How to reform the economics Ph.D

This has been bothering me, so I’m putting it out there – The shift to 6 yrs for an Econ PhD is a TERRIBLE trend for female PhD students – & also some men, obviously – but especially for women. This issue warrants much more attention.

So says the wise Melissa S. Kearney.

Along those lines, I have a modest proposal.  Eliminate the economics Ph.D, period.  Offer everyone three years of graduate economics education, and no more (with a clock reset allowed for pregnancy).  Did Smith, Keynes, or Hayek have an economics Ph.D?  This way, no one will assume you know what you are talking about, and the underlying message is that economics learning is lifelong.

After the three years is up, you would be free to look for a job, or alternatively you might find someone to support you to do additional research, such as in the newly structured “post doc without the doc.”  The researchers who absolutely need additional training would try to glom on to a lab or major grant, but six years would not be the default.

Of course, in that setting, schools could take chances on more students, and more students could take a chance on trying economics as a profession.  Furthermore, for most of the most accomplished students, it is already clear they deserve a top job by the time their third year rolls around, usually well before then.  Women would hit their tenure clocks much earlier, also, easing childbearing constraints.  A dissertation truly would become just a job market paper, which has already been the trend for a long time.  Why obsess over the non-convexity of “finishing”?  Finish everyone, and throw them into the maws of some mix of AI and human evaluators sooner rather than later.

Over time, I would expect that more people would take the first-year sequence in their senior year of undergraduate study, and more first-year jobs would have zero or very low teaching loads.  All to the better.

And if you’re mainly going to teach Principles at a state university, three years of graduate study really is enough.  You’ll learn more your first year teaching anyway.

Which other fields might benefit from such a reform?

People, you have nothing to lose but your chains.


"Furthermore, for most of the most accomplished students, it is already clear they deserve a top job by the time their third year rolls around, usually well before then."

Why doesn't competition among econ departments drive some, especially say those just outside the top tier, to recruit these pre-PhD obvious stars as faculty members? Suppose your department is ranked no. 20-30. If you wait until an obvious star receives her PhD to recruit her as a faculty member, she may take a faculty position at a top-5 school instead. If you offer her a position when she is a 4th-yr grad student, however, then you might get her if the top-5 schools are unwilling to match that offer.

We see elite football and basketball programs recruiting kids well before their senior year in high school, sometimes even in middle school. If the rules allowed it, I'm sure some of these schools would lock-in these kids well before their senior year. (I think some of these schools to offer scholarships to kids early, but I don't think they can prevent the kids from changing their minds and going somewhere else by the time they graduate from high school.)

Regarding the maternity angle, why is it considered so much harder to have a child during grad school than when someone is working? Sure, grad students are busy building their own human capital, but full-time employees are also supposed to be busy earning their paychecks. Is this a kind of principal-agent problem? Grad students are essentially principals in that everything they do is in service of their future selves. They bear the full opportunity cost if they mix grad studies with child rearing. An employee, however, can shift some of that cost to her employer. I'm not sure anyone has looked at it this way before.

"recruit these pre-PhD obvious stars as faculty members"
This is ripe for corruption. What's the point of making 99% of the people grind it out when a lucky few can queue jump? I'm sure Felicity Huffman's kids are a shoo-in for the elite colleges. Amazon is an obvious star so let's give them tax breaks instead of the numerous small businesses that make up your small city.

"Amazon is an obvious star so let's give them tax breaks instead of the numerous small businesses that make up your small city."
The suggestion is more akin to "amazon is obviously a star, do let's buy from there without comparing prices with other options"

Having faculty without PhDs hurts your institution's US News rankings: 3% of the score is determined by "proportion of full-time faculty with the highest degree in their fields".

The situation is even worse in sciences, with ~6 year Ph.D.s and postdoctoral positions of increasing length -- in biology especially 4-6 years as a postdoc is common. The delay to the start of "real life" drives many talented people away, especially women or those without quite flexible personal situations, and is remarkably under-discussed compared to ineffective "feel good" programs for increasing diversity.

Four years of undergraduate (some of it perhaps rather childish?), six for the PhD and six more for post-docs. It's a wonder that any intelligent, spirited people bother with it.

For new readers: I am inordinately proud of persuading my daughter not to do a PhD.

The Ph.D. will be gone by 2025 in all fields. Higher education as we know it
is also heading out the door and this will benefit everyone except tenured professors, but they will still do well, on average.

You underestimate inertia.

Maybe, but I don't think by that much.

Let me crack open the first beer of the day to drink to that. Well, 10h00 and still in the office, I'll have to wait for lunch ;)

I finished my PhD on 2017. If after 8 years the offer of graduates goes to zero...I can negotiate even better payment terms for sure.

To him, technology will fix everything, with no downsides, in the next 5-10 years.

Todd, just because I'm collecting ideas, and because I partly with you, probably for different reasons, (a) why will it disappear (I hope you don't say tech, won't happen), and (b) what will substitute?

Agree with the OP. Why is not the market taking care of this problem?

"This way, no one will assume you know what you are talking about, and the underlying message is that economics learning is lifelong." - in other words, formal degrees are indeed signaling. I myself, a lifetime lerner, lerned that money is largely neutral, short-term and long (I used to believe in the textbook models). Now if only I can persuade Dr. Sumner of my views...

If you outlaw PhDs then only outlaws will have PhDs.

