
Bryan Caplan on Discrimination and Labor Markets 12/4/2006
What this covers
Bryan Caplan and Russ Roberts examine the relationship between labor market competition and discrimination, asking whether profit incentives naturally erode bias or whether government regulation is necessary. Caplan's central claim is economic: if wage gaps reflected widespread discrimination, hiring underpaid workers would present an obvious arbitrage opportunity, yet no such scheme pervades the business landscape. He argues instead that a distribution of discriminatory tastes across employers means the least-biased firms gain a competitive edge, gradually driving out discrimination over time. Roberts and Caplan explore how this mechanism actually functioned in historical cases—Jim Crow laws that criminalized worker recruitment, apartheid regulations protecting white workers from market competition, and the Davis-Bacon Act designed to exclude non-unionized black labor—treating these as evidence that governments often suppressed rather than created anti-discrimination competition.
The conversation ranges across labor regulation in America and Europe, comparing unemployment outcomes and job security claims. Caplan argues that European regulations create a bifurcated labor market: secure jobs for insiders paired with persistent unemployment for outsiders, while the lighter American regulatory regime produces lower joblessness despite weaker enforcement. He contends that since unemployment damages wellbeing beyond its income loss, the American system is ultimately more humane. The speakers also address whether additional income drives happiness (their answer: beyond subsistence, minimally), consider libertarian paternalism as a middle path between coercion and free choice, and examine how wage floors like minimum wages can enable discrimination by letting employers select workers on non-price criteria. A recurring tension surfaces: whether regulations' stated intentions match their actual effects, and who bears the cost of well-meaning but poorly designed policy.
Caplan argues that competitive labor markets systematically punish discrimination because profit-seeking employers can profit by hiring underpaid discriminated-against workers, so most government 'anti-discrimination' regulation historically either entrenched discrimination or harmed the very groups it claimed to protect.
- A distribution of discriminatory tastes means the least-racist employers outcompete the more racist, eroding discrimination over time
- If observed wage gaps were largely discrimination, there would be an obvious get-rich-quick scheme in hiring underpaid groups, which there isn't
- Minimum wages, Davis-Bacon, Jim Crow laws and apartheid show regulation often suppressed competition that favored disadvantaged workers
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
Southern Jim Crow laws (c. 1895-early 1900s) — criminalizing enticing a contracted worker to switch firms, fining recruiters up to $5,000 for recruiting workers out of state, and vagrancy laws sending the unemployed to chain gangs — were designed to suppress competition for black labor that the market was eager to take advantage of, by reducing labor mobility and the attractiveness of black labor.
“It was against the law in many states to entice a laborer who was employed to switch firms... if you were under contract as the worker and you switched jobs, you could actually be arrested.”
The minimum wage, by creating a surplus of labor, lets employers select among equally good workers on non-price margins such as personal liking or who looks most like themselves — analogous to rent control where landlords pick favored tenants — so wage floors enable discrimination.
“When you cannot compete in terms of price, when the wage can't go down, uh then you wind up competing on other margins... you might pick the worker that looks most like what looks most like you.”
Happiness research finds that once people have enough income to meet basic needs, additional income does little for happiness, but unemployment itself makes people deeply miserable even when their income is fully replaced — implying joblessness harms wellbeing beyond its financial cost.
“unemployment per se causes a lot of unhappiness... even if you go and give them the same income they would have had when they were employed, they're still a lot less happy”
Because labor is the main cost of doing business, even a small share of an observed wage gap being due to discrimination implies enormous profit gains — e.g. if 10% of a 33% gap were discrimination, an employer could cut costs by more than 3%, roughly doubling a typical 3% rate of return to 5%, which is the holy grail of business.
“if 10% of a 33% difference is due to discrimination that that means you could cut your what your costs by a little bit more than 3%”
Because there is a distribution of discriminatory preferences, employers who care only about profits will outcompete those who weigh both profits and race, since the profit-focused firm hires the more productive worker, achieves lower costs, and can sell at lower prices or higher quality — so the least racist employers gain a market edge and come to predominate over time.
“if you have people who are people who only care about profits competing about competing with people who care about pro both profits and race, uh you should expect it over time the people only care about profits are going to come to predominate”
If the common view of discrimination were correct, hiring members of underpaid groups would be a trivially easy get-rich-quick scheme requiring no skill, so the absence of such schemes on every corner suggests observed wage gaps are not mainly discrimination — unlike genuinely hard money-making like real estate or alpaca farming.
“it would not take a lot of brains to to set up a firm where your get-rich-quick scheme is hire only women. Anyone could do that.”
Apartheid in South Africa originated partly because white employers were profitably hiring lower-wage black workers over white workers, so regulations were passed to stop employers from doing this and protect the white race from market competition — government working against the market's tendency to reduce discrimination.
