Bryan Caplan  

The Career Consequences of Failing versus Forgetting

A Brief Letter on Signaling... Timothy Taylor's Instant Ec...
If you're reading this blog, you probably didn't fail a lot of classes in school.  But I bet that you've totally forgotten a lot of those classes.  I got A's in junior high and high school Spanish, but barely speak a word of it.

Now ask yourself this:
How would your career have been different if you had failed all the classes you've totally forgotten?
According to the human capital model, failing (i.e., never knowing) course material should have exactly the same career consequences as forgetting (i.e., no longer knowing) course material.  Either way, you lack the skills - and the labor market should treat you accordingly.

According to the signaling model, in contrast, the consequences of failing and forgetting can totally diverge. When you fail to learn useless material, you send a bad signal.  When you demonstrate mastery of useless material, you send a good signal - whether or not retain what you learned.  Employers naturally snub people who fail, yet smile upon those who merely forget.

Take me.  If I'd failed Spanish, I couldn't have gone to a good college, wouldn't have gotten into Princeton's Ph.D. program, and probably wouldn't be a professor.  But since I've merely forgotten my Spanish, I'm sitting in my professorial office, loving life.

How about you?  How would your life have been different if you had failed all the classes you've totally forgotten?

Comments and Sharing

COMMENTS (36 to date)
Andrew Hanson writes:

Part of the benefit of learning is that it's easier to remember things or expand your knowledge when you relearn it the second or third time around. You would be in a better position to learn Spanish now than I am (that which we learn but forget is forged in some brain cells), but never having succeeded in the first place means you really will be starting from square one.

nzgsw writes:

I've mentioned this before, but it's affected me in a very odd way. I studied engineering as an undergrad at one of the most challenging institutions in the country. I failed a quantum mechanics course, but got a job as an engineer applying quantum mechanical concepts (signaling boost due to the reputation of the institution?).

After a decade, I then got my MBA, and last year got a job where I don't use anything remotely resembling quantum mechanics. However, I am a temporary (probationary) employee due to that single failed course (signaling beatdown to that F?).

I would think it would work the opposite way, but there you go.

Andy writes:

Learning about your skills in different fields may be an important return to diversity in educational curricula.

Neal's "Complexity of Job Mobility among Young Men" would be the relevant paper on this subject. I haven't seen anyone rigorously apply that logic to class or major selection though.

Lupis42 writes:

It would be very different. On the other hand, if I failed those classes, rather than forgetting them, I would not have been demonstrating the ability to learn non-mechanical skills and information rapidly. Since the ability to learn whatever was necessary has had a great deal more influence on my career than knowing any one subject well, I'd say that the signal in question was still a very useful one.

ChacoKevy writes:

¡Yo todavía recuerdo mi español, pero no tengo la oficina!

It's a fun thing to consider, though. If we all were forced back into exams from decades past, we would all probably carry 2.0 GPAs. Only A's or F's, very little kinda-sorta recall, I suspect.

Of course, no credits for F's, so none of us graduate.

loveactuary writes:

... they wouldn't let me into the actuary club

Brian Moore writes:

While I'm very sympathetic to the signalling model, I'm not sure this is the best metaphor.

There are lots of things that only matter if you know them at the time they're being tested. I probably was much better at actually doing calculus when I graduated from high school than I am now, because I was using it each day.

Now all the calculus I "use" is wrapped up in code such that I don't need to think about the calculus actually involved. I validated that it worked properly in the past and probably couldn't do so again today without lots of memory-jogging.

But your larger point is valid. Certainly in my profession, my computer science degree is pure signalling. If any budding code monkey only knows what he learned from his CS degree, then he is not going to be very good at his job, and certainly vastly less productive than the amateur who learned it the right (in my opinion) way of just doing it yourself.

I do not use what I learned from college. I certainly use a lot of things I learned IN college, but they weren't always "taught" in "classes." Honestly, the best thing a college CS degree gives you is a general framework of what is considered "things you should" even if the classes aren't particularly well suited to actually teach you them.

