Sunday, 15 January 2012

Think(?)tank advocates shorter working hours as cure for unemployment.

OMG & WTF: The New Economics Foundation (NEF) is advocating shortening the working week to about 20 hours as a way of reducing unemployment! Not that one again!!!

To be fair, the NEF do advocate a number of other advantages to a shorter working week, such as reduced stress and reduced consumption of the world’s scarce resources. I have no quarrel with that. Indeed Jesus, 2000 years ago, and Buddha six hundred years before that were advocating people should be less concerned with worldly wealth. Perhaps they were right.

The NEF’s (hopefully temporary) departure from sanity would not matter if the rest of the world was clued up as to the basic flaw in the “shorter working week as a cure for unemployment” myth. Unfortunately it’s not.

Certainly the above Huffington article doesn’t get it, nor does this respectable economics tutorial site.

Even Robert Skidelsky, an economist I normally greatly respect, seems to have fallen for this nonsense.

The fallacious lump of labour fallacy.

The most popular rebuttal of the shorter working week argument is to cite the so called “lump of labour” fallacy. Personally I find the lump of labour fallacy very muddled and unconvincing. Here is just one reason.

The lump of labour argument normally accuses advocates of the shorter working week of assuming (to quote Wiki) that “the number of hours of labour per day that are demanded by the market is constant”.

Well it’s blindingly obvious that advocates of the shorter working week are NOT MAKING that assumption. To illustrate, suppose the entire workforce of a country cuts working hours from 40 to 20 hours a week, and let’s say unemployment is at the maximum it has reached in the US in the last three years: about 10%.

Let’s also make the over-simple assumption that the aim is to reduce unemployment to zero.

Achieving this end in the above hypothetical scenario quite clearly DOES NOT require there to be a CONSTANT demand for labour hours. Indeed, if there were such a constant demand, then demand would be far too high!!! Here’s the maths.

The original number of hours worked involved 0.9 of the workforce working 40 hours. 0.9 x 40 = 36. The new number of hours will be 100% of the workforce working 20 hours. 1.0 x 20 = 20. Thus the new total demand for labour hours needs to be 20/36 = 0.55 of the original total number. QED.

Conclusion: the lump of labour fallacy is itself fallacious!

The real flaw.

Now for the real flaw in the shorter working week argument.

The REAL flaw in in the shorter working hours is not actually a hundred miles from the above Wiki one, and it is thus. The shorter working week argument assumes aggregate labour supply can be reduced relative to aggregate demand for labour with no inflationary consequences. Now if that’s the case, then aggregate demand for labour can presumably be increased relative to aggregate labour supply with no inflationary consequences! In short, if the objective is to reduce unemployment, why not just bump up demand?

Note that what I called the “real flaw” in the shorter working week argument is actually the same as the Wiki argument, but put in more general terms. I.e. the Wiki argument is a PARTICULAR CASE of my “real flaw”.

Anyway, if you are convinced, then read no further. But for those who want a more detailed explanation, read on.

More details.

As unemployment falls it becomes increasingly difficult for employers to find the types of labour they want. And when it falls far enough, too many employers resort to poaching staff from other firms rather than take labour from the ranks of the unemployed.

This poaching may be conscious or it may be entirely unconscious. E.g. the more difficult it is to find specific types of labour from the dole queue, the more an employer is likely to spend on advertising for such labour, which inevitably tends to draw a labour from other firms, all of which tends result in the price of labour rising too fast in nominal terms, and that equals inflation.

That lack of the right type of labour on each local labour market is a simple statistical phenomenon: that is, the smaller the number of unemployed, the less the likelihood of any given employer being able to find the labour they want from the dole queue.

Now if one cuts working hours, that has NO EFFECT WHATEVER on the likelihood of employers being able to find the labour they want from the dole queue at any given level of unemployment. Ergo NAIRU or the “natural level of unemployment” or the “inflation barrier level of unemployment” to paraphrase Bill Mitchell remains exactly where it was.

I’ll put that differently just to clarify. If at unemployment level X, the last plumber disappears from the dole queue in a particular town when everyone is doing 40 hours a week, then the last plumber will also disappear from the dole queue at exactly the same level of unemployment when everyone does 30 or 20 hours a week. (That assumes of course that the PATTERN of demand for labour on each local labour market remains unaltered by the working hours reduction: i.e. I’m assuming that the demand for plumbers RELATIVE TO the demand for other skills remains unaltered, which is not a wildly unrealistic assumption.)

