The myth of studying to language to build mental muscles – Bryan Caplan


The dark underbelly of the signalling value of engagement rings


Source: Newmark’s Door: Some notes on comparable worth.

Drug testing helps minority job applicants @jacindaardern

Statistical discrimination is a harsh mistress. If reliable measures of the quality of job applicants are unavailable for short-listing, such as credit checks, coarser, less reliable screening devices will be employed. That was the case when credit checks were prohibited in employment recruitment:

Looking at 74 million job listings between 2007 and 2013, Clifford and Shoag found that employers started to become pickier, especially in cities where there were a lot of workers with low credit scores. If a credit-check ban went into effect, job postings were more likely to ask for a bachelor’s degree, and to require additional years of experience.

There are other ways that employers could have also become more discerning, Shoag says. They might have started to rely on referrals or recommendations to make sure that applicants were high-quality. In the absence of credit information to establish trustworthiness, they may even have fallen back on racial stereotypes to screen candidates. The researchers couldn’t measure these tactics, but they’re possibilities.

Drug testing allows employers to dispel less accurate stereotypes about drug use among different ethnic and social groups. They increased hiring of minorities because a reliable measure became available of their drug use:

…after a pro-testing law is passed in a state, African-American employment increases in sectors that have high testing rates (mining, manufacturing, transportation, utilities, and government).

These increases are substantial: African-American employment in these industries increases by 7-30%. Because these industries tend to pay wage premia and to have larger firms offering better benefits, African-American wages and benefits coverage also increase. Real wages increase by 1.4-13% relative to whites. The largest shifts in employment and wages occur for low skilled African-American men.

I also find suggestive evidence that employers substitute white women for African-Americans in the absence of testing. Gains in hiring African-Americans in these sectors may have come at the expense of women, particularly in states with large African-American populations.

Employers test for drug use both for health and safety reasons and as a way of screening out less reliable employees. People who break the rules are not reliable employees and that includes taking drugs. In low skill jobs, what employers seek is a recruit who is friendly and reliable.


Testing of the skills of workers also showed similar results. What happened is that the ratio of black to white hirings do not change. The administration of these skills tests allowed the more productive of both white and black job applicants to be identified and hired.

Employers already had an accurate stereotype of the average skills of different ethnic groups. Administration of tests allow them to identify which members of each group were the most productive.

It is a standard result that statistical discrimination improves the chances of below-average applicants subject to the stereotype but harms those of above average quality. For that reason, applicants look for what methods of counter-signalling to show that they are indeed a quality applicant – make themselves stand out from the crowd.

Employers profit from developing screening devices that go beyond stereotypes to identify above-average applicants. They want screening devices that find those who do not otherwise stand out from the crowd because of difficulties in transferring credible information about their quality. This is a special difficulty with low-skilled vacancies because hiring is made based much more than on character than experience.

There is an economic literature on feuds and duelling

Charles Murray and the OECD’s Trends in Income Inequality and its Impact on Economic Growth – IQ, signalling, over-education and plain bad career advice

Charles Murray has been cooking with gas lately – on fire. One of his points is too many go to college. Murray points out that succeeded at college requires an IQ of at least 115 but 84% of the population don’t have this:

Historically, an IQ of 115 or higher was deemed to make someone “prime college material.”

That range comprises about 16 per cent of the population.

Since 28 per cent of all adults have BAs, the IQ required to get a degree these days is obviously a lot lower than 115.

Those on the margins of this IQ are getting poor advice to go to college. Murray argues that other occupational and educational choices would serve them better in light of their abilities and likelihood of succeeding at college. Moreover, Murray is keen on replacing college degrees with certification after shorter periods of study such as in the certified public accountants exam.

Murray believes a lot of students make poor investments by going on to College, in part, because many of them don’t complete their degrees:

…even though college has been dumbed down, it is still too intellectually demanding for a large majority of students, in an age when about 50 per cent of all high school graduates are heading to four-year colleges the next fall.

The result is lots of failure. Of those who entered a four-year college in 1995, only 58 per cent had gotten their BA five academic years later.

Murray does not want to abandon these teenagers:

Recognizing the fact that most young people do not have ability and/or the interest to succeed on the conventional academic track does not mean spending less effort on the education of some children than of others.

…Too few counsellors tell work-bound high-school students how much money crane operators or master stonemasons make (a lot).

Too few tell them about the well-paying technical specialties that are being produced by a changing job market.

