly evident
in
Liverpool in 1864 and in London in 1868. and which emerges also
mildly in the index of thirty-four towns in 1863, is the only important
exception to general conformity.
It
appears to have been
a
fluctuation
of
more local than national incidence, although
it
was
not
confined
to
any specific geographical
area.
There is
no
similar movement
in
Birmingham and Newport
(Mon.)
although both towns underwent a
minor fluctuation. Glasgow suffered a serious depression in the middle
of the
1860s
and Bradford in the early 1860s. Recovery in both towns
was rapid from 1865-6 and
in
Glasgow activity increased steadily up
10
the building-cycle peak
in
1876. In Bradford
a
vcry high lcvcl of
activity was reached
by
1867 and continucd on that planc throughout
most of the 1870s but there was no boom in the 1860s distinct from
that entered upon
in
the next decade. Finally, the level of house-
building in Hull
rose
steadily from about 1855
;
there is no indication
of a major peak
in
the 1860s although
in
this town also there was a
minor boom
in
1867. The rnovcrnent there
is
more
in
the nature of
a
long secular rise
to
the peak
of
the building cycle
in
1875.
Practically all towns for which statistics are available shared
in
the
building boom
of
the 1870s.
Thc
only exceptions are one or
two
of
the smaller towns. Thus. neither Exeter nor Newport
(Mon.)
were
much affected. but
it
cannot
be said that the fluctuations were confined
I
I0
B.
WEBER
mainly
to
the larger industrial towns. Swindon, Gloucester and
Burton-upon-Trent are cases in
point
which participated fully
in
the
upswing
of
the 1870s. Burton,
in
fact, enjoyed at that time the greatest
hoom
of
its statistically recorded history
of
house-building.
Tlic downswing that followed and lasted in
most
towns up
to
thc
early-
to
mid-1880s is equally evident in most areas. although
Hull
was
relatively little aliected and activity in Cardiff and Newcastle rose
sharply
from 1882-the
first
year for which statistics for these towns
;Ire iivailablt-to a peak
in
1884
and 1888 respectively.
The niinor boom shown by the aggregate index
in
1889 appears
to
have rcached.
in
some towns, amplitudes
as
great
or
greater than
wcrc attained
in
the peaks
of
the
major building cycles in the
1870s
or
;inwild the
turn
of
the century. Thus Newcastle enjoyed a sharp
boom
in
1888. Exeter followed
in
1889,
Burnley in 1890 and Swindon
in
1891. Doncaster had
a
siniilar boom
in
1893. Bradford,
too,
enjoyed
iii
1890
;I
boom
only
a
little below the peak in 1903. Many other
I~\VIIS,
Iiowever. experienced
only
a minor fluctuation
at
the time and
thc
widcr lluctuations
are
best regarded as rather violent aberrations
of
a
gelleritl minor movement rather than
as
an
additional cycle
io
those
described.
Except
for
Coventry and Macclesfield, all towns participated in
t
he
major
boom around 1900. In Coventry. house-building rose from
I805
to
1898.
as
in
the other towns, but the really important advance
did
not
come until 1905-08 and the actual peak is located in
1911.
t
lousc-building activity
in
Coventry exhibits therefore
a
rhythm
opposite
to
that
of
the other towns
in
that it rises
to
boom conditions
;ti
;I
tinie when activity elsewhere drops
into
an acute depression.
~1;icclcsliclrl
follows
a
similar course. The double
peak
in
1899
(possibly 1x98)
and
1903, though very mild, is shown by the figures,
htrf
the
trcrid
throughout
1893-1913
is
gently but consistently upwards
ant1
the
highcst
point is reached
in
191
1.
Othcr cxccptions
to
the depression
in
building
so
widcly felt
in
I
tic
pcriod ininicdiately preceding the First World War were the towns
uf
Stochport
and
Burnlcy
in
Lancashire. Wakefield
in
Yorkshire and
Gourocii
in
Scotland.
All these towns register
a
major peak between
1x07
and
1901.
followed by depression and rapid recovery. with a
rising
trcnd
until
the outbreak
of
war.
In
Wakelield and Goiirock the
1x';ik
in
I91
I
is considerably higher than the
long
cycle peak around
I
')I
If).
t.-iii;iIly,
ill1
towns
took
part
in
the upswing
in
the inter-war period
iiiiil
ttic
tlownturn
brought about
by
the Second World War.
