Can less competition mean more meritocracy?

I am not so sure, but Dawei Fang and Thomas Noe say yes:

Uncompetitive contests for grades, promotions, and job assignments, which feature lax standards or consider only limited talent pools, are often criticized for being unmeritocratic. We show that, when contestants are strategic, lax standards and exclusivity can make selection more meritocratic. Strategic contestants take more risks in more competitive contests. Risk taking reduces the correlation between selection and ability. By reducing the noise engendered by strategic risk taking, dialing down competition can produce outcomes that better conform with the meritocratic ideal of selecting the best and only the best.

Here is the working paper, via the excellent Kevin Lewis.


'I am not so sure'

Well, if meritocracy is defined as competent people rising to the top (in contrast to technocracy, which clearly can be gamed through the process in which one earns credentials), why would it be surprising that allowing a larger pool of people the chance to demonstrate their competence would result in an increase in the number of competent people being rewarded for their competence?

Unless, of course, someone has a fear of competent people being allowed to clearly demonstrate their competence, and thus enjoy the claimed advantage of a meritocratic system.

This used to be associated with America's can do spirit, actually, something that seems to have withered with ever increasing number of university students over the last generation.

The paper is claiming the opposite of what you seem to think it says.

You may well be right, because there seem to be several inherent contradictions in the section cited by Prof. Cowen. 'Lax standards' would seem to increase the number of people, while 'limited talent pools' wouldn't. The same applies to 'lax standards' (used a second time) and 'exclusivity.' I went with the idea that lax standards would lead to a larger group of people being considered in the first place, thus allowing the competent to rise to the top.

However, this final sentence seems to me to pretty clearly to fall on the side of expanding the available pool of people - '...dialing down competition can produce outcomes that better conform with the meritocratic ideal of selecting the best and only the best.'

(Admittedly, I pretty much find the idea of an actual meritocracy pretty much delusional in any area of human endeavor that does not have a direct connection to empirical reality. Engineers have a standard that is independent of the judgment of their peers - lawyers not so much. This is also my personal theory why both the Navy and Air Force seem to be more competent organizations in what they do than the Army - ships sink and airplanes crash, regardless of what the highest ranks believe concerning proper maintenance and how long it can be ignored.)

You still don't understand what the paper is arguing.

Try looking at the statements on risk.

It's an interesting finding, but what real-world selection processes allow contestants to take risks to improve measured performance, as modeled in the paper? I'm struggling to think of any.

Most of the inputs to processes like hiring, promotion, and college admissions don't seem to have this property. I guess lying about your qualifications could be one example, but it's an imperfect one (and makes up only a small portion of the inputs to a decision).

I don't think their conclusion discussing applicability to the real world is well-supported, for this reason.

Also, I would add, that there is a simple way to reduce the role of randomness in outcomes: increase the number of observations.

I learned during my college years that it's possible to sort-of game many exams by trying to figure out which questions the lecturer is likely to ask and try to learn the answers to those questions by rote learning; especially in advanced mathematical courses, there is usually only a handful of "standard" exercises that the students could reasonably be expected to be able to solve during a few hour exam. This can be considered a risky strategy since it often results in me being less able to answer those questions that I didn't practice. An alternative, low risk, but potentially lower return strategy would be to try to honestly understand the material of the course instead of just practicing for the exam.

Also, just plain cheating is a high risk strategy.

An example? Easy. Take a look at the brew-ha-ha surrounding Campbell's Soups newly announced CEO. Guy pumped up the numbers prior to an acquisition and jumped ship hoping the hangover would start after he was clear (it is assumed). (see WSJ). If you can't dazzle them with brilliance, baffle them with bullshit. If this behavior is foreign to you, might I suggest you get out more? :)

It's obvious. Tournament winning strategies are by definition different.

Generally, you optimize position relative to competitor rather than absolute expected value.

Poker tourney. Last hand. You trail by X. You might optimally go all in on bad odds, top optimize winning chances.
Soccer. You trail. You send your goalkeeper to attack.

