Price, wage or service provision discrimination depends on two things
1) market power. You can't do it, if your competitors don't do it. This means either there is a monopoly or cartel- formal or informal- or else it doesn't exist.
2) there are sub-populations characterized by cheap to apprehend but costly to disguise traits which have different elasticities of demand and it is possible to segment the market without leakages.
This is Econ 101.
There is something separate called 'Schelling focality' which identifies the 'natural' solution to a coordination game. Generally, this would be the most salient, in-your-face, blindingly obvious solution which everybody would recognize to be the one everybody else would hit on as the solution. It is certainly possible for an industry to indulge in price fixing and other uncompetitive practices. But this is per se illegal even if there is no proof of communication. It is enough that there is a Schelling focal point- i.e. some signal everybody will interpret in the same way which enables them to coordinate their actions without talking to each other. However, even formal, legal, cartels are inherently unstable. The temptation to 'cheat' is too strong. Look at OPEC. Without a 'swing producer' willing to cut production, it lost the ability to keep oil prices high.
Discrimination is a discoordination game. It is a separating equilibrium based on a costly signal. It isn't necessarily a thing anybody would object to. Sometimes different types of people want to be treated differently and to interact more amongst themselves than with those of a different type.
However, when it comes to things everybody wants more off- e.g. money or good service- the victims of discrimination are being deprived so that profits or rents can increase. This is likely to be allocatively inefficient. On the other hand, for some 'club' and 'public' goods, there may have to be discrimination for the thing to exist at all.
Kaushik Basu, with an impartial ignorance of both Schelling focality and the Economic theory of discrimination, has a paper titled 'Discrimination as Focal Point- Markets and Group Identity'
This paper presents a theory of discrimination for markets in which there are complementarities between different tasks.
It is obvious that any type of useful labor is complementary with every other type of useful labor. But discrimination will only exist if the two conditions mentioned above hold. What Kaushik is groping towards is the notion that members of a particular group may be more effective in some other task- e.g. a charming person may be not just a good mechanic but also good at 'customer retention' or 'drumming up business'. But if the charming person is promoted or paid more because of this trait, no discrimination has occurred. There is a substantive difference between the charming person and the gormless fellow.
It is quite true that Society may class some groups above others. But an enterprise will seek to give the impression that its employees are the best at what they do. Diversity in such an enterprise becomes a signal of their superior ability to spot, recruit, and train talented people.
It is shown that, in such a setting, even when groups are a priori identical,
or just identical. Why use the term 'a priori'?
employers will end up discriminating against certain groups.
If groups are identical, you can't distinguish between them. They are one and the same. This is Liebniz's law re. 'identity of indiscernibles'.
Group discrimination serves the purpose of creating a focal point in a market game.
No it does not. If there is an open market and a particular group is discriminated against, then arbitrageurs will get that group a better deal.
In this model, the free market, far from curbing discrimination, nurtures it, and thereby creates the need for purposive policy intervention.
This isn't a model. It is nonsense. A market may claim to be 'free' or 'open' but there may be an informal cartel or other restrictive trade practices. That's why there is a Department of Fair Trading. It is a common law offence to operate a cartel or to act collusively such that some are discriminated against on a systematic basis. Obama understood this well enough. He argued cases before Judge Posner. The way to tackle the problem in the public sector is to do 'pattern and practice' investigation- this may be of a statistical type- and follow up with 'consent decree' based reform. Basu, at Cornell, didn't get the memo. Maybe his colleagues thought he knew about Indian economics. He didn't.
It is argued that with the rise of technology the problem of discrimination as focal point
How is discrimination a focal point? Is Basu saying that when Whites see a darkie, they don't need to talk among themselves to act in a coordinated fashion to enslave the darkie and send her off to pluck cotton on a Southern plantation? But if this is the case, there is nothing a Democracy can do about the problem. The majority will fuck over the minority, if that is what they want to do. This is because majorities elect legislators who make the laws.
Technology is what has made it possible for most transactions to be anonymous. We don't give a shit whether stuff we buy on Ebay or Amazon was manufactured by people in Switzerland or Swaziland. All that matters is value for money. We may rely on customer feedback for this.
will get more acute and we will have to think in terms of affirmative action
Indian tech companies in America have to hire American engineers whom they don't assign any work to. But this is just 'featherbedding' or rent extraction. It collapses as elasticities of supply increase.
or a system of taxation and subsidy to support groups that get excluded.
till the country goes off a fiscal cliff. Welcome to Venezuela!
For good or for bad, group identity matters in determining market outcomes (Akerlof and Kranton, 2010, Sen, 2006).
Only where economic rent exists by reason of inelastic demand and supply. But such rent is 'quasi'- i.e. tends to disappear as factors of production become more mobile.
The problem with discrimination is that the cost of enforcing barriers between sub-populations rises relative to the reward. Slavery disappears when it becomes too expensive to catch slaves who run-away.
Discrimination against certain groups of people and, by its converse, in favor of other groups has been common practice, observed in different societies and through different periods of history.
Throughout history, it has been a common practice for Professors to talk garbled nonsense.
India’s caste system, with its attendant practice of intolerance and effort to marginalize large groups of people, is an example,
No. India's caste system is a solution to the stable marriage problem. The Price equation explains why endogamous groups will seek to promote their own people at the expense of members of other groups.
as is the history of apartheid in South Africa
Apartheid operated only between 1948 and 1994- i.e. after it had become obvious that White rule in Africa and Asia was doomed.
and racial discrimination and slavery in the United States.
Why not mention Hitler? Didn't he discriminate against Jews? As for Jack the Ripper- there is some evidence that he wasn't a committed Suffragette.
From an analyst’s point of view, they are often troubling because norms and laws often merge into each other.
