This article was originally written in July, 2013. I think it is still mostly accurate.
What follows are predictions for future employment in STEM fields–science, technology, engineering, and math–over the next ten or fifteen years. They are based on my observations during my career as a science professor and college administrator.
These sorts of predictions are like predicting the stock market. Many details will almost certainly turn out to be wrong. I do hope that I get the larger picture approximately correct. You can let me know in 2025 or so, assuming I’m still alive.
Three trends will dominate:
- Automation is going mainstream, and will eliminate many, many jobs. This trend will disproportionately affect the most expensive employees, such as doctors and professors. Conversely, the least expensive employees are at greatest risk of having their jobs moved off-shore.
- Globalization results in a global marketplace, resulting in a trend popularly known as outsourcing. In occupations where Americans produce the most value, jobs will be in-sourced. I expect in-sourcing will be a major trend in sales, project management, finance, custom manufacturing, and entrepreneurial activity. Large, commodity businesses will continue to move to cheaper locations.
- Governments are bankrupt, and government employees will be laid off. Kevin Williamson estimates that for all levels of government in the US, total liabilities (social security, pensions, bonds, etc.) add to about $200 trillion. This is more money than exists in the entire universe, and it will never get paid. Governments in Europe and Asia are similarly in the red.
Employment in the physical sciences–chemistry and physics–will decline. This is true for all degree levels above the associates degree. The primary reason is automation, and to a lesser extent, outsourcing. Many lab tech and quality control jobs have already been automated, and the process will accelerate. Full-time chemists will be replaced by part-time technicians. B.S. chemists who do find jobs will work in R&D, food processing, and the pharmaceutical industry, among others. Businesses downstream from increased oil and gas production may employ chemists. Physicists will be employed in defense and aerospace, and as engineers and computer scientists. Sales positions are open to both chemists and physicists, and will likely be the most lucrative avenue for them. The outlook for masters degree graduates in the physical sciences is similar.
Geologists–from what little I know–should do better, especially if the oil/gas boom continues. Unlike chemistry and physics, these jobs cannot be as easily computerized or outsourced.
Baccalaureate biology majors already face an incredibly weak job market. Many biologists work for the government, e.g., in environmental enforcement. The number of jobs in the government sector will continue to shrink. But most importantly, employment in healthcare will stop growing, and indeed, I predict it will shrink dramatically. Computerization and automation will make huge inroads in this industry, along with some outsourcing.
A stunning graph appeared in The Atlantic (h/t Via Meadia) showing how health-related job growth has been much larger than in any other sector. This obviously cannot continue. Indeed, it looks to me rather like a “bubble,” and I predict a large number of doctors, technicians, and even nurses will be laid off. Doctors, especially, are at high risk of being automated out of business, as is happening to lawyers today. Salaries for all healthcare jobs will go down. It looks to me as though everybody and her kid sister wants to study nursing these days. This will not end well.
Another popular career for science majors has been teaching. Opportunities here will decline sharply because of demographics, more flexible, on-line opportunities for students, and because government school districts are broke. The problem is most acute in the Northeast and Midwest.
PhD grads in the sciences will fare worse than their baccalaureate counterparts. PhDs typically get two kinds of jobs: college teaching and research. The prospects for both are pretty dire. College faculty are at high risk from automation, in the form of on-line delivery. The on-line platform will significantly reduce employment, and accordingly save students money. Similar to health care, the cost of college has gone up much faster than inflation over the past three decades. This cannot continue, and colleges are already under increasing pressure to lower their prices. This is compounded by the fact that most state governments are broke, so most colleges will see public support decline. The net result is that the market for professors will collapse, and many of those currently employed will be laid off.
Most research in the physical sciences is funded by government, and governments are bankrupt. Research monies will not be growing, and will likely be shrinking. Additional funding comes from universities, which use revenue from undergraduate education to subsidize the research effort. This extra revenue will disappear. Finally, chemistry and physics are mature disciplines, and the marginal return on research is getting smaller and smaller.
Research in the biological sciences will remain healthier because of medical applications. Still, because of automation, outsourcing, and general government insolvency, it is doubtful that research opportunities for biology PhDs will grow.
The market for PhDs in the sciences is already glutted, and that problem will only get worse. I strongly advise against anybody getting a science PhD, unless you plan on attending either CalTech or MIT.
