Who whips the whippers? An infinite regress of political economy

The other day we shared this amusing-but-false story that appeared in the Journal of Political Economy a few years ago:

“On a boat trip up China’s Yangtze River in the 19th Century, a titled English woman complained to her host of the cruelty to the oarsmen. One burly coolie stood over the rowers with a whip, making sure there were no laggards. Her host explained that the boat was jointly owned by the oarsmen, and that they hired the man responsible for flogging.”

Paul Campos shares some further background here about the original source of this tale.

But I want to step back a moment and think of this from the economics point of view. The argument is that the “oarsmen” hired the burly man because it was too difficult to assess their individual contributions to the effort. As the academic economists charmingly put it:

Discipline at the workplace—such as the coolie in Cheung’s story—may reflect demand for arrangements to help avoid the temptation to shirk.

But here’s my question: what about the man with the whip? What if he was tempted to shirk? You might think that a whipmaster would just naturally enjoy the task—but it’s a job like any other, and there’s no reason to think that all of them get a positive utility from hurting people, any more than we should assume that all University of California law professors are into torture.

Indeed, many whippers could well feel a . . . temptation to shirk. What, then, could allow them to resist?

The answer is obvious. The whippers need a whipper of their own, an even burlier “coolie” whose job is to whip the whippers if they are not whipping enough. But then these meta-whippers may need some encouragement . . . this will go on forever.

It seems that we have a potentially infinite regress of floggers, each burlier than the last. And if you think this might be a bit of deadweight loss, that’s just you being a naive Westerner, “horrified by this cruel treatment of the obviously overworked coolies. That is, horrified until it was pointed out . . .”

P.S. Could somebody please write up a few formulas for this idea? We could whip something up and get it published in a top-5 journal!

P.P.S. You might say that I shouldn’t be making fun of these poor scholars who innocently took a dubious story, elaborated it with a bunch of incoherent details, and then inserted it into the scholarly literature. All I can say is . . . Nobody forced them to do this! They could’ve played it on the straight and narrow and only included true, or at least well-sourced, stories in their article. If they and others are going to use this story as backing for their theoretical arguments, I think it’s only fair to take the story at face value and consider its full implications, infinite regress and all. If you want, you can make fun of me for writing a paper, way back when, called Enhancing democracy through legislative redistricting. In retrospect, that was a terrible title, and the article itself was flawed in only looking backward with no thoughts about how things could change. Writing that paper may have done all sorts of damage, much more than the riverboat story.

P.P.P.S. Henry Farrell points to this 1994 article, “Why Politics is More Fundamental Than Economics: Incentive-Compatible Mechanisms Are Not Credible,” by Gary Miller and Thomas Hammond, that is relevant to the above discussion. In their framing, the central question “the political problem of ‘constraining the king,'” which indeed seems more realistic than stories about laborers paying to be whipped.

25 thoughts on “Who whips the whippers? An infinite regress of political economy

  1. The problem you are referring too is far more serious. Many situations involve principal-agent problems. The principal cannot directly observe the effort of the agent, so they must expend resources to (partially) mitigate the problem. These measures might be changing the compensation scheme to better align the interests, or hiring other agents (i.e., the whipmaster) to monitor the agent’s actions. As you point out, the whipmaster is also an agent whose actions cannot be observed directly, so they need some intervention as well.

    The infinite regress is not really the problem. Almost all principal agency problems lack perfect solutions – just the expenditure of resources to address the problem makes the result somewhat imperfect. The relevant things to evaluate include the costs of monitoring or influencing performance at different levels (it may be easier to monitor the whipmaster than the people they are whipping) and the relative effectiveness of different interventions (e.g., performance-based contracts vs threats of whippings). I think the only general result is that when you have a principal agency problem, some resources will necessarily be expended to address it. And, in some cases, it is simply not worth the cost for the effectiveness obtained.

    This situation arises in almost all employment scenarios, whether they be oarsmen or professors. What bothers me about this particular research example is what bothers me about so much economic research. The research serves a particular political agenda – it can be used to justify a variety of practices that appear to harm people (such as workers) by “showing” that it makes them better off.

    While such a finding can be accurate, it is selectively employed by economists. This example was used to highlight practices to discipline workers. It does not appear to be used to justify regulatory practices to discipline firms for bad behavior (I could be wrong – if so, please provide references).