We're all clever economists here, right? Presumably there's a reason why we ended up with 6 year phds, it didn't just happen by magic. Suppose that a school offered a 3 year program, how would job seekers and employers view that? (especially in an age where average is over, remember)

In Australia phds are 3-4 years. It's been that way for a long time. And most phds I meet tell me to stay away from them like the plague.

There are abbreviated law and medical schools. I believe Boston College had a six year BA/MD program once. Some schools have MD/PhD and JD/PhD programs that are abbreviated.

Anyone can finish an econ PhD in four years if they can pass the gut check of the dissertation. For various reasons, that is difficult.

I haven't seen any evidence that it takes less than six to nine years to master a subject. The question is, whether teaching undergraduate education in math, statistics, micro, Beckerian Social, Macro, (and Contract Law and philosophy of science please), requires a PhD and post doc. Obviously not.

Can we condense graduate and PhD into six years (certainly). If we fix the compensation system of researchers, then we won't have to use very expensive equivalent of adult apprenticeship in research of questionable value other than publication points - and if we end the publication system, by increasing the minimum criteria for publishing (require warranty of due diligence), we could fix the economy of the higher education and research. Malincentives are everywhere.

All this is well discussed - at least in my circles (educational reform, political reform, legal reform). Conversely it's not clear we need more PhD's, or even that the production of the academy is meaningful compared to industry. The clown world of philosophy, education psychology and social science phd's is absurd. Just randomly read rejected papers. Heck, randomly read SSRN papers.

Military was a better source of funding research throughout history, and our migration to applied rather than basic research has been a disaster.

no one will assume you know what you are talking about
If we are equally ignorant we have a Nash equilibrium. Can we know at least that much?

You touch on this a little, but the PhD is essentially an occupational license for higher-ed teaching. Ergo the incentive of academic PhDs is to make obtaining one difficult enough, and thus rare enough, to keep employment secure and remunerative. This is especially true in research universities where so many who have no place teaching nonetheless earn quite a bit by doing so at the expense of both students and, say, an MA with considerable teaching ability.

Accreditation standards are rising. Rankings are based on faculty academic publications. Undergrad education is almost an afterthought.

The profession that understands specialization and trade better than anyone else sure seems to not be implementing specialization and trade when it comes to teaching and research. A lot more great teachers with MAs could improve undergrad education immensely. Instead we massage the egos of researchers who are horrible teachers and entice parents with "your kid is taught by a Nobel Prize winner."

"You touch on this a little, but the PhD is essentially an occupational license for higher-ed teaching"


"the incentive of academic PhDs is to make obtaining one difficult enough, and thus rare enough"

Have you looked at the job market for people with PhDs lately (meaning, really, the last 20-30 years)? Except for economics and maybe a few other fields, perhaps computer science, the job market for PhDs has been lousy for decades. The problem is too many people with PhDs, not too few.

Suckers. Go for your 3 year PhD and you will lose out to 6 year PhDs. There's a reason why its like this. A few years from now, 10 year PhDs will be the norm.

One of my teachers (in computational linguistics), the late David Hays, maintained that anyone who had any business getting a doctorate should be able to complete the work in their early 20s. He also thought that long Ph.D, programs kept young adults in an adolescent role too long.

“He also thought that long Ph.D, programs kept young adults in an adolescent role too long.“

This. When 34 year old men have only known living in apartments with 4 other dudes playing Smash Brothers on Saturday night, then become lecturers of students who are an age they last year weee on the young end of their PhD-peers... can help explain why so few women in long PhD fields

'People, you have nothing to lose but your chains.'

And professors, you have nothing to lose but your jobs.

Good luck.

In medicine we have known for decades that the current 4/4/3-7 is too much for literally decades. We even had, and continue to have, places that churn out MDs which are not significantly different even while lopping multiple years off the process (some places just require 2 yrs of "undergrad", some med schools have only 3 years of medical school). And lest we forget, those residency years have hour limits (no more than 28 hours in a row, 80 hours per week) and about half of residency programs are estimated to be in compliance.

For more funzies, ever more disciplines are requiring fellowship, i.e. another 1-2 years of intensive training to get top flight positions when everything is finally done.

Medical education as we know it is actively banned in Europe. Fertility issues for female physicians are horrid and we do basically jack all to make it easier for women. The most intensive, and lucrative, specialties historically defended their inane training requirements specifically to softly discriminate against women and particularly mothers.

Advanced education has nothing to do with learning. It has nothing to do with preparing learnings for the real world. It has everything to do with signaling commitment by requiring real and costly sacrifices. The whole point is to make it terrible so that we can allocate positional outcomes based on willingness to sacrifice.

Any attempt to change this will be fought tooth and nail by those who like the fact that they are willing to sacrifice everything for the position. They would most likely have lost out if it had been just a question of ability and devaluing the work they put in back when is a very strong ego blow.

Long, arduous training processes are great at perpetuating elites. Those in the know can work out how to cut corners, they can afford years of reduced income, and they have the social standing to get their deficiencies overlooked. Reducing training time would be a massive step towards opening up the petty fiefdoms of the 9.9% and they will always resist that.

'Medical education as we know it is actively banned in Europe.'

Those horrible rationers. And on top of not having American style medical education, they pay their doctors about half of what American doctors receive.

'Long, arduous training processes are great at perpetuating elites.'