Conclusion: shortening the working week does absolutely nothing to improve the unemployment / inflation trade-off. Therefor unemployment will be

Immigration raises labour supply – shock horror.

A curious feature of those who advocate a shorter working week is that my hunch is they are the same sort of people who claim that immigration does not raise unemployment: (left of centre / liberal / politically correct, etc).
So they’re saying that in the case of immigration, increasing aggregate labour supply DOES NOT raise unemployment, but were we to increase labour supply by increasing the working week from say 20 hours to 40 hours, that WOULD INCREASE unemployment. A slight self-contradiction there, I think.


Afterthought (same day): Here’s Tim Worstall’s take on the NEF. Tim Worstall’s blog is worth following. His style is very sarcastic, abusive and witty. He uses three and four letter words with regularity. But he does understand economics, so I keep an eye on his blog with a view to knicking his ideas and presenting them as my own :-). Imitation is the sincerest form of flattery.

Afterthought (22nd Jan). I said above that cutting working hours has no effect on the “likelihood of employers being able to find the labour they want from the dole queue”. On second thoughts, reducing working hours involves those who work for employment agencies (private and public) ALSO reducing their working hours. Such agencies will thus devote less time per week to each of their customers. Thus the above mentioned “likelihood” will DETERIORATE. Hence reduced working hours will actually RAISE unemployment all else equal.



  1. Ralph,

    Some good stuff here. What you identify as the "real flaw" in the case for shorter hours is what Richard Layard et al. called "the lump of output fallacy" back in 1990. I would agree that the critique comes closer to what most advocates of shorter hours believe than the "fixed amount of work" canard. However, it too relies on theoretical arguments that are less than robust.

    I tried to address the inadequacies of Layard's lump-of-output critique in my 2000 article, "The 'lump of labor' case against work-sharing" as did Bob LaJeunesse (more exhaustively) in his 2007 book "Work Time Regulation as Sustainable Full Employment Strategy."

    The argument is pretty complicated but basically it boils down to the observation that Layard (and you) implicitly assume a theoretical optimization of the given hours of work, yet the actual theoretical argument in neo-classical analysis (Sydney Chapman's) and two centuries of empirical observation suggest otherwise. In short, the market does not optimize hours for output and thus no conclusions can be theoretically deduced about the inflationary effect of shorter hours.

    There is one notorious instance in which, in Keynes's view, the established practice of short-time working in the Lancashire cotton manufacturing industry had the consequences you anticipate here. It has even been argued that the Lancashire case was the empirical foundation for Keynes's General Theory. But Keynes himself was clear that it was not short-time working per se that he objected to but the particular circumstances in which it was being resorted to.

    To sum up, the question of whether work time reduction or demand stimulus would be superior with regard to price stability, unemployment and output is an empirical one that depends on the circumstances at the time. I believe there is substantial evidence to support the view that AT THIS TIME, work time reduction would be the most prudent policy. That evidence has been presented elsewhere (see Working Time: International trends, theory and policy perspectives) but it boils down to the fact that demand management has played itself out as "one-trick pony" policy and thus is too vulnerable to being gamed. It's time to try a new, unexpected trick.

  2. Sandwichman, I’ve just been thru your paper. Chapman, whose ideas seem to be the basis of your argument, claims that working hours are excessive in that a 20% or so reduction in working hours would not cause much of a reduction in output because employees would not get so tired.

    That point is backed by recent evidence. I can’t quote chapter and verse, but someone looked at the output from employees switching from about 40 hrs/wk to about 60hrs/wk recently. They found that initially output rose, but after about two weeks, the employees were so knackered, that output per employee per week reverted to the initial 40 hr level.

    Thus on social grounds there might be a case for limiting working hours. On the other hand I don’t like the idea of governments telling people what they can and cannot do, so I have mixed feelings about that.

    But none of that contradicts my basic point, which is that the real problem is LOCATING specific types of labour on each local labour market. To repeat, if the last plumber disappears from the dole queue when unemployment is at X% and everyone is doing 40 hrs a week, then the last plumber will also disappear from the dole queue when unemployment is X% when everyone is doing 30 or 20 hours a week. Therefor NAIRU remains unchanged.