No they don't. They are separate but influence each other.
Some of these heinous practices were explicitly backed by the law
because elected legislators passed the laws their voters wanted
as in the case of South African apartheid and US slavery. At other times, such as with India’s caste system through history and now (see Deshpande, 2010) or with racial discrimination faced by African Americans in contemporary USA, it was not backed by the law but by social norms, customs and individual beliefs and preferences.
No. It was backed by informal cartels which had the power to punish- e.g. by 'boycott'- those who broke the rules. In particular, 'outcastes' would find it difficult to get their children suitably married.
The focus in this paper will be on discrimination which does not have the backing of law.
If it involves any economic good or service then the thing is per se illegal. However, who you want to marry or hang out with is not an economic good or service.
As a mirror image of this, we often see certain groups benefiting from discrimination in their favor.
Basu is the lackey of an Indian political Dynasty.
This has been true of men through long stretches of history and even now in most societies. Similarly, in the United States, United Kingdom, in India during colonial times or South Africa till recently, if you could choose your skin color, I would strongly recommend white.
Mayhew tells us that in early nineteenth century London, white beggars disguised themselves as Indians to make more money. Then, during the Mutiny, they got stomped. It has always been better to be rich and black than poor and white in the UK.
Where do these discriminatory preferences come from and why have they been so persistent?
The Price Equation. Like other animals, we favor those we are related to. On the other hand, we also favor groups which enable us to have more or better quality progeny. The sad truth about India is that Indians did better with smart Brits in charge. After they left, India spent two decades unable to feed or defend itself.
Discrimination isn't discrimination if one group really is better than another group at doing a particular type of job.
Without doubt there must be many explanations for this, ranging from plain, simple bigotry and prejudice,
or plain, simple, inferiority
to various forms of statistical discrimination that economists have written extensively about.
Granger causality- i.e. correlation- is important. We try to imitate the outward appearance and appurtenances of those who are successful not because we think they cause success but because they are known to correlate with success. This is 'Tardean mimetics'.
One problem with discrimination is that we don't know what is causation and what is correlation. Creating mechanisms which are better for screening and signaling genuine causal factors is what reduces discrimination and the allocative inefficiency it gives rise to.
The problem with 'affirmative action' is that it can distort these mechanisms and end up being self-defeating. It may seem a swell idea to promote brown moneys like Sen and Basu as great Economists. But that worsens outcomes for brown people.
While not wanting to take away from those standard explanations, this paper presents a novel argument whereby discrimination has no innate origins but arises naturally in markets where there happens to be some complementarity between the different tasks we do. This kind of discrimination is a product of the free market and the beliefs of ordinary people.
No. Ordinary people access markets for information- i.e. providers of signaling or screening mechanisms- so as to improve outcomes for themselves.
The upshot is the group identities of people come to matter in such settings.
No. Group identities dissolve. I want to hire an Accountant to save me from bankruptcy. Like the fat slob on 'Family Guy', I have a hazy notion that Jews are good Accountants. But, it occurs to me that some Jews may be crap at Accountancy. So, I consult my local Chamber of Commerce. They point me to a qualified Accountant with a great track-record helping people like me. If the chap is Jewish, no doubt, this will be pointed out to me. If not some other trait might be pointed out which 'correlates' with being a good Accountant- e.g. the warning that the guy is as boring as shit. 'Dynamite at double entry, just don't get chatting with him. He will literally bore the pants off you.'
It is, as I will show, closely linked to the idea of ‘focal point,’ used in game theory (Schelling, 1960).
Nonsense! Focal points are about solving coordination problems where there is no direct communication.
In particular, race, caste and gender becomes important in equilibrium because they acquire the salience of the focal point.
No. Schelling focal points are 'inter-subjectively' salient for all agents who seek to coordinate their actions. Race, Caste, Gender, Religion etc may define a 'discoordination game' for some agents. All women may want to use a female only bathroom. People of a particular Religion may want to worship together according to their accustomed rituals. But these are not 'equilibrium' situations. Indeed, they are not 'economic'. Scarcity does not feature.
In some ways, the focal point theory of discrimination developed in this paper is more alarming than other forms of discrimination where one can point to the source, be it human mendacity or the distortions of statistical information. In my analysis removing government interference and allowing competition in the market to flourish does not remove discrimination, as standard economics had suggested, because it is in fact a product of precisely the free market.
Nothing is the product of the free market because no such beastie has ever existed.
See Arrow (1973, 1998),
utterly vacuous
Phelps (1972), Spence (1974), Stiglitz (1974).
What these guys didn't get is that there can be no discrimination as elasticity becomes perfect. Thus raising elasticity (which also gets rid of 'economic rent) is the solution. Precisely because this is obvious (though ordinary people may not use the term elasticity), those with market or institutional power will invest in political correctness and the appearance of affirmative action. This secures rents.
A polar case of this in a development context occurs in Kremer (1993).
Nonsense! It has no relevance whatsoever.
The model is based on sufficiently realistic assumptions
Basu wouldn't know reality if it bit him in the ass.
to make me believe that, while other forms of discrimination no doubt occur, the focal point theory of discrimination
which says that discrimination of type x is explained by the fact that x is a focal point. One could just as easily say 'discrimination of type x is explained by the fact that x is a trait that can become the basis of discrimination'. This is not a theory. It is stupidity.
does play an important role and, as such, deserves greater attention. In the last section I will argue why, given trends in contemporary labor markets, this is likely to become even more important.
The explanation for why it will become more important is that importance is a focal point which is becoming more important because of its importance. I am not a brown monkey. I'm actually very clever.