Too few assess the non-academic abilities of work-bound students and direct them toward occupations in which they can reasonably expect to succeed.

Worst of all: As these students approach the age at which they can legally drop out of school, they are urged to take more courses in mathematics, literature, history and science so that they can pursue the college fantasy. Is it any wonder that so many of them drop out?

To add to that, he is in the Bryan Caplan School: education is often an elaborate former of signalling for many degrees. Murray says that college is a waste of time because:

Outside a handful of majors — engineering and some of the sciences — a bachelor’s degree tells an employer nothing except that the applicant has a certain amount of intellectual ability and perseverance.

Even a degree in a vocational major like business administration can mean anything from a solid base of knowledge to four years of barely remembered gut courses.

If the OECD is to be believed, that not enough people are going to college from lower middle class families, obviously IQ is not one of the constraints on access to college Charles Murray suggested it to be.


The growing strength of the case that education is a form of signalling is a literature that the now famous OECD paper reviewed, found wanting, but did not have time to discuss in the working paper.

Another contemporary theme the OECD paper reviewed, found wanting, but did not have time to discuss is a large number of graduates who end up holding jobs that do not require a university education – going to college:

About 48 per cent of employed U.S. college graduates are in jobs that the Bureau of Labor Statistics (BLS) suggests requires less than a four-year college education.

Eleven per cent of employed college graduates are in occupations requiring more than a high-school diploma but less than a bachelor’s, and 37 per cent are in occupations requiring no more than a high-school diploma.

The proportion of overeducated workers in occupations appears to have grown substantially; in 1970, fewer than one per cent of taxi drivers and two per cent of fire-fighters had college degrees, while now more than 15 per cent do in both jobs

All in all, the OECD has gone into the dragons den by backing the accumulation of human capital as its mechanism to link inequality with lower growth. No matter how you spin it, this linking of lower economic growth to greater inequality through financial constraints on the accumulation of human capital by the lower middle class was a bold hypothesis.

The case for investing more in education is not a slam dunk. Higher education – university or polytechnic – is a rat race that many don’t need to join.The case for the government paying a great many more to join that rat race is rather weak.

Claudia Goldin’s pollution theory of sex discrimination

Claudia Goldin argues that it is  difficult to rationalise sex segregation and wage discrimination on the basis of men’s taste for women in the same way as discrimination based on race or ethnicity. Goldin developed a pollution theory of discrimination in which new female hires may reduce the prestige of a previously all-male occupation.

When work took more brawn than brain, the distributions of skills and natural talents of men and women were further apart. Women were not as physically strong as men. This counted for more both before the Industrial Revolution and at the height of the Industrial Revolution when most factory work involved a considerable amount of brawn.

As machines substituted for strength, as brain replaced brawn and as educational attainment increased, the distributions of attributes, skills and natural talents narrowed by sex.

Because there is asymmetric information regarding the value of the characteristic of an individual woman, a new female hire may reduce the prestige of a previously all-male occupation.

Prestige is conferred by some portion of society and is based on the level of a productivity-related characteristic (e.g., strength, skill, education, ability) that originally defines the minimum needed to enter a particular occupation. People had to have a minimum amount of the socially prestigious strength or skill before they are hired.

Male fire fighters or police officers, to take two examples, may perceive their occupational status to depend on the sex composition of their police station or firehouse. These occupations are socially prestigious because of the strength and courage of police and fire-fighters. Men in an all-male occupation might be hostile to allowing a woman to enter their occupation even if the woman meets the qualifications for entry.

A reason for this hostility of the existing male members of the occupation is the rest of society may be slow to learn of the qualifications of these female newcomers. Their entry against this background of ignorance in the wider society  may downgrade the occupation as still carrying prestigious characteristics such as physical strength. As Goldin explains:

Because they feel that the entry of women into their occupations would pollute their prestige or status in that occupation. Very simply, some external group is the arbiter of prestige and status.

Let’s take an example of firemen, and let’s say we begin not that long ago when there were no women who were firemen—which is why they’re called firemen.

And to become a fireman you have to take a test, lifting a very heavy hose and running up many flights of stairs. And every night, the firemen get off from work and go to the local bar.

Everyone slaps them on the back and says what great brawny guys they are and what a great occupation they are in, and everybody knows that to be a fireman requires certain brawny traits and lots of courage.

But nobody knows when there’s a technological shock to this occupation. And in this case it might be that fire hoses become really light or the local fire department changes the test. There are information asymmetries. But they do note that for this “brawny” characteristic, the median woman is much lower.