“the origin of a lot of apartheid in South Africa was that you have uh white workers who were losing jobs to black workers because white employers... said... the black worker is willing to work for less. And so I will make more money if I'll hire a black worker.”
Thaler's Swedish-retirement example is misleading because he compared private choice against the single best-designed government retirement plan ever (which had low costs and avoided common investor errors), rather than against the average of government-chosen plans, many of which have been quite bad — including the US's own.
“The key thing to remember is he went and picked the very best designed government retire plan ever. And then compared private choice to it rather than averaging over all the government retirement plans... many of which have been really quite bad.”
Where the least-racist local employer is still quite racist, those who want discrimination to persist must block exit to less-racist regions; this explains why the more-racist South feared northern (less racist) employers bidding black labor away, which ultimately produced the Great Migration.
“part of the concern of people who like that situation and want to make sure that discrimination persists is to make sure that there isn't any exit to a place where the least racist person is less racist”
The Davis-Bacon Act (1931), requiring government-project employers to pay the prevailing (union) wage, was historically motivated by a desire to discriminate against unskilled, non-unionized black labor, despite being championed as protecting workers from exploitation.
“if you look at the history of it... the motivation for keeping out keeping wages high was to discriminate against unskilled black labor, which was not unionized at the time”
Europe's apparent job security is largely an illusion: regulation gives more security to those who already hold jobs but leaves many others 'securely unemployed' with little prospect of work, so claiming Europeans in general have more security is wrong.
“what you have in Europe is if you are lucky enough to currently have a job, you have more security than you do have in the US... there's a lot of people who don't have jobs who would like to get them. And they are securely unemployed.”
US labor regulations are not only less restrictive than Europe's but also much more weakly enforced, illustrated by union-organizing law: the US and Canada have identical statutes making it illegal to fire someone for organizing, but in Canada cases are fast and remedied with reinstatement, while in the US they are slow, often lost, and even when won yield only the wage difference net of subsequent earnings — which paradoxically helps explain the lower US unemployment rate.
“In both countries, it's illegal to fire someone for trying to organize a union. But... in Canada... it's very easy for him to get the case in court... he gets reinstated... In the US even though the law is exactly the same... all that you get is the difference between the earnings”
A common policy error is neglecting the dynamic effects of regulations and presuming that the intention of a law equals its outcome.
“it's a common problem, people neglecting the dynamic effects of various regulations, and just looking at the immediate outcomes, and presuming that the intention of a law is equal to its outcome”
No good economist favors total freedom of choice; the real question is who controls the restrictions on choice and what their incentives are — we routinely and rationally defer to experts (e.g. Google narrowing options, a surgeon calling for a scalpel), and the choice to defer is itself a choice, so true freedom means being able to choose to make up your own mind when you want to.
“certainly no good economist is interested in total freedom of choice. It's question is who controls the restrictions on choice.”
The market punishes discrimination because hiring workers of equal ability based on race or sex rather than merit means missing profit opportunities, so over time discrimination tends to die out because it is costly.
“the marketplace will punish discrimination, and discrimination will eventually die out because it's costly”
Most of the gap in standard of living between Americans and Europeans is attributable to the US having traditionally more free-market policy while Europe is more regulated and statist.
“a very big part of the reason is the US has traditionally had more free market policy than Europe. Europe is more regulated, more statist, and I think that accounts for most of the difference”
High minimum wages and similar wage-floor regulations in much of Europe produce persistent decades-long high unemployment and make it especially hard for lower-skilled immigrants to find work, because wages cannot fall to a level that makes hiring uncertain or lower-productivity workers worthwhile.
“in a lot of European countries, you have very very high unemployment, which has persisted for decades... immigrants in Europe have a tougher time finding a job because if the wages can't come down, then what's the point of hiring someone”
Libertarian paternalism — setting a default option chosen as best for the person while leaving them free to opt out — would be an enormous improvement over coercive paternalism, since anyone who genuinely cares can easily change the default, so only those who don't care much are affected.
“if you are opposed to paternalism, I see every reason to think see libertarian paternalism as an enormous improvement... you would only be affecting people who really don't care that much anyway”
Because unemployment causes misery independent of income, the American system of lighter labor regulation — which lowers unemployment and eases job-finding — is more humanistic and concerned with the whole person than the European system, which assumes that sending an unemployed person a check makes up for joblessness.
“the American system of less labor market regulation, which leads to lower unemployment and makes it easier to find a job, is really more concerned about human happiness... than the European system, which basically assumes that if you just give someone a check, they have nothing to complain about”
When you control for education, IQ score, family structure, marital status, number of children and age, the entire black-white gap in annual labor earnings disappears in 1990s data, and the 'other races' category actually shifts from earning less to earning slightly more once controls are added.