Markets and Discrimination A good starting point is the celebrated paper by Bertrand and Mullainathan (2004).Titled ' Are Emily and Greg More Employable Than Lakisha and Jamal?' If this is research what is a Richord Pryor set circa 1973?
This is often treated as the most compelling empirical demonstration of pure racial bias in labor markets.
Fuck off. It is a demonstration of the power of TV stereotypes. We want to go to a party hosted by Lakisha and Jamal. We don't want to employ them doing boring shite in the sort of dead end job our Company has on offer. Hopefully, Emily and Greg are members of a Satanic Cult. They deserve a foretaste of the tedium of Hades.
When we see discrimination, the question always arises as to whether it is bias or reflection of something else that correlates with race or gender or caste, whatever it is one is studying.
Sadly, we don't get that Professors at Cornell may be as stupid as shit. Positive discrimination towards them is a bad idea.
Thus if an employer hires more whites than blacks, is it really a preference for whites, or is it merely a reflection of the fact that the employer needs PhDs and white job applicants are more likely to have PhDs? Bertrand and Mullainathan corrected for this by sending out job applications with fictitious resumes to help-wanted advertisements that appeared in Chicago and Boston newspapers. It was soon evident that, controlling for all other things, candidates with white names were far more likely to get callbacks for interviews than those with black names.
But 'black names' are a recent phenomenon closely connected to a specific sub-culture which has been stereotyped by TV in a manner many African-Americans feel has been deeply unhelpful.
There were striking results, such as how having a white name is equivalent to 8 years of work experience with a black name.
Was the study replicated? Junk Social Science has a replication problem, not to mention 'p-hacking' and so forth. The fact is first names tell you about parental attitudes. If they call their son Tarquin and their daughter Philomela they are likely to invest more in their upbringing than if they called them Kevin and Karen.
In brief, they had engineered the celebrated “ceteris paribus” condition that traditional economists so often talked about but were seldom able to demonstrate.
No they hadn't. Names aren't a good proxy for Race. They may be a proxy for affiliation to a particular sub-culture.
And the findings were striking .
No. They confirmed the point of a joke first made back in the Seventies.
What I wish to do here is to question whether this necessarily demonstrates racial bias. Note that for most tasks in life, to conduct them effectively, you need to successfully do other tasks. If you work for a firm’s sales department to promote sales, you need to be able to successfully interact with buyers’ groups and delivery services units.
That is all part of the same task. A different task would involve managing petty cash and keeping track of expense claims.
If the buyers’ groups and delivery services units try to shun you, you will not be able to do the sales work you are supposed to do well.
Very true. It is difficult to sell stuff to people who run the fuck away from you. But the task of selling stuff involves getting on with customers not shoving your boot up their arse.
And of course the problem is similar for the buyers’ group. When they reach out to you they know they will get better services from you if you are trusted and used by the sales department and the delivery services unit. And likewise for the delivery services unit. They have to gauge how successful you will be with the sales department and the buyers’ group.
This is all part of the same task of promoting sales. Doing accountancy work for the Sales Department is a separate task.
This is where race can come to acquire significance even in the absence of any innate racial preference. If you feel Emily—a common white name—is more likely to do your task more effectively, you will prefer to hire Emily over Lakisha.
Surely, this is a case where you don't want to hire Lakisha because she will be constantly fighting with everybody who looks at her funny.
If all three units do that, this becomes self-fulfilling. The white name provides a focal point
no. The name does not solve a coordination problem. It merely correlates with a specific preconceived notion or stereotype.
in a labor market for tasks that exhibit ‘strategic complementarity’—
Decisions of two or more players are called strategic complements if they mutually reinforce one another.
economists’ term for work contexts where doing one task raises your productivity in another.
in which case the opportunity cost has risen. A person with skill x, can, as a direct consequence, also do task y. People thought Basu, thanks to the skill which had made him a Professor of useless shite, would also be able to give good Economic Advise to the Government of India. I'm kidding. Nobody thought that. Manmohan just wanted nutters like Sen and Basu and Rajan inside the tent pissing out, rather than outside the tent pissing in. Sen, of course, defeated expectations by shitting on everything.
Since my paper is a methodological intrusion into that rather barren terrain between economics and sociology, and is, at the same time, at deviance from the so-called Chicago school, it may be apt to quote Ken Arrow from an interview he gave to Richard Swedberg in 1988, in which he distances himself from the Beckerian approach but stresses the importance of interaction between economics and sociology: “[A] lot of the environment in which economic transactions take place is social and historical in nature. I do not know exactly how to fit these pieces together but there are, for example, accounts of how special groups have played a distinct role in, say, trade. You have Chinese middlemen in Asia; Jews at certain times; Quakers during one period; and so on. It is clearly their social characteristics that matter.” (Swedberg, 1990, p. 136-7).
No. What mattered was whether those middlemen were protected by the powers that be. Otherwise they'd have been robbed and murdered.
What I am arguing is that this is true but the social characteristics may well be endogenous, the product of equilibrium,
Endogenous change is a disequilibrium phenomenon. It has no 'Noetherian' because it is dissipative. I suppose Basu means that some groups adopt new social characteristics in response to a shifting fitness landscape.
however we got there. The basic idea, which shows the relation between discrimination and focal point, can be illustrated with a simple example .
Basu's example merely shows that people with transferable skills have lower elasticity of supply. This means, ceteris paribus, less 'producer surplus' or 'rent' can be extracted by employing them. This is Econ 101. It has nothing to do with 'focal points' because there is no coordination game. The market is itself the coordination device.