So if we observe a woman entering the occupation and we don’t know how to judge women, we’re going to assume that her skills are those of the median woman. Or it may be that we can observe something having to do with her muscles and that may up it a little bit.

But chances are we’re going to assume that some technological shock has happened to this occupation. And so her entry into the occupation is going to pollute it.

Then when they go to the bar, people will say, “oh you’ve got a woman in the firehouse; now fire fighting has become women’s work.” That’s where the pollution comes in.

Union rules also played a role in preventing the entry of women into some occupations

Many occupations have changed sex over time e.g., librarians, bank tellers, teachers, telephone operators, and sales positions. New occupations  and industries are less like to be segregated on the basis of sex  because they have not developed a social image regarding the prestige of workers.

Occupational segregation came to an end because credentialisation, which spreads information about individual women’s productivities and shatters old stereotypes, can help expunge this pollution of the prestige of specific occupations and jobs both within the industry and in wider society .

The visibility successful women today and in the past may help shatter old stereotypes and increase knowledge about the true distribution of female attributes in this prestigious occupation.

Goldin found that  when typists were primarily men, it was claimed that typing required physical stamina so woman need not apply.

But later, when the occupational sex segregation reversed, when typing became a female occupation, it was said that typing required a woman’s dexterity, which men did not have! When I was at school, only women were taught to type.

At least 20% of New Zealand workers are subject to occupational regulation

There are at least 98 regulated occupations in New Zealand covering about 20% of the workforce. In 2011, this amounts to 440,371 workers. The skills that are regulated range across all skill sets and many occupations:

  • 49% of regulation is in the form of a licence;
  • 18% of regulated work is in the form of licensing of tasks;
  • 31% of regulated workers require a certificate; and
  • 4% of regulated workers require registration.

There are 32 different governing Acts that regulated occupations in New Zealand with 55% of the workers subject to occupational regulation are employed in just five occupations:

  • 98,000 teachers;
  • 48,500 nurses;
  • 42,730 bar managers;
  • 32,733 chartered accountants; and
  • 22,749 electricians.

The Health Practitioners Competency Assurance Act 2003 regulates 22 occupations and a total of 89,807 workers. The next best is the 10 occupations regulated by the Health and Safety in Employment Act 2002 which regulates an unknown number of occupations. The Civil Aviation Act 1990 regulates eight occupations and 19,095 workers, the Building Act 2004 regulates seven occupations and 21,101 workers and the Maritime Transport Act 1994 regulates six occupations and 20,500 workers. 12 of the regulated occupations are regulated under laws passed since 2007.

The purpose of occupational regulation is to protect buyers from quacks and lemons – to overcome asymmetric information about the quality of the provider of the service.

Adverse selection occurs when  the seller knows more than the buyer about the true quality of the product or service on offer. This can make it difficult for the two people to do business together. Buyers cannot tell the good from the bad products on offer so many they do not buy to all and withdraw from the market.

Goods and services divide into inspection, experience and credence goods.

  • Inspection goods are goods or services was quality can be determined before purchase price inspecting them;
  • Experience goods are goods whose quality is determined after  purchase in the course of consuming them; and
  • Credence goods are goods  whose quality may never be known for sure  as to whether the good or service actually worked – was that car repair or medical procedure really necessary?

The problem of adverse selection over experience and credence goods present many potentially profitable but as yet unconsummated wealth-creating transactions because of the uncertainty about quality and reliability.

Buyers are reluctant to buy if they are unsure of quality, but if such assurances can be given in a credible manner, a significant increase in demand is possible.

Any entrepreneur who finds ways of providing credible assurances of the quality of this service or work stands to profit handsomely. Brand names and warranties are examples of market generated institutions that overcome these information gaps through screening and signalling.

Screening is the less informed party’s effort, usually the buyer, to learn the information that the more informed party has. Successful screens have the characteristic that it is unprofitable for bad types of sellers to mimic the behaviour of good types.

Signalling is an informed party’s effort, usually the seller, to communicate information to the less informed party.

The main issue with quacks in the labour market is whether there are a large cost of less than average quality service, and is there a sub-market who will buy less than average quality products in the presence of competing sellers competing on the basis of quality assurance. This demand for assurance creates opportunities for entrepreneurs to profit by providing assurance.

