The Path Not Studied
Schools of Dreams More Education Is Not an Economic Elixir
No compelling evidence exists for the thesis that giving people more formal academic education will result in a restructuring of jobs and increased productivity and growth in the economy at large.
The idea that education is the key to economic success is widely and rightly popular—for individuals. A person who makes it through high school, then college, and then graduate school is likely to enjoy increased earning power with each new degree.
The idea that education is the key to the success of an economy as a whole is a relatively new concept, and the evidence for it is not especially compelling. In developing countries, many people lack even the fundamental literacy and numeracy skills necessary to participate in the modern economy, and that does hold back economies. But beyond acknowledging the value of achieving that basic level of educational competency, is there evidence that higher levels of education help national economic performance? The conventional economic wisdom has been that a nation’s economic strength is linked to physical capital, not human capital: A U.S. engineer is seen as much more productive than an equivalent engineer in India because the U.S. engineer is supported by an abundance of equipment, management systems, and infrastructure. Has something in the global economic system changed so that the overall education level of a country is now a key determinant of its economic prowess?
One attempt to link education levels to economic success relies on the common logical error known as “the fallacy of composition,” which assumes that what is true for an individual will also be true for large groups or for the society as a whole. To illustrate, the average U.S. software engineer earns about $90,000 per year. A high-school graduate with adequate ability who acquired the requisite training in software engineering could expect to earn that much. But if this year’s entire class of high-school graduates decided to become software engineers, there is no way that they would all earn $90,000 per year. Most of them would be unemployed initially because there are not enough jobs for software engineers, then we would see wages fall for those who could get jobs as software engineers, and eventually most would end up working in other fields.
Yes, more education enables an individual to make more money and be less subject to unemployment, but if everyone had that professional degree, they couldn’t all make more money. The reason, of course, is because wages are set by supply and demand. Those with professional degrees make more money only when there is a lot of demand for their skills relative to the supply available. Consider someone with a Ph.D. in history. Although earning this degree requires intelligence and a great deal of hard work, there are a great many such Ph.D.s relative to the demand for them, and as a result, the degree adds almost nothing to the degree holder’s earning power.
Some education boosters argue that the entry of a large number of well-educated baby boomers into the labor force resulted in tremendous growth in the U.S. economy. That was not the case. The 1970s were famous as the period of stagflation, marked by declines in economic growth accompanied by inflation. My colleague Michael Wachter became well-known among economists in that period for his evidence that the entry of the baby boomers into the workforce actually reduced productivity. Although better educated than older workers, their lack of experience made the boomers less productive on average. We baby boomers remember that in those years, college graduates who were lucky enough to find a job could expect to be underemployed for a long, long time. No one is arguing that the greater education level of the boomers caused the collapse of labor productivity, but rather that greater education is not a sufficient condition for improving economic performance.
Others worry that the movement of the boomers into retirement will create a shortage of educated workers, but that is not the case. The workforce as a whole is not shrinking, nor is the number of young people entering the workforce declining. The largest high-school class in U.S. history will leave school this year, and the class next year will be even larger. The U.S. labor force is projected to keep growing as far into the future as we can predict (not true in all countries, of course); only the rate of increase is expected to slow, trivially, in about a decade. It is impossible to say what effect, if any, that modest decline in the rate of increase will have on the economy. The situation regarding education levels is similar. The rate of increase in the average level of education achieved has slowed, but it is still rising every year and according to projections from the National Center for Educational Statistics is projected to keep rising through 2017.
The more important question is whether average education levels affect the economy, and if they do, how. Economies can get into trouble if there is a demand for skills that can’t be met by the labor force, and education can supply some of those skills. Again, we can see this problem in many developing countries such as India and China, where a lack of basic education makes many individuals unemployable. But is there any evidence that the United States will soon face a problem like that? Consider this sobering statistic: The General Social Survey of the United States reports that 30% of adult workers have education levels that exceed the requirements for their current job, a figure that has actually gone up over time. Aside from a small handful of jobs such as nursing where there are infrastructure bottlenecks in training, there is no credible evidence of any overall skill shortage in the United States and certainly no evidence of an education shortfall.
Still, wouldn’t it be good for the economy to raise the level of education in the workforce? Let’s start with a simplifying but clearly wrong assumption that education is free and that it would cost nothing to raise the level of education in the workforce. What would happen to the economy if the average level of education in the workforce went up?