As demonstrated by a number of nation's Olympic programs. OK, that was tongue in cheek - some things actually require long, arduous training processes, and the elite are those able to follow them.

+1, to both Sure and clockwork. The current system is a guild that doubles the wages of American doctors for small differences in quality.

Any Fix for American health care costs has to involve reductions in the industries compensation, staffing & hospital expenditures. That's where the bulk of the money goes.

Doubles is pretty strong wages.

Switzerland, for instance, has more foreign born physicians than the US, but still has higher physician salaries. Regardless, you will run into trouble with the fact that American wages are across the board higher and much higher for comparable difficulty of positions.

The other big thing to remember is that the mean US physician works for 55 hours a week. European regulations limit the ceiling for their physicians to 48 hours per week. Eating the overhead for 25% more physicians is a pretty hefty cost, so while I support a large increase in the number of physicians (with massive changes in medical school and residency) ... it is not going to halve physician salaries. Worse, the US has been increasing the number of nurse practitioners for decades and they have not decreased physician salaries. Given how much of the old GP and hospitalist work they have taken up, I just do not see a big bang for the buck by increasing the supply, certainly not halving. And remember at the end of the day increasing the supply of physicians does not actually lower total net costs.

Even if it did, it would be a drop in the bucket. The biggest labor costs are not the MDs, it lies with the nurses. US nursing ratios are substantially below European norms and any attempt to bring down the staffing costs means firing people, lots of them.

The real reasons that European medicine is cheaper is that the costs are hidden off the books. Being treated sooner has a real value, if only in getting back to work sooner. Europe pays stupendous amounts more due to wait times. More depressingly, Europe pays with dead bodies. Herceptin, for instance, has NNT of 7 but was delayed 4 years in the UK relative to the US. During that timeframe thousands of women died who would have lived. If we cull the specialties down to European levels, we will not have the physicians to offer a lot of the new procedures that offer significant decreases in specific mortalities (e.g. the UK is finally getting around to building proton radiation therapy centers even though we have had data for quite some time showing that they offer improved survival for a variety of pediatric brain tumors).

Then, of course, there are the death of thousand cuts things running around: Europe hides the costs of medical education off the books in the public education budget (e.g. France) whereas in the US we make physicians pay for their own training up front and then double or treble the cost over the course of their careers; ultimately requiring a small but substantial increase in physician's fees to pay off the loans. Likewise, Europe subsidizes public transportation a lot more than US which makes it cheaper to transport patients; here your medical transport gets billed by the hospital instead of the paid for out of transit funds. And on it goes.

Frankly, the biggest thing that makes me doubt that you can provide medical care for cheap is the fact that no large US system can manage it. Not Kaiser. Not Vermont. Not TriCare. Various networks in the US are larger than entire European countries and they cannot appreciably decrease the costs.

If we are lucky, massive overhaul of the physician side of thing will decrease dollar outlays maybe 20%. I more want to overhaul so that we can keep enough providers in the game to actually serve all the patients. Cost savings are a bad reason to restructure physician training.

You didn't mention that single payer in Europe means an entire layer of useless, rent-seeking bureaucrats called insurance companies and other ecosystem parasites like pharmacy benefits managers are eliminated. That will also save tremendously in costs.

You mentioned that Europe has lower costs but you did not mention that they have higher life expectancy than the US.

The number one comfort found in the European system that Americans want is the lack of surprise billing. Going to an in-network hospital in the US still presents a risk as some of the physicians there may be out-of-network which turns medical billing into a game of financial roulette. It is so bad in the US that even Texas passed a law to prevent it for certain plans.

And can anyone explain why giving birth in the US which should be bog standard routine is the world's most expensive[1]? For this expensive privilege(?), the US has the highest infant mortality rate in the developed world.

Not all is rosy of course in Euroland, but the trade offs are worth it for government provided universal health care.


Of course I did not mention that, real life makes it pretty much irrelevant.

Vermont attempted to do as you suggest, eliminating all such players, and they were unable to even mock up a budget comparable to Europe. And Vermont is larger than Malta. Yet every single estimate put forward by the advocates of single payer in one of the most liberal states in the country on many measures was still closer to the US average than anything in Europe.

Then, of course, there is TriCare. TriCare covers around 9.4 million enrollees. It is a mandatory benefit, administered centrally, with no extraneous profit centers. Congress can, and does actively manage the program. They are vertically integrated through 55 hospitals. And lest we forget their population is among the healthiest and most fit in the country. At a gross level, TriCare is maybe cheaper than healthcare in Luxembourg. On a procedure level, forget it, TriCare is often more expensive than American alternatives.

I do not mention life expectancy because that is a fallacious comparison. Estimates, from Europe, put healthcare's effect on life expectancy at no more than 1 part in 4. Given that the health indicators that are responsible for the majority of non-health care outcomes are so much worse in the US (smoking history, obesity, drug use, etc.) the US healthcare system appears to be achieving drastically more than the European one.

Which is why, shockingly, we have better outcomes for diseases with matched patient populations. You know, the actual way to effectively measure the impact of healthcare.

There are several reasons why American healthcare is more expensive. Lack of government control is not one of them. We have such a system. It covers more people than the Austrian healthcare system. It is vastly more expansive than Austria's.