It
is
iittcrchtiiig
to
note, Iwwever. that quite
a
number of towns
(c.{/.
A
NEW INDEX
OF
RESIDENTIAL CONSTRUCTION,
1838-1950
117
Manchester, Derby, Glasgow, Bradford and Liverpool) reached their
peak
of
activity as early
as
1926-9,
almost ten years earlier than
the
peak in the national index."
J
L
i
0
i
0
0
.D
0 0 0
d
U
d
u;
0 0
0
0
0
0
L
A
selection
of
house-building indices for individual towns
are
graphed in Figure
2.
In order
to
reduce the effect
of
minor fluctuations
and to emphasise the long cycles, they are given
in
the
form
of
nine-
year moving averages. The actual figures have not been reproduced
in
Is
This
is
very
largely
due
to
Corporation
building.
118
B.
WEBER
the
Appendix
as
it
is
hoped to publish the material in full in due
course.
V
AMPLITUDE
OF
LONG
CYCLES
One
of
the
most
spectacular and well-known features
of
residential
buildirig cycles is the severity of their fluctuations.
In
his examination
of
104
American industries A.
F.
Burns found only thirteen with
trend-cycle amplitudesis higher than those evident in building. Three
of
these were building materials such
as
cement and asphalt, and the
remainder included
a
variety
of
industries such as beet-sugar. raisins.
sulphur,
vessels. aluminium. cigarettes and locomotives-none
of
which
is
an
industry
of
comparable magnitude to the building industry.
Changes
in
the amplitude of fluctuation
of
residential construction
oicr
tiriie arc
no
lcss
interesting than coniparisons with other industries
a1~1
sollie indication
ol
the broad movement can be derived from the
clata
on
hand.
The nleiiSureS
of
amplitude used for that purpose were
obtained
by
adding. for each cycle. the rise
from
the initial trough tb
the
peak
aiid
tlis
subsequent
fall
from
the peak
to
the terminal trough.
III
order
11)
niakc
comparable the severity
of
cycles
of
different length,
ilic
sims
wcrc dividcd by the duration of the cycles. As we
are
not
concerncd
with
minor fluctuations
it
is desirable to eliminate
or
rctliicr the
inllucncc
of
the
latter
and
this has been done by smoothing
the
iiidex
with
nine-year moving averages. The method described
)iclds
;I
iiic'asurc
of what may be termed
absolztte
amplitude.
A
I'ui.ihcr
iiieasure
of
relatiue
amplitude was calculated by adjusting
for
ilic
Irvcl
01
house-building which prevailed during each cycle. The
nicasurcs
of
amplitude are set out
in
Table
111.
'I'lie results suggest
a
rather wider amplitude (absolute and relative)
for
the
cyclc centred
on
1902 than
for
the preceding one and a much
wicicr
aiiipliturlc
for
the lluctuation centred
on
the inter-war period.
7'lic
;ibsolutc
;iniplitude
of
the
last cycle rises much niore sharply than
tllc
rclatiic aniplitude arid this can be attributed
to
the higher level
of
building
activity
and
to the irnpact of the two wars. The
wars
pruviclc
also
a
partial explanation
for
the
sharp
increase
in
the relative
aiiiplitudc
of
the cycle
with
a
peak
in
1935.
and the First World
War
lic!ps
to
accxiiiiit
for
~hc
wider amplitude
of
thc second cyclc centred
'"A.
1'.
tliirns.
I'ndt~c/;ot~
Ttwttds
iti
Ihe
U.S.
sinro
IX70,
(National Bureai
III
I'conoiiric Kcrearch,
1934).
pp.
230-3.
Burns'
'
trend-cycles
'
are
deviations
tit
i!ie
dccadc rates
ot
series
froiii
cxponsntial curve trends. Standard devia-
\ion,
ais
ud
il~
il
nieasure
of
nniplitudc.
A
NEW INDEX
OF
RESIDENTIAL CONSTRUCTION,
1838-1950 119
on 1902. Not the whole
of
the increase in amplitude in 1887-1919
can, however, be attributed
to
the
First
World War.