David Friedman wrote a paper about contract enforcement in cyberspace where the buyer and seller is in different countries so conventional mechanisms such as the courts are futile in cases where the quality of the good is not as promised or there is a failure to deliver at all:

Public enforcement of contracts between parties in different countries is more costly and uncertain than public enforcement within a single jurisdiction.

Furthermore, in a world where geographical lines are invisible, parties to publicly enforced contracts will frequently not know what law those contracts are likely to fall under. Hence public enforcement, while still possible for future online contracts, will be less workable than for the realspace contracts of the past.

A second and perhaps more serious problem may arise in the future as a result of technological developments that already exist and are now going into common use. These technologies, of which the most fundamental is public key encryption, make possible an online world where many people do business anonymously, with reputations attached to their cyberspace, not their realspace, identities

Online auction and sales sites address adverse selection with authentication and escrow services, insurance, and on-line reputations through the rating of sellers by buyers.

E-commerce is flourishing despite been supposedly plagued by adverse selection and weak contract enforcement against overseas venders.

In the labour market, screening and signalling take the form of probationary periods,  promotion ladders, promotion tournaments, incentive pay and the back loading of pay in the form of pension investing and other prizes and bonds for good performance over a long period.

In the case of the labour force, there are good arguments that a major reason for investments in education is as a to signal quality, reliability, diligence as well as investment in a credential that is of no value the case of misconduct or incompetence. Lower quality workers will find it very difficult if not impossible to fake quality and reliability in this way – through investing in higher education.

In the case of teacher registration, for example, does a teacher registration system screen out any more low quality candidates for recruitment than do proper reference checks and a police check for a criminal record.

Mostly disciplinary investigations and deregistrations under the auspices of occupational regulation is for gross misconduct  and criminal convictions rather than just shading of quality.

Much of personnel  and organisational economics is about the screening and sorting of applicants, recruits and workers by quality and the assurance of performance.

Alert entrepreneurs have every incentive to find more profitable ways to manage the quality of their workforce and sort their recruitment pools.

Baron and Kreps (1999) developed the recruitment taxonomy made up of stars, guardians and foot-soldiers.

Stars hold jobs with limited downside risk but high performance is very good for the firm – the costs of hiring errors for stars such as an R&D worker are small: mostly their salary. Foot-soldiers are employees with narrow ranges of good and bad possible outcomes.

Guardians have jobs where bad performance can be a calamity but good job performance is only slightly better than an average performance.

Airline pilots and safety, compliance, finance and controller jobs are all examples of guardian jobs where risk is all downside. Bad performance of these jobs can  bring the company down. Dual control is common in guardian jobs.

The employer’s focus when recruiting and supervising guardians is low job performance and not associating rewards and promotions with risky behaviours. Employers will closely screen applicants for guardian jobs, impose long apprenticeships and may limit recruiting to port-of-entry jobs.

The private sector has ample experience in handling risk in recruitment for guardian jobs. Firms and entrepreneurs are subject to a hard budget constraints that apply immediately if they hire quacks and duds.

Blackboard economics says that governments may be able to improve on market performance but as Coase warned that actually implement regulatory changes in real life is another matter:

The policy under consideration is one which is implemented on the blackboard.

All the information needed is assumed to be available and the teacher plays all the parts. He fixes prices, imposes taxes, and distributes subsidies (on the blackboard) to promote the general welfare.

But there is no counterpart to the teacher within the real economic system

Occupational regulation  comes with the real risk of the regulation turning into an anti-competitive barrier to entry as Milton Friedman (1962) warned:

The most obvious social cost is that any one of these measures, whether it be registration, certification, or licensure, almost inevitably becomes a tool in the hands of a special producer group to obtain a monopoly position at the expense of the rest of the public.

There is no way to avoid this result. One can devise one or another set of procedural controls designed to avert this outcome, but none is likely to overcome the problem that arises out of the greater concentration of producer than of consumer interest.

The people who are most concerned with any such arrangement, who will press most for its enforcement and be most concerned with its administration, will be the people in the particular occupation or trade involved.

They will inevitably press for the extension of registration to certification and of certification to licensure. Once licensure is attained, the people who might develop an interest in undermining the regulations are kept from exerting their influence. They don’t get a license, must therefore go into other occupations, and will lose interest.

The result is invariably control over entry by members of the occupation itself and hence the establishment of a monopoly position.

Friedman’s PhD was published in 1945 as Income from Independent Professional Practice. With co-author Simon Kuznets, he argued that licensing procedures limited entry into the medical profession allowing doctors to charge higher fees than if competition were more open.

