Jump to ContentJump to Main Navigation
British Economic Growth 1856-1973The Post-War Period in Historical Perspective$

R. C. O. Matthews, C. H. Feinstein, and J. Odling-Smee

Print publication date: 1982

Print ISBN-13: 9780198284536

Published to Oxford Scholarship Online: November 2003

DOI: 10.1093/0198284535.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2017. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 26 February 2017

(p.565) Appendix D Hours Worked

(p.565) Appendix D Hours Worked

Source:
British Economic Growth 1856-1973
Publisher:
Oxford University Press

New estimates of hours worked in 11 benchmark years were prepared for this study. They refer to the hours actually worked after allowing for overtime, short‐time, and part‐time working; holidays; sickness; and strikes. The estimates and their sources are shown in Tables D.1 and D.2, and the methods used to arrive at them are outlined in this appendix.

We believe that the estimates of average hours worked per year are accurate enough to indicate the broad changes that have taken place over a number of years. However, some of the detailed components of the estimates are unreliable, and no doubt improvements could be made. This is especially true of the estimates for pre–World War I years.

The estimation of total hours worked required the separate estimation of seven variables:

Variable

Column in Table D.1

(a) Total in employment

(1)

(b) Average hours worked per week by full‐time workers

(2)

(c) Percentage of part‐time workers in the total in employment

(3)

(d) Average hours worked per week by part‐time workers

(4)

(e) Average weeks lost per year due to holidays

(6)

(f) Average weeks lost per year due to sickness

(7)

(g) Average weeks lost per year due to strikes

(8)

Total hours worked can then be calculated in the following stages:

(h) Average hours worked per week by all workers

(5) = [(2) × [1 − (3)] + (4) × (3)]/100

(i) Average weeks worked per year

(9) = 52 − (6) − (7) − (8)

(j) Average hours worked per year

(10) = (5) × (9)

(k) Total hours worked

(11) = (1) × (10)

The definition of each of these variables should be clear from the name. One detail might be noted here. Absenteeism is in principle allowed for in the (p.566)

Table D.1 The Total in Employment and Total Hours Worked Per Year, 1856–1973

Average weeks lost per year due to:

Year

Total in employment (Thousands) (1)

Average full‐time hours per week (2)

Percentage of part‐timers (3)

Average part‐time hours per week (4)

Average hours per week (5)

Holidays (6)

Sickness (7)

Strikes (8)

Average weeks per year (9)

Average hours per year (10)

Total hours per year (millions) (11)

1856

12,170

65.00

65.00

49.00

3,185

38,760

1873

14,100

56.00

56.00

49.00

2,744

38,690

1913

20,310

56.40

56.40

1.42

1.70

0.07

48.81

2,753

55,913

1924

18,238

47.00

1.8%

24.50

46.60

2.11

2.19

0.08

47.62

2,219

40,475

1937

21,272

48.60

1.5

25.40

48.20

2.31

2.11

0.03

47.55

2,293

48,786

1951

23,589

45.63

4.1

20.15

44.59

3.13

2.45

0.01

46.41

2,071

48,856

1955

24,347

45.88

6.1

20.18

44.32

3.45

2.31

0.02

46.22

2,051

49,941

1960

24,823

44.75

8.6

20.89

42.69

3.56

2.35

0.02

46.07

1,969

48,879

1964

24,999

43.85

10.9

19.96

41.24

3.54

2.34

0.01

46.11

1,904

47,591

1968

24,854

42.16

12.8

19.85

39.31

3.84

2.31

0.03

45.82

1,804

44,832

1973

25,030

41.40

15.5

19.88

38.06

4.67

2.29

0.06

44.98

1,715

42,926

Source:

  1. ((1)) See Appendix N.