There are two entrepreneurs, 1 and 2, who have need for certain tasks to be done, and there are n (>2) service operators or laborers who can do these tasks. Suppose, for instance, entrepreneur 1 needs a person to look after his lawn—buy and apply fertilizer, sow seeds, mow, and so on;
He doesn't want a person to look after his lawn. He wants a contractor to do it for him. He doesn't care if the contractor uses a different laborer on different days of the week.
and entrepreneur 2 wants to lend money to someone.
who can offer security or has a high credit score and thus is likely to repay the loan.
For a large enough market, specialist firms will do the 'screening' or provide 'signals' for a price.
The person who is able to borrow the money, can buy fertilizers and seeds easily and so do the lawn work better.
That's not how these things work. Guys who mow lawns may be bad at running the business side of things. Indeed, they may be recent immigrants or people with poor literacy and numeracy.
And the laborer who gets the lawn contract will be more likely to pay back loans that she takes.
If you can get contracts, you should be hiring laborers to do the heavy lifting.
The entrepreneurs do not know the underlying causation of what makes a laborer more productive, to wit, the fact that if both reach out to the same laborers, they get better value.
This is mad. The lawn guy just hires someone who already cares for a bunch of lawns or else someone recommended to him to be highly likely to do a good job. But part of the task of running a lawn-care business is negotiating business loans. The entrepreneur who lends money will already have a good 'screening' mechanism or else will go bust very quickly.
This is not unlikely in a real setting where thousands of entrepreneurs reach out to hundreds of thousands of laborers.
Laborers work for a wage. They don't borrow money so as to set up as independent contractors. You hire a guy with a broad back and order a delivery of fertilizer or seeds or whatever. Alternatively, you can pay a little more and have a contractor source the labor.
They realize some are more productive than others and may search for markers of that without quite knowing the fundamental model that drives this. The above paragraph may be summed up as follows. Each of the two entrepreneurs picks one citizen for the task he needs to get done.
No. One is hiring a guy for a task. The other is making a loan. He doesn't care if the guy is shit at earning money provided he can offer security for the loan.
If he picks a citizen who is not picked by the other entrepreneur, he gets a benefit of x and if he picks someone the other entrepreneur also picks he gets y.
No. The benefit is the same whether or not, by chance, both had a relationship with a particular contractor. NatWest lent me money to buy a property from Chesterton's. There was no 'strategic complementarity' between the Bank and the Estate Agent. One got their pay-off immediately. So did the other (because of securitization). The outcome would have been different if I'd bought the property for cash at auction.
It is a different matter that there can be 'external effects' in the market for credit, housing etc. But externalities are defined as costs and benefits received outside the market or, indeed, any type of interaction. But such externalities can be 'internalized' by a merger or joint venture of some sort.
Given what we said about strategic complementarity, y > x (1)
The thing doesn't arise on open markets. It could arise where there is oligopoly and all sorts of complicated Cartel agreements. But, in that case, 'externalities will be internalized' through merger or profit sharing agreements which, however, may be per se illegal.
The entrepreneurs are not aware of this strategic complementarity.
They would be if they had market power or else would be displaced by smarter, more Machiavellian, operators.
All they know is that they may get x or y, without being aware of what drives the difference. The only critical assumption in this exercise is (1).
It is absurd.
All other structures of the model can be varied and we will still get the same essential result. I should clarify that I am not claiming that strategic complementarity is always the case but simply that it is realistic in many situations, and when that happens, I want to show that a kind of discrimination happens which requires no innate bias, no differences in ability or entirely through natural market processes with the concept of the focal point playing a critical role. To convert this to a game I need to put in a little more structure to the model. Assume that the n laborers are of two different races: w (>1) of them are whites and b (>1) blacks.
In that case, a result from Chichilnisky & Heal applies. There will be no coordination game unless preference and endowment diversity meets a Goldilocks condition.
Hence w + b = n. Each entrepreneur, in selecting a laborer to get his or her task done, has to use one of the following rules: No discrimination (strategy N), discrimination in favor of whites (strategy W), and discrimination in favor of blacks (strategy B).
This is foolish. You choose the cheapest alternative, ceteris paribus, unless there is some penalty for doing so. Of course, if you gain hedonic utility by employing a particular type of person- e.g. a gorgeous young woman as your Secretary- then ceteris aint paribus.
If she chooses N, it means she randomly picks one of the n citizens, with 1/n probability of each citizen being chosen.
No. It just means she had a 'stopping rule' in a sequential search. This would be like the solution to the 'Secretary problem'. You first check out three of four applicants and choose the next one who is at least as good. Alternatively, you phone the two or three guys who were recommended to you as reliable workers and pick the one who can schedule your job within a reasonable period.
Consider what happens when I order an Uber. It may be that because of my racist beliefs, I wait till I find a driver of my own ethnicity. Still, it isn't really the case that I make a choice among 45,000 drivers. I picked the one whom it suited to pick me up.
If she choose W, it means each white person faces a probability 1/w of being selected.
No it doesn't. There might be a dozen people who can provide a particular service in a particular area. Scheduling conflicts mean that the choice only comes down to one or two. True, for certain services, race and gender may be a big factor. But, in that case, the trait affects utility.
Likewise for strategy B. It is easy to work out the payoffs of the two entrepreneurs depending on the choice each of them makes. This is displayed in the payoff matrix described below, in what I call the Discrimination Game.
The payoff matrix for the two entrepreneurs can't be the same because money-lending is very different from lawn mowing. A guy who can mow lawns very well may be terrible at repaying loans. Equally, a successful garden maintenance contractor might not be able to mow anything at all.
Basu is saying, 'suppose Black peeps be totes shit at mowing lawns, repaying loans, and not eating their own babies. In that case, I can show using Game Theory, that the free market will produce discriminatory outcomes. Nice, non-Racist, people will prefer to hire Whites because 'not being so utterly shitty as to eat your own babies' is a 'focal point''- at least, for brown monkeys who teach at Cornell.