The most lucrative specialties are not always the most intense. Dermatology and ophthalmology (particularly retina) are, per hour worked, very lucrative and their residencies are much less taxing than lesser paid specialties like general surgery and internal medicine. And both derm and opht. residencies are full of women and the competition for training spots is fierce. Discrimination and the “bro” culture may play a role in discouraging women from more intense fields like ortho and neurosurg. But, the primary women avoid them is they realize the nature of those specialties- during and AFTER residency- make them incompatible with childbearing. Head trauma and bone trauma tend to occur at odd hours and their repair often requires many hours in the OR.

I agree medical school could be shortened. Duke, Baylor, and Vandy have compressed the classroom years significantly and their board scores are amongst the highest in the country. But, I’d be very wary of cutting residency times back in surgical specialties. Many residents are graduating with not nearly enough operative experience as it is. And this lack of OR time is even a bigger problem in much of Europe (especially Italy) where many graduates have to work as indentured servants to a senior doc because they just have no experience to operate independently.

" Dermatology and ophthalmology (particularly retina) are, per hour worked, very lucrative and their residencies are much less taxing than lesser paid specialties like general surgery and internal medicine. "

A.I. outperforms ophthalmologists and dermatologists:

" I suspect that these algorithms will play a disruptive role in the delivery of health care as well as the way we practice ophthalmology,” Lee said.

In 2018, A.I. started to outperform dermatologists at diagnosing skin cancer. This year, Google's A.I. outperformed radiologists at detecting lung cancer and A.I. has outperformed cardiologists in a few head to head test like detecting arrhythmia.
"Many published papers show that deep learning models can achieve human expert performance, and in many cases in which an objective ground truth is available, the models can even surpass expert human performance," Lee said. "...

GIGO- AI does well if fed perfect images and perfect medical histories. Unfortunately, in the real world of clinical medicine those conditions are rarely met. A big part of the art of medicine is eliciting histories from afraid, confused, and uneducated patients. And getting good scans from older patients and those with co-morbidities is often very difficult.

But, in ideal conditions AI helps remove some of the drudgery in dealing with boatloads of scans. It increasingly will free dermatologist and ophthalmologist to concentrate on areas where their skill set is more valued. They will become more productive and thus even more highly compensate relative to primary care than they are now.

What can a dermatologist or ophthalmologist do that a nurse practitioner with 2025 level A.I. couldn't do in six years?

Also, virtual reality will be commonly used by 2025, so why couldn't someone consult a dermatologist (or nurse practitioner) outside of the U.S. where prices are cheaper about a skin problem?

What can a dermatologist or ophthalmologist do that a nurse practitioner with 2025 level A.I. couldn't do in six years?

Microsurgery. Perhaps you will be comfortable with an AI enhanced NP doing your macular hole surgery, but I would venture to say you may be in the minority.

Robots will do it better by 2025 if they don't already. This is obviously the final 5 to 10 years of specialists and GPs. There will be the nurse practitioner for a while.

The only robot widely used to assist in surgery is the DaVinci. And it used primarily in urology and gyn which are nowhere as delicate as retinal surgery. And there is very little evidence that using the DaVinci- which merely assists in surgery- improves outcomes. And using the robot increases costs substantially. We are a long way from having autonomous robots do microsurgery. If it happens in the next 30 years I will be surprised.

From your comments it’s pretty clear you do not have a background in medicine, computer science, or robotics.

Well for one get a tactile examination.

For another we have been hearing the same thing about radiologist for literally decades. Turns out that everybody is leery of sending medical decisions overseas. We have been able to all but instantly move chest X-rays overseas for literally decades and nobody does it. Everyone reads local with the exception of night coverage and the majority of that is within jurisdictions or something like the Canada/Australia swap.

Maybe derm will fall to technology, but most of the time technology means new opportunities. Maybe derm will become more procedure based with AI enabled micro treatment of diseases with lasers, cryotherapy, or some other stuff and likely combinations.Maybe derm will start doing some sort of high end collagen remodeling. What I do know is that even absent AI, we have drastically reduced the time it takes a dermatologist to diagnose skin cancer. Just using cellphone cameras and other tools not extant 30 years ago has drastically increased the throughput of dermatologists. Yet their RVUs continue to sail upward. It is almost like once you decrease the time sunk into drudgery the professionals find something else to use their skills.

"For another we have been hearing the same thing about radiologist for literally decades."

No. This is completely false. Go find the quotes and post them. It was this year that Google A.I. detected lung cancer more accurately than radiologists. A.I. can only get better as radiologists have maxed out as have Chess and Go players.

We have been hearing for decades that radiologists would lose out because we would just use the internet to send the images to India and have them read in India.

And you can. For a fraction of a US read (or a British or German one) a nice Indian radiologist will look at your CXR and tell you what it means.

Yet nobody does this. Not the US. Not the NHS. Heck, often not even India. Some of this is a liability question, but more is the fact that radiologists are not interchangeable and the biggest value of a radiology read is not the primary read of the image. Rather we pay radiologists to spot things on surveys that were not suspected. Nearly 100% of survivable pancreatic cancers are due to a sharp eyed radiologist finding a tumor when the patient is imaged for something else. Likewise, child abuse is often found in subtleties that need more than task specific minimums.

Will AI eat radiology? Maybe. But being able to send images off to India has failed to make a dent over decades. Likewise, read rates are orders of magnitude higher than they were even two decades ago, yet radiologists have only seen increased pay.

