Shorter working week redux

by Chris Bertram on January 19, 2012

Last week’s nef event on shorter working week, which I blogged about a few days ago, is now available to watch via the LSE channel. Enjoy.



Sandwichman 01.19.12 at 5:38 pm

The Sandwichman is back from his pilgrimage to London and points North and, jet lag permitting will soon be posting some reflections at Ecological Headstand. It was a pleasure meeting Chris, Daniel Davies and a host of luminaries in the shorter working time racket.

While I was away, I received a remarkable reply to my critique (see http://www.lump-of-labor) of the lump-of-labour fallacy. My correspondent (who will remain anonymous pending a response to my request to cite him or her) offered the following explanation for not presenting evidence to support his or her convictions regarding the optimality of current hours arrangements:

“Of course I don’t provide evidence. It is self-evident. Suppose I work for you for 30 hours per week. Suppose you then find out that it would be much more effective to hire 2 people working for 15 hours each instead. Say, because marginal productivity declines very fast. Or because of complementary skills. Or whatever. Either way, what would you do? -You would split the job, of course. Why on earth would you need a government bureaucrat telling you to do you what is good for you? How likely is it that you don’t know how to achieve a productivity gain in your business, while some distant bureaucrat does know?”

Is the above “self evident”? I don’t think so.


Anspen 01.19.12 at 6:29 pm

“Suppose you then find out that it would be much more effective to hire 2 people working for 15 hours each instead.”

His example show why it’s not self evident: how would you find this out? (And that’s of course ignoring all the human foibles, such as unwilingness to change etc., which stop people from acting the way economist dream about).


SusanC 01.19.12 at 6:46 pm

It just occurs to me that “The Mythical Man Month” (see, e.g. the book of that title by Fred Brooks) is the software engineering equivalent of the econmists’ lump-of-labour fallacy:

if a computer program will take one programmer 2 months to write, it does not follow that two programmers can do it in one month. (e.g. there is added overhead because the programmers will need to communicate with each other, and some overheads will be doubled, like both of them will need to read through the projects specification documents).

I think the quote from Brooks is something like “Putting more programmers on a late project is like trying to put out a fire with kerosone.”


QB 01.19.12 at 7:31 pm

SusanC: I can’t find that particular quote in the book, but Brook’s Law, stated there, is: “Adding manpower to a late software project makes it later.”


bianca steele 01.19.12 at 9:48 pm

In fact, (snark follows) if they have Ph.D.’s (even more if they have Ph.D.’s in math or physics, and you may make substitutions as you like) two of them will take well over two months, because each of them will do the whole thing himself, at the end of which time they will discover it doesn’t work. It’s certainly not true that one person is always better than two, and three are always better than five, but Brooks’ Law mostly holds.


Antti Nannimus 01.19.12 at 10:25 pm


Never mind what Brooks said about programming. The classic statement in this regard is “nine women cannot have one child in a month.” It has never been said better, and it never will be.

Have a nice day!


Sandwichman 01.20.12 at 1:24 am

While it is undoubtedly true that there are some tasks that are indivisible, it doesn’t follow that there are no tasks that are divisible. Otherwise there would be no division of labour. In fact, the substitution of machines for workers is founded on the principle that some tasks are so divisible. Since the issue of the shorter work week pertains especially to the effects of labour-saving technology in automating tasks, witty allusions to nine women gestating a child in a month are strictly irrelevant digressions.


billie 01.20.12 at 2:18 am

Another observation from a lifetime in software engineering is that working harder is not as effective as working smarter. There is no point building an elaborate system which does everything the user could possibly want if the the user doesn’t understand what the system does. I could name corporations who have had this problem.

In fact one system gets a passing mention in the Australian film Red Dog – you might have to be an insider to notice it.

Often the best way to develop software is to throw away your first attempt, that said I am not confident that hiring people with no knowledge of the business will lead to a good eHealth outcome.

From a non-software perspective, despite the increases in productivity achieved in manufacturing over the past century, here in Australia domestic housing is still built using craftsmen techniques, which uses a lot of labour, adding to the costs of new housing, about 3 times a couple’s annual income. The alternative has been around since the 1950s, prefabricate house components in a factory and assemble on site.

