Despite a few flaws -- including a brief descent into credulous lump-of-labour rehash -- there is an interesting new paper on reduced work time by two Harvard economists Alberto Alesina and Edward Glaeser, and a Dartmouth economist, Bruce Sacerdote. It is titled Work and Leisure in the U.S. and Europe: Why so Different? and its major innovation is to introduce the notion of a "social multiplier" to the analysis of the utility of leisure. There was a brief discussion of it at catallaxy.
If I understand the paper's argument correctly, it makes an important contribution to the understanding of the working time issue by raising the possibility of a more desirable equilibrium that might be achieved through the regulation of working time. There is an important precedent for a such a conclusion that the authors have overlooked, S.J. Chapman's 1909 article, "Hours of Labour," published in The Economic Journal. Chapman concluded that under competition both workers and firms would tend to accept hours that were longer than optimal, both from the perspective of the workers' wellbeing and the firms' goal of maximizing total output per worker.
Chapman's theory reflected empirical evidence from the late 19th and early 20th centuries, later cited by Philip Sargent Florence, that the change from a 12-hour day to a 10-hour day actually increased total output per worker and that further reduction to an 8-hour day at least maintained output at the level of the 10-hour day. A.C. Pigou also discussed this phenomenon in the chapter on the hours of labour in his 1924 textbook, Economics of Welfare.
Obviously, it would be impossible to certify how much of the increases in output per worker resulted directly from the decrease in the length of the working day. However, Chapman did provide a theoretical explanation from why such a paradoxical increase in output could occur.
Alesina et al. caution that an increase in the cost of labor input per hour would lead to input substitution and thus reduce employment. However, Chapman's theory makes it not so obvious that reduced hours of work at given total wages would increase the cost of labor input per unit of output, which is the salient issue. Of course, it is possible that firms will make the mistake of viewing an increase in labour rates (per hour) as implying an increase in labour costs (per unit) and will behave "irrationally" as Jeffery Pfeffer has suggested (1998, "Six Dangerous Myths about Pay," Harvard Business Review, 76[3], 109-121) but consideration of such behavior would take us away from the province of economic analysis. So it would be appropriate to reconsider the authors' assumptions about labor costs and input substitutions stated on pages 17 and 22 of the paper.
Also on page 22, they make reference to unions' stated policies allegedly being "based on the assumption that the total amount of work to be performed was somehow fixed." I take it that they picked up this characterization of the unions' stated policies from Jennifer Hunt's articles on work-sharing in Germany, which the cite in the sentence immediately preceding. Several years ago I read Hunt's articles carefully and was unable to find any specific evidence in support of several general statements she made about "wide-spread popular beliefs" in a fixed amount of work. The alleged belief in a fixed amount of work is commonly referred to as the "lump-of-labor fallacy". I read Hunt's articles in connection with an investigation of the intellectual history, credentials and legacy of that oddly-named entity (2000, "The 'lump-of-labor' case againt work-sharing: populist fallacy or marginalist throwback" in _Working Time: International trends, theory and policy perspectives_, Routledge). It turns out that there is no substance to the claim that advocates of reduced working time routinely commit such a fallacy and very questionable grounds for the assumption often advanced by neoclassical economists that there is little or no prospect for employment gains from reduced work time. As Gerhard Bosch argues in an earlier chapter in the same book, the prospect for job creation depends as much on how as on whether working time is reduced.
On page 25 of the article, the authors come close to addressing the how rather than the whether of working time reduction and job creation when they remark that "Had the unions accepted a constant hourly wage that [job creation] might have worked..." I would suggest a minor modification. Again, it is not the hourly wage that needs to be held constant but the cost per unit of output, so it could be argued that wage increases that did not exceed productivity gains from reduced hours would offer positive job creation prospects. Of course, it needs at some point to be recognized that those job creation prospects would not be from the "redistribution" of an existing _quantity_ of work but from the dynamic creation of a qualitatively new regime of work and working time -- a regime in which hours are shorter, productivity is higher, costs are lower and employment is greater. That said, talk about redistributing work may be closer to the mark regarding the relation between hours of work and the ideals of life than are pedantically punctilious objections to "views, fallacious or otherwise, concerning the mechanics of distribution. Bad arguments have been used to justify good ends (Chapman, p. 365)."
I do think that the paper's suggestion of a "happier equilibrium" is a welcome step away from what tends to be a sterile debate between two solitudes. I would add, though, that in light of Chapman's theory such a happier equilibrium may also be closer to being economically optimal than the dismal competitive one that is today usually assumed to be optimal.
Posted by sandwichman at May 16, 2005 02:23 PM