- “I’m not sure I follow the arithmetic here.”
- “It’s all down to the numbers – something the article avoids and so is just pie-in-the-sky.”
- “That clearly does not add up.”
- “If you produce X in 30 hours you will produce > X in 40 – unless you are just sitting on your arse for the extra 10 hours.”
- “If you work 40 hours your total output will be higher than if you work 30 hours – unless you are actually destroying output in those extra 10 hours.”
“Just doesn’t add up'”
“Just doesn’t add up'”
Opponents of shorter working week were confident that the relationship between hours of work and units of output was “a simple sum in arithmetic” the result of which was patently obvious to them: “If you produce X in 30 hours you will produce > X in 40 – unless you are just sitting on your arse for the extra 10 hours.” Some critics explicitly assumed that a 20% reduction in hours would have to be made up by a 25% increase in output to be economically feasible. Others conceded that a reduction in hours could result in an increase in hourly productivity but insisted that total output would inevitably be substantially less in the shorter week.
There are several subtleties to the mathematics of shorter hours that opponents systematically overlook. One is that the relationship between hours and output is dynamic and cumulative, not static and instantaneous (“fatigue and unrest”). Second is that the value of aggregate output isn’t necessarily proportional to the quantity of output (“diminishing marginal utility”). Third is that the lost value of foregone leisure and additional stress and “wear and tear” to a worker has to be reckoned against the value of additional income (“opportunity cost”). Fourth, a shorter standard working week may shift proportions of income going to labor and to capital, respectively. And, finally, that shift in the distribution of income is likely to have an impact on final demand (“propensity to consume”). All of these subtleties are in addition to the more generally — albeit, not universally — acknowledged fact that physical output doesn’t necessarily increase in proportion to hours of work even in the static case.
To make a long story short, critics of the shorter work week proposal commit the same errors that adherents to the wages-fund doctrine did in the 19th century and that led Nassau Senior to insist that the profits of enterprise were entirely due to the last hour of the working day. It is none other than the original, the actual, the one-and-only “lump-of-labor fallacy” committed obsessively and unabashedly by opponents of shorter working time!
Very interesting looking report. The cult of long hours has always baffled me, do people really lack awareness of the relationship between the quality of their work and how many hours they have put in? I know from my own work that if circumstances push me to pull a few long days if I look back over it after having several days to unwind I end up asking myself what the hell I was thinking and having to go back and waste time correcting the relatively poor job I did, assuming I have the opportunity. I haven’t seen any evidence that I am unusual in this. My general impression is that very few businesses have good metrics regarding the hours spent correcting mistakes. When I talk to the really hard driving type of people they always seem to be complaining about how they spend all their time putting out fires rather than doing their actual job.
Now, if they are talking about aggregate output I am sceptical that there are gains to be made from dropping from 40 to 30 hours. I haven’t read a ton in this area, but what I have read generally shows the per hour productivity gains of going from 40 to 30 hours is not large enough to fully make up for the loss in hours. This would still mean that the aggregate productivity losses would be substantially less than arithmetic would indicate. What I have read elsewhere does generally agree that productivity drops off sharply after 40 hours a week so it makes sense here to push back hard against the tendency for business to push salaried workers to longer hours.
Though it has been a few years since I read on the subject and my reading was serendipitous since it’s not my actual field so I may be either out of date or just wrong. I would also add that it seems the effect would depend on the job and its specific duties, fatigue impacts capabilities differently, for instance creative work and decision making seems to decline more quickly than routine work (I have never understood why CEOs brag about 80 hour work weeks, I wouldn’t trust a waiter working 80 hours to get my order right, why do we think it’s a good thing for someone making million dollar decisions to be working 80 hours?)
On the whole, however, 4 day work weeks are clearly worth the trade offs from a societal standpoint. It will take legislation, however, since employers will not capture enough from the per hour productivity increase to offset the loss in labor intensity for them to do this on their own.
There was an interesting study on high level decision making work. They used parole boards. Found that later in the day and I think later in the week boards were more likely to deny. This for in with other findings that people who are more tired will, unconsciously, avoid making decisions. When forced they will prefer the less revokable option. In parole that was denial, the person would just come up again in three months.
It has pretty dire implications for project management during overtime. When a team is behind they need to make cuts and commit to specifics so they can finish. But managers working too many hours may have a stronger tendency to ‘keep their options open’ by not cutting or putting off decisions.
“No-one on his death bed ever said, ‘I wish I’d spent more time at the office.'”-anonymous
Does anyone else remember those full page, mainly text ads for the 4Day Tire company in Los Angeles. They were a tire chain only open four days a week, and they made a big point of this in their advertising. I think they finally folded when the founder died, but they were competitive until then.