    • It sounds far easier to monitor a hypothetical whipper than a rower. A rower can literally go through the motions and look like they are rowing without providing any impetus, hiding among the many rowers down below. A solo whipper up top and visible can’t; you either whip someone or you don’t. If you don’t swing hard, your whip won’t even leave the ground or touch them, much less make the crack from the tip going supersonic. A good whipper can aim precisely and snap just above the skin, but that is still quite frightening (ask me how I know) and the other rowers will be motivated because they can’t see it immediately. We can also later observe the individual whip lash marks, one by one, corresponding to each whipping; whereas, you can’t later observe the individual pushes of a rower. Have you ever seen or read about a fake public flogging? Neither have I. Lots of group-work shirkers or fake rowers, though.

  2. Elinor Ostrom was awarded a Nobel prize for her studies of how people overcome precisely these types of problems. See, e.g. her book “Governing the Commons: The Evolution of Institutions for Collective Action.”

  3. Dale, Sandro:

    Yeah, just to be clear, my above post is a joke. There need be no infinite regress of paid whippers; indeed, there were no oarsmen at all in the original version of this story, and there’s no good reason to think there were any laborers paying people to whip them. The problems of measuring work output are real, but I agree with Dale that any discussion of this problem is degraded by grounding it in a fake story.

  4. Actually, the problem AG points out was well known by the turn of the twentieth century and played an important role in the history of business organization. It was called “the straw boss problem” or something similar, and it went like this: the guys with the whips overwhipped and generated discontent and resistance in the ranks. They were overzealous for multiple reasons: the job attracted sadists, it was easier for them to lord it over a mass of terrified underlings, it gave them ancillary benefits when workers would do favors in order to reduce their chance of being punished, etc. (Of course, literal whips were rarely involved; we’re talking about abusive language, docking pay or firing.)

    As management in large-scale enterprise was professionalized, it became clear that rule-by-fear at the level of direct operation was often counterproductive. The firm needed to stop outsourcing the discipline function (straw bosses were often independent contractors), hire these positions from within, and adopt some form of rule of law, with an internal code of regulations that applied to both front line workers and their overseers, as well as an office to monitor and resolve problems. That was central to the emergence of human relations/resources departments. Economists, in my experience, are mostly unaware of this history, but if you study management, at least in a historical/institutional framework, you’re likely to encounter it.

    But there’s an even deeper problem, the reduction of workplace discipline to matters of “shirking”. Ironically, I hold Marx at least partly responsible for this, since he put disputes over the pace of work at the center of his account of the replacement of putting out by the factory system. This put the question of shirking at center stage, and thinkers critical of Marx and socialism have been motivated to develop theories about why it is socially optimal for workers to cede control over the pace of work to managers. Oliver Williamson’s Nobel-winning stuff on the theory of the firm does this quite directly.

    I’ve worked on jobs where pace of work was the main issue, so I won’t deny it can be a flashpoint. But almost every job I’ve had, and this includes academic ones, involved at least some tension over differences concerning how the job was to be done: methods, criteria for evaluating performance, when to make exceptions, etc. The qualitative dimension, in my view, is fundamental in a way that pace of work isn’t, considering that there is often an overlap between the pace set by intrinsic vs extrinsic motivation. But economists have no theory at all about work organization as planning, and how this introduces friction between different layers and perspectives in the production process. (My reading of Stafford Beer, for instance, is that he is deeply concerned with this problem.)

    One could go on and on, but I’ll stop here, except to say that enjoyed every one of Andrew’s puns I was able to catch.

  5. When Mike Munger presented it as an imaginary problem/parable he introduced the possibility of rotating which member of the crew gets to be the whipper for each shift, precisely in response to the problem of “who monitors the monitor”. Each member would then remember the lack of any whipping when a slacker had the whip, and that slacker can then receive their punishment when someone else holds it! Munger phrased it as the whipper receiving punishment if they whipped too much on their shift, but it really works either way. It does however run into the psychological problem that we only perceive the whipping through our sense-impressions rather than an objective measurement, and that could make it difficult for the whippers & whipped to agree (even implicitly via tit-for-tat) on what amount is appropriate.

    • I don’t know where in the linked article you see Munger calling it a “imaginary problem” or “parable.” He calls it an “example.” Moreover, as I pointed out in the last thread on this–in response to Munger’s claim there that it “Never occurred to me anyone would believe it was a TRUE story. It’s clearly a parable, at most.”–he explicitly calls it a true story in the podcast below.