Professor Tim Watson noted that the service sector provides the opportunity to build an economy that isn’t going to consume non-renewable resources but the workers are burdened by onerous levels of paperwork that audits their activity on top of their primary task, care of patients or education of students.


reason 01.20.12 at 9:31 am

I think you are in danger of hijacking the thread – but isn’t the answer that there is not only a supply side of the market there is also a demand side. That is a calculation based on a single firm and a small set of workers is not the end of the story. There are externalities and there is the problem that a shorter work week for an individual means less income – no matter how efficient they are in that time. And there are co-ordination issues with other workers, and overhead costs in creating and maintaining an employment contract (so short term marginal costs may be a poor guide to total costs).


reason 01.20.12 at 9:32 am

This reminds a bit of Steve Keen’s point that the problem is not such much that macro doesn’t have micro underpinnings, but that micro falsely ignores macro effects.


edo 01.20.12 at 9:40 am

Sandwichman. Let me comment on your here briefly.

First thanks for writing on and researching the history of that peculiar claim. I accept your basic objection against the fallacy claim:

“Whether the demand for labour increases, decreases or stays the same, the work can be divided up differently. Whether it would be advantageous to do so under some particular circumstances is an empirical question, not an a priori certainty or fallacy.”

But I don’t quite follow your reasoning in II on “explanations of the supposed fallacy”.

You there write:
“They are the assumption that, if true (but only if true), would vindicate the claim that the case for work-sharing is based on the assumption of a fixed amount of work” Followed by four bullet points (negations of the four points in your part I).

I don’t see how those assumptions (if they were true, which they aren’t) could do such vindication.

Rather, if any of those four points where true all that would show is that the specific case for work sharing you describe in part I would have one incorrect premise. But that premise would not be a fixed amount of work claim. It would be one of the four points in I.

Or am I missing something here?

As a final thought it seems to me that the “fallacy” talk by work time shortening critics must assume that all arguments for work time shortening are the same or all share one core feature. Isn’t that very assumption also important to challenge by more markedly pointing out the possible diversity of such arguments? If so then your “The case for shorter working time is based on a cluster of core propositions …” could better be phrased “One possible and prominent case for shorter working time is based on…”. In other words, I think you’re right to press that a case for work time shortening will in part depend on empirical questions. But it will also in part depend on a wide range of evaluative stances on welfare, distributive justice, the social bases of self respect, enviromental entities, community, meaningful work and much more. Different mixes of evaluative stances will lead to quite different arguments for work time shortening it would seem. And that is worth stressing as doing so might help in combatting the fallacy claim.


Chris Bertram 01.20.12 at 12:21 pm

_Sandwichman, I think you are in danger of hijacking the thread _

The real discussion thread was the earlier one. I posted the video as an addendum to that (and would have posted it as an update if it had appeared sooner than it did). So I don’t think there’s a thread to hijack here, and I, the author of the OP, am keen the hear what Sandwichman has to say.


Sandwichman 01.20.12 at 3:48 pm

edo: “I don’t see how those assumptions (if they were true, which they aren’t) could do such vindication.”

I’m relying here on analysis I did back in 1999, so forgive me if I am a bit vague on it. Charles Beardsley wrote an article published in the 1890s refuting the fallacy claim made by John Rae, which, according to Beardsley’s critique, itself relied on the exact same assumption of a “fixed wages fund” that Rae purported to discover in the proposition he was criticizing (i.e., that work time reduction could relieve unemployment). My own argument followed similar lines.

The first bullet points in my analysis specify the environmental constraints of the standard production function — automatic adjustment of the economy, unlimited natural resources (due to frictionless substitution), given hours are optimal for output and a labor supply model based on income/leisure choice. Based on all of those assumptions being true, any movement away from the market-determined hours would hypothetically produce a sub-optimal outcome and thus reduce the demand for labor (the amount of work to be done) — with one exception, which is if the demand for labor doesn’t vary with change in the cost of labor (an economically absurd proposition). Relaxing any one of the four constraints invalidates the fallacy claim, however. I presented a mathematical model in my 2000 article to demonstrate.

Now, it would be possible to argue that work time reduction could still create more jobs even if labor demand was adversely affected — as long as people were willing to accept reduced aggregate consumption. That would be a matter of dividing up a smaller pie more equally, though, and thus would be criticized as economically inefficient. I don’t worship efficiency but I still prefer to stick to the economic case when refuting the fallacy claim.


