Wage forms, costs of living and the urban–rural wage gap: Southern Sweden, 1881–1930

Abstract

One common finding of studies of industrialising economies in the nineteenth and twentieth centuries is the existence of an urban–rural nominal wage gap. For the analysis of its causes and context it is however essential to know the extent to which the gap is due to urban–rural differences in payments in kind, cost of living and consumption patterns. Only a few studies have tried to estimate the real urban–rural wage gap though. This paper belongs to this stream of literature, aiming to develop a method for the estimation of the real wage gap given the kind of data available in the Swedish context, and to use it to estimate the real wage gap for Malmö County in 1881–1930. The main result is that the gap is reduced to about one-half of its nominal size when differences in payments in kind, cost of living and consumption patterns are accounted for. The real wage gap was still substantial though, and the trend was increasing. While urban real wages were ranging from 10% less to 40% more than agrarian wages in the nineteenth century, urban real wages were 30–100% above agrarian wages in the 1920s.

Keywords:

• wage gap
• urban–rural
• real wage
• cost of living
• payments in kind

Acknowledgements

The author acknowledges financial support from Riksbankens Jubileumsfond for the research project ‘Swedish Wages in Comparative Perspective, 1860–2008’, and is also grateful for constructive comments of participants at the Higher Seminar in Economic History, University of Gothenburg, and from three anonymous reviewers.

Notes

¹Squire, Employment (1981), 102. Clark, Conditions (1957), 526–531. Williamson, ‘British’ (1987), 641–678.

²Williamson, ‘Structure’ (1982), 1–54. Williamson, ‘British’ (1987). Hatton/Williamson, ‘Wage’ (1991), 381–408. Hatton/Williamson, ‘Integrated’ (1991), 413–425. Hatton/Williamson, ‘Unemployment’ (1991), 605–632. Hatton/Williamson, ‘Explains’ (1992), 267–294. Hatton/Williamson, ‘Labour’ (1993), 89–109.

³For other contributions to this discussion, see Heikkinen, Labour (1997). Margo, North-South (2002). Mora-Sitja, ‘Labour’ (2007), i156–i177. Borodkin et al., ‘Rural/Urban’ (2008), 67–95.

⁴Hatton/Williamson, ‘Wage’ (1991), 383.

⁵Mora-Sitja, ‘Labour’ (2007), i163–i164. Williamson, ‘British’ (1987), 60. Sicsic, ‘City-Farm’ (1992), 685–686. Heikkinen, Labour (1997), 124. Hatton/Williamson, ‘Wage’ (1991), 401. Alston/Hatton, ‘Earnings’ (1991), 91–95.

⁶Eriksson/Rogers, Rural (1978), 26–36. Olsson, ‘Sockerkapitalets’ (2002), 9–41. Lundh, ‘Statarnas’ (2008), 113–119. Lundh/Olsson, ‘Contract-Workers’ (2011), 298–329.

⁷ Lantarbetarnas(1915), 52–53. Till belysning(1911), 17–20. Sommarin, Skånska (1939), 95–96.

⁸Adlercreutz, Kollektivavtalet (1954), 152–153. Gårdlund, ‘Industrins’ (1966), 315–324. Lundh, Spelets (2010), 49–50, 65–66, 86–87.

⁹ Historisk(1969), tables 6, 7, 12, 13.

¹⁰The agrarian wages are from Hushållnings-sällskapens(1882–1912), Arbetaretillgång(1912–1928) and Lönestatistisk (1931–1932). Workshop wages in Malmö are from Sveriges Verkstadsförening's archives: ‘Ser. E. Medeltiminkomst för samtliga arbetare med specification för manliga 21-59 år samt yrkes-, tempo- och grovarbetare 1914–1930’, Centrum för näringslivshistoria, Stockholm.

¹¹The firm-specific wages are from Bagge et al., Wages (1935) and Kommunalarbetarnas (1933).

¹²Payments in kind are included in the official statistics.

¹³See Table 1. For a detailed discussion, see the text below.

¹⁴Jörberg, History (1972). Konsumentpriser (1961), 20.

¹⁵Those including this information are 1912–1914 (1920) and Allmänna bostadsräkningen (1936).

¹⁶ Levnadskostnaderna(1923); Levnadskostnaderna (1929).

¹⁷The populations of the early household surveys are not fully representative, e.g. nuclear and stable households are overrepresented (Johansson, Levebrödet (1996), 52–55). For this paper, however, this shortcoming of this source is not a major issue, since the surveys are used only to calculate urban and rural prices and there is no reason to believe that the chosen households were biased when it came to reporting the volumes and costs of specific goods consumed.

¹⁸1881–1910: Hushållnings-sällskapens berättelser. 1911–1928: Arbetaretillgång, arbetstid och arbetslön. 1929–1930: Lönestatistisk årsbok.

¹⁹For the period 1911–1930 daily wages of workers with permanent employment are taken, and for the period 1881–1920, where statistics are not specific on the employment type, a series is estimated based on the relation of the pay for temporary and all day labourers in 1910–1911. The series refer to workers without board included in the payments.