The truth is that for a while, AI will simply increase the demand for imaging. If per-image reading costs fall, along with their continuing fall in per-acquisition costs, it is quite likely that AI will simply result in ever more images to read. With the way ultrasound is progressing, I suspect that we will end up just routinely imaging every patient in a couple of decades. At that point we would have enough work product for the current crop of radiologists even if they only ever read the challenging 5% that give the computer fits.

Derm and optho are certainly good, but I have a hard time buying the hours after residency argument. After all Ob/Gyn has truly horrid hours but is dominated by women in the current residencies. Further within Ob/Gyn women are even more disproportionately on the Ob side of life working worse hours. Neonatology is also pretty crappy with the attending hours, but again has lots of women. Both of these, however, have better setups for residency than a lot of other more demanding specialties. They also have a lot more points in them where you can easily take some maternity leave and get on with life.

More what I go from is what I was told by the dinosaurs themselves, that they did not want part time or frankly even regular full time lifestyles to become habits in their disciplines so they kept the hours and responsibilities terrible precisely to keep out women.

As far as our surgeons, for most hospitals you will be able to specialize and we could drop a year by allowing for specialization prior to fellowship. We could likely stand to break neurosurgery up and just go for spine surgeons and cranial surgeons much earlier on. Thoracic desperately needs to be broken down into cardiac and general thoracic with less OR time outside of focus.

As is we seem to be lengthening residencies, even without increasing OR times (e.g. the ever popular "research year(s)").

Frankly I think we can compress the entire BSc, preclinical years, and dedicated study time(s) down to around 2.5 - 3 years. This would require rejiggering USMLE to something functional instead of a giant "do thousands of prep questions" game and getting rid of much of the stratification from grades ... but having taught residents from Israel, Europe, and the rest I see no real deficits.

Well, maybe it is the culture in ortho and neurosurgery that keeps women out. But, decisions to enter these specialties is often made in the first year of medical school before female students are even faced with the dinosaur clinicians you describe. To get into these and other competitive specialties you need to get your ducks in order from day one: research publications, high grades and most of all stellar USMLE Step 1 scores.

The surgical specialties are already doing what you recommend. You can go straight into vascular, CT, and plastic surgery without suffering through five to six years of general surgery first. Although I think this is more so those fields can sweep up the talent right out of medical school rather than a sincere attempt to shorten training duration.

I agree about spine. This should be a freestanding specialty. Having someone do five years of ortho or seven years of neurosurgery before doing a spine fellowship is insane. But, there are no shortage of takers because nothing brings in the $$$ like a spine surgeon.

First year medical students are not dumb. The average M1 has family members in medicine and the grueling aspects are the stuff of legends. If anything, I suspect that Med students views of surgery and the like improve during medical school as there has been some, slow, progress towards less inane training and their priors update from whatever mom saw in the 90s.

If you plan to have a family, pregnancy is going to happen in a couple of windows. You can try in medical school, which risks your Step 1 score and possibly wreaks havoc on your finances with loan interruptions. You can try in residency which makes elite programs very angry as any substantial leave will open holes in their training progression, often not timed well with Julys, and of course is an utter bear for long hours on the feet. Or you can elect to try afterwards when you are, most likely, over 30 and quickly losing fertility (female physicians are diagnosed with infertility disorders at double the population rate, likely due to waiting to attempt conception). This is all reasonably well known and if you know this going in, I suspect a lot of women just opt out from the start. Being on your feet for 30 hours is hard enough when you are not running high blood volume with mild tachypnea and resulting alkalosis.

I have simply known too many women who opt for less intensive residencies, sometimes even within a field, because they do not want to have to manage the sorts of all-in commitments demanded by residencies.

As far as the specialties, I am glad that we are slowly making progress, but when I parcel out my patients to surgeons very few of them are actually generalists even with their subspecialties. I have one ENT who does pretty much only nasal work and another who gets all the neck cases. In pretty much all the surgical areas we could stand to specialize sooner, even if not creating new specialties, and devote more OR time to a narrower range of tasks.

I mean seriously, why do breast surgeons need to wait for a fellowship? Will they really be missing all that much if they never do a Hartman? Why do colorectal surgeons need to manage breast patients? How many hepatic cases do they really need?

There should certainly be some generalists (e.g. trauma) but so much of medicine is now based on large health organizations with hyperspecialization. We can and should continue the process you already noted to a lot more areas of subspecialization.

Every bit of data I have ever seen suggests that most of training is just not that useful for learning in the US, training is too long for too little payout.

I'll just note that three years of graduate study is of course the accepted training period for lawyers (and law professors).

In the US. In saner countries you'd do a law degree as an undergraduate and then, perhaps, add a year of professional training immediately after.

Correct, but so too is DearieMe.

I was no more prepared for the legal profession after 3L than 2L. It is really on the job training that matters. This is unfortunate for clients who are test subjects. The legal system really should not rely so heavily on esoteric skills. The system is created by lawyers, for lawyers. If anyone could be a lawyer (and most could), then we would be a dime a dozen and be paid accordingly.

Both medicine and law, in my opinion, could be undergrad degrees. The JD could still exist to educate judges who would take a more active role in ensuring fair and accurate trials rather than relying on the adversarial process.

PhDs are as much about learning discipline, focus, patience, and the management of repeated defeat as they are about developing substantive expertise and research skills. To shorten the PhD is to shorten this other vital dimension of academic preparation. That might be a good thing. Or it might be a bad thing. But it needs to be part of the discussion on reforming and truncating graduate education.