  2. ((2)) Principally Department of Employment Gazette, largely reprinted in British Labour Statistics Historical Abstract, UK [48]. Also, for 1856–1913. Clapham 1932: 667–68; Chief Inspector of Factories, UK [26], 1910; R. C. on the Civil Service, UK [82], 1914: 37 and Q.36, 105–15; R.C. on the Coal Industry, UK [84]: 166; Select Committee on Early Closing of Shops, UK [91], especially Q.629, 2568–73, 114–19. For 1924–37, Fuel and Power Digest, especially 1953, Table 30; Balfour Committee, UK [31]: 111–12; 19th Abstract of Labor Statistics, UK [64]; Select Committee Shop Assistants, UK [92]; H. L. Smith et al. 1930–35; Klingender 1935: 89. For 1951–73, New Earnings Survey, UK [47]; Digest of UK Energy Statistics, UK [49]; Time Rates of Wages and Hours of Work, UK [68]; R. C. on the Civil Service, UK [83], 1995; National Board for Prices and Incomes, UK [74], [75], [76]; Civil Service Clerical Association (n.d.); Industrial Society 1966; G. S. Bain 1970.

  3. ((3)) For pre‐1914, there were assumed to be no part‐timers. For 1924–37, Chapman 1952. For 1951–73, estimates from Census of Population, UK [52], 1961, 1966, and census of employment (for 1968, 1973) in Department of Employment Gazette and British Labour Statistics Year Book, UK [39].

  4. ((4)) For pre‐1914, there were assumed to be no part‐timers. For 1924–37, Chapman 1952. For 1951–73, estimates from Census of Population, UK [52], 1961, 1966, and New Earnings Survey, UK [47].

  5. ((6)) Manual workers: For pre‐1914, Report of an Enquiry by the Board of Trade, 1906, UK [51]. For 1924–37, rough estimates from various sources, including Ministry of Labour Gazette, March 1938, pp. 86–87, and Committee on Holidays with Pay, UK [30]. For 1951–73, Time Rates of Wages and Hours of Work, UK [68]. Non‐manual workers: New Earnings Survey, UK [47]; Cameron 1965: 273–99; G. S. Bain 1970: 63–68.

  6. ((7)) For pre‐1914, A. L. Bowley 1919: 28. For 1924–37, rough estimates from Report by the Government Actuary, 1955–56, UK [97]; Table 7. For 1951–73, Social Security Statistics, UK [51]; Ministry of Pensions and National Insurance, UK [69], various years; and UK [70], 1964: Tables A13, B13; Report by the Government Actuary, UK [97], [98], and others.

  7. ((8)) For 1893–1973, British Labour Statistics Historical Abstract, UK [48]: Table 197.

Note: The definition of the working population changed in 1951 and 1964 (see Appendix N); the estimates of the total in employment (column 1) and total hours worked (column 11) are consistent with later years but not with earlier years. For the total in employment, the 1951 and 1964 estimates comparable with earlier years are 23,748 and 25,541, respectively. For the total hours worked, the comparable estimates are 49,110 and 48,559. Other columns are not significantly affected by the change in definition.

(p.567)

Table D.2 Total Hours Worked Per Year, by Industry, 1924–1964

(Millions)

