Defining discrimination as basically preferences based on something other than productivity, it makes little sense for employers to reject applicants for reasons that have nothing to do with their capacity to do the job, particularly when they will work for less. But still, there exists a wage premium for whites in particular, albeit one that has been vanishing for Asians, who have gained the status of "honorary whites". So the situation looks strange from the outside.
So, viewing discrimination as an acquired taste that employers would pay to enjoy, Becker's model develops a discrimination coefficient, based on the wage differential between the minority worker and other workers, and the numbers of minorities hired. It makes for an interesting graph: a line holds straight across for employers who do not discriminate - they value minorities and other workers the same, basically the red line until you get to the drop-off - then dog-legs down as the willingness of increasingly prejudiced employers to pay them fades. As a model, it's workable and probably fairly accurate, basically saying that discrimination is costly to the discriminator because he passes up capable labor and pays more for labor in the preferred demographic.
We need not get into the conversation about just how much of this discrimination is pure bigotry and how much of it might be justifiable good sense in hiring. If employers don't want to hire someone, they frequently have a reason that the statistics don't show, and it's their prerogative; if they don't want to pay someone equally, those people have the right to leave. If they don't leave, it just proves the power of money and the hierarchical nature of economics, and that's okay. I prefer the attitude that sees "discriminating taste" as a compliment instead of an insult, and business cannot function without the employer having authority over the employee.
Becker is a much more interesting character when it comes to family. The taste for discrimination model is applied to women as well as racial minorities, which may or may not have actually held for the guy who created it. Becker researched family economics a lot; his book on the subject is one of the most widely quoted works in economics, and I wonder just how strongly Becker understood the situation facing family and the reasoning behind those who do not hire equally.
The Conservative Family Model
It was basically impossible for me to take the taste for discrimination model at face value when applied to women, because I have a very different idea as to why employers have hesitated to hire and pay women as if they were men.
The old-world model of family is one in which the genders have different roles within the marriage structure. This is probably best displayed by my great-grandmother, an extremely intelligent women who spent World War 2 as an important supervisor for Western Union. This was the Rosie the Riveter period, but when the war ended, she gave up her job and went back to keeping the home. She is one of the few people I know of who explicitly stated the arrangement: the man works to provide for the family, and a woman taking a serious career position from a man is denying some other family of an income, not just an individual. She knew the arrangement, and did not complain when it was time to move to the side. Today, feminists would probably call her a coward for supporting the patriarchy.
But anyway, I think the male premium stems from employers paying men more because they have the same social model in mind, and believe a family is being supported on that paycheck. There's a marriage premium, too, that exists for the same reason. The marriage premium is massive, and I don't think that performance fully accounts for it; the increased performance of married men is probably overstated by their bosses, who likely respect married men more. The betrothed probably work harder, but not that much harder. That breadwinner image and status creates a difference between the way an employer pays a single or second-income versus a married man, and if you hold a traditionally structured view of society, this is simply the right thing to do. It's a positive prejudice, one that looks an awful lot like a sensible one and not simply a "taste". Employers, believe it or not, don't hate the people that work for them, and they try to cut a balance in their pay between paying well but not draining the bank account and destroying the business. I think that's particularly true of small businesses, which still hire most people.
So what did Becker think? He didn't explicitly say that more women in the workforce was a bad idea, but at one point he basically said we had done enough to promote equality, and a few different commentators took jabs at him for it. He seems to be the type that prefers to see the family as a good institution and prefer that the structure remain intact; it says a lot that he truly thought of it as a structure at all, enough to research it as one. But I doubt he would be the first researcher to establish concepts with a tongue-in-cheek mentality. Marriage would still be valuable if children were seen as the point of it and the parents' happiness was secondary or irrelevant, and it becomes fairly clear when you live around both married people and single people just how radically all of society will have to shift as the norms shift for no good reason. The numbers and the observed outcomes look better for traditional families, if not always for the individuals in them, usually for the children. Becker might well understand this even now. The children are simply better served by one parent being at home and another at work, and the genders had centuries of refinement of the roles to fall back on before they abandoned them.
Freedom versus Structure
Today, this view is so obsolete as to be laughable. For one thing, economists promoting growth have no respect for the zero-sum understanding of labor displayed by my great-grandmother. They do not see jobs as a resource which can be over-exploited; if anything, they think the more everyone works, the more everyone creates economic growth, the more the market for labor expands into infinity. This attitude is one reason I sometimes question the entire field of economics: you cannot look at the labor situation and not see a relative hierarchy, but economists pretend that they don't, refusing to look at economics as anything but a world of individuals maximizing their welfare. The more familiar reality of economics, that of a hierarchical social environment where individuals try to survive by joining groups and maintaining the relative value they have to others, never gets real play. The ideas economists have come up with to explain away the gaps between their models and reality are stunning in their elegance and absolutely ridiculous in practice.
For another thing, society will not ideologically accept the structured view anymore. The modern, liberal world wants to establish a playing field that is totally individualistic and as equal as possible. That's justice in their eyes. The family, as a hierarchical institution, holds no agency or validity except as a source of happiness for the individuals in it. This is probably the greatest sea change in social roles in the history of man, it's fairly recent, and it's taken as absolute dogma that such an individualistic model is morally correct. The alternative notion, that justice is about people meeting their obligations in proven and workable social roles, seems more likely to invite scorn as oppression, and of course the arrangements die if society as a whole has no faith in them.
In any case, the male premium will probably not last but another generation or two, and I'd gamble that the marriage premium will slide a bit over the same period as marriage changes as an institution. I've said in the gay marriage post why I think marriage is becoming a purely romantic and pragmatically pointless arrangement. The decline of gender roles and even greater decline in the expectation to stay married add to that perception, while the safety net and absolute cultural permissiveness make permanent single parenthood viable. Marriage is no longer needed, so while it's going to take a long time for the image of romance as lifelong to fall apart, the importance of marriage structurally is all but gone already.
To lots of people, the elimination of the pay gap is progress. To me, it's the de-structuring of society, and if I were to open a business and hire employees, anti-discrimination laws would prevent me from practicing what I believe to be good citizenship. Not only is the positive prejudice laughable, but it's now also illegal. I would be pissed if I thought that freedom held any validity as a concept. It's frustrating enough to be an anachronism, believe me.