Steve writes:

I was an EE major at a well respected school and was a horrible student, coming close to failing a number of courses, however I retain many of the core concepts seemingly as well as some who got much better grades. The details, however, eluded me then as now, but that's the case also (now, not then) for my more academically successful peers.

I've come to accept much or all of Bryan's signaling theory based on my undergraduate EE experience. It seems that the educational institution existed primarily to offer to industry tranches of students ranked by their ability and or willingness to navigate an arduous obstacle course. Advances in pedagogy that would have made it easier for the lower tranches to understand the subject matter would probably have been rejected if they were to upset the ranking machine, in my opinion. Needlessly and intentionally obfuscatory and adversarial teaching methods appeared to be chosen in order to ensure that only the brightest or most dedicated came out on top. If this is in fact the case, there's nothing wrong with that - it serves a purpose, and for the elite is perhaps best, but I'd argue that the net knowledge transfer across the student population was much lower than it could have been. I hope to see alternative forms of technical education that focus on efficient knowledge transfer that 'raises all boats' for those not suited for the gladiatorial approach that I observed.

Brian writes:

How do you know which skills you're learning will be useless and forgotten, versus those that are useful?

A friend of mine took Spanish in high school because he thought it was the easiest language to satisfy the language requirement for graduation. He majored in math in college and as far as I know didn't need to use Spanish at all. But after graduating college, he became a missionary in Venezuela for a decade, and came back to our high school to give a lecture in fluent Spanish about how useful the class was for him.

But in addition to this, knowing something and forgetting it at least signals to someone that you are capable of learning. Whereas being unable to learn something clearly signals you're unable to learn that skill. And if you could learn it once and forget it, you can probably learn it again if required. But passing or failing classes has less to do with demonstrating a mastery of specific knowledge, but rather the ability to learn something well enough to pass a test that is presumably designed to indicate whether or not you understand the subject.

Eric writes:

I've wanted to say this for a while but kept waiting for an on-topic post. A big smoking gun for the signalling vs. human capital model is that universities don't typically allow a course to be retaken and the old grade replaced by a higher one. If all people cared was what you know (your human capital), they'd care about the final "A" and not the path. As it is, they care about the fact that you were only able to get a "C+" on the standard 15-week schedule. Even if you can retake the course, you usually keep both the C+ and the A.

MikeP writes:

If you're reading this blog, you probably didn't fail a lot of classes in school.

Talk about signalling!

Are you planning to open the Bryan Caplan Institution for Advanced Signaling?

Read the blog. Get a diploma!

Mike writes:

Eric, as a college/university administrator, I've had numerous arguments with the records/registrar people on your point. If they won't give you the new grade, why did they let you take the class? I'm left to believe it's about the money.

For your amusement: Father Guido Sarducci's Five Minute University. If you are not familiar with this, it's on point. 4:00 minutes.

John Thacker writes:

There are a lot of studies suggesting that even when classes are totally forgotten, A students can pick back up the material quicker than C students or smart people who have never studied it. It is thus not truly totally forgotten.

JFK writes:

I flunked Spanish at Harvard, but I got a decent job.

Teddy writes:

Jack, nice of you to cover for me, but that was me.

Lee Kelly writes:

I actually remember more of what I learned in school than my peers, because I was actually interested in the material. However, I hated school itself; it was just a part-time prison to me. I failed many classes, and I usually struggle in the workplace. It's not a coincidence.

I score very low on conscientiousness, extroversion, and relatively low on agreeableness (which is kinda misleading, since I'm very agreeable in some respects but then not in others). I score very high on openness to experience and very low on neuroticism, which might seem like pros, but my experience of the workplace that these traits are at least as much as a hinderence as a help--they certainly didn't help in school.

To say that school is mostly about signaling personality, with a nod toward intelligence, is not to say that school isn't working. Even public schools work surprisingly well, considering they're basically a government monopoly, at providing opportunities for signaling.