In this category I include all STEM jobs traditionally offered at the associates degree level. A few baccalaureate grads will muscle their way in, as will people with requisite experience but no degrees. I prefer to think of these jobs as skilled trades, and I think they’re the only category of STEM jobs that will grow significantly.
The reason is automation–new machines need workers to take care of them, and most of those employees will be in the skilled trades category. People who learn how to fix and maintain 3D printers, for example, have a secure future ahead of them.
While fewer chemists will be employed, chemical instrumentation will be more widely used than ever. Spectrometers, chromatographs, PCR machines, etc., will all require technicians who know how to maintain these instruments. People who can replace columns in an HPLC, change the bulb in a spectrometer, or calibrate the detector in a mass spectrometer will find gainful employment. The software for many of these instruments is complex, and technicians who can set it up, debug it, and offer tech support will be in demand. Similarly, materials handlers will earn a good living. These are folks who know how to safely move, store, and use chemicals, and then to properly dispose of the waste.
There used to be a degree known to as chem tech–I haven’t heard much about those programs recently. I think they’ve fallen out of favor, partly because they’re expensive to teach, but I anticipate that something like that will come back. Unfortunately, the academically-oriented chemistry major doesn’t provide a very good background for these jobs.
I’ve used chemistry trades as an example, but similar opportunities exist in biology, physics, and engineering disciplines. Many of these people will work as contractors, supporting small and medium-sized businesses that need the equipment, but can’t justify full-time staff.
Computer technicians will also be in high demand, but this is a much riskier path. Technology changes so rapidly that whole job descriptions can disappear from one day to the next. Still, tech support people–help desks, software support, security, etc.–will always be needed. Today network technicians and systems operators are hot. People who can keep their skills current will earn a good living.
Unlike jobs for baccalaureate scientists, the trades will not be automated. To the contrary, they benefit from automation. Also, apart from phone support, their jobs will not be outsourced. There is currently a shortage of skilled tradespeople, and I believe these will make for good careers in the future.
Engineering & Computer Science:
The job market for engineers and computer scientists at the baccalaureate level is stronger than for the sciences, and this will continue to be the case. Part of the reason is that a much smaller percentage of them work for the government, and thus are not as subject to the debt problem. Indeed, the private sector part of the American economy is doing well, and accordingly engineers have benefited. I anticipate this trend to continue, and so foresee a continued strong economy for engineering and CS grads.
There are caveats. It’s not all good news.
The extreme case is computer science. Computer software is tradable at almost zero cost, which means there is no such thing as a domestic market for computer scientists. American graduates are not just in competition with their classmates, but also with similar students from India, Israel, Ireland, or elsewhere. Work will go to the lowest cost provider. If wages were the sole cost driver, then almost all of this work would go overseas. But fortunately for Americans, wages may not even be the most important issue. Here are some items that will keep computer science jobs in the United States:
- Customers will remain disproportionately American. This is because of our political stability, entrepreneurial culture, deep and liquid financial markets, and secure property rights, including intellectual property. Thus sales, project management, top-level design, and key proprietary software development, among other items, will always be delivered in the US.
- American computer geeks still have an edge on skills, and thus command premium wages. That, at least, is my anecdotal opinion.
- Infrastructure matters. Cheap electricity, high-speed networking, ability to tele-work and/or reasonable commuting times–these are all important cost factors. India can’t compete in anything on this list, which lowers the value of their labor.
So I predict a healthy market for computer science grads. But I note that wages have been largely stagnant over the recent past, despite a tech boom. That means that the American competitive advantage is approximately cancelled by cheap foreign competition. For consumers that’s very good news. For computer scientists, it’s not bad news.
Some engineering disciplines are also highly tradable, and thus behave like computer science. Computer engineering, mechanical engineering, and electrical engineering are all more or less tradable, and the arguments above hold. On the other hand, chemical engineering and civil engineering are more accurately described as domestic markets.
If earning money is your goal in life, chemical engineering is probably the best discipline to major in. The US is blessed with two important resources, for both of which chemical engineering is a key skill. The US is the world’s granary, and hence food processing is a vital industry. Companies like Monsanto and Archer-Daniels-Midland hire a large number of chemical engineers. Second, chemical engineers are employed by oil refineries and gas distributors across the country. Downstream industries that use gas and oil as feedstocks also hire chemical engineers.
The average starting salary for chemical engineers is now $67,000. I see no dark clouds on the horizon for this career path.