Data Source: Martin Jenkins 2012, Review of Occupational Regulation, released by the Ministry of Business, Innovation and Employment under the Official Information Act.

The importance of bidding for the Olympics, losing that bid, but putting on a good show

Both successful and unsuccessful bids to host the Olympics  have a similar positive impact on exports according to Andrew Rose  – no relation.

The Olympic effect on trade is from a signal that a country sends when bidding to host the games, rather than actually hosting the event.

A country that wishes to liberalise its trade may want to signal this by bidding to host a mega-event. It generates extra trade-related investment and creates a political atmosphere where back-sliding on trade liberalisation or the mega-event becomes difficult.

  • Rome was awarded the 1960 games in 1955: the same year that Italy started to move towards currency convertibility, joined the UN, and began the negotiations that lead two years later to the Treaty of Rome and the creation of the European Economic Community.
  • The Tokyo games of 1964 coincided with Japanese entry into the IMF and the OECD.
  • Barcelona was awarded the 1992 games in 1986, the same year Spain joined the European economic community.
  • The decision to award Korea the 1988 games coincided with Korea’s political liberalisation.

Many of the countries that are hosted the Olympic Games in recent years such as China have done so as part of showing to the world that they have made it and there’s no going back.

The trick  then for the  taxpayer is  for your country’s bid  to host the Olympics to come a close second without anyone knowing you really want to lose.

The trouble with treating the Olympic bid is all show to boost your image  as an investment destination and a liberalising economy is  your beard might actually win. Throwing a fight is never easy as many a boxer knows.

The Case Against Education – Bryan Caplan–updated with Japanese evidence

Bryan Caplan  says that:

When you actually experience education, though, it’s hard not to notice that most classes teach no job skills.

The labour market heavily rewards educational credentials even though academic curriculum is seriously disconnected from the jobs people actually do.

The best explanation for this strange fact is that education is a strong signal of pre-existing worker productivity.

Caplan argues with annoying persuasiveness that education signals desirable employee traits such as intelligence, conscientiousness, conformity and a willingness to learn boring things:

  • Most education is for sending a signal to employers that you can jump through hoops to show off your IQ, work ethic, and conformity.
  • Schools and universities do not to produce wisdom, information, critical thinking or human capital.
  • Subsidising education creates an arms race of credentialism as each student attempts to acquire more and more education than their rival job applicants.

His particular focus is the educational psychology literature on the transfer of learning. That literature started long ago with the question did learning Latin give you muscle to learn other subjects. The educational psychology literature  has been looked at the transfer of learning for 100 years.

Educational psychologist found that Latin does not help much in studying other languages and other subjects. No significant differences were found in deductive and inductive reasoning or text comprehension among students with 4 years of Latin, 2 years of Latin, and no Latin at all.

The trouble is you do this in a race and many try to win the race by lengthening the race by going to and spending more time at university such as taking honours and master’s degrees etc.

Grades do not signal anything in Japan because everyone graduates with an A. It is the lecturer’s fault if you fail.

Japanese universities and employers make up for this everyone gets a A with strict entrance exams.

Getting into a top university signals intelligence and conscientiousness in preparing for their entrance exam. Few go to graduate school in Japan, preferring to learn more on the job.

Japanese students are lazy because everyone passes and therefore grades signal little in the way of intelligence, conscientiousness, conformity to employers.

I had great trouble getting my Japanese students to come to class. Other lecturers got around this by giving marks for attendance and replacing final exams with a pop quiz at the start of every class.

Nonetheless, something of value is acquired through 4-years at a Japanese university because otherwise why not skip straight from passing a university entrance exam to the employer exams.

The crucial objection to Caplan is that if most education expenditures are primarily about signalling, it should be possible to find other, cheaper ways to signal desirable traits to employers. As Bill Dickens noted:

For one thing I find it very hard to believe that we would waste so many resources on a nearly unproductive enterprise.

There are plenty of entrepreneurs out there trying to make money by selling cheaper, in time and money, versions of education and they aren’t very successful.

Mainstream schools have experimented with programmed learning, lectures on video, self-paced learning, etc. and none of the methods have caught on. Why wouldn’t they if they worked?

The spread of charter schools is an example of the rapid diffusion of an educational innovation valued by parents.