I mean, even right now, I am writing these comments at work. Granted, there isn't much else to do right now, and my employer doesn't expect much, but I could be doing something. I'm not, instead, I'm being typically unconscientious and pursuing whatever interesting topic happens to flutter through my easily distracted mind. This is (mostly) why I've always (so far) worked crappy jobs for low pay.

Scott writes:

Treating knowledge like it's a piece of plant equipment has got to be one of the stupidest ideas economists have ever come up with. The prevalence of the human capital model really demonstrates the lengths economists will go to avoid having to come up with an original idea.

finlawyer writes:

For lawyers, it is almost all about signaling. If you go to a top 10 school, you basically just have to graduate to get a prestigious job. The mere fact that you gained admittance to those institutions is largely what matters. At lower-tier law schools, you have to get a good GPA and have a high class rank to have a chance at a job with a large, high-prestige firm, but you don't necessarily need to learn a lot in order to get good grades. I had many classmates in law school who, as evidenced by their comments in class, knew much more law than I did at the time, but they didn't figure out the rules of the "game" of law school in time to get good grades. I didn't learn the material very well, but I figured out the game and my high GPA got me a high paying job at a good firm, while my former classmates in many cases struggled to get a job at all.

Thomas DeMeo writes:

Your point only holds if you equate skills with the particular knowledge from any given class.

Passing a course is a kind of exercise that has a cumulative effect. If you took two people of identical intelligence, and sent one to a decent school for a decade, while the other was not exposed to a regular learning regime, I would think that the individual who was schooled would be able to figure complex new things out far more quickly, not because they remembered particular facts from a course, but because they had practiced the process of figuring things out on a regular basis.

People who fail a course are simply not exercising as effectively as A students are. Ten years from now, neither may know any Spanish, but one person is more practiced at learning it going forward.

An Honest Moose writes:

Something I have learned as an adult, that I very much undervalued when I was a student - a lot of the signalling that goes on is not about raw intellectual ability - a very few tests and courses could establish that. It's about a establishing a track record of personal characteristics over time that are likely to be demonstrated in the future.

Someone who gets all A's over 4 years of college is ambitious, competitive, and has a drive to "win" over his or her peers. The same goes for someone who has good grades and is also a successful (say Div 1) athlete or competitive chess player, etc. The most competitive firms value these characteristics, which are easy to claim but hard to demonstrate in an interview. Top law firms, consultancies, investment banks, and universities hiring tenure track faculty, etc. all want staff who are smart AND have a great desire to be among the best in their field. They value the latter category so much that they want to see proof you are willing to go to extreme lengths to achieve success.

Someone who gets a mix of A's and B's, signals as smart but relatively lazy. They can signal intelligence and decent work ethic easily enough in an interview, can get a good but much less competitive job.

A good natural experiment is non-traditional students who go back to get professional degrees. Lots of relatively low status teachers, journalists, lab techs, etc. go back to law or medical school, where they demonstrate their ability and competitiveness by getting good grades (or fail to do so) and thus have the potential to move into positions that are much more greatly compensated. Raw ability is signaled via the MCAT, LSAT, or GMAT, etc. especially for students from non-traditional background. But firms want more than the test results, they want the professional degrees too. Thus, you can change your signalling but it can't be overnight - the signal has to be sustained over a long period of time and over changing circumstances to be credible in the long term.

RJB writes:

The first comment on the Marginal Revolution post is the thread winner: if you learned it well, you are in a much better position to relearn it when you need to. Someday you will go to Spain, and find that you need much less study to relearn the language. You will find it easier to pick up Italian and French as well.

On top of that, you are probably overestimating how much you have forgotten. You probably know your English grammar better having studied another language, and have better study habits because you learned and practiced (and probably haven't forgotten) how to learn a foreign language.

tom writes:

If we are assuming Bryan failed Spanish, we have to know more about the causes and effects:

1. Was it intentional or accidental? Each says something.

2. Even if it was unintentional, did Bryan fail despite knowing that it would affect his future, or was he (somehow) unaware that it might affect his future?