²⁰In the late nineteenth century, the typical number of working days was 300 including five days of absence for sickness, etc. (von Feilitzen, Tjänare (1892), 10. Sommarin, Skånska (1939), 118–119).

²¹Nyström, Arbetarfrågan (1932), 59–69.

²²Olsson, ‘Storjordbruk’ (2008), 62–63.

²³Nyström, Arbetarfrågan (1932), 33–49.

²⁴ Levnadskostnaderna(1923), table 3, 16–17. A social report from the region in 1891 gives the same picture. Day labourers faced a higher risk of unemployment, but their factual annual earnings exceeded those of contract workers (von Feilitzen, Tjänare (1892), 9).

²⁵ Lantarbetarnas(1915), 196–197, 210–211, 252–253. Own calculation.

²⁶ Arbetaretillgång, tables 2–3, 86–87, 106–107.

²⁷Malmö County in 1910: 11% (Till belysning(1911), 82–83); two contract workers’ households in 1900: 7–8% (Juhlin Dannfelt, ‘Jordbruksarbetare’ (1906–1911), 415); one contract worker's household in 1891: 11% (von Feilitzen, Tjänare (1892), 59–61. Own calculation).

²⁸Bagge et al., Wages (1935), tables 123, 125, 143, 148, 513–523, 556–560, 565–567. In the section on Helsingborg Rubber Factory, calendar men are grouped among unskilled workers (559). For Reymersholm Copper Work only daily wages are reported. In order to estimate the annual earnings, the daily wage is multiplied by 300, the assumed number of working days, and to that sum another 50 kronor per year are added for perquisites (fuel and medical care) (566).

²⁹ Kommunalarbetarnas (1933), 86–89.

³⁰ Arbetsstatistik(1901), 334. The study contains hourly wages and information about the number of working hours per week. The annual earnings are estimated as the product of hourly payment, 59.5 hours per week for 50 weeks per year, plus payment for 10 hours’ overtime or shift work per working week (own calculation based on tables 15–17 and 24, 238–243, 278–279).

³¹Sveriges Verkstadsförening's archives: ‘Ser. E’. The source records hourly wages. In order to be able to estimate the annual earnings, information about the normal/average working hours in the engineering industry was used: the yearly number of working hours was set at 2850 in 1914–1918, 2600 in 1919 and 2400 in 1920–1930 (Johansson, Den effektiva (1977), tables B3-6, 172–175. Isidorsson, Striden (2001), 51–57).

³² Levnadskostnaderna 3. Malmö (1917), table 1, 74. Levnadskostnaderna 6. Hälsingborg (1918), table 1, 74. Levnadskostnaderna (1923), table 3, 16. The official statistics for Malmö County are 59% in 1920 (Arbetaretillgång (1921), table R, 54–55).

³³Lundh, ‘Statarnas’ (2008), Table C, 158. Lönestatistisk (1931), Table 2, 14–15.

³⁴ Till belysning(1911).

³⁵ Till belysning (1911), 46.

³⁶For a similar observation for Finland, see Haapala, Yhteiskunta (1995), 284.

³⁷ Detaljpriser (1933), table 1, 104–143.

³⁸ Levnadskostnaderna (1923), tables 1 and 3, 2–5, 16–25.

³⁹ Levnadskostnaderna (1923), 7–12. The source does not report statistics separately for Malmö County, only grouped together with Kristianstad and Blekinge Counties under the label ‘Southern Sweden’. Of a total of 50 households from ‘Southern Sweden’ 44 were from Malmö County. Here the analysis is based on records for ‘Southern Sweden’ as a proxy for Malmö County.

⁴⁰Included in the survey were 51 and 48 households in Malmö and Helsingborg respectively (Levnadskostnaderna (1929), tables 6 and 12, 178–181, 196–197). Prices are from Detaljpriser (1933), table 1, 104–143.

⁴¹For rural households the basket represents 86% of the total expenditures for food (1920), and for urban households 75% (1923). Since ‘milk and cream’ is reported as a lump sum in the rural survey in 1920, while the official price statistics includes milk but not cream, the urban price of ‘milk and cream’ is taken from the urban cost-of-living survey in 1923 where a lump sum is reported (Levnadskostnaderna (1929), table 1 and 12, 160, 196). The price of ‘milk and cream’ was adjusted to the price level of 1920 based on the change in the price of milk (Detaljpriser(1933), table 1, 104).

⁴² Allmänna bostadsräkningen (1924), table GG, 117*; Hyresräkningen (1925), table C, 24–25. Own calculations.

⁴³ 1912–1914 års allmänna bostadsräkningar (1920), 242 and table 9.

⁴⁴ Allmänna bostadsräkningen (1936), table 39, 100–101.

⁴⁵Myrdal/Bouvin, Cost (1933), 15. Bagge et al., Wages (1935), 255–267. Detaljpriser (1933), 15. Konsumentpriser (1961), 111–116.

⁴⁶Williamson, ‘British’ (1987), 60. Sicsic, ‘City-Farm’ (1992), 685–686.

⁴⁷Heikkinen, Labour (1997), 124.

⁴⁸Hatton/Williamson, ‘Wage’ (1991), 401.

⁴⁹Alston/Hatton, ‘Earnings’ (1991), 91–95.