If
the absolute
amplitude
of
the second cycle was measured over
a
period excluding
the war years (1887-1910) it.would
still
amount to
3.8-0r
approxi-
mately 80 per cent. above the amplitude of the cycle centred on 1874.
(
1)
Long
mclei
trough
to
trough
1860
-
1875
-
1887
1887
-
1902
-
1919
1919
-
1935
-
1944
TABLe
111
(2)
(3)
(4)
(5)
Sum
of
Duration Absolute Relative
rises
and
of
cycles aniplituds amplitude
falls
years
,
58 27 2.1 3.3
10.6
141
31-
44
372 25
14.9
*Based on
9-year
moving
averages of index of thirty-four
towns
(Great
Britain after
1923).
VI
RURAL
HOUSE-BUILDING
The discussion
so
far has been concerned entirely with residential
construction in towns. It
is
equally instructive to survey the course of
house-building in the countryside and
to
determine in particular
whether the long cycles described occur there
also
or
whether they
are entirely an urban phenomenon.
Annual statistics of rural house-building are unfortunately not
available but some idea of the progress
of
house-building can be
obtained from the net increase
in
rural housing stock between one
Census and the next. This can be estimated by extracting from the
Census the number
of
houses
in
all Registration Districts
(or
Sub-
Districts) which are predominantly rural. The boundaries
of
the
Registration Districts do not.
of
course, conform to any precise rural-
urban division, but they
can
be
so
grouped as to approximate roughly
to
the two sectors.
For
the purpose
in
hand. all Registration Districts
other than those including a town with a population of 20,000
or
more
120
B.
WEBER
in
IXYI
were defined as rural. The resulting estinlates include there-
fore
Iiouses not only
in
entirely rural areas but also in small towns.
On
the other Iiand. some houses in the rural fringes
of
large towns are
excluded.
But
the picture of long
run
changes that emerges is on that
count
probably little affected."
A
further dilliculty is that the boundaries of Registration Districts
wcre
changed from the to time. Fortunately. there were relatively
fcw
large
changes. Some of these do not affect
our
calculations since
it
is
with
the
housing stock in the
total
of rural areas that we are
cmccriied
and changes
in
boundaries between rural districts alone
do
not
affect
the totals. In the few cases when there were large changes
in
ths
boundaries between
rural
and urban areas. the necessary adjust-
ti:stits
to
ensure
coniparability could be carried
out.
Smaller changes
(1)
Total
(000'8)
315
491
596
698
606
886
TABLE
IV
(2)
Rural
(000'8)
81
98
113
93
85
I57
NET INCREASES
IN
RURAL, URBAN AND TOTAL
HOUSING
STOCK
IN
ENGLAND
AND WALES,
1841
-
1911'
.-
-
-.
-
-
.
1x41
-
51
1x51
-
61
1861
-
71
1x71
-HI
1x81
-91
1891
-
IYOI
1')OI
-
II
(3)
Urban
(000's)
234
394
483
605
52
I
7
ZY
599
0
Tlic
Scottish figures are excluded because
of
a
change in the definition
of
:I
IICIUW
in
1881
which
makes
comparability with previous years inipossible
I
srfllfiS/t
(.:rrrsrts
NI*/,fJ'1
1x81,
Vol.
1,
p.
X).
':
7'11~
rural sector
of
the country
is
more accurately defined by the Rural
S;irliIltry Ljibtrictr (later Rural Districts). first used in the
1881
Census, but they
.ire
kit
no
use
for our purposes since the housing stock cannot
be
estimated
lor
these
arcas heyonJ
1901.
Their population coverage in
1891
amounted to
t).502.W0
cciiripnrcd with
I
1,203,000 in the grouping
of
Rural Registration
I)i\tii~!b
u\cJ
here.
A
NEW INDEX
OF
RESIDENTIAL CONSTRUCT'ION,
1838-1950 121
in boundaries were more plentiful but their effect is hardly important
enough
to
upset the general conclusions drawn from the
figures."
The increases in the number
of
rural houses together with the
corresponding urban figures-obtained by simple subtraction
of
the
rural estimates from the totals-are set out in Table
1V
and Figure
3.
000s
800
700.
boo
500.
LOO.
*
-
300.