Since it is completely symmetric, there is no need to show the payoffs of both entrepreneurs. I show the payoff earned by entrepreneur 1; and, in this game, entrepreneur 2 gets the same.
Because Black peeps are shitty little cannibals who eat their own babies. No wonder they won't do a stroke of honest work. Lend them money and they'll just buy crack in between gorging on fried chicken and watermelon.
Game 1. The Discrimination Game N W B N 𝑦 𝑛 + 𝑥(𝑛−1) 𝑛 𝑦 𝑛 + 𝑥(𝑛−1) 𝑛 𝑦 𝑛 + 𝑥(𝑛−1) 𝑛 W 𝑦 𝑛 + 𝑥(𝑛−1) 𝑛 𝑦 𝑛 + 𝑥(𝑤−1) 𝑤 X B 𝑦 𝑛 + 𝑥(𝑛−1) 𝑛 x 𝑦 𝑛 + 𝑥(𝑏−1) 𝑏 To understand the payoff, let us check the top left hand box.
It is nonsense. The payoff from hiring a guy or lending to a guy depends wholly on what that guy does or doesn't do. No doubt, social attitudes may affect what you can do. But that has nothing to do with 'strategic complementarity'.
It is certainly true that when you hire a guy or lend to a guy, there are all sorts of things which can affect the outcome. The man could be hit by a bus. His wife might leave him and cause him to become an alcoholic. But stuff like that doesn't feature in the expected pay-off matrix which is what game theory must concentrate on. We may say to a chess player- 'don't move your knight to defend your Queen because then your opponent can checkmate you'. That is part of game theory. But to say 'don't move your knight because that is the pre-arranged signal for Putin pressing the nuclear button' isn't part of Game theory. It is part of a Spy novel.
Both entrepreneur chooses N, that is, pick a laborer with no attention to race. After one has chosen, the probability that the other person will choose the same laborer is 1/n.
This is the case, whether or not anybody else does any choosing whatsoever.
When that happens, each gets a payoff of y. The probability that the other entrepreneur will choose someone else is (n-1)/n. When that happens each gets x.
But neither knows this and so their decisions are not affected. It is also the case that if the guy gets bitten by a radio-active spider and joins the Avengers, both will get a different pay-off.
So the expected payoff is y/n +x(n-1)/n.
No it isn't. It is p.y + (1-p)x, where p is the probability of y. The 'Muth Rational' solution- which is also 'focal' for 'strategic complementarity'- depends on plugging in the correct economic theory to get p. But Basu never has the correct economic theory- which is why he was Chief Economic Advisor of India's most dysfunctional Administration.
It is easy to work out the payoffs in the other boxes by a similar reasoning. It is simple to check that this game has three Nash equilibria: (N, N), (B, B), (W, W),
Not if there is free entry and exit, merger, de-merger etc.
that is, no one discriminates, everybody discriminates in favor of whites, and everybody discriminates in favor of blacks. To check this, note that if the other person chooses N, no matter what you do you will get the same payoff.
Because of an arbitrary stipulation. But you can exit the market or merge with the other guy or develop a threat point or change the rules of the game.
So you cannot do better by unilaterally deviating from N.
But you can worsen the outcome for the other party. That's a threat point.
Next check that, as long as y exceeds x, as assumed, and given that, by definition, n > w, the following are true: y/w + x(w-1)/w > y/n + x(n-1)/n, and y/w + x(w-1)/w > x. In other words, if others discriminate in favor or whites, whites will be on average more productive and so it is in your interest to choose a white to do your task.
But if wages align with marginal product than x rises. It is a fact that where there is legal or pervasive discrimination, there will be wage discrimination. At the margin, you substitute the discriminated group for the, more +
favored group. Consider the Alfred Sloan, the legendary head of General Motors. During the Depression, he reversed the policy of not selling Cadillacs to African Americans and thus saved his Company. He also brought in African American women workers from the South during the War. This meant that the pay differential got squeezed even though Jim Crow was in full swing. Indeed, Hitler's Nuremberg laws were inspired by America.
In other words, (W, W) is a Nash equilibrium.
In which case, there would have been no need for Jim Crow or the KKK. Basu is talking nonsense. Where the market is segmented, arbitrage circumventing the barrier between markets will be highly profitable.
For the same reasons, (B, B) is an equilibrium as well.
Unless B and W aren't homogenous. This is the problem with trait based discrimination. You never know if you have got a truly 'representative' agent. This is why it is worth doing 'discovery'. There's a chance you get lucky and gain a superior outcome more cheaply. But, assuming uncertainty aversion, what's even better is somebody else doing discovery for you. Where there is a change in 'status quo' (though conventionally game theory does not privilege any particular state in this way) without a 'concrete model' of the new state, there may be exit. After all, it is substantive, not strategic, considerations which determine the profitability of an industry or, indeed, the economy as a whole. Suppose, currently, babies don't run the economy. It is not the case that if we appoint babies to run all our major industries, that we would have a Nash, or any other sort of equilibrium. We may have a shitty economy or no economy at all. But sensible people will be running away.
For those with an aversion to symbols, let me convert the above game to a society in which there are 4 laborers, 2 whites and 2 blacks. And suppose y = 2 and x = 1. By inserting these values, the above discrimination game collapses into the special case illustrated below. Game 2. The Discrimination Game: A Special Case N W B N 5/4, 5/4 5/4, 5/4 5/4, 5/4 W 5/4, 5/4 3/2, 3/2 1, 1 B 5/4, 5/4 1, 1 3/2, 3/2 The three Nash equilibria are now obvious. If others discriminate, you had better do the same. But as always with games with many equilibria, there is a need for a focal point which allows players to coordinate their behavior. What I am claiming is that in markets with strategic complementarity, as just described, race or gender or caste can be the focal point.