There are some types of work that are not compressible. An assisted living center or fire department promising 24 hour coverage cannot do that with 16 hour days. A one hour massage takes about an hour including set up and break down. Interestingly, this kind of work is non-exempt, so pay is by the hour, and the length of the non-exempt work week has been falling for decades. Exempt work, salaried work, however, is not charged by the hour, so the work week for more educated and better paid workers has been increasing. If the marginal cost of an additional hour of work is zero, it doesn’t matter if actual productivity is zero as well.
That “25% increase in productivity” came from automation. The output that used to require 40 hours of labor now requires only only 32 – because of automation. The 8 hours freed MAY result in added output IF demand for that output exists. That’s the individual businessperson’s decision to make. Any added earnings from any added output rightly belongs to the businessperson – because they invested in the automation. (However, the enlightened businessperson will share some of those added earnings with their employees.)
All I will say is Kellogg.
And yet it is disingenuous to start straight from the assumption that every {employer, employee} pair has failed to identify this particular mutually advantageous improvement. Out of your 5 “subtleties to the mathematics of shorter hours”, none of the first three remotely qualifies as an obstacle to bilateral contracting, so to buy the argument I must believe that everyone is sitting on a free lunch.
Points 4 and 5 sketch a potential “multiple equilibria / coordination failure” story: if *everyone* switched to a shorter work week, the resulting distributional changes toward labour would trigger large income effects so that workers are happy with more leisure and the cap is no longer needed once we all get there. That story is plausible and does not rely on us all buying that every boss in the corporate world has ever thought about “fatigue and unrest” and that no worker has ever thought that they might be happier with fewer hours.
However, I fail to see why it is obvious that reduced hours should lead to a higher labour share (as opposed to more employment at reduced hours, less production overall, or other effects due to the granularity of hours choice and leisure and the details of how you implement the cap), and the wealth effect would have to be really large.
LucB:
Welcome to Angry Bear. First time comments go to Moderation to weed out spammers and advertising.
No, it’s not “disingenuous” because it is consistent with S.J. Chapman’s analysis of the non-optimizing aspects of the determination of hours under competition. In game theoretic terms what you have here is similar to a prisoner’s dilemma (or tragedy of the commons) in which each individual acting in their own “self interest” results in a less than optimal outcome.
I don’t think it is “obvious” that reduced hours should lead to a higher labour share. I think it is a possibility that is routinely overlooked or ignored. Ira Steward’s eight-hour theory made a case for why it would and in the early 20th century John R. Commons concurred with that expectation, as did Dorothy W. Douglas.
My position is that the relationship between output and the hours of work is INDETERMINATE, which is not, however, to say the relationship is UNIMAGINABLE. The arguments against shorter hours are unimaginative and SYSTEMATICALLY disregard otherwise standard economic arguments such as diminishing returns and diminishing marginal utility.
“I’m not sure I follow the arithmetic here.”
“Unless productivity goes up by at least 25% to compensate, everyone will be worse off.”
“Dropping hours from 5 days a week to 4 means that the work that would have been done in 5 days now needs to be done in 4, which means each day needs a 25% increase in productivity. Where on earth do you think such an increase is going to come from?”
“OK, but to arrive at the same output in 4 days rather than 5 means that people have to become 25% more productive than they are today. That’s an awfully big jump in productivity. I’m not convinced that people today are that unproductive. Certainly when I think of my past workplaces, I don’t think my colleagues were that sub-optimal. A 10% increase in productivity seems more reasonable.”
Following up on the above post about comments in the Guardian, here are some thoughts about “where on earth” a productivity increase of 25 percent might come from:
Let’s start from a 40-hour week in which the rate of output declines somewhat toward the end of the day when workers are beginning to tire. Let\s assume the least productive eight hours of work produce only 75 percent of the output of the most productive 32 hours of work. Call the average output of the most productive 32 hours “one unit” of output. Total output for 40 hours work is 38 units.
Now, reduce the weekly hours to 32. Better rested, more motivated workers result in a “reasonable” 6.25 percent increase in average hourly productivity above and beyond the productivity gain from eliminating the least productive hours. Total output in 32 hours is now 34 units compared with 38 units previously produced in 40 hours.
Those are physical units of output not the monetary value of that output. Assume diminishing marginal utility of the total output. The last four units of output add less value per unit than the first 32 units. So let’s say in value terms 34 units = $34 but 38 units = $37. We are now producing 92 percent of the value previously produced in 80 percent of the time.