I'm not sure that the current PhD structure really does all that much for the substantive expertise and research skills, though. To the extent that students learn things, I think that it is largely through teaching themselves because the professors are skilled researchers or charismatic lecturers for undergrad courses/intuitive topics, but not skilled enough teachers to effectively teach the level of material that they are supposed to be teaching. Contrast what educational best practices are in life/death subjects (e.g. medicine, engineering) with the stagnation in most liberal arts pedagogy. It is pretty astounding to me just how little thought goes into effective pedagogy in most higher education subjects, especially given the price tag.

I imagine many graduate professors believe themselves to be teaching the students how to teach themselves or the "discipline, etc." that you mention, but my experience is that the professors do not even manage this; anything that the students learn to this end is self-taught or learned from peers.

Sorry, the below sentence came out too hard to parse. The idea is that students mostly teach themselves in grad school. While professors may be good researchers or undergrad teachers, it is very very rare for professors to be effective teachers of graduate level material.

*To the extent that students learn things, I think that it is largely through teaching themselves because the professors are skilled researchers or charismatic lecturers for undergrad courses/intuitive topics, but not skilled enough teachers to effectively teach the level of material that they are supposed to be teaching.

I do not mean that students learn *the* discipline. I mean, rather, that they learn *how* to be disciplined. In my experience, many PhD students can handle the coursework and comps, but fall apart at the dissertation stage. It's the marathon that kills them. That's where you learn discipline, patience, and how to manage setbacks. And to me, that's the value of a 6-year degree: it requires massive independent learning. Because if you can't handle a PhD, you certainly can't handle the tenure track.

Yes, this exactly. Tests are easy. Taking one class after another is easy. Writing a dissertation is a completely different skill.

I know what you meant by discipline. I disagree with you that professors are as successful as they may imagine themselves to be at instilling that skillset. I think that poor pedagogy turns graduate school into a filter rather than the crucible it should be (i.e. it does not support student growth and merely assesses who is already at the needed level), even when taking the what I think is an extremely generous viewpoint that poor teaching of "the discipline" is okay if accompanied by effective teaching of discipline skills.

Nah, medical school professors are terrible too. Most medical school classes these days are trending towards just pounding rote memorization for USMLE Step exams. Classes are taught by PhDs, often without special knowledge of their subject matter, and the real learning takes place when the students sit down with independent resources. Typically the worst of it are the anatomists who seem to end up in control of huge amounts of med school curricula who spend the vast majority of student time on stupid things (i.e. differentiate this nerve, artery, or vein on a cadaver by position, which is subject to substantial anatomic variation, whereas in real life blood flow is the real determinant).

For most of medicine the real teaching is, as always, in the clinical side of life. Managing patients is the important thing, but I find the first two years to be ever less preparation for new residents and ever more geared toward having higher Step 1 scores.

Frankly, I expect this overvaluation of examination scores to kill thousands of patients. Ne physicians are relying too much on their memorization (and not developing good habits of when to use an iPhone to look things up), have trouble working outside of the multiple choice format, and have a much harder time relating to patients as they have spent their entire adult lives preparing for tests (often quite literally).

That is disappointing. I have always understood medical schools to be at the forefront of pedagogy, pioneering, for example, the completion of a concrete task as the examination.

Attainment of the NBE’s Certified Business Economist credential possibly is a better predictor of candidates’ likelihood of on-the-job success than any degree. It requires passage of an objective examination. Education credentials primarily reflect an individual’s ability to regurgitate a professor’s hobby horses, shop Indian essay writers, and produce whatever significant research findings the lead asks for. Employers who signal their reliance upon objective examination based credentials rather than academic credentials should see competitive advantages in hiring, attracting investors, and winning contracts.

Link to NBE’s Certified Business Economist page:

This whole concept smacks of socialist planning and SJW virtue signalling. But if we are going to play Dictator over a couple of beers, I'll join in:

1. Screw women. Since when are we modifying academic programs and disciplines for the particularized concerns and wants of a subgroup defined by an arbitrary characteristic? A profession rewards performance. It doesnt provide permanent training wheels and mommy running alongside you to make sure you dont fall over.

2. Questioning the worth of a doctorate is valid. Or any degree for that matter. But credentialism is alive and well. Schools could eliminate doctoral programs, but that doesnt imply that employers will. Whether the degree demonstrates skill acquisition or is signalling, employers find it useful and necessary.

3. We already have such a program. It is called a "Masters degree."

4. The average time to PhD in economics is over 7 years, regardless of whether a masters is obtained prior. How can 6 years be too long when 7.2 is approximately the median? Fact is, some people have interruptions and delays from pregnancy, divorce, military service, burnout, laziness, depression, etc. Pregnancy isnt special and shouldnt be preferred. While not removing blame from the student, it appears that professors arent exactly mentoring their grad students well.

5. Core courses are done in two to three semesters. Field courses are done in two to three more. So that's three years. What causes the difference from 6 or 7.2 or 11? The dissertation, of course. IMO, schools do a horrid job of teaching and motivating PhD students to write dissertations. As one person put it, academia is the only profession where someone throws a pile of work products in front of you and then says, "Do stuff like this, or you're fired."

John Forbes Nash's dissertation was what, 30 pages or so? In the past, scholars would write one magnum opus their entire career. The problems with the PhD are the research preparation and expectations. From what I've seen of the research of older PhDs, theyd never get hired today with the crap they produced 30 years ago. All the low hanging fruit is gone, the methods are much harder, and support nonexistent.