No. What is focal is deemed superiority or inferiority. But there is no equilibrium unless deemed superiority is substantive, not strategic.
It is important only because others think it is important.
and think it will endure. But this means there must be a substantive difference. Otherwise, people expect abrupt shifts. This is what happens in the fashion industry. One moment everybody is wearing kipper ties. Then, nobody would be seen dead in them. People in the Rag Trade knew that one type of tie wasn't substantively better than another.
You prefer Emily to Lakisha not because you have a preference for white over black but because
you think Lakisha is likely to have 'attitude'. She has been brought up to assert herself. Emily is a mousy thing- whatever her color. Her one dissipation is singing with the Choir on a Sunday. Emily Washington, daughter of an African American Minister of Religion is your ideal employee. Lakisha is the name of the attorney who will get you acquitted for murdering your wife. Alternatively, she is the PI who will catch and kill the real murderer who, it turns out, is actually your wife. You'd been living with a clone who, lacking a soul, loved you truly. That's against he law for married women.
all of you need to zero in on some group
why? There is no such need unless the two conditions for wage discrimination are met. If you aren't a natural monopoly and there isn't a different elasticity of supply for trait defined sub-populations, there is neither need nor any fucking point to any 'zeroing in'.
It is a different matter if there is a substantive difference between sub-populations. It is possible that such a difference is artificially created- e.g. people with one trait keep getting lynched or knifed- but that has nothing to do with focal points.
and it so happens, for reasons of history or whatever, you have settled on whites. One important implication of this is that one popular view, namely, if you leave it all to the market, with no government regulations and intervention, racial and caste discrimination would go away, is not valid.
And yet, that is precisely what happens. It was the market which fucked up caste and gender and race based discrimination. Government regulations were and are counter-productive if they merely redistribute rent in some tokenist manner.
Discrimination arises from the free market.
No. It arises through restrictive Trade practices which may be per se illegal in Common Law jurisdictions. Consider what happens if a black female worker of yours is treated less favorably by an enterprise with whom you have a contract. You can sue for damages unless the contract explicitly specifies otherwise. But if that is conscionable and legal, the fact is it is the Law, not the market, which is responsible for creating a substantive, not strategic, difference in 'complementarity'.
If you want to stop discrimination, you may, in fact, need regulation, and conscious affirmative action.
This is even more true if you want to decrease competition and secure rents accruing to vested interests.
And when we go for affirmative action we must not indulge in the politically correct banter, so often heard, that by doing affirmative action you do not hurt your returns. The truth is that your returns may indeed be diminished by such action. The appeal has to be that even if your return drops, there are certain actions in life which ought to be indulged in for its innate moral goodness.
Destroying the economy, for a Bengali buddhijivi, is an innately good action. Why stop there? Why not say, 'we must do affirmative action in the Army so that we can get invaded and enslaved. That is an innately good and moral outcome. '
The problem here is that even the beneficiaries of affirmative action suffer if, for example, surgeons butcher their patients and soldiers are unable to defend the borders.
Affirmative action may be one of those. I shall return to this in the last section.
Affirmative action is a way to buy support for expanding a useless program.
The Art of Creating Optimal Groups As a digression I may point out that this focal point model of discrimination can be put to some rather Machiavellian uses, such as that of creating your own group and then promoting it.
Whereas what you should do is create your own group and go around saying it is totally shit.
And indeed, such practices are not unknown in the world. Alumni associations and fraternities are good examples of this.
In America, alumni associations help Colleges raise money to improve facilities. Basu is against this. Alumni associations should burn down the campuses on which they were educated.
They are often used to promote the college or frat label. Thus we are told how Harvard students are smarter than others, or how Berkeley students are more productive than others, or how Cornell graduates are more creative than others (this one happens to be true), and so on. What the analysis in this paper shows is that, once such beliefs catch on, they can become self-fulfilling because that belief then serves as a focal point.
I suppose, if people believe that alumni help each other, then there is a bonus in hiring a guy from a College with a large and active alumni association. In England this is known as the 'Old School Tie' effect.
Most of us, human beings, have multiple identities, race, nationality, language-group, ethnicity, gender and so on (see Sen, 2006),
We have a single identity but multiple predicates may be applied to us.
and as I just argued we can also create new identities (see also Basu, 2011).
More particularly if we are spies or criminals seeking to evade arrest.
Now. If you want to deliberately nurture the view that one of these identities is a mark of greater productivity, this model suggests it may be worthwhile to pick on or create a group that is relatively less populous. To prove this, consider the payoff an entrepreneur gets from the equilibrium in favor of whites. This is given by [y+(w-1)x]/w. The payoff in an equilibrium in which blacks are favored is given by [y+(b-1)x]/b. Let us call the former whincome, and the latter blincome. Recall b = n – w. Hence, as w becomes smaller, whincome rises and blincome falls. Both (W, W) and (B, B) are still Nash equilibria but the former becomes more and more dominant as the white population becomes smaller. In brief, if you want to promote the idea that a particular group you belong to is more productive, you will be better off if you choose a small group.
People with severe brain damage are a very small group. Why not promote the idea that they are very productive?
Among other things, this explains why raising the profile of women is such a hard task.
It has happened by itself all over the world.
They constitute roughly half the population. Considering the case of nationalities, if you promote the idea that the British are more productive and the idea that Chinese are more productive and people buy into this belief, the British will turn out to be even more productive in a world in which they are believed to be more productive than the Chinese will be in a world in which the Chinese are believed to be more productive.
There are vastly more Chinese people than British people. We Brits think the Chinese are more industrious and thus more productive than we are because, we believe, they suffered great poverty till quite recently and thus are keen to build up their wealth and to give their children a better start in life.