  3. Ralph,

    I also don't like the idea of governments telling people what to do. But there is an alternative to the market vs. government regulation dichotomy. It is what Elinor Ostrom calls "common pool resource management". There are extremely good grounds for regarding labour as a common pool resource rather than an private good, namely the social aspect of nurturing and education of the young.

    I agree that there has to be co-ordination between the demand and supply of specific types of labour. But that is hardly an insoluble problem, especially if you're not relying on some kind of one-size-fits-all, inflexible rule. Let's not exaggerate either the difficulty of coordination or the ease with which the market 'automatically' accomplishes it.

  4. Re matching supply and demand for different types of labour being “hardly an insoluble problem”, anyone who can substantially improve the efficiency of the matching process at minimal cost will get a Nobel Prize. Put another way, anyone who can substantially reduce NAIRU deserves a Nobel prize. I think that problem is a lot more difficult that “hardly insoluable”.

    But to repeat my basic point, there is no reason to suppose that cutting working hours improves the matching efficiency or “labour market efficiency”, or to put it a third way, that it reduces NAIRU. Therefor there is no reason to suppose that cutting working hours reduces unemployment.

  5. "there is no reason to suppose that cutting working hours improves the matching efficiency..."

    This is perhaps an instance of the Andy Rooney maxim that "People will generally accept facts as truth only if the facts agree with what they already believe."

    1. the effects of reducing working time on efficiency CANNOT be deduced from axiomatic principles.

    2. the effects are an empirical, not an abstract theoretical matter (the same is true for "NAIRU" -- see especially Jamie Galbraith on this point).

    3. there are substantial empirical grounds for supposing that, at the present time, an intervention to facilitate work time reduction would lead to improvements in both efficiency and aggregate employment.

    4. one of those supporting grounds has to do with the artificial barriers to work time reduction that have been imposed through the statutory and customary framework governing per-employee "quasi-fixed" labour costs, such as employer-paid medical insurance premiums, administrative overhead, etc. Quasi-fixed costs are NOT fixed. They are QUASI-fixed.

    Please be aware that Oi's analysis of quasi-fixed costs had its foundation in John Maurice Clark's analysis of overhead costs, which in turn was influenced by Cecil Pigou's analysis of market failure in labour costs and specifically working time, which analysis was copped from S.J. Chapman's "Hours of Labour."

    To summarize: Oi <- Clark <- Pigou <- Chapman. The irony is that you are assuming a market outcome of optimization that is antithetical to the ultimate premises of your analytical argument, which is market failure. Let me repeat: Chapman's "Hours of Labour" begets Pigou's Economics of Welfare, which begets Clark's Studies in the Economics of Overhead Costs, which beget's Oi's quasi-fixed costs. It's market failure all the way down. Ptolemy, meet Copernicus.

    There is AMPLE reason to suppose that cutting working hours improves the matching efficiency.

  6. As a postscript I should add that it is not just market failure I am talking about but market failure AMPLIFIED by perverse policy incentives.

  7. Sandwichman,

    Re your No 1, the sciences are full of examples of major breakthroughs made by purely theoretical reasoning based on axioms. As long as the axioms are not patently unrealistic, there is much to be said for that approach. Numerous advances in Physics have been made this way. Einstein’s theory of relativity and the existence of numerous subatomic particles were inferred by theory before the supporting empirical evidence was available.

    Re No. 2 and 3 and what you claim to be “substantial empirical grounds” for your ideas, I can’t find a single reference to empirical evidence in your papers that supports the idea that shorter working hours cuts unemployment. But I’m sure you’ll put me right if I’ve overlooked something. The two papers are, 1, “Missing: the strange disappearance…” and 2, “The Lump of Labour case….”.

    Moreover, it is notoriously difficult in many cases in economics to make any empirical evidence stick. E.g. how does one prove or disprove the effect of the recent French experiment with shorter hours? Darned if I know.

    Next, you claim that Chapman had something significant to say about working hours and unemployment. Just one slight problem there: the word “unemployment” does not appear at all in the first paper! Nor does the word appear in your summary of Chapman’s views in the second paper (p.11-14).