“if you just put in uh really a fairly short list, but in particular the IQ measure, you can actually see the entire black-white gap goes away”
Libertarian paternalism is a worthwhile improvement worth trying if it replaces existing coercive paternalism, but the real worry is that it will be layered on top of existing paternalism or act as a slippery slope toward real paternalism (Glaeser's critique).
“If we could step one, get rid of all the actual non-libertarian paternalism we have. Step two, replace it with some libertarian paternalism. I wouldn't be losing sleep over that.”
Recently the minimum wage has become academically acceptable largely because of two studies suggesting it doesn't hurt employment, reversing an earlier large body of studies showing it reduced employment among unskilled workers and was therefore discriminatory against blacks.
“minimum wage now has suddenly become okay, perhaps because of I think two studies that suggested that it doesn't hurt employment. But there was before that an enormous number of studies that suggested that the minimum wage did reduce employment”
Free movement of workers within the European Union is legally permitted but rarely exercised, partly because wage floors keep wages up so natives are preferred over newcomers, and partly because people simply prefer those from their own country — regulations lowering the cost of acting on that preference.
“people in Europe in the European Union are now generally allowed legally to migrate to other countries... but you don't see so much of it. And probably part of it is just because wages are kept up”
FDA over-caution makes American over-the-counter drugs mild and discourages innovation due to compliance costs, so drugs in America are probably 'too safe,' while less restrictive regimes (e.g. Chile) offer more powerful drugs that consumers may be inexperienced at evaluating trade-offs for.
“as the podcast with Sam Peltzman illustrates, drugs in America are probably too safe. The FDA is so careful it discourages innovation because of the costs of compliance.”
Europeans have a markedly lower standard of living than Americans on concrete measures — car ownership, meat consumption, and home/lot square footage — which official purchasing-power-parity per-capita income figures fail to capture; Japan scores even worse on these measures despite high PPP income.
“there's a number of ways in which the living standards of most Europeans are markedly worse than those in the United States in terms of car ownership, eating meat, uh square footage of your home”
Caplan's argument that discrimination is easily arbitraged away is a bit of a straw man because the realistic question is not whether all observed wage differences are discrimination but whether some residual remains; the residual could be small but still persist.
“it's a bit of a a straw person scarecrow example you're using because... you're arguing that the common view of discrimination is... all due to discrimination”
It is paradoxical to call it pessimistic that markets fail to eliminate discrimination, because the usual critique of markets is that businesses care only about profits; yet caring only about profits is precisely what eliminates discrimination, so the case where profit-focus removes discrimination is being treated as too optimistic.
“here's a case when if if people only care about profits, discrimination won't happen. And then this is seen as being too optimistic.”
When less-racist employers visibly profit more, even formerly racist people may revise their preferences, deciding they don't actually care about race as much as they thought once they see the money to be made by hiring on merit.
“now that I see all this money that I can make by hiring based on merit”
Atrophy of decision-making from accepting defaults can be optimal: like a restaurant's default burger toppings, defaults save effort on things people don't care about, and on things they do care about the default won't matter because they have a genuine preference and will override it.
“it's not so clear to me that atrophy can't be optimal... the burger comes with certain toppings by default and if you don't want that then you have to go and tell them”
Anticipatory market behavior can erode discrimination quickly because, like the internet, people may jump the gun and over-invest in hiring discriminated-against groups before it is proven profitable, potentially even going overboard so that hiring only the underpaid group becomes a bad idea.
“What we saw with the internet is when people thought there was a new business opportunity in the future, they started doing it before it was would actually work.”
The Netherlands mitigates the harm of wage regulations through a grandfather provision under which newly hired workers are not subject to the same regulations as existing workers, encouraging employers to take a chance on new hires.
“the Netherlands often has a grandfather provision where if you hire a new worker, he won't be subject to the same regulations as an existing worker”
Economists agree on Becker's basic point but debate at the margin how completely the market eliminates discrimination — whether it removes 100%, 90%, or 80% of it.
“does the market actually eliminate discrimination 100%? Does it eliminate 90%? 80%?”
Apartheid began coming in during the 1920s, so there is unlikely to be good data on whether the black-white wage gap was shrinking as employers hired cheaper black labor before the regulations.
“I think apartheid was coming in in the '20s. So it's unlikely that there's any very good data on that unfortunately.”
There has been a narrowing over time of the absolute uncontrolled wage gap between men and women and between blacks and whites.
“there has been certainly over time uh a narrowing of the absolute uncontrolled for wage gap, say between men and women or blacks and whites”