Sandwichman 01.20.12 at 3:50 pm

correction: those standard constraints were the second set of bullet points.


edo 01.21.12 at 12:07 pm

Sandwichman: Thanks. So, to see if I get things right here, you argue that it would only make sense to acribe a “lump of labor” claim to work hour shortening proponents if
(1) the four claims in II are true (which they aren’t, we both agree on) AND
(2) if a background set of normative optimality assumptions from mainstream economics should be decisive norms in this context (which is highly debatable)

You argue that since (1) is false the fallacy ascription is wrong. I agree. I’d still like to see (2) made more explicit in the original reasoning. Since (2) involve contestable normative claims which I’d imagine at least some work hour shortening proponents wouldn’t accept.

Here’s how I’d like to phrase things. Imagine a society with two individuals A and B with this sequence of events:
(t1) A works 10 hours, B 0 hours.
(t2) society introduces a work hour shortening mandate capped at 5 hours
(t3) A works 4 hours, B works 4 hours (not 5/5 because of factor X in the economic system)

State t3 might be better than t1 for a number of reasons. The number of hours worked is less but since maximization of the sum total number of hours worked in the world is not the ultimate moral standard that is no decisive reason for t1 over t3. The t3 state may score higher in sum total hedonistic wellbeing, community, self-respect and social status.

(I think it is also a live possibility that t3 may score higher in revealed preference satisfaction if we compare satisfaction at t1 of preferences generated at t1 with satisfaction at t3 of preferences generated at t3. Which is not the same as investigating how the preferences generated at t1 are satisfied at t1 compare to at t3.)


Sandwichman 01.21.12 at 5:49 pm


Indeed state t3 may be better than state t1 but there are also three output (and thus consumption) possibilities to take into account. Output from the eight hours of aggregate labour in state t3 may be less than, equal to or more than output in state t1. The naive assumption is that aggregate output will be less. Qualitatively, this lesser output of state t3 may still be superior to the greater output of state t1 if it better matches the real needs of the population. Of course if output is greater or equal in state t3’s eight hours to state t1’s ten hours there is no contest about which arrangement is better.

Mainstream economics ignores this for a very simple reason, which Marx revealed and which a great deal of derision and distraction is devoted to concealing: the mainstream is concerned not primarily with output and consumption but with revenue (returns to capital) and accumulation. Production is considered only as a means to accumulation (which in turn is a means to further production and so on ad infinitum). Accumulation will be greater in state t1 precisely because value is accounted for in labour hours (exchange value), not in units of usable output (use values). This value has nothing to do with setting relative prices or “the transformation problem”. It is a macro phenomenon the “microfoundations” of which are legal, cultural, social, political, historical and ideological not “economic”.

Mainstream economics wants us to swallow the fable that wages are determined tautologically by the “laws” of supply and demand (or the marginal revenue product of labour). Social domination has nothing to do with it. Pay no attention to the man behind the screen. Also, please ignore the fact that the “laws” of supply and demand were repealed back in 1869 when it was revealed they confused stocks and flows. And while you’re at it, don’t mention the embarrassment that even in theory the wage is determined by marginal revenue product only under the assumption of full employment.

I’m afraid that the full answer to your question, edo, requires holding two distinct analytical frameworks in mind simultaneously — Marxian AND neo-classical. This is a feat that I find difficult to sustain myself and that would be almost impossible to convey to either a Marxist with little neo-classical grounding or a neo-classical with little Marxist grounding, let alone to folks with only a vague familiarity with either or both. But, thankfully, such a feat of mental gymnastics isn’t even necessary because the fallacy claim is defeated strictly within the terms of the neo-classical analysis. Yes, there is more to the story but you don’t have to know it to clear away the most immediate obstacles.


john c. halasz 01.21.12 at 6:07 pm

“the “laws” of supply and demand were repealed back in 1869 when it was revealed they confused stocks and flows”

Further elaboration please. What happened in 1869?


Sandwichman 01.21.12 at 8:17 pm

1869 is the year of John Stuart Mill’s famous recantation of the wages-fund doctrine of classical political economy in response to William Thornton’s devastating critique of that doctrine. It can be taken as the death rattle of the classical view. Marginalist analysis began to take hold in the 1870s with Jevons, Marshall and Menger’s contributions. There is some question whether Mill totally comprehended Thornton’s fundamental critique — which extended to the laws of supply and demand — or just accepted some of the conclusions that flowed from it. Actually, Thornton’s critique of the laws of supply and demand was presented earlier in the 1860s, it’s just the the Mill recantation was a milestone.