It is a big deal for a British kid to get into Cambridge or Oxford. But getting into Beijing University is a very much bigger deal. Cambridge has a 20 percent acceptance rate. Beijing's is 2 per cent.
It simply isn't the case that pretending that some small group- e.g. Jaravas from the Andamans- is more productive will fool anybody.
There is often surprise expressed at the fact that Britain was such a small nation that once ruled virtually the world.
No. It ruled places which were shit at fighting. Britain prevailed over its West European rivals because, being an island, it could concentrate more resources on the Navy. Its policy was to have a bigger navy than all its potential enemies combined.
What is being said is that we should not be surprised.
Britain was genuinely more productive than any other country when it came to having a global Naval presence.
One counter argument to this needs to be kept in mind. If smaller groups are more effective, why not go all the way and proclaim that you as an individual, or one-person group, is more productive?
Elon Musk isn't saying all Musks are very productive. He is saying he is a business genius.
The reason must be that as groups get too fragmented people cannot hold information about group characteristics in their heads because there are too many groups.
British people come from Britain. Japanese people come from Japan. It is easy to hold that information in your head.
This can set a lower bound to how small groups can be without losing advantage.
No. A large group which is genuinely more productive in a field- e.g. literate people vs illiterate people- will retain absolute advantage. Indeed, as literacy rises, the more essential it is to be literate because more and more communication will be written.
To model this formally will entail a more elaborate analysis but the intuition behind what I am arguing should be evident.
This is imbecility, not intuition.
In closing this section, I want to remind the reader that while this focal point model of discrimination is very important, as with all theory, to take it to the real world and to put it to use, we must enrich it with our commonsense and reasoned intuition.
which is precisely what Basu lacks.
. Hence, the above theory must be combined with other ideas and our own experience before it is put to use or employed in designing policy. It is, for instance, worth reminding ourselves that productivity and even intelligence are also dependent on how a person is treated, on how society views this person’s group. Even if the discrimination is purely a focal point at work, it can leave scars on people, making the ones believed to be less intelligent actually behave less intelligently. Unlike in a strategic-form game where a switch from one equilibrium to another can be effected in the twinkling of an eye, in reality, these changes are likely to take time and involve economics and psychology.
No. Such changes can occur overnight as a result of a revolution or a change in Government policy. Idi Amin expelled Asians over a period of 90 days. It was literally the case that those running a lot of businesses were replaced overnight.
Policy Implications
Basu thinks affirmative action will worsen economic outcomes. This means everybody suffers though a few token hires secure a rent. Ergo, don't do affirmative action. Implement targeted policies to raise productivity among specific sections of society.
The model presented here opens up some policy dilemmas. It should be evident from the above model that GDP or the aggregate payoff earned by all is higher when there is group discrimination.
Which is why we know Basu is wrong. Discrimination means different outcomes despite substantive equality. Getting rid of it boosts GNP. There was a time when women were paid much less than men in certain monopolistic enterprises. This had a disincentive effect. It reduced the participation rate. This meant lower per capita Income. Look at Bangladesh, it has overtaken Pakistan because it has higher female participation in the work force.
Since a certain amount of multi-tasking enhances productivity, it is better to have a subset of the population multi-task, that is, do all the work, rather than spread tasks thinly across all.
Basu thinks there is only a fixed quantity of useful work which can be done at any given moment. This is not true. There is a trade-off between leisure and work. Getting rid of discrimination increases work incentives and boosts output. Leisure may have fallen but it is likely that the quality of leisure goods has increased so everybody's utility has increased though they work more hours.
Some people may find this result troubling, namely, the fact that discrimination against a group can enhance GDP.
It is not a 'result'. It is nonsense. Discrimination means a group is paid less than its marginal product. This has a disincentive effect. It may also contribute to crime. If you can barely feed yourself doing a dead-end job, why not mug people instead? The police may catch you and send you to prison where the tax-payer pays for your meals and medical care and so forth. In Japan, elderly people have started committing crimes simply because they get better care in jail.
In my view, this result is disturbing only if you consider a higher GDP to be sacrosanct.
Bengali buddhijivis think the path to virtue is to be very poor and very weak. Hopefully, your country will be invaded and you yourself will be enslaved or killed.
All this result suggests is that you should be willing to forego some GDP to achieve fairness and greater equity across groups.
Till the country is unable to feed or defend itself and gets invaded. The problem with listening to Bengali buddhijivis is that your country may become as shitty as West Bengal.
It is not a matter to go into here but this reference to commonsense and ‘reasoned intuition’ is not a casual side remark. I have argued at length elsewhere that for science to be useful we must combine it with these skills.
But what Basu is doing is not 'science'.
Pure analysis of data or pure theory cannot help us help the world till we combine them with reasoned intuition (Basu, 2014). 6 Among the most notable findings on this are studies by Ambadi, Shih, Kim and Pittinsky (2001) and Hoff and Pande (2006). See, also, Field and Nolen (2005), Hoff (2015) and World Bank (2015). 9 Stewart (2005) pointed out that, in contrast to vertical inequality (that between persons), the normative economics of horizontal inequality (inequality across groups), remains a rather neglected subject.
But all these silly people are utterly useless and unproductive.
Exactly how one develops this and characterizes the various trade-offs between aggregate well-being and horizontal equality can matter a lot in how choices are made, as we just saw (see also Jayadev and Reddy, 2011, and Subramanian, 2011).
But nobody listens to these useless nutters when it comes to actually making choices.
Fortunately, the context is simple enough in the discussion that follows that the exact trade-offs will not matter but this is indeed a subject that deserves greater attention in the future.