6. Dont let theorists run departments. Give equal weight to theory, applied, and historical work. Focus on skills the market for economists really wants...the whole market.

7. Bridge the gaping chasm between undergrad economics and grad economics. Students who took Statistics and Econometrics three years ago, sometimes poorly taught or not focused on grad school, arent prepared for graduate level Econometrics.

8. Give phd students a light course load of refresher and catchup material and let them TEACH principles courses. This will solidify their understanding much better than the course they took as undergraduates. Let them earn tuition waivers with that year of service and then have them teach low course loads (with their prep already done) while they step up to more difficult material.

1. Performance has jack all to do with length of program. Again my gig is medicine and we have condensed programs that churn out doctors that are not appreciably worse than the longer run stuff. If we care about performance, who gives a rat's ass if you "finish" the program "early"?
2. Just because the house on fire doesn't mean throwing on gasoline will be fine. Credentialism is a brutal feedback cycle, sparing even one turn of the screw has its own, innate value.
4. It is not just about pregnancy, as you note, too long of training programs make it harder to do all the things that human society has used for ages to protect the psyche. People who are unable to start families, buy homes, establish long-term communities all are more likely to have depression or other psychiatric illness. It is utterly obnoxious that we consider it perfectly normal to have credentialing eat up a decade of life to no real purpose.

"A profession rewards performance."

This is laughable when the profession is "economist" where performance ranges from bad to middling and they still remain an economist. A bad lawyer gets disbarred, a bad doctor is decertified, a bad engineer loses their license, but a bad economist continues to be quoted by their favorite media outlets.

The PhD is a credential that graduate students need, and which the supervising professor uses to hold power over them. Doing away with the PhD as a valuable credential takes power away from professors. That may not bother Tyler, but most professors covet and zealously protect whatever institutional and personal power they can grasp.

As a side note, the argument for doing away with the PhD as a credential is no different than doing away with the bachelors degree as a credential. Outside of professional, accredited programs, the definition of what a bachelors degree represents is unclear to the student or future employer and wildly variable from student to student, school to school. It remains because HR departments are unwilling to find other methods to rank applicants, and because universities need a valuable credential to sell for their ever-spiralling tuition costs.

I support this idea. In fact, I have said similar in the past, that too many degrees have fixed year cycles. It is just too suspicious a circumstance that so many Bachelors degrees require 4 years.

Maybe eliminate engineering PhDs as well. A Master's is enough. And then of course unlimited research.

The things you choose to get excited about .. too funny.

I'm surprised to learn that there is a fixed time span to a Ph.D in economics.

In science programs there are 1-2 years of classes plus a research project, usually suggested by one's advisor. Usually the criteria for the Ph.D. is that you've written 2-3 papers on your research project, which then becomes your thesis. Whether that takes 3 years or 6 years is about 50% individual talent/effort and 50% other issues like how long it takes to build an experimental apparatus, NASA launching the satellite to get the data you need on time, your cell lines not becoming contaminated so you need to start over, etc.

It seems to me that economics is the most similar of the social sciences to the natural sciences. I would, naively, think that since the objective of the Ph.D. is the same - to acquire the skills needed to do independent research - that an economics Ph.D program could follow the same model.

Economics PhD programs do follow the same model, roughly: around 2 years of courses followed by a preliminary paper guided by an adviser. You graduate once you have 2-3 papers. For the sake of actually finding a job, one of the papers you produce has to be of exceedingly high quality--the job market paper--and you can graduate as soon as this paper is done. In my limited experience, people are staying longer when they realize an additional year of school increases the "quality" (read: length and complexity) of the job market paper by more than the cost of staying. Even if you have a passable set of papers after 4 or 5 years, you will be compared to people with another year of polish on their job market paper, so you stay another year to compete. It seems a little bit like a signaling arms race driving the longer stents in grad school rather than a pedagogical benefit.

There are similarly valid reasons in econ to stay longer (development experiments often have large setbacks, etc.) but most econ students don't have those sorts of issues. Compared to how long it takes to build and launch a satellite or to design and construct a high powered laser that will be used for many years, I would argue economists generally have far less roadblocks to putting together a decent paper.

People stay a long time in the PhD because in economics the capacity to write and present a paper is also important. And after a couple of years of PhD their paper writing and presentation skills are non-existant. They need to stay 4 years after doing the coursework to just hone these skills. In addition, these skills get honed over time after they finish the PhD.

The PhD is essentially the beginning of the academic career and not the end of education: in my last 4 years of PhD I was teaching, presenting seminars and doing research, doing exactly the same as I am doing now as an assistant professor.

I did a PhD (Engineering) in the US and work in UK academia. Though I was initially sceptical, I have come to appreciate the 3-4 year PhD in Europe. Those who stay in academia do a postdoc anyway, and those who don't still get some valuable research experience.

FWIW when my wife would go to conferences other economists would often express surprise when they found out she only had a masters in economics and not a Ph.D.

I get the same reaction, although I don't always correct people who call me Dr. Siegel :-)

Tyler, I have great respect for you, I love reading MR, and I appreciate the value of advancing contrary positions that force us to reflect on commonly-held principles, but you make several incoherent remarks here.