Industry

1924

1937

1951a

1951b

1955

1960

1964c

1964d

1968

1973

Sector

Agriculture, forestry, fishing

3,288

2,829

3,013

2,745

2,658

2,362

2,014

2,216

1,779

1,426

Mining, quarrying

2,492

1,930

1,775

1,781

1,778

1,451

1,249

1,226

859

627

Manufacturing

13,333

16,054

18,473

17,152

17,984

17,867

17,520

16,687

15,611

14,430

Construction

1,819

2,858

3,332

3,368

3,515

3,685

4,123

3,997

3,788

3,710

Gas, electricity, water

438

652

791

791

821

788

847

853

798

627

Transport, communications

3,474

3,779

4,058

3,954

3,831

3,673

3,626

3,625

3,435

3,212

Distributive trades

5,023

6,984

5,074

5,456

5,925

6,297

6,384

6,157

5,466

5,192

Insurance, banking, finance

767

975

869

871

930

1,030

1,126

1,407

1,495

1,721

Professional and scientific services

1,790

2,245

3,612

3,241

3,453

3,739

4,006

4,038

4,444

4,645

Miscellaneous services

6,042

7,956

3,583

5,097

4,791

4,412

4,388

4,050

3,767

3,970

Public administration and defense

2,009

2,524

4,530

4,400

4,255

3,575

3,276

3,335

3,390

3,366

All sectors

40,475

48,786

49,110

48,856

49,941

48,879

48,559

47,591

44,832

42,926

Manufacturing

Food, drink, tobacco

1,367

1,729

1,775

1,562

1,648

1,627

1,582

1,515

1,422

1,357

Chemicals and allied

475

588

1,004

963

1,026

1,065

995

932

876

845

Iron and steel

780

888

966

976

998

997

978

985

869

744

Electrical engineering

382

756

1,166

1,149

1,416

1,540

1,697

1,563

1,485

1,412

Mechanical eng., shipbuilding

1,621

1,876

2,822

2,942

3,194

3,165

3,142

2,844

2,805

2,470

Vehicles

901

1,349

2,232

1,558

1,883

1,886

1,770

1,752

1,557

1,483

Other metal industries

926

1,246

1,591

1,305

1,350

1,402

1,417

1,452

1,388

1,299

Textiles

2,747

2,645

2,236

2,250

2,001

1,745

1,562

1,463

1,249

1,061

Clothing and footwear

1,685

1,784

1,416

1,245

1,164

1,085

1,008

936

805

751

Bricks, pottery, glass, cement

492

701

727

727

761

731

744

713

650

597

Timber, furniture

589

761

716

694

679

660

641

626

633

644

Paper, printing, publishing

861

1,114

1,088

1,063

1,138

1,230

1,234

1,192

1,163

1,049

Leather and other manufacturing

507

617

734

718

726

734

750

714

709

718

All manufacturing

13,333

16,054

18,473

17,152

17,984

17,867

17,520

16,687

15,611

14,430

Source: See Table D.1.

(a) Comparable with interwar years.

(b) Comparable with 1955–64: there was a change in 1951 in both the definition of the economically active population (TableD.1) and the SIC.

(c) Comparable with 1951–60.

(d) Comparable with 1968–73: there was a change in 1964 in both the definition of the economically active population (Table D.1) and the SIC.

(p.568) estimates of the average number of hours worked per week (both full‐time and part‐time), and not in the estimates of the average number of weeks worked per year.

The estimates for the three earliest years, 1856, 1873, and 1913, were made for the whole economy only; those for the two interwar years, 1924 and 1937, were broken down by industry; and those for the six postwar years, 1951, 1955, 1960, 1964, 1968, and 1973, were broken down by industry and sex. The industrial breakdown of total hours worked in 1924–73 is shown in Table D.2. The seven primary variables for each industry and, in postwar years, for each sex, are not shown here, but they have been used for some of the tables in the text (e.g. Table 3.13).

The following notes briefly outline the methods used for the seven primary variables.

  1. (a) Total in employment. See Appendix N.

  2. (b) Average hours worked per week by full‐time workers. Separate estimates were made for broad groups of workers, and then weighted averages were obtained. The hours worked by full‐time manual workers were estimated from Ministry of Labour inquiries, supplemented by information on a few specific industries (agriculture, coal mining, and railways). There were no satisfactory inquiries in the years before 1914, and so the actual hours worked by full‐time manual workers were assumed to be two hours per week longer than the normal hours.* The full‐time hours of non‐manual workers were estimated on the basis of various scanty pieces of information relating to shop assistants, civil service clerks, and other categories of non‐manual workers.

The main source for estimates of hours worked is the Department of Employment Gazette. Most of the relevant information is reprinted in British Labour Statistics Historical Abstract 1886–1968. Other sources used are shown in Table D.1.

The weights for combining the estimates of the average weekly hours worked by various groups of full‐time workers were derived from the Census of Population either directly or indirectly using data in Routh 1965.

This method underestimates the average hours worked per week by full‐time workers to the extent that there was any double job‐holding, or moonlighting, to use the American term. No attempt was made to allow for moonlighting at any stage of the estimation of total hours worked per year, and hence our estimate of the total labor input is biased downward in all years on this account.