      @58:40
      Interviewer: “It’s a true story, right?”
      Munger: “Well, Steven claimed that it was, and I’m confident that it was.”

      https://www.econtalk.org/munger-on-the-nature-of-the-firm/#audio-highlights

      So two levels of revisionist history, really.

      • In the link I used Munger says “Imagine a group of workers pulling a barge upstream.”, so it’s supposed to be in your imagination. He does not say where or when this takes place because in that AIER piece (not the EconTalk podcast) he does not present it as something that really happened but instead just an example of the general idea of pulling-your-weight in team production.

      • Dl:

        I clicked on the link and went to 58:40, and I think you’re right but I’m not 100% sure. Here’s how it goes:

        Roberts: The coolie story, the whipping, is remarkable, and it’s true story, right?

        Munger: Yeah, well, Steven claimed that it was, and I’m confident that it was—uh, it actually makes perfect sense . . .

        There’s just one thing. That bit where he says, “I’m confident that it was—uh”: it could be “I’m confident that it wasn’t.” There’s no way for me to tell from the audio.

        That said, I think you’re right, as “I’m confident that it wasn’t” would not fit in with what Munger says right before and right after.

        My guess that at the time of the interview, Munger thought the story was true, but then later he reflected upon the matter and reassessed the story as a parable. Perhaps he forgets that he ever insisted the story was true.

        • I’m not sure that “Well, Steven claimed that it was, and I’m confident that it wasn’t” works syntactically (the “and” should be “but”), but your effort at charitable reading (listening) is admirable.

  6. Andy:
    The great Leo Hurwicz studied the problem you laid out in all the seriousness it deserves. The paper was appropriately titled “But who will guard the guardians” published in AER.
    https://www.aeaweb.org/articles?id=10.1257/aer.98.3.577
    Someone should look at the data within Kaur et al. I have serious reservations about the empirical finding in the paper that Indian programmers were deliberately picking up bad (dominated – in authors’ terminology) contracts to discipline their future selves.

  7. Andrew said:

    “What if he was tempted to shirk. . . this will go on forever ”

    The coolies could just fire him, right? They hired him so presumably they can fire him. It’s not necessarily turtles all the way down.

    Not that I put any particular stock in this story, much less the just-so economic analysis. But it likely has at least some kernel of truth. During construction of the transcontinental railroad, Chinese workers were well known for organizing themselves into highly productive units with their own leaders, which earned them a reputation as excellent workers among the RR foremen and execs. (But tragically also the enmity of many Americans)

    • Observer:

      I disagree with your statement that the story has a “kernel of truth.” If it’s just that people working in a team might hire a manager, then, sure, but in that case the principle is so obvious that no story would be needed! What makes the story work is the whip. No whip, no story.

      If you try to tell the story without the whip, it falls apart: some Chinese dudes were working really efficiently, and then the titled English woman complains about . . . what, exactly? They’re working too well? Also then there’s no need for the economist surrogate to patiently explain the counterintuitive phenomenon, indeed there’s nothing counterintuitive at all, and then you also lose the meta-moral of the story, which is that softie Westerners don’t understand the real world and they need hard-nosed social scientists to explain it to them.

      • Andrew:

        The story is obviously false. What really happened was that each of the coolies takes a turn with the whip. If the current whipper takes it easy and the team loses time, the next whipper just gives that coolie extra lashes. This situation clearly settles into an equilibrium where all coolies whip and are whipped the optimal number of times. I’ve done the calculations and it comes to 17 lashes per kg of freight per km.

        • That’s one of the possibilities that Munger mentions in his AIER piece, although he doesn’t present it as “what really happened”.

      • “Not that I put any particular stock in this story”

        Chinese workers self-organized into groups to increase their productivity. That’s a historical fact. That is a kernel of truth around which the story about the whip *may have been* built.

        “If it’s just that people working in a team might hire a manager…..the principle is so obvious that no story would be needed!”

        “Team” “Labor”. Labor workers are normally overseen by whip-holding management, not a supervisor of their own choice, much less one they chose to maximize productivity.

        I like Steve’s Just-So story the best. A shining example of economic story-telling!

  8. Just assume everyone has rational expectations and there’s no need even for the first whipper. The n whipper’s whipper expects to be flogged if he shirks, so he doesn’t. The n-1 whipper’s whipper expects to be flogged if he shirks, so he doesn’t… and so on until you reach the oarsmen, who expect to be flogged if they shirk, so they don’t.

Leave a Reply

Your email address will not be published. Required fields are marked *