2w
100
1
1
I
1
I
1
18S1
1861
1871
1881 1891 1901 lV11
Net increases in
rural.
urban and total housing stock in Census decades in
England and Wales.
1851
-
19
I
I.
FIG.
3
Rural house-building appears to have described
a
slight upswing
from
the 1840s
to
1871 and
a
downward movement in the next
twenty years,
to
a
level almost the same as that
of
the 1840s. There
was
a
more pronounced rise in the
1890s
and 1900s. The striking
characteristic
to
emerge, therefore, is the absence
of
major fluctuations
in
parallel with the long cycles
in
the urban and national indices.
Figure
3
illustrates also the close conformity of movement between
the national and urban increases
in
housing stock.
The
latter provided
by
far
the greater proportion
of
the national total and the long
fluctuations were characteristic of this sector alone.
so
that
it
dominated
I'
A
check on the e!iect
of
these minor boundary changes has been obtained
as
follows. Thc Census provides information on houses
for
each Registration
District
in
the current and preceding Census date within current boundaries.
A
comparison
of
the number
of
houses within current boundaries and within the
boundaries.
ten
years hence provides
a
measure
of
the error introduced into
comparability. The
sum
of
the divergences was in each decade
less
than
I
Der
cent.
-
-
Rural
-
122
4.
WEBER
the
course of total house-building. In
1841
-
91.
when there is almost
110
trend in the rural sector, the two series were almost parallel. In
1891
-
1901
the rise in the urban index is somewhat milder and in
1901
-
11
the fall is sharper. The effect
of
rural
building in
this
period
was
therefore to raise the trend of national house-building and an
;ldjusIment
to
our index is required
if
it
is
to
serve as an indicator
for
the
country as a whole.
An
adjustment was camed
out
on the basis
of
an index expressing
the
ratio of the decennial increases
in
total housing stock
to
the
iricreases
in
urban houses. The index was based on
1881
-91
=
100
aiid
the deviations from
100
measure the extent to which the movement
of
the national increases in housing stock diverges from the urban
figures.
Except
in
1901
-
I
I
the deviations are slight throughout the
pcriod for which
a
coniparison with
our
house-building index
is
rcq
ti
ired.
1851
-61
1861-71
1871-81
1881-91
--_
!--
107
106
1891-1901 1901
-
I1
I9
An
element
of
error may still be retained in
tlie
corrected index
on
~ICL‘OUIII
of
tlir
iiiipossibility
of
miking
any
allowance
for
possible changer in
Jcinolilions
and conversions
of
houses into and
from
other
uses.
A
NEW INDEX
OF
RESIDENTIAL CONSTRUCTION,
1838-1950
123
VII
COMPARISONS
WITH
ALTERNATIVE
INDICES
The index of thirty-four towns can be compared,
for
part
of
the
period covered, with a number
of
indices based
on
other data, and
this
has been done as fully
as
possible to obtain some measure
of
its
reliability.
All
statistics used
in
the comparisons
are
set out in the
Appendix.
i
.
-
-
.
~
lot21
nunlbrr
01
hours
rrrctrd
In
trcit
Brwn
-...-.
-.
.-.-./
-
>4Tawm
.
.
.
.
.
.
13
Towns
(rxcl.
London)
-.
-.-
Bulldin;
plmr
Indices
of
house-building in Great Britain.
1912
-
50.
(1930=
100.)
FIG.
4
For
the first comparison the index
of
thirty-four towns has been
continued
to
1937 and in Figure 4
it
is
set
out alongside the index
calculated from the statistics of all houses erected
in
Great Britain
in 1924-37. Agreement is very close throughout and the only
discrepancy
worth
noting is the rather lower level of activity indicated
by the thirty-four towns in 1935.
A
longer comparison with the official
statistics
of
house-building can be carried
out
only with thirty-three
towns. excluding London. This has been done
for
the period 1924
-
50
(Fig.
4). The movements
of
both indices conform very closely
except for the same divergence
in
1935 and again in 1948
-
50.
The
only important discrepancy in turning-points occurs
in
the late 1930s.
with the index
of
thirty-three towns indicating
il
peak in 1938
as
124
B.
WEBER
against I936
in
the comprehensive data. This divergence between
turning-points can be attributed to the exclusion of London where
huilding activity in the 1930s pursued
a
different course.