The bias is likely to be small, but only an order of magnitude can be indicated because of the conflicting evidence provided by three different sources. In the 1966 Census of Population 3.3 percent of males in employment in Great Britain, and 1.7 percent of females, were found to have a second job (as cited in Alden (p.569) 1971: Table 2; the census estimate was not officially published). In a survey in Midlothian in 1968, Alden found a proportion of 5.8 percent (males and females together); he suggests (p. 104) that part of the discrepancy between this figure and the 2.7 percent found in the 1966 census resulted from a higher rate of admission among moonlighters in the Midlothian survey. It was estimated from the Family Expenditure Survey that 6.2 percent of males and 9.1 percent of females had second jobs in 1970.* Finally, the General Household Survey found that multiple‐job holders were 3.1 percent of all workers in 1971 and 1972 (3.5 percent for males and 2.6 percent for females in 1972), and 4 percent in 1974 (4 percent for males and 3 percent for females). (UK [77]; the relevant question was not asked in the 1973 survey.) It is difficult to believe that the discrepancies between the various sources are all due to changes over time. More likely than not, they are due to inaccurate responses and to differences in definition (for example, the census recorded only moonlighting that occurred in the week ending April 23, 1966, whereas the Midlothian and Family Expenditure surveys in principle recorded moonlighting at any time during the year).

To illustrate the order of magnitude of the bias arising from the exclusion of moonlighting, we assume that in 1964 5 percent of males and 7.5 percent of females had second jobs. We assume further that on average moonlighters worked eight hours per week at their second jobs. These assumptions imply that our estimate of total hours worked per year in 1964 should be raised by 0.9 percent for males and 1.7 percent for females. In the absence of any information referring to years before 1966 it is not possible to make similar estimates for earlier years.

The failure to allow for moonlighting in the estimates of total hours worked may bias the interperiod comparison of labor input growth. Though the evidence is inadequate, most writers believe that moonlighting expanded during the postwar period, at least until the 1960's, after which there may have been little change. The biggest error from this source in our measure of the growth of labor input is probably therefore in the period 1951–64. However, there are two reasons for believing that this point is not important quantitatively at the economy‐wide level. (1) As shown in the preceding paragraph, the underestimate of the total labor input in the mid‐1960's on this account was probably little more than 1 percent. Assuming there had been no moonlighting in 1951, the downward bias in the 1951–64 growth rate would be only 0.1 percent per (p.570) year. And (2) it is not even obvious that there was less moonlighting in 1951 than in 1964. On the one hand, full‐timers worked fewer average hours in 1964 than in 1951, and there were more part‐timers, so that on both accounts workers had more time in which to do a second job. On the other hand, real incomes were higher, so that there was less financial incentive (unless income aspirations were even higher) to seek a second job.

Moonlighting tends to be concentrated in a few sectors. In Midlothian in 1968, for example, Alden (1977: Table 7) found that almost 60 percent of second jobs were as teachers (19 percent), in entertainment (29 percent), or as painters and decorators (10 percent). In the United Kingdom in 1969, 43 percent of second jobs were in the distributive trades, 11 percent in professional services, and 24 percent in miscellaneous services (“Family expenditure survey,” UK [45]: Table 5; these figures refer only to workers who were employees [not selfemployed] in their main jobs). The bias might therefore be more serious in particular sectors. But even if one assumes that 50 percent, for example, of moonlighters in 1964 were in the distributive trades, the downward bias in that sector is only 4.4 percent. The 1951–64 growth rate of the labor input in the distributive trades would be underestimated by at most 0.3 percent, and in other sectors the bias would be less.

The failure to allow for moonlighting in the labor statistics carries over to the output statistics in those sectors, such as professional and miscellaneous services, where output growth is measured by employment growth. Productivity growth is therefore less likely to be estimated inaccurately in these sectors on this account than output growth or employment growth.