Instead of earning $15 an hour for a 40-hour week, a worker now earns $17.23 an hour for a 32 hour week. That’s a 15 percent increase in hourly wage coupled with a 20 percent decrease in weekly hours. Not too shabby! Considering that there are costs associated with commuting to work, etc., the 92 percent retention of weekly income might effectively be closer to full compensation.
All of the above, of course, is simply the fleshing out of assumptions. We assumed diminishing productivity in the last hours, we assumed heightened productivity from a shorter working week and we assumed declining marginal utility of goods and services produced. Finally, we assumed a preference for free time over a vanishingly small increment of total income. The point is that each of these assumptions were relatively modest but when combined “add up” to a rather substantial cumulative result.
“All of the above, of course, is simply the fleshing out of assumptions. We assumed diminishing productivity in the last hours, we assumed heightened productivity from a shorter working week and we assumed declining marginal utility of goods and services produced. Finally, we assumed a preference for free time over a vanishingly small increment of total income. The point is that each of these assumptions were relatively modest but when combined “add up” to a rather substantial cumulative result.”
Yes.
I can hardly image my explaining the hunt hare or stag dilemma. Hell, they can barely understand cutting lot size to free up capacity and increase shop floor flow to deliver sooner and lower cost. Great conversation.
The unasked question is the optimum leisure period. If the goal of work is leisure, then the authors say that two days is not enough. We get nine day weeks and 41 weeks a year. What!
The economy adapts to reduce transaction counts, maximize inventory use. The season is fixed, the number of repeated trials is undetermined, but likely synchronous with occasional holidays. Don’t need economics to tell you this, you are selecting the number of weeks per year, take it from that angle, optimize that. If nothing else, we are seasonal.
Sandwichman: thanks for replying.
I think we have the same idea in mind where I talked about “coordination failure / multiple equilibria” and you talked of a “prisoner’s dilemma / tragedy of commons”: a situation where the move to a different labour hours level could make everyone better off, yet no one (rather, no {employer-employee pair} deviates unilaterally. However I don’t think the term Prisoners dilemma is appropriate in this context because PD is a dominance-solvable game, where incentives to act do not depend on what you expect others to do, and where the equilibrium is unique, whereas I think what you want to decribe involves multiple equilibria.
The “stringency of the hypothetical regulation” test tells you what type of underlying rationale for regulation you have:
– if you think people are just failing to grab a free lunch, only persuasion is needed to move toward lower hours.
– if you think there are multiple equilibria (one with low and one with high hours, presumably welfare-inferior), coercion is needed to move away from the current equilibrium but once we all settle at lower hours, we will not need the cap anymore as no one will want to unilaterally move toward larger hours.
– if you think that coercion will still be needed then, you have a genuine Prisoners’ Dilemma, but then it seems to me your story would have to rely on some externalities. I don’t see them in your argument.
My standard economic argument for dismissing the “free lunch” angle is basic symmetric information contract theory. There is nothing stopping an employer/employee pair from reaping the rewards of that diminishing marginal productivity and increasing marginal opportunity cost of work, even with large monopsony power. If you are going to argue that all employer/employee pairs fail to optimise their gains from trade, you have a high burden of proof for why people are not merely failing to optimise but are failing to optimise *SYSTEMATICALLY AND ALL IN THE SAME DIRECTION*.
If instead your story is a multiple equilibrium one, as discussed above, you need to back it up with an argument for why both the redistribution toward labour income AND the income effects of that redistribution (increasing massively the demand for leisure) are large. You need both for a multiple equilibrium story to be plausible.
If you are arguing that current hours represent coordination failure and you want to shift toward lower hours, it’s going to require coercion, temporary for a multiple equilibrium story and permanent for a PD story. If you want to convice people that the coercion is worth it, you will need a lot more than calling them unimaginative.
LucB,
You make a good case for the multiple equilibrium angle. Robert Prasch developed that argument compellingly in “Revising the Labor Supply Curve: implications for work time and minimum wage legislation.” I’m not sure I follow your point about PD, though, “where incentives to act do not depend on what you expect others to do.” In PD incentives DO depend on what you expect others to do. Similarly, working time arrangements also depend on what the actors expect others to do.
John Pencavel also has done studies of worker cooperatives in which hours of work adjusted more optimally to market conditions than in employer/employee firms.
On the question of “coercion,” I would stake my claim with Robert Hale that there is no “non-coercive” transaction (“Coercion and distribution is a supposedly non-coercive state”). There is only coercion of which one approves and coercion of which one disapproves.