Civil engineering is probably the least lucrative of the engineering disciplines. Many civil engineers work for government, with all the disadvantages that entails. Still, because they provide essential services, the career is likely to be secure.
Here are three additional “facts” about engineering disciplines. I wouldn’t go to the mat on all of these, but they’re interesting.
- Tyler Cowen has written on how autistic people respond differently to the economy. I can’t find the quote, but I think he has claimed that autistics are the golden boys of the new economy. After all, an obsessive-compulsive who can happily program computers for 16 hours per day is potentially a very valuable employee. Both Bill Gates and Steve Jobs each spent a couple of years of their youth doing just that–an experience that was crucial to their subsequent success. So if you’re majoring in computer science, just remember that autistics are part of the competition.
- Unemployment among computer scientists and engineers is relatively low, but one hears that it is highly concentrated among older workers. Fifty-year-old engineers are much more likely to be laid off than their twenty-five year old counterparts. Given a field with rapid technical change, this is not surprising. But it is very different from, say, professors, where the older you are, the higher your salary and the greater your job security. Cruel as it may sound, the engineering model probably has greater social value. So I don’t know what happens to Google employees twenty years from now. Moral: Save as much money as you can right now, because you may be forced into retirement at an early age.
- Also from Google, they have seriously studied their hiring process. They found that–apart from people less than two years out of college–college degrees, courses, GPAs and transcripts just don’t matter. Therefore they no longer request that information except from brand new grads. As a result, they now have some teams where 14% of the members have never gone to college.For computer scientists it is no longer necessary to attend college (though it may be beneficial). It remains to be seen if and how quickly this trend spreads to other disciplines. College is very expensive, and the social value is maximized if people are hired at the lowest possible education level. Hence the trend previously described of replacing degreed chemists with machines and technicians.
I’ll end this section with a word on prospects for engineering PhDs. The professor gig is ending for them as well, just as for the scientists. There are more opportunities for them in the private sector, but maybe not a lot more. As there are far fewer graduate students coming up through the ranks, the market is nowhere near as glutted. But my bottom line is still the same: unless you’ve been admitted to either CalTech or MIT, a PhD program is a bad investment.
Pity the poor mathematician. He will be unemployed.
Of course nobody needs to feel sorry for math majors. A math degree is an excellent signal for both intelligence and work ethic, and recipients will find many doors open to them. Some of them will work as computer scientists, others as engineers, and many outside of STEM altogether. But almost none of them will work as mathematicians.
The most common career for mathematicians has been teaching–both K-12 and at the college level. The latter typically requires a PhD. But math education is ground zero for on-line learning, and especially at the college level the labor force will be decimated. If your ambition is to be a math teacher or professor, make sure you have a Plan B.
There are several reasons for this. First, math education has moved on-line for many years, and people have lots of experience. The Khan Academy was founded in 2006, and is gradually revolutionizing math education around the world. Their efforts are steadily improved over time.
Second, learning math is intrinsically a self-paced endeavor. This is enhanced on-line, in a way that is not true in a classroom. Math is very different from, say, literature, where class discussion is an important component of education.
Finally, on-line education gives students immediate feedback, and can enforce rules that require students demonstrate proficiency in topic A before moving on to topic B. That’s much better than just an exam at the end of the third week of class (which takes a week to grade before it is handed back to you).
The world will still need a few math professors (very few), and there will be some demand for people with expertise in certain sub-disciplines, e.g., statistics or operations research. But my advice to avoid the PhD is doubled for math. A math PhD is probably the worst career move you can possibly make.
Mathematicians have historically supported themselves by teaching. As this option is foreclosed, I predict that they will gradually leave academia altogether. Instead they’ll retreat into their own foundations, whose function will be to raise money to support mathematics. The best mathematicians will receive a salary, and millions more people will engage in math as a labor of love.
I am very optimistic about the American economy. There is little doubt in my mind that fifteen years from now, our standard of living will be noticeably higher than it is today. Just the two examples I’ve discussed above–education and health care–make that point. Both will be much cheaper and much more widely accessible than they are today, substantially improving our lives. Multiply that across the entire economy, and we are all going to be getting very, very rich.
The big drag on the future is government debt. But this isn’t really an economic problem as much as it is a political problem. The debt will not be paid in full. The fight will be over who takes the biggest haircut. Let them fight. You need to avoid working for the government, and you need to make sure you stay away from defined benefit pension plans, where people make you promises that have no chance of ever being kept. And then don’t count on social security or medicare (or medicaid or Obamacare, or any other government service).