  1. (c) Percentage of part‐time workers in the total in employment. Because the Census of Population concept of part‐time work is broader than that employed for Ministry of Labour inquiries,* the present estimates of average hours worked per week for years up to 1964 may be biased downward to the extent that the estimates of average hours worked per week by full‐time workers have already been reduced by the inclusion in the samples on which the estimates were based of workers who would be classified as part‐time workers according to the Census of Population concept. No attempt has been made to correct for this potential source of bias, but rough tests suggest that it is not serious.

  2. (d) Average hours worked per week by part‐time workers. The same assumptions about the relationship between the hours worked by part‐time workers and those worked by full‐time workers as were made by Chapman (1952) were used for the estimates for 1924 and 1937. As with the estimate of average full‐time hours, no allowance was made for the holding of second jobs by part‐timers.

  3. (e) Average weeks lost per year due to holidays. The holiday entitlements of manual workers covered by collective agreements were published for the postwar (p.571) years by the Ministry of Labour. For interwar years the information is less complete, and some rough estimation from various sources was required. No comparable figures for non‐manual workers were available, but it appears that their holidays were normally longer than those of manual workers, and that they benefited from longer holidays to the same extent. (New Earnings Survey, UK [47]; Cameron 1965: 273–99; G. S. Bain 1970: 63–68.) For pre‐1914 years there was only a single estimate for manual workers derived from a Board of Trade inquiry. This was assumed to apply to 1913. For 1856 and 1873, since there was also no information on time lost due to sickness and strikes, it was assumed that the total time lost due to all three causes was slightly less than in 1913: 3 weeks, as compared with 3.2 weeks.

  4. (f) Average weeks lost per year due to sickness. The basic source of sickness data was National Insurance statistics. Time lost due to accidents or prescribed diseases qualified for Industrial Injury Benefit, and statistics were available in the annual Reports of the Ministry of Pensions and National Insurance. Time lost due to other sickness qualified for Sickness Benefit, and statistics were taken from Government Actuary sources. Again the information was less satisfactory for interwar than for postwar years, and again the only pre–1914 estimate, referring to sickness among members of a benevolent society in 1910, was assumed to apply to 1913.

The figures of time lost due to other sickness exclude absences of less than four workdays. The inaccuracy introduced by this omission is probably not too serious because the Ministry of Labour estimates of actual hours worked by manual workers—used for the estimates in (b) above—exclude time lost through sickness (or absenteeism) if the person is absent for only part of the week. A short illness of only a few days will therefore be reflected in our figures in a shorter working week rather than under sickness.

  1. (g) Average weeks lost per year due to strikes.* Use was made of the Ministry of Labour's annual series of the total number of working days lost per year due to industrial disputes. The series commenced in 1893, and so it was possible to estimate the average number of weeks lost due to strikes in 1913, but not in 1856 or 1873.

Notes:

(*) Normal weekly hours represent the period recognized, in agreements or by custom, as the basis for the payment of the weekly wage rate. It is exclusive of mealtimes. Actual weekly hours may exceed or fall below this according to the amount of overtime or short‐time working.

(*) UK [45]: Table 1. An estimate for males and females together for 1969 of 7.1% is also given (Table 3), for comparison with the 1970 figure for males and females together of 7.4%.

() In the Midlothian survey 38.7% worked 0–5 hours, 16.1% 11–15 hours, and 6.3% over 15 hours (Alden 1971: Table 12). In the Family Expenditure Survey 70.2% of males and 77.9% of females in 1969 worked 1–8 hours, 25.2% of males and 14.7% of females 9–20 hours, and 4.6% of males and 7.4% of females over 20 hours (UK [45]: Table 8).

() These figures were calculated by multiplying the proportion of moonlighters by the ratio of eight hours to the average hours worked per week by all workers (both full‐ and part‐time) in 1964.

(*) In the population censuses of 1961 and 1966 part‐time appears to mean anything that is not normally full‐time; in the 1968 New Earnings Survey it means working less than 30 hours per week (see British Labour Statistics Historical Abstract, UK [48]: Table 143).

(*) For convenience the reference is to strikes; in fact the desired concept, and that which the statistics measure, is all industrial disputes (including lock‐outs, etc.).