Selection also modes things up.
Maybe higher quality contestants aren't naturally inclined to sacrifice quality for positional optimization. And you select against them!

Quality can be improved, aversion to positional optimisation probably can't.

Yes, but after 100 tournaments, the best poker players are going to have the best records. And that is more relevant to real-world selection processes they want to apply their findings to.


Best poker players that don't know tournament strategy will not win.

Tournament strategies are a must.

Of course, a bad player will not win by tournament strategies alone.

Anyway, of you wish to study it hers one source. I'm not sure I'm able to cover the subject via a blog comment tbh

I'm not sure what you are saying now. The best players are the best at winning by definition, so if they are not good at tournament winning strategy, they are not the best players.

I thought you meant that the tendency to go all in on a losing cause meant that a weaker player might win more often than expected. This is similar to the claim the paper makes, but over a large number of tournaments, the best players would still come out ahead, because they will be in a position to win more often, and because the "all in" strategies tend to be long shots.

Because of the discrepancy between tournament and expected value strategies, a tournament test will create strength winners.

And humans aren't automatically strategy maximizers.

This you'll get better players that are doing worse on tournaments. Logical definition of better isn't everything

The disagreement is how to define the best. A tournament winner will be the best player of that sort of tournament. That player may not be the best at another cooperative structure.

This is very common. Trump won the electoral college and lost the popular vote. Two very different competitions. I follow F1 racing. Last year the driver who made the fewest mistakes won. Is he the best driver? He lost out many times in specific races or duels with other drivers, but over the season won.

Do art competitions produce the best art? Do photograph or design competitions produce the best design or competition? Does a competitive bid on a construction project produce the best contractor able to perform the task?

It depends on how the competition was structured.

Another instance. Did the desktop computer competition in the 90's to 2010 say produce the best producer? Was Microsoft the producer of the best software? Emphatically no, and it was evident very quickly when everything became connected. Their whole market strategy and development structures were flawed in the new game with different rules.

A game produces a winner of the game. The best at that game. That is all.

One good book is here. Missed lining in above comment

My reaction is similar to this. I see more than one kind of meritocracy. In the hard form you want (or fantasize) a perfect top down ordering, with the most excellent candidate accepted first, then the second, and so on.

In a softer meritocracy you only want everyone selected to be capable. In that softer form you are only trying to avoid false positives, failures, collateral damage. If this guy can be a good butcher, baker or candlestick maker, accept him.

Or randomly select from the eligible group.

(Some related cognitive dissonance and hilarity.. how many who think they support hard meritocracy voted Trump? Dude doesn't even fit the soft model.)

I think the soft model is sufficient, if you actually hold to it.

"The designer’s welfare function is purely meritocratic"

Nobody believes that. Maybe at some ultra-high level, but everyone believes that there are big principal-agent problems at the level controlling the procedure. Agents rig selection pools, hence, large selection pools.

+1 Hit the nail on the head. As the Ivies have shown, using more lax standards has allowed them to be more strategic (as well as conniving, dissembling, and nepotistic) in their selection process while claiming that "all students are well qualified." (David Hogg being the most obvious recent example). One of the reasons Chinese people don't want universities to use non-objective criteria is that people are convinced that a "holistic system" will lead to even more corruption than they see now. Judging from US top schools, they're right.

Harvard thinks there is a fair chance Hogg will do interesting things in life, and you oppose him for the same reason.

"Risk taking reduces the correlation between selection and ability." So what the author is suggesting is that by reducing competition, it reduces risk taking, and by reducing risk taking, it increases the likelihood of selecting the best. As I understand the current situation, there's too much complacency and not enough risk taking. But that depends on the meaning of risk taking. Does it mean only a moonshot or does it include a well-considered plan that might appear a moonshot but isn't. Listening to Andreessen and Horowitz describe how they arrive at a decision to invest (or hire or whatever) is the opposite of a moonshot but it is risk taking nevertheless (they didn't become wealthy investing in government bonds). What about higher education? Consider law school. Many selective but non-elite law schools promote competition by requiring professors to distribute grades on a bell curve as the method to identify the best. Elite law schools for the most part don't impose a curve because, well, all of their students are the best. Some believe the approach of the elite law schools promotes complacency. Indeed, some believe elite colleges (as opposed to elite law schools) suffer from complacency because, once admitted, little is expected from the students - many students don't even bother attending classes.