2. Was it because of a lack of ability of a certain kind, like difficulty with languages, couldn't manage other courses, drug problems? Do those same problems affect some of his other classes, or somehow just Spanish?

3. Did it hold him back in Spanish for the next term so that it had more effect on his overall education? Did it stop him from graduating on time?

There is no Bryan who is the same as the one we know who would have gotten an F instead of an A in Spanish.

John Townsend writes:

Related question: Is there any difference between recently acquired knowledge and that acquired many years ago - say in college ? Which sends the more positive signal ? I suspect the latter.

Mr. Econotarian writes:

Although I am a big believer in the "college as signalling" model, I have to admit that while I have forgotten much of my engineering education, it is easier for me to recall that information when prompted.

For example, I technically forgot most of linear algebra, but recently when I re-read about Gram-Schmidt orthogonalization, it all started to come back to me!

Also some classes are required just to understand work-relevant knowledge. I needed fields to understand why coaxial cables work, although I would have to work hard to remember how to mathematically prove it formally.

Hume writes:

Unless you are hoping to obtain a PhD in philosophy (oh the joys of the foreign language requirement).

Never mind college: has anyone tried helping his kids with school homework?

GIVCO writes:

That's an unfair, impoverished view of the human capital argument. A fair evaluation of capital expenditures requires an assessment as of the date of investment; as usual, a gambling axiom helps: "just because you lost doesn't mean it was a bad bet."

Your Spanish class could have helped you meet a Spanish beauty or Carlos Slim's son and created a Latin joint venture of love or business; you could've leveraged it into untold opportunities.

The human capital idea, properly explained, should mention time: the young do not know their plans, how best to keep doors of opportunity open? One idea, the broad education: (1) exposes youth to things that might capture their interest, (2) teaches them how to meaningfully assess opportunities (a freshman can't knowledgeably want to be an economist without high school math and, hopefully, literature/philosophy), (3) buys time in the maturation process (i.e., you are not doing harmful things with harmful people); (4) just might offer cross-pollination that fuels the next great innovation (Gutenberg studied wine presses and apprenticed setting type.); (5) provides a foundation upon which you can rebuild that knowledge; (6) makes you less of an insufferable bore at the office; etc.

Historiography, had you studied it, wouldn't help your professorship but it would have prevented you from characterizing historical events as inevitably leading to your current situation.

JLA writes:

I think there's a big difference between failing and forgetting.

When I fail a class I demonstrate an inability to pick up new skills or an unwillingness to do an unpleasant task.

When I get an A in a class and then forget the material, I've demonstrated that I'm capable of handling new problems or am willing to suck it up and doing something unpleasant. These are two *very* valuable traits to employers.

Enrique writes:

In my last year in college, I had to write a compiler for a make believe programming language, the language was pretty much useless and never got to be used for anything, nearly 10 years later while working on a totally unrelated problem we had a problem with a 2000+ line program that needed a "small change", while trying to fix it, thanks to the skills learned on that "useless" project, I came with a 40 line script that executed much faster, had more options and was several times more robust than the original solution..... just sayin....

Curt Doolittle writes:


We are not paid for our knowledge. We are paid for the rate at which we assimilate and adapt to information and circumstances. We are paid to quickly and inexpensively solve problems in dynamic economy.

Universities successfully filter for those people able to assimilate and adapt to information and circumstances. People who pass the filter are more likely to adapt to the shock of entering the work force and quickly learn the nuances of both organizations and business processes.

Since IQ is largely an expression of the RATE someone is capable of learning, the data should show that universities essentially sort by IQ. And it appears to show just that.

I am not convinced (and I think you've come to the same conclusion) that people learn anything of value in university other than work discipline. (Sowell has been saying this for years.)

It also appears that people eventually sort by IQ in the work force regardless of their education. So, it would seem that an education is a means of temporarily increasing your earning capacity at the median, and a way of shortening your access to income at the top. But at the bottom higher education's a waste of time, and a burdensome debt.