That's because the subject shat the bed and deserves the attention of those who like looking at shitty beds.
With this in the background, note that there are two ways of achieving equity or an equitable distribution of incomes or payoffs in the present model. The first is straightforward ‘affirmative action.’ Incentivize employers to choose a diverse workforce so that the total work is spread equitably across all individuals.
But this may reduce the 'total work' that is done. There will be exit to other jurisdictions.
The second is to let a limited number of people do all the work and then tax them and subsidize those who did not get work.
The more productive workers will emigrate or find ways to evade taxes by getting fringe benefits or just slacking off.
This latter will result in a higher per capita income since the workforce (that is, the people who find work or are called upon to do tasks) will be more productive.
Till the emigrate or start slacking off.
There are social scientists who have objected to the latter on the ground that work in itself gives people dignity; so even if one were to get the same ultimate income but without having to work, this may cause a diminished sense of self.
Not having a pot to piss in diminishes 'sense of self'. Winning the lottery and giving up your dead-end job has the opposite effect.
I prefer to be cautious with this argument. Through history there have been the leisure classes—the British landed aristocracy, the Indian Zamindars (a brainchild of the British landed aristocracy)
Fuck off! There were jagirdars and zamindars and taluqdars in India long before the Brits got there.
—who did precious little work and lived lives of luxury, and there is no evidence of them feeling diminished by the experience.
To be fair, they did spend a lot of time trying to kill their brothers or cousins- till one of their wives poisoned them.
There are also cases, especially relevant to women, where voluntary non-work is, in fact, a statement of empowerment and an act that enhances agency (Alaka Basu, 2016). Basically, what people need is a sense of legitimacy for what they earn.
I will supply it to them for the low low price of $9.99.
There are contexts where to get a dole without getting to work is offensive;
you are welcome to do voluntary work.
it is like being told that you are not fit for work.
That hasn't stopped Basu and his ilk.
This never troubled the aristocracy and the leisure classes because they had a sense of entitlement for their ample incomes, even though where that entitlement came from is a puzzle.
The Clergy offered a convenient explanation. Malthus was a clergyman. Come to think of it, he was also a Professor at the East India Company College.
In a world where, either one group gets to work or the other does, and there is no essential a priori difference between the two groups, for one group to work and the other to be subsidized is a legitimate strategy.
It is not sustainable. There will be emigration or tax evasion or just plain plummeting productivity. The country goes off a fiscal cliff. There is entitlement collapse.
Someone has to not work to make those who work more productive;
This is like Malthus's argument that there will be an under-consumption crisis unless the idle rich fritter away their fortunes.
in this setting there is no indignity in not working. It is in fact a contribution to society.
Only in the sense that rape is a contribution to increasing the population.
There is no need to resolve the policy dilemma in this paper but I want to point out that this problem is going to get more acute,since with the march of technology, relatively-unskilled work is steadily shrinking in the world.
but there are shortages for skilled workers. One reason for this is that young people are fooled into studying worthless shite at Uni.
The share of GDP that accrues to workers as aggregate wage bill has been falling over the last four or five decades.
Which means that retirees can live longer with a better quality of life. Pension funds own a big chunk of industry. Sadly, this may not be sustainable. Us boomers may face a harsher old age than we anticipated.
The trend is quite alarming. As I point out in Basu (2016), in 1975, total wage bill as share of GDP in United States, Japan and the European Union, were 61%, 77% and 66%. Now they are 57%, 60% and 56%, respectively (see also Karabarbounis and Neiman, 2014). This trend is true almost without exception in all high and middle income countries; and this is a challenge we will have to face up to sooner rather than later.
Migration may offset the higher dependency ratio arising from an ageing population.
What the model developed in this paper tells us is that we face a choice--whether to forcefully distribute the limited work
work isn't limited
thinly across the entire labor force (thereby impairing productivity)
there will be negative net marginal product.
or let few people work (and be productive) and then tax them to subsidize the ones without adequate work and with all the time in the world to read philosophy and swim.
This is not sustainable. We need the workers and can do little to stop them increasing their fringe benefits and spending more time reading philosophy in the office or taking the afternoon off to go swimming.
In itself, the advance of technology is a matter for celebration. What makes it worrying is that, as machines and robots displace workers, the incomes of workers become profits for the owners of the machines and the robots.
this would be institutional investors- e.g. pension funds. We can rob the old to help the young but there is a limit to what we can do in the other direction because the young can up-sticks and emigrate. Greek pensioners found this out the hard way.
I have discussed in Basu (2016) how we need to take on this problem head on.
The rest of us were discussing how useless Basu & Co have always been.
What this paper points to is an additional problem. As work becomes scarce, in markets with strategic complementarity, there will likely be an exacerbation in the problem of group discrimination, with large groups being kept out of the labor market.
By Government policies. People with no access to benefits can't be kept out of the labor market.
This paper provided an explanation of the mechanics of how this will happen and drew attention to the kinds of policy questions that, sooner or later, we will have to confront
This paper is utterly foolish. 'Complementarity' can be engineered. There are smartphone apps which do so. A worker may know zero about Accountancy or Engineering, yet as part of their work routine, they can be very accurately updating diagnostic databases for the Accountancy or Engineering Department.
It is quite true that some groups have very effective communicative and cooperative mechanisms. But what a good enterprise does is it actively fosters a workplace culture in which just such mechanisms operate even more effectively not just within the enterprise but also with suppliers and clients.
Discrimination depends on segmenting the market. As Information becomes cheaper and more ubiquitous, profit opportunities are created so the thing is reduced or eliminated. This is not to say that Competition Policy or the prosecution of unfair trade practices can be wholly given up. However, the market, by itself, has an incentive to tackle the problem.
No comments:
Post a Comment