Beveridge's analysis of unemployment in 1909: the reserve of labour

Abstract

The aim of this paper is to provide a rational reconstruction of Beveridge's theory of unemployment published in 1909. First and foremost, it shows that his theory of unemployment is coherent – what Beveridge refers to as ‘the reserve of labour’ represents ‘unemployment’ as a whole; unemployment is due to the imperfection of the labour market and associated friction and the organisation of the labour market is necessary. Second, it suggests that as early as 1909, a negative relationship already existed between unemployment and job vacancies and that the segmentation of the labour market and imperfect information are key factors of friction. The first part of the paper provides a reconstruction of Beveridge's theory of the reserve of labour (1909) including causes and factors of unemployment and unemployment policies. The second part shows that certain founding principles of the ‘Beveridge curve’ (Beveridge 1944 [1953]) were already to be found in his 1909 book and that links can be established between Beveridge (1909), Phelps (1970) and Pissarides (2000).

Keywords:

• Beveridge
• the reserve of labour
• labour market frictions
• organisation of the labour market
• Beveridge curve
• search unemployment

Notes

 1 On this subject, cf. Hills et al. (eds.) (1994) and Harris (1977 [1997]).

 2 Komine (2001a, 2001b, 2004) positioned Beveridge (1909) within the history of economic thought. He analysed the theoretical structure of it (Komine 2001a: 5–15, 2001b: 6–7, 2004: 268–9), the reciprocal influences between Beveridge's predecessors and contemporaries (Komine 2001a: 16–24, 2001b: 7–28, 2004: 269–71) and the making of Beveridge's theory of unemployment by focusing on pre-1909 papers (Komine 2004: 256–68). Our article could be seen as a complement to Komine's work.

 3 Hutchison (1953: 415–6), Schumpeter (1954 [1994]: 944), Garraty (1978: 136–41), Backhouse (1985: 247–48), Topalov (1994: 370–75), Harris (1977 [1997]: 166–7), Laidler (1999: 155–7), Komine (2001a: 5–15, 2001b: 6–7, 2004: 268–9) and Boianovsky and Trautwein (2003: 389–91).

 4 According to Hutchison (1953: 415–6), Beveridge's theory of unemployment is generally incoherent.

 5 The phrase ‘the reserve of labour’ can be found in Marx (1867 [1967]: 641), Booth (1892, vol.1: 152) and Hobson (1895a [2006]: 142, 1896 [2006]: 7). According to Garraty (1978: 136), ‘In his book, Beveridge had nothing original to say about the economics of unemployment’. We believe that Beveridge offered a new explanation of the reserve of labour. Although the links between fluctuations in activity, changes in industrial structure and the reserve of labour played an important role in Beveridge (1909) – as they did in Marx, Booth and Hobson – he presented a new thesis that unemployment is due to imperfections in the labour market.

 6 Laidler (1999: 156–7) did not distinguish the irreducible minimum of unemployment from the reserve of labour. And Komine (2004: 268–9) said that, ‘Beveridge could point out the most significant element in modern industry, the irreducible minimum of unemployment, or the reserve of labour’.

 7 Harris (1977 [1997]) provides the only full biography of Beveridge. When he analyses Beveridge (1909), it is on the basis that, ‘By far the most important of Beveridge's activities in the 1900s was his involvement in schemes for the relief of the unemployed’ (Harris 1977 [1997]: 138). Consequently, Harris (1977 [1997]: 166–7) is interested more in the political character of Unemployment (1909) rather than its theoretical structure. We provide a rational reconstruction of Beveridge (1909). This helps to determine why Beveridge suggests organising the labour market through the establishment of Labour Exchanges (the organised fluidity of labour) and Unemployment Insurance (the averaging of work and earnings) (Beveridge 1909: 192–234).

 8 On this subject, cf. Béraud (2008) and Batyra and De Vroey (2009).

 9 On this subject, cf. Garraty (1978), Mansfield et al. (1994), Topalov (1994) and Corry (1996). Beveridge's approach is ‘unique because his recognition of the unemployment problem is exclusively modern, here “modern” means that his analysis goes far beyond those typical of the 19th century’ (Komine 2001a: 6). Beveridge (1909) marks the end of the period of the discovery of unemployment (Garraty 1978: 129–45).

10 Through his role (1909–1919) on the Board of Trade, Beveridge participated in drafting the Labour Exchange Act (1909) and the National Insurance Act (1911); cf. Beveridge 1930: vii-xii. On this subject, cf. Harris (1972, 1977 [1997], 1994, 1996, 1997). Beveridge's concern with political action was based on Mill's argument that ‘there cannot be a more legitimate object of the legislator's care than the interests’ of the unemployed (1848 [1965]: 99); cf. Beveridge 1909: 12–3. Beveridge's method consisted of three steps: (1) empirical work; (2) theoretical analysis and (3) economic policy. Beveridge reworked Marshall's point of view (1890: 24): ‘It is the business of economics…to collect facts, to arrange and interpret them, and to draw inferences from them’.

11 Komine (2001a: 5–15) presents three conclusions concerning Beveridge's Unemployment (1909): it is a unique approach, Beveridge has a clear-cut classification of unemployment and his remedy for unemployment is the organisation of the labour market. Furthermore, Komine (2004: 268–9) considers that the book of 1909 is the culmination of Beveridge's evolved ideas on unemployment. However, Komine does not provide a rational reconstruction of it.

