Thursday, August 11, 2011

Intellectual Progress?

Nick Rowe (2011): "because my policy increases output, because there's 1% more person-hours of *government* employment."

Lionel Robbins (1929): "The days are gone when it was necessary to combat the naïve assumption that the connection between hours and output is one of direct variation, that it is necessarily true that a lengthening of the working day increases output and a curtailment diminishes it."  

Nick Rowe (2011): "Sandwichman: This post is not about work-sharing. I am holding hours per worker constant. You are off-topic. Do not hijack this post. Stop now. Not everybody shares your interests."


  1. Sandwichman: But I am assuming that hours per worker stays the same. So whether a cut in hours per worker would or would not increase output per worker-hour is irrelevant to my policy experiment. I'm not talking about hours per worker. I am talking about number of workers.

    You are a living example of applied orthogonality.

    If I wanted to do a post on the effect of changing hours on productivity I would have done a post on that subject. But I didn't!

    It would be like me coming onto one of your posts about the benefits of work sharing and saying that you are totally wrong because you are totally ignoring the fact that free gym memberships would make workers healthier and more productive.

    And then going into a total wobbly because you wouldn't let me hijack your thread to talk about the benefits of free gym membership.

    What have you got against healthy workers? Why are you totally ignoring all the literature on exercise and health and productivity?


  2. Nick,

    I plead not guilty. Although it may be an inference from what I did say, I didn't say anything about "work sharing." In a subsequent comment, Frances Woolley made exactly the same point that I was making:

    "Now I think you're suggesting that people put in a 100% work week for 90% of their salary - but isn't that a stronger assumption than you need for your results to go through?"

    Actually, you were suggesting a 100% work week for 99% of their salary -- but that's not important. The key point is that output and hours are not linked to each other in a simple arithmetical formula. There is always and everywhere an "effort bargain" -- that is as long as you are dealing with human beings rather than some robotic carriers of an abstract model. So, it's not merely that you assume "the wrong thing", when you assume direct proportionality between hours and output -- as Woolley pointed out, you assume TOO MUCH. If one can loosen the constraints and get the same result why have more constraints than necessary? Occam's razor.

    I hesitate to point this out but the "simple sum in arithmetic" assumption was John Rae's version of the lump-of-labor fallacy (although he didn't use that term). So let me get this straight, then. You're saying the lump of labor is only a fallacy when it involves a decrease in hours per worker but is not a fallacy when hours remain constant or increase? What a strange fallacy!

    Since you brought up the topic of work-sharing, I'll take the opportunity to link to two posts by colleagues addressing that issue:

    Dean Baker: Work Sharing is the Answer

    Peter Frase: The Productivity of Unemployment

  3. Frances Woolley was making a point about marginal propensity to spend vs save income as one's income goes up. The only reason she brings up a 100% work week is because only 90% of workers work that in one hypothetical, with the rest (10%) working a 0% work week. I hope we can agree that those not employed by the government will not be producing anything like their prior output! There are no work weeks considered other than 0% and 100% in her comment.

  4. TGGP: "There are no work weeks considered other than 0% and 100% in her comment..."

    Wouldn't it be best to let Woolley speak for herself, TGGP? Frances Woolley: "So if you have 100% of the people putting in a 90% work week (and getting paid 90% of their salary), you'll have less saving and more spending than if you have 90% of the people putting in a 100% work week (and getting paid 100% of their salary)."

  5. You're right, I misread her comment. But her point is not about worker-hour productivity, but propensity to consume. The same amount of government payroll is being divided over a larger number of people, and their reduced salary means they save less.

  6. I agree, her main point was about propensity to consume, but in making that point she alluded to the more or less common sense notion of an hours reduction equivalent to the reduction in weekly pay. Her final sentence may be ambiguous but I interpret it to mean that the two assumptions -- of 100% hours and output proportional to hours, regardless of wage cuts -- are two assumptions too many. You can get the same result from assuming that a small hours reduction will not affect total output per worker, especially if workers are getting something in return, i.e. a tiny bit more leisure.

    The reason I mention this is that Nick's "off topic" reaction to my comment was way out of proportion. I didn't mention "work-sharing." That's an inference he drew from my comment. But by the same token, his formula of a wage cut with 100% hours and no change in output lends itself to the inference of a deliberate a priori exclusion. Why? Why would one go out of one's way to set up a counter-intuitive condition (wage cuts won't affect output) at the same time as avoiding an obvious alternative and objecting strenuously when that alternative is mentioned?

    Over at WCI I mentioned a precedent for Nick's idea in a 1945 article by W.H. Hutt, "Full Employment and the Future of Industry." Hutt goes a little further than Rowe. Instead of 100% of hours for less pay, he suggests something like 100%+ hours for the same pay. Presumably then you get more output from the current workers without even having to bother hiring more on the first round. I parodied Hutt's argument in "Thinking Along the Right Lines: Creating Jobs Through Longer Workweeks and Later Retirement" and would be very interested to hear how Nick distinguishes the key assumption in his proposal -- output remains proportional to hours worked -- from Hutt's.