Construction Peaks At 60Hr/Wk

Yesterday I puzzled over:

I have found many studies, conducted by businesses, universities, industry associations and the military, that support the basic notion that, for most people, eight hours a day, five days per week, is the best sustainable long-term balance point between output and exhaustion. Throughout the 30s, 40s, and 50s, these studies were apparently conducted by the hundreds. … Somehow, Silicon Valley didn’t get the memo. … Five-day weeks of eight-hour days maximize long-term output in every industry that has been studied over the past century. (more)

The most recent study this cites is from ’80 on construction. But I just dug a bit deeper, found some recent papers, and I can now say that this claim is just wrong, at least for construction.

First, an ’01 review found total product peaking at 60 hours per weak:

Now some claim that added product only comes in the first few weeks of long hours, after which exhaustion sets in. But an ’05 paper looked at 88 long construction projects, many over fifteen weeks duration, and it still found max product at 60 hr/wk. Also, an ’11 paper gives an integrated model that explicitly includes exhaustion effects, and it also has max product at 60 hr/wk.

So averaging over many construction projects, and including long-run exhaustion effects, total product tends to be highest at sixty hours of work per week. This leaves plenty of room for higher hours-per-week peaks 1) for especially hardy individuals, and 2) in less physically demanding industries.

GD Star Rating
a WordPress rating system
Tagged as:
Trackback URL:
  • http://www.andreasmoser.wordpress.com Andreas Moser

    Guys, you are all working too much!

  • http://www.gwern.net gwern

    > So averaging over many construction projects, and including long-run exhaustion effects, total product tends to be highest at sixty hours of work per week. This leaves plenty of room for higher hours-per-week peaks 1) for especially hardy individuals, and 2) in less physically demanding industries.

    In lieu of additional data, yeah. But we have a great deal of anecdotal data that thinking-intensive occupations burn out even quicker than 60 hours a week, like with mathematics.

    (And I’m sure I’ve read recent studies for programming, game and otherwise, covering deathmarches versus bug rates, but don’t ask me where.)

    • Jeffrey Soreff

      But we have a great deal of anecdotal data that thinking-intensive occupations burn out even quicker than 60 hours a week, like with mathematics.

      Robin himself quotes, in his previous post:

      Performance for knowledge workers declines after 35 hours, not 40.

    • http://hanson.gmu.edu Robin Hanson

      I have lots of anecdotal data of very productive folks in thinking-intensive occupations working 60 hours a week or more.

      • http://bur.sk/ Viliam Búr

        Maybe some factors other than “what kind of work” and “how many hours” are important for productivity.

        I think that when you do physical work, you get tired by working; with better tools you need to use less of your muscle strength, so you can work longer. When you do mental work, you get tired by frustrations that you encounter while working. If the work is fun, you could do it 16 hours a day and not be tired.

        (Note to all managers reading this: regardless of what you think, working at your company is almost surely not fun, and your standard ideas about making it more fun don’t work either.)

        I would guess that the people who work long mentally and don’t get tired, somehow don’t have so much frustration. Maybe they have more autonomy on the workplace or maybe their work is related to their hobbies. If you made other people work for the same time, the effect would not be the same.

  • http://daedalus2u.blogspot.com/ daedalus2u

    What this says is that making wages so low that people have to work two jobs actually costs the economy (and employers) more in terms of lost productivity. In other words, a worker’s productivity for 80 hours labor at minimum wage is less than the same worker for 40 hours labor at 2x minimum wage.

    The figure shown (figure 16), is only for short duration periods of overtime. For longer periods the reduction in efficiency is more apparent (see figure 11).

    What this tells me is that objections to minimum wages are not based on practicality, they are based on signaling. The employer who pays workers minimum wages that are below a living wage such that the employee has to work two jobs isn’t concerned with the low productivity of the worker (because the worker would be more productive if paid 2x for half as much time), rather the employer is concerned about signaling to the employee that the employee is low status and the employer will do whatever it takes to keep the employee from increasing his/her status, even at some cost to the employer.

    • Jayson Virissimo

      How much higher is a living wage than the minimum wage?

    • http://un-thought.blogspot.com/ Floccina

      What this says is that making wages so low that people have to work two jobs

      How could you determine that wages are so low that people need to work 2 jobs? One only needs a warm place to lay his head, enough cheap simple food to survive and 2 sets of clothing. Some people in the USA live on $100/week so one could say anymore than that is more than enough. One might argue that with income below some level one would need to live in a bad area, but a bad area is an area where the least wealthy live and that is not solvable by people at the bottom making more money.

      I think that most of us might all be better off if we worked less but who is to say?

  • Robert Koslover

    My own anecdotal experience as a project manager supports the notion that working well in excess of 40 hrs per week yields greater productivity overall, provided that the people doing the work are sufficiently motivated. I’m not just talking about earning a living here, but working on a project with a well-defined goal that, if achieved, will bring substantial satisfaction to the people working on it. Under those conditions, I once participated in a small group that worked intensely, for several weeks in a row, at roughly 72-80 hrs per week. We were very tired at the end, but thrilled to do the job.