One immediate outcome of such a move would be to lower the wages associated with education, effectively lowering the price of educated labor to employers. The economist Richard Freeman described how the returns from education collapsed for baby boomers when their more-educated cohort hit the labor market. In this period, the talk was of “the overeducated American.” Lower wages could allow employers to hire more educated workers for the same jobs, making it possible to maintain an even more overeducated workforce at the same wage rate. Remember, employers today have little trouble finding workers with adequate education at all skill levels. The question is just about what that costs and whether being able to hire workers more cheaply would help them.
How would this move affect the economy? Adding more educated workers will reduce the wages associated with education, and that will allow employers to hire a more educated labor force at the same price. The first point to note here is the obvious one: that expanding the supply of education, a least in the short run, hurts the wages of educated workers even though it will raise the education level in the economy. Will that, in turn, improve productivity? The evidence on this front is not at all encouraging. The key ingredients of increased worker productivity appear to have little to do with education levels. At the level of the company, the biggest improvements in performance come from closing old facilities and opening new ones with better equipment and more efficient production systems. It is not obvious what role, if any, a lower price for educated workers would play in that process, as it seems to have much more to do with physical than human capital. At the level of the establishment, many and perhaps most of the big improvements in productivity come through changes that eliminate jobs altogether: rationalizing assignments and consolidating roles, replacing secretaries with word processing software, etc. Interestingly, making educated workers cheaper actually reduces the incentive to replace them with the technology that would raise productivity.
It is at the level of the individual job where we might expect to see the real contribution of additional education to economic performance. In the years I spent co-directing the National Center for the Educational Quality of the Workforce for the U.S. Department of Education, we scoured research for evidence on the relationship between education and individual job performance. Clearly, additional education allows individuals to take on more difficult and better-paying jobs, but that’s the fallacy of composition argument if we’re talking about overall economic performance. When we look at workers doing exactly the same jobs, additional years of education have very little, if any, effect on improving job performance. It does matter a lot when levels of education are below basic job requirements: Teaching illiterate workers how to read instructions, for example, has a huge payoff. But increasing education beyond the level of the job requirements doesn’t have much effect on job performance.
To see why, consider the following experiment: Let’s take the current employees on a typical assembly line, put them through a Ph.D. program, and then place them back on the line. Should we expect their job performance to go up? Unless job requirements change and the work is redesigned to allow them to make use of their education, there is no reason to believe that more education will have any effect on job performance. The good news in the economy as a whole is that employers are giving employees more scope to contribute and that skill requirements, on average, have risen. However, there is no evidence that the additional requirements are taxing education levels and lots of evidence that education levels still exceed job requirements.
This is the fundamental problem with the “if we build it, they will use it” approach to increasing education levels. It assumes that once employers hired better-educated workers, they would redesign jobs to make full use of them; that they are just waiting to have more-educated workers so that they can put in place more productive work systems. It’s an appealing prospect, but not one with any evidence behind it, at least in the developed world. If employers acted in this way, we would not have overeducated workers. There is no evidence from the research on work organization that a shortfall in overall education levels has kept employers from empowering their employees and expanding job requirements. Most employers do not have formal mechanisms to empower workers to contribute their ideas to workplace decisions.
A disheartening best case
For the sake of making the most optimistic case, let’s assume not only that education is free but that employers would be willing to raise job requirements if their workers brought more skill to their job. The next problem is that education per se does not equal work-related skill. No matter how flexible and accommodating the employer is, a master’s degree graduate in astrophysics is unlikely to be able to contribute much to the performance of a typical service worker. This is an obvious point but one easy to lose in discussions about overall education levels: Beyond a basic level, well below what we typically think of as postsecondary education, what matters to job performance is not generic education but education specific to the performance of particular jobs.
So let’s make yet another heroic assumption: that our education system can somehow match up in a very close way to job requirements; that we could produce graduates with more occupationally specific education that mapped precisely onto the occupations that were in demand. If we could get candidates with more industrial and mechanical engineering coursework to production jobs, for example., and employers would redesign the jobs in ways to make use of those additional skills, would job performance and overall performance improve? Certainly. But given that the country already has a workforce that is overqualified for the jobs they are doing, the more useful intervention would be to find ways to get employers to make greater use of the skills they already have by changing the nature of current jobs to take better advantage of existing skills rather than to load the employer up with even more redundant skill.