Sandwichman 01.22.12 at 2:59 pm

“…the Mill recantation was a milestone.”

I suppose you could also say it was a Millstone… around the neck of the classical wages-fund doctrine.


David 01.25.12 at 7:18 am

Today job sharing is happening in higher education. It is called part time employment. It threatens the full time employees and dispirits the part timers as they see no way up the ladder to full time. Full time is required to meet basic requirements of health care, loan payment for education and yes food. I’m sure many of those posting and reading here would never make a case that this “sharing of jobs” is desirable.

This is not an academic discussion about a novel theory.
Promoting a shorter work week is a stalking horse for income


Sandwichman 01.25.12 at 10:45 pm

“Promoting a shorter work week is a stalking horse for income Depression!”

My, my, someone’s been reading “Mr. Tufnell’s Report from Lancashire”!


edo 01.25.12 at 11:11 pm

David: Could you illustrate you point with a case? Might it be that you value state (1) over (2)?

(1) person A works 10 hours with the productive output 10X. For the salary A affords full health care coverage and great food. Person B has no job, bad health care and bad food.

(2) person A and B split the job, they work 4 hours each with the productive output 8X (lets just assume that 2 hours, and 2X, are somehow lost through the split). They each afford the same health care coverage and food. It is not bad. But worse than for A in (1). But much better than for B in (1).

I, speaking from the person A position, see a moral reason to choose (2) over (1).


eddie 01.25.12 at 11:51 pm


(t1) Person A works 10 hrs, person B also works 10 hrs, person C works no hours, with A and B having equal contitions including opportunity for advancement.

(t3) Person A still works 10 hrs, person B now works 5 hrs and person C now works 5 hrs, with B and C having little opportunity for advancement.


edo 01.26.12 at 6:44 pm

eddie: If those where the options then I’d favor (t3) on prioritarian grounds. But I doubt that would be the complete option set. To partly channel G.A. Cohen, if it is possible to move from (t1) to (t3) where B’s work is shared between B and C then it is also possible to move to a third state of affairs (x) where both A and B share work with C. Perhaps A currently has enough power, and will use that power, to block a move to (x). But that is no argument against the fundamental desirability of a move to (x). And A himself can’t appeal to his own power (ab)use as a moral reason against moving to (x).


David 01.27.12 at 2:22 pm

Let us step back and look at the trend in consolidation of the number of employees and vendors providing productivity to business.

Employer as controller

There is an assumption that the issue in sharing jobs is the personnel employed accepting the arrangement. No, it is the Employer who is the controller of the issue. The hunt for efficiency (don’t give me more tasks to perform) and profit (less spending today makes management bonuses this quarter and this year) is reducing full time employment and pay in full and part time positions.

Management in private industry and government want less people and companies providing services that they use or are responsible for regulating. This is killing our economy by actively reducing the smaller businesses. Smaller businesses are more flexible but require more effort and oversight to receive services from and for the regulators to regulate. I reduced my departments budget this year give me my bonus! The economy, company, and livelihoods of employed may go to hell in the future but if I do not do this now my employment as manager is gone now.

The smaller businesses that fight taxes actually are putting themselves out of business. They should be fighting corruption, the breaking of law to give extraordinary profits to campaign contributors. There is no law that needs be observered if one has the right people in their pockets.

The government from top to bottom is making their departments more efficient. They are encouraging larger companies providing services to reduce their department costs of supervision and regulation.

Put reduction of full time jobs in you macroeconomic models and see the deflation of income and jobs that result while maximizing profits. I can also see this argument being used to defend income inequality in the political science papers arguing Oligarchy is compatible with Democracy.

Terms to remember:

One who takes resources from others passing on the effort and costs to another’s budget to make their budget look better
X is so efficient that the rest of us are overloaded with work.

It is a situation for maximizing benefits and/or money to the one speaking the word and passing expenses that do not come out of their budget, company or pockets to others.
This term is usually mixed up with the term I Deal as in I deal or I control the situation; if another person or company attempts to negotiate it the response to ideal there usually no deal.

Let me state that this discussion is antithetical to achieving employment that allows a health living for the society. One can pick and chose examples but the issue is not the example but the whole of the masses of employees expected to work (call it sharecroping or dare one call it slavery) in these arrangements.

Comments on this entry are closed.