The
link
with pre-1914 years can be tested by comparison with
an
index calculated from the estimated cost
of
dwelling-houses on
approved building plans (Fig.
4).20
These figures are available
for
eighty towns in 191
1
-
30
(but not 1922) and
for
146 towns in 1923
-
38.
They have been lagged by one year to allow
for
the time that elapses
between the passing
of
building plans and the completion
of
houses.21
lhe
figures
were then deflated by an index
of
building costs and
chained together
in
1924
-
6.
The movement
of
both indices
is
broadly
similar
but divergences arise in 1921
-
2, 1926 -9 and 1935
-
6. In
1Y21
the plan statistics show
a
sharp boom which has no corresponding
~ii~)vciiient
in
the thirty-four towns.
On
the other hand, the thirty-four
towiis
indicate
a
slight peak in the following year, while building plans
have
reversed into
a
severe depression. The reliability of plan statistics
as
an
indicator of actual building activity in these years
is,
however,
cxtrenicly doubtful." In 1926
-
9 the building plans indicate a lower
Icvel
of
building activity than the thirty-four towns but the latter appear
LO
be
the
niorc
accurate when compared with the total number of
Iiuuscs
erected.
On
the other hand, the lack of agreement in 1935
-
6
confirm
the
inipression that the thirty-four towns understate the
vc)ltiiiie
of
activity
in
those years.
A
further comparison between a pre-1914 year and
a
post-war
year
is possible by nuking use of the Census of Production and Inland
lievenue
Assessments. Using these sources, Colin Clark estimated the
valuc.
of
new
dwellings
in
1907
at E35m. and in 1924 at f72m.13 When
deflated
by
an index of building costsz'
the
figures suggest
a
level
of
Iiousc-building
in
1907
slightly higher than in 1924. In contrast. our
index
indicates
a
level
of
house-building in 1907 lower by
approxi-
malcly
8
per cent. The divergence, as
it
stands, is not excessively
2u
lhe statistics
of
building plans cover Greater London but not the London
(.'ounty Council area; the statistics
for
London used in this paper, however,
rrlittc
to
Greater London.
inclusive
of
the L.C.C. area.
A
onc-year lag appears more appropriate in this period than
a
shorter lag,
as
ilic
turning-points then coincide with the turn in all houses erected In Great
Hrilain.
>'.
It
is for this reason that the information was not published for 1922
(1923
if
1;rgjied).
It
was
held onicially that the rapid changes in the costs
of
building
2nd
the considerahle number
of
plans approved upon which work was not
prmwding
'
considerably reduced the value
of
these statistics
as
an indicator
of
huildiiip operetions.
Minialrv
of
Labour
Gazette
(March
1923).
p.
85.
C.
Clark.
ltii*efihi,nt
ill
Fixed Capital
it1
Great
Britaitt,
(Memorandum
No.
4Y.
Royal
Economic Society and London and Cambridge Economic Service,
Octolwr
1924).
p.
6.
K.
blaiuald,
'An
Index
of
Bui!ding
Costs
in the
U.K..
1845
-
1938
',
lIIinfury
Review
(2nd Series), Vol. V11,
No.
2
(December
IY54),
p.
191.
A
NEW
INDEX
OF
RESIDENTIAL CONSTRUCTION,
1838-1950
125
great, but there are some grounds
for
believing that Colin Clark's
estimate
for
1907
is
too
high. Cairncross has a similar estimate
for
that year
of
€28
i
2
m."
In
real terms,
and
expressed
as
a relative
of
Clark's
1924
estimate.*' this figure diverges from our index by approxi-
mately the same amount
as
Colin Clark's. but
it
indicates
a
level
of
house-building in
1907
below
our index.
If
one, however, takes
f30-2
m.,
a figure which is well supported by some
of
the checks
made by Caimcrossz' and approximates to
the
mean
of
the Cairncross
and Clark estimates, agreement with our index
is
most
satisfactory.
The comparisons are set out in Table
VI.
35
28
53.0
66.0
101.7
92.4
81.4
TABLE
V1
HOUSE-BUILDING IN
1907
AND
1924
72
......
1
10.9
64.9
100.0
100.0
......
1.
C.
Clark. Value
of
new dwellings. Current
fm.