Further, if we wanted to raise the average skill level in the workforce, it is not obvious that putting individuals through more academic coursework is the best way to do that. The evidence suggests that other forms of human capital—training and experience—are likely to pay off better in terms of performance and wages in the same job than does more traditional academic education. Anecdotal accounts from employers suggest that their biggest challenge is finding candidates with work-based competencies, the kind that are learned in employer training programs and through experience.
The society benefits in many ways from having a more educated population, as do individuals, and as a society, we have a collective interest in maintaining, even increasing, education levels. And there is no doubt that the country’s postsecondary education system is failing many young people. It is remarkably inefficient at getting students through to graduation. The rapidly rising cost of education is becoming an increasing constraint, exacerbating problems of access and inequality. For the individual, education remains the most important avenue of opportunity in society. One could also make a strong case that there is a segment of the workforce whose level of education is so low that it places them below the job requirements of even the lowest level of jobs. The entire economy could improve if additional education could make them employable.
When we move to a discussion about the effects of overall education levels in society on the aggregate economy, however, we need to put a big caveat around the idea that increasing the level of education in the workforce as a whole will improve the overall economy. When employers are asked what factors are limiting improvements in productivity in their organization, they often don’t have obvious responses. If they did, they would have acted on them. That’s why they spend so much time and money seeking help from consultants in the quest to find better ways to manage their operations. In other words, it’s a mistake to assume that they are being held back in important ways from obvious improvements they would like to make. When they are asked specifically to identify workforce concerns, education issues are almost never at the top of the list. Instead, the top complaints focus on work-related behaviors and attitudes such as conscientiousness, motivation, and social skills. Issues associated with academic skills are far down the list, especially when one moves away from lower-level, front-line work.
Beyond the classroom
The best way to tackle concerns about skills is not by adding more years of traditional classroom education. It is to expand work-based education through programs at the workplace or those that attempt to combine work and classroom experiences. The nation will derive much more economic benefit from apprenticeships, school-to-work programs, and the close associations that many community colleges have developed between employers and classroom topics. If there is a skills problem in the United States, it lies in the area of work-based skills: Every employer wants someone who already has three to five years of experience, who already knows how to do the job. The cold reality is that many employers have either abandoned or cut back on training and work-based learning programs. Union apprenticeship programs have dried up so much that the U.S. Bureau of Labor Statistics no longer even reports them. Employers in industries such as information technology now expect to do no training of any kind, to hire just-in-time candidates who have exactly the skills they need precisely when they need them.
As employers have backed away from training and developing their own talent, they are effectively pushing the problem off onto the job candidates, who in turn look to traditional education providers for help. The shift toward more vocationally oriented degree programs among postsecondary institutions is unmistakable and is a clear response to the hiring practices of employers. Consider, for example, the fact that undergraduates majoring in business have increased from about one in eight students in 1970 to roughly one in four in 2005. This development has also pushed the costs of acquiring skills onto job candidates, who now have to pay for them up front, rather than through employer-based programs (the individuals often pay for those as well through lower wages, but at least they don’t have to advance the costs). My colleagues Ivar Berg and Randall Collins described the now famous finding that individuals achieve more credentials in education to signal to employers that they are better than the other candidates, even though the extra expenditure on additional education has little payoff. If everyone has a high-school degree, then I can differentiate myself by getting a college degree. But then everyone does the same thing, and we all end up doing the same jobs we would have done, jobs that require only high-school degrees, even though we have all been through college. Although it may be true that more educated workers are more conscientious and have better work attitudes, we have to question whether the education is creating more positive attitudes or is it simply that being able to finish a postsecondary degree requires that one already have those work attitudes?
To the extent to which postsecondary education is becoming the substitute for employer-provided training and development, then education truly has become more important to the economy. But it is not education in the traditional sense that matters. It is the highly vocational programs that we see especially in community colleges and the fast-growing for-profit schools. There are lots of reasons to support education in the United States, especially to focus on improving it at the lower end of the distribution. But there are also risks to overpromising what it can do, particularly when we assert that individuals, paying a lot of the costs themselves, can solve the problems just by getting more education. If the goal is to improve economic performance and workplace outcomes, we should think about all the options, including those that might change what employers do, especially given that education is not free. If we want to increase educational expenditures for the purpose of improving economic performance, the way to do that is through work-based learning, not expanding postsecondary education per se.
Peter Cappelli ([email protected]) is George W. Taylor Professor of Management and director of the Center for Human Resources at the Wharton School of the University of Pennsylvania.