Americans try to educate everyone to join the upper middle class, and it's a waste of effort and produces an incompetent working class. instead, we should, as the Germans do, focus on creating a superior working class, because the upper 20% will succeed as long as we don't impede them too much.

As you've stated elsewhere, and as the economic evidence shows, the German model is a superior education system, and perhaps the Finnish model is the best primary school system. For certain, boys should start school later than girls. and should be physically active despite the risk of 'being boys'.

Paul writes:

Microsoft's hiring requirement was (is) to hire people who are "smart, and get things done." Funnily enough, many PhDs do really well on the smart side but are hopeless at getting things done (in the sense of being able to work in a team and produce a lot of high-quality code). So there is a big difference between IQ and what is needed in industry.

The courses are very different too. Most undergraduate CS majors have never written a program much longer than maybe 1000 lines. Then when they get employed they may find themselves part of a team working on a million lines of code.

But they will use far more of what they learned than a physics major who doesn't stay in academia.

Seth writes:

"Employers naturally snub people who fail, yet smile upon those who merely forget."

I think you are basing this observation on a biased sample of employers. I know of employers who look for folks who failed a few classes for various reasons. One reason is that it shows the person didn't take himself too seriously. Another is that it shows the person may have good cocktail party skills which comes in handy in sales and leadership roles. Another is that the person has demonstrated the ability to deal with setbacks.

And many of the folks I know with lower GPAs have done pretty well for themselves.

George writes:

I came to Canada from Czechoslovakia at the age of 19 and got tested to see how many high schools credits my knowledge was worth. I was given 30 credits, enough to obtain the high school diploma, and only needed to take some ESL classes and a few grade 13 classes to get me to a university.

If I had failed, say, Russian at the age of eight, it would have had no effect on my life today. If, however, I had forgotten the trigonometric functions, I would have not received the necessary credits to allow me to continue my education. The fact that I was able to retain most of what I had learned as a kid did impact my life to a great degree. But I suppose my case is an exception.

Lorne M. writes:

There is one funny crude drawn image on the Internet with an obviously white-collar man sitting behind a desk, there are few differential equations showed next to him and the text reads "I am still waiting for that day, when I am going to use it in everyday life."

Maybe none of us will ever need to use the differential equations, Spanish language or other subjects that we now look back at and think of them as a waste of time, but high school and college courses are mainly aimed at broadening our minds and they try to improve our conceptual thinking.

Therefore failing is unacceptable but forgetting is almost a necessity.

Joshua writes:


But in addition to this, knowing something and forgetting it at least signals to someone that you are capable of learning. Whereas being unable to learn something clearly signals you're unable to learn that skill.

Not necessarily. It may indicate a death in the family, pregnancy/abortion woes, winter depression, stress over losing a job in a bad economy, a teacher with too thick an accent to properly teach a course in a way that particular student could understand, or any number of other things.

An "F" indicates poor performance in a particular interval, it is not indicative of persistently limited intellect or competence.

Bmaryott writes:

Failing course material isn't "never knowing" it's "never LEARNING" and that's a crucial difference.

Learning and forgetting shows that you have the ability to retain material, at least for a period of time. This is actually an incredibly important skill. As an example, in my younger days I had to be able to read punch cards on sight (Stop laughing you kids and get of my lawn!) Whether I learned in in a school or on the job was immaterial; it's something I needed to know how to do.

Today, I only remember the number of rows and columns on the cards. The skill has been forgotten. Before you think that's apples and oranges, consider that to get that decent grade,you had to study, learn and retain - at least long enough to pass the classes. That you don't remember it now merely shows that the skill has atrophied through lack of use, not that it was irrelevant.

The Human Capital model factors this in (or is supposed to factor this in) by allowing that individuals that have shown tenancies to learn one thing in schools can learn more; those that fail in schools are more likely not to be able to increase their own value.

Yes, I know it doesn't work EXACTLY like that, but the models are based on large numbers and overtime, Your Mileage May Vary etc. Especially since I nearly failed French.

Comments for this entry have been closed
Return to top