  8. Ralph,

    You are right about it being notoriously difficult in economics to make empirical evidence stick. Gerhard Bosch had a discussion of this in the same volume as my "lump of labor case" chapter appears. Anders Hayden wrote about the French 35-hours laws and the methodological issues regarding the measurement of effects. In fact, Chapman did address unemployment and "short time working" but not in his Hours of Labour paper. See his The Lancashire Cotton Industry (1904) and Unemployment (1914) by Chapman and Hallsworth.

    Look, Ralph, I am under no illusions about persuading those who are predisposed to dismiss the employment/shorter working time relationship. Plausibility is in the eye of the beholder. I just want to disabuse you of the farcical notion that, based on a glib formula, you know more about what advocates of shorter work time know and think than they do.

    If you had done ten or fifteen years of archival research on the thought of shorter work time advocates and could rattle off the top of your head what is wrong with Dorothy Douglas's assessment of Ira Steward's Political Economy of Eight Hours, you'd have more credibility with me. But I don't suppose you're familiar either with the thought of Dorning Rasbotham or E.C. Tufnell, whose writings were formative of the lump-of-labor claim.

    Perhaps you have an explanation for why, if the case against shorter working time as an employment generator is so cut and dried, economists keep falling back on an archaic and utterly baseless lump of labor fallacy slogan that you admit is itself a fallacy?

    I think it was Bertrand Russell who said something along the lines of "all scientific theories are false; some are less false than others." I suppose it would be impossible to PROVE that shorter working time can create jobs. But I also think it would be MORE impossible to prove that it doesn't.

  9. Why do economists keep “falling back on an archaic and utterly baseless lump of labor fallacy slogan”? Good question.

    First, I don’t think the idea is “utterly useless”. In the above post I said I thought that the lump of labour idea was nearly right. I said it is a PARTICULAR INSTANCE of what I think is the correct formulation.

    Another factor leading to the chaos in this area is that economics is a huge subject. You can probably do a three year economics course at university without ever seeing the lump of labour idea mentioned. That means that 90% of economists do not have time to think about it, and get all the details right.

    Plus there is the fact that 90% of the human race are robots: they’ll do and think whatever some authoritative voice tells them to do and think. The standard and “authoritative” rebuttal of the shorter working week idea is to chant the phrase “lump and labour”. So every time the shorter working week is mentioned, a chorus of voices intone the phrase “lump of labour”.

  10. My hunch is different than yours. It seems to me that the "aristocratic" principle, being much older than scientific inquiry prevails. Economists are respecters of persons, first and only secondarily engaged with ideas.

    "And then we told them wealth would 'trickle down,'" is a variation on "there is not so great a difference between the real interests of the rich and the poor," which was Dorning Rasbotham's 2nd principle (Thoughts on the Use of Machines in the Cotton Manufacture [1780]). Number 11 was that "there is only a certain quantity of labour to be performed" is a false principle. Rasbotham, the country squire and worthy magistrate, strikes me as a humane and enlightened upper-class Englishman of the 18th century. So too, Josiah Wedgwood, who, as his biographer explained, entertained, along with many of his class, "a strange kind of quasi-religious belief that changes in economic conditions were divinely inspired and that to fight against such forces was a form of impiety."

    Edward Carleton Tufnell was a bird of a different feather. He was an implacable foe of trade unionism and spared no excess of overblown rhetoric in opposing the Ten Hour Bill (see "Mr. Tufnell's Report from Lancashire" (1834) Factories Inquiry Commission. Supplementary report of the Central Board of His Majesty's commissioners appointed to collect information in the manufacturing districts, as to the employment of children in factories, and as to the propriety and means of curtailing the hours of their labour.)

    These are the loci classicus of what came to be known as the lump-of-labour fallacy. Everything after was variations on a theme and mighty thin gruel. Tufnell and Rasbotham are meat and potatoes.

    I'm not saying you should immediately abandon your view that the fallacy claim is a particular instance of "the correct formulation." I just think you should have a solid idea of what the core argument is before attempting your own Rube Goldberg improvisations on it.

  11. Actually, maybe my hunch is the same as yours: "90% of the human race are robots: they’ll do and think whatever some authoritative voice tells them to do and think. "


Post a comment.