I don’t expect to persuade anybody of anything they disagree with. I only want to lend encouragement and give ammunition to those who suspect it would be a good idea but don’t know how to articulate it. If in the process I happen to change some minds, that’s a bonus.
A PD is dominance solvable. If I expect the other prisoner to squeal, I should squeal. If I expect him to keep quiet, I should squeal anyway. So it does not matter what I expect him to do, I have a dominant strategy and I should squeal. A stag hunt game is one where n players choose whether to hunt stag or hunt hare. If all hunt stag, no one wants to hunt hare instead, so “All stag” is an equilibrium. If you expect enough partners to hunt hare, your band won’t catch the stag anyway so you should hunt hare too. “All hare” is an equilibrium, where everyone is worse off than in “All stag”. That is different from a PD where the pareto optimal profile of strategies (quiet, quiet) is not an equilibrium,
In such a context, policy has to do more than move people to the new equilibrium, it has to permanently change incentives. Or we’re back to squealing.
Based on your comment on illustrative numbers about productivity, it seems you are in the ” free lunch” camp, even though you might dislike the label.
“A PD is dominance solvable. If I expect the other prisoner to squeal, I should squeal. If I expect him to keep quiet, I should squeal anyway.”
Only if one assumes “all else is equal.”
Oh, I am definitely in the “free lunch” camp. I’ve had more than my share of free lunches and I’ve witnessed free lunches that I could never hope to taste. Have you never heard of “inheritance”?
To clarify. PD implies a multi-stage decision process. The first stage is the evaluation of the incentive structure. “If I expect the other prisoner to squeal, I should squeal. If I expect him to keep quiet, I should squeal anyway.” Second stage is projection of this incentive structure onto the other prisoner’s choice: The other prisoner will squeal because “If he expect me to squeal, he should squeal. If he expects me to keep quiet, he should squeal anyway.” So it is no longer a choice between the best outcome for me and the second best; but a choice between the worst outcome for me and the second worst. My expectation of the other prisoner’s most likely choice has changed my incentive structure from opportunism to self-protection.
Incidentally — and this is a bit of a digression — experimental PD games, iterative and multiplayer games are interesting in the ways that they deviate from the presumed outcome based on incentive structures of two-player games. Anatol Rapoport contributed a great deal to the experimental and iterative games literature. I take it the stags and hares game is a variety of multi-player PD.
Sandwichman:
If this is directed to me, thank you. If not, thank you anyway for the explanation.
As I explained it to one manufacturer, there is a sphere of costs you fit into and must be less than in order to be successful. Each manufacturer has its own sphere within the greater sphere of competition or costs. If we are all the same than we are equal and compete on an even plane. If one lessens their costs, you can still exist if you have a uniqueness to your product, better quality, or can deliver sooner; but, what happens if the products are similar and the business moves? You are forced into making a change in your cost sphere by making a similar change in technology, process, costs, or take less profit.
The over riding factor today or the larger sphere of costs is global competition and whether it is worthwhile to make the change or get out of the market altogether as the US Overhead is too high and usually for worthy reasons.
So solve the problem, using big data.
You have as a firs axis, the natural work relaxation cycle of the human.
The next axis is geographical variation in sunlight
Then it is the seasonal placement of holidays.
Finally self selected vacation
Get strings of measurements that make a good proxy anywhere you can measure n these axis. Then search and shrink the data on a long, monotonous searching algorithm. Find a good distribution for the first axis, modify it with the second, add in seasonal holiday to bound the result. Abstract the resulting distribution as the number of week per year, week end size, and a whole bunch of holidays spread about. You have limited space, holiday ruing the effect of long weekends, take up space. You are always looking to match the cardio cycle for the typical human everywhere, leaving enough channel space in selected vacation time to stay in rational bounds.
OK, so we solve the problem, solve it in time and space so we get a distribution across the USA in bounded intervals. We see variations from optimum, in real time; our big data search keeps up.
Our big data model tell WalMart when to load up with camping gear in California, or if it needs more skiing gear up North. That is a billion dollar algorithm, WalMart buys you teams as soon as you assemble the team.
Where could we get that 25% increase in productivity? According to the BLS, hourly output per worker has doubled between the 1980s and 2010s. That’s 100%. Even in completely inelastic theory, we could have gone from a 40 hour week to a 20 hour week with pay remaining the same. Alternatively, wages could have been doubled over that period. Instead, we have the same or longer hours and wages lower than ever for most workers.
We don’t need a 25% increase in productivity to cut working hours. We’ve had at least three 25% increases in productivity and minimal cutting of hours and major cutting of wages.
Too many people in this debate seem to have been living on Ultima Thule for the last 50 years.