A major driver of the doubling of college tuition fees in the U.S. is demand for greater quality. As Becker and Murphy explain:

Indeed, it appears that the increases in tuition were partly induced by the greater return to col­lege education. Pablo Peña, in a Ph.D. dissertation in progress at the University of Chicago, argues con­vincingly that tuition rose in part because students want to invest more in the quality of their education, and increased spending per student by colleges is partly financed by higher tuition levels

What specific and general skills are learnt at school and at university matters too, as Bill Dickens explains:

Education isn’t mainly about learning specific subject matter.

Rather education is mainly about practicing the sort of self-discipline that is necessary to be productive in a modern work environment.

High school allows you to practice showing up on time and doing what you are told.

College allows you to practice and work out techniques that work for you that allow you to take on and complete on time complicated multi-part tasks in an environment where you have considerable freedom about how you spend your time.

Some people may be more talented than others at this sort of thing (you come to mind as someone who is particularly talented at self-discipline), but this is also an acquired skill that one can develop with practice, and everyone needs to develop certain work habits that make one more productive at both types of tasks.

The debate really turns on the extent to which it is possible to find easier and cheaper ways to signal conscientiousness and conformity. As Bill Dickens noted as his fall-back position, which is based on comparative institutional analysis:

most of the return to education is due to it signalling desirable characteristics, but that there is no more efficient way to sort the capable from the incapable.

I also think that signalling performs a valuable sorting function that no alternative process can out-compete. But, as Caplan notes, a conventional education benefits from large government and private subsidies as compared to other sorting devices.

Table of Contents – The Case Against Education – Bryan Caplan


Chapter 1: The Magic of Education

Chapter 2: Useless Studies with Big Payoffs: The Puzzle Is Real

Chapter 3: Signalling Explained

Chapter 4: Measuring Signalling

Chapter 5: Who Cares If It’s Signalling? The Private, Familial, and Social Returns to Education

Chapter 6: Is Education Good for the Soul?

Chapter 7: We Need Lots Less Education

Chapter 8: We Need More Vocational Education


The Book’s basic plot:

The labor market heavily rewards educational credentials even though academic curriculum is seriously disconnected from the jobs people actually do.  The best explanation for this strange fact is that education is a strong signal of pre-existing worker productivity. (chapter 1)

While the return to education is often overstated, it remains high after making various statistical adjustments.  Degrees in useless subjects really do substantially raise wages. (chapter 2)

Education signals a package of desirable employee traits: intelligence of course, but also conscientiousness and conformity.  Many people dismiss the signalling model on a priori grounds, but educational signalling is at least as plausible as many widely accepted forms of of statistical discrimination. (chapter 3)

Empirically distinguishing signalling from human capital is notoriously difficult.  But literatures on the sheepskin effect, employer learning, and the international return to education confirm that signalling is moderately to highly important. (chapter 4)

How much education should you get?  The human capital-signalling distinction isn’t important at the individual level, but the policy implications are enormous. (chapter 5)

The non-pecuniary benefits of education are over-rated, and the non-pecuniary costs (especially boredom) are under-rated.  There’s a massive selection bias because the kind of people who hate school rarely publicize their complaints. (chapter 6)

The most important implication of the signalling model is that we spend way too much money on education.  Education spending at all levels should be drastically reduced, and people should enter the labor force at much younger ages. (chapter 7)

The education we offer should be more vocational.  Especially for weaker students, vocational education has a higher private and social return than traditional academic education. (chapter 8)

Caplan has also posted this nice topology below to allow you to select your starting point:

Model Effect of Education on Income Effect of Education on Productivity Notes
Pure Human Capital WYSIWYG

(What You See Is What You Get)

WYSIWYG Education may raise productivity by directly teaching job skills, but character formation, acculturation, etc. also count.
Pure Ability Bias Zero Zero “Ability” includes not just pre-existing intelligence, but pre-existing character, acculturation, etc.

Pure Ability Bias is observationally equivalent to a Pure Consumption model of education.

Pure Signalling WYSIWYG Zero Pure educational signalling can consist in (a) learning and retaining useless material, (b) learning but not retaining material regardless of usefulness, (c) simply wasting time in ways that less productive workers find relatively painful, leading to a positive correlation between education and productivity.
1/3 Pure Human Capital,
1/3 Pure Ability Bias,
1/3 Pure Signalling
2/3*WYSIWYG 1/3*WYSIWYG A good starting position for agnostics.
0.1 Pure Human Capital,
0.5 Pure Ability Bias,
0.4 Pure Signalling
.5*WYSIWYG .1*WYSIWYG Caplan’s preferred point estimates.  He knows they’re extreme, but his book will explain his reasons and try to win you over.