.
2.
A.
K.
Cairncross. Value
of
new dwellings.
3.
Index of building costs.
1930
=
100
. .
.
4.
C.
Clark
(1)
in
1930
prices.
(l)+(3)
fm.
.
.
5.
C. Clark
(4)
with
1924
=
100
. . .
. .
6.
B,
Weber. Index
of
house-building.
1924
=
100
.
Current
fm.
7.
A.
K.
Cairncross
(2)
deflated by index of building
costs
(3)
and expressed as a percentage
of
C.
Clark's estimate
for 1924 (4)
For
years before
1914
house-building
in
the thirty-four towns
can
be compared with Cairncross's estimates
of
the number
of
houses
erected annually
in
Great Britain
in
1870
-
1914.2*
For
1901
-
14
these
estimates are based on the number
of
houses erected
in
Greater London
aiid in the
forty
largest niunicipal areas in England and Wales (plus
an allowance for Scotland). and for
I870
-
1900
on the Inhabited House
Duty statistics supplemented by Census and other niiscellaneous data.
25
Cairncross,
op.
cit.,
p.
109.
26There
is
no reason
10
query C. Clark's figure for
1924.
The
1924
Census
of
Production
is
easier
to
work with and the same figure is arrived at also
by
1.
Bowen
(op.
cit.,
p.
118).
Cairncross,
op.
cit.,
p.
108
-
9.
za
Ibkl.,
p.
157.
I26
B.
WEBER
A
NEW INDEX
OF
RESIDENTIAL CONSTRUCTION.
1838-1950 127
The House Duty statistics were made use
of
only in revaluation years
and the year-to-year changes were smoothed in accord with
other
evidence on hand.
The major fluctuation
in
the 1870s and around
1900
emerges clearly
In both series and there are only
minor
divergences in turning-points.
Cairncross sets his turning-points
in
1877. 1885 and 1899
as
against
ours in 1876, 1886 and 1898. The amplitude
of
the boom.in the 1870s
is milder in Cairncross's series and there
is
no fluctuation whatsoever
in
1870
-
7. In contrast the town indices suggest a depression. varying
in severity
from
place to place, in
or
before 1873. Further divergences
arise in the minor boom in 1889, which is less pronounced in the
thirty-four towns, and in 1903. when the second peak
of
the boom
is
only mildly evident in the Cairncross index.
Its
near disappearance
may be the
result
of
the link in 1901
of
the House Duty statistics with
the other material. The
House
Duty figures themselves
point
to
a
rather stronger upward movement at the time.a'
col.
(1)
1901
-
11
=
100
TABLE
VII
Houses
in
course
o
erection
on
dwui
day
af
end
of
period
NET INCREASES
iN
HOUSING
STOCK
AND
NUMBER
OF
HOUSES
IN
COURSE OF ERECTION ON CENSUS
DAY
IN
ENGLAND AND WALES.
I851
-
1911
1851
-
61
1861
-
71
1871
-81
1881
-
91
1891
-
1901
1901
-
11
--
--
(I)
Net
increases
itr
hurtsing
sfucl;
492,666
596,263
697,733
605,486
886, I03
840,649
58.6
I
27,305
709
I
37,803
80.9
I
46,414
72.0
I
38.387
105.4
I
61,909
100.0
1
38.178
(4)
Col.
(3)
1901
=
107.0
47.2
65-3
802
66.3
107.0
66.0
The niaterials
in
the Census
of
Population enable us to make two
further comparisons. The index
ol
thirty-four towns
can
be compared
with, first, an index based on the number
of
houses
in
the course
of
2*
Cairncross,
up.
cit.,
p.
157.
I28
D.
WEDER
crcction
on Census dates
;
and, second, with. the net inter-Census
incrcases
in
total housing
stock.
The
figures
art
set out in Table
VII
and
charted
in
Figure
5.
The gradually mounting level
of
activity in
illc
1860s
and
1870s.
the reversal in the
1880s,
the renewed upswing
in
Ihc
1890s.
and finally, the downswing after the turn
of
the century,
arc
;dl
confirrncd
by the increases
in
housing stock. Agreement with
thc
number
of
houses being built, with the possible exception
of
1881,
is
equally satisfactory.
To continue reading
Request your trial