12 On Laidler's opinion (1999: 155), it is impossible to establish any precise date on which unemployment emerged as a separate and distinct topic for economic analysis.

13 On this subject, cf. Corry (1996).

14 According to Komine (2001a: 4, n. 3), ‘Schumpeter simply referred to Beveridge in a single sentence’. On the contrary, Schumpeter (1954 [1994]: 944) positioned Beveridge (1909) within the history of economic analysis: unemployment was treated using the tools that had served the ‘classical economists’; it was a frictional problem (a permanent phenomenon due to its recurrent nature); it included various types of unemployment (due to imperfections, fluctuations, etc.). Theorists working between 1870 and 1914 knew of no unemployment other than the frictional variety. ‘For the period, cf. especially Beveridge, Unemployment (1909)’ (Schumpeter 1954 [1994]: 944, n. 57).

15 ‘The concept of “involuntary” unemployment, as well as the controversies surrounding it, became part and parcel of macroeconomic theory with the publication of Keynes's General Theory in 1936’ (Boianovsky and Trautwein 2003: 385). Yet, economists had made the connection between unemployment and involuntary action before Keynes made it one of the central ideas of his book: George (1879), Marshall (1890), Hobson (1895b, 1896 [2006]), Pareto (1896, Vol. 2), Wicksell (1901, Vol. 1), Fagnot (1905), Beveridge (1909), Webb and Webb (1911), Pigou (1913), Mitchell (1913), Robertson (1915). On this subject, cf. Hutchison (1953), Kahn (1976), Garraty (1978), Corry (1996, 1997), Laidler (1999), Boianovsky and Trautwein (2003: 385) and Béraud (2008).

16 Beveridge's understanding of unemployment as an excess labour supply is taken from Hobson (1896 [2006]). The Problem of the Unemployed (Hobson 1896 [2006]) is the first consistent analysis of unemployment (Hutchison 1953; Topalov 1994; Komine 2004). More generally, it was the standard way to define unemployment in England at the end of the 19th century and at the start of the 20th century (Boianovsky and Trautwein 2003). On the influence of previous literature on Beveridge's analysis, cf. Garraty (1978), Topalov (1994), Boianovsky and Trautwein (2003) and Komine (2004).

17 Beveridge reworked the idea that in economic science, there is always a multitude of interrelated causes (Mill 1867 [1973]: 877–8; Marshall 1890: 26–7).

18 Marshall, 1890: bk. v, ch. ii and bk. vi, chs. i and iv.

19 Beveridge took Mill's ideas (1848 [1965]: 96–7; 570–6) that there is no lack in the demand for the products of labour, and that the demand for commodities is entirely different from the demand for labour. He accepted Mill's version of Say's Law.

20 There are also certain industrial activities in which production levels depend on seasonal factors, such as climatic conditions or social habits, although seasonal fluctuations are only a surface movement which does not deeply affect the aggregate level of unemployment (Beveridge 1909: 24–30).

21 According to Laidler (1999: 155), ‘Beveridge treated the cycle as a factor which alternatingly intensified and mitigated the effects on employment of forces that were always present on the labour market’. In our opinion, Beveridge treated cyclical fluctuations as the primary cause, which, due to the factors of friction at work in the labour market, always implied the existence of unemployment.

22 Although there are two other phenomena (fluctuations of industrial activity and changes in industrial activity), they result in a need for the reserve of labour (Komine 2001a: 12).

23 Beveridge highlights here the existence of a second form of segmentation: the segmentation of the labour market between the different activities (a’).

24 ‘The natural rate of unemployment […] is the level that would be ground out by the Walrasian system of general equilibrium equations, provided there is in them the actual structural characteristics of the labour and commodity markets, including market imperfections, stochastic variability in demands and supplies, and the cost gathering about job vacancies and labour availabilities, the costs of mobility, and so on’ (Friedman 1968: 8).

25 Then, there is the global segmentation of the labour market (between the different activities), which explains the existence of the irreducible minimum of unemployment and there is the local segmentation of labour markets (in each activity), which explains why the reserve of labour is above its minimum level. The first form of segmentation (the irreducible minimum of unemployment) is inevitable; the second form (the glutting of the labour market) is avoidable, and it is the source of the frictions in the movement of labour on the labour market.

26 More precisely, Beveridge wants to reconcile the statement that unemployment is due to labour market frictions with the statement that unemployment is due to general factors affecting demand for the products of industry (Beveridge 1944 [1953]: 59). Unemployment is due to the disorganisation of the labour market and insufficient total demand (Ibid.: 59).

27 Thus, Beveridge argued in favour not only of an ‘adequate total outlay’ (1944 [1953]: 131–66) through state spending (Ibid.: 134) but also of an ‘organised mobility of labour’ (Ibid.: 170) through Labour Exchanges and Unemployment Insurance (Ibid.: 170–5).

28 Pissarides assumes that there is a successful matching function that gives the number of jobs formed at any moment in time as a function of the number of workers looking for jobs and the number of firms looking for workers. The matching function gives the outcome of the investment of resources by firms and workers in the trading process as a function of the inputs. It is a modelling device that captures the implications of the costly trading process without the need to make the heterogeneities and other features that give rise to it explicit. The matching function summarises a trading technology between heterogeneous agents which again is not made explicit (Pissarides 2000: 4).