  • Ed

    Possibly in modern environments a lot of time during the day is unproductive time spent multi-tasking trivialities, email, ISO9000 meetings and the like.

    The four hours a day after the HR drones go home and leave you alone were always my most productive hours.

    Actual productive hours…rather less than 40, overall.

    • Dave

      Words of wisdom. Yes, just what is work? To answer the question,drive a car past a construction site.

  • Lord

    My experience is that in construction a lot of the time is not working but waiting for someone else to complete their tasks before preceding. It is very possible longer hours just translate into being available when needed.

    This seems to occur in other areas to some extent. The extra time is not spent working but relaxing, socializing, communicating. Often those working the longest hours don’t work during the day at all but only early or late when they won’t be disturbed. Usually long hours are associated with unpleasant family lives or deficient social lives where work is relief from otherwise unsatisfactory areas of life.

  • http://disputedissues.blogspot.com Stephen R Diamond

    Construction work is seasonal, and fifteen week construction project is short-term when you’re talking about physical exhaustion. Surely it’s wrong to reject the optimality of the 40-hour week based on experiences lasting less than four months.

    Intellectual work raises different issues. The most fatiguing aspect of intellectual work seems to be making decisions, following recent work on decision fatigue. (See references in my review at http://tinyurl.com/7lnoxne) The detrimental effect of making too many decisions was demonstrated in a study of Israeli judges, whose rate of ruling favorably on petitions for parole went from 65% in the morning to 10% at the day’s end. But decision fatigue research doesn’t seem to augur so much for a shorter day as for frequent breaks (including snacks). The brain workers who put in 60-hour days are unlikely to spend that many hours in actual work. But if they’re employees, maybe they must _appear_ busy all the time.

    • http://hanson.gmu.edu Robin Hanson

      If 15 weeks is short term in construction, that further calls into question the previous studies that supposedly found huge exhaustion effects over the course of six weeks.

  • Ian

    Thinking of robins recent post on “what work cuts” it seems likely that you’ll be productive at more than 40 hours if you value yor work more than what you’re cutting to work. I work 55 hours/week at a retail sales job where I make commission. 15 hours of that is illegal where I’m off the clock. My manager lets me do it because I’m more productive by myself than the 3 other sales people are put together. I also love working that many hours. 40 hours seems entirely too short of a week, especially for what I’d like to earn.

    So I don’t believe you have to find your job fun in order to be more productive (unless you can show me some data).

    Also I think you can easily negotiate less hours if you value that without sending bad signals.

  • http://m.facebook.com/jeff.cliff themusicgod1

    Although I couldn’t read 2 of the papers linked to here, one factor seems to be left out of the discussion so far – - transportation time. It seems assumed that employees simply ‘show up’ out of nowhere to work and dissapear to some undefined place afterwards where they simply recuperate from work and have nothing outside of work that occupies them. Perhaps that’s a little extreme way of putting it but it is something that is important to that ’40 hour’ number.

    In the past I’ve been OK with working 60-70+ hour weeks for extended periods with minimal loss of productivity but I have a ~7 minute commute and fairly little nonsense at home to deal with. Likewise I’ve lived in a small town where I worked 2 fulltime jobs and a couple part time jobs(at the same time) but I felt much better (and seemed to be more productive) than another job where I had an 2 hour commute, even at less than half the hours, and even though the other job was much simpler and less physically and mentally demanding. There’s always other factors but this seems to be a big ommission.

    I know people who commute for 4+ hours a day with families and responsibilities at home and still manage to work 45+ hour weeks…all else considered equal, the just sheer amount of work they are capable is going to be less than what I’m capable of merely because a lot of their ‘good’ hours are spent on the road, whereas mine are spent in front of a workstation.

    Sure some of that will be evened out in large numbers of datapoints but the point is that if you’re looking for a place to measure exhaustion point, the part to start measuring isn’t the one that’s available(ie the one where people start the clock) but the part where they leave what they are doing outside of work and start down the path of doing something else related to the interests of The Bureaucracy.

    And then *that* line blurs very quickly into shiftwork and on-call 24/7ness where really, you’re never not working on some level.

  • Alistair Morley

    Even if total product is maximised at 60, marginal product must be falling drastically way before that. What if marginal costs (for the company) are linear or near-linear?

    Whilst some industries allow folks to work extra hours for no extra money, for many industries those extra hours will be overtime pay. A company will find itself paying the same amount for less and less useful work as the worker approaches 60 hours. As the company has a supply-curve to control, this will not be in its interests where overtime payments have to be made.

    We might test this by comparing industries with standard overtime payments (e.g. blue collar) vs those without (some whitecollar/IT?).