But this economic boom is unlike any other. As previous booms have depended on new technology (electricity, the automobile, the container ship), this one depends on computerization and the Internet. The principle result is an unprecedented tsunami of automation. We’ve had automation in the past–sharecroppers were displaced by mechanical cotton pickers, ditch diggers by diesel backhoes, autoworkers by robots, and so on. But that all pales by comparison with what will happen now. Vast swathes of job descriptions will be rendered obsolete, or nearly so–“professor” and “doctor” being just two that we’ve mentioned. Add to that “lawyer,” “accountant,” “travel agent,” “stock broker,” “truck driver,” “airline pilot,”–and the list goes on and on. Computers, rather than displacing menial and manual labor, cheapen intellectual labor.
Anything you can do, a computer can do better. And cheaper.
- Any repetitive task can be done better by a robot. Think Kiva.
- Any mathematical task can be done more accurately and better by computer. Think Mathematica.
- Anything that requires a prodigious memory, such as medical diagnosis, a computer can do better. Think Watson, of Jeopardy fame.
- Anything that requires infinite patience, such as teaching math to 3rd graders, a computer can do better. Think Khan Academy.
- Anything that requires quick feedback, such as flying an airplane, a computer can do better. Think military drones.
- Anything that benefits from node to node communication, such as maneuvering in traffic, a computer can do better. Think Google Car.
So what’s left for people to do? Not much.
- Non-routine, non-repetitive tasks, such as repairing equipment at arbitrary locations on the customer’s property. This is why tech workers will be in demand. Similarly, surgery, nursing, a lot of engineering, and computer programming will remain, at some level, human tasks.
- What I call emotional work, i.e., work that makes other human beings feel good, cannot be done by computers. This includes waitressing, the caring part of teaching, sales, certain kinds of tech support, counseling and religious work, art and performance, and front-line medical staff, such as the primary care physician.
- Top-level programming of computers, because ultimately computers work for human beings. It takes a human to tell them what to do.
These are the jobs of the future. You will immediately notice that the modern university–designed in the mid-20th Century–is not well adapted for this new world. It is busy teaching facts, math, and process–all things computers can do better than humans. A new, more suitable college curriculum will evolve slowly. In the meantime, it is up to you to make sure you’re getting the education you need. Here is some advice:
- Separate the education you need for your career from the education you want for personal fulfillment. The former you need while you’re still young; the latter you can pursue at any point in life (and probably for free). So now you should concentrate narrowly on your career education. (Colleges will tell you that general education is important for your career. That was never all that true, and it is not true at all today.)
- Formal education really only prepares you for your first job. You’re not going to school for a career so much as for a job. A career will grow out of that first job.
- Set realistic career expectations. Do not prepare for jobs that don’t/won’t exist, or for jobs that you have a 0.0001% chance of getting (e.g., a biologist at the Department of Environmental Conservation). (Colleges will tell you to do what you love. Ideally, yes, but we don’t live in an ideal world. Make sure you prepare for a job that leads to a career that can earn you a living. That said, calculated risks are OK. If you’re really talented and passionate (or stupid), go for broke.)
- Get only the education you need to land that first job you want. Take the minimum number of classes necessary. Education costs money (time is money)–an extra expense for either you or your employer. The term over-educated really means that you’re too expensive. You make yourself more competitive by delivering the goods as cheaply as possible. (Colleges will tell you that if some education is good, then more education is better. That is terrible advice. Think of college education as an expense to be minimized.)
- Do not go to grad school unless you need to acquire a particular license. Likewise, don’t waste time on a second major or minor unless you have some specific job goal in mind.
- Take on-line classes. You will need that skill-set in the future, and you should learn how to succeed in that environment as soon as you can. (Colleges will tell you that the classroom environment is superior. They’re wrong, at least much of the time.)
- After you get your first job, keep taking classes. THEN is the time to learn new things that weren’t part of your narrow, job objective. Register for at least one class a year for the rest of your life. This is where those on-line skills come in handy. (Colleges will tell you if you don’t go to school when you’re young, you’ll never get another chance. That’s not true anymore. Take advantage of new technology to always improve your skills and your life.)
- Don’t be afraid to change careers. (Colleges will tell you that the decisions you make in school are final. Horsefeathers. Remember, college prepares you for a job, not a career.)
Have fun. Work hard. Get rich.