>By reducing the noise engendered by strategic risk taking, dialing
>down competition can produce outcomes that better conform with
>the meritocratic ideal of selecting the best and only the best.

Define "the best". "The best" really means "that which serves the interests of the managerial class". There is no other definition of "best".

The point is making the members of the professional class brutally compete with each other for tenure, partnership etc. Doesn't matter if the guy who wins the end is a little less smart or has a little less "merit" (whatever that means).

Noe says yes?

Who's on first!

"Risk taking reduces the correlation between selection and ability."

This is just an argument over the definition of "ability."

A problem I see with this analysis is that it rates competitors on a binary scale. In the real world people are on a continuum, and people higher on the continuum are more likely to win competitions. So even the “unqualified” people who win competitions tend to be on the higher end of the ability distribution from the “unqualified” pool.

For an ecologist this is not a complex problem. Basically, the question is: what is the scale of the system being examined? By narrowing the parameters of the environment, one can optimize against those stable parameters more effectively. More external controls means more focused competition, more reliable payoffs. If those controls ever relax, ... resilience in the resulting system is low. Merit is not some Platonic reality. Merit is matching the performer to the environment; and other performers are part of that environment, so relative merit translates into absolute merit ...

Economists seem to be worse than most people in understanding that the map is not the terrain. There seems to be poor understanding that the world is noisy. Do they (the people who study "merit") know what part of their metric(s) is noise and what the signal is? I doubt it. TL;DR the paper, but I wonder if their "lax" just means increasing the gain while not improving S/N ratio? Also, I'm a bit confused about their use of risk taking. Isn't that, for many jobs, a qualifying characteristic? So, if their measure of merit doesn't include a measure of it, how realistic can it be? I wonder if there's anything in the paper which advances the science? It can't be that there is any inherent value in their grossly unrealistic and overly-simplified model, and I just ain't seeing any novel finding. People who don't take risks are less likely to win (some/many) competitions, who knew?

If I'm applying for a job, especially if I've made the first and second cuts, I like to know if there is an internal candidate as the finalist and anything about that person. If there is an internal candidate, my strategy for the interview changes. I try to win over the rank and file with my vision, leadership, and charm. For the hiring manager, I explain how I intend on using the internal candidate (assuming I know his name and background) as my deputy and the contributions the internal candidate can make to help me do the job.

Regarding the paper itself, it would have been nice if they had included some actual examples of the types of risk taking and strategy that are utilized by candidates.

Anybody who has been through a pressure cooker school knows this intuitively. The difficulty here is the tautological nature of meritocracy. Merit, defined as the winner of a tournament, is not merit in any objective sense. Returning to the Ancients on this topic would serve us well . The original idea of merit was one's deserved portion, or destiny. This is very different than how the Moderns interpret merit.

In observing members of my age cohort attend professional schools, I noticed a big distinction between lawyers/doctors and MBAs. While the latter is much maligned, most elite programs, which have grade nondisclosure, seem to produce a wider range of outcomes than law and medicine (although, disappointingly 70% of my class went to traditional finance and consulting jobs). Although less than 3% of my business school class joined or started a startup, the numbers in law and medicine are probably less. The biggest issue I saw was that by the point of entering business school, the class had already been indoctrinated into the cult of meritocratic hoop jumping.

> less than 3% of my business school class joined or started a startup

Amazing statistic worth highlighting.

I think, judging by numerous economics papers I've recently read, that teaching or arguing economics is paradoxical or counter-intuitive isn't such a good idea. Apparently, economic explanations must now be paradoxical or counter-intuitive to be published or praised. No wonder the profession is so addled.

Comments for this post are closed