1. Smith, Keynes, Hayek operated in a world where the best mathematics was done by people before they were 25. That is no longer true in mathematics, for the obvious reason that experience matters more now as fields deepen. This is not an instructive example and it misleads.

2. Security for five to six years helps to provide the kind of stability for these human capital investments. Look at the University of Chicago: its PhD program is a nightmare, in part because the first two years is a mad rush not be in the bottom quartile and get kicked out after two years. This kills cooperation, kills the spirit of positive inquiry that should animate good research. It inflicts arguably needless, inarguably substantial suffering and anxiety on individuals.

3. "for most of the most accomplished students, it is already clear they deserve a top job by the time their third year rolls around, usually well before then" is a ludicrous claim. It is also a sad one. In my view, it is very possible for someone to accomplish a tremendous amount in years G3--G6. People come from dramatically different initial conditions, and most of the success in years G1-G2 reflect initial preparation (e.g., having taken the full micro core already at Princeton or two years of PhD courses at Bocconi, etc.) rather than the persistence and drive that are defining qualities of good researchers in the longer term.

4. These are not chains. There is no hard rule that you must finish in 6. There are many examples of people who finish faster. Some of them go on to make valuable contributions to the field and some never properly learn what it is to do truly good economics and are ruined forever (Glen Weyl). Anyone can go on the market whenever they want. The constraint is whether or not they have written a paper that is good enough for what they want to do. That is a physical constraint on the quality of output, as mediated, of course, through the viewpoint of junior hiring committees. But it is wrong to think about this as an imaginary constraint imposed by arbitrary institutions.

5. The true violence of increasingly long PhD programs is simply the violence of a more competitive labor market. The pressure induces longer hours, more intensive work, and exposes individuals to greater personal risk while foreclosing their outside options. Whether or not this is an inevitable cost of scientific progress is not obvious to me.

Full disclosure: I am in the sixth and final year of an economics PhD and therefore am naturally quite biased towards the status quo.

"Look at the University of Chicago: its PhD program is a nightmare, in part because the first two years is a mad rush not be in the bottom quartile and get kicked out after two years. "

That's better odds that they offered a few decades ago, when I was looking at econ graduate programs. The story, perhaps an urban legend as I don't know the real stats, was that Chicago's entering class was three times larger than other universities' -- and they would flunk 2/3 of them at the end of the first year, giving them a masters degree as a consolation prize or send-off prize.

My late career experience of interviewing people for tenure track jobs was that they had better credentials than their equivalents decades earlier, but seemed less intelligent and less high-spirited.

I suppose the system had simply selected against those merits (if merits they be).

First, the Ph.D. is useful for signaling to parents of college students. Parents of college students are not going to read out papers, no matter how amazing they might be.
Second, moving cities is a huge cost on mothers or prospective mothers. I actually had an offer to leave my Ph.D. office and go do funded research in a lab across the country. I elected to do 6-year traditional Ph.D. instead, precisely because I had a baby.
Third, I'm very close to someone who is going through IVF treatment right now. It's TERRIBLE. I had no idea how costly (I don't just mean $, I had known about the $ before) it is to the women who undergo it, just to increase their chance to getting pregnant. Those who think it's no big deal to ask women to delay pregnancy until mid-30s should know this. So, I do appreciate Tyler's attempt to think through a better path for female academics.

Wouldn't the obvious solution be then to not admit women into any PhD program? As Roderick Kaine explains in his book "Smart and SeXY," the likelihood of a woman having anything to contribute in an intellectual field is highly unlikely anyway.

I assume six years is based on the American way of combining what in other countries would be a Masters degree and a separate PhD degree. I never understood that system as it means that a lot of people are in six years programs that could otherwise get a masters degree to qualify themselves for work in Government or something like that. Also, after a master's degree, many people learn that graduate school and an academic life is not for them. Why make them take six year programs?

Most (all?) American econ PhD programs have a bail-out option or consolation prize where you can drop out of the program and be awarded a masters degree.

There are programs where it's a two-stage process, you have to earn a masters first and then go for the PhD. It's my impression that these are rare in economics though.

Finally there are some programs that are strictly for a masters degree in econ. But I don't think any of the top departments do this.

"Eliminate the Economics Ph.D."
I suggest the University of Maryland go first.

I wonder if the first school to do such a thing will see the quality of its applicants go dramatically up.

the PhD is not going to disappear, no way. How long it takes to get one is a serous matter, but not under the control of anybody.

The more serious matter implicit in this post is what has happened to the PhD dissertation. For quite some time in economics at least it has been three papers, with one of them the "job market paper." These may have nothing to do with each other and may even have coauthors. This is not how it used to be.

So, once upon a time a PhD thesis was a unified and serious long work on a topic. In principle, if it was good enough, it could be publishable as a book, and some were, with Samuelson's _Foundations of Economic Analysis_ perhaps the ultimate model of this. In many disciplines this is still how it is, such as history, or so I am told. But it is now in only oddball heterodox places or perhaps abroad that this old model still exists.

So indeed I appreciate the logic of Tyler's post, which amounts to it being that once one gets beyond the 2-3 years of grad course work, which in most programs is sufficient to get a Masters degree, whhat really matters beyond that is the quality of papers one produces as measured by publishability, and ultimately citability, is what matters for the job market and longer term academic professional success.

Actually it is nice idea. Thanks for sharing nice information with us. I like your post and it is quite informative.

-Mike Noble

Comments for this post are closed